So Remdesivir, some good news, some bad. Not a game changer, but it may help COVID patients requiring supplemental O2 *before* they become critically ill. Some regional confounding in the data. There also are significant logistical issues. https://www.nejm.org/doi/full/10.1056/NEJMoa2007764

Right off, I’ll say the biggest disappointment is the drug’s performance in the sickest patients.

It seems that once the disease has progressed beyond the patient needing supplemental O2, there does not appear to be any utility in its use. This is likely because, at this point, ongoing viral replication is not the issue as much as the body’s response to overwheling infection.

This is also borne out by the ordinal scoring in table 2 (a very confusing table) but the long and shirt of it seems to be if the drug was started in more severe illness, less benefit was seen

Finally the third figure shows the “time to recovery” by each group. This raises all kinds of questions and again is concerning as even for the (new) primary endpoint it was not achieved outside of North America, black or asian patients, and was blunted or absent in older pts!

What is going on here? I am unsure. This kind of heterogeneity makes me worry, very much about confounding.

What can we conclude? Remdesivir may shorten time to recovery in younger (white North American?) patients with COVID requiring supplemental O2. It doesn’t clearly affect mortality. It does *not* rescue severe disease. There is likely confounding heterogeneity between centers.

This presents big problems in my view. This is a disease that results in severe fatal illness in older patients, and there is a definite disparity in outcomes with higher mortality in AA patients. This drug appears to help *neither* group at the primary end point.

Then think about the logistics. This drug is not in wide supply, and to show a benefit we would need to give it to a *huge* patient population - COVID requiring supplemental O2 - without a clear idea of who is most likely to progress.

Finally with the clear racial, age, and geographic confounding, utility only after *changing a primary endpoint* and even then, only working in that subgroup of patients on supplemental O2 - my skepticism is way up.

This drug is not going to be useful.

The best news will be for Gilead. Every dose of this drug will be bought and dosed, useful or not, for the immediate future.

Bad news for those of us looking for a clear evidence of benefit that can be generalized to *all* patients and we will now never know the answer.

It was a *terrible* decision to stop this placebo control arm. This is not convincing data. This makes me, if anything, far more angry over that decision.

This is deeply confounded data.