Proceedings of Reanimation 2023, the French Intensive Care Society International Congress (2024)

Correspondence: Manon Passard (manon.passard@chu-lyon.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-01

Rationale: Central lines (CL) are frequent in pediatrics but implicate some severe complications. CL infection is the main cause of nosocomial infection in pediatrics. Prevalence of CL infection varies a lot depending on the populations studied and there are no recent data regarding the prevalence of severe sepsis due to CL.

Objective: To estimate the prevalence and identify factors associated with occurrence of severe sepsis due to CL associated bloodstream infection (CLABSI) in pediatrics.

Patients and methods/materials and methods: We conducted a 4-year multicentric retrospective cohort study including all children with confirmed CLABSI from birth to 18years old, hospitalized in all pediatric units in two university hospitals. Neonatalogy units were excluded. Patients were screened from bacteriological datas.Epidemiology and bacteriology of CLABSI were described. To find potential factors associated with severe sepsis due to CLABSI we performed univariate and multivariate analysis.

Results: Over 4476 suspected CL infections, we included 372 CLABSI. The prevalence of severe sepsis due to CLABSI was 17.2% (15.25–19.16%) among confirmed CLABSI and 1.43% (0.02–2.84%) among suspected CL infections. 12 children died because of severe sepsis due to CLABSI. In univariate analysis, factors associated with the occurrence of severe sepsis due to CLABSI were antibiotic therapy within 3weeks before, known bacterial colonization, CL daily use and hospitalization in intensive care unit when the CLABSI occurred. Tachycardia and hypotension at the initial presentation, augmentation of lactate and procalcitonin at the initial biological sample were associated with the occurrence of severe sepsis due to CLABSI. In multivariate analysis, daily use of CL, know bacterial colonization, tachycardia and hypotension were associated with the occurrence of severe sepsis due to CLABSI. We couldn’t evaluate the dosage of lactate and procalcitonin due to a lack of power. There was an augmentation of gram-negative bacillus proportion among the severe sepsis due to CLABSI.

Conclusion: Severe sepsis prevalence in children with documented CLABSI is close to 20%. We identified factors associated with the occurrence of severe sepsis complicating CL infection. Those factors should be tested in a prospective multicenter study, to improve our ability to identify patients at risk of severe sepsis and to prevent the evolution to septic shock. This retrospective study is the first step to a prospective multicenter study aiming to optimize the initial treatment and identify children at risk to develop severe sepsis due to CLABSI, in order to improve sepsis bundles application.

Reference 1

Niedner MF, Huskins WC, Colantuoni E, Muschelli J, Harris JM, Rice TB, et al. Epidemiology of Central Line-Associated Bloodstream Infections in the Pediatric Intensive Care Unit. Infect Control Hosp Epidemiol. déc 2011;32(12):1200?8.

Reference 2

Ziegler MJ, Pellegrini DC, Safdar N. Attributable mortality of central line associated bloodstream infection: systematic review and meta-analysis. Infection. févr 2015;43(1):29?36.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Julian San Geroteo (julian.sangeroteo@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):CO-02

Rationale: Pediatric sepsis is the leading cause of mortality in children under five years of age. The latest international guidelines of the Surviving Sepsis Campaign 2020 (SSC-2020) have revised the management of paediatric sepsis, particularly concerning fluid resuscitation. Although these guidelines are important, no study has evaluated their application in pediatric emergency departments. The aim of the study was toassess physician adherence to the SSC-2020 fluid resuscitation guidelines in infants and children with suspected sepsis in Pediatric Emergency Departments (PED).

Patients and methods/materials and methods: Fluid REsuscitation for Suspected Sepsis in Pediatric Emergency Departments (FRESSPED) was a prospective multicenter observational study conducted in 21 French hospitals over 5 sequential weeks, between November 2021 and March 2022. Children with suspected sepsis and who received anti-infectious treatment within 72h were included. Severe sepsis was confirmed using the 2005 consensus criteria in patients admitted to PICU according to a review of each medical file by pediatric intensivist. Primary outcome was SSC-2020 fluid resuscitation guidelines adherence (low 0–24%, moderate 25–74% or high 75–100%) according to: volume of 10–20ml/kg for each bolus; exclusive administration of balanced crystalloids at 1h and 24h of management; and initiation of fluid resuscitation within one hour of sepsis recognition. An electronic survey was sent to all the PEDs’physicians working to report their adherence to SSC-2020.

Results: A total of 63 patients were included, 10 of whom had confirmed sepsis. Median age [IQR] was 7.9months [2.8–35.9] and a comorbidity was present in 19 children (30.1%). Primary site of infection was respiratory in half of the cases (50.8%) and unknown in 8 (12.7%) subjects.Microbiological documentation was found in 47 (74.6%) cases. Compared to the SSC-2020 guidelines: 43 (68.3%) patients received boluses of 10 to 20ml/kg; exclusive administration of balanced crystalloids was found in 35 (55.6%) and 34 (54.8%) children at one hour and twenty-four hours of management respectively; andfluid resuscitation was initiated within one hour of sepsis recognition in 42 (76.4%) cases. Main barriers reported by physicians to such adherence were: difficult intravenous access (43.2%); lack of team training (28.8%); workload constraints (28.0%); absence or out-of-date protocols (24.0%).

Conclusion: FRESSPED study found high adherence for fluid resuscitation initiation and moderate adherence for bolus volume and fluid choice. Implementation of up-to-date standardized management protocols and teaching programs could improve adherence and outcomes of suspected sepsis children.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Michael Levy (michael.levy@aphp.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-03

Rationale: Sickle cell disease (SCD) is one of the most frequent inherited diseases in France and in the world. Over the last decades, in high-income countries, an important decrease in mortality have been observed due to the improved of care. However, children with SCD can become critically ill and require admission in Pediatric Intensive Care Units (PICU). Only a few studies with small samples have focused on children with SCD admitted to PICU and robust data regarding the epidemiology of these patients are missing. The aim of this study was to describe the epidemiology of children admitted to PICU for acute complications of SCD and their outcomes.

Patients and methods/materials and methods: We conducted a single center retrospective study in a tertiary center PICU. Consecutive children with SCD admitted to PICU between January 1st, 2009 and December 31, 2019 were included. Patients with sickle cell trait (heterozygous for the causative gene) or that have been admitted following bone-marrow transplant were not included.

Results: 582 children with SCD were admitted to PICU during the study period. The median age was 9.2 [5.5–13.4] years and 60% were boys. 538 (93%) were of SS genotype and the baseline hemoglobin was 8.4g/dl. 42% had a history of PICU admission and 51% of acute chest syndrome (ACS). Patients were mainly admitted from pediatric ward (n = 440, 76%) and emergency units (n = 114, 19.5%). ACS was the main reason of admission (54%) followed by vaso-occlusive crisis (17%) and stroke (17%). 492 (85%) were treated with antibiotics, 293 (50%) patients required non-invasive ventilation and 47 (8%) invasive ventilation. Cardiovascular failure was rare (2%). 223 (38%) patients required blood transfusion and 155 (27%) exchange transfusion. The median duration of PICU stay was 3 [1–4] days. The mortality rate at PICU discharge was 1% (n = 7) and 1% (n = 8) after 6months.

Conclusion: Children with SCD hospitalized in PICU were mainly admitted for acute chest syndrome and half of patients required non-invasive ventilation. The median length of stay was short and the mortality low.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Omar Hanafia (omarhanafia@yahoo.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-04

Rationale: Extracorporeal Membrane Oxygenation (ECMO) is a rescue technique that allows the replacement of circulatory and/or respiratory functions. The pharmacokinetic modifications generated by this circulatory assistance require adosage adaptation of certain drugs. The objective of the study was to compare the drug prescription of patients under ECMO and to propose appropriate dosages adjustments.

Patients and methods/materials and methods: Our 6-months prospective observational monocentric study focuses on critically ill patients assisted by ECMO. Clinico-biological data were collected from the computerized patient record and by the daily presence of the pharmacist in the department. We noted the type and indication of ECMO, complications and adequacy of dosages compared to the literature for relevance.

Results: Fourteen patients with ECMO were included: mean age 18months [0 to 168months], sex ratio = 1. Renal function was impaired in 8 patients (57%). The average duration of ECMO was 15days [3–24days]. 6 patients were weaned, 4 of whom were still hospitalized on the ward (43%) and 8 patients died (57%). 13 patients (93%) were on veno-arterial ECMO, following acute respiratory distress syndrome (8 cases or 61%), refractory cardiac arrest (3 cases 23%), cardiogenic shock (8%) or septic shock (8%). 1 patient (7%) was on veno-venous ECMO following acute respiratory distress syndrome (ARDS). Elevent patients (79%) developed complications related to ECMO (9 hemorrhages, 8 hemolysis, 6 oxygenation difficulties, 5 PAO, 4 stroke). Concerning the drug management of these patients, we counted 16 overdosages and 2 underdosages not justified according to the literature. There was also no dosage that would explain these discrepancies, i.e. 18 nonconformities out of 73 lines analyzed concerning Vancomycin, Gentamicin, Fluconazole, Caspofungin, Voriconazole, Ganciclovir, Heparin, Morphine, Sufentanil, Midazolam, Cisatracurium, Hydrocortisone Hemisuccinate, Methadone.

Discussion: The populations studied in the literature remain different from ours, making it difficult to discuss our clinical results.

Conclusion: However, following the non-conformities of dosage noted, the pharmacist proposed a table of dosage adaptation under ECMO synthesizing the literature for the studied molecules which is systematically accompanied by instructions to make a therapeutic drug monitoring.

Compliance with ethics regulations: N/A.

Correspondence: Morgan Recher (morgan.recher@chu-lille.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-05

Rationale: Trauma is an important cause of paediatric out-of-hospital cardiac arrest with a high mortality rate. The first aim of this study was to compare the survival rate at day 30 and at hospital discharge following paediatric traumatic and medical out-of-hospital cardiac arrest. The second aim was to compare the rates of return of spontaneous circulation and survival rates at hospital admission (day 0).

Patients and methods/materials and methods: This multicentre comparative post-hoc study was conducted between July 2011 and February 2022 based on the French National Cardiac Arrest Registry data. All patients aged < 18years with out-of-hospital cardiac arrest were included in the study. Patients with traumatic aetiology were matched with those with medical aetiology using propensity score matching.

Results: There were 398 traumatic and 1,061 medical out-of-hospital cardiac arrests. Matching yielded 227 pairs. In non-adjusted comparisons, days 0 and 30 survival rates were lower in the traumatic aetiology group than in the medical aetiology group (19.1% vs. 24.0%, odds ratio (OR) 0.75, 95% confidence interval (CI) 0.56–0.99, and 2.0% vs. 4.5%, OR 0.43, 95% CI 0.20–0.92, respectively). In adjusted comparisons, day 30 survival rate was lower in the traumatic aetiology group than in the medical aetiology group (2.2% vs. 6.2%, OR 0.36, 95% CI 0.13–0.99).

Conclusion: In paediatric out-of-hospital cardiac arrest patients, the adjusted survival rates at day 30 for those with a traumatic aetiology were lower than those with a medical aetiology. Prioritising trauma-specific interventions that address possible reversible causes of cardiac arrest may lead to improved survival in this population.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Kaoutar El Fakhr (kaoutarelfakhr2013@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):CO-06

Rationale: Refeeding syndrome (RFS) describes a potentially fatal physiological response to the increase in energy from any source including: intravenous, oral food, enteral nutrition or parenteral nutrition following a period of starvation or severely reduced energy intake, reflecting a change from catabolism to anabolism following the release of insulin. However, data on its incidence is lacking, and the heterogeneity of diagnostic criteria and frequent electrolyte disorders in this population make its diagnosis complex.In 2020, the American Society for Parenteral and Enteral Nutrition (ASPEN) developed consensus recommendations for identifying patients at risk and with refeeding syndrome. The aim of this study was to determine the incidence of refeeding syndrome in critically ill children on nutritional support.

Patients and methods/materials and methods: A prospective cohort spread over three years from January 2020-January 2023, conducted in a tertiary pediatric intensive care unit was undertaken, and additional data were retrospectively collected.

Results: A total of 2058 children were included in the study, with 350 children (17%) classified as undernourished, who were at risk of refeeding syndrome (142F, 158M). Over a period of 3years, 50 children of these were identified as having refeeding syndrome, giving a cumulative incidence of 14.28%; IC95% [0.106–0.179], aged 12days -17years (median 24SD + -62.7months). Patients with refeeding syndrome have a higher risk of mortality (odds ratio [OR], 6.75; CI, 3.18–14.3; P = 0.001). Children who were fed late > 5days are more likely to have refeeding syndrome (odds ratio [OR], 5.33; CI, 3.13–9.8; P = 0.001), children who received early feeding < 5days are less likely to have refeeding syndrome (odds ratio [OR], 0.18; CI, 0.11–0.32; P = 0.001).

Conclusion: Refeeding syndrome is a serious, life-threatening metabolic complication that can lead to increased morbidity and mortality.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Michael Thy (michael245thy@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):CO-07

Rationale: Sepsis is complicated in about 50% of cases by an acute encephalopathy ranging from delirium to coma (sepsis-associated encephalopathy, SAE), which is associated with poor outcome. Systemic causes of secondary brain injury (SSBI), that are frequently observed in ICU patients, may worsen outcomes of SAE. We aimed to investigate the association between the presence of at least one SSBI within the first two days of intensive care unit (ICU) admission with outcomes of SAE patients.

Patients and methods/materials and methods: We performed a retrospective analysis using data from the French OUTCOMEREA multicenter prospective database. We included consecutive SAE patients (defined by a score on the Glasgow coma scale ≤ 13 and severe sepsis or septic shock (SEPSIS 2.0 definition) criteria) requiring invasive mechanical ventilation at ICU admission. We excluded patients with infectious and non-infectious acute cerebral pathologies. For each patient, we analyzed the SSBI events (abnormal glycemia, hypotension, temperature abnormalities, anemia, dysnatremia, oxygenation abnormalities, dyscapnia) present within the first 2days of hospitalization in ICU and the impact of their control at day 3 on day-14 mortality using a Cox-model with adjusted hazard ratio (aHR) and 95% confidence interval [95% CI].

Results: From 1997 to 2020, 995 patients were included in the study. Compared to decedents at D14, alive patients had significantly less SSBI events, including less hypoglycemia (< 3mmol/L), less hypothermia (< 36°C), less hypoxemia (PaO2 < 60mmHg), less hypotension (DAP < 50mmHg) and anemia (hematocrit < 21%) within the 2 first days hospitalization in ICU. After adjusting for confounders (SOFA score, type of admission, comorbid conditions), the control of the following SSBI at day 3 was significantly associated with lower day-14 mortality for blood pressure (diastolic pressure > 50mmHg: aHR = 1.997 [1.42–2.8], p < 0.001), oxygenation (PaO2 between 60 and 200 mHg: aHR = 2.076 [1.36–3.16], p < 0.001), glycemia (glycemia between 3 and 11mM/l: aHR = 1.604 [1.21–2.12], p < 0.001), temperature (36°C-38.3°C: aHR = 1.55 [1.14–2.11], p = 0.006).

Conclusion: We demonstrate that the control of SSBI (particularly oxygenation, arterial pressure, glycemia, temperature) improves prognosis of patients with SAE. SSBI monitoring and control should be systematically done in all septic patients with SAE.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Chloé Tridon (chloe.tridon12@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):CO-08

Rationale: Neurologic outcomes of patients under venoarterial ECMO (VA-ECMO) may be worsened by secondary insults of systemic origin. We aimed to assess whether sepsis, commonly observed during ECMO support, is associated with brain injury and outcomes in these patients.

Patients and methods/materials and methods: We conducted a single-center cohort "exposed-non-exposed" type study (2013–2020) in a university hospital in consecutive adult patients treated by VA-ECMO. Patients with "sepsis" at the time of VA-ECMO cannulation were compared with "non-sepsis" patients. Main outcome measures included clinical status, NSE and S100β blood concentrations, and EEG parameters measured within 72h of VA-ECMO cannulation. We investigated the association between sepsis and poor functional outcome at 90days (modified Rankin scale (mRS) score ≥ 4) in adjusted (logistic regression, inverse probability of treatment weighting (IPTW)) analyses. Subgroup analyses were conducted in patients with or without pre-ECMO cardiac arrest.

Results: A total of 196 patients was included (“sepsis”, n = 128; “non-sepsis”, n = 68), of whom 87 (44.4%) presented cardiac arrest before VA-ECMO cannulation. Compared to the “non-sepsis” group, “sepsis” patients presented a significant increase in the S100β protein concentrations at day 1 (0.94μg/L versus 0.52μg/L, p = 0.03), and more frequent EEG alterations (i.e., severe slowing, discontinuous background, and a lower prevalence of sleep patterns). A poor functional outcome was observed in 99/128 (77.3%) patients of the "sepsis" group and 46/68 (67.6%) of the "non-sepsis" group (adjusted logistic regression OR 1.21, 95%CI 0.58–2.47; IPTW OR 1.24, 95%CI 0.79–1.95). Sensitivity analyses suggested that sepsis was independently associated with poorer functional outcomes in the “pre-ECMO cardiac arrest” subgroup (adjusted logistic regression OR 3.44, 95%CI 1.06–11.40; IPTW OR 3.52, 95%CI 1.68 -7.73), whereas no such association was observed in patients without pre-ECMO cardiac arrest (adjusted logistic regression OR 0.69, 95%CI 0.27 -1.69; IPTW OR 0.76, 95%CI 0.42–1.35).

Conclusion: Sepsis at the time of VA-ECMO initiation is associated with early biochemical and electrophysiological alterations. We observed a detrimental role of sepsis on neurologic outcomes in the subgroup of patients who had experienced pre-ECMO cardiac arrest, but not in other patients.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Ahlem Trifi (trifiahlem2@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):CO-09

Rationale: In sepsis, the imbalance of the immune-inflammatory response is a source of organ dysfunction. Immunomodulatory therapies have been tried in order to moderate this imbalance and improve the outcome of critically septic patients. We sought to determine whether the adjunction of erythromycin (a macrolide antibiotic known to have immunomodulatory effects) in septic patients has a regulatory effect on the infection biomarkers and decreases mortality.

Patients and methods/materials and methods: We performed a single-blind randomized trial comparing erythromycin 1 gx3/d intravenously for 5days versus placebo for 5days in adult patients admitted to intensive care with. The exclusion criteria was death within 5days of the protocol. Inflammatory markers (whole blood count, CRP and procalcitonin) were assayed on inclusion (D0) and on D6 (end of prescription). The primary outcome was 28-day mortality and the secondary outcomes were changes in infection biomarkers between D0-D6, vasopressor requirements, ventilation use and duration, and length of stay. [ClinicalTrials.gov ID: NCT04665089].

Results: One hundrer and nine patients were enrolled, of whom 80 were included (erythromycin group, n = 40 and placebo group, n = 40). The male sex was predominant in the first group (31 (77.5%) vs 20 (50%), p = 0.01). The other baseline characteristics were similar (age = 63 [51–78] vs 60 [50–84], p = 0.78; SOFA = 9 vs 7.5, p = 0.19; diabetes: 30% vs 50%, p = 0.11 and cardiovascular diseases: 32.5% vs 60%, p = 0.37) in erythromycin versus placebo groups respectively. There was no difference on the time of occurrence of sepsis (6.5 [5–10] d vs 5.5 [2–16] d, p = 0.16) nor on the nosocomial nature (47,5% vs 58%, p = 0.37). No favorable effect of adding erythromycin in septic patients was found compared to placebo on mortality (45% vs 37.5%, p = 0.63) nor on the other parameters. Concerning the biomarkers, they were similar at baseline. At the end of the protocol, no difference was showed. In the intragroup analysis, both procalcitonin and CRP decreased significantly from day 0 to day 6 in the erythromycin arm (from 2.47 to 1.14 ug/l, p < 10–3 and from 172 to 109mg/l, p = 0.001 respectively). In the placebo arm, only procalcitonin decreased (attached table). No conductive disorders or QT segment abnormalities were described in the erythromycin group.

Conclusion: the adjunction of erythromycin did not reduce mortality but it conducted to a better control of the inflammatory response, in particular the reduction of procalcitonin and CRP.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Baptiste Michaux (baptiste.michaux@chu-rouen.fr)

Rationale: Acquired weakness in intensive care unit (ICU), particularly favored by multi-organ failure caused by septic shock and prolonged bed rest, is responsible for an increase in duration of weaning from mechanical ventilation, duration of hospitalization and morbi-mortality. In addition to the usual physiotherapy, cycloergometry sessions are proposed. A recent meta-analysis however did not show superiority of these sessions (1). The main objective of this study was to determine impact on ICU length of stay in patients with septic shock of early mobilization by cycloergometer in association with standard physiotherapy.

Patients and methods/materials and methods: This is a prospective, randomized, open multicenter study, which included patients with a septic shock, invasive ventilation and sedation, hemodynamically stabilized. At inclusion, patients were randomized in intervention group (cycloergometer (20min/day) added to standard physiotherapy) or in control group (standard physiotherapy alone). In both groups, physiotherapy started as soon as the patients were included and was adapted to the patient's condition. Patients were randomized a second time at the first lifting of sedation. This double randomization constituted a dataset for the endpoints analysis from hemodynamic stability to the first lifting of sedations (phase I) and another dataset for unbiased endpoints analysis from lifting of sedations to the discharge from ICU (phase II).

Results: We included 119 patients in eight ICU, with median SAPS II and SOFA (ICU admission) of respectively 60 [95%RR 23.9–93.0] and 9 [95%RR 2.00–15.00]. Characteristics at inclusion were similar. We followed 109 patients until the cessation of sedation (Phase I). Then, we re-randomized 107 patients until ICU discharge (Phase II). There was no difference regarding length of stay in ICU at each phase between the two groups (Figure 1A, B). The phase I median duration is 4days for both groups. For phase II, the median is 9days [6–15] days for cycloergometer group and 12 [6–28] days for control group. Patients with standard physiotherapy during the two phases spent 20days in ICU [11–66] versus 13days [8–33] for patients with cycloergonometry in phase I and/or phase II (p = 0.052) (Figure 1C). Similarly, the median duration of invasive mechanical ventilation was significantly higher for patients without cycloergometry (14 vs 9days, p = 0,04). Another abstract proposes the results concerning the safety and feasibility.

Conclusion: Although we did not show significant difference between groups, a trend emerges concerning the increase in the length of stay for patients who do not benefit from a cycloergometry in either of the two phases.

Reference 1

Takaoka A, Utgikar R, Rochwerg B, Cook DJ, Kho ME. The Efficacy and Safety of In-Intensive Care Unit Leg-Cycle Ergometry in Critically Ill Adults. A Systematic Review and Meta-analysis. Ann Am Thorac Soc. oct 2020;17(10):1289?307.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Juliette Quintin (quintin.juliette@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):CO-011

Rationale: Living with dogs or cats is common. Adverse events in relationship with pets include bites, infections and areincreasing. Bacteria most frequently involved arePasteurella, Bartonella, Tularemia, orCapnocytophaga. The goal of the PETSEPSIS study was toreportcharacteristics, clinical symptomsand outcomesof patientsrequiring intensive care unit(ICU)admission forinfectiontransmitted by cats or dogs.

Patients and methods/materials and methods: PETSEPSIS was a retrospective multicenter study conducted in 46 ICUs in France. Consecutive adults admitted to the ICU from January 2009 to December 2019 for sepsisafter bitesby cats or dogs, or due to Pasteurellaspp, Bartonellaspp, Tularemiaspp, orCapnocytophagaspp, were included. Data analysis was descriptive, and univariate and multivariate analysis was performed to identify risks factors of mortality.

Results: 174 patientswere included; median age was64.5years[50.3–73.7]andSAPSIIscore was41.5 [28–57.5].51.2% ofthepatients were mechanically ventilated, 53.49% needed cardiovascular drugs, 18.97% renal replacement therapy and 3.6%had limbs amputation.Duration of mechanical ventilation was7 [4–13] days.The length of stay inICUwas 7days [4–16].Hospital mortality was 24.1%.Clinicalpresentation or severityof sepsisdid not vary with bacteria, exceptfor infections due toPasteurella spp, for which patients seem to be older and to have more comorbidities. Contactswith cats were more often reported in Pasteurella infections whereas contactswith dogs were more often reported inCapnocytophagaand Bartonella infections.Risks factors of death were older age, worse SAPSII score, bacteriemia, chronic alcohol abuse and chronic hepatic dysfunction, need for endotracheal ventilation, vasoactive drugs and renal replacement therapy. Initial antimicrobial therapy had no impact on outcome.

Conclusion: This is the first and largest multicenterobservational study onseveresepsis transmitted by cats or dogs.Mortality was associated withpreexisting diseaseandseverity of organ failure,and not withbacteria identified.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Nicholas Heming (nicholas.heming@aphp.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-012

Rationale: The role of corticosteroids in the management of non COVID-19 related acute respiratory distress syndrome (ARDS) remains controversial.

Patients and methods/materials and methods: The Activated Protein C and Corticosteroids for Human Septic Shock (APROCCHSS) trial was a placebo, controlled trial of the efficacy and safety of the corticosteroids hydrocortisone and fludrocortisone, in adults with septic shock. Patients with ARDS were identified a priori as a subgroup of interest. The primary outcome was 90-day mortality. Key secondary outcomes included mortality at intensive care unit (ICU) discharge, at hospital discharge, at day 28 and day 180 as well as mechanical ventilation free-days within 28days.

Results: Among 1241 APROCCHSS participants, 648 met the criteria for ARDS at the time of inclusion. In ARDS patients, the mean ± SD time between ICU admission and randomization was 2 ± 4days. There were 155/320 (48.4%) deaths at day 90 in the corticosteroid arm and 186/328 (56.7%) in the placebo arm. The risk ratio (RR) for death at day 90 was 0.85 (95%CI: 0.74 to 0.99; P = 0.04) and 0.91 (95% CI: 0.74 to 1.12; P = 0.36) in patients with and without ARDS, respectively (Breslow-Day homogeneity test P = 0.45). In the ARDS subset, the RR of dying was 0.83 (95%CI: 0.69 to 0.99; P = 0.04) at 28day, 0.82 (95%CI: 0.70 to 0.97; P = 0.02) at ICU discharge, 0.84 (95%CI: 0.72 to 0.98; P = 0.03) at hospital discharge, and 0.87 (95%CI: 0.76 to 1.00; P = 0.05) at day 180. In the ARDS subset, there was no significant difference regarding ventilator free days 8 ± 10 vs. 7 ± 9 in the corticosteroid vs placebo arm (P = 0.24).

Conclusion: The effects of hydrocortisone plus fludrocortisone on 90-day mortality are consistent in septic shock with and without early ARDS.

Reference 1

Annane D et al. NEJM 2018.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Tai Pham (tai.pham@aphp.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-13

Rationale: Intubated patients with the acute respiratory distress syndrome (ARDS) usually receive protective ventilation limiting plateau pressure below 30 cmH2O and, if possible, a driving pressure under 15 cmH2O. However, airway pressures might not reflect the actual pressure applied to the lung. Transpulmonary pressure is the difference between airway pressure and pleural pressure, the latter is estimated by the esophageal pressure (1). This technique and measurements had been validated, initially, in healthy patients in half-sitting position, and then confirmed in different positions (2). Few studies have evaluated the impact of patient positioning on pleural pressure and transpulmonary pressure in intubated patients with ARDS. We aimed to assess the changes in esophageal and transpulmonary pressures at different positions (supine at 0° and semi-recumbent at 45°) in intubated patients with ARDS.

Patients and methods/materials and methods: Single-center, prospective study performed in a medical ICU of a tertiary care center. All patients diagnosed with ARDS in the study period and monitored with an esophageal catheter were included. We collected end-expiratory and end-inspiratory esophageal pressures, airway pressures and intra-abdominal pressure at the two positions. Inspiratory transpulmonary pressure was calculated using the elastance ratio method. Positive end expiratory pressure (PEEP) was not modified and remained at the level set by the clinician in charge of the patient. Comparisons of values at 0° and 45° were performed using paired Wilcoxon tests.

Results: We included 36 patients admitted from January to August 2022. Their median (IQR) age, SOFA score and SAPS II were 63years (58–72), 9 (5–12) and 52 (33–69) respectively and 77% were men. Twenty-nine of them (80%) were intubated for pneumonia including seven positive to SARS-CoV2 (19%). Their median PaO2/FiO2 on the day of intubation was 96 (72–111) mmHg and 33 (92%) had moderate or severe ARDS. At the time of measurement, set and total PEEP were 12 [10;14] and 13 [11;15] cmH2O respectively. Comparisons of the airway pressures, esophageal pressures and transpulmonary pressures are shown in Table 1. Intra-abdominal pressure also significantly increased from 10 [7;13] to 20 [16;25] cmH2O when changing position from 0° to 45° (p = 0.001).

Conclusion: Keeping PEEP stable and changing the patient position from supine at 0° to semi-recumbent at 45° significantly increases all airway pressures and decreases the esophageal pressure. The relative changes led to a non-significant increase in transpulmonary plateau pressure but a significant increase all transpulmonary pressures (end-inspiratory; end-expiratory) and driving pressures.

Reference 1

Coudroy R, Chen L, Pham T, Piraino T, Telias I, Brochard L. Acute Respiratory Distress Syndrome: Respiratory Monitoring and Pulmonary Physiology. Semin Respir Crit Care Med. 2019;40(01):066?80.

Reference 2

Milic-Emili J, Mead J, Turner JM. Topography of esophageal pressure as a function of posture in man. J Appl Physiol. 1964;19(2):212?6.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Damien Barrau (djrbarrau@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):CO-014

Rationale: Complete airway closure (CAC) is a frequent phenomenon in ventilated acute respiratory distress syndrome (ARDS) patients. Setting external positive end-expiratory pressure (PEEP) regarding the presence of a CAC and the level of airway opening pressure (AOP) is still debated. The objective of this study was to assess the prevalence of CAC in ventilated-ARDS patients and its evolution over time.

Patients and methods/materials and methods: We conducted a bicentric study including all adult patients with moderate or severe ARDS between August 2021 and August 2022. We detected Complete airway closureand measured airway opening pressure using low-flow pressure–volume curve (from PEEP 0 cmH2O) every day, in the absence of patient’s inspiratory effort.

Results: Here are preliminary data of 25 patients, corresponding to 89 PV curves. The prevalence of CAC with AOP > 5 cmH2O was 32% (8/25 patients) and 2 patients (6%) developed once during the follow up. The median AOP was 5.3 cmH2O on day 1, 6.2 cmH2O on day 2, and 4.7 cmH2O on day 3. The average weight gain during the follow-up was significantly higher in patients with CAC compared to those without CAC (8.0 [0.5–16.5] kg vs. 1.5 [0.0–9.0] kg, p = 0.019). CAC with an AOP > 5 cmH2O was more common in patients with focal than in non-focal ARDS (50% vs. 6%, p = 0.025).

Conclusion: The CAC is a frequent and dynamic phenomenon in ARDS patients. The occurrence of CAC seems to be influenced by hydrolytic balance and chest CT pattern. Repeated measurements seem to be a requisite for personalized lung protective ventilation.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Cassilia Dei (cassilia.dei@live.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-015

Rationale: A non-invasive method called "Recruitment to inflation ratio" (R/I ratio), has been proposed to assess the potential of recruitment in patients with acute respiratory distress syndrome (ARDS) and then to titrate and personalize the level of positive end-expiratory pressure (PEEP) to ensure lung recruitment while limiting lung hyperinflation. Prone positioning (PP) is recommended in most severe patients with ARDS. The effects of PP on oxygenation and respiratory mechanics according to the R/I ratio remain to be determined.

Patients and methods/materials and methods: This multicenter and prospective study was conducted in 7 French intensive care units and included patients with ARDS requiring prone positioning. The R/I ratio as well as oxygenation parameters and respiratory mechanics were recorded just before and during the 16-h PP session and after supine repositioning. The respiratory rate, the PEEP level and the inspired oxygen fraction (FiO2) remained unchanged during the study period. The median R/I ratio before PP in our cohort was used to classify patients as high and low recruiters.

Results: Overall, 41 patients were included: 24 (58%) were men, the median age was 63 (57–70) years old and 32 (78%) had ARDS related to SARS-CoV-2 pneumoniae. The median R/I was 0.50 (0.30–0.84): R/I ratio was 0.24 (0.06–0.35) in low recruiters and 0.83 (0.58–1.00) in high recruiters (p < 0.001). Baseline characteristics, partial arterial pressure of oxygen over FiO2 (PaO2/FiO2) ratio, partial arterial pressure of carbon dioxide (PaCO2) and respiratory mechanics did not significantly differ between high and low recruiters. In high recruiters, PP induced a significant decrease in R/I ratio (p < 0.0001) and increase in PaO2/FiO2 ratio p < 0.0001), while the driving pressure, the compliance of the respiratory system and PaCO2 were unchanged. The PP-induced decrease in R/I ratio persisted after supine repositioning, while the PP-induced increase in PaO2/FiO2 diminished as early as 2h after supine repositioning but the PaO2/FiO2 ratio did not differ between 2 and 4h after supine repositioning. In low recruiters, PP induced a significant increase in PaO2/FiO2 ratio (p < 0.0001), while the R/I ratio, the driving pressure, the compliance of the respiratory system and PaCO2 were unchanged. As in high recruiters, the PP-induced increase in PaO2/FiO2 diminished as early as 2h after supine repositioning but the PaO2/FiO2 ratio did not differ between 2 and 4h after supine repositioning.

Conclusion: The effects of PP on oxygenation and respiratory mechanics appear to differ in high and low recruiters. Inclusions are in progress.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Christopher Lai (christopher.lai@aphp.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-016

Rationale: One of the beneficial respiratory effects of prone position in patients with acute respiratory distress syndrome related to Coronavirus disease 2019 (C-ARDS) could be secondary to improvement in lung mechanics. However, these improvements are inconstant. The objective of the study was to investigate whether an improvement of the compliance of the respiratory system (Crs) and reduction of the driving pressure at the end of a prone position session could be predicted by the potential of lung recruitment at baseline in supine position.

Patients and methods/materials and methods: In this prospective, observational, monocentric study in patients who underwent prone position for C-ARDS, from January to May 2021, respiratory variables were assessed just before prone positioning and at the end of the session. Respiratory variables included mechanical ventilation settings and variables of respiratory mechanics, including the recruitment-to-inflation ratio (R/I), an estimate of the potential of lung recruitment compared to lung overinflation.

Results: In 41 patients, 156 prone position sessions lasting 18.9 ± 2.7h, were evaluated. Mortality at day 60 was 49%. Neuromuscular blockade agents were used in 91(58%) sessions. At baseline, tidal volume was 6.1 ± 0.5mL/kg of predicted body weight, PEEP was 14 ± 3 cmH₂O and plateau pressure was 29 ± 4 cmH₂O. The median of the R/I ratio was 0.53 (0.31–0.79), separating low- and high-recruiters. The airway opening pressure was present in 76 (49%) patients and was 6 (5–9) cmH2O in these patients. At the end of the prone position session PaO₂/FiO₂ increased from 111 ± 31mmHg to 165 ± 67mmHg (p < 0.001), with an increase > 20mmHg in 96 (62%) sessions. The driving pressure decreased from 14.3 ± 4.4 to 13.7 ± 4.7 cmH₂O (p = 0.045) and Crs increased from 32 ± 11 to 34 ± 11 (p = 0.037) ml/cmH2O. Whereas the PaO₂/FiO₂ improved to the same extent in both low- and high-recruiters, driving pressure and Crs improved only in high-recruiters (from 14 ± 4 to 12 ± 4, p = 0.027, and from 34 ± 11 to 38 ± 13, p = 0.014, respectively).

Conclusion: Whereas oxygenation improves to the same extent in low- and high-recruiters with prone position in patients with C-ARDS, driving pressure and Crs improve only in high-recruiter patients. The benefits of prone position could be thus even greater in these patients.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Mathilde Taillantou-Candau (m.taillantou@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):CO-017

Rationale: The optimal ventilation mode after the early phase of acute respiratory distress syndrome (ARDS) when spontaneous ventilation (SV) is resumed is still debated. Volume Assist Control (VAC) is the most frequently used mode, but Pressure regulated volume control (PRVC) has been proposed to limit dyspnea and asynchrony and Airway Pressure Release Ventilation (APRV), a non-synchronized mode, has been suggested to decrease the risk of ventilator induced lung injury (VILI) and improve compliance, gas exchange and respiratory variability. We hypothesized that APRV may promote SV and respiratory variability without altering gas exchange, dyspnea and respiratory effort.

Patients and methods/materials and methods: In this prospective, cross-over study, APRV, PRVC and VAC were randomly applied as soon as SV reached 20 to 30% of minute ventilation. For each mode, the ventilator was set to obtain similar positive end-expiratory pressure, tidal volume and respiratory rate. Airway and esophageal pressure and volume distribution in electrical impedance tomography (EIT) were continuously recorded. Gas exchange, hemodynamics, dyspnea (IC-RDOS score), P0.1 and pressure–time product (PTP) were assessed at the end of application of each mode. Tidal volume and PTP variability were assessed by the respective coefficients of variation (standard deviation divided by the mean value).

Results: Twenty ARDS patients were included and data were analyzed in 19 patients (one patient discontinued the study because of clinical deterioration). Main ARDS cause was pneumonia in 17 patients (89%). The duration of ventilatory support was 12 (5–19) days before inclusion. Minute ventilation did not differ between the three modes. Sedative drugs were unchanged during the study and RASS (Richmond agitation sedation scale) was similar among the three modes. Gas exchange, hemodynamics, IC-RDOS score, ventilation distribution, P0.1 and PTP did not significantly differ among the three modes. Tidal volume variability was significantly higher in APRV than in VAC and PRVC, PTP variability was significantly higher in APRV than in VAC but did not differ between APRV and PRVC.

Conclusion: When SV is resumed after the early phase of ARDS, gas exchange, hemodynamics, dyspnea and respiratory effort did not differ among APRV, VAC and PRVC. Respiratory variability was higher in APRV than in VAC but did not differ between APRV and PRVC.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Pierre-Nicolas Bris (pierrenicolasbris@free.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-018

Rationale: Acute respiratory distress syndrome (ARDS) occurring in patients with hematological malignancies (HM) is a life-threatening condition with specific features. Mortality rate remains high but improvement has been described over the past several years. We aimed to describe the characteristics and outcomes of ARDS in HM patients admitted during one year in French ICUs (Intensive Care Units).

Patients and methods/materials and methods: Data for this nationwide cohort study were collected from the French national hospital database (Programme de Médicalisation des Systèmes d’Information (PMSI)). All patients (18years or older) admitted to French ICUs in 2017 and with an ICD-10 diagnosis of ARDS were included. Diagnosis of HM or solid cancer was detected using an algorithm provided by the Institut National du Cancer (INCa). Three groups were determined and compared according to the presence of an HM, a solid cancer or absence of cancer. Primary endpoint was ARDS mortality (in ICU or at day 90). Secondary endpoints were ICU characteristics (duration of mechanical ventilation, prone positioning, the use of high flow nasal cannula or vasopressors, the use of renal replacement therapy), ARDS characteristics (associated infections) and mortality predictors.

Results: A total of 12 846 ARDS patients were included. Among them, 990 had HM and 2744 had a solid cancer. The main malignancies were non-Hodgkin lymphoma (NHL) (28.5%), acute myeloid leukemia (AML) (20.4%) and multiple myeloma (19.7%). Day-90 mortality in patients with HM was higher than patients with no cancer (64.4% vs 46.6% p = 0.0001) but there was no significant difference when comparing to patients with solid cancer (64.4% vs 61.4%, p = 0.091). Intubation rate was lower in patients with HM in comparison with both groups (87.7% vs 90.4% p = 0.017 for patients with solid cancer and 87.7% vs 91.3% p = 0.0002 with no cancer) and SAPSII at admission was higher in this population (49 vs 42.3 p < 0.0001 in solid cancer patients and 49 vs 42.7 p < 0.0001 in no cancer patients). Independent predictors of mortality for patients with HM were a diagnosis of NHL or AML, age, a high SAPS II score, a renal replacement therapy and a septic shock. Bacterial pneumonia and extrapulmonary sepsis causing ARDS were protective.

Conclusion: Mortality remains high in patients with HM admitted in ICU with ARDS but patients’ and ARDS characteristics evolved in comparison of older studies.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Guillaume Dumas (GDumasgalant@chu-grenoble.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-019

Rationale: Patients with hematological malignancies are at high risk for life-threatening complications. In this context, most of the studies have focused on the initial oxygenation strategy but little attention has been paid to the impact of hyperoxemia and excess oxygen use in this population. We sought to investigate the impact of partial pressure of arterial oxygen (PaO2) and administered oxygen (FiO2) on day-28 mortality in critically ill patients with hematologic malignancies.

Patients and methods/materials and methods: Clinical and oxygenation data from three international cohorts (Europe, Canada, Oceania) of critically-ill patients with hematologic malignancies who required respiratory support (noninvasive ventilation, high-flow nasal cannula, or invasive mechanical ventilation) were obtained. We descriptively evaluated the incidence of normoxemia (paO2 between 60-120mmHg), hypoxemia (paO2 ≤ 60) and hyperoxemia (paO2 ≥ 120). The association between PaO2 on day 1 and 28days-mortality was first modeled using a restricted cubic spline. We used mixed-effect Cox models to estimate the effect of hyperoxemia and excess oxygen use (defined as a FiO2 ≥ 0.6 with a PaO2 > 100mmHg on day 1 of ICU admission) on day-28 mortality.

Results: 11,249 patients were included in the analysis. On day 1, 6911 patients (61.4%) had normoxemia, 1454 (12.9%) hypoxemia, and 2884 patients (25.6%) hyperoxemia. The crude day-28 mortality rate was 40.6%. Excess oxygen use was noticed in 2,201 patients (20 After adjustment, there was a significant association between PaO2 and day-28 mortality which demonstrated a U-shape relationship (p < 0.001). We found an independent association between day-1 excess oxygen use and subsequent day-28 mortality (aHR: 1.11[1.03–1.20], p-value = 0.006). On a propensity-matched analysis (n = 4,402 patients), excess oxygen use remained associated with day-28 mortality (HR:1.27[1.16–1.39], p-value < 0.001). In subgroup analysis, higher PaO2 values were associated with poor outcomes in patients admitted with neurological disorders or cardiac arrest contrary to those with sepsis.

Conclusion: In critically-ill patients with hematological malignancies, excess oxygen use, and high PaO2 exposure on day 1 were associated with increased mortality. These results suggest a new way to potentially improve cancer patients’ prognosis and should be confirmed in the future.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Antoine Lafarge (antoine.lafarge@aphp.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-20

Rationale: Despite notable changes in the practice of allogeneic hematopoietic stem cell transplantation (Allo-HSCT) in the past years, intensive care unit (ICU) admission is still associated with poor outcome in recipients. Since multicenter longitudinal data on this topic remain scarce, we analyzed reasons for ICU admission, critical care management as well as short- and long-term outcome of critically-ill Allo-HSCT recipients.

Patients and methods/materials and methods: The medical records of Allo-HSCT recipients admitted to 14 French ICU during a 6-year period from 1 January 2015 to 31 December 2020 were retrospectively analyzed. Acute graft-versus-host disease (aGVHD) trajectory prior to ICU admission was specifically assessed.

Results: Data from 1164 Allo-HSCT recipients admitted to the ICU during the study period were analyzed. Median age at ICU admission was 56.4 [42.7–63.6] years, 61.4% were male. Median time from Allo-HSCT to ICU admission was 84 [15–369] days and 463 (39.8%) patients were transferred to the ICU during the engraftment period. Respiratory dysfunction (461 patients, 39.6%) and shock (228 patients, 19.6%) were the main reasons for ICU admission. Seventy-nine (6.8%) patients were admitted for acute kidney injury. Median SOFA at ICU admission was 6.0 [4.0–8.0]. Six hundred and eight (52.2%) patients displayed ≥ 1 infection(s) during the ICU stay, including opportunistic infections in 228 (37.5%) patients. In the whole cohort, 221 (19.0%) patients underwent renal replacement therapy, 478 (41.1%) received vasopressors and 440 (37.8%) required mechanical ventilation. Eight hundred and sixty-two (74.1%) patients could be discharged alive from the ICU and 590 (57.4%) from the hospital. Day-90, 1-year and 3-year survival rates were 54.1%, 41.3% and 33.8% respectively (Figure 1). Age (hazard ratio (HR) 1.02 [1.01–1.03]; P < 0.001), glucocorticoid-refractory GVHD (HR 1.46 [1.22–1.75]; P < 0.001), male sex (HR 1.50 [1.14–1.98]; P < 0.001), glucocorticoids posology (HR 1.61 [1.18–2.18]; P < 0.001), vasopressors (HR 1.70 [1.27–2.29]; P < 0.001), mechanical ventilation (HR 2.67 [1.95–3.64]; P < 0.001) and renal replacement therapy (HR 3.66 [2.48–5.42]; P < 0.001) were independent predictive factors of day-90 mortality.

Conclusion: In Allo-HSCT recipients, ICU admission remains associated with a dismal prognosis. Beyond host factors, glucocorticoid-refractory aGVHD and organ dysfunctions requiring live sustaining therapies are the major determinants of death. Further analyses are warranted to stratify Allo-HSCT recipients based on a high risk of mortality and identify patients for whom critical care management may be aggressive but nonbeneficial.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Mickael Lescroart (dr.lescroart@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):CO-21

Rationale: Despite significant advancements in oncology/hematology over the past decades, cancer patients (CP) continue to face a significant cardiovascular (CV) burden due to both their malignancy and the toxicity of anti-cancer therapy. No data is currently available regarding the characteristics and outcomes of cancer patients (CP) admitted in intensive care unit (ICU) for cardiogenic shock (CS). We aimed to provide a contemporary cohort of cardiogenic shock in CP and determine risk factors associated with ICU mortality.

Patients and methods/materials and methods: 214 oncological patients with CS from 21 French ICUs were included between 2010 and 2022. Clinical data were extracted from medical records. The primary outcome was overall ICU survival. Secondary outcomes were the need for renal replacement therapy (RRT), mechanical ventilation, and ICU length of stay. Risk factors for ICU mortality were determined through logistic regression.

Results: Hemopathy (68.7%) was the most represented type of malignancy. CV risk factors were prevalent as one third of patients had a history of smoking or hypertension. The overall ICU mortality was 50% with a 37-days median survival. One hundred sixty-eight patients (78.5%) required vasopressor therapy, 140 (65.7%) mechanical ventilation and 52 (24.9%) renal replacement therapy. Independent risk factors for ICU mortality were: mechanical ventilation (OR 2.55 [1.16–5.72], p = 0.02), vasopressor therapy (OR 3.36 [1.15–11.31], p = 0.03), home medication with renin–angiotensin–aldosterone system inhibitors (OR 2.84 [1.06–8.17], p = 0.04), and the inotropic drug (OR for dobutamine, versus epinephrine: 0.27 [0.08–0.78], p = 0.02). Thirteen patients (6.0%) required VA-ECMO with similar survival rates (n = 7/13, p = 0.95). Among patients alive at discharge, the median survival was 418days.

Conclusion: The prognosis of cancer patients admitted in ICU for cardiogenic shock remains poor with a high mortality rate and frequent need for invasive organ support. The population that could benefit from intensive care has yet to be determined.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Clara Vigneron (claravigneron@hotmail.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-22

Rationale: Increasing evidence argue for promotion of tumorigenesis through activation of renin-angiotensin system pathway. Accordingly, a benefit of renin-angiotensin system blockers (RABs) treatments has been suggested in patients with solid cancers in terms of survival. Our study aims to evaluate in-ICU survival and one-year survival in cancer patients admitted to the ICU with respect to the use of RABs.

Patients and methods/materials and methods: We conducted a retrospective observational single-center study in a 24-bed medical ICU. Consecutive adult patients (age ≥ 18years) with a diagnosis of solid tumour (previously known or diagnosed during the ICU stay) requiring unplanned ICU admission were included.

Results: From 2007 to 2020, 1845 patients with solid malignancies were admitted (median age 67years [59–75], males 61,7%). The most frequent primary tumor sites were the gastrointestinal tract (26.8%), the lung (24.7%), the urological tract (20.1%), and gynecologic and breast cancers (13.9%). RABs were used in 414 patients, distributed into 220 (53.1%) with angiotensin-receptor blockers (ARBs) and 194 (46.9%) with angiotensin-converting enzyme inhibitors (ACEi). After adjustment in a multivariate model, ARBs use (OR = 0.62 (0.40–0.92), p = 0.03) and ACEi use (OR = 0.52, 95%CI (0.32–0.82), p = 0.006) were both associated with improved in-ICU survival (Figure 1). Treatment with ARBs was independently associated with decreased one-year mortality (OR = 0.6 95% CI (0.4–0.9), p = 0.02), whereas treatment with ACEi was not (Figure 1).

