Top News

Colby Cosh: COVID and steroids — no 'magic bullet,' but a bullet nonetheless

This week the Journal of the American Medical Association published a meta-analysis of studies on the use of steroids in critically ill COVID-19 patients. The analysis was carried out by a group with the only slightly contrived acronym REACT: it’s the World Health Organization’s Rapid Evidence Appraisal for COVID-19 Therapies crew.

REACT, as both versions of its name suggest, has the job of monitoring the clinical evidence for COVID-19 treatments. It has given the world’s doctors the all-clear to use cheap, familiar steroids on patients in serious respiratory trouble, who can expect their odds of survival to improve by about a third.

Steroids have been a special problem in the COVID-19 fog of war since this whole business kicked off. Anybody who has been prescribed a steroid for any kind of inflammation, respiratory or otherwise, knows that they can feel a bit like wizardry. But there is no free lunch in medicine: steroids also suppress the underlying immune response, and they can allow a pathogen that would otherwise be driven off to linger in the tissues. They also come with weird, bad side effects.

In the early days of the pandemic, physicians had nothing much to go on besides the evidence for steroid use in COVID-19’s coronavirus precursors — namely, the original SARS, and MERS, the camel-borne Middle East respiratory syndrome that the world has mostly managed to contain since the first outbreak in 2012.

A Lancet review from February shows what researchers made of those rehearsals. Steroids didn’t seem to save lives, they delayed the clearance of the viruses and they created freaky complications including diabetes and avascular necrosis. Steroids also have an established poor reputation as a treatment for influenza-derived pneumonia, even though it is very natural to try using steroids to prevent lung injury or win major-league batting titles.

So the initial advice for COVID-19 was to leave the steroids aside for ordinary patients. In Lancet’s weighty words, “no unique reason exists to expect that patients with 2019-nCoV infection will benefit from corticosteroids, and they might be more likely to be harmed with such treatment.” But steroids were still included in research agendas, because of that anti-inflammatory magic, and because with a novel coronavirus you have a positive obligation to try everything.

In late July, a big research consortium in the United Kingdom called Recovery published findings from a large randomized trial of several leading candidates for COVID-19 treatments, including many of the ones that have become unexpected household names to news-reading laymen in 2020: hydrochloroquine, the antibiotic azithromycin, antivirals used in HIV treatment and even convalescent plasma.

The Recovery group hastened to publish in the New England Journal of Medicine because steroids had done better than anybody anticipated in patients with severe cases, and particularly those receiving mechanical ventilation. The Kaplan-Meier “survival curves” in the ventilation group were yoinked upward as if somebody had attached them to a crane, and that’s what doctors look for. (They’re mostly too busy to read, and when it comes to looking at studies, focusing on the graphs isn’t a bad thing. I wouldn’t recommend choosing a physician specifically for illiteracy, but you would be better off than you might imagine.)

Needless to say, this caused a problem (see “no free lunch,” supra). Other researchers who had been studying steroids and COVID-19 had been proceeding on an ethical assumption of “equipoise.” The direct evidence concerning the use of steroids didn’t exist for a disease whose existence was first announced on New Year’s Eve, and the indirect evidence from other illnesses wasn’t exactly decisive, so it was deemed ethical to hold randomized trials and withhold steroids from COVID-19 patients in the control groups.

But the Recovery paper had a pretty sizable and varied sample, and above all it was coming from the world capital of evidence-based medicine at Oxford. Equipoise was thrown out the window at once: the ethical fulcrum, so to speak, had shifted in favour of steroids.

That left REACT with existing data from seven strong international trials, including Recovery, that had each followed up serious ill patients (prospectively, and with randomization) for about a month. When the results were pooled, the numbers were almost unnervingly consistent, with odds ratios for mortality hovering around two-thirds.

Setting Recovery aside didn’t make any difference to the numbers, follow-up was fine all around (with only a handful of patients giving up the fight and withdrawing consent) and nobody picked up any adverse effects to speak of. It appears that there are important differences between SARS-CoV-2 and its cousins that make steroids more effective against 2020’s hip, happening new virus.

(Either that, or the previously existing consensus about steroids and the other diseases was just wrong, which is possible. Hopefully SARS-1 and MERS won’t rear their heads again and force us to re-open that question.)

As the lay press has been careful to observe, this isn’t a “magic bullet,” although for the last century or so, “magic bullets” have only been mentioned when someone is denying that they exist, and we all go on visiting doctors anyway. Moreover, steroids seem to be unhelpful in “mild” cases that are still pretty nightmarish.

But the validated knowledge that will help patients leave hospital at body temperature instead of room temperature is finally arriving. At last the evidence base is coming into focus, with details that are more than just very intelligent guesswork. In this particular case, after all, educated guessing turned out to be misleading.

National Post
Twitter.com/colbycosh

Copyright Postmedia Network Inc., 2020

Did this story inform or enhance your perspective on this subject?
1 being least likely, and 10 being most likely

Recent Stories