Meta-Analysis: COVID19 drug treatment
Alright, now we are talking! I knew this had to be coming out soon... A meta-analysis on COVID19 drug treatments (yes, including hydroxychloroquine).
What is a meta-analysis? One massive study that combines the results of ALL previous studies. This is a really powerful tool because it takes into account whether past studies were "strong" or "weak". It also takes into account different populations (think different genes, different environments, different cultures, different confounders). Basically, this allows us to get an overarching idea of what is working and what is not working.
Why aren't more meta-analyses done? It takes a LOT of time to do this. Typically years (sorry graduate students). We also need "enough" studies on one topic so we can combine them.
Fortunately, with the onslaught of COVID19 publications, we already have enough COVID19 drug treatment studies that we can do a meta-analysis.
This was published over the weekend. These scientists combined ALL randomized control studies (the gold standard) on COVID19 drug treatments. They looked at which drugs impact mortality, ventilation use, and symptoms by combining 23 studies.
What did they find?
Glucocorticoids was the ONLY intervention that reduced mortality and the need for mechanical ventilation
The effects of hydroxychloroquine, remdesivir, and lopinavir-ritonavir is highly uncertain because studies were small, have serious imprecision, and have concerning limitations (like lack of double-blinding)
Considering this, though, authors concluded…
Hydroxychloroquine, remdesivir, and lopinavir-ritonavir MIGHT reduce symptom duration
Hydroxychloroquine might INCREASE the risk of adverse events when coupled with other interventions
Remdesivir might NOT increase the risk of adverse events
Translation? We need more studies. Specifically ones with more than 100 patients. Not very exciting, I know, but a stark reality of science. The science is not strong enough to say that, as a population, we can use anything but glucocorticoids to improve COVID19.
Next Steps? The even cooler thing about this publication is that the scientists created a "living" meta-analysis. I've never seen this before. As new studies come in, this analysis will be automatically adjusted each time. Giving us a better and better picture of COVID19 drug treatments in real-time. 9 studies are already in the queue to be included in the next update. Check out the study link for their visualization.
Love, your local epidemiologist
Data source: Table and meta-analysis conducted by the brilliant: Siemieniuk et al. (2020). Drug Treatments for Covid-19: Living Systematic Review and Network Meta-Analysis. BMJ.