Conclusion: The use of ARBs, but not of ACEi, is associated with improved one-year survival in patients with solid malignancies admitted to the ICU.

Compliance with ethics regulations: Yes in clinical research.Whole cohort survival analysis

Correspondence: Clément Devautour (cdevautour@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):CO-23

Rationale: Bleeding events are common complications in critically ill patients with haematological malignancies. We aimed to describe acquired major bleeding events during their ICU stay.

Patients and methods/materials and methods: We conducted an observational monocenter retrospective study including all adult patients with a history of haematological malignancy requiring unplanned ICU admission over a 12-year period (2007–2018). Non-inclusion criteria were the following: admission to secure a procedure, planned admissions following elective surgery and patients with haematological malignancies cured for more than 5years. Endpoints were major bleeding events (WHO classification grade 3–4) acquired during their ICU stay.

Results: A total of 1012 patients (age = 65 [55–76] years) were included. The main haematological malignancies were non-Hodgkin lymphoma (n = 339, 33.5%), acute myeloid leukemia (n = 183, 18.1%), and multiple myeloma (n = 171, 16.9%). Most patients were recently diagnosed (n = 339, 33.5%) or in progression under treatment (n = 262, 25.9%). Five hundred ninety-six (58.9%) patients were under treatment. The leading causes of admission were circulatory (n = 347, 34.2%) and acute respiratory failures (n = 287, 28.4%). Among them, 109 (10.8%) presented at least one episode of acquired major bleeding event in the ICU, mainly gastrointestinal (n = 45, 41.3%), airway (n = 13, 11.9%) and vascular catheter (n = 13, 11.9%) bleeding events. Compared to patients without haemorrhagic event, they were more likely to have been admitted to the ICU for acute kidney injury (56.9% vs. 38.1% vs p < 0.001) and haemorrhagic event (28.4% vs 2.5%, p < 0.001). Biological exams performed at ICU admission showed multiple alterations in coagulation measurements including decreased platelet count (54 vs 80 G/L, p = 0.002), prolonged prothrombin time (62.0 vs 72.5%, p < 0.001) and increased activated partial thromboplastin ratio (1.25 vs 1.18, p = 0.002), acute kidney dysfunction as assessed by increased urea (10.6 vs 8.7mmol/L, p = 0.011) and creatinine levels (129.0 vs 102.5µmol/L, p = 0.007), lower albumin level (23.0 vs 28.0g/L, p = 0.026) and increased bilirubin level (17.0 vs 12.0µmol/L, p < 0.001). They required more frequently invasive mechanical ventilation (85.3% vs 43.4%, p < 0.001), inotropes/vasopressors (67.9% vs 32.6%, p < 0.001) or renal replacement therapy (46.8% vs 21.5%, p < 0.001) during their ICU stay. Accordingly, they exhibited poorer prognosis: ICU, hospital and one-year mortality rates were 55.0% vs 18.0% (p < 0.001), 59.6% vs 23.7% (p < 0.001) and 80.4% vs 54.7% (p < 0.001), respectively.

Conclusion: In patients with haematological malignancies, ICU-acquired major bleeding events are associated with acute kidney injury and recent history of bleeding event. Our study revealed biological disturbances associated with those events. Patients developing major bleeding events exhibit a dismal short and long-term prognosis.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Lina Jeantin (lina.jeantin@aphp.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-24

Rationale: Electroencephalography (EEG) is recommended for the practical approach to the diagnosis and prognosis of encephalitis. We aimed to investigate the prognostic value of standard EEG (_stdEEG) in adult patients with severe herpes simplex encephalitis (HSE).

Patients and methods/materials and methods: We performed a retrospective analysis of consecutive ICU patients with severe HSE in 38 French centers between 2006 and 2016. Patients with at least one _stdEEG study performed at ICU admission were included. _stdEEG findings were reviewed independently by two investigators. The association between _stdEEG findings (i.e., background activity, lateralized periodic discharges, seizures/status epilepticus, and reactivity to painful/auditory stimuli) and poor functional outcome, defined by a score on the modified Rankin Scale (mRS) of 3 to 6 (moderate to severe disability or death) at 90days, were investigated.

Results: We included 214 patients with at least one available _stdEEG study. The first stdEEG was performed after a median time of one (IQR 0 to 2) day from ICU admission. At the time of recording, 138 (64.5%) patients were under invasive mechanical ventilation. Lateralized periodic discharges were recorded in 91 (42.5%) patients, seizures in 21 (9.8%) and status epilepticus in 16 (7.5%). In the whole population, reactivity to auditory/noxious stimuli was tested in 140/214 (65.4%) patients and was absent in 71/140 (33.2%) cases. In mechanically ventilated patients, _stdEEG reactivity was tested in 91/138 (65.9%) subjects, and was absent in 53/91 (58.2%) cases. Absence of reactivity was the only independent _stdEEG finding associated with poor functional outcome in the whole population (OR 2.80, 95%CI 1.19 to 6.58) and in the subgroup of mechanically ventilated patients (OR 4.99, 95%CI 1.6 to 15.59). adjusted analyses for common clinical predictors of outcome and sedation at time of _stdEEG revealed similar findings in the whole population (OR 2.03, 95% CI 1.18 to 3.49) and in mechanically ventilated patients (OR 2.62, 95%CI 1.25 to 5.50).

Discussion: The lack of EEG reactivity to external stimuli in critically ill patients with HSE is associated with poor outcome, independently from the presence of sedation and common clinical prognostic factors in this population (including age, coma, and fever at admission). To date, little data is available about the prognostic value of _stdEEG and our study assembled one of the largest cohorts of patients with HSE requiring ICU admission to assess the value of _stdEEG markers for functional outcome.

Conclusion: Absence of _stdEEG reactivity to auditory/noxious stimuli is an independent marker of poor functional outcome in severe herpes simplex encephalitis.

Reference 1

ENCEPHALITICA study group, Jaquet P, de Montmollin E, Dupuis C, Sazio C, Conrad M, et al. Functional outcomes in adult patients with herpes simplex encephalitis admitted to the ICU: a multicenter cohort study. Intensive Care Med. 2019.

Reference 2

Azabou E, Navarro V, Kubis N, Gavaret M, Heming N, Cariou A, et al. Value and mechanisms of EEG reactivity in the prognosis of patients with impaired consciousness: a systematic review. Crit Care. 2018.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Camille Legouy (c.legouy@ghu-paris.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-25

Rationale: Pneumococcal meningitis is life-threatening infection with high morbidity and mortality rates, likely resulting from the occurrence of neurological complications, such as brain oedema, ischemic stroke, and intracranial hemorrhage. To investigate the association of neurological complications diagnosed on admission neuroimaging with functional outcomes of adults with severe pneumococcal meningitis, requiring care in the intensive care unit (ICU).

Patients and methods/materials and methods: We performed a retrospective multicenter study in 11 centers in France, between February 2005 and September 2021. We included consecutive adults with pneumococcal meningitis who required at least 48h of ICU stay and who underwent neuroimaging assessment. Neurological complications were defined as 1) ischemic stroke, 2) intracranial hemorrhage (i.e. intracerebral hematoma, subdural hematoma, or subarachnoid hemorrhage), 3) abscess/empyema, 4) ventriculitis, 5) thrombophlebitis, 6) hydrocephalus, and 7) cerebral oedema on neuroimaging at admission. Exposure: Neuroimaging at admission. Main outcome and measures: AUnfavorable neurological outcome at 90days after ICU admission, defined by the modified Rankin Scale (mRS) > 2, indicating moderate-to-severe disability or death.

Results: Among the 237 patients included, 103 (44%, 95%CI 37–50%) had an unfavorable outcome at 90days, including 71 (30%, 95%CI 25–37%) deaths. Neurological complications at admission occurred in 68 of the 237 patients (31%) and in 70 patients (64%) during ICU stay with a majority of ischemic stroke. Using a multivariate logistic model at ICU admission, we identified four predictive markers independently associated with unfavorable outcome: chronic alcohol consumption (adjusted Odds Ratio (aOR 2.37, 95%CI 1.09–5.26), focal neurological sign(s) (aOR 2.11 95%CI 1.06–4.24), neurological complications on ICU admission (aOR 3.65, 95%CI 1.83–7.47) and cerebrospinal fluid (CSF) leukocyte count < 1000 cell/microL, (aOR 4.17, 95%CI 2.22–8.34). Competing risk analysis with death as competing risk and showed that chronic alcohol consumption was the only significant marker for the risk of persistent disability (mRS > 2) at 90days (Cause specific Hazard Ratio 4.26, 95%CI 1.83–9.91).

Conclusion: In adult patients with severe pneumococcal meningitis, the identification of neurological complications on admission neuroimaging is an independent predictor of unfavorable neurological outcome. Chronic alcohol consumption is the only predictive risk of persistent disability at three months in survivors.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Sarah Benghanem (sarah.benghanem@aphp.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-26

Rationale: We aimed to investigate the prevalence of early EEG abnormalities in adult stroke patients requiring mechanical ventilation and their predictive value for functional outcome.

Patients and methods/materials and methods: We conducted a pre-planned ancillary study of the prospective multicenter SPICE cohort study (ClinicalTrials.gov Identifier: NCT03335995, study protocol PMID:32026446). We included patients from the SPICE cohort with at least one EEG study performed during intensive care unit (ICU) stay. The primary endpoint was poor functional outcome at one year, defined by a score of 4–6 (severe disability or death) on the modified Rankin scale (mRS).

Results: Among the 364 patients prospectively enrolled in SPICE, 153 patients (49 ischemic strokes, 52 intracranial hemorrhages, and 52 subarachnoid hemorrhages) had at least one EEG performed, after a median time of 4 (IQR, 2–7)) days. Diffuse slowing (70% vs 63%, p = 0.37), focal slowing (38% vs 32%, p = 0.15), periodic discharge (2.3% vs 3.7%, p = 0.9), and electrographic seizure (4.5% vs 3.7%, p = 0.4) rates were comparable between poor outcome and good outcome patients. After adjustment for clinical confounders (i.e. age, stroke subtype, non-neurological organ failure, Glasgow coma score), an unreactive EEG background to noise (OR 4.8, 95%CI (1.84–12.53)) or pain (OR 4.76, 95%CI (1.89–11.96)) remained independently associated with poor functional outcome. Conversely, a benign EEG (defined as continuous and reactive background, without seizure, periodic discharges, and burst suppression) was protective (OR 0.29, 95%CI (0.12–0.73)).

Conclusion: Unreactive and reactive EEG predicts respectively poor and good one-year functional outcomes in mechanically ventilated stroke patients.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Bertrand Hermann (bertrand.hermann@aphp.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-27

Rationale: Approximately 50% of post-cardiac arrest survivors remain comatose after 72h, a substantial proportion of which will have a poor neurological outcome, predominantly due to irreversible hypoxic-ischemic brain injury. Recent findings in healthy subjects and patients suggested that autonomic nervous system activity measured by brain–heart interactions could be reliable markers of consciousness and cognitive processing. Thus, we hypothesized that brain–heart interactions are associated with the severity of hypoxic-ischemic brain injury and the prognosis of these patients.

Patients and methods/materials and methods: In post-cardiac arrest patients still comatose 48h after sedation weaning, brain–heart interaction markers were computed on 5min of continuous EEG/ECG recording using a synthetic data generation model, gathering bidirectional interactions between EEG frequency bands (delta, theta and alpha) and heart-rate variability frequency bands (low and high frequency—LF and HF). The strength and complexity of the interactions were quantified using medians and refined composite multiscale entropy (RCMSE). Primary outcome was the severity of brain injury, assessed by: (i) standardized qualitative EEG classification, (ii) somatosensory evoked potentials (N20), and (iii) neuron-specific enolase levels. Secondary outcome was the 3-month neurological status, assessed by the Cerebral Performance Category score [good (1–2) vs. poor outcome (3–4-5)].

Results: Between January 2007 and July 2021, 181 patients [116 males (64%), median age 61years, age range 49–72years] were admitted to ICU for a resuscitated cardiac arrest (76% out-of-hospital, 69% non-shockable rhythm). Poor neurological outcome was observed in 134 patients (74%). Qualitative EEG patterns suggesting high severity were associated with a decreased sympatho-vagal balance. Severity of EEG changes were proportional to higher absolute values of brain-to-heart coupling strength (p < 2 × 10–3 for all brain-to-heart frequencies) and lower values of complexity (all p-values < 0.05 except for alpha-to-low frequency). Brain-to-heart coupling strength was significantly higher in patients with bilateral absent N20 and correlated with neuron-specific enolase levels at day 3. This aberrant brain-to-heart coupling (increased strength, decreased complexity) was also associated with 3-month poor neurological outcome (Table).

Conclusion: Our results suggest that autonomic dysfunctions may well represent hypoxic-ischemic brain injury post-cardiac arrest pathophysiology. These results open avenues for integrative monitoring of autonomic functioning in critical care patients with potential prognostic applications.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Benjamine Sarton (benjamine.sarton@hotmail.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-28

Rationale: Coma resulting from traumatic brain injury (TBI) and anoxia, remains a major issue for prognostication and care. Accumulative evidence suggests that coma is due to a disconnection within a “consciousness network”: fronto-parietal cortex and subcortical structures (mesocircuit)(1). Inflammation is potentially one of the leading causes of this connectome disruption. PET imaging allows in-vivo brain inflammation insight, triggering TSPO receptor expressed by activated microglia (MA). We aimed to study acute neuroinflammation in coma and its relationship with neurological outcome. We hypothesised that (i) brain injuries triggers MA within brain structures responsible for conscious processing (ii) with distinct MA pattern for anoxic and traumatic mechanism (iii) there is a significant relationship between in-vivo MA features and patients 3month’s neurological outcome.

Patients and methods/materials and methods: We conducted a cross sectional study of coma patients compared to controls. We included comatose adult patient, GCS < 9 with motor score < 6 induced by traumatic or anoxic brain injury (< 30days), free from sedations. Patients and controls were excluded in case of neurological illness, anti-inflammatory medication, low affinity binding phenotype for TSPO receptor. Control group was gender and age matched. The protocol included a neurological examination, T1-MRI and TSPO PET with18F-DPA-714 radioligand. Brain18F-DPA-714 binding was quantified with a supervised clustering algorithm (2). Neurological outcome was assessed with Coma recovery scale (CRS-r) at 3months. The ethic committee approved the study.

Results: Seventeen coma (6 anoxic and 13 TBI) patients and 20 controls were included (2019–2022). Significant increased binding potential (BP) of18F-DPA-714 was observed in coma patient compared to control, with differentintensity and spatial distribution between anoxic and traumatic groups (Figure 1A). Anoxic patients have showed significant increased BPin the thalamus, pallidum, putamen, posterior cingular cortex, medial prefrontal cortex (mPFC), precuneus; and in mPFC for trauma patients.A significant higher18F-DPA-714 binding was observed in coma patients with unfavorable outcome (Figure 1B) notably in the pallidum (p = 0,001), putamen (p = 0,02) and PCC (p = 0,004).

Conclusion: For the first time we have in vivo identified significant neuroinflammation in acute coma. MA is identified into networks implicated in conscious processing. Interestingly, MA profile was different between anoxic and traumatic coma. MA intensity and its spatial distribution were associated with patient’s further neurological outcome. Those results provide new insights into the characterization of neuroinflammation after severe brain injury and hold promise for personalized treatment.

Reference 1

Giacino JT, Fins JJ, Laureys S, Schiff ND. Disorders of consciousness after acquired brain injury: the state of the science. Nat Rev Neurol. févr 2014;10(2):99?114.

Reference 2

2. Schubert J, Tonietto M, Turkheimer F, Zanotti-Fregonara P, Veronese M. Supervised clustering for TSPO PET imaging. Eur J Nucl Med Mol Imaging. déc 2021;49(1):257?68.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: William Buffieres (buffieres.w@chu-toulouse.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-29

Rationale Delirium is a common problem in critically ill patients but has only recently been recognized as associated with important clinical outcomes such as long-term cognitive impairment. Currently, there is a major knowledge gap related to the covert yet potentially incapacitating higher-order cognitive impairment underpinning this disorder of consciousness.

Patients and methods/materials and methods: The aim of the current study was to analyze the electrophysiological signatures of delirium by using a specifically designed multidimensional auditory event-related potentials (ERPs) battery to probe hierarchical cognitive processes, including automatic and voluntary attention, self-related processing and detection of arithmetic incongruences. Clinical variables and electrophysiological data were prospectively collected in controls subjects (n = 14) and in critically ill patients with (n = 18) and without (n = 20) delirium.

Results: Overall, we have specifically identified in patients with delirium, both a preservation of low-level central auditory processing—notably the detection of local regularity violation (MMN and P3a or ‘local effect’)—and a coherent ensemble of covert higher-order cognitive dysfunctions encompassing the processing of global regularity violation (P3b or ‘global effect’), self-relevant stimulus (P3 to subject's own name), and simple arithmetic facts violation (P3 to incorrect results) (spatio-temporal clustering, p-cluster ≤ 0.05). In this setting, estimation of the cortical generators of the P3 response to target stimuli by source modeling identified the predominant role of the frontal lobe in control subjects, whose activation was modulated downward by delirium.

Conclusion: Multidomain cognitive processes involving P3 ERP responses to auditory protocols are impaired in patients with delirium and could be due to frontal brain dysfunction. We suggest that our results shed new light on the neuropsychological disorders associated with ICU-related delirium and may constitute a valuable method for patient’s bedside diagnosis and monitoring in this clinically challenging setting.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Meryl Vedrenne-Cloquet (meryl.vedrenne@aphp.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-30

Rationale: Accurate control of low tidal volumes (Vt) is crucial during pediatric protective ventilation. Vt delivery can be influenced by gas compression, humidity, temperature, and ventilatory circuit compliance, thus might be altered by the humidification system. The impact of the humidification system on the accuracy of Vt delivery during pediatric volume-controlled ventilation is unknown. The primary objective of this study was to assess the accuracy of Vt delivery by pediatric intensive care ventilators according to the humidification system. Secondary objectives were to assess: i) the accuracy of Vt delivery according to the presence or absence of an integrated Y-piece pneumotachograph; ii) the ability of ventilators to deliver and maintain the preset positive end-expiratory pressure (PEEP).

Patients and methods/materials and methods: Six modern intensive care ventilators equipped with a pediatric mode were tested on the ASL5000 test lung. Ventilators were tested in volume-controlled mode under 4 simulated pediatric bench models (full-term neonate (NN), infant (I), pre-school kid (PS), school kid (S)) with a heated humidifier (HH) or a heat-moisturer exchanger (HME), under various loading conditions (healthy, obstructive, restrictive), and with two levels of PEEP. Three ventilators equipped with Y-piece pneumotachograph were tested with or without the integrated pneumotachograph in the NN and I models. “Accurate Vt” delivery was defined as the volume error (percentage of the preset Vt under body temperature and pressure and saturated water vapour (BTPS) conditions) ≤ 10% of absolute preset value.

Results: Vt accuracy varied significantly across ventilators. The use of a HH or HME was associated with a significant difference in volume error for almost all the ventilators (p < 0.05). Ventilation with HH tended to deliver a lower-than-set Vt, the opposite was observed with HME, especially in the NN and I models. The use of an integrated pneumotachograph was associated with lower volume error in only one ventilator (p < 0.001). All the tested ventilators were able to maintain adequate PEEP levels.

Conclusion: The accuracy of Vt delivery is highly variable across pediatric intensive care ventilators during volume-controlled ventilation, and is altered by the humidification system, especially in the youngest patients.

Reference 1

Lyazidi A, Thille AW, Carteaux G, et al.: Bench test evaluation of volume delivered by modern ICU ventilators during volume-controlled ventilation. Intensive Care Med 2010; 36:2074–2080.

Reference 2

Lellouche F, Taillé S, Maggiore SM, et al.: Influence of ambient and ventilator output temperatures on performance of heated-wire humidifiers. Am J Respir Crit Care Med 2004; 170:1073–1079.

Compliance with ethics regulations: N/A.

Correspondence: Sirine Boussena (boussenasirine@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):CO-31

Rationale: Electrical impedance tomography (EIT) is a non-invasive real-time bedside lung ventilation monitoring. The aim of the study was to describe the changes in PEEP (positive end expiratory pressure) values induced by an EIT-guided PEEP trial.

Patients and methods/materials and methods: We conducted a prospective study (Peluche 1) with IRB approval n°2022-A00505-38. All patient admitted in the PICU and monitored with EIT (Pulmovista 500, Draeger, France) were included after informed consent. Demographic and clinical data, indications and recordings from the EIT were collected and analyzed.

Results: Between June and December 2020, ten patients, median [IQR] age 4.2years [1.7–6.9], were included in the study and nine had at least one EIT-guided PEEP trial. Among the 16 EIT-guided PEEP trials performed, ten led to a change of the level of PEEP: seven increases with a median change of 3 cmH20 [2–5] and 3 decreases (median [IQR] −2 cmH2O [−2; −4]). For the 7 patients with a diagnosis of ARDS, the PEEP value set after the EIT-guided PEEP trial was (median [IQR]) 3 cmH20 [0–6] upper than the lowest value proposed by the PEEP/FIO2 table.

Conclusion: In pediatric intensive care unit, EIT-guided PEEP trial is feasible and resulted in change in PEEP for 2 third of patients.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Christophe Milesi (c-milesi@chu-montpellier.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-32

Rationale: The moderate forms of acute viral bronchiolitis (AVB) benefit from ventilatory support by High-Flow Nasal Cannul (HFNC) or Continuous Positive Airway Pressure (CPAP). HFNC can be used in the paediatric wards unlike CPAP, which requires ICU monitoring. Early identification of this group is an essential organisational element. The ROX index [pulsea oximetry/FiO2/respiratory rate] is a relevant tool in predicting the HFNC failure in hypoxic adult population. Our objective was to validate the ROX index before (H0) and 1h (H1) after HFNC initiation to determine the failure in children with AVB.

Patients and methods/materials and methods: This is an ancillary study of a RCT performed in 16 pediatric intensive care units (TRAMONTANE 2), focusing on patients younger than 6months with moderate to severe bronchiolitis ()defined by mWCAS score > 3)Data were collected at H0 and H1 (ROX index, heart and respiratory ratemWCAS score, EDIN scale).

Results: From November 2016 to March 2017, 286 infants were included. Failure occurred early, within the first 6h for half of children. The ROXindex at H0 and H1 does not predict failure. H0: Area Under the Curve 0.56 (0.48; 0.63); H1: AUC 0.60 (0.53; 0.67). mWCAS was higher at H1 in the failure group (3.3 (± 1.0) vs 4.0 (± 1.5); p < 0.01). Comfort under HFNC improved more significantly in the successful group (Delta EDIN: −2.18 (± 2.38) vs −1.54 (± 2.34): p = 0.02).

Conclusion: ROX index cannot predict HFNC failure in patients with moderate bronchiolitis. Compliance with ethics regulations: Yes in clinical research.

Correspondence: Charlotte Pierron (pierron.charlotte@chl.lu)

Annals of Intensive Care 2013, 13(Suppl 1):CO-33

Rationale: Few studies reported the end-of-life (EOL) signs in pediatric intensive care unit (PICU), while these signs can be impressive or distressful for parents (1). The aim of this study was to assess the impact of EOL signs on parents' experience of their child's death.

Patients and methods/materials and methods: To evaluate the emotional state of parents confronted to EOL signs and their wishes on information they would like to receive, a psychologist interviewed parents, whose child had died in PICU during the previous months. Parents were informed of the study before their child's death and gave their consent to be contacted afterword. The interview was conducted face-to-face or by telephone and was semi-directive to avoid suggested answers. The interviews were transcribed and analyzed by a sociologist according to a qualitative method.

Results: We interviewed 11 parents, 2 to 27months after the death of their child. As parents thought about their child's end of life, the EOL signs were not in the foreground. They talked about it only as the psychologist mentioned it. Changes on physical appearance, e.g. oedemas, pallor…, were the EOL signs the most often mentioned by parents. Those signs were connected to a symbolism for the parents, even when they received information from the healthcare professionals about these signs. They were associated with the disembodiment of their child, the child became a stranger to them, they no longer recognized him. Furthermore, signs related to respiratory movements evoked a battle between the child and death, as if the child was fighting to survive. About the information they would like to receive on EOL signs, they suggested that it should be given while the signs were occurring and not before, to prevent parents from actively searching for it. And above all, that this information should not be provided in writing, as it would seem violent for them. Our study has some limitations, the interviews did not occur at the same time for all parents, making it challenging to compare them and most of the parents had lost their child during the first week of life.

Conclusion: Parents who lost their child in PICU did not spontaneously mention EOL perceptible signs. The most common signs they cited were signs that modified appearance, which are the most common in critically ill children (2). The most appropriate information about these signs, according to the parents, should be given at the time of their occurrence, orally.

Reference 1

1. Perkin RM: The agony of agonal respiration: is the last gasp necessary? J Med Ethics 2002; 28:164–169

Reference 2

2. Pierron C, Levy M, Mattioni V, et al.: Perceptible Signs of End of Life in Pediatric Intensive Care Patients. J Palliat Med 2022; 25:1829–1834

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Fleur Le Bourgeois (fleurlebourgeois@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):CO-34

Rationale: In France, decision to withdraw or limit life-sustaining treatment (WLST) in pediatrics is a medical decision discussed collectively, it is based mainly on medical criteria. However, subjective factors can influence this decision. To what extent do pediatric intensive care unit (PICU) healthcare professionals' (HCP) perceptions of the possibility of future parental presence and involvement influence the decision that is considered most appropriate?

Patients and methods/materials and methods: In a PICU, HCP were asked to fill out an anonymous electronic survey. It presented a vignette, where a decision of WLST is discussed. It was a 2-month-old patient with multi-organ disease, whose neurological, digestive, and respiratory prognosis were uncertain. The patient was dependent on mechanical ventilation and parenteral nutrition. HCP were asked to choose between continuation of artificial organ support in long-term care institution (LTCI), palliative support care until death or no decision. Five vignettes were presented: unknown parental location, parents living near the institution, parents living abroad, a single mother living 1h from the institution and the child was an orphan. Qualitative data were described as numbers and percentages.

Results: Fifty-eight HCP completed the questionnaire, HCP’s decisions changed dramatically according to scenario (Figure 1). When the parental background was unknown, the decision was balanced between all alternatives. When the parents were living nearby the LTCI, 64% of HCP decided continuation of treatments, conversely, if the parents were living far away or completely absent, a WLST was decided by 64 and 79% of HCP respectively.

Discussion: Although the decision of WLST mainly relies on medical factors, it considers non-measurable as the parents ‘ability to easily visit and accompany their multi-disabled child in a LTCI. The quality of the bond between parents and their child is difficult to predict. However, the possible absence or the difficulty in developing one, due to distance from the LTCI, or even the absence of a parent, seems to be a pejorative factor in HCP’s projection, of the child’s future situation. In this decision-making process, HCPs’ perceptions of the possibility of future parental involvement seem to be oversimplified and translated into a stereotypical assumption: parents will take better care of their child because they live near the LTCI, and conversely. To limit these biases, multidisciplinary collegial meetings should allow reasoned discussion before a decision of WLST.

Conclusion: This study highlights the need to be aware of these biases and their impact, and to know how to identify them.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Simon Macdonald (simon.macdonald@umontreal.ca)

Annals of Intensive Care 2013, 13(Suppl 1):CO-35

Rationale: Organizing follow-up of Pediatric Intensive Care Unit (PICU) survivors is challenging. Many children have complex medical conditions that require multiple appointments. Scheduling and transportation constraints are common barriers to follow-up. Participation rates to PICU follow-up have been very variable with some studies reporting rates as low as 30%. We evaluated the participation rate of families to a PICU follow-up clinic offering flexible schedules, dedicated research personnel and the option to choose in-person versus telemedicine follow-ups.

Patients and methods/materials and methods: This is a retrospective descriptive study of the feasibility of a post-PICU clinic in a pediatric tertiary care university hospital in the year 2021 (post-pandemic). Inclusion criteria to the clinic were: age < 18years at PICU admission, PICU hospitalization > 96h, AND invasive ventilation > 48h OR non-invasive ventilation > 96h. Follow-up is planned 2months post-PICU discharge with a dedicated PICU provider. To optimize patient participation, the following strategies were put into place: (1) patient enrolment obtained prior to PICU discharge, (2) information on post-PICU clinic and post-PICU complications provided prior to discharge with an informative pamphlet, (3) dedicated research personnel responsible to track participants and organize follow-ups, (4) flexible days and schedule options offered for accommodation, including rescheduling, (5) six PICU providers available at various time periods, (6) option to choose between in person vs. telemedicine appointments, (7) visits planned at same time as other hospital appointments when possible, (8) research personnel meeting with families at hospital entrance to guide to appointment, (9) phone call reminders done 1week prior to appointment. A subgroup of children followed at the clinic underwent 15–60min neurocognitive evaluations. They were performed either in-person or virtually.

Results: From 01/2021 to 12/2021, 103/123 (83.8%) patients were enrolled. Most common reasons for non-enrollment were: (1) difficulties to reach parents, (2) no interest in additional follow-up, (3) distance from hospital, (4) language barrier. Of those enrolled, 82/103 (79.6%) patients were evaluated at 2-months follow-up. 23.2% of appointments were performed virtually. 13.6% of appointments needed to be rescheduled to accommodate parents. Of 23 children that were offered neurocognitive evaluations, 30.4% were performed in-person and 69.6% were performed virtually and all evaluations were completed.

Conclusion: Performing PICU follow-up clinic is feasible with a strong participation rate when strategies including a dedicated team, scheduling flexibility and virtual options are set into place.

Reference 1

Maddux A.B., Pinto N., Fink E.L., Hartman M.E., Nett S., Biagas K., Killien E.Y., Dervan L.A., Christie L.M., Luckett P.M., et al. Postdischarge Outcome Domains in Pediatric Critical Care and the Instruments Used to Evaluate Them: A Scoping Review. Crit.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Jean Baptiste Lascarrou (jeanbaptiste.lascarrou@chu-nantes.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-36

Rationale: Out-of-hospital cardiac arrest (OHCA) is a heterogeneous entity with multiple origins and various prognosis. An early and reliable assessment of the prognosis is useful for various reasons, in particular to adapt the therapeutic strategy and to inform relatives. Our aim was to evaluate a large panel of dedicated scores that predict outcome after OHCA in a prospective multicentric study.

Patients and methods/materials and methods: We prospectively collected data from 24 French ICUs between August 2020 and June 2022. All cases of non-traumatic OHCA (both cardiac and non-cardiac cause) patients with stable return of spontaneous circulation (ROSC) and comatose at ICU admission (defined by Glasgow coma score ≤ 8) were included. Primary outcome was the modified Rankin scale at day 90 after cardiac arrest assessed by phone interview. A large panel of developed scores (CAHP; OHCA; CREST; C-Graph; TTM; CAST; NULL-Please; MIRACLE2) were included in the analysis. Their accuracies in predicting poor outcome at 3months after OHCA were determined using the area under the receiving operating characteristic curve (AUROC) and calibration belt with Utstein style criteria as a reference.

Results: During the study period (22months), 907 patients were screened and 658 were included in the study. Patients were mainly male (72%), 61 ± 15years old and mostly collapsed from supposed cardiac cause (64%). Mortality rate at day 90 was 63% and an unfavorable neurological outcome was observed in 66%. Performance (AUROC) of Utstein criteria for poor outcome prediction was 0.79 [0.76–0.83] whereas AUROC from others score included varied from 0.79 [0.75–0.83] to 0.88 [0.86–0.91]. Considering their respective AUROC, the 3 most performant scores were the TTM(0.88 [0.86–0.91]), the CAHP (0.87 [0.84–0.90] and the mCAHP (0.86 [0.83–0.89]). At the opposite, the 3 less performant scores were the C-GRAPH (0.76 [0.71–0.80]), the CREST(0.79 [0.75–0.83]) and the NULL-PLEASE (0.81 [0.77–0.84]). For each score, the proportion of patients for whom individual value cannot be determinable varied from 1.4% to 17.4%.

Conclusion: In patients admitted to intensive care after a cardiac arrest, most of the scores available for the evaluation of the subsequent prognosis are more efficient than the usual Utstein criteria. Some of these scores revealed superior performance to others, even if the observed differences are modest. If a choice must be made among these different scores, the immediate availability of all the variables in the environment should guide this choice.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Romain Griffon (rgriffon@chu-reims.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-37

Rationale: Diastolic arterial pressure (DAP) is mainly determined by vascular tone and less by heart rate (HR) and arterial stiffness. Consequently, in septic shock patients, vasoplegia may be diagnosed simply by a low value of DAP. However, the doses of vasopressors (VP) should be taken into account when correlating values of DAP to the degree of vasoplegia. Due to the lack of clinical data, DAP tends to be neglected and underused by the clinicians. We hypothesized that if the DAP/VP dose ratio is a better indicator of the degree of vasoplegia than DAP alone, it would be more associated with mortality than DAP alone.

Patients and methods/materials and methods: This is a post-hoc analysis of ANDROMEDA-SHOCK trial, a multicenter randomized controlled trial conducted between March 2017 to April 2018 in 28 hospitals in 5 countries. All patients included were admitted in intensive care unit (ICU) for septic shock. Systolic arterial pressure (SAP) and DAP were directly measured using an arterial catheter, which also displayed the mean arterial pressure (MAP) by calculating it from the SAP and DAP values. We calculated the DAP/VP dose ratio at inclusion time and compared it between survivors and non-survivors. We also calculated the SAP/VP dose ratio. A ROC curve was performed to define the statistical performance of the DAP/VP dose ratio and compared it to ROC curves constructed for DAP/HR and DAP alone.

Results: A total of 424 patients were included in the analysis. Median age was 66 [52–76;IQR]. Median SOFA score and SAPSII at baseline were 10 [7–12;IQR] and 21 [17–28;IQR], respectively. Median ICU length of stay was 6 [3–12;IQR]. In-hospital mortality rate was 43%. There was a significant difference between survivors and non survivors for DAP/VP dose (426 ± 24 in the survivors group and 283 ± 22 in the non-survivors group; p = 0.001), but not for the SAP/VP dose (800 ± 44 in the survivors group and 542 ± 44 in the non-survivors group; p = 0.055) (Figure 1). The AUC of DAP/VP ratio was 0.661 (Se = 54, Sp = 73; p = 0.001) for predicting in-hospital mortality and was significantly higher than the AUC of DAP/HR (p = 0.0002) and DAP alone (p = 0.0002) (Figure 1).

Conclusion: The DAP/VP dose ratio calculated at the early phase of septic shock is probably a better indicator of the degree of vasoplegia than DAP alone as it is better associated with poor outcome than DAP alone.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Grégoire Muller (muller.gregoire.chro@protonmail.com)

Annals of Intensive Care 2013, 13(1): CO-38

Rationale: To determine whether current management of severe acute pulmonary oedema (APE) is in line with international guidelines and differs according to medical specialty and recognized experts in the field, we designed a survey withfour case vignettes presenting different phenotypes of APE [1].

Patients and methods/materials and methods: Four clinical cases vignettes were designed by a scientific committee representing the scientific societies of emergency medicine, cardiology, and intensive care. In addition, 20 experts were enrolled according to their publications and involvement in the scientific societies. Target participants and designated experts were French physicians in cardiology, intensive care medicine and emergency medicine. From 06/2022 to 09/2022, the vignettes were submitted to the physicians and experts using an open online survey.

Results: Among the 1048 respondents 60% emergency physicians, 22% intensivists and 18% cardiologists), 781 completed the 4 cases whereas the 20 experts completed the whole survey.

• SpO₂ < 90% threshold for O₂ administration (ESC Guidelines I-C recommendation [1]): only 22% of physicians acted according to guidelines (SpO₂ < 90%), 47% administered O₂ for SpO₂ < 95%, and 31% gave O₂ in case of respiratory distress. The results were homogeneous according to physician specialty. Among experts, 53% followed the guidelines.

• Non-invasive ventilation (NIV) for respiratory distress (Respiratory rate > 25/mn, SpO₂ < 90%) (IIa-B recommendation): 54% used NIV because of respiratory rate > 30/mn, 83% because of respiratory distress signs, 43% according to oxygen flow rate, and 42% in case of past COPD history. Except for past COPD history (cardiologists 32%, intensivists 52%, emergency physicians 41%), results were homogeneous according to physician specialty. Among the experts, the respective figures were 93%, 87%, 67%, 47%. BiPAP was chosen in 66% (respectively 55%, 84%, 61%, 73%, in cardiologists, intensivists, emergency physicians, experts), whereas CPAP in 34%.

• Intubation for progressive respiratory failure persisting despite O₂ or NIV (I-C recommendation): in case of pre-hospital NSTEMI worsening despite NIV and medications, only 55% used mechanical ventilation (respectively 65%, 73%, 43%, 85% in cardiologists, intensivists, emergency physicians, experts). In case of STEMI worsening while waiting for coronary angiography, 44% used mechanical ventilation (respectively 45%, 70%, 33%, 70% in cardiologists, intensivists, emergency physician, experts).

Conclusion: Ventilation strategies regarding classical clinical cases of severe APE remain highly variable, and variations exist among medical specialties. The low level of evidence of the ESC Guidelines may explain these heterogeneities [2]. Further research is needed to fill the current gap in evidence.

Reference 1

McDonagh TA, Metra M, Adamo M, Gardner RS, Baumbach A, Böhm M, et al. 2021 ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure. Eur Heart J. 2021;42:3599–726.

Reference 2

Aissaoui N, Hamzaoui O, Price S. Ten questions ICU specialists should address when managing cardiogenic acute pulmonary oedema. Intensive Care Med. 2022;48:482–5.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Florent Huang (florenth27@hotmail.com)

Annals of Intensive Care 2013, 13(1): CO-39

Rationale: While endomyocardial biopsy (EMB) is recommended in adult patients with fulminant myocarditis (FM), the clinical impact of its timing is still unclear.

Patients and methods/materials and methods: Data were collected from 419 adult patients with clinically suspected FM admitted to intensive care units (ICUs) across 36 tertiary centers in 15 countries worldwide. Histology or cardiac magnetic resonance imaging confirmed the diagnosis in 306 (73%) cases. The primary outcome of survival free of heart transplantation (HTx) or left ventricular assist device (LVAD) at one year was specifically compared between patients with early EMB (within two days after ICU admission, n = 103) and delayed EMB (n = 80). A propensity score-weighted analysis was done to control for confounders.

Results: Median age on admission was 40 [29–52] years and 322 (77%) patients received temporary mechanical circulatory support. 273 (65%) patients survived without HTx/LVAD. The primary outcome was significantly different between patients with early and delayed EMB (70% vs. 49%, p = 0.004). After propensity score weighting, the early EMB group still significantly differed from the delayed EMB group in terms of survival free of HTx/LVAD (63% vs. 40%, p = 0.021). Moreover, early EMB was independently associated with a lower rate of death or HTx/LVAD at one year (odds ratio of 0.44; 95% confidence of interval: 0.22–0.86; p = 0.016).

Conclusion: EMB should be broadly and promptly used in patients admitted to ICU for clinically suspected FM.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Lisa Raia (lr.lisa.raia@gmail.com)

Annals of Intensive Care 2013, 13(1): CO-40

Rationale: Intravenous fluid therapy is a daily concern in the management of ICU patients, especially in patients with acute circulatory failure. Several studies reported the consistent association between positive fluid balance and outcome. However, the assessment of intravenous fluids is most often restricted to fluids given for resuscitation fluids, whereas the contribution of alternative fluid infusions to fluid balance has been poorly estimated. We herein aimed at quantifying the whole determinants and trends in fluid balance in septic shock, as well as the impact on outcome.

Patients and methods/materials and methods: We conducted a retrospective study in a medical ICU over a 13-year period (2008–2020). Adult patients with septic shock fulfilling the Sepsis-3 definition within 48h of ICU admission were included. Three periods of admission (2008–2011, 2012–2015 and 2016–2020) were predefined. The components of fluid balance were extracted from the patient data management system from admission to day-7. We also assessed the impact of fluid balance on durations under organ supports.

Results: Among 1756 patients admitted to the ICU for septic shock, 954 patients were included. Pneumonia was the first source of infection (42.3%), and 793 (83.2%) patients required invasive mechanical ventilation. ICU and hospital mortality rates were 38.9% and 46.6% respectively. From their admission to day7, patients received a median of 1.8 [1.0–3.3] L of intravenous fluids per day. Fluid intake was 5.3 [3.6–7.5] L at day-1, and then rapidly decreased from day-2 and thereafter. Resuscitation fluids accounted for an important proportion (45.2%) of intravenous intake on day-1 only. From day-2, maintenance and vehicle fluids were mostly responsible for daily fluid intake, and therefore for fluid balance. There was a significant inverse relationship between the cumulative fluid balance within the first three days and the number of days alive without organ support (invasive mechanical ventilation and vasopressors). Patients admitted after 2011 received significantly less fluid between day-2 and day-6, associated with increased use in albumin during the initial resuscitation phase.

Conclusion: Fluid balance in septic shock is mainly determined by maintenance and vehicle fluids from day-2, whereas the overall contribution of resuscitation fluids is limited. These data suggest that targeted interventions based on non-resuscitation fluids are likely to result in major impact on fluid balance.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Philippe Vignon (philippe.vignon@unilim.fr)

Annals of Intensive Care 2013, 13(1): CO-41

Rationale: To evaluate the temporal course of right ventricular (RV) injury during the ICU stay of patients with COVID-19 acute respiratory distress syndrome (ARDS) and its association with mortality, according to currently proposed definitions.

Patients and methods/materials and methods: This is a post-hoc analysis of longitudinal data collected in the multicenter ECHO-COVID observational study in patients who underwent at least two echocardiography assessments during ICU stay. Screened RV injuries were acute cor pulmonale (ACP: RV cavity dilatation and paradoxical septal motion), RV failure (RVF: RV cavity dilatation and systemic venous congestion), and RV dysfunction (tricuspid annular plane systolic excursion ≤ 16mm).

Results: Of 281 patients who underwent 948 echocardiography studies during ICU stay, 189 (67%) were diagnosed with at least one type of RV injury during one or several examinations: ACP (37.4%), RVF (54.7%) and/or RV dysfunction (29.0%). Keeping the other RV injuries constant, patients with all examinations displaying ACP had shortened survival time by 0.479 times [0.284–0.803] (P = 0.005) when compared to patients with all examinations depicting no ACP. RVF showed a trend towards shortened survival time (0.642 [0.405–1.018], P = 0.059), whereas RV dysfunction had no significant effect (0.0834 [0.521–1.336], P = 0.451).

Conclusion: RV injury is prevalent in patients ventilated for COVID-19 ARDS. Unlike RV dysfunction, the development of ACP during ICU stay was associated with an increased risk of death. RVF showed a trend towards reduced survival time but was not significantly associated with ICU mortality. Determining consensus definitions of different types of RV injury is warranted since their impact on outcome appears variable.

Compliance with ethics regulations: N/A.

Correspondence: Jeremie Joffre (jeremie.joffre@aphp.fr)

Annals of Intensive Care 2013, 13(1): CO-42

Rationale: Suspected upper-gastrointestinal bleeding (SUGIB) is a common issue for critically ill patients during ICU stay. In the absence of specific guidelines regarding the indication and timing of oesogastroduodenoscopy (EGD), there is substantial variability in EGD practice policy between ICUs, depending partly on EGD accessibility. This study aimed to investigate factors associated with the need for per-EGD hemostatic therapy and to develop a point-based score predicting the therapeutic value of emergency bedside EGD in ICU patients presenting a SUGIB.

Patients and methods/materials and methods: We conducted a retrospective cohort study in our university hospital ICU, a high-volume and reference center for acute GI hemorrhage in Paris (France), with 24/7 access to EGD. Two hundred fifty-five patients, not primarily admitted for GI bleeding, who underwent a bedside EGD for SUGIB during their ICU stay were analyzed. Clinicals, biologicals, EGD-related data, and outcomes were collected. We followed the TRIPOD Checklist to develop a prediction model relying on a simple point-based score.

Results: Amongst the 255 patients who underwent EGD for SUGIB during their ICU stay, the preeminent EGD indication was anemia (79%), melena (19%), shock (14%), and hematemesis (13%). EGD was normal in 24.3%, and the primary lesions reported were ulcers (23.1%), esophagitis (18.8%), and gastritis (12.5%). Only 12.9% of patients underwent hemostatic endotherapy during EGD. Based on predictive factors of the need for hemostasis during EGD, we developed the SUspected GIB in Icu score (SUGIBI): Male gender, smoking, and history of cirrhosis were attributed one point. RRT, shock and/or hematemesis indicating EGD were attributed two points. No external bleeding was associated with minus one point. A SUGIBI score < 4 has a negative predictive value of 97% (0.94 to 0.99) for hemostatic endotherapy and a positive predictive value of 53% (0.39 to 0.66) if ≥ 4 (AUC of the model: 0.82; 0.73–0.9 (P < 0.0001)).

Conclusion: Although bedside emergency EGD for SUGIB has an indisputable diagnostic value in critically ill patients, hemostatic therapy is performed in a minority of patients. We propose the SUGIBI score based on simple criteria to help the clinician stratify the probability of a therapeutic EGD. In case of a SUGIBI score < 4, we advocate that EGD might safely be postponed or reconsidered, especially if low accessibility or frail patients are at risk of EGD complications. Conversely, if the SUGIBI score > 4 EGD must be performed without delay, given the high probability of hemostatic endotherapy. This score requires prospective external validation.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Maxime Gasperment (maxime.gasperment@gmail.com)

Annals of Intensive Care 2013, 13(1): CO-43

Rationale: Alcoholic hepatitis (AH) is a clinical syndrome responsible for high morbidity and mortality, but little is known about the management of AH patients admitted in Intensive Care Units (ICUs). Our study aimed to describe the clinical and biological characteristics of these patients and to analyze mortality risk factors.

Patients and methods/materials and methods: We conducted a retrospective study of all AH patients hospitalized in 9 French ICUs (Jan 2006–Dec 2017). AH was defined histologically or by an association of clinical and biological criteria according to current guidelines. Data collected included usual clinical and biological features, overall and disease severity, organ failures and patient management.

Results: 187 patients (median age: 53 [43;60]; male: 69%) were included. Median Child–Pugh score at admission was 12 [11;13]. Patients were admitted for impaired consciousness (71%), sepsis (64%), shock (44%), respiratory failure (37%) or hemorrhage (8%). At admission, median SOFA, SAPS II and MELD scores were 10 [7;13], 52 [39;74] and 31 [26;40] respectively. During ICU stay, main liver-related complications were encephalopathy (76%), ascites (67%) and hepato-renal syndrome (60%). 62% of patients received vasopressors, 63% invasive mechanical ventilation and 36% renal replacement therapy. 66% received steroids, and liver transplantation was performed in 16 patients (8.5%). ICU death was 37%, in-hospital death 53%. In univariate analysis, ICU mortality was associated with unemployment (p = 0.034), and admission respiratory failure (p < 0.001), shock (p < 0.001), sepsis (p < 0.001), acute kidney injury (p < 0.001), lower Glasgow score (p = 0.006) and higher SOFA (p < 0.001), SAPS II (p < 0.001) and MELD scores (p < 0.001). Need for supportive therapy (vasopressors (p < 0.001), ventilation (p < 0.001), renal (p < 0.001)) was associated with worse outcomes, as well as hepato-renal syndrome (p < 0.001) and encephalopathy (p = 0.002). In multivariate analysis, ICU mortality was associated with admission SOFA score (HR 1.08 [1.02;1.14]), arterial lactate (HR 1.08 [1.03;1.13]) and MELD score (HR 1.09 [1.04;1.14]), while employment was a protective factor (HR 0.49 [0.28;0.86]). Overall, use of steroids did not affect ICU mortality (HR 0.96 [0.48;1.91]); however, significant differences were found between steroids responders and non-responders as assessed by Lille score at day 7 (Lille score > 0.45: OR 6.67 [2.44;20.15], p < 0.001).

