ESICM LIVES 2016: part one - M.MOAM.INFO (2025)

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27 DOI 10.1186/s40635-016-0098-x

Intensive Care Medicine Experimental

MEETING ABSTRACTS

Open Access

ESICM LIVES 2016: part one Milan, Italy. 1-5 October 2016 Published: 29 September 2016

About this supplement These abstracts have been published as part of Intensive Care Medicine Experimental Volume 4 Suppl 1, 2016. The full contents of the supplement are available online at http://icm-experimental.springeropen.com/articles/supplements/volume-3-supplement-1. Please note that this is part 1 of 3.

Oral Sessions. ARDS: CLINICAL STUDIES A1 Identification of distinct endophenotypes in patients with acute respiratory distress syndrome by unbiased cluster analysis, and their association with mortality L. Bos1, L. Schouten1, L. van Vught1, M. Wiewel1, D. Ong2, O. Cremer2, A. Artigas3, I. Martin-Loeches4, A. Hoogendijk1, T. van der Poll1, J. Horn1, N. Juffermans1, M. Schultz1 1 Academic Medical Center, University of Amsterdam, Amsterdam, Netherlands; 2UMCU, Utrecht, Netherlands; 3Autonomous University of Barcelona, Barcelona, Spain; 4Hospital São Francisco Xavier, Lisbon, Portugal Correspondence: L. Bos – Academic Medical Center, University of Amsterdam, Amsterdam, Netherlands Intensive Care Medicine Experimental 2016, 4(Suppl 1):A1 Introduction: Pharmacological immunomodulatory interventions in 'acute respiratory distress syndrome' (ARDS) have been unsuccessful in clinical trials [1-3] despite promising results in preclinical studies using animals [4-5]. Poor phenotyping of patients could be responsible for these disappointing results. Objectives: We hypothesized that ARDS patients can be clustered based on concentrations of plasma biomarkers and that such biological endophenotypes are association with clinical outcomes. Methods: Patients were screened for presence of ARDS. Unbiased cluster analysis of plasma concentrations of 20 biomarkers of inflammation, coagulation and endothelial activation at diagnosis of ARDS provided the endophenotypes. A decision tree was then used to predict cluster membership based on a more restricted set of biomarkers. The independent association of endophenotypes with ICU mortality was studied by multivariate logistic regression. Results: Three endophenotypes of ARDS were identified in 771 patients, which we named 'impassive' (N = 383), 'intermediate' (N = 224) and 'reactive' (N = 164), had mortality rates of 16 %, 26 % and 47 %, respectively (P < 0.01). Patients with a 'reactive' endophenotype were younger, had higher disease severity scores, more failing organs and more frequently had an indirect cause for ARDS than patients with an 'impassive' or 'intermediate' endophenotype. A 'reactive endophenotype' was independent from confounders associated with ICU mortality (OR 1.18 [95 % confidence interval: 1.09-1.28]). The concentration of interleukin 10, interleukin 8 and matrix metalloproteinase 8 were sufficient to predict the three endophenotypes. Conclusions: ARDS patients can be clustered into three biological endophenotypes, with different mortality rates. Three easy to measure biomarkers can be used to predict the endophenotype. References 1. Takeda S. Pulm Pharmacol Ther 2005. 2. Boyle AJ. Expert Opin Biol Ther 2014. 3. Cepkova M. J Intensive Care Med 2006. 4. Calfee CS. Chest 2007. 5. Beitler JR. Chest 2014.

Grant acknowledgement This study is supported by the MARS consortium, a public-private partnership.

Table 1 (abstract A1). Endophenotypes versus clinical characteristics Impassive endophenotype

N=383

Intermediate endophenotype

N=224

Reactive endophenotype

N=164

P-value

Age

63

(53.5-72)

61.5

(51.8-72)

58

(45-66)

7.5 ug / ml).

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Methods: A prospective, observational and descriptive study in one ICU. All patients who received intravenous LNZ (60-minutes infusion) at doses of 600 mg every 12 hours, for treatment of a suspected or confirmed CGP-MR infection in which it was possible to obtain a blood sample (5 ml) after the third day (steady state) for the determination of the Cmin, just before administering the next dose, were included. Patients under renal replacement therapy were excluded. Demographic variables, comorbidities, severity on admission and analytical data were evaluated. LNZ quantification was performed using a high performance liquid chromatography technique (HPLC). Differences between groups were assessed using chi-square for categorical variables and Student's t-test or Mann-Whitney test for continuous variables. Significant variables in the univariate analysis were included in a multivariate model (logistic regression) to identify the variables related to sub/supratherapeutic levels. We considered p < 0.05 to be significant. The predictive value of each formula was calculated using a receiver operator characteristic curve (ROC), and the area under the curve (AUC) was also computed. Results: A total of 103 patients were included. In 58 of them (56.3 %) the Cmin was < 2 ug / ml and in 30 (29.1 %) the Cmin was >7.5 ug / ml. Glomerular filtration rate (OR 1.01, 95 % CI 1.00- 1.02; p = 0.002) and diabetes mellitus (OR 0.24, 95 % CI 0.07-0.83; p = 0.02) were independently variables associated with subtherapeutic concentrations. The discrimination ability model obtained an AUC (95 % CI) of 0.808 (0.72 to 0.89). Liver cirrhosis (OR 14.51, 95 % CI 1.35-156.40; p = 0.03) and glomerular filtration rate (OR 0.98, 95 % CI 0.97 to 0.99 p = 0.002) were independently variables associated with supratherapeutic concentrations. The discrimination ability model obtained an AUC (95 % CI) 0.81 (0.72 to 0.898). Conclusions: Variables related to sub and supra-therapeutic LNZ concentrations in critically ill patients admitted to the ICU have been identified. Presence of these variables would require the determination of plasma levels of LNZ. Glomerular filtration rate is the variable that influences both groups.

A19 Validation of a checklist for high risk for multirresistant bacteria according to the “resistencia zero” project S. Carvalho Brugger, G. Jimenez Jimenez, M. Miralbés Torner, J. Trujillano Cabello, B. Balsera Garrido, X. Nuvials Casals, F. Barcenilla Gaite, M. Vallverdú Vidal, M. Palomar Martínez Hospital Universitari Arnau de Vilanova, Lleida, Spain Correspondence: S. Carvalho Brugger – Hospital Universitari Arnau de Vilanova, Lleida, Spain Intensive Care Medicine Experimental 2016, 4(Suppl 1):A19 Introduction: Multirresistant bacteria (MRB) development is a growing phenomenon. In 2013, the “Zero Resistance” (RZ) program was launched in Spain, to help prevent the emergence of MRB in critically ill patients. One of its recommendations is to complete a checklist upon patient admission in Intensive Care Unit (ICU) to identify those patients at high risk for colonization or infection by MRB. AIMS. To analyse the effectiveness of the checklist for risk factors (RF) proposed by the RZ project as a way of MRB early detection. Methods: A prospective study from March/2014 to January/2016. All patients admitted to a polyvalent ICU of a general hospital were submitted to the checklist proposed, with the application of contact precaution (CP) strategies for patients with RF for MRB. Bacteriologic swabs (nasal, pharyngeal, axillary and rectal) were routinely performed on all patients admitted, besides diagnostic cultures when necessary. Furthermore, we analysed other pathological variables and comorbidities (diabetes, renal failure, immunosuppression state, neoplasia, cirrhosis, chronic obstructive pulmonary disease -COPD-, organ transplantation, malnutrition or type of admission to ICU - urgent or programmed). Univariate and multivariate analysis of RF for MRB with binary logistic regression were performed. Statistical significance was set at CI 95 %. Results: 1651 patients were admitted. 532 (32,2 %) met some CP criteria. In 136 (8,2 %) were detected one or more MRB, 87 of these (64 %) presented CP criteria according to the checklist. 37 met 1

Page 12 of 607

criteria, 31 met 2 criteria and 19 met 3 or more criteria with accumulation of risk (p < 0,001). In 49 (36 %) MRB carriers it was not identified any of the RF from the checklist. Tables 13 and 14 show risk factors and comorbidities that were significant as added risk for MRB. Conclusion: After comparing to previous analysis, it was seen that, increasing the sample size, almost all RF included in the RZ checklist were predictors of MRB. Despite this, it could not detect 36 % of the patients infected or colonized by MRB. Because of that, we should consider other factors to predict the presence of a MRB on admitted patients to ICU. Reference J Montero et al (Scientific Expert Committee for the “Zero Resistance” Project). Combatting resistance in intensive care: the multimodal approach of the Spanish ICU “Zero Resistance” program. Critical Care (2015) 19:114.

Table 13 (abstract A19). Risk Factors for MRB RZ CHECK LIST

OR (IC 95%) UNIVARIATE

OR (IC 95%) MULTIVARIATE

Hospitalization >5 days in past 3 months

3,1 (2,1-4,4)

1,7 (1,1- 2,6)

Institutionalized patient

4,2 (2,1-8,6)

3,0 (1,3- 6,9)

Prior MRB colonization/ infection

17,8 (9,4-33,8)

11,3 (5,5-23,2)

Antibiotherapy >7 days in the past month

4,2 (2,9-6,2)

2,1 (1,3- 3,6)

Chronic kidney disease with dialysis

2,0 (0,4-9,3)

NS

Colonization susceptibility (bronchiectasis, cystic fibrosis)

2,4 (1,0-5,4)

NS

Table 14 (abstract A19). Relation between comorbidities and MRB COMORBIDITY

OR (IC 95%) UNIVARIATE

OR (IC 95%) MULTIVARIATE

DM type II

1,5 (1,1-2,2

NS

COPD

1,9 (1,3-3,0)

NS

Malnutrition

1,8 (1,2-2,6)

NS

Cirrhosis

3,9 (2,1-7,1)

2,7 (1,4-5,4)

APACHE II >25

2,46 (1,6-3,8)

2,2 (1,4-3,6)

A20 Influence of antibiotic stewardship program on etiology and clinical outcomes in bacteriemic patients in a multidisciplinary surgical hospital V. Gusarov1, D. Shilkin1, M. Dementienko1, E. Nesterova1, N. Lashenkova2, A. Kuzovlev1, M. Zamyatin1 1 N.I. Pirogov National Medical Surgical Center, Anestesiology and Intensive Care, Moscow, Russian Federation; 2N.I. Pirogov National Medical Surgical Center, Microbiology Laboratory, Moscow, Russian Federation Correspondence: V. Gusarov – N.I. Pirogov National Medical Surgical Center, Anestesiology and Intensive Care, Moscow, Russian Federation Intensive Care Medicine Experimental 2016, 4(Suppl 1):A20 Introduction: Development of candidemia and bacteremia caused by multidrug-resistant (MDR) pathogens in hospitalized patients significantly increases the risk of adverse outcome [1-3]. The main task of the antibiotic stewardship program (ASP) is to reduce the microbial resistance and decrease incidence of infectious complications caused by MDR pathogens. Objectives: To estimate the etiological structure and clinical outcome in patients with bacteremia in a multidisciplinary hospital in the period before and after the implementation of ASP. Methods: Interventional study with historical controls. The intervention: ASP, including a group of experts on antimicrobial therapy (AMT), AMT and prophylaxis protocols, infection control, internal audit. Study onset: June 2013. Analysis of pre-intervention (Jan. 2011-June 2013) and

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

intervention (July 2013-Dec. 2015) periods was carried out. We assessed incidence of bacteremia caused by methicillin-resistant S. aureus (MRSA), vancomycin-resistant Enterococci (VRE), ESBL-producing Enterobacteriaceae, MDR Gram-negative non-fermenting bacteria (MDRNFGB), incidence of candidemia. The hospital length of stay (LOS) and mortality of patients with bacteremia were secondary end-points. Results: In the pre-intervention period 2905 blood samples were analyzed, 448 (15.4 %) positive blood cultures from 206 patients were obtained; in the interventional period 2742 blood samples were analyzed, 470 (17.1 %) positive blood cultures from 230 patients were obtained. In the intervention period a significant decrease in the incidence of bacteremia caused by MDR-NFGB from 55 (12.3 %) to 20 (4.3 %), p < 0.001, VRE from 10 (2.2 %) to 0, p = 0.0011 and incidence of candidemia from 29 (6.5 %) to 13 (2.85), p < 0.01, was detected. The frequency of bacteremia caused by ESBL-producing Enterobacteriae did not significantly change [91 (20.3 %) vs. 108 (23.0 %), p > 0.05]. Number of MRSA bacteremias remained low [6 (1.3 %) vs. 4 (0.9 %), p > 0.05]. There was a significant reduction of mortality from 30.6 % to 20.3 %, p < 0.025 in the group of patients with bacteremia. The LOS did not change significantly [29 (IQI 19-52) vs. 27 (IQI 17-53), p > 0.05]. Conclusions: Implementation of ASP in multidisciplinary surgical hospital allows significantly decrease incidence of candidemia and bacteremia caused by MDR-NFGNB and VRE. As a result - reduction in mortality in patients with bacteremia. References 1. Bassetti M., Merelli M., Ansaldi F. et al. Clinical and Therapeutic Aspects of Candidemia: A Five Year Single Centre Study. PLoS One. 2015;10(5): e0127534. 2. Tam V., Rogers C., Chang K. et al. Impact of multidrug resistant Pseudomonas aeruginosa bacteremia on patient outcomes. Antimicrob Agents Chemother 2010;54(9):3717-22. 3. Blot S., Vandewoude K., Colardyn F. Nosocomial bacteremia involving Acinetobacter baumannii in critically ill patients: a matched cohort study. Intensive Care Med. 2003 Mar;29(3):471-5.

ADJUNCTIVE INTERVENTIONS IN ACUTE RESPIRATORY FAILURE A21 Impact of earplugs and eye mask on sleep in critically ill patients: a prospective polysomnographic study A. Demoule, S. Carreira, S. Lavault, O. Palancca, E. Morawiec, J. Mayaux, I. Arnulf, T. Similowski Pitie-Salpetriere Hospital and Pierre and Marie Curie University, Paris, France Correspondence: A. Demoule – Pitie-Salpetriere Hospital and Pierre and Marie Curie University, Paris, France Intensive Care Medicine Experimental 2016, 4(Suppl 1):A21 Rationale: Poor sleep is common in intensive care unit (ICU) patients and environmental factors contribute to sleep deprivation and alterations of sleep architecture. The objective of the present study was to evaluate the impact of earplugs and eye masks on sleep architecture in ICU patients. Patients and methods: A single center randomized controlled trial of 64 ICU patients was conducted from July 2011 to December 2013. Patients were randomly assigned to sleep with or without earplugs and a facemask. A polysomnography was performed on the first day and night following inclusion. The primary end point was the proportion of sleep stage 3 + 4. Secondary end points were other descriptors of sleep and major outcome variables. Results: In the intervention group, 33 % of patients did not wear earplugs all night long. The proportion of sleep stage 3 + 4 was 11 [3-23]% in the control group and 13 [6-23]% in the protective group (p = 0.72). Other descriptors of sleep were not different between the two groups except the number of long awaking that was lower in the protective group than in the control group (21 [19-26] vs. 31 [21-47], p = 0.02). There was no difference among the two groups in terms of sleep quality, occurrence of a delirium, ICU length of stay and

Page 13 of 607

mortality, anxiety and depression on ICU discharge and day-90 and the incidence of post-traumatic stress disorder. Conclusion: In ICU patients, earplugs and eye mask are not well accepted by patients, do not increase the proportion of sleep stage 3 + 4 but decrease the number of prolonged awakenings. They had no impact on the outcome. Grant acknowledgement French Ministry of Health (PHRC) Note: This abstract has been previously published and is available at [1]. It is included here as a complete record of the abstracts from the conference. References 1. Demoule A, Carreira S, Lavault S, Pallanca O, Morawiec E, Mayaux J, Arnulf I, Similowski T (2016) Impact of earplugs and eye mask on sleep in critically ill patients: a prospective randomized polysomnographic study. Annals of Intensive Care 6(Suppl 1): S9

A22 Metabonomics identifies early molecular changes associated with progression into postoperative hypoxemia in cardiac surgery patient: a human model that can provide new insights into the pathophysiology of acute lung injury and potentially identify specific biomarkers of lung tissue injury B.S. Rasmussen1,2, R.G. Maltesen2, M. Hanifa2, S. Pedersen2,3, S.R. Kristensen2,3, R. Wimmer4 1 Aalborg University Hospital, Department of Anaesthesia and Intensive Care Medicine, Aalborg, Denmark; 2Aalborg University, Department of Clinical Medicine, Aalborg, Denmark; 3Aalborg University Hospital, Department of Clinical Biochemistry, Aalborg, Denmark; 4Aalborg University, Department of Chemistry and Bioscience, Section of Biotechnology, Aalborg, Denmark Correspondence: B.S. Rasmussen – Aalborg University Hospital, Department of Anaesthesia and Intensive Care Medicine, Aalborg, Denmark Intensive Care Medicine Experimental 2016, 4(Suppl 1):A22 Introduction: Postoperative pulmonary dysfunction after cardiac surgery with the use of cardiopulmonary bypass (CPB) is common, ranging from transient hypoxemia to acute respiratory distress syndrome (ARDS). It is triggered by an inflammatory response, disrupted coagulation, ischemia-reperfusion injury, and oxidative stress. However, as the nadir of partial pressure of arterial oxygen (PaO2) appears on the second to third postoperative day, it is of paramount importance to identify patients at risk at an early stage. Determining the progression into hypoxemia is challenging, as no early diagnostic test exists. We hypothesized that metobonomics can provide new insights into the pathogenesis of pulmonary dysfunction, thus potentially enabling the identification of specific biomarkers of lung injury. Objectives: To use a human model of cardiac surgery to study the progression into postoperative hypoxemia by means of blood metabonomics. Methods: Fifty consecutive patients undergoing cardiac surgery were included. Arterial blood samples were collected the day before surgery, and 48 and 72 hours (h) postoperatively. In addition, samples from the pulmonary artery and the left atrium were collected at seven different time points: just before CPB, straight after CPB (0 h), and 2, 4, 8, 20, 48, and 78 h after CPB. Samples were analyzed by nuclear magnetic resonance (NMR) spectroscopy. Statistical methods, including principal component analysis (PCA), partial least-square regression (PLS) and discriminant analysis (PLS-DA), were applied to find metabolite patterns related to surgical trauma and postoperative hypoxemia. Venetian-Blinds cross-validation and permutation testing were used for validation. Results: All patients had a normal preoperative PaO2; 11.3 ± 1.3 kPa (mean ± SD). PaO2 decreased to 7.9 ± 1.5 kPa at 48 h, and to 7.8 ± 1.4 kPa at 72 h, respectively. Twenty-three patients developed moderate hypoxemia (PaO2 ≤ 8 kPa) and 9 patients had severe hypoxemia (PaO2 ≤ 6 kPa). NMR showed major shifts in the metabolome during and after surgery. Different metabolites were identified at different

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

timepoints, of which several were directly correlated to the progression into hypoxemia. PLS-DA modeling predicted postoperative hypoxemia at 72 h with 72.2 % sensitivity and 81.0 % specificity based on blood samples collected just before CPB, while PLS models revealed a correlation of r2 equal to 0.7. Higher predictive values (>88 % sensitivity, >92 % specificity, and r2 > 0.91) were achieved when analyzing blood collected 0, 2, 4, 8, and 20 h postoperatively, demonstrating the ability of metabonomics in early diagnosis. Conclusion: We found that metabonomics may contribute to the detection of early signs of pulmonary dysfunction two-three days before significant postoperative hypoxemia. The study provides novel insights into the underlying mechanisms that trigger progression into hypoxemia, facilitating new hypotheses and treatment options. A23 Multicenter randomized clinical trial of lateral-trendelenburg vs. semi recumbent position for the prevention of ventilatorassociated pneumonia - the gravity-VAP trial M. Panigada1, G. Li Bassi2, O.T. Ranzani2, T. Kolobow3, A. Zanella1, M. Cressoni1, L. Berra4, V. Parrini5, H. Kandil6, G. Salati7, S. Livigni8, A. Amatu9, A. Andreotti10, F. Tagliaferri11, G. Moise12, G. Mercurio13, A. Costa14, A. Vezzani15, S. Lindau16, J. Babel17, M. Cavana18, D. Consonni1, A. Pesenti1, L. Gattinoni19, A. Torres2, for the GRAVITY-VAP TRIAL NETWORK 1 Policlinico Di Milano, Milan, Italy; 2Hospital Clinic, Barcelona, Spain; 3 National Institutes of Health, Bethesda, United States; 4Massachusetts General Hospital, Boston, United States; 5Ospedale Nuovo del Mugello, Borgo San Lorenzo (FI), Italy; 6Gruppo Ospedaliero San Donato, San Donato M.se, Italy; 7Arcispedale S. Maria Nuova, Reggio Emilia, Italy; 8 Ospedale San Giovanni Bosco, Torino, Italy; 9Policlinico San Matteo, Pavia, Italy; 10Policlinico di Modena, Modena, Italy; 11Azienda Ospedaliero - Universitaria di Parma, Parma, Italy; 12Ospedale Città di Sesto San Giovanni, Sesto San Giovanni, Italy; 13Policlinico Gemelli, Roma, Italy; 14. Azienda Ospedaliero-Universitaria di Parma, Parma, Italy; 15Azienda Ospedaliero-Universitaria di Parma, Parma, Italy; 16University Hospital Frankfurt, Frankfurt, Germany; 17University Hospital Center Zagreb, Zagreb, Croatia; 18Ospedale Santa Chiara, Trento, Italy; 19University of Göttingen, Göttingen, Germany Correspondence: M. Panigada – Policlinico Di Milano, Milan, Italy Intensive Care Medicine Experimental 2016, 4(Suppl 1):A23 Introduction: Gravity plays a pivotal role in the pathogenesis of ventilator-associated pneumonia (VAP) (1). In previous laboratory studies (2) the semi-lateral Trendelenburg position (LTP) hindered gravity-driven pulmonary aspiration and avoided VAP. Objectives: To determine whether the LTP vs. the semi-recumbent position (SRP) would reduce the incidence of microbiologically confirmed VAP and to appraise patient's compliance and safety. Methods: We conducted a randomized, single-blind, controlled study in 17 European centers and 1 in North America. A total of 2019 adult patients were screened between 2010 and 2015. 395 patients were randomized - 194 in LTP and 201 in SRP - and analyzed in an intention to treat approach. Patients in LTP were placed in semi-lateral (60°) Trendelenburg position to achieve an orientation, from the sternal notch toward the mouth, slightly below horizontal, and turned from one side to the other every 6 hours. LTP was encouraged during the first days of mechanical ventilation, but always in compliance with the patient's wish. In the SRP group, the head of the bed was elevated ≥ 30°. Primary outcome was VAP incidence rate, based on quantitative bronchoalveolar lavage fluid culture with ≥ 104 colonyforming units/mL. Secondary outcomes were compliance to the randomized position, length of intubation, duration of intensive care unit and hospital stay, mortality, and adverse events. Results: The trial was stopped after the planned interim analysis for achieving efficacy endpoints and owing to safety concerns. Patients

Page 14 of 607

in the LTP and SRP group were kept in the randomized position for 38 % and 90 % of the study time, respectively (p = 0.001). Yet, during the first 48 hours, LTP patients were kept in the randomized position for 50 % of the study time, and SRP patients for 88 % (p = 0.001). In the LTP, the bed was angulated 5.6° in Trendelenburg; while, the head of the bed was elevated 34.1° in the SRP group. Incidence rates of microbiologically confirmed VAP were 0.88 (1/1136 patient-days; 95 % confidence interval [CI], 0.12-6.25) in the LTP group, and 7.19 (8/1113 patient-days; CI 95 %, 3.60-14.37) in the SRP (p = 0.020), relative risk reduction of 0.12 (95 % CI, 0.01-0.91). No statistically significant differences were observed in durations of mechanical ventilation, intensive care unit and hospital stay, and mortality. Vomiting was more common in LTP patients (8.3 % vs. 2.5 % in the SRP, p = 0.013). Conclusions: Critically ill patients positioned in the LTP had a statistically significant reduction in the incidence of VAP, compared with those positioned in the SRP. A comprehensive evaluation of potential LTP contraindications is warranted to enhance safety. References 1) Li Bassi G et al. Crit Care Med 2014; 42: e620-7 2) Zanella et al. Intensive Care Med 2012; 33: 677-85 Grant acknowledgement ClinicalTrials.gov ID: NCT01138540. The study was endorsed by ECCRN/ESICM. Funding by 2013 ECCRN Clinical Trial Award and Hill-Rom

A24 The effects of oral rinse with 0.2 % and 2 % chlorhexidine on oropharyngeal colonization and ventilator associated pneumonia in adults' intensive care units P. Mansouri1, F. Zand2, L. Zahed3, F. Dehghanrad3, M. Bahrani3, M. Ghorbani3 1 School of Nursing, Shiraz University of Medical Science, Shiraz, Islamic Republic of Iran; 2Shiraz University of Medical Sciences, Anesthesiology and Critical Care Research Center, Shiraz, Islamic Republic of Iran; 3Shiraz University of Medical Sciences, Shiraz, Islamic Republic of Iran Correspondence: F. Zand – Shiraz University of Medical Sciences, Anesthesiology and Critical Care Research Center, Shiraz, Islamic Republic of Iran Intensive Care Medicine Experimental 2016, 4(Suppl 1):A24 Introduction: Ventilator Associated Pneumonia (VAP) is the most common nosocomial infection in Intensive Care Units (ICUs), which increases the length of ICU stay, duration of mechanical ventilation, and mortality. Objectives: The present study used an oral care protocol and compared the effects of two different concentrations of chlorhexidine on reduction of oropharyngeal colonization and VAP. Methods: This study was performed on 114 patients from trauma, surgery, neurosurgery, and general ICUs randomly allocated to two groups under oral care with 0.2 % and 2 % chlorhexidine solution. A multidisciplinary team approved the oral care protocol. The data were collected using a demographic information form, Apache IV form, Beck oral assessment scale, mucosal-plaque assessment scale, and oropharyngeal swab culture. Results: The results showed a significant reduction in VAP (p = 0.007) and oropharyngeal colonization (p = 0.007) in the group under oral care with 2 % chlorhexidine solution compared to the other group. However, no significant difference was found between the two groups in terms of oropharyngeal adverse effects (p = 0.361). Conclusions: Oral decontamination with 2 % compared to 0.2 % chlorhexidine is a more effective method in prevention of VAP and reduction of oropharyngeal colonization (especially gram-positive). Grant acknowledgement This study supported by a grant (grant No. 7363) from Vice-chancellor in Research, Shiraz University of Medical Sciences

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

A25 Acute consequences of lobar unilateral pulmonary perfusion block in an animal model of high tidal volume ventilation and lavage B. Cambiaghi1, O. Moerer2, T. Mauri3,4, N. Kunze-Szikszay2, C. Ritter2, A. Pesenti5, M. Quintel2 1 University of Milan - Bicocca, Monza, Italy; 2University of Goettingen, Goettingen, Germany; 3Fondazione IRCCS Ospedale Maggiore Policlinico, Anesthesia and Critical Care, Milan, Italy; 4Plug Working Group, ESICM, Milan, Italy; 5University of Milan, Milan, Italy Correspondence: B. Cambiaghi – University of Milan - Bicocca, Monza, Italy Intensive Care Medicine Experimental 2016, 4(Suppl 1):A25 Introduction: Previous studies showed that, in the presence of preserved ventilation, interruption of regional pulmonary blood flow rapidly leads to severe ventilation-induced lung injury (VILI) [1-2]. These data suggest that tissue alkalosis might be a multiplication factor for VILI [3], especially in the presence of high tidal volume (Vt) ventilation and/or pre-existing lung injury. Objectives: We investigated the alterations in blood gases, respiratory mechanics and CT-scan imaging in an animal model of controlled unilateral lobar pulmonary artery occlusion followed by lung lavage and high Vt ventilation. Methods: We report data from 20 pigs (weight 61 ± 2 Kg). Animals were randomly assigned to the following 5 study groups: 1. Five animals underwent right lower lobar endovascular embolization, lung lavage with 500 ml of normal saline solution and high Vt ventilation for 8 hours; 2. Four animals underwent left lower lobar embolization, lavage and high Vt ventilation; 3. Five pigs received lung lavage and high Vt ventilation; 4. Three animals received right lower lobar embolization and low Vt ventilation; 5. Three animals received only low Vt ventilation. In groups 1, 2 and 4 selective embolization of the pulmonary artery branch perfusing the lower lobe of the right or left lung was performed under fluoroscopy guidance by expert radiologist; in group 1-3, Vt was set at 20-25 ml/Kg with plateau pressure ≤40 cmH2O and peak pressure ≤50 cmH2O, without positive end-expiratory pressure (PEEP) and respiratory rate of 7-9 breaths/minute to control pH. In groups 3 and 5, Vt was 6-8 ml/Kg with PEEP 5 cmH2O. After 8 hours, the animals underwent physiologic data collection and chest CT scan at zero PEEP. CT images were analysed offline by custom-made software. Data were analysed by one-way ANOVA and Tukey post-hoc tests. Results: Main parameters describing pathophysiological characteristics of each group after 8 hours are reported in Tables 15 and 16. Lung lavage and high Vt ventilation following regional embolization led to more severe impairment of oxygenation (p < 0.001), especially in the left embolization group. Lung mechanics at the end of the experiment showed similar trend: plateau pressure was higher and respiratory system compliance lower (p < 0.05) in the embolization + lavage + high Vt groups. Finally, CTscan revealed higher lung weight in groups 1 and 2 (p < 0.05). Interestingly, in the same groups, the non-embolized lungs were heavier, as if embolization triggered more severe ventilationinduced lung oedema that could develop mainly where perfusion was preserved. Conclusions: Regional block of lung perfusion might amplify lung injury due to lavage and high Vt ventilation. References [1] Slutsky AS N Engl J Med 2013 [2] Kolobow T Inter J Artif Organs 1981 [3] Ando T J Int Med Res 2007 Grant acknowledgement Departmental fundings.

Page 15 of 607

Table 15 (abstract A25). Ventilation pattern and gas exchange Right embolization + lavage + high Vt (n= 5)

Left embolization + lavage + high Vt (n= 4)

Lavage + high Vt (n= 5)

Right embolization + low Vt ventilation (n= 3)

Low Vt ventilation (n=3)

P-value ANOVA

Vt (ml/kg)

20.1 ± 2.5°#

18.7 ± 2.7°#

23.5 ± 2.3*°#

8.1 ± 0.2

8.0 ± 0.3

90 % purity post sorting). Genes of interest were quantified using polymerase chain reaction (PCR) using TaqMan® labelled primers and real time PCR (ABI HT7900). Flow cytometry data was analysed using FlowJo. Continuous variables were analysed using a Wilcoxin signed-rank test (JMP (version 11) statistical software). Results: PMBCs incubated with post-operative serum demonstrated a significant reduction in mHLA-DR membrane density (Fig. 19, p = 0.001). The reduction in mHLA-DR density was prevented when co-incubated with GM-CSF and IFNγ (Fig. 19). Incubation with IFNγ but not GM-CSF increased expression of HLA-DRα chain (p = 0.01), Cathepsin S (CTSS) (p = 0.001), suppressor of cytokine signalling 3 (SOCS3) (p = 0.01) and March 1 (p = 0.002) (Fig. 20). Conclusions: These results suggest that these immune stimulants, GM-CSF and IFNγ, exert their effects on monocyte antigen presentation through different signalling pathways. The increased gene expression associated with IFNγ may be indicative of potential therapeutic benefit in reversing post-operative immune suppression. References 1. Torrance et al., Curr Opin Anaesthesiol. 2016 Mar 9.[Epub ahead of print] 2. Fragkou et al., EJA Volume 31, e-supplemental 52, pg 141AP4-2, June 2014. Grant acknowledgement NIAA (BJA/Royal College of Anaesthetists) and the European Society of Anaesthesiologists.

Fig. 19 (abstract A32). Effects of GM-CSG, INFγ on mHLA-DR geometric mean intesity in post-operative serum. Results are depicted in a dot plot with media and IQR, n=12, 3 independent experiments

Fig. 20 (abstract A32). mRNA levels in monocytes incubated with post-operative serum containing either GM-CSF or INFγ. Dot plots, with median and IQR represented, n=12

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

A33 Impact of a tailored multifaceted implementation of the pain, agitation and delirium guidelines in adult critically ill patients on guideline adherence, delirium and clinical outcomes: a prospective multicenter before-after study Z. Trogrlic1, M. van der Jagt1, H. Lingsma2, H.H. Ponssen3, J.F. Schoonderbeek4, F. Schreiner5, S.J. Verbrugge6, S. Duran7, T. van Achterberg8, J. Bakker1, D.A.M.P.J. Gommers1, E. Ista9 1 Erasmus Medical Center, Department of Intensive Care, Rotterdam, Netherlands; 2Erasmus Medical Centre, Department of Public Health, Rotterdam, Netherlands; 3Albert Schweitzer Hospital, Department of Intensive Care, Dordrecht, Netherlands; 4Ikazia Hospital, Department of Intensive Care, Rotterdam, Netherlands; 5IJsselland Hospital, Department of Intensive Care, Rotterdam, Netherlands; 6Sint Franciscus Gasthuis, Department of Intensive Care, Rotterdam, Netherlands; 7Maasstad Hospital, Department of Intensive Care, Rotterdam, Netherlands; 8KU Leuven, Academic Centre for Nursing and Midwifery, Leuven, Belgium; 9 Erasmus Medical Center - Sophia Children's Hospital, Department of Pediatric Surgery, Intensive Care Unit, Rotterdam, Netherlands Correspondence: Z. Trogrlic – Erasmus Medical Center, Department of Intensive Care, Rotterdam, Netherlands Intensive Care Medicine Experimental 2016, 4(Suppl 1):A33 Introduction: Delirium in critically ill patients is associated with increased mortality and long-term cognitive decline. The recent Pain, Agitation and Delirium (PAD) guidelines include recommendations for delirium screening, prevention and management, which may help to improve clinical outcomes and reduce delirium burden. Objectives: We aimed to measure the effectiveness of a multifaceted implementation program for improved adherence to the PAD guidelines and associated changes in delirium incidence and duration, length of ICU stay and hospital mortality. Methods: A prospective multicenter before-after study was conducted in six ICUs in the Netherlands between March 2012 and April 2015. The intervention consisted of a two-phase multifaceted tailored implementation of the PAD guidelines. Multiple implementation strategies were applied to change clinical practice (Table 18). Data of all adult ICU patients were collected during three four-month periods: 1) before implementation, 2) after implementation of delirium screening, and 3) after implementation of other PAD guidelines. The difference in adherence and clinical outcomes between the three periods was assessed with random effects Poisson and logistic regression with a random intercept for patients for outcomes on patient-day level and with logistic and linear regression for outcomes on patient level, adjusted for APACHE II, hospital, age and admission type. Differences were expressed as adjusted rate ratios (aRR), odds ratios (aOR) or beta's with the “before”-period as the reference. Results: A total of 4727 patients were enrolled in the study with a total of 23958 ICU days. Adherence to most PAD guideline recommendations improved significantly whereas early mobilization and reduced benzodiazepine sedation improved only in the last period (Table 19). The incidence of delirium increased from 22 % before to 30 % after implementation (aOR = 1.5, p < 0.01) whereas delirium duration decreased from 6.3 before to 3.6 days after implementation (aBeta = -2.6, p < 0.01). There were no statistically significant differences in ICU length of stay (aBeta = 0.001, p = 0.99); ICU mortality (aOR = 1.2, p = 0.17); and hospital mortality (aOR = 1.2, p = 0.21) after vs. before the implementation. Only length of mechanical ventilation increased with half a day (aBeta = 0.55, p = 0.01). Conclusions: This multifaceted implementation program was effective in improving adherences to multiple key PAD guideline recommendations. Delirium duration decreased significantly in spite of increased incidence probably due to improved screening. We observed differential effects of screening vs. further guideline implementation. However, these improved adherences to guideline recommendations did not translate into measurable improvements of short-term clinical outcomes. To improve clinical outcomes, future investigations on PAD

Page 21 of 607

guideline implementation should consider focusing on specific recommendations or targeting long-term outcomes. Table 18 (abstract A33). Implementation strategies useda Category

Guideline implementation

Professional

1. Distribution of educational materials 2. Educational meetings 3. Local consensus processes 4.Outreach visits 5.Local opinion leader 6. Patient-mediated intervention 7. Audit and feedback 8. Reminders 9.Marketing / Tailored interventions 10.Mass media

Organizational 11. Provider oriented interventions 12. Patient oriented interventions 13. Structural interventions Financial

14.Provider or patient interventions

Regulatory

15.Changes in medical liability 16.Peer review or Licensure

a

Implementation strategies used are described according to EPOC (Cochrane Effective Practice and Organization of Care) classification system. Only "Management of patients complaints" strategy was not used in this study

Table 19 (abstract A33). Adherence to guideline recommendations Performance indicator

Period I Before Implementation

Period II After Screening Implementation

Period III After Guideline Implementation

OR or RR (95% CI)

p

no. patient-days (n= 4727 patients)

8005

8207

7746

Delirium screening (assessments / patient-days)

0.80

2.4

2.6

4.6 (4.4, 4.9)

0.05

A38 The effectiveness of a recovery program aimed at improving quality of life and sense of coherence in post intensive care patients: a pragmatic multicenter randomized controlled trial, the recovery and aftercare of post intensive care patients (RAPIT) study J.F. Jensen1, I. Egerod2,3, M.H. Bestle1, D.F. Christensen1, A. Alklit4, R.L. Hansen1, H. Knudsen5, L.B. Grode6, D. Overgaard7,8, The RAPIT group 1 Nordsjællands Hospital, University of Copenhagen, Department of Anesthesiology, Hillerød, Denmark; 2University of Copenhagen, Health & Medical Sciences, Copenhagen, Denmark; 3University of Copenhagen, Rigshospitalet, Department of Neuroanesthesiology, Copenhagen, Denmark; 4University of Southern Denmark, Department of Psychology, Odense, Denmark; 5Herlev Hospital, University of Copenhagen, Department of Anesthesiology, Herlev, Denmark; 6Hospital of Horsens, Department of Anesthesiology, Horsens, Denmark; 7Metropolitan University, Department of Nursing, Copenhagen, Denmark; 8 Nordsjællands Hospital, Research Unit, Hillerød, Denmark Correspondence: J.F. Jensen – Nordsjællands Hospital, University of Copenhagen, Department of Anesthesiology, Hillerød, Denmark Intensive Care Medicine Experimental 2016, 4(Suppl 1):A38 Introduction: More ICU programs are emerging to promote psychological recovery after a stay in intensive care unit (ICU) (1). In the Scandinavian countries, patients' discharge rehabilitation plan usually includes physical training, but seldom psychological rehabilitation. To address this gap, we developed a post-ICU recovery program to improve psychological health after intensive care. We hypothesized that a nurse-led recovery program, that helped to construct a coherent illness narrative including person-centered communication, would improve health-related quality of life (HRQOL), sense of coherence (SOC), reduce symptoms of anxiety, depression, and post-traumatic stress (PTSD) in the first year after ICU discharge. Objectives: To investigate the effectiveness of a post-ICU recovery program compared to standard care in improving quality of life in the first year after ICU. Methods: We randomly assigned 386 patients adult (≥18 years) survivors after receiving mechanical ventilation (≥48 hours) to standard care (SC) plus a recovery program or SC alone after discharge from intensive care. It was a nurse-led intervention consisted of patient photographs, three follow-up consultations and reflection sheets. Results: Primary outcome was HRQOL, and secondary outcomes were SOC, anxiety, depression, PTSD assessed at 3 and 12 months after intensive care discharge using t-tests. We supplemented outcomes with rehabilitation services within the first year. At 12 months after intensive care the HRQOL scores were unchanged by the intervention (mean difference in the Mental Component Summary score, 0.9 [95 % CI, −1.5 to 3.3; P =0.47]; and in the Physical Component Summary score, 1.1 [95 %

References 1. Jensen JF et al.,(2015) Impact of follow-up consultations for ICU survivors on post-ICU syndrome: a systematic review and meta-analysis. Intensive Care Med 41: 763-775 2. Egerod I et al., (2011) Constructing the illness narrative: a grounded theory exploring patients´ and relatives´ use of intensive care diaries. Crit Care Med 39: 1922-1928 Grant acknowledgement The Danish Nursing Organization, The Novo Nordisk Foundation and Nordsjællands Hospital, University of Copenhagen, DK

A39 The value of progressively accrued information during initial postadmission hours in forecasting future cardiorespiratory instability M. Hravnak1, L. Chen2, A. Dubrawski2, G. Clermont3, M.R. Pinsky3 1 University of Pittsburgh, School of Nursing, Pittsburgh, United States; 2 Carnegie Mellon University, Robotics Institute; Auton Lab, Pittsburgh, United States; 3University of Pittsburgh, Critical Care Medicine, Pittsburgh, Pennsylvania, United States Correspondence: M. Hravnak – University of Pittsburgh, School of Nursing, Pittsburgh, United States Intensive Care Medicine Experimental 2016, 4(Suppl 1):A39 Introduction: Enabling clinicians to prospectively identify patients who will later become unstable would enable targeting resources to patients most in need as well as potential application of preventive care. Objectives: To determine the incremental contribution of information progressively available within the first 4 hours of (SDU) admission to improve models forecasting later development of cardiorespiratory instability (CRI), including a novel CRI relative risk score. Methods: Continuous noninvasive vital sign (VS) monitoring data (heart rate [HR], respiratory rate [RR; bioimpedance], oscillometric blood pressure [BP], peripheral oximetry [SpO2]) were collected from 1971 stepdown unit (SDU) patients, and CRI episodes defined as VS deviation beyond stability thresholds. Patients with any CRI (cases, n = 918) and those never displaying CRI (controls, n = 1053) were identified. We computed a minute-by-minute integrated CRI risk score based on the method described in [1], using features computed from VS data streams during trailing 15 minute rolling windows and a trained random forest machine learning model. We then computed for each patient a mean risk score aggregated from the risk scores during first 4 hours of SDU stay. Next we built a logistic regression model to forecast whether or not there will be a CRI event in the future. To mimic the temporal availability of data following patient admission, we first entered demographics available at patient admission (age, gender, Charlson Comorbidity Index score) into the model, and then the initial VS (5-minute average of continuous VS data accrued from minutes 10 to 15 after admission), and finally the relative risk score derived in the first 4 hours. We assessed the predictive contribution of information from these 3 progressively accrued categories (demographics, initial VS, 4-hr risk score) by the Area Under Receiver Operating Curve (AUC) in a 10-fold cross validation experiment setup. Results: The risk score derived from admission demographics alone yielded an AUC of 58 ± 0.002 % to forecast future CRI. Adding the initial VS improved the AUC to 64 ± 0.003 %, and with further adding the 4-hr risk score the AUC became 67 ± 0.002 %. Conclusions: A predictive model which incorporates patient data as it becomes available, including a risk score derived within the first

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

4 hours, progressively improves the models ability to forecast future CRI development. Such forecasting information could enable clinicians to identify those patients who will become unstable in future very soon after admission in order to triage patients needing closer surveillance and potentially apply preemptive interventions. References [1] Chen L et. Al. Modelling Risk of Cardio-Respiratory Instability as a Heterogeneous Process. AMIA Annu Symp Proc. 2015 Nov 5;2015:1841-50. Grant acknowledgement NIH NINR R01NR013912

A40 What are the factors that impact on physical activity and rehabilitation for survivors of critical illness: a systematic review of quantitative and qualitative studies S.M. Parry1, L.D. Knight2, B.C. Connolly3,4, C.E. Baldwin5,6, Z.A. Puthucheary7,8,9, L. Denehy1, N. Hart3,4,10, P.E. Morris11, J. Mortimore4, C.L. Granger1 1 The University of Melbourne, Department of Physiotherapy, Melbourne, Australia; 2Royal Melbourne Hospital, Department of Physiotherapy, Melbourne, Australia; 3Guy's and St Thomas' NHS Foundation Trust and Kings College, London, United Kingdom; 4Guy's & St Thomas' NHS Foundation Trust, Lane Fox Respiratory Unit, London, United Kingdom; 5 University of South Australia, Member of the International Centre for Allied Health Evidence (iCAHE) and the Sansom Institute, Adelaide, Australia; 6 Flinders Medical Centre, Department of Physiotherapy, Adelaide, Australia; 7 National University Hospital, Division of Respiratory and Critical Care, Singapore, Singapore; 8University College London Hospitals, Division of Critical Care, London, United Kingdom; 9University College London Hospitals, Institute of Sports and Exercise Health, London, United Kingdom; 10 Kings College London, Division of Asthma, Allergy and Lung Biology, London, United Kingdom; 11University of Kentucky, Division of Critical Care and Pulmonology, Kentucky, United States Correspondence: S.M. Parry – The University of Melbourne, Department of Physiotherapy, Melbourne, Australia Intensive Care Medicine Experimental 2016, 4(Suppl 1):A40 Introduction: Physical activity / rehabilitation forms a pivotal aspect of recovery after critical illness and studies have demonstrated it is safe, feasible and potentially efficacious at improving patient outcomes [1,2]. However, international data demonstrate low levels of mobilisation occur in the ICU[3,4]. A current gap exists between the perceived need and actual practice of implementing physical activity across the recovery continuum. Objectives: To identify, evaluate and synthesise studies examining the barriers and enablers for patients with critical illness to participate in physical activity from the perspective of healthcare providers, patients and caregivers. Methods: Systematic review of articles using electronic databases: MEDLINE, CINAHL, EMBASE, Scopus and Cochrane. Quantitative and qualitative studies which assessed the barriers, or enablers to physical activity for patients with critical illness were included. Registered on PROSPERO (number: CRD42016035454). Results: 79 studies were included. Studies included primarily ICU survivors (69 %, n = 54 studies), healthcare providers (29 %, n = 23 studies) with only one study specifically examining caregivers and patients. Barriers and enablers to physical activity were identified (5 major themes and 28 sub-themes). Patient-level barriers included physical capability (physiological stability, illness severity, sedation, weakness, delirium), psychological influences (fear/motivation) and perceived relevance. Healthcare provider barriers included lack of time/ knowledge and expertise, communication, and concern for line safety. Environmental barriers included lack of resources (staffing and equipment), lower prioritisation, and lack of an established rehabilitation pathway post ICU. Enablers included: presence of mobility teams/ protocols, designated discipline and overall leaders, teamwork and development of daily care plans. Conclusions: This systematic review has identified the volume of literature demonstrating that barriers and enablers to physical activity are multi-dimensional and span diverse factors. These factors need to be

Page 25 of 607

considered when developing rehabilitation interventions to facilitate cultural change in rehabilitation practices across the recovery continuum. References 1. Kayambu, G., R. Boots, and J. Paratz, Physical Therapy for the Critically Ill in the ICU: A Systematic Review and Meta-Analysis. Critical Care Medicine, 2013. 41(6): p. 1543-54. 2. Morris, P.E., et al. Early intensive care unit mobility therapy in the treatment of acute respiratory failure (Structured abstract). Critical Care Medicine, 2008. 36, 2238-2243. 3. Berney, S., et al., Intensive care unit mobility practices in Australia and New Zealand: a point prevalence study. Critical Care and Resuscitation, 2013. 15(4): p. 260-265. 4. Nydahl, P., et al., Early mobilisation of mechanically ventilated patients: a one day point prevalence study in Germany. Critical Care Medicine, 2014. 42(5): p. 1178-1186.

END-OF-LIFE-CARE: GET IT RIGHT THE FIRST TIME A41 Perceptions of end-of-life decision-making climate among healthcare providers working in european and us icus: differences between nurses and physicians H.I. Jensen1,2, R. Piers3, B. Van den Bulcke4, J. Malmgren5, V. Metaxa6, A.K. Reyners7, M. Darmon8, K. Rusinova9, D. Talmor10, A.-P. Meert11, L. Cancelliere12, L. Zubek13, P. Maia14, A. Michalsen15, J. Decruyenaere4, E. Kompanje16, S. Vanheule17, E. Azoulay18, S. Vansteelandt19, D. Benoit4 1 Lillebaelt Hospital, Department of Anaesthesiology and Intensive Care, Vejle and Kolding, Denmark; 2University of Southern Denmark, Institute of Regional Health Research, Odense, Denmark; 3Ghent University Hospital, Department of Geriatrics, Gent, Belgium; 4Ghent University Hospital, Department of Intensive Care, Gent, Belgium; 5Sahlgrenska University Hospital, Department of Anaesthesiology and Intensive Care, Gothenburg, Sweden; 6King's College Hospital, Critical Care and Major Trauma, London, United Kingdom; 7University of Groningen, University Medical Center Groningen, Department of Internal Medicine, Groningen, Netherlands; 8 Saint-Etienne University Hospital and Jacques Lisfranc Medical School, Saint-Etienne, France; 9General University Hospital, 1st Faculty of Medicine, Charles University in Prague, Department of Anaesthesia and Intensive Care Medicine, Prague, Czech Republic; 10Beth Israel Deaconess Medical Center and Harvard Medical School, Department of Anesthesia, Critical Care, and Pain Medicine, Boston, United States; 11Institut Jules Bordet, ULB, Service des soins Intensifs et Ergences Oncologiques, Bruxelles, Belgium; 12 Ospedale Maggiore della Carità Novara, Dipartimento Emergenza Urgenza Anestesia e Rianimazione, Novara, Italy; 13Semmelweis University, Budapest, Hungary; 14Hospital S.António, Intensive Care Department, Porto, Portugal; 15Tettnang Hospital, Department of Anesthesiology and Critical Care, Tettnang, Germany; 16Erasmus Hospital, Department of Intensive Care, Rotterdam, Netherlands; 17Ghent University, Department of Psychoanalysis and Clinical Consulting, Gent, Belgium; 18Hôpital Saint Louis, Service de Reanimation Médicale, Paris, France; 19Ghent University, Department of Applied Mathematics and Computer Science and Statistics, Gent, Belgium Intensive Care Medicine Experimental 2016, 4(Suppl 1):A41 Introduction: Literature depicts differences in perceptions of End-OfLife (EOL) decision-making (DM) between nurses and physicians. Objectives: To examine perceptions of nurses and physicians in regard to EOL DM in the ICU and to test the hypothesis that the worse the EOL DM climate, the greater the discordance between nurses' and physicians' rating of EOL DM. Methods: Perceptions of EOL-DM among health care providers of 68 adult ICUs in 13 European countries and the US were measured in AprilMay 2014, using a validated self-assessment questionnaire. The questionnaire existed of 35 questions and was based on the Appropricus questionnaire [1], the IPEQS instrument and the LDBQ questionnaire [2]. Results: A total of 2,275 nurses and 717 physicians participated. Response rates were 63.1 % and 62.9 %, respectively. Using factor analyses and cluster analysis, seven meaningful factors (physician leadership, interdisciplinary reflection, not avoiding EOL decisions, mutual respect

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

within the interdisciplinary team, involvement of nurses in EOL, active DM by physicians, and ethical awareness) yielded 4 climates: good, average with and average without involvement of nurses at EOL, and fourth a poor EOL-DM climate. When looking at overall perceptions, there were for all seven factors significant differences (p < 0.001) between nurses and physicians, with physicians consistently perceiving the EOL-DM climate as more positive as nurses. The largest differences were found in regard to physician leadership (median factor scores -0.09 for nurses and 0.61 for physicians, respectively), interdisciplinary reflection (medians -0.03 and 0.51) and not avoiding EOL decisions (medians -0.07 and 0.37). When looking at differences within the 4 types of EOL-DM, the same general pattern was found, and all differences were statistically significant (p < 0.001). The differences were largest in the two lowest types of EOL-DM: average without nurses and poor EOL-DM climate. Head nurses generally had a more positive perception of EOL-DM climate compared to other nurses, and senior physicians had a more positive perception than junior physicians. When comparing nurses (without including head nurses) and junior physicians, the physicians still had a significantly higher perception of EOL-DM, except for “Mutual respect within the interdisciplinary team” (p = 0.26). Physicians having a more positive perception of EOL-DM compared to nurses were found within all 13 participating countries and within the individual ICU's. Conclusions: The poorer the EOL DM climate, the more physicians were likely to overestimate the EOL DM climate compared to nurses. References [1] Piers, R et al. Inappropriate Care in European ICUs: Confronting Views From Nurses and Junior and Senior Physicians. Chest 2014;146(2):267-275. [2] Van den Bulcke B et al. The perceived quality of inter-professional ICU teamwork: A single centre intervention study. J Interprofessional Care. In press.

A42 Perceptions of end-of-life decision making climate among European and US ICU health care providers: development and validation of a self-assessment tool to differentiate end-of-life decision making climates B. Van den Bulcke1, R. Piers2, H.I. Jensen3,4, J. Malmgren5, V. Metaxa6, A.K. Reyners7, M. Darmon8, K. Rusinova9, D. Talmor10, A.-P. Meert11, L. Cancelliere12, L. Zubek13, P. Maia14, A. Michalsen15, J. Decruyenaere1, E. Kompanje16, S. Vanheule17, E. Azoulay18, S. Vansteelandt19, D. Benoit1 1 Ghent University Hospital, Department of Intensive Care, Gent, Belgium; 2 Ghent University Hospital, Department of Geriatrics, Gent, Belgium; 3 Lillebaelt Hospital, Department of Anaestesiology and Intensive Care, Veijle and Kolding, Denmark; 4University of Southern Denmark, Institute of Regional Health Research, Odense, Denmark; 5Sahlgrenska University Hospital, Department of Anaesthesiology and Intensive Care, Gothenburg, Sweden; 6King's College Hospital, Critical Care and Major Trauma, London, United Kingdom; 7University of Groningen, University Medical Center Groningen, Department of Internal Medicine, Groningen, Netherlands; 8 Saint-Etienne University Hospital and Jacques Lisfranc Medical School, Saint-Etienne, France; 9General University Hospital, 1st Faculty of Medicine, Charles University in Prague, Department of Anaesthesia and Intensive Care, Prague, Czech Republic; 10Beth Israel Deaconess Medical Center and Harvard Medical School, Department of Anesthesia, Critical Care, and Pain Medicine, Boston, United States; 11Institut Jules Bordet, ULB, Service des Soins Intensifs et Urgences Oncologiques, Bruxelles, Belgium; 12Ospedale maggiore della carità Novara, Dipartimento Emergenza Urgenza Anestesia e Rianimazione, Novara, Italy; 13Semmelweis University, Budapest, Hungary; 14 Hospital S. António, Department of Intensive Care, Porto, Portugal; 15 Tettnang Hospital, Department of Anesthesiology and Critical Care, Tettnang, Germany; 16Erasmus Hospital, Department of Intensive Care, Rotterdam, Netherlands; 17Ghent University, Department of Psychoanalysis and Clinical Consulting, Gent, Belgium; 18Hopital Saint Louis, Service de Reanimation Médicale, Paris, France; 19Ghent University, Department of Applied Mathematics and Computer Science and Statistics, Gent, Belgium Correspondence: B. Van den Bulcke – Ghent University Hospital, Department of Intensive Care, Gent, Belgium Intensive Care Medicine Experimental 2016, 4(Suppl 1):A42

Page 26 of 607

Introduction: Literature depicts large differences in end-of-life (EOL) decision making (DM) between countries and ICUs. Objectives: To better conceptualize EOL-DM and to develop and validate a tool to assess EOL-DM climates. Methods: Perceptions of EOL-DM among health care providers (HCPs) of 68 adult ICUs in 13 European countries and the US were measured, using a self-assessment questionnaire in April-May 2014. The EOL-DM climate questionnaire, existing of 35 items, based on the Appropricus questionnaire (11 items concerning EOL care practices) and extended with 24 validated items concerning key conditions to provide good EOL care: interdisciplinary collaboration and communication (Interprofessional Practice and Education Quality Scales (IPEQS, 11 items) and leadership skills (LBDQ,13 items) (1, 2,3). Exploratory and confirmatory factor analysis followed by cluster analysis was used to determine EOL-DM climates. Results: Of the 3610 nurses and 1137 doctors providing ICU bedside care, 63,1 % and 62,9 % participated respectively. Seven meaningful factors were identified (physician leadership, interdisciplinary reflection, not avoiding EOL decisions, mutual respect, involvement of nurses in EOL, active DM by physicians, ethical awareness); which yielded 4 EOLDM climates: good (17.6 % of the ICUs), average with (+) (17.6 %) and without involvement of nurses at EOL (-) (32.3 %), and poor (54.3 %) (factor analysis p < 0.001). According to HCPs working in a good climate leadership is active and facilitates interdisciplinary reflection and decision-making overall. Within the 'average + ' climate, HCPs perceive their leaders as empowering nurses to share interdisciplinary decisionmaking at EOL mainly. Stimulating more open ethical awareness overall and being less hesitant in taking timely EOL decisions could be a main goal for the leader in this climate. HCPs working in an 'average -' climate do perceive their leaders as not hesitant to take important decisions (at EOL) however as insufficiently empowering the involvement of nurses in the DM process. HCPs working in a poor climate perceive a need for improvement on all previous factors. This climate was further characterized by poor communication, distrust and low respect. Conclusions: We identified seven key dimensions, from which 4 meaningful types of ICU EOL-DM climates could be discerned. The climates differ mainly in 2 key dimensions: the way of nurse involvement in EOL and in physicians active DM. Our research offers opportunities to develop tailored ICU team interventions. References [1] Piers, R et al. (2014) Inappropriate Care in European ICUs: Confronting Views From Nurses and Junior and Senior Physicians,Chest,146,(2),267-275. [2] Van den Bulcke B et al. The perceived quality of inter-professional ICU teamwork: A single centre intervention study, J Interprofessional Care. In press. [3] Stogdill, RM &Coons, AE (Eds). (1957). Leader behavior description questionnaire. Oxford, England: Ohio State University.

A43 Nurses' perceptions of aids and obstacles to the provision of optimal end of life care in ICU C. Ryan, D. Dawson, J. Ball, K. Noone, B. Aisling, S. Prudden St Georges NHS Foundation Trust, General Intensive Care, London, United Kingdom Correspondence: C. Ryan – St Georges NHS Foundation Trust, General Intensive Care, London, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A43 Introduction: There is increasing recognition of the need for comprehensive expertise in the management of EOLC in the ICU. However there appear to be many controversies between professionals regarding optimal provision. Objectives: As a component of a local, on-going quality improvement process in this area, we performed a survey of nurses' perceptions of aids and obstacles to the optimal provision of EoLC. Methods: We modified a previously validated survey tool1 and anglicised the language. Nurses were asked to rate both the size and frequency of 20 possible obstacles and 14 possible aids to providing EoLC using a 6 point Likert scale (0-5). The survey was distributed to 120 nursing staff on one adult general critical care unit in March 2015.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Page 27 of 607

Confidentiality was assured. For each obstacle and aid, the median and interquartile ranges were determined for the size and frequency. To determine the effect size of each obstacle and aid, the median of the size was multiplied by the median of the frequency. These were then themed Results: Sixty surveys were returned representing a 50 % return rate. Conclusions: This study has highlighted the need to proactively identify a family liaison to cascade information to friends and relatives to allow nurses to concentrate on care delivery. Despite having a poor unit design and lack of privacy, nurses feel they can provide a dignified death and feel that multidisciplinary agreement is an important part of this process. References 1. Kirchoff and Beckstrand, 2000 Critical care nurses´ perceptions of obstacles and helpful behaviour's in providing end-of-life care to dying patients. Am J Crit Care. 2000 Mar;9(2):96-105. 2. Festic E,Wilson ME, Gajic O, Divertie GD, Rabatin JT, 2012 Perspectives of physicians and nurses regarding end-of-life care in the intensive care unit. J Intensive Care Med. 2012 Feb;27(1):45-54. doi: 10.1177/0885066610393465. Epub 2011 Jan 21

Table 23 (abstract A43). Effect size and themes obstacles LARGE OBSTACLES occurring frequently

EFFECT SIZE

THEME

Family requesting constant updates

16

Workload

Poor unit design restricting privacy

16

Environment

Family unable to accept patient dying

12

Family Coping

Nurse caring for family and patient simultaneously

12

Workload

Not knowing patient's wishes for EOLC

12

Staff Coping

SMALL OBSTACLES occurring infrequently

EFFECT SIZE

THEME

Lack of chaplaincy services

2

Logistics

Lack of support from colleagues

3.75

Staff coping

Restrictive visiting hours

4

Logisrics

Table 24 (abstract A43). Effect size and themes of aids LARGE AIDS occurring frequently

EFFECT SIZE

THEME

Multidisciplinary agreement

15

Communication

Provision of a peaceful and dignified scene

15

Environment

Allowing adequate time after death

15

Logistics

LARGE AIDS that occur infrequently

EFFECT SIZE

THEME

Having a designated family liaison

10

Workload

Having unit design that allows privacy

10

Environment

SMALL AIDS

EFFECT SIZE

THEME

Family helping with care

6

Case Specific

Having support outside work

3

Staff coping

A44 The impact of personality and religiousness of icu personnel on end-of-life decisions A. Ntantana1, D. Matamis1, S. Savvidou1, M. Giannakou2, M. Gouva3, G. Nakos4, V. Koulouras4 1 Papageorgiou General Hospital, ICU, Thessaloniki, Greece; 2AHEPA University Hospital, ICU, Thessaloniki, Greece; 3Technological Educational Institutes of Ipeirus, Ioannina, Greece; 4University Hospital of Ioannina, ICU, Ioannina, Greece Correspondence: A. Ntantana – Papageorgiou General Hospital, ICU, Thessaloniki, Greece Intensive Care Medicine Experimental 2016, 4(Suppl 1):A44 Introduction: It has been recognized that healthcare workers involved in “End of life” (EoL) decisions may be influenced by cultural, geographical, religious and personal characteristics1.

Objectives: To investigate possible associations of ICU personnel's aspects of personality and religiousness with attitude towards EoL. Methods: A cross-sectional, multicenter study was conducted in a national level in Greece during June to December 2015. ICU physicians (n = 149) and ICU nurses (n = 320) participated by answering three questionnaires: the main survey questionnaire investigating EoL attitudes2, the Eysenck Personality Questionnaire (EPQ)3, and the Spiritual and Religious Attitudes Questionnaire (SpREUK)4. Different scores of EPQ-Neuroticism, EPQ-Psychoticism and SpREUK-Trust in a higher guidance were recorded in each participant. Results: A high participation rate was recorded (65.7 %). Eighty-four participants preferred to characterize EoL decisions as “passive euthanasia” over “refusal of futile care” (7.6 % of ICU physicians vs.23.3 % of ICU nurses, p < 0.001), 289 participants declared that removing artificial ventilation represented a different approach from other EoL decisions because of the analogy breath = life (51.4 % of physicians vs.75.8 % of nurses, p < 0.001), and 71 participants acknowledged fear of litigation as the major reason for not informing the family about EoL decisions (17.4 % of doctors vs.14.9 % of nurses, p = 0.004). Statistical analysis with multivariate logistic regression identified that attitude towards passive euthanasia could independently be predicted by high neuroticism scores (Odds ratio 1.6, 95%CI 1.1-2.7, p = 0.048), attitude towards withdrawal of artificial ventilation by trust in a higher guidance (Odds ratio 1.7, 95%CI 1.1-2.5, p = 0.010), and fear of litigation by high psychoticism scores (Odds ratio 2.4, 95%CI 1.3-4.8, p = 0.009). Conclusions: The results of this study indicate that specific attitudes towards EoL decisions may be influenced by aspects of personality and religiousness of ICU personnel. References 1. Sprung Ch, Maia P, Bulow H-H, et al. The importance of religious affiliation and culture on end of life decisions in European intensive care units. Intensive Care Med 2007;33:1732-39. 2. Ferrand E, Lemaire F, Regnier B, et al. Discrepancies between perceptions by physicians and nursing staff of intensive care unit end-of-life decisions. Am J Respir Crit Care Med 2003;167:1310-5. 3. Eysenck H.J Eysenck S.B.G. manual of the Eysenck personality questionnaire (EPQ) Hodder and Stoyghton educational London, UK.: 1975. 4. Bussing A, Ostermann T, Matthiessen PF, et al. Role of religion and spirituality in medical patients: confirmatory results with the SpREUK questionnaire. Health Qual Life Outcomes 2005;3:1-10.

A45 End of life care in the intensive care unit - an audit cycle and quality improvement project J. Aron, G. Lumley, D. Milliken, K. Dhadwal Royal Free Hospital, Intensive Care Unit, London, United Kingdom Correspondence: J. Aron – Royal Free Hospital, Intensive Care Unit, London, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A45 Introduction: The decision to limit or withdraw active treatment in patients admitted to the intensive care unit (ICU) is multi-factorial (1). Prolonged periods of organ support in elderly, frail patients may not be beneficial and has resource implications. International guidelines suggest that goals of treatment should be set within 72 hours of admission to the ICU (2). Objectives: We aim to identify the frequency of end-of-life decisionmaking and the effectiveness in communicating these decisions to the ICU team. Methods: A prospective audit cycle was completed involving unplanned ICU admissions during a one-week period. Demographic information, APACHE II score, clinical frailty score (CFS), limitations of treatment (LOT) and awareness of these decisions were all documented. A re-audit using the same methodology was conducted after CFS and communication proformas were introduced for unplanned admissions. Results: 61 patients were admitted as unplanned admissions during the first audit period. Patients were elderly, had a high predicted mortality and a significant proportion were clinically frail (CFS ≥ 5).

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Mean length of stay (LOS) was prolonged and level of organ support was high. These results are summarised in Table 25. LOT was instigated in 9 (11.5 %) patients. Age and APACHE II scores were similar in patients with LOT and those with none. Frail patients had a higher LOS and received more organ support compared to non-frail patients. The proportion of patients with LOT remained low despite an increasing LOS. Communication between team members was sub-standard with 33 % of bedside nurses and 44 % of residents being unaware of these decisions. The re-audit after intervention demonstrated an increase in patients receiving LOT, with a higher proportion of frail patients being represented. These patients also received less organ support and LOT was instigated earlier in their stay. Awareness regarding these decisions amongst all staff improved to 100 %. Feedback regarding the communication tools was excellent. Conclusions: End of life care was performed infrequently in unplanned admissions to the ICU. When instigated, knowledge of these plans was substandard. Introducing CFS on admission to ICU identified patients who were unlikely to benefit from prolonged organ support, thus increasing LOT decisions. The communication sheets improved team communication. This has now become standard practice in two London ICUs. References 1. Le Maguet et al. Prevalence and impact of frailty on mortality in elderly ICU patients: a prospective, multicenter, observational study. Intensive Care Med. 2014 May;40(5):674-82. 2. Dellinger RP, et al for the Surviving Sepsis Campaign Guidelines Committee including the Pediatric Subgroup. Surviving sepsis campaign: international guidelines for management of severe sepsis and septic shock: 2012. Crit Care Med. 2013 Feb;41(2):580-637

Table 25 (abstract A45). Results overview from the audit cycle No Treatment Limits (first audit)

Treatment Limits (first audit)

Treatment limits after interventions

Changes observed after interventions

Mean Age/range (years)

67.5 (44- 89)

73.33 (47 - 91)

82.5 (62-88)

More elderly population

APACHE II (mean)

22.3

21.8

22.3

No change in disease severity

CFS < 5, n (%)

29 (58.2%)

2 (22%)

2 (12.5%)

CFS > 5, n (%)

23 (41.8%)

7 (78%)

14 (87.5%)

Higher proportion of frail patients with LOT

LOS > 72 hours n/total, (%)

35/39 (90.0%)

4/39 (10.3%)

8/36 (22.2%)

Higher proportion at 72 hours with LOT

LOS > 30 days, n/total (%)

11/14 (78.6%)

3/14 (21.4%)

7/11 (63.6%)

Higher proportion at 30 days with LOT

Awareness of decisions - Nurses, n (%)

49 (89.1%)

6 (66.7%)

16 (100%)

Improved awareness

Awareness of decisions Resident, n (%)

51 (92.7%)

5 (55.6%)

16 (100%)

Improved awareness

Page 28 of 607

that may occur.[1,2] The Global Tracheostomy (Quality Improvement) Collaborative (GTC) offers resources for participating hospital sites. Whilst individual elements of the GTC programme have been shown to be of benefit in individual sites, implementation of the package of resources and the ability of the GTC database to track and benchmark healthcare improvements had not been evaluated. Objectives: The primary aim of the project was to implement the GTC into four diverse NHS hospitals and to evaluate impact on the safety and quality of care delivered. We hypothesised that systematic healthcare improvements would reduce the severity of harm resulting from tracheostomy-related safety incidents and improve surrogate markers of the quality of patient-centred outcomes such as time to first vocalisation. Methods: As part of a Health Foundation funded quality improvement project, we introduced the GTC into four NHS sites in Manchester, England between August 2014 and August 2015. Interventions included multidisciplinary staff educational courses, webinars, standardisation of protocols and development of multidisciplinary tracheostomy teams. Local projects were overseen by local staff and patient champions. Data were collected using the bespoke GTC database and monthly trends in surrogate and patient safety incidents were analysed. Monthly incident rates were analysed using Chi Square test for linear trend. Results: Over the 12 month data collection period 296 tracheostomy patient admissions were tracked across the four sites with similar demographics to previously reported national data.[1] A total of 124 adverse events were identified affecting 29.8 % patients. Analysis of reported incidents over the duration of the project showed a significant reduction in the severity of harm by month (Chi Square p < 0.01, Fig. 22). There was also a significant trend towards lower harm categories for incidents over the duration of the project (Chi Square test for linear trend, r = -0.21, p < 0.01). Monthly analysis of the dataset for percutaneous tracheostomies showed non-significant trend toward earlier speaking valve use and vocalisation (median slope = -0.17, -0.83 to 0.4) associated with improvements in reported patient satisfaction scores. Conclusions: Our study has demonstrated that meaningful improvements in the safety and quality of care for patients with tracheostomies are possible using improvement methodology advocated by the GTC in diverse NHS Trusts. References 1. ‘On the right trach?’ NCEPOD 2014. www.ncepod.org 2. McGrath BA,BJA 2015;115(2):155-8. Grant acknowledgement This work was supported by the Health Foundation, an independent charity committed to bringing about better health and health care for people in the UK.

