To date, few epidemiologic studies have investigated the relationship between pioglitazone use and prostate cancer. The available studies show conflicting findings. This case-control study explored the association between prior pioglitazone usage and prostate cancer using a large, population-based data set. Data were derived from the Longitudinal Health Insurance Database 2005 in Taiwan, a population-based sample of National Health Insurance enrollees with longitudinal claims data since 1995. Cases were 3513 patients with prostate cancer aged over 40 years, and the controls were 3513 patients without prostate cancer, matched with prostate cancer cases on age, and having a medical care utilization episode in the year of the index prostate cancer (1 control per case). We performed conditional logistic regression to examine the odds ratio and 95% confidence intervals for prior pioglitazone use between cases and controls. Analysis showed that 178 of the total sample patients (2.53%) had used pioglitazone prior to the index date. Prior pioglitazone use was found in 72 (2.05%) cases and 106 (3.02%) controls. The crude odds ratio of prior pioglitazone usage for cases was 0.67 (95% confidence interval, 0.50-0.91) compared to controls. After adjusting for demographic characteristics and comorbidities, the negative association of prior pioglitazone with prostate cancer persisted (adjusted odds ratio, 0.59; 95% confidence interval, 0.43-0.80). We concluded that there was an inverse association between prior pioglitazone usage and prostate cancer in this study.
ASJC Scopus subject areas