Conclusion: Alcoholic hepatitis is a severe condition leading to high mortality in ICU patients. Severity of organ failure at admission are mortality risk factors, as well as hepato-renal syndrome during ICU stay. Outcome was significantly worse in non-responders to steroids therapy.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Thomas Frapard (thomas.frapard@aphp.fr)

Annals of Intensive Care 2013, 13(1): CO-44

Rationale: Acute liver failure is a rare but life-threatening critical illness. Among its causes, one of the least reported is infection by Herpes Simplex Virus (HSV). This aetiology is evoked in case of febrile hepatic insufficiency with cutaneous signs, in particular in immunocompromised patients. Previous small case series have reported marked transaminase elevation and high mortality. The aim of the current study is to describe the clinico-biological characteristics and prognosis of patients admitted to an intensive care unit for acute hepatitis and proven HSV infection.

Patients and methods/materials and methods: This is a retrospective multicenter study conducted in 14 intensive care units in France. All patients hospitalized in participating intensive care units between 2006 and 2022 with a compatible history and a proven HSV infection (PCR blood or tissue) were included.

Results: A total of 29 patients were included. The diagnosis was confirmed by histology for 14 (48.3%) patients. There were 3 (10.3%) pregnant patients and 21 (72.4%) immunocompromised patients. Upon ICU admission 27 (93.1%) patients were febrile with a median temperature of 38.7°C [38.4–39.4], a skin rash was noted in 11 (37.9%) patients, the median SOFA score was 15 [11–17], the median platelet counts was 50 G/L [18–82], the median prothrombin rate (PT) was 35% [28–50], and the median ASAT level was 2497IU/L [765.75–6304]. The median time interval between the onset of symptoms and the initiation of acyclovir was 8.5days [5–11]. During ICU stay, 25 (86.2%) patients developed hepatic encephalopathy, 3 (10.3%) developed HSV encephalitis, 4 (13.8%) myocarditis, 10 (34.5%) haemorrhagic shock, 12 (41.4%) disseminated intravascular coagulation (DIC) and 16 (55.2%) a hemophagocytic syndrome. 21 (72.4%) of the patients died in intensive care unit. A liver transplantation was attempted in 4 (13.8%) patients, but the procedure was unsuccessful in all cases (100% mortality). Factors associated with mortality in univariable logistic regression were SAPS II and SOFA scores, temperature, confusion, creatinine at ICU admission, and hepatic encephalopathy, the occurrence of DIC and the nadir of PT during ICU stay.

Conclusion: Herpetic hepatitis in the ICU is an extremely severe condition associated with a very high mortality. Despite previously known risk factors (immunosuppression and pregnancy) and a clinico-biological picture associating a febrile hepatitis with a predominance of ASAT, skin rash and cytopenia, the diagnosis remains difficult with a median delay of 8days until the introduction of aciclovir. The benefit/risk ratio of early initiation of aciclovir in acute febrile hepatitis seems very favourable.

Reference 1

Little L, Rule J, Peng L, Gottfried M, Lee WM, Acute Liver Failure Study Group. Herpes Simplex Virus-Associated Acute Liver Failure Often Goes Unrecognized. Hepatology. févr 2019;69(2):917?9.

Reference 2

Ichai P, Roque Afonso AM, Sebagh M, Gonzalez ME, Codés L, Azoulay D, et al. Herpes simplex virus-associated acute liver failure: a difficult diagnosis with a poor prognosis. Liver Transpl. déc 2005;11(12):1550?5.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Marie Le Goff (legoffmarie1@gmail.com)

Annals of Intensive Care 2013, 13(1): CO-45

Rationale: Liver abscess (LA) is a rare and potentially life-threatening disease. There is no available data on the epidemiology and prognosis of LA patients requiring intensive care unit (ICU) admission in France.

Patients and methods/materials and methods: We conducted a retrospective multicenter study including adults admitted in 27 French intensive care units with a diagnosis of LA from 2010 to 2020. Data are presented as number (%) or median (interquartile range). Risk factors for ICU mortality were identified using univariate analysis.

Results: 400 patients met inclusion criteria. In this preliminary report we present the data of 102 patients, included in 9 centers. Patients’ age was 66years (55–73) and 73% were men. The main causes of admission were sepsis (90%) and respiratory failure (9%). At ICU admission, IGS2 score was 39 (29–53) and SOFA score was 5 [2–9]. The putative origin of LA was biliary (25%), portal (13%), due to pre-existing lesions (16%), arterial (7%) or cryptogenetic (30%). Fifty-eight percent of patients had a solitary LA, the rest having 2 (2–2.25) lesions. LA involved the right part of the liver in 51% of cases, the left in 32% or both in 17%. Larger abscess size was 60mm [32–90]. Eighty-seven percent of LA were microbiologically documented, with the isolation of gram-negative bacilli (68%), gram-positive cocci (34%) and/or other germs (19%) (details in Figure 1). Multiple pathogens identification occurred in 33 (37%) of patients. Fine needle aspiration of LA was performed in 69 (68%) patients and drain insertion in 47 (46%). Duration of antibiotherapy was 28days [18–42]. During ICU stay, 24 (24%) patients developed septic shock, 20 (20%) renal failure and 16 (16%) respiratory failure. In-ICU mortality was 11,7%.Factors associated with mortality were organ dysfunctions, illustrated by SOFA score (OR1.184 [1.065–1.316] IC95%, p = 0.0018) and lack of microbial documentation (OR5.347 [1.27–22.489] IC95%, p = 0.022). Abscess evacuation by puncture was associated with better short-term survival (OR4.375 [1.18–16.216] IC95%, p = 0.0272) with no additional benefit of drainage (OR0.2887 [0.034–2.745] IC95%, p = 0.2887).

Conclusion: In this multicentric study we described the characteristics and short-term outcome of severe LA in France. Based on our preliminary results, we will analyse the full cohort of 400 patients looking for parameters associated with improved short- and long-term survival.

Reference 1

1. Chen W, Chen C-H, Chiu K-L, et al. (2008) Clinical outcome and prognostic factors of patients with pyogenic liver abscess requiring intensive care: Critical Care Medicine 36:1184–1188.

Reference 2

2. Kaplan G, Gregson D, Laupland K (2004) Population-based study of the epidemiology of and the risk factors for pyogenic liver abscess. Clinical Gastroenterology and Hepatology 2:1032–1038

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Thibault Vieille (thibault.vieille91@gmail.com)

Annals of Intensive Care 2013, 13(1): CO-46

Rationale: Abdominal compartment syndrome (ACS) may occur during severe acute pancreatitis (SAP) and is associated with a poor outcome.Guidelines provide to perform invasive decompression in patients with a sustained intra-abdominal pressure > 25mmHg with new onset organ failure refractory to medical therapy. Nevertheless, current criteria triggering surgical decompression are probably too selective, leading to delayed treatment. The primary objective of this retrospective study was to determine predictors associated with 28-day mortality in patients requiring laparostomy for ACS during SAP. Secondary objectives were to evaluate the benefits of surgical decompression on organ failure improvement, and the prevalence of acute mesenteric ischemia.

Patients and methods/materials and methods: Monocenter, retrospective, observational study including all adult patients admitted to the ICU for SAP and requiring surgical decompression for abdominal compartment syndrome between January 2006 and March 2022. Comparison of the patients according to the vital status at 28-day.

Results: 24 patients with SAP and laparostomy for ACS were included. 28-day mortality rate was 38%. Decompressive laparostomy did not improve organ failure (SOFA score, norepinephrine dose, arterial pH, urine output and Pa02/Fi02 ratio), despite a significant decrease of intra-abdominal pressure. At the surgical decompression time, arterial pH, SOFA score, norepinephrine dose, and urinary output, were significantly associated with 28day-mortality. We identified 4 criteria associated with a poor evolution when present in the 24h before laparostomy: arterial pH < 7.25, arterial lactate level > 5mmol/L, urine output (/24h) < 0.5 L/day, and norepinephrine > 0.5mcg/kg/min. These criteria were suggestive of the development of non-occlusive mesenteric ischemia (NOMI). Indeed, 2/3 of the non-survivors developed acute mesenteric ischemia.

Conclusion: We suggest that during severe acute pancreatitis, practitioners should focus not only on the level of intra-abdominal pressure to consider decompressive laparostomy but should also consider the possibility of NOMI. Indeed, NOMI may develop early during SAP, before the threshold of 25mmHg of intra-abdominal pressure has been reached. Suspicion of NOMI during severe AP should lead to early decompressive laparotomy, even if the criteria of ACS are not reached, in order to improve overall survival.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Jean-Baptiste Peretout (jperetout@cognacq-jay.fr)

Annals of Intensive Care 2013, 13(1): CO-47

Rationale: A recent systematic review revealed the strikingly high prevalence of malnutrition in intensive care unit (ICU) patients (ranged from 38 to 78%)¹. Enteral feeding through a nasogastric tube is the most common form of nutrition in ICU. Prolonged immobilization and neuromyopathy also lead to malnutrition and sarcopenia². Reinforced multidisciplinary management in the weaning unit makes it possible to monitor and adapt their nutritional needs and facilitate their reautonomization. The objective of this study is to determine the proportion of patients suffering from malnutrition, their clinico-biological characteristics and the effect of rehabilitation on their renutrition capacities.

Patients and methods/materials and methods: This is a prospective, descriptive cohort study of patients hospitalized in respiratory weaning unit (WU) after a prolonged ICU stay over the year 2022. All patients hospitalized in the WU were included. Demographic characteristics and nutritional characteristics were collected. All inpatients with complete data were analyzed. All patients experienced active dietetic and speech therapy management.

Results: In 2022, 103 patients were admitted in our unit and a total of 97 patients were analyzed. All patients hospitalized in WU were tracheostomized. The main diagnosis at admission in ICU was acute respiratory failure in 58,7%. Characteristics at admission to WU are presented in Table I. 60% of the patients presented malnutrition: moderate 20% and severe 40%. All biological criteria for malnutrition were abnormal: hypoalbumenia, high CRP and low 25(OH)D. At admission, almost all patients had exclusive enteral feeding (92.8%): 74.2% by nasogastric tube and 18,6% by gastrostomy tube. At discharge from WU, 85.6% of patients had resumed oral feeding compared to admission 7.2% (p < 0,001) and only 25.8% of patients still had enteral feeding compared to admission 92.8% (p < 0,001).

Conclusion: After a stay in WU, a vast majority of hospitalized patients were able to resume exclusive oral feeding despite a significant proportion of patients with malnutrition. The dietetic management of these patients remains essential and is fully integrated in the multidisciplinary management of patients in WU.

Reference 1

1. Lew CCH, Yandell R, Fraser RJL, Chua AP, Chong MFF, Miller M. Association between malnutrition and clinical outcomes in the intensive care unit: a systematic review. JPEN J Parenter Enteral Nutr. 2017;41(5):744–58. https://doi.org/10.1177/0148607115625638

Reference 2

2. Hill A, Elke G, Weimann A. Nutrition in the Intensive Care Unit-A Narrative Review. Nutrients. 2021 Aug 19;13(8):2851. https://doi.org/10.3390/nu13082851

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Alexandre Coppens (alex95630@yahoo.fr)

Annals of Intensive Care 2013, 13(1): CO-48

Rationale: Electrical impedance tomography (EIT) allows for determining “optimal” positive end-expiratory pressure (PEEP) at the bedside by identifying the PEEP level with minimal lung overdistension and collapse [1]. On the other hand, the recruitment-to-inflation ratio (R/I) has been developed to evaluate the potential for lung recruitment in patients with ARDS but has never been studied in patients on ECMO, frequently ventilated with very low tidal volume. The aim of this study was to assess the accuracy of the R/I and to compare its performance with EIT-based optimal PEEP in ECMO-treated patients ventilated with low tidal volume.

Patients and methods/materials and methods: Intubated patients with severe ARDS on venovenous ECMO with a tidal volume > 2ml/kg were included for 9months. First, a low-flow insufflation was performed to detect airway opening closure. EIT-derived optimal PEEP was determined during a decremental PEEP trial (20 to 6 cmH2O). Similarly, R/I was calculated between a PEEP of 15 to 5 cmH₂O or the airway opening pressure. The whole cohort was randomly split into two groups. In the first one (regression cohort), the correlation between EIT “optimal” PEEP and R/I was assessed using a mixed model for repeated measures whereas the agreement between the two methods was evaluated by Cohen's kappa method in the second group (validation cohort). If possible, similar measurements were repeated on Day 7 of ECMO.

Results: A total of 54 measurements were obtained in 31 patients (median age 47years, tidal volume 4.7 [range: 3.0–5.8] mL/kg) after a median of 3days on ECMO. An airway opening pressure of 13.5 [10–14.25] cmH₂O was found in 6 patients. In the regression cohort (22 patients and 39 measurements), EIT-derived “optimal” PEEP was correlated with R/I with a correlation coefficient of 0.038 (95% IC, 0.021 to 0.055), p < 0.001 (Figure 1). Based on the regression equation, a R/I of 0.47 corresponds to an optimal PEEP of 11 cmH2O. In the validation cohort (8 patients with 13 measurements), R/I showed moderate agreement with EIT “optimal” PEEP (K = 0.70) to classify recruiter and non-recruiter.

Conclusion: In patients on ECMO ventilated with very low tidal volume, R/I was correlated with “optimal” PEEP determined by EIT. However, beyond identifying the potential for lung recruitment in that population, R/I value had a moderate agreement with EIT optimal PEEP.

Reference 1

Franchineau, G., Bréchot, N., Lebreton, G., Hekimian, G., Nieszkowska, A., Trouillet, J.-L., Leprince, P., Chastre, J., Luyt, C.-E., Combes, A., Schmidt, M., 2017. Bedside Contribution of Electrical Impedance Tomography to Setting Positive End-Expiratory.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Sebastien Clerc (seb.clerc23@gmail.com)

Annals of Intensive Care 2013, 13(1): CO-49

Rationale: The effects of spontaneous breathing efforts during de novo acute respiratory failure (ARF) are debated. While avoiding invasive mechanical ventilation is an important goal, excessive inspiratory efforts and dyspnea may be harmful and associated with worst prognosis, i.e. intubation. This bicentric study sought to investigate dyspnea and diaphragm activity of patients presenting with de novo ARF according to the risk of intubation.

Patients and methods/materials and methods: Patients aged of more than 18years old, hospitalized in the intensive care unit for de novo ARF (respiratory rate > 25/min and hypoxemia defined by PaO2/FiO2 ratio < 300mHg with face mask oxygen ≥ 10L/min O2 or high flow nasal oxygen with FiO2 ≥ 50%) were included. Dyspnea was evaluated with a visual analogic scale (VAS) at inclusion and at H2 and H4. Right ultrasound evaluation of diaphragmatic activity (excursion and thickening fraction) was also performed at inclusion and at H2 and H4. Primary outcome was the intubation rate censored at 7days after enrolment.

Results: Forty-nine patients were included over an 18months period. Patients were mostly men (69%) with a median age of 66years. Main causes of ARF were COVID-19 (49%) and post-operative ARF (32%). Thirty patients (61%) were treated with high flow nasal oxygen and in total 13 (27%) were intubated. Median (25–75 IQR) dyspnea was 21mm (10–40) at inclusion. Dyspnea VAS was 33mm vs 15mm (p = 0.11) at inclusion, 41mm vs 21mm (p = 0.10) at H2 and 31mm vs 16mm at H4 (p = 0.07), in intubated and non-intubated patientsrespectively. The thickening fraction at inclusion was significantly lower in intubated patients (11% vs 30%; p = 0.003) and H4 (8% vs 24%; p = 0.027). The difference was not significative at H2. The diaphragm excursion was 14mm vs 13mm (p = 0.69) at inclusion, 15mm vs 13mm (p = 0.94) and 12mm vs 13mm (p = 0.65), in intubated and non-intubated patients respectively.

Conclusion: These preliminary findings show that dyspnea and lower diaphragm thickening fraction might be associated with the risk of intubation in de novo ARF.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Mélodie Parfait (melodie.parfait@gmail.com)

Annals of Intensive Care 2013, 13(1): CO-50

Rationale: While being life-saving,positive-pressureventilation (PPV) is associated with harmful effects that carry morbidity and mortality in patients with acute respiratory distress syndrome (ARDS). Reducing the amount of pressure delivered by the ventilator inside the thorax may reduce the development of ventilator-induced lung injury and drop in cardiac output. Diaphragm neurostimulation-assisted ventilation (DiSAV) could overcome these issues.

Patients and methods/materials and methods: Patients with moderate ARDS were included after 48h of invasive mechanical ventilation and were equipped with a left subclavian dedicated catheter which was used to deliver phrenic nerve stimulation. Diaphragm neurostimulation was delivered continually, in synchrony with volume-controlled mode ventilation (DiSAV) during two 60-min sessions, interspersed with two 60-min washout sessions (PPV, no stimulation). Lung mechanics were continuously monitored with esophageal and gastric pressures. Regional lung ventilation was continuously monitored with electrical impedance tomography. Cardiac index was continuously monitored with a transpulmonary thermodilution system. Plateau pressure, driving pressure, transdiaphragmatic pressure, dorsal/ventral ventilation surface ratio and cardiac index were compared between DiSAV and PPV.

Results: Thirteen patients were included. Phrenic stimulation was associated with an increase in transdiaphragmatic pressure to 10 cmH2O [8–11] while it was 2 cmH2O [1–4] during PPV (p = 0.003). During stimulation sessions driving pressure was 11 cmH2O while it was 14 cmH2O without stimulation (p < 0.001). Lung compliance improved with a median gain of 35% (p < 0.001) during stimulation sessions. There was also an improvement in ventilation distribution to dorsal areas during DiSAV. Finally, the cardiac index went from 2.7 L/min/m2 [2.3–3.5] without stimulation to 3.3 L/min/m2 [2.5–3.9] during stimulation sessions (p = 0.14).

Conclusion: This pilot study shows that transvenous phrenic stimulation during ARDS 1) is feasible, 2) is associated with an improvement in ventilatory mechanics (increase in compliance) and more homogeneous regional ventilation distribution, suggesting a more protective mechanical ventilation.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Bruno Evrard (bruno.evrard@chu-limoges.fr)

Annals of Intensive Care 2013, 13(1): CO-51

Rationale: Previous studies of patients with Acute Respiratory Distress Syndrome (ARDS) and sepsis have identified two phenotypes: the more prevalent but less well understood hypoinflammatory phenotype, and the hyperinflammatory phenotype which is characterized by elevated levels of inflammatory biomarkers. We aimed (i) to mathematically estimate the attributable mortality from ARDS in hypo-inflammatory and hyper-inflammatory sepsis; and (ii) to determine via chart review the primary cause of death within each phenotype.

Patients and methods/materials and methods: We studied 1816 septic patients from two prospectively enrolled critically ill adult cohorts in United States of America (EARLI and VALID studies) who were previously assigned to the hyperinflammatory or hypoinflammatory phenotype using Latent Class Analysis(1). Population attributable fraction (AF_ARDS) of mortality from hypoinflammatory and hyperinflammatory ARDS were mathematically estimated adjusting for illness severity using APACHE II stratification(2), with the comparator groups being hypoinflammatory and hyperinflammatory sepsis, respectively. Organ dysfunction, severe comorbidities, and withdrawal of life support were collected from the medical record when available on patients who died from the EARLI cohort (n = 129/180) by personnel blinded to ARDS status and phenotype, using a standardized instrument (Stapleton RD, CHEST 2005). The primary cause of death was then defined as the organ system that most directly contributed to death or withdrawal of life support.

Results: Among 1253 (69%) hypoinflammatory patients, 443 (35%) developed ARDS, while among the 563 (31%) hyperinflammatory patients, 258 (46%) developed ARDS. In the hypoinflammatory phenotype, theAF_ARDSwas 19% (95%CI: 10–27%). In the hyperinflammatory group, theAF_ARDSwas 13% (95%CI: 5–21%). In the 129 patients who died with medical records available for review, 34% died without any severe underlying comorbidities in the hyperinflammatory phenotype, versus 20% in the hypoinflammatory phenotype (p = 0.01) (Figure). Cause of death as determined by review of the medical records differed between the two phenotypes (p < 0.001). Respiratory failure (including failure to wean from mechanical ventilation) was the most common cause of death in hypoinflammatory ARDS (63%), whereas circulatory failure was the most common cause of death (65%) in hyperinflammatory ARDS (Figure). These proportions were similar in patients with hypoinflammatory and hyperinflammatory sepsis.

Conclusion: Attributable mortality of ARDS remains higher in the hypoinflammatory phenotype of ARDS, with a high proportion of patients dying primarily from respiratory failure. Conversely, attributable mortality of ARDS was lower in the hyperinflammatory phenotype, with more patients dying from circulatory failure. Further studies of the hypoinflammatory phenotype are needed to determine the factors associated with mortality.

Reference 1

Sinha P, He J, Delucchi K, Zhuo H, Abbott J, Jones C, et al. Latent Class Analysis-Derived Hypoinflammatory and Hyperinflammatory Phenotypes Are Generalisable to Sepsis Patients Requiring Intensive Care. B95 ARDS: WHAT’S THE LATEST AND GREATEST? American.

Reference 2

Auriemma CL, Zhuo H, Delucchi K, Deiss T, Liu T, Jauregui A, et al. Acute respiratory distress syndrome-attributable mortality in critically ill patients with sepsis. Intensive Care Med 2020;46:1222–1231.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Maïa Joos (mjoos@hotmail.fr)

Annals of Intensive Care 2013, 13(1): CO-52

Rationale: According to the Berlin definition of acute respiratory distress syndrome (ARDS), bilateral infiltrates can be diagnosed using chest X-ray or computed tomography (CT)-scan. However, the interindividual agreement for the diagnostic of lung infiltrates on chest X-ray is poor and may explain the low rate of ARDS recognition by clinicians. Indeed, as compared to CT-scan, the accuracy of chest X-ray interpreted by a radiologist to identify ARDS was limited. However, in daily practice, chest X-rays are interpreted at the bedside by intensivists, not radiologists. We hypothesized that the agreement between chest X-ray interpreted by an intensivist and CT-scan interpreted by a radiologist, as reference standard for ARDS diagnosis, was poor.

Patients and methods/materials and methods: From January 2015 to December 2021, we included all CT-scan performed in patients invasively ventilated with PaO₂/FiO₂ratio ≤ 300mmHg with positive end-expiratory pressure ≥ 5 cmH₂O on the day of the CT-scan and a chest X-ray performed within the 24h around the CT-scan. Patients whose hypoxemia was related to cardiogenic pulmonary edema were excluded. CT-scans and chest X-rays were interpreted semi-quantitatively by a senior radiologist and by a senior intensivist, respectively. The primary outcome was the agreement between chest X-ray and CT-scan for ARDS diagnosis using Cohen’s kappa. Secondary outcomes were the sensitivity of chest X-ray for ARDS diagnosis in the overall population, and in the subgroup of focal ARDS, defined as infiltrates in the lower lobes on CT-scan.

Results: Among the 1015 CT-scan performed during the study period, 332 were included in the analysis. Mean age of patients was 62, 74% were males, and 29% were immunocompromised. In-ICU mortality was 41%. CT-scans were performed 3 [1–11] days from intubation. The day of the CT-scan, mean PaO₂/FiO₂ratio was 165mmHg. The prevalence of ARDS on CT-scan was 82%. Consolidations were mostly found in inferior lobes whereas ground glass was found evenly in all lobes. The agreement between chest X-ray and CT-scan for ARDS diagnosis was moderate (Cohen’s kappa 0.41 [95% confidence interval 0.30–0.51]). Sensitivity of chest X-ray for ARDS diagnosis in the overall population was 77%(95%CI 71–82) and 54%(95%CI 40–67) in the subgroup of focal ARDS.

Conclusion: Chest X-ray interpreted by an intensivist and CT-scan interpreted by a radiologist had moderate agreement for ARDS diagnosis. The performance of chest X-ray was particularly poor in focal ARDS. Systematic CT-scan assessment in hypoxemic patients may improve the likelihood of ARDS diagnosis.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Elsa Moncomble (elsa.moncomble@aphp.fr)

Annals of Intensive Care 2013, 13(1): CO-53

Rationale: Recent reports in patients with prolonged acute respiratory distress syndrome (ARDS) have described a paradoxical decrease in plateau pressure (Pplat) and increase in respiratory system compliance (Crs) in response to continuous anterior chest compression (CACC). This effect could be explained by the decrease in overdistension due to the decrease in end-expiratory lung volume. We hypothesized that in the early phase of ARDS, a paradoxical response to CACC may unveil overdistension during protective ventilation.

Patients and methods/materials and methods: Patients with moderate-to-severe ARDS were enrolled in a single-center study. The tidal volume was set to 6mL/kg of predicted body weight. Positive end-expiratory pressure (PEEP) was initially set so that the Pplat reached 28 to 30 cmH20 (High PEEP). A CACC of 80 cmH₂0 was obtained by compression of a 1 L saline bag opposite the sternum. Respiratory mechanics was assessed at High PEEP, before and during CACC. Subsequently, a decremental PEEP trial from PEEP 20 to 4 cmH20 (with steps of 2 cmH₂0) was performed while recording electrical impedance tomography (Enlight 800, Timpel), which allowed estimation of the percentage of overdistended lung at each PEEP level. PEEP was then set to the highest level that minimized overdistension, defined as the first level with < 1% of overdistended lung (Low PEEP). Finally, respiratory mechanics was assessed at Low PEEP, before and during CACC.

Results: 20 patients were included, 16 males and four females, aged 63 [42–65] years, intubated for 1 [0–2] days at the time of enrollment. The etiology of ARDS was pneumonia in 15 (75%) patients. At High PEEP (16 cmH₂O [16–18]), PaO₂/FiO₂ was 123 [93–172] mmHg and Crs was 38 [33–44] ml.cmH₂O⁻¹. When CACC was applied, Pplat increased in only two (10%) patients (delta Pplat: + 1.6 and + 2 cmH₂0). These two patients had a percentage of overdistended lung < 1%. A paradoxical response was observed in the remaining 18 (90%) patients (delta Pplat:—1cmH₂0 [-2–1]). These patients had a percentage of overdistended lung of 25% [15–34], significantly higher than their counterpart (p = 0.01). At Low PEEP (8 cmH₂O [6–10]), when CACC was applied, Pplat did not decrease in 19 (95%) patients (delta Pplat: + 1.8 cmH₂0 [1.6–2]) and decreased in only one (Figure 1).

Conclusion: The paradoxical response to CACC may unveil “occult” overdistension in moderate-to-severe ARDS during protective ventilation. Whether the response to CACC could help to customize ventilator settings deserves further investigation.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Marion Giry (marion.giry7@gmail.com)

Annals of Intensive Care 2013, 13(1): CO-54

Rationale: The increase in antimicrobial resistance (AMR) is a major public health issue. Antibiotic stewardship programs (ASP) are being developed to curb this phenomenon. Implementing ASP in the intensive care unit (ICU) is an acute topic given that AMR is common and antibiotic consumption is the highest in these departments. Different resources are available for ASP (prospective audit and feedback, referees, antibiotic time-out, etc.). Among them, checklists developed in different fields are interesting tools that can be used in the ICU. This study aimed to assess the impact of a checklist for prescribing antibiotics in the ICU.

Patients and methods/materials and methods: We conducted a monocenter before-after study in the medical ICU of a French tertiary hospital. We analyzed patients admitted to the ICU who received non-prophylactic antibiotic therapy for more than 48h over periods of nine months before and after the implementation of the checklist (from June 2020 to February 2021 and from June 2021 to February 2022, respectively). The checklist consisted of a paper sheet compiling all data related to antibiotic prescriptions daily filled by physicians or residents. We evaluated the rate of appropriate antibiotic therapy.

Results: Among the 385 patients included (239 for the baseline period and 146 for the intervention period), there was no statistically significant difference in the percentage of appropriate antibiotic therapy between the two periods (23% versus 22%, p = 0.9). There was a significant increase in the rate of appropriate dosing (81% vs. 91%, p = 0.006), in the rate of appropriate antibiotic therapy duration (59% vs. 69%, p = 0.037), and in the performance of therapeutic drug monitoring by increasing the performance of antibiotic blood measures (31% vs. 49%, p < 0.001). These improvements were achieved without any negative impact on mortality (19% vs. 27%, p = 0.062) or length of stay in the ICU (24days vs. 21days, p = 0.063). Also, a decrease in the ICU antibiotic costs (€113,667 for the baseline period versus €102,266 for the intervention period) and a decrease in antibiotic use in ICU were observed after the checklist implementation (1,350 versus 1,271 Daily Defines Dose/1,000 patients days).

Conclusion: The introduction of a specific checklist for antibiotics in the ICU was associated with an improvement in quality markers for antibiotic prescription without impacting the safety of patients hospitalized in the ICU. To our knowledge, this was the first time a checklist specific to antibiotic prescription has been evaluated in an adult ICU and its encouraging results show the importance of developing such ASP.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Manon Launay (manon.launay@chu-st-etienne.fr)

Annals of Intensive Care 2013, 13(1): CO-55

Rationale: Ceftazidime (CAZ) is commonly used as a pivotal antibiotic against Pseudomonas aeruginosa in critically ill patients. ICU patients have severely altered and variable antibiotic pharmacokinetics, resulting in lower antimicrobial concentrations and potentially poor outcomes. Several factors, including obesity and renal function, may influence pharmacokinetics. Thus, the objective of the study was to evaluate the impact of obesity and renal function on CAZ plasma concentrations and dosing regimens in ICU patients.

Patients and methods/materials and methods: All consecutive adult patients from 6 ICUs, treated with continuous CAZ infusion and under Therapeutic Drug Monitoring evaluation were included. Obesity was defined as body mass index ≥ 30kg/m². Glomerular filtration rate (GFR) was estimated by Chronic Kidney Disease Epidemiology Collaboration formula. CAZ recommended levels for plasma concentrations were between 35 and 80mg/L.

Results: A total of 111 patients (45 obese), weighted 90 (± 24.4) kg, were included. Mean GFR was 82mL/min/1.73m2 (± 40.3). Recommended CAZ plasma concentrations were achieved only for 49.6% of patients, with a median dosing regimen of 6g/d. Obese patients had lower CAZ plasma concentrations compared to non-obese patients (37.8 vs 56.3mg/L; p = 0.0042*) despite similar dosing regimens (5.83g/d vs 5.52g/d, p = 0.2529). Almost all Augmented Renal Clearance patients were underdosed despite CAZ dosing of 6.6g/d (± 0.8).

Conclusion: ICU obese patients required significantly greater CAZ amount to achieve the target range. A ailored dosing regimen may be considered based on weight and GFR. Future prospective studies should be performed to confirm this individualized dosing approach.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Maëlle Martin (maellemart1@gmail.com)

Annals of Intensive Care 2013, 13(1): CO-56

Rationale: Ventilator-associated pneumonia (VAP) is the leading nosocomial infection in intensive care and is associated with prolonged mechanical ventilation (MV) and greater antibiotic use. When VAP is suspected, starting antibiotic therapy (AT) immediately after pulmonary sampling may expose uninfected patients to unnecessary treatment, whereas waiting for bacteriological confirmation may delay AT in infected patients. In absence of robust data, VAP suspicion management remains at the physicians’ discretion, balancing diagnostic probability, patient risks, and risk of selecting resistant bacteria. The aim of this study is to compare an immediate AT (aggressive strategy) to a delayed AT (restrictive strategy) in patients with suspected non-severe VAP, in terms of antibiotic sparing and morbi-mortality.

Patients and methods/materials and methods: We conducted a 2-year retro-prospective before/after monocentric study, comparing an aggressive to a restrictive strategy of VAP suspicion management. All patients under MV longer than 48h, pulmonary sampled for a first VAP suspicion, were included, excepted in case of clinical severity (Onset/worsening shock or hypoxemia with PaO2/FiO2 < 150 or neutropenia < 1G/L) or ongoing AT predicted ≥ 4weeks. In the aggressive period (2019), AT was started immediately after distal lung sampling was performed to assess VAP. In the restrictive period (2022), no AT was given before cultures results were available. If cultures were positive, probabilistic AT was begun with subsequent adaptation to antibiotic susceptibility for a total of 7days. We compared the number of days alive without AT censored at day 28, ventilator-free days (VFD), and Day-28 mortality.

Results: Respectively 44 and 43 patients were included in the aggressive and restrictive phases. Compared to an aggressive strategy, a restrictive strategy was not associated with a reduction in AT consumption (12.8 ± 9.1days alive without AT over 28days vs. 13.4 ± 10.2days, p = 0.58), including no reduction in broad-spectrum AT (17.9 ± 11.2days vs. 17.6 ± 12.3days, p = 0.66). There was no significant difference in MV duration (13.1 ± 11.1 VFD in the restrictive group vs 12.8 ± 11.2, p = 0.92) nor in Day-28 mortality (HR = 0.96 IC95% [0.42–2.23] p = 0.93).Treatment was significantly more appropriate in the restrictive than aggressive strategy (38/43 [88.4%] vs 30/44 [68.2%] patients, p = 0.02).

Conclusion: In the setting of suspected VAP without severity, delaying AT to a confirmed VAP, compared to an immediate AT, was not associated with morbidity or mortality. Nonetheless, it was not associated with antibiotic sparing (including broad spectrum), even if treatment was significantly more appropriate.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Thomas Douillard (douillardt@gmail.com)

Annals of Intensive Care 2013, 13(1): CO-57

Rationale: Most patients admitted to the ICU for severe COVID-19 pneumonia received early empirical antibiotic therapy (1). This is explained by the severity at presentation with profound hypoxemia and a high rate of bacterial co-infection in other viral infections, like influenza pneumonia. In fact, the rate of bacterial co-infection seems to be very low, under 10% of cases (2). Wesought to evaluate the impact of early empiric antibiotic therapy for patients admitted to the ICU with severe SARS-CoV-2 pneumonia.

Patients and methods/materials and methods: Our study is an ancillary analysis of the retrospective cohort “COCOV IRL”, conducted in 13 ICUs in France to assess the impact of corticosteroids. All adult patients admitted between February 1 and December 31, 2021, for severe COVID-19 were included. Two groups were formed: patients who received empiric antibiotic therapy between admission and day 2, and those who did not. The primary outcome was death on day 28. Secondary outcomes included intubation or death on day 28 and the number of ventilator-free days on day 28. Univariate and multivariate analyses were conducted including age, comorbidities, severity at presentation, corticosteroids administration, ventilatory support, and center effect as co-factors.

Results: One thousand and fifty-six patients were included, 852 (80%) received antibiotic therapy and 204 (20%) did not. Patients who received antibiotic therapy were more severe with higher IGS2 scores at admission (35 ± 13.2 vs 30 ± 12.7, p < 0,001). In a multivariate analysis, factors associated with death at day 28 were older age and higher IGS2 scores. There was no association between early antibiotic therapy and mortality on day 28 (HR = 0.78[0.41; 1.47] p = 0.4).After adjustment on confounding factors, the absence of early empiric antibiotic therapy was associated with an increase of ventilator-free days at day 28 (beta = 2.8[0.7; 4.9],p = 0.009), and a risk reduction of intubation or death at day 28 (HR 0.62[0.45; 0.84],p = 0.003).

Conclusion: In our study, early empiric antibiotic therapy in severe SARS-CoV-2 pneumonia was not associated with day 28 mortality. Nevertheless, the absence of early empiric antibiotic therapy was associated with more ventilator-free days and less intubation or death at day 28. Our results do not establish a causal link but open a discussion about the real impact of early empiric antibiotic therapy in severe SARS-CoV-2 pneumonia.

Reference 1

1.Langford BJ, So M, Raybardhan S, Leung V, Soucy J-PR, Westwood D, et al. Antibiotic prescribing in patients with COVID-19: rapid review and meta-analysis. Clin Microbiol Infect. 2021;27:520–31.

Reference 2

2. Rouzé A, Martin-Loeches I, Povoa P, Metzelard M, Du Cheyron D, Lambiotte F, et al. Early Bacterial Identification among Intubated Patients with COVID-19 or Influenza Pneumonia: A European Multicenter Comparative Clinical Trial. Am J Respir Crit Care Me.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Alexis Maillard (alexis.maillard@aphp.fr)

Annals of Intensive Care 2013, 13(1): CO-58

Rationale: In immunocompromised patients with acute respiratory failure (ARF), the clinical significance ofrespiratory virus detection in the nasopharynxremainsuncertain. We aimed to evaluate the association between viral detection in nasopharyngeal swabs and the causes and outcomes of ARF.

Patients and methods/materials and methods: Preplannedpost hocanalysis of a randomized controlled trial that enrolled immunocompromised patients admitted to 32 intensive care units for ARF. Nasal swabs sampled at inclusion were assessed for 23 respiratory pathogens using a multiplex PCR assay. The causes of ARF were established by managing physicians and reviewed by three expert investigators blinded to the multiplex PCR assay results. Associations between virus detection in nasal swabs, causes of ARF, and a composite outcome of Day-28 mortality and/or invasive mechanical ventilation (IMV) were assessed.

Results: Among the 512 sampled patients, 44.9% had hematological malignancies, 35.6% had solid tumors, 11.4% were solid organ transplant (SOT) recipients and 7.3% were allogeneic hematopoietic stem cell transplant (HSCT) recipients. The multiplex PCR assay was positive in 105 patients (20.5%), and a virus was detected in 104, including 36.6% rhinovirus/enteroviruses, 10.7% coronaviruses, and 53.8% flu-like viruses (influenza virus, parainfluenza virus, respiratory syncytial virus, or human metapneumovirus). The etiology of ARF varied significantly according to the results of the multiplex PCR assay, especially the proportion of viral pneumonia:50.0% with flu-like viruses, 18.2% with other viruses, and 3.6% when no virus was detected (p < 0.001). There was no difference in the composite outcome of Day-28 mortality and/or IMV according to assay positivity (53.8% vs. 54.7%, p = 0.882). In a pre-established subgroup analysis, flu-like virus detection was associated with a higher rate of Day-28 mortality and or IMV among allogeneic HSCT recipients.

Conclusion: In immunocompromised patients with ARF,the results of nasopharyngealmultiplex PCR assays are not associated with the need for IMV or mortality. A final diagnosis of viral pneumonia is retained in one-third of patients with a positive assay and in half of the patients carrying a flu-like virus.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Benoît Misset (benoit.misset@chuliege.be)

Annals of Intensive Care 2013, 13(1): CO-59

Rationale: Passive immunization with plasma collected from convalescent patients has been used to treat COVID-19. The CONFIDENT open-label randomized trial found that mortality on day 28 was lower in COVID-ARDS patients treated with convalescent plasma (CP) than in patients treated with standard of care (SOC) (SRLF 2022 and (1,2)). The protocol planned secondary outcomes including a long-term assessment. The objective is to present secondary endpoints of the CONFIDENT trial.

Patients and methods/materials and methods: 475 patients with COVID-19 ARDS mechanically ventilated for less than 5days (stratified at 48h) were randomly assigned in a 1:1 ratio to receive either convalescent plasma with a neutralizing antibodies titer at least 1/320 or standard of care in the trial between October 2020 and March 2022. The primary outcome—day-28 mortality—was 35.4% in the CP group and 44.9% in the SOC group (p = 0.03), and 32.7% and 46.8% in the stratum included < 48h (p = 0.008). Secondary endpoints included the antibody response to COVID-19, ICU-acquired bacteremia and pneumonia, and day-365 mortality and quality of life. Funded by KCE # COV201004.

Results: Total IgG titers (BAU/mL) increased from 262 [31–881] on inclusion to 1180 [621–1,720] on day 7 (p < 0.0001), 1,470 [872–2,260] on day 28, and 1,260 [797–1,960] on day 90 (p > 0.05 between CP and SOC and between days 7, 28 and 90). NAb titers and total IgG titers to SARS-CoV-2 in the patients were associated (p < 0.0001, linear mixed model). In the CP and SOC groups: bacteremia occurred in 85 (35.9%) and 102 (42.9%) (Δ [CI95] = −7.0 [−15.95 to 1.84]), VAP in 182 (76.8%) and 188 (79.0%) (Δ [CI95] = −2.2 [−9.54 to 5.05]) patients. Bacteremia per 1,000 ICU-days were 22 and 23 and occurred at a median [IQR] of 53 [10– > 100] days for the CP group and 22 [9– > 100] days for the SOC group. VAP per 1,000 MV days were 73 and 75 (p > 0.05) and occurred at median [IQR] of 7 [2–18] days for the CP group and 5 [2–13] days for the SOC group. Mortality on day 90 was 102 (44.0%) in the CP group and 121 (50.8%) in the SOC group. Day-365 mortality was 107/237 (45.1%) in the CP group and 123/238 (51.6%) in the SOC group.

Conclusion: In COVID-ARDS patients, anti-SARS-Cov-2 antibodies increased in the first week following inclusion and secondary infections were very high. The beneficial effect of convalescent plasma on day-28 mortality was not associated with a difference in these prespecified secondary outcomes.

Reference 1

Misset, BMC Pulm Med. 2020 Dec 7;20(1):317.

Reference 2

Misset, Intensive Care Medicine Experimental 2022, 10(2):000713.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Hassen Ben Ghezala (hassen.ghezala@gmail.com)

Annals of Intensive Care 2013, 13(1): CO-60

Rationale: Intoxication among detained patients for the purpose of autolysis is still a taboo subject. The aim of the study is to assess the epidemiological indicators of acute intoxications, the factors that affect the clinical course and the outcome in ICU.

Patients and methods/materials and methods: It is a retrospective, descriptive and cross-sectional study, including all critically ill detainees patients admitted in the medical and toxicology intensive care unit of a reference poison center, from 1st January 2020 to 31st January 2023. We collected demographic data of the detained population, exposures, suspected intoxications, clinical and paraclinical manifestations, and treatments.

Results: Eighty-one acute poisoning was included in the study: sixty patients (96% male); twenty-one recurrent episodes [2,5]. Most patients (51%) were unemployed, 10% were a day labourer. The mean age was 34years [24,69]. 75% were single. Diabetes mellitus, chronic heart disease and chronic obstructive pulmonary disease were the most common co-morbidities (40.7%, 17.3%, 11.1%). Psychiatric disorders were recorded in 35% patients. 40% had a personal history of suicide attempts ranging from 2 to 5 attempts,75% of which required admission in ICU. 80% were belonging to low socioeconomic class. Median SAPS II and APACHEII scores at admission were respectively 8[6,50], 2[0,22]. Suicidal attempt leading to acute poisoning was present in 65%22% have intoxication for a secondary interest, 10% were body packer and withdrawal syndrome was recorded in 2.5% of cases. The mean delay of presentation to the emergency department was 15h. For forty-five patients poisoning was due to oral antidiabetics (50% metformin and 22% sulfonamide), 22% due to conversion enzyme inhibitor, 13% due to beta-blockers, 6% to paracetamol. The predominant clinical manifestations were gastrointestinal manifestations (79%), hemodynamic failure (13%), neurologic failure (13%), and only one patient had respiratory failure. 53% had metabolic disorders: metabolic acidosis (79,4%), hyperlactatemia (64%), renal failure(26.5%), rhabdomyolysis (12%). Most patients were treated with supportive measures, one patient was given gastrointestinal decontamination and specific antidotes were used in 20% of patients. Four patients (5%) required invasive mechanical ventilation. Hemodialysis was used for five patients. The median length of hospital stay was 2days [2,31], and the mortality rate was about 1%.

Conclusion: Acute poisoning is an alarming health problem in the prison. It is important to think about how to limit its impact.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Hela Maamouri (helamaamouri@yahoo.fr)

Annals of Intensive Care 2013, 13(1): CO-61

Rationale: Various scores in intensive care are calculated to estimate the severity of the patients, or to help with diagnostic approach. Until now, there is no specific score for intoxicated patients that can be easily calculated and validated to evaluate the severity of the patients. The aim of this work was to validate a pre-established score; the "scoretox" for patients admitted following acute intoxication by cardiotoxic product and to study its performance.

Patients and methods/materials and methods: It is an observational, longitudinal, retrospective, monocentric study including all acute intoxications with cardiotropic agents admitted to the medical intensive care unit Mahmoud Yaacoub Center from January 2018 to October 2021. The primary endpoint was severity. The score is calculated at admission and based on sixteen items including patientage, history, clinical and paraclinical features, and therapeutic behaviors at admission.

Results: We included 2495 patients including 604 patients admitted for poisoning by cardiotropic agents. The median age of the patients was 25years[18–37] with a female predominance and a gender ratio M/F at 0,5. The IGSII, APACHE II and scoretox scores were calculated after patient admission with a respective median of 14[8–28], 5[2–10] and 2[1–4]. Sinus bradycardia was found in 21% of patients (n = 106). Eleven patients (1.8%) showed a membrane stabilizing effect on the electrocardiogram. Seventy-seven patients (12%) were in cardiogenic shock. The performance study had shown that scoretox performed better in patients admitted for cardiotoxic poisoning by comparing it with IGSII and APACHE II. It also had the highest likehood ratio and the highest specificity. There was a linear and positive correlation between scortox and these scores (p < 10⁴) with a stronger relationship between scortox and APACHEII (r = 0.457). In addition, in the study of calibration, the scoretox was not well calibrated.

Conclusion: The new severity prediction score for patients admitted for cardiotropic intoxication was outperformed by the IGSII, APACHEII and PSS scores.This study suggests that the "scoretox" can be useful to assess the severity of patients admitted to intensive care for poisoning with cardiotoxic drugs. An external validation on a multicenter study is however necessary.

Compliance with ethics regulations: N/A.

Correspondence: Dabor Resiere (dabor.resiere@chu-martinique.fr)

Annals of Intensive Care 2013, 13(1): CO-62

Rationale: To evaluate the clinical characteristics and consequences of long-term exposure to noxious gas emissions from decomposing Sargassum in the local population of Martinique.

Patients and methods/materials and methods: We conducted a prospective follow-up of a patient cohort admitted to the emergency department (University Hospital of Martinique) due to exposure to sargassum from March 2018 to December 2022. We assessed the patient's exposure to Sargassum and air pollutants based on recordings of coastline sensors measuring H2S and NH3 levels and data from the Regional Air Quality Observatory. Demographics and clinical data (including cardiovascular, neurological, and respiratory events) were collected.

Results: In the six-year study period, 560 patients were included. Patients mainly arrived with referral letters from their general practitioner (80%) and presented headaches (76%), gastrointestinal disturbances (79%), dizziness (54%), skin lesions (30%), cough (44%), and conjunctivitis (33%). Not all patients were clinically symptomatic. Initial lung function tests were regular (50%), and twelve patients were admitted to the intensive care unit.

Conclusion: Our study indicates that the magnitude of health effects following long-term exposure to Sargassum may be larger than previously recognized. Efforts to limit long-term exposure are mandatory.

Compliance with ethics regulations: N/A.

Correspondence: Salma Ghalloussi (salmaghalloussi93@gmail.com)

Annals of Intensive Care 2013, 13(1): CO-63

Rationale: Datura stramonium is a ubiquitous solanaceous plant found in urban and rural areas. It has become popular for its hallucinogenic effect and is consumed in various forms. All parts of the plant are toxic because they contain substances (alkaloids) that oppose the action of a neurotransmitter, acetylcholine. These compounds are atropine, scopolamine, hyoscyamine.

Patients and methods/materials and methods: We report a series of cases of acute intoxication with Datura hospitalized in the department of intensive care and toxicology of Mahmoud Yaacoub center of Tunis (CAMU) from October 2008 to February 2022.

Results: Sixteen patients were eligible, among them nine were male (56%). The median age was 39years with extremes of 17 and 79years. The main season of intoxication was spring (37%). The median time to onset of symptoms was 1h [1,8] and the median time to emergency room visit was 6h [2,24]. Only one patient was followed in psychiatry for a bipolar disorder and four patients were addicted to drugs (25%). Intoxication was voluntary in six patients and accidental in six others. Clinical signs were mydriasis with photophobia and blurred vision (93%), mucocutaneous dryness (43%), intense thirst (40%), fever (10%), dysphagia (25%), nausea (35%) and difficulty in speaking (13%). The presence of a bladder globe was noted in five patients (31%). The neuropsychological signs were almost constant including agitation (88%), hallucinations (88%), confusion (81%) reminiscent of alcoholic drunkenness, somnolence (63%), coma in three patients (19%) and seizures in two patients who required mechanical ventilation. Cardiovascular signs were palpitations (49%), sinus tachycardia (41%), bradycardia (13%), and hypertension (39%). Acute renal failure was present in 12% of cases. No digestive decontamination was performed in this series. Symptomatic treatment was effective in most cases. All intubated patients were extubated and evolved favorably.