PATIENT SAFETY IN THE ICU A46 Improvements in the safety and quality of care in four UK NHS hospitals participating in the global tracheostomy collaborative B.A. McGrath1,2, S.J. Lynch1, B. Bovento1, G. Sharpe1, E. Grainger1, S. Pieri-Davies1, S. Wallace1 1 University Hospital South Manchester, Manchester, United Kingdom, 2 Manchester Academic Health Sciences Centre, University of Manchester, Manchester, United Kingdom Correspondence: B.A. McGrath – University Hospital South Manchester, Manchester, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A46 Introduction: Tracheostomies are used in the management of around 10 % of ICU admissions. Increasingly complex patients require truly multidisciplinary care to reduce the well described preventable harm

Fig. 22 (abstract A46). Trends in reported patient safety incidents

A47 The impact of multidisciplinary tracheostomy safety teams on hospital length of stay for patients with tracheostomy B. McGrath1,2, S.J. Lynch1, B. Bovento1, E. Grainger1, S. Pieri-Davies1, G. Sharpe1, S. Wallace1 1 University Hospital South Manchester, Manchester, United Kingdom; 2 Manchester Academic Health Sciences Centre, University of Manchester, Manchester, United Kingdom Correspondence: B. McGrath – University Hospital South Manchester, Manchester, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A47

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Introduction: Around 12-14,000 tracheostomies are performed by intensivists in critically ill patients in England and Wales annually. Patients are increasingly complex and high quality care requires the input of a wide range of multidisciplinary healthcare professionals, including medical, nursing, speech and language pathology and specialist physiotherapy staff. National reports have identified disjointed and uncoordinated care that leads to delays in progressing care.[1,2,3] The Global Tracheotomy (Quality Improvement) Collaborative (GTC) brings together resources from international exemplar centres. Multidisciplinary tracheostomy teams (MDTT) are a key intervention. We introduced GTC MDTT and resources into four diverse NHS sites and performed a detailed evaluation of their impact. Objectives: We hypothesised that systematic healthcare improvements that raised the quality and safety of tracheostomy services would lead to more efficient care, measured by earlier tracheostomy decannulation times and reduced hospital lengths of stay. Methods: We introduced MDTT into four diverse NHS sites in Manchester, England as part of a Health Foundation funded project between August 2014 and August 2015, measuring outcomes using the GTC database. Each site set up a MDTT, supported by the resources of the GTC (via www.globaltrach.org). Webinars offered opportunities for question and answer sessions around how obstacles to changes in care delivery had been overcome. MDTTs reviewed tracheostomy patients weekly and provided a point of contact and education for bedside ICU and ward staff. The largest site (UHSM) kept detailed MDTT records over 40 weeks, with all four sites recording individual patient metrics via the GTC database. Length of stay (LoS) data were plotted monthly with box-and-whisker plots. All data points were retained for non-parametric linear regression analysis. Results: UHSM MDTT undertook 155 reviews, making 184 interventions. There was a significant month-by-month trend towards reducing LoS across all sites for all 296 patients, with median hospital LoS reduced by 6 days over the 12 month project (95%CI 9.96-3.96). For the 214 newly inserted tracheostomies, there was a non-significant trend towards reduced tracheostomy time (median slope -0.05, -0.17 to 0.25). ICU LoS reduced significantly over the duration of the project with a median slope of -0.11 (-0.25 to 0), equating to a reduction of 1.3 days per patient. Conclusions: Introducing MDTT can coordinate and improve the quality and safety of care provided to ICU and ward tracheostomy patients. Improvements are reflected in reductions in ICU and hospital LoS. References 1. McGrath BA, Otolaryngology-Head&Neck Surgery 2015;153(2):167-9 2. McGrath B, Clin Otolaryngol. 2013;38(6):541-545 3. Hettige R, BJIC 2013(Autumn):89-92 Grant acknowledgement Funded by the Health Foundation, an independent charity committed to bringing about better health and health care for people in the UK.

A49 The relationship between nurse staffing levels and mortality in the pediatric intensive care unit M. Jung, J. Cho, H. Park, G. Suh Samsung Medical Center, Sungkyunkwan University School of Medicine, Department of Critical Care Medicine, Seoul, Korea, Republic of Correspondence: M. Jung – Samsung Medical Center, Sungkyunkwan University School of Medicine, Department of Critical Care Medicine, Seoul, Korea, Republic of Intensive Care Medicine Experimental 2016, 4(Suppl 1):A49 Introduction: Previous studies showed that low nurse staffing levels were associated with poor outcomes and increased occurrence of adverse events. However, national data are scarce regarding the relationship between the nurse staffing ratio and pediatric patient outcomes. Objectives: To investigate the association between nurse staffing levels and the outcome of pediatric patients in Korean intensive care units (ICUs). Methods: This retrospective cohort study used data from National Health Insurance claims. The included patients were under 18 years

Page 29 of 607

old and admitted to ICUs (but not neonatal ICUs) between August 2009 and September 2014. The ICU nurse staffing was graded into nine levels according to the total bed to nurse ratio for each hospital (National Health Insurance policy). The lowest and highest bed-to-nurse ratios were 0.5:1 and 2:1, respectively. We inspected the differences between overall patient mortality and the nursing staff grades at discharge. We also analyzed the mortality of patients who underwent mechanical ventilator care for more than 3 hours according to the nursing staff grades. Results: During the study period, 39,917 medical and surgical pediatric patients were admitted to the ICUs. Mechanical ventilation was administered to 45.9 % of these patients. The median (interquartile range) ICU length of stay was 2 (1-6) days. The overall hospital mortality rate was 5.5 %, and the mortality rate of mechanically ventilated patients was 10.8 %. About half of the patients (18,053; 45.2 %) received the highest grade ICU treatment. The overall mortality rate did not increase with a higher bed-to-nurse ratio (i.e., lower nurse staffing level). However, a trend toward increased mortality was observed in mechanically ventilated patients as the bed-to-nurse ratio increased. The mortality of mechanically ventilated patients was 7.6 % for a bed-to-nurse ratio < 0.5:1, 16.7 % for a 0.63-0.77:1 ratio, 25.2 % for a 0.88-1:1 ratio, and 38.1 % for a > 2:1 ratio. Conclusions: The nursing staff to bed ratio may not be associated with the overall mortality of pediatric patients at discharge. However, a low nursing staff level was associated with a higher mortality in pediatric patients who required mechanical ventilators.

A50 Exposure keratopathy in critically ill adults: incidence, risk factors and impact of protocolised care O. Kousha, J. Paddle Royal Cornwall Hospital, Intensive Care Unit, Truro, United Kingdom J. Paddle – Royal Cornwall Hospital, Intensive Care Unit, Truro, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A50 Introduction: Exposure keratopathy (EK) is a clinical syndrome characterised by incomplete eyelid closure and corneal wetting leading to corneal damage of variable severity and extent. It has an incidence in critically ill patients of between 10 and 60 % (1). Objectives: This study aimed to determine the rate of EK in patients admitted to critical care (ICU) and identify risk factors for developing EK. Using the identified risk factors and experience from the first part of the study, we developed an eye-care protocol. Finally, we studied the effectiveness of the protocol to prevent EK. Methods: We undertook a two-phase prospective cohort singlecentre study between November 2014 and August 2015 in a general adult ICU in the United Kingdom. The first phase of the study was observational. In the second phase of the study an eye-care protocol was introduced. Ethical approval was waived by the Trust ethics committee. Inclusion criteria were all patients admitted to ICU. Exclusion criteria were age < 16, known eye disease, patient agitation, and refusal to participate. The patient remained in the study until ICU discharge or withdrawal of active therapy. An eye-care protocol was developed and incorporated into the electronic patient record.All data were collected by a single investigator. Data were APACHE II score, daily SOFA, mechanical ventilation, Richmond Agitation-Sedation Scale, and level of eye care. Ophthalmic assessment was examination of the external eye, eyelids, and ocular surface using a portable slit lamp preand post-fluorescein dye instillation. Statistical tests were two-tailed (α = 0.05). Student t-test was used for continuous data and χ2 or Fisher's exact test for binary data. Logistic regression was used to analyse the relationship between EK and independent variables. Relative risk (RR) and Odds ratio (OR) were calculated with 95 % confidence intervals. Results: We studied 371 patients. The overall rate of EK was 21 %. Among mechanically ventilated patients the rate was 54.3 % compared to 5.1 % in patients receiving non-invasive or no ventilatory support: RR = 10.6 (5.5-20.7), p < 0.001. OR for development of EK was 6.8 (3.2-8.0), p = 0.028 with mechanical ventilation and 32.5 (15.3-45.1), p < 0.001 with lagophthalmos. SOFA was associated with EK with OR

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

1.3 (1.1-1.5), p = 0.016. We found no independent effect of age, sedation or APACHE II score. Following the introduction of the protocol in the second phase of the study, the rate of EK reduced to 2.6 % (3 cases): RR = 7.8 (3.5-0.5), p < 0.001. The association with mechanical ventilation and lagophthalmos remained. Compliance with the protocol was 97 %. Conclusions: EK is common in critically ill patients, and is associated with mechanical ventilation and lagophthalmos. A simple protocol substantially reduced the incidence of EK and was easily achieved in clinical practice. Reference 1. Kuruvilla S, Peter J, David S et al. J Crit Care. 2015:30;400-404 Grant acknowledgement None received

Poster Corner Sessions: NUTRITION AND HORMONAL DISTURBANCES A51 An attenuated rate of leg muscle protein depletion over time is seen in ICU longstayers L. Gamrin Gripenberg, M. Sundström Rehal, J. Wernerman, O. Rooyackers Karolinska University Hospital Huddinge, Stockholm, Sweden Correspondence: L. Gamrin Gripenberg – Karolinska University Hospital Huddinge, Stockholm, Sweden Intensive Care Medicine Experimental 2016, 4(Suppl 1):A51 Introduction: The loss of muscle mass in critical illness is caused by a mismatch between synthesis and degradation, where synthesis is near-normal, while degradation is enhanced, resulting in a negative protein balance. A muscle mass loss of 10 % per week is reported. Studies supporting this mechanism are almost exclusively from patients during the initial two weeks of critical illness. Objectives: To investigate if this mechanism is valid also in a longer time perspective we studied critical ill patients during days 10-40 of ICU stay. Methods: Critically ill patients on mechanical ventilation (n = 20) were included, multiple times if possible. In total 30 measurements of muscle protein turnover were performed employing a 3-pool model technique with a constant infusion of d5-phenylalanine, leg blood flow measurements, and biopsies to determine tissue enrichments. Results: Net protein balance of leg mixed muscle showed a pattern of becoming less negative over time with regression analyses. This pattern was totally attributable to an increase in protein synthesis rate, while protein breakdown rate, although higher than in healthy subjects, was completely unaltered over time. Net protein balance during days 10-20 (-21 ± 21 nmol phenylalanine/min/100 g muscle) was lower (p = 0.002; T-test) than during days 30-40 (1 ± 11 nmol phenylalanine/min/100 g muscle). Conclusions: The temporal pattern of protein turnover in leg muscle showed a pattern of diminished protein losses over time in surviving patients staying in the ICU. How this pattern relates to nutrition, mobilization, and pharmacology remains to be established. A52 Pharmacokinetics of four high-dose regimes of intravenous vitamin C in critically ill patients H.-J. de Grooth1, W.-P. Choo2, A.M. Spoelstra - de Man1, E.L. Swart2, H.M. Oudemans-van Straaten1 1 VU University Medical Center, Department of Intensive Care, Amsterdam, Netherlands; 2VU University Medical Center, Department of Clinical Pharmacology and Pharmacy, Amsterdam, Netherlands Correspondence: H.-J. de Grooth – VU University Medical Center, Department of Intensive Care, Amsterdam, Netherlands Intensive Care Medicine Experimental 2016, 4(Suppl 1):A52

Page 30 of 607

Introduction: Critically ill patients exhibit a high degree of vitamin C deficiency at ICU admission and plasma concentrations decrease even more during the following days [1, 2]. High intravenous (iv) doses of vitamin C are required to increase plasma concentrations to normal and supra-normal ranges [2], but the optimal dosage regime in this population remains unclear. Prolonged vitamin C administration may be associated with oxalate kidney stone formation, but the effect of short-term high-dose supplementation on urinary oxalate excretion is unknown. Study design and objectives: We conducted a prospective randomized controlled trial to determine the pharmacokinetics of four high dose regimes of iv Vitamin C in critically ill patients and to measure oxalate excretion. Patients:. Adult patients admitted to the ICU with sepsis or SIRS, with a non-neurological sequential organ failure assessment (SOFA) score >6 and an expected length of ICU stay >96 hours. Intervention: Patients were randomized to either 1 g or 5 g vitamin C twice daily as a 30-min bolus infusion, or to 2 g or 10 g daily as a continuous infusion. Vitamin C administration was continued for 48 hours, so that all patients received a total dose of either 4 g or 20 g vitamin C. Measurements: Concentrations of vitamin C were determined in plasma at t = 0, 1, 2, 4, 8, 12, 24, 36, 48, 72 and 96 hours. Urine vitamin C and oxalate concentrations were determined during the first and last 12 hours of vitamin C administration. NONMEM was used for the pharmacokinetic analysis. Results: Fourteen patients were included: 5 patients received 1 g vitamin C and 5 patients 5 g vitamin C iv twice daily for two days. Two patients received 2 g/day vitamin C and 2 patients received 10 g/day by continuous infusion for two days (Table 26). Four patients (28 %) were vitamin C deficient on admission (15 μg/dl were followed and cortisol testing was repeated in patients with ongoing vasopressor dependency with no other underlying cause. Results: A total of 37 septic shock patients were screened on ICU admission. Of these, 19 patients had cortisol levels >15 μg/dl and underwent cortisol retesting for ongoing vasopressor need during the same ICU admission. Eleven (%58) had serum cortisol levels < 15 μg/dl. Mean (±SD) age was 70 ± 13.5 years. APACHEII and Sequential Organ Failure Assessment scores on ICU admission were 25.3 ± 6.5 and 10.4 ± 5.2, respectively. Age, gender, APACHEII and SOFA scores, serum albumin, protein, CRP and procalcitonin levels were similar between the groups.

References 1. Jorge A, Guzman MD, et al.Adrenal exhaustion in septic patients with vasopressor dependency. J Cri Care 2007; 22, 319-323. 2. Wu JY, Hsu SC, et al.Adrenal insufficiency in prolonged critical illness. Crit Care 2008;12:R65. Grant acknowledgement None.

A54 Use of nurse-driven feeding protocol improves enteral caloric intake: observational single centre before-and-after study M. Padar1, G. Uusvel1, L. Starkopf2, J. Starkopf1,3, A. Reintam Blaser3,4 1 Tartu University Hospital, Department of Anaesthesiology and Intensive Care, Tartu, Estonia; 2University of Copenhagen, Faculty of Health and Medical Sciences, Section of Biostatistics, Institute of Public Health, Copenhagen, Denmark; 3University of Tartu, Department of Anaesthesiology and Intensive Care, Tartu, Estonia; 4Lucerne Cantonal Hospital, Department of Anaesthesiology, Intensive Care, Emergency and Pain Medicine, Lucerne, Switzerland Correspondence: M. Padar – Tartu University Hospital, Department of Anaesthesiology and Intensive Care, Tartu, Estonia Intensive Care Medicine Experimental 2016, 4(Suppl 1):A54 Introduction: Early and sufficient enteral intake is associated with reduced ICU morbidity and mortality. Limited data suggests that use of nurse-driven feeding protocol with defined targets may facilitate nutrition and improve the outcomes. Objective: To investigate whether implementation of nurse-driven feeding protocol results in increased enteral caloric intake in critically ill patients during their first week in the ICU. Methods: We performed an uncontrolled before-and-after study. Data of consecutive adult patients, readmissions excluded, treated for at least 7 days in General ICU of Tartu University Hospital, were extracted from existing database. In 2013, nurse-driven feeding protocol was instituted in the department [1]. To analyze outcomes of this implementation, patients treated from 2011 to 2012 were included in Before, and patients from 2014 to 2015 in After group. Results: 231 patients were included in Before and 225 in After group, respectively. The groups are comparable regarding demographics, case-mix and severity of illness. Instalment of feeding protocol resulted in significantly higher cumulative amount of enterally provided calories by day 7 [3165 (1165-5215) kcal in After vs 2360 (450-5075) kcal in Before group, median (IQR), p = 0.043], while less calories were given parenterally [2600 (712-4287) vs 3900 (1725-6645) kcal, p < 0.001]. Cumulative proportion of patients who did not receive any enteral feed was significantly smaller in After group (Fig. 24). Percentage of enterally received calories from caloric needs was significantly higher in After group (Fig. 25). Prevalence of GI symptoms and intra-abdominal hypertension was not different between the groups. ICU length of stay was significantly shorter in After group (11 vs 13 days, respectively, p < 0.001), but no difference was noted in duration of mechanical ventilation and ICU mortality. Conclusions: Use of nurse-driven feeding protocol in ICU patients is associated with improved enteral nutrition without an increase of GI complications.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

References 1. Kuslapuu M, Jõgela K, Starkopf J, Reintam Blaser A. The reasons for insufficient enteral feeding in an intensive care unit: A prospective observational study. Intensive Crit Care Nurs. 2015 Oct;31(5):309-14 Grant acknowledgement: Ministry of Education and Research of Estonia (IUT34-24)

Fig. 24 (abstract A54). Percentage of patients not fed enterally

Page 32 of 607

on protocol driven continuous enteral nutrition, targeted to achieve 25 Kcal/kg/day of energy and 1 gm/kg/day of protein. Data was collected on demography, admission severity of illness, hemodynamics, subjective global nutritional assessment (SGA), and lead time to initiation and achievement of full feeds, adequacy of energy and proteins supplied reasons for feeds interruptions and complications of enteral nutrition. All data on nutrition was collected till the patient stayed on ventilator. Primary outcome was time to achieve full feeds; secondary endpoints were time to initiate feeds and reason for interrupting feeds. Results: 374 patients fit into the inclusion criteria. Majority were males (64 %), mean age of patients was 55.1(±17.8) years and their BMI was 24(±3.5). APACHE II 22.3 (±5.9) and SOFA 7.2(±2.3) scores were high. 35.4 % of patients were malnourished on nutritional screening(SGA). Respiratory failure (59 %) was the most common reason for intubation followed by neurological deterioration and majority of patients were medical patients Two patients(2/374) had nasogastric tube malpositioning on X-ray, which required re-insertion. 86.8 % received enteral feeds, 2.4 % received parenteral nutrition and 10.4 % received no feeds Time to initiate enteral feeds was 12 hours (median) (IQR)(7.5-24) hours and it ranged from 1-130 hours 66.86 % of patients received 100 % of target calories in 39.2(±19.4) hours and 33.1 % of patients didn´t achieve their target calories. 372 interruptions of feeds occurred in 335 patients enterally fed; most common reasons for interruption was weaning (43.8 %) followed by airway related procedures (26.6 %) and these patients were restarted on feeds after 9.8(±5.1) hours following interruption of feeds. Patients received 70 % (mean-1069 Kcal) of their prescribed calories during their ventilator days. Mortality rate was 42.2 %, ICU ALOS and ventilator days were 7.7(±5.3) and 5.9(±4.2) days respectively. Conclusions: Protocol driven enteral nutrition allows early feeding, with minimum interruptions of feeds in mechanically ventilated patients and their is still room for improvement. References Keng F Yip, Vineya Rai, Kang K Wong.Evaluation of delivery of enteral nutrition in mechanically ventilated Malaysian ICU patients. BMC Anesthesiology 2014, 14:127.

Fig. 25 (abstract A54). Daily proportions of enteral calories

A55 Nutritional support in mechanically ventilated patients: are we doing enough? M.S. Kalaiselvan1, A.S. Arunkumar1, M.K. Renuka2, R.L. Shivkumar1 1 Sri Ramachandra University, Department of Critical Care Medicine, Chennai, India; 2Sri Ramachandra University, Department of Anesthesiology, Chennai, India Correspondence: M.S. Kalaiselvan – Sri Ramachandra University, Department of Critical Care Medicine, Chennai, India Intensive Care Medicine Experimental 2016, 4(Suppl 1):A55 Introduction: Enteral nutritional is an essential component of critical care. Adequate nutrition is essential for recovery from critical illness; malnutrition has been associated with poor patient outcomes in intensive care units. Objectives: To estimate the adequacy of nutritional in mechanically ventilated patients and its effect on patient outcomes. Methods: This was a prospective observational study, done over a period of one year(Jan.-Dec. 2015}). All adult patients mechanically ventilated for >48 hours were included in the study. Patients were started

Fig. 26 (abstract A55). Case mix of patients

Table 28 (abstract A55). Enteral nutrition practices in mechanically ventil Routes of nutrition

n=374(%)

Enteral nutrition

325(86.8)

Parenteral nutrition

9(2.4)

Enteral + parenteral nutrition

1(0.3)

No nutrition

39(10.5)

Reasons for not feeding

n=39(%)

Intestinal obstruction/Post-operative bowel surgery

10(25.6)

Hemodynamic instability

10(25.6)

Poisoning

9(23)

GI bleed

8(20.5)

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Page 33 of 607

A56 Prognostic value of post-operative cortisol in cardiothoracic surgery patients with or without high-dose dexamethasone M. Volbeda1, D. ten Kate1, M. Hoekstra1, J.M. van der Maaten2, M.W. Nijsten1 1 University Medical Center Groningen, University of Groningen, Department of Intensive Care, Groningen, Netherlands; 2University Medical Center Groningen, University of Groningen, Departments of Intensive Care and Anesthesiology, Groningen, Netherlands Correspondence: M. Volbeda – University Medical Center Groningen, University of Groningen, Department of Intensive Care, Groningen, Netherlands Intensive Care Medicine Experimental 2016, 4(Suppl 1):A56

Fig. 27 (abstract A55). Lead time to initiate feeding(hours)

Table 29 (abstract A55). Calories achieved over days No.of patients (n)

Calories achieved mean(±SD) Kcal

Percentage of target calories achieved (%)

Day 1

309

656(424)

42.3

Day 2

284

1591(512)

88.7

Day 3

221

1517(512)

90

Day 4

176

1400(517)

73.4

42.3% of targeted calories were achieved on day1

Introduction: The recent randomized controlled DECS-trial demonstrated that high-dose dexamethasone before cardiac surgery with cardiopulmonary bypass decreased ICU length of stay [1]. In our center many patients undergo off-pump cardiac surgery for which no dexamethasone is administered. Post-operative cortisol levels might be associated with outcome [2, 3]. In this study we prospectively measured cortisol levels after cardiothoracic surgery in relation with outcome for patients who did or did not receive dexamethasone. Methods: In 2013 and 2014 we prospectively collected post-operative cortisol levels obtained on the first morning after cardiothoracic surgery. Cortisol was determined in serum with a chemi-luminiscence assay. The use of dexamethasone (1 mg/kg) was recorded as well ICU and hospital length of stay and hospital and long-term mortality after an observation period of at least 1.5 years. Patients were divided into low or high cortisol according to the median cortisol measured in the dexamethasone or control patients. Results: We included 927 patients, 70 % males with a mean ± SD age of 63 ± 13 years. Cardiopulmonary bypass was used in 488 (53 %) and high-dose dexamethasone was administered in 439 (47 %) of the operations. ICU and hospital length of stay were 2.3 ± 5.0 and 13.7 ± 12.5 days respectively. Hospital and long-term mortality were 2.2 % and 6.5 % respectively. Median (IQR) cortisol on the first day postsurgery was 600 (195-965) nmol/L (or 22 ug/dL). Median cortisol levels were lower in the dexa group than the control group (P < 0.0001). Patients in the dexa group had a median (IQR) cortisol of 205 (75-535) and were divided into low-cortisol and high cortisol groups (cortisol < =205 and >205 nmol/L respectively). Likewise control patients had a median cortisol of 830 (585-1125) and were divided into < =830 and >830 nmol/L groups. No significant differences in length of stay or mortality were observed between patients with a low or high cortisol level within the dexa and control groups. Conclusions: Cortisol levels obtained one day after cardiac surgery are far lower in dexamethasone treated patients. Both in controls and dexamethasone-treated patients, lower cortisol levels were not associated with a worse outcome. References [1] Dieleman JM, Nierich AP, Rosseel PM, et al. Intraoperative high-dose dexamethasone for cardiac surgery: a randomized controlled trial. JAMA 2012; 308(17):1761-7. [2] Venkatesh B, Cohen J, Cooper M. Ten false beliefs about cortisol in critically ill patients. Intensive Care Med 2015; 41:1817-19. [3] Henzen C, Kobza R, Schwaller-Protzmann B, et al. Adrenal function during coronary artery bypass grafting. Eur J Endocrinol. 2003; 148(6):663-8.

Fig. 28 (abstract A55). % of feeds received over days

Table 30 (abstract A55). Reasons for interrupting feeds Reasons for interrupting feeds

n=372(%)

Weaning

163(43.8)

Airway-related procedures

99(26.6)

Free water ( for treatment of hypernatremia)

64(17.2)

Preparation for surgery /procedures

39(10.4)

GRV>500ml

10(2.6)

Diarrhoea

10(2.6)

Vomiting

6(1.6)

A57 Comparison of three techniques to estimate albumin catabolic rate A. Komaromi, O. Rooyackers, J. Wernerman, Å. Norberg Karolinska University Hospital Huddinge, Stockholm, Sweden Correspondence: A. Komaromi – Karolinska University Hospital Huddinge, Stockholm, Sweden Intensive Care Medicine Experimental 2016, 4(Suppl 1):A57 Introduction: Plasma albumin concentration is low in critically ill patients despite a higher than normal synthesis rate. It is demonstrated

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

that the capillary escape rate of albumin is increased, related to an enhanced capillary leakage, but the return to plasma through the lymphatic system is not characterized. Nevertheless it is necessary to postulate an enhanced degradation rate or increased volume of distribution, to explain this finding. Objectives: We explored 3 different techniques to assess albumin turnover and degradation. Methods: Healthy volunteers (n = 10) were studied over 42 days. The first day, albumin synthesis rate was measured in the postabsorptive state by the in vivo incorporation of d5-phenylalanine, and a dose of 125 I-albumin was given to assess albumin degradation. Degradation rate was measured and calculated by the decay in radioactivity in 2 ways; in urine collected over 24 hours at 3 separate days, and in plasma repeatedly over 42 days. Results: Basal albumin synthesis rate was 118 ± 16 mg/kg/day, equal to degradation rate in steady state. Catabolic rate by urine sampling was 142 ± 31 mg/kg/day (P = 0.022 v. basal), and by plasma 177 ± 29 mg/kg/day (P = 0.001 v. basal). Conclusions: The differences in calculated turnover rates may correspond to effects of feeding or differences in half-lives between exogenous and endogenous albumin. The exact explanation needs to be further investigated. A58 The relation between plasma glutamine concentration and endogenous glutamine production M. Smedberg, M. Mori, L. Pettersson, Å. Norberg, O. Rooyackers, J. Wernerman Karolinska University Hospital Huddinge, Stockholm, Sweden Correspondence: M. Smedberg – Karolinska University Hospital Huddinge, Stockholm, Sweden Intensive Care Medicine Experimental 2016, 4(Suppl 1):A58 Introduction: A low plasma glutamine concentration at ICU admittance is associated with an unfavorable outcome. A number of studies have shown beneficial effects of exogenous glutamine supplementation to unselected critically ill patients. However, recently harm is reported when pharmacological doses of exogenous glutamine is given to underfed unselected ICU patients with ≥ 2 organ failures. The mechanism behind this finding is not understood and more specifically the reason for the low levels is not known. Objectives: The relation between plasma glutamine concentration and glutamine production rates. Methods: Critically ill patients (n = 17) with normal or low plasma glutamine concentration were studied. Glutamine rate of appearance reflection glutamine production rates was determined employing a bolus dose technique, and the decay curves were fitted into a single pool or a 2-pool model. Results: Glutamine rate of appearance was in the range 5 - 15 umol/ kg/min in ICU patients with plasma concentrations in the range300700 umol/L. There was a statistical significant (P < 0.05) relation, but the coefficient of determination was low R2 = 0.15. Also the correlation between the 1-pool and 2-pool models was statistically significant (P < 0.01), but the coefficient of determination intermediary R2 = 0.60. Conclusions: To understand the results of recent glutamine supplementation studies in the critically ill, the relevance of plasma glutamine levels must be under stood. In healthy volunteers there is a high concordance between 1-pool and 2-pool models to calculate the rate of appearance. The lower agreement in ICU patients indicates that a more extensive modeling may be necessary.

A59 Effect of enteral versus parenteral nutrition on outcome of mechanically ventilated septic ICU patients M. Theodorakopoulou, T. Christodoulopoulou, A. Diamantakis, F. Frantzeskaki, M. Kontogiorgi, E. Chrysanthopoulou, M. Lygnos, C. Diakaki, A. Armaganidis Attikon University Hospital, ICU, Athens, Greece Correspondence: M. Theodorakopoulou – Attikon University Hospital, ICU, Athens, Greece Intensive Care Medicine Experimental 2016, 4(Suppl 1):A59

Page 34 of 607

Introduction: The prevalence of malnutrition in ICU has been estimated at up to 40 % with the majority of critically ill patients requiring nutritional support. Traditional teaching suggests that early enteral feeding in mechanically ventilated septic patients is superior to parenteral feeding due to a lower complication rate but neither form of support is without risk. Objectives: In this study, we compared the outcomes in mechanically ventilated septic ICU patients receiving enteral versus parenteral nutrition. Methods: A single centre study of patients admitted to a 25 bed University Hospital ICU over a period of three years. Demographics, severity of illness scores (APACHE and SOFA), BMI and MUST were measured upon admission. Daily nutrition requirements were calculated for each patient. Patients were randomized to enteral(EN) or parenteral nutrition(PN) group within 48 hours of intubation and admission to the unit. Duration of mechanical ventilation, ICU and hospital length of stay (LOS), and mortality rates were recorded. Results: A total of 148 patients (76 men) mechanically ventilated septic patients having a mean(±SD) age of 69.6 ± 19.4 years were studied. All patients met the consensus criteria for sepsis. Baseline characteristics were similar in the two groups. APACHE II and SOFA at study entry were 24 ± 5 and 8 ± 3 respectively. The mean(±SD) BMI was ≈ 21.5 ± 3.4 kg/m2. Seventy seven (52,02 %) patients received EN, and sixty nine (47,26 %) received PN. There was no difference between the two groups for age, sex, BMI, and scores. ICU and hospital LOS were similar for both groups. ICU mortality rate was 29.4 % for PN group vs. 27.2 % for EN group indicating no significant difference. Hospital mortality was similar for both groups as well. In the PN fed group however, duration of mechanical ventilation was longer (p = .018), but the feeding goal was attained earlier (p = .009). Conclusions: In mechanically ventilated septic ICU patients the ICU LOS and the hospital LOS, as well as the ICU and hospital mortality rates of patients receiving PN are not significantly different than those in patients receiving EN. Furthermore, feeding goals can be attained much easier by PN. Duration of mechanical ventilation however appears to be longer in patients receiving PN.

A60 Association between the route of nutrition and adipokine hormones levels in critically ill patients: a pilot study K. Gundogan1, E. Dogan2, R. Coskun1, S. Muhtaroglu3, M. Sungur1, T. Ziegler4, M. Guven1 1 Erciyes University, Intensive Care Unit, Kayseri, Turkey; 2Erciyes University, Internal Medicine Department, Kayseri, Turkey; 3Erciyes University, Clinical Biochemistry Department, Kayseri, Turkey; 4Emory University, Department of Medicine, Division of Endocrinology, Metabolism and Lipids, Atlanta, United States Correspondence: K. Gundogan – Erciyes University, Intensive Care Unit, Kayseri, Turkey Intensive Care Medicine Experimental 2016, 4(Suppl 1):A60 Introduction: Adipokine hormones play an important role in regulation of insulin metabolism, body fat distribution and regulation of appetite and satiety. Some adipokine hormones have effects on inflammation and insulin resistance but the relation between these hormones and critical illness and the route of nutrition is not known. Objectives: The aim of this study to determine association between nutrition route and adipokine hormones levels in critically ill patients Methods: This study was performed prospectively in Medical and Surgical ICU at Erciyes University. Patients expected to stay in ICU at least 72 hours and received either parenteral or enteral nutrition included into the study. Results: Total of 26 patients were included into the study and 17 of them were male (65 %). The mean age was 62.8 ± 18.2 years. Total of 14 patients (54 %) were fed via enteral route and 12 patients (46 %) were received parenteral nutrition. The mean APACHE II score was 22.7 ± 7.1. Resistin levels were lower in enteral nutrition group at 24th (p = 0.015) and 72nd hours (p = 0.014) compared to parenteral nutrition group. Baseline, 24.th hours and 72.th hours GLP-1 levels were found to be higher in enteral nutrition group than parenteral

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Page 35 of 607

nutrition group (p = 0.031, p = 0.006 and p = 0.001 respectively). Adiponectin levels were significantly higher in enteral nutrition group compared to parenteral nutrition group at 72th hour (p = 0.014). Conclusions: Our study showed that enteral nutrition helped to reverse abnormal process in critically ill patients with increasing adiponectin and GLP-1 levels and decreasing resistin levels.

References (1) Rudiger A. et al. Clin Sci 2013; 124:391-401 Grant acknowledgement ESICM Basic Science Award, Intensive Care Foundation (UK) Young Investigator Award, NIHR

References 1. Kwon H, Pessin JE. Adipokines mediate inflammation and insulin resistance. Frontiers in endocrinology. 2013;4:71. 2. Vassiliadi DA, Tzanela M, Kotanidou A, Orfanos SE, Nikitas N, Armaganidis A, et al. Serial changes in adiponectin and resistin in critically ill patients with sepsis: associations with sepsis phase, severity, and circulating cytokine levels. Journal of critical care. 2012;27(4):400-9.

Table 31 (abstract A61). Sham

Survivors

Non-survivors

Phospho PDH

1.4±0.32

2.85±0.42

5.79±0.42a,b

PDH

0.99±0.11

0.78±0.31

0,24±0.03a,b

0.37±0.01

1.37±0.012a

1.51±0.012a

0.25±0.001

0.32±0.005

0.37±0.004

Soleus

Heart Phospho ACC

A61 Molecular mechanisms providing a switch from carbohydrate to fat metabolism in different organs in sepsis A. Kleyman1,2, W. Khaliq1, D. Andreas1, M. Singer1 1 University College London, Bloomsbury Institute of Intensive Care Medicine, London, United Kingdom; 2University Hospital Jena, SG Sepsis Research, Department of Anaesthesiology and Intensive Care Medicine, Jena, Germany Correspondence: A. Kleyman – University College London, Bloomsbury Institute of Intensive Care Medicine, London, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A61 Introduction: A switch from carbohydrate to fat utilization is a hallmark of systemic inflammation. While this adaptive response occurs quickly and allows animals to survive under restricted food supply conditions, it will markedly affect cell metabolism. Objectives: To elucidate molecular mechanisms underlying the substrate switch from carbohydrate to fat in different organs, we studied early alterations in (i) phosphorylation of enzymes involved in energy metabolism: AMP kinase (AMPK, thr172), acetylCoA carboxylase (ACC ser179), pyruvate dehydrogenase (PDK ser293), hormone sensitive lipase (HSL ser563); and (ii) expression of mitochondrial uncoupling proteins 2 and 3 in soleus and gastrocnemius skeletal muscle, liver, kidney and heart in sham-operated and septic rats at 6 h in our wellcharacterized 72 h fluid-resuscitated rat model of faecal peritonitis Methods: Awake, instrumented yet fully mobile male Wistar rats (325 ± 15 g) received an i.p. injection of 4 μl/g faecal slurry. Fluid resuscitation (50:50 mix of 5 % glucose/Hartmann´s; 10 ml/kg/h) was commenced at 2 h. Control animals were treated identically except for slurry injection. At 6 h, an echo-measured heart rate cut-off of 460 bpm was used to classify animals into predicted survivors SR or non-survivors NSR.(1) Animals were killed and organs were immediately collected into liquid nitrogen. Alterations in protein phosphorylation and expression were studied by Western blot. Normalization was performed to loading control: actin or PFK. Results were presented as mean +/- SE, analyzed using Student's t-test and considered statistically significant when p < 0.05. Results: At 6 h post-sepsis no differences were seen in renal and hepatic phosphorylation of AMPK, ACC, PDH and HSL between sham and septic animals. While cardiac ACC phosphorylation was strongly increased in septic rats, AMPK phosphorylation did not differ, suggesting that ACC phosphorylation was mediated not by AMPK but rather via the glucagon-PKA pathway. The biggest changes were observed in skeletal muscle. AMPK phosphorylation was increased in gastrocnemius and even more so in soleus in septic rats but this was not accompanied by a corresponding increase in ACC phosphorylation. In both muscles PDH phosphorylation markedly increased while PDH fell, suggesting a fall in pyruvate oxidative decarboxylation and glucose usage as a fuel. HSL phosphorylation was strongly increased in soleus in non-survivors. UCP2 and UCP3 levels were not altered in any organ. Table 31. Conclusions: 1) Each organ has its own program for the switch from carbohydrate to fat metabolism. 2) The consequences of these changes on the development of organ dysfunction merit further investigation, as this may lead to novel directed therapeutics.

Phospho AMPK a

p< 0.05 vs control,

b

p< 0.05 vs survivors

A62 Estimation of energy requirements using a standard ventilator R. Meierhans, R. Schuepbach University Hospital Zurich, University Zurich, Surgical Intensive Care Unit, Zurich, Switzerland Correspondence: R. Schuepbach – University Hospital Zurich, University Zurich, Surgical Intensive Care Unit, Zurich, Switzerland Intensive Care Medicine Experimental 2016, 4(Suppl 1):A62 Introduction: Albeit practical implementation remains a concern, indirect calorimetry (IC) is the first choice technique for estimating energy requirements in critically ill patients. Methods: In our adult mainly surgical intensive care unit (ICU) we retrospectively assessed in 25 patients the utility and practical aspects of estimating the daily energy requirements. We determined energy requirements by IC (Quark, COSMED) and compared the estimates with those calculated by measuring VCO2 alone using a ventilator (Evita XL; Draeger) or using formulas based on body mass of body mass index. Results: Our study population was found to have an average body mass of 69.7 kg and according to IC and energy requirement of 1987 kcal/day on average. No correlation (R2 = 6E-05) was found when estimates of IC were compared to estimates based on body mass alone. However using an adjusted mass (based on hight and an assumed body mass index of 23) correlation improved (R2 = 0.3358). VCO2 based estimation of energy requirement only weakly correlated with IC (R2 = 0.414). Weak correlation is explained by assuming a fixed respiratory coefficient for ventilator based energy estimates as well as by imperfect correlation of VCO2 estimates reported by calorimeter and ventilator (R2 = 0.6189) Conclusions: Body parameters such as hight or mass aren't suited to predict the patient's energy needs. Albeit promising and simple, estimation of energy requirement based on ventilator derived VCO2 also fails precision. Whether technical improvements on how ventilators calculate VCO2 can render them more suited in daily praxis will have to be determined in future studies. A63 Nutritional adequacy of ECMO patients I. De Brito-Ashurst Royal Brompton and Harefield NHS Foundation Trust, London, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A63 Introduction: Extracorporeal membrane oxygenation (ECMO) is an advanced treatment to support the critically ill patients with cardiac and/or with severe respiratory failure. It is suggested that ECMO patients are at risk of splanchnic ischemia and thus, enteral feeding is poorly tolerated needing nasojejunal feeding and is potentially unsafe. This study aims to investigate the nutritional adequacy of this patient group.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Objectives: To establish feeding tolerance, caloric and protein intake during the initial 5 days of feeding. Methods: A retrospective review of all the patients that received ECMO between January/2000 and December/2014 was conducted in a tertiary critical care unit/ECMO referral centre. Patients were all fed as per unit feeding protocol that recommends early enteral feeding instead of parenteral or delayed enteral nutrition. Results: Data were reviewed for 80 patients for the initial 5 feeding days and is reported as mean and (SD). The average duration of ECMO was 12.7 (9.9) days. Sixty-five patients received venovenous ECMO for respiratory failure whilst the remainder received venoarterial ECMO for cardiac failure. Patients age, Body mass index (BMI) and sofa scores were 44.1 (15.7), 29.1 (8.5) and 15.7 (2.8) respectively. Enteral feeding started at 14.2 (5.8) hrs and reached 83 % target within 72 hrs. Overall, patients had good tolerance to the feed with a mean gastric residual volume (GRV) of 176.6 ml/24 hrs (167.75) during the 5 days. Feeding intolerance, defined as GRV >250 ml, were observed every day with 7, 13, 15, 10 and 8 episodes on days 1, 2, 3, 4 and 5 accordingly. Diarrhoea was observed on 16 % of the patients and mainly on days 3 (15/80), 4 (16/80) and 5 (17/80). Forty-two patients needed prokinectics (metoclopramide) and eight of those were also on erythromycin. Eight patients were on parenteral nutrition to supplement inadequate enteral feeding. Overall patients were receiving >80 % by day 3 onwards. All patients were on nasogastric feeding and nasojejunal feeding was not necessary. Conclusion: Enteral nutrition is well tolerated by patients receiving ECMO, whether in venovenous or venoarterial mode. No serious adverse events were attributable to enteral nutrition in these patients.

RESPIRATORY INFECTIONS A64 Incidence of ventilator associated pneumonia when pantoprazole or ranitidine is used for stress ulcer prophylaxis in critically ill adult patients: a pilot study F. Zand1, G. Sabetian2, R. Nikandish3, F. Hagar3, M. Masjedi1, B. Maghsudi1, A. Vazin3, M. Ghorbani1, E. Asadpour1 1 Anesthesiology and Critical Care Research Center, Shiraz University of Medical Sciences, Shiraz, Islamic Republic of Iran; 2Shiraz University of Medical Sciences, Trauma Research Center, Shiraz, Islamic Republic of Iran; 3 Shiraz University of Medical Sciences, Shiraz, Islamic Republic of Iran Correspondence: G. Sabetian – Shiraz University of Medical Sciences, Trauma Research Center, Shiraz, Islamic Republic of Iran Intensive Care Medicine Experimental 2016, 4(Suppl 1):A64 Introduction: Nowadays, hospital acquired infections are the most widespread phenomena among the critically ill patients. Prophylaxis of gastric stress ulcer with acid-suppressive therapy increases the risk of ventilator associated pneumonia (VAP) in critically ill patients(1). While, two meta- analyses comparing PPI and H2RA in terms of prophylactic effectiveness for gastric stress ulcer in critical patients have showed no apparent differences in both groups as regard to upper gastric ulcer prophylaxis, pneumonia and mortality in patients admitted to ICU (2,3). Objectives: Prophylaxis against stress ulcer in mechanically ventilated patients has been considered as a culprit in the development of ventilator associated pneumonia (VAP) in retrospective studies. To test this hypothesis in a mixed medical-surgical adult ICU, we designed a randomized clinical pilot study to determine the necessary sample size for comparing the effect of intravenous ranitidine and pentoprazole on the incidence of VAP. Methods: Patients with at least 48 hours of expected mechanical ventilation were allocated randomly to receive either 50 mg ranitidine (R) every 8 hours or 40 mg pentoprazole (P) every 12 hours intravenously from admission till 48 hours after extubation. VAP diagnosis was according to Clinical Pneumonia Infection Score and positive culture of endotracheal tube aspiration. Results: After recruiting 86 patients during 15 months period and a preliminary analysis, the study was terminated due to very low

Page 36 of 607

difference between VAP incidence in R (32.6 %) and P groups (27.9 %; P value 0.63, Odds ratio; 1.24, confidence interval; 0.493.13). No statistically significant difference was observed in terms of gastro-intestinal bleeding, ICU and hospital length of stay and mortality between the two groups. Conclusions: VAP incidence is hardly related to the type of stress ulcer prophylaxis agent in our ICU setting with high rate of VAP and low utilization of VAP prophylaxis bundle. To find any such effect a huge number of patients are needed to be recruited in a large randomized clinical trial References 1. Prescott HC, O´Brien JM. Prevention of ventilator-associated pneumonia in adults. F1000 medicine reports. 2010 2. Lin PC, Chang CH, Hsu PI, Tseng PL, Huang YB. The efficacy and safety of proton pump inhibitors vs histamine-2 receptor antagonists for stress ulcer bleeding prophylaxis among critical care patients: a meta-analysis. Critical care medicine. 2010;38(4):1197-205 3. Alhazzani W, Alenezi F, Jaeschke RZ, Moayyedi P, Cook DJ. Proton pump inhibitors versus histamine 2 receptor antagonists for stress ulcer prophylaxis in critically ill patients: a systematic review and meta-analysis. Critical care medicine. 2013;41(3):693-705. Grant acknowledgement This study was funded by a grant number 4227 from Vice-Chancellery of Research and Technology in Shiraz University of Medical Sciences, Shiraz, Iran.

A65 Coinfection associated mortality in pneumonia-induced acute respiratory distress syndrome (ARDS) K.-C. Kao1, L.-C. Chiu2, C.-Y. Hung2, C.-H. Chang2, S.-H. Li2, H.-C. Hu2 1 Chang Gung Memorial Hospital, Thoacic Medicine, Kwei-Shan, Taoyuan, Taiwan, Province of China; 2Chang Gung Memorial Hospital, Taoyuan, Taiwan, Province of China Correspondence: K.-C. Kao – Chang Gung Memorial Hospital, Thoacic Medicine, Kwei-Shan, Taoyuan, Taiwan, Province of China Intensive Care Medicine Experimental 2016, 4(Suppl 1):A65 Rationale: Pneumonia is the leading risk factor of acute respiratory distress syndrome (ARDS). For the critically ill patients with pneumonia, coinfection had higher mortality and longer length of stay in the intensive care unit (ICU). Little is known about the impact of coinfection on the outcomes of pneumonia-induced ARDS patients. Objectives: To evaluate the role of coinfection from bronchoalveolar lavage (BAL) examination on the outcomes of pneumonia-induced ARDS. Methods: We performed a prospective observational study in adult ICUs at the Chang Gung Memorial Hospital from October 2012 to May 2015. Patients were included if they met the Berlin definition of ARDS. The BAL indication were pneumonia-induced ARDS was suspected clinically and no definite microbial was indentified from tracheal aspirate. The BAL specimen was sent for comprehensive microbiological study including bacteria, fungi and virus. Demographics and baseline clinical characteristics were collected on enrollment. The final pathogen identified results and clinical outcomes were analyzed. Results: Totally 19936 receiving invasive mechanical ventilation patients admitted to ICUs with PaO2/FiO2 < 300 mmHg screened, 902 (4.5 %) patients fulfilled the Berlin definition of ARDS. Of these ARDS patients, 205 (22.7 %) patients who were pneumonia-induced ARDS with BAL for pathogen survey were included for analysis. One hundred and forty two (55.7 %) patients were identified with microbiological pathogen. According the identified virus pathogen, these 142 patients were divided as only virus identified group (n = 41, 28.9 %), no virus identified group (n = 60, 42.2 %) and coinfection with other pathogen group (n = 41, 28.9 %). The distribution of ARDS severity were no significant difference between these 3 group patients (p = 0.43). The coinfection group had significantly higher hospital mortality rate than only virus group and no virus group (80.5 % vs 53.7 % and 63.3 %; p = 0.03).

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Conclusions: The coinfection with virus and other pathogen from BAL was associated with increased mortality in pneumonia-induced ARDS patients. A66 Kallistatin level as a novel prognostic marker for community acquired pneumonia (CAP) in critically ill patients S. El Maraghi1, M. Ali1, D. Rageb2, M. Helmy1 1 Faculty of Medicine - Beni Suef University, Critical Care Department, Cairo, Egypt; 2Faculty of Medicine - Cairo University, Critical Care Department, Cairo, Egypt Correspondence: S. El Maraghi – Faculty of Medicine - Beni Suef University, Critical Care Department, Cairo, Egypt Intensive Care Medicine Experimental 2016, 4(Suppl 1):A66 Introduction: CAP is a potentially life threatening disorder despite the advent of potent antibiotics & commonly causes acute respiratory failure with high mortality1.Assessment of disease severity & prediction of outcome is essential for appropriate allocation of health care resources & for optimized treatment decisions2. Kallistatin, an endogenous tissue kallikrein inhibitor; which protects against inflammation, fibrosis & oxidative stress may be involved in CAP pathogenesis through anti-inflammatory effect3. Objectives: To determine whether kallistatin levels have a prognostic value in severe CAP & to correlate it with other biomarkers as CRP,APACHE II ,SOFA, CURB-65 scores & Pneumonia severity index (PSI). Methods: Plasma samples and clinical data were prospectively collected from 30 patients admitted to the Critical Care Department with severe CAP according to IDSA/ATS Criteria4. Serum Kallistatin levels were collected on days 1 and 4 of ICU admission. It was correlated with APACHE II, SOFA, CURB-65 scores & PSI on admission also with PO2/FiO2 ratio and CRP on days 1 and 4. Results: Lower kallistatin levels on days 1 & 4 showed a strong trend toward increased mortality with a p-value of 0.043 & 0.04 respectively. Its level was negatively correlated with APACHE II,SOFA,CURB-65 scores & PSI and with a p-value of 0.036,0.001, 0.102 & 0.001 respectively while it showed a positive linear correlation with Po2/Fio2 ratio on days 1 & 4 with a P-value of 0.001 & 0.005 respectively. It also showed a negative correlation with CRP on days 1 and 4 with a p-value of 0.001. Its level on days 1 & 4 of ICU admission were significantly decreased in patients who developed septic shock with a p-value of 0.044 & 0.043 respectively and who had ARDS requiring mechanical ventilation with a P-value of 0.001 & 0.005 respectively. Kallistatin cut-off value to predict mortality determined by ROC curve was 6.65ug/ ml on day1 with sensitivity 90 % and specificity 80 % & a p-value 0.022 while was 8.6ug/ml on day 4 with sensitivity 85 % and specificity 77 % & a p-value 0.035. Conclusions: kallistatin may serve as a promising novel marker for prognosis of severe CAP & may be involved in its pathogenesis References 1. Restrepo MI, Mortensen EM, Velez JA, et al: A comparative study of community-acquired pneumonia patients admitted to the ward and the ICU. Chest 2008, 133:610-617. 2. Lim WS, van der Eerden MM, Laing R, Boersma et al: Defining community acquired pneumonia severity on presentation to hospital: an international derivation and validation study. Thorax 2003, 58:377-382. 3. Wei-Chieh Lin, Shiou-Ling Lu, Chiou-Feng Lin, et al: Plasma kallistatin levels in patients with severe community-acquired pneumonia.Critical Care 2013, 17:R27. 4. Mandell LA, Wunderink RG, Anzueto A, et al. Infectious Diseases Society of America/American Thoracic Society consensus guidelines on the management of community-acquired pneumonia in adults. Clin Infect Dis 2007;44(Suppl 2): S27-72.

Page 37 of 607

A67 Clinical impact of delayed diagnosis of influenza A (H1N1)pdm09 in critically ill patients J. Marin-Corral1, C. Vilà1, J.R. Masclans1,2,3, A. Vàzquez1, I. Martín-Loeches4, E. Díaz5, J.C. Yébenes6, A. Rodriguez7, F. Álvarez-Lerma1,8, H1N1 SEMICYUC/GETGAG Working Group 1 Hospital Parc de Salut Mar - GREPAC, IMIM, Critical Care Department, Barcelona, Spain; 2Pompeu Fabra University (UPF) - CEXS, Barcelona, Spain; 3CIBERES. Instituto de Salud Carlos III, Madrid, Spain; 4 Multidisciplinary Intensive Care Research Organization (MICRO). St James's; University Hospital. Trinity Center for Health Sciences, Department of Anesthesia and Critical Care, Dublin, Ireland; 5Hospital Parc Tauli, Critical Care Department, Sabadell, Spain; 6Hospital de Mataró, Critical Care Department, Mataró, Spain; 7University Hospital Joan XXIII IISPV-URV, Critical Care Department, Tarragona, Spain; 8Universitat Autònoma de Barcelona, Barcelona, Spain Correspondence: J. Marin-Corral – Hospital Parc de Salut Mar GREPAC, IMIM, Critical Care Department, Barcelona, Spain Intensive Care Medicine Experimental 2016, 4(Suppl 1):A67 Objectives: To assess the clinical and developmental implications of delayed diagnosis of community-acquired influenza A (H1N1) pdm09 infection in critically ill patients admitted to the ICU. Methods: Prospective, observational and multicenter study in 148 Spanish ICU. Data were obtained from GETGAG / SEMICYUC (20092015). All patients infected by influenza A (H1N1) confirmed by RT-PCR in the first week of hospital stay were included (viral communityacquired pneumonia). Patients were classified according to the moment of flu diagnosis: early (in the first two days of hospitalization) and late (between the 3rd and 7th day of admission). Demographic, temporary variables, comorbidities, severity at ICU admission, treatment and mortality were evaluated intra-ICU. Logistic regression was used to identify related factors to late diagnosis and Cox regression to determine whether the late diagnosis was an independently variable associated with mortality. The results are presented as odds ratios (OR) and 95 % confidence intervals (CI). We considered p < 0.05 to be significant. Results: 2059 ICU patients diagnosed in the first 7 days of hospital stay were evaluated. 1314 (63.8 %) were early diagnosed and 745 (36.2 %) were late diagnosed. Independent variables associated to late diagnosis were: days of hospital stay until ICU admission (OR 1.254, 95 % CI 1.169 to 1.345; p < 0.001), mechanical ventilation (OR 1.690, 95 % CI 1.259 to 2.268; p < 0.001) and the need of renal replacement (OR 1.509, 95 % CI 1.071 to 2.125; p = 0.019). Clinical presentation as viral pneumonia was a protective variable related to the delay (OR 0.654, 95 % CI 0.495 to 0.881; p = 0.005). Patients with late diagnosis presented higher ICU [8(4-17) vs 10(5-20), p = 0.000] and hospital days of stay [14(8-25) vs 18(10-30), p = 0.000] as well as higher intra-ICU mortality (20.5 % vs 32.6 %, p = 0.000). The diagnosis delay was an independently variable associated with mortality in the multivariate analysis (OR 1.364, 95 % CI 1.029 to 1.808). Conclusions: Influenza A (H1N1)pdm09 diagnosis delay in critically ill patients admitted to the ICU is associated with delayed ICU admission, increased respiratory failure or acute renal failure and increased mortality. Delayed diagnosis is an independently variable associated with mortality. A68 Influenza A in severe acute respiratory infection. Timing for ICU admission is important N. Varga, A. Cortina-Gutiérrez, L. Dono, M. Martínez-Martínez, C. Maldonado, E. Papiol, M. Pérez-Carrasco, R. Ferrer Hospital Universitari Vall d'Hebron, Critical Care, Barcelona, Spain Correspondence: N. Varga – Hospital Universitari Vall d'Hebron, Critical Care, Barcelona, Spain Intensive Care Medicine Experimental 2016, 4(Suppl 1):A68

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Introduction: Influenza viruses is one of the main cause of severe acute respiratory infection (SARI). To know its clinical features and prognosis factors are essential for management and outcome. Objectives: To determine the incidence and mortality of patients with SARI and influenza A infection admitted to the ICU; and to identify the most common symptoms and comorbidities among patients affected by SARI and influenza A infection. Methods: A prospective observational study was performed including patients with SARI admitted in ICU of a tertiary care Hospital from november 2014 to march 2015. Influenza infection was diagnosed by RT-PCR. Patients´ demographic, clinical and radiologic features and outcomes were recorded. Pearsoń s chi-square, Fisher, Mann Whitney U tests and logistic regression have been used for statistical analysis. Data are expressed as frequency (percentage) or median (25th-75th interquartile range). Results: During study period 425 patients were admitted to ICU. Seventy-five (17.6 %) were included as SARI (57.3 % male; median age 57(47-67) years old; APACHE II 21(13-27) and SOFA 8(5-10); ICU LOS 8(4-23) days; days from symptom onset to hospital admission 4(1-6)). Main co-morbidities were lung diseases in 44(33 %) patients and cardiovascular diseases in 19(25.3 %) patients. Seventeen (22.7 %) patients presented Influenza A infection, only one did not presented viral pneumonia. Cough, shortness of breath and muscle pain were the most common symptoms in influenza A patients (87.5 %, 68.8 % and 43.8 % respectively). Influenza-like illness was more frequent in influenza A patients (52.9 % vs. 32.8 %; p = 0.13), furthermore most patients with influenza A infection were diagnosed during influenza epidemic period (88.2 % vs. 37.9 %; p < 0.001); all patients with influenza A infection were treated with oseltamivir. High-flow oxygen therapy was required in 53(70.7 %) SARI patients, 11(14.7 %) non-invasive mechanical ventilation, 46(61.3 %) invasive mechanical ventilation, 30(40 %) inotropic drugs and 11(14.7 %) renal replacement therapy. Influenza A patients presented a similar distribution. Overall SARI mortality was 20 %, only one with influenza A infection (5.6 % vs. 26.7 %; p = 0.1). APACHE II (25(19-33) vs. 19(11-26); p = 0.03) and SOFA score (12(8-14) vs. 8(5-10); p = 0.003) showed to be a good predictor of mortality at ICU admission. Also, delayed admission to ICU (4(1-16) vs. 1(0-2); p = 0.01) and renal replacement therapy (40 % vs. 6.9 %; p = 0.04) were associated with increased mortality. Previous renal diseases (OR 10.28 CI 95 % (2.25-46.96); p = 0.003) and oncology diseases (OR 6.8 CI 95 % (1.59-29.11); p = 0.01) were independently associated with mortality. Conclusions: A nonsignificant decrease in mortality was observed in patients with influenza A infection. Early ICU admission could improve SARI prognosis.

A69 Do routine clinical markers predict discharge in patients with H1N1 influenza? K. Nweze1,2, B. Morton2, I. Welters2 1 University of Liverpool, Institute of Infection and Global Health, Liverpool, United Kingdom; 2Royal Liverpool University Hospital, Intensive Care Unit, Liverpool, United Kingdom Correspondence: K. Nweze – University of Liverpool, Institute of Infection and Global Health, Liverpool, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A69 Introduction: H1N1 influenza can cause both self-limiting and lifethreatening illnesses. Triaging patients who require hospital admission and those who can be safely discharged can be difficult in a pandemic situation. Existing triage tools focus on prediction of hospital admission in H1N1. We assessed the ability of routine clinical markers to predict early discharge in patients with confirmed H1N1 influenza. Objectives: 1) Assess the ability of routine clinical and laboratory variables to predict patients who could safely be discharged home within 24 hours. 2) Compare the predictive value of clinical variables to triaging tools (STSS and SOFA) and the qSOFA criteria. Methods: We reviewed an existing database of patients who presented to the Royal Liverpool University hospital from 2010 - 2011. Inclusion

Page 38 of 607

criteria were: patients with H1N1 confirmed by reverse transcriptase polymerase chain reaction who were ≥18 years of age. Exclusion criteria were: unconfirmed cases, patients not seen in the hospital and with missing case notes. Differences in clinical parameters between patients discharged within 24 hours of medical assessment and those admitted to hospital were analysed. A chest X-ray scoring system was also employed to assess the ability of radiographic findings to predict likelihood of discharge. Results: Eighty-six patients were eligible for the study. 17 patients were discharged early and 69 patients were admitted to hospital. P/F ratio and CRP predicted discharge with area under receiver operating characteristic (ROC) curves of 0.788 (CI 0.681-0.894) and 0.763 (CI 0.6377-0.889) respectively, which was higher than triage and bedside tools. The chest radiograph scoring tool did not predict patient discharge (p = 0.191-0.999), but demonstrated very good inter-rater reliability (Cohen's kappa statistic >0.8). Conclusions: P/F ratio and CRP predicted discharge better other clinical parameters. Both were superior to H1N1 specific triage tools described in the literature. P/F ratio is a simple and effective method to determine oxygen exchange. We recommend this tool in the assessment of patients during influenza pandemic to guide management decisions and future work would involve validation in prospective cohorts. References 1. Singanayagam et al (2011) ‘Factors associated with severe illness in pandemic 2009 influenza a (H1N1) infection: implications for triage in primary and secondary care’, J Infect, 63(4), 243-51. 2. Singer et al. (2016) ‘The third international consensus definitions for sepsis and septic shock (sepsis-3)’, JAMA, 315(8), 801-810. 3. Taylor, E. (2015) ‘A chest radiograph scoring system in patients with severe acute respiratory infection: a validation study’, BMC Med Imaging, 15, 61.

A70 Relationship between digestive tract colonization and subsequent ventilator-associated pneumonia related to ESBL-producing enterobacteriaceae M. Houard, B. Voisin, G. Ledoux, S. Six, E. Jaillette, S. Nseir Lille University Hospital, ICU, Lille, France Correspondence: M. Houard – Lille University Hospital, ICU, Lille, France Intensive Care Medicine Experimental 2016, 4(Suppl 1):A70 Introduction: Ventilator-associated pneumonia (VAP) is the most common ICU-acquired infection. Recently, the incidence of extended-spectrum beta-lactamase producing Enterobacteriaceae (ESBLE) has substantially increased in critically ill patients. Identifying patients at risk for VAP related to ESBLE could be helpful to improve the rate of appropriate initial antibiotic treatment, and reduce unnecessary exposure to carbapenems. Objectives: The primary objective was to identify risk factors for VAP related to ESBLE. Secondary objective was to determine the impact of ESBLE on outcome of VAP patients. Methods: This retrospective study was conducted in a single mixed ICU, during a 4-year period. All patients with confirmed VAP were included. VAP was defined using clinical, radiologic and quantitative microbiological data. VAP first episodes were prospectively identified using the continuous surveillance data. Exposure to different risk factors was taken into account until the diagnosis of ESBLE VAP or until ICU discharge, in patients with ESBLE VAP and VAP related to other bacteria, respectively. In all patients, routine screening for ESBL (rectal swab) was performed at ICU admission and once a week. Patients with ESBLE VAP were compared with those with VAP related to other bacteria using univariate analysis. All significant factors were included in the multivariate logistic regression model. Results: Among the 410 patients with VAP, 43 (10.5 %) had ESBLE VAP. 76 (18.5 %) patients had polymicrobial VAP, and 189 (46 %) had VAP related to multidrug resistant bacteria. The following factors were significantly associated with higher rates of ESBLE VAP by univariate analysis: female gender, medical admission, ARDS, shock, or infection at admission, antibiotic treatment during the

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

30 days preceding ICU admission, large-spectrum antibiotic treatment during ICU stay, and digestive tract colonization related to ESBLE. Acute exacerbation of COPD at ICU admission was significantly associated with lower rate of ESBLE VAP by univariate analysis. Multivariate analysis identified prior ESBLE colonization of the digestive tract as the only independent risk factor for ESBLE VAP (OR [95 % CI] = 23 [10-55], p < 0.001). Whilst the positive predictive value of ESBL digestive colonization (43.6 %) was low, its negative predictive value was excellent (97.3 %) in predicting ESBLE VAP. Duration of mechanical ventilation (median [IQR], 28 [18, 42] vs 23 [15, 42] d, p = 0.4), length of ICU stay 31 [19, 53] vs 29 [18, 46] d, p = 0.6), and mortality rates (55.8 % vs 50 %, p = 0.48) were similar in ESBLE VAP, compared with VAP related to other bacteria. Conclusions: Digestive tract colonization related to ESBLE is independently associated with ESBLE VAP. Its excellent negative predictive value suggests that patients without ESBLE colonization should not receive carbapenems as part of their initial empirical antibiotic treatment to cover ESBLE. A71 Severe acute respiratory infections in a Tunisian ICU S. Romdhani1, R. Bouneb1, D. Loghmari1, N. Ben Aicha1, J. Ayachi1, K. Meddeb1, I. Chouchène1, A. Khedher1, M. Boussarsar1,2 1 Farhat Hached University Hospital, Medical Intensive Care Unit, Sousse, Tunisia; 2Ibn Al Jazzar Faculty of Medicine, University of Sousse, Research Laboratory N° LR14ES05, Interactions of the Cardiopulmonary System, Sousse, Tunisia Correspondence: S. Romdhani – Farhat Hached University Hospital, Medical Intensive Care Unit, Sousse, Tunisia Intensive Care Medicine Experimental 2016, 4(Suppl 1):A71 Introduction: Severe acute respiratory syndrome (SARS) and H1N1 influenza infection (1) have activated an interest in the surveillance of patients with severe acute respiratory infections (SARI). Objectives: To investigate the local epidemiology, patterns of infections, severity and outcome in patients admitted to the intensive care unit (ICU) as a result of severe acute respiratory infections (SARI). Methods: This is a prospective observational study. All patients admitted to the ICU, from October 1st, 2015 to march 31st, 2016 were screened. Were studied, demographic characteristics, underlying conditions, clinical presentation, therapeutic intervention and outcome. SARI was defined as an acute respiratory illness of recent onset (within 7 days) that includes fever (≥38 °C), cough, and dyspnea requiring overnight hospitalization (2). Results: 33 patients were screened over the study period. Median (IQR) age was 68(18-87) years. 63.6 % were male. Median (IQR) SAPS II, 26(13-64) and SOFA score, 5(2-9). 24(77.2 %) had underlying diseases. The symptoms' onset occurred at a median (IQR) of 4(3-7) days before admission to the ICU. Chest X-ray displayed alveolar consolidation in 19(86.4 %). On admission to the ICU, Gram-positive and Gram-negative bacteria were found in 5(15 %) and 8(24 %) of SARI patients but rarely atypical bacteria in 1(3 %). Viruses were present in 6(18 %) of the patients (influenza A-H1N1, 3; A-H3N2, 3). 22(66.6 %) required invasive mechanical ventilation, 18(54.5 %) vasopressors, 10(45.4 %) corticosteroids, 21(63.6 %) antimicrobial therapy and none was treated with antiviral agents. Organ failure occurred in 26 (80 %) patients, (respiratory, 14 (53,9 %) ; hemodynamic, 9(34,6 %) and renal, 3(11,5 %)). Median (IQR) length of stay was 6(1-30) days and duration of mechanical ventilation was 4(1-20) days. ICU mortality rates in patients with SARI were, overall, 18(54.4 %) ; 7(45 %) in severe and prolonged hypoxemia and 11 (54 %) with refractory shock. Univariate analysis identified respiratory chronic failure and greater severity scores at ICU admission to be associated with an increased risk of ICU death. Conclusions: Since the pandemic influenza A (H1N1) in 2009, the proportion of viral infections had decreased. Whatever, admission to

Page 39 of 607

the ICU for SARI remains associated with high morbidity and mortality rates References 1. Rello J, and al. H1N1 SEMICYUC Working Group (2009) Intensive care adult patients with severe respiratory failure caused by Influenza A (H1N1) in Spain. Crit Care 13:R148 2. Sakr Y, Ferrer R, Reinhart K, Beale R, Rhodes A, Moreno R, et al. The Intensive Care Global Study on Severe Acute Respiratory Infection (ICGLOSSARI): a multicenter, multinational, 14-day inception cohort study. Intensive care medicine. 2016;42(5):817-28.

A72 Concurrent epidemics of influenza and aspergillosis in Taiwan, 2016 K.S. Chan1, W.L. Yu1,2 1 Chi Mei Medical Center, Tainan, Taiwan, Province of China; 2Taipei Medical University, Taipei, Taiwan, Province of China Correspondence: K.S. Chan – Chi Mei Medical Center, Tainan, Taiwan, Province of China Intensive Care Medicine Experimental 2016, 4(Suppl 1):A72 Introduction: The epidemic flu in 2016, predominantly influenza A (H1N1), is causing chaotic situations to the health care facilities in Taiwan, including emergency department overcrowding, difficult for patients to find beds in emergency rooms or in ICUs, and shortage of respiratory ventilators as well as extracorporeal membrane oxygenation (ECMO) machines. Death toll continues to rise and is record-breaking. The reasons why so many patients suffered from such a severe influenza that required intensive care is unclear. Some studies have emphasized the importance of Aspergillus infection during severe influenza attack. Objectives: To show if there is any correlation between the activity of aspergillosis and influenza infection Methods: Confirmed influenza case was defined at least one positive assay for testing influenza included influenza A and B rapid antigens, real time polymerase chain reaction (PCR), viral isolation and identification for specimens of nasopharyngeal swab and/or lower respiratory tract aspirates. The detection of Aspergillus galactomannan (GM) antigen by Platelia Aspergillus Ag assay in human serum and bronchoalveolar lavage fluid was used as an aid in the assessment of patients with suspected aspergillosis. A GM index ≥ 0.5 was defined positive for Aspergillus antigenemia. Decision of testing Aspergillus antigen was made by attending physicians, mostly for patients with worsening pneumonia under oseltamivir and/or routine antibiotic therapy. Results: A total of 1,640 hospitalized medical ICU patients were identified during the observed 7-month period, including 32 (1.95 %) patients had a diagnosis of severe influenza and 24 (1.46 %) patients were positive for Aspergillus GM antigen test. Meanwhile, 12 (37.5 %) patients of severe influenza with worsening pneumonia had Aspergillus antigenemia. However, only 13 influenza patients were tested for Aspergillus GM antigen. The physician-dependent patient selection bias resulted in 92.3 % accuracy. The incidence of patients with influenza and/or positive Aspergillus antigenemia in medical ICU patients significantly increased in February 2016 in comparison to previous 6 months in the hospital. Conclusions: Medical imaging and serum galactomannan antigen currently constitute the basis of the screening approach for invasive pulmonary aspergillosis. The presence of Aspergillus antigenemia might not represent true invasive Aspergillus infection, but a prompt diagnostic workup should be encouraged as most of these patients suffering life-threatening ARDS. The clinical impact of invasive aspergillosis on severe influenza may be more important than previous thought. In conclusion, our study revealed a possible link between invasive aspergillosis and severe influenza patients. Recognition of this potential might help physicians to initiate a prompt diagnostic workup incorporated into strategies of preparedness for the epidemic crisis.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

A73 Prognostic value of nosocomial influenza A (H1N1)pdm09 infection J. Marin-Corral1, C. Vilà1, J.R. Masclans1,2,3, J. Nolla2,4, L. Vidaur5, J. Bonastre6, B. Suberbiola7, J.E. Guerrero8, A. Rodriguez3,9, H1N1 SEMICYUC/GETGAG working group 1 Hospital Parc de Salut Mar - GREPAC, IMIM, Critical Care Department, Barcelona, Spain; 2Pompeu Fabra University (UPF) - CEXS, Barcelona, Spain; 3CIBERES. Instituto de Salud Carlos III, Madrid, Spain; 4Hospital Parc de Salut Mar - GREPAC, IMIM, Barcelona, Spain; 5Hospital de Donosti, Critical care Department, San Sebastian, Spain; 6Hospital la Fe, Critical care Department, Valencia, Spain; 7Hospital de Valdecillas, Critical care Department, Santander, Spain; 8Hospital Gregorio Marañon, Critical care Department, Madrid, Spain; 9University Joan XXIII Hospital - IISPV - URV, Critical care Department, Tarragona, Spain Correspondence: J. Marin-Corral – Hospital Parc de Salut Mar GREPAC, IMIM, Critical Care Department, Barcelona, Spain Intensive Care Medicine Experimental 2016, 4(Suppl 1):A73 Objectives: To assess the prognostic implications of hospital acquisition of influenza A (H1N1) pdm09 virus in a population of critically ill patients admitted to the ICU. Methods: prospective, observational and multi-center study in 148 Spanish ICU. Data were obtained from the GETGAG / SEMICYUC (20092015) registry. All patients infected by influenza A (H1N1) confirmed by RT-PCR were included. Patients were classified into two groups depending on the day of diagnosis of flu: community-acquired (in the first 48 h of hospital admission) and nosocomial infection (after the 7th day of admission and without treatment with oseltamivir). Demographic and temporary variables, comorbidities, severity on admission, treatment and mortality were evaluated. Differences between groups were assessed using chi-square for categorical variables and Student's t-test or Mann-Whitney test for continuous variables. Significant variables in the univariate analysis were included in a multivariate model (logistic regression). We considered p < 0.05 to be significant. Results: 2421 patients with influenza A (H1N1) were included of which 2035 (84 %) were evaluables. 1103 (54.2 %) were classified as influenza A community-acquired infection and 224 (11.0 %) as nosocomial infection. 708 (34.8 %) could'nt be classified in either groups of patients. Patients with nosocomial pneumonia presented: older age (53.47 ± 15.15 vs 48.86 ± 15.33; p < 0.000), higher rates of severity on admission (APACHE II 17 (7) vs 15 (7); p = 0.005 and SOFA 7 (3) vs 5 (3); p = 0.004), more immunosuppression and hematological diseases (20.5 % vs 9.1 %, p = 0.000; 12.5 % vs 5.5 % p = 0.000), more requirements of invasive mechanical ventilatilation (82.7 % vs 66.1 %, p < 0.001) and more days of mechanical ventilation (12 [5-20] vs 8 [315], p = 0.000). ICU stay (8 (4-17) vs 12 (5-22), p = 0.001), hospital stay (14 (8-25) vs 20 (12-30), p = 0.000) and mortality (18.8 % vs 39.2 %, p = 0.000) also increased significantly in the group of nosocomial infection. Furthermore, the hospital acquisition of influenza A H1N1 (OR = 1.63, 95%IC 1.33-1.99, p = 0.000), the APACHEII (OR = 1.08, 95%IC 1.06-1.11, p = 0.000), the hematological diseases (OR = 3.19, 95%IC 1.77-5.73, p = 0.000) and the need of renal replacement (OR = 4.20, 95%IC 2.60-6.77, p = 0.000) and mechanical ventilation (OR = 4.34, 95%IC 2.62-7.20, p = 0.000) were independently variables associated with mortality in the multivariate analysis. Conclusions: Patients with nosocomial influenza A (H1N1) pdm09 admitted in the ICU are more seriously ill requiring more resources consumption. The hospital acquisition of influenza A (H1N1) pdm09 in critically ill patients who need to be admitted in the ICU is associated with increased mortality.

