Metformin use was not associated with any improved outcomes among patients with newly diagnosed diffuse large B-cell lymphoma (DLBCL) or follicular lymphoma, according to a single-center study out of the Mayo Clinic.
In fact, patients with diabetes mellitus who were using metformin as treatment had inferior overall survival compared with patients with no diabetes and no metformin use (hazard ratio [HR], 2.17; 95% CI, 1.19–3.95).
These results are in contrast to previous findings that showed that metformin had anti-lymphoma activity.
“The in vivo potency of metformin in blocking mTOR and the clinical anti-lymphoma activity may be limited,” wrote Yucai Wang, MD, PhD, of the Mayo Clinic in Rochester, Minnesota, and colleagues, in the British Journal of Hematology.
Another possible explanation for the lack of activity could be that patients with diabetes had worse prognosis due to other comorbidities or complications, limiting any possible benefit of metformin.
The study looked at 869 patients with DLBCL and 895 patients with follicular lymphoma newly diagnosed with disease and enrolled in the Mayo component of the Molecular Epidemiology Resource cohort study between 2002 and 2015. Information on diabetes and metformin use was collected from electronic medical records.
Overall, compared with patients with no diabetes or metformin use, there was no association with diabetes/metformin use (HR, 1.16) or diabetes/no metformin use (HR, 1.16) with event-free survival. Similarly, there were no associations found for lymphoma-specific survival.
Among only those patients with DLBCL, patients with diabetes and no metformin use and those with diabetes and metformin use both had inferior event-free survival compared with no diabetes and no metformin use (median, 72.5 vs 63.8 vs 127.3 months). Patients with diabetes with and without metformin use also had inferior overall survival compared with no diabetes and no metformin use (113.7 vs 90.0 vs 165.0 months).
Among only those patients with follicular lymphoma, there was no significant difference in event-free survival between the three patient groups; however, overall survival was inferior for diabetes without metformin use and diabetes with metformin use compared with no diabetes or metformin use (median, not reached vs 139.5 months vs not reached).
“Metformin did not demonstrate ‘single agent activity’ in follicular lymphoma patients not receiving other active therapy, suggesting that its therapeutic role is probably limited,” the researchers wrote. “However, these patients did not require treatment initially and were expected to have an excellent outcome, so any potential benefit of metformin may not be apparent.”
The researchers noted that large randomized clinical trials could help to better define the role of metformin in DLBCL and follicular lymphoma, but added that “such studies are unlikely to be conducted considering small if any potential benefit of metformin and competition from other more promising agents targeting similar pathways.”
Commenting on the study, Kenneth Carson, MD, PhD, of Washington University in St. Louis, said the results were interesting and add to the body of evidence available to understand the potential effect of metformin in patients with the two most common types of non-Hodgkin lymphoma.
"While the authors did caution about over-interpretation of the negative results due to small patient numbers, those who read the study need to remember that absence of evidence is not the same as evidence of absence," Carson said. "There were only 48 patients who received metformin in the DLBCL group and 37 in the follicular lymphoma group. Thus, roughly 5% of the patients in the cohort received the exposure of interest. With such small numbers, even potentially large outcome differences are not statistically significant. The best evidence of this is the hazard ratio of 1.41 for event-free survival in patients with diabetes who didn't receive metformin (compared to those without diabetes). The hazard ratio for those who received metformin was only 1.05. If these numbers held in a larger study, it would be suggestive of what is potentially an important signal in patients with diabetes. Would this actually happen in another, larger population? Who knows! But it highlights the limitations imposed by the small number of patients exposed to metformin."
Carson noted other concerns with the study, such as the lower than expected prevalence of diabetes in the study population, which was less than 10% compared with 25% in those older than 65 in the United States; he added that DLBCL occurs disproportionately in the elderly, making this rate surprising.
"I think the most important point is that it is hard to draw conclusions of 'no effect' in a cohort study when the exposed subgroup is so small," Carson said.