Time to catch up with the latest clinical news in the field. The big story this morning is apparently some real data out of a controlled trial of chloroquine from China. There have been anecdotal reports, but to the best of my knowledge we have seen no actual numbers. These are not exactly from the peer-reviewed literature, either, but were unearthed by market analyst Umer Raffat at Evercore/ISI. It’s another small study, understandably, with 15 patients in the control group and 15 randomized to a treatment group getting 400mg/day of hydroxychloroquine for five days. At the end of this period, the treatment group showed 13/15 negative for viral RNA via throat swab. . .and the control group showed 14/15 negative. Other parameters were also so similar between the treatment and control groups as to be uninterpretable. So those numbers are unfortunately not too useful. It’s tempting to run with a “hydroxychloroquine fails” take, but we can’t even say that with such a strong control group response. It’s basically a blown trial that can’t tell us anything – we have no idea what an earlier endpoint would have told us, for example, although I have to note that this is about the time course of treatment in the widely-discussed Marseille study. We need more and better numbers (which is the same conclusion as the authors of this study have), and I hope that we get them soon.
Now for the rest of the trial news. This study from China isn’t large either: 7 control patients, 21 getting a lopinavir/ritonavir combination, and 16 getting the Russian antiviral compound umifenovir (Arbidol). That second regimen showed some efficacy against SARS and was certainly a logical thing to try, but it has already failed one trial, and by gosh it failed this time, too: no benefit, and a trend towards more adverse events. It’s hard to see why this regimen should have more work done on it, to be honest. Now, umifenovir I hadn’t heard of, and I can’t say that I’m in love with its structure, either. Those alpha-aminomethyl phenols are often trouble; they can eliminate to form a reactive quinone methide which can go on to do all sorts of things. I have no problem with covalent drugs per se, but quinone methides are, I think, more towards the hot end of the spectrum and more liable to go for the first nucleophile they see. As far as I can see, the drug’s targets are unknown – it seems to affect the earlier stages of viral entry and has been tried against a wide range of viruses (DNA and RNA). It has sold very well in Russia and China over the years, but is hardly approved anywhere else, and apparently even the Russian Academy of Medical Sciences came out in 2007 saying that there was no evidence that it was useful. It wasn’t this time – no difference from the control group.
That was a small trial, to be sure, but the drug shows up in this trial as well, with 120 patients getting Arbidol and 116 getting favipiravir, which I’ve also mentioned before. That first trial looked as if there might be some benefit, but it was open-label with only 35 patients. In this case, we don’t have a control group, and all that we can say is that favipiravir looks better than Arbidol. If that smaller trial above turns out to be accurate, then the latter is pretty much the same as getting nothing, in which case perhaps favipiravir helps. Mechanistically I have trouble seeing how it can, though, and have since the beginning of this whole epidemic, since it barely does anything in the in vitro assays. It definitely has some adverse reactions as well (liver function tests, psychiatric symptoms, and others), but a patient in bad shape would, you’d think, rather get this drug than nothing.
There’s a lot of noise in this area, of course. According to this tweet, the Chinese State Council, for example, has apparently announced that Traditional Chinese Medicine has a 90% success rate in treating the virus. What I’m supposed to make of that, I do not know – I hope that this is a mistranslation of earlier reports that 90% of the outbreak patients have been given some form of TCM, and even that number is hard to evaluate. The Chinese government has of course been promoting this stuff for some time now, and I suppose that we can’t expect them to ignore an opportunity like this, but geez. Here’s the South China Morning Post on the issue. Papers on this stuff are already appearing, and already being ripped to pieces for having things like duplicated data in them. Interestingly, it appears that earlier WHO guidelines recommended against herbal remedies for the Covid-19 epidemic, but that language appears to have disappeared, in what some speculate is pressure from the Chinese authorities. My own opinion is simple to state: bring clinical data or shut up.
We still have several remdesivir trials waiting to read out, and of course, everyone’s object of interest: hydroxychloroquine (and its combination with azithromycin). The full paper on that one has published, though, and Leonid Schneider noticed that the curves for the hydroxychloroquine/azithromycin treatment group are different (and slightly better looking) than those in the preprint. Others have gone through the data tables and found that the evidence is not as solid as it might seem from just looking at the now-famous chart, either. My take continues to be that this work definitely justifies a larger, more controlled study, but that it also is nowhere near large enough or well-controlled enough to prove anything by itself. I understand that many physicians are going ahead and giving patients the HCQ/AZ combination anyway, and I can see why they’re doing that if they believe that the chances of benefit outweigh the chances of harm. But we don’t quite know if they do, to be honest. Giving it only to people who are in bad shape (already on ventilators) is probably the best way to split that difference, and from what I’m hearing, that’s the usual call that physicians are making. I hope it does some good, and I hope that we find out as soon as possible if it does.