Conclusion: The diagnosis of Datura intoxication is clinical. Its low cost and easy access are the factors of an increase in its consumption as a psychoactive substance. The prognosis is related to clinical disorders: hospitalization is always necessary.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Boudour Ben Dhia (bidourabdbg2014@gmail.com)

Annals of Intensive Care 2013, 13(1): CO-64

Rationale: White henbane «jusquiame» is a very rare plant, of the family Solanaceae, its use is reserved for the medical profession since it is a powerful narcotic, used as antispasmodic, sedative and analgesic or even anesthetic. The alkaloids contained in this plant have toxic power both dermal and ingested form. We reported this series of cases of jusquiame poisoning to emphasize on clinical signs and a potentially fatal prognosis.

Patients and methods/materials and methods: It is a retrospective study conducted in our intensive care unit during a period of ten-year (2012–2022). Were included all patients admitted for henbane intoxication.

Results: Seven patients were included in our study, with gender-ratio (F/M) = 1.3 (ages 24–83years). Intoxications were consequent to accidental ingestion. One patient used this herb from self-medication and one for recreative purpose. All the cases have been severely intoxicated and they were presented with an anticholinergic toxidrome two to six hours after ingestion. Neurological manifestations of intoxication were mydriasis and exaggeration ofosteo-tendon reflexes for all cases. Four patients had loss of consciousness with CGS (from 14 to 8), 67% with confusion, agitated delirium in two cases and one patient presented seizures. The conscious patients (33%) had described excruciating headaches. All patients had mild tachycardiawith a mean of 90 ± DS 20. One patient had respiratory failure. The other symptoms were mucocutaneous dryness (83%), flushed skin (32%), urinary retention (50%) and nausea, vomiting (50%). One patient had acute renal failure. Management was symptomatic for all patients, one case required mechanical ventilation and sedation by benzodiazepine. The median length of stay was 1day [1–9]. There were no death.

Conclusion: Health care providers and physicians, in particular emergency physicians and clinical toxicologists, should be aware of the nature, medical uses, clinical characteristics, diagnosis and management of White henbane intoxication.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Hela Maamouri (helamaamouri@yahoo.fr)

Annals of Intensive Care 2013, 13(1): CO-65

Rationale: Metformin is a biguanide oral antidiabetic agent used as a first line treatment of type II diabetes. Metformin toxicity is rare but may lead to life threatening lactic acidosis with high mortality rate. Hemodialysis is the only method allowing removing of metformin and lactate clearance. The aim of this study is to determin the best moment of hemodialysis initiation and its predictive factors.

Patients and methods/materials and methods: It is a retrospective monocentric study on a series of clinical cases with metformin poisoning admitted to intensive care unit of CMYAMU conducted from January 2011 to February 2021. We analyzed anamnestic, clinical and gasometrical data (kinetics of pH, HCO3 and lactate at H8 and H12 of admission) with SPSS version 24.0

Results: We included 200 patients admitted for metformin poisoning. Most of cases (99,5%) had voluntary intake metformin. Median ingested dose was 14,45g[8.5–22.5]. Co-ingestions were found in 38.5% of cases. Median age was 25years[20–33] with female predominance. Psychiatric history was frequently identified (27,5%, n = 55). The mean delay between ingestion and admission was about 3h. Most of patients consulted for digestive symptoms. Eight patients needed vasoactive amines treatment and four of them required mechanical ventilation. At admission, average of pH was at 7.36, bicarbonates at 19.83 and lactate at 4.45mmol/l. Hemodialysis was required for thirty-two patients (16%) and was based on kinetics of pH, bicarbonates and serum lactates at H8 and H12 from admission. Median duration of dialysis was about 4h[3–6]. Independent predictive hemodialysis factors identified by multivariate analysis were serum lactate 11mmol/l[95% CI 6.03–715], ingested dose superior to 35.843g[95% CI 3.1–50.1] and anion gap superior to 27.5[95% CI 4.3–284].

Conclusion: Metformin poisoning prognosis depends on patient comorbidities, circumstances that leaded to toxicity and moment of extracorporeal treatment. Kinetics of pH, bicarbonates and serum lactates may help clinicans to indicate the best moment of hemodialysis. Four hours of conventional hemodialysis may be effective for patients without comorbidities.

Compliance with ethics regulations: N/A.

Correspondence: Hana Fredj (fredjhana@yahoo.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-66

Rationale: Extubation failure is a complication that increases morbidity and mortality. Non invasive ventilation (NIV) and High flow nasal cannula (HFNC) are two commonly used methods to provide respiratory support after. However, the optimal approach for post-extubation support in severe burn patients remains controversial. The aim of this study was to assess the effect of postextubation High-Flow Nasal Oxygen vs Noninvasive Ventilation on reintubation among burn patients at high risk of extubation failure.

Patients and methods/materials and methods: Retrospective case–control study was conducted in intensive burn care in Tunisia over a period of two years (2021–2022). Two groups of patients were assigned: (G1: HFNC), and (G2: NIV). The primary outcome was the proportion of patients reintubated at day 3; secondary outcomes were length of stay and ICU mortality.

Results: During this period, 235 were undergoing mechanical ventilation. Sixty-two were extubated. Thirty-four patients were included and divided in 2 groups: G1 (HFNC; n = 17), and G2 (: NIV; n = 17). The majority of patients were intubated for facial burns (N = 21). The two groups were comparable in terms of age, TBSA and severity of burns. Our study shows that there was no significant difference between the use of HFNC and NIV on the prevention of reintubation. The number of reintubated patients was comparable in two groups (3/17 versus 4/17, p: ns). Reintubation was within less than 72h in all cases. There was no difference between the two groups in terms of interface tolerance (Table1). The length of stay and the mortality were comparable in the two groups (Table1).

Conclusion: In mechanically ventilated burns at high risk of extubation failure, HFNC therapy was comparable to NIV for preventing reintubation.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Vincent Bonny (vincent.bonny@aphp.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-67

Rationale: The main risk of extubation is a relapse of respiratory distress with the need for reintubation. This is associated with increased morbidity and mortality. The mechanisms implicated in weaning failure are not fully understood and determining when a patient can be separated from the ventilator remains challenging.

Patients and methods/materials and methods: In patients who presented a successful spontaneous breathing trial on a T piece, we performed a quantitative measure of sound (dB) induced by three coughing efforts. This measure was done by using the sonometer Pulsar 14, two centimeters away from the distal extremity of the endotracheal tube.

Results: 106 patients were analyzed. Extubation failure occurred in 15 patients. The median [interquartile range] age was 65 [51–75], 36% of patients were women, 43% required intubation for acute respiratory failure, 25% for coma, 17% for shock, 6% for cardiac arrest and 9% for other reasons. The median duration of mechanical ventilation at enrollment was 4 [3–7] days. The sonoscore was significantly higher in patients who succeeded extubation, the median score being 222 [197–233] against 174 [156–191] dB, p < 0.001 in patients presenting extubation failure. In patients who succeeded extubation, the sound intensity produced by the last cough was also significantly higher than the one produced by the first, whereas there was no difference in sound intensity in patients with extubation failure. The area under the Receiver Operating Characteristic curve of the sonoscore to predict extubation failure was 0.911 IC 95% [0.829–0.993], p < 0.001. With a cut-off value of sonoscore at 180dB the relative risk to present extubation failure was 16.38 [5.2–51], p < 0.001, and with a cut-off value at 200dB the relative risk was at 5.3 [3.2–8.6], p < 0.001.

Conclusion: In weaning patients, undergoing mechanical ventilation for at least 24h in the intensive care unit, the intensity of the sound induced by coughing measured with a sonometer seems to predict extubation failure.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Jean-Baptiste Peretout (jperetout@cognacq-jay.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-68

Rationale: Weaning units have shown their value in the management of tracheostomized patients after prolonged intensive care¹. The global and multidisciplinary approach contributes to accelerate and improve weaning². The objective of this study is to describe the results of weaning and decannulation in these patients.

Patients and methods/materials and methods: This is a retrospective, descriptive study in a respiratory weaning unit over the year 2022. All patients hospitalized in the weaning unit were included. Demographic characteristics and weaning unit outcomes were collected. All inpatients with complete data were analyzed. The analysis was completed by a subgroup analysis of patients on 24-h spontaneous ventilation at admission and patients still with mechanical ventilation (MV) at admission (intermittent or continuous).

Results: In 2022, 103 tracheostomized patients were admitted. A total of 97 patients were analyzed with male 62%, age 62 ± 12years and length of stay 58 ± 27days in intensive care unit. The main diagnosis at admission in ICU was acute respiratory failure. The goal of admission on weaning unit was: ventilatory and tracheostomy weaning in 77% and tracheostomy weaning in 23%. On admission, patients were on 24-h spontaneous ventilation in 24% (n = 23) and on mechanical ventilation in 76% (n = 74): 80% intermittent MV (n = 59); 20% continuous MV (n = 15). In overall patients, 24-h spontaneous ventilation was achieved in 10.2 ± 13.9days, decannulation in 19 ± 16.3days and the length of stay was 35.8 ± 27days. The overall decannulation success rate was 86.6%. Subgroup analyses are presented in Table I.

Conclusion: Patients with prolonged intensive care and complex medical needs treated in a specialized weaning unit had high weaning and decannulation rates. This was achieved with a multidisciplinary team approach to continued intensive care and simultaneous rehabilitation with physiotherapists, speech therapists, nutritionists… Cares in a weaning unit remains relevant despite a prolonged stay in ICU.

Reference 1

1. Boles JM, Bion J, Connors A, Herridge M, Marsh B, Melot C, Pearl R, Silverman H, Stanchina M, Vieillard-Baron A, Welte T. Weaning from mechanical ventilation. Eur Respir J. 2007 May;29(5):1033–56.

Reference 2

2. Nelson JE, Cox CE, Hope AA, Carson SS. Chronic critical illness. Am J Respir Crit Care Med. 2010 Aug 15;182(4):446–54.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Romain Lombardi (lombardi.romain@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):CO-69

Rationale: Machine learning has been illustrated in various medical fields, such as fluid management in sepsis, prediction of renal failure and others. The weaning period is the key to the management of a patient on mechanical ventilation. Weaning can take up to half the time spent in intensive care. Up to 20% of patients do not pass their first withdrawal test. However, mortality can reach 38% in patients with the most difficult weaning. [1] Only a small number of studies have looked at the application of machine learning to the weaning process.Being able to predict the success of spontaneous breathing test and extubation is essential to improve morbi-mortality and reduce length of stay. We propose to develop a predictive algorithm for the success of a weaning test and identify the different factors that determine this success.

Patients and methods/materials and methods: It is a critical care, single-centre and retrospective study. Most of the variables are taken from the literature and extracted from the computerized patient record. We designed several machine learning algorithms: Logistic Regression, Random Forest Classifier, Support Vector Classifier (SVC), K-Nearest Neighbors algorithm (KNN),XGBoost, and Light Gradient Boosting Method (LGBM). In order to maximize the accuracy of the prediction and to reduce the risk of overfitting, we computeddifferent methods: multiple imputation with K-Nearest Neighbors for missing data, the SMOTE method (Synthetic Minority Oversampling technique) and K-fold cross validation. The results are expressed in terms of AUC, and the factors’ importance is determined by the SHAP (Shapley added explanations) method.

Results: Our cohort included 80 patients (60 successful, 20 unsuccessful). The best algorithm was LGBM with AUC 0.873 (F1score: 0.87, AUC-pr: 0.95, accuracy: 0.813), Figure 1. The other algorithms had a worse predictive performance (Random Forest: 0.727, Logistic Regression: 0.8, XGBoost: 0.836). Weight gain since admission, presence of VAP, protidemia and fibrinogenemia at the time of testing, and duration of invasive ventilation were considered to have the greatest impact on the model.

Discussion: Despite the small number of patients, all the techniques in place allow good prediction and reproducibility. However, the cohort will continue to grow until the congress (to a minimum of 250 patients). The main determinants affecting the success of the test are similar to those found in other studies in this field and are factors that can be directly influenced.

Conclusion: Machine learning will be useful in predicting weaning success and will be a real help in daily clinical practice.

Reference 1

Béduneau, G., 2017. Epidemiology of Weaning Outcome according to a New Definition. The WIND Study. Am J Respir Crit Care Med 195, 772–783. https://doi.org/10.1164/rccm.201602-0320OC

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Kahaia De Longeaux (delongeaux.kahaia@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):CO-70

Rationale: Despite a successful spontaneous breathing trial (SBT), until 30% of the patients will need a new intubation in the 72h following extubation, defining weaning failure. SBT failure and weaning failure are associated with poor outcomes. The aim of this study was to evaluate tidal volume, EWS.O2 score (including respiratory rate, pulsed saturation of oxygen (SpO2), inspired fraction of oxygen and heart rate) and heart rate variability (HRV) in patients during weaning process.

Patients and methods/materials and methods: We conducted a prospective cohort study in the medical critical care department of a teaching hospital. Patients ventilated for more than 24h and meeting predefined criteria for weaning were included. SBTs were performed with a T-piece during 30min. Tidal volume, EWS.O2 score and HRV parameters were collected, continuously and non-invasively, before, during the test and after extubation. Tidal volume was measured using a non-contact time-of-flight camera. EWS.O2 was recorded automatically and continuously with a closed-loop device used for oxygen flow titration according to SpO2 target set by the clinician.

Results: 52 spontaneous breathing trials were analyzed. We observed an increase of the EWS.O2 score and no increase of tidal volume in patients failing SBT. In patients extubated after a successful test but needing reintubation in the 72h, an increase in EWS.O2 during and after the test was observed, no increase in tidal volume was observed and multiple scale entropy (an HRV nonlinear parameter) increased in the first ten minutes of the test. A > 7,8ml/kg and a > 450mL cutoffs for tidal volume at 5min from the start of the SBT have a 100% sensitivity to predict weaning success (AUC = 0,824, p < 0,001 and AUC = 0,789, p = 0,002 respectively).

Conclusion: Non-invasive and continuous evaluation, easy to acquire, of physiological parameters during T-piece SBTs could help us better and early identify patients at risk of weaning failure. We aim to develop an artificial intelligence dynamic model to predict weaning outcome using those parameters.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Christophe Rault (christophe.rault@chu-poitiers.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-71

Rationale: Introduction: In intensive care unit, prediction of spontaneous breathing trial outcome and weaning from assisted ventilation remains difficult (1). Since the inspiratory cortical command is strongly involved during respiratory efforts (2), the amplitude of this inspiratory premotor cortical command could be associated with the success or failure of weaning from the ventilator.

Patients and methods/materials and methods: Material and methods: In 68 intubated and ventilated patients in intensive care unit, the amplitude of inspiratory premotor potentials was measured by quantified electroencephalography. The evolution (ΔIPP) of this amplitude between breathing under mechanical ventilation and spontaneous breathing through a T-piece was calculated. ΔIPP was compared between patients who failed and those who passed the spontaneous breathing test.

Results: Results: ΔIPP was significantly different between the failure and success groups (p < 0.0001). ΔIPP increased (i.e. inspiratory premotor potentials amplitude increased in spontaneous breathing compared to mechanical ventilation) in the failure group and decreased in the success group. An increase in ΔIPP above the threshold of 0.425µV was associated with failure of the T-piece test with a sensitivity of 100% and a specificity of 87%(Figure).

Conclusion: Conclusion: Premotor inspiratory potentials amplitude is a reliable reflection of the difficulties experienced by intubated patients when stopping the ventilator and could be useful in clinical practice if the measurement method was simplified.

Reference 1

Thille AW, Richard JC, Brochard L. The decision to extubate in the intensive care unit. Am J Respir Crit Care Med. 2013;187(12):1294–302.

Reference 2

Rault C, Sangare A, Diaz V, Ragot S, Frat JP, Raux M, et al. Impact of Sleep Deprivation on Respiratory Motor Output and Endurance. A Physiological Study. Am J Respir Crit Care Med. 2020;201(8):976–83.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Anaïs Dartevel (adartevel@chu-grenoble.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-72

Rationale: Legionella Pneumophiliais a cause of pneumonia in ICU (Intensive Care Unit). We aimed for the first end-point to describe the epidemiology and outcomes of Legionella pneumonia in French ICU. Secondary, we aimed to explore potentials death risk factors.

Patients and methods/materials and methods: We performed a multi-center retrospective study in 11 different French medical or medico-chirurgical ICUs between January 2014 andDecember 2019.

Results: Legionella pneumonia was diagnosed in 157 patients in the study period. Invasive ventilation was necessary for 93 patients (58.6%). Among these, 73 presented the criteria of acute respiratory distress syndrome (ARDS). Most of these patients were treated with bi-therapy antibiotics (124, patients; 79%). The most frequent bi-therapy was fluoroquinolone and macrolide (116 patients).The median duration of ICU stay was 11days (Q1–Q3 = 5–19).Among the 157 patients,19 patients did not survive at 28days. In multivariate analyses, only age(HR, 1.07; 95% CI, 1.01–1.13)SOFA score(HR, 1.30; 95% CI, 1.04–1.63)and septic shock at baseline(HR, 4.37; 95% CI, 1.05–18.32)were significantly associated with mortality.

Conclusion: Mortality at 28days in ICU of legionella was 12.1%. The septic shock inLegionella Pneumophiliain ICU wasindependently associated with mortality.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Nicholas Heming (nicholas.heming@aphp.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-73

Rationale: The role of corticosteroids in the management of non COVID-19 community acquire pneumonia (CAP) remains controversial.

Patients and methods/materials and methods: The Activated Protein C and Corticosteroids for Human Septic Shock (APROCCHSS) trial was a placebo, controlled trial of the efficacy and safety of the corticosteroids hydrocortisone and fludrocortisone, in adults with septic shock. Patients with CAP were identified a priori as a subgroup of interest. The primary outcome was 90-day mortality.

Results: In patients with CAP, there were 109/283 (38.5%) deaths at 90-day in the hydrocortisone plus fludrocortisone arm and 143/279 (51.3%) in the placebo arm (RR [95% CI]: 0.75 [0.62; 0.91] P = 0.002). In patients without CAP, there were 148/319 (46.4%) deaths at 90-day in the hydrocortisone and fludrocortisone arm and 157/329 (47.7%) in the placebo arm (RR [95% CI]: 0.97 [0.83; 1.14] P = 0.73). There was a significant statistical heterogeneity in corticosteroids effect on 90-day mortality across subgroups with or without CAP (Breslow-Day homogeneity test P = 0.045).

Conclusion: Hydrocortisone plus fludrocortisone was associated with reduced 90-day mortality in community acquired pneumonia associated with septic shock.

Reference 1

Annane D et al. NEJM 2018.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Guillaume Millot (millot.gui@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):CO-74

Rationale: The relationship between ventilator-associated events (VAE) and micro aspiration in intubated patients has not been studied to this day. The objective was to study the relationship between abundant micro aspiration of oropharyngeal secretions or gastric content and the incidence of VAE.

Patients and methods/materials and methods: This was a post hoc analysis of the BESTCUFF study, which was a multicenter, cluster randomized, cross-over, controlled, open-label trial conducted in French intensive care units (ICU) from June 2014 to October 2015 to study the impact of tracheal tube cuff shape on micro aspiration (1). In adult patients ventilated for over 48h, tracheal aspirates were sampled repeatedly for 48h following enrollment, with measurement of pepsin and amylase in said samples. VAE were identified using NHSN VAE criteria (2) based on PEEP or FiO2 variations compared to stable parameters in previous days. The primary outcome was the impact of overall abundant micro aspiration on the incidence of VAE using a survival analysis, adjusted for potential confounding factors.

Results: Among 326 patients included in the BESTCUFF study, at least one measurement of either pepsin or amylase was available for 261 patients, of which 31 (11.9%) developed VAE, with an overall mean age of 62.7 ± 14.7, a majority of male patients (164, 62.8%), a median SAPS II score of 50 [40–61], a median SOFA score of 8 [5–11]. Patients were mainly admitted for acute respiratory failure (117, 44.8%). There was no statistically significant impact of overall abundant micro aspiration on the incidence of VAE at day 28 (adjusted hazard ratio (HR): 1.55 [0.46–5.17], p = 0.48) nor of gastric abundant micro aspiration (adjusted HR: 1.24 [0.61–2.53], p = 0.55) or oropharyngeal abundant micro aspiration (adjusted HR: 1.07 [0.47–2.42], p = 0.88). Similarly, we could not find an association between overall abundant micro aspiration and day-28 mortality (adjusted HR: 0.79 [0.35–1.73], p = 0.55), 28-day ICU length-of-say (adjusted HR: 1.21 [0.74–1.98], p = 0.44) and day-28 duration of mechanical ventilation (adjusted HR: 1.12 [0.75–1.66]).

Conclusion: In this post hoc analysis, there was no significant association between abundant micro aspiration and the incidence of VAE, nor on secondary outcomes including mortality, ICU length-of-stay and duration of mechanical ventilation. Given the small amount of patients who presented with VAE in this study, the analysis might have been underpowered. A dedicated study enrolling more patients could be required to rule out a relationship between those events.

Reference 1

Jaillette, E., Girault, C., Brunin, G. et al. Impact of tapered-cuff tracheal tube on microaspiration of gastric contents in intubated critically ill patients: a multicenter cluster-randomized cross-over controlled trial. Intensive Care Med 43, 1562–1571.

Reference 2

Magill, S., Klompas, M., Balk, R. et al. Developing a New, National Approach to Surveillance for Ventilator-Associated Events. Am J Crit Care 22 (6), 469–473 (2013). https://doi.org/10.4037/ajcc2013893

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Maxime Vaconsin (maximevaconsin@hotmail.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-75

Rationale: Pneumocystis jiroveccii is a ubiquitous fungus responsible for Pneumocystis pneumonia (PCP) in immunocompromised patients. As the number of non-HIV immunocompromised patients escalates in intensive care units, predominantly admitted foracute respiratory failure,identifying fungal infection is crucial. According to a recent systematic review, no data was published on the diagnostic value of PCR in the specific subgroup of HIV-negative immunocompromised patients in critical care. The diagnosis of PCP is challenged by cysts scarcely recovered from respiratory samples in this population, and few DNA copies often found in healthy subjects. Also, the added value of ß-D-glucan (BDG) assays on blood samples in this population is imperfectly known.

Our aim is to describe the non-HIV critical care population with a positive Pneumocystis PCR, and to evaluate its relevance to diagnose PCP.

Patients and methods/materials and methods: Data were collected retrospectively in one ICU specialized in infectious diseases between 2013 to 2022. Non-HIV patients with a positive Pneumocystis PCR were divided into three groups at the time of care, confirmed PCP, uncertain PCP and ruled-out PCP, according to factors like the presence of cysts, radiographic findings, or the likelihood of other diagnoses. We compared the PCR cycles and BDG assays of patients in each of these groups.

Results: Forty-four non-HIV patients had a positive Pneumocystis PCR, 18 were categorized as having a confirmed PCP, 10 as uncertain PCP, and 16 as ruled-out PCP. The main characteristics of the patients can be found in the attached table. Immunosuppression was mostly secondary to autoimmune conditions (45%) and solid organ transplantation (20%). No clinical or biological characteristics were significantly different between groups, except cycle threshold (CT) and BDG values. PCR < = 29 cycles was the best diagnostic threshold, with an area under ROC curve of 0,845 (95% IC = [0,674–0,948],p < 0.001),a sensitivity of 78% (95% IC = [52–94]), and a specificity of 79% (95% IC = [49–95]). BDG > 80pg/ml increased the diagnostic value in this population. Cysts were only found in respiratory samples of 7 patients. From the 16 patients in the ruled-out PCP group, 11 received no treatment and did not later develop PCP.

Conclusion: Diagnosing PCP in non-HIV patients can be challenging. Pneumocystis PCR can be positive in immunocompromised patients who do not have PCP. Using a threshold of 29 cycles and combining it with a BDG > 80pg/ml increases its diagnostic value.

Reference 1

Senécal et al., Non-invasive diagnosis of Pneumocystis jirovecii pneumonia: a systematic rev.

Reference 2

Giacobbe et al., Performance of Existing Definitions and Tests for the Diagnosis of Invasive Fungal Dise.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Matthieu Petit (matthieu.petit@aphp.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-76

Rationale: Ventilator associated pneumonia (VAP) due to wild-type AmpC-producing Enterobacterales (wtAE) is frequent in critically ill patients. Despite a low level of evidence, definitive treatment withcefepime is recommended to treat wtAE VAP over the use of third generation cephalosporin (3GC) or piperacillin, because the latter may fail or induce selection of AmpC-overproducing mutants. Our aim was to assess the risk of treatment failure and relapse of VAP of a strategy ofβ-lactams de-escalation in these patients.

Patients and methods/materials and methods: Our prospective, observational study included consecutive wtAE VAP patients in 20 French intensive care units between February 2021 and June 2022. The primary outcome was treatment failure, defined asan inadequate response to antimicrobial therapy leading to death or to the necessity to switch to a broader-spectrum antibiotic at day-7. Relapse of infection was collected as a secondary outcome and defined as a new AE pneumoniae within the first 28days after inclusion. De-escalation ofβ-lactams was defined as definitive treatment by piperacillin or 3GCs. Characteristics were compared across de-escalation and full-treatment groups. Factors associated with the primary outcome were evaluated using a multivariable mixed logistic regression was performed to take into account the center effect.

Results: Two-hundred and eighty patients were included. Patients were mostly ventilated for respiratory failure (66%), and 54% had COVID-19 pneumonia. At VAP diagnosis, the median SOFA was 7 [5;10] and 41 (15%) patients had ECMO. Enterobacter cloacae was the most prevalent specie (30%), followed by Klebsiella aerogenes (25%) and Serratia marcescens (22%). De-escalation was performed in 87 (31%) patients while 193 (69%) received a full-treatment with broad spectrum antibiotics. Treatment failure was similar in the de-escalation and full-treatment group (26 vs. 34%, p = 0.32). At day-28, relapse was as frequently observed in the de-escalation group as in the full-treatment group (31 vs. 24%). In multivariable analysis, de-escalation to 3GCs or piperacillin was not associated with treatment failure or relapse. There was no difference in vital status at day-28, mechanical ventilation duration or ICU length of stay between the two groups. ESBL acquisition or Clostridium difficile infection during hospitalization were also similar.

Conclusion: In patients presenting wtAE VAP, treatment failure and risk of relapse at day 28 were not higher when a de-escalation ofβ-lactams was used. A randomized controlled trial could provide more data in support of such a strategy.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Pierre Bay (pierre.bay@aphp.fr)

Annals of Intensive Care 2013, 13(Suppl 1):CO-77

Rationale: Antibiotherapy stewardship for suspected ventilator-associated pneumonia (VAP)/ventilated hospital-acquired pneumonia (vHAP) in ESBL-E carriers is challenging. BioFire® FilmArray® Pneumonia plus Panel (mPCR) allows the detection of 18 bacteria, 9 viruses, and 7 antibiotic resistance genes, including blaCTX-M, the most common ESBL gene.

Patients and methods/materials and methods: This monocentric, prospective study was conducted from March 2020 to August 2022. First, we evaluated the performance of mPCR on ESBL-E related pneumonia detection through blaCTX-M on respiratory samples performed for suspicion of ventilated hospital-acquired pneumonia. Over the study period, 228 mPCR were performed for a suspected VAP/hVAP, 41 in patients with ESBL-E rectal carriage and 187 in non-carriers. Second, we assessed the impact on antibiotherapy stewardship of mPCR, in 95 ESBL-E carriers (mPCR performed n = 22 or non-performed n = 73) with confirmed ventilated hospital-acquired pneumonia (confirmed from quantitative cultures of lower respiratory tract secretions sampled).

Results: mPCR performance: Among ESBL-E carriers, mPCR was positive for 24 patients (59%) including 15 (62%) with blaCTX-M detected. Among carriers, 9 (22%) developed an ESBL-E related pneumonia. The negative predictive value of mPCR to identify ESBL-E related pneumonia was 100% in carriers. In patients without ESBL-E carriage, sensitivity and positive predictive value of mPCR to identify ESBL-E related pneumonia were respectively 100% (95% CI 2–100) and 33% (95% CI 11–66). Antibiotic assessment: ESBL-E carriers developed 95 VAP/vHAP, 47 (49%) episodes were related to ESBL-E, and 24 (25%) to carbapenem-resistant bacteria. A mPCR-based approach was independently associated with an adequate empiric antibiotic therapy in the multivariable logistic regression (adjusted odds ratio (aOR) 7.8, 95% confidence interval (95% CI) 2–30.9, p = 0.003) and the propensity-weighted (aOR 10.1, 95% CI 2.7–37.9, p = 0.001) models.

Conclusion: mPCR showedgood microbiological performances with a negative predictive value of 100% to rule out the diagnosis of ESBL-E related VAP/vHAP. A mPCR-based antibiotherapy stewardship was associated with increased adequate empiric antibiotic therapy. mPCR on respiratory sample seems to be a promising tool in ESBL-E carriers with suspected VAP/vHAP.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Aïcha Rebai (aicha.rebai@fmt.utm.tn)

Annals of Intensive Care 2013, 13(Suppl 1):FC-001

Rationale: The primary cause of gas exchange abnormalities during anesthesia is pulmonary atelectasis.The aim of our study is to assess the potential effects of the recruitment maneuver on blood pressure and respiratory parameters. The primary end point was the change in lung compliance after alveolar recruitment.

Patients and methods/materials and methods: This is a prospective observational study conducted in the operating room of our Military Hospital. Patients from ASA1 to ASA 3 proposed exclusively for elective major abdominal surgery were included in the current study. The alveolar recruitment maneuver was performed to characterize a continuous positive airway pressure of 30cm H₂O for 30s. Then, respiratory and hemodynamic parameters before, during, and after alveolar recruitment were measured.

Results: Thirty-six patients underwent 79 alveolar recruitment maneuvers. No significant difference was detected in heart rate during and after alveolar recruitment (p = 0.63). In addition, we found a significant decrease in the minimum systolic blood pressure values during alveolar recruitment as compared to the values measured before and after alveolar recruitment (100mmHg [83; 113],104mmHg [94; 121], and 102mmHg [90; 113] respectively. Diastolic blood pressure did not drop during recruitment (p = 0.16). However, we detected a significant decrease in mean arterial pressure during alveolar recruitment as compared to values detected before and after the alveolar recruitment maneuver (68mmHg [58; 80], 77mmHg [67; 86], and 76mmHg [66; 84] respectively). The pulse pressure variation values were significantly higher during alveolar recruitment (p = 0.0001). Evenly, lung and tele-expiratory CO 2 pressure compliance were significantly increased after the alveolar recruitment maneuver.

Conclusion: The alveolar recruitment maneuver highly improves lung compliance. This ventilation strategy was relatively well tolerated on the hemodynamic level. However, further studies are necessary to specify the appropriate protocols for this approach.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Bruno Evrard (bruno.evrard@chu-limoges.fr)

Annals of Intensive Care 2013, 13(Suppl 1):FC-002

Rationale: Accurate diagnosis of acute cor pulmonale is key since it is prognostic in acute respiratory distress syndrome (ARDS). Identification of paradoxical septal motion is highly subjective. LV radial strain has not yet been used to help identification of paradoxical septal motion. Machine learning approaches provide individual prediction which can be reused for further assessment. We aimed to evaluate the accuracy of left ventricular (LV) radial strain using a supervised machine learning approach for the diagnosis of paradoxical septal motion, using conventional two-dimensional assessment as reference.

Patients and methods/materials and methods: In this retrospective study, a multicenter prospective cohort of ventilated patients with moderate-to-severe acute respiratory distress syndrome (ARDS) related to COVID-19 (n = 327) was used as the train dataset (75% of the cohort) and as test dataset (25% of the cohort). A second multicenter cohort of patients with moderate-to-severe ARDS was used as external validation dataset (n = 85). A supervised machine learning approach was compared with conventional visual identification of paradoxical septal motion in the two-dimensional short-axis view used as reference by two independents abstractors. The same digital loop used for conventional identification were used for LV short-axis strain assessment. Data were extracted and converted to perform data engineering using previous described method (Evrard B preprint 2023). Four algorithms (Extreme gradient boosting (XGBoost), Generalized Linear Model-Lasso (GLM-Lasso), Support Vector Machines (SVM) and random forest (RF)) were used and benchmarked to develop a diagnostic algorithm using 20 variables featured from LV radial strain curves assessment which were previously described for paradoxical septal motion diagnosis (ref).

Results: In the train set, the accuracy to identify paradoxical septal motion obtained by tenfold cross validation reached 0.73, 0.77, 0.78 and 0.75 for XG Boost, GLM-Lasso, SVM and RF, respectively. In the test set, similar performances were obtained (Figure 1A) with the best performances exhibited for the SVM and GLM-Lasso algorithm (area under the receiver operating characteristic curve [AUC], 0.84 and 0.90, respectively). Finally, all the models accurately classified the patients in the external validation (Figure 1B) with the best performances obtained by SVM algorithm (AUC, 0.85).

Conclusion: LV radial strain promises to accurately diagnose paradoxical septal motion with a machine learning approach.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Bruno Evrard (bruno.evrard@chu-limoges.fr)

Annals of Intensive Care 2013, 13(Suppl 1):FC-003

Rationale: Accurate diagnosis of acute cor pulmonale (ACP) is key since it is prognostic in patients ventilated for an acute respiratory distress syndrome (ARDS). ACP is characterized by a paradoxical septal motion (PSM) which identification using two-dimensional echocardiography is highly subjective. Left ventricular (LV) radial strain has not yet been used to help the identification and assess the severity of PSM. We sought to describe LV radial strain changes related to PSM in patients at risk of sustaining ACP according to its severity.

Patients and methods/materials and methods: This prospective bicentric study included patients ventilated for an ARDS related to COVID-19 who were assessed using transesophageal echocardiography between March 2020 and June 2021. Two-dimensional transgastric short-axis view at mid-papillary level was used to grade septal motion: normal (grade 0), transient end-systolic septal flattening (grade 1), prolonged end-systolic septal flattening or reversed septal curvature (grade 2). LV radial strain analysis was performed off-line on the same digital loops and six LV segments were distinguished: mid-infero-septal and mid antero-septal, their opposite segments (mid-infero-lateral and mid-antero-lateral, respectively), and the remaining two segments (mid-anterior and mid-inferior). After having confirmed visually that LV radial strain curves were altered in certain segments when a PSM was present (Figure), we performed feature engineering. We calculated the “time-to-peak” defined as the time lag required to reach the maximal value of strain, which was normalized by the length of cardiac cycle and the “partial area under segmental strain curves” which was calculated as the area under each LV segmental strain curve between 33 and 66% of the cardiac cycle length (time period selected on graphical examination of the strain curves where most alterations of strain pattern occurred) to standardize the measurement.

Results: Overall, 318 echocardiography examinations performed in 184 patients were analyzed. Two-dimensional assessment identified a grade 1 and a grade 2 PSM at end-systole in 106 (33%) and 43 (14%) examinations, respectively. When compared with mid-anterior or mid-inferior segments, the time-to-peak of mid-septal and mid-lateral segments occurred significantly later in systole and this delay gradually increased with the grade of PSM (Figure). Similarly, the area under the strain curve of mid-septal and mid-lateral segments increased significantly with the grade of PSM, compared with mid-anterior or mid-inferior segments (Figure).

Conclusion: LV radial strain can objectively depict PSM and quantitatively assess its severity.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Rosalie Schoux (rschoux@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-004

Rationale: During Acute respiratory distress syndrome (ARDS), high positive end expiratory pressure (PEEP) may have multiple and complex hemodynamic deleterious effects. To aim of this study was to evaluate the effect of high PEEP on coronary blood flow (CBF) measured at pressure 5 and 15 cmH2O.

Patients and methods/materials and methods: Thirty-two patients have been included. A multicenter prospective observational study of ARDS patients under mechanical ventilation underwent a transthoracic (TTE) and transesophageal echocardiography (TEE).

Results: Compared to PEEP 5 cmH2O, respiratory system compliance (Crs) (32[25–45]vs38 [27–57]mL/cm H2O; p < 0.0001),mean arterial pressure (MAP) (77[67–84] vs 83 [73–90]mmHg; p < 0.0001)and cardiac output (CO) (4.6 [3.8–5.8] vs 5.1 [3.8–5.8] l/mn, p = 0.0009) were lower at 15cmH2O. Left ventricular (LV) diastolic function assessed using early mitral annulus velocity (6 [5–8] vs 8 [7–11] cm/s, p = 0.004) and right ventricular (RV) systolic function assessed using systolic tricuspid annulus velocity (11[9–14] vs 13[11–15] cm/s, p = 0.02) were impaired at PEEP 15 cmH2O.Assessed using Doppler velocity time integral, diastolic and systolic left CBF (16.2[11.9–22] vs 18.1[13.1–23.2] cm/s, p = 0.03) (5.1[3.4–7.7] vs 6.3[4.4–8.7] cm/s, p = 0.006) and systolic right CBF (7.5[5.3–10.6] vs 9.6[8–14.6] cm/s, p = 0.02) were lower at PEEP 15cmH2O comparedto PEEP 5cmH2O. Impairment of left CBF was associated with Crs and CO impairement.

Conclusion: Left diastolic and systolic and right systolic CBF were lower at PEEP 15cmH2O in comparison with PEEP 5 cmH2O and associated with decreased Crs, CO, MAP, and impaired LV diastolic function and RV systolic function.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Alexis Lambour (alexis.lambour@hotmail.fr)

Annals of Intensive Care 2013, 13(Suppl 1):FC-005

Rationale: Mechanical ventilation with high PEEP in ARDS patients could induce cardiac dysfunction mainly due to the cardiac effect of lung overdistension. We hypothesized that lung recruitment maneuver (LRM) by opening and preventing lung over distension may improve the hemodynamic tolerance of high PEEP. Then, we decided to apply the same high PEEP to ARDS patients before and after recruitment maneuver to assess hemodynamics.

Patients and methods/materials and methods: Thirty-two patients with moderate to severe ARDS hospitalized in our intensive care unit. Respiratory and hemodynamic evaluation by transthoracic echocardiography were done at the same incremental (before LRM) and decremental (after LRM) PEEP = 25cmH₂O and driving pressure = 15cmH₂O. Patients were separated into two groups based on median compliance gain (“responders” and “non-responders”).

Results: Left ventricular (LV) ejection fraction (EF) (58% [51.3–71] vs 52.6% [44.4–67.3]; p = 0.04), mitral annulus plane systolic excursion (1.29 [1.13–1.37] vs 1.01 [0.85–1.13]; p = 0.007) and the product of heart rate and aortic velocity time integral (a surrogate of cardiac output) (1600 [1310–1789] vs 1370 [1215–1689]; p = 0.02) were improved after LRM. After LRM, we observed a decrease in right ventricular (RV) to the LV ratio (RV/LV ratio 0.88 [0.79–1] vs 1.01 [0.93–1.1]; p = 0.01) and an improvement of RV systolic function (tricuspid annulus plane systolic excursion 1.93 [1.65-0.2.19] vs. 1.6 [1.24–1.8]; p = 0.0002). The subgroup analysis revealed an improvement in LVEF and a decreased RV/LV in the "responders" group, and not in the "non-responders" group.

Conclusion: LRM induced an improvement of the cardiac function in patients ventilated with high PEEP particularly in the responders group of patients.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Simone Cappio Borlino (simone.cappio@unimi.it)

Annals of Intensive Care 2013, 13(Suppl 1):FC-006

Rationale: During acute respiratory distress syndrome (ARDS), positive end-expiratory pressure (PEEP) might increase pulmonary vascular resistance (PVR). In theory, according to the U-shape of the relationship between PVR and lung volume, which nadir corresponds to the functional residual capacity, PEEP should increase PVR in case of alveolar overdistention and lower PVR if alveolar recruitment predominates. This has never been proven in patients with ARDS. We evaluated PVR changes during PEEP modifications depending on the potential of lung recruitability, estimated through the recruitment-to-inflation (R/I) ratio.

Patients and methods/materials and methods: In intubated ARDS patients, monitored by a pulmonary artery catheter, an esophageal balloon and transthoracic echocardiography, PEEP was lowered by 10 cmH₂O from baseline. At the two levels of PEEP, hemodynamic, echocardiographic and respiratory variables were measured and preload dependance was evaluated through the passive leg raising test. An R/I ≥ 0.5 defined high recruitment potential (HRP) and R/I < 0.5 defined low recruitment potential (LRP).

Results: Forty measurements were performed in 14 patients (2 [1; 3] measures per patient): mean age was 65 ± 11years, 12 (86%) were male and median (IQR) baseline PaO₂/FiO₂ was 110 (74; 124) mmHg. The R/I ratio was ≥ 0.5 (0.70 [0.59; 0.85]) in 23 (57%) measurements. PEEP was lowered from 15 (12; 15) to 5 (2; 5) cmH₂O. By lowering PEEP, the PVR decreased in case of LRP (-25 [-30; -14]%) while they remained unvaried in case of HRP (0 [-11; 13]%, p < 0.0001) (figure). The gradient between mean pulmonary arterial pressure and pulmonary arterial occlusion pressure decreased in case of LRP (-8 [-18; -4]%) but increased in case of HRP (9 [0; 17]%, p = 0.0001). A similar effect was observed for right to left ventricular end-diastolic area ratio (-16 [-25; -14]% vs. 8 [-13; 20]%, respectively, p = 0.0008). The magnitude of PVR change induced by PEEP lowering was also correlated with R/I ratio (r = 0.52 [0.25; 0.71], p = 0.0006). In case of LRP, the variation in cardiac index induced by PEEP reduction was not different in preload dependent and preload non-dependent cases (23 [5; 32]% vs. -1 [-7; 26]%, respectively, p = 0.1).

Conclusion: Effects of PEEP on PVR in mechanically ventilated patients with ARDS might depend on alveolar recruitment. A significant risk of PVR increase exists when PEEP leads to pulmonary overdistention. This suggests that the potential of lung recruitment should be considered during PEEP titration, in order to better individualize respiratory support. The study is ongoing.

Reference 1

Chen L, Del Sorbo L, Grieco DL, et al. Potential for Lung Recruitment Estimated by the Recruitment-to-Inflation Ratio in Acute Respiratory Distress Syndrome. A Clinical Trial. Am J Respir Crit Care Med. 2020;201(2):178–187. https://doi.org/10.1164/rccm.201902-0334OC

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Daniela Rosalba (danielarosalba245@gmail.com).

Annals of Intensive Care 2013, 13(Suppl 1):FC-007

Rationale: The Starling system is a non-invasive hemodynamic monitoring device that measures cardiac output through bioreactance. Through bioimpedance, it also measures the mean transthoracic electric impedance (Z0), from which the thoracic fluid content (TFC) can be derived, which should reflect extravascular lung water, fluid in the large thoracic vessels and the cardiac cavities. We compared TFC to the extravascular lung water indexed for ideal body weight (EVLWI) and the global end-diastolic volume index for body surface (GEDVI) estimated by transpulmonary thermodilution (TPTD).

Patients and methods/materials and methods: A group of patients with shock and another one with acute respiratory distress syndrome (ARDS) were included. All the patients were equipped with calibrated TPTD device and the Starling system. Variables measured by TPTD and by the Starling system were collected at the same time before and after volume expansion (500mL saline) for patients with shock, and once a day during the course of the disease for ARDS patients.

Results: In the 42 patients with shock receiving a fluid bolus, TPTD-derived cardiac output increased by 0.44 [−0.16 to 1.06] (22.5%) L/min/m2 and bioreactance-derived cardiac output by 0.22 [−0.72 to 1.16] (11.8%) L/min/m2. GEDVI increased by 43.8 [−11.9–99.5] (7.6%) ml/m2 and TFC by 2.4 [−0.3 to 5.1] (3%). EWLVI decreased by 0.02 [−1.83 to 1.79] (0.25%) ml/kg. There was no correlation between the fluid-induced changes in GEDVI or in EWLVI on the one side and in TFC on the other (p = 0.95 and p = 0.58). In the 23 patients with ARDS, 124 measurements were performed. Between two successive measurements, EVLWI changed by −0.1 (6.3%) ml/kg and the TFC by −0.4 (4.6%). There was no correlation between the changes in EVLWI and in TFC between successive timepoints (p = 0.42). Considering all measurement together (n = 166), the was no correlation between the absolute values of GEDVI and TFC (p = 0.16), EVLWI and TFC (p = 0.06) and the sum of GEDVI and EVLWI on the one side and TFC on the other (p = 0.19). Still considering all measurements, there was no correlation between the bias between the sum of GEDVI and EVLWI vs. TFC on the one side and the bias between TPTD- and bioreactance-derived cardiac output on the other (p = 0.29).

Conclusion: In patients with shock or ARDS, there is no significant relationship between TFC measured by bioimpedance and GEDVI or EVLWI measured by TPTD, neither in absolute values, nor in relative changes.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Florence Daviet (florence.daviet@ap-hm.fr)

Annals of Intensive Care 2013, 13(Suppl 1):FC-008

Rationale: Lung transplant is associated with poorer survival than that observed in other solid organ transplants. Primary graft dysfunction (PGD) and infections are the leading causes of death during the first month of transplantation. The main objective was to assess whether the Fluid balance of the first 24h after lung transplantation had an impact on the occurrence of PGD at 72h.

Patients and methods/materials and methods: This is an ambispective observational study including all adult patients who received a lung transplant in our hospital from November 1^st 2020. The primary endpoint was the occurrence of stage 3 PGD 72h after lung transplantation, defined by the presence of pulmonary edema on the chest X-ray, associated with a PaO2/FiO2 ratio < 2001. The fluid balance was calculated by making the difference between fluid intake and fluid loss from the operating room to 24h after clamp release from the second lung (corresponding to Day1) then every day for 72h.

Results: We present the preliminary results on the first 30 patients included, transplanted between 1/11/2020 and 31/10/2021. They had a median age of 55years (45–63), 53% were men. Half of the patients presented with PGD whatever the stage during the first 72h, including 11 patients (37%) with stage 3 PGD.The median Fluid balance on Day 1 was 1278ml (259–3087) and 8341ml (5218–11,861) at 72h. In univariate analysis, the total Fluid balance at Day 1 (3024ml (675–6880) versus 678ml (−200 to 2323)), as well as the total Fluid balance at 72h (11,945ml (5853–14,595) versus 7596ml (4516–10,415)) were significantly higher in patients with stage 3 PGD (p = 0.006 and 0.03 respectively) (Figure 1). The other main factors associated with the occurrence of stage 3 PGD were: the recipient's body mass index (p = 0.005), the total lung capacity on the recipient's pre-transplant Exploration Respiratory Tests (p = 0.008) and the worst Donor PaO2/FiO2 (p = 0.04). There was a positive correlation between PGD stage and total Fluid balance at Day 1 (r = 0.43, p = 0.019).

Conclusion: These preliminary results support a correlation between the occurrence of stage 3 PGD and fluid balance in the first 24h after lung transplantation. They suggest that optimized management of fluid balance in the first hours after transplantation could improve patient prognosis.

Reference 1

Snell, G. I. et al. Report of the ISHLT Working Group on Primary Lung Graft Dysfunction, part I: Definition and grading—A 2016 Consensus Group statement of the International Society for Heart and Lung Transplantation. J. Heart Lung Transplant. 36, 1097–11.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Iyed Matouk (maatouk.yed@gmail.com)

Rationale: Phenobarbital (PB) is one of the oldest anti-epileptic drug. It is effective in all types of epilepsy except typical absences. However, its pharmacokinetic profile and its numerous adverse effects limit its use as a first-line drug in the long-term treatment of epilepsy. Barbiturates, particularly phenobarbital, the main anti-seizure drug, are an important cause of intoxication and mortality. The objective of this study was to describe the clinical, para-clinical, therapeutic and outcome characteristics of phenobarbital intoxications.

Patients and methods/materials and methods: We conducted a retrospective study including all cases of phenobarbital poisoning confirmed by serum assay and admitted in our ICU from 2004 to 2021.

Results: One hundred thirty-seven patients were included for voluntary poisoning with suicidal attempts. They were female in 66.4%. The mean age was 29.02 ± 12.36years. A history of psychiatric disorders were found in 21.9%, and 27.6% had previous suicide attempts. The mean estimated ingested dose was 1250mg (Interquartile range: 912.5–2000). The mean Apache II score was 4 (Interquartile range: 6–19). The mean SAPSII was 11 (Interquartile range 2–7). Neurological, cardiovascular and respiratory distress were present respectively in 59.3%, 12%, 11% of the patients respectively. The most frequent neurological signs were drowsiness (76.3%) followed by coma (12%), agitation (9%) and confusion (3%). Hypotension was observed in 7% and acute respiratory failure due to aspiration pneumonia in 11%. The serum level (phenobarbitolaemia) on admission was 56.95 ± 19.43mg. Mechanical ventilation was required in (37.2%), fluid expansion (12%), activated charcoal (17.5%), alkaline diuresis (8%), gastric lavage (7.3%). All patients recovered without any sequelae.