A74 Descriptive analysis of influenza affected patients admitted in ICU from 2010 to 2016 N. Ramon Coll, G. Jiménez Jiménez, S. Carvalho Brugger, J. Codina Calero, B. Balsera Garrido, M. García, M. Palomar Martínez, M. Vallverdú Vidal Hospital Universitari Arnau de Vilanova, Lleida, Spain Correspondence: S. Carvalho Brugger – Hospital Universitari Arnau de Vilanova, Lleida, Spain Intensive Care Medicine Experimental 2016, 4(Suppl 1):A74

Page 40 of 607

Introduction: In 2009 was described the first pandemic of XXI century caused by influenza virus A (H1N1). This virus did not affect only patients in extreme ages of life and significant comorbidity, it also affected young and immunocompetent patients. The antigenic variations these viruses suffer, cause seasonal outbreaks of different extension. Objectives: Describe the characteristics of patients diagnosed by influenza who required admission to an intensive care unit (ICU). METHODOLOGY: Descriptive analysis including all patients diagnosed with influenza admitted from October 2010 to April 2016 in the ICU of a referral hospital for a population of 450,000 inhabitants. The diagnosis was obtained by PCR on nasal or pharyngeal swabs, sputum or tracheal aspirate. Epidemiological variables, risk factors, severity on admission, administered treatments, life support therapies and mortality were collected. Results: During this period, 52 patients with influenza were admitted with an average of 8.67 per year (0-22). The mean age was 55.38 years (16-86) and 65.38 % were men. Mean APACHE II was 17.56 and mean SOFA was 6.15. Only 16 patients (30.76 %) had been vaccinated. Influenza A was identified in 94.23 % cases (78.85 % influenza A H1N1). Comorbidities presented by the patients were: COPD 14 (26.92 %), hematologic diseases 10 (19.23 %), diabetes mellitus 10 (19.23 %), obesity 9 (17.3 %) and pregnancy 4 (7.69 %). Radiological affectation was observed in 40 patients (76.92 %) at admission. 32 patients required vasoactive drugs, 7 required renal replacement and 4 patients were moved into the prone position (applied in our center since 2013). 94.23 % patients received oseltamivir for an average of 7.7 days (5.98 days on average between onset of symptoms and first dose). 94.23 % received empirical antibiotic therapy. There was respiratory coinfection in 18 cases (44.44 % S. pneumoniae). Ventilator-associated pneumonia was not observed in any patient. In 48 patients (92.3 %) was required mechanical ventilation (MV). Initially, 37 patients received non-invasive mechanical ventilation (NIMV) and 11 invasive mechanical ventilation (IMV). NIMV failure occurred in 21 patients (59.96 %) and 20 of them were connected to IMV. Only 4 patients did not require ventilatory support. The mean ICU stay was 12 days. A total of 11 patients (22.92 %) died during their hospital stay, all of them in the ICU. In this group, the mean APACHE II was 18.1 and mean SOFA 7.5, and the days between onset of symptoms and the administration of the first dose of oseltamivir were, on average, 8.72 Conclusions: Influenza A H1N1 was identified in most of the patients. The non-survivor subgroup presented a higher rate of immunosuppression, an increased severity at admission and a longer period of time between onset of symptoms and the administration of the first dose of oseltamivir.

A75 IgG2 as an independent risk factor for mortality in patients with community-acquired pneumonia M.C. de la Torre1, E. Vendrell1, E. Palomera2, E. Güell1, J.C. Yébenes1, M. Serra-Prat2, J.F. Bermejo-Martín3, J. Almirall1 1 Consorci Sanitari del Maresme, Intensive Care Unit, Mataró, Spain; 2 Consorci Sanitari del Maresme, Investigation Unit, Mataró, Spain; 3 Hospital Clínico Universitario de Valladolid, Unidad de Investigación Médica en Infección e Immunidad, Valladolid, Spain Correspondence: M.C. de la Torre – Consorci Sanitari del Maresme, Intensive Care Unit, Mataró, Spain Intensive Care Medicine Experimental 2016, 4(Suppl 1):A75 Introduction: Mortality in patients with community-acquired pneumonia (CAP) remains high despite improvements in treatment. OBJECTIVE. To determine immunoglobulin levels in patients with CAP and impact on disease severity and mortality Methods: Observational study. Hospitalized patients with CAP were followed up 30-days. Levels of immunoglobulin G (IgG) and subclasses, immunoglobulin A (IgA) and immunoglobulin M (IgM) were measured in serum on the first 24 hours of CAP diagnosis.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Page 41 of 607

Results: 362 patients with CAP — 172 ward-treated and 190 ICUtreated — were enrolled. ICU-treated patients had significantly lower values of IgG1, IgG2, IgG3 subclasses and IgA than ward-treated patients. 38 patients died before 30 days. Levels of IgG2 were significantly lower in non-survivors than survivors (p = .004) and a level of IgG2 < 301 mg/dL was associated with poorer survival according to both the bivariate (HR 4.47; p < .001) and multivariate (HR 3.48; p = .003) analyses. Conclusions: Patients with CAP with IgG2 levels < 301 mg/dL had a poorer prognosis and a higher risk of death. Our study suggests the utility of IgG2 to predict the evolution of CAP and increase support measures or additional treatment.

A76 VAP: incidence, risk factors and the antimicrobial resistance pattern in an Angolan ICU E. Tomas1, A. Escoval2, F. Froe2, M.H. Vitoria Pereira3, N. Velez4, E. Viegas4, E. Filipe1 1 Clinica Sagrada Esperança, ICU, Luanda, Angola; 2Escola Nacional de Saude Publica/ UNL, Lisboa, Portugal; 3Clinica Sagrada Esperança, Infectiology, Luanda, Angola; 4Clinica Sagrada Esperança, Luanda, Angola Correspondence: E. Tomas – Clinica Sagrada Esperança, ICU, Luanda, Angola Intensive Care Medicine Experimental 2016, 4(Suppl 1):A76 Introduction: Ventilator-Associated Pneumonia (VAP) is one of the most frequent nosocomial infections in ICU and results in increased mortality, prolonged hospital stay and greater healthcare costs. Objectives: To analyze the frequency, risk factors and the antimicrobial resistance pattern of microbiological agents responsible for VAP in patients admitted in the ICU at Clínica Sagrada Esperança, in 2015. Methods: A retrospective cohort of 99 patients mechanically ventilated for ≥48 h. Data was collected from ICU admission to 28 days or death. The VAP diagnosis was defined according to CDC but required microbiological confirmation. We analyzed the association between risk factors and the occurrence of VAP. Results: The mechanical ventilation utilization ratio was 0.67. The incidence of VAP was 14.9 episodes per 1,000 days of mechanical ventilation. 69 % of cases of VAP were late onset VAP. The reintubation OR ( 95 % CI) 10,714 ( 2.320 to 49.490 ) was identified as independent risk factor for VAP. VAP was not associated with attributable mortality in the ICU or 28 days mortality (8%vs.49 %; p = 0.013 )vs.(15%vs.51 %; p = 0.011), but was related to a threefold increase in duration of mechanical ventilation (29.8 ± 13.7vs.10.7 ± 6.5; p = 0.000) and increase in hospital stay (20.6 ± 13.7vs.7.0 ± 4.2; p = 0.000). Gram-negative bacteria were the most common agents in our ICU. Conclusions: To know and close monitoring of the incidence of VAP, microbiological agents associated and its antimicrobial resistance pattern is the key to the adoption of more specific standards ICU rules. References 1- Alp, E., & Voss, A. (2006). Annals of Clinical Microbiology and Antimicrobials, 5 (1), 7. 2- Rosenthal, V. D., Maki, D. G., Mehta, Y., Leblebicioglu, H., Memish, Z. A., Al- Mousa, H. H., … & Apisarnthanarak, A. (2014). American journal of infection control,42(9), 942-956

Fig. 29 (abstract A76) Potential risk factors for VAP

Fig. 30 (abstract A76). Outcome A77 Predictors of positive microbiology using a vap electronic triggering system C. Groves1, M. Reay2 1 University Hospital, UHCW NHS Trust, General Critical Care, Coventry, United Kingdom; 2Dudley Group of Hospitals NHS Foundation Trust, Department of Anaesthetics and Intensive Care, Dudley, United Kingdom Correspondence: C. Groves – University Hospital, UHCW NHS Trust, General Critical Care, Coventry, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A77 Introduction: Ventilator Associated Pneumonia (VAP) is the most common hospital acquired infection in those requiring mechanical ventilation. It is associated with prolonged length of stay on the ICU, poorer outcomes and increased cost1.There is currently no 'gold standard' definition but a high clinical suspicion is often enough to begin treatment. Diagnosis can be complicated as the signs that are seen with VAP are not uncommon in those who are critically ill in ITU, potentially delaying treatment1,2. Objectives: The aim of this project was to establish if there are significant differences between the Triggers which predicted VAPs (VAP

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

triggers) as opposed to triggers which did not predict a VAP (Non VAP triggers) where both groups had triggered an alert. Methods: All data from the Intensive Care Unit electronic recording system (ICIP Phillips) was queried to extract all VAP alerts triggered between June 2009 and March 2012. During this time period there were 28 VAP triggers and 130 non-VAP cases. We looked at acknowledgment time, differences in values and order of the index parameters between VAP triggers and Non VAP triggers to look for differences suggesting early predictors for positive microbiology and radiology. Microsoft Excel ® was used to manipulate data for analysis and SPSS Statistics 17.0 ® was used to calculate summary statistics, generate tables, perform the student's t-test analysis and produce graphs. Student's t-test was used to analyse statistical significance between groups. Results: Of the quantitative variables measured, there was no statistically significant difference between the confirmed VAP cases and the non-VAP cases apart from in temperature with non-VAP cases having a higher average temperature, 38.2 °C, compared with 37.89 °C in confirmed VAP. This notwithstanding we noted a lower mean C-rective protein (CRP), a higher oxygen requirement and a lower White Cell Count (WCC) in confirmed VAP cases. As would be expected those in the VAP group had secretions most commonly described as creamy with moderate to copious amount. In terms of trigger order in the confirmed VAP cases, secretion amount was the most common first and second trigger in the VAP alert followed by nonspecific inflammatory markers such as WBC, CRP and temperature. In the non VAP group, temperature, oxygen requirement (PaO2/FiO2) and WBC were the most common first trigger factors. Conclusions: In conclusion early Secretion quantity and description predict positive microbiology and radiology compared with other parameters such as temperature, WBC, CRP which do not. References 1. Valencia M, Torres L. Ventilator-associated pneumonia. Curr Opin Crit Care Feb 2009; 15: (30-35) 2. Hunter JD, Ventilator Associated Pneumonia, BMJ May 2012; 344: (e3325e3331) Grant acknowledgement No funding to declare.

ECMO FOR MANAGEMENT OF RESPIRATORY FAILURE A78 Driving pressure associated mortality in acute respiratory distress syndrome with extracorporeal membrane oxygenation L.-C. Chiu1, H.-C. Hu1,2,3, C.-Y. Hung1, C.-H. Chang1, S.-H. Li1, K.-C. Kao1,2,3 1 Chang Gung Memorial Hospital, Division of Thoracic Medicine, Taoyuan, Taiwan, Province of China; 2Chang Gung Memorial Hospital, Department of Respiratory Therapy, Taoyuan, Taiwan, Province of China; 3 Chang Gung University College of Medicine, Department of Respiratory Therapy, Taoyuan, Taiwan, Province of China Correspondence: K.-C. Kao – Chang Gung Memorial Hospital, Division of Thoracic Medicine, Taoyuan, Taiwan, Province of China Intensive Care Medicine Experimental 2016, 4(Suppl 1):A78 Introduction: The survival predictors and optimal mechanical ventilator (MV) settings in patients with severe acute respiratory distress syndrome (ARDS) undergoing extracorporeal membrane oxygenation (ECMO) are uncertain. Objectives: To evaluate the influence of clinical variables and MV settings on intensive care unit (ICU) mortality for severe ARDS patients treated with ECMO. Methods: We retrospectively reviewed severe ARDS patients who received ECMO due to refractory hypoxemia between May 2006 and October 2015. Serial MV settings during ECMO and factors associated with ICU mortality were analyzed.

Page 42 of 607

Results: A total of 158 severe ARDS patients received ECMO were analyzed. Overall ICU mortality was 55.1 %. After ECMO initiation, nonsurvivors had significantly higher peak inspiratory pressure and driving pressure than survivors on day 2, day 3, and day 7 (32.8 ± 6.4 vs 30.6 ± 5.2; 32.9 ± 6.8 vs 30.4 ± 6.0; 33.1 ± 7.1 vs 29.9 ± 5.8 cm H2O, p < 0.05; 20.9 ± 6.8 vs 18.3 ± 5.9; 21.3 ± 7.2 vs 17.9 ± 6.8; 21.4 ± 7.5 vs 17.7 ± 6.8 cm H2O, p < 0.05). After multivariate analysis, mean driving pressure above 21 cm H2O during first 3 days on ECMO were independently associated with higher death (odds ratio, 2.968; 95 % confidence interval, 1.312-6.712; p = 0.009). Other factors independently associated with ICU mortality included ARDS duration before ECMO (odds ratio, 1.005; 95 % confidence interval, 1.001-1.010; p = 0.009), and Acute Physiology and Chronic Health Evaluation II score before ECMO initiation (odds ratio, 1.092; 95 % confidence interval, 1.0321.156; p = 0.002). Conclusions: Driving pressure during first 3 days of ECMO support in severe ARDS was independently associated with ICU mortality and level above 21 cm H2O was related to higher death. Further large multicenter, prospective, randomized, controlled trials are warranted to confirm our findings.

A79 Circulating microparticles in patients with severe ARDS undergoing veno-venous ECMO: another piece of the inflammation puzzle? A. Ballin1, F. Facchin1, G. Sartori1, F. Zarantonello1, E. Campello2, C.M. Radu2, S. Rossi1, C. Ori3, P. Simioni2 1 Azienda Ospedaliera di Padova, UOC Anesthesia and Intensive Care Unit, Padua, Italy; 2University of Padua, Thrombotic and Hemorrhagic Diseases Unit, Department of Medicine, Padua, Italy; 3Azienda Ospedaliera di Padova, UOC Anesthesia and Intensive Care Unit, Department of Medicine-DIMED, Padua, Italy Correspondence: A. Ballin – Azienda Ospedaliera di Padova, UOC Anesthesia and Intensive Care Unit, Padua, Italy Intensive Care Medicine Experimental 2016, 4(Suppl 1):A79 Introduction: Acute Respiratory Distress Syndrome (ARDS) is a severe acute inflammatory lung injury with still high mortality, despite current lung protective ventilatory strategy and rescue therapies, such as veno-venous extracorporeal membrane oxygenation (VV-ECMO)1. During ARDS, immune system and inflammatory pathways are strongly activated, resulting in a systemic disease2. Microparticles (MPs), tiny cell-derived vescicles released from a variety of activated or apoptotic cells, play a role in several disease processes3. However, MPs role during vv-ECMO has not been described yet. Objectives: To assess the presence, the origin of circulating microparticles in patients with severe ARDS treated with VV-ECMO, and their relation with the clinical course. Methods: 15 consecutive patients admitted to our ICU for severe ARDS that required VV-ECMO were enrolled. Immediately before (T0), in the 2nd day after placement (T1) and removal (T2) of VV-ECMO, arterial platelet free plasma was collected and analysed by Flow Cytometry to evaluate endothelial (EMPs), platelet-derived (PMPs), leucocyte-derived (LMPs), and tissue factor-bearing (TF-MPs) microparticles levels. Comparisons were drawn between different time points and survivors (n = 8) and non survivors (n = 7) with analysis by two-way ANOVA. P values < 0.05 were considered significantly different. Results: We found a progressive and significant reduction in both circulating EMPs (median 207, IQR 95-404 è median 64, IQR 45-90 MPs/ mL, p = 0.013) and LMPs (median 432, IQR 93-479 èmedian 39, IQR 35-282 MPs/mL, p = 0.044) between the start of ECMO support (T0) and soon after the decannulation (T2), as shown in Fig. 31. Survivors showed higer levels of LMPs (p = 0.034) compared to non survivors and almost significantly higher levels of EMPs (p = 0.062). Standard laboratory tests (i.e. CRP, PCT, and WBC) did not show any significant difference between the same time points, and between different outcomes. There was no correlation between level of any kind of MPs and heparin dose, bleeding episodes, and mortality.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Conclusions: In patients with severe ARDS treated with VV-ECMO, plasma levels of both EMPs and LMPs were significantly elevated just before ECMO support, then gradually reduced during treatment and after decannulation. Moreover, blood passage through the ECMO circuit did not seem to generate more procoagulant, endothelial, leucocyte-derived or platelet-derived MPs, compared to baseline. References 1 Bellani G. JAMA 2016, 315:8, 788-800. 2 Fujishima S. Journal of Intensive Care. 2014, 2:32. 3 McVey M. et al. Am J Physiol Lung Cell Mol Physiol 2012, 303: 364-81.

Fig. 31 (abstract A79). Endothelial and Leucocyte-derived circulating Microparticles (MPs) at T0: just before ECMO, T1: two days after ECMO start, and T2: two days after decannulation

A80 Factors affecting the outcome of extracorporeal membrane oxygenation: a single institution experience N. Umei, I. Shingo Nippon Medical School Hospital, Tokyo, Japan Correspondence: N. Umei – Nippon Medical School Hospital, Tokyo, Japan Intensive Care Medicine Experimental 2016, 4(Suppl 1):A80 Introduction: The use of extracorporeal membrane oxygenation (ECMO) for acute respiratory distress syndrome (ARDS) that is unresponsive to conventional ventilatory management has been increasing rapidly. Therefore, in April 2015, we established a new ECMO program under the Japanese health care system without regionalization. Objectives: This study aimed to determine whether the outcome of our ECMO program is associated with the inter-hospital transfer of patients. Methods: The clinical data of patients treated with ECMO, from April 2015 to March 2016, were collected and analyzed retrospectively. The clinical data of patients who were transferred from other hospitals (transferred group) and those who were not (non-transferred group) were compared before and after the venovenous (VV) ECMO treatment Results: During the study period, 20 patients were treated with VV-ECMO. The median age of the patients was 60 years. The median duration of mechanical ventilation prior to introducing ECMO was 2 days. The median duration of VV-ECMO was 7.2 days, and the total duration of intensive care unit stay was 23 days. Twelve (60 %) patients were successfully weaned from ECMO, and 10 (50 %) patients survived to discharge. However, the median respiratory ECMO survival prediction (RESP) score was -2.5, indicating that the estimated survival rate was low (35 %). Moreover, 70 % of patients were transferred from other hospitals. The results showed that the survival rate was higher for the transferred group than for the non-transferred group (57.1 % vs. 33.3 %, P = 0.6). The proportion of immunocompromised patients in the transferred group was lower than that in the nontransferred group (14.2 % vs. 50.0 %, P = 0.13). Before VV-ECMO treatment, the duration of mechanical ventilation in the transferred group was shorter than that in the non-transferred group (2 vs. 5.5 d, P = 0.047). Moreover, the RESP score for the

Page 43 of 607

transferred group was lower than that for the non-transferred group (-1 vs. -4, P = 0.0076). During the treatment with VV-ECMO, the duration of ECMO support in the transferred group was longer than that in the non-transferred group (8.6 vs. 7.0 d, P = 0.030) Conclusions: The clinical outcomes of our program were inferior to those of advanced ECMO centers in Europe and North America. To improve the outcomes of ECMO for ARDS in Japan, our results suggest that regionalization should be employed such that ARDS patients with better RESP scores are transferred to an ECMO center. References 1. Combes, Brodie D, Bartlett R et al. Position paper for the organization of extracorporeal membrane oxygenationprograms for acute respiratory failure in adult patients. Am J Respir Crit Care Med. 2014 Sep 1;190(5):488-96.

A81 Comparing the prognosis of H1N1-associated acute respiratory distress syndrome with ARDS from other causes treated with ECMO support A.C. Santos, C. Candeias, I. Moniz, R. Marçal, Z. Costa e Silva, J.M. Ribeiro University Hospital of Santa Maria, CHLN, Intensive Care Department, Lisbon, Portugal Correspondence: A.C. Santos – University Hospital of Santa Maria, CHLN, Intensive Care Department, Lisbon, Portugal Intensive Care Medicine Experimental 2016, 4(Suppl 1):A81 Introduction: Influenza A (H1N1)pdm09 pandemic was a determinant event for development of modern extracorporeal life support techniques. It still represents a frequent cause of conventional respiratory support failure, usually presenting with rapidly evolving critical hypoxemia demanding extracorporeal oxygenation (ECMO) rescue treatment. Objectives: Identify characteristics of patients with the most severe forms of H1N1-associated ARDS in order to promote early identification of patients with putative need for ECMO treatment. Comparison of the ECMO H1N1-infected population with the ECMO treated population with ARDS from other causes in terms of physiological, functional and biomechanical ventilation parameters to rule prognosis prediction. Methods: Retrospective review of prospectively collected data from a protocol-driven ECMO referral centre, with inter-hospital patient rescue capability and ELSO registered activity. Results: Between 2011 and 2015, sixty one patients were treated with ECMO support for severe acute respiratory distress syndrome (Quadrox HLS or PLS Oxygenator System, Maquet). There were 25 patients with H1N1-associated ARDS and 34 patients with ARDS from other causes (two patients excluded because of insufficient data). H1N1-patients, when compared with non H1N1-patients, had similar age (50.0 ± 12.1 vs 46.0 ± 11.3; p > 0.05), SAPS II (41.8 ± 14.4 vs 39.3 ± 14.5; p > 0.05), pre-ECMO mechanical ventilation duration (4.2 ± 2.0 vs 6.4 ± 5.0 days; p > 0.05), hypoxemia degree (65.1 ± 17.9 vs 63.1 ± 9.0 mmHg; p > 0.05), static lung compliance at day 1 (22.3 ± 8.3 vs 23.6 ± 9.8 ml/cmH2O; p > 0.05) and average ECMO run duration (14.8 ± 9.7 vs 12.8 ± 9.4 days; p > 0.05). Patients with H1N1-ARDS had higher body mass index (32.2 ± 7.3 vs 29.6 ± 4.5; p < 0.05), higher Murray scores from more diffuse lung involvement (3.56 ± 0.30 vs 3.28 ± 0.34; p < 0.05), lower levels of PaCO2 (56.6 ± 18.3 vs 67.0 ± 20.2 mmHg; p < 0.05) and higher levels of pH (7.33 ± 0.8 vs 7.26 ± 0,11; p < 0.05) translating less severe alveolar ventilation compromise. Global survival was 73 % with no difference between the groups, and prognosis prediction scores were also similar (ECMOnet: 4.0 ± 1.6 vs 4.7 ± 2.1; p > 0,05; LIPS: 7.5 ± 1.8 vs 6.3 ± 2.4; p > 0,05). Conclusions: Patients with H1N1-associated severe ARDS presented with predominantly hypoxemic respiratory failure, with more diffuse bilateral lung disease, higher Murray scores and less effective ventilation compromise, but those characteristics did not result in worse outcomes when compared with patients with severe ARDS from other causes.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Page 44 of 607

A82 In moderate to severe ARDS patients with severe respiratory acidosis, can we improve the arterial pH and make ultra-protective ventilation with the introduction of an extra-corporeal circulation CO2 removal (ECCO2r) technique J.F. Georger, J.P. Ponthus, M. Tchir, V. Amilien, M. Ayoub, E. Barsam Centre Hospitalier Intercommunal de Villeneuve Saint Georges, Lucie et Raymond AUBRAC, Reanimation Polyvalente - Surveillance Continue, Villeneuve Saint Georges, France Correspondence: J.F. Georger – Centre Hospitalier Intercommunal de Villeneuve Saint Georges, Lucie et Raymond AUBRAC, Reanimation Polyvalente - Surveillance Continue, Villeneuve Saint Georges, France Intensive Care Medicine Experimental 2016, 4(Suppl 1):A82

A83 Restrictive transfusion strategy in VV-ECMO for ARDS: ISMETT experience 2011-2015 G. Martucci1, G. Panarello1, F. Tuzzolino2, G. Capitanio1, V. Ferrazza1, T. Carollo1, L. Giovanni1, A. Arcadipane1 1 IRCCS-ISMETT Mediterranean Institute for Transplantation and Advanced Therapies, Department of Anesthesia and Intensive Care, Palermo, Italy; 2 IRCCS-ISMETT Mediterranean Institute for Transplantation and Advanced Therapies, Statistics, Research Office, Palermo, Italy Correspondence: G. Martucci – IRCCS-ISMETT Mediterranean Institute for Transplantation and Advanced Therapies, Department of Anesthesia and Intensive Care, Palermo, Italy Intensive Care Medicine Experimental 2016, 4(Suppl 1):A83

Objective: In patients with moderate to severe ARDS and respiratory acidosis we can introduce ECCO2r to enable protective ventilation or ultra-protective ventilation. We don't know the results we can obtain after the introduction of ECCO2R in a population of ARDS with respiratory acidosis.The objective of this study is to describe if we can at the same time improved the blood pH and allow ultra-protective ventilation. Methods: We retrospectively included patients who received ECCO2r for ARDS with respiratory acidosis between august 2014 and march 2016 in our ICU. The ECCO2R was performed with ILAACTIVE® device (Novalung®) with a Minilung® or ILA® membrane. The sweep gas was oxygen at 10 l/min. The vascular access was a 24 F dual light catheter in femoral position. The blood flow in the membrane was around 1.5 l/min. The ECCO2R was introduced in patients with PaO2/fiO2 ratio between 80 and 150 and acidosis. All the patients was ventilated in controled ventilation with 6 ml/kg (PBW) of tidal volume (Vt) and a respiratory rate (RR) above 25/min. All the patient was sedated with midazolam and Sufentanyl and if necessary we used neuromuscular blocking agent. If is necessary the clinician in charge of the patient performed prone position session during 16H.We collect pH, PaCO2, the PaO2/FiO2ratio, Vt, RR, PEEPtotal and driving pressure, before ECCO2r and at 4 h the initiation of ECCO2r then at J1, J2, J3. We compared the parameters with repeated measures ANOVA test. Results: We included 16 patients, 9 males and 7 females. The average for the age was 67 years (36-84), for the BMI was 32.2 (22-60). The cause of the ARDS was a pneumonia for 14 patients, a cellulitis for 1 patient and a septicemia for the last patient. On eleven patients, we performed at least one prone position session, during 16 hours, with ECCO2R. The average duration of treatment by ECCO2R was 11 days (2-28).The mortality was 50 %, none patient died by a complication of the ECCO2R. We didn't have any hemorrhage complication on the catheter for ECCO2R. The evolution of parameters was in the Table 32. Conclusion: The ECCO2r for ARDS patients with respiratory acidosis helps to normalize the pH and decreased Vt and RR to make ultraprotective ventilation and decrease the driving pressure. In this type of ARDS with severe hypercapnia, others studies are necessary to know if this kind of procedure can improve the mortality of this patients.

Introduction: In critically ill patients evidence suggests a conservative transfusion strategy. Some paper questioned the standard use of Hb 7 mg/dl trigger considering comorbidities and clinical picture. [1] Thresholds for transfusion of PRBC in ECMO are still a matter of debate, because PRBC given to increase DO2 act on a critical point of patients with ARDS and consumption coagulopathy and bleeding are frequent in ECMO. ELSO guidelines suggest to keep Hb at a normal value (1214 g/dl) and frequently 2-3 PRBC units are transfused daily, but also in this setting have been reported lower Hb trigger for transfusion. [2] OBJECTIVE. Describe transfusion practice in VV-ECMO for severe ARDS (Berlin definition) at ISMETT (2011-2015) Methods: Our blood management strategy is based on:

Table 32 (abstract A82). Evolution of the parameters before ECCO2R

H4

J1

J2

pH

7.20 (7.14-7.25)

7.33±(7.29-7.37)

7,33± (7.28-7.38)

7.37± (7.32-7.41)

7.35± (7.32-7.38)

PaCO2 (mmHg)

67 (59-75)

44± (39-49)

47± (39-54)

46± (40-52)

51± (43-59)

Hb level 8-10 according to SvO2, metabolic and perfusion data transfusion using antileukocyte filters low dose anticoagulation by heparin (aPTT range 40-50) anticoagulation stop in case of bleeding autotransfusion of the blood in the circuit at decannulation if possible.

Retrospective observation analysis from electronic medical charts. Data management performed by STATA 13.1 Results: Results are reported as mean value +/- standard deviation or total number and percentage. In case of high variability median and interquartile range are reported and marked as *. During the selected period we run 59 ECMO for severe ARDS due to: H1N1 (28), Bacterial Pneumonia 15), Politrauma (8), Post-pneumonectomy (2), Pneumocystis Pneumonia (2), Lung graft failure and pneumonia (2), Complicated pleural empyema (1), Chemical pneumonia (1). Main results are 78 % of ECMO weaning and 71.2 % of ICU discharge. Conclusion: Transfusion practice at or institution has not a definite trigger, and since several years we abandoned the suggested Hb value tacking more into account functional data. This is possible thanks to high blood flow reached and high efficiency of oxygenators. Decision is left to the care team considering actual DO2, SvO2, hemodynamic parameters, lactates, comorbidities, sepsis, circuit function. We reach a 78 % weaning from ECMO (2016 ELSO registry 66 %) with main result of 5 patients without transfusion and a median rate of PRBC transfusion of 125 ml/day of ECMO support that is higher than reports that use lower Hb trigger, but probably are quite restrictive considering the severity of patients supported. Related to the transfusion strategy we report a 18 % of bleeding needing a treatment (stop anticoagulation or surgical/endoscopic) and 8.5 % of patients needing FFP and 23 % needing PLT.

J3

PaO2/fiO2

111 (91-132)

126 (103-149)

161 (136-186)

134 (110-157)

151 (123-179)

Vt (ml/kg (PBW))

6.1 (5.5-6.7)

4.4± (3.5-5.2)

4.4± (3.6-5.2)

4.2± (3.1-5.2)

4.3± (3.5-5.1)

RR (/min)

30.6 (28.3-32.8)

25.5 (21.1-29.9)

24.9± (21.4-28.5)

23.9± (20.3-27.5)

23.4± (19.7-27.2)

PEEP total (cmH2O)

14.8 (12.8-16.7)

15.4 (13.0-17.7)

15.2 (13.1-17.4)

13.9 (11-16.1)

14.9 (12,4-17,3)

Driving pressure(cmH2O)

11.9 (9.4-14.3)

8.6± (6.2-11.1)

8.6± (6.3-10.9)

9.6± (7.7-11.6)

8.8± (6.8-10.8)

(± mean that p< 0.05 compared to before ECCO2R)

1. 2. 3. 4. 5.

References 1. Lelubre C, et al. Red blood transfusion strategies in critically ill patients: lessons from recent randomized clinical studies. Minerva Anestesiol 2016; Jan Epub 2. Agerstrand CL, et al. Blood conservation in extracorporeal membrane oxygenation for acute respiratory distress syndrome. Ann Thorac Surg 2015;99(2):590-5

Table 33 (abstract A83). Demographics N.

Male N (%)

Age years

BMI

SAPS II

Sofa

LOS preECMO

Mech. Vent. preECMO

PaO2/ FiO2

Murray score

59

49 (83)

43±12

29±5

42±12

9±3

5 (6)*

4 (7)*

63±17

3.5±0.25

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Page 45 of 607

Table 34 (abstract A83). Transfusion parameters Mean HTC

ECMO Days

PRBC total ml/pt

PRBC Day ml/pt/day

No PRBC N. (%)

FFP N. (%)

PLT N. (%)

Bleeding N. (%)

30±3

13 (14)*

1500(3100)*

125(141)*

5 (8.5)

5 (8.5)

14(23)

11(18)

Table 35 (abstract A83). Predictive score and outcomes Preserve score

ECMOnet score

RESPSCORE

ECMO survivors N. (%)

ICU discharge N. (%)

CRRT N. (%)

Mech.Vent. post ECMO

LOS ICU post ECMO

LOS TOT Hospital

4±2

5.8±1.5

1(6)*

46(78)

42 (71.2)

32 (54)

10 (12)*

14.5 (12)*

30 (24)*

A84 Prealbumin level and body mass index before ECMO initiation as prognostic factors in patients with lung transplantation M. López Sánchez1, M.A. González-Gay2, F.J. Llorca Díaz3, M.I. Rubio López1 1 Hospital Universitario Marqués de Valdecilla, Intensive Care Unit, Santander, Spain; 2Hospital Universitario Marqués de Valdecilla, Rheumathology Service, Santander, Spain; 3Universidad de Cantabria, Preventive Medicine and Public Health, Santander, Spain Correspondence: M. López Sánchez – Hospital Universitario Marqués de Valdecilla, Intensive Care Unit, Santander, Spain Intensive Care Medicine Experimental 2016, 4(Suppl 1):A84 Introduction: Indications for extracorporeal membrane oxygenation (ECMO) use in lung transplantation (LT) are: bridge to transplantation, intraoperative extracorporeal respiratory and/or circulatory support and treatment of primary graft dysfunction (PGD) in postoperative period. Several mortality risk factors preECMO initiation has been investigated. Prealbumin and body mass index (BMI) are predictors of mortality in critically ill patients. Objectives: To estimate if nutritional state at the beginning of ECMO is associated with mortality. Methods: Retrospective observational study between January 2009 March 2016, in 12 beds intensive care unit (ICU) of a tertiary hospital center. Inclusion criteria for ECMO entry: patients listed for LT as a bridge to transplant, intraoperative extracorporeal respiratory and/or circulatory support and PGD. ECMO systems: centrifugal pump and polymethylpentene membrane oxygenation with Bioline® coated circuits and cannulas. Demographic data (included BMI), level of prealbumin and time of mechanical ventilation (MV) were collected at the time of ECMO initiation. ECMO indications, type of ECMO support and ICU mortality were also recorded. Prealbumin level and BMI were compared between survivors and non-survivors. Continuous variables, reported as mean ± standard deviation (SD) or median with range were compared using the Student t-test. Categorical variables were compared using the X2 test. Analysis was performed using SPSS ver. 22.0 and value of p ≤ 0.005 was considered to be statistically significant. Results: 33 patients (p), 22 male and 11 females were included with a median age of 49.63 years (SD ± 13.37, interval 16-66. Median simplified Acute Physiology Score (SAPS II) was 38.63 ± 12.97. Pulmonary fibrosis was the most frequent diagnosis (13 p). In 16 p (50 %) MV was started before ECMO support, with a median of 24.27 hours (interval 0-146). Intraoperative support was used in 17 p (2 p respiratory, 10 p cardiac and 5 p both support); bridge to LT in 8 p (4 p respiratory, 3 p cardiac and 1 p both support); PGD was the indication in 8 p. In 4 p ECMO was placed after cardiopulmonary bypass. Type of ECMO support was peripheral venous-arterial in 19 p, central

venous- arterial in 2 p and venous-venous in 12 p. ICU mortality was 42.42 %. Median BMI in survivors was 24.85 ± 5.19 and 24.99 ± 4.16 in nonsurvivors. Median prealbumin level in survivors was 14.09 ± 4.69 and 14.28 ± 10.66 in nonsurvivors. No statistical significance was observed in BMI or prealbumin level between survivors and nonsurvivors. Conclusions: In this review with ECMO patients and LT, there are not statistically significant differences in prealbumin level and BMI between survivors and nonsurvivors at the time ECMO initiation. References 1. Wagner K, Risnes I, Abdelnoor M, Karlsen HM, Svennevig JL. Is It possible predict outcome in pulmonary ECMO? Analysis of pre-operative risk factors. Perfusion 2008;23:95-9.

A85 VV-ECMO in ARDS: toward a lower anticoagulation ratio? E. Zogheib1,2, L. Villeret1, J. Nader3, M. Bernasinski1, P. Besserve1, T. Caus2,3, H. Dupont1,2 1 CHU Amiens - Picardie, Cardio Thoracic and Vascular Intensive Care Unit, Amiens, France; 2Université de Picardie Jules Verne, CURS, Amiens, France; 3CHU Amiens - Picardie, Cardiac Surgery, Amiens, France Correspondence: E. Zogheib – CHU Amiens - Picardie, Cardio Thoracic and Vascular Intensive Care Unit, Amiens, France Intensive Care Medicine Experimental 2016, 4(Suppl 1):A85 Introduction: Venovenous ECMO (VV-ECMO) is a rescue therapy in refractory ARDS. Despite technological progress and enhanced biocompatibility, anticoagulation with unfractionated heparin remains the most often used as recommended. Bleeding complications remains a main cause of increasing morbidity and mortality in patients under VV-ECMO. Objectives: We aimed to study the effectiveness of the dosage of unfractionated heparin (CAW and heparinization, anti-Xa) in patients on VV-ECMO for ARDS. Methods: We performed a retrospective observational study from 2008 to the first quarter 2015. Patients' baseline characteristics, the dose of UFH anticoagulation and biological data of anticoagulation (APTT, anti-Xa, HEMOCHRON ACT, ATIII, fibrinogen) were collected and we studied the incidence of haemorragic or thromboembolic events under UFH anticoagulation. Results: 48 patients with ARDS requiring VV-ECMO were included,with a majority of men (35 (63.6 %) vs 20 (36.4 %) (p < 0.5) The mean age was. 47 years (+/- 15 years). In 35.4 % of cases, a double-lumen cannula (Avalon®) was used (mean diameter of 27GB), while the others had a femorà-jugular cannulation with an average diameter of femoral vein to 25GB. UFH anticoagulant was used in most patients. 33 % of patients did not have UFH during installation of VV-ECMO. The average value of APTT was 35 +/- 8 secondes, the anti-Xa to 0.14 +/- 0.06. The average of the Hemochron was 171 +/- 23 sec. The values of UFH administered averaged 778 +/- 601 IU / h IVSE. Bleeding complications were noted in 56.25 % of cases. None of our patients had ECMO thrombosis cannulas or oxygenator. 39 % of patients had positive anti-PF4 and only 11 patients had a change in anticoagulation. Conclusions: Anticoagulation used in our study appears lower than recommendations without increasing thromboembolic complications, with persistence of a high rate of bleeding complications with or without a direct connection with ECMO. Does these results reinforce the idea of lowering the anticoagulation ratio on patients under VVECMO?

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Page 46 of 607

Note: This abstract has been previously published and is available at [2]. It is included here as a complete record of the abstracts from the conference. References 1. Extracorporeal life support for patients with acute respiratory distress syndrome: report of a Consensus Conference. Ann Intensive Care. 2014 May 24;4:15 2. Zogheib E, Nader J, Villeret L, Guilbart M, Pesserve P, Caus T, Dupont H (2016) VV-ECMO on ARDS: towards a lower coagulation ration? Ann Intensive Care 6(Suppl 1):P10 Grant acknowledgement ECMO 80

A86 Effect of citrate anticoagulation on CO2 extraction during low flow extracorporeal veno-venous CO2 removal therapy P. Morimont1,2, S. Habran2, R. Hubert2, T. Desaive2, F. Blaffart3, N. Janssen2, J. Guiot1, A. Pironet2, P. Dauby2, B. Lambermont1 1 University Hospital of Liège, Medical and Coronary Intensive Care, Liege, Belgium; 2University of Liège, GIGA Research, Liège, Belgium; 3University Hospital of Liège, Perfusion School, Liege, Belgium Correspondence: P. Morimont – University Hospital of Liège, Medical and Coronary Intensive Care, Liege, Belgium Intensive Care Medicine Experimental 2016, 4(Suppl 1):A86 Introduction: Low flow extracorporeal veno-venous CO2 removal therapy (ECCO2RT) in addition to mechanical ventilation is used to remove CO2 while allowing protective ventilation (PV) during ARDS. However, this technique requires anticoagulation that may induce severe bleeding in critically ill patients (1). An alternative method consists in using citrate anticoagulation. Objectives: The aim of this study was to assess the effect of citrate anticoagulation on CO2 extraction during ECCO2RT. Methods: This study was conducted on an experimental model of severe hypercapnic acidosis performed in 2 groups of 3 pigs. In the first group (heparin group), pigs were anticoagulated with a standard protocol of unfractionated heparin while citrate was used for ECCO2RT device anticoagulation in the second group (citrate group). After sedation, analgesia and endotracheal intubation via a cervical tracheostomy, pigs were connected to a volume-cycled ventilator (tidal volume 10 mL/Kg, respiratory rate 20/min, FiO2 1.0, PEEP 5 cmH2O). Severe hypercapnic acidosis was obtained by reducing tidal volume by half. ECCO2RT was started in both groups when arterial pH was lower than 7.2. Pump Assisted Lung Protection (PALP, Maquet, Germany) system was used with two small cannulas to remove CO2. Blood flow in the PALP was successively set at 200, 400, 600 and 0 mL/min, each setting lasting 60 minutes. Sweep gas flow was set at 10 L/ min. CO2 extraction, arterial pH, PaCO2 as well as systemic and pulmonary pressures were continuously followed. Results: CO2 extraction is depicted in graph 1. Mean arterial pH was normalized to 7.37 ± 1.4 at an extracorporeal blood flow of 400 mL/ min, coming from 7.11 ± 1.3. Arterial pH did not significantly changed in the citrate group as compared to the heparin group. Conclusions: Using citrate anticoagulation during ECCO2RT is feasible. A trend toward better CO2 extraction was observed in the citrate group but not statistically significant as compared to the heparin group. References (1) Kluge S et al., Intensive Care Medicine 2012, 38: 1632-1639. Avoiding invasive mechanical ventilation by extracorporeal carbon dioxide removal in patients failing noninvasive ventilation. Grant acknowledgement Leon Fredericq Foundation of the University of Liège, Belgium

Fig. 32 (abstract A86).

A87 Point-of-care rotational thromboelastometry and platelet aggregometry during ECMO for ARDS F. Zarantonello1, A. Ballin1, F. Facchin1, G. Sartori1, E. Campello1, T. Pettenuzzo1, G. Citton1, S. Rossi2, P. Simioni1, C. Ori1 1 University of Padova, Department of Medicine (DIMED), Padova, Italy; 2Azienda Ospedaliera di Padova, Emergency Department, Padova, Italy Correspondence: F. Zarantonello – University of Padova, Department of Medicine (DIMED), Padova, Italy Intensive Care Medicine Experimental 2016, 4(Suppl 1):A87 Introduction: Extracorporeal Membrane Oxygenation (ECMO) is one of the rescue therapies in severe acute respiratory distress syndrome (ARDS)1. The need for anticoagulation and the activation of clotting cascade caused by ECMO circuit put patients at risk of bleeding and thrombosis. Conventional coagulation tests fail to fully describe such a complicated process, but the use of point of care (POC) methods like thromboelastometry and platelet aggregometry may provide additional information on clot development, lysis and platelet activity. However, the role of POC tests during veno-venous (VV) ECMO for ARDS still remains unclear2. Objectives: To assess the association between POC tests and episodes of hemorrhage during ECMO. Methods: Blood samples for rotational thromboelastometry (ROTEM®) and aggregometry (Multiplate®) were collected immediately before starting ECMO, every 2 days for the first 10 days and 48 hours after the removal of the circuit. Prothrombin time (PT), activated partial thromboplastin time (aPTT), platelet count (PLT), anti-Xa activity, plasma fibrinogen levels and activated clotting time (ACT) were recorded. Bleeding episodes were defined as severe when associated with the administration of at least 2 units of packed red blood cells (RBC). Results: We enrolled 15 severe ARDS patients requiring VV-ECMO. ECMO median duration was 9 (interquartile range 8-14.5) days. Mortality was 53.3 %. We found a strong correlation between maximum clot firmness (MCF) and amplitude of the curve at 10 minutes (A10) in all ROTEM tests (Rho > 0.96, p < 0.001). 67 % of patients had at least one episode of bleeding. 30.5 % of the hemorrhages were severe. PLT was lower in bleeding compared to non-bleeding group (p = 0.02). Among the ROTEM tests, heparinase modified thromboelastometry clotting time and clot formation time (HEPTEM CT and CFT) were significantly higher in bleeding group (p < 0.05) (Fig. 33). The POC tests evaluating the intrinsic and extrinsic coagulation pathways, as well as the thrombin receptor activating peptide-6 test (TRAP-AUC) assessing platelet aggregometry, were different between groups, even though not statistically significant (Fig. 34). There was no difference in PT, aPTT, anti-Xa activity, fibrinogen and ACT between groups.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Conclusions: ROTEM A10 is strongly correlated with the MCF, providing a fast coagulation assessment during ECMO. Bleeding episodes are not predictable using conventional laboratory coagulation tests, except for PLT. We found an association only between HEPTEM CT and CFT and hemorrhage. Increasing the sample size might identify other POC tests useful to predict and monitor hemorrhagic and thrombotic complications. References 1. A-NZ ECMO influenza investigators. Extracorporeal Membrane Oxygenation for 2009 Influenza A(H1N1) Acute Respiratory Distress Syndrome. JAMA 2009; 302: 1888-95 2. Bolliger D et al. Point-of-care coagulation management algorithms during ECMO support: are we there yet? Minerva Anestesiol. 2016 Mar 30.

Fig. 33 (abstract A87).

Page 47 of 607

A88 Pilot study of extracorporeal carbon dioxide removal in chronic obstructive pulmonary disease patients with late NIV failure C. Kirakli, O. Ediboglu, S. Ataman, M. Yarici, F. Tuksavul Dr. Suat Seren Chest Diseases and Surgery Training Hospital, Intensive Care Unit, Izmir, Turkey Correspondence: C. Kirakli – Dr. Suat Seren Chest Diseases and Surgery Training Hospital, Intensive Care Unit, Izmir, Turkey Intensive Care Medicine Experimental 2016, 4(Suppl 1):A88 Introduction: Noninvasive ventilation (NIV) is routinely used in hypercapnic patients with an acute exacerbation of COPD. Extracorporeal CO2 removal (ECCO2R) techniques are being evaluated in patients who have NIV failure but the feasibility of these techniques in late NIV failure (after 24 hours) is unknown (1,2). Objectives: We tried to assess the feasibility of a single site ECCO2R system in hypercapnic patients with an acute exacerbation of COPD who had late NIV failure. Methods: Five hypercapnic COPD patients who were admitted to the ICU with an acute exacerbation whose respiratory acidosis did not improve despite 24 hours of NIV were enrolled in this single center, prospective, feasibility trial. A low flow single site ECCO2R system (ALung Technologies, Pittsburgh, Pa) was used. The primary endpoint was the time to normalization of arterial blood gas pH levels (above 7.35). Data are expressed as median (25th-75th percentiles). Results: The pH and PaCO2 levels at ICU admission were 7.24 (7.177.27) and 92 (75-125) mmHg respectively. pH and PaCO2 levels just before the initiation of ECCO2R were 7.29 (7.27-7.33) and 93 (70-104) mmHg respectively. ECCO2R was started after 72 (36-170) hours of NIV treatment. Arterial blood gas pH levels came to normal ranges after 22 (12-43) hours of ECCO2R treatment. In one patient ECCO2R was stopped due to clotting in the circuit. No other complications were observed regarding the use of ECCO2R system. Conclusions: ECCO2R systems can be an alternative in COPD patients who are having late NIV failure. The system seems to be safe due to its low flow characteristics but clotting complication must be kept in mind. Further studies are needed to emphasize these results and also test the impact of this procedure on other outcomes like ventilator free days, duration of ICU stay, cost and mortality. References 1. Burki NK, Mani RK, Herth FJ, Schmidt W, Teschler H, Bonin F, Becker H, Randerath WJ, Stieglitz S, Hagmeyer L, Priegnitz C, Pfeifer M, Blaas SH, Putensen C, Theuerkauf N, Quintel M, Moerer O. A novel extracorporeal CO(2) removal system: results of a pilot study of hypercapnic respiratory failure in patients with COPD. Chest. 2013 Mar;143(3):678-86. doi: 10.1378/chest.12-0228. 2. Kluge S, Braune SA, Engel M, Nierhaus A, Frings D, Ebelt H, Uhrig A, Metschke M, Wegscheider K, Suttorp N, Rousseau S. Avoiding invasive mechanical ventilation by extracorporeal carbon dioxide removal in patients failing noninvasive ventilation. Intensive Care Med. 2012 Oct;38(10):1632-9.

A89 Experience of extracorporeal support using pumped veno-veno extracorporeal carbon dioxide removal in ventilated patients: with severe acute respiratory failure S. Keating1, A. Gibson2, M. Gilles2, M. Dunn2, G. Price2, N. Young2 1 Royal Infirmary of Edinburgh, Ward 118, Edinburgh, United Kingdom; 2 Royal Infirmary of Edinburgh, Edinburgh, United Kingdom Correspondence: N. Young – Royal Infirmary of Edinburgh, Edinburgh, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A89

Fig. 34 (abstract A87).

Introduction: Severe acute respiratory distress syndrome (ARDS) is associated with a mortality of 46.1 %[1]. Currently only low tidal volume ventilation has been proven to reduce mortality in all patients with ARDS[2]. This is frequently difficult to achieve in the presence of hypercapnia. Decreases in ventilatory driving pressure

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

have been strongly associated with ICU survival[3]. It has been hypothesised that ventilation combining reduced tidal volumes with extracorporeal carbon dioxide removal (ECCOR) may result in further improvements in mortality[4], and prospective randomised trials are planned [5]. Objectives: To assess in patients with severe ARDS and acidosis secondary to hypercapnia if by using veno-veno ECCOR there can be a normalisation of pH and reduction of peak airway pressures. Secondary outcomes - survival to hospital discharge and complications of therapy. Methods: Data on ventilatory and arterial blood gas parameters before and during therapy was prospectively collected and entered into the Extracorporeal Life Support Organisation (ELSO) registry in line with national guidance[6]. The Hemolung RAS (A Lung technologies) was used to provide ECCOR. Results: 4 patients received ECCOR using the Hemolung. All patients had an improvement in the degree of acidosis with mean hydrogen ion concentration falling from 72.65 nM (pH 7.14) to 46.59 nM (pH 7.33) over the first 24 hours, with a mean PaCO2 fall from 13.5 kPa to 8.7 kPa. Peak airway pressures dropped from 31.5 cmH2O to 29.1 cmH2O over the same time period. There were no direct complications of therapy. 3 patients survived to hospital discharge. Discussion: Using ECCOR locally in patients with severe ARDS it is possible to lower the hydrogen ion concentration improving the degree of acidosis. This corresponds to a trend in lower peak airway pressures, which in this small case series does not reach statistical significance. There were no complications of the therapy and mortality was shown to be lower then quoted in other ARDS trials [1]. References 1. Bellani et al.: Epidemiology, Patterns of Care, and Mortality for Patients With Acute Respiratory Distress Syndrome in Intensive Care Units in 50 Countries. JAMA. 2016;315(8):788-800 2. ARDS Network: Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. New England Journal of Medicine 2000, 342:1301-1308 3. Amato et al.: Driving Pressure and Survival in the Acute Respiratory Distress Syndrome. New England Journal of Medicine 2015; 372:747-755 4. Bein et al.: Lower tidal volume strategy (3 ml/kg) combined with extracorporeal CO2 removal versus 'conventional' protective ventilation (6 ml/kg) in severe ARDS. Intensive Care Med 2013 39:847-856 5. pRotective vEntilation with veno-venouS lung assisT in respiratory failure http://www.nets.nihr.ac.uk/projects/hta/1314302 (accessed 2/4/16) 6. https://www.nice.org.uk/guidance/ipg428 (accessed 2/4/2016)

A90 Our experience with A-V ECCO2r device (Novalung iLA) in the district general hospital setting P. Remeta1, P. Bishop2 1 PAH Harlow, Intensive Care, Harlow, United Kingdom; 2Colchester NHSFT, Intensive Care, Colchester, United Kingdom Correspondence: P. Remeta – PAH Harlow, Intensive Care, Harlow, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A90 Introduction: We would like to present our experience with use of arterio-venous extracorporeal membrane carbon dioxide removal (ECCO2R) device- Novalung iLA in the District General Hospital setting. Objectives: We analysed our data to review our practice in use of ECCO2R in respiratory failure. Methods: Over the period of 5 years we used arterio-venous ECCO2R device Novalung iLA in 10 patients, with overall 50 % mortality. We analysed retrospectively only 8 patients due to lack of complete documentation in 2 patients (survivors).

Page 48 of 607

Results: From a reviewed sample of 8 patients three patients (37 %) survived. Most common diagnosis in our population was pneumonia (6), others were ARDS post cardiac arrest (1) and refractory asthma (1). We compared our patients in groups of survivors and nonsurvivors. Mean age of survivors was 29 (18-39) vs 70.8 (59-80) of nonsurvivors. Mean APACHE score was 13.6 vs 21. Two of three survivors had single organ failure, one required CRRT for AKI. All patients in non-surviving group had multi-organ failure. Mean time to initiation of ECCO2R support post ITU admission was 39 vs 80 hours (11-75 vs 6.5-174 hours). In all patients mean paCO2 before initiation of ECCO2R was 10.89 kPa, after the first hour 7.59 kPa and after 24 hours 5.9 kPa. No ECCO2R device associated complications in survivors dgroup, one limb ischaemia in non-survivors group. Conclusion: Our analysis showed that the use of ECCO2R device significantly improved CO2 elimination in all our patients, allowed us to use lung protective ventilation strategy. Survival of 37 % in our small group is clearly influenced by selection of our patients. We tried to identify a group of patients who would most likely benefit from this invasive intervention. In our small group of patients were early initiation, lower APACHE score, single organ failure and young age factors predicting good outcome. We continue to audit use of ECCO2R device on our unit and currently using veno-venous ECCOR2 device which eliminates the risk of significant complication of A-V ECCOR2 - limb ischaemia.

A91 Resolving dudes:IABP before CABG or levosimendan after? M.D. Fernández Zamora1, J. Muñoz-Bono1, E. Curiel-Balsera1, E. Aguilar-Alonso2, R. Hinojosa3, A. Gordillo-Brenes4, J.A. ArboledaSánchez5, ARIAM-CARDIAC SURGERY PROJECT AUTHORS 1 Hospital Regional Málaga, Intensive Care, Málaga, Spain; 2Hospital Infanta Margarita, Intensive Care, Cabra, Spain; 3Hospital Virgen del Rocio, Intensive Care, Sevilla, Spain; 4Hospital Puerta del Mar, Intensive Care, Cádiz, Spain; 5Hospital Regional Málaga, Málaga, Spain Correspondence: M.D. Fernández Zamora – Hospital Regional Málaga, Intensive Care, Málaga, Spain Intensive Care Medicine Experimental 2016, 4(Suppl 1):A91 Introduction: Meta-analyses suggest that levosimendan is superior to traditional inotropes, with decreased postoperative morbidity, improved hemodynamic function and decreased myocardial injury but at no time has been compared to the balloon counterpulsation. Methods: Observational prospective admitted patients in ARIAM registry from February2010 to September 2012 in 13 public and private hospitals with CCV. We retrospectively compared the patients who underwent CABG who were treated with IABP before surgery and those who received levosimendan during surgery or immediately after surgery. We analysed clinical and epidemiological data of both groups as well as complications, hospital stays and mortality. We used the Mann Whitney U test and Fischer´s exact test with an alpha maximum of 5 % for the comparison of variables according to necessity. Results: 949 patients were involved CABG in the time period studied. Of these, 17 patients were treated with BCIAO before surgery and 96 received levosimendan. No differences in age or comorbidities between the two groups analysed or extracorporeal circulation time (p > 0.05). However, functional status at the time of surgery was worse in the group BCIAo (NYHA III or IV 43.8 % from 22.2 %) p = 0.011, and the risk score used (EuroSCORE p = 0, 0001), 30 % of patients had no postoperative complications, similar in both groups. There were no differences in hospital stay or mortality in both groups. The group of patients treated with levosimendan developed lower rate of heart failure, renal failure (Creat > 2 mg / dl) and prolonged mechanical ventilation than the group that received BCIAO (p < 0.05). However, the

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Page 49 of 607

levosimendan group required in 4.2 % of patients while it was not necessary in the other group. Conclusions: levosimendan in heart coronary artery bypass surgery decreases postoperative morbidity (heart failure, renal failure or prolonged mechanical ventilation)with an increase in the need for CRRT. No differences were found in mortality or hospital stay. References Levosimendan Versus an Intra-aortic Balloon Pump in Adult Cardiac Surgery Patients. JCTVA 25:1154.(Wessex Cardiothoracic Centre, University of Southampton, Southampton General Hospital, Southampton, United Kingdom). The Calcium Sensitizer Levosimendan Gives Superior Results to Dobutamine in Postoperative Low Cardiac Output Syndrome R L Levina et al Servicios de Cirugía y Recuperación Cardiovascular, Hospital Universitario de la Universidad Abierta InterAmericana y del Hospital Francés, Buenos Aires, Argentina

Fig. 35 (abstract A92).

FLUID BALANCE, THERAPIES AND OUTCOMES IN AKI Table 36 (abstract A92). A92 Association between fluid overload at initiation of renal replacement therapy and outcome in critically ill patients: with acute kidney injury I. Skorniakov1,2, D. Vikulova2, C. Whiteley1, O. Shaikh1, A. Jones1, M. Ostermann3 1 Guy's & St Thomas' NHS Foundation Trust, London, United Kingdom; 2 Sverdlovsk Regional Clinical Hospital 1, Ekaterinburg, Russian Federation, 3Guys and St Thomas' NHS Foundation Trust, London, United Kingdom Correspondence: M. Ostermann – Guys and St Thomas' NHS Foundation Trust, London, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A92 Introduction: Fluid overload is associated with worse outcomes in critically ill patients. Objectives: The aim of this study was to investigate the impact of fluid overload (FO) at initiation of renal replacement therapy (RRT) in critically ill patients with acute kidney injury (AKI). Methods: We performed a retrospective analysis of all patients who were treated with RRT for AKI in the multi-disciplinary Intensive Care Unit (ICU) at a university hospital in London (UK) between 2012 - 2015.Total cumulative fluid balance on day of initiation of RRT was used to describe fluid accumulation and estimated in % of baseline body weight (BW). FO was defined as fluid accumulation greater than 10 % of BW. We collected data related to patient demographics, anthropometrics and SOFA score. Outcomes were hospital mortality and length of stay in ICU. Results: 1129 patients received RRT for AKI of whom 42 % died in hospital. There was a significant difference in cumulative fluid balance at initiation of RRT between hospital survivors and nonsurvivors. (Fig. 35) 108 patients (9.6 %) had FO >10 %. They had a significantly higher hospital mortality (X2 = 5,89; p = 0,015) and longer stay in ICU (19.6 days versus 12.8; p < 0,001) but also higher SOFA score compared to patients with FO ≤10 %. (Table 36) Conclusions: In critically ill patients with AKI, fluid overload at initiation of RRT was associated with a significantly higher mortality and longer stay in ICU. Grant acknowledgement: This research was solely supported by the European Renal Association-European Dialysis and Transplantation Association (ERA-EDTA): short-term Young Fellowship grant.

Patients with FO (n=108)

Patients without FO (n=1018)

p

Days in ICU, mean (SD)

19.6 (15.7)

12.8 (14.9)

5 % BW fluid overload the HR for death was 1.07 (p < 0.001). In contrast, there was no association between fluid accumulation and renal recovery. Conclusion: In our cohort of 864 patients with AKI, 255 developed fluid overload (10 % BW) during the first 5 days in ICU. Fluid overload, even at 5 % BW, was associated with increased mortality after adjusting for risk factors, but fluid overload was not associated with renal recovery. References [1] Hoste EAJ, Bagshaw SM et al. Epidemiology of acute kidney injury in critically ill patients: the multinational AKI-EPI study. Intensive Care Med 2015;41:1411-23. [2] Bellomo R, Kellum J a, Ronco C. Acute kidney injury. Lancet 2012;380:756-66. [3] Rizopoulos D. Joint models for longitudinal and time-to-event data. 1st ed. London: Chapman & Hall/CRC Biostatistics Series; 2012. Grant acknowledgement This study was funded by the Department of Anaesthesiology, Nordsjællands Hospital.

Table 37 (abstract A95). Patient characteristics All patients Fluid Balance > 5% Fluid Balance < 5% (N = 864) BW (N = 461) BW (N = 367) Age, median (IQR), years

68 (59-76)

68 (59-76)

68 (58-77)

Male gender, N(%)

528 (61 %)

267 (58 %)

237 (65 %)

SAPSS II, median (IQR)

50 (39-63)

54 (43-66)

46 (36-59)

KDIGO score ≥ 2, N(%)

405 (47%)

237 (51 %)

155 (42 %)

Mechanical ventilation, N(%)

728 (84 %)

401 (87 %)

300 (82 %)

Renal replacement therapy, N(%)

277 (32 %)

171 (37 %)

97 (26 %)

Vasopressor use, N(%)

593 (69 %)

347 (75 %)

232 (63 %)

Renal recovery, N (%)

514 (60 %)

247 (54 %)

246 (67 %)

Mortality, N(%)

282 (33 %)

192 (42 %)

85 (23 %)

Page 51 of 607

Fig. 37 (abstract A95). Study flowchart

Fig. 38 (abstract A95). FB = fluid balance, MV = mechanical ventilation, RRT = renal replacement therapy

A96 Staging of cardiorenal syndrome for outcome prediction in pediatric acute decompensated heart failure D.J. Gebhard1, J. Price2, C.E. Kennedy3, A. Akcan-Arikan3,4 1 University of Texas Health Science Center San Antonio, Pediatric Critical Care Medicine, San Antonio, United States; 2Baylor College of Medicine, Pediatric Cardiology, Houston, United States; 3Baylor College of Medicine, Pediatric Critical Care Medicine, Houston, United States; 4 Baylor College of Medicine, Pediatric Nephrology, Houston, United States Correspondence: A. Akcan-Arikan – Baylor College of Medicine, Pediatric Critical Care Medicine, Houston, United States Intensive Care Medicine Experimental 2016, 4(Suppl 1):A96 Introduction: Subtle worsening of renal function is associated with adverse outcomes in pediatric patients (pts) with heart failure. Cardiorenal syndrome (CRS) is a recently coined term underscoring the co-existence of cardiac and renal dysfunction and stresses the bidirectional nature of the heart-kidney interactions. While well defined in adult populations, available pediatric data is scarce. Acute kidney injury (AKI) consensus definitions could

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

offer a standardized approach to CRS definition and could help stratify patients. Objectives: To apply three consensus definitions of AKI previously validated in heterogeneous pediatric populations (pediatric Risk, Injury, Failure (pRIFLE), Acute Kidney Injury Network (AKIN), and Kidney Disease Improving Global Outcomes (KDIGO) to assess CRS burden and association with outcomes Methods: Retrospective study of all pts admitted with acute decompensated heart failure (ADHF).Failure of medical therapy was defined as mechanical support, transplantation or death. CRS was defined as admission with ADHF and concomitant AKI according to KDIGO, AKIN or pRIFLE creatinine criteria. Results: 75 pts (47 % male, 5.5 years (IQR 0.8-15.4)) were included. Hospital LOS was 48.9 ± 59.1 days, 33 pts (44 %) had a prior diagnosis of heart failure, 39 (52 %) did not have a baseline creatinine. FMT happened in 25 pts(33 %), mortality was 13 % (10/75). AKIN and KDIGO were identical in diagnosing and staging CRS so only KDIGO data are presented. CRS on admission was present in 63 %(47/75) by pRIFLE vs 27 % (20/75) by KDIGO (p < 0.001). pRIFLE identified 27 additional pts with CRS on admission compared to KDIGO (21 R, 5 I, 1 F). CRS on admission either by pRIFLE or KDIGO did not have good prediction for FMT (AUC 0.58 vs 0.64, respectively), similar to any stage of CRS reached during admission by either definition (AUC 0.58 any pRIFLE stage vs 0.61 any KDIGO stage), whereas peak CRS stage reached during admission had good predictive ability for FMT for both (AUC 0.79 for both). 35 and 24 pts still fulfilled CRS criteria by pRIFLE and KDIGO, respectively, at discharge, Of the 47 pRIFLE and 20 KDIGO admission CRS pts, 25 (53 %) and 15 (75 %) had persistent CRS at discharge. KDIGO stage of CRS on discharge was associated with readmission (p < 0.001) but pRIFLE stage was not (p = 0.25)(AUC 0.63 for KDIGO vs 0.5 for pRIFLE for readmission). Conclusions: CRS was very common in pediatric pts with ADH, persisted at discharge, and renal function at discharge predicted readmission. Peak KDIGO stage predicted composite outcome of death, transplant, or mechanical support; discharge KDIGO stage predicted readmission. KDIGO staging outperformed pRIFLE staging for outcome prediction in pediatric CRS. This finding needs to be validated in a prospective multicenter study.

Page 52 of 607

compression of tubular lumen and of the peritubular microvessels, suggesting an obstructive phenomenon by cellular edema. Histology. After 6 hours of sepsis, animals treated with aggressive fluid infusionshowed less peritubular congestion, preserved mesangial space and lower areas with hyaline degenerations and better preservation of the tubular lumen compared to sepsis without aggressive overhydrating. PAS staining showed a very slight hyaline degeneration, better preservation of tissue architecture, lower occurrence of cellular death, suggesting that the aggressive fluid therapy in the early stage of sepsis minimizes the severity of vascular and tissue injury in the kidney. The live animals treated with the aggressive fluid showed less peritubularmicrovesselscongestion compared to S8, better preservation of the mesangial space, minor hyaline degeneration, and less cell death in deeper regions of the cortexafter a month of recovery. However, the general appearance showing a kidney limitation for venous blood drainage. Clearly, the aggressive-fluid therapy attenuates renal damage compared to S8. The results suggest that animals treated with aggressive fluid therapy, at the early stage of sepsis, have better conditions to respond against new harmful challenges to the kidney. Besides, the recover process after sepsis seems to be partial, justifying the occurrence of the post-sepsis syndrome. Grant acknowledgement FAPESP 2011/20401-4.

A97 Early aggressive fluid therapy and kidney: consequences in sepsis survivors. An experimental study A.M.A. Liberatore1, R.B. Souza2, A.M.C.R.P.F. Martins3, J.C.F. Vieira1, Y.R. Kang1, M.N. Nakamae1, I.H.J. Koh1 1 Federal University of São Paulo, Surgery, São Paulo, Brazil; 2Federal University of São Paulo, Morphology and Genetics, São Paulo, Brazil; 3 Biological Institute of São Paulo, Sao Paulo, Brazil Correspondence: A.M.A. Liberatore – Federal University of São Paulo, Surgery, São Paulo, Brazil Intensive Care Medicine Experimental 2016, 4(Suppl 1):A97 Introduction: Microcirculatory dysfunction is an important triggering event of organ dysfunction in sepsis, and the fluid therapy is essential for improvement of the hemodynamic, but the ideal fluid strategy in sepsis has not been developed yet. Aggressive-fluid therapy is controversial and the consequences for the kidney dysfunction are little known. This study investigated the effect of an early phase aggressive-fluid therapy at the kidney microcirculation and tubular structures. Methods: Adult Wistar rats (200 g) were submitted to sepsis {iv. 2 mL E. coli 108 CFU/mL (S8), DL60 in 26 hours}, or sepsis and Ringer Lactate infusion (30 mL/kg/20 min), 30 minutes after the sepsis challenge. Under general anesthesia, the microcirculation of the renal cortical area was monitored by Sidestream Dark Field Imaging (SDF) video-microscopy, at 6 hours (T6h - n = 3/group) and 30 days (T30d n = 3/group) after sepsis. The tissue samples were evaluated by histology (T6h,T30d) by HE and PAS staining. Results and conclusions: SDF at T6h showed broadly distributed microcirculation and tissue dysfunction at both groups (Fig. 39). The outlining of tubules became blurred by their enlargement, with

Fig. 39 (abstract A97).

