Comparing Pioglitazone to Insulin with Respect to Cancer, Cardiovascular and Bone Fracture Endpoints, Using Propensity Score Weights

BACKGROUND Diabetes is an important global disease, associated with significant morbidity and an increased risk of death due to chronic end-organ complications. The thiazolidinediones, used mainly as third-line agents in type 2 diabetes mellitus (T2DM), have been associated with some safety concerns, such as an increased risk of bladder cancer, an increased risk of bone fracture and heterogeneous effects on cardiovascular events. OBJECTIVE This study aimed to evaluate safety data on pioglitazone for several outcomes and examine them in context with each other as well as with insulin, another third-line treatment for T2DM. METHODS This retrospective cohort study extracted data from May 1, 2000 until June 30, 2010, from the i3 InVision Data Mart™ database. To adjust for the testing of multiple hypotheses, the Holm method was applied to endpoints representing potential harm from pioglitazone treatment, separately from those representing potential benefit from pioglitazone. The study population included patients with T2DM ≥ 45 years old who were new users of either pioglitazone or insulin. Key outcomes were incident cases of a composite of myocardial infarction (MI) or stroke requiring hospitalization; bone fracture requiring hospitalization; bladder cancer; and a composite of nine other selected cancers. Kaplan-Meier curves were generated and hazard ratios (HRs) for pioglitazone versus insulin were estimated from Cox proportional hazards models adjusted with inverse probability of treatment weights derived from propensity scores. RESULTS A total of 56,536 patients (pioglitazone group 38,588; insulin group 17,948) qualified for the study. The mean follow-up was 2.2 years for pioglitazone and 1.9 years for insulin patients. Weighted survival analysis of the composite of MI and stroke, as well as the composite of nine cancers, yielded significant differences in favour of pioglitazone. For the composite of MI and stroke, the HR for pioglitazone versus insulin was 0.44 (95 % confidence interval [CI] 0.39-0.50, p < 0.0001). Modelling of the composite of nine selected cancers produced an HR of 0.78 (95 % CI 0.71-0.85, p < 0.0001). A non-statistically significant difference in favour of pioglitazone was observed in the incidence rate of bone fracture requiring hospitalization (HR 0.86, 95 % CI 0.74-1.01, p = 0.058). For bladder cancer, the overall incidence rates were relatively low and showed no significant difference between the two groups; the HR for pioglitazone versus insulin was 0.92 (95 % CI 0.63-1.33, p = 0.64). CONCLUSION Compared with insulin, pioglitazone was associated with a significant reduction in the risk of MI and stroke requiring hospitalization, and a significant reduction in the risk of other selected cancers. While pioglitazone treatment may be linked with a lower risk of bladder cancer and bone fracture relative to insulin, these differences were not statistically significant.