COVID-19 research briefs: Remdesivir speeds recovery in patients with COVID-19

Reference

Beigel JH, Tomashek KM, Dodd LE, et al, for the ACTT-1 Study Group Members. Remdesivir for the treatment of Covid-19: preliminary report. N Engl J Med 2020; DOI: 10.1056/NEJMoa2007764

Synopsis

Research brief #35: These authors identified patients who were hospitalized with COVID-19 at 60 sites in 10 countries (45 of the sites were in the United States). Patients with any duration of symptoms were included if they met any of the following criteria: infiltrates on imaging; room air SpO2 94% or less; or requiring supplemental oxygen, mechanical ventilation, or ECMO. Patients (N = 1063) were randomized to receive remdesivir (200 mg loading dose on day 1, followed by 100 mg intravenously once daily on days 2 through 10) or matching placebo. Medication was stopped at hospital discharge. When the study was designed, the primary outcome was the patient's clinical status on an 8-category functional scale, ranging from discharged home and asymptomatic to death. The trial statisticians (without knowledge of treatment assignment; only 72 patients had been recruited) changed the the primary outcome to time to recovery defined as hospitalized but not requiring medical care (some patients stayed in the hospital because they were still testing positive) or discharged home. This is likely because they remembered that you don't need as large a sample size to demonstrate a difference in continuous variables compared with ordinal or dichotomous variables. The analysis appears to have been by intention to treat, though the authors don't explicitly say so. The groups were balanced at baseline, with a mean age of 58 years, 64% were men, 21% African-American, and 26% receiving mechanical ventilation or ECMO on enrollment. The study was halted by a data safety monitoring committee, at which time 391 people in the remdesivir group and 340 in the placebo group had either recovered or died. The median time to recovery was significantly shorter in the remdesivir group (11 vs 15 days; rate ratio for recovery 1.32; 95% CI 1.12 - 1.55). Put another way, recovery was more likely in the remdesivir group (62% vs 52%; P < .001; number needed to treat = 10). There were numerically fewer deaths in the remdesivir group, but this difference was not statistically significant (7.1% vs 11.9%; hazard ratio 0.70; 0.47 - 1.04). Using the 8-point ordinal scale with 8 different stages of illness, there was a greater likelihood of improvement in the remdesivir group (odds ratio 1.50; 1.18 - 1.91). There was no difference in efficacy by sex or by days of symptoms. White patients, patients who were younger than 40 years, and those not receiving mechanical ventilation appeared to do better than their counterparts. However, the subgroup analyses had broad confidence intervals and should be interpreted cautiously; in fact, all of the trends described above could be due to chance as the confidence intervals of all subgroups overlapped