A98 Non invasive adjustment of fluid status in critically ill patients on renal replacement therapy. Role of electrical cardiometry K. Hamed, M. M Khaled, R. Aly Soliman, M. Sherif Mokhtar Faculty of Medicine, Cairo University, Critical Care Medicine, Cairo, Egypt Correspondence: K. Hamed – Faculty of Medicine, Cairo University, Critical Care Medicine, Cairo, Egypt Intensive Care Medicine Experimental 2016, 4(Suppl 1):A98 Introduction: Electrical Cardiometry allows measurement of fluid status using thoracic fluid content(TFC),cardiac output,cardiac index,systemic vascular resistance index which could be ideal noninvasive hemodynamic monitoring for patients undergoing hemodialysis(HD. Objectives: Investigating relation between changes in TFC and amount of fluid removal during HD session and to monitor

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

hemodynamic parametersto avoid episodes of hemodynamic compromise during HD session. Methods: Thirty critically ill patients on HD were enrolled.Clinical assessment of volume overload and hemodynamics(BP,MAP,CVP),monitored by Electrical Cardiometry ICON® before HD and all through sessions. Results: Out of studied patients males represented 46.7%n = 14with mean age 48 ± 16 years.There was positive correlation between UFvolume and TFC (r = 0.410, P = 0.025). Out of the 30pts studied 18pts60%were hemodynamically stable vs 12pts40% had hypotension represented non responders group and had lower TFC compared to hemodunamically stable group (26.45kohm-1 vs 37.8kohm-1) P value 0.004 indicating that they were hypovolemic. Out of the 30pts studied 18pts60%weren´t congested vs 12pts40%were remained persistently congested after accomplishing HD session with significantly higher TFC when compared to those get red off congestion (43.14 ± 9.9kohm-1 vs 25.44 ± 5.5kohm-1) P value 0.0001 indicating that they were still hypervolemic. Using analysis of ROC curve TFC at 25.34kohm-1 was significantly predictor of hypotension with P value0.002, AUC83.4 %, sensitivity67%, specificity100%. Also TFC cutoff value predicting persistent congestion was 37.02kohm-1 with P value0.0001, AUC95.8 %, sensitivity83%, specificity100%. Conclusions: Electrical Cardiometry is evolving noninvasive tool for adjusting fluid status of critically ill patient on RRT using thoracic fluid content as indicator of fluid status that could be used to avoid hemodynamic instability and persistent volume overload and congestion during and after HD session. References 1. Palmer B.Fand Henrich W.L,Recent advances in the prevention and management of intradialytic hypotension.J Am Soc Nephrol,2008.19(1): p.8-11 2. Rosner M.Hand Ronco C,Techniques for the assessment of volume status in patients with end stage renal disease.Semin Dial,2014.27(6):p.538-41 3. Malik V,Subramanian A,et al,Correlation of Electric Cardiometry and Continuous Thermodilution Cardiac Output Monitoring Systems.World Journal of Cardiovascular Surgery,2014.4:p.101-108 4. Wynne J.L,Ovadje L.O,et al,Impedance cardiography:a potential monitor for hemodialysis.J Surg Res,2006.133(1):p.55-60 5. Osypka M.J,An introduction to Electrical cardiometry c.a.O. company, Editor 2009 6. De Nicola A and Sucre M.J,Impedance cardiography in the estimation of hemodynamic and fluid status of coma patients during continuous venovenous hemodiafiltration Critical Care,2009.13(1):p.202 7. van de Water J.M,Mount B.E,et al,TFC(thoracic fluid content):a new parameter for assessment of changes in chest fluid volume.Am Surg,2005.71(1):p.81-6

A99 Design of a protocol for the estimation of functional renal reserve in critical care patients G. Seller-Pérez1, D. Arias-Verdú1, E. Llopar-Valdor2, I. De-Diós-Chacón3, G. Quesada-García1, M.E. Herrera-Gutierrez1 1 Complejo Universitario Carlos Haya, Málaga, Spain; 2Hospital Parc Tauli, Sabadell, Spain, 3Hospital Son Llatzer, Palma de Mallorca, Spain Correspondence: D. Arias-Verdú – Complejo Universitario Carlos Haya, Málaga, Spain Intensive Care Medicine Experimental 2016, 4(Suppl 1):A99 Introduction: Acute Kidney Injury (AKI) is believed to carry a good prognosis but recent reports have raised concern about long-term outcome. Assessment of renal functional reserve after AKI can be of aid for ascertaining recovery of kidney function. Objective: To develop a method for the estimation of renal reserve in the critical care setting. Methods: Exploratory study. We selected 8 patients (4 men/women) between 20 to 50 years, without known previous renal disease who did not develop AKI during ICU stay. Patients were not receiving drugs that could interfere with renal function, were stable and already recovered from their initial problem, but still with a bladder catheter, a IV line and a nasogastric tube for enteral feeding. After

Page 53 of 607

administration of a load of 20 gr of proteins by enteral route, a creatinine clearance (CrCl) was calculated for each of the next 6 hours, with a timely collection of urine and a sample of serum creatinine at the end of every period. For the analysis we performed u-MannWhitney and Kruskal-Wallis non-parametric test. Data as mean (mean error standard) or median (percentiles 25-75). Results: Age 41.13 (3.2) years, 50 % men and base CrCl 163 (19.51) mL/min. Median percentage of change between base and maximum CrCl was 123.5 % (78.2-143.2) and because a pick was detected between 3 and four hours after protein load we analysed these two hours together, finding a median change of 80.7 % (69.95-144). Changes between the first and third hours were significant either for absolute values (p 0.023) or % (p 0.031). Hourly changes in CrCl are presented in Fig. 40 Conclusions: A protein load by enteral route is followed by an early rise in creatinine clearance. The profile of this response lets us propose a creatinine clearance 3-4 hours after a load of 20 gr of proteins as a quick and easy way to estimate renal reserve. Our next goal should be to define the profile of response for different kind of patients and degrees of renal dysfunction.

Fig. 40 (abstract A99). A100 Nephrology follow-up of patients with acute kidney injury requiring renal replacement therapy in critical care R. Hafes1, G. Carroll2, P. Doherty2, C. Wright2 1 University of Glasgow, Glasgow, United Kingdom; 2Queen Elizabeth Hospital, Critical Care, Glasgow, United Kingdom Correspondence: R. Hafes – University of Glasgow, Glasgow, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A100 Introduction: Evidence has established that there is strong link with Acute Kidney Injury (AKI) and progression to chronic kidney disease (CKD) and end stage renal failure (ESRF) [1]. After AKI, 9 % have permenant loss of kidney function and 9-13 % are dialysis dependent post hospital discharge[2]. Renal followup is recommended for patients who have had AKI whilst critically ill [3]. Objectives: To assess whether patients that received renal replacement therapy (RRT) in ICU had a nephrology follow up, and determine if there was a need for the service.To assess if there were any difference in mortality between the groups. Methods: We performed a retrospective longitudinal cohort study analysis of all patients that received RRT after an AKI in the Greater Glasgow and Clyde Hospitals over one year. Renal function at 3-6 months pre-admission and post discharge were examined by comparing urea, creatinine and eGFR values. Nephrology follow up status was also investigated. We excluded patients that had renal baselines out with the 3-6 month period.Paired t-test analysis was used to analyze preadmission vs post discharge renal baselines in normal eGFR patients

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Results: We sampled 130 admissions with 68 patients surviving discharge from ICU. 28 patients (41 %) had a nephrology follow-up. The remaining 49 patients (59 %) were not followed up. 39 patients (57 %) had an abnormal eGFR at 3-6 months, and of these 24 (61 %) were followed up. 4 patients (6 %) died in each group . Conclusions: Renal follow up rate after RRT in critical care is low despite 37 % having an abnormal eGFR. An efficient followup service is required. References 1. Uchino S, Kellum JA, Bellomo R, et al. Acute renal failure in critically ill patients: a multinational, multicenter study. JAMA 2005; 294:813-8.3. 2. Acute kidney injury leading to chronic kidney disease and long-term outcomes of acute kidney injury: the best opportunity to mitigate acute kidney injury? Contrib Nephrol 2011; 174:182-90.7. 3. Kirwan CJ, Blunden MJ, Dobbie H, James A, Nedungadi A, Prowle JR. Critically ill patients requiring acute renal replacement therapy are at an increased risk of long-term renal dysfunction, but rarely receive specialist nephrology follow-up. Nephron. 2015; 129(3): 164-70.

A101 Chronic kidney disease by stage secondary to diabetes I.G. Guerra Vera Centro Universitario de La Costa, Puerto Vallarta Jalisco, Mexico Intensive Care Medicine Experimental 2016, 4(Suppl 1):A101 Introduction The risk of chronic kidney disease increases with the time of evolution of type 2 diabetes and chronic metabolic control. In people over 40 years occur a progressive loss of glomerular filtration rate corresponding to 1 mL per year. This is associated with progressive deterioration of renal tissue replacement by fibrous tissue, which involves progressive glomeruroesclerosis, tubulointerstitial fibrosis and nephrosclerosis. Objectives To determine the evolution time of diabetes and prevalence for stages on chronic kidney disease. Material and method: A cross-sectional and descriptive study was done on 150 patients diabetic type 2. There were included all of them that have more than 5 years of evolution on the diabetes type 2, the sample was calculated with the averages formula for finite population and the selection was simple random. Sociodemographic variables and health variables were studied, the stage of chronic kidney disease were estimated by the Cockcroft-Gault equation. The statistic analysis included averages, percentages and confidence intervals. Methods A cross- sectional, descriptive, observational, retrospective study was conducted which included adult patients that were 50 years and older, and had 10 year diabetes mellitus type 2 diagnosis, without complications mentioned in the Nephrology Unit in Hospital General de Mexico, Mexico City. Results Regarding the study population, 57 % are female, the average age was 62.12 years and mean glucose was 165.23 mg / dL. The time evolution of diabetes 2 patients in stage 5 was 20.05 years and in patients with stage 1 was 11.05 years. The average creatinine clearance in stage 2 was 75.10 mL / min and in stage 5 10.33 mL / min. 13 % of the population was in stage 4 and a similar percentage (15 %) in stage 5. The time evolution of stage 1 to stage 5 was 10.10 years and stage 3 to 4, 1.5 years. Conclusions Chronic kidney disease is a public health problem that affects health systems around the world. Today his studio is preferably focused on the population undergoing dialysis treatment in its various forms which lies in stage 5; However, a comprehensive approach to chronic kidney disease in all its stages is necessary to have information about the condition; hence the importance of this study, in which the time evolution of diabetes and population analyzed by stage of chronic kidney disease. The diabetic patient with chronic kidney disease is not flattering; it is distinguished by short evolution times between the stages and high population percentage on stages 4 and 5. Reference(s) Wild S, Roglic G, Green A, Sicree R, King H. Globla prevalence of Diabetes: estimates for the year 2000 and projections for 2030. Diabetes care 2004;27:1047-1053. Eknoyan G, LAmeire N, EckardKU, Kasiske BL, et al. KDIGO 2012, clinical practice guideline for the evaluation and Management of Chronic Kidney Disease. 2013;3.

Page 54 of 607

Fluid balance, therapies and outcomes in aki A102 Outcomes of patients requiring renal replacement therapy in intensive care: a ten year retrospective study M. Ralston, L. Gemmell, A. MacKay, E. Black, C. Wright, R.I. Docking, R. Appleton Queen Elizabeth University Hospital, Anaesthetics and Intensive Care, Glasgow, United Kingdom Correspondence: M. Ralston – Queen Elizabeth University Hospital, Anaesthetics and Intensive Care, Glasgow, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A102 Introduction Approximately 4 % of patients in Intensive Care Units (ICU´s) require renal replacement therapy (RRT) during their admission, though this figure ranges from 1-25 % depending on the particular hospital (1). Previous studies suggest that patients requiring RRT in ICU have a 60 % in-hospital mortality rate, compared with an overall mortality rate of 19 % for patients admitted to ICU. (1.2). Objectives The aim of the study was to compare demographic and physiological characteristics of patients requiring RRT in ICU with those who did not require it, and to investigate the impact of receiving RRT on length of unit stay and mortality on ultimate hospital discharge. Methods The study is a multi-centre retrospective observational cohort study. It uses data from January 1st 2005 to December 31st 2014 from three teaching hospital intensive care units with 18 combined beds and an associated tertiary referral renal service in Glasgow. The data used was collected prospectively from the Wardwatcher service. Demographic details, severity scores, physiological parameters and information on length of unit stay and in-hospital mortality was gathered. Standardised Mortality Ratios (SMRs) were calculated using the APACHE II predicted mortality scores as the denominator. Results Data for a total of 10549 patients was collected, of whom 13.9% (1471) received RRT during their admission. The mean duration of RRT was 5.0 +/- 0.2days. Standardised mortality ratios between the two groups is comparable at 1.04 for the RRT group, and 10.6 for the non RRT group. The results for the two groups are shown below. Conclusions Patients receiving RRT had a mortality rate on ultimate hospital discharge of 52.6 %, roughly in line with previous studies, and nearly twice the rate in those not requiring RRT. (1). These patient also stayed in ICU for over twice as many days as those not requiring RRT. It is interesting to note, however that the SMR is actually marginally lower for those receiving RRT than not. Patients who require RRT at some point during their ICU stay also have significantly more deranged physiology during their first 24hours of admission, as evidenced by the urine output, potassium, urea and creatinine values for the two groups. Reference(s) 1. Uchino S, Kellum JA. Acute renal failure in critically ill patients: a multinational, multicenter study. JAMA:2005:294(7): 813-8. 2. Scottish Intensive Care Society Audit group. Audit of Critical Care in Scotland 2014: reporting on 2013. Edinburgh, Uk: NHS National Services Scotland; 2014.

Table 38 (abstract A102). RRT (n=1471)

No RRT (n=9078)

p value

Age (years)

59.3 ± 0.8

56.1 ± 0.4

15; OR 19.0, p < 0.001). Mortality risk of delirium in the afterperiod compared with the before-period and with CAM-ICU versus ICDSC (after-period) did not differ (before versus afterperiod, OR 0.82, p = 0.079 and ICDSC versus CAM-ICU, OR 1.03, p = 0.831).

Page 80 of 607

Conclusions In this large multicentre prospective study, we confirmed the independent association of ICU delirium with hospital mortality but found significant interaction with APACHE II. The screening instrument used (CAM-ICU versus ICDSC) did not influence the delirium-associated risk of in-hospital death.

Table 55 (abstract A148). Association delirium - hospital mortality (1) Hospital mortality Delirium during ICU admission1

(2) Hospital mortality

1.95***

8.61***

(1.62 - 2.34)

(4.46 - 16.63)

APACHE II score

1.19*** (1.17 - 1.21)

Delirium during ICU * APACHE II (interaction term) Number of patients (n)

0.91*** (0.89 - 0.94) 4033

4017*

1 Any delirium diagnosis in before- and after-period (with or without screening) Coefficients are odds ratios (ORs) and their 95% confidence intervals (95% CI). (1)=crude analysis, (2) =adjusted analyses with the following covariates next to those shown in table: age, hospital (n=6), admission type *** p VAS < 60 mm) and NSCR was measured throughout the experiment. All possible combinations of conditions resulted in twelve 60 second sessions, each followed up by NRS for pain, and ratings of arousal and valence as indicators of experienced anxiety. Results Both NRS and NSCR increased in the pain conditions (NRS: M = 3.95, [SD = 1.78]; NSCR: 0.22 [0.09]), compared to no pain (NRS:0.00 [0.00]; NSCR: 0.09 [0.08]), t(18) = 9.45, p < .001 and t(18) = 8.72, p < .001, respectively. There was no change over time in NRS nor NSCR. Significant positive correlations were found between NSCR and the magnitude of the electrical stimulation, r(18) = .48, p = .046, and also between NRS and the anxiety index (mean of arousal and valence ratings), r(18) = .60, p = .009. In the pain condition, 3 (emotion) x 2 (noise) ANOVAs of NSCR, NRS, and anxiety showed that both NRS and anxiety were sensitive to the experimental manipulations, shown by main effects of both emotion, F(2,34) = 7.54, p = .002 and F(2,34) = 16.89, p < .001 (for NRS and anxiety respectively), and main effects of noise, F(1,17) = 8.67, p = .009 and F(1,17) = 9.78, p = .006 (for NRS and anxiety respectively), with elevated ratings of both pain and anxiety in negative conditions compared to ratings in both positive and neutral conditions. NSCR was not influenced by emotion or noise in the pain condition. However, Helmert's contrasts of emotion state only, showed significantly larger NSCR in emotional conditions compared to the neutral, F(1,17) = 5.21, p = .036. Conclusions Both NRS and NSCR are reliable indicators of pain, and the correspondence between NSCR and actual pain stimulation moreover validates the use of NSCR as a measure of pain in patients. However, NRS is also sensitive to the contextual setting and anxiety, which NSCR is not. A discussion, whether to administer analgesic or ataractic drugs during the perioperative stage when NRS is moderate or higher, is warranted. Reference(s) 1. Günther A, Schandl A, Bernhardsson J, Bjärtå A, Wållgren M, Sundin Ö, Alvarsson J, Bottai M, Martling CR, Sackey P. Pain rather than induced emotions and ICU sound increases skin conductance variability in healthy volunteers. Submitted. Grant acknowledgement None

A159 The pharmacokinetics of propofol in ICU patients undergoing long-term sedation A. Bienert1, P. Smuszkiewicz1, P. Wiczling2, K. Przybylowski1, A. Borsuk2, I. Trojanowska1, J. Matysiak1, Z. Kokot1, M. Paterska1, E. Grzeskowiak1 1 Poznan Unversity of Medical Sciences, Poznan, Poland; 2Medical University of Gdansk, Gdansk, Poland Correspondence: A. Bienert – Poznan Unversity of Medical Sciences, Poznan, Poland Intensive Care Medicine Experimental 2016, 4(Suppl 1):A159 Introduction According to the recent guideline (2013), nonbenzodiazepine drugs, like propofol, are preferred to benzodiazepines in sedation of ICU patients undergoing mechanical ventilation due to decreased duration of mechanical ventilation, shortened ICU stay, lower risk of patients death and decreased costs of treatment. Propofol is a relatively well known drug, nevertheless the influence of various factors connected with patients, like demographics, health

Page 85 of 607

status or co-administered drugs on the propofol pharmacokinetics has not been fully understood. Objectives The aim of our study was to examine the pharmacokinetics of propofol in a heterogeneous group of patients sedated in an ICU. The specific objective was to investigate the influence of different variables monitored and patients' health status descriptors, like SOFA or presence of sepsis, on the PK of propofol. Methods The propofol concentration-time profiles were obtained from 29 patients. All the subjects were evaluated according to APACHE II score and SOFA score, whereas the level of sedation was applied according to modified Ramsay Sedation score to achieve a sedation score of 3-4. Non-linear mixed-effects modelling in NONMEM (Version 7.3.0, Icon Development Solutions, Ellicott City, MD, USA) was used to analyse the observed data. Blood samples for propofol assay were collected from the patients' arteries on every day of the infusion, at the selected time points after its termination. The propofol concentration in the plasma was measured within eight weeks by HPLC method with a fluorescence detector. Nonparametric bootstrap and visual predictive check were conducted to evaluate the adequacy of the induced model to describe the observations. Results Propofol pharmacokinetics was best described with a threecompartment disposition model. A typical value of propofol clearance (1.46 L/min) approximated liver blood flow. The volume of distribution at steady state was high 955.1 L, but consistent with other studies on ICU patients. We were unable to identify any statistically significant covariate relationships between PK parameters and opioid type, SOFA score at admission, APACHE II, predicted death rate, reason for admission to the ICU (sepsis, trauma or surgery), gender, body weight, age, infusion duration and C-reactive protein. Conclusions The population PK model was successfully developed to describe the time course of propofol concentration in ICU patients undergoing prolonged sedation. Despite a very heterogeneous group of patients, consistent PK profiles were observed. Reference(s) Eleveld DJ, Proost JH, Cortínez LI, Absalom AR, Struys MM. General purpose pharmacokinetic model for propofol. Anesth Analg 2014; 118: 1221-37. Gradwohl-Matis I, Mehta S, Dünser MW. What´s new in sedationstrategies? Intensive Care Med. 2015; 41: 1696-9. Grant acknowledgement This project was supported by the grant 2015/17/B/NZ7/03032 founded by Polish National Science Centre.

Assessment of preload and fluid responsiveness A160 Early variations of pulse pressure variation, cardiac cycle efficiency and dicrotic pressure to predict fluid challenge success A. Messina1, E. Bonicolini2, D. Colombo1, G. Moro1, S. Romagnoli2, A.R. De Gaudio2, F. Della Corte3, S.M. Romano2 1 AOU Maggiore della Carità, Novara, Italy; 2Azienda OspedalieroUniversitaria Careggi, Firenze, Italy; 3Azienda Ospedaliero-Universitaria Careggi, Novara, Italy Correspondence: A. Messina – AOU Maggiore della Carità, Novara, Italy Intensive Care Medicine Experimental 2016, 4(Suppl 1):A160 Introduction Fluid challenge (FC) is commonly used to increase cardiac index and oxygen delivery. A Tidal volume (Vt) > 8 ml/kg, which is required to guarantee the efficacy of dynamic index of fluid responsiveness in predicting fluid responsiveness, is nowadays not recommended in operating room (1). The identification of early variation of hemodynamic variables may act as a clinical target or a safety limit to stop infusion. This approach has been successfully used to assess fluid responsiveness in critically ill patients ventilated with low tidal volume (2). We evaluated the early variations after a FC in operating room of cardiac cycle efficacy (CCE) and dicrotic pressure (Pdic) to improve

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

baseline pulse pressure variation (PPV) reliability and predict fluid responsiveness. Methods 40 consecutive adult patients scheduled for elective abdominal surgery and ventilated with Vt 8 ml/kg, unconventional modes of mechanical ventilation. Statistical analysis was performed using SPSS program version 19. Results We included 66 patients, mean age of 47.19 years, female gender 53 % male 47 %; the mean APACHE II and SOFA were 14 and 8 points respectively; more frequent admission diagnosis was sepsis (31.8 %); the most frequent hypoperfusion criteria was lactate >2 mmol/L (33.3 %); We found a positive correlation between dIVC and PPV with r = 0.642(p = < 0.05). After FC there was an increase of cardiac index 0.6196 ± 0.53 l/min/m2, a decrease of lactate 0.68 ± 0.49 mmol/L, an increase urine output of 0.26 ± 0.23 ml/kg/hour, and decrease ΔPCO2 3.37 ± 1.7 mmHg with T 5.214(p = < 0.05), U 4.834(p = 0.009), U 3.057(p = 0.07) and U -5.863(p = < 0.05) respectively. Conclusions There is a positive correlation between dIVC and PPV, in patients with shock which received FC in a conventional manner. Reference(s) 1. Marik PE, Monnet X, Teboul JL. Hemodynamic parameters to guide fluid therapy. Marik et al. Annals of Intensive Care 2011, 1:1 2. Monnet X, Teboul JL. Assessment of volume responsiveness during mechanical ventilation: recent advances. Critical Care 2013, 17:217

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

3. Cecconi M, De Backer D, Antonelli M, Beale R, Bakker J, Hofer C. Consensus on circulatory shock and hemodynamic monitoring. Task force of the European Society of Intensive Care Medicine. Intensive Care Med (2014) 40:1795-1815 Grant acknowledgement This study did not receive any grant from any funding agency.

Page 88 of 607

MAP:0.653, PP: 0.557, PCWP:0.644). In contrast, increase in SV after PEEP challenge was significantly higher in fluid-responders than nonresponders (8.6 ± 7.7 vs. 1.3 ± 12.1 ml, p = 0.023) predicting FR with an AUC of 0.750 (p < 0.001). Conclusion Our data suggest that PEEP-induced changes of CVP, MAP, PP and PCWP in mechanically ventilated patients after cardiac surgery cannot predict fluid responsiveness whereas an increase of SV under PEEP elevation appears to be a good parameter to predict fluid responsiveness.

A164 Fluid administration and mortality in septic shock patients: role of fluid balance and PH and electrolytes alterations A. Palmaccio, A.M. Dell'Anna, D.L. Grieco, F. Torrini, C. Iaquaniello, F. Bongiovanni, M. Antonelli Policlinico Universitario A. Gemelli, Anesthesia and Intensive Care Medicine, Roma, Italy Correspondence: A. Palmaccio – Policlinico Universitario A. Gemelli, Anesthesia and Intensive Care Medicine, Roma, Italy Intensive Care Medicine Experimental 2016, 4(Suppl 1):A164

Fig. 73 (abstract A162). Correlation between distensibility index of inferior vena cava and pulse pressure variation A163 Prediction of fluid responsiveness after cardiac surgery with hemodynamic changes induced by peep-elevation K. Pilarczyk1,2, J. Lubarksi2, D. Wendt2, F. Dusse2,3, J. Günter2, B. Huschens2, E. Demircioglu2, H. Jakob2 1 Imland Klinik Rendsburg, Department for Intensive Care Medicine, Rendsburg, Germany; 2University Hospital Essen, Department for Thoracic and Cardiovascular Surgery, Essen, Germany; 3University Witten/ Herdecke, Medical Centre Cologne-Merheim, Department of Anaesthesiology and Intensive Care Medicine, Cologne, Germany Correspondence: K. Pilarczyk – Imland Klinik Rendsburg, Department for Intensive Care Medicine, Rendsburg, Germany Intensive Care Medicine Experimental 2016, 4(Suppl 1):A163 Introduction In contrast to static parameters, e.g. central venous pressure (CVP), dynamic variables representing cardiorespiratory interactions, e.g. stroke volume variation, allow excellent prediction of fluid responsiveness (FR). A recently published study suggested that positive end-expiratory pressure (PEEP) induced changes of CVP or mean arterial pressure (MAP) might represent a promising tool to evaluate FR in ventilated patients with sepsis. Objectives We evaluated the ability of hemodynamic changes produced by an increase in PEEP to predict FR in ventilated adult patients after cardiac surgery. Methods 50 consecutive hemodynamically stable patients undergoing cardiac surgery with use of cardiopulmonary bypass were enrolled in this study. Hemodynamic monitoring included pulmonary artery catheter (PAC) and transesophageal echocardiography (TEE). Hemodynamic variables were assessed at 4 different time points: 1. Baseline (PEEP 6.0 ± 1.2 cm H2O), 2. PEEP challenge (PEEP + 10 cm H2O for 5 minutes), 3. Baseline 2 (with reduction of PEEP to initial value), 4. After a crystalloid fluid challenge of 6 ml/kgBW. Fluid responsiveness was defined as increase of stroke volume (SV) >15 % after fluid challenge measured by PAC. Results 58 % of patients (n = 29) were fluid responders. The PEEP induced changes in CVP, MAP, pulse pressure (PP) and pulmonary capillary wedge pressure (PWP) were comparable between nonresponders (Δ CVP 2.5 ± 5.2 mmHg, ΔMAP 4.1 ± 10.3 mmHg, ΔPP: 11.4 ± 15.2 mmHg, Δ PCWP 1.9 ± 3.4 mmHg ) and responders (Δ CVP 3.3 ± 4.5 mmHg, Δ MAP 7.6 ± 8.6 mmHg, ΔPP: 7.5 ± 17.1, ΔPCWP 3.7 ± 4.,1 mmHg) (p = n.s.). Accordingly, there was no correlation between PEEP-induced changes in CVP, MAP, PP and PCWP and changes in SV after a fluid challenge and the area under the curve (AUC) for predicting fluid responsiveness was low (CVP:0.584,

Introduction Fluid administration is one of the principal therapy adopted in order to achieve early haemodynamic stabilization in septic patients. Despite international guidelines promote rapid volume expansion in the early stages of shock [1], several studies have recently shown that excessive fluid balance in septic patients is correlated with increased mortality [1] Objectives This study aims to investigate the impact of fluid balance on mortality in septic patients along with any associated electrolytes (strong ion difference-SID) and pH imbalance Methods This pilot retrospective study enrolled approximately 10 % of the adult patients admitted to the intensive care unit (ICU) of the Gemelli University Hospital with a diagnosis of severe sepsis/septic shock. We excluded pregnant women and those who survived less than 48 hours. We collected all the data concerning daily fluid balance, arterial blood gases (ABG), the hemodynamics and daily SOFA, until day 28 of ICU stay or discharge. Fluid balance at 24 hours was divided into quartiles for subsequent analysis. Mortality was assessed at 90 days. Results 61 patients were included in this pilot analysis from the 1st January 2008 to 30 June 2015, age 66 ± 14 years , SAPS II 51 [IQR 38.75-64]. The 36 patients who died had higher blood lactate (4.3 ± 3.7 vs. 2.1 ± 1.7 mmol/L, p = 0.007), developed more frequently acute kidney injury (AKI) (44.4 % vs. 16 %, p = 0.027) and presented a more positive fluid balance at 48, 72 and 96 hours (2228 ± 539 vs. 784 ± 465 ml, p = 0.048; 2225 ± 824 vs. -181 ± 479 ml, p = 0.016; 2189 ± 922 vs. -475 ± 154.7 ml, p < 0.01, respectively). A non significant trend toward increased mortality in the 3rd (dead 40 % vs. alive 60 %) and 4th quartile (dead 73.3 % vs alive 26.7 %) of fluid balance at 24 hours, was also found. No significant differences in terms of pH and electrolytic alterations have emerged, though a trend towards higher SID in the 2nd quartile was evident (28 ± 6 vs. 31 ± 5 vs. 28 ± 5 vs. 27 ± 6 vs. 29 ± 6, respectively p = 0.13). Multivariate analysis, including SAPS II on, Charlson's comorbidity score AKI and blood lactate on admission confirmed the association between positive fluid balance and 90-day survival even after adjustment for several confounding factors (OR 0.54 96 % CI 0.30-0.96, p = 0.035). Conclusions In our population positive fluid balance was independently associated with increased mortality at 90 days, but without any significant differences in terms of SID and pH. Reference(s) [1] Dellinger RP, et al., Surviving Sepsis Campaign Guidelines Committee including the Pediatric S, (2013) Surviving sepsis campaign: international guidelines for management of severe sepsis and septic shock: 2012. Critical care medicine 41: 580-637 [2] Boyd JH, et al., (2011) Fluid resuscitation in septic shock: a positive fluid balance and elevated central venous pressure are associated with increased mortality. Critical care medicine 39: 259-265

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

A165 A systematic review and meta-analysis of the fluid challenge technique in anaesthesia and intensive care L. Toscani1,2, D. Antonakaki1, D. Bastoni1,3, H.D. Aya1, A. Rhodes1, M. Cecconi1 1 St George's University Hospital, London, United Kingdom; 2Università La Sapienza di Roma, Rome, Italy; 3Università degli Studi di Parma, Parma, Italy Correspondence: L. Toscani – St George's University Hospital, London, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A165 Introduction The gold standard to evaluate fluid responsiveness and guide fluid administration, in both critically ill and surgical patients, is the fluid challenge technique (FC). Recent evidence has highlighted an important variability in the current practice of this technique. Different techniques may lead to different results and different clinical decisions. Objectives The aim of this study is to describe the fluid challenge techniques reported in fluid responsiveness studies and clinical trials and to assess whether there is a difference in the proportion of “responders” depending on the type of fluid, volume and duration of infusion. Methods We conducted a systematic review using Medline and EMBASE. The inclusion criteria for the studies were: the use of the FC as a test of cardiac preload; the inclusion of a description of the FC ( reporting volume, type of fluid and duration of infusion); a reported definition of fluid responsiveness; a reported proportion of responders and non-responders;clinical setting(ICU or operative theatre).Included studies were examined in full and subjected to quantifiable analysis.The predictors were categorized as follows: volume (500 ml); fluid type (colloids or crystalloids); infusion rate ( 500 mL is .57 (median = .50, IQR = .38-.79). There is no evidence to suggest that the proportion of responders changes with the volume used (H = 1,79 p = .40). 23 (31,5 %)studies reported the use of crystalloids, whilst 50 (68,5 %) studies used colloids. In 3 studies the type of fluid was not clearly reported. There is no evidence to suggest that the proportion of responders changes between colloids and crystalloids (M = .53, SD = .15 vs M = .56, SD = .15, F(1,71) = .85, p = .35). The duration of infusion was 4 % (sensitivity = 94 % [95 % confidence interval, CI: 70-100 %]; specificity = 100 % [95%CI: 82100 %]). If the absolute values of SV changes recorded during endexpiratory plus end-inspiratory occlusions were added, volume responsiveness was predicted by an increase in SV >12 % (sensitivity = 94 % [95%CI: 70-100 %]; specificity = 95 % [95%CI: 74-100 %]). This predictive accuracy for volume responsiveness was not better than for the SV changes recorded during the only end-expiratory occlusion (area under the ROC curve: 0.98 (95%CI: 0.86-1.00) vs. 0.94 (95%CI: 0.81-0.99), respectively, p = 0.46). Conclusions The sum of the absolute values of changes in SV recorded during consecutive end-expiratory and end-inspiratory occlusions reliably predict volume responsiveness. This prediction is not better than for the changes in SV during the only end-expiratory occlusion. Nevertheless, combining end-expiratory and end-inspiratory occlusions induces larger SV changes and increases the value of the threshold, what is more compatible with the precision of echocardiography and oesophageal Doppler. A167 Physiological volume replacement ratio can be reached in experimental hemorrhage model I. László1, G. Demeter1, N. Öveges1, K. Tánczos1, M. Németh1, D. Trásy1, I. Kertmegi2, D. Érces2, B. Tudor3, J. Kaszaki2, Z. Molnár1 1 University of Szeged, Department of Anesthesiology and Intensive Therapy, Szeged, Hungary; 2University of Szeged, Institute of Surgical Research, Szeged, Hungary; 3Medical University of Vienna, Department of Anaesthesiology and General Intensive Care Medicine, Vienna, Austria Correspondence: I. László – University of Szeged, Department of Anesthesiology and Intensive Therapy, Szeged, Hungary Intensive Care Medicine Experimental 2016, 4(Suppl 1):A167

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Introduction Acute bleeding is a life threatening condition requiring immediate and adequate interventions. Adequate fluid resuscitation is the cornerstone of maintaining and correcting oxygen delivery (1). According to Starling's “3-compartment model”, 4-times more crystalloids should have the same volume effect as colloids. However, this volume-replacement ratio remains a controversial issue as this may be affected by the degradation of the endothelial glycocalyx layer often found in the critically ill. Objectives Our aim was to compare colloid and crystalloid based fluid resuscitation during an experimental stroke volume index (SVI) guided hemorrhage and resuscitation model in Vietnamese minipigs. Methods In this experiment 15 anesthetized and mechanically ventilated pigs were randomized to receive colloid (Voluven®,HES, n = 7) or crystalloid (Ringerfundin®,RF, n = 8) infusion. Animals were bled till baseline SVI (Tbsl) dropped by 50 % (T0), followed by resuscitation until initial SVI was reached (T4) in four steps. Statistics were performed by using SPSS® 23.0 and statistical analysis were tested by General Linear Model, Independent samples T-test and MannWhitney U test, as appropriate. Results Hemodynamic changes during the experiment did not show clinically relevant differences between the groups. At Tbsl the SVI values were similar (HES: 34 ± 8 ml/m2, RF: 33 ± 4 ml/m2), after bleeding, the SVI decreased by the planned 50 % to T0 (HES: 17 ± 4, RF: 15 ± 2 ml/m2) and returned to its initial value by T4 (HES: 34 ± 7, RF: 32 ± 3 ml/m2). Cardiac index(CI) also decreased (Tbsl, HES: 3.25 ± 0.23, RF: 3.14 ± 0.19 l/min/m2; T0, HES: 1.58 ± 0.27, RF: 1.84 ± 0.40 l/min/m2) and reached a higher value by T4 (HES: 3.99 ± 0.54, RF: 3.39 ± 0.36 l/ min/m2, p = 0,006). There was a significant increase in heart rate over time (Tbsl, HES: 95 ± 19; RF: 97 ± 18 beats/min; T4, HES: 117 ± 17, RF: 102 ± 14 beats/min, p‹0.05). Similar amount of blood was shed in both groups (HES: 553 ± 206 ml, RF: 506 ± 107 ml) but the animals received significantly less resuscitation fluid (623 ± 208, 1754 ± 602 ml, p = 0.002, respectively) and total fluid in the HES-group (1029 ± 252, 2010 ± 600 ml, p = 0.006, respectively). The volume replacement ratio was significantly different between the HES median and RF-groups: median = 0.87 [IQR: 0.78-1.83], 3.06 [2.93-4.57], p = 0,002, respectively). Conclusions Our results showed that in healthy pigs the volumereplacement ratio follows the Starling's principle. This indicates that in acute bleeding events, like in trauma and during surgery when the glycocalix is most likely to be still intact, colloids may be beneficial as hemodynamic stability can be achieved faster than with crystalloids.

Page 90 of 607

with PLR has been reported as a surrogate of increased CO in prediction of FR (1), however, the use of end-tidal Co2 is limited to mechanically ventilated patients. Increased mixed venous oxygen saturation was another surrogate of increased CO in prediction of FR with PLR (2), however, this parameter needs insertion of pulmonary artery catheter. Although central venous oxygen saturation (Scvo2) has been a surrogate of mixed venous oxygen saturation, however Scvo2 was not investigated in prediction of FR. Venous-Arterial Co2 gap (VA-Co2 gap) is defined as the difference between central venous and arterial Co2 partial pressure. VA-Co2 gap increases in cases of tissue hypoperfusion (3) and decreases after CO improvement. Objectives We investigated the role of the change in VA-Co2 gap and the change in Scvo2 after PLR in prediction of FR Methods We included 42 patients with acute circulatory failure with elevated serum lactate. PLR test was performed to detect FR. Fluid responders were defined as patients with increased stroke volume by 10 % (measured by ICON cardiometry device) after PLR test. In addition to demographic and hemodynamic data, arterial and central venous blood gases were obtained before and after PLR. The predictive ability of the change in VA-Co2 gap and the change in Scvo2 with PLR to predict FR was obtained using area under receiver operating characteristic (AUROC) curve. Results Fluid responders were 11 patients (26 %). AUROC for the change in VA-Co2 gap in prediction of FR was 0.786(95 % CI: 0.6250.895). Sensitivity was 73 % and specificity was 65 % at cutoff value of 0.5 mmHg increase in VA-Co2 gap. AUROC for the change in Scvo2 in prediction of FR was 0.608(95 % CI: 0.443-0.756). Sensitivity was 64 % and specificity was 63 % at cutoff value of 2 % decrease in Scvo2. Conclusions The change in VA-Co2 gap with PLR is superior to the change in Scvo2 in prediction of FR in patients with acute circulatory failure. VA-Co2 gap can be considered in settings where CO monitoring is not feasible. References 1. Monnet X, Bataille A, Magalhaes E, Barrois J, Le Corre M, Gosset C, et al. End-tidal carbon dioxide is better than arterial pressure for predicting volume responsiveness by the passive leg raising test. Intensive Care Med 2013;39:93-100. 2. Kuiper et al. Mixed venous O2 saturation and fluid responsiveness after cardiac or major vascular surgery. J Cardiothorac Surg 2013;8:189 3. Robin E, Futier E, Pires O, Fleyfel M, Tavernier B, Lebuffe G, et al. Central venous-to-arterial carbon dioxide difference as a prognostic tool in highrisk surgical patients. Crit Care 2015;19:227.

Reference(s) 1. Tánczos K, Németh M, Trásy D, et al. Goal-Directed Resuscitation Aiming Cardiac Index Masks Residual Hypovolemia: An Animal Experiment. Biomed Res Int 2015;2015:160979. Grant acknowledgement Supported by NKFIH K116689.

A168 The ability of venous-arterial carbon dioxide gap and central venous oxygen saturation to predict fluid responsiveness by passive leg raising test A. Hasanin, A. Lotfy, A. El-adawy, H. Nassar, S. Mahmoud, A. Abougabal, A. Mukhtar Cairo University, Department of Anesthesia and Critical Care Medicine, Cairo, Egypt Correspondence: A. Abougabal – Cairo University, Department of Anesthesia and Critical Care Medicine, Cairo, Egypt Intensive Care Medicine Experimental 2016, 4(Suppl 1):A168 Introduction Passive leg raising (PLR) test has been a popular method used for detection of fluid responsiveness (FR) in patients with acute circulatory failure. The most important limitation with PLR test is the need to a real-time cardiac output (CO) monitor to trace the patient hemodynamic response to PLR. Increased end-tidal Co2

Fig. 74 (abstract A168). AUROC for Co2 Gap and ScVo2 to predict FR

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

A169 Carotid doppler to predict fluid responsiveness in intensive care unit: a mini fluid challenge study F. Quinty1, S. Habchi2, A. Luzi3, E. Antok1 1 University of la Reunion, Intensive Care Unit, Saint Pierre de la Reunion, France; 2University of la Reunion, Neurocritical Care Unit, Saint Pierre de la Reunion, France; 3University of la Reunion, Anesthesiology Department, Saint Pierre de la Reunion, France Correspondence: F. Quinty – University of la Reunion, Intensive Care Unit, Saint Pierre de la Reunion, France Intensive Care Medicine Experimental 2016, 4(Suppl 1):A169 Introduction Fluid management is one of the most difficult tasks in critical care medicine. A recent study shows that in current practice, fluid administration is not evidence based despite a lot of tools available for physician (1). However many of them are invasive or not always applicable depending on clinical context. Carotid doppler, a non invasive and easy-to-use method, has shown excellent predictive values to monitor fluid responsiveness after passive leg raising (2). Otherwise, mini fluid challenge is described as a reliable alternative to classical fluid challenge to prevent fluid overload. Objectives To assess if variation of carotid doppler flow can predict fluid responsiveness after a mini fluid challenge. Methods This prospective observational study was performed from September to December 2015. Patients requiring volume expansion were eligible for enrollment. Patients less than 18 years old, with carotid stenosis or poor insonation were excluded. We recorded the variation of carotid doppler flow after 100 cc of cristalloids (ΔDc100) over 1 min and the variation of velocity time integral (VTI) after an additional infusion of 400 cc over 14 min assessed by transthoracic echocardiography. The cardiac output (CO) was calculated as CO = VTI x Heart Rate x Aortic Surface . A patient whose CO increased by 15 % following 500 cc (ΔCO500) was consider as a fluid responder. If mechanical ventilation was required, a lung protective strategy was applied. Statistics: Spearman's correlation test was used. The receiver operating curve (ROC) and grey zone were defined for ΔDc100. Results In total, 30 patients were included. Diagnosis admission were severe sepsis/septic shock (22), brain injury (4) and post operative (4). Sixty five percent of patients were ventilated and 45 % required vasopressor support. There was no difference between responders and non responders. Fourteen patients (45 %) were volume responders. Dc100 increased by 32 % +/- 24 % in the responders compared with 10 % +/- 8 % in the non responders (p < 0.001). ΔDc100 was strongly correlated with ΔCO500 (r 0,78 ; p < 0,001). The best threshold of ΔDc100 was 14 % with a sensitivity and specificity of 93 % and 82 % respectively. The areas under the ROC curve of ΔDc100 was 0,91 +/- 0,01. After bootstrapping (1000 resamples) a grey zone ranging between 11 and 20 % was observed in up to 29 % of patients. Conclusion Our study shows that variation of carotid doppler after a mini fluid challenge can predict fluid responsiveness in heterogeneous critical care population. Reference(s) (1) Cecconi M, Hofer C, Teboul JL et al (2015) Fluid challenges in intensive care: the FENICE study: A global inception cohort study. Intensive Care Med ; 41(9):1529-37 (2) Marik PE, Levitov A, Young A et al (2013). The use of bioreactance and carotid Doppler to determine volume responsiveness and blood flow redistribution following passive leg raising in hemodynamically unstable patients. Chest ; 143:364-70

Page 91 of 607

A170 Peripheral perfusion response to the first fluid resuscitation predicts mortality in patients with sepsis-related acute circulatory dysfunction admitted to the emergency department G. Hernandez1, B. Lara2, L. Enberg2, M. Ortega1, P. Leon1, C. Kripper2, P. Aguilera2, E. Kattan1, J. Bakker1 1 Pontificia Universidad Catolica de Chile, Facultad de Medicina, Departamento de Medicina Intensiva, Santiago, Chile; 2Pontificia Universidad Catolica de Chile, Programa de Medicina de Urgencia, Facultad de Medicina, Santiago, Chile Correspondence: G. Hernandez – Pontificia Universidad Catolica de Chile, Facultad de Medicina, Departamento de Medicina Intensiva, Santiago, Chile Intensive Care Medicine Experimental 2016, 4(Suppl 1):A170 Background Sepsis-related acute circulatory dysfunction is a life threatening condition. Peripheral perfusion as a marker of hypoperfusion could be used to trigger initial fluid resuscitation (FR). The response of peripheral perfusion to FR in patients with abnormal perfusion could potentially identify high-risk patients. Objective Our aim was to study the effects of the first protocolized FR on capillary refill time (CRT) and other perfusion parameters, and the relationship of the response with outcome. Methods Prospective observational study including patients with sepsis and acute circulatory dysfunction just admitted to the Emergency department (ED) and in whom an initial FR was indicated according to standard care. Peripheral perfusion and laboratory assessments were performed before and after protocolized FR. Follow-up of patients until hospital discharge. CRT responders were defined as patients who were able to maintain normal CRT or to normalize abnormal CRT values after FR. Results One hundred consecutive patients were included. Of 30 patients with an abnormal CRT at admission, 23 (77 %) normalized CRT after initial FR. CRT responders showed a significant decrease in heart rate and lactate, presented less organ dysfunction and requirement of mechanical ventilation. Hospital mortality was significantly lower in CRT responders when compared to non-responders (9.6 % vs. 55.6 %; p < 0.001). In logistic regression analysis only CRT was significantly related to hospital mortality. This association was maintained after adjusting for baseline severity. Conclusions Patients with sepsis-related acute circulatory dysfunction that exhibit normal CRT after early FR have low mortality rates. In contrast, failure to improve peripheral perfusion in response to initial FR is a strong predictor of mortality. This finding could be very important for the ED or limited-resource settings since it could help to decide on additional diagnostic and treatment options. A171 Prediction of fluid responsiveness in patients with assisted mechanical ventilation: a comparison of the "fluid responsiveness index" FRI to CVP, global end-diastolic volume index gedvi and stroke volume variation SVV W. Huber1, M. Lehmann1, S. Sakka2, B. Bein3, R.M. Schmid1 1 Technische Universität München, II. Medizinische Klinik, Munich, Germany; 2Klinikum Köln-Merheim, Cologne, Germany; 3Asklepios Klinik St. Georg, Hamburg, Ghana Correspondence: W. Huber – Technische Universität München, II. Medizinische Klinik, Munich, Germany Intensive Care Medicine Experimental 2016, 4(Suppl 1):A171 Introduction Appropriate fluid support is crucial in critical care. To guide fluid resuscitation filling pressures such as central venous pressure CVP, volumetric parameters (e.g. global end-diastolic volume index GEDVI) and dynamic parameters of fluid responsiveness (FR) like stroke volume variation SVV are used. Prediction of FR is

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

particular difficult in patients with spontaneous breathing, since the use of SVV is usually restricted to patients under controlled ventilation. Under these circumstances passive leg raising PLR and “mini volume challenges” can be used for intermittent assessment of FR. Recently “fluid responsiveness index” FRI has been introduced as experimental marker of FR. FRI is derived from a proprietary algorithm based on the continuous analysis of arterial and central-venous pressures by a special software provided for scientific use with the PiCCO-2 device (Pulsion Medical Systems SE, Feldkirchen, Germany). Objectives It was the aim of our study to compare the prediction of FR (defined as an increase in stroke volume index ≥10 %) by FRI, GEDVI, CVP and SVV after a volume challenge (VC) with 7 mL/kg crystalloid within ≤30 minutes in 27 patients under assisted mechanical ventilation and equipped with TPTD-monitoring irrespective of the study. Methods All haemodynamic parameters were measured immediately before and after the VC with a modified PiCCO-2-device providing continuous FRI. The study was approved by the local Ethics Committee. Statistics: SPSS 23. Results 20 male, 7 female patients; APACHE-II 25 ± 6. FRI (ROC-AUC 0.881; p = 0.005) and GEDVI (AUC 0.778; p = 0.041) significantly predicted an increase in stroke volume index ≥10 %, whereas neither CVP (AUC 0.675; p = 0.199) nor SVV were predictive (AUC 0.702; p = 0.137). FRI was significantly higher in fluid-responders than in nonresponders (0.60 ± 0.20 vs 0.20 ± 024; p = 0.003), while GEDVI was significantly lower in responders (657 ± 109 vs. 797 ± 170 mL/m2; p = 0.042). CVP and SVV were not different between responders and non-responders. A cut-off of 0.42 for FRI-provided a sensitivity of 83 % and a specificity of 81 % regarding an increase in stroke volume index ≥10 %. GEDVI values below 800 mL/m2 predicted FR with a sensitivity of 83 % and a specificity of 62 %. In binary regression analysis only FRI (p = 0.015) was independently associated to the primary endpoint. Similarly, FRI (AUC 0.827; p = 0.025) and GEDVI (AUC 0.809; p = 0.034) predicted an increase in CI ≥10 % after VC, whereas CVP and SVV were not predictive. Finally, only FRI predicted an increase in cardiac power index CPI ≥10 % (AUC 0.812; p = 0.008). Conclusions FRI significantly and independently predicts FR in patients with spontaneous assisted ventilation. Furthermore, GEDVI is significantly associated to FR in these patients, while CVP and SVV were not predictive. With regard to the limited number of patients included, confirmatory validation studies are required.

A172 The use of venous return gradient to predict fluid responsiveness in critically ill patients J. Preti, J. Creteur, A. Herpain Université Libre de Bruxelles, Brussels, Belgium Correspondence: J. Preti – Université Libre de Bruxelles, Brussels, Belgium Intensive Care Medicine Experimental 2016, 4(Suppl 1):A172 Introduction Fluid expansion (FE) is a major issue in the management of critically ill patients. Accurate and practical tools to predict fluid responsiveness are still lacking. Venous return gradient (dVR) is a major component of venous return and therefore of right heart preload and cardiac output. dVR is the difference between mean systemic filling pressure (Pmsf) and central venous pressure (CVP) or right atrial pressure (RAP). Pmsf can be estimated by measuring the static filling pressure of the arm (Parm). This later has shown to be a useful predictor of fluid responsiveness in cardiac surgery patients [1]. Objective The aim of the study was to evaluate if dVR is a reliable variable to predict fluid responsiveness. Methods Twenty-nine critically ill patients requiring FE were included in this prospective observational study. In each patient, Parm was measured three times before FE from the arterial pressure curve obtained from a radial artery catheter 30 sec after occluding the arterial flow by a cuff inflated up to 50 mmHg above systolic blood pressure. The average value of triplicate measurements was calculated. Stroke

Page 92 of 607

volume was estimated either by Swan-Ganz catheter or transthoracic echocardiography. A positive response to the fluid expansion was defined as a SV increased by at least 10 %. The optimal cut offs were chosen using a receiver operating characteristic curve (ROC) analysis and identifying the maximal Youden's index. Sensitivity, specificity, and their approximate 95 % confidence intervals were computed. Results In the group of patients who responded to the fluid expansion (n = 19), the mean CVP/RAP was 8 mmHg ±3, the mean Parm was 21 mmHg ±7 and the mean dVR was 13 mmHg ±7. In nonresponders patients (n = 10), the mean CVP/RAP was 8 mmHg ±3, the mean Parm was 26 mmHg ±6 and the mean dVR was 18 mmHg ±6. The best parameter in order to predict fluid responsiveness was dVR, with an area under the curve of 0,77 (95 % confidence interval 0,57-0,97); a dVR cut-off of < 15 mmHg predicting a successful FE with a sensitivity of 74 % and a specificity of 90 %. In comparison, for Parm, the area under the curve was 0,73 (95 % confidence interval 0,54-0,92); a Parm cut-off of < 24 mmHg predicting a successful FE with a sensitivity of 74 % and a specificity of 70 %. Conclusion Venous return gradient estimated by arm occlusion pressure is at least as good as Parm for the prediction of fluid responsiveness in critically ill patients. Reference(s) [1] Geerts, B.F., et al., Arm occlusion pressure is a useful predictor of an increase in cardiac output after fluid loading following cardiac surgery. Eur J Anaesthesiol, 2011. Grant acknowledgement None

A173 The effectiveness of echocardiography to evaluate preload dependency in septic shock with left systolic heart failure dyspred study J. Marc1, E. Zogheib1, F. Trojette1, S. Bar1, L. Kontar1, D. Titeca1, J. Richecoeur2, B. Gelee2, N. Verrier2, R. Mercier2, E. Lorne1, J. Maizel1, H. Dupont1, M. Slama1 1 Univ. Hospital of Amiens, Amiens, France; 2CH Beauvais, Beauvais, France Correspondence: J. Marc – Univ. Hospital of Amiens, Amiens, France Intensive Care Medicine Experimental 2016, 4(Suppl 1):A173 Introduction The ability of echocardiography and analysis of mitral profile to predict fluid responsiveness in a septic shock with left systolic heart failure is difficult to manage. Objective The objective of the study was to evaluate the ability of mitral profile and its evolution with a test of passive leg raising to discriminate fluid responsiveness in septic shock with left systolic heart failure. Methods 60 patients in septic shock and left systolic failure, monitored by transthoracic echocardiography (TTE) and continuous measurement of cardiac output (CO) (catheter pulmonary artery catheter (PAC) or transpulmonary thermodilution (TPTD)) were included. Mitral profile (E, A, E/A, E´ lateral, E´ septal, E´ average, E/E´), CO, pulmonary artery occlusion pressure (PAOP) and extravascular lung water (EVLW)) were collected before volume expansion (VE), after passive leg raising (PLR) and after VE with 500 ml of crystalloid solution. Variation of each hemodynamic values after VE (Δ(VE)) and after PLR (Δ(PLR)) was performed. The left systolic heart failure was defined with left ventricle ejection fraction (LVEF) ≤ 40 %. Patients were classified into two groups according to their response after VE measured by thermodilution: responders (R) defined by an increase ≥ 15 % of CO, and non-responders (NR). Results Of the 60 patients monitored by TTE, 28 (46 %) with also TPTD and 32 (54 %) with PAC. 25 (42 %) were R and 35 (58 %) NR. All were under norepinephrine (1.1 gamma/kg/min +/- 0.4). There was no significant difference between R and NR with E, A, E/A (1.67 +/- 0.85 vs 1.40 +/- 0.97), E´ average (7.5 +/- 2.5 vs 7.1 +/- 2.8 ), E/E´ average (12.5 +/- 6.1 vs 13.5 + / 7.5). It was not found correlation between the E/E ´and PAOP (p = 0.48) nor with the EVLW (p = 0.78).

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

33 % (10/30) of NR had an E/E´ lat < 8 and 28 % (9/25) of R had an E/ E´ lat > 15. ΔCO(VE) (p < 0.05) and ΔE ´ average(VE) (p < 0.05) were higher in the R while ΔE/E´ lat (VE) (p < 0.05) and ΔE/E ´ average (VE) (p < 0.05) were higher in NR. ΔE´(VE) was correlated with the ΔCO(VE) independently with LVEF (r = 0.39 p < 0.05). He was found a positive correlation between ΔCO(EV) and ΔCO(PLR) (r = 0.58 p < 0.05). The ΔCO(PLR) AUC with an optimal threshold at 12 % to predicted fluid responsiveness was 0.890 p < 0.05) with 92 % sensitivity and 85 % specificity. ΔE/E ´average(PLR) was correlated with ΔE/E´ average(VE) (r = 0.29 p < 0.05). Conclusions In decompensated patients with septic shock and left systolic heart failure, analysis of mitral profile alone not be reliable in predicting left filling pressure nor in predicting fluid responsiveness. The PLR and the variation of the mitral profile (ΔE/E´ average and ΔE´ lat) would secure VE in this population.

Cardiovascular monitoring 2 A174 The effect of fluid balance on extra-vascular lung water assessed by lung ultrasound & electrical cardiometry: a prospective cohort study M.E. Abdelfattah1, A. Eladawy2, M.A. Ali Elsayed1, A. Mukhtar1 1 Cairo University, Cairo, Egypt; 2Kasr Alainy Medical School, Cairo University, Cairo, Egypt Correspondence: A. Eladawy – Kasr Alainy Medical School, Cairo University, Cairo, Egypt Intensive Care Medicine Experimental 2016, 4(Suppl 1):A174 Introduction Fluid balance is one of the most frequently manipulated clinical care variables in the ICU. The risks associated with invasive monitoring and its relationship to heightened mortality make the evaluation and utilization of other modalities for tracking volume status in critically ill patients vitally important.1 Objectives To correlate the three day cumulative fluid balance (CFB) with extravascular lung water assessed by LUS & electrical cardiometry by the end of the 3rd day & to assess the diagnostic accuracy of electrical cardiometry in assessment of thoracic fluid content (TFC) in comparison to lung ultrasound. Methods Three day CFB was measured in the 1st three days of ICU stay. Lung ultrasound score was obtained by scanning 12-rib interspaces. & the sum of B-lines yielded a score (0-36) 2. Electrical cardiometry was used to assess TFC (considered abnormal if exceeding a predetermined value). E/e´ ratio was done to assess fluid load; for further correlation with CFB, lung score, & thoracic fluid content (TFC). All measurements were done on day 1 & day 3. Other variables collected were age, gender, cardiac output, APACHE II score, medical or surgical patient, mechanical ventilation, length of ICU stay & ICU mortality. Results 30 patients were enrolled. The median (IQR) cumulative 3day fluid balance was -600 (-2225, 437). Ten patients (33.3 %) had positive fluid balance & 20 patients (67.7 %) had negative fluid balance. The median (IQR) of TFC at day 1 was 52 (35.5, 58) and at day 3 was 46 (37, 52.75). There was no significant correlation between 3-day cumulative fluid balance with either LUS, Ee´ ratio, or TFC (Table 56). There was a significant correlation between TFC & LUS on day 1 & day 3 (r = 0.610, p < 0.01), (r = 0.4, p = 0.05) respectively. There was no relationship between Ee´ ratio & both TFC & LUS on day 1, it became significant on day 3 (Table 57). TFC correlated significantly with LUS in patients with negative fluid balance but not in those with positive fluid balance. By the same token, LUS was correlated significantly with Ee´ in patients with negative balance only. However, no correlation was found between TFC & Ee´ in patients with either negative or positive fluid balance (Table 58). Conclusions There is a good relationship between TFC & LUS for assessment of extra-vascular lung water. However the relationship between CFB & extra-vascular lung water is poor. Future larger studies are warranted to develop predictive equation for assessment of the amount of extravascular lung water from TFC.

Page 93 of 607

Reference(s) 1. Martin GS, Ely EW, Carroll FE, Bernard GR. Findings on the Portable Chest Radiograph Correlate With Fluid Balance in Critically Ill Patients. 2007; 2087-95. 2. Volpicelli G, Mussa A, Garofalo G, Cardinale L, Casoli G, Perotto F, et al. Bedside lung ultrasound in the assessment of alveolar-interstitial syndrome. Am J Emerg Med. 2006 Oct;24(6):689-96.

Table 56 (abstract A174). Correlation between CFB, TFC, Ee' (Day 3) Cumulative Fluid balance r-value

p-value

TFC

0.25

0.18

LUS

-0.107

0.57

Ee´

-0.002

0.99

Table 57 (abstract A174). Correlation between TFC, Ee' & LUS (Day 1 & Day 3) Day 1

Day 3

r-value

p-value

r-value

TFC & LUS

0.610

19 %) had significantly lower MAP and higher lactate values than the mild severity group (Fig. 87, top panel). Temporal changes in cardiac contractility

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

and output did not however relate to insult severity (Fig. 87, bottom panel). Conclusions In an experimental rat myocardial ischaemia/reperfusion injury model, blood pressure and metabolic impairments related to insult severity. By contrast, cardiac contractility and cardiac output showed no significant temporal change. References [1] Frank A et al. Semin Cardiothorac Vasc Anesth. 2012; 16:123-32 [2] van den Bos EJ et al. Am J Physiol Heart Circ Physiol. 2005; 289:H1291-300

Page 107 of 607

Results Time of ECG done was not significantly different between the groups, 9.8 ± 3.3 and 7.8 ± 2.3 minutesThe D2B time was 119.3 ± 16.3 minutes in group 1 compared with 90.4 ± 9.8 minutes in group 2 (p = 0.13). The reduction in the D2B time was driven by a reduction in time from ECG interpretation to activate CCL (28.3 ± 4.1 in the group 1 and 17.6 ± 2.3 minutes in the group 2) (p = .001) shown in Table 69. Mortality rate in group 1 is 12.5 % compare to 2.2 % in group 2 (p = .07), shown in Table 69. Conclusions Usage smartphone to transmit ECG in the ED to interventional cardiologists can facilitate communication and reduce the time of CCL activation and balloon inflation. Reference(s) 1. Peterson MC, Syndergaard T, Bowler J, Doxey R. A systematic review of factors predicting door to balloon time in ST-segment elevation myocardial infarction treated with percutaneous intervention. International journal of cardiology. 2012;157(1):8-23.

Table 69 (abstract A200). Group 1 using telephone; group 2 using smartphone

Fig. 87 (abstract A199). Timecourse of haemodynamic, contractility and perfusion markers in mild (IA/LV < 19 %, green dots) and severe (IA/LV > 19 %, blue dots) animals. Ischaemia time shown as red area A200 Using smartphone application for transmitting ECG to reduce time of percutaneous intervention C.C. Chao1,2, P. Hou3,4 1 Taipei Medical University Hospital, Emergency Department, Taipei, United States; 2Taipei Medical University, Taipei, Taiwan, Province of China; 3Harvard Medical School, Boston, United States, 4Brigham and Women's Hospital, Boston, United States Correspondence: C.C. Chao – Taipei Medical University Hospital, Emergency Department, Taipei, United States Intensive Care Medicine Experimental 2016, 4(Suppl 1):A200 Introduction How to reduce the time of percutaneous intervention in patients diagnosed with ST elevation myocardial infarction (STEMI) is a critical issue for care process. Many interventions such as prehospital ECG, cardiac catheterization laboratory (CCL) activated by emergency physician (EP) have been proved to improve the result(1). However discordance in diagnosis between the EP and the interventional cardiologist may result in unnecessary activation of the CCL. Communication between EP and interventional cardiologist is often verbal report via telephone. After Development of the smartphones, EPs can use application to transmit ECG images to interventional cardiologists when STEMIs were diagnosed. Objectives This study aims to evaluate the use of smartphone application to facilitate communication between the EP and the interventional cardiologist in order to minimize the time to activate CCL and time of percutaneous intervention. Methods We retrospectively collected time-point of every step in management and outcome of patients diagnosed with STEMI in the ED. Total 84 patients were enrolled. In group 1, patient's ECG was described by traditional verbal communication via telephone. In group 2, we use smartphone application for transmitting ECG to interventional cardiologist. Time-points of intervention were recorded for analysis.

Group 1 (n=40) (95 % CI)

Group2 (n=44) (95 % CI)

P value

Sex, Male

34

40

0.40

Age, years

60.0±1.8 (56.2~63.7)

64.4±2.1 (60.2~68.5)

0.11

Time to ECG Time to ECG done, minutes

9.8±3.3 (3.1~16.5)

7.8±2.3 (3.6~11.5)

0.57

Time to ECG interpretation by EP, minutes

11.8±3.5 (4.6~19.0)

8.6±2.0 (4.6~12.7)

0.44

Activate CCL time, minutes

54.5±12.7 (28.8~8.02)

34.7±9.0 (16.5~52.8)

0.10

EP read ECG to CCL activation time, minutes

28.3±4.1 (19.9~36.7)

17.6±2.3 (13.1~22.1)

0.01

EP read ECG to balloon inflation time, minutes

93.1±9.5 (73.9~112.4)

73.4±3.9 (65.6~81.2)

0.025

D2B, minutes

119.3±16.3 (86.4~152.2)

90.4±9.8 (70.7~110.1)

0.13

Mortality rate

Mortality rate 12.5% (Expire 5; Survive 35)

Mortality rate 2.2% (Expire 1; Survive 43)

0.07

A201 The application of innovative design of an electrocardiogram exam accessory device to improve ambulance prehospital electrocardiogram implantation rate in a city based multicenter trial W.-C. Huang1, C.-C. Hung1, C.-H. Chiang2, W.-T. Hung1, K.-C. Lin1, S.-C. Lin1, Y.-J. Liou3, S.-M. Hung3, Y.-S. Lin3, C.-C. Cheng2, F.-Y. Kuo2, K.-R. Chiou2, C.-J. Chen4, L.-S. Yan4, C.-Y. Liu4, H.-H. Wang4, P.-L. Kang2, H.-L. Chen3, C.-K. Ho4, G.-Y. Mar2, C.-P. Liu4 1 Kaohsiung Veterans General Hospital, Critical Care Division, Kaohsiung City, Taiwan, Province of China; 2Kaohsiung Veterans General Hospital, Cardiovascular Division, Kaohsiung City, Taiwan, Province of China; 3 Kaohsiung City Government, Fire Bureau, Kaohsiung City, Taiwan, Province of China; 4Kaohsiung City Government, Department of Health, Kaohsiung City, Taiwan, Province of China Correspondence: C.-H. Chiang – Kaohsiung Veterans General Hospital, Cardiovascular Division, Kaohsiung City, Taiwan, Province of China Intensive Care Medicine Experimental 2016, 4(Suppl 1):A201 Introduction The prehospital electrocardiogram (ECG) was identified as an critical part of treatment for patients with STEMI. However, it remained a challenging issue to set up prehospital ECG in Asia. Objectives This study is to investigate the application of innovative design of an electrocardiogram exam accessory device to improve ambulance prehospital ambulance electrocardiogram implantation rate via in a city based multicenter trail.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Page 108 of 607

Method This study started since Sep, 2011 via a multidisciplinary team among Kaohsiung veterans General hospital, fire bureau and department of health, Kaohsiung city government. The unique accessory device for 12 lead electrocardiography apparatus has 10 holes, which are arranged according to a standard electrode placement for the 12 lead ECG measurements. The design of inter-nipple line and mid-sternum line on the device can assist the staffs to perform ECG shortly. This breakthrough innovation designed to address the core issue of the efficiency of the ambulance pre-hospital ECG system. Therefore, the invention successfully promoted Kaohsiung city council to set up Asian first ambulance prehospital telemetry electrocardiogram system. The innovative design of a ECG exam accessory device was patented in Taiwan and won golden award in Geneva and Korean international invention. The consecutative chest pain patients received ambulance ECG exam were enrolled from Jan. 2011 to September. 2015 in 18 different fire brigades at Kaohsiung city. The ECG implementation rate is defined as chest pain patients received ambulance ECG exam divided by all patients with chest pain. Results The ECG implementation rate increased from 0 % in preinterventional to 62.2 % in post-interventional group (p < 0.001). Total 66 patients with STEMI was detected in 1205 chest pain patients received ambulance ECG exam. In these STEMI patients, average door to balloon time was 51 minutes, average ischemia to balloon time was 125 minutes and in-hospital mortality was 0 %. Conclusions This study demonstrates that application of innovative design of an electrocardiogram exam accessory device can solve the main problem of system and assist to set up first ambulance prehospital telemetry electrocardiogram system in Asia and further improve ambulance prehospital electrocardiogram implantation rate in Kaohsiung city.

ranged from 14 to 48 hours. Following readmission, the length of stay ranged from 10 days to 17 days. In 2015, there were 702 discharges with 30 readmissions. 14 of these were within 48 hours (2.0 %). The time of readmission varied from 9 to 48 hours (median = 25 hours). Cardio-respiratory failure was the most common cause for readmission. The length of the readmission stay on ICU ranged from 1 day to 68 days (median = 8 days). Over the two year period the ICU mortality rate remained unchanged at 16 % however patients readmitted within 48 hours showed considerably higher hospital mortality rates of 50 % in 2014 and 42 % in 2015. Conclusions Over the two year period, there was an increase in admissions and discharges from ICU. This was associated with an increase in the midnight bed occupancy rate from 79.5 % in 2014 to 84.3 % in 2015. However, this was linked with a 7 fold rise in readmissions within 48 hours and an associated hospital mortality rate higher than both the previous year and national average. An increasing demand for ICU services can have an adverse effect on the quality of care delivered. Further research is needed to inform clinicians and commissioners how quality can be maintained with increasing demand.

