; for the CORIMUNO-19 Collaborative Group IMPORTANCE Severe pneumonia with hyperinflammation and elevated interleukin-6 is a common presentation of coronavirus disease 2019 (COVID-19). OBJECTIVE To determine whether tocilizumab (TCZ) improves outcomes of patients hospitalized with moderate-to-severe COVID-19 pneumonia. DESIGN, SETTING, AND PARTICPANTS This cohort-embedded, investigator-initiated, multicenter, open-label, bayesian randomized clinical trial investigating patients with COVID-19 and moderate or severe pneumonia requiring at least 3 L/min of oxygen but without ventilation or admission to the intensive care unit was conducted between March 31, 2020, to April 18, 2020, with follow-up through 28 days. Patients were recruited from 9 university hospitals in France. Analyses were performed on an intention-to-treat basis with no correction for multiplicity for secondary outcomes. INTERVENTIONS Patients were randomly assigned to receive TCZ, 8 mg/kg, intravenously plus usual care on day 1 and on day 3 if clinically indicated (TCZ group) or to receive usual care alone (UC group). Usual care included antibiotic agents, antiviral agents, corticosteroids, vasopressor support, and anticoagulants. MAIN OUTCOMES AND MEASURES Primary outcomes were scores higher than 5 on the World Health Organization 10-point Clinical Progression Scale (WHO-CPS) on day 4 and survival without need of ventilation (including noninvasive ventilation) at day 14. Secondary outcomes were clinical status assessed with the WHO-CPS scores at day 7 and day 14, overall survival, time to discharge, time to oxygen supply independency, biological factors such as C-reactive protein level, and adverse events. RESULTS Of 131 patients, 64 patients were randomly assigned to the TCZ group and 67 to UC group; 1 patient in the TCZ group withdrew consent and was not included in the analysis. Of the 130 patients, 42 were women (32%), and median (interquartile range) age was 64 (57.1-74.3) years. In the TCZ group, 12 patients had a WHO-CPS score greater than 5 at day 4 vs 19 in the UC group (median posterior absolute risk difference [ARD] −9.0%; 90% credible interval [CrI], −21.0 to 3.1), with a posterior probability of negative ARD of 89.0% not achieving the 95% predefined efficacy threshold. At day 14, 12% (95% CI −28% to 4%) fewer patients needed noninvasive ventilation (NIV) or mechanical ventilation (MV) or died in the TCZ group than in the UC group (24% vs 36%, median posterior hazard ratio [HR] 0.58; 90% CrI, 0.33-1.00), with a posterior probability of HR less than 1 of 95.0%, achieving the predefined efficacy threshold. The HR for MV or death was 0.58 (90% CrI, 0.30 to 1.09). At day 28, 7 patients had died in the TCZ group and 8 in the UC group (adjusted HR, 0.92; 95% CI 0.33-2.53). Serious adverse events occurred in 20 (32%) patients in the TCZ group and 29 (43%) in the UC group (P = .21). CONCLUSIONS AND RELEVANCE In this randomized clinical trial of patients with COVID-19 and pneumonia requiring oxygen support but not admitted to the intensive care...

designed the study, obtained the funding and all the authorizations. Michael Darmon and Frédéric Pochard were part of the steering committee and helped to design the project and to select all the metrics and mental health outcomes.

