key: cord-0894829-2xmxq8ty
authors: Chocron, Richard; Galand, Vincent; Cellier, Joffrey; Gendron, Nicolas; Pommier, Thibaut; Bory, Olivier; Khider, Lina; Trimaille, Antonin; Goudot, Guillaume; Weizman, Orianne; Alsac, Jean Marc; Geneste, Laura; Schmeltz, Armand; Panagides, Vassili; Philippe, Aurélien; Marsou, Wassima; Ben Abdallah, Iannis; Deney, Antoine; El Batti, Salma; Attou, Sabir; Juvin, Philippe; Delmotte, Thomas; Messas, Emmanuel; Pezel, Théo; Planquette, Benjamin; Duceau, Baptiste; Gaussem, Pascale; Sutter, Willy; Sanchez, Olivier; Waldman, Victor; Diehl, Jean‐Luc; Mirault, Tristan; Bonnet, Guillaume; Cohen, Ariel; Smadja, David M.
title: Anticoagulation Before Hospitalization Is a Potential Protective Factor for COVID‐19: Insight From a French Multicenter Cohort Study
date: 2021-04-16
journal: J Am Heart Assoc
DOI: 10.1161/jaha.120.018624
sha: 361c237865e58dcd39944b45d8a2c064f170b1d3
doc_id: 894829
cord_uid: 2xmxq8ty

BACKGROUND: Coronavirus disease 2019 (COVID‐19) is a respiratory disease associated with thrombotic outcomes with coagulation and endothelial disorders. Based on that, several anticoagulation guidelines have been proposed. We aimed to determine whether anticoagulation therapy modifies the risk of developing severe COVID‐19. METHODS AND RESULTS: Patients with COVID‐19 initially admitted in medical wards of 24 French hospitals were included prospectively from February 26 to April 20, 2020. We used a Poisson regression model, Cox proportional hazard model, and matched propensity score to assess the effect of anticoagulation on outcomes (intensive care unit admission or in‐hospital mortality). The study enrolled 2878 patients with COVID‐19, among whom 382 (13.2%) were treated with oral anticoagulation therapy before hospitalization. After adjustment, anticoagulation therapy before hospitalization was associated with a better prognosis with an adjusted hazard ratio of 0.70 (95% CI, 0.55–0.88). Analyses performed using propensity score matching confirmed that anticoagulation therapy before hospitalization was associated with a better prognosis, with an adjusted hazard ratio of 0.43 (95% CI, 0.29–0.63) for intensive care unit admission and adjusted hazard ratio of 0.76 (95% CI, 0.61–0.98) for composite criteria intensive care unit admission or death. In contrast, therapeutic or prophylactic low‐ or high‐dose anticoagulation started during hospitalization were not associated with any of the outcomes. CONCLUSIONS: Anticoagulation therapy used before hospitalization in medical wards was associated with a better prognosis in contrast with anticoagulation initiated during hospitalization. Anticoagulation therapy introduced in early disease could better prevent COVID‐19–associated coagulopathy and endotheliopathy, and lead to a better prognosis.

