Metformin Linked to Reduced Death, Long COVID Post-COVID-19

Researchers from the University of Minnesota, led by Steven G. Johnson, Ph.D., have identified a potentially beneficial connection between metformin use and reduced long-term impacts of SARS-CoV-2 infection. Their findings were published in the journal Diabetes Care on September 17, 2024.

The study focuses on the effects of metformin, a medication commonly used to manage diabetes, on the incidence of death and post-acute sequelae of SARS-CoV-2 (PASC), often referred to as long COVID. The researchers conducted a retrospective cohort analysis utilizing two electronic health record databases, N3C and PCORnet, to compare outcomes in individuals taking metformin against those on other diabetes medications.

The team defined PASC using two distinct methods: first, by employing the U09.9 diagnostic code assigned within 180 days following the initial COVID-19 event (termed PASC-U09.9), and second, by applying a computable phenotype specific to each dataset (labeled PASC-N3C and PASC-PCORnet).

The analysis revealed that in the N3C database, the hazard ratio for death or PASC-U09.9 among those using metformin was 0.79, indicating a 21% reduction in risk compared to the comparator group. The confidence interval for this result was between 0.71 and 0.88. Similarly, the hazard ratio for death or PASC-N3C was 0.85, with a confidence interval ranging from 0.78 to 0.92, showing a 15% decrease in risk.

In the PCORnet database, the hazard ratio for death or PASC-U09.9 was found to be 0.87, though this result was not statistically significant, as the confidence interval spanned from 0.66 to 1.14. The hazard ratio for death or PASC-PCORnet was 1.04, with a confidence interval between 0.97 and 1.11, indicating no significant difference between metformin users and non-users in this dataset.

When examining the incidence of PASC using diagnostic codes, the rates were 1.6% for metformin users compared to 2.0% for the comparator group in the N3C database. In the PCORnet database, these rates were 2.1% and 2.5%, respectively. Using the computable phenotype method, the incidence rates in the N3C database were 4.8% for metformin users versus 5.2% for the comparator group. In the PCORnet database, the rates were notably higher at 24.7% for metformin users and 26.1% for the comparator group.

The authors of the study suggest that these findings indicate a modestly beneficial effect of metformin on long-term COVID-19 outcomes in individuals with diabetes. However, they also note the importance of considering the variation in results across different datasets.

The study acknowledges the potential conflicts of interest, with several authors disclosing financial ties to the pharmaceutical industry. Despite this, the research contributes valuable insights into the management of post-acute sequelae of SARS-CoV-2 infection, particularly in the context of diabetes treatment.