Background Patients with COVID-19 pneumonia have an excess of inflammation and increased concentrations of cytokines including interleukin-1 (IL-1). We aimed to determine whether anakinra, a recombinant human IL-1 receptor antagonist, could improve outcomes in patients in hospital with mild-to-moderate COVID-19 pneumonia.Methods In this multicentre, open-label, Bayesian randomised clinical trial (CORIMUNO-ANA-1), nested within the CORIMUNO-19 cohort, we recruited patients from 16 University hospitals in France with mild-to-moderate COVID-19 pneumonia, severe acute respiratory syndrome coronavirus 2 infection confirmed by real-time RT-PCR, requiring at least 3 L/min of oxygen by mask or nasal cannula but without ventilation assistance, a score of 5 on the WHO Clinical Progression Scale (WHO-CPS), and a C-reactive protein serum concentration of more than 25 mg/L not requiring admission to the intensive care unit at admission to hospital. Eligible patients were randomly assigned (1:1) using a web-based secure centralised system, stratified by centre and blocked with varying block sizes (randomly of size two or four), to either usual care plus anakinra (200 mg twice a day on days 1-3, 100 mg twice on day 4, 100 mg once on day 5) or usual care alone. Usual care was provided at the discretion of the site clinicians. The two coprimary outcomes were the proportion of patients who had died or needed non-invasive or mechanical ventilation by day 4 (ie, a score of >5 on the WHO-CPS) and survival without need for mechanical or non-invasive ventilation (including high-flow oxygen) at day 14. All analyses were done on an intention-to-treat basis. The trial is registered with ClinicalTrials.gov, NCT04341584, and is now closed to accrual. FindingsBetween April 8 and April 26, 2020, we screened 153 patients. The study was stopped early following the recommendation of the data and safety monitoring board, after the recruitment of 116 patients: 59 were assigned to the anakinra group, and 57 were assigned to the usual care group. Two patients in the usual care group withdrew consent and were not analysed. In the analysable population, the median age was 66 years (IQR 59 to 76) and 80 (70%) participants were men. In the anakinra group, 21 (36%) of 59 patients had a WHO-CPS score of more than 5 at day 4 versus 21 (38%) of 55 in the usual care group (median posterior absolute risk difference [ARD] -2•5%, 90% credible interval [CrI] -17•1 to 12•0), with a posterior probability of ARD of less than 0 (ie, anakinra better than usual care) of 61•2%. At day 14, 28 (47%; 95% CI 33 to 59) patients in the anakinra group and 28 (51%; 95% CI 36 to 62) in the usual care group needed ventilation or died, with a posterior probability of any efficacy of anakinra (hazard ratio [HR] being less than 1) of 54•5% (median posterior HR 0•97; 90% CrI 0•62 to 1•52). At day 90, 16 (27%) patients in the anakinra group and 15 (27%) in the usual care group had died. Serious adverse events occurred in 27 (46%) patients in the anakinra group and 21 (38%) in the usu...

We present and compare multiple imputation methods for multilevel continuous and binary data where variables are systematically and sporadically missing.The methods are compared from a theoretical point of view and through an extensive simulation study motivated by a real dataset comprising multiple studies. Simulations are reproducible. The comparisons show why these multiple imputation methods are the most appropriate to handle missing values in a multilevel setting and why their relative performances can vary according to the missing data pattern, the multilevel structure and the type of missing variables.This study shows that valid inferences can only be obtained if the dataset gathers a large number of clusters. In addition, it highlights that heteroscedastic MI methods provide more accurate inferences than homoscedastic methods, which should be reserved for data with few individuals per cluster. Finally, the method of Quartagno and Carpenter (2016a) appears generally accurate for binary variables, the method of Resche-Rigon and White (2016) with large clusters, and the approach of Jolani et al. (2015) with small clusters.