QUALITY AND SAFETY I

A203 'Ventsafe': a quality improvement project A. Wilson, J. Capps, W. Ayoub, A. Lomas, S. Ghani, J. Moore, D. Atkinson, M. Sharman Central Manchester Foundation NHS Trust, Adult Critical Care, Manchester, United Kingdom Correspondence: A. Wilson – Central Manchester Foundation NHS Trust, Adult Critical Care, Manchester, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A203

A202 To investigate the effect of increased ICU demand on the quality of care provided S. Grewal1, S. Gopal1, C. Corbett2 1 New Cross Hospital, Integrated Critical Care Unit, Wolverhamptom, United Kingdom; 2New Cross Hospital, Critical Care Services, Research and Audit Office, Wolverhamptom, United Kingdom Correspondence: S. Grewal – New Cross Hospital, Integrated Critical Care Unit, Wolverhamptom, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A202 Introduction The demand for critical care services is increasing steadily both nationally and internationally. It is important that with increasing demand the quality of care delivered is maintained. A well-recognised quality marker for intensive care units(ICU) are readmissions within 48 hours of discharge[1]. ICU readmission is associated with significantly increased morbidity, mortality, prolonged hospital admissions and increased cost[2,3]. The reported average of unplanned readmissions is 1.4 % in the UK[4]. Globally this varies between 1.3-13.7 %[2]. There have been no specific causal factors found[2]. Objectives Identify the 48 hour readmission rate over two consecutive years and its impact on patient care. Method This was a retrospective study in a 16 bedded adult general medicosurgical ICU in a large acute hospital in England. The sample was collated from the ICU discharges in 2014 and 2015, identifying those patients who were readmitted within 48 hours of discharge. Specific times of discharge and readmission, underlying cause, length of readmission stay and final outcome data was collected. The results of 2014 and 2015 were compared. Results In 2014, 672 patients were discharged from ICU, with 21 readmissions and only 2 within 48 hours (0.3 %). The time of re-admission

References 1. The Faculty of Intensive Care Medicine: Core Standards for Intensive Care, Edition1 (2013) 2. Elliott M, et al. Intensive Care Readmission: A contemporary review of literature. Intensive Crit Care (Nurse) 2013. http://dx.doi.org/10.1016/ j.iccn.2013.10.005 3. Roseberg AL et al. Patients readmitted to ICUs: A systematic review of risk factors and outcomes. Chest 2000;118(2):492-502. 4. ICNARC Case Mix Programme 2012-13

Introduction There is growing evidence that a protective lung ventilation strategy offers benefits to all patients undergoing mechanical ventilation[1]. In addition, the deleterious effects of excess oxygen in critically ill patients are becoming increasingly well recognised[2]. We present the results of a 10-month-long quality improvement project that has improved adherence to a lung protective ventilation strategy on our unit. The project has also yielded a culture shift in the way that oxygen is prescribed and titrated for all our patients. Objectives 1. Record the height of all critical care patients and calculate their ideal body weight (IBW). 2. Adopt a lung protective (6-8 ml/kg IBW) ventilation strategy for all patients receiving controlled ventilation. 3. Titrate oxygen delivery according to prescribed saturation targets (90 % or 95 %). Method Our project consisted of a series of tools to raise awareness, educational interventions, daily prompts and practical measures to promote adherence. In addition we adopted a pressure-control, volumeguarantee mode of ventilation as the initial standard for all patients ventilated on our unit. Our critical care technologists offered ongoing practice education for members of the clinical team whilst a run-chart of week-by-week performance offered immediate feedback on the status of the project.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Results We have reviewed over 4200 patient-days of data in a rolling audit of the effectiveness of our project. Fig. 88 demonstrates the percentage of patients for whom an IBW was recorded and for whom a target oxygen saturation was prescribed each month. Our latest results (April 2016) showed that 88 % (75/85) of patients on our unit had their IBW recorded and 100 % (14/14) of those receiving controlled ventilation had an appropriate tidal volume prescribed. Ventilation matched the prescription in 71 % (10/14) of cases. Routinely knowing the IBW of every patient on the unit has also been useful for the safe prescribing of haemofiltration and critical care drugs. 76 % (65/85) of patients had target oxygen saturations prescribed although only 38 % (25/65) were 'on target' at the time of data collection. Most importantly, prescribing an oxygen saturation target has empowered our nursing staff to wean inspired oxygen concentrations without waiting for an arterial blood gas. Conclusions Our project has demonstrated that a series of simple interventions can help to optimise mechanical ventilation and oxygen titration within critical care. There have been associated benefits in terms of safe prescribing and a reduction in our reliance on arterial blood gas analysis. References [1] Serpa Neto A. Association between use of lung-protective ventilation with lower tidal volumes and clinical outcomes among patients without acute respiratory distress syndrome. JAMA. 2012;308(16):1651-1659. [2] Ridler N et al. Oxygen therapy in critical illness: friend or foe? JICS. 2014;15(3):190-196.

Page 109 of 607

control, 30° head elevation, ETT fixation, gastric residue control, ETT with polyurethane cuff, closed airway suctioning 2009: stop closed airway suctioning, ETT with PVC tapered shaped cuff and subglottic suctioning, digital continuous cuff pressure controller, automated subglottic suctioning pump 2011: oral care with chlorhexidine 0,2 % and oral care system permitting social control 2012: Belgian VAP-bundle, with continuous compliance measurement 2014: RASS to adjust the depth of sedation Results 2008 measures were ineffective. Nurses identified ineffective tracheal suctioning by closed airway suctioning as a major problem. In 2009, stopping closed airway suctioning and the Introduction of 3 new technologies, reduced VAP-rate by half. In 2011, oral care with chlorhexidine 0,2 % and an oral care system permitting social control did not improve results. Although almost all VAP-bundle elements already were used in our ICU, the formal implementation of the Belgian VAP-bundle in 2012 caused an important additional decline in VAP-rate, emphasizing the importance of sedation control and compliance measurement. In 2014, adjusting depth of sedation by RASS resulted in the lowest VAP-rate in 10 years. Conclusions This continuous quality improvement program shows that the systematic use of evidence based and PDCA-cycle driven process improvements may contribute to reduce VAP-rate. References 1. Blot S, Poelaert J, Kollef M. How to avoid microaspiration? A key element for the prevention of ventilator-associated pneumonia in intubated ICU patients. BMC Infectious Diseases. 2014;14:119. 2. How to improve. Institute for Healthcare Improvement. http://www.ihi.org/ resources/Pages/HowtoImprove/default.aspx Accessed 19/4/2016.

Fig. 88 (abstract A203).

A204 From zero to hero: a history of 10 years of quality improvement in VAP prevention W. Swinnen, J. Pauwels, K. Mignolet, E. Pannier, A. Koch, T. Sarens, W. Temmerman az Sint-Blasius, Dendermonde, Belgium Correspondence: W. Swinnen – az Sint-Blasius, Dendermonde, Belgium Intensive Care Medicine Experimental 2016, 4(Suppl 1):A204 Introduction VAP is an ICU-specific nosocomial infection, causing important additional costs in healthcare, extending ICU and hospital length of stay, with its own attributable mortality. A sudden rise in VAP-rate in 2007 urged the ICU of the az SintBlasius to improve quality of care. Objectives To reduce VAP-rate by systematical Introduction of evidence based measures supported by PDCA-cycle Methodology. Method Process improvements included: 2008: tracheal suction protocol, prevention of colonization of ventilator tubing, HME change, oral care with hexetidin, cuff pressure

Fig. 89 (abstract A204). az Sint-Blasius VAP-rate A205 The impact of goal directed morbidity and mortality conferences on patient outcome, 3 year study A.M. Elmenshawy, A.M. Fayed, M. Elboriuny, E. Hamdy, E. Zakaria Alexandria University, Critical Care Medicine, Alexandria, Egypt Correspondence: A.M. Elmenshawy – Alexandria University, Critical Care Medicine, Alexandria, Egypt Intensive Care Medicine Experimental 2016, 4(Suppl 1):A205 Introduction Morbidity and mortality conferences (MMCs) are a traditional tool of improving local care management, and clinical management education especially in high risk specialty, but they lack a precise format for practice in intensive care units (ICU).

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Objectives To explore the impact of goal directed MMCs on mortality rate and adverse events. Method 3 year prospective study in 2 ICUs (15 bed each) for evaluating a systematized MMCs with a clear goal of improving local care (how to avoid this) through case discussion, analysis, brainstorming and clear recommendations in a multidisciplinary meeting involving all related caregivers in a blame free environment. The present study included 4 phases; pre-intervention (Jan-12-April-13), intervention (May-July 2013), post-intervention (Aug-13-Oct-14), washout period (Nov-14Dec-15). A period prevalence of ICU acquired adverse events were audited in March and December 2013. Results 4589 patients were included in the study, from which 1456 died (31.7 %). During the intervention and post-intervention phases, 18 MMCs were held which discussed 36 died cases (28 from ICU3 and 8 from ICU1), reviewed other causes of death and unit performance indicators and made 96 recommendations (80 % accomplished in ICU3 versus 60 % in ICU1). The mortality rate decreased in postintervention phase from 44.2 % to 25.1 % in ICU3 (p < 0.001) and from 21 % to 20.5 % in ICU1 (p = 0.748) then increased in washout period to 28.3 % in ICU3 (p = 0.189) and to 25.2 % in ICU1 (p = 0.041). Conversely, the mortality rate in washout period (in relation to pre-intervention phase) decreased in ICU3 (p < 0.001) but increased in ICU1(p = 0.036). Subanalysis in ICU1 mortality revealed significant decrease in early post-intervention (17.3 % with p = 0.031) and significant increase in late post-intervention (24.4 % with p = 0.028). Adverse events decreased insignificantly in December 2013 (as compared to April 2013); unplanned extubation from 71 to 48/ 1000 ventilator days ( 24 % to 15.5 %) in ICU3, unexpected cardiac arrest from 17.7 to 12.7/1000 patient days (8.6 % to 6.2 %) in ICU1, stress ulcer bleeding from 17 to 4.5/ 1000 patient days (12.6 % to 4.4 %) in ICU3 and 25 to 18/1000 patient days (12 % to 8.7 %) in ICU1, Deep vein thrombosis from 5.7 to 0/1000 patient days (4 % to 0 %) in ICU3 and from 12.6 to 5/1000 patient days (6 % to 2.5 %) in ICU1, iatrogenic pneumothorax from 5.7 to 0/1000 patient days (4 to 0 %) in ICU3 and from 5 to 7.6/1000 patient days (from 2.5 to 3.8 %) in ICU1. However, unexpected cardiac arrest (from 26.7 to 6.8/1000 patient days, p = 0.019) and VAP rate (from 10 to 2.5/1000 ventilator days, p = 0.002) in ICU3 and unplanned extubation (from 105 to 33/1000 ventilator days, p = 0.003) in ICU1 decreased significantly in Dec-13. Conclusions Although confounding factors were not controlled, goal oriented MMC reduced some adverse event and probably mortality rate especially with continuous monitoring and improvement.

Fig. 90 (abstract A205). [ICU3 mortality]

Page 110 of 607

Fig. 91 (abstract A205). [ICU1 mortality]

A206 Quality of nursing and care burdened measurement during intensive care in relation to patient outcome A.-C. Falk Karolinska University Hospital, CIVA, Stockholm, Sweden Intensive Care Medicine Experimental 2016, 4(Suppl 1):A206 Introduction Intensive care is one of the most resource-intensive forms of medical care due to severely ill patients. During recent years the quality of intensive care has been in focus however there is still lacking result from nursing point of view. Objectives To describe nurse/patient ratio in relation to Care burdened measurement (VTS) and optimal medical and nursing-related result usually used indicators as mortality and complications during intensive care. Method This is a retrospective registry study includes a survey of critical care of registry data (all patients > 15 years) receiving care in two general Level I critical care units with similar rate of admissions during 20102014. Data of nurse/patient ratio is collected from each unit. The data is analyzed by descriptive and comparative statistical Method Results The result showed differences in specialized nurse/patient ratio of 0,5:1 to 1:1 ratio and Care burdened measurement (VTS) despite similarities in admission rate. Differences in cause of admission (surgicalv.s medical) and in the amount of unexpected surgery patients were found. Differences were also found in mean time on noninvasive ventilation and mean time on ventilator. Complications during critical care was measured by readmission and unplanned reintubation and showed that unplanned reintubation varied between 2.41.6 percent. ICU mortality showed differences with the lowest ICU mortality in the hospital with lower nurse/patient ratio. However, 30 days mortality was lower in the hospital with higher nurse/patient ratio. Further analysis is needed. Conclusions Preliminary results show differences in nurse/patient ratios and Care burdened measurement (VTS) with differences in and quality measurements in general critical care units. References 1. Soini K, Stiernström H (red.Larsson & Rubertsson) (2005) Intensivvård. Liber. Torino, Italien 2. Institute of medicine (2001). Crossing the quality chasm: a new health system for the 21 st century. Washington DC: National Academy Press

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

3. Aiken L et al (2014). Nurse staffing and education and hospital mortality in nine European countries: a retrospective observational study. Lancet http://dx.doi.org/10.1016/s014-6736(13)62631-8 4. McGahan M, Kucharski G, Coyer F (2012). Nursing staffing levels and the incidence of mortality and morbidity in the adult intensive care unit: a literature review. Australian critical care 2012; 25,64-77

A207 Adverse events measured by quality indicators in intensive care, an observational study A. Petosic1, K. Olafsen2, H. Wøien3, H. Flaatten4, K. Sunde2 1 Oslo University Hospital, General Intensive Care Unit - Ullevål, Oslo, Norway; 2Oslo University Hospital, Department of Anaesthesiology, Oslo, Norway; 3Oslo University Hospital, General Intensive Care Unit 1 -Rikshospitalet, Oslo, Norway; 4Haukeland University Hospital, General Intensive Care Unit, Bergen, Norway Correspondence: A. Petosic – Oslo University Hospital, General Intensive Care Unit - Ullevål, Oslo, Norway Intensive Care Medicine Experimental 2016, 4(Suppl 1):A207 Introduction The use of quality indicators (QIs) for improvement of care in the intensive care unit (ICU) is increasing. Incidence of adverse events like ventilator associated pneumonia (VAP) and pressure ulcers are not routinely measured in norwegian ICUs. Objectives The aim of this pilot study was to evaluate the frequency of these two adverse events measured by specific QIs in two ICUs at the Oslo University Hospital Ullevål (OUHU). Method ICU-patients at two different ICUs at OUHU were included in the study in a predefined time period during Autumn 2015. Data for determining VAP and pressure ulcers were retrieved from the daily electronic patient record, journal and/ or bedside on a specific study sheet. Pneumonia (PN) was identified using the European Centre for Disease Prevention and Control (ECDC) protocol; -a combination of imaging, clinical and laboratory criteria. VAP was defined as a PN where the patient had been on mechanical ventilation for more than 48 hours and ventilator in place on the day of event or the day before. Pressure ulcers were identified using the European Pressure Ulcer Advisory Panel (EPUAP) classification system. To determine the severity of the pressure ulcer, four categories were used. Descriptive statistics are presented using SPSS version 21. Results are presented as mean ± standard deviation (SD). Results We included 58 adult ICU patients, all mechanically ventilated for a minimum of 48 hours, of whom 93 % were surgical patients. Among those, 69 % were trauma patients and 12 % acute surgery patients. Mean SAPSII score was 32,3 ± 14,1 and mean time on mechanical ventilation was 11,2 days ± 9,6. PN was present in 28 patients (48 %), and 19 (33 %) had per definition VAP. Per ECDC classification 19 of the 28 PNs were classified as PN1 with microbiology confirmation from BronchoAlveolar Lavage (BAL), five PN2 with microbiology confirmation from endotracheal aspirates, and four PN5 without microbiology confirmation (clinical pneumonia). The incidence of VAP was 29/1000 ventilator days. Pressure ulcers were present in 18 (31 %) of the patients, and the majority of patients had more than one ulcer. The ulcers were classified as EPUAP category one (57 %), two (34 %), three (4 %) and four (1 %). The incidence of pressure ulcers was 70/ 1000 ICU days. Conclusions At OUHU there is a potential for reducing the incidence of VAP and pressure ulcers if we are to reach the current standard of care. Identification of specific QIs is important for future improving quality of care. Grant acknowledgment Oslo University Hospital.

Page 111 of 607

A208 Analysis of the perception and evaluation of the heathcare quality and the process of death: an empirical study from the perspective of families of deceased patients in an intensive care unit (ICU) J.J. Cáceres Agra1, J.L. Santana Cabrera1, J.D. Martín Santana2, L. Melián Alzola2, H. Rodríguez Pérez1 1 Hospital Insular Las Palmas GC, Las Palmas de Gran Canaria, Spain; 2 Hospital Insular Las Palmas GC, Department of Economy. University of Las Palmas de Gran Canaria, Las Palmas de Gran Canaria, Spain Correspondence: H. Rodríguez Pérez – Hospital Insular Las Palmas GC, Las Palmas de Gran Canaria, Spain Intensive Care Medicine Experimental 2016, 4(Suppl 1):A208 Introduction Care that takes the needs of families into account is very important, but in order to offer family-centered care, it is necessary to understand families experiences. Objectives To develop a valid and reliable tool to measure the perceived quality of care, the quality of the process of death and the satisfaction generated from the perspective of relatives of deceased patients in an ICU. Method Elaboration of an ad hoc questionnaire after literature review and validated by a multidisciplinary panel of experts. The construct of quality of service (QS) was measured by using three different constructs: quality of the communication (QS1), with two dimensions (“kindness and respect” and “sincerity and empathy”); quality of the information (QS2), with four dimensions (“welcome”, “information about disease”, “shared decission making” and “empathy with family needs”); technical and infraestructural quality (QS3), with six dimensions (“human aproach, “professionalism”, “waiting room”, “facilities”, “visiting hours” and “meals”). The construct about the process of death was measured by using three dimensions (“technical quality”, “human quality” and “quality of the information given”. Finally, the construct about overall satisfaction with the service was measured with three dimensions about satisfaction with “form and content of the information”, “technical and structural quality”, and “process of death”. 95 relatives of deceased patients participated with a post mail questionnaire (response rate: 38,9 %). After a month since the death, recruitment by telephone was carried out. Analysis of psychometric properties: convergent validity with correlation coefficient of Pearson, predictive validity with multiple regression models in order to predict the dimensions of satisfaction from the scales of quality, and reliability with α Cronbach and test of two halves. Results Nearly all of the correlation coefficients between the items of each construct were > 0,5, which meant, therefore, that all constructs had convergent validity. The three regression models performed to evaluate the predictive validity showed high determination coefficients (78,2 %. 94,9 % and 75,4 %). The evaluation of the items of all dimensions were high, with mean values around 6, except dimensions “empathy with family needs” and “waiting room”, with scores < 5. Conclusions The ad hoc tool developed was easy to use and showed adequate psychometric properties of validity and reliability which could be improved by removing some items that showed low correlations. References 1. Van den Broek JM, Brunsveld-Reinders AH, Zedlitz AM, et al. Questionnaires on Family Satisfaction in the Adult ICU: A Systematic Review Including Psychometric Properties. Crit Care Med. 2015; 43(8):1731-44. 2. Wall RJ, Curtis JR, Cooke CR, et al. Family satisfaction in the between families of survivors and nonsurvivors. Chest. 2007; 132(5): 1425-33.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Page 112 of 607

A209 Reducing antibiotic resistance - watch out for protocols! T. Castro Pires1, H. Calderón2, A. Pereira3, S. Castro2, C. Granja1,2,4 1 University of Algarve, Departmant of Biomedical Sciences and Medicine, Faro, Portugal; 2Centro Hospitalar do Algarve, Hospital de Faro, Emergency and Intensive Care Departmant, Faro, Portugal; 3Centro Hospitalar do Algarve, Hospital de Faro, Faro, Portugal; 4Faculty of Medicine of Porto, CINTESIS, Porto, Portugal Correspondence: T. Castro Pires – University of Algarve, Departmant of Biomedical Sciences and Medicine, Faro, Portugal Intensive Care Medicine Experimental 2016, 4(Suppl 1):A209

A210 Impact of preoperative health related quality of life on outcomes after cardiac surgery I. Norkiene1, I. Urbanaviciute2, G. Kezyte2, D. Ringaitiene1, T. Jovaisa3 1 Vilnius University, Clinic of Anaesthesiology and Intensive care, Vilnius, Lithuania; 2Vilnius University, Faculty of Medicine, Vilnius, Lithuania; 3 Lithuanian University of Health Sciences, Clinic of Anaesthesiology, Kaunas, Lithuania Correspondence: I. Urbanaviciute – Vilnius University, Faculty of Medicine, Vilnius, Lithuania Intensive Care Medicine Experimental 2016, 4(Suppl 1):A210

Objectives The purpose of this study was to evaluate the profile of antibiotic resistance and the effect of the Introduction of an antibiotic protocol in the ICU length of stay and mortality, costs and consumption of antibiotics. Method We conducted a retrospective study from a total of 476 patients admitted with in the ICU between January 2015 and December 2015. For the purpose of this study, and to obtain a profile of the microorganisms in our ICU, resistant microorganisms (RM) were defined as those with non-susceptibility to one (RM1) or more than one class of antibiotics (RM2).We also compared our results whit previous studies conducted in the ICU. Results Most prevalent microorganisms were: Staphylococcus aureus, Escherichia coli, Pseudomonas aeruginosa and Klebsiella pneumonia, 33 % of all microorganisms identified were resistant to more than one class of antibiotics (RM2) and 17 % of the patients had a RM1.Resistance was statistically significantly (p < 0,05) associated with:Previous stay in wards - in particular patients coming from surgical wards or Emergency;Number of days from admission in hospital to ICU admission - patients with RM2 had a mean hospital stay before ICU admission of 12 days;Number of previous surgeries in the same hospital event - RM2 were identified in patients with an average of two previous surgeries;Admission type - mainly patients coming from urgent surgery and medical causes;Number of days in ICU - patients with RM2 stay in average at least 5 days more in ICU than patients with microorganisms with no resistance;Antibiotic treatment previous to ICU admission (for at least 5 days before ICU admission), more than 30 % of patients with RM2 were submitted to antibiotic treatment before ICU admission.After the implementation of the antibiotic protocols we verified that the ICU length of stay was reduced in almost 3 days; the number of agents isolated increased 33 %; microbiological analyses increased 53 % which allowed us the practice of de-escalation consumption of antibiotics, by category, decreased by 82 % for the carbapenems, 33 % for antifungals and 35 % for antipseudomonal beta-lactams.There was a significant decrease in mortality, from 31,8 % to 26,8 %;There was a 22 % reduction in antibiotics consumption, with a total saving of 65,862€ in one year. Conclusions Resistance was associated with previous stay in wards, previous surgery and antibiotic treatment previous to surgery. The implementation of antibiotic protocols has had a positive impact as it was significantly associated with reduction on the length of ICU stay and ICU mortality, as well as consumption and costs associated with antibiotic therapy

Introduction Advances in cardiac surgery techniques and postoperative care had led to decrease in major postoperative morbidity and mortality. Since operative risk declines, improvement of quality of life became an important patient centred outcome. It depends not only on physical but also on mental status and individual perception of health and rehabilitation. Objectives The primary aim of our study was to define the impact of preoperative patient characteristics and quality of preoperative life on Health Related Quality Of Life (HRQOL) dynamics one year after cardiac surgery. Secondary aim was to identify factors influencing long-term HRQOL postoperatively. Method A prospective cohort study in a tertiary referral university hospital. Study protocol was approved by institutional bioethics committee. The 36-item Short Form Health Survey (SF-36) was used to assess HRQOL amongst the study participants. An inpatient SF-36 questionnaire was completed a day before the elective surgery and repeated one year after surgery via telephonic interview. SF-36 is composed of 8 domains covering physical and mental aspects of health. Answers are converted to scale from 0 to 100 where higher values represent better health status. Effect Size Method was used to establish clinically significant change, therefore improvement was defined as positive change exceeding 1 Standard Deviation between baseline and follow up score in particular domain. Summary scores of Physical Component (PCS) and Mental Component (MCS) were used to identify patients with overall improvement in HRQOL. Based on these results study group was divided into improvers and non-improvers. Clinical data was collected from medical records. Independent samples and paired samples T-tests were used to compare baseline and follow-up data and differences between two groups. Results 210 patients were enrolled in the study and underwent surgery in 2013-2014. After one year we were able to contact 53.8 % patients, hence final analysis included 105 patients. Overall positive significant change was identified in half of the domains, with significant improvement in PCS and MCS. Mean PCS at baseline was 42.5 (13.13-91.25) and 58.75 (23.13-92.5) at followup; MCS 52.75 (24.38-91) at baseline and 62 (24.5-99) at followup, p < 0,001 for both. There were 51 patients in non-improvers group and 54 patients in improvers group. No significant differences were identified between groups in demographic and peri-operative variables. Non-improvers had significantly higher preoperative scores in all domains including MCS (60.59 ± 14.28 vs 44.09 ± 12.09, p < 0,001) and PCS (54.36 ± 14.94 vs 37.93 ± 11.83 p < 0,001). Conclusions Our findings suggest that patients with worse HRQOL at baseline are more likely to experience long-term improvement following cardiac surgery. We did not identify any other factors influencing long-term HRQOL outcomes.

References 1. Spellberg B. (2014). The future of antibiotics. Critical Care 18:228; 2. European Centre for Disease Prevention and Control. Antimicrobial resistance surveillance in Europe 2014. Annual Report of the European Antimicrobial Resistance Surveillance Network (EARS-Net). Stockholm: ECDC; 2015

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

A211 Health-related quality of life after surgical intensive care G. Vogel1, U.-B. Johansson1,2, A. Sandgren3, C. Svensen1, E. Joelsson-Alm1 1 Karolinska Institutet, Department of Clinical Science and Education, Södersjukhuset, Stockholm, Sweden; 2Sophiahemmet University, Stockholm, Sweden; 3Linnaeus University, Kalmar/Växjö, Sweden Correspondence: G. Vogel – Karolinska Institutet, Department of Clinical Science and Education, Södersjukhuset, Stockholm, Sweden Intensive Care Medicine Experimental 2016, 4(Suppl 1):A211 Introduction Patients who have been treated at an ICU often have affected mental and physical health with complications such as depression, anxiety, posttraumatic stress disorder (PTSD) and, sleep disturbances[1-4]. Decreased health-related quality of life (HRQoL) after ICU stay have been shown in mixed ICU-populations but there is a lack of knowledge of HRQoL after surgical intensive care. Objectives The aim of this study was to describe HRQoL 3, 6 and 12 months after discharge from a general surgical ICU and to analyze factors associated with impaired HRQoL. Method Prospective cohort study in a general surgical ICU in Sweden. Included are patients with an ICU length of stay ≥ 96 hrs, < 18 years old, 20042012. HRQoL was measured with SF-36 at 3, 6 and 12 months after discharge from the ICU. Age, gender, APACHE II, LoS, mechanical ventilation, admission diagnosis, preexisting disease and marital status were recorded. Wilcoxon Signed Ranks Test was used for comparing HRQoL and changes over time. A general linear regression analysis was performed to analyze impact of background and ICU-related factors on HRQoL at 12-months, standardized to a general population in the same age in Sweden. Results A total of 276 (62 %) patients 18 - 89 yrs were included. HRQoL significantly improved between 3-12 months after ICU in six domains; PF (p = 0.00), RP (p = 0.00), GH (p = 0.03), VT (p = 0.00), SF (p = 0.00) and RE (p = 0.01), but were still lower in all eight domains compared to standard population. Age < 75 years was a risk factor. Conclusions Surgical intensive care patients have lower HRQoL compared to standard population one year after stay in the ICU. Age < 75 yrs is a risk factor for worse HRQoL. This study also contributes to the knowledge of risk factors for impaired HRQoL which can be important when considering selection of patients to the ICU follow-up clinics. References 1. Jackson JC, Pandharipande PP, Girard TD, et al. Depression, post-traumatic stress disorder, and functional disability in survivors of critical illness in the BRAIN-ICU study: a longitudinal cohort study. The Lancet Respiratory medicine 2014; 2(5): 369-79. 2. Deja M, Denke C, Weber-Carstens S, et al. Social support during intensive care unit stay might improve mental impairment and consequently health-related quality of life in survivors of severe acute respiratory distress syndrome. Critical care (London, England) 2006; 10(5): R147. 3. Jones C, Backman C, Capuzzo M, Flaatten H, Rylander C, Griffiths RD. Precipitants of post-traumatic stress disorder following intensive care: a hypothesis generating study of diversity in care. Intensive care medicine 2007; 33(6): 978-85. 4. Parsons EC, Hough CL, Vitiello MV, Zatzick D, Davydow DS. Insomnia is associated with quality of life impairment in medical-surgical intensive care unit survivors. Heart & lung: the journal of critical care 2015; 44(2): 89-94.

A212 Pulmonary function and quality of life in a Brazilian post-ICU follow-up outpatient clinic M.A. Leite, L.D. Murbach, E.F. Osaku, C.R.L.M. Costa, M. Pelenz, N.M. Neitzke, M.M. Moraes, J.L. Jaskowiak, M.M.M. Silva, R.S. Zaponi, L.R.L. Abentroth, S.M. Ogasawara, A.C. Jorge, P.A.D. Duarte Western Parana State University Hospital, Cascavel, Brazil Correspondence: M.A. Leite – Western Parana State University Hospital, Cascavel, Brazil Intensive Care Medicine Experimental 2016, 4(Suppl 1):A212

Page 113 of 607

Introduction Critical illness survivors have physical impairment, reduced functional capacity and change in quality of life after discharge from the Intensive Care Unit (ICU). Objectives To evaluate pulmonary function and quality of life of critically ill patients three months after discharge from the ICU. Method Cohort study conducted in the Post-ICU Follow-Up Outpatient Clinic of an University Hospital, Southern Brazil, from April 2012 to June 2013. It was performed spirometry test and the patients answered SF-36 after three months of discharge from ICU. The patients were divided into 3 groups: normal spirometry (G1), restrictive (G2) and obstructive (G3). The variables that were in accordance with the assumptions of normality and homoscedasticity were evaluated by analysis of variance (ANOVA single-factor) and those that were not in accordance with the assumptions were evaluated by Kruskal-Wallis test. Results 609 patients were admitted, 440 were discharged alive from the hospital. In the Outpatient Clinic, 126 patients were evaluated, 61 % male, age 44 ± 18.19 years, 48 % were smokers and 8 % COPD. The main causes of admission were postoperative elective surgery (26 %) and trauma with head injury (20 %), mean APACHE II 19.08, ICU length of time 7.23 days, hospital length of time 17.75 days, MV > 24 hours 47.62 %, and mean MV time 86.45 hours. The groups showed significant differences in spirometric data, shown in Table 70. In the SF-36 only the areas vitality and social aspects showed statistical difference (Table 71). Conclusions Most of the patients had normal spirometry. The main pulmonary impairment was obstructive feature. Some areas of quality of life after ICU discharge present relationship with worsening of lung function. Table 70 (abstract A212). Spirometry Results Spirometric variables

Normal (n=79) Restrictive (n=10)

Obstructive (n=35)

p

FEV1

2.98 + 0.81

2.21 + 0.67

2.24 + 1.01

0.00002

PEF

375.84 + 128.92

321.79 + 130.04 282.37 + 139.70

0.001

FVC

3.60 + 1.01

2.48 + 0.69

0.001

3.48 + 1.36

Table 71 (abstract A212). Results of SF-36 SF-36

Normal (n=79)

Restrictive (n=10)

Obstructive (n=35)

p

Functional Capacity

65.57 ± 110.34

47.86 ± 27.99

33.50 ± 30.32

0.09

Physical Limitations 23.73 ± 35.33

25.00 ± 34.26

22.50 ± 36.88

0.91

Pain

24,17 ± 49.93

42.20 ± 33.10

0.23

34.53 ± 38.81

General Health

68.51 ± 24.07

62.57 ± 25.76

62.40 ± 25.84

0.20

Vitality

68.10 ± 25.87

58.00 ± 27.11

57.50 ± 26.46

0.04*

Social Aspects

72.66 ± 34.12

64.40 ± 33.92

46.40 ± 35.21

0.04*

Emotional Aspects

35.43 ± 40.81

38.11 ± 47.14

33.30 ± 45.14

0.96

Mental Health

67.44 ± 26.75

65.14 ± 26.94

65.20 ± 30.49

0.80

A213 Quality of life of critical patients three months after ICU discharge L.D. Murbach, M.A. Leite, E.F. Osaku, J. Barreto, S.T. Duarte, S. Taba, D. Miglioranza, D.P. Gund, C.F. Lordani, C.R.L.M. Costa, S.M. Ogasawara, A.C. Jorge, P.A.D. Duarte Western Parana State University Hospital, Cascavel, Brazil Correspondence: L.D. Murbach – Western Parana State University Hospital, Cascavel, Brazil Intensive Care Medicine Experimental 2016, 4(Suppl 1):A213

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Introduction The evolution of care to critically ill patients has provided a considerable number of survivors after admission to the ICU, but several patients have impaired quality of life owing to changes in physical, functional, social and emotional disorders that often culminate in a prolonged recovery. In low-income countries this picture is poorly described. Objective To evaluate health-related quality of life in patients that survived to ICU stay three months after ICU discharge. Method Prospective cohort study conducted in the period 2012-2014, at the Post-ICU Multiprofessional Follow-up outpatient Clinic of a teaching hospital, in Southern Brazil. Included patients > 24 hours in the ICU, ≥ 18 years old and who attended the outpatient evaluation. To evaluate the quality of life it was applied the questionnaire SF-36. Results In the 2-year period, 275 patients were evaluated: 63.6 % male, mean age 44.0 years. Trauma was the leading cause of hospitalization (37.1 %). 45.5 % had comorbidities: hypertension (43.7 %), diabetes (32.4 %) and alcohol / tobacco (18.7 %). Mean ICU length of stay was 9.9 days; 74.9 % remained more than 48 hours with mechanical ventilation; 63.6 % were sedated and 25.8 % tracheostomy. Mean APACHE II 20.2.There was significant impairment in relation to the limitations of physical and emotional aspects. The main factors associated with change in QOL were: Physical Component: Mechanical Ventilation, Sedation and the presence of post-ICU sequelae; Mental Component: Admission by neuro-trauma and the presence of post-ICU sequelae. Conclusion There was impairment in quality of life related to the limitations of physical, social and emotional aspects, justifying the need for monitoring and rehabilitation after discharge, mainly the patients victim of neurological trauma and underwent sedation and prolonged MV.

Page 114 of 607

82.8. The highest score was in Honesty of Information (85.7); the lowest score of 76.9 was in Frequency of Communication with Doctors. 11 % of relatives felt they could have used more time when making decisions. 77 % felt they had been supported through the decisionmaking process; 6 % felt overwhelmed. 8 % felt out of control over the care of the patient, although this was not always negative - 'leaving it to the experts'. The free text responses were insightful and a valuable contribution to the data. Conclusions The low response rate and the potential for response bias due to the most dissatisfied relatives declining to respond limit this study. Its strengths are that it provides individualised feedback which can be used to enhance quality of care locally. Future quality improvement projects will target improving communication and include disseminating a study summary to staff members to highlight the importance of communication, more information in the waiting room on expectations in the intensive care unit and the use of family ward rounds. References 1) Wall RJ, Engelberg RA, Downey L et al. Refinement, scoring, and validation of the Family Satisfaction in the Intensive Care Unit (FS-ICU) survey. Crit Care Med 2007;35(1):271-9. 2) Wright SE, Walmsley E, Harvey SE et al. Family-Reported Experiences Evaluation study: a mixed-Method study to evaluate families' satisfaction with adult critical care services in the NHS. Health Services and Delivery Research 2015;3(45):ISSN 2050-4349

A214 Quality of care in the intensive care unit: more than just survival statistics. Measurement of family satisfaction using the FS(ICU) 24 questionnaire in a single UK centre H. Vollmer, M. Gager, C. Waldmann Royal Berkshire Hospital, Intensive Care Unit, Reading, United Kingdom Correspondence: H. Vollmer – Royal Berkshire Hospital, Intensive Care Unit, Reading, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A214 Introduction Patient experience is a key marker of healthcare quality and is a focus of the Department of Health. 1 in 5 critically ill patients die, those that survive often have little memory of their intensive care stay. Family members are a valuable proxy for patient experience in ICU. The Family Satisfaction in the Intensive Care Unit FS(ICU) 24 questionnaire was developed in Canada and is widely validated. It has 2 domains: Satisfaction with Care and Satisfaction with Decision Making[1]. The UK FREE study adapted the questionnaire and found it to have high internal consistency and criterion validity, but that the free text responses demonstrated scope for improvement [2]. Objectives To assess levels of satisfaction in family members of ICU patients to guide improvement in quality of care. Method 213 patients were admitted between 01/10/15 and 01/01/16. 210 FS(ICU)-24 questionnaires were sent to their next of kin in the month after death or ICU discharge. 19 questions were analysed using the 5-point Likert scale. Results 96 questionnaires were returned (45.7 %). 62.5 % of responders were female; the mean age was 60.4 (±15.7) yrs. 25 % of relatives had previous ICU experience and 71.6 % lived with the patient prior to their illness. 66 % were the patient's partner. Overall satisfaction was 86.1. Satisfaction with Care score was 87.6 The highest score was in Concern and Caring (95.3) and the lowest in Waiting Room Atmosphere (67.4). Satisfaction with Decision Making scored

Fig. 92 (abstract A214). Family satisfaction with decision making

Fig. 93 (abstract A214). Support during the decision making process

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Page 115 of 607

were evaluated as predictor of ICU mortality with OR 1.0737 (C.I. 0.8897; 1.2956) and OR = 1.5204 (C.I. 1.0962; 2.1086), respectively. Conclusions In critical ill patients of different etiology the multivariate projection Method represents a valuable technique to identify inflammatory patterns as predictor of outcome. References Mickiewicz B et al:Integration of metabolic and inflammatory mediator profiles as a potential prognostic approach for septic shock in the intensive care unit. Critical Care 2015;19:11 Helmy A et al:Recombinant human IL-I receptor antagonist promotes MI microglia biased cytokines and chemokines following human traumatic brain injury.JCBFM 2015;Dec1 Grant acknowledgment University of Torino MAZA_RILO_1601

Fig. 94 (abstract A214). Control over the care of your family member

A214 A multivariate projection method to investigate inflammatory mediator profiles in the early phase of critical illness A.T. Mazzeo1, R. Tesio1, C. Filippini1, M.E. Vallero1, C. Giolitti1, S. Caccia1, M. Medugno1, T. Tenaglia1, R. Rosato2, I. Mastromauro1, L. Brazzi1, P.P. Terragni1, R. Urbino1, V. Fanelli1, V.M. Ranieri3, L. Mascia4 1 University of Turin, Anesthesia and Intensive Care, Turin, Italy; 2 University of Turin, Psychology, Torino, Italy; 3Sapienza University of Rome, Anesthesia and Intensive Care, Rome, Italy; 4Sapienza University of Rome, Scienze e Biotecnologie Medico Chirurgiche, Rome, Italy Correspondence: A.T. Mazzeo – University of Turin, Anesthesia and Intensive Care, Turin, Italy Intensive Care Medicine Experimental 2016, 4(Suppl 1):A214 Introduction Critical illness of different etiology triggers an inflammatory cascade associated with organ dysfunction. However it is not known if the inflammatory pattern is disease specific or related to the severity of injury. We hypothesized that in critical ill patients of different etiology a specific pattern of cytokines measured in the early phase after ICU admission may predict mortality. Objectives Aim of this study was to investigate the early inflammatory profile in critical ill patients and its association with ICU mortality. Method Local Ethic Committee approved the study protocol. Critical ill patients admitted to ICU for sepsis-associated acute respiratory distress syndrome (ARDS), severe traumatic brain injury (TBI) and subarachnoid hemorrhage (SAH) were studied. Demographic data, severity indexes at admission and physiological variables were recorded. Blood samples for cytokines analysis were collected at days 1, 2 and 4 after admission. The cytokine analysis was performed with Bioplex technology and 27 cytokines were analysed. Statistical analysis: Multivariate projection technique was applied to analyse variation and collinearity within the cytokines dataset without a priori selecting potential relevant molecules. Principal component analysis (PCA) was used to identify principal components (PC) which account for the majority of the variation within the dataset. Results Eighty-six critical ill patients admitted for sepsis-associated ARDS (n = 36), severe TBI (n = 29), SAH (n = 21) were studied. Mean age was 52.8 ± 18, APACHE II 17.6 ± 2.9, SAPS II 44.9 ± 11.7, mean SOFA in the first week 8.3 ± 2.9; median ICU-length of stay was 19 days (range 14-31), median Hospital-LOS was 34 days (range 17-50), ICU mortality 31.7 %. Using PCA, the first five PCs generated by the model explained 65 % of the variation within the dataset. The first component is strongly correlated with the following cytokines - IL4, IL12, IL10, IL-1ra, FGF basic, IL13, GCSF, IL-1β, IL7, TNFα, IL5, MIP-1α, INFγ, IL8, IL2, IL17, IL6. Moreover the second component is strongly correlated with VEGF, IP-10, IL9. After correcting for age, ApacheII, SAPS and SOFA, the two principal components PC1 and PC2

CONTEMPORARY ISSUES IN ARF 0215 Inappropriate coding of chronic severe respiratory disease in a general adult intensive care unit J. Ballantyne, L. Paton, A. Mackay Queen Elizabeth University Hospital, NHS GG&C, Intensive Care Medicine, Glasgow, United Kingdom Correspondence: J. Ballantyne – Queen Elizabeth University Hospital, NHS GG&C, Intensive Care Medicine, Glasgow, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):0215 Introduction High quality clinical databases facilitate repeated evaluation of clinical practice and comparative audit. [1] However, inaccurate data entry can lead to inaccurate conclusions. Objectives We sought to determine whether patients in our intensive care unit (ICU) were appropriately coded as having chronic severe respiratory disease. Method We conducted a retrospective case note review of patients admitted to a general adult ICU between January 2009 and December 2014 and labelled with chronic severe respiratory disease. Severe respiratory disease was defined as “chronic restrictive, obstructive or vascular disease resulting in severe exercise restriction or documented chronic hypoxia, hypercapnia, secondary polycythaemia, severe pulmonary hypertension or respirator dependence”. Patients were identified from the WardWatcher™ database, electronic records interrogated and statistical significance assessed using the Paired T-test and the Chi Squared test. Results There were 112 ICU patients (48 men, 43 %) labelled with chronic severe respiratory disease over this six year period. These patients had a median age of 62 years (IQR 54-70 years) and median APACHE II score of 20 (IQR 16-28). Electronic records were insufficient to confirm or refute the label of chronic severe respiratory disease in 13 cases and these patients were excluded from analyses. Forty-two of the remaining 99 patients (42 %) did not have chronic severe respiratory disease by the standardised definition. There was no significant difference in APACHE-II, ICU mortality, ICU length of stay, hospital mortality or hospital stay. The standardised mortality ratio was 0.752 (95 % CI 0.748-0.756) for appropriately coded patients and 0.855 (95 % CI 0.848-0.862) for inappropriately coded patients. Conclusions Almost half of patients are wrongly coded as having chronic severe respiratory disease. Moreover, this clearly impacts the SMR for this group of patients. Such errors in data entry compromise conclusions drawn from this cohort and comparisons made with prior cohorts and other ICUs. In order to improve this resource, stringent documentation by clinicians and education of those inputting the data is required. References 1. Harrision D, Brady A, Rowan K. Case mix, outcomes and length of stay for admissions to adult, general critical care units in England, Wales and Northern Ireland. Critical care. 2004; 8:R99-R111

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

A216 Prospective validation of right ventricular role in primary graft dysfunction after lung transplantation P. Perez-Teran1, O. Roca2, J.C. Ruiz-Rodriguez2, A. Zapatero1, J. Serra2, J.R. Masclans1 1 Hospital del Mar/CIBERES/UPF, Intensive Care, Barcelona, Spain; 2Vall d´ Hebrón University Hospital, Intensive Care, Barcelona, Spain Correspondence: P. Perez-Teran – Hospital del Mar/CIBERES/UPF, Intensive Care, Barcelona, Spain Intensive Care Medicine Experimental 2016, 4(Suppl 1):A216 Introduction Primary graft dysfunction is a significant cause of lung transplant morbidity and mortality, but its underlying mechanisms are not completely understood. Objectives Aims of the study: 1) to confirm that right ventricular function is a risk factor for severe primary graft dysfunction; 2) to propose a clinical model for predicting the development of severe primary graft dysfunction. Method A prospective cohort study was performed over 14 months. The primary outcome was development of primary graft dysfunction grade 3. An echocardiogram was performed immediately before transplantation, measuring conventional and speckle-tracking parameters. Pulmonary artery catheter data were also measured. A classification and a regression tree were made to identify prognostic models for the development of severe graft dysfunction. Results Seventy lung transplant recipients were included. Patients who developed severe primary graft dysfunction had better right ventricular function, as estimated by cardiac index (3.5 ± 0.8 vs. 2.6 ± 0.7 l/min*m2; p < 0.01) and basal longitudinal strain (-25.7 ± 7.3 vs. -19.5 ± 6.6 %; p < 0.01). Regression tree analysis provided an algorithm based on the combined use of three variables (Basal longitudinal strain, pulmonary fibrosis disease and ischemia time), allowing accurate preoperative discrimination of three distinct subgroups with low (11 to 20 %), intermediate (54 %) and high (75 %) risk of severe primary graft dysfunction (AUROC 0.81). Conclusions Better right ventricular function is a risk factor for the development of severe primary graft dysfunction. Preoperative estimation of right ventricular function could allow early identification of recipients at increased risk, who would benefit the most from careful perioperative management in order to limit pulmonary overflow. References 1. Pérez-Terán P, et al. Influence of right ventricular function on the development of primary graft dysfunction after lung transplantation. J Hear Lung Transplant 2015. 2. Liu Y, et al. Recipient-related clinical risk factors for primary graft dysfunction after lung transplantation: a systematic review and meta-analysis. PLoS One 2014;9(3). 3. Champion HC, et al. Comprehensive invasive and noninvasive approach to the right ventricle-pulmonary circulation unit: state of the art and clinical and research implications. Circulation 2009;120(11):992-1007. 4. Wrobel JP, et al. Preoperative echocardiographic-defined moderate-severe pulmonary hypertension predicts prolonged duration of mechanical ventilation following lung transplantation for patients with COPD. Lung 2012;190(6):635-643. Grant acknowledgment Dr. Pérez-Terán is the recipient of a “Rio Hortega” grant (Ref. CM12/00216) from the Instituto de Salud Carlos III, Ministerio de Ciencia e Innovación.

A217 Gas exchanges during whole lung lavage S. Bianzina, P. Cornara, G. Rodi, G. Tavazzi, M. Pozzi, G.A. Iotti, F. Mojoli, A. Braschi Anesthesia and Intensive Care, Fondazione IRCCS Policlinico S. Matteo, University of Pavia, Pavia, Italy Correspondence: S. Bianzina – Anesthesia and Intensive Care, Fondazione IRCCS Policlinico S. Matteo, University of Pavia, Pavia, Italy Intensive Care Medicine Experimental 2016, 4(Suppl 1):A217

Page 116 of 607

Introduction Pulmonary alveolar proteinosis (PAP) is a rare disorder characterized by a perturbation in surfactant homeostasis, resulting in its accumulation within alveolar spaces, with a consequent development of severe hypoxemia. Whole lung lavage (WLL) is a complex procedure, dedicated to those patients affected by a severe condition not responsive to medical treatment[1]. Objective To analyze the evolution of gas exchanges during WLL, evaluating PaO2 variations during the different phases of the procedure. Patients and method We enrolled 27 patients with PAP (16 males, age 15-64 years), who underwent WLL in our ICU between 2010 and 2015. WLL was conducted in general anesthesia, using selective endotracheal tubes. It consisted of different phases, for each lung: 1) bipulmonary mechanical ventilation in supine position at FiO2 1 and ZEEP; 2) bipulmonary ventilation in supine position with PEEP; 3) monopulmonary ventilation in supine position, in order to achieve complete atelectasis of the contralateral lung; 4) monopulmonary ventilation in lateral position, with the ventilated lung downward; 5) lavage of the whole atelectatic lung with liquid tidal ventilation at different levels of hydrostatic lavage pressure. We collected data regarding patient gas exchanges by performing several blood gas analysis during the different steps of WLL. We expressed data as mean ± standard deviation (SD). Results Figure 95 shows the mean values of PaO2 during the different phases of WLL. During bipulmonary ventilation gas exchanges improved in response to FiO2 1 and PEEP. Monopulmonary ventilation, instead, induced a clear reduction of PaO2, which increased in lateral position and during liquid tidal ventilation, with a substantial effect of elevated hydrostatic lavage pressures. The wide SD indicates an uneven response of gas exchanges in the studied population. Conclusions The pathophysiology of PAP is characterized by altered alveolocapillary diffusion and intrapulmonary shunt, thus significantly responsive to FiO2 and PEEP, respectively. During monopulmonary ventilation, shunt is reduced by the lateral position, which provides a better perfusion of the ventilated lung, and the elevated hydrostatic lavage pressures, which are able to limit perfusion in the contralateral lung under WLL. Reference 1. Pulmonary alveolar proteinosis: diagnostic and therapeutic challenges. Campo et al. Multidisciplinary Respiratory Medicine 2012, 7:4.

Fig. 95 (abstract A217).

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

A218 Study of clinico-epidemiological profile of patients during swine flu epidemic in 2015 at a tertiary care hospital in new Delhi A. Vishnu, D. Buche, R. Pande Blkapoor Super Speciality Hospital, BLK Centre for Critical Care, New Delhi, India Correspondence: A. Vishnu – Blkapoor Super Speciality Hospital, BLK Centre for Critical Care, New Delhi, India Intensive Care Medicine Experimental 2016, 4(Suppl 1):A218 Introduction WHO reported that influenza A(H1N1) pandemic in 2009 caused 4100 deaths globally[1]. New Delhi reported 2241 cases in 2015 with 8 confirmed deaths till March 2015 [2]. The mortality in severe refractory ARDS related to H1N1 influenza A pneumonia is high [3]. During 2015 swine flu epidemic our hospital admitted 162 adult patients who were either suspected or confirmed cases. Objectives To study the epidemiological and clinical profile of patients admitted to BLK Superspeciality Hospital, New Delhi during 2015 H1N1 influenza epidemic. Method Data was collected retrospectively from hospital records after IRB approval, for patients admitted either as a suspected or proven case (outside) of Influenza A (H1N1) pneumonia. Demographical details, category of presentation, hospital and ICU course and ventilation strategies applied were collected.18 patients were excluded as significant data was missing from records. Results We analysed 144 patients out of which, 66 % (n = 96) were from an urban background and 33 % (n = 48) were from a rural background. 56.9 % (n = 82) patients were H1N1 positive by RT-PCR technique. 24.4 % (n = 20) had a positive history of contact with established H1N1 cases, whereas history of travel to affected areas was present in 54.6 % (n = 45) patients. Maximum admissions were reported in the month of February (n = 47) and March( n = 28). 21.8 % patients ( n = 21 ) were belonging to age group 30 - 35 years and > 60 years age group.15.85 % (n = 13) of all the positive patients were of Category C, who were admitted to ICU directly. 71.6 % (n = 58) were managed with non invasive ventilatory support only, and 14.6 % (n = 12) were intubated and mechanically ventilated. 10 out of 12 patients requiring intubation and mechanical ventilation belonged to Category C. 75 % (n = 9) of intubated patients had refractory hypoxemia (mean PaO2/FIO2 ratio 61.33 ± 4.05), and 6 patients required early prone ventilation. 8.5 % (n = 7) patients died and all of them were mechanically ventilated. 89 % (n = 73) were successfully discharged from the hospital. Mean ICU and hospital LOS was 3.83 ± 4.06 days and 7.53 ± 4.2 days respectively. Higher mortality rates was also seen in those patients who had more than 3 days of time lag for initiation of oseltamivir after the symptom onset (n = 5). Conclusions Influenza A H1N1 pneumonia is a significant burden during epidemics associated with a high morbidity. The mortality among severe refractory ARDS patients who were invasively ventilated is very high. References 1. MOHFW, Press information Bureau Release, March 2015 2. CDC guidelines, 2009 3. Domínguez-Cherit G1, Lapinsky SE, Macias A et al. Critically Ill patients with 2009 influenza A(H1N1) in Mexico. JAMA. 2009 Nov 4;302(17):1880-7.

A219 Structural and organizational factors influencing mortality of severe community-acquired pneumonia patients in the intensive care D.L.J. Moolenaar1, F. Bakhshi-Raiez2,3, D.A. Dongelmans2,3, N.F. de Keizer2,3, D.W. de Lange1,3 1 University Medical Center Utrecht, Intensive Care, Utrecht, Netherlands; 2 Academic Medical Center, University of Amsterdam, Intensive Care, Amsterdam, Netherlands; 3Dutch National Intensive Care Evaluation Foundation, Amsterdam, Netherlands Correspondence: D.L.J. Moolenaar – University Medical Center Utrecht, Intensive Care, Utrecht, Netherlands Intensive Care Medicine Experimental 2016, 4(Suppl 1):A219

Page 117 of 607

Introduction Severe community-acquired pneumonia patients bear the highest morbidity, in-hospital mortality, and cost of all patients with community acquired pneumonia[1]. The appropriate management of these patients has received close attention in the current era of rising health care costs. Nevertheless, the outcome of these patients remains poor[2] and it is unknown which patient characteristics or treatment modalities are associated with a better outcome. By combining a national registry with a specifically designed questionnaire we were able to examine the factors related to organizational characteristics and treatment policies that might explain the variation in mortality outcomes in sCAP patients in the ICU. Method This study used a dataset from a national registry containing data on patient and ICU level combined with a web-based survey on treatment policies. The relationship between in-hospital mortality and determinants was analyzed using multivariable logistic regression analysis. Results From January 1st , 2008 to January 1st 2013, 79 Dutch ICUs provided data to the registry, from which 62 (78.5 %) returned the questionnaire. The responding ICUs admitted 11,245 patients with severe community-acquired pneumonia. These patients had a higher severity of illness than the general medical ICU population and their mortality was also higher (ICU mortality 17.8 %; hospital mortality 25.1 % versus ICU mortality 15.8 %; hospital mortality 22.5 %, respectively). Severe community-acquired pneumonia patients had more comorbidities and stayed longer at the ICU. Rare use of tracheostomies for the weaning process (OR:0.784(0.662-0.928)) was associated with better outcomes in these patients. The need for mechanical ventilation (OR:1.385(1.194-1.607)) was associated with poor outcomes as was the case for the mean number of ICU admissions per year (OR: 1.006 ( 1.001-1.010). Conclusions CAP is a disease with a high in-hospital mortality. We have shown that, after correction for confounding factors some treatment and organizational factors are related with outcome in patients with sCAP. The need for mechanical ventilation remained an independent risk factor for mortality, and liberal use of tracheostomies for the weaning process was also associated with higher mortality. The mean number of ICU admissions with sCAP per year was inversely correlated with hospital mortality. Many other parameters that are often claimed to be associated with better or worse outcome (like differences in antibiotic therapy) did not correlate with outcome. References 1) Mandell, L. A. et al. 2007. Infectious Diseases Society of America/American Thoracic Society consensus guidelines on the management of communityacquired pneumonia in adults. Clin.Infect.Dis. 44 Suppl 2:S27-S72. 2) Brinkman, S. et al. Mortality after hospital discharge in ICU patients. Crit Care Med. 41[5], 1229-1236. 2013. Grant acknowledgment None.

A220 Status asthma in intensive care unit. Experience of a third level center I. Fuentes Fernández, D. Martínez Baño, J.L. Buendía Moreno, R. Jara Rubio Hospital Virgen de la Arrixaca, ICU, El Palmar, Spain Correspondence: I. Fuentes Fernández – Hospital Virgen de la Arrixaca, ICU, El Palmar, Spain Intensive Care Medicine Experimental 2016, 4(Suppl 1):A220 Background and objectives Severe asthmatic patient who requires the use of invasive mechanical ventilation in an intensive care unit is, despite its low incidence, a potentially very serious case, which requires early and very specific respiratory care. This care will largely determine the average ICU stay, occurrence of secondary injuries or even death. Our objective is to describe the characteristics of severe asthmatic patient admitted to ICU and to analyze the initial treatment.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Method Descriptive study of patients admitted to the ICU of the Clinical Hospital Virgin of Arrixaca in the period between January 2011 and September 2015. We analyzed different variables, age, sex, APACHE II, SAPS II, parameters of mechanical ventilation, ICU stay and laboratory parameters. Results 18 patients (6 women) admitted to the ICU with a diagnosis of severe asthma. The average age was 35.06. The average stay in the ICU was 14.39 ± 25.55 days, with 16.89 ± 6.67 APACHE II and SAPS II 33.72 ± 13.45 points. Eight patients needed non invasive mechanical ventilation (44.4 %) with a failure rate of 50 % (4 patients). The use of invasive mechanical ventilation was required in 10 patients, 3 of them needed tracheostomy for prolonged mechanical ventilation and sevoflurane Anaconda® device was used in 2 of them. Data respirator parameters were collected during the first 48 hours, highlighting VT (ml / kg) 6.46 ± 1.23, 2.25 ± 2.26 initial PEEP, Ppeak 37.0 ± 13.32 and 12.57 ± 4.68 AutoPEEP cm H2O. The average pH at admission was 7.21 ± 0.14 with a range of [6.99 to 7.40], PCO2 of 64.5 ± 24.22 mm Hg and a lactate concentration of 2.91 ± 2.11 mMol/L. All patients were administered inhaled beta-agonists (two patients salbutamol endovenous) and corticosteroids. 5 patients required magnesium sulfate, 11 patients required sedation; 6 of these patients required relaxation with atracurium or cisatracurium. One patient presented barotrauma while using noninvasive mechanical ventilation prior to endotracheal intubation. Conclusions The clinical profile of patients admitted to the ICU with severe/status asthmaticus is a young male, with requirements of invasive mechanical ventilation after the failure in using medical gases such as helium and noninvasive mechanical ventilation. The strategy of ventilation is protective with permissive hypercapnia. Table 72 (abstract A220). Characteristics (n = 18) Average age

35,06

Woman

6 (33,3%)

Average stay (days)

14,39±25,55

APACHE II

16,89±6,67

INVASIVE MECHANICAL VENTILATION

10 patients

Tidal Volume (ml/kg)

6,46±1,23

AutoPEEP (cm H2O)

12,57±4,68)

pH at admission

7,21±0,14. Range [6,99-7,40]

pCO2 (mm Hg)

64,5±24,22

Ion lactate at admission

2,91±2,11

A221 2015/16 influenza pneumonitis: mater misericordiae university hospital intensive care unit J. Scott1, D. Phelan2, D. Morely2, J. O'Flynn2, P. Stapleton3, M. Lynch3, B. Marsh1, E. Carton1, C. O'Loughlin1 1 Mater Misericordiae University Hospital, ICU, Dublin, Ireland; 2Mater Misericordiae University Hospital, Dublin, Ireland, 3Mater Misericordiae University Hospital, Microbiology, Dublin, Ireland Correspondence: J. Scott – Mater Misericordiae University Hospital, ICU, Dublin, Ireland Intensive Care Medicine Experimental 2016, 4(Suppl 1):A221 Introduction Viral influenza, especially influenza A (H1N1) pandemic influenza, is associated with a significant increase in morbidity and Intensive Care Unit (ICU) admissions. Objectives The aim of this study is to characterise the burden of illness and secondary infection among critically unwell patients admitted to our ICU this year.

Page 118 of 607

Method A retrospective, observational, cohort study of critically ill adult patients with influenza admitted to a Level 3 ICU Dublin, between Decemeber 2015 and March 2016. IntelliVue Clinical Information Portfolio (ICIP) was used to obtain data. Results Critical illness occurred in twenty patients with confirmed influenza; 12 A(H1N1) (60 %), 1 A(H3N2) (5 %), 4 A(non-subtyped) (20 %) and 3 B (15 %). The median age was 50 years (42.5-66.5); 14 patients (70 %) were under 65; 11 (55 %) were male. Thirteen (65 %) had comorbidities, including respiratory disease 5 (25 %), morbid obesity 3 (15 %) and malignancy or immunosuppression 3 (15 %). One was pregnant (5 %) and 8 (40 %) had a smoking history. Nine (45 %) were retrieved from outside the Ireland East Hospital Group. The mean APACHE (Acute Physiology and Chronic Health Evaluation) II Score and the mean SOFA (Sequential Organ Failure Assessment) score, on day one, were 17.85 ± 5.78 and 9.85 ± 3.33 respectively. Twenty (100 %) were mechanically ventilated, for a median of 17.5 days (9.5-32). Eighteen (90 %) satisfied the criteria for Acute Respiratory Distress Syndrome, with a mean PaO2/FIO2 ratio on day one of 13.35 ± 6.46. Fifteen (83.3 %) required rescue therapies for severe hypoxaemia, including extracorporeal life support (ECLS) in five (25 %), prone ventilation and inhaled nitric oxide. Sixteen (80 %) received vasopressors, and 12 (60 %) required renal replacement therapy. The median ICU length of stay was 18.5 days (11-42) and as of 31st March 2016, three (15 %) had died. Four (20 %) had documented early secondary infection, at less than 48 hours; streptococcus pneumoniae was the sole isolate (100 %). Seven (35 %) were treated for presumed late (greater than 48 hours) secondary infection; aspergillus fumigatus (28.6 %) and pansensitive staphylococcus aureus (42.9 %) were the most prevalent. All patients received a neuraminidase inhibitor: oseltamivir was prescribed in 19 (95 %), for a median of 7 days (7-9.75), with 3 (20 %) patients receiving the higher, 150 mg twice daily, dosing regime. Conclusion Seasonal influenza is a major public health concern. It is associated with severe morbidity, resulting in significant economic consequences, as well as a substantial burden on tertiary ICUs. In keeping with national trends, the predominant circulating virus was influenza A(H1N1) and secondary coinfection was common. Although traditional teaching emphasises S.aureus as a common coinfection in viral illness, our results highlight the importance of considering a broad spectrum of bacterial, viral and fungal microorganisms when prescribing empirically in the critically ill patient. A222 Predictors of successful endotracheal extubation K.-C. Cheng, M.-I. Sung Chi Mei Medical Center, Internal Medicine, Tainan, Taiwan, Province of China Correspondence: K.-C. Cheng – Chi Mei Medical Center, Internal Medicine, Tainan, Taiwan, Province of China Intensive Care Medicine Experimental 2016, 4(Suppl 1):A222 Introduction About 10-16 % patients are reintubated after planned endotracheal intubation for 48-72 hours. Reintubation will increase the risk of pneumonia, ventilator and ICU days, and resulting in 25-50 % mortality. Many studies had explored the risk factors of failed extubation, but the integrated indexes for successfully planned endotracheal extubation are inadequate. Objectives To establish useful predictors for successfully-planned extubation which can be followed by medical personnel. Method The patients admitted to the adult ICUs of a tertiary hospital in southern Taiwan, who met the criteria of intubated over 48 hours and prepared for extubation were collected retrospectively between January 2005 to December 2015. Patient's characteristics, disease severity, rapid shallow breath index (RSBI), maximal inspiratory pressure

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

(MIP), maximal expiratory pressure (MEP), cuff leak test (CLT) before extubation, ICU days and outcomes were recorded retrospectively. CLT was classified as 2+ with audible flow without stethoscope, 1+ with audible flow by stethoscope and negative (N) with no audible flow even by stethoscope. Failure of extubation was defined as reintubated within 48 hours. Results Totally 7460 patients were enrolled with 436 (5.8 %) patients failure for extubation. The extubation failure rate of female was higher than male (7.1 % vs 5.1 %, P < 0.001). Failure group patients were noted to have older age (66.6 ± 14.5 vs 64.4 ± 16.4, P = 0.002), higher APACHE II score (17.0 ± 7.5 vs 16.2 ± 7.9, P = 0.044), lower coma scale (10.1 ± 3.8 vs 10.6 ± 3.8, P = 0.019), higher RSBI (69.2 ± 38.0 vs 58.6 ± 30.2, P < 0.001), lower MIP and MEP (-35.5 ± 15.1 vs -38.0 ± 14.7, P = 0.0001; 49.5 ± 28.2 vs 58.9 ± 30.5,P < 0.001 respectively), and higher mortality rate (24.6 % vs 10.7 %, P < 0.001) as compared with successful group. After multivariate logistic regression, cuff leak test 2 + (OR 1.85, P = 0.002), MEP≧55 cmH2O(OR 1.48, P = 0.001), RSBI < 68 (OR 1.47, P < 0.001) and MIP < -40cmH2O(OR 1.26, P = 0.048) were found as predictors for successful extubation. Conclusions Cuff leak test, RSBI < 58, MEP≧60 cmH2O and MIP < -40cmH2O are identified as effective predictors for successfully planned endotracheal extubation. References 1. Penuelas O, Frutos-Vivar F, Fernandez C, et al. Characteristics and outcomes of ventilated patients according to time to liberation from mechanical ventilation. American journal of respiratory and critical care medicine 2011;184(4): 430-7. 2. Thille AW, Harrois A, Schortgen F, Brun-Buisson C, Brochard L. Outcomes of extubation failure in medical intensive care unit patients. Critical care medicine 2011; 39(12): 2612-8. 3. Thille AW, Boissier F, Ben Ghezala H, Razazi K, Mekontso-Dessap A, BrunBuisson C. Risk factors for and prediction by caregivers of extubation failure in ICU patients: a prospective study. Critical care medicine 2015; 43(3): 613-20 Grant acknowledgment Financial support by Chi Mei Medical Center

A223 Measuring intra-abdominal pressure during spontaneous breathing trial, does it help? M.O. Elghonemi, M.H. Saleh Kasr Alainy Medical School, Cairo University, Critical Care Medicine Department, Cairo, Egypt Correspondence: M.O. Elghonemi – Kasr Alainy Medical School, Cairo University, Critical Care Medicine Department, Cairo, Egypt Intensive Care Medicine Experimental 2016, 4(Suppl 1):A223 Background Respiratory system impairment may be caused by an increase of the intra-abdominal pressure. Aim of work: To assess the role of measuring intra-abdominal pressure in predicting successful weaning from mechanical ventilation. Method 124 patients with ARF fulfilling the criteria for weaning were included. Each underwent a 1-hour SBT. All clinical, respiratory parameters and mechanics were recorded. IAP was measured using Kron`s technique at the beginning and every 15 minutes till the end of SBT. The mean of IAP during SBT was calculated. Results Of 124 patients included in the study, 94 patients achieved successful SBT and extubation, while 31 patients needed re-intubation within 48 hours. Mean IAP was lower in patients achieved successful SBT compared to patients who didn`t, 7.25 ± 2.28 vs 9.96 ± 2.6, p value < 0.001. Moreover, patients who needed re-intubation within 48 hours had higher mean IAP compared to patients who didn`t,

Page 119 of 607

9.96 ± 1.4 vs 5.92 ± 1.17, P value < 0.001. Using Spearman´s rank correlation coefficient, it was found that mean IAP was positively correlated with auto PEEP that measured at the beginning and at the end of SBT, and admission APACHE II score, with correlation coefficient measuring 0.515, 0.595, and 0.4 respectively. Conclusion High IAP predicts failure of SBT and need for re-intubation within 48 hours. References 1 -Torquato Jamili Anbar, Lucato Jeanette Janaina Jaber, Antunes Telma, Barbas Carmen Valente. Interaction between intra-abdominal pressure and positive-end expiratory pressure. Clinics. 2009 Feb; 64(2): 105-112. 2- Malbrain ML. Different Techniques to measure intra-abdominal pressure (IAP): time for a critical re-appraisal. Int Care Med. 2004;30:357-71.

A224 The association between the use of life support and 90-day mortality in an international cohort of adult ICU patients T.S. Meyhoff, M. Krag, P.B. Hjortrup, A. Perner, M.H. Møller Copenhagen University Hospital, Department of Intensive Care 4131, Copenhagen, Denmark Correspondence: T.S. Meyhoff – Copenhagen University Hospital, Department of Intensive Care 4131, Copenhagen, Denmark Intensive Care Medicine Experimental 2016, 4(Suppl 1):A224 Introduction The use of life support in intensive care units (ICUs) is common and associated with a high risk of poor outcome. However, the prognostic importance of the duration of life support is less studied. Objectives We aimed to assess the use of life support and the association between its duration and 90-day mortality in adult ICU patients. Method We performed a post-hoc analysis of the SUP-ICU 7-day inception cohort study (1) conducted from Dec 2013 till April 2014 in 97 ICUs in 11 countries. From this cohort (n = 1034), we included adult general ICU patients with an ICU stay of ≥3 days. We assessed the use of life support day 1-3 in the ICU and the crude and adjusted association between its duration and 90-day mortality using logistic regression analysis. Results We included 690 patients with a 90-day mortality rate of 23 %. During the first 3 days in ICU, 65 % of the patients received respiratory support, 57 % circulatory support and 13 % renal replacement therapy (RRT). Patients receiving 3 days of RRT had the worst outcome (OR 6.5 [95 % CI 1.3 - 32.8]) as compared to patients receiving 1 day. For respiratoryand circulatory support the odds ratios were 2.2 [0.9 - 5.3] and 1.2 [0.5 2.6], respectively (Fig. 96). Conclusions The outcome of adult ICU patients was associated with both type and duration of life support. RRT seemed to be associated with worst outcome, potentially because kidney failure often occurs concomitantly to respiratory and circulatory failure. References (1) Krag M, Perner A, Wetterslev J, et al. Prevalence and outcome of gastrointestinal bleeding and use of acid suppressants in acutely ill adult intensive care patients. Intensive Care Med 2015; 41:833-45. Grant acknowledgment We received support from The Novo Nordisk Foundation (NNF15OC0017574). The funding source had no influence on the design of the study.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Page 120 of 607

References 1. Capnodynamic assessment of effective lung volume during cardiac output manipulations in a porcine model. Hällsjö Sander C, Lönnqvist PA, Hallbäck M, Sipmann FS, Wallin M, Oldner A, Björne H. J Clin monit Comput 2015 Sept: 1-9 Grant acknowledgment Håkan Björne recives grants from Maquet critical care

Fig. 96 (abstract A224).

A225 Evaluation of a capnodynamic method for assessment of effective lung volume in intubated pigs during hypercapnia T. Öhman1, T. Sigmundsson1, E. Redondo2, M. Hallbäck3, F. SuarezSipmann4, H. Björne1, C. Hällsjö Sander1, KARISMA 1 Karolinska University Hospital, Department of Anaesthesiology, Surgical Services and Intensive Care Medicine, Stockholm, Sweden; 2Hospital de Navarra, Department of Intensive Care Medicine, Pamplona, Spain; 3 Maquet Critical Care AB, Solna, Sweden; 4Uppsala University, Hedenstierna Laboratory, Department of Surgical Sciences, Uppsala, Sweden Correspondence: T. Öhman – Karolinska University Hospital, Department of Anaesthesiology, Surgical Services and Intensive Care Medicine, Stockholm, Sweden Intensive Care Medicine Experimental 2016, 4(Suppl 1):A225 Introduction Effective lung volume (ELV) can be calculated continuously using a capnodynamic equation, and correlates well with functional residual capacity (FRC) in healthy lungs in a porcine model [1]. Hypercapnia is common in the ICU during lung protective mechanical ventilation, and may affect the capnodynamic method. Objectives The aim of this study was to evaluate ELV during induced hypercapnia in pigs, and to confirm its stability during hemodynamic challenges. Method A cyclic sequence altering breaths with expiratory holds with normal breaths induces periodic changes in alveolar concentration of carbon dioxide. By integrating these variations into the capnodynamic equation ELV can be calculated. Hypercapnia was induced by increasing instrumental dead space in eight anaesthetized, relaxed and mechanically ventilated pigs. FRC was measured with a Sulfur-hexafluoride wash out technique. Cardiac output (CO) was measured using an ultrasonic flow probe placed around the pulmonary artery trunk. Hemodynamic measurements and blood gas analysis were obtained during normocapnia and during hypercapnia at baseline, preload reduction (cava balloon inflation) and dobutamine stimulation. Results Carbon dioxide levels raised from (mean (SD)) 5.6 kPa (0.40) to 9.2 kPa (0.47) during hypercapnia. The bias (limits of agreement, LoA) for ELV at normocapnia was 303 (131 to 476) ml, and percentage error (PE) was 31 %. During hypercapnia, bias (LoA) decreased to -75 (-188 to 39) ml, and PE to 20 %. The hemodynamic interventions resulted in significant changes in CO, i.e. a decrease by 41 % (caval occlusion) followed by a 59 % increase (dobutamine inf.). ELV and FRC remained stable throughout these changes (Fig. 97). Conclusions ELV showed good performance during hypercapnia. The Method shows good stability during severe changes in cardiac output. This indicates that it would be interesting to further evaluate if the Method could be suitable for monitoring lung function in the ICU for instance during protective lung ventilation with permissive hypercapnia or in septic patient with hyper dynamic hemodynamics.

