Metformin Is Not Associated with Incidence Risk of Non-Hodgkin Lymphomas among Diabetic Patients

Epidemiological studies have consistently shown associations between diabetes mellitus (DM) and an increased risk of some cancers (e.g., liver, pancreas, endometrium, colon and rectum, breast, bladder; ref. 1). Potential preventative effects of antidiabetic drugs, such as metformin, on cancer have thus been hypothesized. Epidemiological evidence exists on the association between metformin and decreased incidence risk of overall and specific cancers (e.g., pancreatic and hepatocellular cancers) among diabetic patients (2).

Immunosuppression, chronic inflammation, and lymphocyte dysfunction occurring in DM patients have been associated with the development of lymphoma, indicating a plausible link between DM and non-Hodgkin lymphoma (NHL; ref. 3). Several epidemiological studies have found that DM was associated with increased NHL incidence risk (3, 4). We have not found studies on the association between metformin and NHL. We aimed to study the association between metformin use and NHL risk in Manitoba, Canada.

We conducted a population-based nested case–control study using cancer registry and health service administrative databases in Manitoba, Canada. The methods have been described in detail previously. Cases and controls were included if they were 40 years or older and had been diagnosed with diabetes (based on ICD-8/9 code 250.x and ICD-10 code E11-E14). Participants must have been registered with Manitoba Health for no less than five years before the index date to be included in the study, to ensure that the histories of exposure, if any, are of adequate length. NHL cases were identified from the Manitoba Cancer Registry. For each case, five controls are matched on gender, age, region of residence, date of diabetes diagnosis, and length of registration with the provincial health insurance, using the risk set sampling approach. Case and control records were linked to prescription drug, hospital discharge, and physician claim databases. Medication use histories were ascertained for the period between the index date and April 1, 1995, or the coverage initiation date, whichever was later. Metformin use during the 1-year period before the index date was excluded to minimize protopathic bias—a drug can be used to target early disease symptoms.

We used conditional logistic regression models to calculate odds ratios (ORs) and 95% confidence intervals (CIs). The sample size ensured a greater than 80% statistical power to detect an OR between 0.6 and 1.5. We assessed potential confounders using the change-in-estimation method and clinical relevance. We adjusted for chronic cardiovascular diseases, use of statin and non-statin lipid-lowering drugs, aspirin, non-aspirin anti-steroidal anti-inflammatory drugs (NSAIDs), income, and number of physician visits 5 years before diagnosis in the final analysis models. Separate analyses were undertaken for Total NHL and the main NHL subtypes including chronic lymphocytic leukemia/small lymphocytic leukemia (CLL/SLL), diffuse large B-cell lymphoma (DLBCL), plasma cell neoplasms (PCN), follicular lymphoma (FL), and other NHLs. All data analysis was performed in Stata 14 (StataCorp LLC).

In total 878 diabetic NHL cases were included. Cases and controls were similar in terms of demographics, medical condition and medication use histories, but cases tended to have more physician visits during the past 5 years before the index date (Table 1). About half of diabetes patients (cases and controls) were prescribed metformin. Approximately 16% and 49% of patients used insulin and other oral hypoglycemic drugs, respectively. As shown in Table 2, metformin use was not associated with incidence risk of Total NHL (OR, 0.93; 95% CI, 0.79–1.10). There were also no associations between metformin use and incidence risk of NHL subtypes.

We did not find an association between metformin use and NHL incidence risk among diabetic patients. A recent study reported an inverse association between metformin use and NHL risk among females, with an OR of 0.76 (95% CI, 0.64–0.91; ref. 5). Unlike the present study, which is conducted among diabetic patients only, the recent study included all types of patients. Caution in interpreting the finding from the recent study is also warranted due to the lack of clarity for epidemiological methodology and the simultaneous testing of multiple hypotheses in the study.

Epidemiological studies demonstrated that diabetic patients treated with metformin had lower incidence or mortality risk for several cancers, including colorectal, pancreatic, and hepatocellular cancers (6). Metformin has been associated with a reduced risk of progression from monoclonal gammopathy of undetermined significance to multiple myeloma (7). Metformin's anti-cancer mechanisms are unclear, but may involve multiple pathways, including suppressing tumor cell growth via AMPK activation and inhibiting proliferation via the mTOR pathway. Those effects were demonstrated in lymphoma cells (8), indicating the biological possibility for metformin's effect in preventing lymphoma and in improving lymphoma progression. Despite the experimental study evidence, our epidemiological study findings do not support the association in NHLs.

V. Banerji reports receiving commercial research grants from Lundbeck, Gilead, and Janssen and is a consultant/advisory board member for Gilead, Abbvie, and Janssen. No potential conflicts of interest were disclosed by the other authors.

The results and conclusions are those of the authors and no official endorsement by the Manitoba Centre for Health Policy, Manitoba Health, or other data providers is intended or should be inferred. The content is solely the responsibility of the authors and does not necessarily represent the official views of Research Manitoba.

Conception and design: X. Ye, V. Banerji, S.B. Gibson, S.M. Mahmud

Development of methodology: X. Ye, C. Righolt

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): X. Ye

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): X. Ye, G. Zhang, C. Righolt, V. Banerji, S.M. Mahmud

Writing, review, and/or revision of the manuscript: X. Ye, G. Zhang, C. Righolt, J.B. Johnston, V. Banerji, S.M. Mahmud

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): X. Ye, C. Righolt

Study supervision: X. Ye, C. Righolt, S.M. Mahmud

The authors acknowledge the Manitoba Center for Health Policy for use of data contained in the Manitoba Population Research Data Repository under project #2015-042 (HIPC#_2015/16-25; HREB # HS18835 (H2015:293); RRIC # 2015-044). Data used in this study are from the Manitoba Population Research Data Repository housed at the Manitoba Centre for Health Policy, University of Manitoba and were derived from data provided by Manitoba Health and CancerCare Manitoba. Research reported in this publication was supported by Research Manitoba New Investigator Operating Grant awarded to X. Ye (#1286 YEX NI) and Research Manitoba CLL Research Cluster Grant awarded to S.B. Gibson (#761090863).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.