Yes — in this metformin long covid study 2026, a new metformin prescription given within days of a documented SARS-CoV-2 infection was associated with a 53% lower risk of developing long COVID or dying over the following months, compared with people given certain other medicines in the same early window. That is a striking headline, and it is also a headline that needs careful reading, because the research was an observational emulation based on health records, not a randomised trial.
So what should an ordinary reader take from it? The short version is hopeful but restrained: early metformin looked protective in this particular group of patients, the pattern lines up with evidence from elsewhere, and yet nobody should treat this as a reason to seek out or start metformin on their own. The longer version — the one that actually matters — is about who was studied, how the comparison was made, what the numbers really mean, and where the gaps still are. That is what the rest of this article walks through, in plain language, without hype and without medical advice.
What the Metformin Long COVID Study 2026 Actually Found
The headline finding is simple to state and important not to overstate. People who received a new prescription for metformin within six days of a documented SARS-CoV-2 infection had a substantially lower combined risk of long COVID or death over six months than a matched group of people who received a different prescription in that same window. The size of the difference, expressed in relative terms, was a 53% reduction.
Relative reductions can sound abstract, so it helps to look at the absolute figures alongside them. Across the whole study, the weighted incidence of long COVID or death was 6.3%. In the metformin group it was 4.0%, while in the comparison group it was 8.5%. The gap between those two figures — an absolute risk reduction of 4.5 percentage points — is the more grounded way to picture what happened. Out of every hundred people in a situation like the comparison group, roughly eight or nine went on to develop long COVID or die within six months; in the metformin group, that figure was roughly four in every hundred. Both ways of expressing the result describe the same pattern. The relative figure tells you how much lower the risk was in proportional terms; the absolute figure tells you how many fewer people, in a hundred, were affected.
The study was published in the Journal of Medical Virology, and it was covered by MedicalXpress on 6 November 2026. It was led by researchers from the University of Minnesota Medical School, with Carolyn T. Bramante as lead author. Those details matter, not because a journal name settles an argument on its own, but because they tell you where the work can be checked, challenged, and built upon. A published study with a named team, a clear method, and a public record invites exactly the kind of scrutiny that good science depends on.
How the Metformin Long COVID Study 2026 Was Done
Understanding the method is the key to understanding both the strength and the limits of this finding. This was a target trial emulation using electronic health records. That phrase sounds technical, but the idea behind it is straightforward and rather elegant. A randomised trial is the gold standard for testing whether a treatment works, because random assignment balances out the differences between groups — age, other illnesses, severity of infection, and all the things researchers may not even have thought to measure. But randomised trials take time, money, and willing participants, and they cannot always be run quickly when a new question emerges. A target trial emulation tries to answer the same kind of question using records that already exist, by designing the analysis as carefully as if it were a trial, and then applying that design to real-world health data.
In this case, the researchers identified people with a documented SARS-CoV-2 infection and then looked at what prescription they received in the days immediately afterwards. One group — 248 people — received a new prescription for metformin within six days of that documented infection. The comparison group — also 248 people — received a prescription in the same window for one of four other medicines: fluvoxamine, fluticasone, ivermectin, or montelukast. The researchers then followed both groups and compared how often long COVID or death occurred over six months.
And that word, association, is doing a lot of work. An observational study, even a carefully emulated one, can show that two things travelled together. It cannot, on its own, prove that one caused the other. People who were prescribed metformin may have differed from people prescribed the comparison medicines in ways the records did not fully capture — in the severity of their infection, in their underlying health, in the care they received afterwards, or in things that never made it into a coded record at all. The researchers used weighting and careful design to balance what they could measure. What they could not measure remains a source of uncertainty, and honest reporting has to say so.
Who was in the study — and who was not
The sample was primarily made up of people infected with the omicron variant. About half of the metformin group and 52% of controls had the omicron variant, which places the study firmly in the more recent phase of the pandemic, rather than in the earliest waves when the virus, the population's immunity, and the available treatments were all very different. That context matters for anyone wondering whether the finding applies to them. A result found mainly in omicron infections is, first and foremost, a result about omicron infections. Whether the same pattern would hold in other variants, in vaccinated and unvaccinated people in different proportions, or in people with very different health backgrounds, is a question for further research rather than something this study alone can settle.