Conclusion: Phenobarbital poisoning can lead to serious neurological, cardiovascular and respiratory distress. Coma can last for some days. Physicians should be aware of the severity of this drug poisoning which has a good prognostic if symptomatic treatment is started in time.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Soumaya Saad (saadsoumaya1994@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-010

Rationale: Amitriptyline is one of the major tricyclic antidepressants that is used widely and is still one of the common causes of fatal drug poisoning. It manifests iespecially with electrocardiographic (ECG) abnormalities, arrhythmias, and hypotension. Predictive factors for the development of life-threatening complications are limited and not well known. The objective of our study is to describe clinical and prognosis findings in patients presenting with acute amitriptyline intoxication.

Patients and methods/materials and methods: This is an observational, retrospective single-center study, over a period of 6years (2017 to 2022). Inclusion of patients over the age of 14years admitted to ICU for acute amitriptyline intoxication. Demographics, clinical, laboratory, electrocardiographic (ECG) findings and patient’s progress were evaluated retrospectively. Predictive parameters for the development of serious complications were studied.

Results: 91 patients were included. Mean age was 32 ± 13years. Sixty-eight patients were female. Ingested amitriptyline median dose was 400 [240, 500] mg. 56 patients were followed for psychiatric disorders (psychosis (50%), depression (28.5%), bipolar disorder (14.3%), and personality disorder (7.1%). The main clinical neurological symptoms on admission were drowsiness in 84.6% of cases followed by coma in 56 patients (61.5%). Acute respiratory failure was reported in 3 patients. Hemodynamically, tachycardia was reported in 54(59.3%) patients, hypotension in 9(9.9%) patients and shock in 3 patients. Electrocardiographic abnormalities reported in our study series were: widening of the QT space in 6 patients, widening of the QRS complex in 4 cases, ventricular tachycardia in 4 patients and asystole in 2 patients. Therapeutic management was characterized by oxygen in 57(62.6%), volume expansion in 7 patients, catecholamines in 5 patients, bowel accelerator in 21 patients, anticonvulsant therapy in 8 patients. Invasive mechanical ventilation was used in 61.5% of patients. The median hospital stay at ICU was 2days and only one patient died.

Conclusion: Amitriptyline intoxication in our series was associated with a good prognosis but still remains serious due to its cardiac toxicity. Early and rapid management is necessary to anticipate and control potential complications.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Hassen Ben Ghezala (hassen.ghezala@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-011

Rationale: An overdose of Olanzapine, an atypical antipsychotic, is increasingly used in intentional drug poisoning. It can lead to serious or fatal poisoning. Objective:The objective of this study was to describe the clinical, para-clinical, therapeutic and outcome characteristics of Olanzapine intoxications observed at the Reference Tunisian poison center.

Patients and methods/materials and methods: We conducted a retrospective study including all patients admitted for Olanzapine poisoning in our Intensive Care Unit of Urgent Medical Assistance Center Mahmoud Yaacoub (Tunisia) from January 2011 to June 2022.

Results: In total, 80 patients were included. In our study, 63,8% patients were female and 36.3% patients were male. The mean age was 34.1 ± 13.06years. Mental health problems were found in 87.5% of patients. The most frequent was schizophrenia (51.9%). All cases of Olanzapine intoxication were related to suicide attempts. The mean estimated ingested dose of Olanzapine was 166.47mg ± 112.9. Poison Severity Score was minor in one patient, moderate in 40 patients, and severe in 39 patients. In our series, no death was reported. The intoxication was concomitant to benzodiazepine in 50% of cases. In our study, neurological symptoms were the most common adverse effect. Coma was found in 49% with Glasgow Coma Score (GCS) less than 8 in 33% and drowsiness in 45%. Cardiovascular symptoms were tachycardia (25%), hypotension (3.8%) and shock (1.3%). The membrane stabilizing effect was not noted in any of the patients. Respiratory symptoms were ARF (1.3%) whose etiology was aspiration pneumonia. The most frequent biological abnormality was hypokalemia (35.4%). At the electrocardiogram, the rhythm was regular and sinus (67.5%) except a tachycardia in 24.7% and bradycardia in 3.8%. The screening of olanzapine was performed in the urine (26.3%), the gastric fluid (17.5%) and in blood (1.3%). The treatment was mainly symptomatic: mechanical ventilation (41.3%), gastric lavage (1.3%), vascular expansion (2.5%) and Correction of hypokalemia (35.4%). The different complications noted were: decreased level of consciousness (76.8%), agitation (20.3%), aspiration pneumonia (11,4%), hypotension (2.5%) and bradycardia (1.3%). In our study, no death was reported and all cases recovered without any consequences.

Conclusion: In our study, Olanzapine Poisoning was associated with considerable toxicity and a significant number of complications. Therefore, physicians should be aware of the necessity of a careful clinical monitoring in severe cases.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Hassen Ben Ghezala (hassen.ghezala@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-012

Rationale: An overdose of Quetiapine can lead to serious or fatal complications. There are few data on quetiapine poisoning in the literature. Objective:The objective of this study was to describe the clinical, para-clinical, therapeutic and outcome characteristics of quetiapine intoxications observed at the Tunisian reference poison center.

Patients and methods/materials and methods: We conducted a retrospective case series monocentric study including all quetiapine poisoning admitted the Toxicology and medical Intensive Care Unit from March 2018 to July 2022.

Results: Fifteen patients were enrolled. Thirteen (13) patients were female and 2 patients were male. The mean age was 38 ± 14years. Twelve patients were known to have bipolar disorder. All cases of quetiapine intoxication were related to suicide attempts. The mean estimated ingested dose of Quetiapine was 883mg ± 665.2. Poison Severity Score was minor in 4 patients, moderate in 5 patients and severe in 6 patients. In our series, no death was reported. The intoxication was concomitant to benzodiazepine consumption in 8 cases. The most frequent neurological signs were coma (6/15) with Glasgow Coma Score (GCS) less than 8 (6/15) and drowsiness (4/15). Cardiovascular symptoms observed were hypotension (4/15), shock (1/15), tachycardia (1/15) and a long QT (2/15). We did not observe any membrane stabilizing effect in our patients. Respiratory symptoms observed were acute respiratory failure (7/15) with mainly aspiration pneumonia in 5 cases. The most frequent laboratory disorder was hypokalemia (8/15). At the electrocardiogram, we noticed regular sinus rhythm (12/15) except one case with tachycardia (1/15) and two others with a long QT (2/15). The qualitative serum screening of Quetiapine was performed only in one patient. The treatment was mainly symptomatic: mechanical ventilation (5/15) gastric lavage (2/15), fluid expansion (2/15), transit accelerator (2/15) and Correction of hypokalemia (8/15). The different complications seen were: hypotension in 4 cases with one case of shock, aspiration pneumonia in 5 cases, hypokalemia in 8 cases, decreased level of consciousness in 6 cases, delirium, agitation and irritability in one case. In our study, no death was reported and all cases recovered without any sequelae.

Conclusion: In our study, quetiapine overdoses were associated with considerable neurological and cardiovascular toxicity. Emergency department physicians should be aware of the increasing number and the severity of this emerging drug poisoning by quetiapine.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Hassen Ben Ghezala (hassen.ghezala@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-013

Rationale: Risperidone intoxication is uncommon and known to have a good prognosis with a low mortality rate. Very few reports on this intoxication exist.

Patients and methods/materials and methods: Retrospective observational study that included all patients admitted to the medical and toxicology intensive care unit of a reference poison center for management of Risperidone intoxication. The study period was 10years (from 2012 to 2022).

Results: Forty-four patients were included during the study period, a female predominance was noted with a gender ratio (1.93), the mean age was 33years ± 14. Psychiatric history was reported in 86% of the patients (22.7% depressed; 18.2% bipolar disorder; 18.2% schizophrenic and 40.9% had unspecified psychiatric history). In 9% of the cases the intoxication was related to a psychiatric decompensation and a therapeutic overdose and in 91% of cases was for suicidal purposes. Multi-drug intoxication was reported in the majority of cases: 62% associated with benzodiazepines, 20% associated with another neuroleptic and in 10% antidepressant. The supposed dose ingested was 40mg [8,120] with an average consultation time of 3h [1,6]. The neurological symptoms reported were: 50% somnolence, 25% extrapyramidal syndrome, 15.9% agitation, 4.1% seizures. A decreased level of consciousness (average Glasgow score 11 ± 4) was reported with a delay after intoxication of 5 ± 2h and required mechanical ventilation in 36%. Pupils were reflexive in 45.5%, mydriatic in 11.4% and miosis in 43.2%. No hyperthermia was reported. Cardiovascular manifestations reported: tachycardia 75%, collapsus 16% (43% requiring catecholamines and 47% responding to fluid expansion), rarely an ECG change was reported: long QT 6.8%, right bundle branch block 4.5%. The most frequent metabolic disorders were metabolic acidosis (29.5%), rhabdomyolysis (9.1%) and acute renal failure (4.5%), of which only one patient required hemodialysis. Activated charcoal was used in two patients. 42% of the patients presented pneumonia: 18.8% ventilator-associated pneumonia; 22.7% aspiration pneumonia. The median duration of mechanical ventilation was 2days [1; 24]. Among intubated patients, 32% presented an agitated awakening and 93.8% were subsequently extubated. The median length of stay in intensive care was 2days [2,32]. The reported death rate was 2.3%.

Conclusion: This study shows that acute risperidone intoxication can cause severe neurological and cardiovascular manifestations. Treatment is mainly symptomatic with no specific antidote.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Hassen Ben Ghezala (hassen.ghezala@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-014

Rationale: Minoxidil was initially an antihypertensive drug, which had restricted therapeutic indications. It is currently used in topical form for the treatment of hair loss.

Patients and methods/materials and methods: Monocentric retrospective observational study which included all patients admitted for minoxidil intoxication at the medical and toxicological intensive care unit of the Tunis reference poison center over a period of 13years (2009–2022). Inclusion criteria were mainly: ingestion of minoxidil associated with clinical and biological signs of intoxication.

Results: During the study period, 16 patients were admitted for the management of minoxidil intoxication and included in the study. The mean age was 43 ± 21years with a female predominance of 81.3%. Three patients (18.8%) had a history of hypertension and one (6.3%) patient had a psychiatric history. Thirteen patients (81.3%) mistook the treatment for a cough syrup and ingested it accidentally, while 3 patients (18.8%) took the treatment voluntarily for suicide attempt. The average quantity ingested was 16 ± 14cc with an average time to the emergency room of 10 ± 9h. At presentation, all patients were conscious, among which 6 (37.5%) presented with hypotension, 1 of which had peripheral signs of shock. Other signs of intoxication were dizziness with nausea in 7 patients (43.8%), abdominal pain and vomiting. Blurred vision was noted in only one patient and palpitations in half of the patients. During hospitalization, 10 patients (62.5%) presented shock that lasted on average 2.5 ± 1.4 with a maximum vasoactive drug dose of 5mg/h. On ECG, 10 patients (62.5%) had repolarization disorders, and troponins were elevated in 4 patients (25%). The major complications in ICU were mainly cardiogenic pulmonary edema in 3 patients (18.75%) and infectious pneumonia in 1 patient. The average length of stay was 3 ± 1days.

Conclusion: Although minoxidil intoxication is rare, it can be serious and life-threatening. In fact and as demonstrated by this work, several hemodynamic and cardiovascular complications can occur, Treatment is mainly symptomatic.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Salma Ghalloussi (salmaghalloussi93@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-015

Rationale: Colchicine is a drug with a narrow therapeutic margin. The intoxication is rare but can be serious, even fatal, due to multivisceral failure.

Patients and methods/materials and methods: Retrospective observational study from 2007 to 2022 including all patients admitted for colchicine intoxication in the department of intensive care and toxicology of CAMU. Were included poisoned patients with supposedly toxic ingested dose ≥ 0.15mg/kg. We collected demographic, clinical, biological and prognostic data.

Results: During the study period, 25 patients were admitted for acute colchicine voluntary intoxication with a female predominance (88%) and an average age of 27 ± 11years. Co-morbidities were Gout disease (8%), Behcet's disease (8%), pericarditis (4%), lupus in (4%), rheumatic disease (4%), hypertension and dyslipidemia (4%). No psychiatric history was found. The intoxication was poly-drug in 28% of cases. The digestive signs were predominant with vomiting (84%), nausea (82%), abdominal pain (72%), profuse diarrhoea (64%) and digestive haemorrhage (8%). Cardiovascular signs were also frequent including tachycardia (52%), hypotension (12%), QT and ST segment prolongation (8%) with repolarization disorders (8%). Fever was present in 16% of cases. Neurological signs were rare as confusion (8%) or delirium (4%). Hydroelectrolytic disorders were frequent (48%) dominated by hypokalemia (44%), thrombocytopenia (16%), cytolysis (8%), cholestasis (4%) and lactic acidosis (24%). The poisoning was complicated by hepatocellular insufficiency (12%), rhabdomyolysis (12%), acute renal failure (8%) and pancytopenia in two patients. Excessive alopecia occurred in one patient. Mechanical ventilation was necessary in three patients (12%) with fatal outcome for the three intubated patients. Prognostic factors were an assumed ingested dose of 0.3mg/kg [0.1, 1.5], a mean delay to initial consultation (7h [1,47]), the decrease in prothrombin level (12%), the occurrence of cardiogenic shock (20%), an acute respiratory distress syndrome (8%), a sepsis (8%) or a multivisceral failure (16%).

Conclusion: Colchicine intoxication is potentially lethal and there is no specific antidote to date. It is recommended to observe the patient for at least 24h. Close monitoring should be performed and supportive treatment should be initiated early.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Boudour Ben Dhia (bidourabdbg2014@gmail.com)

Rationale: Theophylline belongs to the xanthine family and is used for its bronchodilator effect. However, because of its very narrow therapeutic margin, it is less and less prescribed. Theophylline intoxications, whether accidental or voluntary, are rare but can be serious in some cases due to life-threatening cardiac and neurological complications.

Patients and methods/materials and methods: It was a retrospective study between 2017 and 31st January 2023. We followed all patients who had been referred to our ICU for theophylline intoxication. Information collected included age, sex, amount of drug supposedly ingested and any initial clinical manifestations of toxicity. A blood sample was also collected for determination of plasma theophylline concentration. A plasma theophylline concentration ≥ 15mg/L was considered a reference for the diagnosis of poisoning.

Results: Forty-five patients were included. Their median age was 20 [14;84] years; they were predominantly female 35 (79,5%). The co-morbidities were in 27% chronic obstructive pulmonary disease (asthmatic n = 7; 58% and chronic obstructive pulmonary disease n = 5; 42%). A history of psychiatry depressive disorder was noted in 12%. The mean delay in presenting to the emergency department was of 6h [1;25]. The median theophylline supposed ingested dose was 3000mg [300; 12000]. It was associated with other substances in 23 cases. Intoxication was voluntary in 91% and the overdose was reported in 9%. Clinical features were digestive disorders in 36 cases (80%), tachycardia in 28 patients (62.2%), tremors in 12 patients (27%), agitation for two patients (4%), seizures in one case and mean of CGS was 15 [14–15]. Six patients had a hemodynamic distress requiring the use of cathecolamines. Metabolic acidosis was present in 36 patients (80%) with a mean bicarbonate level of 18.8 ± 3.66mmol/L. Lactate measurement was performed in 16 patients with a mean of 4.69 ± 2.02mmol/L. Hypokalemia was found on admission in 27 patients (60%), with a medium level of 3.18 ± 0.53mmol/l. Theophylline measurement was performed at admission with a median of 40[20;86] mg/l. Theuse of beta blockers as antidote was required in 15 cases (33%). The average length of stay in hospital was 2days [1,6]. The evolution was favorable in all cases.

Conclusion: Doctors should be aware of theophylline toxicity. Theophylline poisoning can be fatal for serum level above 100mg/l. Recognition of clinical signs of poisoning and rapid management are the only guarantees of the prognosis.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Manel Kallel (manel.kallel@fmt.utm.tn)

Annals of Intensive Care 2013, 13(Suppl 1):FC-017

Rationale: Early assessment of the prognosis of patients admitted to emergency departments (EDs) is a major objective of the emergency physician. Management and referral depend on this assessment. Numerous severity scores based on the patient's clinical and biological parameters have been developed and validated. The Rapid Emergency Medicine Score (REMS) is one of these scores and has been studied in trauma emergencies and in septic patients. The main objective of this study is to evaluate the prognostic value of REMS in the prediction of intra-hospital mortality at 24h in patients admitted to the emergency room.

Patients and methods/materials and methods: This is a prospective, observational, single-centre study conducted in an emergency department of a regional hospital over a six-month period. We included all patients managed in the life-saving emergency room and calculated their REMS.

Results: During the study period, we included 367 patients. The mean age of the patients was 55 with a sex ratio of 1.75. The in-hospital mortality rate was estimated to be 5.7%. The REMS score was statistically higher in the group of deceased patients (p = 0.02). The mean REMS in survivors was 4.75versus 9.43 (p = 0.02); patients deceased with a REMS inferior to 6 were six and those with a REMS superior to 6 were fifteen (p < 0.001). A REMS superior to 6 was also significantly associated with admission to intensive care (p = 0.031). In univariate study, it was predictive of mortality (OR = 5.98, 95% CI [2.25—15.84], p = 0.001). The characteristics of the ROC curve of the REMS score for predicting mortality were therefore: AUC 0.79, Sensitivity 71.43%, Specificity 70%, positive predictive value 12.82%, negative predictive value 97.6%.

Conclusion: The REMS is a good predictive tool for in-hospital mortality at the 24th hour of admission of patients in the emergency department. It is simple and easy to calculate.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Manel Kallel (manel.kallel@fmt.utm.tn)

Annals of Intensive Care 2013, 13(Suppl 1):FC-018

Rationale: Cardiovascular pathology is one of the most frequent causes of admission of patients to the emergency room. Classically associated with the male gender and advanced age, its prevalence in women and young adults is nevertheless clearly increasing. The aim of our work was to study the epidemiological, clinical and evolutionary characteristics of cardiovascular pathologies requiring admission to life-saving emergency room.

Patients and methods/materials and methods: Retrospective, descriptive, comparative and monocentric study conducted over a period of four months in a multipurpose emergency unit of a regional hospital. We included patients aged 18years and older with a cardiovascular pathology requiring a temporary hospitalization in the emergency room. We excluded pregnant women and any post-traumatic context.

Results: During the study period, 397 patients required admission to the life-saving emergency room, and cardiovascular pathology was identified in 93 patients (23.4%). Among them, there were 63 men (67.7%) and 30 women (32.3%). The average age of the women was 68 ± 15years, while that of the men was 59 ± 16years (p = 0.03). The diagnosis was acute coronary syndrome (ACS) in 43% of the cases (of which 20% were STEMI), acute heart failure (AHF) in 24.7%, a rhythm disorder in 17.2%, aorto-vascular pathology in 8.7%, pericarditis and cardiorespiratory arrest in 3.2%. In the ACS subgroup, the average age of onset was 57 ± 15years (p = 0.03). Pain was particularly severe (p = 0.046). Blood glucose was initially higher at 2 ± 1.4g/l (p = 0.048). However, the heart rate was significantly lower at 89 ± 17batt/mn (p = 0.05). The platelet count was significantly higher at 267,840 ± 67,000 el/mm3. In the CIA subgroup, the mean age was 74 ± 9years (10–3), the MEWS score was significantly elevated (10–3) as was the SAP (10–3). In the rhythm disorder subgroup, the mean age was 60 ± 15years (p = 0.7). There was no significant association with NEWS score, MEWS or pain intensity. Systolic blood pressure was significantly lower at 100mmHg (10–3) with no particular elevation in lactate levels (p = 0.8). Five patients died, including two in the ACS subgroup.

Discussion: Cardiovascular pathologies require scores to assess the severity and choose the orientation of patients admitted to the emergency room.

Conclusion: Cardiovascular pathology concerned elderly men and young women. The ACS subgroup was distinguished by high adrenergic stimulation, which was higher than in the rhythm disorder subgroup.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Badra Bahri (bahribadra@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-019

Rationale: Acute heart failure (AHF) is a frequent cause of admission in the emergency departement (ED). The prognosis evaluation is an essential step for the emergency physician. D- dimers are frequently collected in the ED given the importance of venous thromboembolic disease among the causes of decompensation of AHF. The purpose of our study was to identify the prognostic value of D-dimers in patients admitted to the ED for AHF.

Patients and methods/materials and methods: Monocentric observational study. Prospective collection of clinical, biological and evolution demographic data from computerized medical records. The diagnosis of AHF was retained on clinical, electrocardiographic, radiological, biological and gazometric data. The dosage of D-dimers was performed on admission to the ED. The primary endpoint was intra-hospital mortality.

Results: One hundred and twenty patients were included. The mean age was 72 ± 11.4years, the sex ratio was 0.96. The comorbidities were high blood pressure (78.3%), diabetes (49.2%), coronary artery disease (45.8%), heart failure (53.3%). 70% of patients consulted for orthopnea. The average length of stay was 6.36 ± 5.03days. The use of continuous positive airway pressure (CPAP) support was indicated in 36% of patients. The use of ventilatory support with inspiratory support was necessary in 37% of patients. The rate of admission to intensive care was 11.8%. The average D-dimer value was 1966 ± 2627ug/L. Intra-hospital mortality was 14.2%. Comparing the two groups of patients, the deceased (G1) versus the survivors (G2), the mean value of the D-dimers was (ug/L): G1: 4262 ± 5370 vs G2: 1550 ± 1390; p = 0.000. In multivariate analysis, a D-dimer value greater than 913 ug/L was an independent factor of intra-hospital mortality (AUC = 0.810; Se = 84%; Sp 60%; OR 3; 95% CI: 0.64–14.6; p = 0.017).

Conclusion: A Ddimer value greater than 913ug/L is an independent factor associated with in-hospital mortality in patients admitted to the emergency department for acute heart failure.

Compliance with ethics regulations: N/A.

Correspondence: Aïcha Rebai (aicha.rebai@fmt.utm.tn)

Annals of Intensive Care 2013, 13(Suppl 1):FC-020

Rationale: Terrorism is an international scourge that is the source of non-negligible mortality among victims. However, some deaths can be avoided through pre-hospital care. This study aimed to classify the deaths among victims of terrorism.

Patients and methods/materials and methods: We conducted a retrospective study from 2012 to 2017. We included victims of terrorist acts in our country who underwent autopsies at our Military Institution. A group of experts separately classified the deceased patients based on autopsy reports and images of injuries using the following classification: A preventable death, when a lifesaving act could be performed at the scene of the event by the victim, a person present next to the victim, or a first-line medical or paramedical professional. A potentially preventable death, when a potential lifesaving act is possible but unavailable due to lack of means or skills and can only be performed by a highly qualified person in a well-developed hospital setting. An unavoidable death: when the injury is beyond all therapeutic resources and cannot be treated no matter the means and skills available. Death was defined as preventable, potentially preventable, or unavoidable if at least two experts established this consideration.

Results: Our study included forty-six victims who died during a terrorist event, of which 93.4% died at the scene of the event and 6.6% in-hospital. A group of experts classified the patients using the following classification: preventable death, potentially preventable death, and unavoidable death. The three experts agreed on 82.6% of the deceased victims. The experts considered ten patients as having had preventable and potentially preventable death. The others were considered to have had an unavoidable death. We analyzed the degree of agreement among the experts' opinions. The intraclass coefficient was 0.9, indicating excellent agreement. The Kendal and Fleiss Kappa coefficients favored a strong agreement between the experts. The Kendal coefficients between the experts were all greater than 0.7, in favor of a strong agreement among the experts. The Kappa coefficient was greater than or equal to 0.66, indicating a strong to excellent agreement between the experts. All victims whose death was preventable and potentially preventable died from shock and hemorrhage.

Conclusion: All preventable deaths were secondary to a state of shock and hemorrhage. Classifying the deaths of victims of terrorism allows for future strategies to be put in place for both in-hospital and pre-hospital care to reduce preventable and potentially preventable deaths.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Alaeddine Zouari (zouari.aladin@gmail.com)

Rationale: Chest trauma is the third leading cause of trauma mortality. They are rarely isolated and are most often associated with other life-threatening lesions. The aim of our study is to report the experience of our department concerning chest trauma.

Patients and methods/materials and methods: This is a prospective study including patients who are admitted for chest trauma, over a period of one year, from 1^st August, 2020 to 31^st July, 2021 at the emergency department in the CHU Habib Bourguiba in Sfax.

Results: The total number of polytrauma patients admitted to our department during this period was 470 patients. Patients with chest trauma counted 203 patients. The average age was 37.6years with extremes ranging from 1 to 80years. They were 172 men and 31 women with a sex ratio of 5/1. It was a blunt trauma in 199 cases. On clinical examination, the respiratory rate was 18.4bpm with extremes ranging from 12 to 33bpm. Pulmonary auscultation was abnormal in 80 cases. The average oxygen saturation was 96.16% with extremes ranging from 66 to 99%. The average shock index was 0.85.The average Glasgow score was 13.2. On clinical examination of the chest, forty-nine patients presented pain and forty-seven patients presented chest contusion and bruises. Sixteen patients had rib fractures. Subcutaneous emphysema was identified in twenty cases. A chest X-ray was performed in one hundred and seven patients and was abnormal in ninety-five cases. An E-FAST ultrasound was performed in sixty-nine cases and showed the presence of pneumothorax in seventeen cases and hemothorax in fifteen cases. A Bodyscan was performed. It showed that the most frequent lesions in decreasing order were a rib fracture in 129 cases, a pulmonary contusion in 126 cases, a pneumothorax in 126 cases, an hemothorax in fifty-six cases. A flail chest was noted in eleven patients. Thirteen patients had a sternum fracture. A scapula fracture was noted in thirty-eight patients. Twenty-five patients had a clavicle fracture.Dorsal spinal trauma was associated in twenty-nine cases. Fifty drains were inserted. Fifty-three patients were intubated and ventilated. The evolution was marked by the death of nine patients, the transfer to another department of 127 patients and sixty-seven patients were discharged.

Conclusion: High-energy accident should always lead to the examination of the chest, even if the initial presentation is reassuring. Chest trauma should be considered as an evolving lesion and should therefore benefit from a reevaluation.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Karima Naanani (naananik@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-022

Rationale: The organization of the practice of sedation analgesia imposes a strict rigor, to reach an optimal and rationalized level of sedation. This study allowed us to review the practice of sedation analgesia in the emergency surgical unit, to evaluate the complications of prolonged sedation and in particular its impact on the duration of mechanical ventilation (MV) and on the length of stay.

Patients and methods/materials and methods: Our prospective observational study lasted 17months and included 230 patients sedated in the emergency surgical unit of the CHU Ibn Rochd.

Results: Our study included 230 sedated patients, 79% were male, the mean age was 46.2 ‡ 19.3years. Post-traumatic consciousness disorders represented the first indication for admission to the intensive care unit (43.5%). The mean duration of sedation was 3.6days [1–36days], the median duration of VM and the median length of stay were 5 and 6.5days, respectively. Mortality was more common in the sedated group superior to 48H. The introduction of sedation analgesia protocol and a better understanding of the pharmacology of sedation means, as well as the application of optimal and rationalized sedation significantly reduced the duration of MV, the length of stay, the total duration of sedation, the duration of intubation, the risks of tracheotomy and ventilator-associated pneumonia (all P value less than 0.001). This also led to better management of the weaning syndrome, delayed recovery, and the occurrence of critical illness neuromyopathy.

Conclusion: Sedation-analgesia, although frequently used, involves certain risks. Its practice has been evolving for several years towards a rationalization of its administration according to the needs of the patient. Reducing the duration of sedation could considerably reduce the duration of MV, and consequently the length of stay.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Romain Jouffroy (romain.jouffroy@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-023

Rationale: International recommendations for the management of sepsis and septic shock recommend early recognition, diagnosis and treatment to reduce associated mortality [1,2]. More than a single treatment, a set of care including antibiotic therapy (ABT) and hemodynamic optimization has demonstrated its effectiveness in a hospital setting. This study aims to study the association between 30-day mortality in patients with septic shock and the administration of ABT combined with hemodynamic optimization defined by a volume expansion of at least 10mlkg⁻¹h⁻¹.

Patients and methods/materials and methods: From May 2016 to March 2021, patients with septic shock requiring prehospital intervention by an Out-of-Hospital Mobile Emergency Unit (SMUR) were retrospectively analyzed. To assess the association with 30-day mortality, the Inverse Probability Treatment Weighting (IPTW) propensity method was applied.

Results: Among the 529 patients included, 354 (67%) were analyzed. Suspected pulmonary, digestive, and urinary infections were the cause of septic shock in 49%, 25%, and 13% of cases, respectively. The overall 30-day mortality was 32%. Seventy-one patients (20%) received prehospital ABT (C3G or Tazocillin) and volume expansion of at least 10ml.kg-1.h-1. Binomial log regression weighted by the IPTW found a significant association between 30-day mortality and prehospital administration of antibiotic therapy combined with prehospital hemodynamic optimization by optimal vascular replacement (RR of 0.56 [0.33—0.89], p = 0.02 and adjusted RR 0.52 [0.27–0.93], p = 0.03).

Conclusion: The administration of antibiotic therapy combined with prehospital hemodynamic optimization is associated with a reduction in 30-day mortality in patients with septic shock treated by a SMUR team.

Reference 1

Singer M. JAMA. 2016.

Reference 2

Chen AX. N Engl J Med. 2019.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Badra Bahri (bahribadra@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-024

Rationale: Acute dyspnea is a frequent complaint in the emergency depatement (ED). The use of biological markers is a tool that can help the emergency physician for an appropriate orientation of patients according to severity. The prognostic value of procalcitonine (PCT) has been described in many pathologies. The purpose of our study was to assess the prognostic value of PCT in patients admitted to the ED for acute dyspnea.

Patients and methods/materials and methods: It was a single-center study between January and March 2022. We collected prospective collection of demographics, clinical and biological data from a computerized medical record. We included all patients over the age of 18 admitted to ED for non-traumatic acute dyspnea. A PCT assay was performed on admission. The primary endpoint was intra-hospital mortality. The secondary endpoint was the use of mechanical ventilation.

Results: 211 patients were included. The average age was 72 ± 13.46years. The sex ratio 1.04. The comorbidities were (%): high blood pressure (n = 115, 54%), diabetes (n = 83, 39%), coronary artery disease (n = 47, 22.2%), obstructive pulmonary disease (COPD) (n = 48, 22.7%). The main causes of dyspnea were n (%): pneumonia (n = 46, 21.8%), heart failure (n = 49, 23%), pulmonary embolism (n = 11, 5.2%), covid (n = 75, 35.5%), exacerbation of COPD (n = 11, 5,2%). In-hospital mortality was 23%. The mechanical ventilation rate was 8%. The average PCT value was 5.5ng/L. We compared the PCT value of the group of deceased patients (G1) versus G2 survivors: G1: 12.96ng/L; G2: 3.35ng/l; p = 0.001. PCT AUC for predicting mortality was 0.734. In multivariate analysis, a PCT value greater than 0.60ng/l was an independent factor of mortality (OR 3, 74; 95% CI: 0.96–14.8; p = 0.000); with a specificity of 82%. By comparing the PCT value in patients who required mechanical ventilation (GA) vs those not ventilated (GB): GA 22.4ng/L; GB 4.11ng/l; p = 0.000. The AUC for the prediction of mechanical ventilation was 0.600. A PCT greater than 0.8ng/l was an independent factor for the use of mechanical ventilation (OR 1.12; 95% CI 1.06–1.19, p = 0.004).

Conclusion :A PCT greater than 0.6ng/L is an independent factor of intra-hospital mortality and a value greater than 0.8ng/L is a factor associated with mechanical ventilation in patients admitted to the ED for acute dyspnea.

Compliance with ethics regulations: N/A.

Correspondence: Mariem Cheikhrouhou (cheikhrouhou.mariem@yahoo.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-025

Rationale: Chronic critical illness (CCI) characterizes patients admitted to the intensive care unit who, after an acute illness, remain dependent on care. Its incidence varies from 7 to 33% depending on the studied population. The aim of our study was to determine the incidence of CCI in burn patients and its risk factors.

Patients and methods/materials and methods: Retrospective observational case–control study, conducted over a period of 1year (2021) in a severely burned patients’ intensive care unit of Tunisia. The RTI defined CCI by a length of stay more than eight days and one of the following criteria: sepsis, mechanical ventilation (MV) of more than 96h, severe injury, head trauma or stroke and tracheostomy (1). All patients aged over 16years, burned and having been hospitalized for more than eight days were included.

Results: Out of 463 hospitalized patients, 103 patients were included. The average age was 43 ± 18years, the sex ratio was 1.5, the median total burned skin area (TBSA) was 30%. The history was dominated by diabetes (n = 22; 21,4%) and psychiatric disorders (n = 12; 11,7%). The population was subdivided into two groups: 85 patients in the CCI group and eight patients in the non-CCI group. The incidence of CCI was 27%. The eligible criteria for CCI were mainly sepsis (n = 83; 97,6%) and MV (n = 57; 67%). Risk factors for CCI in multivariate analysis were: APACHE II, SOFA, UBS, TBSA, development of acute renal failure and neuropsychiatric disorders, certain hematologic disorders (anemia, thrombocytopenia, lymphopenia) and metabolic disorders (hypernatremia, hypophosphatemia and metabolic acidosis). A TBSA greater than or equal to 21.5% (AUC = 0,759; Se = 68,2% and Sp = 72,2%) and a UBS score greater than or equal to 22 (AUC = 0,768; Se = 76,5% and Sp = 72,2%) predicted CCI. The median of hospital stay was 12days [8 to 128days], longer in the CCI group (p = 0.02). The mortality rate of the CCI group was 55.3%.

Conclusion: Our study showed that CCI is frequent in burn patients with an incidence of 27% and it is associated with a high mortality rate.

Reference 1

1. Kahn JM, Le T, Angus DC, Cox CE, Hough CL, White DB, et al. The Epidemiology of Chronic Critical Illness in the United States. Crit Care Med. févr 2015;43(2):282?7.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Rémy Marnai (rmarnai@ch-lemans.fr)

Annals of Intensive Care 2013, 13(Suppl 1):FC-026

Rationale: The admission and level of care of the very old (≥ 80years) in the ICU, require lengthy upstream discussions and should be focused on the expected benefit to the patient. The main objective of this work was to study the one-year prognosis of patients aged 80years and over, admitted to the ICU with at least one organ failure treated.

Patients and Methods/Materials and Methods: From January 2015 to December 2019, we conducted an observational, retrospective, single-center study in the ICU of the Le Mans Hospital. All patients aged 80years and older admitted during the inclusion period were eligible. Exclusion criteria were: the absence of organ failure during the stay, an absence of specific ICU therapy implemented during the stay (exclusive invasive or non-invasive mechanical ventilation, vasopressor support, high flow oxygen therapy, renal replacement therapy), an admission for scheduled surgery or for a complication related to scheduled surgery, the need for transfer to an expert center (neurosurgery, cardiac surgery) or an admission in anticipation of an organ donation.

Results: From January 2015 to December 2019, among 569 patients aged 80years and older, admitted to the ICU of our hospital, 393 were included. The mean age was 83.71 ± 3.07years; the mean SAPS-2 score was 55.66 ± 19.99; 351 patients (90.93%) lived at home; 223 patients (62.29%) had a Clinical Frailty Score (CFS) ≤ 4; 47 patients (13.62%) had a home care nurse; the mean Charlson score was 5.70 ± 1.67 with a median of 5 (IQR 4–7). Overall, one-year mortality rate was 57.25%. The ICU mortality and hospital mortality rates were 41.73% and 49.56% respectively. In multivariate analysis, SAPS-2 score (HR 1.039 [1.028–1.050], p < 0.0001), age (HR 1.077 [1.024–1.133], p = 0.004) and passage of a home care nurse (HR 1.766 [1.089–2.866], p = 0.021) were significantly associated with an increase of one-year mortality rate. All patients were treated for at least one organ failure, of which 181 (46.06%) received vasopressor support, 275 (69.98%) received invasive mechanical ventilation (19.59%) and 77 patients (19.59%) received exclusive non-invasive mechanical ventilation.

Conclusion: In this observational study conducted in a medical-surgical intensive care unit between 2015 and 2019, the one-year mortality rate was 57.25% in patients aged 80years and over with atleast one organ failure. This mortality rate was significantly correlated with age, high severity score and loss of autonomy characterized by a home care nurse.

Reference 1

Guidet B,et al. The contribution of frailty, cognition, activity of daily life and comorbidities on outcome in acutely admitted patients over 80years in European ICUs: the VIP2 study. Intensive Care Med.2020.

Reference 2

Demiselle J et al. Determinants of hospital and one-year mortality among older patients admitted to intensive care units: results from the multicentric SENIOREA cohort. Annals of Intensive Care. 2021.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Alice Mathieu (alicewanda@wanadoo.fr)

Annals of Intensive Care 2013, 13(Suppl 1):FC-028

Rationale: Patients admitted to intensive care unit (ICU) for severe critical illness experience multi-organ failure with high risk of death. This condition can lead to a potentially long-lasting post-traumatic stress disorder (PTSD). The RésiRéa study aimed to identify patients with PTSD after ICU stay, to assess resilience skills and to have a better understanding in the involvement of some factors in the resilience process such as social support, illness perception or quality of life.

Patients and methods/materials and methods: RésiRéa was a prospective and longitudinal study involving patients admitted in 42 ICUs in France. RESIREA was a parallel study of the NUTRIREA-3 trial, which included patients treated with mechanical ventilation and vasoactive amine for shock. Three months and one year after ICU admission, patients were asked to complete five scalesassessing PTSD (IES-R), resilience (CDRISC), quality of life (SF-36), illness perception (BIPQ) and social support (MSPSS). Resilience was defined by a score at CDRISC ≥ 80 and a low IESR score (absence of post-traumatic stress).

Results: 380 patients were included: 125 women (32.9%), 61 ± 13years old, SAPS-2 56 ± 18, SOFA 10 ± 3. 22.1% [17.9; 26.9] had PTSD at three months and 17.3% [12.6; 22.9] at one year. The rate of patients with CDRISC score ≥ 80 was 18.0% [14.0; 22.6] at three months and 24.3% [18.8; 30.5] at one year. The rate of resilient patients was 12.6% [9.5; 16.4]. In comparison to non-resilient patients, resilient patients had a higher score of social support at three months (MSPSS 78.9 ± 9.3 vs. 70.9 ± 14.3, p < 0.001), a better perception of illness (BIPQ 27.7 ± 8.8 vs 38.2 ± 12.6, p < 0.001) and a better quality of life (PCS 44.1 ± 10.6 vs 38.9 ± 10.8, p = 0.009; MCS 55.3 ± 8.1 vs 47.6 ± 12.4, p < 0.001). These differences still remained one year after inclusion in the study.

Conclusion: Resilience seems to be an important indicator to consider when focusing on how to better support patients after an ICU stay and to reduce the risks of PTSD or increase quality of life. Resilience in ICU patients seems to be influenced by social support and illness perception.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Maite Lacou Agbakou (maite.agbakou@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-029

Rationale: Critical illness survivors are likely to suffer from Intensive Care Unit (ICU) acquired weakness as part of the Post Intensive Care Syndrome (PICS), and poor quality of life (QOL). Handgrip strength (HGS) measured by dynamometer is an easy and reproducible test which reflects global muscle status. We aim to assess the association between HGS at three months after ICU discharge and QOL.

Patients and methods/materials and methods: This monocentric retrospective descriptive study was set in a French University Hospital medical ICU. Our team holds post ICU medical visits (PIMV) three months after ICU discharge for each patient who suffered from a cardiac arrest or underwent mechanical ventilation for at least 48h. During this PMIV, a multimodal evaluation is performed, including a HGS measure with a handheld Jamar hydraulic dynamometer, as well as a QOL questionnaire, the Short Form 36 Health Survey (SF36). We retrospectively collected data from our PIMV register.

Results: 152 patients were seen in PIMV between 8th of August 2020 and 16th of September 2022. There were 108 (71%) men, with a median age of 58 15years old. Main reason of the PIMV was MV longer than 48h (n = 131) followed by cardiac arrest (n = 21; most of them received MV > 48h). The HGS was performed for 121 patients with a median value index of 28.8kg [22–37.8]. We split our population into two subgroups: strong (HGS ≥ 30kg) and weak patients (HGS < 30kg). Differences between strong and weak patients appear in Table 1.

Conclusion: Muscular strength assessed by HGS at three months after ICU discharge is associated with lower cognition but increased physical component of SF36. This result emphasizes the need for interventions promoting physical recovery after ICU discharge in order to improve patients centered outcome.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Estelle Burban (stl.burban@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-030

Rationale: After a stay in intensive care unit (ICU), patients can experience physical, psychological and cognitive impairments. To assess this post intensive care syndrom (PICS), post-ICU consultations have been developed. Since 5-year mortality is higher in ICU survivors, it might be interesting to address those patients' views relating to illness and end of life through a discussion on advance directives (AD), during post-ICU consultation. The aim of this study is to assess the possibility of discussing AD during post-ICU consultation.

Patients and methods/materials and methods: This is a single-center prospective, observational study conducted in a French medical ICU. All patients with at least 48h of mechanical ventilation and/or a 5-day stay in ICU, and then summoned to post-ICU consultation, were included. The primary endpoint was the number of patients willing to discuss AD after an ICU stay. Secondary endpoints were the number of patients knowing about AD before their ICU stay, the factors influencing their writing and the evaluation of the privileged interlocutor to discuss and then complete AD.

Results: A total of 47 patients were included: 72% had never heard of AD and among those knowing about AD, only three had written theirs. For 29 patients (66%), post-ICU consultation appears to be an appropriate time for the discussion of AD. Twenty-one patients (47.7%) relied on a written form to inform them about AD; 15 (34%) on a specialist and seven (15,9%) on their general practitioner. They planned to write AD with their loved one (40.9%) or their attending physician (38.6%) before a specialist, including an ICU practitioner (29.5%). At the end of the consultation, 12 patients (27%) were considering writing AD, 18 patients (40.9%) had no intention to write theirs; the 14 others could not position themselves.Univariate or multivariate analysis did not make it possible to define factors promoting the will to write AD (Figure 1).

Conclusion: As 66% of patients thought the discussion of AD conceivable, the implementation of this discussion should be considered during the post-ICU consultation. However, few patients seem to contemplate their writing following this appointment. Today, more than encouraging the writing of AD, the reflection should evolve towards the development of "Advance Care Planning": a possibility for the patient to reflect gradually on his therapeutic project and end of life, with a reflection which could be initiated in the post-ICU consultation.

Reference 1

Rigaud, J.-P., GELINOTTE, S., BEUZELIN, M., MARCHALOT, A., ERALDI, J.-P., BOUGEROL, F., ECARNOT, F., QUENOT, J.-P., & DECLERCQ, P.-L. (2022). Post-ICU consultation: an indispensable tool in 2022. Médecine Intensive Réanimation, 31(Hors-série1), 79–86.

Reference 2

Lone NI, Walsh TS. Impact of intensive care unit organ failures on mortality during the five years after a critical illness. Am J Respir Crit Care Med. 2012 Oct 1;186(7):640–7. https://doi.org/10.1164/rccm.201201-0059OC. Epub 2012 Jul 26. PMID: 22837381.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Nacim Benchabane (nacim.benchabane@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-031

Rationale: Acute kidney injury (AKI) is associated with high morbidity and mortality and 75% of patients with hematological malignancies (HM) are at higher risk. We aimed to evaluate the impact on long-term prognosis of the occurrence of AKI in patients with HM who survived their critical care admission.

Patients and methods: This is a single-center retrospective cohort study that included patients with HM, admitted to the Intensive Care Unit (ICU) between January 1, 2012 and December 31, 2020. Two groups were distinguished according to the occurrence of AKI. AKI was defined according to the Kidney Disease: Improving Global Outcomes (KDIGO) criterias. The primary outcome was 3-year mortality.

Results: A cohort of 251 patients was analyzed: 64.5% were males and the median age was 63 [52; 69] years old. The median Sequential Organ Failure Assessment was 5 [4; 8]. AKI occurred in 58.6% of the patients, including 36.7% KDIGO stage 1, 31.3% KDIGO stage 2 and 32% KDIGO stage 3. Among these, 25.6% required renal replacement therapy (RRT). Three-year mortality was 57.7% in patients without AKI and 66.7% in patients with AKI (p = 0.27). There was a non-significant trend towards excessive 3-year mortality in patients who required RRT compared to those without AKI or KDIGO stage 1 (78.9% vs. 58.9%, p = 0.07). Recovery of previous renal function was 86.8% without AKI and 69.9% with AKI (p < 0.01). In patients developing AKI, 59.9% were able to continue their hematological course versus 73% without AKI (p = 0.03).

Conclusion: AKI affects more than half of the patients with HM admitted to the ICU. No significant impact on mortality was observed until three years after ICU discharge. Most of the patients recovered their previous renal function. The occurrence of AKI during the ICU stay did not significantly impact the long course of the hematological disease.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Guillaume Rigault (grigault@chu-grenoble.fr)

Annals of Intensive Care 2013, 13(Suppl 1):FC-032

Rationale: Patients with Acute myeloid leukemia (AML) are at high risk of life-threatening complications, leading to intensive care unit (ICU) admission. In this context, mortality remains high, and identifying patients who will benefit the most from early ICU admission is of major concern. In this study, we sought to identify different clusters of patients using their clinical and biological characteristics and investigate their association with outcome.

Patients and methods/materials and methods: We designed an observational, retrospective, monocentric study including consecutive patients with an AML diagnosis admitted in ICU between 2010 and 2022 in a tertiary teaching Hospital. We used a hierarchical clustering method including clinical and biological characteristics available at admission to identify patient sub-phenotypes. The associations between clusters and both 28-day mortality and the need for life-sustaining therapies during ICU stay were evaluated.

Results: Overall, 73 patients with a diagnosis of AML (age 58.2 [47.25—65.82] years, 47.9% male) were included. Two different clusters were identified. Patients in Cluster 1 (N = 37) referred from another hospital, were significantly younger with less comorbidities than patients in Cluster 2 (N = 36) exhibiting higher proportion of hyperleukocytosis. Patients in cluster 2 presented more hypotension, acute renal failure, or coma and demonstrated higher risk for oxygen therapy requirement. Specific management of AML did not differ across clusters, respected to leukocytapheresis use (73 vs 63.9%, respectively for Cluster 1 and 2, p 0 0.457) or hydroxycarbamide administration before intravenous chemotherapy initiation (29.7 vs 33.3, p 0 0.804) and the median time to intravenous chemotherapy initiation (6.4 [-1.1—19.2] vs 7.6 [3.4—15.2] hours, p 0.548) was similar. During ICU stay, patients in Cluster 1 were less treated by vasopressor or renal replacement therapy, without any differences about the use of invasive mechanical ventilation. Twenty-eight days mortality was 13.5% and 47.2% for Cluster 1 and 2 respectively (p = 0.002). This difference is still significant after adjustment on SAPSII and the use of invasive mechanical ventilation (HR4.3 [95%IC, 1.5 to 12]).

Conclusion: Among an ICU population with a diagnosis of AML, two different clusters with different ICU trajectories and outcomes have been identified. Such variability could help to improve the management of AML patients and guide intensivist decisions most suitable for each patient at the bedside.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Louise Baucher (louise.baucher@hotmail.fr)

Rationale: In the last decade, ibrutinib has become the standard of care in the treatment of several lymphoproliferative diseases. Beyond Bruton tyrosine kinase inhibition, ibrutinib shows broad immunomodulatory effects that may promote the occurrence of infectious complications, including opportunistic infections. The infectious burden has been shown to vary by disease status, neutropenia and prior therapy but data focusing on severe infections requiring intensive care unit (ICU) admission remain scarce.

Patients and methods/materials and methods: The medical records of hematological patients receiving ibrutinib admitted to 10 ICU departments in France from January 1st, 2015 to December 31st, 2020 were retrospectively analyzed. Severe infections in patients receiving ibrutinib were analyzed and compared to severe infections in patients receiving conventional chemotherapies from an historic cohort (TRIAL-OH¹) matched on age and underlying malignancy.

