Remdesivir and survival outcomes in critically ill patients with COVID-19: A multicentre observational cohort study

The role of remdesivir (RDV) in the treatment of critically ill COVID-19 patients remains ill-defined.1Kelly Ansems Felicitas Grundeis Karolina Dahms Agata Mikolajewska Volker Thieme Vanessa Piechotta et al.Remdesivir for the treatment of COVID-19.Cochrane Database Syst Rev. 2021; 8CD014962https://doi.org/10.1002/14651858.CD014962Crossref Scopus (56) Google Scholar The impact of the systemic inflammation and other aspects adding to the potential severity of patients in the SARS-CoV-2 viral dynamics are not well elucidated, particularly in critically ill patients.2Kemp Steven A. Collier Dami A. Datir Rawlings P. Ferreira Isabella A.T.M. Salma Gayed Aminu Jahun et al.SARS-CoV-2 evolution during treatment of chronic infection.Nature. 2021; 592: 277-282https://doi.org/10.1038/s41586-021-03291-yCrossref PubMed Scopus (489) Google Scholar, 3Sefik E Qu R Junqueira C Kaffe E Mirza H Zhao J et al.Inflammasome activation in infected macrophages drives COVID-19 pathology.Nature. 2022; 606: 585-593https://doi.org/10.1038/s41586-022-04802-1Crossref PubMed Scopus (97) Google Scholar, 4Leisman Daniel E. Lukas Ronner Rachel Pinotti Taylor Matthew D. Pratik Sinha Calfee Carolyn S. et al.Cytokine elevation in severe and critical COVID-19: a rapid systematic review, meta-analysis, and comparison with other inflammatory syndromes.Lancet Respir Med. 2020; 8: 1233-1244https://doi.org/10.1016/S2213-2600(20)30404-5Abstract Full Text Full Text PDF PubMed Scopus (498) Google Scholar Thus, we aimed to evaluate the effect of RDV on the outcomes of critically ill patients with severe COVID-19 and analyze such outcomes in certain subsets of patients to investigate whether RDV proved particularly beneficial in a particular group of critically ill patients. Multicentre, observational cohort study including consecutive COVID-19 patients admitted to 55 Spanish ICUs between 5 February 2020 and 21 December 2021. We compared two main groups: patients receiving RDV versus those who did not receive RDV at any moment. The primary outcome was all-cause 90-day mortality. We examined outcomes according to the following categories: (1) overall population; (2) several patient subgroups based on baseline data at ICU admission: age, illness severity and organ damage (Acute Physiology and Chronic Health Evaluation [APACHE] -II and Sequential Organ Failure Assessment [SOFA] score, respectively), laboratory findings (lymphocyte count and C-reactive protein [C-RP]), respiratory support; (3) by corticosteroids therapy; (4) by SARS-CoV-2 viral load (viral RNAemia, viral antigenemia and viral RNA load [N1 region]); and (5) by RDV administration timing. To evaluate the effect of remdesivir on 90-day mortality, we used a Cox regression model stratified on the center variable, tested in univariable and propensity-adjusted multivariable analyses. A further analysis using the propensity score matching method was performed to confirm the results obtained with the propensity-adjusted multivariable model. On the other hand, effect modification by factors potentially associate with patient outcomes and remdesivir use were assessed by an interaction term. We also analysed the association between remdesivir therapy and the following variables: in-hospital and 30-day mortality (by means of a Fine-Gray competing risks model stratified on the center variable and a Cox regression model stratified on the center variable, respectively); and nosocomial bacterial pneumonia (by means of generalised estimating equations, considering a binomial distribution and accounting for the effect raised by the clustering of patients from the same center). Among