key: cord-0753569-3ry876lw
authors: Garibaldi, Brian T; Wang, Kunbo; Robinson, Matthew L; Betz, Joshua; Alexander, G Caleb; Andersen, Kathleen M; Joseph, Corey S; Mehta, Hemalkumar B; Korwek, Kimberly; Sands, Kenneth E; Fisher, Arielle M; Bollinger, Robert C; Xu, Yanxun
title: Real-World Effectiveness Of Remdesivir In Adults Hospitalized With Covid-19: A Retrospective, Multicenter Comparative Effectiveness Study
date: 2021-12-15
journal: Clin Infect Dis
DOI: 10.1093/cid/ciab1035
sha: c5aaf7b56a7d2cd21b89f406b3762dce914f6e87
doc_id: 753569
cord_uid: 3ry876lw

BACKGROUND: There is an urgent need to understand the real-world effectiveness of remdesivir in the treatment of SARS-CoV-2. METHODS: This was a retrospective comparative effectiveness study. Individuals hospitalized in a large private healthcare network in the US from February 23, 2020 through February 11, 2021 with a positive test for SARS-CoV-2 and ICD-10 diagnosis codes consistent with symptomatic COVID-19 were included. Remdesivir recipients were matched to controls using time-dependent propensity scores. The primary outcome was time to improvement with a secondary outcome of time to death. RESULTS: Of 96,859 COVID-19 patients, 42,473 (43.9%) received at least one remdesivir dose. The median age of remdesivir recipients was 65 years, 23,701 (55.8%) were male and 22,819 (53.7%) were non-white. Matches were found for 18,328 patients (43.2%). Remdesivir recipients were significantly more likely to achieve clinical improvement by 28 days (adjusted hazard ratio [1.19, 95% confidence interval (CI), 1.16-1.22]). Remdesivir patients on no oxygen (aHR 1.30, 95% CI 1.22-1.38) or low-flow oxygen (aHR 1.23, 95% CI 1.19-1.27) were significantly more likely to achieve clinical improvement by 28 days. There was no significant impact on the likelihood of mortality overall (aHR 1.02, 95% CI 0.97-1.08). Remdesivir recipients on low-flow oxygen were significantly less likely to die than controls (aHR 0.85, 95% CI 0.77-0.92; 28-day mortality 8.4% [865 deaths] for remdesivir patients, 12.5% [1,334 deaths] for controls). CONCLUSIONS: These results support the use of remdesivir for hospitalized COVID-19 patients on no or low-flow oxygen. Routine initiation of remdesivir in more severely ill patients is unlikely to be beneficial.

As of December 1, 2021, there have been over 260 million cases of SARS-CoV-2 infection, the virus that causes COVID-19, with greater than 5 million deaths. 1 There remains an urgent need to deploy therapeutics to improve COVID-19 outcomes. Remdesivir, a nucleoside analog that inhibits SARS-CoV-2 replication, reduced the time to clinical improvement in 1,062 patients in the National Institutes of Health (NIH) Adaptive COVID-19 Treatment Trial (ACTT). 2 A study of 237 patients in Hubei, China did not show a statistically significant benefit for remdesivir, but the trial was stopped early due to a lack of COVID-19 cases. 3 The open-label Solidarity trial (2,743 remdesivir recipients), sponsored by the World Health Organization (WHO;) 4 , did not show a mortality benefit and the open-label DisCoVeRy study (429 remdesivir recipients) 5 did not show a difference in outcomes.

Data on the real-world effectiveness of remdesivir has also yielded conflicting results. One retrospective study from a single health system showed that among 342 recipients, remdesivir reduced the time to clinical improvement by 2 days but did not decrease mortality. 6 A retrospective study of 1,172 remdesivir patients from the Veterans Health Administration (VA), 7 a UK retrospective study of 1,549 remdesivir recipients 8 and two small multicenter retrospective studies (including 368 and 286 remdesivir patients) 9,10 did not demonstrate a mortality benefit. The VA study also found an increased length of stay for remdesivir recipients. 7 A larger industry-sponsored retrospective study of 28,555 remdesivir patients from the Premier Healthcare Database showed an overall improvement in mortality at 14 and 28 days, including a benefit in some critically ill patients. 11 A post-hoc analysis of the ACTT-1 trial suggested that remdesivir may reduce progression to invasive mechanical ventilation (IMV) and death, even among patients who were already receiving IMV. 12 These discordant findings have contributed to variation in remdesivir use 13 and while the US Food and Drug Administration (FDA) granted full approval for remdesivir, 14 the World Health Organization (WHO) recommends against its routine administration. 15 Given the ongoing pandemic, it is critical to examine the real-world effectiveness of remdesivir using large populations, which can M a n u s c r i p t 5 allow for analyses within distinct clinical subgroups. Using a dataset from a large, geographically diverse multi-hospital health system in the United States, we quantified the effectiveness of remdesivir in the treatment of hospitalized patients with COVID-19, with a focus on patients with different disease severity at time of treatment initiation.