What Is Long COVID, and Why Does Early Treatment Matter?
Before going further, it is worth being clear about what long COVID actually means, because the term covers a wide range of experiences. In general, it refers to symptoms that persist or appear after the acute phase of a COVID infection has passed — symptoms that cannot be explained by another diagnosis and that interfere with daily life. The most commonly described include overwhelming fatigue that does not lift with rest, difficulty concentrating or remembering things — often called brain fog — breathlessness on exertion, chest discomfort, a racing or irregular heartbeat, disturbed sleep, headaches, muscle and joint pain, and a general sense that the body has not returned to its previous baseline.
Why Researchers Looked at Metformin in the First Place
Metformin is a prescription medicine used for prediabetes and type 2 diabetes, and it has been prescribed for decades. Doctors know its safety profile well, they know who should not take it, and they know how to monitor people who do. That long history is part of what makes it an attractive candidate for repurposing: if a medicine is already in common use, researchers do not have to start from scratch in understanding how it behaves in the human body.
But familiarity alone would not be enough. Researchers were interested in metformin and long covid for a more specific reason: earlier signals had already suggested that it might have a role in COVID. The study's authors note that their finding is consistent with a UK sequential target trial analysis, in which metformin started within 90 days of infection was associated with a hazard ratio of 0.36. In plain terms, that UK analysis pointed in the same direction — towards a lower risk in people who received metformin — using a different dataset, a different health system, and a longer window for starting the medicine. When two independent analyses, conducted in different places and with different designs, point the same way, it strengthens the case for taking the idea seriously, even while it falls short of proof.
What the Numbers Mean in Plain Language
Numbers in medical studies are often reported in ways that make a result sound larger or smaller than it really is, so it is worth taking a moment to translate this one carefully. The study reports a 53% reduction in the combined outcome of long COVID or death. That is a relative risk reduction: it compares the risk in the metformin group with the risk in the comparison group, and expresses the difference as a proportion of the comparison group's risk.
The absolute figures give the same story a more concrete shape, and they are worth keeping side by side with the headline. The table below sets out the key figures as reported, so that readers can see them together rather than relying on the headline alone.
| Measure | Result |
|---|---|
| People in the metformin group | 248 |
| People in the comparison group | 248 |
| Weighted incidence of long COVID or death, overall | 6.3% |
| Weighted incidence in the metformin group | 4.0% |
| Weighted incidence in the comparison group | 8.5% |
| Absolute risk reduction | 4.5 percentage points |
| Relative reduction in long COVID or death | 53% |
| Prescription window after documented infection | Within 6 days |
| Follow-up period | 6 months |
Metformin and Long COVID: What This Does Not Prove
It would be easy, and wrong, to read this study as proof that metformin prevents long COVID. It does not prove that, and the distinction between association and causation is not a pedantic footnote here — it is the heart of the matter. The researchers did something careful and valuable: they emulated a trial as closely as records allow, they balanced the groups on the factors they could measure, and they found a clear association. But records, however detailed, cannot capture everything. They cannot fully capture how unwell someone felt, how a clinician weighed competing options in the moment, what other care a patient received, or the many small decisions that shape recovery and are never written down in a coded field.
For readers who follow health news closely, this pattern will feel familiar. A promising observational finding appears, it is discussed widely, and then the harder work begins: larger studies, randomised trials, and careful replication. Sometimes the early promise holds up. Sometimes it shrinks, or disappears, or turns out to apply only to a subgroup. That is not a failure of science; it is science working as it should. The responsible way to respond to a study like this is with interest, with patience, and with a clear-eyed view of what would be needed to move from "associated with" to "shown to cause."
