Purpose: Metformin, a first-line antidiabetic drug in type 2 diabetes, is being evaluated for prevention and treatment of various solid tumors. Metformin has also been shown to act against lymphoid malignancies (LM) in preclinical models, although population-based data are lacking. Thus, we examined whether metformin use is associated with reduced risk of LM in patients with type 2 diabetes.

Patients and Methods: This was a population-based cohort study using data from the United Kingdom Clinical Practice Research Datalink. We included patients > 18 years of age newly-treated with non-insulin antidiabetic drugs between 1998 and 2014. Time-dependent Cox proportional hazards models were used to estimate hazard ratios (HR) with 95% confidence intervals (CI) of LM (inpatient diagnoses of non-Hodgkin and Hodgkin lymphoma, chronic and acute lymphocytic leukemia, and multiple myeloma) comparing the use of metformin with the use of other antidiabetic drugs, and lagged by one year for cancer latency. Secondary analyses included stratifications of metformin use by cumulative duration and dose, as well as of LM by disease type. All estimates were adjusted to ethnicity, socioeconomic status, time-updated use of other antidiabetic drugs, known risk factors (autoimmune diseases and solid organ transplant), and important comorbidities.

Results: The cohort included a total of 100,091 patients, which generated 502 LM events during 500,774 person-years of follow-up (rate, 1/1000 person-years). Overall, metformin use was not associated with the incidence of LM (adjusted HR, 0.86; 95% CI, 0.63-1.16). Furthermore, there was no significant trend for duration (P=0.46) or cumulative dose (P= 0.60) of metformin use with LM incidence, with the highest categories being > 5 years and > 2,492 grams (Table 1). When stratified by disease type, metformin use was not associated with the incidence of multiple myeloma, chronic lymphocytic leukemia acute, and non-Hodgkin lymphoma (Table 2). Similar findings were seen for acute lymphoblastic leukemia (adjusted HR, 0.36; 95% CI, 0.04-3.41), and Hodgkin lymphoma (adjusted HR, 0.94; 95% CI, 0.23-3.91).

Conclusions To our knowledge, this is the first population-based study to examine the association between metformin use and LM incidence. We observed no duration- or dose-response of metformin with LM incidence, which are paramount in the concept of chemoprevention. There were no important differences between LM types, although precision for acute lymphoblastic leukemia and Hodgkin lymphoma was limited due to sample size. While our findings are observational and apply to the diabetic population, our findings provide no rationale for the prospective investigation of metformin to prevent LM.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.