Age-dependent association of thyroid peroxidase antibodies with incident cardiovascular disease events: Insights from a longitudinal cohort study

Abstract

Background The relationship between thyroid autoimmunity and cardiovascular disease (CVD) is controversial, and age-specific effects are not well characterized. We examined associations between thyroid peroxidase antibodies (TPOAb) status and incident CVD. Methods In this prospective population-based cohort, 4825 adults aged ≥20 years without baseline CVD were identified. Survival analyses were conducted using a 1.5-year landmark approach. Mean follow-up after the landmark was 14.8 ± 3.3 years. TPOAb was analyzed as a log-transformed continuous variable, dichotomous status, and categorical levels. Cox proportional hazards models were used to estimate hazard ratios (HRs) with progressive adjustment for demographic, lifestyle, cardiometabolic, and thyroid-related factors. Effect modification by age was assessed using interaction and spline analyses. Results A total of 4825 participants (56.5% female and mean age of 39.7 ± 14.0 years) were included. During follow-up, 576 incident CVD events occurred. In the overall population, TPOAb was not associated with CVD risk (adjusted HR per 1-SD increase: 1.04, 95%CI 0.95-1.13). The primary continuous-age spline analysis demonstrated significant effect modification by age (global interaction P = 0.024). In exploratory age-stratified analyses, higher TPOAb levels were associated with increased CVD risk among participants aged <40 years (adjusted HR: 1.35, 95%CI 1.10-1.66), but not among those aged ≥40 years (HR, 1.00; 95% CI, 0.92–1.10). The association among younger participants remained significant in the euthyroid subgroup and in sensitivity analyses using Firth penalized regression and hard cardiovascular outcomes. Dose-response analyses further suggested increased risk at higher TPOAb levels among younger participants. Conclusions Thyroid autoimmunity was not associated with incident CVD in the overall cohort but demonstrated a significant age-dependent association. Exploratory analyses suggested increased risk among participants aged <40 years. These hypothesis-generating findings identify age as a potentially important modifier of the thyroid autoimmunity–CVD association and warrant validation in independent cohorts.