The UK Evidence That Points in the Same Direction
One of the more reassuring features of this research is that it does not stand alone. The authors note consistency with a UK sequential target trial analysis, in which metformin started within 90 days of infection was associated with a hazard ratio of 0.36. A hazard ratio below one points towards a lower risk in the group receiving the medicine, and a figure of that size points towards a substantially lower risk in that analysis.
What to Ask Your Doctor — and What Not to Do
The most important practical message of this article is also the simplest: do not start metformin, or any prescription medicine, on the basis of a news story or a study summary. Metformin is a prescription medicine for prediabetes and type 2 diabetes. It is not a supplement, not an over-the-counter remedy, and not something to borrow from a family member's medicine cabinet. It has known contraindications, it interacts with other medicines and conditions, and it requires a clinician who knows a patient's full history to decide whether it is appropriate — and, if so, how it should be used and monitored.
What readers can do, quite reasonably, is talk to their own doctor — not to demand a prescription, but to ask informed questions. If you test positive, or if you are worried about long COVID, it is entirely reasonable to ask what the current guidance says about early care, what follow-up is available if symptoms persist, and what the evidence does and does not support. A good conversation with a clinician is not about quoting a single study; it is about placing that study in the context of everything else that is known about a person's health, and making a decision together.
Related reading on View Storys may help put this kind of health research in context. A recent report on a new triglyceride-lowering drug trial and what its results mean for heart health shows how trial findings are interpreted step by step, while coverage of hormone replacement therapy and heart risk in recent research is a useful reminder that even familiar treatments need careful, long-term evidence. Readers interested in how large trials report their results can also look at the Horizon trial results for pelacarsen and lipoprotein(a), which illustrates the same careful distinction between a promising signal and a proven benefit.
Frequently Asked Questions
Does this study prove that metformin prevents long COVID?
No. The study found an association between a new early metformin prescription and a lower combined risk of long COVID or death over six months. Because it was an observational target trial emulation using health records, rather than a randomised trial, it cannot prove causation. Larger randomised trials are needed before anyone can say that metformin itself caused the lower risk.
Should I ask my doctor for metformin if I test positive for COVID?
You should talk to your own doctor about your individual situation, and never self-medicate. Metformin is a prescription medicine for prediabetes and type 2 diabetes, with known contraindications and interactions. Only a clinician who knows your medical history can judge whether any medicine is appropriate for you, and this study alone is not a basis for starting it.
What is a target trial emulation?
It is a method that designs an observational analysis as carefully as if it were a randomised trial, and then applies that design to existing health records. It is a serious and valuable approach, especially when a randomised trial has not yet been run, but it cannot balance unmeasured differences between groups in the way that random assignment can.
Why does the study focus on the first six days after infection?
Because the researchers were asking whether starting metformin early, in the acute phase of infection, was associated with a better outcome months later. The six-day window is central to the question they tested, and the findings speak to that early period rather than to starting treatment much later or to treating long COVID that is already established.
The Bottom Line
This research gives us a genuinely interesting lead, and it gives us a clear sense of what would be needed to turn that lead into something doctors could act on. A new metformin prescription within six days of a documented SARS-CoV-2 infection was associated with a 53% lower risk of long COVID or death over six months, with a weighted incidence of 4.0% in the metformin group against 8.5% in the comparison group, and an absolute risk reduction of 4.5 percentage points. The pattern is consistent with a UK analysis that pointed in the same direction, and the study was conducted in a primarily omicron-infected sample, with 248 people in each group.
Those are real findings, from a real study, published in the Journal of Medical Virology and led by researchers from the University of Minnesota Medical School. They deserve attention. They do not yet deserve to change how anyone is treated outside of a trial or a clinician's careful judgement. The difference between those two statements is the difference between good science reporting and hype — and it is a difference worth holding on to.
For the full study record, see the published paper in the Journal of Medical Virology, and for an accessible summary of the findings, see the MedicalXpress report from 6 November 2026. Both are worth reading alongside this article, because they let you check the details for yourself — and checking the details, rather than taking any single summary on trust, is exactly how good health decisions begin.