There’s a lot of talk about the Chinese experience with Remdesivir (RDS) vs placebo (PLC) as reported by Wang, et al. It was the first double-blind, placebo-controlled, multi-center randomized clinical trial (DB, PC, MCRCT) on RDS that was done in February and March 2020 but was not published in Lancet until the end of April 2020. There was unintended almost malicious early release of this study’s results on the WHO website that was immediately taken down after few hours.
Remdesivir is a broad spectrum antiviral drug with decent success against SARS and MERS; not so good one with Ebola, though. The study did not show conclusive benefit of the drug against severe COVID-19, despite a trend toward faster clinical improvement for treatment within 10 days of symptom onset.
The study was terminated before attaining the prespecified sample size of 453 (302 on RDS and 151 on PLC) patients because of the control of the outbreak in Wuhan after March 12. There were also restrictions on bed availability that time that resulted in most patients being enrolled later in the disease course, so they were unable to adequately assess whether earlier treatment with RDS might have provided clinical benefit.
They actually specifically mentioned that in their statistical analysis (“The original design required a total of 325 events across both groups, which would provide 80% power under a one-sided type I error of 2·5% if the hazard ratio comparing remdesivir to placebo is 1·4, corresponding to a change in time to clinical improvement of 6 days assuming that time to clinical improvement is 21 days on placebo”) on page 1572.
They ended up having 236 eligible patients randomized on 2:1 ratio (158 RDS and 78 PLC). There lies the proverbial smoking gun on why, although it is a well designed study (DB, PC, MCRCT) and well conducted (high protocol adherence and no loss-to-follow up), the study is somewhat flawed as the absence of statistical significance in an underpowered trial means that the findings are INCONCLUSIVE.
As I always tell the medical students, in most if not all clinical trials, N (number of study subjects) is King.
For example: every time somebody asks me how many bikes do i have, I always answer N + 1 (N: bikes I already have at home and + 1 is the bike I’m looking at and salivating for in the bike store that I am forbidden to buy).