Standardized incidence and associated factors for thyroid cancer in diabetic patients: a population-based analysis

Abstract

BACKGROUND: Although recent studies have suggested that the rise in thyroid carcinoma (TC) might be linked to the rising incidence of diabetes mellitus (DM), larger epidemiological evidence and proposed mechanism are lacking. Our study aimed to examine the standardized-incidence and associated factors of TC, after taking into account of DM control and therapy, in a large DM cohort. METHODS: Using a prospectively-collected population database, 144,084 diabetic patients managed in primary care without history of TC were identified. The time at-risk of TC was calculated from the date of cohort entry to the date of TC, date of death or last follow-up, whichever came first. The TC incidence in this cohort was then compared to that of the general population using the standardized incidence ratio (SIR) after stratifying by age (≤19, 20-44, 45-64 and ≥65 years old) and sex. To examine the association between clinical factors (socio-demographics, BMI, average HbA1c level over time, duration of DM and type of anti-DM medications) and incidence of TC, the Cox proportional hazards regression model was used for univariate and multivariate analyses. RESULTS: After a 323,844 person-years of observation, 36 new TCs were diagnosed. Compared to the general population, the TC incidence in the cohort was not significantly increased regardless of age and sex. For males, the overall SIR was 0.665 (0.095-1.234) while for females, it was 0.905 (0.562-1.248). In the univariate analysis, females (p = 0.003), lower education level (p = 0.043) and more frequent primary care DM visits (p < 0.001) were significant associated factors. After adjusting for education level, females (HR = 4.50,95%CI=1.75-11.63, p = 0.002) and more frequent primary care DM visits (HR = 1.13,95%CI=1.101-1.160, p < 0.001) were independent associated factors for TC. CONCLUSIONS: Contrary to recent evidence, diabetic patients managed in primary care were not at greater risk of developing TC relative to the non-diabetic normal population. Females and more frequent primary care DM visits were independent associated factors for TC. The latter finding implies increase detection/surveillance during unrelated primary care visit might be responsible for the recent surge of TC in diabetic patients.