An ongoing dialogue on HIV/AIDS, infectious diseases,
October 18th, 2020
Does Remdesivir Actually Work?
Quick answer — it’s complicated.
Let’s start with a clinical anecdote — rightfully considered the weakest form of evidence, yet paradoxically holding great power over us because we’re imperfect humans. It’s the way we’re wired.
In April, a patient of mine with stable HIV came into the hospital with COVID-19 pneumonia. (Certain details changed for privacy.)
She works cutting hair, in a community hard-hit from the pandemic and where mask-wearing was inconsistent. She knew as soon as she developed fever, chills, and back pain that this is what she had.
She must have had symptoms for no more than 36 hours when she arrived at the hospital, and because she has HIV (who knew whether this worsened outcomes?) and is quite overweight, she was admitted.
She enrolled in the SIMPLE study — which compared remdesivir for 5 or 10 days to standard of care, all open-label — and was randomized to the 5-day course. She received her first dose of the drug the night of admission.
Again, she had been symptomatic for no more than 2 days.
The next day, she looked like a new person. Her fever was down, she was breathing more easily, and she told me her back pain went away as soon as the first dose had completed its infusion. She left the hospital on day 3, and made a complete recovery.
Of course, she could have recovered just as quickly without remdesivir — that’s the problem with an anecdote.
But based on this and other cases I saw — and the extensive experience of my indefatigable colleagues Dr. Francisco Marty and his team, who enrolled dozens of patients into this study — I was not surprised when in May, a different and more rigorous remdesivir clinical trial reported significantly faster recovery in the treatment arm than in the controls.
And because this study — called ACTT-1, now with final results — included a placebo arm and was blinded, this provided much stronger evidence that remdesivir actually works. (That’s in the title of this post.) It worked particularly well in people with shorter duration of symptoms and in those requiring oxygen. It didn’t help people so sick that they needed mechanical ventilation or ECMO.
When the data from ACTT-1 became available, we created a construct about these critically ill patients who didn’t benefit from remdesivir. In this view, they were in the immune phase of the illness, where the body’s immunologic response to the infection drove more disease than viral replication.
We can’t expect an antiviral to control these processes. Just like oseltamivir or baloxavir for influenza, you have to act early with remdesivir when treating SARS-CoV-2. Let dexamethasone or some other immunomodulator do the late work.
See, it all fits together perfectly.
But there were always holes in this neat little package.
First, the original remdesivir study from China showed no benefit of treatment. Yes, it was underpowered due to dropping case numbers, but the drug didn’t lower viral loads in recipients either. Concerning.
Second, the open-label study my patient enrolled in had a funny result in the 10-day arm — no apparent benefit compared with standard of care. (The 5-day arm did show benefit.) How do we explain that?
Now we have the interim results of the SOLIDARITY study, at least in preprint form, and that neat little package has even more holes.
In SOLIDARITY, over 11,000 hospitalized patients with COVID-19 (from 405 hospitals and 30 countries) were randomized between whichever study drugs were locally available and open control. This included up to five options — four active treatments versus local standard-of-care. The drugs were lopinavir/ritonavir, hydroxychloroquine (remember those?), interferon beta-1a, or remdesivir.
(Based on the limited availability of some of the drugs across countries, the study arms differed in size.)
The results for all the interventions failed to show a survival benefit. And the survival curves for the remdesivir arm versus standard of care look depressingly the same:
So how do we explain these discordant results? We can’t do so completely given the different study designs and populations. But just as how ACTT-1’s benefits can’t overrule the SOLIDARITY results, nor can SOLIDARITY negate ACTT-1 or the 5-day results from SIMPLE.
So let’s put all the studies together, as shown here in this colorful meta-analysis, and exclude patients who are on mechanical ventilation since no study demonstrated benefit in this population:
Remdesivir Meta-Analysis (Sensitivity Analyses #1 / Version #6)
*This one excludes mechanically ventilated patients (There is one from the Lancet, 2020 paper that I couldn't remove).
Random effects used. pic.twitter.com/GQZHibkcdn
— mike johansen (@mikejohansenmd) October 16, 2020
First, we’ll note that SOLIDARITY’s much larger sample size trumps (ouch) the other studies. Second, the point estimate just crosses 1 (no benefit), but falls to the left of the line, suggesting (if you squint) some benefit.
If I had to postulate where we’d see the greatest benefit for remdesivir, it would be in patients with shorter duration of symptoms. Even in the negative underpowered study from China, those with fewer days of illness did better than controls. Could it be that the favorable results in ACTT-1 were from the fact that 25% of patients were enrolled with symptoms for 6 days or fewer?
Related, SOLIDARITY began enrollment in March, and for much of the enrollment period, patients with COVID-19 did everything they could to avoid hospitalization. For many, I suspect the short window of time for this antiviral to benefit had closed by the time they were admitted. Duration of symptoms is not reported in the preprint, a critical piece of information.
So for now, the answer to the question, “Does remdesivir actually work?” is a cautious maybe. Sometimes. For some people.
Which, given the absence of anything else right now and its low toxicity, means I’d still recommend it for most hospitalized people with COVID-19 — with the hope of giving it sooner rather than later, especially for those on oxygen at high risk for disease progression.
But if we can learn anything from the mental gyrations required to square these conflicting study results, it’s that we definitely need more effective options.