The Effect of Remdesivir for The Treatment of COVID-19 on Mortality Among Hospitalized Patients

Background: Guidelines recommend remdesivir for certain patients hospitalized with COVID-19. However, the evidence underpinning these guidelines is limited, based on subgroup analyses from randomized trials, early observational studies, or expert opinion. Further investigation is needed to better inform the guidelines with regards to the effectiveness of remdesivir in these patients.

Objectives: To estimate the effect of a remdesivir initiation within two days of hospital admission compared with no remdesivir treatment on mortality among hospitalized patients with a diagnosis of COVID-19.

Methods: The data include inpatient claims for billable products, procedures, and services from 297 hospitals across the United States. Adult patients admitted to the hospital with a COVID-19 diagnosis between May 1, 2020 and August 15, 2021 and no prior emergent care were included in the study. Baseline comorbidities were recorded from hospital encounters at or up to 12 months prior to admission. Timing and use of in-hospital medications, procedures, and oxygen were identified from hospital charges. Mortality was defined as a discharge disposition of death or transfer to hospice. Discharge with any other disposition treated was considered a competing event. The data were analyzed using a clone-censor-weight approach designed to emulate a randomized clinical trial. First, two copies of each patient were created, one for each treatment protocol. Second, patient copies were artificially censored at the time of deviation from their group’s protocol. Finally, a Cox proportional hazard model was fit, conditional on baseline and time-varying covariates, to estimate the probability of remaining uncensored at each time point. Cumulative risks and risk differences were estimated with inverse probability of censoring weighted estimators.

Results: Overall, 223081 patients were included, contributing 1265692 person-days to the remdesivir protocol and 1152472 to the no remdesivir protocol. The estimated counterfactual 30-day mortality risks were 13.0% (95% CI: 12.8%, 13.3%) and 18.4% (95% CI 17.6%, 19.2%) for the remdesivir and no remdesivir protocols, respectively, for a risk difference of -5.4% (95% CI: -6.3%, -4.5%). The risk differences were -3.7% (95% CI: -5.9%, -1.5%), -10.1% (95% CI: -15.8%, -4.4%), and -7.3% (95% CI: -12.2%, -2.3%) for patients on low-flow oxygen, high-flow oxygen, and mechanical ventilation, respectively.

Conclusions: In this large study of patients hospitalized with COVID-19, we found that remdesivir treatment within two days of hospital admission reduced mortality compared with no remdesivir use during the hospital stay. These findings support guidelines suggesting remdesivir for these patients, particularly those receiving high-flow oxygen support.