the 6225 COVID-19 patients that were admitted to 55 ICUs. We included 5004 patients in this analysis, of whom 4209 (84%) did not receive RDV and 795 (16%) received RDV during hospitalization (Fig. 1). Ninety-day mortality rate was lower in patients who received RDV than in those who did not receive RDV (34 % vs. 29%, p = 0.012). Also, in-hospital (31 % vs. 27%, p = 0.025) and 30-day mortality (25 % vs. 18%, p<0.001) were lower in the RDV group (Supplementary Table 1). In the propensity-adjusted multivariable analysis, RDV use was not significantly associated with 90-day mortality (HR 1.00, 95% CI 0.85 to 1.17; p = 0.970), nor with in-hospital (sHR 0.94, 95% CI 0.81 to 1.09, p = 0.41) and 30-day mortality (HR 0.86, 95% CI 0.71 to 1.05, p = 0.145) (Table 1). In the analysis using the propensity score matching (777 patients received RDV and 777 did not receive RDV), the Kaplan-Meier curves showed that there was no statistical difference between both groups for 90-day mortality (p = 0.280) (Supplementary Figure 1). Furthermore, Cox regression showed that RDV use was not associated with the risk of 90-day mortality (HR 0.95, 95% CI 0.79 to 1.16; p = 0.636), nor with in-hospital mortality (sHR 0.92, 95% CI 0.76 to 1.11; p = 0.36) and 30-day mortality (HR 0.81, 95% CI 0.65 to 1.02; p = 0.075).Table 1Association of remdesivir therapy and 90-day mortality (Panel A) and early administration of remdesivir treatment from first symptoms and 90-day mortality (Panel B).Univariable analysisAdjusted analysisaAdjusted for variables (age, sex, body mass index, diabetes mellitus, chronic liver disease, chronic heart disease, chronic lung disease, chronic renal failure, immunosuppression, APACHE-II score at ICU admission, lymphocyte count at ICU admission, C-reactive protein at ICU admission, respiratory support at ICU admission, septic shock at ICU admission, lopinavir/ritonavir administration, tocilizumab administration, other antiviral administration, corticosteroids, COVID-19 wave and the propensity score).HR (95% CI)P-valueAdjusted HR (95% CI)P-valuePanel A - Remdesivir therapyAll patients (N = 5004)0.78 (0.67 to 0.91)0.0011.00 (0.85 to 1.17)0.970Subgroup analysesbAPACHE-II score was assessed in 2887 patients; SOFA score in 3470 patients; lymphocyte count in 4834 patients; C-reactive protein in 4675 patients; respiratory support in 4988 patients; corticosteroids in 4956 patients; viral RNAemia in plasma in 751 patients; viral antigenemia in plasma in 735 patients; and viral RNA load in plasma (N1 region) in 751 patients.Age group0.284cInteraction effect for the subgroup and treatment group.Age <40 years (n = 281)0.13 (0.02 to 1.05)0.055--dEstimation failed due to numerical problem. Because the coefficients did not converge, no further models were fitted.Age ≥40 - <65 years (n = 2419)0.75 (0.58 to 0.98)0.0331.02 (0.77 to 1.36)0.885Age ≥65 years (n = 2304)0.89 (0.74 to 1.08)0.2380.96 (0.78 to 1.17)0.681Severity of illness at ICU admission group0.368cInteraction effect for the subgroup and treatment group.APACHE-II score <12 (n = 1411)0.57 (0.37 to 0.89)0.0120.97 (0.60 to 1.57)0.893APACHE-II score ≥12 (n = 1476)0.86 (0.65 to 1.12)0.2600.96 (0.72 to 1.29)0.793Organ dysfunction and failure at ICU admission group0.465cInteraction effect for the subgroup and treatment group.SOFA score <5 (n = 1667)0.76 (0.55 to 1.05)0.0950.88 (0.62 to 1.26)0.494SOFA score ≥5 (n = 1803)0.89 (0.70 to 1.13)0.3331.01 (0.79 to 1.31)0.922Laboratory findings at ICU admissionLymphocyte count group0.694cInteraction effect for the subgroup and treatment group.Lymphocyte count <0.724 × 109/L (n = 2613)0.84 (0.69 to 1.03)0.0870.97 (0.78 to 1.20)0.768Lymphocyte count ≥0.724 × 109/L (n = 2221)0.73 (0.57 to 0.93)0.0101.04 (0.80 to 