# Incidence, prevalence and mortality of anorexia nervosa in individuals with childhood-onset type 1 diabetes: a nationwide retrospective cohort study in Sweden

## TL;DR

This study finds that females with childhood-onset type 1 diabetes in Sweden have a higher risk of anorexia nervosa and higher mortality when both conditions are present.

## Contribution

The study provides the first nationwide analysis of anorexia nervosa incidence, prevalence, and mortality in individuals with childhood-onset type 1 diabetes.

## Key findings

- Females with childhood-onset T1D had a 1.9% period prevalence of AN compared to 1.1% in controls.
- Mortality was significantly higher in females with both T1D and AN compared to controls.
- The 10-year incidence rate of AN was 74.7 per 100,000 person-years in T1D females versus 45.2 in controls.

## Abstract

To investigate the incidence, prevalence and mortality of anorexia nervosa (AN) among individuals with childhood-onset type 1 diabetes (T1D) compared with matched controls in Sweden.

Retrospective nationwide cohort study using linked registry data.

Nationwide, Sweden; population-based registers (covering the period 1977–2019).

12 202 individuals diagnosed with T1D before age 15 years (5618 females; 6584 males) and 48 484 age-matched, sex-matched and municipality-matched controls without diabetes (23 618 females; 24 866 males).

AN diagnoses (International Classification of Diseases-10 codes F50.0 and F50.1) identified via the National Patient Register. Outcomes were period prevalence, point prevalence at ages 15 and 20 years, 10-year incidence rates and proportional mortality ratios (PMR), stratified by sex. ORs and incidence rate ratios (IRR) with 95% CIs were estimated using Mantel-Haenszel methods; Kaplan-Meier analysis compared time to AN diagnosis between groups.

The period prevalence of AN among females with T1D was 1.9% compared with 1.1% in controls (OR 1.64, 95% CI 1.31 to 2.06; p<0.001). The 10-year incidence rate for females with T1D was 74.7 per 100 000 person-years vs 45.2 per 100 000 person-years in controls (IRR 1.77, 95% CI 1.35 to 2.32). Point prevalence at age 15 years was 0.87% (T1D) vs 0.53% (controls) (IRR 1.65, 95% CI 1.16 to 2.35), and at age 20 years was 1.73% (T1D) vs 1.11% (controls) (IRR 1.55, 95% CI 1.20 to 1.99). The PMR for females with both T1D and AN compared with controls without either condition was 20.4 (95% CI 6.6 to 47.6). Male cases were few (n=4 in the T1D group; n=12 in controls).

Females with childhood-onset T1D in Sweden have an elevated risk of AN and markedly higher mortality when both conditions are present. Despite the increased relative risk, the absolute risk of AN in females with T1D remained below 2%. These findings support routine screening for eating disorders in the T1D population, particularly among adolescent and young adult females.