Abstract
The reimbursement database of the Taiwan's National Health Insurance was used to evaluate oral cancer risk after sitagliptin use. Patients newly diagnosed of type 2 diabetes during 1999-2008 were recruited. A 1:1 propensity score matched-pair sample of 39195 ever users and 39195 never users were followed up until December 31, 2011. Cox regression incorporated with the inverse probability of treatment weighting using propensity score was used to estimate hazard ratios. Results showed that the overall hazard ratio was not statistically significant (0.956, 95% confidence interval: 0.652-1.401). However, in tertile analyses, the hazard ratio for the first (< 7.47 months), second (7.47-15.63 months) and third (> 15.63 months) tertile of cumulative duration was 1.563 (0.963-2.537), 1.236 (0.738-2.071) and 0.345 (0.164-0.725), respectively; and was 1.575 (0.963-2.575), 1.224 (0.738-2.033) and 0.347 (0.165-0.731), respectively, for the first (< 19,600 mg), second (19,600-42,200 mg) and third (> 42,200 mg) tertile of cumulative dose. Sensitivity analyses after excluding patients who developed any other cancer during follow-up did not change the results substantially. Additionally, the risk of oral diseases that may predispose to oral cancer (i.e., "gingival and periodontal diseases" and/or "oral mucosal lesions") paralleled the risk pattern of oral cancer, suggesting a possible explanation for the risk change of oral cancer related to sitagliptin. In conclusion, sitagliptin may reduce oral cancer risk when the cumulative duration is > 15.63 months or the cumulative dose is > 42,200 mg.