Fig. 97 (abstract A225). ELV during induced hypercapnia

A226 Collapse and decollapse in acute respiratory distress syndrome M. Cressoni1, D. Chiumello2,3, C. Chiurazzi4, M. Brioni1, I. Algieri1, M. Guanziroli1, G. Vergani1, T. Tonetti5, I. Tomic6, A. Colombo1, F. Crimella1, E. Carlesso1, A. Colombo2, V. Gasparovic6, L. Gattinoni5 1 Università degli Studi di Milano, Dipartimento di Fisiopatologia Medico Chirurgica e dei Trapianti, Milano, Italy; 2Policlinico Di Milano, Dipartimento di Anestesia, Rianimazione, Urgenza ed Emergenza, Milano, Italy; 3Plug Working Group, Milan, Italy; 4Università degli Studi di Milano, Dipartimento di Fisiopatologia Medico-Chirurgica e dei Trapianti, Milano, Italy; 5Georg-August-University Goettingen, Anesthesiology and Intensive Care Medicine, Goettingen, Germany; 6University of Zagreb, Department of Intensive Care Medicine, Rebro, Croatia Correspondence: C. Chiurazzi – Università degli Studi di Milano, Dipartimento di Fisiopatologia Medico-Chirurgica e dei Trapianti, Milano, Italy Intensive Care Medicine Experimental 2016, 4(Suppl 1):A226 Introduction In ARDS, independent of severity, lung protective strategy implies a high-PEEP ventilation setting to prevent end-expiratory collapse and to prevent cyclic alveolar opening and closing. [1] [2] Objectives Measure the intratidal collapse and decollapse at similar tidal volumes at positive end-expiratory pressure (PEEP) of 5 and 15 cmH2O.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Method ARDS patients [3] underwent expiratory and inspiratory CT scans at 5-15 cmH2O PEEP during inspiratory and expiratory pause keeping constant the tidal volume (6-8 ml/kg IBW). In each of the CT slices, lung profiles were manually delineated, excluding hilar structures. Thereafter, quantitative analysis of CT scan images was performed with a dedicated software package (Soft-E-Film, www.elekton.it). Lung tissue was classified, according to its gas/tissue content, as not inflated when CT number between +100 and -100. [4] Lung collapsedecollapse was estimated as: 1. Collapse-decollapse PEEP 5 cmH2O = not inflated 5 insp (g) not inflated 5 esp (g) Collapse-decollapse PEEP 15 cmH2O = not inflated 15 insp (g) - not inflated 15 esp (g) Results Thirty-three ARDS patients were enrolled (5 mild, 10 moderate and 18 severe). As shown in Fig. 98, within a given class of severity, the grams of tissue undergoing the intratidal collapse did not change significantly between PEEP 5 or 15 cmH2O (63 ± 26 vs 39 ± 32, 92 ± 53 vs 78 ± 142 and 123 ± 94 vs 96 ± 84 in mild, moderate and severe ARDS respectively). We observed a clear tendency to decrease from PEEP 5 to 15 cmH2O, though it was not statistically significant (p = 0.23, 0.76 and 0.27 respectively in mild, moderate and severe ARDS - paired t-test). Conclusions A consistent intratidal collapse and decollapse is still present at 15 cmH2O PEEP. We observed a clear tendency to decrease at 15 cmH2O PEEP, though it was not significant. References [1] Muscedere JG, Am J Respir Crit Care Med 1994. [2] Caironi P, Am J Respir Crit Care Med 2010 [3] ARDS Definition Task Force,JAMA J Am Med Assoc 2012. [4] Gattinoni L, Am Rev Respir Dis 1987

Fig. 98 (abstract A226). Intratidal collapse and decollapse A227 Low level laser therapy in chronic obstructive pulmonary disease R. El-Sherif1, M. Abd Al-Basser1, A. Raafat2, A. El-Sherif1 1 Kasr Al-Aini Hospitals, Cairo University, Critical Care Department, Cairo, Egypt; 2Cairo University, Physiotherapy Department, Cairo, Egypt Correspondence: A. Raafat – Cairo University, Physiotherapy Department, Cairo, Egypt Intensive Care Medicine Experimental 2016, 4(Suppl 1):A227 Introduction Chronic obstructive pulmonary disease (COPD) is a common preventable and treatable disease. Low level laser (LLL) therapy appears to be a promising noninvasive modality in COPD management. Objectives Study the short-term effects of LLL therapy on clinical and cardiac status in stable COPD patients.

Page 121 of 607

Method After exclusion of patients with Impaired left ventricular systolic function, atrial fibrillation, pulmonary hypertension due to causes other than COPD, and those who had any contraindication to exercise test, thirty stable COPD patients were divided into laser and control groups (15 pts each). Medical treatment was optimized in each group with the addition of LLL in the laser group. In addition to history and physical examination, MMRC scale, 6 MWT, echocardiography with measurements of RV dimensions, TAPSE, and lateral tricuspid annulus tissue Doppler velocities were assessed in each patient before and after LLL. The LLL was done using the following parameters: Wave length: 905 nm, Output 5-20 mw & Frequency 500 HZ. Laser probe was placed on intercostal space corresponding to the site of lesion both anteriorly and posteriorly on chest wall and arm with standardized laser acupuncture points of application with a frequency of 5 sessions/week for 2 successive weeks. Results No significant differences in both groups regarding demographic data. The laser group had higher PASP, lower E´, and higher A´ velocities versus control with p 0.009, 0.03, and < 0.0001 respectively. The laser group showed more improvement in MMRC scale and 6 MWT versus control. 100 % of laser patients showed improvement in MMRC scale by at least one grade versus 46 % in the control. In laser group, 6 MWT was 24.4 ± 10.4 before the study versus 52.9 ± 14.7 meters at the end of the study, p 0.001. In control, 6 MWT was 32.4 ± 14.9 versus 40.1 ± 19.2, p 0.003. No significant changes between any of the echocardiographic criteria before and after the study. Conclusions The use of LLL was associated with more clinical improvement. No echocardiographic changes were noticed after LLL. A228 Physiotherapy assessment of extubation suitability L.E. Brock1, L. Osman2, G. Cork3 1 Guys and St Thomas' NHS Foundation Trust, Physiotherapy, London, United Kingdom; 2Guys and St Thomas' NHS Foundation Trust, London, United Kingdom; 3Guy's and St Thomas' NHS Foundation Trust, London, United Kingdom Correspondence: L.E. Brock – Guys and St Thomas' NHS Foundation Trust, Physiotherapy, London, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A228 Introduction Timing of extubation is of clinical importance as extended periods of intubation and premature extubation resulting in reintubation are both associated with negative patient outcomes[1]. Adult intensive care unit (AICU) extubation failure rates of 10-20 % have been reported[2]. Within a large, UK teaching hospital it was identified that extubation assessment was a multidisciplinary decision but different clinicians applied varying criteria and assessment was not standardised. Following a review of extubation failure rates, a quality improvement initiative was instigated including the development of local ventilator weaning guidelines and physiotherapy (PT) led assessment of extubation suitability. Objectives The aims of this evaluation were to describe the PT assessment of extubation suitability in the AICU and to report extubation failure rates. Method All patients under consideration for extubation by the AICU physicians, who underwent a PT assessment of extubation were included. This assessment was documented within the patient's electronic casenotes. Data was collected by retrospective casenote review during a 3 week period in August 2015. The project was registered as a service evaluation and therefore ethics requirements were waived. Demographics, details of the PT assessment ,and outcomes following extubation were collected (Table 72). Extubation failure was defined as reintubation up to one week following extubation.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Page 122 of 607

Results Data was collected from 45 PT assessments. These assessments most frequently included neurological status, Rapid Shallow Breathing index (RSBi), occlusion pressure during initial 100 ms of inspiration (P0.1), Negative Inspiratory Force (NIF), peak cough flow (PCF) and secretion load. Range of values for the assessed parameters are shown below. When providing opinion regarding suitability for extubation, PTs gave more weight to neurological status, PCF and secretion load than other weaning parameters (see image below). They were more likely to recommend against extubation in the presence of low PCF, inappropriate neurology and large secretion load. Conclusions Physiotherapists frequently include neurological status, work of breathing, PCF, secretion load and NIF in their assessment for extubation. They predominantly use neurological status, cough strength and secretion load to inform recommendations regarding extubation. The extubation failure rate of 15 % is in keeping with current literature although the studied sample size was small. References 1. Menon N, Joffe AM, Deem S, Yanez ND, Grabinsky A, Dagal AH, Daniel S, Treggiari MM. Occurrence and complications of tracheal reintubation in critically ill adults. Respir Care. 2012 Oct;57(10):1555-63. 2. Thille AW, Richard JC, Brochard L. The decision to extubate in the intensive care unit. Am J Respir Crit Care Med. 2013 Jun 15;187(12):1294-302.

Table 73 (abstract A228). Demographics Demographic Age (median)

56.5

% male

68

Number of patients

33

Number of assessments

45

Days ET ventilated (median)

3

Extubationn events

33

Failure rate

15%

Table 74 (abstract A228). Parameters assessed Parameter

Frequency of assessment (%, n=45)

Neurology

88

RSBi

90

P01

95

NIF

90

PCF

100

Secretion load

90

Fig. 99 (abstract A228). Ranges of parameters assessed

Fig. 100 (abstract A228). Parameters in relation to PT advice given

CLINICAL STUDIES ON ARDS AND MECHANICAL VENTILATION A229 Improved prognostication of patients with mild ARDS based on P/F and PEEP thresholds 24 hours after presentation F.D. Simonis1, L.R.A. Schouten1, O.L. Cremer2, D.S.Y. Ong2, G. Amoruso3, G. Cinnella3, M.J. Schultz1, L.D.J. Bos1 1 Academic Medical Center, Amsterdam, Netherlands; 2University Medical Center Utrecht, Utrecht, Netherlands, 3University of Foggia, Foggia, Italy Correspondence: F.D. Simonis – Academic Medical Center, Amsterdam, Netherlands Intensive Care Medicine Experimental 2016, 4(Suppl 1):A229 Introduction Reclassification after 24 hours using PaO2/FiO2 and PEEP thresholds improves the prognostication of patients with moderate/severe ARDS [1,2] but it is uncertain if this also holds true for patients with mild ARDS. Objectives The aim of this investigation was to determine if reclassification after 24 hours using PaO2/FiO2 and PEEP thresholds improves prognostication of mortality in a cohort of prospectively identified patients with mild ARDS in two intensive care units in the Netherlands. Method Patients with mild ARDS, according to the Berlin definition [3], were categorized into 4 groups based on measurements obtained at presentation of ARDS or 24 hours after: PaO2/FiO2 ≥ 250 mm Hg and PEEP = 5 cm H2O (group I); PaO2/FiO2 ≥ 250 mm Hg and PEEP > 5 cm H2O (group II); PaO2/FiO2 < 250 mm Hg and PEEP = 5 cm H2O (group III); PaO2/FiO2 < 250 mm Hg and PEEP > 5 cm H2O (group IV). Patients no longer receiving mechanical ventilation after 24 hours were classified as 'extubated' (group 0). No patients died within 24 hours. The primary outcome was all-cause in hospital mortality. Secondary outcomes were , ICU- and 90-day mortality and the number of ventilator-free days and alive at day 28. Results Of 7,784 patients, 693 patients had ARDS of which 164 patients with mild ARDS and on invasive ventilation were included in the analysis. Table 75 shows outcomes per group at the moment mild ARDS was diagnosed, and after 24 hours. Reclassification after 24 hours showed an improved prognostication with regard to hospital mortality, ICU- and 90-day mortality and the number of ventilator-free days and alive at day 28. Conclusions Reclassification after 24 hours using two simple cutoffs improves prognostication in mild ARDS patients. References 1. Villar J, Fernández RL, Ambrós A, et al. Crit Care Med 2015; 43: 346 2. Bos LD, Cremer OL, Ong DSY, et al. Intensive Care Med 2015; 41:2004 3. The ARDS Definition Task Force. JAMA 2012; 307:2526 Grant acknowledgment This research was performed within the framework of CTMM, the Center for Translational Molecular Medicine (www.ctmm.nl) project MARS (grant 04I-201).

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Page 123 of 607

Table 75 (abstract A229). Distribution of outcome in patients with mild ARDS at onset

Group 0 Group I Group II (n=0) (n=28) (n=29)

Group III (n=38)

Group IV (n=69)

P-value

ICU mortality

7%

17%

18%

26%

0.03

All-cause hospital mortality

39%

38%

29%

36%

0.73

90-day mortality

39%

38%

42%

40.5%

0.83

VFD-28

22 [1525]

23 [7-25]

22 [8-25]

17 [0-24]

0.09

Group III (n=42)

Group IV (n=60)

P-value

after 24 hours

-

Group 0 Group I Group II (n=20) (n=27) (n=15) ICU 0% mortality

7%

20%

12%

37%

60mmHg

MAP40 mL/min/1.73 m2 or not). Correlation analysis was also performed between serum NGAL and others. Results Median age was 64 y.o., 61 % of patients were male. Median value of serum NGAL was 83 ng/mL. The value of serum NGAL was significant difference between the groups of eGFR > 40 ng/mL and eGFR < 40 ng/ mL (80 ng/mL v.s. 163 ng/mL, P = 0.045). Serum NGAL and CRP were significantly correlated in the group of eGFR > 40 ng/mL (R = 0.650, P < 0.001). Conclusions Serum NGAL might reflects systemic inflammation with patient in emergency department.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

References 1. Wang B, Chen G, Zhang J, et al. Increased neutrophil gelatinase-associated lipocalin is associated with mortality and multiple organ dysfunction syndrome in severe sepsis and septic shock. Shock 2015;44:234-8. Grant acknowledgment None.

A296 Noble gas xenon lessens myocardial injury after out-of-hospital cardiac arrest O. Arola1, R. Laitio1, A. Saraste2, J. Airaksinen2, M. Pietilä2, M. Hynninen3, J. Wennervirta3, M. Bäcklund3, E. Ylikoski3, P. Silvasti3, E. Nukarinen3, J. Grönlund1, V.-P. Harjola4, J. Niiranen5, K. Korpi5, M. Varpula5, R.O. Roine6, T. Laitio1, for the Xe-HYPOTHECA study group 1 Turku University, Turku University Hospital, Division of Perioperative Services, Intensive Care Medicine and Pain Management, Turku, Finland; 2 Turku University, Turku University Hospital, Heart Center, Turku, Finland; 3 Helsinki University, Helsinki University Hospital, Division of Intensive Care Medicine, Department of Anaesthesiology, Intensive Care and Pain Medicine, Helsinki, Finland; 4Helsinki University, Helsinki University Hospital, Department of Emergency Medicine and Services, Helsinki, Finland; 5Helsinki University, Helsinki University Hospital, Department of Cardiology, Helsinki, Finland; 6Turku University, Turku University Hospital, Division of Clinical Neurosciences, Turku, Finland Correspondence: O. Arola – Turku University, Turku University Hospital, Division of Perioperative Services, Intensive Care Medicine and Pain Management, Turku, Finland Intensive Care Medicine Experimental 2016, 4(Suppl 1):A296 Introduction In comatose cardiac arrest survivors, the extent of myocardial damage and cardiovascular instability have an important role in the course of developing post cardiac arrest syndrome and predicting long-term outcome. Recent clinical studies have revealed that inhaled xenon provides beneficial cardiovascular effect and mitigate ischemic brain injury in out-of-hospital cardiac arrest (OHCA) patients [1,2]. Objectives The purpose of this study was to assess the effect of xenon inhalation on myocardial ischemic damage and left ventricular function after OHCA. Method A total of 110 comatose patients who had experienced OHCA were randomized to receive either inhaled xenon combined with hypothermia (33 °C) for 24 hours (n = 55 in the xenon group) or hypothermia treatment alone (n = 55 in the control group). Whenever indicated, coronary angiography and percutaneous interventions were performed before intensive care unit admission or later during hospital stay. Xenon was administered with at least 40 % end-tidal concentration and completed at start of rewarming. Troponin-T (TnT) was measured at hospital admission, and at 24 h, 48 h and 72 h post cardiac arrest. Left ventricular function was assessed with echocardiography by cardiologist at intensive care arrival and at 24 hours after completing rewarming. Results Among the 110 patients comprehensive TnT measurements were available from 51 xenon patients (median age 63) and 53 control patients (median age 60). Complete echocardiographic data was available on 18 xenon and 20 control patients. The number of STelevation myocardial infarction and primary coronary intervention, time for return of spontaneous circulation, cardiovascular medication among other baseline characteristics did not differ significantly between the groups. A median (interquartile range, IQR) post-arrival incremental change in TnT at 72 hours was 0.05 μg/l (-0.03 μg/l -0.61 μg/l) in the xenon group and 0.28 μg/l (0.04 μg/l -1.48 μg/l) in the control group (P = 0.014 for the difference between the groups). A mean (95 % confidence interval) absolute post-arrival incremental change in left ventricular ejection fraction at 24 hours after rewarming was 10.4 % (5.6 %-15.3 %) in the

Page 156 of 607

xenon group and 4.3 % (1.4 %-7.2 %) in the control group (P = 0.028 for the difference between the groups). Conclusions Inhaled xenon in combination with mild therapeutic hypothermia may protect against ischemic myocardial injury as demonstrated by significantly attenuated short-term TnT release and better left ventricular function when compared with hypothermia treatment alone. These observations suggest that inhaled xenon may provide cardioprotective effect in OHCA patients. References 1. Arola O et al. Crit Care Med 2013;41:2116-24; 2. Laitio R et al. JAMA 2016;315:1120-28 Grant acknowledgment This study was funded by Academy of Finland and the Clinical Research Funding (EVO) of Hospital District of South-West Finland.

A297 NT-proBNP levels and echocardiography findings in the etiologic diagnosis of acute dyspnea S. Salah, B.G. Hassen, A. Mohamed Fehmi Regional Hospital of Zaghouan, Zaghouan, Tunisia Correspondence: S. Salah – Regional Hospital of Zaghouan, Zaghouan, Tunisia Intensive Care Medicine Experimental 2016, 4(Suppl 1):A297 Introduction Both NT- proBNP and Doppler echocardiography have been approved in the diagnosis of heart failure. In our Study, we compared the contribution of the NT-proBNP levels with the Doppler echocardiography findings in the diagnosis of decompensated congestive leftheart failure (CHF) in patients with acute dyspnea. Patients and method It was a prospective, observation al study at the teaching department of emergency and intensive care in the regional hospital of Zaghouan, including patients with severe dyspnea over six months. All patients underwent physical examination, 12-lead ECG, RX Thorax, NT-ProBNP essay and echocardiography by an attending cardiologist on admission. The accuracy of the two Method for etiologic diagnosis was compared on the basis of the final diagnoses established by the medical staff. Results 65 patients were enrolled, including 45 (69 %) with CHF. Diagnosis of CHF was due to coronary artery disease, hypertension, valve disease, arrhythmia and dilated cardiomyopathy. Non-CHF was due to decompensated chronic obstructive pulmonary disease, pneumonia, and severe asthma. Fifteen patients (23 %) were misdiagnosed at admission. The mean NT-proBNP concentration was 8989 [769 to 18945] pg/ml in the CHF subgroup and 462 [22 to 1589] pg/ml in the other patients (p < 0.01). Systolic LV dysfunction (LVEF < 0.45) was found in 31 patients with CHF (60 %) and in 7 patients with other causes of dyspnea (15 %) (p < 0.01). The E/A ratio and the deceleration time of E-wave (DT) were respectively 1,85 ± 0,77 and 12o ± 23 in the group CHF; 0,81 ± 0,44 and 208 ± 47 in the group NonCHF (p < 0.01). Impaired relaxation and Restrictive mitral pattern were observed respectively in 28 % and 30 % of the patients with CHF and in only two patients and tree patients in the other group. Conclusion Both NT-proBNP assay and echocardiography can be used for the diagnosis of CHF in acutely dyspneic patients. However, the echocardiography is more accurate in patients with intermediate BNP levels. Note: This abstract has been previously published and is available at [1]. It is included here as a complete record of the abstracts from the conference. References 1. Snouda S, Ben Ghezala H, Abbes MF, Daoudi R, Kaddour M, Benchiekh I (2016). NT-proBNP levels and echocardiography findings in the etiologic diagnosis of acute dyspnea. Annals of Intensive Care 6(Suppl 1): P221.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

A298 Role of in-house trauma surgeon in the initial resuscitation of severe major trauma patients in the South Korea S. Kim Catholic University of Korea, Seoul, Republic of Korea Intensive Care Medicine Experimental 2016, 4(Suppl 1):A298 Introduction The predictive mortality of major trauma patients in the South Korea is 35-40 %. High mortality late of major trauma in the South Korea is due to underdevelopment of prehospital management and transport system, hospital care system, and insufficient supporting system of government policy. Establishment and operation of level I trauma center and management of major trauma patients in the South Korea is at an early stage. In house trauma surgeon´s role in the initial resuscitation of patient with major trauma is very important. Objectives This study will evaluate the effect of in house trauma surgeon (IHTS) on time for decision making about major procedure (TD), time for hospitalization (TH) and time to operation (TO) in patients with major trauma. Method This is a retrospective cohort study using trauma database in Uijeonbu St. Mary´s hospital. According to the hospital system, IHTS takes on trauma patients at emergency room for three days on a week. On the other four days, trauma patients were managed by on-call surgeon. Between January, 2013 and December, 2013, 372 of major trauma patients were consecutively enrolled in the study. TD, TH, TO, trauma team activation (TTA). Data were analyzed with presence of IHTS. Results In patients who were admitted to the department of trauma surgery of general surgery, TD and TH with IHTS took significantly less than those without IHTS (TD: 137 minutes vs. 283 minutes, p = 0.002; TH:302 minutes vs. 635 minutes, p < 0.001). Also TO with IHTS took shorter than that without IHTS (200 minutes vs. 256 minutes), although it shows no statistical significance (p = 0.202). In patients who were admitted to the department of orthopedic surgery, neurosurgery, thoracic surgery or plastic surgery, TD and TH with IHTS took 173,331 minutes. They were lower than those without IHTS (289,642 minutes) and showed statistical significance (p = 0.001, 0.001), Likewise TO with IHTS took shorter than that without IHTS (238 minutes vs. 283 minutes) and showed no statistical difference (p = 0.497). In patients with TTA, TD and TH took statistically less time than those without TTA (TD: 168 minutes vs. 320 minutes, p < 0.001; TH: 329 minutes vs. 773 minutes, P < 0.001). However, TO showed no statistical difference patient with an without TTA (2016 minutes vs. 271 minutes, p = 0.200) Conclusions IHTS proved it important role by deducing TD and TH in initial trauma resuscitation. References 1. Luchette F, Kelly B, Davis K, Johanningman J, et al. Impact of the inhouse trauma surgeon on initial patient care, outcome, and cost. J Trauma. 1997 Mar;42(3):490-5. 2. Helling TS, Nelson PW, Shook JW, Lainhart K, Kintigh D. The presence of in-house attending trauma surgeons does not improve management or outcome of critically injured patients.J Trauma. 2003 Jul;55(1):20-5.

A299 Early tapering down of resuscitation fluids shortens the periods of mechanical ventilation in major burn patients Y.-C. Hsu National Taiwan University Hospital, Anesthesiology, Taipei, Taiwan, Province of China Intensive Care Medicine Experimental 2016, 4(Suppl 1):A299

Page 157 of 607

Introduction Fluid resuscitation is a crucial component of initial resuscitation in major burn patients to avoid early mortality. Large amount of intravenous fluids are given during this period of time no matter Parkland formula or other formulae are used as the fluid resuscitation guidelines. Furthermore, it is a trend to administrate even larger amount of intravenous fluids in these years, which is known as “fluid creep” phenomenon [1]. However, excessive intravenous fluids would produce generalized and lung edema, which may preclude wound healing and prolong intubation. Besides, no guideline tells us how to adjust intravenous fluids after the first 48 hours. Objectives We investigate the impact of total intravenous fluids given in the first week to prognosis. Method This is a retrospective cohort study. Patients who injured in Taiwan Formosa Fun Coast explosion event and admitted to the intensive care units (ICUs) in our hospital in 7 days were included. Group A were patients who admitted to the ICU where a more restrictive fluid management Method was conducted; Group B were patients who admitted to other ICUs where more liberal fluid management approach was applied. We compare the outcome such as sepsis insults, organ dysfunctions, intubation days, ICU days, and hospital days between these 2 groups of patient of different fluid management strategies in major burn patients. Results 18 patients were included. The mean age and total body surface area of burn (TBSA burn) were 21.0 ± 3.0 years old and 56.2 ± 15.0 %, respectively [Table 93]. The total intravenous fluids in the first week in Group A patients (57,363.1 ml) were much less than in Group B patients (68,437.9 ml), and the daily fluids of Group A patients were significantly less at day 4 and day 5 [Table 94] comparing to Group B patients. Group A patients also had less body weight gain than Group B patients [Fig. 114]. As the results, Group A patients had significantly less days of mechanical ventilation than Group B patients, while length of ICU stay, length of hospitalization, sepsis insults, and organ dysfunctions were all similar in 2 groups [Table 95]. Conclusions We concluded that a more restrictive fluid resuscitation strategy might shorten the period of mechanical ventilation in major burn patients. The aggressive resuscitation fluids should be tapered down quickly after 72 hours. References 1. Saffle JI: The phenomenon of "fluid creep" in acute burn resuscitation. J Burn Care Res 2007, 28(3):382-395 Grant acknowledgment The emergent burn team in Taiwan Fun Coast explosion event, 3A1, 3A2, and 4FI ICUs in National Taiwan University Hospital

Table 93 (abstract A299). Patient demographic data Patients

Group A

Group B

10

8

p-value

Gender (male/female)

2/8

5/3

Age (years old)

21.2±2.6

20.7±3.7

0.145 0.756

TBSA (%)

55.4±9.4

57.1±20.6

0.832

Inhalation injury

6

5

1.000

Intubation

7

5

1.000

Body Weight (kg)

52.9±7.3

60.1±10.7

0.109

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Page 158 of 607

Table 94 (abstract A299). Daily intravenous fluids in the 1st week Day

Group A

Group B

p-value

1

12501±4509

15069±5647

0.369

2

13183±3446

13437±2904

0.890

3

10074±2338

10720±4340

0.702

4

5870±1518

8702±2288

0.010*

5

5465±1017

9013±2765

0.023*

6

5049±834

6260±2235

0.214

7

5521±863

5237±1328

0.975

Fig. 114 (abstract A299). The trend of body weight

Table 95 (abstract A299). Patient Outcomes Group A

Group B

p-value

Days with mechanical ventilation (all patients)

8.7±9.0

18.4±16.2

0.127

Days with mechanical ventilation (patients with mechanical ventilation)

12.4±8.2

29.4±7.5

0.004*

Length of ICU stay

42.7±12.5

42.9±15.9

0.979

Length of hospitalization

82.7±36.5

82.4±28.8

0.984

ARDS

3/10

4/8

0.630

Sepsis

7/10

6/8

1.000

Acute kidney injury

1/10

1/8

1.000

Mortality

0/10

0/8

A300 Time to computer tomography for trauma patients: uncentric prospective study J. Barea-Mendoza1, C. García-Fuentes1, M. Castillo-Jaramillo1, H. Dominguez-Aguado1, R. Viejo-Moreno1, L. Terceros-Almanza1, S. Bermejo Aznárez1, C. Mudarra-Reche1, W. Xu2, M. Chico-Fernández1, J.C. MontejoGonzález1 1 Hospital 12 de Octubre, Trauma ICU, Critical Care Department, Madrid, Spain; 2Maryland, Bethesda, United States Correspondence: J. Barea-Mendoza – Hospital 12 de Octubre, Trauma ICU, Critical Care Department, Madrid, Spain Intensive Care Medicine Experimental 2016, 4(Suppl 1):A300

Introduction Time is critical in trauma patient care. In other critical diseases (sepsis or IM), decreasing time to definitive treatment have improved patient outcome. Computed Tomography (CT) help us to detect life-threatening lesions. The time until definitive treatment is highly related with CT acquisition times. Due to variations in current literature, there is no consistent arrival to CT times established. Objectives To describe the time related with Computer Tomography for trauma patients as well as determining associated factors. Method The design was a prospective and unicentric with a previous protocol. The inclusion criteria were trauma patients accepted in ICU from February to December 2015. Exclusion criteria were previous admission in other hospital. We collected information about demographic characteristics, severity trauma scores and treatment in the Trauma Room (TR). A form was used to collect the time-points (in minutes). Five time-period were defined: T1 from admission (TR) until the start of transferring to CT; T2 from the start of transferring until first image; T3 between the first and the last CT image; T4 from admission (TR) until first oral report by radiologist, T5 from admission (TR) until definitive destination ( ICU, operating room or interventional radiology). Continuous data are presented as medians (interquartile range; IQR) and categorical data in percents. Analyses were performed with Wilcoxon or Fisher´s test depending on data. We perform a multivariate logistic analysis to asses the factors implicated in the times. Using STATA 12, we calculated 2 tailed P-values setting significance at 0.05. Results 113 patients were admitted during study period with a median age of 39 years; 82.2 % were men. The injury severity score (ISS) was > 16 in 53.9 % of patients and urgent surgery was required in 38 % of them. The most frequent mechanism of injury was precipitation (23.3 %). For the described time periods in minutes (T1-T5) median were T1: 30 (25-41), T2: 8 (6-11), T3: 15 (9-20), T4: 50 (30-65), T5: 65 (52-87). Total body strategy was achieved in 78.9 %. In this subgroup, times were longer at T3 (15 vs 9 ) and at T5 (65 vs 50 ) ; p < 0.05. In the multivariate analysis the factors included were: sex, age, GCS and the related with the resuscitation. At T1 the factors associated with an increased time to definitive treatment were shock and thoracic drainage in the TR (p < 0.05). For T5 the factors were shock , thoracic drainage and whole body strategy (p < 0.05). Conclusions The CT acquisition times in our center were similar to other centers. Prolonged times were associated with shock patients who needed more aggressive therapies. Awareness about time-wasting activities in trauma patients could help us to detect correctable delays. References 1. Huber-Wagner S et al. Effect of the localisation of the CT scanner during trauma resuscitation on survival. Injury . 2014 Oct.

Fig. 115 (abstract A300). Times vs CT strategy

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

A301 Apnoeic preoxygenation for emergency rapid sequence induction of anaesthesia of critically ill patients K. Crewdson, M. Thomas, M. Merghani, L. Fenner, P. Morgan, D. Lockey North Bristol NHS Trust, Anaesthetics and Intensive Care Medicine, Bristol, United Kingdom Correspondence: K. Crewdson – North Bristol NHS Trust, Anaesthetics and Intensive Care Medicine, Bristol, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A301 Introduction Tracheal intubation is associated with significant complications including death and neurological injury. Patients who require rapid sequence induction (RSI) for emergency tracheal intubation are often critically ill or injured with little physiological reserve. Desaturation is a common problem and studies have demonstrated severe hypoxaemia (arterial oxygen saturation, SaO2 < 80 %), in up to 26 % of patients during emergency RSI performed on Intensive Care Units (ICU). Episodes of hypoxaemia during RSI are associated with worse morbidity and increased mortality. Preoxygenation is intended to reduce desaturation and hypoxaemia in the drug-induced apnoeic phase of RSI. Application of 100 % oxygen via nasal prongs promotes movement of oxygen-enriched air into the alveoli during apnoea, which can help sustain SaO2. The majority of studies, though of variable methodology suggest a benefit with apnoeic oxygenation with improvement in SaO2 during the apnoeic phase or difficult laryngoscopy. One RCT comparing apnoeic oxygenation to standard facemask oxygen prior to laryngoscopy failed to show any benefit of apnoeic oxygenation. Objectives This is a clinical effectiveness study of apnoeic oxygenation as a preoxygenation strategy to reduce episodes of hypoxaemia for patients undergoing emergency RSI. The primary endpoint was the number of patients who desaturate to less than 90 % in the peri-RSI period. Methods Patients undergoing emergency RSI in the Emergency Department (ED) and ICU at Southmead Hospital, Bristol were included in the study. Southmead Hospital is the regional major trauma, neuroscience and vascular centre and has an additional large unselected medical and surgical take. Data collected included patient demographics, observations before, during and after RSI, and indication for RSI. Data from current standard practice (preoxygenation via a face-mask) was collected over a 3-month period to establish current rates of hypoxaemia. The same data were collected for a subsequent 3 months using apneic oxygenation. Results In total, 71 patients underwent emergency RSI in the study period. Common reasons for intubation were clinical course (39 %), failure to oxygenate (34 %) and failure to ventilate (14 %). Twenty-one complications were experienced during RSI in 20 patients; hypoxia was the most common complication, occurring in 23 % of patients. In the study period, 24 patients received apnoeic oxygenation. Nine of 47 patients who did not receive apnoeic oxygenation experienced hypoxia (19 %), compared with 7 of 24 patients (24 %) did, p = 0.377. Summary: The incidence of hypoxia associated with emergency RSI in unselected patients is high. The use of apnoeic oxygenation to reduce hypoxia during the drug-induced apnoeic phase of RSI remains a much debated intervention. Having established feasibility, we now intend to investigate the utility of apnoeic oxygenation with larger patient numbers.

A302 Performance of new generation transport ventilators in simulated critical care conditions E.J. van Lieshout1, B. Oomen1, J.M. Binnekade1, D.A. Dongelmans1, R.J. de Haan2, N.P. Juffermans1, M.B. Vroom1 1 Academic Medical Center, University of Amsterdam, Intensive Care, Amsterdam, Netherlands; 2Academic Medical Center, University of Amsterdam, Clinical Research Unit, Amsterdam, Netherlands Correspondence: E.J. van Lieshout – Academic Medical Center, University of Amsterdam, Intensive Care, Amsterdam, Netherlands Intensive Care Medicine Experimental 2016, 4(Suppl 1):A302

Page 159 of 607

Introduction Given the increase in intra- and inter-hospital transport of critically ill patients it is of vital importance that transport ventilators perform at the same level as ICU ventilators, despite their compact design and challenging conditions during transport 1. Objectives To determine accuracy of tidal volume (VT), plateau pressure (Pplat) and positive expiratory pressure (PEEP) delivery by gasdriven and turbine equipped transports ventilators under different simulated pulmonary conditions, ventilator settings and oxygen supply modes. Methods Six transport ventilators (gas-driven Hamilton Raphael, Oxylog 3000, Medumat Transport and turbine-equipped Hamilton C1, C2, Elisée 350) and two ICU ventilators (Servo-I , Hamilton G5) were tested under pulmonary conditions simulating healthy lungs, Acute Respiratory Distress Syndrome (ARDS) and Chronic Obstructive Pulmonary Disease (COPD). Accuracy of VT, Pplat and PEEP were measured by a calibrated pneumotachograph. A percentage difference between actual and displayed values of more than ±10 % was defined as inaccurate. Results Inaccuracy in VT delivery was demonstrated in gas-driven transport ventilators Medumat Transport 66010 (in 8 of 10 experiments), Oxylog 3000 (in 7 of 10 experiments), Hamilton Raphael 350 (in 4 of 10 experiments) as well in the turbine-equipped ventilator Elisée 350 (in 6 of 10 experiments). Inaccuracy in PEEP was present mainly in Medumat Transport 66010 (in 5 of 10 experiments) and turbine-equipped Hamilton C2 (in 4 of 10 experiments). No Pplat inaccuracies were detected in any ventilator. Pulmonary conditions as ARDS or COPD or ventilation settings with high PEEP and respiratory rate did not consistently influence inaccuracy in VT. The influence of delivery of oxygen from a cylinder on VT inaccuracy was present only in gas-driven ventilators Oxylog 3000 and Medumat Transport 66010 under ARDS-conditions. Conclusions Transport ventilators differ in accuracy of delivering tidal volume demonstrating better performance in turbine equipped models. Oxygen supply by cylinder was of limited influence in two gas-driven ventilators only. Two of three turbine equipped transport ventilators (Hamilton C1 & C2) tested showed accuracy comparable to ICU ventilators and therefore are suitable for critical care transport. The use of many gas-driven ventilators in critical care transport should be questioned considering their inaccurate performance. References 1. Boussen S, Gainnier M, Michelet P. Evaluation of ventilators used during transport of critically ill patients: a bench study. Respiratory care. 2013;58(11):1911-1922.

A303 Clinical warning capability prior to in-hospital cardiac arrest R. Algarte1, L. Martínez1, B. Sánchez1, I. Romero2, F. Martínez1, S. Quintana1, J. Trenado1 1 Hospital Universitari Mutua Terrassa, Critical Care Department, Terrassa, Spain; 2Hospital Universitari Mutua Terrassa, Terrassa, Spain Correspondence: L. Martínez – Hospital Universitari Mutua Terrassa, Critical Care Department, Terrassa, Spain Intensive Care Medicine Experimental 2016, 4(Suppl 1):A303 Introduction Early detection of clinical response is determinants of clinical outcome in people with acute illness. IHCA (intrahospital cardiac arrest) often is preceded by a clinical deterioration that could be identified and treated by trained personnel. The NEWS (National Early Warning Score) from NHS (National Health System)1, is based on a simple scoring system in which a score is allocated to six simple vital signs parameters (respiratory rate, oxygen saturations, temperature, systolic blood pressure, pulse rate, level of consciousness. Figure 116 Objectives To describe the patient clinical state prior to IHCA and the relationship with prognosis, in a University Hospital. Methods A retrospective, single-center and descriptive study was conducted during 2014 and 2015. We analyzed all patients admitted on hospital ward that were assisted by the IHCA team. Patients admitted less than 24 hours on Ward and patients not eligible for

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

resuscitation were excluded. Demographic data (age and gender) were collected. We described the type of patient (medical or surgical), features of IHCA attention. The NEWS has 3 levels of scoring, low (1-4), medium (5 and 6) and high (>6) that are related to clinical risk. It was performed by evaluating the data from the vital signs registered on the Ward in the 24 hours prior IHCA. Statistics.Qualitative variables are expressed as percentages and compared using the X2-test; quantitative ones are expressed as means and standard deviations (± S.D), and analyzed using Student´s t-test. The level of significance was placed at p < 0.05. The statistical analysis was performed using specific software (IBM SPSS Statistics for Windows, Version 19.0. Armonk, NY: IBM Corp). Results 85 patients were included. In Table 96 we described the characteristics of the study population. The beginning of cardiopulmonary resuscitation (CPR) maneuvers were immediate on ward, according to IHCA protocol. The arrival of IHCA team was less than 5 minutes in all cases. Figure 117 shows the distribution of NEWS in the IHCA analyzed. In 43.5 % of cases there are not enough information to make a NEWS. In the remaining patients it was able to perform the NEWS although we could only obtain all data in 6.5 % of patients, in the remaining 50 % of IHCA in which was calculated the NEWS some data (vital sings) was missing. Respiratory rate and oxygen saturations were the most frequent missed data. It could be probably underestimating the NEWS performed. Figure 118 shows the relationship between a greater hospital mortality and a higher NEWS. Conclusions Abnormal vital signs are common within 24 hours before IHCA events on hospital wards. A suitable recording vital signs could be useful to alert patients at risk and anticipate in the detection of IHCA.

Page 160 of 607

Table 96 (abstract A303). Characteristics of the study population Characteristics of the study population

All patients (n=85)

Age. Years mean (SD)

75.1 (SD 11.7).

Male %

65.9

Pre-cardiac arrest situation at IHCA team´s arrival %

37.8

ROSC %

41.5

Hospital mortality %

63.5

IHCA witnessed %

60

IHCA in holiday schedule %

61.5

Recording frequency of vital signs at least every 8 hours %

85

Medical patient %

71.8

References 1- Royal College of Physicians. National Early Warning Score (NEWS): Standardising the assessment of acuteillness severity in the NHS. Report of a working party. London: RCP, 2012.

Fig. 117 (abstract A303). Distribution of NEWS in the total IHCA a

Fig. 118 (abstract A303). Analysis between hospital mortality A304 Audit of OOHVF post resuscitation care in Portsmouth post TTM O. Sheikh, D. Pogson, R. Clinton, F. Riccio Queen Alexandra Hospital, Critical Care, Portsmouth, United Kingdom Correspondence: O. Sheikh – Queen Alexandra Hospital, Critical Care, Portsmouth, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A304

Fig. 116 (abstract A303). National Early Warning Score (NEWS)

Introduction We conducted a retrospective review of all our ICU admissions post cardiac arrest to identify OOHVF survivors and compare our own care in the post TTM era with ILCOR standards of care and unit guidelines. Objectives To analyse the electronic record of all OOHVF patients admitted to our unit between October 2014 and July 2015 and identify

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

our performance in achieving key components of current standard post-resuscitation care having adopted 36C as our target for 30H post ROSC. Unit LOS and outcomes were also collected. Methods We identified all admissions to our unit post cardiac arrest using the ICNARC database. Our electronic record was then analysed to identify all those patients admitted post OOHVF. Electronic records were used to identify age, sex, unit LOS(d), estimated ROSC duration, cooling method, time to achieve TTM target 36C, max temp recorded 30H post ROSC, presence of rebound hyperthermia >38C within 48H post TTM, incidence of seizures, use of anticonvulsants and NMB drugs, SSEP N20 and EEG, GCS on unit discharge and unit outcome. Results 66 OOHVF survivors were admitted to our ICU in the audit period. Median age was 62.7 (17-85) M 54, F 12. Median unit LOS was 5.7 days. 56 % of admissions were discharged alive from ICU. 25.7 % patients had seizures on EEG and SSEP was used to prognosticate in 19.7 %. 38 % patients required atracurium to prevent shivering during the TTM period. TTM target 36C was achieved within 4 h post ROSC in 87 %. 78 % patients received cooling via the Icycath and the remainder via surface methods. 37 % of patients had rebound hyperthermia over 38C within 48 h post TTM. One patient was discharged with GCS9. All other survivors had a GCS of 14 or 15 at unit discharge. Conclusions Our unit adopted controlled temperature management to 36C for 30 h post ROSC after publication of the TTM trial. Our admission numbers and demographics have remained unchanged and our adherence to our guideline is good, though only 78 % received IV cooling catheters as stipulated. Our unit survival of 56 % is identical to audits previously performed and is in line with published outcomes. Despite using cooling catheters to control temperature after the TTM period, 37 % of patients suffered rebound hyperthermia in that 48H window. This does not seem to have adversely affected mortality. Use of EEG and SSEP to assist prognostication common on our unit in prolonged coma. The incidence of seizures is common in the post TTM phase. References Nolan J et al. ILCOR consensus statement. Resuscitation (2008) 79; 350-379 Nielsen et al.TTM at 33c vs. 35c after cardiac arrest. NEJM (2013) 369; 21972206 GRANT ACKNOWLEDGMENT The help of Mr M Lympany, IT support in our Dept.

A305 A review of outcome after out of hospital cardiac arrest admitted to a tertiary centre L. Gemmell1, A. MacKay1, A. Arthur2, L. Young1, A. Sinclair2 1 Queen Elizabeth University hospital, Anaesthetics and Intensive Care, Glasgow, United Kingdom; 2Golden Jubilee National Hospital, Anaesthetics and Intensive Care, Glasgow, United Kingdom Correspondence: L. Gemmell – Queen Elizabeth University hospital, Anaesthetics and Intensive Care, Glasgow, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A305 Introduction Sudden cardiac death represents a major health problem. In adults, the prevalence of out of hospital cardiac arrests (OHCA) attended by the emergency medical services is 75 per 100,000. Mortality remains high, and exceeds 90 %. It is well documented that patients with a shockable rhythm have a consistently higher survival than those whose initial cardiac rhythm is non-shockable (1). Objectives The aim of this study was to look at outcome from cardiac arrests admitted to a tertiary referral service and assess whether the presentation of a shockable rhythm was associated with a better prognosis. Methods A retrospective case note review over a one year period of all patients admitted to the Golden Jubilee National hospital with an OHCA with initial rhythm being ventricular tachycardia (VT) or ventricular fibrillation (VF) were included. We looked at cause of cardiac arrest, ICU survival and dependance upon survival, age and length of stay.

Page 161 of 607

Results Sixty three patients were identified, and a full data set obtained for 59. Median age of admission was 59 years, with age range from 26 years to 83 years. All presenting rhythms were either VF or VT. 58 % of this cohort of patients survived post cardiac arrest. Median length of stay was 5 days (IQ range 1-14days). Of the 34patients that survived, 32 of these patients went on to live an independant life. Of the 25patients that died, all of these patients died during their admission to Intensive Care and did not die post ICU discharge. Interestingly, patients whose cause of cardiac arrest was purely arrhythmogenic all survived to hospital discharge, although these numbers are small (n = 11). Conclusions Mortality from cardiac arrest is high and places a huge burden on ICU services. The median length of stay for OHCA patients is 5 days, and with patient numbers in this study is attributable to one ICU bed per day. Their length of stay is thought to quantify the difficulties in prognostication of survival in this cohort of patients, particularly when it comes to the secondary brain injury. Although our sample size is small, it is demonstrated that survival after OHCA where a shockable rhythm is the presenting rhythm, is higher than quoted in literature. Interestingly, patients that survive OHCA are likely to lead an independant life post discharge. If the cardiac arrest cause is arrhythmogenic, although the numbers are small, the survival for this cohort is 100 %. References 1. Temple et al. Predicting neurological outcome and survival after cardiac arrest. Anaesthesia 2012.

A306 Increase of lipopolysaccharide binding protein is associated with reduction of circulating endotoxin with the onset of fever in trauma patients hospitalized in intensive care unit. Preliminary results D. Markopoulou, K. Venetsanou, L. Filippou, E. Salla, S. Stratouli, I. Alamanos KAT Hospital, B ICU and Research Center, Kifisia, Greece Correspondence: D. Markopoulou – KAT Hospital, B ICU and Research Center, Kifisia, Greece Intensive Care Medicine Experimental 2016, 4(Suppl 1):A306 The occurrence of septic shock and sepsis in critically ill patients, hospitalized in ICU, continues to be among the most serious complications, despite the new methods of diagnosis and treatment. The aim of the study is to investigate the effect of fever onset on endotoxin markers. Materials & methods Eighteen polytrauma patients admitted in ICU and 18 healthy volunteers enrolled in the study. 10 ml of blood collected from i) each patient within 24 h of admission (A) and the onset of fever >38° (F) and ii) from healthy individuals (H). Clinical and demographic data recorded on admission. Serum/plasma samples were isolated with centrifugation and stored at -70 ° C. Lipopolysacharide binding protein (LBP), measured with ELISA and circulating endotoxin (LAL) with chromatometric assay. Results On admission, the levels of LBP and endotoxin had no significant differences between patients (A) and healthy (H), P > 0.05. The onset of fever (F) was accompanied by abundant significant LBP release (PFA >0.001 and PFH >0.001) parallel reduction of circulating endotoxin (PFA >0.001 and PFH >0.001) Conclusions Increased LBP release at the onset of fever could account for marker of circulating endotoxin and the following inflammatory implications, in trauma patients hospitalized in ICU.

Table 97 (abstract A306). Data are presented as Median± IQR, LBP and endotox A (n=18)

F (n=18)

H (n=18)

p A~H

p F~H

ENDOTOXIN (EU/ml)

0.32(0.257, 0,422)

0,14±0,03

0,3±0,01

NS

18 yrs) spontaneously breathing patients. Two researchers selected studies using inclusion criteria and then assessed their quality using the STARD and QUADAS guidelines. The key words for literature search were: inferior vena cava, ultrasonography, volume status, central venous pressure. Results We collected 593 studies: 148 excluded with reasons, 433 because duplicates. 12 studies were included for the final analysis with 7 reports which included spontaneously breathing patients and 5 a mixed population (spontaneous and ventilated): 3 on reliability, 10 on correlation with CVP, 5 on validity. The IVC ultrasound measures (IVC ratio, IVC MAX diameter, IVC MIN diameter , Caval index [IVC-CI]) showed an inter-rater agreement range from moderate to very good. The IVC MAX diameter had a significant high correlation with CVP; there were divergent conclusions on IVC-CI and poorly correlation for IVC ratio. The IVC-MAX and IVC-CI showed a good validity in predicting low CVP; the IVC Max and IVCCI showed good validity in prediction high CVP in one study. Eight studies respected more than 60 % of the STARD items and five more than 80 % of QUADAS items. Conclusions Because few reports have been published on the reliability and validity of IVC ultrasound measures the conclusions of this review should be confirmed. Anyway the quality of reporting and methodology of the studies collected were good. The IVC max and min diameter seem to correlate with CVP. All IVC measures show a good accuracy in predicting low or high CVP.

A313 Transesophageal echocardiography in sedated patients requiring noninvasive ventilation by face mask M. Möller, J. Müller- Engelmann, G. Montag, P. Adams, C. Lange, J. Neuzner, R. Gradaus Klinikum Kassel, Cardiology and Intensive Care, Kassel, Germany Correspondence: M. Möller – Klinikum Kassel, Cardiology and Intensive Care, Kassel, Germany Intensive Care Medicine Experimental 2016, 4(Suppl 1):A313 Introduction Transesophageal Echocardiography (TEE) is a standard procedure in Intensive Care. In patients with respiratory insufficiency, TEE with often necessary sedation may lead to further deterioration. We performed TEE in pts. already under NIV via a full face mask. Methods Consecutive pts. already under NIV because of respiratory failure (15 pts, 8 m/7f , age 72 ± 11,2 years, SAPS II score 44 ± 12, PaO2/ FiO2 ≤ 250) with a clinically given indication for immediate TEE were equipped with a full face mask (“endoscopy mask“, VBM Medizintechnik GmbH, Sulz a.N.,Germany). Ventilation was uniformely set (BiPAP f = 20/ min, 30-10 cmH2O, I:E 1:1, FiO2 1,0). For sedation, disoprivan was given as needed. Heart rate, arterial blood pressures and transcutaneous SaO2 were monitored continuously. Blood gases were taken 15 min before, on start and then every five min troughout TEE and 15 and 60 min thereafter. Indications for TEE included endocarditis/ sepsis- focus, n = 8; intracavitary thrombus in tachyarrhythmia, n = 5; or quantification of vituim cordis, n = 2. Results TEE was completed in all pts. Mean duration of the procedure was 11,6 ± 2,9 min, mean ejectionfraction was 45 ± 15 %. Percutaneous

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Page 165 of 607

SaO2 never fell below 92 %. No pt. had to be intubated within 24 h. Mean HR (99,5 ± 13,7 / min), MAP (74,3 ± 3,2 mmHg), pH (7,41 ± 0,01), PaO2/ FiO2 - ratio (212 ± 21) und pCO2 (42,9 ± 1,8 mmHg) did not change significantly (p < 0,5, ANOVA). In 11 pts. TEE lead to definite diagnosis. In 5 pts. immediate therapy (i.e. electrical cardioversion) was be performed Conclusions Performed by experienced physicians, TEE at the bedside during NIV in sedated pts. was feasible and save.

A314 Identification of the aorta by electrical impedance tomography K.H. Wodack1, F. Thürk2, A.D. Waldmann3, M.F. Grässler1, S. Nishimoto1, S.H. Böhm3, E. Kaniusas2, D.A. Reuter1, C.J. Trepte1 1 University Medical Center Hamburg-Eppendorf, Department of Anesthesiology, Hamburg, Germany; 2Vienna University of Technology, Institute of Electrodynamics, Microwave and Circuit Engineering, Vienna, Austria; 3Swisstom AG, Landquart, Switzerland Correspondence: K.H. Wodack – University Medical Center HamburgEppendorf, Department of Anesthesiology, Hamburg, Germany Intensive Care Medicine Experimental 2016, 4(Suppl 1):A314 Introduction Electrical impedance tomography (EIT) is a noninvasive and radiation free bedside monitoring technology, primarily used to detect ventilation disorders. First experimental data in animals suggests that measurement of central hemodynamics within the descending aorta might become possible with EIT.1 To achieve this goal, it is first necessary to determine within the EIT images the exact location of an individual´s aorta. Objectives The aim of this study was to improve and validate an algorithm to automatically detect the aorta by EIT using a hypertonic saline bolus.2 Methods Ten domestic pigs were anesthetized and mechanically ventilated. A bolus of hypertonic saline (10 mL, 20 %), with a higher conductivity than blood was administered into the ascending aorta while EIT data were recorded. The resulting EIT images were analyzed pixel by pixel to identify the aortic pixel (pA), in which the bolus caused the highest transient impedance peak in time (Fig. 120). After completion of the EIT measurements a thoracic computed tomography scan (CT) was performed for each pig. The CT-images were segmented individually for the relevant anatomical structures. EIT images were reconstructed using the GREIT model, based on the individual´s thoracic contours derived from the segmented CTimages.3The resulting spatial resolution of EIT images was 3 mm / pixel. Results The location of the aorta could be detected by EIT in all animals, showing a mean offset of 15 ± 7.5 mm when compared to the center of the true anatomical location identified by CT (Fig. 121). Conclusions It is possible to detect the aorta by EIT using an intraaortic bolus of hypertonic saline. There is a misalignment between the location of the aorta identified by EIT and CT. The significance of this offset for an accurate measurement of fluid responsiveness needs to be determined in further studies. References 1. Maisch S, Bohm SH, Solà J et al: Heart-lung interactions measured by electrical impedance tomography.Crit Care Med 2011; 39: 2173-6 2. Thürk F, Waldmann A et al: Hypertonic saline injection to detect aorta in porcine EIT. Accepted for 17th conference on EIT 2016 3. Adler A, Arnold JH, Bayford R et al.: GREIT: a unified approach to 2D linear EIT reconstruction of lung images.Physiol Meas 2; 30: 35-55 Grant acknowledgement The study was supported by departmental funds of the Department of Anesthesiology, University Medical Center Hamburg-Eppendorf, Germany.

Fig. 120 (abstract A314). Example of peak in conductivity caused by the hypertonic saline bolus within the pixel representing the aorta; the dark blue line shows the filtered conductivity signal, while the light blue line marks the bolus event

Fig. 121 (abstract A314). Examplary EIT image with superimposed CT-derived outer contours of thorax and lungs. The red circle marks the location of the aorta according to CT. The red asterisk indicates the location of aorta as detected by EIT using a bolus of hypertonic saline as contrast agent A315 A capnodynamic method for monitoring effective pulmonary blood flow - evaluation during hypercapnia T. Sigmundsson1,2, T. Öhman1,2, E. Redondo3, M. Hallbäck4, M. Wallin4, F. Suarez Sipman5, A. Oldner1,2, C. Hällsjö Sander1,2, H. Björne1,2 1 Karolinska University Hospital, Department of Anaesthesiology, Surgical Services and Intensive Care Medicine, Stockholm, Sweden; 2Karolinska Institute, Department of Physiology and Pharmacology, Stockholm, Sweden; 3Hospital de Navarra, Department of Intensive Care Medicine, Pamplona, Spain; 4Maquet Critical Care, Stockholm, Sweden; 5 Hedenstierna Laboratory, Department of Surgical Sciences, Uppsala, Sweden Correspondence: T. Sigmundsson – Karolinska University Hospital, Department of Anaesthesiology, Surgical Services and Intensive Care Medicine, Stockholm, Sweden Intensive Care Medicine Experimental 2016, 4(Suppl 1):A315

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Page 166 of 607

Introduction A capnodynamic equation can be used to continuously calculate non shunted pulmonary blood flow (COEPBF) during severe hemodynamic changes ADDIN EN.CITE ADDIN EN.CITE.DATA (1). Hypercapnia is a common clinical state both in the perioperative period and the ICU. Theoretically, elevated carbon dioxide levels could affect the performance of the capnodynamic method. Objectives The aim of the current study was to evaluate the performance of COEPBF during elevated PvCO2 in a porcine model. Methods The required alterations of alveolar concentration of carbon dioxide were created by a ventilatory pattern containing cyclic reoccurring expiratory holds during controlled ventilation. The mathematical model used to calculate COEPBF assumes a steady state in PvCO2 levels. Hypercapnia was induced by three means; decreasing minute ventilation by either lowering respiratory rate or tidal volume, and finally, increasing dead space with preserved minute ventilation. COEPBF was compared to a reference method for CO, an ultrasonic flow probe around the pulmonary trunk. Hemodynamic measurements and blood gas analysis were obtained at baseline before hypercapnia and during the three different types of hypercapnia. Preload reduction (inflated cava balloon) and dobutamine stimulation was performed during low respiratory rate and increased dead space. Results During hypercapnia the PaCO2 and PvCO2 levels were raised on average 58 % and 40 % (+/- 17 and 16 %) from within normal limits, respectively. Bias (LoA) att baseline before induction of hypercapnia was 0.5 L/min (-0.5 to 1.5) and percentage error (PE) 28 %. During hypercapnia, bias (LoA) was 1.4 L/min (0.2 to 2.7), PE 26 % following lower respiratory rate, 0.7 L/min (-0.7 to 2.2), PE 30 % when tidal volumes were decreased, and 0.5 L/min (-0.4 to 1.4), PE 19 % when dead space was increased. During hemodynamic changes the PE was slightly increased (see Table 100 for all values) and the concordance rate was 100 % (see Fig. 122). Conclusions COEPBF performed well during hypercapnia, both during different types of low minute ventilation and increased dead space. The performance was maintained during major changes in cardiac output and trending was excellent. These results indicate that the capnodynamic method should be tested during lung protective ventilation with permissive hypercapnia and even in laparoscopic surgery. References 1. Hallsjo Sander C, Hallback M, Wallin M, Emtell P, Oldner A, Bjorne H. Novel continuous capnodynamic method for cardiac output assessment during mechanical ventilation. British journal of anaesthesia. 2014;112(5):824-31. Grant acknowledgement Håkan Björne received research grant from Maquet Critical Care.

Table 100 (abstract A315). Intervention

COEPBF (L/min)

COTS (L/min)

Shunt (%)

Dead space (%)

PaCO2 (kPa)

PvCO2 (kPa)

Bias (L/ min)

LoA (L/min)

PE (%)

Normocapnia

3.8 (0.5)

3.3 (0.5)

8 (3)

55 (3)

5.56 (0.42)

7.41 (0.69)

0.5

-0.5 to 1.5

28

Hypercapnia low RR

5.5 (0.9)

4.1 (0.8)

13 (4)

47 (9)

8.52 (0.79)

9.83 (1.22)

1.4

0.2 to 2.7

26

>Caval occlusion

3.1 (0.5)

2.3 (0.4)

7 (2)

53 (4)

8.05 (0.84)

10.46 (1.00)

0.8

0.1 to 1.4

24

>Dobutamine

7.0 (0.7)

5.5 (0.7)

19 (5)

45 (10)

9.40 (0.74)

10.65 (1.02)

1.6

-0.1 to 3.2

27

Hypercapnia low TV

5.6 (1.0)

4.9 (0.6)

13 (4)

66 (4)

9.23 (0.32)

10.94 (0.46)

0.7

-0.8 to 2.2

30

Hypercapnia dead space

5.0 (0.7)

4.5 (0.5)

13 (4)

78 (2)

9.06 (0.49)

10.50 (0.91)

0.5

-0.4 to 1.4

19

>Caval occlusion

3.1 (0.5)

2.7 (0.2)

7 (3)

80 (2)

8.34 (0.42)

10.33 (0.95)

0.4

-0.5 to 1.3

31

>Dobutamine

5.0 (0.7)

4.4 (0.5)

14 (4)

78 (3)

9.91 (0.43)

10.46 (1.32)

-0.04

-2.3 to 2.2

33

Fig. 122 (abstract A315).

A316 Ecocardiographyc measurement of cardiac output through modified subcostal window: consistency analysis with conventional methods in critically-ill patients L. Colinas1, G. Hernandez1, R. Vicho2, M. Serna3, R. Cuena4, A. Canabal1, ECOCRITIC group 1 Hospital Virgen de la Salud, SESCAM, Intensive Care Medicine, Toledo, Spain; 2Hospital Quironsalud Palmaplanas, Intensive Care Medicine, Palma de Mallorca, Spain; 3Hospital Marina Salud Denia, Intensive Care Medicine, Murcia, Spain; 4Hospital Virgen de la Salud, SESCAM, Research Unit, Medical Council, Toledo, Spain Correspondence: L. Colinas – Hospital Virgen de la Salud, SESCAM, Intensive Care Medicine, Toledo, Spain Intensive Care Medicine Experimental 2016, 4(Suppl 1):A316 Introduction Echocardiography in the setting of the critically ill patient might be hindered due to poor transthoracic acoustic window. The modified subcostal window, which is obtained from an extrathoracic region, gives us a short-axis parasternal-like view, at the level of the great vessels. Objectives To address the consistency of cardiac output values, measured from the modified subcostal view as compared with those obtained from a transthoracic approach. Methods In 54 consecutive critically-ill patients undergoing transthoracic ecocardiography (Philips Sparq) for both initial diagnosis of shock or subsequent hemodynamic monitoring, velocity time integral of pulsed wave Doppler of the left ventricular outflow tract (VTILVOT) and the pulmonary flow (VTIPu-Ps and VTIPu-Ms) were measured from the apical four-chamber view, the short-axis paraesternal and modified subcostal views respectively. Results were analyzed with predictive performance test and the consistency analysis. Interobserver reproducibility analysis was assessed in 20 patients with repeated measurements by two experienced physicians using the intraclass correlation coefficient (ICC). Results Baseline characteristics included: mean heart rate 79 beats per minute (15.7), age 59.2 years (15.6), male 29 (53.7 %), previus cardiac disease 9 (16.7 %), medical diagnosis at admission in 40 patients (74.1 %, with primary cardiac diagnosis in 8 [14.8 %]); 19 mechanically ventilated patients (35.2 %, mean PEEP 6.8 [2.8]), with vasoactive drugs in 13 patients (24.1 %, mean norepinephrine dose .38 [.29] mcg/Kg/min). Median VTILVOT was 24.3 cm (interquartile range [IQR] 8-36), VTIPu-Ps 24.5 cm (IQR: 6-57) and VTIPu-Ms 18.9 cm (IQR: 7-43). The comparison between VTILVOT and VTIPu-Ms revealed a consistency of .52 (95%CI .29 to .69, p < .001), see Fig. 123; and between VTILVOT and VTIPu-Ps .44 (95%CI .13 to .67, p = .004), see Fig. 124. The ICC observed was .93. Conclusions The consistencies between VTILVOT and VTIPu-Ms and VTILVOT and VTIPu-Ps were moderate but VTILVOT and VTIPu-Ms consistency improved after excluding patients with VTILVOT >29 cm. VTIPu-Ms underestimate VTILVOT a mean 4.45 cm. Further research is required in order to assess it.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Page 167 of 607

echocardiography was performed simultaneously with transpulmonary thermodilution. Cardiac function index and left ventricle ejection fraction were recorded simultaneously. Results Twenty-one patients were included in our study. Median [quartile] age was 66 [57-81] years. Sex-ratio (M/F) was 12/9. Fourteen patients (66.7 %) were admitted with septic shock whereas 7 patients (33.3 %) had cardiogenic shock. Mechanical ventilation was required for 13 patients (61.9 %). Median [quartile] CFI was 4.5 [2.75 - 6] min-1. Median [quartile] LVEF was 60 [45.75 - 66] %. Spearman coefficient was 0.844 (p < 0.001). Moreover, a CFI > 4.5 min-1 predicted a LVEF > 60 % with a sensitivity of 100 % and a specificity of 82 %. Conclusion CFI can be used as a reliable tool to predict LVEF in patients with acute circulatory failure.

Fig. 123 (abstract A316).

References 1. Jabot J, Monnet X, Bouchra L, Chemla D, Richard C, Teboul JL. Cardiac function index provided by transpulmonary thermodilution behaves as an indicator of left ventricular systolic function. Crit Care Med. 2009; 37(11):2913-8. 2. Combes A, Berneau JB, Luyt CE, Trouillet JL. Estimation of left ventricular systolic function by single transpulmonary thermodilution. Intensive Care Med. 2004;30(7):1377-83. Grant acknowledgment None.

A318 Correlation between microcirculatory perfusion and arterial elastance O. Bond1, P. De Santis1, E. Iesu1, F. Franchi2, J.-L. Vincent1, J. Creteur1, S. Scolletta2, F.S. Taccone1 1 Hopital Erasme, Université Libre de Bruxelles, Department of Intensive Care, Brussels, Belgium; 2University of Siena, Anesthesia and Intensive Care Unit, Department of Medical Biotechnologies, Siena, Italy Correspondence: O. Bond – Hopital Erasme, Université Libre de Bruxelles, Department of Intensive Care, Brussels, Belgium Intensive Care Medicine Experimental 2016, 4(Suppl 1):A318

Fig. 124 (abstract A316).

A317 Usefulness of cardiac function index as a surrogate of left ventricle ejection fraction in patients with acute circulatory failure A. Chaari1, K. Abdel Hakim2, M. Etman2, M. El Bahr2, A. El Sakka2, K. Bousselmi2, A. Arali2, V. Kauts2, W.F. Casey2 1 King Hamad University Hospital, Intensive Care, Muharaq, Bahrain; 2King Hamad University Hospital, Muharaq, Bahrain Correspondence: A. Chaari – King Hamad University Hospital, Intensive Care, Muharaq, Bahrain Intensive Care Medicine Experimental 2016, 4(Suppl 1):A317 Introduction The assessment of the left ventricle function is of paramount importance for the management of patients with acute circulatory failure. In this regard, the measurement of the left ventricle ejection fraction (LVEF) by echocardiography is considered to be the gold standard. Only few studies investigated the correlation between the LVEF and the Cardiac Function Index (CFI) calculated through transpulmonary thermodilution [1, 2]. Objective To assess the usefulness of cardiac function index (CFI) as a surrogate of left ventricle ejection fraction (LVEF) in patients with acute circulatory failure. Methods We conducted a prospective study in a 12-bed medical surgical intensive care unit. All patients admitted with acute circulatory failure and monitored with a transpulmonary thermodilution device were included in our study. We excluded the patients with acute core pulmonale and those with poor echogenicity. A bedside trans-thoracic

Objectives Hypotension is a common problem in critically ill patients. Arterial blood pressure (ABP) is influenced by changes in intravascular volume, cardiac output and vascular tone. In particular, systemic vascular resistance (SVR) is a main determinant of arterial elastance (Ea); SVR is also influenced by peripheral endothelial function (e.g. microcirculatory flow). However, few data are available on the correlation of these parameters in this setting. Methods Prospective study conducted in a 35-bed medico-surgical ICU since January 2016. Patients with an invasive ABP monitoring and requiring a cardiac output (CO) monitoring during a fluid challenge (FC) were simultaneously assessed with a pulse wave analysis (PWA) system (MostCare, Vygon, France) to estimate Ea and with an Incident Dark Field (IDF) handheld device (Braedius Medical BV, The Netherlands) to evaluate sublingual microcirculation. Microvascular perfusion was assessed using the proportion of small-perfused vessels (PPV). SVR was calculated according to standard formulas. Relative changes in each variable were calculated before and after FC; fluid responders had a CO increase of at least 10 % from baseline. Results We included 13 patients (age 64 [54-73] years; 7 male) requiring a fluid challenge (n = 6 for hypotension; n = 5 for oliguria; n = 2 for hypovolemia). At baseline, mean arterial pressure was 72 [66-84] mmHg, heart rate (HR) 83 [80-106] bpm and CO 4.2 [3.5-4.4] L/min; after fluid challenge, MAP was 73 [71-74] mmHg, HR 85 [77-103] bpm and CO 4.8 [3.9-5.4] L/min. Seven patients were fluid responders. There was no correlation between PPV and SVR (ρ = 0.12; p = 0.23) or Ea (ρ = 0.03; p = 0.56). Likewise changes in SVR and Ea during fluid challenge were not correlated with changes of PPV (-ρ = 0.01; p = 0.91). Conclusions No correlation was found between either SVR or elastance and indexes of microvascular perfusion in the sublingual region. The impact of microcirculatory perfusion on the arterial load should be further defined.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

A319 Comparison of the transthoracic doppler sonography and echocardiography in cardiac output measurement in severe trauma patients Z. Marutyan, L. Hamidova, A. Shakotko, V. Movsisyan, I. Uysupova, A. Evdokimov, S. Petrikov N.V. Sklifosovsky Research Institute of Emergency Medicine of the Moscow Healthcare Department, Moscow, Russian Federation Correspondence: Z. Marutyan – N.V. Sklifosovsky Research Institute of Emergency Medicine of the Moscow Healthcare Department, Moscow, Russian Federation Intensive Care Medicine Experimental 2016, 4(Suppl 1):A319 Introduction Monitoring of a systemic hemodynamic at victims with the severe combined injury allows to outline the volume and structure of an infusion therapy. At this moment invasive methods are typically used for monitoring of a systemic hemodynamic. However, invasive monitoring is impossible or extremely limited in some patients because of a mode of failure (for example, pelvic fracture) and a consumable cost. In such cases non-invasive monitoring of a systemic hemodynamic can be used as an optional method. Objective to compare cardiac output measurements between transthoracic Doppler sonography and echocardiography (EchoKG). Material and methods Twenty four patients at the age of 41 ± 14 enrolled in the study (women/men - 3/19, traumatic brain injury - 3, a spinal trauma - 2, a concomitant injury- 19). We determined a stroke volume (SV) and a cardiac output (CO) with the use of the transthoracic Doppler sonography ("USCOM", Australia) and the transthoracic EchoKG (GE Vivid Q, USA). Transthoracic doppler sonography was provided by doctors of the intensive care unit, EchoKG was performed by doctors of the functional diagnostics. Results We didn't find differences between the average values SV and CO evaluated by DS and EchoKG: 67,2 ± 21,4 ml vs 64,2 ± 18 ml, and 6,5 ± 2,4 l/min vs 6,4 ± 2 l/min. Conclusion Usage of transthoracic doppler sonography for determining systemic hemodynamic values at victims with a severe injury allows to receive the measured data of the systolic discharge and cardiac output comparable to data of transthoracic EchoKG. TDS provided a comparable results of SV and CO measurements with EchoKG and could be used in trauma care practice.