1.36)0.753C-reactive protein group0.333cInteraction effect for the subgroup and treatment group.C-reactive protein <150 mg/L (n = 2627)0.84 (0.68 to 1.03)0.0991.00 (0.80 to 1.26)0.974C-reactive protein ≥150 mg/L (n = 2048)0.71 (0.56 to 0.91)0.0060.93 (0.72 to 1.21)0.614Respiratory support at ICU admission group0.133cInteraction effect for the subgroup and treatment group.Conventional oxygen therapy (n = 371)0.85 (0.45 to 1.63)0.6330.69 (0.28 to 1.74)0.436High-flow nasal cannula / Non-invasive mechanical ventilation (n = 2046)0.99 (0.78 to 1.24)0.9021.25 (0.97 to 1.62)0.088Invasive mechanical ventilation (n = 2571)0.77 (0.62 to 0.96)0.0200.90 (0.72 to 1.14)0.397Corticosteroid therapy during ICU admission group0.225cInteraction effect for the subgroup and treatment group.No (n = 684)0.71 (0.38 to 1.33)0.2850.73 (0.36 to 1.46)0.370Yes (n = 4271)0.81 (0.69 to 0.95)0.0081.05 (0.89 to 1.24)0.561SARS-CoV-2 viral loadViral RNAemia in plasma group0.367cInteraction effect for the subgroup and treatment group.No (n = 167)0.70 (0.20 to 2.51)0.5883.32 (0.14 to 80.96)0.462Yes (n = 584)0.88 (0.60 to 1.29)0.5070.83 (0.55 to 1.26)0.384Viral antigenemia in plasma group0.238cInteraction effect for the subgroup and treatment group.No (n = 400)1.08 (0.62 to 1.87)0.7911.51 (0.81 to 2.84)0.195Yes (n = 335)0.77 (0.47 to 1.27)0.3100.74 (0.41 to 1.34)0.320Viral RNA load in plasma (N1 region) groupeCut-off value obtained from ROC curve for 90-day mortality.0.753cInteraction effect for the subgroup and treatment group.0 copies/mL (n = 130)1.33 (0.27 to 6.53)0.7230.97 (0.04 to 26.90)0.987>0 - <3255 copies/mL (n = 376)0.89 (0.50 to 1.59)0.6921.08 (0.55 to 2.13)0.824≥3255 copies/mL (n = 245)0.90 (0.53 to 1.54)0.7021.25 (0.68 to 2.28)0.469Panel B - Early administration of remdesivir treatment from first symptomsAll patients receiving remdesivir (N = 733)fThe time of administration of remdesivir treatment from first symptoms was not available for 62 patients.<5 days1.65 (1.13 to 2.40)0.0091.53 (1.02 to 2.31)0.042<7 days1.19 (0.88 to 1.60)0.2531.21 (0.85 to 1.72)0.285Abbreviations: HR indicates hazard ratio; CI, confidence interval; APACHE, acute physiology and chronic health evaluation; SOFA, sequential organ failure assessment; RNA, ribonucleic acid.a Adjusted for variables (age, sex, body mass index, diabetes mellitus, chronic liver disease, chronic heart disease, chronic lung disease, chronic renal failure, immunosuppression, APACHE-II score at ICU admission, lymphocyte count at ICU admission, C-reactive protein at ICU admission, respiratory support at ICU admission, septic shock at ICU admission, lopinavir/ritonavir administration, tocilizumab administration, other antiviral administration, corticosteroids, COVID-19 wave and the propensity score).b APACHE-II score was assessed in 2887 patients; SOFA score in 3470 patients; lymphocyte count in 4834 patients; C-reactive protein in 4675 patients; respiratory support in 4988 patients; corticosteroids in 4956 patients; viral RNAemia in plasma in 751 patients; viral antigenemia in plasma in 735 patients; and viral RNA load in plasma (N1 region) in 751 patients.c Interaction effect for the subgroup and treatment group.d Estimation failed due to numerical problem. Because the coefficients did not converge, no further models were fitted.e Cut-off value obtained from ROC curve for 90-day mortality.f The time of administration of remdesivir treatment from first symptoms was not available for 62 patients. Open table in a new tab Abbreviations: HR indicates hazard ratio; CI, confidence interval; APACHE, acute physiology and chronic health evaluation; SOFA, sequential organ failure assessment; RNA, ribonucleic acid. To examine mortality risk for particular types of patients, we