Results: Among 92 critically-ill patients receiving ibrutinib, 69 (75%) patients were admitted to the ICU for severe infections. The most common underlying malignancy was chronic lymphocytic leukemia (59%). Forty-seven (68%) patients received ibrutinib as monotherapy. The median interval from ibrutinib initiation to ICU admission was 6.6 [3–18] months. Severe infections were mostly bacterial infections (71%), while viral and fungal infections accounted for 17% and 12% respectively. Opportunistic infections included invasive aspergillosis in six (9%) patients and Pneumocystis pneumonia in two (3%) patients. Thirty-nine (57%) patients received vasopressors and 28 (41%) required mechanical ventilation. Twenty (29%) patients died in the ICU and day-90 mortality reached 55%. In comparison with decedents, survivors displayed more severe organ dysfunctions at ICU admission (SOFA 7 [5–11] vs. 4 [3–7]; P = 0.004) but there was no difference regarding age, underlying malignancy, prior therapy and infection type. Sixty-three patients receiving ibrutinib were matched with 63 patients receiving conventional chemotherapy. Age, sex ratio and underlying malignancies did not differ between the two groups. Among these patients, chronic lymphocytic leukemia was the most common underlying disease (65% and 64% respectively). Despite a higher median number of prior treatment regimens (3 vs. 1; P < 0.001) and a higher rate of fungal [13% vs. 8%] and viral [19% vs. 5%] infections; [P = 0.023], patients receiving ibrutinib displayed similar outcomes in comparison with patients receiving conventional chemotherapies (ICU mortality: 27% vs. 38% [P = 0.254]; day-90 mortality: 52% vs. 48% [P = 0.785]).

Conclusion: In hematological patients receiving ibrutinib, severe infections requiring ICU admission are associated with a dismal prognosis, mostly impacted by initial organ dysfunctions. Due to the high proportion of fungal infections, opportunistic agents should be routinely screened by ICU clinicians in this immunocompromised population.

Reference 1

Azoulay E, Mokart D, Pène F, Lambert J, Kouatchet A, Mayaux J, Vincent F, Nyunga M, Bruneel F, Laisne LM, Rabbat A, Lebert C, Perez P, Chaize M, Renault A, Meert AP, Benoit D, Hamidfar R, Jourdain M, Darmon M, Schlemmer B, Chevret S, Lemiale V. Outcomes o.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Antoine Herault (antoine-herault@hotmail.fr)

Annals of Intensive Care 2013, 13(Suppl 1):FC-034

Rationale: Severe infections requiring admission to the intensive care unit (ICU) often lead to rapid loss of skeletal muscle mass. These changes in body composition can be measured by computed tomography (CT) images. The objective of this study was to evaluate the prognostic impact of pre-existing sarcopenia measured by CT in patients with hematologic malignancies (HM) admitted to the ICU for sepsis or septic shock.

Patients and methods/materials and methods: We performed a retrospective study of 186 patients with HM who had a first hospitalization in ICU between 2013 and 2020 at the Rouen University Hospital for sepsis or septic shock. We used CT images collected as part of routine care in the 2months preceding ICU admission to assess baseline body composition by manual segmentation of slices on the third lumbar vertebra (L3) and twelfth thoracic vertebra (T12).

Results: The mean age of patients was 54.1 (15.5) years. The sex ratio (M/F) was 1.90. Patients displayed severity criteria at admission: median SAPS II [q1; q3] of 52[40;66]and a median SOFA score of 8[5;12]. We classified as sarcopenic 64.9% of patients on L3 slices and 83% on T12 slices on baseline CT images. There was a good correlation between skeletal muscle measurements on L3 and T12 sections (r = 0.77). The raw ICU mortality rate was 45.7%. The overall survival rate at 1month from ICU admission in patients with pre-existing sarcopenia versus patients without pre-existing sarcopenia was 47.9% (95% IC[37.6; 61.0]) versus 55.0% (95% IC[41.6; 72.8]), p = 0.99) for measurements performed in L3; and 48.4% (95% IC[40.4; 58.0]) versus 66.7% (95% IC[51.1; 87.0], p = 0.062) for measurements performed in T12, respectively.

Discussion: Mortality rate in our cohort of HM patients with severe infection was particularly high. The majority of patients in the cohort had pre-existing sarcopenia before ICU admission. Patients with T12-assessed pre-existing sarcopenia had a lower overall survival. Pre-existing sarcopenia might be influenced by the underlying HM. We observed hydro sodic inflation responsible for tissue edema that occasionally hampered the manual segmentation and body composition assessment on L3 images.

Conclusion: Pre-existing sarcopenia may be regarded as a risk factor for severe infection in HM patients and contributed to the high mortality observed in this ICU population.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Julien Schmidt (julien.schmidt@aphp.fr)

Annals of Intensive Care 2013, 13(Suppl 1):FC-035

Rationale: Hemophagocytic syndrome (HS), a rare and life-threatening disease, encompasses a wide spectrum of heterogeneous conditions, all resulting in severe systemic inflammation. Systemic lupus erythematosus (SLE) flare-up can mimic reactive HS, both entities sharing common features can frequently lead to ICU admission.

Patients and Methods: This study aims to focus on SLE patients with reactive HS admitted to ICU, regarding clinical, laboratory and pathological features as well as treatment strategies and prognosis. It is a retrospective multicenter study including SLE patients with reactive HS in 9 French and 1 Belgian ICUs over 12years.

Results: Twenty-one patients were included in the study. Main reason for ICU transfer was hemodynamic failure in 52.4% of patients, acute respiratory failure in 19%, altered mental state in 14.3%, acute liver failure in 4.8% and monitoring in 9.6%. Median HS score was 201 and was > 168 in 100% of patients. Besides SLE, another associated condition was found in 12 (57.1%) patients. 95.2% of patients were treated with corticosteroids. Etoposide was used in 47.6% of patients, cyclophosphamide in 33.3%, MMF in 9.5%, plasma exchange therapy in 14.3%, tacrolimus, rituximab and polyvalent immunoglobulins in 4.8% each. Catecholamine infusion was used in 52.4% of patients; 52.4% of patients underwent invasive mechanical ventilation; 19% of patients required renal replacement therapy. Overall ICU mortality was 9.6% and overall 28-days mortality was 14.3%.

Discussion: Reactive HS is a rare complication of SLE, inaugural in more than half of patients; distinction between SLE flare-up, sepsis and non SLE-related reactive HS is tricky, as most clinical and biological symptoms are non-specific. Yet, given the significant need for organ support therapy among young patients without comorbidities, early recognition is crucial. To that extent, multi-organ failure in young patients with suggestive clinical and biological signs should promptly lead to a biological and immunological workup to assess for both SLE and reactive HS; ferritin assessment may be especially useful in this setting when cytopenia tends to be less frequent. Severe bleeding events seem to be frequent and should be considered before performing non-urgent invasive procedures. Front-line therapy needs to be clarified, are both SLE and reactive HS with severe presentation require distinct kinds of immunosuppressive therapy; randomized controlled trials and prospective data are most needed.

Conclusion: Reactive HS is a rare complication of SLE. Yet, given the significant need for organ support therapy among young patients without comorbidities, early recognition is crucial. Front-line therapy needs to be clarified.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Aïcha Rebai (aicha.rebai@fmt.utm.tn)

Annals of Intensive Care 2013, 13(Suppl 1):FC-036

Rationale: Since the beginning COVID-19 pandemic, an increased risk of venous or arterial thrombotic events has been observed in infected patients [1]. Many studies have described the presence of anti-phospholipid antibodies (aPL). Aim of the study: This study aimed to determine the prevalence and the kinetics of APL in COVID-19 patients and their correlation to thromboembolic disease.

Patients and methods/materials and methods: In a prospective study, 43 critically-ill COVID-19 admitted to the intensive care unit of the Military Hospital of Tunis have been consecutively enrolled from October 01, 2020 to April 30, 2021. Anti-cardiolipin antibodies (aCL), antiphosphatidylserine antibodies (aPS), D-Dimers, fibrinogen and C-reactive protein (CRP) have been measured with enzyme-linked immunosorbent assay and the SOFA score, the Disseminated Intravascular Coagulation score (DIC) and the Sepsis Induced Coagulopathy score (SIC) were estimated on day 1, 3, 5, 8 and 10 for each patient. At a final point, patients were compared based on the occurrence of a thromboembolic event. Data were analyzed with SPSS version 25.0.

Results: 58% (25/43) of the study population were positive for APL distributed as follow: 23 had positive aCL, 12 positive aPS and 10 were positive for aCL and aPS. When comparing patients based on the presence of APL, there was no significant difference between the demographic criteria between the two groups. The length of stay in intensive care was 19 ± 9days and the mortality rate was 79%. The severity scores on admission, at D3, D5, D8 and D10 were comparable between the two groups. Thirteen patients presented a thromboembolic event, ie 30.23% of the population, with no significant difference between the two populations.

Discussion: The incidence of APL and their correlation to thromboembolic events in COVID19 patients varies from one population to another; while our results are consistent with some previous studies, others have reported a cumulative incidence of thrombotic complications in patients with a severe form of COVID-19 admitted to intensive care.

Conclusion: In our study, there was no correlation between APL status and the occurrence of thromboembolic events. To be confirmed, it would be essential to carry out a multicenter study with a larger workforce, by measuring the different subtypes of APL and correlating them with the occurrence of thrombosis, but also with the evolution of the pro-coagulant inflammatory status of COVID19 patients.

Reference 1

Bilaloglu S, Aphinyanaphongs Y, Jones S, Iturrate E, Hochman J, Berger JS. Thrombosis in Hospitalized Patients With COVID-19 in a New York City Health System. JAMA. 2020;324(8):799–801.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Jeremie Joffre (jeremie.joffre@aphp.fr)

Annals of Intensive Care 2013, 13(Suppl 1):FC-037

Rationale: Immune-mediated thrombotic thrombocytopenic purpura (iTTP) is a rare disease characterized by arteriolar and capillary microthrombosis precipitating organ failure. However, the contribution of endothelial dysfunction on impaired microvascular blood flow in iTTP patients has been poorly explored. This pilot observational study aimed to explore endothelial-mediated vasoreactivity in iTTP patients at admission and its changes after plasma exchange therapy (PE).

Patients and methods/materials and methods: We conducted a prospective observational study in patients (> 18-yr old) admitted in ICU for iTTP. Using laser Doppler flowmetry and iontophoretic acetylcholine (Ach) in the forearm, we recorded the skin microvascular blood flow and the endothelium-mediated vasoreactivity at admission and after PE. Demographics, biological, clinical courses, and outcomes were also collected. As a control group, we used a previously published cohort of young diabetic patients after correction of ketoacidosis.

Results: Eighteen confirmed iTTP patients and 34 controls were included in the study, mainly female (72%) aged 43 ± 16-year-old. At admission, 55% had neurological abnormalities, 50% cardiac issues and 27.8% an acute kidney injury. Median platelet count was 19 G/mL [10–37]. Baseline microvascular blood flow was decreased in iTTP patients when compared to controls (5.97 ± 4.5 vs. 10.1 ± 6.3 PU, P = 0.03), associated with markedly impaired endothelial-mediated skin microvascular reactivity (AUC: 9627 ± 8122 vs. 16,475 ± 11,738, P = 0.03). Microvascular reactivity improved after the first PE session (AUC: 9627 ± 8122 vs 16,558 ± 10,699, P = 0.007, respectively baseline and post PE1) and much more after the second session (26,431 ± 23,181, P = 0.04 post PE1 vs post PE2). Hemolysis biomarkers (LDH and bilirubin) negatively correlates with microvascular flow and vasoreactivity.

Conclusion: We highlighted a reversible marked endothelium-mediated microvascular hyporeactivity in iTTP patients that could participate in organ injury pathophysiology.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Francesco Pini (francesco.pini@hotmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-038

Rationale: With the rising use of immune checkpoint inhibitors (ICIs) in cancer treatment, emergency physicians are increasingly confronted with their adverse effects. These immune related adverse events (irAEs) can affect multiple organs and present in a variety of ways, making their diagnosis complex. This study aims to evaluate the acute management of patients undergoing an ICI treatment and presenting to the emergency department (ED).

Patients and Methods: A retrospective analysis of emergency room visits of patients on ICI was performed in an oncology referral center. The study was spread over a period of four years, from 15/12/2016 to 06/12/2020. The primary objective was to evaluate the chief complaint for consultation at the ED. The secondary objectives were to characterize irAEs, to describe the therapeutic management, as well as to evaluate survival 7-day and 30-day post-consultation.

Results: 227 patients on ICI presented to the ED, with a total of 452 visits. 55 (12.2%) of the visits resulted in a diagnosis of irAE, from which 44 (80%) were suspected by emergency physicians. The most common irAEs were colitis and diarrhea (25.5%), followed by pneumonia (23.6%), hepatitis (9.1%), and dermatitis (9.1%). Two chief complaints were associated with an irAE: gastro-intestinal complaints (p < 0.006) and skin rashes (p < 0.02). Another association was found between an irAE and three different factors: a cancer status in remission (OR = 5.43, 95% CI 2.75–10.7), a combination of two ICIs (OR = 4.41, 95% 2.08–9.37) and any medical history of irAE (OR = 2.34, 95% 1.26–4.3).

Conclusion: Diagnosing an irAE can be particularly complicated in the emergency department. Two chief complaints should raise the suspicion of physicians of facing an adverse event: gastro-intestinal and dermatological ones. It is also essential to look for predictive factors of developing an adverse effect: any history of irAE, a concomitant use of two ICI or patients with a cancer status in remission.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Jihene Ayachi (ayachijihen@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-039

Rationale: Morbidity and mortality associated with COVID-19 infection has prompted us to understand potential risk factors that can predict patient outcomes. Little is known about most common underlying diseases in COVID-19 patients and were they really associated with poor outcomes. Aim: To describe the prevalence of underlying diseases in critically ill COVID-19 patients and determine underlying diseases predictors of poor outcomes of COVID-19 infection.

Patients and methods/materials and methods: A retrospective analytical cohort study was performed from September 2020 to December 2021 in a 9-bed medical ICU of a university hospital including patients admitted for hypoxemic acute respiratory failure secondary to COVID-19 pneumonia. Characteristics, underling diseases, severity scores, management and outcomes are recorded. Univariate and multivariate analysis were used to identify underlying diseases independently associated with poor outcomes. The primary outcomes were adverse events and mortality.

Results: During the study period, 172 patients were included. Mean age was 51.5 ± 14.4years with male predominance 104 (60.5%). Median SAPSII and Charlson scores were respectively 24 [18–32] and 1 [0–2]. Eighty (46.5%) patients had underlying conditions. The most common comorbidities were obesity 60 (34.9%), diabetes 48 (28.5%) and hypertension 30 (17.4%). Twelve (7%) patients had underlying lung diseases and 5 (2.9%) patients had immunodeficiency. The median PaO₂/FiO₂ ratio was 92 [69–136]. Invasive mechanical ventilation was performed in 69 (40.1%) cases. The most frequent adverse events were: shock 68 (39.5%), nosocomial infection 58 (33.7%), acute renal failure 51 (29.7%) and acute metabolic complications of diabetes 31 (18%). Median ICU length of stay was 9 [6–14] days and death rate was at 40.7%. On univariate analysis, two underlying diseases were associated with poor outcomes. Mortality is associated with hypertension (p = 0.005) and diabetes (p = 0.037) and occurrence of adverse events is also associated with hypertension (p = 0.007) and diabetes (p = 0.004). On multivariate analysis, hypertension is an independent predictive factor of mortality (OR, 3; 95%CI, [1.36–6.98]; p = 0.007) and diabetes is independently associated with the occurrence of adverse events (HR, 2.64; 95%CI, [1.34–5.21]; p = 0.005) in critically ill COVID-19 patients.

Conclusion: This study showed that patients with pre-existing hypertension and diabetes should take all necessary precautions to avoid getting infected with COVID-19, as they seem to be more vulnerable to developing severe outcomes as occurrence of adverse events and mortality.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Yahyaoui Hajer (hajer.yahyaoui94@hotmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-040

Rationale: Emerging SARS COV2 has been challenging intensive care units all over the world, not only by the increasing number of admissions, but also by the management of severe forms that require special knowledge of this disease’s characteristics. Our objective is to study epidemiological and clinical features and outcomes of COVID 19 critically ill patients and to identify mortality associated factors.

Patients and methods/materials and methods: Descriptive, retrospective and monocentric study including all patients admitted in an intensive care units (ICU) department between 01/09/2020 and 30/06/2022. All patient’s data including demographics, clinical, bacteriological, therapeutical and outcomes were collected.

Results: During the period of the study, 450 patients infected by the SARS COV2 were admitted to our unit. Sex ratio was 1.14 and mean age was 63 ± 15. 46.8% and 42.6% of patients had respectively arterial hypertension and diabetes mellitus and 51% had obesity. Mean SAPSII, APACHEII and SOFA score were respectively 30 ± 39, 30 ± 39 and 8,6 ± 5. Diagnosis of COVID 19 infection was made by RT-PCR in 85 cases (19%), by an antigenic test in 250 cases (55.5%), and by a chest CT scan in 115 cases (25.5%). The percentage of lung parenchyma affected in the chest CT scan was less than 25% in 26 patients (6%), 25 to 50% in 118 patients (26.5%), 50 to 75% in 153 patients (34.5) and above 75% in 147 cases (37%). It revealed a pulmonary embolism (PE) in 44 cases. 19.5% of our patients needed invasive mechanical ventilation (IV) (n = 88). Among ICU complications, at least one nosocomial infection occurred in 95 patients, and 77 patients had septic shock. Mean duration of stay was 10.7days ± 6 and mortality were 35.5%. Several factors associated significantly to mortality were identified: Delay of admission > 12days (p = 0.049, OR:1.81; CI: 1.O1-321), sever ARDS at admission (p = 0.010, OR:2,74; CI: 1.2–5,7), IV (p < 10^–3, OR:75,7; CI: 14–407), nosocomial infection (p < 10^–3, OR: 6,3; 95% CI: 2,5–15,7), sever affection at CT scan ( p = 0,003, OR:18.1; 95% CI: 1.01–12) and APACHE II score > 6 (p = < 10–3, OR: 8,4; 95% CI: 4,6–19,4).

Conclusion: Patients admitted in ICU for a SARS COV2 infection are mostly patients with at least one comorbidity. Sever forms predispose to ICU complications and are associated to a higher mortality.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Yahyaoui Hajer (hajer.yahyaoui94@hotmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-041

Rationale: In the COVID 19 pandemic, and since the RT-PCR tests are less available and take more time, it was primordial to find an efficient and rapid way to make the diagnosis of SARS COV2 infection. The chest computed tomography (CT) scan revealing a typical aspect of the COVID 19 had a crucial contribution in this pandemic. Our objective isto describe the role of the chest CT in diagnosis and prognosis in COVID 19 patients admitted in the intensive care unit (ICU).

Patients and methods/materials and methods: Descriptive and prospective study including patients admitted in an ICU department between 01/09/2020 and 30/06/2022. All patients data including demographics, clinical, biological, radiographic and outcomes were taken from patient’s medical files.

Results: In the period of the study, a total of 450 patients infected by the SARS COV2 were admitted in our department. Sex ratio was 1.14. Mean age, SAPSII, APACHE II and SOFA score were respectively 63years 30 ± 39, 30 ± 39 and 8.6 ± 5. At admission, 444 patients had a chest CT scan. The diagnosis of COVD 19 pneumonia was made with the contribution of the chest CT in 115 cases. Ground-glass opacities and crazy paving were found respectively in 82% and 70% of cases. Radiological signs of associated bacterial infection were found in 49 cases (11%). Quantification of lung lesions was moderate, important, severe, and critical in 26 (6%), 118 (26.5%), 153 (34.5%), and 147(33%) patients respectively. Other findings were reported including consolidation (409%).The chest CT revealed 47 cases (10.4%) of pulmonary at admission. Sever ARDS at admission and consultation delay > 12days were identified as an independent risk factor predictive of a severe to critical pulmonary damage (superior to 50%) on the chest CT with respective (OR: 2.4; 95% CI: 1.5–4.1; p < 10–³) and (OR: 7.7; 95% CI: 3.2–18.2; p < 10^–3). Mortality was 35.5%. Sever to critical pulmonary damage (superior to 50%) on the imaging was associated to a higher mortality (p < 10^–3).

Conclusion: The chest CT was identified as a swift, accessible, and reliable diagnostic modality that served not only as analternative to the RT-PCR but also helped in assessing prognosis by quantifying lung lesions and revealing complications in critical ill patients.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Amira Marouf (amira.marouf.pro@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-042

Rationale: Thymic activation improves the outcome of COVID-19 patients with severe pneumonia.¹ The rs2204985 genetic polymorphism within the TCRA-TCRD locus, which affects thymic output in healthy individuals, was found here to influence SARS-CoV-2-specific immunity and disease severity in COVID-19 patients with severe pneumonia.²

Patients and methods/materials and methods: Forty consecutive patients hospitalized in intensive care unit for severe COVID-19 with pulmonary involvement, who gave written consent for TCRA-TCRD locus DNA sequencing, were included. Thoracic CT-scans were performed for the initial evaluation of pulmonary status. All images were independently reviewed and classified by two radiologists. DNA sequencing of the rs2204985-containing region in Covid-19 patients was performed and correlated with the extent of lung involvement, the level of thymic function and the strength of immune response in those patients. Thymic function was estimated through the quantification of T cell receptor excision circles (TRECs) and peripheral T cells phenotyping. Patient’s SARS-CoV-2 specific immune response was assessed by quantification of anti-SARS-CoV-2T-cell responses and plasma containing SARS-CoV-2 neutralizing antibodies.

Results: The GG genotype at the rs2204985 locus was associated with less severe lung involvement: only 54.5% of GG patients showed a CT-scan score of 3 or higher, compared to 83.3% and 100% in GA and AA patients, respectively (logistic regression model of AA + GA to GG, odds ratio = 0.16, 95% confidence interval: [0.03–0.8], p = 0.032; Figure 1A). Despite a comparable age distribution according to the genotype, the GG patients displayed a higher median sj/βTREC ratio compared to the combined AA + GA group (p = 0.042; Figure 1B) suggesting an increased thymic production of new lymphocytes. The GG genotype, independently of age and sex, was also associated with higher counts of blood naïve T lymphocytes including recent thymic emigrants, higher frequency of neutralizing antibodies during acute infection and with stronger anti-SARS-CoV-2 helper and cytotoxic T-cell responses six month after recovery.

Conclusion: We report the first evidence that rs2204985 polymorphism can impact the clinical and immunological responses to severe viral infections, by affecting both the baseline production of T-cells by the thymus and its enhancement during the acute phase of the infection. Patients with the GG genotype, who display a greater basal thymic activity and a larger naïve T-cell pool, were able to mount stronger and more efficient T-cell responses against SARS-CoV-2. The analysis of the rs2204985 sequence variants might be included among the predictive tests in critically ill COVID-19 patients and more generally in severe infections and other clinical situations where stimulated thymopoïesis has mechanistic and/or prognostic importance.

Reference 1

Cuvelier, P. et al. Protective reactive thymus hyperplasia in COVID-19 acute respiratory distress syndrome. Crit Care 25, 4 (2021).

Reference 2

Clave, E. et al. Human thymopoiesis is influenced by a common genetic variant within the TCRA-TCRD locus. Science Translational Medicine 10, eaao2966 (2018).

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Laurent Camous (laurent.camous@chu-guadeloupe.fr)

Annals of Intensive Care 2013, 13(Suppl 1):FC-043

Rationale: Describe epidemiology, outcome of and death risk factors of Solid Organ Transplant (SOT) patients admitted in the intensive care unit (ICU) of the Guadeloupe University Hospital. Six months graft function and post ICU immunologic and infectious events were also recorded.

Patients and methods/materials and methods: All SOT patients admitted in the ICU between January2014 to December 2022 were included in the study. Three groups were defined according to time from organ transplantation (< 1year, 1–5years and > 5years). Using univariate and multi-variate analysis, we assessed risk factors associated with death.

Results: Among the 199 patients admitted to the ICU on the study period, 57 (29%) were assigned to the early group (< 1year), 80 (40%) to themiddle group (1–5year), and 62 (31%) to the late group (> 5years from transplant). Overall median age and SAPS II were 60years and 42 respectively. In the early group, life threatening events leading to ICU admission were mostly linked to transplantation procedure related complications (including immunosuppressors toxicity) differing statistically from the two other groups. In the middle and late group, infectious diseases were the leading cause for ICU admission (65% and 56% respectively). Severe cardiovascular events were mostly seen in the early and middle group. Overall mortality rate at 3months after ICU admission was 30%. Risk factors for death in multivariate analysis were shock, mechanical ventilation during ICU course and SARS Cov2 infection as cause for ICU admission. Original nephropathy and previous immunosuppression before transplantation were not associated with a worse outcome. Immunosuppressors discontinuation during ICU course was not associated with graft rejection at 6months.

Conclusion: ICU admission in solid organ transplant recipients is frequent, with several etiologies. Pattern of ICU admission is different, according to time to transplantation time. In the early period (< 1year), transplantation procedures related etiologies are predominant and after 1-year, infectious events are predominant. In multivariate analysis, mechanical ventilation, shock and SARS COV2 infection were independently predictive of death. The burden of SARS COV 2 infection in SOT patients was major as the pandemic period spanned less than 1/3 of the study period.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Iyed Matouk (maatouk.yed@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-044

Rationale: Because of the frequent association between COVID-19 infection and thromboembolic events in patients hospitalized in the ICU, anticoagulation was often prescribed. Nevertheless, anticoagulation treatment can lead to serious complications such as bleeding. We aimed to describe the characteristics of patients hospitalized with COVID-19 pneumonia who developed a hemorrhagic complication.

Patients and methods/materials and methods: We conducted a retrospective study between 2020 and 2022 in the Intensive Care Unit (ICU) of a Tunisian University Hospital. We included COVID-19 patients aged more than 18years admitted to ICU for severe ARDS due to SARS-COV2.

Results: During the study period, 433 patients were admitted to our ICU for COVID-19 ARDS. The majority were male (66.7%). Mean age was 59.6 ± 13years. Obesity was found among 38.3% of patients. Median SAPSII was 29 (Interquartile range (IQR): 23.7- 35). Median APACHE was 6 (IQR: 9–12). The mean SOFA score was 3 (IIQ: 3–4). Median length of stay was 10 (IQR: 6–14) days. Invasive Mechanical ventilation was required in 43.3% of patients. In total, 20 patients presented hemorrhagic complications secondary to anticoagulation (5%). Among them, 60% were male. Their mean age was 65.6 ± 11years. Regarding COVID-19 disease, 60% of the patients had moderate to severe lung injury. All patients received anticoagulation, of which 90% was preventive. Low molecular weight heparin was prescribed in almost all patients (80%). Out of the 20 cases, 9 patients had a digestive hemorrhage (45%), one patient had a muscle hematoma (5%) and one patient had a retroperitoneal hematoma (5%). Other patients had minor complications (epistaxis, gingivorrhage or ecchymosis). Only one patient presented a hemorrhagic shock and 8 patients presented a deglobulation without hemodynamic repercussions. No patient had interventional treatment, and 40% required transfusion. Eighty percent of patients died secondary to septic shock or refractory hypoxemia.

Conclusion: The prevalence of hemorrhagic complications among COVID-19 patients was low in our study. They are dominated by severe digestive hemorrhage and epistaxis. Other studies should be conducted to assess the risk–benefit ratio of this treatment among COVID-19 patients known by their state of hypercoaguability.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Geoffrey Garcia (geoffrey.garcia@inserm.fr).

Annals of Intensive Care 2013, 13(Suppl 1):FC-045

Rationale: Immunothrombosis is increased in severe COVID-19, and potentially involved in its pathogenesis but the mechanisms explaining this deregulation are currently not understood. We hypothesized that decreased DNase activity was associated with increased NETosis and clinical worsening. We aimed to compare the balance between NETs and DNase activity according to COVID-19 severity and investigate the mechanisms responsible for the imbalance between NETosis and DNase activity.

Patients and methods/materials and methods: We studied 93 non-severe (outpatients), 27 severe (hospitalized) and 37 critical COVID-19 patients from COLCOV19-BX and COVERAGE studies. We quantified NET markers (MPO-DNA, H3Cit, H3cit-DNA), total DNase activity and the levels of DNase 1 and DNase1L3 proteins in plasmas from the 157 patients. In the 54 patients that were hospitalized, we sequenced DNase1 and DNase1L3 genes and used flow cytometry to quantify the numbers of plasmacytoid dendritic cells (pDC) and dendritic cells (DC) among peripheral blood mononuclear cells. Lastly, we analyzed scRNAseq public data from COVID-19 patients.

Results: Plasma NET markers were higher in the most severe patients while DNase activity was lower. Critical patients showed decreased level of DNase 1 & 1L3 proteins compared to severe patients. We did not find any recurrent mutations in DNase1 and DNase1L3 that was associated with decreased amount of DNase1 or DNase1L3. We observed a quantitative defect of DC and pDCs in critical patients. Besides, analysis of public scRNaseq data revealed an inverse correlation between DNase1L3 RNA expression in pDC and disease severity.

Conclusion: Altogether, our study shows that severe and critical COVID-19 is associated with an imbalance between NETs and DNase activity, with decreased amount of DNase 1 and DNase1L3. We speculate that the defect in DNase1L3 may be due to a defect in DCs and pDCs. Early identification of patients with NETosis imbalance could allow administration of targeted therapies, such as DNase, to prevent clinical worsening.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Salma Ghalloussi (salmaghalloussi93@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-046

Rationale: The resurgence of COVID-19 cases since June 2021, referred to as the fourth COVID-19 wave, has led to the approval and administration of repeated vaccine doses. Our study aims to investigate the association between vaccination status and clinical presentation, biological features, amount of oxygen support and outcome among patients admitted in ICU for critical SARS-COV-2 infection.

Patients and methods/materials and methods: A retrospective, observational study conducted from January 2021 to 30 August 2022 carried out in the medical ICU of the University Hospital Center of la Rabta-Tunisia. Were included all adult patients admitted for critical SARS-CoV-2 infection after generalization of vaccination. Patients were divided into four groups according to their vaccination status: Unvaccinated, receiving one dose, two doses and three doses or more. Data related to clinical and biological features, oxygen support needed and outcome were collected and compared according to vaccination status.

Results: 237 patients were included. Between groups, Age > 60years (p = 0.007) and co- morbidities including hypertension (p = 0.001), immunodeficient status (p = 0.037), chronic heart failure (p = 0.034), chronic respiratory failure (p = 0.011) and chronic kidney disease (p = 0.013) were significantly more reported among vaccinated patients with 3 doses or more. PaO₂/FiO₂ ratio was significantly correlated to the dose received (p < 10^–3). Severe to critical CT scan lesions (p = 0.002), Severe ARDS (p < 10^–3), mechanical ventilation (p = 0.003) and mortality (p = 0.04) were lower among vaccinated patients with 3 doses or more.

Conclusion: Vaccination was found to be a protective factor, since vaccinated patients presented moderate symptoms, less biological disorders, CT scan extent, critical forms and need of mechanical ventilation. This protective effect was, for some parameters, more found among patients vaccinated with 3 doses or more.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Manel Lahmar (firassmal4@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-047

Rationale: The clinical presentation during respiratory infection with SARS-CoV-2 is similar to the infection with other respiratory pathogens; however, the management and prognosis are different. The search for these agents has been reduced during the covid pandemic but does not prevent their co-existence with SARS-CoV-2. The objectives of this study were to identify the bacteria and viruses responsible for respiratory infections using the multiplex PCR diagnostic test and characterize clinical profiles during the COVID-19 pandemic.

Patients and methods/materials and methods: An observational study was performed between January 2020 and September 2022, in the Intensive Care Unit (Monastir, Tunisia), including patients admitted with hypoxemic respiratory failure. Respiratory samples were analyzed using a qualitative multiplex PCR technique (RespiFinder). Diagnosis of Covid-19 infection was based on COVID PCR test or chest CT scan.

Results: During the study period, 352 COVID-19 patients with hypoxemic respiratory failure were admitted to our ICU. PCR multiplex was made in 60 patients given the unavailability of this technique during the COVID-19 pandemic. The median age of these patients was 50years [IQR: 41–70] and the median SAPS II score was 31 [IQR: 20–39]. The main clinical signs presented by the patients were dyspnea and cough. The multiplex PCR was positive in 45% of cases. When the PCR panel was performed, 86% of patients were receiving antibiotic treatment. All germs identified were viruses and no bacterial infection was detected in our study: 14 cases of Influenza A (H1N1), 3 cases of VRS B, 3 cases of coronavirus 229E, 1 case of coronavirus HKU1, 3 cases of coinfection Rhinovirus/Enterovirus, 1 case of coinfection coronavirus NL63/HKU1, 1 case of coinfection coronavirus 229E/VRS B and 1 case of coinfection Influenza B/Adenovirus. The diagnosis of SARS-Cov2 infection was observed in 28.3% of cases, 23.5% of them were coinfections (4 cases). During hospitalization, 69% of these patients required invasive mechanical ventilation.

Conclusion: Pathogenic microorganisms with respiratory tropism still exist during.

The COVID-19 pandemic and may be responsible for severe respiratory infections. Their research must be performed simultaneously with the SARS-CoV-2 diagnostic test to adapt the therapeutic management of patients.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Anne-Françoise Rousseau (afrousseau@chuliege.be)

Annals of Intensive Care 2013, 13(Suppl 1):FC-048

Rationale: Alterations in oropharyngeal (OP) microbiota have been associated with viral infections, with potential bi-directional interactions. In published studies, the OP of COVID-19 patients was mostly compared to healthy patients, with inconsistent observations. The aim of this monocenter prospective study was to compare the OP bacterial microbiota profile in patients admitted to the hospital for COVID-19 and for stroke.

Patients and methods/materials and methods: Patients admitted in our tertiary hospital either for COVID-19 pneumonia diagnosed with a positive polymerase chain reaction (PCR) for SARS-CoV-2 in nasal swab or for an ischemic or hemorrhagic stroke (with negative SARS-CoV-2 PCR) were enrolled respectively during the second and third waves of the pandemic, and between September and December 2021. COVID-19 patients were admitted to general wards or directly to the intensive care unit (ICU). Exclusion criteria were antibiotic exposure during the past 15days, coagulopathy, endotracheal intubation and pregnancy. OP swabs were taken from the hard palate within the first 24h after admission and used for a 16S rRNA gene-targeted metagenomic analysis. Demographic data, medical history, and in-hospital deaths were recorded.

Results: A total of 135 swabs were taken, and 83 were analyzed: 55 in the COVID group (65 [64–69] years, 22 women), including 16 patients in an ICU subgroup, and 28 in the stroke group (72 [64–80] years, 11 women). The multivariate analysis showed that the reason for admission significantly influenced the beta diversity (p < 0.001). The beta diversity in the two subgroups of COVID patients was similar. The beta diversity was not influenced by age, smoking habits, or proton pump inhibitor treatment and was not associated with in-hospital mortality. Alpha diversity in COVID patients (population richness and population diversity) and the stroke group was not significantly different. At the genus level, 38 genera were identified with significant relative abundance differences between COVID and stroke groups. Among them, 34 taxa, including the genusVeillonellaand familyPrevotellaceaewere significantly increased in COVID patients, while 4 taxa, including the genusStreptococcuswere significantly depleted. The microbiota was dominated by taxa belonging to 4 phyla: Firmicutes (Streptococcus, Veillonella, Gemella), Bacteroidota (Prevotella, Prophyromonas), Fusobacteriota (Fusobacterium, Leptotrichia) and Proteobacteria (Neisseria).

Conclusion: In this study, a significant difference in OP bacterial microbiota was observed in COVID-19 patients, including more pro-inflammatory genera, compared to other newly admitted patients for a stroke. It can be assumed that this dysbiosis occurred earlier during viral infection. Moreover, this dysbiosis was not associated with respiratory failure severity.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Renaud Prével (renaud.prevel@hotmail.fr)

Annals of Intensive Care 2013, 13(Suppl 1):FC-049

Rationale: Acute respiratory distress syndrome (ARDS) is still responsible for about 400,000 deaths per year worldwide, requiring new areas of research to improve patients’ care. The lung microbiota is involved in numerous chronic inflammatory pulmonary diseases and recent data suggest that it could play a role in ARDS as well. Nevertheless, the only improvements made despite five decades of intensive research consist in limiting the adverse effects of supportive care. Failure in the development of new therapeutics may be partly explained by the fact that ARDS is a highly heterogeneous syndrome involving different types of aetiologies and extremely clinically and biologically diverse patients. The aim of this study is to compare lung microbiota (both bacterial and fungal kingdoms) composition between specific ARDS subphenotypes, i.e., pulmonary ARDS due to influenza, SARS-CoV-2, or bacterial infection.

Patients and methods/materials and methods: Consecutive ARDS patients according to Berlin’s classification requiring invasive mechanical ventilation were screened in 2 ICUs. Influenza and SARS-CoV-2 infections were confirmed by PCR, and bacterial infection when > 10⁵CFU/mL bacteria were isolated on endotracheal aspirate. Endotracheal aspirate was collected at admission. After DNA extraction, V3-V4 and ITS2 regions were amplified by PCR. Deep sequencing was performed on MiSeq sequencer (Illumina®); data were analysed using DADA2 pipeline.

Results: Twenty-four COVID-19-, 18 consecutive influenza-, and 11 bacteria-related ARDS patients were included. Bacteria-related ARDS patients had lower lung bacteriobiota α-diversity compared with COVID-19-related ARDS patients (p = 0.42 and p = 0.029 for Shannon and Simpson indices, respectively). Influenza-related ARDS patients had higher lung mycobiota α-diversity compared with COVID-19-related ARDS ones (p = 0.01 and 0.05 for Shannon and Simpson indices, respectively). Lung bacteriobiota composition was dissimilar between COVID-19-related and both influenza- and bacteria-related ARDS patients (β-diversity, Permanova p = 0.05 for each). Lung mycobiota composition was dissimilar between influenza-related and both COVID-19- and bacteria-related ARDS patients (β-diversity, Permanova p = 0.01 and p = 0.04, respectively).

Conclusion: Compositions of both lung bacteriobiota and lung mycobiota are different between influenza-, COVID-19- and bacteria-related ARDS. Future studies investigating the role of lung microbiota in ARDS pathophysiology should take aetiology into account.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Michael Ejzenberg (ejzenberg.michael@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-050

Rationale: During the first months of the COVID pandemic, we observed a high rate of ventilator-associated pneumonia (VAP) and nosocomial bacteremia. In accordance with formal expert recommendations on VAP, we decided to implement selective digestive decontamination (SDD) for all intubated patients with SARS-CoV-2 infection. We evaluated the effectiveness of this measure on day-60 mortality, rates of acquired bloodstream infection (BSI) and VAP.

Patients and methods/materials and methods: We performed a retrospective, monocentric cohort study based on the analysis of data collected in the OUTCOMEREA database. In addition to standard hygiene and nosocomial infection prevention measures, the SDD regimen consisted in enteral administration 4 times daily of colimycin 135mg, gentamicin 100mg, and amphotericin B 500mg, without intravenous antibiotic prophylaxis. ICU mortality, rates of bacteremia and VAP were assessed over 2 periods, before and during SDD administration. Multivariate analyses used Cox models for survival and the Fine and Gray model for the risk of nosocomial infection.

Results: 104 patients were included from may 2020 to may 2021: 59 patients in the cohort without SDD, and 45 patients in the SDD cohort (Table). Nearly ¾ of the patients in the 2 groups were treated with IV antibiotic therapy within 48h of intubation. After adjustment on case-mix and confounders, Day-60 mortality was not different between periods (adj HR:0.61 95% CI 0.29–1.27), p = 0.19). BSI rates were similar between groups (Table), without significant difference after controlling for confounders (sHR1.29, [0.65—2.52], p = 0.47). VAP rates were also similar after adjustment (sHR 1.74, [0.94—3.2]) and the VAP incidence densities were also similar (11.69/1000 vs 11.68/1000 ventilator-days, without and with SDD respectively). Except for 2 bezoars, SDD was not associated with adverse events.

Conclusion: In a before-after study, the use of systematic SDD in mechanically ventilated patients with SARS-CoV-2 pneumonia was not associated with a significant reduction in the rate of VAP and ICU-BSI and had no beneficial effect on day-60 survival.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Antoine Piantoni (antoine.piantoni@hotmail.fr)

Annals of Intensive Care 2013, 13(Suppl 1):FC-051

Rationale: Bloodstream infection (BSI)is a common ICU-acquired infection. A growing proportion of them is caused by multidrug-resistant (MDR) bacteria. Recently, the SARS-CoV-2 pandemic led to an increased rate of critical illness and seems to be associated with a high rate of secondary infections, such as pneumonia and BSI. However, there is a lack of data on the influence of SARS-CoV-2 infections on the incidence of BSI-related MDR. In this study, our aim was to evaluate the influence of SARS-CoV-2 infection on MDR BSI occurrence. The secondary objectives were to study the impact of COVID on outcomes, including mortality and length of ICU stay, in MDR BSI patients.

Patients and methods/materials and methods: This retrospective study was conducted in a single mixed intensive care unit (ICU) from January 1 to December 31, 2020. All adult patients admitted to our ICU for more than 48h were included. Collected data included COVID-19 status, MDR BSI, demographic characteristics, main admission diagnosis, comorbidities, prior antibiotic exposure, invasive life support procedures, and immunomodulators. Univariate and multivariate analyses were conducted with a Cox regression analysis. Exposure to COVID-19 was considered as a time-dependent covariate in the Cox model.

Results: Among the 1327 patients included in the analysis, 502 had COVID-19. MDR BSI occurred in 36 (7%) COVID-19 patients and in 14 (1.6%)non-COVID-19 patients. SARS-CoV-2 infection was associated with a significantly higher risk of MDR BSI starting on the 15^th day of ICU hospitalization in univariate (HR 3.74 (1.43–9.79)) and multivariate analysis (HR 4.09 (1.50–11.1)), adjusting for age, SAPS II, immunosuppression, prior hospitalization, prior antibiotic treatment, MDR colonization, and dialysis.No significant impact of COVID-19 was found on mortality (HR 2.11 (1.13–3.93) vs 1.36 (0.50–3.70), interaction p-value 0.55) or length of ICU stay (HR 1.06 (0.73–1.52) vs 0.76 (0.43–1.32), interaction p-value 0.31) in COVID-19 and non-COVID-19 patients, respectively.

Conclusion: In this study, COVID-19 was independently associated with a higher risk of MDR BSI starting on the 15th day of ICU hospitalization.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Louis Kreitmann (louis.kreitmann@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-052

Rationale: Patients presenting the most severe form of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) disease 2019 (COVID-19) pneumonia have a prolonged intensive care unit (ICU) stay and are exposed to broad-spectrum antibiotics, but the impact of COVID-19 on antimicrobial resistance is unknown. The main objective of this study was toinvestigate the association of COVID-19 with the cumulative incidence of ICU-acquired colonization and infection related to multidrug-resistant (MDR) bacteria (ICU-MDR-col and ICU-MDR-inf, respectively).

Patients and methods/materials and methods: Observational prospective before-after study in 7 ICUs in France. All consecutive patients with an ICU stay longer than 48h and a confirmed SARS-CoV-2 infection were included prospectively and followed for 28days. Patients underwent systematic screening for colonization with MDR bacteria upon admission and every week subsequently. COVID-19 patients were compared to a recent prospective cohort of control patients from the same ICUs [1].

Results: During the first and second waves of the pandemic, 367 COVID-19 patients were included and compared to 680 controls. After adjustment for prespecified baseline confounders, the cumulative incidence of a composite outcome including ICU-MDR-col and/or ICU-MDR-inf was not significantly different between groups (adjusted sub-hazard ratio [sHR] 1.39, 95% confidence interval [CI] 0.91–2.09). When considering both outcomes separately, COVID-19 patients had a higher incidence of ICU-MDR-inf than controls (adjusted sHR 2.50, 95%CI 1.90–3.28), but the incidence of ICU-MDR-col was not significantly different between groups (adjusted sHR 1.27, 95%CI 0.85–1.88).

Conclusion: COVID-19 patients had an increased incidence of ICU-MDR-inf compared to controls, but the difference was not significant when considering a composite outcome including ICU-MDR-col and/or ICU-MDR-inf. If confirmed, our findingscould have important practical implications regarding the choice of empiric antibacterial regimens in COVID-19 patients.

Reference 1

1. Kreitmann L, Vasseur M, Jermoumi S, Perche J, Richard J-C, Wallet F, et al. Relationship between immunosuppression and intensive care unit-acquired colonization and infection related to multidrug-resistant bacteria: a prospective multicenter cohort stu.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Lucie Le Fevre (lucie.lefevre@aphp.fr)

Annals of Intensive Care 2013, 13(Suppl 1):FC-053

Rationale: As compared to other etiologies of acute respiratory distress syndrome (ARDS), COVID-19 is associated with a higher incidence of ventilator-associated pneumonia (VAP) and a higher incidence of VAP recurrences. We aimed to describe successive VAP episodes and factors associated with VAP recurrence in patients with COVID-19 ARDS requiring veno-venous extracorporeal membrane oxygenation (VV-ECMO) support.

Patients and methods/materials and methods: Consecutive adult patients with COVID-19 ARDS requiring VV-ECMO support admitted to our ICU between March 2020 and January 2022 were included. Microbiological data and antimicrobial regimen were collected for each VAP episode. All VAP episodes, when clinically suspected, were bacteriologically confirmed (bronchoalveolar lavage fluid yielding one or more pathogens growing ≥ 10⁴ UFC/ml). Each VAP recurrence was categorized as persistent infection (same pathogen growing at a significant concentration while receiving antimicrobial treatment for a previous VAP), relapse (same pathogen growing at a significant concentration and occurring at least 48h after treatment cessation of a previous VAP), or superinfection (new pathogen growing at a significant concentration).

Results: Of the 252 patients included during the study period, 226 (89.7%) developed a first VAP (VAP1), of which 182 (80.5%) had at least one VAP recurrence. Enterobacteriaceae were the most frequent pathogens identified at VAP1 (50.4%), followed by non-fermenting gram-negative bacilli (NF-GNB) (36.3%). At first recurrence, Enterobacteriaceae and NF-GNB were identified in similar numbers of patients (60.4% and 64.8%, respectively). Sixty-three percent of first recurrences were relapses (30.2%) or persistent infections (33%), demonstrating the difficulty of durably eradicating pathogens from the lungs. Of the 226 patients who experienced one VAP episode, 168 (74%) had an antimicrobial course of less than 7days for the first VAP (median duration 6days, IQR 5–7) and 58 (26%) had a prolonged course of 8days or more (median duration 9days, IQR 8–11). In a survival model comparing short and long treatment groups of VAP1 considering extubation and death as competing events, we found no significant difference in terms of VAP recurrence rate.

Conclusion: In COVID-19 ARDS requiring VV-ECMO support, rates of VAP relapses and persistent infections were extremely high, with Enterobacteriaceae and NF-GNB being predominant causative microorganisms. Antimicrobial treatment of a first VAP episode for 8days or more did not significantly reduce VAP recurrence rate.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Yahyaoui Hajer (hajer.yahyaoui94@hotmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-054

Rationale: The SARS-CoV-2 pandemic was characterized by increased antibiotic prescribing. This was initially justified by bacterial co-infection and superinfection risk and the antiviral and immunomodulator effects of macrolide antibiotics. Rapidly, this prescription became controversial due to a higher risk of nosocomial infection and multidrug-resistant bacterial selection. Our objective was tostudy empiric antibiotic therapy in COVID-19 critically ill patients, and to investigate its supporting arguments and evolving features.

Patients and methods/materials and methods: Descriptive, retrospective, and monocentric study including all patients admitted to an intensive care unit (ICU) between September 1, 2020, and June 30, 2022. All patient’s data including demographics, clinical, biological, radiological, bacteriological, therapeutic data and outcomes were taken from patient’s medical records.