The COVID-19 Consortium of HCA Healthcare and Academia for Research GEneration (CHARGE) is a group of 10 academic centers in partnership with HCA Healthcare and the federal Agency for Health Research and Quality (AHRQ). 16 HCA Healthcare comprises over 2,000 care sites including more than 180 acute-care facilities. This system conducts over 32 million annual patient encounters, including approximately 6% of all inpatient care in the US. 17 As of July 2021, over 180,000 COVID-19 patients had been admitted to HCA facilities.

We included patients hospitalized for COVID-19 at HCA hospitals in the US between February 23, 2020 and February 11, 2021. Diagnosis of COVID-19 was determined by the detection of SARS-CoV-2 using any nucleic acid test with an FDA Emergency Use Authorization (EUA) combined with specific ICD-10 codes that indicate symptomatic infection (Appendix Table 1 Research Institute, to aggregate electronic health record (EHR) data in an enterprise data warehouse. Only data from facilities using a single EHR system were included, accounting for >90% of affiliated facilities. Data included socio-demographics, past medical history, ICD-10 codes, laboratory data, vital signs, medications, oxygen support (e.g. low-flow nasal cannula, high-flow nasal cannula (HFNC), non-invasive positive pressure ventilation (NIPPV), IMV and extracorporeal membrane oxygenation (ECMO)), length of stay, location of discharge, and death. Limited data sets were accessible via a secure platform hosted within a private virtual network.

HCA guidelines for the use of remdesivir consistent with the initial FDA EUA were established. 19 Guidelines were updated to align with FDA recommendations following full approval. 14,20 At the time of analysis, guidelines recommended a 5-day treatment course for patients with an oxygen saturation less than 94% or the need for oxygen.

The primary outcome was time to clinical improvement from the first day of remdesivir treatment or the matched day, defined as a 2-point decrease in the 8-point WHO severity score or discharged alive from the hospital without worsening of the WHO severity score within 28 days (see Appendix Table 2 ). 6, 21 Failure of clinical improvement was censored at the last day of follow-up or 28-days, whichever came first. The secondary outcome was time to death from the first day of remdesivir treatment or the matched day. Patients who were discharged alive to "home" or "selfcare" were censored at 28 days. 22 Patients who were discharged to another health care facility without a known death date were censored at last follow-up. Patients discharged to hospice with a recorded death date were included in the death group.

A c c e p t e d M a n u s c r i p t 7

To account for the non-randomized use of remdesivir and the variable timing of administration, we used time-dependent propensity score (PS) matching to create pairs of individuals, one patient treated with remdesivir and the other the most similar patient eligible for treatment at the time of remdesivir initiation but who did not receive remdesivir. The PS was computed using a time-dependent Cox proportional hazards regression model with the time from admission to the first dose of remdesivir being the outcome. Dexamethasone was included as a matching variable (see Supplement). 6, 23, 24 In order to account for changes in the pandemic over time, an individual that received remdesivir prior to October 1, 2020 had to be matched to a control patient hospitalized before October 1, 2020 (Appendix Figure 1 ). An additional time constraint was imposed such that a patient who received k days of treatment with remdesivir was matched to a control patient who stayed in the hospital at least k days (up to a maximum of 5) beyond the matching day. 6 This condition avoids matching remdesivir patients to individuals who were healthy enough to be discharged soon after the matching day and would not have been considered candidates for remdesivir treatment.

