Initial metformin or sulphonylurea exposure and cancer occurrence among patients with type 2 diabetes mellitus

Aim This was a retrospective cohort study of type 2 diabetes patients, to evaluate the association between initial metformin or sulphonylurea treatment and cancer incidence. Methods Patients identified in the UK Clinical Practice Research Datalink (CPRD), previously General Practice Research Database, during 1995–2008 who were initially stabilized on OHA monotherapy, including metformin, sulphonylurea, thiazolidinediones (TZDs) or meglitinides, were included in the cohort. New diagnoses of cancer, including malignant solid tumours and haematological malignancies, occurring during the follow‐up were identified from the cohort. Age‐standardized incidence rates were estimated and compared between metformin and sulphonylurea exposure groups. Results The age standardized incidences of cancer were 7.5 and 8.5 per 1000 person‐years for the metformin and sulphonylurea exposure groups, respectively. After adjusting for potential confounders, the hazard ratios (HR) for malignant solid tumours and haematological malignancies were 1.06 (95% CI: 0.98, 1.15) and 0.98 (95% CI: 0.67, 1.43) for sulphonylurea group as compared to the metformin group, respectively. For individual cancers, the HRs were 1.17 (95% CI: 0.95, 1.44), 1.04 (95% CI: 0.83, 1.31) and 0.88 (95% CI: 0.71, 1.11) for colorectal cancer, breast cancer and prostate cancer, respectively. Conclusion This study provides evidence that cancer incidence in the first few years after starting metformin or sulphonylurea therapy in type 2 diabetes patients is not much affected by choice of hypoglycaemic drug class.