Results: 450 patients infected with SARS-CoV-2 were included. The sex ratio was 1.14 and the mean age was 63 ± 15years. Mean SAPSII, APACHEII and SOFA score were respectively 30 ± 39, 30 ± 39 and 8.6 ± 5. 83 patients had a severe form of acute respiratory distress syndrome (ARDS) on admission. The mean lymphocytes count, D-dimer, lactates, CRP and procalcitonin levels were 1295 ± 1721 cells/mm³, 3049 ± 10,090ng/µL, 4 ± 14mg/L, 137 ± 96mg/L and 2.8 ± 10ng/µL, respectively. 333 patients received empiric antibiotics in combination with parenteral corticosteroid therapy. Monotherapy was used in 6 cases (2%), dual therapy in 241 cases (73%) and tritherapy in 83 cases (25%). Betalactams were the most prescribed antibiotic (269 patients (81.5%)). Patients receiving antibiotics had a higher CRP ((151 vs 99), p = 0.046), procalcitonin ((3.6 vs 0.3), p < 10^–3) and a lower count of lymphocytes ((1222 vs 1498) and p = 0.053)). Global mortality was 49% (n = 33). Empiric antibiotic prescription was justified by high procalcitonin levels in 58 cases (17.5%), by parenchymal opacities on chest CT scan in 18 cases (5.5%) and systematically in 254 cases (77%). No microbiologic evidence of a community bacterial superinfection was found. Empiric antibiotic therapy on admission was associated with nosocomial infection (p = 0.07), invasive mechanical ventilation (P = 0.001) and mortality (p < 10^–3).

Conclusion: The COVID-19 pandemic is associated with an over-consumption of antibiotics, especially in the ICU. This systematic prescription is not justified and is associated with nosocomial infections and mortality.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Jolan Malherbe (malherbe-j@chu-caen.fr)

Annals of Intensive Care 2013, 13(Suppl 1):FC-055

Rationale: Varicella-zoster virus (VZV) is a ubiquitous herpesvirus known to cause infections in humans. The incidence of Herpes zoster is increasing, probably due to the aging population and the increasing number of immunocompromised patients. Therefore, an increase in life-threatening organ damage related to VZV can be expected. VZV is recognized as the second most common agent of encephalitis in France (the first one in immunocompromised patients) and is a well-known cause of viral pneumonia, but there are no data on VZV-related severe organ involvement. The main objective of this study was to describe the clinical features and outcome of all life-threatening VZV manifestations requiring admission to an intensive care unit (ICU).

Patients and methods/materials and methods: This retrospective cohort study was conducted in 25 French ICUs and included adult patients with life-threatening VZV-related events requiring ICU admission or occurring in ICU between 2010 and 2019.

Results: One-hundred and sixteen patients were included with a median age of 66years and a median SOFA score of 6. Forty-three were primo-infected (37.1%), and 58 (50%) were immunocompromised. Encephalitis was the most prominent organ involvement (64 patients, 55.2%), followed by pneumonia (51, 44%) and hepatitis (11, 9.5%). Fifty-two patients (44.8%) received norepinephrine, and respiratory support was used in 96 patients (83.5%) including invasive mechanical ventilation in 70 of them (72.9%). Thirty-one patients (27%) received renal replacement therapy. In-hospital mortality was 36.2% and was significantly associated with 3 independent risk factors: immunosuppression, VZV reactivation and alcohol abuse in multivariate logistic regression. Hierarchical clustering on principal components revealed 4 phenotypically distinct clusters of patients (Figure): VZV-related pneumonia, mild-encephalitis, severe encephalitis predominantly due to VZV reactivation, and diffuse VZV reactivation occurring in profoundly immunocompromised patients. Phenotypes differed by medical history, clinical manifestations, and prognosis, with in-hospital mortality ranging from 5.6% to 70.4% (p < 0.001).

Conclusion: Overall, severe VZV manifestations still result in high mortality in the ICU. In-hospital mortality appears to be driven by immunosuppression status rather than by the predominance of organ involvement. Deciphering the different phenotypes may help clinicians assess prognosis.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Julien Vaidie (julien.vaidie@chu-limoges.fr)

Annals of Intensive Care 2013, 13(Suppl 1):FC-056

Rationale: With the exception of a few retrospective studies mainly including patients under chemotherapy, information regarding the impact of immunosuppressive therapy on the prognosis of patients admitted to the intensive care unit (ICU) for septic shock is scarce. Accordingly, the PACIFIC study aimed to assess if immunosuppressive therapy was associated with an increased mortality in patients admitted to the ICU for septic shock.

Patients and methods/materials and methods: This was a multicenter retrospective exposure/no exposure study. To better identify the effect of immunosuppressive treatment on outcome, we used an original selection approach to reduce limitations of a retrospective design. Patients in the “exposed” group were selected from the screen failure logs of seven recent randomised controlled trials (RCT) since they had not been enrolled because of their immunosuppressive treatment (e.g.: steroids, calcineurin, mTOR or TNF inhibitors, immunosuppressive monoclonal antibodies…). The “unexposed” patients were those included in the placebo arm of the same RCTs. We used a ratio of three "unexposed" patients to one "exposed" patient to increase power. Eight high enroller centers in septic shock RCTs participated in the study. A multivariate logistic regression model was used to estimate the risk of death.

Results: Among the 433 patients enrolled, 103 were included in the “exposed” group and 330 in the “unexposed” group. Reasons for immunosuppressive therapy included organ transplantation (n = 45 [44%]) or systemic disease (n = 58 [56%]). Baseline characteristics were similar between the two groups, with a mean SAPS 2 of 56 in the "exposed" group and 60 in the "unexposed" group (p = 0.14). ICU mortality rate was 24% in the “exposed” group and 25% in the “unexposed” group (p = 0.9). In univariate and multivariate analysis, immunosuppressive therapy was not associated with a higher ICU mortality (OR: 0.95; [95%CI: 0.56–1.58]: p = 0.86 and 1.21 [95%CI: 0.69–2.08]: p = 0.5, respectively) or 3-month mortality (OR: 1.13; [95%CI: 0.69–1.82]: p = 0.27 and OR: 1.37 [95%CI: 0.78–2.38]: p = 0.27, respectively). ICU length of stay was similar in the “exposed” and “non-exposed” groups (8 ± 4 vs. 9 ± 5days: p = 0.2), as well as the duration of ventilation (4 ± 2 vs. 4 ± 2days: p = 0.7) and the number of vasopressor-free days within 30days (27 ± 2days vs 26 ± 3days: p = 0.11).

Conclusion: In this multicenter retrospective study, long-term immunosuppressive therapy excluding chemotherapy was not associated with a significantly higher ICU and 3-month mortality rate in patients admitted to the ICU for septic shock.

Compliance with ethics regulations: N/A.

Correspondence: Margot Combet (margotcombet@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-057

Rationale: Human Herpesvirus 6 (HHV-6) DNAemia is not rare in intensive care unit (ICU) patients, especially in the most severely ill and in immunocompromised patients. However, evidence for a causal association of HHV-6 with diseases and link with mortality is limited. We sought 1) to measure the incidence of HHV-6 related disease in ICU patients with HHV-6 DNAemia and 2) to explore its association with ICU mortality.

Patients and methods/materials and methods: Retrospective multicenter cohort study in three ICUs from January 2009 to January 2022. All consecutive patients with at least one positive viral load measured by real-time PCR (expressed as log10 HHV-6 DNA copies/106 cells [log/106 cells]) in whole blood samples, performed during the ICU stay, were included. Patients with positive HHV-6 viral load were classified as reactivation (DNAemia without HHV-6 related organ failure) or HHV6 related disease (DNAemia with HHV-6 related organ failure) based on guidelines and expert adjudication committee. Patients with inherited chromosomally integrated HHV-6 were excluded. Logistic regression multivariate analysis was performed on ICU mortality.

Results: One hundred and sixty-eight patients were included (100 (60%) males, 46 [37 − 65] median [interquartile interval] years old, simplified acute physiology score 2 45 [32–68]). Invasive mechanical ventilation and extracorporeal lung circulation were initiated in 119 (71%) and 61 (36%) patients, respectively. Among these 168 patients, 140 (83%) patients were classified as HHV-6 reactivation and 28 (17%) as HHV-6 related disease. The two main HHV-6 diseases were encephalitis (n = 14/28, 50%) and pneumonitis (n = 12/28, 43%). Patients with HHV-6 related disease were more frequently immunocompromised than patients with HHV-6 reactivation (93% vs. 46%, p < 0.001), whereas veno-arterial extracorporeal membrane oxygenation (VA-ECMO) was more frequently observed in patients with HHV-6 reactivation (32% vs. 0%, p < 0.001). Maximum viral load was higher in patients with HHV-6 related disease than their counterpart (4.63 [3.30, 5.55] vs. 1.86 [1.55, 2.38] log/106 cells, p < 0.001). Among the 28 patients with HHV-6 related disease, only 12 (43%) received anti-HHV-6 therapies during the ICU stay. ICU mortality was 32% (n = 53). Using multivariate analysis, HHV-6 related disease (OR 3.1 95% confidence interval (CI) 1.10–8.80, p 0.03), being immunocompromised (OR 35 [5.80–218], p < 0.001) and VA-ECMO (OR 13.35 95%CI 2.30–76.8, p < 0.001) were independently associated with ICU mortality.

Conclusion: Organ disease could be attributable in up to 20% of ICU patients with HHV-6 DNAemia. Such HHV-6 related disease is an independent predictor of ICU mortality.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Morgane Snacken (snackenmorgane@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-058

Rationale: Post-aggressive acquired immunosuppression after septic or sterile insults is characterized among other immunological change: i) by a monocytic functional deactivation highlighted by a decrease of HLA-DR expression and ii) by the overexpression of PD1 and PDL-1 in CD4 + cells and monocytes. These markers have been shown to be associated with the development of nosocomial infections and are thus attractive therapeutic target of post-aggressive acquired immunosuppression. In this setting, the question whether these markers are concomitantly impaired has been poorly studied. In particular, the ability of HLA-DR expression to predict overexpression of PD1/PDL1 is unclear. A clear association would indicate that testing for HLA-DR expression would be sufficient both for prognostic tool (risks of nosocomial infections, mortality) and to decide a treatment even targeting PD-1/PDL-1. The aim of this work was to study the association between HLA-DR expression at the monocyte membrane and PD1-PDL 1 expression at the lymphocyte and monocyte membrane.

Patients and methods/materials and methods: This observational, prospective, monocentric study controlled by healthy volunteers was performed in the medico-surgical intensive care unit. HLA DR, PD1, PDL1 to J2-3 and J5-7, among other immunological markers were analyzed by FACS.

Results: We recruited 21 controls to monitor and calibrate the cytometry technique and define the patient analysis strategy. We also included 63 patients admitted to our ICU between November 2019 and September 2020 and a total of 47 were analyzed. Main patients characteristics are reported (table 1). We did not find any correlation between PD1/PDL1 and HLADR expression (y = 0,856x + 16,812. R² = 0,0621, y = 0,1897x + 19,015. R² = 0,0355.) We compared septic versus non septic profiles. Septic patients had lower HLADR and CD86 (monocyte inactivation profile) but no difference in PD1/PDL1 was found. Finally, we compared PD1 in low HLA DR (< 30%) patients with normal HLA DR (Chi2 test. P = 0,754) with no significant difference.

Conclusion: The study of HLA DR alone is not sufficient to characterize the immunological profile of severely assaulted patients. There is an interest in studying PD1/PDL 1 markers to better select candidates to PD-1/PDL-1 inhibition therapies.

Reference 1

Guignant and al. Critical Care 2011, 15:R99.

Reference 2

Morris and al. British Journal of Anaesthesia 111 (5): 778–87 (2013).

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Chloé Tridon (chloe.tridon12@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-059

Rationale: In endemic areas, intravenous (IV) artesunate is now firmly established as the treatment of choice for severe malaria. However, during severe imported malaria treated in France with IV artesunate, little is known about the management in the intensive care unit (ICU) of the most severe forms. As a first step of PALUSTAR Study, we sought to describe the characteristics, management and outcome of patients with a very severe malaria treated with IV artesunate over the period 2011–2019.

Patients and methods/materials and methods: Multicenter observational retrospective study in adults admitted for very severe malaria treated by IV artesunate. Data were collected from medical charts using standardized case-report forms, in 58 French ICU in 2011–2019. In a patient with a Plasmodium falciparum asexual parasitaemia, very severe malaria was defined by at least one criterion, during the first 72h of ICU stay, among the following criteria: coma with Glasgow Coma Scale score < 11 and/or repeated convulsions, shock, respiratory distress, acidosis, hyperlactatemia > 5mmol/L, and/or hospital death.

Results: Over the 2011–2019 period, 259 patients were included. Most were male (62.9%), and median age was 52 (IQR, 38—62). On the first day (D0) of ICU admission, 97 (37.5%) patients had impaired consciousness, 118 (45.6%) were in respiratory distress, 136 (52.5%) had shock, 100 (38.6%) had acidosis, and 98 (37.8%) had hyperlactatemia. Median SAPS II and SOFA score was 41 (IQR, 30—58) and 10 (IQR, 7—14), respectively. During the ICU stay, mechanical ventilation was required in 84 (32.4%) patients, 137 (53.1%) received vasopressive drugs and 77 (29.8%) required renal replacement therapy. Post Artesunate Delayed Hemolysis (PADH) was reported in 47 (18.1%) patients. D28 mortality was 11.2%. By multivariate analysis, three variables at ICU admission were independently associated with death at D28: SOFA score, arterial lactatemia, Glasgow Coma Scale score.

Conclusion: In a large cohort of patients hospitalized in the ICU with very severe imported malaria and treated by IV artesunate, the D28 mortality was 11.2%. As a second step of PALUSTAR Study, we plan to compare these patients (treated by IV artesunate) with those of an historical cohort (treated with IV quinine).

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Laurent Camous (laurent.camous@chu-guadeloupe.fr)

Annals of Intensive Care 2013, 13(Suppl 1):FC-060

Rationale: Study characterics and outcome of patients with biologically confirmed Leptospirosis admitted in the intensive care unit (ICU) of the Guadeloupe Teaching Hospital.Identify patients’ clusters by multiCorrespondence analysis.

Patients and methods/materials and methods: All patients admitted in the ICU for biologically Leptospirosis between January2014 to December 2022 were included in the study. Multiple Correspondence statistical analysis (MCA)was realized on the overall population. Using univariate and bi-variate analysis, we assessed risk factors at ICU admission and during ICU course associated with death.

Results: On the study period, 130 patients admitted in the ICU had confirmed leptospirosis. Median age and SAPS II were 56years and 48 respectively. During ICU course, acute respiratory failure was present in 34% of the patients and 26% were mechanically ventilated. Patients had shock in 52% (n = 67) and 21% (n = 27) had neurological involvement. Myocardial involvement was present in 41% (n = 53) of the patients. Based on clinical and biological parameters at ICU admission, MCA identified three patient clusters: “mild leptospirosis” (n = 62), “neurological leptospirosis” (n = 27) and “multiple organ failure (MOF)” (n = 41) with different ICU courses and outcomes. Overall mortality rate at day 28 was 13%. Risk factors associated with death in uni/bivariate analysis were shock, acute respiratory failure, myocardial damage, neurological involvement at ICU admission and need for dialysis during ICU course.

Conclusion: Leptospirosis is an endemic tropical area zoonosis, responsible for severe manifestations with frequent organ failures. Identification in endemic areas of patients with potential life-threatening worsening is essential. In our study, one of the determinent at ICU admission for evolution to a severe form was cardiac involvement which helped to identify two different sub-populations with different outcomes. A third cluster, corresponding to neurological forms seems to represent a different, late form of the disease.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Giacomo Rotini (giacomo.rotini17@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-061

Rationale:Acinetobacter baumannii (Ab) is a well-known nosocomial pathogen that has emerged in the last decades as a cause of community-acquired pneumoniae (CAP) in tropical and subtropical regions. Few global epidemiological studies on CAP-Ab have been published to date and, to our knowledge, no data are available on this pathology in France. The objective of this work is to perform an epidemiological analysis of CAP caused by Ab in Reunion Island and compare them with severe CAP due to other pathogens.

Patients and methods/materials and methods: Case–control study listing severe CAP-Ab cases hospitalized in one of the two intensive care unit (ICU) of Réunion’ University Hospital, during the period 2014–2022; and compare them with an historical cohort (PAC_RUN), obtained by retrospective chart (2016–2021) of severe community-acquired pneumonia, in which PAC-Ab were excluded (n = 4). CAP-Ab was confirmed by microbiological respiratory samples and/or blood cultures realized at the admission in ICUs. All patients who presented with with nosocomial infection were excluded (n = 50).

Results: During the study period, eight CAP-Ab cases were identified, giving an incidence of 0.1 case for 100 000 people/year. By comparing with CAP no Ab (n = 761), patients had more alcohol disorders (p = 0.005) and a lower BMI (p = 0.007). Six CAP-Ab occurred during wet season (p = 0.06). Mortality was higher (62,5% vs 24,3%, p = 0.02) and time to death was shorter (median 2days vs 7, p = 0.009) in CAP-Ab group. Bacteremic pneumonia was strongly associated with CAP-Ab (p = 0.004). Significant differences were found for need for renal replacement therapy (p < 0.001), catecholamines (p = 0.01) and mechanical ventilation (p = 0.03). Ab strains isolated had low resistance pattern, all strains were found sensitive to ceftazidime, cefepime, piperacillin-tazobactam, ciprofloxacine, gentamicin, and imipenem. All patients with CAP-Ab received first line inappropriate antibiotic therapy.

Conclusion: Severe CAP-Ab have a low incidence but a fulminant course with a high mortality. Some demographics and clinical features, such as a fulminant pneumonia during the rainy season among an alcoholic patient, should suggest this pathogen prompting an antibiotic therapy targeting A. baumannii.

Reference 1

Dexter C, Murray GL, Paulsen IT, Peleg AY. Community-acquired Acinetobacter baumannii?: clinical characteristics, epidemiology and pathogenesis. Expert Rev Anti Infect Ther. 4 mai 2015;13(5):567?73.

Reference 2

Davis JS, McMillan M, Swaminathan A, Kelly JA, Piera KE, Baird RW, et al. A 16-Year Prospective Study of Community-Onset Bacteremic Acinetobacter Pneumonia. Chest. oct 2014;146(4):1038?45.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Keyvan Razazi (keyvan.razazi@aphp.fr)

Annals of Intensive Care 2013, 13(Suppl 1):FC-062

Rationale: Emphysematous pyelonephritis (EP) is a rare and severe necrotizing infection of the urinary tract, with gas accumulation, probably produced through a glucose fermentation process within the collecting system, the renal parenchyma, and/or the perirenal tissue. Data on emphysematous pyelonephritis in ICU are scarce. We aimed at reporting the clinical features and outcomes of adult patients admitted in the intensive care unit (ICU) for an infectious EP,as well as the predictive factors for nephrectomy and mortality.

Patients and methods/materials and methods: A 21-year national multicenter retrospective cohort study in 48 ICUs in France from 2001 to 2021, including adult patients admitted for an infectious EP. Primary outcome variables included ICU mortality and nephrectomy during the follow-up period (time between ICU admission and amputation, death or end of follow-up).

Results: A total of 115 patients were included in the study. We herein present the preliminary results of 107 patients. Patients had a median age of 62years [54; 72], with a female predominance (64%), and 74% of them were either overweight or obese. Diabetes mellitus was present in 62% of cases. The median SAPS2 score at ICU admission was 50 [37; 66] and the median SOFA was 8 [5; 12]. Most patients fulfilled septic shock criteria (68%). A urinary tract obstruction was found in 50 patients (48%). Regarding surgical procedures, urine drainage or percutaneous drainage was performed in 60 patients (56%) and nephrectomy was performed in 23 patients (21%).Enterobacteriaceaerepresented 93% of microbiological documentation,E. colibeing the most frequently identified species (57%). Mortality at 90days was 21%. Factors associated with day 90 mortality were SOFA, and SAPS2 scores, and presence of a septic shock.

Conclusion: Emphysematous pyelonephritiscarries a high mortality and morbidity. Need for nephrectomy was not infrequent.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Sarah Benghanem (sarah.benghanem@aphp.fr)

Annals of Intensive Care 2013, 13(Suppl 1):FC-063

Rationale: Mismatch negativity (MMN) is elicited during an oddball paradigm, using frequent sounds (“standards”) mixed with rare sounds (“deviants”). MMN response is mainly used for awakening prediction of disorder of consciousness(1). The use of emotional sounds, as looming sound intensity, might improve the MMN prognostic performance although they are usually not used for MMN. We hypothesize that the acoustics characteristics of deviant auditory stimulus could modify the neurophysiological characteristics of auditory evoked potentials.

Patients and methods/materials and methods: We assess 18 healthy volunteers with 64-channels EEG and compared MMN responses with three modalities of deviant (looming, receding and flat) during an oddball paradigm (80% standard, 6.7% of each deviant). Deviant sounds duration was 600ms with a gradual linear increased (looming) or decreased (receding) intensity of 15db. Deviant flat duration was 600ms without intensity variation, standard sounds were shorter (300ms, no intensity variation).

Results: Compared to receding sounds, looming deviants elicited a later (206 vs 175ms, p = 0.02) and a wider (area under curve AUC 0.116 vs 0.053, p = 0.002) MMN. Looming, receding and flat deviants also elicited a later potential at 600ms, with a higher peak amplitude (-1.96µV vs -1.43µV, p = 0.01) and a wider AUC (0.26 vs 0.18, p = 0.02) for looming than for flat sounds. Our interpretation is that the first component purely reflects the physical characteristics of the stimuli, while the second component reflects a higher-level, cognitive emphasis of looming sounds.

Conclusion: Looming and receding “3D” sounds elicit different MMN responses compared to flat. Looming sounds may act as a warning clue (2) and may constitute a trigger to elicit higher level of consciousness processing.

Reference 1

André-Obadia N, Zyss J, Gavaret M, Lefaucheur JP, Azabou E, Boulogne S, et al. Recommendations for the use of electroencephalography and evoked potentials in comatose patients. Neurophysiol Clin. juin 2018;48(3):143?69.

Reference 2

Bach DR, Furl N, Barnes G, Dolan RJ. Sustained Magnetic Responses in Temporal Cortex Reflect Instantaneous Significance of Approaching and Receding Sounds. Snyder J, éditeur. PLOS ONE. 30 juill 2015;10(7):e0134060.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Juliette Pocquet (juliette@pocquet.fr)

Annals of Intensive Care 2013, 13(Suppl 1):FC-064

Rationale: Posterior reversible encephalopathy syndrome (PRES) is a clinico-radiological entity gathering unspecific symptoms (headache, confusion, visual disturbances, seizures or coma), classically associated with a sudden rise in blood pressure, related to white matter edema. Brain MRI is the gold standard exam, with bilateral and symmetrical T2 and FLAIR hypersignals and T1 hypo and/or isosignals, predominant in the parieto- occipital cortico-subcortical regions. These symptoms and radiological abnormalities are most often reversible. PRES is a rare syndrome, but with technical advances in neuroradiology, more and more cases are being diagnosed. More than 40% of PRES require hospitalization in intensive care unit (ICU). Very few studies focus on ICU patients in the literature (clinical cases or cases series). The largest retrospective cohort was conducted by S. Legriel and included 70 patients (1). Our goal was to characterize the population and the prognostic factors.

Patients and methods/materials and methods: We conducted an observational, retrospective and multicentric study. From January 1, 2009 to December 31, 2020, we selected all the medical records of adult patients hospitalized in 6 ICUs in a French region with PRES via the CIM 10 coding and the brain MRI reports. 50 patients were included in our study.

Results: Most of our patients were women (66%), with a median age of 52years-old. PRES was mainly attributed to drugs (44%), hypertension (28%), and pre-eclampsia (14%); 24% did not have arterial hypertension; a sepsis was associated in 46% of our cases. More than half of the patients required mechanical ventilation, 38% had acute kidney injury; 80% had neurological symptoms on admission, and 70% consciousness disorders. Thirty-two percent had sequelae at day 90 (seizures, persistent deficit, cognitive alteration), 28% had an unfavorable outcome (Glasgow Outcome Scale < 5) at day 180 and 10% died (table 1). The risk factors associated with an unfavorable outcome were extra neurologic organ failure at the time of admission (OR 6,66 [1,7–27]) and the occurrence of acute renal failure requiring dialysis (OR 24 [2,5–229]). These two factors were also risk factors for the persistence of sequelae (OR 4.58 [1.25–16.8] and 6.4 [1.07–38], respectively). Pre-eclampsia seemed to be a protective factor for good outcome, but the difference was not significant.

Conclusion: Our epidemiological data are similar to those described in the literature. The term "reversible" seems inappropriate since we found a significant rate of neurologic sequelae and unfavorable outcomes. Our results plead for a specific follow-up in PRES patients.

Reference 1

Legriel S, Schraub O, Azoulay E, Hantson P, Magalhaes E, Coquet I, Bretonniere C, Gilhodes O, Anguel N, Megarbane B, Benayoun L, Schnell D, Plantefeve G, Charpentier J, Argaud L, Mourvillier B, Galbois A, Chalumeau-Lemoine L, Rivoal M, Durand F, Geffroy A.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Adam Celier (adam.celier@aphp.fr)

Annals of Intensive Care 2013, 13(Suppl 1):FC-065

Rationale: While electroencephalogram (EEG) plays an important role in the management of many resuscitation conditions, few studies have evaluated its use and impact on patient management in intensive care units (ICU). The prevalence of pathologies requiring EEG in a general ICU and most frequently identified EEG abnormalities is poorly documented.

Patients and methods/materials and methods: This project is a retrospective study of the RéVE (RéaÉValuationEeg) cohort. This cohort includes all patients admitted to an ICU and who had a bedside EEG recording between 1 January 2017 and 1 April 2020. For each patient, demographic, clinical, paraclinical and evolution data were collected. For each report, many elements concerning the EEG background activity and EEG abnormalities were collected, as well as the main diagnosis.

Results: Over the period, 305 patients had an EEG recording in this ICU. Based on the number of hospitalizations in the department in 2018, 11.5% of patients admitted to the ICU have an EEG during their stay. The reasons for admission to the ICU resulting in the performance of an EEG during the patient's stay were mainly represented by three reasons: disease of consciousness (DOC) (25%), status epilepticus (20%) and cardiac arrest (18%). Mean age was 57years old. History of epilepsy was known for 20% of the patients. EEGs were mostly abnormal, with disparities between the 3 main groups. A majority (44%) of EEG findings were of the slowing-sedation type, followed by metabolic encephalopathy (16%), post-anoxic encephalopathy (14%). Epilepsy was reported in only 7% of EEGs. EEG trace was continuous for 87% of the patients, but only for 43% of the cardiac arrest patients. Slow triphasic waves were seen in 15% of the DOC patients, epileptic abnormalities in 18% of the epilepsy group. Only 14 seizures (5%) were recorded. Two-thirds of the EEGs were considered to have had an impact on patient management.

Discussion: The analysis of all EEG reports allowed us to have an overview of the EEG findings in the ICU. This work sheds light on the impact of EEG in ICU and provides a better understanding of intensivists expectations with regard to this examination.

Conclusion: EEG is now an important part of the management of many intensive care patients. In this single-centre observational study, the three main indications of an EEG in the ICU were disorder of consciousness, status epilepticus and cardiac arrest.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Maeva Le Goic (maevalgo@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-066

Rationale: Status epilepticus (SE) is an acute brain condition of extreme severity which is associated with significant morbimortality. Its severity depends on many factors including age, aetiology, and duration of SE. Some patients may reach a prolonged super-refractory SE, requiring the maintenance of a prolonged induced coma and continuous EEG monitoring to early diagnose and treat non-convulsive SE. The aim of this study was to develop a descriptive and comprehensive approach of a specific electric pattern of seizures that has a periodic/cyclic mode of recurrence.

Patients and methods/materials and methods: In a retrospective single-centre study, we screened 138 patients admitted with SE in a neurological ICU, between 2016 and 2022. A subgroup of SE patients with cyclic electrographic seizures i.e. displaying recurrent seizures at regular intervals was compared to SE patients without cyclic seizures (controls) on several pre-set clinical, biological and electrophysiological variables, along with their on-going medication (sedatives, antiepileptic drugs and concomitant treatments that may lower the seizures threshold). Their clinical outcomes (rate of mortality, functional status, and sequel of epilepsy at ICU discharge and after six months) were also compared. A computerized seizure detection method was used in complement to the standard visual EEG analysis and allowed characterization of seizures features as well as their dynamics of occurrence during the whole continuous EEG monitoring period (42days on average).

Results: Our results show that patients with cyclic seizures had a more important seizures burden, were younger, stayed longer in ICU and all reached a prolonged super refractory status epilepticus condition in comparison to the control population. Moreover, they displayed a poorer clinical outcome at discharge, with a higher rate of sequelar refractory epilepsy. In majority, cyclic seizures had no clinical correlates, a focal ictal onset and a shorter duration and, strikingly often organized into clusters. When comparing the first and terminal intra-cluster duration of seizure, this latter was longer which may imply a better capacity to activate inhibitory mechanisms and terminate a cluster. In addition to these endogenous mechanisms, our assumption that multiple associated factors contribute to their cessation and/or generation was confirmed as metabolic disturbances such as renal failure, hypomagnesemia, higher vigilance, clinical severity and a more aggressive treatment were associated with a higher risk of cyclic seizures.

Conclusion: This study may help provide insight into the underlying pathophysiology of cyclic seizures and may help to improve the management of super-refractory SE patients.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Ines Sdiri (sdiri.ynes@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-067

Rationale: Phenytoin and valproate are commonly used antiepileptic drugs. It's generally believed that phenytoin is more effective for partial onset seizures and valproate is more effective in generalized onset seizures. Although there'is no evidence between this two antiepileptic molecules in the management of status epilepticus. We aim to compare phenytoin and valproate when used as antiepileptic drug in the management of status epilepticus in the prognostic.

Patients and methods/materials and methods: We conducted a retrospective study in an intensive care unit including critically ill patients with status epilepticus between January 2016 and December 2022. We divided into 2 groups: those who received valproate acid (G1) and those who received phenytoin (G2). Groups were matched on age, gender, SAPSII and APACHE scores, severity of distress on admission and etiologies for status epilepticus.

Results: Eighty patients with status epilepticus were enrolled in our institution. The median age was 53 [16–90] years and sex-ratio was 4,71. 37,5% had epilepsy, 27,2% had diabetes and 25% had HTA. Only 4,3% had renal failure. The mean APACHEII and SAPS II scores were respectively 18 ± 2 and 37 ± 4. Only 8,5% of patients had an electroencephalogram which demonstrated a comiality in third cases. Generalized tonic–clonic movement was the most frequent type of seizure (87,2%). The differents etiologies of status epilepticus were: cerebrovascular accidents (hemorrhagic stroke in 21,7% and ischemic stroke in 10,9%); switching off the epileptic drug (19,6%); central nervous system infection (meningitis 10,9%); cerebral thrombophlebitis (6,5%); drug toxicity (6,8%) and metabolic abnormalities (2,2%). The management of secondary neurological injuries was achieved in 93,6% of cases. Invasive mechanical ventilation with neurosedation were needed for 76,6% of patients and 38,3% of patients had shock. The complications of status epilepticus were observed in 13,2% cases: hemiplegia (6,6%), frontal lobe syndrome (1,7%), vegetative state (6,6%) and facial palsy (1,7%). ICU mortality was 23,5%. ICU length of stay was 13 ± 22days [1–132]. There were 37 patients in G1 versus 43 in G2. There was no difference in the etiologies (p = 0,329) or the type of seizure (p = 0,11) between the two groups. A repeat crisis was observed in 10,6% of cases without difference between the two groups (p = 0,6). We noticed a higher incidence of the need of mechanical ventilation, nosocomial infection, need of vaso-active drugs and mortality in G1 without signification (p:0,373, p = 0,56, p = 0,34 and p = 0,34 respectively).

Conclusion: We did not find any evidence that a significant difference exists between phenytoin and valproate for the outcomes. However, a propective study is needed to identify bias and to confirm these results.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Rabeb Hammami (hammamirabeb2@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-068

Rationale: Myasthenia gravis is a disease of the neuromuscular junction that progresses in bouts called myasthenic attacks. The latter leads to acute respiratory failure which may require respiratory assistance and a prolonged stay in the intensive care unit.The aim of this study was to describe the epidemiological, therapeutic, and evolutionary characteristics of myasthenic attacks managed in an intensive care unit and to study the factors associated with mortality.

Patients and methods/materials and methods: This was a retrospective, descriptive study conducted in our intensive care unit over a 10-year period between 2012 and 2021. Patients admitted to the intensive care unit for the myasthenic crisis, aged over 18years, and who required mechanical ventilation were included. Patients followed for congenital myasthenia were not included. The epidemiological (age, gender, and history), therapeutic (therapeutic adjustment; plasmapheresis and immunoglobulins), and evolutionary (infectious and thromboembolic complications) characteristics of the patients retained for the study were described. Factors associated with mortality were also studied.

Results: Fifty-seven patients were selected for the study. The average age was 41.21 ± 13.57years. The sex ratio was 0.23: 11 men (19.29%) and 46 women (80.70%). The average age of discovery of myasthenia was 38.44 ± 13.50years, and that of the first myasthenia crisis at 40years ± 13.78. Diabetes dominated the medical history (10.5%). The majority of myasthenic attacks were triggered by an infectious episode (47.36%) or a therapeutic deviation (31.57%). Regarding the specific treatment, 17 attacks were treated with intravenous immunoglobulin, and 17 attacks were treated with plasmapheresis. 23 attacks did not require specific treatment because of the effectiveness of the therapeutic adjustment or because of Clinical instability contraindicating immunoglobulins and plasmapheresis. The mean duration of mechanical ventilation was 15 ± 15.94days, and the mean length of stay was 20.42days ± 22.49. The mortality rate was 17.54%. The most frequent direct cause of death was a septic shock in 80% of cases, followed by cardiogenic shock. The factors associated with mortality—shown in table 1—in univariate analysis were a high APACHE III score on admission and the presence of a thymoma. The use of plasmapheresis was significantly associated with a higher survival rate.

Conclusion: The advent of resuscitation has markedly changed the prognosis of myasthenia gravis, with respiratory support and the development of specific immunomodulatory therapy having considerably reduced the mortality rate.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Badreddine Errouane (badreddine468@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-069

Rationale: Guillain-Barré syndrome (GBS) is the most common cause of acute extensive paralysis. 30% of patients will require mechanical ventilation and 10% of them will have very debilitating motor sequelae and motor disabilities that can affect their quality of life. Quality of life is a multidimensional concept defined by an individual's perception of their position in life, in the context of the culture and value system in which they live, in relation to their goals, expectations, standards and concerns.

Patients and methods/materials and methods: Our study is prospective observational, carried out in a medical intensive care unit. Included were patients over 16years of age with acute polyradiculoneuritis, GBS type. On their discharge, the patients were either reviewed as part of the post-resuscitation consultation at 03months, 06months and one year. Quality of life was measured by a standardized test (WHO in 1995), namely the “Medical Outcome Study Short Form-36 Health Survey” (MOS SF-36). Our results will be compared to the reference values in terms of quality of life of the national study "TAHINA": Transition and Health Impact In North Africa.

Results: 32 patients were collected with an average age of 40.75years (16–72years) and a sex ratio of 1.67. The average Hughes handicap score was 4.03. The extension phase lasted an average of 8.06days and the plateau phase 10.90days. 14 patients required mechanical ventilation, the average duration of which was 16.69days. Plasma exchanges were performed in all patients. The average time to resume walking with assistance was 44.19days and walking without assistance was 89.62days. For quality of life, the results were compared to the national reference, the “TAHINA” study (11). At 03months and 06months, the evaluation scales are all lowered significantly compared to the national reference, both in physical and psychological dimensions. At 01year, the calculated means are not all significantly different from the national reference means.

Conclusion: GBS is a serious pathology that sometimes leads to significant and lasting motor disability (up to 01year). This motor handicap with its psychic repercussions will alter the quality of life of these patients. The study shows that the Values of the quality of life (MOS-SF36), tend to improve gradually and the values become comparable to the general population only after 01year of evolution. Overall, our results are comparable to data from the international literature.

Reference 1

Forsberg, A., et al., Disability and health-related quality of life in Guillain-Barré syndrome during the first two years after onset: a prospective study. Clin Rehabil, 2005. 19(8): p. 900–9.

Reference 2

Institut national de santé publique (INSP)—Programme INCOMED. (2007). « Etude TAHINA—La transition épidémiologique et son impact sur la santé dans les pays nord africains». Alger.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Amir Bedhiafi (bedhiafi.emir@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-070

Rationale: Non-invasive respiratory support (NIRS) has always been controversial in the management of acute hypoxemic respiratory failure. During the COVID-19 outbreak, the use of NIRS evolved progressively in line with the disease spread and the pathophysiological processes understanding. This study aims to describe trends in the use of NIRS over the consecutive COVID-19 waves.

Patients and methods/materials and methods: This is a single-center retrospective observational study conducted in the medical ICU of a Tunisian University Hospital over two years from March 2020 to March 2022 covering four consecutive waves of the disease spread. COVID-19 waves were defined as first, from March 2020 to April 2020, second, from August 2020 to May 2021, third, from June 2021 to September 2021 and fourth, from December 2021 to March 2022. All consecutive patients admitted for COVID-19-related acute respiratory failure (ARF) and requiring NIRS were included. Baseline demographics and clinical data were collected. Univariate descriptive analysis was performed.

Results: During the study period, 407 patients were included out of 512 ICU-admission for COVID-19-related ARF. During the first-wave, eleven critical patients were admitted, among them only two received Non-invasive positive pressure ventilation (NIPPV). In the second-wave, 149 patients were included, median age 66 [60–74], PaO2/FiO2112 [79–166] mmHg, and SAPS II 31[26–37]. NIPPVuse was predominant (66%), either through NIV alone (44%) or alternating with High-Flow Nasal Cannula (HFNC) (22%). HFNC was used alone in 34% of cases and awake proning was performed in more than half of patients (62%). 61% required intubation within 2 [1–4] days with an overall mortality of 55.7%. Over the third-wave, 203 patients were included, median age 59 [48–70], PaO2/FiO2107 [77–164] mmHg, and SAPS II 27[22–32]. Respiratory management strategies have evolved towards the predominantly combined use of HFNC (74%) and awake proning (86%). NIPPV was used in 20% of cases as rescue therapy through an attempt of alveolar recruitment before escalation to invasive mechanical ventilation. The intubation rate was lower at 41.5% (p = 0.001) with a shorter delay of 1 [1–2] days (p = 0.07). The mortality rate was also lowered to 32.3% (p = 0.001). About the fourth-wave, 57 patients were included. Respiratory management was quite similar to the third wave. HFNC was the first-line treatment in 63% of patients. Awake proning was performed in 63%. NIPPV attempt was required in 20% of cases.

Conclusion: NIRS use in the management of COVID-19-related ARF has evolved gradually as the disease has progressed. Combined HFNC with awake proning as first-line treatment might be associated with better outcomes.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Inass Arhoun El Haddad (iarhounelhaddad@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-071

Rationale: If early and prolonged prone positioning reduces mortality in patients on invasive ventilation with acute respiratory distress syndrome, its role in conscious patients remains a subject of debate and research. The objective of our study was to evaluate the effectiveness of awake prone positioning in preventing intubation or death in spontaneously breathing patients with COVID-19 associated acute respiratory failure. The primary endpoint was death within 28days of hospitalization, while the secondary endpoints were: intubation rate, length of intensive care stay, time between hospitalization and death, between hospitalization and intubation and the factors influencing the survival.

Patients and methods/materials and methods: This is a retrospective, monocentric, descriptive and analytical study carried out over a period of 22months from March 2020 to December 2021 and involving 1069 patients hospitalized in the intensive care unit of our hospital for the management of COVID-19 associated acute respiratory failure, of which, 681 patients have been put on prone position ( PP (prone position) group), and 388 have received standard care (SP (supine position) group). We defined the treatment failure as the proportion of patients who died within 28days as a major endpoint. To assess the effectiveness of our procedure, we have conducted a univariate analysis by the kaplan meier methode for the survival rate, and a multivariate analysis by cox regression.

Results: Mortality in our total sample was 26.28% (15.12% in the PP group and 46.13% in the SP group) The intubation rate on day 28 was 33.20% for the general population, (25.58% in the PP group and 46.16% in the SP group). The major factors influencing survival werethe age, the degry of pulmonary involvement on CT scan and the lactate rate.

Conclusion: Awake prone position is significantly associated with a reduction in 28-day mortality. Its use in acute respiratory distress syndrome with and without the context of Covid-19 deserves to be studied in the context of randomized controlled trials.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Ines Sedghiani (sedghiani.ines@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-072

Rationale: Covid-19 can be complicated by pneumonia of varying severity up to acute respiratory distress syndrome (C-ARDS). The study aim was to describe the clinical, paraclinical and evolutionary data of C-ARDS according to the severity.

Patients and methods/materials and methods: We conducted a retrospective study between March 2020 and February 2021 including all patients hospitalized for C-ARDS (according to the Berlin criteria with a positive virological test). We included patients hospitalized via the hospital emergency department for C-ARDS. Patients who died within 48h of admission and those admitted from another hospital were not included in the study. Patients without C-ARDS and those with ARDS of other etiologies were excluded. We compared the initial clinical and paraclinical data and patients outcomes according to the severity of the ARDS.

Results: We enrolled 215 patients with a mean age of 66 ± 12years and a sex ratio of 1.38. The main comorbidities were hypertension (55.8%) and diabetes (45.1%). The median consultation time was seven days and was shorter in mild C-ARDS (5days). Oxygen saturation was lower in severe and moderate C-ARDS (80 and 84% respectively) than in mild C-ARDS (88%). Lung injury extension was < 10% in four cases (1.9%), 10 to 25% in 29 cases (13.5%), 25 to 50% in 67 cases (31.2%), 50 to 75% in 83 cases (38.6%) and > 75% in 32 cases (14.9%). C-ARDS was mild in 121 cases (56.3%), moderate in 84 cases (39.1%) and severe in 10 cases (4.7%). Demographic characteristics, comorbidities, initial biological and CT data were comparable between the different severity stages of C-ARDS. Non-invasive ventilation was the most used oxygenation modality; indicated in 204 cases (94.9%) alternating with High-Flow Nasal Cannula (HFNC) oxygen therapy in 30 cases. It was more indicated in mild (98.3%) and moderate (91.7%) ARDS than in severe ARDS (80%). Invasive ventilation was indicated in 28.8% of cases (n = 62) and 80% of severe ARDS; after a median delay of five days [2–10]. Ventilation was performed in the prone position in 81 patients (37.7%), 76 of whom received spontaneous ventilation. ICU length of stay was comparable between the three stages of C-ARDS. The mortality rate was 34.9% (n = 75); higher in severe C-ARDS.

Conclusion: Patients with C-ARDS have a mean age of 66years. Oxygen saturation and lung injury extension define the C-ARDS severity. The most frequent modes needed in C-ARDS are non-invasive ventilation. The use of invasive ventilation and mortality are high in severe forms.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Radhouane Toumi (Radhouane.toumi@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-073

Rationale: Most severe cases of COVID-19-related ARDS (CARDS) often require invasive mechanical ventilation. CARDS has been described as presenting a dissociation between hypoxemia’s severity and relatively good respiratory mechanics while a minority of severe patients develop decreased pulmonary compliance. The aim of this study was to describe airway pressures and ventilatory efficiency trends in CARDS patients and their association to mortality.

Patients and methods/materials and methods: This is a retrospective observational monocentric cohort study carried out in the medical ICU of a University hospital in a low-middle income country, between March 3rd 2020 and December 31st 2021. All consecutive patients with confirmed SARS-CoV-2 infection admitted to the ICU and requiring invasive mechanical ventilation (IMV) were included. Demographic, clinical, physiological and ventilator data were collected. Univariate regression models were used to identify risk factors of mortality.

Results: 465 patients were admitted to the ICU for a confirmed COVID-19 infection, 247(53.1%) required IMV. Median [IQR] age was 67 [59–74] years. At admission, 15 (6.1%) patients were already on IMV; SAPS II was 31 [27–37] and P/F ratio was 96 [76.3–139.8] mmHg. At days 1 and 5 of IMV, respectively, plateau pressure was 27[24–28] and 28[26–30] cmH2O; driving pressures, 16[14–19] and 18[15–20] cmH2O; modified oxygenation index, 29.8 ± 13.2 and 24 ± 13.7 cmH2O; mechanical power, 36.0 [29.3–45.2] and 41.6[31.5–49.6] J/min; P/F ratio, 126.3 [90.3–170.8] and 142 [99.3–199.3] mmHg and ventilatory ratio, 2.36 ± 0.89 and 2.55 ± 1.01. At day 1 of IMV, higher plateau and driving pressures, mechanical power and ventilatory ratio were significantly associated to mortality. These were also significantly associated to mortality at day 5 of IMV along with higher modified oxygenation index and lower P/F ratio.

Conclusion: In CARDS, it is important to rely on trends of indexes of poor respiratory mechanics properties (plateau pressure, driving pressure, modified oxygenation index and mechanical power) and indexes of poor ventilatory efficiency (oxygen exchange, P/F and carbon-dioxide clearance, ventilatory ratio).

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Inass Arhoun El Haddad (iarhounelhaddad@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-074

Rationale: COVID-19 first appeared in China in Wuhan, Hubei province,and then rapidly spread worldwide due to the highly contagious nature of this virus. On March 2, 2020, we experienced the first confirmed case of covid-19, and since then, the number of cases has continued to increase exponentially. The management of a case of COVID-19 differs according to the protocol of each country. In this study we were interested in the protocol of our intensive care unit.

Patients and methods/materials and methods: This was a single-center descriptive retrospective study of the records of 1381 patients admitted to intensive care at the Oujda University Hospital in Morocco. This work was carried out in the anesthesia and resuscitation department of our hospital. It runs from March 22, 2020, to September 30, 2022, including all patients with confirmed Covid-19 admitted to the intensive care unit during the study period, and excluding the patients who died within 48h of admission.

Results: The median period of observation following hospital admission was 8.52days for survivors and 10.89days for non-survivors. The median age of patients was 63.48years. Men were more represented than women with a percentage of 33% against 31.5%. Two hundred and three (14.6%) patients had obesity (defined as body-mass index [BMI] ≥ 30kg/m2). Patients presented to hospital a median of 6days (2–7) after symptom onset. The most common presenting symptoms were shortness of breath, fever, cough, myalgia, and diarrhea. Lymphocytopenia was common. Concentrations of IL-6, high-sensitivity C-reactive protein, ferritin, D-dimer, high-sensitivity troponin, and procalcitonin were elevated in most patients. Invasive mechanical ventilation (IMV) during hospitalization had been necessary in 306 cases (27.22%). Six hundred and ninety-three (61.65%) patients were discharged alive versus 431 patients (38.34%) were died. Analyzing the survival curve as a function of Kaplan–Meier days of hospitalization, we note that survival was increasingly nil in patients who exceed hospitalization days and this constitutes a mortality factor (figure). The mortality rate was higher in patients with a degree of lung parenchymal involvement between [75%-100%] on the chest CT scan on admission compared to patients with a lower degree of parenchymal involvement [50–75%] (p-value = 0.015).

Conclusion: Our study shows that age, sex, obesity, as well as certain comorbidities, and biological assessments on admission of COVID-19 patients are among the criteria that make COVID-19 disease serious as well as increase the mortality rate.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Ines Sedghiani (sedghiani.ines@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-075

Rationale: Since the beginning of the COVID-19, several studies showed a high incidence of venous thromboembolic events (VTE). They are frequently asymptomatic and they have adverse impact on patient prognosis. Our study objective was to identify predictive factors of VTE in COVID-19 patients.

Patients and methods/materials and methods: We conducted a retrospective study over a period of one year from March 01, 2020 to February 28, 2021. We included patients hospitalized for COVID-19 confirmed by virological and/or CT scan data. All patients had chest CT scan at admission. We have analyzed risk factors for VTE diagnosed at the first radiological imaging.

Results: Three hundred and twenty-three patients were included. They had a mean age of 65 ± 13years and a sex ratio of 1.31. The common comorbidities were hypertension (52%) and diabetes (44%). Respiratory symptoms were the principal reason for consultation (99.1%). The initial chest CT scan exam revealed an extent of parenchyma lung ≥ 50% in 45.5% of patients. Forty-six patients had a confirmed diagnosis of pulmonary embolism with an incidence of 14.2%. Age over 65years (p = 0.003; OR = 2; 95% CI [2–10]), female sex (p = 0.03), hypertension (p = 0.01; OR = 2.1), chronic heart failure (p = 0.01; OR = 2.17) and long-term therapeutic anticoagulation (p = 0.05; OR = 1.9) were identified as the risk factors of VTE among COVID-19 patients. On admission, clinical respiratory deterioration signs (p = 0.05) and an oxygen saturation lower than 86% (p = 0.01) were more common in patients with VTE. A Wells score above 2 and a SOFA score above 7 were associated with a higher risk of VTE (p < 10^-3). Patients developing VTE had higher white blood cell count (10.3 vs 8.6 × 10^9/L), platelets count (267 vs 227*10^9/L) and LDH levels (492 vs 360 UI/L) than non-VTE patients. D-dimer > 1600µg/L was predictive of a higher risk of VTE. Patients with pulmonary embolism had longer hospital stay (11 vs 8days) and higher in-hospital mortality (49% vs 27.5%).

