Coronavirus: Some Clinical Trial Data

We're starting to see some clinical trial data on possible Covid-19 therapies, specifically some of the ones mentioned in this earlier post on small-molecule approaches. First, the bad news (although I have to say it's not unexpected). A trial of the lopinavir/ritonavir combination has shown no benefit. I say "not unexpected" because these antiretroviral drugs are not at all targeted to coronavirus proteins - this was always something of a long shot.

The second report is a bit secondhand; no paper or preprint with full data has yet appeared. The Chinese authorities report that favipiravir shows useful effects in trials of the drug. I'll give people the choice of links for that one: the original is China Daily, but you'll have to deal with the little box up in the right corner labled "Xi's Moments", where a smiling Xi Jinping urges joint efforts to fight the pandemic. The report at the Guardian is actually more complete, and is Xi-free (as far as I can tell).

. . .Patients who were given the medicine in Shenzhen turned negative for the virus after a median of four days after becoming positive, compared with a median of 11 days for those who were not treated with the drug, public broadcaster NHK said.In addition, X-rays confirmed improvements in lung condition in about 91% of the patients who were treated with favipiravir, compared to 62% or those without the drug. . .. . .But a Japanese health ministry source suggested the drug was not as effective in people with more severe symptoms. "We've given Avigan to 70 to 80 people, but it doesn't seem to work that well when the virus has already multiplied," the source told the Mainichi Shimbun.

I'm surprised that it works at all, but the data should speak for themselves, when we can see all of it. The good news is that the compound (an RNA polymerase inhibitor) seems to have a good safety profile.

Update: here's the paper itself. I note that it is open-label, 35 patients in the treatment arm, so like the study below, not enough to base a large decision on yet, other than one to do a bigger and more controlled trial. Second update: it's been pointed out (I missed this) that the control patients were enrolled in the last week of January, while the treatment group was enrolled during the first two weeks of February. That's not exactly randomization, but the effect of this on the data is up for debate). Here's the explanation for this, and I can see their point:

. . .given the high number of patients presenting simultaneously and the very high infectivity of the disease, it was ethically unacceptable to allocate patients to receive a different experimental drug using a randomization process impossible for most of the patients to understand. Furthermore, in the context of rumors and distrust of hospital isolation, using a randomized design at the outset might have led even more patients to refuse being isolated.

Next up is remdesivir. There are scattered reports from on-the-spot use in several countries, but we need better quality data on this one. Trials are underway from the NIH, from a WHO consortium, and from Gilead itself, among others, so we should have a clearer picture soon. Do not expect a cure. But it would be good if this compound can slow things down.

Finally, there are some potentially very interesting results from France on hydroxychloroquine. That compound (and chloroquine itself) have been the subject of much interest, and these are the first trial data that I've seen. A number of things need to be said up front: first of all, this was a small trial. Second, it was open-label. Third, there were significant patient drop-outs in the treatment group, making the sample even smaller. Under normal circumstances, to be honest, I would be looking askance at this, but (1) these ain't normal circumstances and (2) the effect size seen in this work may be significant.

In summary, 26 patients were enrolled in the treatment group, with 16 controls. Six patients dropped out of the treatment group: 3 went to the ICU, one dropped out due to nausea, one left the hospital (apparently recovered?) and one died. No one left the control group. There were 15 male and 21 female patients. 6 of them were asymptomatic, 22 had upper respiratory symptoms, and 8 had lower respiratory tract symptoms (all of those had confirmed pneumonia by imaging).

The treatment group got 200mg of hydroxychloroquine sulfate three times a day, and six of those patients were also given 500mg azithromycin in addition. The paper says that this was the deal with possible bacterial superinfection, and the lead author also makes mention of possible antiviral effects of the compound. I hadn't heard of these - azithromycin is, of course, more famous as an antibacterial - but there seems to be a pretty established literature on this, although the mechanism doesn't seem to be well worked out.

The results are shown at right. As you can see, there appears to be an effect of hydroxychloroquine (although I would like to see some error bars), and a notably stronger effect (down to zero virus as measured by nasal swab) of the hydroxychloroquine - azithromycin combination. That's the result that's getting the attention, and justifiably so.

I would expect this to start some larger trials, and that looks completely justified. This by itself is not enough to recommend that people start using this combination - again, it's a very small trial and open-label at that. But it points the way to something larger and more controlled. These are two inexpensive generic drugs with a long history of use in humans; if they can be repurposed in this manner we need to know as soon as possible. Chloroquine and hydroxychloroquine both can have notable side effects, but this is not a long course of treatment, either. Let's see if this is real!