AIMS: To analyse the association between cancer incidence and oral diabetes therapy (biguanide, sulphonylurea, thiazolidinedione and meglitinide) in men and women with type 2 diabetes mellitus. METHODS: A retrospective analysis of the electronic health record-based Cleveland Clinic Diabetes Registry (25 613 patients) was cross-indexed with the histology-based tumour registry (48 051 cancer occurrences) over an 8-year period (1998-2006). Multiple imputations were used to account for missing data. Cox regression with propensity scores was used to model time for the development of incident cancer in each of the imputed datasets and the results were pooled. RESULTS: During 51 994 person follow-up years, 892 incident cancer cases were identified; prostate (14.5%) and breast (11.7%) malignancies were most frequent. In women, thiazolidinedione use was associated with a 32% decreased cancer risk compared with sulphonylurea use [hazard ratio (HR) 0.68; 95% confidence interval (CI) 0.48-0.97, in the adjusted analysis]. Comparison of insulin secretagogues (sulphonylurea and meglitinide) versus insulin sensitizers (biguanide and thiazolidinedione) demonstrated a 21% decreased cancer risk in insulin sensitizers [HR 0.79 (95% CI 0.64-0.98) in the adjusted analysis]. Oral diabetes therapy showed no significant difference in men. Adjustments were made for age, body mass index (BMI), low-density lipoprotein (LDL), high-density lipoprotein (HDL), triglycerides, coronary heart disease (CHD), diabetes oral monotherapy, race, gender, haemoglobin A1c, statin use, income, insulin use, glomerular filtration rate (GFR), new diabetes status, prior cancer, prior cerebrovascular accident (stroke or transient ischaemic event), systolic/diastolic blood pressure, tobacco use (ever/never) and the propensity score for receiving a biguanide. CONCLUSIONS: Oral insulin sensitizers, particularly thiazolidinedione, are associated with decreased malignancy risk in women with type 2 diabetes mellitus.
AIMS: To analyse the association between cancer incidence and oral diabetes therapy (biguanide, sulphonylurea, thiazolidinedione and meglitinide) in men and women with type 2 diabetes mellitus. METHODS: A retrospective analysis of the electronic health record-based Cleveland Clinic Diabetes Registry (25 613 patients) was cross-indexed with the histology-based tumour registry (48 051 cancer occurrences) over an 8-year period (1998-2006). Multiple imputations were used to account for missing data. Cox regression with propensity scores was used to model time for the development of incident cancer in each of the imputed datasets and the results were pooled. RESULTS: During 51 994 person follow-up years, 892 incident cancer cases were identified; prostate (14.5%) and breast (11.7%) malignancies were most frequent. In women, thiazolidinedione use was associated with a 32% decreased cancer risk compared with sulphonylurea use [hazard ratio (HR) 0.68; 95% confidence interval (CI) 0.48-0.97, in the adjusted analysis]. Comparison of insulin secretagogues (sulphonylurea and meglitinide) versus insulin sensitizers (biguanide and thiazolidinedione) demonstrated a 21% decreased cancer risk in insulin sensitizers [HR 0.79 (95% CI 0.64-0.98) in the adjusted analysis]. Oral diabetes therapy showed no significant difference in men. Adjustments were made for age, body mass index (BMI), low-density lipoprotein (LDL), high-density lipoprotein (HDL), triglycerides, coronary heart disease (CHD), diabetes oral monotherapy, race, gender, haemoglobin A1c, statin use, income, insulin use, glomerular filtration rate (GFR), new diabetes status, prior cancer, prior cerebrovascular accident (stroke or transient ischaemic event), systolic/diastolic blood pressure, tobacco use (ever/never) and the propensity score for receiving a biguanide. CONCLUSIONS: Oral insulin sensitizers, particularly thiazolidinedione, are associated with decreased malignancy risk in women with type 2 diabetes mellitus.