Investigating the association between Incretin-based therapies and thyroid cancer incidence using United States Medicare data 2007-2019

Background: The potential risk of thyroid cancer with glucagon-like peptide 1 receptor agonists (GLP-1RA) use is a public health concern. Preclinical studies in rodents have suggested a potential thyroid cancer link with GLP-1RA use, yet it is unclear if this risk translates to humans.

Objectives: To estimate the comparative effect of incretin-based therapies (GLP-1RA and dipeptidyl-peptidase-4 inhibitors [DPP4i]) versus sodium-glucose cotransporter-2 inhibitors (SGLT-2i) on thyroid cancer incidence among US older adults with type 2 diabetes.

Methods: We implemented an active comparator new user design using a random 20% sample of Medicare Fee-for-Service beneficiaries aged 66+ years with two consecutive prescription claims for GLP-1RA, DPP4i or SGLT-2i between Jan 1, 2008, and Dec 31, 2018. Three pair-wise new user cohorts were created to compare: GLP-1RA and SGLT-2i (cohort I); DPP4i and SGLT-2i (cohort II); and GLP-1RA and DPP4i (cohort III). Follow-up began six months after the second prescription date and continued until the first occurrence of the outcome, treatment discontinuation, switching to or augmentation with comparator, death from any cause, termination of Medicare enrollment, or Dec 31, 2019. We required ≥2 separate diagnoses for malignant neoplasm of thyroid gland (ICD-9 193 or ICD-10 C73, D09.3 or D44.0) within 60 days to establish a thyroid cancer diagnosis. We estimated adjusted 3-year cumulative risk differences of thyroid cancer (aRDs) with 95% confidence intervals (CIs) using standardized morbidity ratio (SMR) weighted Kaplan-Meier survival functions, and adjusted hazard ratios (aHRs) and 95% CIs using SMR weighted Cox models.

Results: We included 73,388 new users in cohort I (mean age 72.4 years, male: 48.3%), 106,274 new users in cohort II (mean age 74.6 years, male: 44.9%) and 191,143 new users in cohort III (mean age 74.5 years, male: 41.3%). For cohort I, the 3-year crude risks of thyroid cancer were 0.06% vs 0.25%, aRD was -0.29% (95% CI: -0.57%, -0.01%), and the aHR was 0.37 (95% CI: 0.16, 0.87). The 3-year crude risks of thyroid cancer were 0.14% vs 0.17%, and 0.12% vs 0.11% for cohorts II and III respectively. Neither cohort II (aRD = -0.01% [95% CI: -0.16%, 0.14%]) nor cohort III (aRD = -0.02% [95% CI: -0.15%, 0.10%]) was associated with an increased 3-year risk of thyroid cancer. Requiring a thyroidectomy in addition to ≥2 separate diagnoses codes for thyroid cancer yielded aRD = -0.23% [-0.51%, 0.04%], cohort I; aRD = -0.03% [-0.19%, 0.14%], cohort II; and aRD = -0.07% [-0.11%, -0.03%], cohort III.

Conclusions: Our study of U.S. Medicare beneficiaries with diabetes suggests that the initiation of incretin-based therapies is not associated with an increased 3-year risk of thyroid cancer, but the results are too imprecise to exclude a moderate increase in risk.