S ince December 2019, the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic spread rapidly throughout China 1,2 with a higher transmission capacity and a higher mortality compared with SARS-CoV-1 epidemic of 2003. 3 Coronavirus disease 2019 (COVID- 19) clinical manifestations are first dominated by respiratory symptoms. Patients may experience severe cardiovascular damages 4,5 associated with a coagulopathy characterized by an increase in procoagulant factors such as fibrinogen, and high levels of d-dimer that have been associated with a worsening of the disease and higher mortality. [6] [7] [8] Increased incidence of venous thromboembolism, in particular pulmonary embolism, has been observed in several reports, [9] [10] [11] [12] and microvascular thrombosis in the lungs has been observed in autopsy series and during COVID-19 acute respiratory distress syndrome. [13] [14] [15] [16] Moreover, SARS-CoV-2 has been shown to induce vascular damage, and we recently described that circulating markers of endothelial lesion are a predictive factor for intensive care unit (ICU) referral and mortality, reinforcing the hypothesis of a COVID-19-associated microvascular dysfunction. 17, 18 The COVID-19-associated coagulopathy has led many medical societies to propose different anticoagulation strategies. 19, 20 Results showing that prophylactic anticoagulation with lowmolecular-weight heparin (LMWH) improved critically ill patients' prognosis argued for the use of prophylactic anticoagulation in patients with COVID-19. [20] [21] [22] [23] No protective effect of anticoagulation before hospitalization for COVID-19 on survival has been observed until now, in 2 small retrospective cohorts. 24, 25 Using data from a multicenter French cohort (n=2848), we aimed to determine whether anticoagulation therapy before hospitalization was an independent predictor for developing a severe form of the disease. Moreover, since several anticoagulation strategies based on recent national and international guidelines 19, 26 were used and still remain a matter of debate, we also aimed to analyze different prophylactic or therapeutic anticoagulation strategies introduced during hospitalization on patients' clinical outcomes.

Detailed data are available on request from the authors.

From February 26 to April 20, 2020, consecutive patients with a diagnosis of SARS-CoV-2 infection and hospitalized in medical wards were included (none of the patients were directly admitted to the ICU). Patients were aged over 18 years old and were included in a retrospective multicentric (24 centers) observational study, which was named the CCF (Critical COVID-19 France) study and initiated by the French Society of Cardiology. Following World Health Organization criteria, SARS-CoV-2 infection was determined by positive results from real-time reverse transcriptase-polymerase chain reaction of nasal or pharyngeal swabs or lower respiratory tract aspirates (confirmed case) or by typical imaging characteristics on chest computed tomography (CT) when laboratory testing was inconclusive (probable case). 27 

The CCF study was declared and authorized at the French data protection committee (authorization no. 2207326v0) and conducted in accordance with the ethical standards established in the 1964 Declaration of Helsinki and its later amendments (NCT04344327). Because an anonymized hospital database was used, informed consent from individual patients was not obtained. 

The time from diagnosis to death or ICU admission for ventilation support (including invasive, noninvasive, or high-flow oxygen) were used as a composite primary outcome to define ICU-free survival. The secondary outcome was the time from diagnosis to death only for the selected population of patients secondarily referred to the ICU after the medical ward. Outcomes were assessed using the electronic medical record.