explored modification effects by age, APACHE-II score, SOFA score, lymphocyte count, C-RP, respiratory support, corticosteroids, viral RNAemia, viral antigenemia and viral RNA load in plasma. No significant effect modification was found after adjustment for covariates using propensity score (Table 1). Among the overall population receiving RDV, there was significant association observed between early administration (<5 days since symptoms’ onset) and the propensity-adjusted risk of 90-day mortality (HR 1.53, 95% CI 1.02 to 2.31, p = 0.042) (Table 1). In contrast, there were no significant association between <7 days since symptoms’ onset and the risk of 90-day mortality (HR 1.21, 95% CI 0.85 to 1.72, p = 0.285). When compared to the non-RDV group, patients in the RDV group less frequently presented myocarditis, cardiac ischemia, delirium, coagulation disorder, anemia, acute renal failure and liver dysfunction (Supplementary Table 2). The propensity-adjusted analysis showed no association between RDV treatment and nosocomial bacterial pneumonia (OR 1.06, 95 CI 0.84 to 1.33, p = 0.640). In the analysis using the propensity score matching, RDV use was not associated with nosocomial bacterial pneumonia (OR 1.22, 95 CI 0.98 to 1.52, p = 0.068) (Supplementary Table 3). In this large, multicentre study involving over five thousand critically ill patients with COVID-19 admitted to 55 Spanish ICUs, we observed that the use of RDV was not associated with an overall reduced risk of 90-day mortality, nor when analyzing by subgroup populations; patients treated with RDV overall presented longer lengths of ICU stay, which seems to be due to longer length of mechanical ventilation; early administration of RDV from symptoms’ onset was associated with a higher risk of 90-day mortality in the overall population; and patients treated with RDV showed lower significant rates of organic damage associated with severe COVID-19 such as cardiac, neurological, coagulation, renal and liver complications. The underlying pathophysiological mechanisms associated with severe COVID-19 are not fully elucidated,5Kousathanas A Pairo-Castineira E Rawlik K Stuckey A Odhams CA Walker S et al.Whole-genome sequencing reveals host factors underlying critical COVID-19.Nature. 2022; 607: 97-103https://doi.org/10.1038/s41586-022-04576-6Crossref PubMed Scopus (83) Google Scholar,6Erola Pairo-Castineira Sara Clohisey Lucija, Bretherick, Klaric Andrew D. Konrad Rawlik Dorota Pasko et al.Genetic mechanisms of critical illness in COVID-19.Nature. 2021; 591: 92-98https://doi.org/10.1038/s41586-020-03065-yCrossref PubMed Scopus (640) Google Scholar nor are therefore the tools to identify early phenotypes with high risk of developing more severe cases of COVID-19 that might potentially benefit more of early and intensive antiviral treatment.7Elie Azoulay Lara Zafrani Adrien Mirouse Etienne Lengliné Michael Darmon Sylvie Chevret Clinical phenotypes of critically ill COVID-19 patients.Intensive Care Med. 2020; 46: 1651-1652https://doi.org/10.1007/s00134-020-06120-4Crossref PubMed Scopus (39) Google Scholar This has become a major challenge for the scientific community, as patients developing severe COVID-19, and in particular those requiring ICU admission have poor prognosis.8Arbov E Tayara A Wu S Rich TC Wagener BM COVID-19 and Long-Term Outcomes: Lessons from Other Critical Care Illnesses and Potential Mechanisms.Am J Respir Cell Mol Biol. 2022; 67: 275-283https://doi.org/10.1165/rcmb.2021-0374PSCrossref PubMed Scopus (3) Google Scholar Our hypothesis that certain subsets of patients with either pro-inflammatory phenotypes or presenting with high viral loads9Bermejo-Martin Jesús F. Milagros González-Rivera Raquel Almansa Dariela Micheloud