A320 Cerebral oximetry assessed by near -infrared spectrometry in patients with postresuscitation syndrome C. Gonen1, E. Haftacı2, C. Balci1,2 1 Kocaeli Derince Training Hospital, Intensive Care, Kocaeli, Turkey; 2 Kocaeli Derince Education and Research Hospital, Kocaeli, Turkey Correspondence: C. Balci – Kocaeli Derince Training Hospital, Intensive Care, Kocaeli, Turkey Intensive Care Medicine Experimental 2016, 4(Suppl 1):A320 Introduction Transcranial cerebral oximetry is a non-invasive method to monitor the changes in the cerebral oxygen metabolism. It is a near-infrared spectroscopy method that uses multi-wavelength radiation between 690-1100 nm spectrum. This photons can pass skin, bone, brain and cerebrospinal fluid. Objectives The aim of the study is to analyze the association of monitoring transcranial cerebral oximertry with morbidity and mortality in patients with postresuscitation syndrome. Methods In this study we retrospectively analyzed the data of 23 patients with postresuscitation syndrome. The data included age, sex,

Page 168 of 607

arrest time, location of arrest, light reflex, Glascow coma scale and cerebral oximetry values compared with mean arterial pressures, SpO2 values and survival. Results Glascow coma scale and survival are well correlated with higher values of cerebral oximetry (p < 0.05). Conclusions This study suggests that, high cerebral oxygen saturation values in patients with postresuscitation syndrome are associated with lower morbidity and mortality. References Association between hemoglobin, cerebral oxygenation and neurologic outcome in postcardiac arrest patients I Meex, K Ameloot, C Genbrugge, M Dupont, B Ferdinande, J Dens, C Dedeyne Crit Care. 2015; 19(Suppl 1): P430. Published online 2015 March 16. doi: 10.1186/cc14510.

A321 Goal-directed therapy guiaded by dynamics preload variables (SVV and VVP) after major hepatic resection can help intraoperative fluid optimization and reduce post-surgical complications F.J. Redondo Calvo1, N. Bejarano2, V. Baladron3, R. Villazala3, J. Redondo3, D. Padilla4, P. Villarejo4 1 Facultad de Medicina Ciudad Real, Hospital General Universitario de Ciudad Real, Anestesiologia y Reanimacion, Ciudad Real, Spain; 2Facultad de Medicina Ciudad Real, Hospital General Universitario Ciudad Real, Cuidados Criticos Pediatricos, Ciudad Real, Spain; 3Hospital General Universitario Ciudad Real, Anestesiologia y Reanimacion, Ciudad Real, Spain; 4Facultad de Medicina Ciudad Real, Hospital General Universitario Ciudad Real, Cirugía Hepatobiliar, Ciudad Real, Spain Correspondence: F.J. Redondo Calvo – Facultad de Medicina Ciudad Real, Hospital General Universitario de Ciudad Real, Anestesiologia y Reanimacion, Ciudad Real, Spain Intensive Care Medicine Experimental 2016, 4(Suppl 1):A321 Introduction Classically, central venous pressure (CVP) and pulmonary artery occlusion pressures (PAOPs) have been used as surrogates for volume measurements.However dynamic preload variables like as pulse pressure variations (PPV) and stroke volumen variations (SVV) could be very useful in hepatic postresection phase to optimize the volumen needed for the patients. Objectives The aim of this study is to assess if is better to reduce postoperative complications, optimizing hemodynamic situation after hepatic resection guided by dynamic variables preload (PPV and SVV) versus using liberal fluid management. Methods Experimental clinical trial, controlled, randomized, single blind, in patients undergoing hepatic resection. In both groups perioperative fluid restriction was done (5 ml/kg/hour of Ringer Lactate) until removal of the surgical specimen was performed. Affer that, two randomized groups were established. In the control group fluids (colloids) were administered until hemodiámica stability was achiveved and standard pressures were got (MAP > 65 mmHg CVP 8-14, urine output > 0.5 ml / kg / h). In the other group volume was administered until a SVV < 12 and a PPV < 14 and MAP > 65 mmHg were achieved. In both groups the volume administered to achieve quantified objectives and postoperative complications (nausea and vomiting, respiratory and infectious complications) was registred. Results 9 patients were enrolled in the GDT group and 10 patients in the control group. There were no statistically significant differences in preoperative variables. A statistically significant difference in the volume administered after resection in both groups (1290 +/- 375 vs 128 +/485.55, p < 0.01) was found. In the surgical time we found no correlation between VPP and CVP (r = 0,172, p = 0.656) and between the VSS and CVP (r = 0,243, p =0.492). We found very good ability to predict the

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

response to volume with both the VPP (ROC curve: 0.96) as the VSS (ROC curve: 0.92), defined as the 20 % improvement in cardiac output. We found a decrease in complications but no statistically significant differences in respiratory complications (4 vs 2, p = 0.62), infectious complications (5 vs 2, p = 0.35) and nausea and vomiting (6 vs 1, p = 0.057). Conclusions Guided therapy goals (GTD) is able to decrease the volume of liquid provided after hepatic resection. This fact optimize hemodynamics values in these patients (cardiac output, mean arterial pressure) and therefore, reduce postoperative complications due to excessive intake volumen. References 1. De Wolf Am, Aggarwal S. Monitoring preload during liver transplantation. Liver Traspl 2008; 14: 268-269 Grant acknowledgement We express our gratitudes to Mutua Madrileña Fundation (Madrid, Spain) for its grant collaboration by without which this work could not have been completes.

RRT NEW DEVELOPMENTS FOR AKI A322 Impact of cumulative nephrotoxin exposure on acute kidney injury in the pediatric critically ill patients A. Akcan-Arikan1,2, C.E. Kennedy1 1 Baylor College of Medicine, Pediatric Critical Care, Houston, TX, United States; 2Baylor College of Medicine, Pediatric Nephrology, Houston, TX, United States Correspondence: A. Akcan-Arikan – Baylor College of Medicine, Pediatric Critical Care, Houston, TX, United States Intensive Care Medicine Experimental 2016, 4(Suppl 1):A322 Introduction Critically ill children are exposed to multiple nephrotoxic medications due to the nature of the underlying disease process as well as comorbid conditions. Acute kidney injury (AKI) is prevalent in the pediatric critically ill. Nephrotoxin exposure is a potentially modifiable risk factor for AKI in non-critically ill children, with exposure to three or more discrete nephrotoxic agents significantly increasing odds of AKI. Despite higher inherent risk, data on AKI and nephrotoxin use in the pediatric intensive care unit (PICU) setting is scarce. Fluid Overload Kidney Injury Score (FOKIS) is a decision support tool piloted in our PICU as a daily score incorporating subscores for AKI (pRIFLE creatinine and urine output), fluid overload (total fluid (in- out)/ICU admission weight), and exposure to nephrotoxic medications (a priori determined list of medications). Objectives We aimed to investigate AKI prevalence and frequency of nephrotoxin exposure using FOKIS subscores. Methods Retrospective analysis of daily FOKIS subscores in PICU patients (pts) over 18 months. AKI was defined and staged using pRIFLE creatinine criteria. Nephrotoxin exposure (Exp) was defined as exposure to three or more nephrotoxic agents in 24 hours. Each additional medication exposure was also recorded. Results 2830 pts (median age 5.5 years (IQR 1.3-12.9 years), 55 % male) were included over 18 months. 246 pts (8.7 %) had Exp during PICU stay (120 (49 %) to 3, 73 (30 %) to 4, 30 (12 %) to 5, 14 (5.7 %) to 6, 9 (3.7 %) to 7 different nephrotoxins). Fifty eight percent (142/246) of pts with Exp had AKI (41 (16.7 %) R, 29 (11.8 %) I, 72 (29.3 %) F) compared to 16 % (413/2584) without (201 (7.8 %) R, 90 (3.5 %) I, 122 (4.7 %) F) (p < 0.001). Conclusions A majority of PICU pts with AKI had exposure to nephrotoxins. Patients who had exposure to three or more nephrotoxins had increased AKI compared to patients who did not. This association needs to be further studied prospectively to determine

Page 169 of 607

causality, specifically exploring timing of AKI onset in relation to nephrotoxin exposure in critically ill pediatric patients. Cumulative nephrotoxin exposure is a possible modifiable AKI risk factor in critically ill children. Table 101 (abstract A322). Cumulative nephrotoxin exposure and AKI Exp

No AKI

Risk

Injury

Failure

No exposure

2171, (84%)

201, (7.8%)

90, (3.5%)

122, (4.7%)

3 meds

59, (49.2%)

22, (18.3%)

12, (10%)

27, (22.5%)

4 meds

28, (38.4%)

11, (15.1%)

11, (15.1%)

23, (31.5%)

5 meds

12, (40%)

6, (20%)

2, (6.7%)

10, (33.3%)

6 meds

5, (35.7%)

1, (7.1%)

2, (14.3%)

6, (42.9%)

7 meds

0, (0%)

1, (11.1%)

2, (22.2%)

6, (66.7%)

A323 Effectiveness of renal angina index score predicting acute kidney injury on critically ill patients M.F. Aguilar Arzapalo SSA UADY, Mérida, Mexico Intensive Care Medicine Experimental 2016, 4(Suppl 1):A323 Introduction Until two thirds of critically ill patients develop Acute Kidney Injury (AKI) and it is associated with an increased risk of death. Renal Angina Index is a score that evaluates the risk of presenting AKI. Objectives This study was used to determine the effectiveness of the Renal Angina Index as a developing predictor of AKI in 3 days. Methods The study was based on a prospective cohort of critically ill patients in whom Renal Angina Index score was completed with a 72 hour follow up of serum creatinine levels, water balance and urinary output, establishing after patients that develop AKI. Afterwards a relationship was established between Renal Angina Index score and its predictive capacity of developing AKI, determining sensibility, specificity, positive predictive value and negative predictive value for the score. Results A final sample of 206 patients was obtained at the end of the study. The incidence of AKI in the studied population was 27.2 % (n = 56), and the average scores of Renal Angina Index were 20.52 in those who develop AKI and 4.35 in those who didn't develop AKI. This score offers a 90.7 % sensibility, 95.4 % specificity with an area under de curve of 0.963 (0.934-0.991). A positive predictive value of 0.88 was obtained and a negative predictive value 0.97 Conclusions Renal Angina Index score is effective predicting AKI in critically ill adult patients. Note: This abstract has been previously published and is available at [4]. It is included here as a complete record of the abstracts from the conference. References 1. Chawla L, Goldstein S, Kellum J, Ronco C. Renal Agina: concept and development pretest probability assessment in acute kidney injury. Critical Care. 2015; 19: 93. 2. Basu R, Zappitelli M, Brunner L, Wang Y, Wong H et al. Kidney International. 2013; 85, 659-667. 3. Basu R, Wang Y, Wong H, Chawla L, Wheeler D, Goldstein S, Incorporation of biomarkers with the renal angina index for prediction of severe AKI in critically ill children. Clin J Am Soc Nephrol. 2014; 9: 654-662. 4. Aguilar Arzapalo M, Barradas L, Lopez V, Escalante A, Jimmy G, Cetina M (2016) Effectiveness of renal angina index score predicting acute kidney injury on critically ill patients. Critical Care 20(Suppl 2): 94.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Grant acknowledgement To Hospital O´Horán.

Page 170 of 607

Results 12 % of the patients developed CA-AKI. Predicting factors were elevated APACHE II test score, hemoglobin and baseline serum creatinine, shock or acute myocardium infarct at admission, vasoactive drugs and diuretics at the moment of the contrast administration, and the following comorbidities: chronic heart failure and chronic kidney failure. Significant risk factors in the univariate analysis were selected for the predictive model (Table 102). A bootstrap method was used to select the best subset of risk factors to avoid overfitting the data. The corresponding ROC curve of the model (Fig. 126) has an area under the curve of AUC = 0.75 (range 0.71-0-79). Conclusions A predictive model of CA-AKI has been developed. Predicting variables with prognostic value are the hemoglobin content, the APACHE II test score on admission and the use of vasoactive drugs and of diuretics. The corresponding nomogram allows for easy evaluation of the probability of developing CA-AKI in critical patients. References 1. Hoste EA, Doom S, De Waele J et al. Epidemiology of contrast-associated acute kidney injury in patients: a retrospective cohorte analisis. Intensive Care Med 2011; 37(12):1921-31. 2. McDonald J, MacDonald R, Comin J et al. Frequency of Acute Kidney Injury Following Intravenous Contrast Medium Administration: A Systematic Review and Meta-Analysis. Radiology 2013; 267(1):119-128.

Fig. 125 (abstract A323). ROC Curve ''Angina Renal Index Score''

A324 Predictive model for contrast-associated acute kidney injury in critical patients C. Gomez-Gonzalez1, S. Mas-Font2, A. Puppo-Moreno1, M. HerreraGutierrez3, M. Garcia-Garcia4, S. Aldunate-Calvo5, NEFROCON Investigators 1 Hospital U.V. Rocío, Seville, Spain; 2Hospital General Universitario de Castellon, Castellon, Spain; 3H Carlos Haya, Malaga, Spain; 4Hospital de Sagunto, Valencia, Spain; 5Complejo Hospitalario de Navarra, Pamplona, Spain Correspondence: C. Gomez-Gonzalez – Hospital U.V. Rocío, Seville, Spain Intensive Care Medicine Experimental 2016, 4(Suppl 1):A324 Introduction Incidence of acute kidney injury in critically ill patients increases with chronic diseases, nephrotoxic drugs, and the use of contrast in diagnostic/therapeutic techniques. Contrast-associated acute kidney injury (CA-AKI) is strongly predictive of adverse outcomes. Objectives To build a statistical predictive model to evaluate the probability of developing contrast-associated acute kidney injury in critical patients. Methods This study has been endorsed by the Spanish Society of Intensive Critical and Emergency Care Medicine (SEMICYUC). Data were obtained in a prospective multicenter study in 33 Spanish Intensive Care Units, with a total of 1009 patients. The criteria used to define CA-AKI was the AKIN criteria: a rise of serum creatinine of ≥0.5 mg/dl or a 50 % relative rise in creatinine at 48-72 hours after contrast exposure. The predictive model has been developed employing a binary logistic regression using the software R. The ROC curve was obtained (Fig. 126) and the model was calibrated using this graph. From the model, we have generated a graphical nomogram (Fig. 127) to facilitate its use in a clinical environment. The nomogram includes the 4 variables shown to have prognostic value.

Table 102 (abstract A324). Risk factors selected for the predictive model APACHE II

Significance (p)

Odds Ratio

Confidence Interval 95%

0.05), anyway yielding decreased minute ventilation (MV) (p < 0.01) and corrected minute ventilation (MVcorr = MV*PaCO2/40 mmHg) (p < 0.01), indicating enhanced CO2 removal; finally, esophageal pressure swing (ΔPes) and pressure-time product (PTPes) decreased (p < 0.01 for both). AIC for linear correlation with flow rates was lower for PaO2, RR, ΔEELV and PEF, as if improved aeration and its effects on oxygenation constantly increase with flow; while non-linear AIC

Facial mask (12 L/min)

HFNC (30 L/min)

HFNC (45 L/min)

HFNC (60 L/min)

P-value (ANOVA)

Fitting (AIC) Linear

Fitting (AIC) Non-linear

PaO2 (mmHg)

72.8±13.5

86.7±14.4*

92.4±11.3*

95.9±13.9*

0.05); The respiratory frequency of the patients in the group has been ameliorated in the 48 h due to the HFNC treatment(20.57 ± 2.37vs.23.87 ± 2.64,P < 0.05)and the result can be better in the 72 h.As long as the treatment of the HFNC, Pa02/FiO2 was significantly higher at 72 h(276.00 ± 108.15vs.177.75 ± 64.23, P < 0.05). In the severe ARDS group, the failure rate is 40 percent and there is no statistics significance about Pa02/FiO2 before and after the treatment of the HFNC.The respiratory frequency could be ameliorated in the proceeding of HFNC treatment(48 h:19.25 ± 3.77vs.25.60 ± 2.50, P < 0.05;72 h 20.75 ± 3.59vs.25.60 ± 2.50,P < 0.05),but the PaCO2 has achieved to the normal level(39.75 ± 7.13vs.30.40 ± 6.98,P < 0.05). Conclusions HFNC can ameliorate the symptom of the mild to moderate ARDS, but the effect in the severe ARDS isn't certain. So, the symptom of patients must have been monitored frequently in order to avoid missing the chance of intubation. References 1) Nishimura M. High-Flow Nasal Cannula Oxygen Therapy in Adults: Physiological Benefits, Indication, Clinical Benefits, and Adverse Effects. Respir Care, 2016, 61(4):529-541. 2) Masaji Nishimura. High-flow nasal cannula oxygen therapy in adults [J]. Nishimura Journal of Intensive Care, 2015, 3:15. 3) D. Chiumello , M. Gotti, C. Chiurazzi. High-Flow Nasal Cannula Oxygen Therapy: Physiological Effects and Clinical Data. J.-L. Vincent (ed.), Annual Update in Intensive Care and Emergency Medicine 2016,DOI 10.1007/ 978-3-319-27349-5_21 4) Laurent Papazian, Amanda Corley, Dean Hess, et al.Use of high-flow nasal cannula oxygenation in ICU adults: a narrative review[J]. Intensive Care Med DOI 10.1007/s00134-016-4277-8. 5) Lemiale V, Mokart D, Resche-Rigon M, et al. The effect of noninvasive ventilation vs oxygen therapy on mortality among immunocompromised patients with acute respiratory failure: a randomized clinical trial[J]. JAMA, 2015, 314:1711-1719.

A342 Comparison of an ordinary mask and an open face mask: is there a better way to provide oxygen? T. Kobayashi, Y. Onodera, R. Akimoto, A. Sugiura, H. Suzuki, M. Iwabuchi, M. Nakane, K. Kawamae Yamagata University Faculty of Medicine, Anaesthesiology and Critical Care Medicine, Yamagata, Japan

Page 180 of 607

Correspondence: T. Kobayashi – Yamagata University Faculty of Medicine, Anaesthesiology and Critical Care Medicine, Yamagata, Japan Intensive Care Medicine Experimental 2016, 4(Suppl 1):A342 Introduction Oxygen therapy is one of the most popular treatments in the ICU. We typically use oxygen masks to give oxygen to patients who breathe spontaneously. However, ordinary oxygen masks have a risk of CO2 rebreathing—especially at lower flow rates (less than 5 l/ min). For our bench study, we investigated the use of an open face mask (Atom Medical, Japan) in alleviating the risk of CO2 rebreathing while still delivering a relatively high FIO2. The result showed that the open face mask mitigated this safety concern. It provided a 10 % higher of inspired oxygen (FIO2) and a lower inspired carbon dioxide (FICO2) as 60-100 % reduction than an ordinary mask (IMJ, Japan). That result was presented in the congress of WFSICCM. Objective We hypothesize that the use of an open face mask is better for our spontaneous breathing patients who need oxygen than an ordinary mask because it can give patients better oxygenation and CO2 clearance. Methods This study was performed among 22 ICU patients who suffered from mild respiratory failure. We applied both masks (ordinary and open face mask) with various flow rates: 5, 4, 3, 2, 1 and 0.5 l/ min for three minutes, respectively. First, we tried ordinary masks with 5 l/min for 15 minutes and decreased the flow to 0.5 l/min every three minutes. The patients whose SpO2 achieved 100 % with 5 l/min of oxygen were excluded. Respiratory parameters (SpO2, respiratory rate: RR; FICO2, end-tidal CO2: EtCO2) were recorded at the end of each flow rate. FICO2 and EtCO2 were measured via CapnostreamTM (Medtronic, US). Next, we investigated an open face mask in the same manner: applying 5 l/min of oxygen for 15 minutes. Finally, we checked patient satisfaction with each mask using a visual analogue scale (VAS: 0-best, 5-worst) at the end of each trial. We used the Wilcoxon matched-pairs signedranks test to detect significant differences in the two groups. Results Data from 12 patients were analysed. The median age was 66 (range: 50-87). Most of the patents (92 %) underwent an operation (cardiovascular surgery: 67 %, orthopaedic surgery: 17 %, abdominal surgery: 8.3 %). The patient who did not receive an operation had congestive heart failure. The group using open face masks showed higher SpO2 (Mean: 97.8 ± 0.47 % vs 97.5 ± 0.71 %, p < 0.01) and lower FICO2 (0.76 ± 0.16 % vs 2.00 ± 0.77 %, p < 0.01) and EtCO2 (35.5 ± 0.36 % vs 37.0 ± 0.78 %, p < 0.01). The respiratory rate did not reveal any significant difference (20.19 ± 0.75/min vs 20.85 ± 0.68/min, p = 0.15). Patient satisfaction with the open face mask was superior to an ordinary oxygen mask (VAS 2.42 ± 1.08 vs 3.25 ± 1.14, p < 0.01). Conclusion Open face masks can provide better oxygenation and reduce CO2 rebreathing for the ICU patients in comparison with ordinary masks. Reference Lamb K et al. Can J Respir Ther. 2016; 52: 13-5.

A343 Prophylactic assistance for low cough peak expiratory flow at extubation P. Carmona Sanchez, M.D. Bautista Rodriguez, M. Rodriguez Delgado, V. Martínez de Pinillos Sánchez, A. Mula Gómez, J.M. Serrano Simón Hospital Universitario Reina Sofia, Intensive Care Unit, Córdoba, Spain Correspondence: P. Carmona Sanchez – Hospital Universitario Reina Sofia, Intensive Care Unit, Córdoba, Spain Intensive Care Medicine Experimental 2016, 4(Suppl 1):A343 Introduction The peak expiratory flow (PEF) is a strong predictor of success or failure of extubation. The risk of extubation failure for PEF < 60 L/min is to 6.3 times higher than PEF > 60 L/min (1,2). Objectives To evaluate the impact on extubation outcome of prophylactic noninvasive assistance at extubation of patients with weak cough, and to identify optimal device for assistance.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Methods Prospective collected database was conducted of from December 2014 to April 2016. Weak cough was defined by PEF < 60 L/min. The PEF was measured with Cosmed Pony Graphic® spirometer v.4.0 SCZ before extubation, for the patients mechanically ventilated > 24 h, and without tracheostomy, who passed successfully a sponyaneus breathing trial (SBT) at least of 30 min of pressure support at 5-8 cmH2O, CPAP, or T-T. The patients were then extubated regardless the PEF. The patients with PEF > 60 L/min conventional oxygen therapy was applied, and groups at risk of extubation failure (PEF < 60 L/min) was applied randomly prophylactic CPAP, noninvasive ventilation (VNI BiLevel), or humidified high flow nasal cannula (HFNC). Extubation failure was defined by the need of reintubation within 48 h following extubation. We compared both groups of patients according to the PEF, and prophylactic assistance on groups patients with weak cough, on outcome extubation. Continuous variables were expressed as mean ± SD or median (IRQ) and categorical variables as absolute value and percentage. The comparison of continuous variables was performed by Student t test and MannWhitney test and comparison between categorical variables was performed by Fisher's exact test and Chi-square test. Results 137 patients were studied, 89 males (65 %). The two groups of patients according to the PEF, were similar regarding age, APACHE II, underlying chronic disease and duration of mechanical ventilation before extubation. Prophylatic assistance was effectively applied to 72,1 % of patients with PEF < 60 L/min and 87,2 % for PEF > 60 L/ min. VNI was applied at 51,2 % (CPAP 18,6 % and BiLevel 32,6 %), Optiflow® 48,8 %. No significant differences were found between the prophylactic devices of assistance at risk group of extubation failure. In the patients with PEF > 60 L/min, extubation failure rate was 12,8 %. There were not differences between diagnosis and extubation failure. Not differences in total stay in ICU between both groups. Conclusions Prophylactic non invasive assistance at extubation could reduce the risk of extubation failure in patients with a weak cough strength without increasing length of stay. It's could be applied to any mode and device (HFNC, CPAP or VNI). References 1. Smina M. Cough peak flows and extubation outcomes. Chest. 2003 Jul;124:262-8. 2. Beuret P. Interest of an objective evaluation of cough during weaning from mechanical ventilation. Intensive Care Med 2009; 35:1090-1093.

A344 Measure of cough strength at extubation on the screen of the ventilator P. Beuret, C. Fortes, M. Lauer, M. Reboul, J.-C. Chakarian, X. Fabre, B. Philippon-Jouve, S. Devillez, M. Clerc Centre Hospitalier, Intensive Care Unit, Roanne, France Correspondence: P. Beuret – Centre Hospitalier, Intensive Care Unit, Roanne, France Intensive Care Medicine Experimental 2016, 4(Suppl 1):A344 Introduction Numerous studies have shown that a weak cough strength, if evaluated objectively before extubation by the measure of peak cough expiratory flow (PCEF), is a strong predictor of extubation failure. However, reported cut-off values depend on the device used for the measure. Measuring PCEF on the screen of the ventilator by its flow sensor seems easy to perform, without disconnecting the patient from the ventilator, and could provide a method available everywhere at the bedside. Objectives This prospective study aimed to compare the measure of PCEF on the screen of the ventilator Servo i (Maquet, Solna, Sweden) to the measure obtained with an electronic flowmeter, the Piko-1 (Ferraris Respiratory, Hertford, UK), whose accuracy to predict extubation outcome has been previously reported, with an optimal cut-off value of 35 l/min [1]. Methods The PCEF was measured by the respiratory therapist just before extubation for the patients mechanically ventilated for more than 24 hours and who passed successfully a spontaneous breathing trial of 30 minutes of pressure support at 8 cm H2O. The order of the measures with the two devices was changed after inclusion of half of

Page 181 of 607

the patients. The best value obtained from two measures was kept for the analysis. Results Over one year 87 patients were eligible for the study; the measure of PCEF was impossible to achieve because of lack of understanding in 14 patients (16 %). Among the 73 patients included for the analysis, there was a significant correlation between the measures obtained with the Piko-1 and the Servo i (rs = 0.831; p < 0.01) (Fig. 135). The linear regression line obtained predicts a cut-off value of PCEF on the Servo i at 60 l/min, corresponding to the value at 35 l/min previously determined with the Piko-1 [1]. Conclusions The measure of PCEF just before extubation on the screen of the ventilator Servo i is easy to perform and well correlated with the measure performed by an electronic flowmeter Piko-1. This allows to propose a cut-off value of 60 l/min with the ventilator, below which the cough strength may be judged as weak. References 1. Beuret P. Interest of an objective evaluation of cough during weaning from mechanical ventilation. Intensive Care Med 2009; 35: 1090-1093.

Fig. 136 (abstract A344).

A345 A randomized cross-over physiological study of high flow nasal oxygen cannula versus non-invasive ventilation in adult patients with cystic fibrosis: the HIFEN study N. Rittayamai1,2,3, M. Sklar1,2, M. Dres1,2,4, M. Rauseo1,2, C. Campbell1, B. West5, D.E. Tullis1,5, L. Brochard1,2 1 Keenan Research Centre, Li Ka Shing Knowledge Institute, St. Michael's Hospital, Toronto, Canada; 2Interdepartmental Division of Critical Care Medicine, University of Toronto, Toronto, Canada; 3Division of Respiratory Diseases and Tuberculosis, Department of Medicine, Faculty of Medicine Siriraj Hospital, Bangkok, Thailand; 4Neurophysiologie Respiratoire Expérimentale et Clinique, Sorbonne Universités, Paris, France; 5Division of Respirology, St. Michael's Hospital, Toronto, Canada Correspondence: N. Rittayamai – Keenan Research Centre, Li Ka Shing Knowledge Institute, St. Michael's Hospital, Toronto, Canada Intensive Care Medicine Experimental 2016, 4(Suppl 1):A345 Introduction Non-invasive ventilation (NIV) is the first option for the treatment of cystic fibrosis (CF) patients with acute exacerbation. High flow nasal oxygen cannula (HFNC) is a heated humidified, high flow oxygen delivery system that has demonstrated benefits in terms of survival in patients with acute hypoxemic respiratory failure and in preventing postextubation failure. This device may also have benefits in patients with hypercapneic respiratory failure including CF patients. We hypothesize that HFNC would not be inferior to NIV in terms of reducing work of breathing and improving breathing pattern in CF patients requiring ventilator support.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Page 182 of 607

Objectives To compare HFNC vs. NIV induced changes in inspiratory work of breathing assessed by the thickening fraction of the diaphragm (TFdi), and breathing pattern, hemodynamics, dyspnea and comfort. Methods CF patients with acute exacerbation requiring ventilator support were ventilated with HFNC and NIV for 30 minutes in random order. TFdi was measured using ultrasound at baseline and at 25 minutes with each device. Pulse oximetry (SpO2), transcutaneous CO2 (PtcCO2) were continuously recorded and respiratory rate, tidal volume (VT) and minute ventilation (MV) measured by bio-impedance techniques; hemodynamics, dyspnea and comfort assessed by visual analog scales were also recorded. Results were compared using a Mann Whitney, 2 tailed test, and are expressed as mean (SD) with each intervention compared to baseline conditions. Results 12 patients were enrolled (mean age 31.3 years, mean FEV1/ FVC 49.9 %, mean FEV1 28.4 % predicted). TFdi was similar with the two techniques, but HFNC, compared to NIV, resulted in a significant decrease in respiratory rate (-20.2 % (18.0) vs -0.2 % (18.7), p = 0.024) and a lower mean arterial pressure (0.3 % (5.6) vs 5.8 % (4.9), p = 0.017). No significant differences were found in heart rate, SpO2, PtcCO2, VT, MV, comfort and dyspnea (Table 111). Conclusions HFNC was not inferior to NIV with respect to diaphragmatic work in CF patients who had an indication for ventilator support. These preliminary data suggest that HFNC may confer physiological benefits by decreasing respiratory rate, and constitute an interesting alternative to NIV. References 1. Madden BP, Kariyawasam H, Siddiqi AJ, Machin A, Pryor JA, Hodson ME. Noninvasive ventilation in cystic fibrosis patients with acute or chronic respiratory failure. Eur Respir J 2002;19(2):310-3. 2. Vivier E, Mekontso Dessap A, Dimassi S, Vargas F, Lyazidi A, Thille AW, et al. Diaphragm ultrasonography to estimate the work of breathing during non-invasive ventilation. Intensive Care Med 2012;38(5):796-803. Grant acknowledgement This study was supported by Cystic Fibrosis Canada and by a grant from Siriraj Hospital in Bangkok. LB's laboratory received a grant and equipment from Fisher Paykel for the study. LB holds the Keenan Chair in Critical Care and Acute Respiratory Failure.

Table 111 (abstract A345). Percentage of change from baseline and 25 minutes HFNC; Mean (SD)

NIV; Mean (SD)

p-value

Thickening fraction of the diaphragm

-1.8 (32.7)

-2.0 (32.8)

0.880

Respiratory rate

-20.2 (18.0)

-0.2 (18.7)

0.024

SpO2

1.5 (3.2)

0.3 (2.9)

0.450

PtcCO2

1.2 (4.7)

0.8 (4.0)

0.667

Minute ventilation

-17.7 (33.2)

-11.3 (23.1)

0.627

Mean arterial pressure

0.3 (5.6)

5.8 (4.9)

0.019

Heart rate

-3.8 (5.7)

-3.7 (4.4)

0.793

Dyspnea

11.1 (63.0)

-11.1 (42.8)

0.338

Comfort

-22.9 (13.9)

-15.8 (27.6)

0.310

A346 High-flow nasal cannula effectively washes out CO2 from anatomical dead space in a sophisticated respiratory model made by 3D printer Y. Onodera, R. Akimoto, H. Suzuki, M. Okada, M. Nakane, K. Kawamae Yamagata University, Department of Anesthesiology, Faculty of Medicine, Yamagata, Japan Correspondence: Y. Onodera – Yamagata University, Department of Anesthesiology, Faculty of Medicine, Yamagata, Japan Intensive Care Medicine Experimental 2016, 4(Suppl 1):A346 Introduction The washout effect of HFNC has not been well evaluated. Our prior study, presented at the ESICM congress 2014, evaluated the reduction of PETCO2 with HFNC using a tracheal intubation trainer and a test lung. Although we concluded that the washout effect is most effective with a relatively low flow, the model had a large upper airway dead space of 200 mL, which was thought to have influenced PETCO2. We also were not able to measure PEEP, which is also thought to have interaction between HFNC flows. Therefore, we developed a more sophisticated artificial respiratory model using a 3D printer, and used a lung model equipped with a pressure sensor to quantitatively evaluate the washout effect of HFNC and the interaction with PEEP. Objectives To quantitatively evaluate the washout effect and interaction with PEEP using different levels of HFNC flow. Methods The airway model was made by a 3D printer using the craniocervical CT data of a healthy 32-year-old male. The total anatomical dead space was adjusted to 180 mL (3 mL/kg). The model lung (LUNGOO : Air water safety service Inc., Kobe, Japan) had the following settings: normal (Compliance (C) 50 ml/cmH2O, resistance (R) 5 cmH2O/L/s, tidal volume (Vt) of 500 mL, respiratory rate (RR) 16 /min), obstructive (C 70, R 20, Vt 700, RR 10), restrictive (C 30, R 5, Vt 300, RR 30) with inspiratory time at 1 second, and residual volume of 1000 mL. CO2 was infused into respiratory lung models to reach PETCO2 of 40 mmHg without HFNC. After setting PETCO2 with each lung model, HFNC with flows of 10 to 60 L/min were applied and the change in PETCO2 in the subglottic area and the inlet of the lung model was measured. PEEP inside the model was also recorded. Results With the normal lung open-mouth model, 10 L/min of HFNC flow decreased the PETCO2 of the subglottic area and the inlet in a lung model to 30 mmHg. Increasing HFNC flow did not decrease PETCO2 in either area. With the normal lung closedmouth model, PETCO2 of all sites required a HFNC flow of 40 L/ min to decrease, and reached 30 mmHg with HFNC flow of 60 L/ min. With the obstructive lung open mouth model, PETCO2 of all sites had the same trends as the normal lung open-mouth model. With the restrictive lung open mouth model, 20 L/min of HFNC flow decreased the PETCO2 of the subglottic area and the inlet to 25 mmHg, and did not decrease thereafter. As HFNC flow was increased, PEEP gradually generated up to around 8 to 5 cm H2O with open mouth models and up to 17 cmH2O with the normal lung closed mouth model. Conclusions The washout effect of HFNC is thought to reduce the PETCO2 enough to have a clinical effect. Contrary to the relation of HFNC flow and generated PEEP, the HFNC reduced PETCO2 with a relatively low flow in open mouth models. HFNC required more flow to reduce PETCO2 with a closed mouth model, which is thought to be due to a less efficient washing out of the dead space than in the open model. Grant acknowledgment Nothing to Declare.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Page 183 of 607

of stay in CC or in hospital or duration of NIV use between the two groups. Conclusion Our data suggests that a Rockwood frailty score of 5 or above is associated with reduced CC and hospital survival in patients with ARF requiring NIV. Use of a frailty index may be a useful predictive marker in patients admitted to critical care and further work is warranted to define the prognostic value of frailty scoring. References 1. Le Maguet, P.et al. Prevalence and impact of frailty on mortality in elderly ICU patients: a prospective, multicenter, observational study. Intensive Care Med, 2014.4(5):p.674-82. 2. Roberts, C.M.et al. Acidosis, non-invasive ventilation and mortality in hospitalised COPD exacerbations. Thorax, 2011.66(1):p.43-8. 3. Masip, J.et al. Noninvasive ventilation in acute cardiogenic pulmonary edema: systematic review and meta-analysis. JAMA, 2005.294(24):p.3124-30.

Fig. 137 (abstract A346). Relation of PETCO2, PEEP and HFNC flows

A347 Impact of frailty score on survival rate in patients admitted to critical care with acute respiratory failure treated with bilevel noninvasive ventilation (NIV) in a large teaching hospital critical care department N. Ahmad, M. Wood, A. Glossop Northern General Hospital, Sheffield Teaching Hospital NHS Trust, Anaesthesia and Critical Care, Sheffield, United Kingdom Correspondence: N. Ahmad – Northern General Hospital, Sheffield Teaching Hospital NHS Trust, Anaesthesia and Critical Care, Sheffield, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A347 Introduction Clinical frailty is increasingly used and recognised to predict mortality and functional dependence following critical admission1 although less is known regarding its impact on specific patient groups. Bilevel noninvasive ventilation (NIV) is frequently used to treat acute respiratory failure (ARF) of various aetiologies. Whilst NIV is known to be beneficial in exacerbations of COPD2, acute cardiogenic pulmonary oedema3 and other presentations of ARF, there is scarce data available on the impact frailty may have on the success of NIV in this patient group. Objective We examined the impact of frailty score on outcomes in patients receiving NIV for mixed aetiology of ARF within our critical care (CC) unit. Methods Case notes and electronic patient data records (Metavision,iMDsoft,MA,USA) were retrospectively reviewed for adult patients with ARF who received NIV between December 2011 and April 2013. Patients who received continuous positive airway pressure and those with a primary surgical problem were excluded. Patient frailty score was recorded on admission using the Rockwood frailty index in on 55 consecutive patients. Demographic data, body mass index (BMI), primary cause of ARF and hospital mortality were recorded.Patients were divided into two groups according to their frailty score: below 4 (very fit to vulnerable) and 5 and above (frail to severely frail). Results 55 patients were identified. 44 % were male, mean age 63 years. Presenting diagnosis was pneumonia in 32 patients (58 %) and acute exacerbation of COPD in 13 patients (23 %). The overall CC survival rate was 84 % with 73 % surviving to hospital discharge. 31 patients had a frailty score of 5 or above (77 % male, mean age 66 years) and 24 patients had a frailty score of 4 or less (50 % male, mean age of 62 years). Patients with a lower frailty score had a higher overall survival rate (91 % on CC and 83 % to hospital discharge) compared to those with a higher frailty score (70 % on CC and 51 % to hospital discharge). The mean BMI for both groups was identical (30.5 kg/m2). There were no differences in the mean length

A348 Has the high flow therapy changed our approach to the acute respiratory failure? J. Higuera Lucas, A. Blandino Ortiz, D. Cabestrero Alonso, R. De Pablo Sánchez, L. Rey González Hospital Universitario Ramón y Cajal, Unidad de Cuidados Intensivos, Madrid, Spain Correspondence: A. Blandino Ortiz – Hospital Universitario Ramón y Cajal, Unidad de Cuidados Intensivos, Madrid, Spain Intensive Care Medicine Experimental 2016, 4(Suppl 1):A348 Introduction The use of high flow therapy in the critical care units for adults is becoming more common from few years on. This therapy approaches the patient with acute respiratory failure from a new way of treatment. Objectives Our goal is to analyze the impact and the results of the application of high flow therapy in an intensive care unit of a tertiary hospital. Evaluate if the use of such a therapy has avoided the use of mechanical ventilation in the patients. Materials and methods The study has summarized all the patients that have been treated with the oxygen therapy treatment between May 2013-October 2015. The reviewed data has been: SOFA, APACHE II, SAPS II, tabulated diagnosis, etiology of the respiratory failure, length of stay, use of high flow before or after the intubation, evaluate if the intubation has been avoided, mortality and mortality in 30 days. Results The study includes 113 patients, medium average age 58, men 56.5 %,and women 43.5 %. The severity indexes of the study are: SOFA 8,37 ± 4,44(1-19), APACHE II 19,68 ± 8,75(3-44), SAPS2 49,31 ± 20,59(11-95). The average stay of the patients is 14,70 days. The origin of the disease is respiratory in the 55 % of the causes, hematological 22 %, septic 14,5 %, cardiologic 3,8 % and miscellaneous in all the rest. The cause of the ventilation decompensate has been:respiratory 77,1 %, septic 16,8 %, and miscellaneous in all the other cases. The causes of the respiratory failure have been hypoxemic in the 75,6 % of the cases, hipercapnic in the 1,5 % and mixed causes in the rest. 42,7 % of the times high flow therapy is provided before the mechanical ventilation. 19,8 % of the times after the mechanical ventilation. It is considered that avoids mechanical ventilation in 43,5 % of the cases. Applying non parametrical test the group of patients that needs mechanical ventilation has a higher mortality (p < 0, 0001) Conclusions The use of high flow therapy has changed the attitude towards the patients with insufficiency respiratory diseases avoiding in many cases the use of mechanical ventilation. More controlled studies are needed showing which other patients can be beneficed by the high flow therapy to prevent the delays in the start of mechanical ventilation. References [1] Williams R, Rankin N, Smith T, Galler D, Seakins P. Relationes-hip between the humidity and temperature of inspired gas and the function of the airway mucosa. Crit Care Med.1996; 24:1920-9

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

[2] Williams AB, Ritchie JE, Gerard C. Evaluation of a high-flownasal oxygen delivery system: Gas analysis and pharyngeal presure. Intensive Care Med. 2006;32 (S1):S219 [3] Groves N, Tobin A. High flow nasal oxygen generates positive air-way pressure in adult volunteers. Aus Crit. Care2007;20:126-31 [4] Sztrymf B, Messika J, Mayot T, Lenglet H, Dreyfuss D, Ricard JD. Impact of high flow nasal cannula oxygen therapy on intensive care unit pacients with acute respiratory failure: A prospective observacional study. J Crit. Care.2012;27:324,e9-13

A349 Comparative evaluation of three full face masks for delivering NIV R. Costa, G. Spinazzola, A. Pizza, G. Ferrone, M. Rossi, M. Antonelli, G. Conti Catholic University of Rome, Department of Anesthesia and Intensive Care, Rome, Italy Correspondence: R. Costa – Catholic University of Rome, Department of Anesthesia and Intensive Care, Rome, Italy Intensive Care Medicine Experimental 2016, 4(Suppl 1):A349 Introduction Full face masks have been proposed to improve patient comfort during NIV and thus increase NIV success. Objectives To compare three full Face masks (FF) (Dimax, Dimar, Italy; Performax, Respironics,US and BiTrac, Pulmodyne, US). Methods An adult mannequin, connected to an active lung simulator (Compliance : 60 ml/cmH2O; Respiratory Resistances : 4 cmH2O/L/sec) breathing at 20 and 30 breaths/minute was ventilated with the three FF in Pressure Support (PS 10 and 15 cmH2O, PEEP 8 cmH2O) with a fast pressurization ramp (100 %) and 2 expiratory triggers (25 %, 50 %). The data analysis evaluated Patient-ventilator interaction (Delaytrinsp, Delaytrexp, Timesync) and masks performance (Swingtrigger, PTPt, PTP300, PTP500 Index). Results At RR 20, Timesync was longer with FF-Pulmo and FF-Dimar, while, at RR 30, no significant difference was found between the masks. FF-Dimar and FF-Respir significantly increased Swingtrigger and PTPt at RR 20, but both FF showed a better performance in terms of PTP300. No significant difference was found in terms of PTP500 Index. Conclusions Independently from the setting, all FF guaranteed a Timesync above 50 % of the neural inspiratory time. The different shape and material of the FF significantly influenced our results: softer flange and the need for less head-strips tightness required by FF-Dimar caused an initial pressure dissipation with longer Delaytrinsp, Swingtrigger and PTPt. Conversely, the need to strongly pull the headstrips with FF-Pulmo to reduce leaks, explained its superior performance and interaction. Grant acknowledgment None.

OUTCOME ANALYSIS I A350 Three years of experience in pulmonary embolism approach in an intermediate care unit H. Ribeiro, J. Alves, M. Sousa, P. Reis Centro Hospitalar de Entre o Douro e Vouga, Serviço de Medicina Intensiva Polivalente, Santa Maria da Feira, Portugal Correspondence: H. Ribeiro – Centro Hospitalar de Entre o Douro e Vouga, Serviço de Medicina Intensiva Polivalente, Santa Maria da Feira, Portugal Intensive Care Medicine Experimental 2016, 4(Suppl 1):A350 Introduction Venous Thromboembolism encompasses deep venous thrombosis (DVT) and pulmonary embolism (PE) and has a significant impact on morbidity and mortality, being responsible for more than 370000 deaths every year in 6 European countries. Nowadays, clinical

Page 184 of 607

scores and diagnostic non-invasive techniques have improved diagnosis and early treatment: anticoagulant and thrombolytic therapy. Objectives To characterize patients admitted to Intermediate Care Unit (ICU) due to PE and submitted to thrombolytic therapy. Methods Retrospective observational study of patients admitted due to PE in the ICU between January 2013 and December 2015. Data was collected and analysed with SPSS and descriptive and analytic statistic were performed. A value of p ≤ 0,05 was considered statistically significant. Results During 3 years, 82 patients (51 female, median age 69,5, ranging from 24 to 89, 48 with more than 65 years) were admitted due to PE, more frequently from the Emergency Service (86,6 %).The most common symptoms were dyspnea (61 %), chest pain (50 %) and syncope (24,4 %) and 54 patients were classified as low probability PE using Wells Criteria. Computed Tomography was positive in 97,6 % of patients. Using Pulmonary Embolism Severity Index (PESI) Score, 9 patients had very low risk score (mean age 40,4 ± 7,1 years), 13 low risk (mean age 50 ± 16,9 years), 26 moderate risk (mean age 70,4 ± 10,3 years), 14 high risk (mean age 71,4 ± 12,6 years)and 18 very high risk score (mean age 72,6 ± 14,2 years). Age was different between PESI Score classes (p < 0,001). The most common DVT risk factors were age, peripheral venous insufficiency (24,4 %), obesity (23,2 %), previous immobilization (17,1 %). In the ICU, 10 patients were admitted due to massive PE, 23 submassive PE, 11 due to hypoxemia and 23 for other reasons, more frequently due to the need constant monitoring. DVT was confirmed in 17 patients. Thrombolytic therapy was administered to 29 patients (13 submassive PE; 13 massive PE), 3 with registered complications (haemoptysis, haematuria and multiple hematomas and skin haemorrhage). Ventilation and vasopressor support were needed in 7,3 %. One patient was transferred to Intensive Care Unit; 1 died in ICU, 2 after discharge of ICU and 2 in the next 6 months. Conclusions Age is an important risk factor in patients admitted to ICU. Although 21 patients were classified as low/very low risk patients they were clinically unstable, had cardiac repercussion or needed constant monitoring. These patients were younger, with less comorbidities and with better physiological condition that may have underestimated PESI Score. This study highlights the importance of individualized and clinical evaluation, early therapy and the low rate of complications in patients submitted to thrombolytic therapy.

A351 Socioeconomic status in ICU survivors and post-hospital outcomes C.S. Socolovsky1, R.P. Cauley2, J.E. Frankel3, A.L. Beam4, K.O. Olaniran5, F.K. Gibbons6, K.B. Christopher7,8 1 Massachusetts General Hospital, Department of Medicine, Boston, MA, United States; 2Massachusetts General Hospital, Department of Surgery, Boston, MA, United States; 3Spaulding Rehabilitation Hospital, Physical Medicine & Rehabilitation, Boston, MA, United States; 4Harvard Medical School, Biomedical Informatics, Boston, MA, United States; 5Massachusetts General Hospital, Renal Division, Boston, MA, United States; 6Massachusetts General Hospital, Pulmonary and Critical Care Medicine, Boston, MA, United States; 7Brigham and Women's Hospital, Renal Division, Boston, MA, United States; 8Brigham and Women's Hospital, Channing Division of Network Medicine, Boston, MA, United States Correspondence: C.S. Socolovsky – Massachusetts General Hospital, Department of Medicine, Boston, MA, United States Intensive Care Medicine Experimental 2016, 4(Suppl 1):A351 Introduction Complex factors in ICU survivors such as health literacy, socioeconomic status and social environment may be predictive of adverse outcomes following hospital discharge but are not well studied. Objectives We hypothesized that Medicaid Insurance, a proxy for individual low socioeconomic status, would be associated with increased hospital readmission rates following hospital discharge.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Methods We performed a two center observational study of patients treated in medical and surgical intensive care units in Boston, Massachusetts. We studied 82,583 patients, age ≥ 18 years, who received critical care between 1998 and 2012 and survived hospitalization. The exposure of interest was Medicaid Insurance status. The primary outcome was unplanned 30-day hospital readmission. Adjusted odds ratios were estimated by multivariable logistic regression models with inclusion of covariate terms for gender, race, Deyo-Charlson index, type (surgical vs. medical), sepsis, acute organ failure, area deprivation index (proxy for area socioeconomic status and social environment) and age as a restricted cubic spline function. In a subset admitted to Spaulding Rehabilitation Hospital following hospital discharge, (n = 4,232) we evaluated the association of Medicaid Insurance status and rehabilitation hospital length of stay utilizing a negative binomial regression model. Results The cohort patients were 58 % male, 20 % nonwhite and 51 % surgical. 10 % of the cohort had sepsis and the mean age was 61.2 years. Medicaid Insurance was present in 10 %. Those with Medicaid Insurance were significantly younger, more frequently non-white, with higher sepsis and acute lung injury rates. Unplanned 30-day readmission rate was 14.1 %. 33.1 % were discharged to a care facility. 90-day post-discharge mortality was 6.5 % in patients with Medicaid Insurance and 7.3 % in patients without. Medicaid Insurance was a robust predictor of 30-day hospital readmission and remained so following multivariable adjustment. Patients with Medicaid Insurance have an adjusted OR of 30day readmission of 1.31 (95%CI, 1.22-1.41; P < 0.001) relative to patients without Medicaid Insurance. Further, patients with Medicaid Insurance have an adjusted OR for discharge to a care facility of 1.34 (95%CI, 1.26-1.42; P < 0.001) relative to patients without Medicaid Insurance. Finally, patients with Medicaid Insurance compared to those without Medicaid Insurance, are expected to have a 1.2 fold times greater rehabilitation hospital length of stay [adjusted IRR = 1.21 (95%CI 1.11, 1.32) P < 0.001]. Conclusions In critical illness survivors, individual socioeconomic status as reflected by Medicaid Insurance is a robust predictor of hospital readmission, placement in a care facility and rehabilitation length of stay. Disparities in ICU survivor outcomes are likely multifactorial involving individual, acute illness, hospital, posthospital, and neighborhood-level factors that alter optimum post ICU recovery.

A352 Outcome after critical care predicted by preceding ward length of stay J. Pennington, P. Zolfaghari Barts Health NHS Trust, Adult Critical Care Unit, London, United Kingdom Correspondence: J. Pennington – Barts Health NHS Trust, Adult Critical Care Unit, London, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A352 Introduction Prolonged hospital stay prior to admission to the intensive care unit has been shown to be independently associated with poorer outcome (1,2). Even a few hour delay in transfer from emergency department to intensive care worsens outcome (3). This may relate to an ongoing deterioration of physiological function while in hospital, potentially influenced by the process or disease state that culminates in admission to critical care. We investigated whether common illness severity scores e.g. Acute Physiology and Chronic Health Evaluation II (APACHE II) or Intensive Care National Audit & Research Centre (ICNARC), are significantly different in patients admitted after a prolonged ward stay. We describe mortality and ICU / hospital length of stay in said patients, in an order to elucidate predictors for outcome. Objectives Show higher length of stay before ICU admission as an independent predictor of outcome · Identify whether ICU scoring systems predict outcome in our population

Page 185 of 607

· Identify whether serum albumin or other surrogate marker of frailty (e.g. Creatinine) can help predict outcome? Methods Retrospective analysis of prospectively collected data of all admissions in ICNARC database to 44 bed adult critical care unit in University Associated London Major Trauma Centre over two-year period (1st January 2013 - 31st December 2014). Demographic data, APACHE II, ICNARC score, ICU mortality, and length of stay on ward preceding ICU, within ICU and overall hospital length of stay were collected. Data was analyzed using ANOVA tests, according to preceding ward length of stay. RESULTS. n = 4340 Mean APACHE 2 score for all admissions was 14.88 (sd 7.0) P < 0.01 ANOVA comparing Length of ICU stay between 0-7 and >28 days. P < 0.01 ANOVA comparing APACHE2 across all ward LOS groups. P < 0.01 ANOVA comparing serum albumin across ward LOS 0-7 days vs >28 days, and non significant across other groups. Conclusions Prolonged pre-ICU hospital admission is associated with longer ICU and hospital admission and generally higher ICU mortality. APACHE2 scoring and serum albumin predict outcome. References 1. Goldhill DR, McNarry AF, Hadjianastassiou VG, Tekkis PP. The longer patients are in hospital before Intensive Care admission the higher their mortality. Intensive Care Med. 2004 Oct;30(10):1908-13. 2. Higgins TL, McGee WT, Steingrub JS, Rapoport J, Lemeshow S, Teres D. Early indicators of prolonged intensive care unit stay: impact of illness severity, physician staffing, and pre-intensive care unit length of stay. Crit Care Med. 2003 Jan;31(1):45-51. 3. Chalfin DB, Trzeciak S, Likourezos A, Baumann BM, Dellinger RP, DELAYED study group. Impact of delayed transfer of critically ill patients from the emergency department to the intensive care unit. Crit Care Med. 2007 Jun;35(6):1477-83.

Table 112 (abstract A352). Demographics & scores LOS prior to ICU admission (days)

Number of patients

0–7

8–14

15–21

21–28

>28

3872

216

93

55

104

Age (years)

56.0 (sd 19.1)

62.8 (sd 15.5)

61.1 (sd 15.1)

60.5 (sd 16.0)

63.2 (sd 14.3)

APACHE II score

14.5 (sd 6.9)

17.5 (sd 7.4)

17.7 (sd 7.0)

18.1 (sd 7.7)

18.4 (sd 6.9)

APACHE II probability of death (%)

18.0 (sd 20.9)

24.1 (sd 22.3)

26.0 (sd 21.7)

24.2 (sd 24.1)

26.2 (sd 22.8)

ICNARC Score

13.9 (sd 8.7)

15.4 (sd 8.6)

15.4 (sd 9.1)

16.9 (sd 9.5)

15.5 (sd 8.5)

ICNARC mortality probability (%)

15.6%

20.9%

20.3%

24.3%

21.4%

Outreach involvement (%)

7.6%

26.4%

29.0%

43.6%

45.2%

LOS on ICU (days)

6.5 (sd 14.3)

8.2 (sd 16.0)

7.3 (sd 11.1)

10.8 (sd 22.3)

11.5 (sd 24.3)

LOS hospital (days)

20.9 (sd 27.1)

43.8 (sd 40.2)

51.3 (sd 41.6)

54.8 (sd 27.5)

108.7 (sd 70.3)

Fig. 138 (abstract A352). ICU Mortality vs Preceding ward Length of stay

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Page 186 of 607

groups. Inter-hospital ECMO retrieval is feasible for further development. The RESP score and SAVE score systems provides a tool for monitoring of our centre´s performance in comparison to ECMO centres worldwide. References 1. Schmidt M, et al.; Survival after Extracorporeal Membrane Oxygenation for Severe Acute Respiratory Failure - The Respiratory Extracorporeal Membrane Oxygenation Survival Prediction (RESP) Score. Am J Respir Crit Care Med Vol 189, Iss 11, pp 1374-1382 2. Schmidt M, et al.; Predicting survival after ECMO for refractory cardiogenic shock: the survival after veno-arterial-ECMO (SAVE) score. European Heart Journal First published online: 1 June 2015 Grant acknowledgment None.

Fig. 139 (abstract A352). Albumin vs Ward LOS

A353 A retrospective cohort study evaluating clinical outcomes of interhospital retrieval ECMO patients H.S. King, H.H.Y. Kong, H.P. Shum, W.W. Yan Pamela Youde Nethersole Eastern Hospital, Department of Intensive Care, Hong Kong, Hong Kong, China Correspondence: H.S. King – Pamela Youde Nethersole Eastern Hospital, Department of Intensive Care, Hong Kong, Hong Kong, China Intensive Care Medicine Experimental 2016, 4(Suppl 1):A353 Introduction A historical cohort study of patients who received Extra-corporeal membrane oxygenation (ECMO) in ICU of a regional hospital. Objectives To compare the clinical outcomes between in-house ECMO patients and inter-hospital retrieval ECMO patients. Methods Primary outcome was the ICU mortality rate between inhouse and inter-hospital retrieved ECMO patients. Mann-Whitney U tests and Fisher's exact tests were used for comparisons of continuous and categorical variables respectively. 2-tailed p-values < 0.05 represented statistical significance. Results Between 2009 to 2015, 112 patients received ECMO. Among them, 101 patients (77 VV-ECMO, 24 VA-ECMO) with complete data for further analysis. Overall mortality rate was 40/101 (39.6 %). Mean Acute Physiology and Chronic Health Evaluation (APACHE) II scores were 30.03 for in-house ECMO vs. 30.80 for inter-hospital retrieval ECMO (p = 0.774). For those patients who received ECMO for respiratory support (VV-ECMO, n = 77), their mean Respiratory Extracorporeal Membrane Oxygenation Survival Prediction (RESP) score (3.64 vs. 3.03, p = 0.060) and ICU mortality (34.8 % vs. 19.4 %, p = 0.199) were not significantly different between in-house and retrieved cases. Predicted hospital survival for RESP score risk classe II (score 3 to 5) was 76 %. For those patients who received ECMO for circulatory support (VAECMO, n = 24), their mean Survival After Veno-arterial-ECMO (SAVE) score (-7.2 vs. -2.75, p = 0.304) and ICU mortality (86.7 % vs. 55.6 %, p = 0.150) were not significantly different between in-house and retrieved cases. Predicted hospital survival for SAVE score risk class IV (score -9 to -5) and class III (score -4 to 0) were 30 % and 18 % respectively. For in-house ECMO patients, higher proportions of non-infective aetiologies for VV-ECMO (RESP score) and post-MI cariogenic shock for VA-ECMO were observed. These conditions have poorer prognosis and lower reversibility, which may account for the poorer RESP/SAVE scores (though statistically insignificant) and higher predicted and observed mortality. Conclusions Our findings did not show a statistically significant difference in ICU mortality rate between in-house and retrieval ECMO

A354 Evaluation of outcome from intensive care units in turkey: a prevalence study C. Kaymak1, N. Okumus2, A. Sari2, B. Erdogdu2, S. Aksun2, H. Basar1, A. Ozcan1, N. Ozcan1, D. Oztuna3 1 Ministry of Health, Ankara Training and Research Hospital, Anesthesiology and Reanimation Department, Intensive Care Unit, Ankara, Turkey; 2Ministry of Health, Departmant of Health Services, Ankara, Turkey; 3University of Ankara, Faculty of Medicine, Medical Biostatistics Department, Ankara, Turkey Intensive Care Medicine Experimental 2016, 4(Suppl 1):A354 Correspondence: C. Kaymak – Ministry of Health, Ankara Training and Research Hospital, Anesthesiology and Reanimation Department, Intensive Care Unit, Ankara, Turkey Intensive Care Medicine Experimental 2016, 4(Suppl 1):A354 Introduction The outcome in critically ill patients regarded with prognosis has many background effects of risk factors. An aging population and chronic diseases may also result in an increased number of patients in intensive care unit (ICU). Clinical results have revealed the need for outcome examination and guidance on the effective use of ICU. Objectives The aim of this study was to evaluate mortality among patients in Turkish ICUs. Regarding this, the present study analyzed APACHE II databases in critically ill patients at secondary and tertiary referral hospital ICUs in Turkey. Methods During the study period, clinical data that were collected concurrently for each patient contained demographic details, diagnostic category leading to ICU admission and APACHE II scores following ICU admission. Patients were followed up during ICU stay. The equation coefficients for APACHE II were supplied by APACHE Medical Systems. The mortality in intensive care units was analyzed according to APACHE II scores. The other attempts performed during ICU stay were also recorded. Results 13.313 patients were enrolled in this study. The 69.9 % of patients were > 60 years old and 53.8 % of them were male. The mean APACHE II score was 21.49. The ICU's mortality rate was 44.5 %. The mechanical ventilation was determined as the most performed attempt in ICU's with a ratio of 55.9 %. The ratio of central venous catheterization was found 53.5 %. The ratios of systemic infection and antibiotic administration were 63.6 % and 80.7 %, respectively. Conclusion In the present study, the patients hospitalized in ICU's of ministry, university, and private hospitals were analyzed all over Turkey. Early identification of patients at risk, both before admission and after discharge from ICU, may allow to prevent some of the physiologic abnormalities contributing to the APACHE II score. There was a wide difference in outcome for patients admitted to different ICU's by using risk adjustment methods.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

A355 Development of an ICU-specific questionnaire for patient-reported outcome measures J.A. Malmgren1, S. Lundin1, K. Torén2, M. Eckerström2, A. Wallin2, A.-C. Waldenström2, for the Section on Ethics of the ESICM 1 Sahlgrenska University Hospital, Dep. of Anaesthesiology and Intensive Care, Gothenburg, Sweden; 2University of Gothenburg, Gothenburg, Sweden Correspondence: J.A. Malmgren – Sahlgrenska University Hospital, Dep. of Anaesthesiology and Intensive Care, Gothenburg, Sweden Intensive Care Medicine Experimental 2016, 4(Suppl 1):A355 Introduction Growing interest in the long-term effects after critical care has formed investigator-led clinical research groups around the world. Work aiming at standardizing core outcome measures and instruments for randomized clinical trials is ongoing. Whether these outcome measures reflect the domains most valued by patients or if important issues are missing from the existing scales is unclear. Commonly used tools like SF-36 and EQ-5D are too unspecific. Objectives To develop and validate a questionnaire for patient-reported outcome measures (PROM) after critical care. Methods During a 24-months qualitative phase, 35 former ICU-patients were interviewed in a semi-structured way, providing detailed information on symptoms and difficulties in all areas of everyday life. Patients were recruited from the post-ICU clinic at Sahlgrenska University Hospital, covering both urban and rural areas. The interviews were recorded, transcribed, and issues were categorized into 13 hypothesized domains: cognitive, executive/fatigue, physical health, pain, mental health, daily activities, sleep, food/drink/smoking, sexuality, hearing/visual/dysphagia, intestinal and urinary problems, and return to work/financial situation. After searching the literature and commonly used assessment tools, additional issues were included. All issues were then rephrased into questions, with care taken to maintain only one conceptual entity per question, and with the recall period usually being the last month. Adequate scales for frequency, intensity and duration were used. All questions were validated face-to-face with another set of former ICUpatients and with non-ICU-treated controls to make sure the wording was easily understood and neither confusing nor upsetting. Results The questionnaire contains 271 questions. It is currently being tested in a pilot study with 650 patients, recruited six months to three years after discharge from the ICU, and 200 controls, matched for age and gender. The questionnaire is sent by mail after an invitation letter followed by a phone call, and returned in a pre-stamped envelope. Returned questionnaires are being scanned and data digitally imported into SPSS, where additional clinical data will be added. After comparison with controls, item reduction will follow, resulting in an ICU-specific PROM questionnaire. Conclusions A patient-centred, ICU-specific questionnaire will be available for long-term follow-up in the post-ICU clinic. Being a postal document, the patients do not have to return to the hospital to provide their information, making it and suitable for large-scale studies. References 2002 Brussels Roundtable Participants; Intensive Care Med; 2003; 29 Needham D, Pronovost et al.; Intensive Care Med; 2005; 31 Lind H, Waldenström AC, Dunberger G et al.; Br J Cancer; 2011; 105(6) Grant acknowledgment Göteborgs Läkaresällskap.

A356 One year review of oesophagogastrectomies in Queen Alexandra Hospital Portsmouth F.C. Riccio, D. Pogson Queen Alexandra Hospital, Critical Care, Portsmouth, United Kingdom Correspondence: F.C. Riccio – Queen Alexandra Hospital, Critical Care, Portsmouth, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A356

Page 187 of 607

Introduction Approximately 2000 oesophagogasterectomies are performed each year, with a 5 year survival of 25 % and 30 day mortality of 10 %1. Secondary analysis of ICNARC data has shown the median length of stay to be 2.8 days, and a readmission rate of 12.2 %2. We conducted a retrospective review of all post Oesophagogastrectomies ICU admissions in our hospital between Jan 2014-15, as part of a quality improvement project to reduce morbidity and length of stay (LOS). Objective To analyse the electronic record and chest Xray of every oesophagogastrectomy in order to compare our own LOS, patient characteristics and identify factors affecting LOS and unit morbidity. Method We analysed the electronic case record for each open or minimal access oesophagogastrectomy patient. Data was gathered on age, gender, analgesic method, pain scores, analgesia failure, vasopressor use, highest arterial lactate within 24 h, unit LOS and readmission rates. Results 42 patients were identified; 77 % male and median age 64.4. 25/42 (58 %) of patients had a combination of epidural and paravertebral analgesia, 17/42 (40 %) epidural alone and 1/42 (2 %) paravertebral alone. Median pain score on D1 was 1.4 (0-3). There was a 13/42 (30 %) epidural failure rate, 10 disconnections and 3 never effective. Vasopressors were used on D1 in 28/42 (67 %) patients. Median base excess on D1 was -1 (0-5.1) and median lactate on D1 was 1.93 (0.1-5). Unit LOS had a range of 3-52 days (median 6 days). Combinations of collapse and atelectasis were identified in 26/42 (62 %) patients on review of CXR, with one apical pneumothorax. Conclusion Epidural disconnection rate was very high, contributing to a longer unit LOS than national average and higher pain scores. We have introduced a training package on epidural care for our nurses and encouraged the use of tunnelled epidurals in combination with paravertebral catheters. The incidence of CXR changes after surgery in this cohort has not been studied previously. A large proportion of our patients had radiologically apparent collapse and even consolidation on arrival on ICU. This may contribute to the development of pneumonia in this setting if analgesia is imperfect as CPAP is relatively contraindicated due to the oesophageal anastomosis. We now encourage a longer period in recovery, with lung toilet and recruitment manoeuvres at the end of one-lung ventilation. References 1. Bartels H, Stein HJ, Siewert JR. Preoperative risk analysis and postoperative mortality of oesophagectomy for resectable oesophageal cancer. Br J Surg 1998; 85: 840-4 2. Park DP, Welch CA, Harrison DA, et al. Outcomes following oesophagectomy in patients with oesophageal cancer: a secondary analysis of the ICNARC Case Mix Programme Database. Critical Care 2009; 13(2):S1 (doi:10.1186/cc7868) Grant acknowledgment The help of Mr M Lympany, IT support in our department.

A357 Characteristics and outcomes of critically ill patients undergoing tracheostomy and transferred to the ward in a Brazilian public hospital A.C.P. Antonio1,2, A.F. Leivas1, F. Kenji1 1 Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil; 2Hospital Moinhos de Vento, Porto Alegre, Brazil Correspondence: A.C.P. Antonio – Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil Intensive Care Medicine Experimental 2016, 4(Suppl 1):A357 Introduction Placement of tracheostomy is commonly thought to allow a more secure and manageable airway and to facilitate weaning from mechanical ventilation. However, current literature suggests that tracheostomy has no impact on survival in unselected ICU patients, and it only transfers the mortality from the ICU to the ward. Moreover, in many circumstances tracheostomies are placed in patients who are at the end of their lives with little hope of meaningful recovery. Objectives To describe main characteristics and outcomes of tracheostomized patients discharged from ICU to ward in a public hospital.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Methods A retrospective descriptive study was conducted to analyze data from the electronic medical record. The setting was four adult medical-surgical ICUs (44 beds) in a tertiary care public hospital in southern Brazil. Data from 71 adult subjects who underwent a tracheostomy as part of their ICU management and were subsequently transferred to ward were obtained. Individuals who died at ICU during first admission were excluded. Demographic data, diagnoses on admission, comorbidities, duration of mechanical ventilation, ICU length of stay, end-of-life decisions and mortality were recorded. Results From January to December 2015, 104 subjects received tracheostomy. Thirty-two died during their initial ICU admission (30.4 %) and therefore were excluded of analysis. Of the remaining 73 individuals, twenty-eight died (38.3 %), and only four of whom were readmitted to the ICU within 48 hours of discharge. Mean age was 56.9 ± 17.7 years, 52.1 % were male and mean APACHE II score was 21.6 ± 6.8 points. Chronic neurologic disorders and cancer were main comorbidities (21.1 % and 14 %, respectively). Most common diagnosis were sepsis (33.8 %) and neurological emergencies (stroke, intracerebral hemorrhage, meningitis) [23.9 %]. Length of ICU stay was 30.7 ± 17 days and duration of mechanical ventilation was 23 (13 - 29) days. Lifesustaining treatments were withheld or withdrawn in twenty-five decedents. Seven subjects died in posterior hospitalizations at our institution over the period recorded. Conclusions Tracheostomy may represent a burden after ICU discharge, requiring high resource use and low survival rate. Indication for tracheostomy should be cautious, and efforts should be made to recognize patients who might clearly benefit from this technique to avoid unnecessary and unwanted prolonged mechanical ventilation. All the decisions we make in the ICU do have an important impact on future care needs. Knowledge of characteristics and outcomes may assist in identifying interventions to reduce the need for tracheostomy or improve outcomes.

A358 ICU readmissions and subsequent outcomes: a ten-year retrospective analysis E. James, P. Morgan East Surrey Hospital, Intensive Care, Redhill, United Kingdom Correspondence: E. James – East Surrey Hospital, Intensive Care, Redhill, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A358 Introduction Readmission to Intensive Care during the same hospital admission has been shown to be associated with a higher risk of mortality (1,2). Objectives To determine the readmission rate over a ten year period and subsequently analyse any effect this may have on mortality, length of ICU stay and whether readmission to ICU requires higher levels of care. Methods Retrospective analysis of our electronic patient information system for admissions from 2005 to 2016. Patients under 18 years old were excluded. Data collected included: mean age, mean stay, proportions of ventilated patients, patients commenced on RRT, unit mortality, hospital mortality and MRSA status. The two patient groups were compared: patients readmitted to ICU within their hospital admission (Group 1), and those patients who were not readmitted (Group 2). Results 7422 patients were identified in the 10-year period; 277 in Group 1 (3.7 %) and 7145 in Group 2 (96.3 %). Mean age in Group 1 was 64 compared to 61 in Group 2. Mean stay was 7.1 days and 5.3 days respectively. Conclusions Patients readmitted to ICU during their hospital stay were at significantly increased risk of dying in hospital, but not ICU itself. MRSA is more likely to be detected in patients who are readmitted to ICU. Further work should be carried out to investigate patients readmitted more thoroughly, with an aim of identifying patients most at risk of readmission and strategies to prevent readmission and improve discharge planning.