Continuous data were expressed as mean (SD) and categorical data as proportion. Patients were compared according to the use of oral anticoagulation therapy (VKA or DOAC) before hospitalization using the Mann-Whitney test for continuous variables and Fisher exact test for categorical variables. 28 In the multivariable analysis, we used Poisson regression to assess the association between the use of oral anticoagulation therapy before hospitalization and outcomes. 29, 30 The model included as covariates sex, age, cardiovascular comorbidities (history of high blood pressure, dyslipidemia, body mass index, type 2 diabetes mellitus, and current smoking), plasma creatinine level (µmoL/L), C-reactive protein (mg/L), fraction of inspired oxygen, and the degree of pulmonary lesions with ground-glass opacities, and areas of consolidation. For the survival analysis, the start of the study was triggered by the diagnosis of SARS-CoV-2 infection and hospitalization in a medical ward. The end of the study was defined either by the death of the patient during the hospitalization or by discharge alive from the hospital. Vital status was ascertained through April 20, 2020, with the potential period of follow-up ranging from 1 to 54 days. If the patient was still hospitalized on April 20, 2020, the vital status was coded as alive. Survival time was calculated as the difference between the date of the diagnosis of SARS-CoV-2 infection and the date of event occurrence (ICU admission and in-hospital death) or the date of hospital discharge or the length of hospital stay for patients still hospitalized at the end of the study (April 20, 2020). We used a Cox proportional hazard (PH) model adjusted for the same potential confounders included in the Poisson regression model to investigate the relationships between the use of oral anticoagulation therapy and outcomes. Kaplan-Meier method was used to represent the Cox PH model results according to the use or not of oral anticoagulation therapy before hospitalization for COVID-19. We performed 3 sensitivity analyses: (1) To take into account the retrospective design and to avoid the bias caused by censored data (n=513), we performed a complete data analysis for mortality risk: the same Poisson regression analysis in the population of patients who were discharged alive from hospital or dead in hospital (total patients analyzed n=2335) and thus excluded patient with censored outcome; (2) to explore the individual effect of in-hospital anticoagulation therapy (unfractionated heparin or LMWH) on the one hand, and of anticoagulation treatment before hospitalization (VKA or DOAC) on the other hand, and the magnitude of each effect on outcome, we decided to rerun the Cox PH models in the selected population without anticoagulation before hospitalization (n=2466) for the subgroup of patients who were hospitalized in the ICU and for those who stayed in the medical ward using the outcome (in-hospital mortality); (3) to adjust for bias caused by nonrandom allocation of potential covariates, we performed a propensitymatched analysis of patients who were on oral anticoagulation therapy (VKA or DOAC) before hospitalization for COVID-19 infection compared with those who were not and rerun the Cox PH model adjusted only for plasma creatinine level, C-reactive protein, fraction of inspired oxygen, and the degree of pulmonary lesions with groundglass opacities and areas of consolidation. Missing data were handled using multiple random forest imputation using chained equation (10 sets of imputations). All analyses were 2-sided and a P<0.05 was considered statistically significant. Statistical analysis was performed using R studio software including R version 3.6.3 (R Foundation for Statistical Computing, Vienna, Austria).

During the study period, 2878 consecutive patients who were hospitalized for SARS-CoV-2 infection in the medical ward were included. The demographic and clinical characteristics of the patients with COVID-19 according to use of anticoagulation before hospitalization are provided in Table 1 . For analysis of anticoagulation before hospitalization assessment, we excluded 30 patients who were treated by heparin or LMWH because the regimen (therapeutic or prophylactic doses) was not known. Among the study population, 382 of 2848 (13.4%) patients were treated with oral anticoagulation therapy before hospitalization, physicians decided to pursue anticoagulation for 341 (90.0%) patients and to use prophylactic low dose for 58 (15.2%), prophylactic high dose for 9 (2.4%), and therapeutic dose for 274 (71.7%) during hospitalization. Among the remaining 2466 patients, 1478 (59.9%) received prophylactic low dose, 261 (10.6%) received prophylactic high dose, and 246 (10.0%) were treated with therapeutic anticoagulation during hospitalization. Patients receiving oral anticoagulation before hospitalization were older (P<0.001) and had significantly more cardiovascular risk factors, that is, more high blood pressure, diabetes mellitus, and dyslipidemia (Table 1) . Overall, 352 of 2848 (12.4%) patients including 84 (2.9%) under oral anticoagulation therapy died in the medical ward or the ICU. The mean time of follow-up was 8.7 days (6.5) ( Table 1) .

