Does taking Metformin during a COVID infection reduce the incidence of long covid

The short version: one good trial suggests metformin cuts long COVID risk by about 40%, but the evidence base is thin and funded by conflicted hands.

The best study cited here is a 2023 trial in the scoping review [6] that gave metformin around the time of acute infection and found fewer cases of long COVID months later. That is the right design: a randomized controlled trial with a real clinical endpoint, not antibodies. The same review notes that evidence for metformin treating already-established long COVID is insufficient [6]. So metformin appears to be a preventative, not a cure.

But look at whose hands the rest of the evidence comes from. The records that actually contain metformin data are mostly about diabetes and heart disease [3, 4]. One is a study of exercise timing in diabetics on metformin [3]. Another is a 2019 cohort study showing metformin use during a heart attack was linked to a higher risk of major cardiac events, driven by cardiovascular death [4]. That is not the same question, but it is a signal that metformin is not a harmless sugar pill in every context.

So what does the one good trial show? It claims a meaningful reduction. But the absolute risk numbers for long COVID in the general population are not in these records. Without them, you cannot know if the reduction is from 10 people per hundred to 6, or from 1 to 0.6. That matters. The scoping review itself says trials in established PASC patients are needed [6] — meaning the evidence for the specific use case (metformin during infection to prevent long COVID) rests on a single trial in a single review.

The rest of the retrieved evidence is noise. A case of bullous pemphigoid after a COVID vaccine [1] has nothing to do with metformin and long COVID. The diabetes preference study [10] asks Singaporeans which pill side effect they hate most. Neither answers your question.

For a healthy person weighing whether to start metformin at the first sign of COVID, the evidence amounts to one trial with a positive result, no replicated study, and no long-term safety data in that specific use. The drug is cheap and generally well-tolerated, but the burden of proof for an intervention on a healthy body is on the people recommending it. They have not met it with more than a single unconfirmed trial.

My call: metformin may reduce long COVID risk based on one trial, but the evidence is not strong enough to recommend it as standard practice. Confidence: low.