Page 188 of 607

References 1. Patients readmitted to the intensive care unit during the same hospitalization: Clinical features and outcomes. Chen LM et al. Critical Care Medicine: Nov 1998. Volume 26(11). pp1834-1841 2. Critically ill patients readmitted to intensive care units—lessons to learn? Metnitz PGH et al. Intensive Care Medicine. Feb 2003, Volume 29(2). pp 241-248

Table 113 (abstract A358). Results Readmitted (Group 1)

Not readmitted (Group 2)

p value (NS = Not significant)

Hospital mortality

100 (36.1)

1889 (26.4%)

0.008

Unit mortality

60 (21.6%)

1351 (18.9%)

NS

Mech. Vent

144 (51.9%)

3433 (48.1%)

NS

RRT

39 (14.1%)

1107 (15.5%)

NS

MRSA

20 (7.2%)

243 (3.4%)

0.002

A359 Pancreatitis in the west of Scotland intensive care population over a 20 year period G. Carroll, L. Gemmell, A. MacKay, C. Wright, J. Ballantyne Queen Elizabeth University hospital, Anaesthetics and Intensive Care, Glasgow, United Kingdom Correspondence: G. Carroll – Queen Elizabeth University hospital, Anaesthetics and Intensive Care, Glasgow, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A359 Introduction Pancreatitis is a common precipitant of critical illness and intensive care admission. Mortality from pancreatitis overall should be under 10 % and in severe pancreatitis under 30 % (1). Mortality risk is mulit-factorial but those at high risk are co-morbid, elderly, develop SIRS or progress to pancreatic necrosis. We sought to look at all of our pancreatitis admissions to ICU over a 21 year period and identify the average demographics and difference between survivors and non survivors. Objectives To identify patients admitted to ICU with a primary diagnosis of acute pancreatitis, and to compare predicted demographics and features of systemic inflammatory response between survivors and non survivors. We hypothesised that if you required Intensive Care for the management of severe pancreatitis, death is likely to occur at the beginning of your ICU stay due to overwhelming organ failure. If you were to survive the initial insult, it was hypothesised that you may survive to hospital discharge, although the length of hospital stay would be prolonged. We sought to test this theory with our patient group. Methods A retrospective audit of patients admitted to ICU in the Glasgow Victoria Infirmary, Southern General and Queen Elizabeth University hospital from 1994 to 2015. Patients were identified on Wardwatcher via a search of APACHE II diagnosis including pancreatitis. Data was collected from patient profiles on the Wardwatcher and TrakCare CIS. Results 182 patients were identified with an admission diagnosis of pancreatitis from 12704 patients admitted giving an incidence of 1.4 % of all ICU admissions. Other results are as demonstrated below with with all data being presented as mean and 95 % confidence intervals with p-values from Student´s unpaired t-test where applicable. Conclusions As could be predicted, pancreatitis is a diagnosis of the older male population in ICU, likely as a result of the concomitant problem of alcohol abuse in the West of Scotland. These patients have a higher than normal APACHE-II score and predicted mortality compared with unit averages. All SIRS criteria were met when looking at average data, hence why their likely admission to ICU. When comparing survivors to non-survivors, survivors were significantly more likely to be younger, with lower APACHE-II scores and predicted mortality. There was no difference in length of stay between groups nor degree of derangement of any of the SIRS criteria.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Page 189 of 607

References 1. Young, Thompson. Severe acute pancreatitis. Cont Ed Anaesthesia Crit Care. 2008: 8(4): 125-8

Table 114 (abstract A359). All patients (n= 182)

Survivors (n=95)

Non-survivors (n=87)

p-value

prescribing IV fluids may help improve patient outcomes in critical care. References 1 Pradeep, A., et al. "High volumes of intravenous fluid during cardiac surgery are associated with increased mortality." HSR Proceedings in Intensive Care and Cardiovascular Anesthesia 2 (2010): 287-296. 2 Lee, J., et al. "Association between fluid balance and survival in critically ill patients." Journal of internal medicine 277.4 (2015): 468-477. GRANT ACKNOWLEDGMENT N/A.

Age (years)

57.4 +/- 2.2

53.8 +/- 3

61.2 +/- 3.1

4 mmol/L were included. SH was defined by a rise in lactate > 4 mmol/L following an initial LC. Univariate and multivariable logistic regression analysis was used to test the association between SH and 28-day mortality. Results A total of 3390 patients were included, of which 341 (10 %) showed a SH, the remainder belonging to the LC group. The average age at admission was 63 (+/- 18) years, 58 % males and frequent comorbidities included hypertension, diabetes mellitus and congestive cardiac failure. The median time of initial LC was 15 hours and the delay between SH and death was 83 hours. The 28-day mortality rate in the SH group was 42 % versus 12 % in the LC group. SH was associated with 28-day mortality in unadjusted analysis (OR 5.41, 95 % CI 4.25 - 6.89; p < 0.001). After multivariable adjustments for patient demographics, co-morbid disease, new organ dysfunction and laboratory data, SH remained independently associated with 28-day mortality (OR 4.87, 95 % CI 3.70 - 6.41; p < 0.001).

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Page 190 of 607

Conclusions In this preliminary study, secondary hyperlactataemia was associated with mortality in critically ill patients. The association remained robust after multivariable adjustments. Secondary hyperlactataemia may serve as an early signal for deterioration in critically ill patients. Future studies should assess additional therapeutic options for patients with secondary hyperlactataemia in the ICU.

better Outcome as well as Hepatocarcinoma int the Tumor Group. Autoimmune/Toxic/Idiopatic is the most frequent in Fulminant Liver Failure Group and has the best Outcome of all LT Indications though, poor Outcome at Retransplant Group leads to a more careful selection of cases to get better Survival Ratios.

References 1. Nguyen HB, et al. Early lactate clearance is associated with improved outcome in severe sepsis and septic shock. Critical care medicine. 2004 Aug 1;32(8):1637-42. 2. Donnino MW, et al. Initial lactate and lactate change in post-cardiac arrest: a multi-center validation study. Critical care medicine. 2014 Aug;42(8):1804.

A363 Length of stay of Spinal Cord Injury (SCI) patients and their outcome while in Neuro ICU C. Hilasque St. George's Healthcare Trust, Neuro ICU, London, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A363

A362 Cumulative survival rates based on liver transplant indication over 20 years experience in Santiago de Compostela (Spain) J.R. Fernández Villanueva1, R. Fernández Garda1, A.M. López Lago1, E. Rodríguez Ruiz1, R. Hernández Vaquero1, C. Galbán Rodríguez1, E. Varo Pérez2 1 Hospital Clínico Universitario de Santiago de Compostela, Critical Care Unit, Santiago de Compostela, Spain; 2Hospital Clínico Universitario de Santiago de Compostela, Abdominal Transplant Unit, Santiago de Compostela, Spain Correspondence: J.R. Fernández Villanueva – Hospital Clínico Universitario de Santiago de Compostela, Critical Care Unit, Santiago de Compostela, Spain Intensive Care Medicine Experimental 2016, 4(Suppl 1):A362 809 Liver Transplants (LT) made at Hospital Clínico Universitario de Santiago de Compostela (Spain) in period 1994-2014, 4.25 % of all LT performed in Spain (19005 cases, according to the Spanish Liver Transplant Registry- RETH). Review and comparison of Cumulative Survival Rates based on LT Indication over 20 Years Experience. Materials and methods Retrospective and descriptive study of 809 LT cases performed from 1994 to 2014 at Hospital Clínico Universitario de Santiago de Compostela (Spain) based in our local LT Registration. Results 809 LT cases,12 of which were Hepatorenal Transplantations,35 required a Re-Transplantation. Media of LT:36 LT/year. Gender:79.35 % Male,20.64 % Female,Mean Age of 51 years old.Blood Group:A positive (49 %) and 0 positive (39 %).Most frequent LT Indication Groups:Liver Cirrhosis (LC): 541 cases,Tumors (T):186 cases,Fulminant Liver Failure (FLF):50 cases and Re-Transplant (RLT):35 cases.In the LC Group the most frequent: Alcohol-Related (ARC): 64 %, Hepatitis Virus C Cirrhosis (HVC): 22 %, Primary Biliary Cirrhosis (PBC): 5 % with Cumulative Survival Rate at 20 Years: ARC: 50 %, HCV: 35 % and PBC: 48 %. In T Group the most frequent indication: Hepatocarcinoma (HC): 90 %, Neuroendocrine Tumor (NET): 4 %, Klatskin Tumor (KT): 3 % with Cumulative Survival Rate at 20 Years: HC: 59 %, NET: 50 % and KT: 30 %. In FLF Group: Autoimmune/ Toxic/ Idiopatic (ATI): 72 %, Hepatitis B Virus (HBV): 10 % and Post-traumatic (PT): 6 % with a Global Cumulative Survival Rate at 20 years: 70 % in the total group. In RLT Group: Hepatic Artery Thrombosis (HAT): 35 %, Primary Allograft Dysfunction (PAD): 29 % and Recurrence of Underlying Disease (RUD): 21 % with a Global Cumulative Survival Rate at 5 years (20082013): 23 %. Conclusions Liver Chirrosis is the most frequent indication towards LT, followed by Tumors, with Hepatocarcinoma predominance and Fulminant Liver Failure. Cumulative Survival Rate globally is similar to the ones found out in other Transplantation Programs. In Liver Cirrhosis Group, Alcohol-Related Cirrhosis is the most frequent and with

Introduction Why do patients with Spinal Cord Injury (SCI) stay in Neuro ICU for a long time? Objectives A survey about the reasons why SCI patients stayed longer in NICU between 2013 and 2015 and their outcome Methods A local database (Wardwatcher) was used to search and collection of data for patients admitted with SCI between 2013 and 2015. Results Out of 3,539 admissions in NICU, 326 cases were related to spinal injury and 137 were considered as spinal cord injury and 93 % of admissions were related to trauma and 6 patients were readmitted while in the hospital between 2013 and 2015. There was an increase trend of admissions in NICU between 2013 and 2015 and the data showed that the average length number of days of patient stay was 3.6 days while spinal patients were 8.6 days. The Guidelines for the Provision of Intensive Care Services (2015) recommends a discharge from critical care to ward must be within 4 hours from the decision of the consultant, the data showed that 71 % of between 9 to 14 hours delayed discharges from NICU to ward are mainly caused by shortage of ward beds followed by delay on ward. 60-66 % of SCI patients required advance cardiac and respiratory organ support and 36 % were on neurological support. 86 % of the patients had tracheostomy. Only 19 patients were qualified for SCI study called Injured Spinal Cord Pressure Evaluation (ISCoPE) where the intraspinal pressure was measured in relation to traumatic spinal cord injury and this requires the patient to be in the spinal monitoring pressure for at least 7 days. On patients outcome, 75 % had improved, 13 % died and 12 % of the patients remain unchanged. While the hospital patient outcome on discharge, 83 % lived and 17 % died. Conclusions Though there were not many SCI patients admitted in the unit compared to other patients. SCI patients stayed longer by 5 days on average. Three main reasons were identified why they stayed longer; one was because of ISCoPE research, the other one was their dependency on advance organ support and lastly, the delayed discharges because of shortage of ward beds. The delayed discharges could indicate that the hospital needs to increase the bed capacity. The patients' outcome appears encouraging with their improved outcome. A further audit is necessary to see any changes. References Guidelines for the Provision of Intensive Care Services (2015). http:// www.ics.ac.uk/ics-homepage/latest-news/guidelines-for-the-provision-ofintensive-care-services/ (Accessed: 12 April 2015) Phang, I. and Papadopoulos, M., Intraspinal Pressure Monitoring in a Patient with Spinal Cord Injury Reveals Different Intradural Compartments: Injured Spinal Cord Pressure Evaluation (ISCoPE) Study. Neurocrit Care. 2015 Dec;23(3):414-8. doi: 10.1007/s12028-015-0153-6

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

QUALITY AND SAFETY II A364 Impact of real time random safety analysis in structure, process and outcome indicators: a multicenter study I. Oliva1, G. Sirgo1, M.C. Martin2, M. Olona3, M.C. Gilavert1, M. Bodí1 1 Hospital Universitari Joan XXIII, Intensive Care Unit, Tarragona, Spain; 2 Hospital Universitario de Torrejón, Intensive Care Unit, Madrid, Spain; 3 Hospital Universitari Joan XXIII, Department of Preventive Medicine, Tarragona, Spain Correspondence: I. Oliva – Hospital Universitari Joan XXIII, Intensive Care Unit, Tarragona, Spain Intensive Care Medicine Experimental 2016, 4(Suppl 1):A364 Introduction The risk of medical errors is high in intensive care medicine. Errors in healthcare may occur due to an unintended act or by omission. Errors of omission are more insidious and more difficult to identify. Our group previously developed and validated a new tool: the real time random safety audits (in Spanish: Análisis Aleatorios de Seguridad en Tiempo Real, AASTRE). It was effective in detecting and remedying errors of omission in real time. Objective The purpose of this study was to investigate the AASTRE impact in structure, process and outcome indicators through a multicenter study. Methods A prospective study was conducted over a period of 12 months in two adult patient ICUs. Safety rounds were conducted three days a week ascertaining the 37 safety measures (grouped into 10 blocks). In each round, 50 % of the patients and 50 % of the measures were randomized. The impact of this safety tool was analysed in indicators of structure (culture of safety, healthcare protocols), process (improvement proportion related to tool application, IPR-AASTRE) and outcome (mortality, average stay, rate of catheter-related bacteremias and rate of ventilator-associated pneumonia, VAP). Results 1241 patients-day were analyzed. Structure indicators: AASTRE was associated with an increased climate of security and creation / modification of protocols (sedation/analgesia and weaning). Process indicators: 12 of the 37 measures had an IPR-AASTRE > 10 %. Seven mesures had an IPR-AASTRE > 10 % in the three quarters analyzed. Six mesures showed a progressive decrease of the IPR over the study period. Nursing workloads and patient severity on the day of analysis were independently associated with a higher IPR-AASTRE in half of the blocks of variables. Outcome indicators: AASTRE was associated with a significant decrease in the rate of NAV. Conclusions AASTRE was associated with improvement in structure, process and outcome indicators. This tool also improved the care process and adherence to the clinical practice guidelines and proved to be most useful in situations of high care load and in patients with more severe disease. References Bodí M, Olona M, Martín MC, Alceaga R, Rodríguez JC, Corral E, Pérez Villares JM, Sirgo G. Feasibility and utility of the use of real time random safety audits in adult ICU patients: a multicentre study. Intensive Care Med. 2015 Jun;41:1089-98. Grant acknowledgement supported by grants from the Fondo de Investigación Sanitaria (Institute of Health Carlos III from Spain, FIS grants, project PI11/02311).

A365 Cost-effectiveness analysis of a quality improvement bundle to reduce mortality after emergency laparotomy C. Ebm1, G. Aggarwal2, S. Huddart2, N. Quiney2, M. Cecconi3 1 WPK Hopsital, Vienna, Austria; 2Royal Surrey County Hospital, Guildford, United Kingdom; 3St. George's Healthcare Trust, London, United Kingdom Correspondence: C. Ebm – WPK Hopsital, Vienna, Austria Intensive Care Medicine Experimental 2016, 4(Suppl 1):A365 Introduction Emergency laparotomy is associated with a high risk of mortality and morbidity, which leads to significant financial expenditures

Page 191 of 607

for the NHS. In a recent study, the Emergency Laparotomy Pathway Quality Improvement Care (ELPQUIC) program has shown that a set of sequential interventions have the potential to improve clinical outcome; however related costs or savings remain unknown. Objective This economic evaluation aims to evaluate costs and the cost-effectiveness of a clinical pathway for patients undergoing emergency laparotomy, compared to a historical cohort receiving standard care. Methods 299 consecutive patients in the control group were compared with 427 patients directed into a predefined pathway. To assess costs and cost-effectiveness, two decision models were constructed; the first model took hospitals management perspective, the second model took a societal perspective and evaluated lifetime costs and quality adjusted life years. Results One time implementation costs of £23,406.7/hospital (£1,399.0-£31,793.0) for training, supervision, purchase of equipment and pharmaceuticals can be expected. However, these costs were offset after treating 26 patients, mainly due to reduced LOS and lower complication rates. The long-term model showed that the intervention is both more effective (2.4 month) and leads to lower costs to society (cost savings of £899.6/patient). The incremental cost-effectiveness ratio is £-4,015.9, meaning the new pathway is the dominant strategy and should be recommended to decision makers. Conclusion A bundled pathway to improve clinical care for patients undergoing emergency laparotomy has shown to reduce mortality, seems cost-effective and has the potential to improve clinical outcome and lower costs for society. Decision makers need to adopt a long-term vision and be prepared to make one off investments to lower future costs. References Huddart S, Peden CJ, Swart M, et al. Use of a pathway quality improvement care bundle to reduce mortality after emergency laparotomy. Br J Surg. 2015; 102(1): 57-66.

A366 Delay in the emergency department: does it matter? S.M. Fernandes1,2, J. Santos Silva1, J. Gouveia1, D. Silva1, R. Marques1, H. Bento3, A. Alvarez1, Z. Costa Silva1 1 Hospital de Santa Maria/CHLN, Serviço de Medicina Intensiva, Lisboa, Portugal; 2Faculdade de Medicina, Universidade de Lisboa, Lisbon, Portugal; 3Hospital de Santa Maria/CHLN, Lisboa, Portugal Correspondence: J. Gouveia – Hospital de Santa Maria/CHLN, Serviço de Medicina Intensiva, Lisboa, Portugal Intensive Care Medicine Experimental 2016, 4(Suppl 1):A366 Introduction Delayed admission from the emergency department (ED) to intensive care unit (ICU) has been associated with increased ICU/hospital mortality, length of stay and total cost of care1,2. However, there is little data to support the use of any particular time frame as an indicator of quality of care. Objectives We aimed to investigate the relationship between time in ED and ICU mortality in a tertiary university affiliated hospital. Methods We conducted a retrospective study of patients admitted from 2011 to 2015. We included patients admitted directly from the ED based upon administrative registries Transfers from other hospitals including administrative passages through the ED and admissions after emergency surgery were excluded. Primary outcome was ICU mortality. Secondary outcomes included nosocomial infection, ventilation time and hospital mortality. Statistical analysis included univariate and multivariate logistic regression. Results We included 555 admissions (548 patients) with a mean SAPSII level of 43.0 ± 18.8 and ICU mortality of 19.8 % (n = 110). Mean age at the time of admission was 57.8 ± 18.2y, and 60.0 % were male. Patients were admitted due to medical conditions (78.2 %) and trauma with no need for emergent surgery (21.8 %). Mean time in the ED was 8.4 hours (±5.7), with a skewed distribution to the left. Significantly, there was a significant decrease of ED time during this five year period (-1.9 (-3.5- -0.5, p = 0.009) hours in 2015 compared to 2011). In the univariate analysis, we found no association of time in

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

the ED with ICU mortality (OR = 1.01; p = 0.73). This result was not altered after adjusting for SAPSII, age, gender, trauma, acute kidney injury or year of admission. Of relevance, there was also no association with ICU mortality and an excess of 12 hours in the emergency department. Finally, there was no association between time spent in the ED and hospital mortality (OR = 1.2; p = 0.13), ventilation time (p = 0.6) or nosocomial infection (p = 0.8). Conclusions Unnecessary time spent in the ED has the potential to adversely affect the care and to overwhelm crucial ED resources. In our single-center retrospective study we did not find any association between the absolute time spent in the ED and ICU mortality, and we cannot recommend any clinically relevant specific time frame. Meanwhile, patients should be transferred to the ICU from the emergency department as soon as possible and further research is needed to develop potential quality indicators. References 1. Chalfin DB, et al. Impact of delayed transfer of critically ill patients from the emergency department to the intensive care unit. Crit Care Med. 2007. 2. Cardoso LT, et al. Impact of delayed admission to intensive care units on mortality of critically ill patients: a cohort study. Critical Care 2011. Grant acknowledgment No grant support.

A367 The utility of a daily checklist in patients under mechanical ventilation to achieve low tidal volumes D. Díaz Diaz, M. Villanova Martínez, E. Palencia Herrejon, A. Martinez de la Gandara, G. Gonzalo, M.A. Lopez Hospital Universitario Infanta Leonor, Intensive Care Unit, Madrid, Spain Correspondence: D. Díaz Diaz – Hospital Universitario Infanta Leonor, Intensive Care Unit, Madrid, Spain Intensive Care Medicine Experimental 2016, 4(Suppl 1):A367 Objectives Several studies have demonstrated a decrease of mortality under mechanical protective ventilation with low tidal volumes (6-8 ml per kilogram of predicted body weight) not only in patients with acute respiratory distress syndrome (ARDS), but also in ICU patients without criteria of pulmonary injury, The aim of this study is to evaluate the utility of a daily checklist applied to all the patients under mechanical ventilation (MV) admitted to an ICU as well as to identify the factors influencing the achievement of the goal to keep the tidal volume at 6-8 ml per kilogram of predicted body weight Methods A prospective study including all the patients admitted to an ICU during a four-months study period. A daily-based register was implemented for all the ICU patients under MV during the morning, the afternoon and the night shifts. We analyse the degree of compliance with the checklist, as well as the influence on the achievement of the goal for tidal volume of the following factors: height, sex, type of MV (Continuous Mandatory Ventilation-CMV, Pressure Control VentilationPCV, Bi-level positive airway pressure-BiPAP, Synchronized Intermittentmandatory Ventilation-SIMV, Pressure Support Ventilation-PSV), gaseous exchange (relation paO2/FiO2), pulmonary mechanics (plateau pressure). Results We registered 883 measurements (36 % in the morning shifts, 31 % in the afternoon and 33 % in the night shift) for all the patients with MV (either with ARDS or without criteria of pulmonary injury). No patient developed ventilator induced lung injury (VILI). The percentage of measurements outside the established goal of tidal volume was 28,7 % (CI 95 % 23.9-33.9 %) without differences between the shifts. The average tidal volume was of 443.9 ml (range 431-457) that corresponds to 7.5 ml per kilogram of predicted body weight (IC 95 %: 7,35-7,66). In univariate analysis the factors associated to being outside of the goal for tidal volume were: female sex, height and the Pressure Support Ventilation-PSV as opposite to mandatory modes with no influence of the shifts of work, the relation paO2/FiO2 and the plateau pressure. The multivariate logistic regression analysis showed that the factors independently associated with being outside of the goal for tidal volume were female sex and the Pressure Support Ventilation-PSV.

Page 192 of 607

Conclusions In almost 30 % of the ICU patients under MV the goal of low tidal volumes is not achieved, especially in female patients and in those under Pressure Support Ventilation. The establishment of a daily checklist for patients under VM is feasible and could influence in a low frequency of VILI in patients with ARDS and also in those without criteria of pulmonary injury.

A368 Analysis of the serious adverse events reported at the heart surgery intensive care unit P. Ruíz de Gopegui Miguelena, C.I. Bernal Matilla, P. Sánchez Chueca, M.D.C. Rodríguez Longares, R. Ramos Abril, A.L. Ruíz Aguilar, R. Garrido López de Murillas Hospital Universitario Miguel Servet, Medicina Intensiva, Zaragoza, Spain Correspondence: P. Ruíz de Gopegui Miguelena – Hospital Universitario Miguel Servet, Medicina Intensiva, Zaragoza, Spain Intensive Care Medicine Experimental 2016, 4(Suppl 1):A368 Introduction Fostering a safety culture is one of the most important landmarks when it comes to provide quality assistance in an Intensive Care Unit (ICU). The detection and analysis of the incidents and adverse events that affect our patients are the cornerstone for the development of programs and protocols leading to safe health assistance with high standards of quality. Objectives Knowing the nature, casuistry and the underlying elements of adverse events (AE), rated as serious, according to the Protocol on Adverse Events Reporting established in our Unit. Methods 266 notified EA, reaching 201 patients, were analysed through the Adverse Events Reporting System (AERS) established in the Heart Surgery Intensive Care Unit (HSICU) from January 2014 to January 2016. Among these, we rated as serious those that caused to the patients a temporary or permanent damage, extending their hospitalization, compromising their lives or needing surgery in order to save their lives and contributing or causing their demises. In our AERS, those rated as F-I according to the catalogue of the National Coordinating Council for Medication Error Reporting and Prevention (NCCMERP). Results 21 AE rated as serious were notified. Among these, 13 were rated H category (incident that compromised the patient's life and needed care to keep him alive) and 2 were rated I category (the incident contributed or caused the patient's demise). The last were related to surgery complications. All of them needed medical attention, being the majority of them discharged to ordinary ward hospitalization. There were not relevant differences as far as the urgency of the hospitalization, the clinical profile and the assistance needed in the moment of the AE. In 30 % of the AE, the equipment and available resources were pointed out as contributing factors; in 35 % of cases it was related to formation and training; in 40 % of cases it was due to elements related to the patient and in 45 % of the cases it was elements related to the performance. The most frequent serious safety issues were surgery complications or damages related to invasive procedures, three of them related to the handling of breathing devices and two of them with reaction after blood transfusions. In 75 % , according to professional advice, the AE was deemed avoidable. Conclusions The AE rated as serious were 7.7 %. The most common feature ,was their avoidability and their relation with formation and training in the techniques used during the assistance. Beyond the cases of death, the serious AE involved an extension in the stay at the hospital and a rise in the morbidity and mortality rate. According to this, the intervention over factors capable of improvement, such as formation and training of all the professionals involved in the healing process, is compulsory. References Merino P, et al. Adverse events in Spanish intensive care units: the SYREC study. Int Qual Health Care. 2012 Apr; 24(2):105-113. Grant acknowledgment None.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

A369 Evaluation of the influence of an adequate briefing on reducing adverse effects in an intensive care unit R. Fernández Fernández1, P. Morales Laborías2, M.A. Díaz Castellanos2, M.E. Morales Laborías2 1 Hospital Comarcal Santa Ana, UCI, Motril, Spain; 2Hospital Comarcal Santa Ana, Motril, Spain Correspondence: R. Fernández Fernández – Hospital Comarcal Santa Ana, UCI, Motril, Spain Intensive Care Medicine Experimental 2016, 4(Suppl 1):A369 Introduction A well-defined briefing can help to improve the exchange of information among professionals in the shift change and strengthens the safety culture. Objectives General: To promote safety culture in the ICU of the HGB of Motril. Specific: Encourage communication between the interprofessional team. Diminish the appearance of adverse effects (EAs). Implement use of safety briefings to increase patient safety. Since 2008 daily meetings are held in the morning shift among medical and nursing professionals, and among nurses on each shift, for exchange of information and communication regarding patients admitted, assessing the risk of occurrence of AEs each patient. Methods Description of briefing protocol designed in 2015 in our ICU wich has 8 beds with a total income of 582 in 2015. The personnel that make the interprofessional team are 7 doctors, 14 nurses and 10 assistants.In 2014 a total of 18 EA were reported 33.33 % being derived from a care or clinical procedure, 22.22 % resulting from an accident patien, 16.67 % of medical equipment and 11.11 % patients´falls. The rest was due to other causes.A working group was created to develop and promote safety culture. In November 2015 implementing security briefing was performed after consensus of all nursing professionals. The variables were entered in the checklist (yes/no) three times a day (in the morning, in the afternoon and at night) and were: mechanical ventilation, vasoactive drugs, fasteners in the weaning, pressure ulcers, drains, revision infusion rate, review alarms, repositioning, mobilization endotracheal tube, nasogastric tube and urine probe, wristband. Mandatory measures of pneumonia Zero program were included: Ambu review and aspirator, strict hand hygiene before handling air, headboard position 30°-40°, check neumotaponamiento pressure, chlorhexidine mouthwash. Results – Regulated exchange of information among professionals. – The number of EA reported has decreased by 60 % since the introduction of the use of the new briefing protocol. – Briefing accepted by 100 % of the professionals. – 97 % of staff completed the checklist successfully. Conclusions It has reduced the number of adverse effects in more than a half, thus we have demonstrated the effectiveness of the protocol, it is also very important that it has been accepted by almost all of the staff, this data shows that it is simple and easy to carry out. With these data we have sensitized the staff of the importance of information for patient safety. We think that with this work we can help other units. Grant acknowledgment To ICU stafffor making this work possible.

A370 Daytime activation of rapid response team might improve patient safety J. Cho1,2, J. Kim1, J. Park1, S. Woo2,3 1 Inha University Hospital, Pulmonary and Critical Care, Incheon, Republic of Korea; 2Inha University, Internal Medicine, Incheon, Republic of Korea; 3Inha University Hospital, Cardiovascular Medicine, Incheon, Republic of Korea Correspondence: J. Cho – Inha University Hospital, Pulmonary and Critical Care, Incheon, Republic of Korea Intensive Care Medicine Experimental 2016, 4(Suppl 1):A370

Page 193 of 607

Introduction Rapid response systems (RRS) may slightly reduce cardiac arrest in the hospital. RRS has an efferent limb such as rapid response team (RRT) and an afferent limb that has a process of timely detection and activation. RRS activates mostly 24 hours/day and 7 days/week by a multidisciplinary team. In middle-income country, unlike EU and USA the personal resources are limited. RRT call were most common during working hours in Australian study1. Objectives Daytime activation of RRT may reduce cardiac arrests rates. Methods From October 2015, Inha university hospital rapid response team (INHART) had organized and activated. The INHART consisted with three doctors (two were pulmonologist, one was cardiologist) and one nurse who has experienced in intensive care unit. The activation time was from 9 AM to 5 PM. We collected the numbers of hospital discharge, cardiac arrest events (CAE) retrospectively for 2 years (before INHART) and prospectively from October 2015 to February 2016 (after INHART). CAE per 1,000 discharge were compared between before and after INHART. Nonparametric test was performed. Results Mean CAE per 1,000 discharge were 4.03 before INHART and 2.93 after INHART (p = 0.041). CAE was reduced by 27 % after INHART. Conclusions Despite of short-term operation, daytime only activation of RRT reduced cardiac arrests events. After long-term application, RRT would ameliorate cardiac arrests and improve patient safety in a general hospital. References 1. Jones D. The epidemiology of adult Rapid Response Team patients in Australia. Anesth Intensive Care 2014; 42:213-219 Grant acknowledgment Inha University Research Grant.

A371 A survey of patient safety culture across two critical care units T. West1, E. Powell2, A. Rimmer2, C. Orford2, N. Jones2, J. Williams2 1 University Hospital of Wales, Anaesthetics, Cardiff, United Kingdom; 2 Royal Gwent Hospital, Newport, United Kingdom Correspondence: T. West – University Hospital of Wales, Anaesthetics, Cardiff, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A371 Introduction Patient safety in the ICU is the prevention of injury or harm arising as a result of the process of care, rather than as a result of the underlying disease process. The incidence of medical errors and adverse events in ICU is surprisingly high1. Systems to improve patient safety, such as checklists, are only effective when implemented within a receptive and positive safety climate. A positive safety climate is associated with a measurable decrease in error rates2. Objectives To evaluate the safety climate across both critical care units within Aneurin Bevan University Health Board (ABUHB). Methods We conducted an anonymous online survey across all staff in critical care within ABUHB. The survey consisted of questions validated by the Agency for Healthcare Research and Quality, part of the United States Department of Health and our own questions regarding staff's incident reporting behaviour. We ran the survey for a month and aimed to get a 50 % return rate from the 180 staff working across both sites. Results We received 114 responses and 101 completed surveys from across the multi-disciplinary team. Eighty eight percent rated overall patient safety as very good or excellent. The survey reported a supportive environment (83 %) where people worked well together as a team (97 %), and treated each other respectfully (90 %). Generally patient safety was not sacrificed to get more work done (66 %), and there was no pressure to take shortcuts to speed up work rate (85 %). "Crisis mode" was precipitated by an inability to discharge wardable patients (52 %), rather than lack of staff (27 %), or the unit being too full (33 %).

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Page 194 of 607

Staff would generally discuss adverse incidents amongst themselves informally (85 %), and in many cases (63 %) implement a solution following this discussion. Feedback as a result of adverse event reports was highlighted as a deficiency with only 33 (31 %) staff feeling they received this, similarly 39 (37 %) of staff did not feel they were adequately informed about any errors that happened on the units. Conclusions Whilst our results indicate a positive work environment, there are areas for improvement, specifically feedback following adverse events. Designated nurses to attend morbidity and mortality meetings and subsequent publication of patient safety bulletins will improve feedback to nursing staff. We are going to instigate a designated patient safety board to publicise patient safety events and resultant changes in practice. Little is currently published regarding UK ICU safety surveys to allow benchmarking of our results. The authors feel UK ICUs should consider collection and publication of this data.

had direct effect on the I/AE. The mean duration of hospitalization stay before the AE was 7.68 days (statistic standard deviation, SSD, 30.1); the mean duration of hospitalisation was 16.45 days (SSD 25.26) and the mean stay in the Intensive Care Unit 8.69 days (SSD 13.48), (p < 0.05). Conclusions The majority of the AE were deemed evitable and neither had they important effects on the evolution of the patient nor extended the patient's stay in the ICU. A longer stay in the ICU was related to a bigger incidence of AE. (p < 0.05). Formation and training are key elements in order to avoid AE and they are part in the measures currently established in our ICU in order to avoid these I/AE.

References 1 ICS, FICM. Guidelines for Provision of Intensive Care Services. 2015. 1stEdition. 167-171. 2 Valentin A, Schiffinger M, Steyrer J et al. Safety climate reduces medication and dislodgement errors in routine intensive care practice. Intensive Care Medicine 2013;39: 391-398.

A373 Reliability of clinical impact grading by health professionals of common prescribing errors and optimisations in critical care patients R.S. Bourne1, R. Shulman2, M. Tomlin3, G.H. Mills1, M. Borthwick4, W. Berry5 1 Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, United Kingdom; 2University College London Hospitals NHS Foundation Trust, London, United Kingdom; 3Southampton University Hospitals NHS Trust, Southampton, United Kingdom; 4Oxford University Hospitals NHS Foundation Trust, Oxford, United Kingdom, 5Guy's and St Thomas' NHS Foundation Trust, London, United Kingdom Correspondence: W. Berry – Guy's and St Thomas' NHS Foundation Trust, London, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A373

A372 Experience of a report protocole on incidents and adverse events in a heart surgery intensive care unit C.I. Bernal Matilla, P. Ruiz de Gopegui Miguelena, P. Sánchez Chueca, R. Ramos Abril, M.D.C. Rodríguez Longares, A.L. Ruíz Aguilar, R. Garrido López de Murillas Hospital Universitario Miguel Servet, Medicina Intensiva, Zaragoza, Spain Correspondence: C.I. Bernal Matilla – Hospital Universitario Miguel Servet, Medicina Intensiva, Zaragoza, Spain Intensive Care Medicine Experimental 2016, 4(Suppl 1):A372 Introduction Clinical safety has become on one side a quality standard and on the other side an ethical obligation in the daily work of an Intensive Care Unit (ICU). The establishment of standardized protocols and appropriate training in the area of safety have become a must when it comes to its development. Also all professional categories have to be involved in this process. Objectives Analysis of the incidents and adverse events (I/AE) reported after the establishment of a report protocol (RPIAE) in a Heart Surgery Intensive Care Unit (HSICU). Methods The Clinical Safety Group of the Intensive Care Service, responsible for the design of the RPIAE, was created in January 2014. Thus, the RPIAE was introduced in the HSICU in March 2014. This methodology reviews the adverse events (AE) detected in every single patient that was looked after in this unit between March 2014 and October 2015.AE: non-intentional damage caused during or as a consequence of the medical attention received and non-related to the evolution or eventual complications of the initial illness. The information was gathered in an Excel data base and the statistical analysis was performed through the SPSS and Epidat programs. Results 254 AE's were reported on 189 patients, 67 % of which were males. 2.7 % of them were admitted to hospital as a result of an AE. 90.9 % were discharged and 23 patients passed away. 50 % of them were reported by the nursing staff; 39.8 % by doctors. 88.9 % took place in the Intensive Care Unit., 44.9 % of which during the morning shift. 79.9 % while the Unit was holding 100 % occupancy. 24 % of the AE affected the patient in a way that they needed further monitoring and/or intervention in order to check that no damage had been caused. 7.1 % compromised the patient's life and 1.6 % contributed or caused the demise of the patient. 57.9 % of them were doubtlessly avoidable. 22.8 % had their origin in the unplugging/removal of accesses/probes/catheters/sensors/tubes and 7.9 % had their origin in the handling of breathing devices and mechanical ventilation. In 43.75 % of the cases the lack of training and formation

References Merino P, et al. Adverse events in Spanish intensive care units: the SYREC study. Int Qual Health Care. 2012 Apr; 24(2):105-113. Grant acknowledgment None.

Introduction Medication errors are common in critically ill patients, [1] and have the potential to adversely affect patient safety [2]. Whilst adverse events, including those caused by medications, have a direct relationship with patient outcomes, [3] the potential clinical impact of individual medication errors remain less clear. Moreover, the complexity of critically ill patient care means frequent optimisation of medications is required to ensure effective therapy. Currently the clinical impact grading of medication errors and optimisations remains subjective in clinical practice and research methodologies. Objectives To identify the reliability of clinical impact grading by health professionals (i) between professions (inter-rater) and (ii) within professions (intra-rater)) of common prescribing errors and optimisations in critical care patients; and (iii)identify representative clinical impact grades for each example. Methods 50 representative medication error and 55 optimisation cases from the PROTECTED UK study, [4] were sent electronically to a purposive sample of 30 healthcare professionals (10 ICU consultants, specialist pharmacists and specialist nurses) in 5 UK sites. Each rater graded the error or optimisation case for severity of clinical impact using a 5-point categorical scale (no, minor, moderate, severe and life threatening/saving clinical impact). Inter-rater reliability was tested using a linear mixed model. Intra-rater reliability was tested using the intra-class correlation (ICC) with a two-way random effects model using absolute agreement. Results The majority of errors and optimisations (both 76 %) had a modal clinical severity grade of moderate or higher. Error cases: doctors graded clinical impact significantly lower than pharmacists (-0.25; p < 0.001) and nurses (-0.53; p < 0.001), with nurses significantly higher than pharmacists (0.28; p < 0.001). Optimisation cases: doctors graded clinical impact significantly lower than nurses and pharmacists (-0.39 and -0.5; p < 0.001 respectively). There was excellent intra-reliability (ICC) grading for pharmacists (0.882 & 0.887; p < 0.001) and doctors (0.794 & 0.829; p < 0.001)) but only fair to good

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

for nurses (0.429 & 0.739; p < 0.001), for optimisations and errors respectively. Conclusions Over one hundred common prescribing errors and optimisations had modal clinical impact grades recorded for potential application in clinical practice and research. The inter-professional variability highlights the importance of multidisciplinary perspectives in assessment of medication errors and optimisations in clinical practice and research. References 1. Rothschild JM, et al. Crit Care Med 2005; 33:1694-700. 2. Manias E, et al. Br J Clin Pharmacol 2012; 74:411-23. 3. Garrouste-Orgeas M, et al. Am J Resp Crit Care Med 2010; 181:134-42. 4. Shulman R, et al. J Crit Care 2015; 30:808-13. Grant acknowledgment Supported by Sheffield Hospitals Charity (Registered Charity Number 1059043).

A374 Incidence of adverse events in adult intensive care unit patients D. García Huertas, F. Manzano, F. Villagrán-Ramírez, A. Ruiz-Perea, C. Rodríguez-Mejías, F. Santiago-Ruiz, M. Colmenero-Ruiz Complejo Hospitalario de Granada, Granada, Spain Correspondence: D. García Huertas – Complejo Hospitalario de Granada, Granada, Spain Intensive Care Medicine Experimental 2016, 4(Suppl 1):A374 Introduction Adverse events (AE) are frequent in patients admitted in intensive care unit (ICU) for different causes. Objective To determine the incidence, characteristics and results of adverse events in critically ill patients requiring mechanical ventilation (MV). Methods An observational, prospective and single-center study, conducted in a medical-surgical ICU over a period of 12 months (2015), in patients who require MV > 24 hours. The variables studied were: demographic data, extubation unscheduled, severe obstructions airway, cardiac arrest, atelectasis, loss of medical devices, septic shock, duration MV, ICU and hospital stay, and ICU and hospital mortality. Statistical analysis: descriptive, bivariate (chi 2 and t-Student) and multivariate logistic regression analysis. Results 330 patients were included, with APACHE II score 23.5 ± 7, age 61 ± 15 years, body mass index 28 ± 5, total SOFA on day 1, 8.9 ± 3.5 pts, and 67 % were males. Renal replacement therapies were 14.6 %, duration of ICU stay 17 ± 18, hospital stay 26 ± 22 days, duration of MV 10.8 ± 11 days, cardiac arrest 4.9 %, septic shock 39.8 %, airway obstructions 33.4 %, unscheduled extubation 9.1 %, loss of medical devices 8.25 %. ICU mortality was 33.7 % and hospital mortality 38.6 %. Total incidence of AE was 42.2 %. Predictive factors associated with hospital mortality were APACHE II score (OR 1.048, 95%CI 1.01 to 1.08), septic shock (OR 1.27, 95%CI 0.76 to 2.11), airway obstruction (OR 0.81, 95%CI 0.49 to 1.35), unscheduled extubation (OR 1.80, 95%CI 0.73 to 4.45), and loss of medical devices (OR 1.60, 95%CI 0.65 to 3.9). Conclusion The occurrence of AE in critically ill patients is high. Strategies to reduce the incidence of adverse events may eventually improve outcomes in these patients.

A375 Development of patient-reported domains describing quality of life after sepsis C. König1,2, B. Matt1, A. Kortgen1, C.S. Hartog1,2 1 Jena University Hospital, Department of Anesthesiology and Intensive Care Medicine, Jena, Germany; 2Jena University Hospital, Integrated Research and Treatment Center for Sepsis Control and Care (CSCC), Jena, Germany Correspondence: C. König – Jena University Hospital, Department of Anesthesiology and Intensive Care Medicine, Jena, Germany Intensive Care Medicine Experimental 2016, 4(Suppl 1):A375

Page 195 of 607

Introduction Sepsis survival rates have improved in recent years and it becomes more and more important to asses long-term outcomes in order to align treatment with what is important for patients. Quality of life (QoL) is an important patient-reported outcome (PRO) that is commonly used as an endpoint in clinical studies. However, the most frequently used instruments, eg. SF-36 and EuroQoL-5D, were developed without input from intensive care survivors and may thus lack validity in these populations. Quality of life domains which matter to sepsis survivors have not yet been elicited. Objectives Identification of the domains of quality of life of sepsis survivors according to their own perception and priorities. Methods Open-ended face-to-face or telephone interviews were conducted with sepsis survivors from interdisciplinary ICUs of a German university hospital. Interviews were transcribed verbatim with a fixed set of transcription rules and analyzed qualitatively with the interpretative phenomenological analysis (IPA) approach. Codes and subcodes were applied to the data and a codebook was developed. A second experienced researcher unfamiliar with the data evaluated the codebook; differences were solved by discussion. Codes were clustered into domains and evaluated for their importance by stakeholders (survivors, relatives of survivors, intensive care physicians, nurses, psychologists, rehabilitation physicians) through a Delphi process. Results Fifteen participants (7 female, 8 male) had a mean age of 62 years (range: 27 - 87 years) with mean time after onset of sepsis of 11 months (range: 5 - 40 months). Mean interview duration was 68 minutes (range: 34 - 95 minutes). The final codebook comprised 16 codes and 99 subcodes; data saturation was already 95 % after 7 interviews. Initial agreement between the two researchers about the codebook was 83 % and could be improved to 100 % after minor changes. 10 core domains of quality of life could be identified. The most relevant domain for sepsis survivors' quality of life is reintegration to normal living. Survivors reported that their quality of life depended more on how much they could participate in everyday life and less on each single persisting physical and psychological deficit. Other domains include: ability to walk, fatigue, family and partnership, self-perception and control over one's life. Conclusions This study highlights the domains that are important for the quality of life of sepsis survivors as assessed by the survivors' own perception. Identified domains are only partly assessed by commonly used quality of life instruments. Reintegration as the most important domain from a patient' perspective is a novel concept in the context of quality of life after sepsis. These domains will help to improve quality of life questionnaires which provide patient-relevant outcomes measures. Grant acknowledgment Federal Ministry of Education and Research (BMBF), Germany, grant number 01EO1502.

A376 Chest drains: is the triangle of safety really safe? An ultrasonographic study of enshrined practice to improve patient outcome and safety A. Wong, C. Balan, G. Barker Oxford University Hospitals NHS Trust, OCCULAR Group, Adult Intensive Care Unit, Oxford, United Kingdom Correspondence: A. Wong – Oxford University Hospitals NHS Trust, OCCULAR Group, Adult Intensive Care Unit, Oxford, United Kingdom Intensive Care Medicine Experimental 2016, 4(Suppl 1):A376 Introduction Intercostal chest drain (ICD) insertion is a relatively common procedure but can be associated with significant complications including bleeding and organ perforation. Indications for insertion include management of hemothorax, pneumothorax and pleural effusions. The "triangle of safety (TOS)" is an often quoted site as being appropriate for ICD insertion; including in trauma patients by ATLS. The space is delineated by lateral border of the pectoralis major, lateral border of the latismus dorsi, line of the 5th intercostal space and the base of the axillaintercostal space. Although ultrasound (US) guidance has been recommended to aid insertion, its use

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

is far from established practice and landmark techniques and the TOS is still widely practiced. Objectives To establish the safety profile of the TOS through the use of US to delineate underlying anatomy including the position of intercostal vessels. Methods 50 consecutive patients on a general ICU underwent bilateral US examination of their TOS. The position of liver, spleen and heart in the respective TOS was noted. Doppler ultrasound was used to identify intercostal vessels within the TOS. Results Overall, the heart, liver and spleen were visible in 60 % of patients within the TOS. This percentage increased in the intubated and ventilated patient population. Intercostal vessels where visible in the minority of patients. Conclusions The British Thoracic Society and the National Patient Safety Agency UK has recommended ultrasound before inserting a drain for fluid. Our study found that the TOS is a misnomer exposing the patient to the risk of underlying organ perforation especially in the patient who is intubated and ventilated. It is therefore not safe and the practice should be abandoned. The routine use of real-time US to guide ICD insertion should be recommended to improve procedural safety. References 1. Havelock T, Teoh R, Laws , et al. Pleural procedures and thoracic ultrasound: BTS pleural disease guideline 2010. Thorax 2010;65(Suppl 2):61-76. 2. National Patient Safety Agency. Rapid response report, chest drains: Risks associated with insertion of chest drains (2008) http://www.nrls.npsa. nhs.uk/resources/patient-safety-topics/medical-device-equipment/? entryid45=59887&p=2

SEPSIS THERAPEUTICS A377 The immunomodulation effect of Polymyxin-B Hemoperfusion in severe sepsis/septic shock: a randomized controlled trial N. Srisawat, S. Peerapornratana, P. Laoveeravat, S. Tachaboon, S. Eiamong Chulalongkorn university, Medicine, Bangkok, Thailand Correspondence: N. Srisawat – Chulalongkorn university, Medicine, Bangkok, Thailand Intensive Care Medicine Experimental 2016, 4(Suppl 1):A377 Introduction Severe sepsis and septic shock are still the common cause of death in ICU. The extracorporeal blood purification therapy could provide a benefit in reduction the inflammatory mediators and bacterial endotoxins. Objectives To study the immunomoudation effect of Polymyxin-B Hemoperfusion (PMX) in severe sepsis/septic shock Methods The adult patients with severe sepsis/septic shock in intensive care units (ICUs) of King Chulalongkorn Memorial Hospital were tested with endotoxin assay. The patients with blood endotoxin level ≥0.6 were randomized into two groups, PMX hemoperfusion group and standard treatment group. The 2-hour PMX hemoperfusion were done in two consecutive days in PMX hemoperfusion group. Data for hemodynamic parameters, Sequential Organ Failure Assessment (SOFA) score, and blood sample were collected on first three day after enrollment The primary outcomes were leukocyte expression of CD11b, HLA-DR, chemotaxis function and blood endotoxin level on day 3 after enrollment. The secondary outcomes were 28-day survival, APACHE II score, change in SOFA score, CVS SOFA score. Results We have enrolled 36 participants into the study, 18 patients for each group. The age, APACHE II score, endotoxin level were comparable between two groups. The percent change of EAA was -19.8 % vs -10.6 % in PMX group and control group, respectively. The APACHE II and total SOFA score was -14.8 % vs +1.5 % and -12.8 % vs +6.1 %, in PMX group and control group, respectively. The CVS SOFA score was -31.4 % vs +24.9 %, in PMX group and control group, respectively. There was the trend increase in HLA-DR expression in PMX group and no change in control group.HLA-DR expression on day 3 in PMX group was significantly higher than control group,

Page 196 of 607

34.5 % vs 24.2 %, p = 0.03. 28-day survival rate was comparable between two groups. Conclusions The PMX hemoperfusion seemed to improve hemodynamic instability and improved immunoparalysis status in severe sepsis/septic shock patients. A378 Early exercise in patients with sepsis syndromes improves tissue oxygenation J. Paratz1,2, G. Kayambu1, R. Boots1 1 University of Queensland, Brisbane, Australia; 2Griffith University, Brisbane, Australia Correspondence: J. Paratz – University of Queensland, Brisbane, Australia Intensive Care Medicine Experimental 2016, 4(Suppl 1):A378 Introduction Sepsis can induce rapid proteolysis and early exercise is desirable to prevent loss of muscle mass. It is not known whether early intervention is safe in critically ill patients with sepsis. Exercise may also act to recruit the microcirculation. Objectives To investigate the effect of 30 minutes of electrical muscle stimulation and passive movements to major muscle groups on the microcirculation in patients in the first 48 hours of sepsis. Methods Changes in tissue oxygenation were assessed at baseline and immediately post exercise using a] a near infrared oxygenation device on the thenar eminence to assess percentage of oxygen in the muscle (StO2) and percentage of muscle oxygen extraction rate (MOER%) under both baseline and hyperaemic conditions. B] an orthogonal polarization spectral imaging device was placed sublingually to measure the microvascular flow index and capillary density of the microcirculation. Results All 20 patients completed the trial. There were significant effects post exercise on StO2 with hyperaemia [63.0 (±3.3)% to 66.2 (±5.3)%, p = 0.02], and an improvement in capillary density [2.45 (±1.01)% to 3.85 (±1.02)%, p = 0.04]. There were no adverse effects. Conclusions Early exercise appears to have a beneficial effect on critically ill patients as tissue oxygenation improved. There was evidence of recruitment of the microcirculation. References Kayambu G, Boots R and Paratz J A Prospective Randomised Controlled Trial investigating Functional and Physiological Outcomes following Early Rehabilitation in Sepsis: The i-PERFORM Trial - Protocol Article, BMC Anesthesiology in press 2011 Oct 31;11:21 Grant acknowledgement Intensive Care Foundation, Australia.

A379 Methylene blue effectiveness as contributory treatment in patients with septic shock M.F. Aguilar Arzapalo SSA UADY, Mérida, Mexico Intensive Care Medicine Experimental 2016, 4(Suppl 1):A379 Introduction Generalized vasodilation with nonresponding hypotension is present in half death cases due to septicaemia. Methylene blue could be used as a valuable complement in refractory hypotension treatment. with septic shock. Objectives The aim of this study was to determine the effectiveness of methylene blue as contributory treatment in patients. Methods A controlled, randomized, double blinded, clinical trial was performed. 60 patients were divided in two groups. A Group received a single dose of methylene blue calculated 2 mg/kg per body weight diluted in 100 cc of 5 % dextrose infused in 60 min. and C Group, (control) received 100 cc of 5 % dextrose infused in 60 min. Basal measurements of study variables were taken (MBP, lactate, base deficit, central venous saturation and CO2 delta) prior blue methylene administration and every hour afterwards, until MBP >65 mmHg without vasopressor or 72 hours passed after shock

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

began. Data about total noradrenaline dose in mg, length of stay, mechanical ventilation length and mortality was recorded. Results MBP increased progressively first 6 hours after methylene blue infusion in A Group 22 % and C Group 9.2 % (p:< 0.05), steadily until 72 hour follow up. Noradrenaline dose decreased in the first 6 hours, on A Group an 86 %, C Group was 56 % (p:< 0.05). Lactate clearance first 6 hours was 62 % in A Group, in contrast with C Group with 33 % clearance (p:< 0.05). Mortality at ICU discharge on A Group was 20.0 % and C Group was 36.6 % (p: < 0.05) without variation at 21 days. Conclusions Methylene blue is effective as contributory in septic shock treatment. Note: This abstract has been previously published and is available at [2]. It is included here as a complete record of the abstracts from the conference. References 1. Edmund S, Kwok M, Daniel H. Use of Methylene Blue in Sepsis: A Systematic Review. Journal of Intensive Care Medicine 2006; 21(6):359363. 2. Alderton W, Cooper C, Knowles G. Nitric Oxide synthases: structure, function and inhibition. Biochem. J. 2001; 357:593-615. 2. Lopez V, Aguilar Arzapalo M, Barradas L, Escalante A, Gongoro J, Cetina M (2016) Methylene blue effectiveness as contributory treatment in patients with septic shock. Critical Care 20(Suppl 2): 94. Grant acknowledgment Hospital "Agustin O´Horan".

Page 197 of 607

ml for IL-8; 350 pg/ml for IL-18; IL-10 was in the range of 50-400 pg/ml. The level of IL-4 was not increased (2.6 pg/ml) and had no significant difference. Levels of cytokines changed individually during hemosorption procedure and had no unambiguous correlation and did not show a clear dependence on the baseline level or the number of the procedures performed. The concentration median of interleukins from blood serum obtained after hemosorption was 380 pg/mlfor IL-6; 165 pg/ml for IL-8; 400 pg/ml for IL-18; 70 pg/ml for IL-10.During the analysis of the results, patients were divided into two groups: those with a moderately high value of IL-6 (250 pg/ml), and with ultra-high value (from 250 pg/ml) before hemosorption procedure. In the first group 70 % of patients after hemosorption procedure showed no significant changes in the concentration of IL6. 30 % of the second group showed a tendency to decrease the level of serum IL-6. In the studied ofAlteco LPSadsorbers eluates sorbed cytokines were found, but significantly high level was only for IL-8. There was an increase in serum IL-10 (33 %) and IL-18 (35 %) after extracorporal detoxification procedure. Conclusions These results indicate that the application of LPS-absorber Alteco in patients with sepsis is not significantly affect of free cytokines level in the systemic circulation. Reduction of IL- 8 and IL-6 observed only in patients with moderate increase of these cytokines concentration (not higher than 250 pg / ml). The presence of significant amounts of IL- 6 and IL-8 in Alteco LPS-adsorbers eluates with may indicate the presence in the blood of patients with sepsis related forms of cytokines capable of dissociation and adsorption on adsorber filter LPS-Alteco. References 1 Kulabukhov V. V. Acta Anaesthesiol. Scand. 2008.;52:P. 1024-5.

A381 Effect of oXiris haemodiafiltration in shock reversal for intraabdominal sepsis and septic shock: a case control series K.B. Tang, C.M. Chau, K.N. Lam North District Hospital, Hong Kong, Hong Kong, China Correspondence: K.B. Tang – North District Hospital, Hong Kong, Hong Kong, China Intensive Care Medicine Experimental 2016, 4(Suppl 1):A381 Fig. 140 (abstract A379). Mean pressure in the two groups

A380 Elimination of cytokines from blood of septic patients with lps adsorber R. Vlasenko1, E. Gromova1, S. Loginov2, M. Kiselevskiy1, Y. Dolgikova1 1 NN Blokhin Russisn Cancer Research Center, Moscow, Russian Federation; 2Moscow Botkin Clinic, Moscow, Russian Federation Correspondence: R. Vlasenko – NN Blokhin Russisn Cancer Research Center, Moscow, Russian Federation Intensive Care Medicine Experimental 2016, 4(Suppl 1):A380 Introduction In clinical practice LPS adsorbers are used on the method of extracorporal detoxification to remove the excess of LPS from the blood of the patients with sepsis. In recent years, there is evidence of the ability of the hemosorption columnsto eliminate cytokines and inflammatory mediators that trigger a cascade reaction of irreversible pathological changes in organs and tissues of the patient [ 1]. Objectives Study of elimination of proinflammatory cytokines from the blood of septic patients during the hemosorption procedureby using the LPS adsorber Alteco Methods The study included 20 patients with confirmed Gramnegative sepsis. Sorption procedure was implemented on Alteco columns (Sweden). Blood serum sampleswere taken before and immediately after the procedure, column lavages were taken as well. Samples were stored at - 70 ° C. The values of IL- 4 , IL- 6 , IL- 8 , IL10 , IL- 18 in serum and lavage was measured by ELISA. Results Baseline level of serum proinflammatory cytokines was significantly increased and the median was 380 pg/ml for IL-6; 122 pg/

Introduction Septic shock induced by gram-negative bacteria infection is a serious condition in intensive care unit (ICU), and endotoxin plays an important role in their pathogenesis. oXiris filter is a polyacrylonitrile haemofiltration membrane treated with polyethyleneimine, which provide a positive charged surface to enhance the absorption of endotoxins and cytokines. Objectives case series to determine oXiris haemodiafiltration could improve the haemodynamic status in patients with gram-negative bacteria related septic shock. Methods Patients admitted to regional ICU with intra-abdominal sepsis that required intervention, who developed severe septic shock required inotropic support and acute renal failure required renal support, was assigned to receive oXiris haemodiafiltration for 72 hours. Matched patients with similar pathology and severity, who received ordinary haemodiafiltration for acute renal failure, were identified for comparison. Primary outcome was the percentage of reduction of inotrope dose within 96 hours after the start of renal replacement therapy, which was calculated according to equivalent Noradrenaline dose (ug/ hr). Second outcomes including duration of MV, duration of CRRT, ICU length of stay, and ICU mortality. Results Three patients were assigned to receive oXiris haemodiafiltration, and five matched patients were identified for comparison. ICU mortality for patient receiving oXiris haemodiafiltration was 33 %, compare to 60 % for patient with ordinary haemodiafiltration. Patient received oXiris haemodiafiltration shown a rapid reversal of inotrope dose compare to the control, and the effect became obvious after 24 hours of treatment (24 hr: 46.21 +/- 8.23 % drop for oXiris group, 24.93 +/- 9.22 % drop for control group, 48 hr: 81.95 +/- 12.34 % drop for oXiris group, 54.85 +/15.24 % drop for control group, 72 hr: 93.98 +/- 2.05 % drop for oXiris group, 65.61 +/- 11.47 % drop for control group, 96 hr: 97.28 +/- 1.85 % drop for oXiris group, 42.29 +/- 36.27 % drop for control group). There was no different between groups in terms duration of MV (10.28 +/-

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

2.67 days in oXiris group vs 11.67 +/- 3.06 days in control group), duration of CRRT (4.55 +/- 1.7 days in oXiris group vs 5.72 +/- 2.55 days in control group), and ICU length of stay (15.17 +/- 6.96 days in oXiris group vs 14.49 +/- 2.33 days in control group). Conclusions oXiris haemodiafiltration shown a rapid reversal of shock and lower mortality in patient with intra-abdominal sepsis and septic shock, while there was no difference in terms of duration of MV, duration of CRRT, and ICU length of stay. References 1. T. Rimmelé, A. et al; High volume haemofiltration with a new haemofiltration membrane having enhanced adsorption properties in septic pigs. Nephrol Dial Transplant, 2008; pp. 1-7 Grant acknowledgment nil

A382 Clinical experience of Polymyxin B hemoperfusion in septic shock: 47 cases in a single center E. Gil1, G.Y. Suh1,2, C.-M. Park1,3, J. Park1, C.R. Chung1 1 Samsung Medical Center, Critical Care Medicine, Seoul, Republic of Korea; 2Samung Medical Center, Medicine, Seoul, Republic of Korea; 3 Samsung Medical Center, Surgery, Seoul, Republic of Korea Correspondence: E. Gil – Samsung Medical Center, Critical Care Medicine, Seoul, Republic of Korea Intensive Care Medicine Experimental 2016, 4(Suppl 1):A382 Introduction Direct hemoperfusion with Polymyxin-B-immobilized fiber column (PMX-DHP) has been successfully used to treat patients with septic shock Objectives The aim of this study was to report clinical experience of PMX-DHP in septic shock patients in a single center, and identifying subgroup of patients who may benefit from this treatment. Methods From July 2014 to March 2016, we performed 56 cases of PMX-DHP in septic shock and severe acute respiratory distress syndrome (ARDS) patients. Among these, septic shock patients who required vasopressor and lactic acid level was more than 2 mg/dl after fluid resuscitation were analyzed. Results We performed 47 cases of PMX-DPH treatment in 31 septic shock patients. Two session of PMX-DPH were performed in 16 patients (51.6 %). Median age was 64 years old (IQR 51-69) and 54.8 % (17/31) was male. When PMX-DHP treatment started, median SOFA score was 17 (IQR 14-20), median lactic acid level was 9.72 mg/dl (IQR 6.47-20.42), and median inotropic score (IS) was 105 (IQR 65300). Major infection focus was intra-abdominal (22/31, 71.0 %) and 5 patients (16.1 %) treated due to respiratory tract infection. The 28day mortality was 48.4 % (15/31), and 90-day mortality was 54.8 % (17/31). SOFA score at PMX-DHP treatment day, infection site, source control, and lactic acid level were associated with 28-day mortality. Cox-regression analysis shows a significant association of SOFA score (OR, 1.33; 95 % CI, 1.095-1.616, p = 0.004) and infection site (OR 0.264; 95 % CI, 0.083-0.845, p = 0.025) in overall survival. Conclusions PMX-DHP application for septic shock patients could be considered as an adjunctive therapy. A383 Heparin dosing score protocol for anticoagulation during hemoperfusion with a polymyxin B-immobilized cartridge C.T. Lee1, A. Chao1, P.-Y. Shih1, Y.-F. Chang2, C.-H. Lai3, Y.-C. Hsu1, Y.-C. Yeh1, Y.-J. Cheng1 1 National Taiwan University Hospital, Department of Anesthesiology, Taipei, Taiwan, Province of China; 2National Taiwan University Hospital, Department of Nursing, Taipei, Taiwan, Province of China; 3National Taiwan University Hospital, Department of Surgery, Taipei, Taiwan, Province of China Correspondence: C.T. Lee – National Taiwan University Hospital, Department of Anesthesiology, Taipei, Taiwan, Province of China Intensive Care Medicine Experimental 2016, 4(Suppl 1):A383

Page 198 of 607

Introduction In recent decades, more and more studies have reported that treatment with polymyxin B-immobilized hemoperfusion cartridge may have beneficial effects on hemodynamics and mortality in patients with severe sepsis or septic shock.(1, 2) Premature cartridge clotting is a common problem during polymyxin B hemoperfusion, and it may decrease therapeutic efficacy and increase cost of therapy. However, excessive anticoagulation in patients with severe disseminated intravascular coagulation or in intra-abdominal infection patients after operation may increase the risk of bleeding. Currently, nafamostat mesilate is used dominantly in Japan,(2) but the experience of using heparin for anticoagulation for polymyxin B hemoperfusion therapy in other countries is less and need further investigation. Objectives We aimed to investigate the effectiveness and safety of a heparin dosing score protocol for anticoagulation during polymyxin B-immobilized cartridge hemoperfusion. Methods This was a retrospective study in 6 ICUs in National Taiwan University Hospital. The medical records of polymyxin B hemoperfusion in 23 septic shock patients from October, 2013 to February, 2016 were reviewed. The heparin dosing score was modified from the instruction of heparin dosing from Toray Industries, Inc. The primary aim was to investigate the completion rate of 2-hour session of hemoperfusion without premature cartridge clotting or any event of significant bleeding. The secondary aim was to investigate the effect of polymyxin B hemoperfusion on hemodynamics and disease severity scores. Results Among the 23 enrolled patients, 34 sessions of polymyxin Bimmobilized hemoperfusion were administered. The completion rate of 2-hour session of hemoperfusion was 97 %. In the only one case of premature cartridge clotting, the hemoperfusion time was 94 minutes. There was no any documented significant bleeding. For surgical patients, the mean time to initiate hemoperfusion was 16.3 ± 15.2 hours after operation. The 28-day mortality of these enrolled patients was 24 %. Significant hemodynamic improvement and reduction of disease severity scores were noted after polymyxin B hemoperfusion. Conclusions Our results support that our heparin dosing score protocol provides effective and safe anticoagulation during polymyxin B hemoperfusion in septic shock patients. References 1. Mitaka C, Tomita M. Polymyxin B-immobilized fiber column hemoperfusion therapy for septic shock. Shock. 2011;36:332-338 2. Cruz DN, Antonelli M, Fumagalli R, Foltran F, Brienza N, Donati A, Malcangi V, Petrini F, Volta G, Pallavicini FMB, et al.: Early use of polymyxin B hemoperfusion in abdominal septic shock: the EUPHAS randomized controlled trial. JAMA. 2009;301(23):2445Y2452 Grant acknowledgment Supported, in part, by research grant NTUH.105-A125 from the National Taiwan University Hospital.

Table 116 (abstract A383). Heparin Dosing Score Score

1

2

6

(A) Platelet count (K/μL) (1000/mL)

>80

50-80

20-49

80

50-80

20-49

10/group) allowed the following indexes to be analyzed: survival of animals, invasive blood pressure (BP), lactate rate, heart function and cytokine profile using plasma, organs (heart, kidney, lung, liver and brain) and RT-PCR 18 hours after surgery. Oxidative stress was assessed using DHE. Results and discussion Survival, 18 h after the onset of surgery, was 0 %, 83 % and 42 % for sham, CLP and CLP-EPO rats (p < 0.0001), respectively. The whole inflammatory response (pro- and anti-inflammatory cytokines) was prevented by EPO in organs and plasma (see figure). Hence, the increases in cytokines as TNFα, Il-6 and Il-10 in the CLP-EPO heart were -50, -75 and -25 % lower than those observed in CLP ones (p = 0.0103 ; 0.0037 and 0.0031 respectively). At the protein level, p62, an inflammatory marker, was increased in the CLP hearts when compare to sham, whereas no difference was seen between sham and CLP-EPO rats (p = 0.0162). There was less oxidative stress in the CLP-EPO hearts when compared to sham and CLP groups (p < 0.05) 18 hours after the onset of sepsis. Regarding function indices, EPO did not prevent systolic dysfunction, but completely prevented tachycardia (p = 0.0248) and the decrease of BP (p = 0.0455). Conclusion During sepsis in rats, EPO prevents inflammatory responses and oxidative stress in organs and plasma. References 1 Rabuel and Mebazaa, Intensive Care Medicine 2006, 32 : 799-807 2 Walden et al., Critical Care 2010, 14 : 227 3 Rittirsch et al., Nature Protocols 2009 ; 4, n°1 : 31-36

Fig. 143 (abstract A389). EPO prevents inflammatory response

A390 Improved cytokine clearance with a high cut-off filter for the supportive treatment of sepsis J. Hartmann, S. Harm, V. Weber Danube University Krems, Center for Biomedical Technology, Krems, Austria Correspondence: J. Hartmann – Danube University Krems, Center for Biomedical Technology, Krems, Austria Intensive Care Medicine Experimental 2016, 4(Suppl 1):A390 Introduction Extracorporeal therapies are considered as an option to modulate cytokine levels in the circulation in systemic inflammatory syndromes. The technologies based on high cut-off (HCO) membranes with larger pore-size and effective clearance of middle molecules are most promising [1, 2]. Objectives Aim of this study was to conduct in vitro experiments in order to compare the Ultraflux® AV 1000S high-flux haemofilter to the Ultraflux® EMiC®2 HCO filter (both from Fresenius Medical Care, Bad Homburg, Germany) regarding cytokine elimination. Methods In these experiments, 2 x 1000 ml human plasma was spiked with IL-6, IL-8, IL-10 and TNF-α and dialyzed with the AV 1000S filter or the EMiC®2 filter. The experiments were conducted using two Fresenius 4008H haemodialysis machines with a plasma- and dialysate flow of 300 ml/min. Samples for clearance measurements were taken at 15, 30, 60, 120, 180 and 240 min. Sieving coefficients were determined at a filtrate flow of 10 % of the plasma flow. All experiments were carried out in triplicates and cytokines were quantified by ELISA. Results The results show that IL-8 is efficiently removed by both filters. Larger cytokines, such as IL-6 and TNF-α were removed only by the EMiC®2 filter with a clearance rate of 4.7 and 2.4 ml/min, respectively. For IL-10, the EMiC®2 shows much higher clearance rates than AV 1000S.

Intensive Care Medicine Experimental 2016, 4(Suppl 1):27

Page 203 of 607

Conclusions Compared to high-flux dialyzers, HCO filters offer significantly higher removal rates for middle molecules such as cytokines and other target molecules with similar molar mass. Therefore, HCO filters should be preferred in extracorporeal therapies for the supportive treatment of systemic inflammatory syndromes.

High Cut-Off Continuous Hemodialysis; ASA Annual Meeting, 17.-21. October 2009, New Orleans. [2] Mathieu Page, Charles-Eric Ber, Davy Hayi-Slayman, Bernard Allaouchiche, Thomas Rimmelé, Clinical Tolerance of Continuous Hemodialysis with a High Cut-Off Membrane; ASA Annual Meeting, 17.-21. October 2009, New Orleans.

References [1] Mathieu Page, Charles-Eric Ber, Davy Hayi-Slayman, Bernard Allaouchiche, Thomas Rimmelé, Removal of Middle-Molecular Weight Molecules with

Grant acknowledgment This work was supported by the government of Lower Austria within the project ID WST3-T-91/036-2014).

Submit your manuscript to a journal and benefit from: 7 Convenient online submission 7 Rigorous peer review 7 Immediate publication on acceptance 7 Open access: articles freely available online 7 High visibility within the field 7 Retaining the copyright to your article

Submit your next manuscript at 7 springeropen.com