We used Cox proportional hazards regression models to estimate the association between remdesivir treatment and outcomes of interest on the matched sets. 25 We included demographics, oxygen delivery device, vital signs, key laboratory data, comorbidities (including the Charlson Comorbidity Index 26 ) and COVID-19-specific medications (e.g. dexamethasone, tocilizumab, etc.) in the models (Table 1) A c c e p t e d M a n u s c r i p t 8

We performed four sensitivity analyses. First, we excluded individuals who received corticosteroids. Second, we excluded individuals treated before July 1, 2020. Third, we reduced the number of days that control patients had to remain in the hospital after the matched day to 4 days and 3 days. Fourth, we repeated analyses after matching patients who received both remdesivir and dexamethasone to patients who received dexamethasone alone. Table 1 shows the demographic and clinical characteristics of remdesivir recipients and control patients at hospital admission and after matching. Appendix Table 3 shows characteristics of unmatched remdesivir and control patients.

Appendix Table 4 Figure 2E ; median of 28 days in both groups; (IQR, 10,28 in remdesivir patients compared to IQR, 9,28 in controls).

There was no significant impact of remdesivir on mortality overall (aHR 1.02, CI 0.97-1.08; Our study has several strengths. We included data from over 160 hospitals across 21 states making the findings generalizable to a wide range of health systems. Our dataset included longitudinal vital signs and laboratory data, allowing for more detailed matching than in the Premier Healthcare Database retrospective study. 11 Our study also included a larger percentage of non-white participants than have previously been studied in remdesivir trials, 2,30,31 further expanding the A c c e p t e d M a n u s c r i p t 12 generalizability of the results to populations that have borne a disproportionate burden during the pandemic. [32] [33] [34] Our results are concordant with the ACTT-1 trial, which was a well-designed doubleblinded placebo-controlled trial that showed a similar decrease in time to clinical improvement in remdesivir treated patients. While ACTT-1 was not powered to detect differences in mortality, patients who were receiving low-flow oxygen at enrollment had a significant reduction in mortality, similar to our results.

Our primary analyses accounted for the use of dexamethasone and other anti-inflammatory therapies that have been shown to improve outcomes in COVID-19. 22, 35, 36 In sensitivity analyses we found that remdesivir alone as well as remdesivir plus dexamethasone (compared to dexamethasone alone) were associated with a statistically significant increase in the likelihood of clinical improvement. Remdesivir plus dexamethasone was also associated with a statistically significant decrease in mortality (compared to dexamethasone alone) in those on low-flow oxygen.

This suggests that the benefits seen in the primary analysis were driven at least in part by remdesivir and not by concomitant anti-inflammatory therapies.

Our study has limitations. Unmeasured variables could affect our treatment effect estimates. We were unable to match approximately half of the remdesivir patients, largely due to the fact that many patients received at least one dose of remdesivir, particularly after October 1, 2020. Symptom onset was not available in the dataset, so we were not able to examine whether or not the benefit of remdesivir differed based on timing of treatment. Since antiviral therapies are likely most effective early in the disease course, differential timing of treatment could bias outcomes towards specific groups. COVID-19 outcomes have improved over time, 37,38 which we accounted for by matching remdesivir patients to control patients based on admission before October 1, 2020. The early months of the pandemic presented unique challenges to health systems beyond broader secular trends. We examined this impact in sensitivity analyses excluding patients hospitalized A c c e p t e d M a n u s c r i p t 13 before July 1, 2020. Our study was conducted prior to the widespread use of vaccines and the emergence of variants such as Delta and Omicron which could impact generalizability.

When health systems are overwhelmed by COVID-19 cases, outcomes for COVID-19 patients are worse 39, 40 and care for non-COVID diseases is limited. 41, 42 The finding that remdesivir patients achieved clinical improvement two days sooner than matched controls could reduce the strain on healthcare systems during COVID-19 surges. This finding contrasts with a VA study that showed an increased length of stay among remdesivir recipients. 7 The majority of discharges among control patients in the VA study occurred within 3 days of matching. Since ACTT-1 excluded patients expected to be discharged within 72 hours of randomization, 2 we imposed a requirement that controls remain in the hospital for the treatment duration of the remdesivir patients (up to 5 days) in order to exclude patients who were well enough to be discharged without COVID-specific therapy.

Our findings were sensitive to this requirement when it was reduced to 3 days.