In the whole population and after adjustment, male sex, body mass index, C-reactive protein, fraction of inspired oxygen, and parenchymal opacification in chest CT scan were significantly associated with ICU admission and in-hospital mortality ( Table 2) . Therapeutic oral anticoagulation before hospitalization was associated with a better outcome, with a relative risk (RR) of 0.45 (95% CI, 0.32-0.62; P<0.001) for ICU admission and relative risk of 0.72 (95% CI, 0.57-0.90; P=0.006) for ICU admission or in-hospital mortality. The use of prophylactic high-dose or therapeutic anticoagulation during hospitalization was significantly associated with higher ICU admission and in-hospital mortality (relative risk, 2.00; 95% CI, 1.49-2.70; P<0.001; and relative risk, 1.62; 95% CI, 1.21-2.18; P=0.001, respectively) ( Table 2 ). To take into account the retrospective design and to avoid the bias caused by censored data (n=513), the same multivariable analysis was performed in the selected population of patients without censored outcome (patients who were discharged alive from the hospital or dead in the hospital; total patients analyzed, n=2335) and found the same protective effect of therapeutic anticoagulation before hospitalization for COVID-19 and worse outcome for receiving prophylactic high-dose or therapeutic-dose anticoagulation during hospitalization (Table S1 ). Effect of oral anticoagulation before hospitalization on outcomes was confirmed by using the Kaplan-Meier curve and multivariable adjusted Cox PH model (Figures 1 and 2 ). In the multivariable adjusted Cox regression model, the adjusted hazard ratio for the effect of oral anticoagulation before hospitalization was 0.70 (95% CI, 0.55-0.88; P=0.003; Figure 2 ). In this multivariable adjusted Cox regression model, we also confirmed that male sex and chest CT scan with ≥50% opacification were associated with a high risk of ICU transfer or in-hospital mortality (Figure 2 ). In sensitivity analysis, the propensity matching yielded 382 patients who received oral anticoagulation and 1528 patients who did not before hospitalization, with balanced variables between the groups (Table S2 ). In this propensity-matched population, there was still a statistically significant difference in outcome between the 2 groups: adjusted hazard ratios for the effect of oral anticoagulation before hospitalization were 0.43 (95% CI, 0.29-0.63; P=<0.001), 0.76 (95% CI, 0.61-0.98; P=0.04) and 1.09 (95% CI, 0.83-1.55; P=0.49), respectively, for ICU admission, ICU admission or in-hospital mortality (combined outcome), and for in-hospital mortality ( Table 3) .

Regarding anticoagulation therapy introduced during hospitalization, we reran multivariable analysis in the subgroup of patients without oral anticoagulation before hospitalization (n=2466). For the entire population, we did not observe a significant association between the use of anticoagulation during hospitalization for any regimen (prophylactic low or high dose and therapeutic dose) and in-hospital mortality ( Table 4) . The results were similar for patients admitted to the ICU or for patients who stayed in the medical ward (Table 4) .

