Abstract
BACKGROUND: Quantify association between the glucagon-like peptide-1 receptor agonist liraglutide and risk of thyroid cancer (TC) compared to other antidiabetics. PATIENTS AND METHODS: Initiators of liraglutide, exenatide, metformin, pioglitazone or groups of dipeptidyl peptidase-4 inhibitors or sulfonylureas were identified in a US health plan (2010-2014) and followed for a median of 17 months. Thyroid cancer cases during follow-up were identified via a validated algorithm. Incidence rates of TC among liraglutide and comparators were assessed using relative risks estimated within propensity score-matched cohorts using intention to treat (ITT) and time on drug analyses. Latency effects and potential surveillance bias were evaluated. RESULTS: Relative risks from ITT analyses ranged from 1.00 (95% confidence interval (CI) 0.56-1.79) versus metformin to 1.70 (95% CI 1.03-2.81) versus all comparators excluding exenatide. Effect estimates from latency analyses were slightly attenuated. Time on drug analyses suggested no increased risk for either longer duration or higher cumulative dose of liraglutide. Medical record review found 85% were papillary or a follicular variant of papillary or both; 46% were microcarcinomas (≤10 millimeters), which were more prevalent in the liraglutide cohort (67% versus 43% in all comparators). CONCLUSION: Relative risks were elevated for several comparisons, which should be interpreted cautiously because of potential residual confounding and surveillance bias. Liraglutide cases had smaller thyroid nodules and shorter time-to-diagnosis, suggesting increased surveillance for TC among liraglutide initiators, especially shortly after the drug´s approval. After adjusting the primary analyses (ITT) for latency, no significant elevated risk of TC was observed among liraglutide initiators.