BackgroundIntensive care unit (ICU) survivors have reduced long-term survival compared to the general population. Identifying parameters at ICU discharge that are associated with poor long-term outcomes may prove useful in targeting an at-risk population. The main objective of the study was to identify clinical and biological determinants of death in the year following ICU discharge.MethodsFROG-ICU was a prospective, observational, multicenter cohort study of ICU survivors followed 1 year after discharge, including 21 medical, surgical or mixed ICUs in France and Belgium. All consecutive patients admitted to intensive care with a requirement for invasive mechanical ventilation and/or vasoactive drug support for more than 24 h following ICU admission and discharged from ICU were included. The main outcome measure was all-cause mortality at 1 year after ICU discharge. Clinical and biological parameters on ICU discharge were measured, including the circulating cardiovascular biomarkers N-terminal pro-B type natriuretic peptide, high-sensitive troponin I, bioactive-adrenomedullin and soluble-ST2. Socioeconomic status was assessed using a validated deprivation index (FDep).ResultsOf 1570 patients discharged alive from the ICU, 333 (21%) died over the following year. Multivariable analysis identified age, comorbidity, red blood cell transfusion, ICU length of stay and abnormalities in common clinical factors at the time of ICU discharge (low systolic blood pressure, temperature, total protein, platelet and white cell count) as independent factors associated with 1-year mortality. Elevated biomarkers of cardiac and vascular failure independently associated with 1-year death when they are added to multivariable model, with an almost 3-fold increase in the risk of death when combined (adjusted odds ratio 2.84 (95% confidence interval 1.73–4.65), p < 0.001).ConclusionsThe FROG-ICU study identified, at the time of ICU discharge, potentially actionable clinical and biological factors associated with poor long-term outcome after ICU discharge. Those factors may guide discharge planning and directed interventions.Trial registrationClinicalTrials.gov NCT01367093. Registered on 6 June 2011.Electronic supplementary materialThe online version of this article (10.1186/s13054-017-1922-8) contains supplementary material, which is available to authorized users.

In multilevel settings such as individual participant data meta-analysis, a variable is 'systematically missing' if it is wholly missing in some clusters and 'sporadically missing' if it is partly missing in some clusters. Previously proposed methods to impute incomplete multilevel data handle either systematically or sporadically missing data, but frequently both patterns are observed. We describe a new multiple imputation by chained equations (MICE) algorithm for multilevel data with arbitrary patterns of systematically and sporadically missing variables. The algorithm is described for multilevel normal data but can easily be extended for other variable types. We first propose two methods for imputing a single incomplete variable: an extension of an existing method and a new two-stage method which conveniently allows for heteroscedastic data. We then discuss the difficulties of imputing missing values in several variables in multilevel data using MICE, and show that even the simplest joint multilevel model implies conditional models which involve cluster means and heteroscedasticity. However, a simulation study finds that the proposed methods can be successfully combined in a multilevel MICE procedure, even when cluster means are not included in the imputation models.