In this retrospective study, we demonstrated that an early anticoagulation before COVID-19 hospitalization improves the outcome of patients with COVID-19. Using a multicenter French study of patients hospitalized for COVID-19, we provide evidence that previous oral anticoagulation with VKA or DOAC significantly decreased ICU admission or in-hospital mortality. Furthermore, in patients without anticoagulation before hospitalization, anticoagulation started during hospitalization (heparin or LMWH) was not associated with a better prognosis. Importantly, this is the first study evaluating anticoagulation in patients with COVID-19 that provides a clear description of baseline patient characteristics. 21 Based on the rationale that SARS-CoV-2 infection is associated with endothelial dysfunction, 17,18 COVID-19-induced coagulopathy might be a consequence of endothelial injury. 2,32 We indeed previously described that patients with COVID-19 treated with therapeutic anticoagulation had a lower level of circulating endothelial cells, a marker of endothelial lesion. 33 This protective effect of anticoagulation therapy on endothelial dysfunction could explain the protective effect of anticoagulation on microvascular thrombosis and coagulopathy observed in patients with COVID-19. Indeed, endotheliitis has been described during COVID-19 and could be at the origin of impaired microcirculatory function affecting particularly the lungs and kidneys. 34 From patients' autopsies, this endotheliitis has been described associated with an angiogenic process in the lungs. 35 Moreover, the central involvement of endothelial compartment in COVID-19 outcome and pathophysiology is supported by the higher level of circulating endothelial cells in patients who are COVID-19 positive versus negative, associated with the increased plasma levels of angiopoietin-2 and E-selectin correlated to ICU transfer. 17, 33 In the present study, we observed that anticoagulation administered before hospitalization for COVID-19 had a significant positive impact on ICU admission or in-hospital mortality by contrast with patients without anticoagulation. Our results are not in line with those of Tremblay et al, 24 who recently reported that they used a propensity score to compare patients who were anticoagulated versus nonanticoagulated before hospitalization. Of note, the logistic regression model they used to calculate the propensity score was not adjusted on relevant cardiovascular comorbidities such as hypertension, diabetes mellitus, smoking, or renal function and could explain the divergent results in the literature. This makes the association between anticoagulation and outcomes difficult to analyze since endothelial dysfunction during COVID-19 mostly results from these comorbidities. 36 Moreover, in a cohort of 449 patients with COVID-19 *Patients treated with LMWH or unfractionated heparin at admission were excluded because the therapeutic or prophylactic regimen was unknown. in Wuhan, China, a prophylactic dose of LMWH used in 99 patients decreased mortality only in a specific subgroup of patients with a sepsis-induced coagulopathy. 21 In our study, we have tried to discern the specific and individual effect of the anticoagulation before hospitalization and anticoagulation initiated during hospitalization. In our population, no beneficial effect of prophylactic anticoagulation was reported, as recently described by Lynn et al. 37 Even for prophylactic high doses or therapeutic anticoagulation introduced In-hospital mortality or ICU admission as a composite event to define ICU-free survival (B) and to in-hospital mortality (C). CT indicates computed tomography; DOAC, direct oral anticoagulant; FiO 2 , fraction of inspired oxygen; ICU, intensive care unit; RR, relative risk; and VKA, vitamin K antagonist. A, Admission to in-hospital mortality, or ICU admission as a composite event to define ICU-free survival (B) and to admission to in hospital mortality (C). Cox proportional hazards model is adjusted for sex, age, cardiovascular comorbidities (history of high blood pressure. dyslipidemia. body mass index, type 2 diabetes mellitus, and current smoking), plasma creatinine level (µmol/L), Creactive protein (mg/L), and fraction of inspired oxygen. The degree of pulmonary lesions with ground-glass opacities and areas of consolidation and the use of in hospital anticoagulation (preventive low or high dose and therapeutic dose anticoagulation). DOAC indicates direct oral anticoagulant; ICU, intensive care unit; and VKA, vitamin K antagonist. during hospitalization, an increase of patients with unfavorable outcome was observed. This finding is supported by the recent description of Paranjpe et al, 31 who found that patients who received anticoagulation were more likely to require invasive mechanical ventilation. We hypothesize that prehospital anticoagulation (VKA or DOAC) is more likely efficient at the early step of disease by preventing COVID-19-associated coagulopathy and endotheliopathy. Once the endotheliitis condition is reached, the COVID-19 worsens and the anticoagulation therapy loses its protective effect.

The clinical course of COVID-19 consists of 2 main phases: viral infection followed by immune and inflammatory response. Both phases are associated with coagulopathy and potential need of anticoagulation but probably with several different objectives. In both phases, anticoagulation could decrease microthrombosis in small lung vessels and delay aggravation in the early phase or decrease the burst of thromboinflammation and associated microthrombosis. Early anticoagulation before hospitalization for COVID-19 could also help, besides the coagulopathy associated with endothelial lesion, directly on the inhibition of the entry of SARS-CoV-2 into endothelial cells. Indeed, cell entry of SARS-CoV-2 depends on the binding of the viral spike proteins to cellular receptors and on spike-protein priming by host cell proteases. It has been demonstrated that serine protease TMPRSS2 is necessary for spike-protein priming. 38 Thus, a TMPRSS2 inhibitor has been proposed as a treatment option. If virus entrance inside cells is dependent on a serine protease action, anticoagulation, by blocking the coagulation cascade, could directly limit the virus entrance in endothelial but also in other cells. This hypothesis needs to be tested in preclinical models of infection.

