COVID-19 treatment may pass muster

On Saturday, the Montreal Heart Institute (MHI) announced promising results from its large trial of colchicine as a treatment for COVID-19. Colchicine is an extract of the autumn crocus, a plant whose anti-inflammatory effects have been exploited since antiquity. It is most famous as a treatment for gout, but it is somewhat dangerous and is only prescribed now for short-term, severe flare-ups. It emerged very early as a hypothetical contender in the race to repurpose old drugs for COVID-19, and if the MHI's press release can be trusted, it is by far the most cheap and convenient drug to pass muster in a serious clinical trial.

What I suppose has to be said first is: don't go sniffing around for a colchicine prescription. The drug is no picnic even when used in correct therapeutic amounts. Gastrointestinal side effects are very common. Its “therapeutic index,” which is a term for the distance between “amount that will heal you” and “amount that will seriously mess you up,” is narrow.

It is classified as an “extremely hazardous substance” in U.S. law. And nobody who has had gout will take its status as a gout medication as a sign that it is not a serious drug. (It's not so easy to think of a disease that more closely resembles literal torture.)

With all that said, colchicine is a drug that is sent home with patients, all the time, in pill form: if it has utility in treating COVID-19 upon the early emergence of symptoms, this feature may make it uniquely valuable. So it is all right to get a little excited about a press release.

What we have here is a case study in some of the unusual features of medicine in a time of pandemic — or even, perhaps, a study in usual features of medicine that the pandemic has intensified. The Montreal Heart Institute's release says that in “patients with a proven diagnosis of COVID-19, colchicine reduced hospitalizations by 25 per cent, the need for mechanical ventilation by 50 per cent and deaths by 44 per cent.”

These are obviously exciting numbers, but there is no peer-reviewed research paper available for anyone to follow up on the details yet. Even the release's advice to go to the website of the ColCorona study for more information will lead to disappointment, unless you like reading National Post clippings from almost a year ago, which would be perfectly understandable.

Hype is always part of the scientific process, for better or worse, and if you have discovered a universally available cheap drug that cuts COVID mortality, you arguably have a positive ethical obligation to get the early word out. But science is supposed to treat negative findings with the same respect it gives to positive ones. Researchers who have ended up debunking promising COVID treatments are working just as hard and benefiting humanity just as much as the ones who take a wild guess and have it pan out — as the MHI may have done in this case.

There is one eyebrow-raising moment in the press release. Those fantastic numbers I quoted above apply to only a subset of the population in the study. MHI's results for all 4,488 of its registered patients, we are told, “approached statistical significance.” That “approached” should have you making the gravelly Marge Simpson disapproval noise.

When MHI analyzed only the 4,159 patients who had PCR-confirmed COVID tests, its positive results, mirabile dictu, attained actual statistical significance. Looking more closely at the people whose infections we are more confident in will seem reasonable to the casual reader, but this chopping of the population into subgroups was not anticipated in the pre-registered prospectus for ColCorona.

Once the data are published properly (ideally with maximum sharing of the data), MHI's spin-doctoring will matter less. Moreover, the institute may have data from testing of its patients that helps to account for the mechanism of action of its miracle drug. “Statistical significance” is not, in itself, a bar to be cleared in a high-jump competition. But if MHI is going to say things like “approached statistical significance” in its press materials, the actual research should get an extra helping of professional-grade skeptical scrutiny.

ALL RIGHT TO GET A LITTLE EXCITED ABOUT A PRESS RELEASE.