Conclusion: The VTE incidence is high in COVID-19 patients. An age over than 65years, comorbidities and a D-dimer value greater than 1600µg/L were significantly associated with a higher risk of VTE among COVID-19 patients.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Nabil Bouguezzi (dr_nabil@live.fr)

Annals of Intensive Care 2013, 13(Suppl 1):FC-076

Rationale: The increasing age in the industrialized countries places significant demands on intensive care unit (ICU) resources and triggers debates about end-of-life care for the elderly. Actual data are leading intensivists to wonder whether breathing tubes should be used as often as they are.We aimed to investigate the patterns of 28-day mortality, by age group, among mechanically ventilated critically ill COVID-19 patients.

Patients and methods/materials and methods: It is a retrospective observational study conducted from March 2020 to December 2021, in a Medical ICU. Information regarding demographic, clinical characteristics and outcomes of Critically Ill COVID-19 patients was obtained from medical records. The population was divided into 3 groups according to age; < 60years old (G1), 60 to 80years old (G2), and > = 80years old (G3). We evaluated mortality rates between age groups. Kaplan Meier curves were used to describe 28-day mortality.

Results: Among the465 COVID-19 patients admitted in the study period, we analyzed 244 patients who had undergone IMV. They were mostly male patients. Two hunded and twelve patients (85.8%) had at least one comorbidity. The mean age was 65.3 ± 11.9. Patients’ presentation on admission was severe, all patients presented with a severe to critical COVID-19, of which 160 (64.8%) had a severe ARDS; median P/F ratio at admission was at 96 [76.3–139.8] mmHg. The mortality rate was high at 84.4%. We had 62, 163 and 19 patients in G1, G2 and G3, respectively. By age group, the mortality rates were 69.4%, 89% and 100% in patients in G1, G2 and G3, respectively. There was no significant difference in length of stay between different groups (9 [5–14], 9 [6–15] and 10 [8–14] days in G1, G2 and G3, respectively). Kaplan- Meier analyses of survival of patients under IMV are displayed in Figure 1.

Conclusion: Patients aged 80 and over seem to be a reasonable situation to discuss for ICU admission denial or at least for intubation denial.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Mathilde Phillips-Houlbracq (mathilde.phillips@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-077

Rationale: Among severe Sars-CoV2 pneumoniae, 20–40% are admitted in Intensive Care Unit for Acute Hypoxemic Respiratory Failure (AHRF) to receive non-invasive oxygen-support (NI-OS) and/or mechanical ventilation. The timing of intubation for COVID-AHRF remained controversial. The aim of this study was to compare day-60 mortality between patients intubated 5days after the hospitalization in ICU and those with similar criteria of AHRF after day 5, maintained under NI-OS.

Patients and methods/materials and methods: This observational retrospective study was performed with data from the French prospective multicenter (n = 11 ICUs) OutcomeReaTM database from February, 1^st 2020 to May, 30th 2022. Patients with HFAR were included if they had an ICU stay of more than 5days without intubation (under NIOS: CPAP, HFNC, NIV). The 60-day mortality rate of intubated patients after day 5 were compared to non-intubated patients (under NIOS) who harbored the same gravity of AHRF (FiO2 of 80% and having a PaO2 of less than 80mmHg) at the same length of stay.

Results: From February 2020 to August 2022, 1206 patients were admitted for COVID-19 pneumonia in the 11 French ICUs of OutcomeRea TM database. 387 patients were still in ICU and non-mechanically ventilated at day 5. Patients were mainly men (70%), median age of 66.4years and at least one comorbidity in 70% cases. 73,6% (n = 285) received steroids during hospitalization. 25.8% were finally intubated during stay of intensive care and the mortality rate was 26.8% at day 60 among the 387 patients. Patients intubated after day 10 (n = 28) had a poor prognosis compared to those still under NIOS after day 10 (n = 10) with 60-day mortalityrespectively 89.3% and 60% (p = 0.04), HR 2.43 (0.96–6.11) after adjustment for BMI and age.

Conclusion: These data suggest a poorer prognosis for patients intubated after 10days under NIOS for COVID AHRF. In this context, invasive ventilation should be avoided.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Tifany Vatignez (tifany.vatignez@hotmail.fr)

Annals of Intensive Care 2013, 13(Suppl 1):FC-078

Rationale: The COVID-19 pandemic provoked a surge of intensive care unit (ICU) patients with profound hypoxemia. Due to the specificities of this pneumonia and risk of viral aerosolization, the different oxygenation techniques have been debated. We used, however, high flow nasal oxygen therapy (HFNO) from the start of the pandemic in our ICU. Only few studies have evaluated the impact of the oxygenation strategy on the outcome of COVID-19 ICU patients during the different waves of the pandemic.

Patients and methods/materials and methods: We conducted a single center retrospective study to assess the impact of HFNO use in COVID-19 patients with severe hypoxemic acute respiratory failure (ARF) during the first 3 waves of the pandemic. All COVID-19 ARF patients admitted in our ICU between March 2020 and May 2021 were evaluated for their clinical characteristics, respiratory management and outcome. We also assessed the prognosis of patients according to the HFNO duration before intubation (HFNO late failure: > 48h and FiO2 ≥ 80% and flow rate ≥ 60L/min).

Results: 198 patients were included during the 3 first waves: 56, 85 and 57 patients, respectively. Despite similar clinical characteristics on admission, the use of HFNO increased significantly between the 3 waves (62.5%, 96.5%, 98.2% respectively; p < 0.001) with a trend toward for a higher HFNO success rate (46.4%, 65.9% and 63.2%; p = 0.057). When HFNO failed, intubation occurred later during the 3rd wave compared to the ICU admission (p < 0.001), duration of HFNO before intubation (1.8, 3.0 and 5.2days, respectively; p = 0.015) and with more severe underlying hypoxemia (PaO2/FiO2 ratio; p = 0.012). The length of stay and overall ICU mortality were similar between the 3 waves. Finally, all waves combined, the intubation time (early n = 19, intermediate n = 24, late n = 31) had no deleterious effect on patients’ outcomes (duration of ventilation, length of stay and ICU mortality).

Conclusion: Our study confirms that HFNO is a useful oxygenation strategy for the management of severe COVID-19 ARF patients. HFNO can avoid intubation in satisfactory respiratory conditions without compromising outcomes, including in patients with prolonged HFNO and late failure.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Jihene Ayachi (ayachijihen@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-079

Rationale: Non-invasive ventilation (NIV) has been largely applied to treat COVID-19 patients with moderate to severe acute respiratory failure (ARF). HACOR score has been used to predict non-invasive ventilation (NIV) failure. Updated HACOR score was proposed to evaluate NIV outcomes. This study aimed to analyze the ability of the HACOR and updated-HACOR score to predict NIV failure of critically ill patients admitted for COVID-19 pneumonia.

Patients and methods/materials and methods: A retrospective observational study including patients receiving NIV with confirmed COVID19, admitted in a 9-bed medical intensive care unit from September 2020 to December 2021. Scores were calculated one hour after NIV initiation. HACOR score associates heart rate, acidosis, consciousness, oxygenation and respiratory rate. Updated HACOR score associates HACOR score, SOFA score, pneumonia, cardiogenic pulmonary oedema, ARDS, immunosuppression and septic shock. Patients with updated-HACOR scores were classified into ≤ 7, 7.5–10.5, 11–14, and > 14. A receiver operating characteristic (ROC) curve was used to evaluate discrimination of HACOR and Updated-HACOR score for NIV failure.

Results: During the study period, 47 patients required NIV (first line and after high flow nasal cannula failure). Main characteristics were mean age 57.1 ± 12.3years, men predominance (53.2%), SAPS II 27.8 ± 8.9, P/F ratio 109 ± 54mmHg. The most common comorbidities were obesity (53.2%), diabetes (34%) and hypertension (31.9%). Mean Pressure support was 16 ± 2 cmH2O, mean PEEP 5.8 ± 0.8 cmH2O, median Fio2 78 and mean NIV use 4 ± 3.9days. HACOR score ≥ 5 was found in 36 (76.6%) patients, and updated HACOR score ≥ 11 points in 42 (89.4%) patients. Updated HACOR scores ≤ 7, 7.5–10.5,11–14, and > 14 one hour after initialization of NIV were found in 2.1%, 8.5%, 42.6%, and 46.8% of cases respectively. Thirty-three patients (70.2%) experienced NIV failure, among them 29 (87.9%) patients have HACOR ≥ 5 and 97% of them have an updated HACOR score ≥ 11 with (p-value at 0.005 and 0.009 respectively). The area under the curve (AUC) for predicting NIV failure was 0.74 (95% CI,0.57–0.91) when it was tested with HACOR score and 0.76 (95% CI,0.61–0.91) when updated HACOR score was used. The AUC in the range of 0.7–0.9 was interpreted as moderate diagnostic accuracy.

Conclusion: The HACOR and the updated HACOR scores are clinically useful bedside tools for the prediction of NIV failure in COVID19 pneumonia with hypoxemic respiratory failure. They could be helpful in decision-making when NIV is used.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Mohamed Ahmed Boujelben (m.ahmedboujelben@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-081

Rationale: Previous studies have shown that continuous anterior chest compression (CACC) can induce a paradoxical decrease in plateau pressure (Pplat) in persistent ARDS with severe impairment of the respiratory mechanics by limiting overdistension. This study aimed to assess the physiological effects of CACC in the early phase of ARDS and to compare them to prone position.

Patients and methods/materials and methods: This was a prospective single-center physiologic study. The patients with moderate-to-severe ARDS who were placed in prone position by the attending physician were eligible. The regional respiratory mechanics and distribution of ventilation were explored by esophageal catheter and electrical impedance tomography (Enlight 800, Timpel) during four consecutive steps: prone position (PP), supine position (SP1), supine position with CACC applying a pressure opposite the sternum equal to that observed during prone position, and supine position again (SP2). The primary endpoint was the end-inspiratory transpulmonary pressure derived from elastance ratio (PLinsp). Secondary endpoints included regional lung compliances.

Results: We present the results of the first seven patients included in this study. The median age was 60years [57–70] and the median BMI was 33kg/m2 [26–39]. All patients had pneumonia-related ARDS. The median tidal volume was 5.9mL/kg [5.9–6.1] and the median PEEP 12cm H2O [10–15]. PaO2/FiO2 was 96mmHg [77–113] and respiratory system compliance was 31.4ml/cm H2O [31.2–39.3]. The pressure exerted by the body weight at the sternum during PP was 72 cmH2O [60–86]. During CACC using the same pressure, two physiological responses were observed: either an increase in Pplat (non-responders, n = 3), or a paradoxical decrease in Pplat (responders, n = 4). In all responders, PLinsp decreased during CACC and both anterior and posterior regional lung compliances increased, suggesting both a reduction in pre-existing anterior overdistension and an improvement in posterior recruitment. The opposite was observed in non-responders. The effects of CACC on Pplat, PLinsp and regional lung compliances appeared comparable to the effects of PP according to the patients’ profiles.

Conclusion: The effects of CACC on PLinsp and regional ventilation distribution were comparable to those observed during prone position and may depend on the pre-existing level of overdistension.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Amazigh Aguersif (aguersif.a@chu-toulouse.fr)

Annals of Intensive Care 2013, 13(Suppl 1):FC-082

Rationale: Lung ultrasonography (LUS) has become an essential component of the evaluation and clinical management of patients admitted to the intensive care unit (ICU). The interpretation of LUS artifacts (A- and B-patterns), analysis of the pleura, and the visualization of real images (C pattern) have demonstrated usefulness for the differential diagnoses of acute respiratory failure (ARF). However, current methods are non-quantitative and have important drawbacks deriving from visually guided assessment of LUS data. Nevertheless, it is commonly acknowledged that these methods, which are based only on medical experts’ analysis, can be time-consuming, are user-dependent, and hold the risk of leading to oversimplified and potentially harmful diagnosis algorithms, particularly in the complex pathophysiological setting of critically ill patients. To accurately cope with these issues, computer vision approaches,built uponmachine learning (ML) algorithms have been recently proposed. These methods have potential to provide new computer-aided diagnosis for LUS data that could transform the way in which ICU practitioners assess and manage critically ill patients. However, it is worth noting that the data available in this field have demonstrated variable classifier’s accuracy and was exclusively limited to the analysis of B-lines.We followed a knowledge transfer approach from satellite to medical imaging based on semantic segmentation and signal processing. In order to do this, we designed a proof-of-concept study built upon one of the largest LUS datasets from severe COVID-19 patients reported to date.

Patients and methods/materials and methods: We prospectively recruited adult COVID-19 patients who were in ARF (pulse oximetry < 90% while breathing room air or respiratory rate > or = 30 breaths/min) at hospital admission. Overall, 5000 LUS frames from 78 patients affected by COVID-19 with different degree of severity were gathered and labeled.

Results: We provided an automatic pixel-wise classifier, which was able to accurately identify for the first time, all the main LUS patterns that can be observed in this setting (overall Area Under the Curve, AUC = 0.97; pleural line AUC = 0.99; A-pattern AUC = 0.96; B pattern AUC = 0.97; C-pattern AUC = 0.95).

Conclusion: The eventual integration of this model into ultrasound hardware seems plausible as a method to ultimately improve outcomes through a combined use of LUS with other components of critical care ultrasonography to yield a more comprehensive and accurate evaluation of the critically ill patient at point of care.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Iyed Matouk (maatouk.yed@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-083

Rationale: Coronavirus disease 2019 (COVID-19) is considered as a global health emergency. It can cause an acute respiratory distress syndrome (ARDS) requiring mechanical ventilation. The prone position is one of the methods used in the management of patients with this syndrome. Our study aimed to determine the impact of the prone position on the oxygenation of COVID-19 patients with ARDS.

Patients and methods/materials and methods: We conducted a retrospective study between March 2020 and May 2022 in the Intensive Care Unit of the University Hospital Fattouma Bourguiba of Monastir (Tunisia). COVID-19 patients who responded to the clinical criteria of the Berlin definition of ARDS and requiring mechanical ventilation were included. Data analysis was performed using the Statistical Package for Social Sciences (SPSS) version 21.0.

Results: In total,182 mechanically ventilated patients diagnosed with COVID-19 ARDS were enrolled in the study. Most patients were male (67.6%). The mean age was 63.5 ± 11years. Obesity was found in 44.9% of the patients. The median length of stay was 10.5 (Interquartile range (IQR): 5.7–16) days. The median SAPS II, APACHE II and SOFA score were 34 (IQR: 28- 4), 11 (IQR: 9–15), and 4 (IQR: 3–4), respectively. Patients were intubated after a median duration of 2days (IQR: 1–5) after admission. Mean compliance was 25.5 ± 9.8ml/cmH2O. The median number of prone position sessions was 3 (IQR: 1.5–6). PaO2/FiO2 ratio was significantly improved from 103.7 ± 44.9mmHg in supine position to 126.61 ± 59.7mmHg in prone position (p < 10^-3).

Conclusion: Our study shows that the change from supine to prone position improves oxygenation among COVID-19 ARDS patients. Further studies should be conducted on a larger number of patients in order to better guide intensivists in the treatment of ARDS due to COVID-19.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Ines Sdiri (sdiri.ynes@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-084

Rationale: Acute respiratory failure is a frequent cause of hospitalization in the elderly. Symptoms are often atypical and responsible for a delay in diagnosis. The aim of our study was to describe the epidemiological, clinical and evolutionary particularities of acute respiratory failure in the elderly who consulted in the emergency room.

Patients and methods/materials and methods: It was a retrospective and descriptive study including elderly subjects admitted to the emergency and medical ICU between January 1, 2022 and August 1, 2022. We collected clinical, paraclinical, therapeutic and evolutionary data. We defined elderly age range over 65years.

Results: We canalyzed 72 patients admitted for acute respiratory failure. The median age was 77.5years [66–100]. The sex ratio was 1.2. We noted 40 (55.6%) patients with chronic hypertension. Nineteen (26.3%) patients had chronic heart failure, 10 (13.9%) had chronic obstructive pulmonary disease, 11 (15.3%)had chronic renal failure, 5 (6.9%) patients were on hemodialysis, and 9 (12.5%) of whom had a history of stroke. The reasons for consultation were dyspnea in 48 patients (66.7%), altered neurological status in 7 cases (9.7%) and asthenia in 17 cases (23.6%). The acute respiratory failure etiology was COVID-19 infection in 63 patients (87.5%) associated with a bacterial superinfection in 4 cases, a pulmonary embolism in 8 cases, a left heart failure in 17 cases and an acute coronary syndrome in 10 cases. Left heart failure was found in 5 patients (7%) and bacterial pneumonia in 4 patients (5.5%). Twelve patients (16.7%) required Continuous Positive Airway Pressure (CPAP), 18 (25%) received non-invasive ventilation in pressure support mode, 8 (11.7%) were put on high flow oxygen, and 7 (9.7%) were intubated and ventilated in controlled mode. The most common complications were acute renal failure in 12 patients (16.7%) and upper gastrointestinal bleeding in 4 cases (5.6%). The evolution was fatal in 23 cases (31.9%). Three patients (4.2%) were discharged on long-term oxygen at home.

Conclusion: Acute respiratory failure in the elderly remains a serious pathology with a high rate of mortality. Its main etiology was Covid-19 during the pandemic. A specific emergency department for geriatric patients must be evaluated to optimize the management of these patients who suffer from several chronic diseases.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Rida Cheikh Youssef (rida.cheikhyoussef@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-085

Rationale: Non-rebreather face masks (NRM) are used in patients with profound hypoxemia, they have a reputation for delivering very high FiO2. (1) However, for some authors, these masks provide the same FiO2as nasal cannulas, i.e., an increase of 3% of FiO2per litre of O2administered. (2) The purpose of the study was to bench analyze the FiO2delivered by the NRM at different oxygen flow rates, and to verify if the increase in FiO2is equal to 3% per litre of O2administered.

Patients and methods/materials and methods: On a bench study, the FiO2delivered by a total NRM was analysed with an adult lung model (Michigan Instrument Inc., Grand Rapids, MI), driven by a mechanical ventilator (Servo I™ Maquet) set to volume control mode. Resistive and elastic characteristics of the test lung were set at 5cm H2O/L/sec and 0.06 L/cm H2O, respectively. Data acquisition was performed via the Labscribe 3TM software (Iworx®, United States) after a stabilization period. The parameters were modified as follows: O2flow: 10 and 15 L/min, minute ventilation (MV) from 6 to 21 L/min, Ti/Ttot: 0.33 and 0.25.

Results: The agreement between calculated and measured FiO2(CMFiO2) was expressed as proposed by Bland and Altman. Thus, the bias and limits of agreement are reported (95% CI for the difference between measurements). Next, an analysis of calculated and measured FiO2as a function of MV was performed. Results: When MV increases, both the FiO2and the difference between CMFiO2decreases. The agreement between CMFiO2was: Bias 1.46/limits of agreement from -13.89 to 16.82. The median value of FiO2increase by litre of oxygen administered is equal to 3.2% (IQR: 2.7 to 3.6). There is not a statistically significant difference between CMFiO2(p = 0.331). When the O2flow rate is equal to 10 L/min, the "3% rule” is reached for MV of 15 L/min, while for an O2flow rate of 15 L/min, the "3% rule” is reached for MV of 10 L/min (Figure 1).

Conclusion: In this study, the total NRM never delivered a FiO2close to 100%. The "3% rule” was statistically verified. However, the limits of agreement between the FiO2obtained with the "3% rule" calculation and those obtained at the bench are very large. We conclude that the calculation of theoretical FiO2with NRM can be highly variable depending on O2flow and MV.

Reference 1

Wagstaff T, Soni N. Performance of six types of oxygen delivery devices at varying respiratory rates. Anaesthesia. May 2007;62(5):492–503.

Reference 2

Frat JP, Thille AW, Mercat A, Girault C, Ragot S, Perbet S, et al. High-flow oxygen through nasal cannula in acute hypoxemic respiratory failure. N Engl J Med. June 4, 2015;372(23):2185–96.

Compliance with ethics regulations: N/A.

Correspondence: Yann Renaud (yann_r@bluewin.ch)

Annals of Intensive Care 2013, 13(Suppl 1):FC-086

Rationale: Patient-ventilator synchrony refers to the matching between the ventilator pressurization and the patient’s neural inspiratory command. During non-invasive ventilation (NIV) synchrony is difficult to achieve because of inherent leaks, even with automatic leak compensation. Automated setting adaptation using algorithms designed to perform real-time analysis of airway pressure and flow-time curves to detect asynchrony is now available (IntelliSync + ®, Hamilton, Bonaduz, Switzerland). This bench study aim was to evaluate the benefits of IntelliSync + ® on patient-ventilator synchrony in case of obstructive respiratory mechanics.

Patients and methods/materials and methods: To simulate inspiratory effort, one chamber of a two-chamber Michigan test lung (Michigan Instruments, Grand Rapids, USA) was connected to a driver ventilator (P0.1 = 5,6 ± 0,2 cmH2O). The second chamber was connected to a Hamitlon-S1 ventilator (Hamilton medical, Bonaduz, Switzerland) equipped with IntelliSync + ® and set in pressure support (PS) with the NIV mode activated. Obstructive respiratory mechanics was obtained by placing a resistance in the ventilator circuit (R = 24,6 ± 0,7 cmH2O/L/sec). Continuous leak was created in the circuit to reveal triggering asynchronies, whereas cycling-off asynchronies were revealed by creating a leak limited to inspiration (leak flow of 20 L/min). PS was set at 14cmH2O. Synchrony was assessed at steady state (number of asynchronies and measurement of trigger delay and inspiratory time in excess). Pressurization capacity was assessed using the triggering pressure–time product (PTP) and the PTP300 and 500. Measurements were made without IntelliSync + ® (A), with IntelliSync + ® activated during inspiration (B), cycling off (C) and both (D). Results of 5 consecutive ventilatory cycles are reported as medians, 1st and 3rd quartile. Comparisons were performed with one-way ANOVA on ranks (Bonferroni correction for pairwise comparisons).

Results: The results are illustrated in the following table.

Conclusion: In this bench study, IntelliSync + ® shortened the trigger delay, improved expiratory synchrony and optimized pressurization capacity in a model of an obstructive patient in respiratory distress ventilated with NIV.

Compliance with ethics regulations: N/A.

Correspondence: Christian Caillard (cricaillard@orange.fr)

Annals of Intensive Care 2013, 13(Suppl 1):FC-087

Rationale: A large choice of recent ventilators, dedicated or not, are available to perform noninvasive ventilation (NIV) in intensive care units (ICU). We conducted a bench study to compare the technical performances of different ventilators for NIV, and particularly their impact on patient-ventilator asynchrony (PVA) according to different respiratory conditions.

Patients and methods/materials and methods: Ventilators have been evaluated on a previously validated bench test for NIV consisting of a 3D mannequin head connected to an artificial lung model via. a dual-limb circuit with a non vented facial mask: Five ICU ventilators with the NIV algorithm activated, 2 dedicated NIV ventilators and 1 transport ventilator were tested according to 3 patient profiles simulated with the lung model (normal, obstructive, and restrictive lung) and different respiratory conditions: 3 levels of non-intentional leak (0, 15 and 30L/min), 2 levels of pressure support (8 and 14 cmH2O) and 2 respiratory rates (15 and 25 cycles/min).

Results: Median total leak was not different between the ventilators (p = 0,09). The asynchrony index (AI), defined as the [number of asynchrony events/(ventilator cycles + wasted efforts) × 100],was higher with ICU ventilators than with dedicated NIV ventilators (4% [0;76] vs 0% [0–15]; p < 0.05) (Figure 1). The AI was different between all ventilators (p < 0.001). The AI was higher with ICU ventilators for the normal and restrictive profiles as compared to dedicated NIV and transport ventilators (p < 0.01) but not different between ventilators for the obstructive profile. The overall median AI was correlated with the mean total leaks (p < 0.01), but several ventilators had a better adaptation to an increase of the level of leaks, the AI increasing less than for others (p < 0.01). Auto-triggering represented 43% of overall PVAs. AI was higher with a pressure support of 14cmH2O than 8cmH2O, there was no difference of AI with a respiratory rate of 25cycles/min than 15cycles/min. Triggering delay, cycling delay, pressure–time product, pressure rise time and pressure in the mask were different between all ventilators (p < 0.01). There was no difference between ventilators for minute ventilation (p = 0.21) and peak pressure (p = 0.24). Dedicated NIV ventilators induced a lower pressure–time product than ICU and transport ventilators (p < 0.01). The rising time was found higher with dedicated NIV ventilators than with others (p < 0.01).

Conclusion: Despite the implementation of NIV algorithm, the most recent ICU ventilators appear to be less efficient than dedicated NIV ventilators, particularly in terms of PVA and work of breathing. These performances could however change according to the underlying respiratory disease and/or the number of leaks.

Compliance with ethics regulations: N/A.

Correspondence: François Lellouche (francois.lellouche@criucpq.ulaval.ca)

Annals of Intensive Care 2013, 13(Suppl 1):FC-088

Rationale: The optimization of the inspiratory and expiratory circuits used with heated humidification is necessary to avoid condensation and related problems, such as auto-triggering, difficulties to trigger the ventilator, or increase in expiratory resistances (1). The objective of the study was to compare different expiratory circuits of different heated wire humidifiers.

Patients and methods/materials and methods: On a bench model simulating humidified expired gas (absolute humidity of 35 mgH2O/L), we have measured hygrometry of expiratory gases at proximal side (corresponding to the Y-piece), and distal side (ventilator inlet) of expiratory circuits. The circuits evaluated were Inspired ref. 51005683 (Vincent medical), FP950 (ref:950A81J, Fisher&Paykel), MR850 (RT210, F&P) and Evaqua2 (RT380, F&P), at different room temperatures (22–24°C and 28–30°C). The ventilator settings were Assist Control, tidal volume 400ml; respiratory rate 25/min, PEEP 5 cmH2O, FiO2 0.21, Flow 60 lpm. Hygrometry was measured with the psychrometric method and 3 measurements were performed for each condition. Expiratory Tidal Volumes in each condition were all recorded from the ventilator screen.

Results: Main results are shown in the figure. The absolute humidity was reduced along the expiratory circuit with the FP950 and the Evaqua circuits (RT380). To avoid condensation in the expiratory limb, the VHB20 increases the temperature along the expiratory limb (leading to a reduction of relative humidity with stable water content). A similar functioning was observed with the MR850/RT210 circuit. No condensation was found along the expiratory limb for the different tested circuits in the bench study conditions. Results at different ambient temperatures were very closes. Despite high temperatures at the inlet ventilator, the tidal volumes measured by the ventilator were very close.

Conclusion: Different strategies are used to avoid the occurrence of condensation in the expiratory limb. Two circuits (FP950 and Evaqua2) use ‘’porous’’ materials allowing a reduction of the humidity along the expiratory line. Two circuits (VHB20 and RT210) use a strategy of increased temperature allowing a reduction of the relative humidity to limit the risk of condensation. However, this strategy may alter the tidal volume measurements (which has not been shown with the ventilator used in the study), and may increase the resistance if HME filters are used instead of filters (with low hygrophobic/hygroscopic properties) (1).

Reference 1

Tonnelier A, et al. Impact of humidification and nebulization during expiratory limb protection: an experimental bench study. Respiratory care 2013; 58(8): 1315–22.

Compliance with ethics regulations: N/A.

Correspondence: Alexandre Broc (alexandre.broc@airliquide.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-089

Rationale: Bag-Valve-Mask (BVM) is the most commonly used device for pre-oxygenation and ventilation during cardiopulmonary resuscitation (CPR). A minimal FiO2 above 85% is expected during pre-oxygenation [1] while insufflated volume (VTi) should not exceed 8ml/kg during CPR [2]. The objective is to compare the performances of BVM in simulated pre-oxygenation and during CPR settings.

Patients and methods/materials and methods:

Nine BVM were evaluated as follows:

1. Pre-oxygenation: Moderate (PEEP 0) and severe (PEEP 0 and PEEP 5)spontaneous breathing profiles were simulated on an active lung model (ASL 5000). Actual FiO2 (inspired fraction of oxygen) was measured and CO2 rebreathing was evaluated.

2. CPR: BVM ventilation was performed by healthcare professionals belonging to three different emergency or ICU departments (n = 36) on a specific manikin. Ventilation was tested during mechanical Interrupted Chest Compressions strategy (30:2). The same procedure was repeated by healthcare professionals (n = 3) who provided manual ventilation on two cadavers with three BVM.

Results:

1. Pre-oxygenation: FiO2 was lower than 85% for three BVM in severe profile and for two BVM in moderate profile (PEEP 0). In eight out of nine BVM for which a PEEP valve was attached (PEEP 5), FiO2 was above 85% but at the expense of a slight increase in resistive load. One BVM induced CO2 rebreathing.

2. Ventilation during CPR: on the bench, mean VTi were above the predefined target range for lung protective ventilation (4–8ml/kg of PBW) for three BVM. Similar results were observed on the cadaver model.

Conclusion: BVM tested demonstrated varying performances in pre-oxygenation and CPR settings. The minimal 85% FiO2 target expected during simulated pre-oxygenation was not guaranteed for three BVM. During simulated CPR, VTi actually delivered varied significantly and was excessive for three BVM. The BVM working principle explained at least in part most of these observations.

Reference 1

Frerk, C., Mitchell, V. S., McNarry, A. F., Mendonca, C., Bhagrath, R., Patel, A., O’Sullivan, E. P., Woodall, N. M., & Ahmad, I. (2015). Difficult Airway Society 2015 guidelines for management of unanticipated difficult intubation in adults. British Jour.

Reference 2

Acute Respiratory Distress Syndrome Network, Brower, R. G., Matthay, M. A., Morris, A., Schoenfeld, D., Thompson, B. T., & Wheeler, A. (2000). Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the ac.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Elise Artaud-Macari (eliseartaudmacari@yahoo.fr)

Annals of Intensive Care 2013, 13(Suppl 1):FC-090

Rationale: In de novoacute respiratory failure(ARF), the worsening of underlying pulmonary injuries could be related to the risk of P-SILI (patientself-inflicted lung injury) during spontaneous breathing or VILI (ventilator- induced lung injury) with mechanical ventilation. To better assess the potential consequences of spontaneous or mechanical oxygenation techniques,we aimed to design a model of de novoARF and assessthe conditions that favor volo- and atelec-trauma.

Patients and methods/materials and methods: Based on our previously described model of ventilation in healthy subjects and patients with lung fibrosis (altered compliance (C)) on a 2-compartment (A1 and A2) artificial mechanical lung simulator, we have simulated conditions of breathing patterns and inspiratory efforts in de novoARF based on physiological data from the literature and our cohort of healthy subjects at exercise. Inspiratory efforts were adjusted to maintain a minute ventilation of 28 L/min. Worsening of ARF was simulated by increasing the proportion of altered alveoli, and the model was therefore composed of a proportion (X) of alveoli with basal compliance (C1 = 200ml/cmH₂O) and a proportion (Y) of alveoli with altered compliance (C2 = 50ml/cmH₂O), where X + Y = 100%.Resistance was kept constant (5 cmH₂O/L/s) in all simulations. The following parameters were measured: tidal volume (TV) distribution, end-expiratory lung volume (EELV), driving pressure (DeltaP), driving transpulmonary pressure (DeltaP_tp), dynamic alveolar strain (Strain_alv), mechanical power (MP), time lag between inspiratory flow in A1 and A2 (Deltat _(Q1-Q2)).

Results: During ARF worsening,TV remained stable but was found higher in the compartment with highestcompliance. EELV decreased dramatically (as low as 167ml) especially in the compartment with the lowest compliance, suggesting a potential risk of alveolar collapse. From mild to very severe ARF, DeltaP increased (from 11.9 to 14.4 cmH₂O), DeltaP_tp increased (5.9 to 14.2 cmH₂O), Strainalv increased (0.8 to 1.2), suggesting a potential overdistension, and MP increased (27.8to 45.1J/min). Deltat _(Q1-Q2) was postitively correlated with the difference of compliance between A1 and A2 (r = 0.98, CI_95% (0.7783; 0.9990), p = 0.002) and increased when this difference increased, suggesting Pendelluft phenomenon.

Conclusion: This physiological bench study designed for the first time a mechanical ventilation model that can reproduce lung inhomogeneity and the mechanisms of volo- and atelec-trauma in de novoARF. This model could be useful in the future to evaluate the impact of various oxygenation strategies in this clinical scenario.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Alice Vuillermoz (alice.vuillermoz@chu-angers.fr)

Annals of Intensive Care 2013, 13(Suppl 1):FC-091

Rationale: Several inhaled nitric oxide (iNO) devices are available with different modes of iNO administration: “basic continuous administration” (MiniKINOX, Cahouet), “guided” (OptiKINOX, ALMS) with an administration only during inspiratory-time and “synchronized-to-ventilator” (EZKINOX, EKU, SoKINOX, iNOsystems) which adapt continuously their administration to the total flow in the inspiratory limb. We hypothesized that the different bias-flows among ICU ventilators may affect iNO delivery, accuracy, NO₂ formation and gas consumption depending on the device's capability to accurately detect and measure bias-flow.

Patients and methods/materials and methods: Four iNO devices were evaluated on a Michigan test lung (ARDS model), connected to four ventilators (V800, Dräger; R860, GE Healthcare; E500, Löwenstein andServo-i, Getinge) on volume-controlled ventilation. We tested four bias flows (2, 2.5, 10 and 30 L/min). NO was administered just before the heated humidifier, on the inspiratory limb. Four NO concentration settings were set: 5, 10, 14 and 20ppm. Actual NO and NO₂ concentrations were measured in the test lung by NO and NO₂ electrochemical cells (SoKINOX). NO-N₂ mixture flow was measured by TSI-flowmeter to assess gas consumption.

Results: With minimal bias flow, almost all the devices are able to achieve target iNO concentration except the OptiKINOX (16 [15; 17] ppm).Increasing the bias flow to 30 L/min resulted in a significant reduction of iNO concentration except for the two “synchronized-to-ventilator” devices that maintained iNO inside the target. As a result, NO/N₂ consumption was higher with these devices (Figure 1). NO₂ was below the toxic threshold in all experimental conditions.

Discussion: Bias flow setting is usually hidden or neglected by the caregiver. By increasing the quantity of fresh gas circulation in the inspiratory (and expiratory) circuit-limb, bias flow may dilute iNO concentration, resulting in iNO under administration. By directly detecting the inspiratory flow, new devices are able to adapt properly to iNO administration, thus maintaining iNO actual administration in the prest target.

Conclusion: Differences in ventilator bias flow can generate large variations on iNO administration accuracy. New devices detecting and compensating the total flow (inspiratory and bias flow) accurately deliver iNO concentration (EZKINOX and SoKINOX). On the contrary, devices that do not assess bias flow (MiniKINOX, OptiKINOX) significantly under deliver iNO.

Compliance with ethics regulations: N/A.

Correspondence: Samira Kalouch (dr.kalouch@gmail.com)

Annals of Intensive Care 2013, 13(Suppl 1):FC-092

Rationale: Severe acute asthma is an asthma attack that does not respond to well-conducted treatment, or whose intensity is unusual in its evolution or its symptomatology, putting the vital prognosis at risk. The aim of our study is to identify the prognostic factors of severe acute asthma in children hospitalized in the pediatric intensive care unit.

Patients and methods/materials and methods: This is a retrospective and analytical study, which involved 81 children, hospitalized for severe acute asthma in the pediatric intensive care unit, during a period from February 2013 to October 2022. The different variables collected at admission were analyzed and compared between two groups: deceased patients and survivors. Statical analysis (univariate and multivariate) by logistic regression was performed using SPSS version 23 software.

Results: The frequency was 2%. The average age was 5.5years, with a predominance of females. There were 2 peaks of admission in April and October (12%). 4% of patients were mechanically ventilated. 46% of the patients reported having a trigger for the current attack, 27% of the patients had a viral infection, and 7% of patients suffer from passive smoking. 77% of the patients were under asthma treatment, however 36% had a bad compliance and 17% stopped the treatment. At admission, 64% of the patients had SpO2 between 80 and 90%. 42% of patients received magnesium sulfate, 74% of patients had antibiotic therapy. The mean duration of intensive care stay was 3.71days. Overall, 15% of patents died. In univariate and multivariate analysis, the factors significantly influencing mortality in our study were: parental smoking, neurological faiure, hypercapnia and acidosis, mechanical ventilation, oxygenation by non-rebreather mask, magnesium sulfate and antibiotic therapy administration.

Conclusion: Comprehensive management of the asthmatic child is essential, as well as educating parents about the value of treatment and the severity of asthma disease, and the advantage of rapid management in acute severe asthma.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Elena Fernandez Fernandez (efernandez@aerogen.com)

Annals of Intensive Care 2023, 13(Suppl 1):FC-093

Rationale: According to the latest World Health Organisation Figures an estimated 2.4 million new-borns died from preventable and treatable diseases [1]. Amongst the leading causes were respiratory infections such as pneumonia. Supplemental oxygen in conjunction with aerosol therapy has become a mainstay in the treatment of respiratory illnesses in paediatric and neonatal patients in the critical care setting. The reliable supply of oxygen to prevent hypoxaemia is critical to improving patient outcomes [2]. Oxygen concentrators are one such means of facilitating reliable, medical grade supplemental oxygen generation and delivery, particularly in regions with lesser resources or where bottled oxygen is not readily available. The objective of this study was to investigate the effects of various low oxygen concentrator flow rates on aerosol drug delivery to a simulated spontaneously breathing neonatal patient.

Patients and methods/materials and methods: 500µg of salbutamol was aerosolised using an Aerogen Solo vibrating mesh nebuliser and Pro-X controller (Aerogen, IRE). The aerosol was delivered using the Bitmos Oxy 6000 oxygen concentrator (Bitoms GmbH, DE), via a nasal cannula (Hudson RCI Teleflex Comfort flo, Flexicare Medical, UK) at flow rates of 0.2, 0.3 & 0.5 LPM to a 3D printed anatomically correct head model of a 4-month-old (RONAN model). A simulated neonatal breathing pattern (BR: 40 BPM, Vt: 25mL, I:E: 1.0:3.0) was generated using a breathing simulator (BRS 2100, Copley, UK). The level quantity of aerosol available at the level of the trachea was measured using UV-spectrophotometry at 276nm. A one-way ANOVA test determined statistical significance at p < 0.05.

Results: The results of this study are presented in Figure 1.

Conclusion: In this study, it was found that low oxygen concentrator gas flow rates had a statistically significant effect on the quantity of aerosol available at the level of the trachea in a simulated spontaneously breathing neonatal patient.

Reference 1

World Health Organisation (n.d.). Child mortality (under 5years) https://www.who.int/news-room/fact-sheets/detail/levels-and-trends-in-child-under-5-mortality-in-2020

Reference 2

Enarson, P., LaVincente, S., Gie, R., Maganga, E. & Chokani, C. Implementation of an oxygen concentrator system in district hospital paediatric wards throughout Malawi. Bulletin of the World Health Organisation, 2008 (86); 344–348.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Simon Mellul (simonmellul@orange.fr)

Annals of Intensive Care 2023, 13(Suppl 1):FC-094

Rationale: In critically ill children, respiratory distress is clinically assessed using child inspection, auscultation and pulse oximetry. This assessment relies on caregivers’ expertise to analyze these signs including retraction signs. The development of supervised machine learning methods has the potential to help in the analysis of retraction signs in real time using video image. However, these methods need video annotation by caregivers to identify retraction signs. The aim of our study was to create an annotated video database of children with respiratory distress.

Patients and methods/materials and methods: In a video database (MEDEVAC database), we classified the videos of children according to the sign of retractions they presented. The database included 30min RGB, 3D and thermography videos of critically ill children admitted in the pediatric intensive care of our institution. To annotate the MEDEVAC database, we collected demographics data (age, gender, disease) and patient’s sedation clinical scores at the time of video collection, from the electronic medical record. We classified seven categories of retractions signs on videos: sous costal, inter costal, sous sternal, supra sternal, supra clavicular, thoraco-abdominal asynchrony and generalized signs. A caregiver clinically scored retraction signs using a qualitative scale (+, ++, +++). The database is approved by the review ethical committee (number 2016-1242) as well as the respiratory distress analysis by machine learning methods (number 2020-2276).

Results: 176 videos were classified. The reason for admission was respiratory failure in 51.1% of cases (including 51 children with bronchiolitis), post cardiac surgery in 24.4% of cases and 24.5% other reasons. 36 children were mechanically ventilated with an endotracvheal tube (iMV), 59 children were on none-invasive mechanical ventilation (NIV), 22 children had high flow nasal canula (HFNC) and 9 were breathing spontaneously without any respiratory support (VS). In children < 6years old (yo), the median (range) comfort B score was 12 (20–6) in iMV group and the median (range) FLACC scale was 0 (8–0) in the other groups. In children > 6 yo, we used a numerical scale of pain named EN which median (range) was 0 (2–0). The distribution of retractions signs was: 73 sous costal, 27 inter costal, 41 sous sternal, 14 supra-sternal, 5 supra clavicular retraction. Only 4 thoraco-abdominal asynchrony were seen as videos were registered when children were stabilized.

Conclusion: We annotated a large video database of critically ill children with respiratory failure. This video database will help for the development of clinical decision support system aimed at early recognition of respiratory distress.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Amal Miraoui (amal.miraoui92@gmail.com)

Annals of Intensive Care 2023, 13(Suppl 1):FC-095

Rationale: Neonatal respiratory distress (NRD) is the most common cause of admission in neonatal intensive care units with high morbidity and mortality. Such an emergency justifies the need of a diagnostic tool that is accessible, fast and reproducible. The aim of this study was to evaluate lung ultrasound (LUS) performance in NRD and interoperator agreement level for LUS results.

Patients and methods/materials and methods: We conducted a prospective, observational and evaluative study in our pediatric intensive care unit between June the 1st, 2020 and June the 30th, 2021 including newborns admitted for NRD aged under 24h of life.

Results: During the study period, we included 76 newborns. The median gestational age was 37weeks [IQR: 34 + 1 − 38 + 3]. The median weight was 2950g [IQR: 2412–3500]. The median age at transfer was 7h [IQR: 5–12]. Clinical findings at admission were a median respiratory rate of 77 cycles/min [IQR: 69–85], a median Silverman score of 4.5 [IQR: 3–5] and a median FiO₂ of 50% [IQR: 30–60]. Transient tachypnea of the newborn (TTN) was the most common etiology in our study. Agreement between LUS and reference diagnosis was moderate (kappa = 0.56; CI 95%: 0.43–0.69). LUS performance was studied for the 3 most common etiologies and concluded to a sensitivity and a specificity of 61.2% and 88.9% in TTN, 76.9% and 90.5% in respiratory distress syndrome (RDS), 100% and 98.5% in meconium aspiration syndrome (MAS). Interoperator agreement was almost perfect (kappa = 0.94; CI 95%: 0.88–1.00).

Conclusion: Lung ultrasound was performant in identifying the etiology of neonatal respiratory distress. A great interoperator agreement has been proved. This diagnostic tool should be introduced in all neonatal and pediatric intensive care units.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Karima Amanzoui (amenkarima@gmail.com)

Annals of Intensive Care 2023, 13(Suppl 1):FC-096

Rationale: Inhalation of the foreign body is one of the main causes of death by domestic accidents in children. Asphyxia remains the main immediate risk, but very serious sequelae may appear if the extraction time is prolonged. With advances in instrumentation and anesthesia, the severity of this incident has been greatly reduced. The aim of this work was to study the epidemiological, clinical, para-clinical, therapeutic criteria as well as the anesthetic modalities of the extraction of inhaled foreign bodies at the ABDERRAHIM HAROUCHI Children’s University Hospital in Casablanca.

Patients and methods/materials and methods: Retrospective study carried out in the pediatric intensive care unit of the CHU IBN ROCHD during the years 2014–2015, which allowed the recruitment of 42 cases of foreign bodies extracted under general anesthesia.

Results: The inhalation of foreign bodies was frequent between 6months and 3years (74%), and the diagnosis was based on the search for the syndrome of penetration which was present in 88% of the cases. Chest X-ray was normal in 50% of cases and foreign bodies were radiopaque in 12% of cases. The first-line treatment was the extraction of foreign bodies by rigid bronchoscopy under general anesthesia. The nature of foreign bodies was dominated by plants including peanuts (28.57%) with a predominance of intra-bronchial location (90.48%). The evolution was favorable for most patients, one patient had a cardiorespiratory arrest following a severe bronchospasm, two patients presented a pneumothorax of great abundance and three patients presented a bronchial spasm.

Conclusion: Early extraction of foreign bodies is essential to avoid the risk of potentially serious medium and long-term sequelae. We must also strengthen the prevention and education of parents, as well as their training on the Heimlich maneuver, in order to ensure good vigilance, and a possible reduction in the incidence of this serious accident.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Jacques Lacroix (jlacroix052@gmail.com)

Annals of Intensive Care 2023, 13(Suppl 1):FC-097

Rationale: There is increasing interest in using electronic medical data monitoring system (eMDMS) to record outcome data for clinical trials. We provide an analysis of the P-OpTTICCA study, a pilot randomized controlled trial (RCT) of red cell transfusion thresholds in children admitted to the pediatric intensive care unit (PICU) of four international hospital sites. The aim was to describe the challenges and lessons learnt for using electronic systems to abstract data directly from an eMDMS and to transfer these data into the case report form (CRF) of the study.

Patients and methods/materials and methods: 120 patients were enrolled from the PICU of four university-affiliated hospitals In Canada, France and United Kingdom. The basic design of P-OpTTICCA is described in the registry “clinicalTrials.gov” (NCT03871244).

Results: All data were extracted manually in one site. Most data (88%) were extracted electronically in the three other sites, using the eMDMS IntelliSpace Critical Care and Anesthesia (ICCA, Philips Medical Systems, Royal Philips Electronics, Netherland). We identified a number of challenges at the three sites supporting use of electronic data. Specific limitations were identified for reports of co-morbidity. Electronic data flows needed to be tailorred to each site; for example, some data were not stored in the same fields in their eMDMS even though the three PICUs used ICCA. Some laboratory data were stored in different locations. Like hematocrit which can be found in hematology and blood gases fields. Different unique patient identifiers were used, including for the same patient. Local algorithms (e.g., to extract data and calculate PIM2 and PELOD-2 scores) were not reliable and needed to be validated manually.

Discussion: Despite the many advantages of electronic systems for collection of routine data for analysis in clinical trials, our study highlighted the importance of understanding the full IT processes at all sites. Multiple validations are necessary to check data collection in the data warehouse, and data transfer from the warehouse into the study database. Such validations must be done in each site because some data were not located in the same field in the eMDMS of different sites even though they used the same software.

Conclusion: These electronic set up and validation processes must be taken into account in the trial budget and execution.

Compliance with ethics regulations: Yes in clinical research.

Correspondence: Maxence Fernandes (maxence.fernandes@chru-strasbourg.fr)

Annals of Intensive Care 2023, 13(Suppl 1):FC-098

Rationale: Despite the life-saving benefit of extracorporeal membrane oxygenation (ECMO), ECMO in Pediatrics iremains a risky procedure with a high mortality rate because of the pathology underneath but also given the complications. This work aims to analyze 10years of ECMO’s practice in our University Hospital of Strasbourg, France.

Patients and methods/materials and methods: We analyzed in a retrospective way data from all patients supported by veno venous or veno-arterial or VA ECMO from 2010 to 2020.

Results: Data from 59 patients were analyzed. Our population had a median age of 5.5months and a median weight of 5.1 kgt. 80% of ECMO were AV. For around 60% of patients, ECMO was implanted after cardiac surgery. Overall, 46% of patients died, and acute kidney injury was the only factor significantly associated with death in our population. The most frequent adverse outcomes were hemorragic (50%), thrombotic (30%), infectious (19%) and neurological (12%).

Conclusion: In the 10years period, we used ECMO for 59 patients for various indication. Acute Kidney Injury was the only risk factor for death in our cohort. Hemorragic and thrombotic adverse outcome are by far the most prevalent.

Compliance with ethics regulations: Yes in clinical research.