Table of contentsP001 - Sepsis impairs the capillary response within hypoxic capillaries and decreases erythrocyte oxygen-dependent ATP effluxR. M. Bateman, M. D. Sharpe, J. E. Jagger, C. G. EllisP002 - Lower serum immunoglobulin G2 level does not predispose to severe flu.J. Solé-Violán, M. López-Rodríguez, E. Herrera-Ramos, J. Ruíz-Hernández, L. Borderías, J. Horcajada, N. González-Quevedo, O. Rajas, M. Briones, F. Rodríguez de Castro, C. Rodríguez GallegoP003 - Brain protective effects of intravenous immunoglobulin through inhibition of complement activation and apoptosis in a rat model of sepsisF. Esen, G. Orhun, P. Ergin Ozcan, E. Senturk, C. Ugur Yilmaz, N. Orhan, N. Arican, M. Kaya, M. Kucukerden, M. Giris, U. Akcan, S. Bilgic Gazioglu, E. TuzunP004 - Adenosine a1 receptor dysfunction is associated with leukopenia: A possible mechanism for sepsis-induced leukopeniaR. Riff, O. Naamani, A. DouvdevaniP005 - Analysis of neutrophil by hyper spectral imaging - A preliminary reportR. Takegawa, H. Yoshida, T. Hirose, N. Yamamoto, H. Hagiya, M. Ojima, Y. Akeda, O. Tasaki, K. Tomono, T. ShimazuP006 - Chemiluminescent intensity assessed by eaa predicts the incidence of postoperative infectious complications following gastrointestinal surgeryS. Ono, T. Kubo, S. Suda, T. Ueno, T. IkedaP007 - Serial change of c1 inhibitor in patients with sepsis – A prospective observational studyT. Hirose, H. Ogura, H. Takahashi, M. Ojima, J. Kang, Y. Nakamura, T. Kojima, T. ShimazuP008 - Comparison of bacteremia and sepsis on sepsis related biomarkersT. Ikeda, S. Suda, Y. Izutani, T. Ueno, S. OnoP009 - The changes of procalcitonin levels in critical patients with abdominal septic shock during blood purificationT. Taniguchi, M. OP010 - Validation of a new sensitive point of care device for rapid measurement of procalcitoninC. Dinter, J. Lotz, B. Eilers, C. Wissmann, R. LottP011 - Infection biomarkers in primary care patients with acute respiratory tract infections – Comparison of procalcitonin and C-reactive proteinM. M. Meili, P. S. SchuetzP012 - Do we need a lower procalcitonin cut off?H. Hawa, M. Sharshir, M. Aburageila, N. SalahuddinP013 - The predictive role of C-reactive protein and procalcitonin biomarkers in central nervous system infections with extensively drug resistant bacteriaV. Chantziara, S. Georgiou, A. Tsimogianni, P. Alexandropoulos, A. Vassi, F. Lagiou, M. Valta, G. Micha, E. Chinou, G. MichaloudisP014 - Changes in endotoxin activity assay and procalcitonin levels after direct hemoperfusion with polymyxin-b immobilized fiberA. Kodaira, T. Ikeda, S. Ono, T. Ueno, S. Suda, Y. Izutani, H. ImaizumiP015 - Diagnostic usefullness of combination biomarkers on ICU admissionM. V. De la Torre-Prados, A. Garcia-De la Torre, A. Enguix-Armada, A. Puerto-Morlan, V. Perez-Valero, A. Garcia-AlcantaraP016 - Platelet function analysis utilising the PFA-100 does not predict infection, bacteraemia, sepsis or outcome in critically ill patientsN. Bolton, J. Dudziak, S. Bonney, A. Tridente, P. NeeP017 - Extracellular histone H3 levels are in...

BACKGROUND: Working in the ICU during the first COVID-19 wave was associated with high levels of mental health disorders.RESEARCH QUESTION: What are the mental health symptoms in health-care providers (HCPs) facing the second wave? STUDY DESIGN AND METHODS: A cross-sectional study (October 30-December 1, 2020) was conducted in 16 ICUs during the second wave in France. HCPs completed the Hospital Anxiety and Depression Scale, the Impact of Event Scale-Revised (for post-traumatic stress disorder), and the Maslach Burnout Inventory.RESULTS: Of 1,203 HCPs, 845 responded (70%) (66% nursing staff, 32% medical staff, 2% other professionals); 487 (57.6%) had treated more than 10 new patients with COVID-19 in the previous week. Insomnia affected 320 (37.9%), and 7.7% were taking a psychotropic drug daily. Symptoms of anxiety, depression, post-traumatic stress disorder, and burnout were reported in 60.0% (95% CI, 56.6%-63.3%), 36.1% (95% CI, 32.9%-39.5%), 28.4% (95% CI, 25.4%-31.6%), and 45.1% (95% CI, 41.7%-48.5%) of respondents, respectively. Independent predictors of such symptoms included respondent characteristics (sex, profession, experience, personality traits), work organization (ability to rest and to care for family), and self-perceptions (fear of becoming infected or of infecting family and friends, feeling pressure related to the surge, intention to leave the ICU, lassitude, working conditions, feeling they had a high-risk profession, and "missing the clapping"). The number of patients with COVID-19 treated in the first wave or over the last week was not associated with symptoms of mental health disorders. INTERPRETATION:The prevalence of symptoms of mental health disorders is high in ICU HCPs managing the second COVID-19 surge. The highest tiers of hospital management urgently need to provide psychological support, peer-support groups, and a communication structure that ensure the well-being of HCPs.