Tedim Ana P. Marta Domínguez-Gil et al.Viral RNA load in plasma is associated with critical illness and a dysregulated host response in COVID-19.Crit Care. 2020; 24: 691https://doi.org/10.1186/s13054-020-03398-0Crossref PubMed Scopus (136) Google Scholar might particularly benefit from RDV treatment seems to be refuted by our findings. The lack of influence of RDV on mortality independently of the viral load in plasma and the inflammatory status could reflect an inability of RDV to mediate a significant inhibitory activity of viral replication and/or clinical benefit in patients already exhibiting a large burden viral replication (both groups, RDV treated and non-treated presented to the ICU with high median levels of viral load in plasma), which in turn is associated to a strong stimulation of the innate immunity leading to exacerbated inflammation. Whether new, more specific/potent antivirals could mediate a beneficial effect in this context remains to be elucidated. Our results are in accordance of the recent published update of a living review about remdesivir in adults hospitalized with COVID-19 that confirm that remdesivir probably results in little to no difference in mortality.10Kaka Anjum S. Roderick MacDonald Linskens Eric J. Lisa Langsetmo Kathryn Vela Wei Duan-Porter et al.Major Update 2: remdesivir for adults with COVID-19: a living systematic review and meta-analysis for the American college of physicians practice points.Ann Intern Med. 2022; 175: 701-709https://doi.org/10.7326/M21-4784Crossref PubMed Scopus (17) Google Scholar A possible limitation of the propensity score methods is their inability to control for unmeasured confounding. Another limitation is the different waves of the pandemic, which could have influenced our results. We have however adjusted all of our analyses for this confounder. In summary, treatment with RDV was not associated with improved outcomes in critically ill patients with severe COVID-19, neither overall nor when stratifying by clinically relevant variables such as age, illness severity, organ damage, laboratory findings, respiratory support or SARS-CoV-2 viral load in plasma. Moreover, RDV treatment was associated with longer lengths of ICU admission. Early administration since symptoms onset may prove harmful. Our study adds to the mounting evidence suggesting that RDV is not efficacious in treating severe COVID-19, although further studies are warranted to elucidate whether certain subsets of patients might benefit from it. Conceptualization: CC, AM, AT; Data Curation: CC, AM, TC, Formal analysis: AG; Methodology: all authors; Investigation: CC, AM, TC, FB, AT; Project administration: CC, AM, AT; critical revision of the manuscript for important intellectual content: CC, AM, FB, and AT; and Funding acquisition: AT; Resources: CC, AM, AT; Sofware: AG; Supervision: AT; Validation:all authors; Visualization: all the authors, Writing-original draft: all authors; Writing-review and editing: CC, AM, FB, TC, AT. AThad full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. All authors read and approved the final manuscript. CiberesUCICOVID consortium participated in data collection. This study was supported by the Instituto de Salud Carlos III de Madrid (COV20/00110, ISCIII); Fondo Europeo de Desarrollo Regional (FEDER); "Una manera de hacer Europa"; and Centro de Investigación Biomédica En Red – Enfermedades Respiratorias (CIBERES). DdGC has received financial support from the Instituto de Salud Carlos III (Miguel Servet 2020: CP20/00041), co-funded by European Social Fund (ESF)/ “Investing in your future”. CC received a grant from the Fondo de Investigación Sanitaria (PI19/00207), Instituto de Salud Carlos III, co-funded by the European Union.