Can taking metformin early for COVID-19 lower the odds of long COVID?
In this large NIH-funded trial, a 14-day course of metformin started within 7 days of COVID-19 symptom onset reduced clinician-diagnosed long COVID, with a risk ratio of 0.495 (95% credible interval 0.155 to 0.995). On the trial’s main measure, self-reported symptoms attributed to COVID-19 at day 180, metformin did not meet the study’s efficacy threshold.
Metformin is an old, cheap, and widely used diabetes pill. Researchers wanted to know if it could do something new: prevent the lingering illness many people develop after COVID-19. This study, called ACTIV-6, tested that idea in a careful, controlled way. The results are mixed, and the details deserve a closer look.
What the data show
The trial enrolled 2,983 outpatient adults aged 30 and older with confirmed SARS-CoV-2 infection and at least two COVID-19 symptoms, enrolled within 7 days of symptom onset. Participants were randomized to metformin or placebo for 14 days. The median age was 47 years, 63% were female, and 47% were Hispanic or Latino.
The primary outcome was post-acute sequelae of SARS-CoV-2 or death, measured by asking participants at day 180 whether they had symptoms they attributed to COVID-19. Overall, 96 people (3.2%) reported COVID-19 symptoms on day 90, 101 (3.4%) on day 120, and 79 (2.6%) on day 180. The adjusted risk of symptoms at day 180 was 0.8 percentage points lower with metformin (95% credible interval -2.2 to 0.6), a risk ratio of 0.79 (95% credible interval 0.474 to 1.230), with a posterior probability of efficacy of 0.83. That fell short of the trial’s prespecified efficacy threshold of 0.975.
The secondary outcome of clinician-diagnosed long COVID told a different story. Compared with placebo, the risk was 0.7 percentage points lower with metformin (95% credible interval -1.5 to 0.1), with a posterior probability of efficacy of 0.96 and a risk ratio of 0.495 (95% credible interval 0.155 to 0.995). There were no deaths in the trial.
So the signal was for a doctor’s formal diagnosis of long COVID, not for the broader self-reported symptom measure the trial was primarily designed to test.
Dr. Kumar’s Take
The practical appeal here is obvious. Metformin has been around for decades, costs very little, and its safety profile is well understood. A familiar, inexpensive pill that lowers the odds of a long COVID diagnosis would help a lot of people if the finding holds up.
I also weigh the fact that 83% of participants reported at least one prior COVID-19 infection or at least two SARS-CoV-2 vaccines. That mirrors the real world today, so the results feel relevant to the patients I actually see. Still, I want to be honest about the limits. The primary outcome did not meet the trial’s own efficacy threshold, and the benefit rests on a secondary outcome. This is one trial, not the final word, and I would want the findings repeated before metformin becomes a routine recommendation after infection.
Study Snapshot
ACTIV-6 was designed to be rigorous. It was quadruple-blinded, randomized, and placebo-controlled. That design helps prevent expectation from coloring the results.
The trial ran across 90 sites and enrolled outpatients with mild to moderate COVID-19 between September 19, 2023 and May 1, 2024. By treating people within a week of symptom onset, the researchers tested metformin during the early phase of infection.
Who might benefit most
This trial points to adults aged 30 and older in the first days of a COVID-19 infection with mild to moderate illness. Because treatment began within 7 days of symptom onset, the finding applies to acting early rather than waiting. The population was low risk, and most already had immunity from prior infection or vaccination.
Practical Takeaways
- If you catch COVID-19, talk to your doctor early about your options, since this trial tested treatment started within the first 7 days of symptoms.
- Do not start metformin on your own. It is a prescription medicine, and this trial did not meet its own efficacy threshold on the primary outcome.
- Watch for lingering symptoms after an infection and bring them to your doctor, since early attention to long COVID can guide better care.
Related Studies and Research
- Normobaric oxygen treatment for mild-to-moderate depression: RCT
- Long-term oxygen treatment trial (LOTT) and health benefits
- Optimizing cardiopulmonary rehabilitation duration for long COVID
- Oxygen therapy during ADL rehabilitation in severe COVID-19
FAQs
How does metformin work to lower long COVID risk?
This trial measured outcomes, not mechanism. It was designed to compare 14 days of metformin against placebo, started within 7 days of symptom onset, and to track symptoms out to day 180. Understanding how any protection might work would require different lab and clinical studies.
Does this mean everyone with COVID-19 should take metformin?
No. The trial’s primary outcome, self-reported symptoms attributed to COVID-19 at day 180, did not meet the prespecified efficacy threshold of 0.975. The reduction in clinician-diagnosed long COVID was a secondary outcome. Medical guidelines usually change only after findings are repeated in more studies. For now, discuss your personal risk and options with your doctor rather than seeking the drug on your own.
Bottom Line
In this NIH trial of 2,983 outpatient adults, 14 days of metformin started within 7 days of COVID-19 symptom onset reduced clinician-diagnosed long COVID, risk ratio 0.495 (95% credible interval 0.155 to 0.995). On the primary outcome, symptoms attributed to COVID-19 at day 180, metformin did not exceed the trial’s efficacy threshold of 0.975. That makes this a hopeful signal rather than a verdict. Because most participants already had prior immunity, the results are relevant to today’s population, and they make a case for more research into this cheap, familiar pill.
