Unfortunately, we’re getting a dose these days of “That’s why you run clinical trials”. Word came Monday evening (Peter Loftus in the WSJ, and a Lilly statement) that the ACTIV-3 trial being run by the NIH has shown lack of efficacy for the combination of the Eli Lilly/AbCellera anti-coronavirus antibody (bamlanivimab, LY-CoV555) when combined with remdesivir in hospitalized patients. This is the trial that had been paused earlier this month for safety concerns, but now no more patients will be enrolled at all. A thorough review of the data showed that there were, in fact, no differences in safety between the arms of the trial with and without the antibody – but there was also no difference in efficacy.
Lilly’s own statement notes that their own BLAZE-1 trial continues. That one is studying patients who are not hospitalized, with this antibody as a monotherapy and in combination with another anti-coronavirus antibody candidate (etesevimab, LY-CoV016). The company provided some data earlier this month on this trial, and indeed has said that they’re asking for an Emergency Use Authorization for high-risk patients on the basis of the numbers so far. There’s also the NIH ACTIV-2 trial, which continues to study the antibody in mild-to-moderate disease, and Lilly’s BLAZE-2, which is looking at using it for prophylaxis in patients and staff at long-term care facilities. It would be a coherent story if these trials show positive data – that way we could say that this therapy needs to be given early, and is of little use after someone has already been hospitalized. That makes a lot of sense, too – but you know what? A lot of things make sense that still aren’t real. We have to get the clinical data to be sure.
As if to underline this point, we’ve also recently had a controlled trial of convalescent plasma report its results (Damian Garde at Stat, and the article in the BMJ). Recall that the earlier Mayo Clinic data on this had no control group, making it frankly impossible to say how much benefit the plasma treatment might have had. But the new PLACID trial was a study of 464 patients across India, randomized versus a control arm, and it found no benefit to the plasma treatment at all. One objection is that many of these patients may have been dosed too late as well – many of these patients, it seems, were already developing their own antibodies, and it’s possible that if there is a benefit, that it will only show up with earlier treatment. But we don’t know that yet. Right now, convalescent plasma treatment is, at the very least, not looking spectacular, and could even be more or less useless, which would be quite disappointing.
You’ll notice a theme here, to which we can add the recent remdesivir data from the SOLIDARITY trial. Earlier data had suggested that the drug has some positive effects (although not dramatic ones) but the larger data set didn’t bear that out. What these things do is set bounds for what you can expect for a given therapy: the level of response in these cases means that even if there is some benefit, it’s going to be limited (or, at the most optimistic, limited in these particular populations of patients). A Wonder Drug would have given a stronger signal, and none of these are wonder drugs. The question is whether there’s any signal left after more waves of well-controlled data come in – and if so, what patients can get what benefits there are.
A month ago, for example, Regeneron said that their own monoclonal antibody therapy showed some signs of benefit, particularly in patients who had not mounted their own immune response. But if that’s not going to be the best-responding group, who is? Perhaps this study ran into the same problem as the PLACID trial above, and that could mean that monoclonals won’t add that much to people who are already producing their own antibodies. But if the Regeneron antibody cocktail worked wonders, we would have seen something more impressive. It might be that prophylaxis is where these therapies will show their worth (can’t dose much earlier than before a person has the disease at all!) I very much hope so. I will admit to having expected better from the mAbs in general, though, and even from convalescent plasma.
Prophylaxis frankly might be an easier public health decision than one that relies on very early treatment. After all, most cases of coronavirus resolve, and most people don’t end up in the hospital. When do you make the call for an expensive bolus of monocolonal antibodies? Your best shot is probably to catch the highest-risk patients as early as you can, and hope that it’s still early enough. But if such treatment can protect high-risk workers (sort of like a temporary vaccine), then you can far more easily pick out the people who could benefit.
Of course, right now we’re heading back into high case loads in many parts of the country, with some (like Utah) already beginning to hit the wall of hospital capacity. Deploying mAbs under these conditions is going to be more complex than ever. Let’s hope that we get some more clinical data soon to guide these decisions. There really does look to be a gap before any vaccine rollout where the antibodies could help. The key word being “could”. Might. Maybe. Clarity has been in short supply in 2020, and here we are calling for more of it.