Our study has several limitations. Because of its retrospective design, it precludes determination of any causal relationship between the use of oral anticoagulants and outcomes. Moreover, this study has included only patients who were initially hospitalized in the medical ward and did not include patients who were directly admitted to the ICU. This selected population would not be representative of all patients with COVID-19. Despite efforts to control confounders by using different analytical strategies, some potential biases may have been disregarded such as the delay to introduction of anticoagulation, the duration of anticoagulation therapy during hospitalization and its indication. All Admission to in-hospital mortality or ICU admission as a composite event to define ICU-free survival (B) and to admission to in hospital mortality (C) in propensity matched population (n=1528). Cox proportional hazards model is adjusted for sex, age, cardiovascular comorbidities (history of high blood pressure, dyslipidemia, body mass index, type 2 diabetes mellitus, and current smoking), plasma creatinine level (µmol/L), C-reactive protein (mg/L), fraction of inspired oxygen, the degree of pulmonary lesions with ground-glass opacities and areas of consolidation, and the use of in-hospital anticoagulation (preventive low or high dose and therapeutic dose). efforts were made to adjust the analyses for relevant variables, including cardiovascular comorbidities, patient characteristics, severity of clinical features, and CT scan results. By design, the study can only report associations and cannot investigate causality.

In conclusion, our study is the first to demonstrate a beneficial effect of oral anticoagulation before hospitalization for COVID-19 on disease outcomes. Our results provide new data to suggest introduction of anticoagulation in ambulatory patients in case of close contacts with patients with confirmed COVID-19. Anticoagulation introduction in early disease, in particular in ambulatory patients, could prevent coagulopathy and endothelialrelated disease, while an anticoagulation introduction too late, during the thrombo-inflammatory phase, do not provide protective effects. This hypothesis needs to be tested in appropriate prospective randomized studies. Cox proportional hazard model is adjusted on sex, age, cardiovascular comorbidities (history of high blood pressure, dyslipidemia, body mass index, type 2 diabetes mellitus, and current smoking). Plasma creatinine level (µmol/L). C-reactive protein (mg/L). FiO 2 . The degree of pulmonary lesions with ground-glass opacities and areas of consolidation. DOAC indicates direct oral anticoagulant; HR, hazard ratio; ICU, intensive care unit; ref, reference; and VKA, vitamin K antagonist.

Georges Pompidou European Hospital

)and Intensive Care Medicine Department and Biosurgical Research Lab (Carpentier Foundation), AP-HP

Is COVID-19 a new hematologic disease?

Aerosol and surface stability of SARS-CoV-2 as compared with SARS-CoV-1

Characteristics of and important lessons from the coronavirus disease 2019 (COVID-19) outbreak in China: summary of a report of 72314 cases from the Chinese Center for Disease Control and Prevention

COVID-19 and the cardiovascular system

Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study

Abnormal coagulation parameters are associated with poor prognosis in patients with novel coronavirus pneumonia

Predictive factor for COVID-19 worsening: insights for high-sensitivity troponin and D-dimer and correlation with right ventricular afterload

Pulmonary embolism in patients with COVID-19: awareness of an increased prevalence

Confirmation of the high cumulative incidence of thrombotic complications in critically ill ICU patients with COVID-19: an updated analysis

Incidence of thrombotic complications in critically ill ICU patients with COVID-19

Prevalence and characteristics of pulmonary embolism in 1042 COVID-19 patients with respiratory symptoms: a nested case-control study

Pulmonary and cardiac pathology in COVID-19: the first autopsy series from New Orleans. medRxiv preprint

Histopathologic changes and SARS-CoV-2 immunostaining in the lung of a patient with COVID-19

Ventilatory ratio in hypercapnic mechanically ventilated patients with COVID-19-associated acute respiratory distress syndrome

Respiratory mechanics and gas exchanges in the early course of COVID-19 ARDS: a hypothesis-generating study

Angiopoietin-2 as a marker of endothelial activation is a good predictor factor for intensive care unit admission of COVID-19 patients

