Adults with type 2 diabetes who took metformin before catching COVID-19 had a 39% lower risk of dying within a year than those on another drug, a Stanford-led study published in Diabetes, Obesity and Metabolism reports.
“This is not enough evidence to change treatment,” said Ziyad Al-Aly, M.D., a senior clinical epidemiologist at Washington University in St. Louis who was not involved in the study.
The comparison group was older and sicker, and the authors say the records-based work cannot prove that metformin caused the difference, according to Fox News.
How the Stanford Team Compared Two Drugs
The researchers analyzed U.S. medical records for 53,332 adults aged 30 to 85 with type 2 diabetes, CIDRAP reported. All were diagnosed with COVID-19 between October 2021 and November 2023. Each had been prescribed either metformin or a dipeptidyl peptidase-4 inhibitor (DPP4i), another diabetes drug, at least 30 days earlier.
Of the patients, 50,965 were taking metformin and 2,367 a DPP4i. Metformin users averaged 60 years old; DPP4i users averaged 64. CIDRAP adds that women made up a majority of the metformin group and of the DPP4i group.
After adjusting for confounding factors, metformin use was associated with a 39% lower 12-month mortality risk (adjusted risk ratio 0.61), CIDRAP said. An adjusted risk ratio compares the risk in two groups after accounting for other differences between them.
On long COVID, Fox News and Mezha both report that an initial 13% drop among metformin users disappeared under further statistical testing. There was no statistically significant link.
Why the Two Drug Groups Are Not a Real Choice
Al-Aly told Fox News Digital that metformin is the first drug most people with type 2 diabetes receive. DPP-4 inhibitors usually come later, or go to people who cannot take metformin, often because they are older or have kidney disease or other health problems.
“People take a DPP-4 inhibitor because they need it to control their diabetes, usually because metformin was not enough or was not an option for them,” Al-Aly said.
He added that no doctor chooses between the two drugs for the same patient. The study’s comparison is not a real choice in any clinic. His advice: if you take either drug, do not change it because of this study.
Authors Name the Limits of Their Work
Patients on DPP4i tended to be older and to have more underlying conditions, which could have raised their risk of death, the researchers said. They also lacked complete COVID-19 vaccination records.
The authors suggest several possible mechanisms. Metformin lowers blood sugar, and the Centers for Disease Control and Prevention says high blood sugar can weaken immune defenses. The drug may also help regulate a pathway involved in blood pressure and fluid balance that affects COVID-19 severity, and earlier studies indicate it can reduce inflammation and viral levels during infection. Fox News reports it is unclear whether blood-sugar effects explain the findings.
A Separate Study Tests Prescriptions Within Six Days
A different analysis, in the Journal of Medical Virology, looked at people without diabetes or prediabetes, News-Medical reported on November 10. It used electronic health records to study patients newly prescribed metformin within six days of documented infection.
The study included 248 patients in the metformin group and a weighted comparison group equivalent to about 248. Over six months, about 4% of the metformin group and 8.5% of the comparison group developed long COVID or died. That is an absolute difference of about 4.5 percentage points. Adjusted for other factors, it corresponds to a 53% lower relative risk.
Fox News and Mezha cite a University of Minnesota Medical School study with a similar finding. The Stanford authors say their work is consistent with it.
The Stanford authors conclude: “These findings provide modest support for prescribing metformin for individuals with [type 2 diabetes] during acute COVID-19 to reduce post-COVID-19 mortality, although further investigation is needed to evaluate its impact on Long COVID and assess causality.”