Background: Pioglitazone has shown an inhibitory effect on the growth of cell lines derived from human salivary gland and human oral squamous cell carcinoma. However, whether this effect can be applied to patients with type 2 diabetes mellitus who use pioglitazone for glycemic control remains unanswered. Methods: The reimbursement records of all Taiwanese diabetic patients under treatment with oral anti-diabetic agents or insulin from 1996 to 2009 were retrieved from the National Health Insurance database. The entry date was set at 1 January 2006 and a total of 1,093,391 patients with type 2 diabetes were followed up for oral cancer incidence till the end of 2009. Incidences for ever-users, never-users and subgroups of pioglitazone dose-responsive exposure (using the parameters of time since starting pioglitazone, duration of therapy and cumulative dose) were calculated and hazard ratios estimated by Cox regression. Results: There were 58,238 ever-users and 1,035,153 never-users, with respective numbers of incident oral cancer of 201 (0.35%) and 4168 (0.40%), and respective incidences of 94.26 and 114.24 per 100,000 person-years. The overall hazard ratios (95% confidence intervals) did show a significantly lower risk in unadjusted [0.832 (0.722-0.958)] and age-sex-adjusted [0.866 (0.752-0.998)] models. However, the multivariable-adjusted hazard ratio (95% confidence interval) was not significant: 0.992 (0.857-1.148). Although significant P-trends for the dose-response parameters could be observed in some of the unadjusted and age-sex-adjusted models, none was significant in the multivariable-adjusted models. Conclusions: Pioglitazone has a null association with oral cancer after adjustment for potential confounders.