Circulating von Willebrand factor and high molecular weight multimers as markers of endothelial injury predict COVID-19 in-hospital mortality

Prevention of thrombotic risk in hospitalized patients with COVID-19 and hemostasis monitoring

ISTH interim guidance on recognition and management of coagulopathy in COVID-19

Anticoagulant treatment is associated with decreased mortality in severe coronavirus disease 2019 patients with coagulopathy

The association between treatment with heparin and survival in patients with Covid-19

Hospital-based use of thromboprophylaxis in patients with COVID-19

Impact of anticoagulation prior to COVID-19 infection: a propensity score-matched cohort study

Clinical impact of pre-admission antithrombotic therapy in hospitalized patients with COVID-19: a multicenter observational study

Scientific and standardization committee communication: clinical guidance on the diagnosis, prevention, and treatment of venous thromboembolism in hospitalized patients with COVID-19

European Society of Radiology, the European Society of Thoracic Imaging. COVID-19 patients and the radiology department-advice from the European Society of Radiology (ESR) and the European Society of Thoracic Imaging (ESTI)

Statistics review 10: further nonparametric methods

Measures and models for causal inference in cross-sectional studies: arguments for the appropriateness of the prevalence odds ratio and related logistic regression

Bias in observational study designs: cross sectional studies

Association of treatment dose anticoagulation with in-hospital survival among hospitalized patients with COVID-19

Multidimensional proteomic approach of endothelial progenitors demonstrate expression of KDR restricted to CD19 cells

Curative anticoagulation prevents endothelial lesion in COVID-19 patients

Endothelial cell infection and endotheliitis in COVID-19

Pulmonary vascular endothelialitis, thrombosis, and angiogenesis in COVID-19

Hypertension, thrombosis, kidney failure, and diabetes: is COVID-19 an endothelial disease? A comprehensive evaluation of clinical and basic evidence

The effect of anticoagulation on clinical outcomes in novel coronavirus (COVID-19) pneumonia in a U

SARS-CoV-2 cell entry depends on ACE2 and TMPRSS2 and is blocked by a clinically proven protease inhibitor

Nouvel Hopital Civil

Poisson regression model according to: ICU admission (A); ICU admission or In-hospital mortality as a composite event to define ICU-free survival (B); and to in-hospital mortality (C) in uncensored population

Author contributions: All authors substantially contributed to the paper. Drs Smadja and Chocron designed the present study and wrote the manuscript. Dr Chocron performed statistical analyses. Drs Cohen and Bonnet designed the trial. All authors reviewed the paper. All authors declare that the submitted work is original and has not been published before (neither in English nor in any other language) and that the work is not under consideration for publication elsewhere.

None.

Drs Chocron, Gendron, Sanchez, Cohen, and Smadja acknowledge the following without any relation to the current manuscript. Dr Chocron received consultant fees from Aspen. Dr Cohen received research grant from RESICARD (research nurses) and consultant and lecture fees from Amgen, AstraZeneca, Bayer, Alliance Bristol-Myers Squibb/Pfizer, Novartis, and Sanofi-Aventis. Dr Smadja received consultant and lecture fees or travel awards from Aspen, Bayer, Carmat, Alliance Bristol-Myers Squibb/Pfizer, LEO-Pharma, and Boehringer-Ingelheim. Dr Sanchez received grants, personal fees, or nonfinancial support from Bayer, Alliance Bristol-Myers Squibb/Pfizer, Sanofi Aventis, Daiichi Sankyo, MSD, Boston Scientifics, and Chiesi. Dr Gendron discloses consulting fees by Boehringer-Ingelheim, Bayer, Alliance Bristol-Myers Squibb/Pfizer, and LEO-Pharma. The remaining authors have no disclosures to report.

Appendix S1 Tables S1-S2 CT= Computerized Tomography, ICU = Intensive Care Unit, FiO2=Fraction of inspired oxygen (%), U/ L= Units per Liter.**Covariates used in the propensity score matching.