Incidence and Risk of Celiac Disease in Individuals with Type 1 Diabetes over the Lifespan—A Population-Based Cohort Study Using the Health Improvement Network Database

Abstract

Objective: Individuals with type 1 diabetes (T1D) are at increased risk for celiac disease (CD), though the incidence for those diagnosed with T1D beyond childhood is unknown. We used survival analysis to investigate the incidence and risk factors for diagnosis of CD over time in a population-based cohort of children and adults with T1D. Methods: Retrospective cohort study using The Health Improvement Network (THIN), a primary care electronic medical records database of >13 million people in the United Kingdom. Subjects were 1-35 years old at T1D diagnosis; subjects with type 2 diabetes or prevalent CD were excluded. We used Cox proportional hazards regression to identify factors associated with diagnosis of CD during follow-up. Risk factors assessed included age at T1D, sex, year of T1D diagnosis, and time-varying covariates of thyroid disease and adrenal insufficiency. Results: Subjects (n=11,537; 43% female) had a median follow-up time of 6.8 years (IQR 2.5-14.1). CD was diagnosed in 234 (2%), with median time to diagnosis of 2.9 years (IQR 1.2-7.8); 37% were diagnosed > 5 years after T1D. Incidence (per 10,000 person-years) was greater in females than males (26.9 [95% CI 22.4-31.9] vs. 16.4 [95% CI 13.4-19.9]) and in subjects <10 years at T1D diagnosis (50.1 [95% CI 41.4-60.2] vs. 13.3 [95% CI 11.0-15.9]). In multivariable Cox regression, the hazard of CD was higher with younger age (HR 0.88 per year of age at T1D diagnosis, 95% CI 0.80-0.96, p=0.006), in females (HR 1.50, 95% CI 1.16-1.94, p=0.002), and with more recent year of diagnosis (HR 1.09, 95% CI 1.07-1.12, p < 0.001). Conclusions: A significant proportion of CD cases were diagnosed more than 5 years after T1D diagnosis, the time frame currently recommended for screening. A more personalized approach to CD screening, accounting for the greater risk for younger-onset and female patients with T1D, may be warranted to minimize over-testing while limiting missed diagnoses. Disclosure M. Vajravelu: None. D. De Leon: Employee; Spouse/Partner; Merck & Co., Inc.. Research Support; Self; Zealand Pharma A/S. Consultant; Self; XOMA Corporation, ProSciento. Research Support; Self; Biomarin Corporation. D.R. Weber: None. R. Keren: None. M.R. Denburg: Research Support; Self; Mallinckrodt Pharmaceuticals. R. Verma: None.