In this large, multicenter retrospective cohort study, treatment with remdesivir significantly 

Remdesivir for the Treatment of Covid-19 -Final Report

Remdesivir in adults with severe COVID-19: a randomised, double-blind, placebo-controlled, multicentre trial

Repurposed Antiviral Drugs for Covid-19 -Interim WHO Solidarity Trial Results

Remdesivir plus standard of care versus standard of care alone for the treatment of patients admitted to hospital with COVID-19 (DisCoVeRy): a phase 3, randomised, controlled, open-label trial. The Lancet Infectious Diseases

Comparison of Time to Clinical Improvement With vs Without Remdesivir Treatment in Hospitalized Patients With COVID-19

Association of Remdesivir Treatment With Survival and Length of Hospital Stay Among US Veterans Hospitalized With COVID-19

Evaluation of the effectiveness of remdesivir in treating severe COVID-19 using data from the ISARIC WHO Clinical Characterisation Protocol UK: a prospective, national cohort study. medRxiv

Remdesivir Versus Standard-of-Care for Severe Coronavirus Disease 2019 Infection: An Analysis of 28-Day Mortality. Open Forum Infectious Diseases

Remdesivir and Mortality in Patients with COVID-19

Remdesivir treatment in hospitalized patients with COVID-19: a comparative analysis of in-hospital all-cause mortality in a large multi-center observational cohort

Remdesivir for the prevention of invasive mechanical ventilation or death in COVID-19 -A post-hoc analysis of the Adaptive COVID-19 Treatment Trial-1 Cohort Data

Annals of Internal Medicine. 2021. 14. FDA. FDA Approves First Treatment for COVID-19. US Food and Drug Administration

19-Registry-by-Forming-Consortium-With-AHRQ-and-Research-Institutions-to-Accelerate-COVID-19-Research/default.aspx

Good research practices for comparative effectiveness research: defining, reporting and interpreting nonrandomized studies of treatment effects using secondary data sources: the ISPOR Good Research Practices for Retrospective Database Analysis Task Force Report--Part I. Value Health

Emergency Use Authorization for Remdesivir

Update: FDA Broadens Emergency Use Authorization for Veklury (remdesivir) to Include All Hospitalized Patients for Treatment of COVID-19

Tocilizumab for the Treatment of COVID-19

Among Hospitalized Patients: A Matched Retrospective Cohort Analysis. Open Forum Infectious Diseases

Risk of death among users of Proton Pump Inhibitors: a longitudinal observational cohort study of United States veterans

Propensity score matching with time-dependent covariates

Regression Models and Life-Tables

Validation of a combined comorbidity index

Infectious Diseases Society of America Guidelines on the Treatment and Management of Patients with COVID-19

Surviving Sepsis Campaign Guidelines on the Management of Adults With Coronavirus Disease 2019 (COVID-19) in the ICU: First Update

Remdesivir for 5 or 10 Days in Patients with Severe Covid-19

Effect of Remdesivir vs Standard Care on Clinical Status at 11 Days in Patients With Moderate COVID-19: A Randomized Clinical Trial

The National COVID Cohort Collaborative: Clinical Characterization and Early Severity Prediction. medRxiv

The COVID racial data tracker

COVID-19 and racial disparities

Dexamethasone in Hospitalized Patients with Covid-19

Interleukin-6 Receptor Antagonists in Critically Ill Patients with Covid-19

Trends in COVID-19 Risk-Adjusted Mortality Rates

Improved outcomes over time for adult COVID-19 patients with acute respiratory distress syndrome or acute respiratory failure

Association of Intensive Care Unit Patient Load and Demand With Mortality Rates in US Department of Veterans Affairs Hospitals During the COVID-19 Pandemic

The association between mechanical ventilator compatible bed occupancy and mortality risk in intensive care patients with COVID-19: a national retrospective cohort study

The Impact Of The COVID-19 Pandemic On Hospital Admissions In The United States. Health Affairs

The Untold Toll -The Pandemic's Effects on Patients without Covid-19

No Supplemental Oxygen Low Flow Supplemental Oxygen High Flow Nasal Cannula CPAP or BiPAP Mechanical Ventilator

Mean (SD): Temperature ( o Celsius)

Pulse (beats per minute) Systolic BP (mmHg)

Diastolic BP (mmHg)

A c c e p t e d M a n u s c r i p t 16 A c c e p t e d M a n u s c r i p t 17 A c c e p t e d M a n u s c r i p t A c c e p t e d M a n u s c r i p t 19 c Comprised non-White, non-Black, and non-Latinx patients.d Only the Charlson Comorbidity Index was used in the Cox proportional hazards models. Individual comorbidities are shown but were not used in matching.