Abstract:Objective: To characterize mortality after hospital discharge in cohorts with and without anorexia nervosa (AN). Methods: We obtained data for all hospitalizations for psychiatric reasons in Canada (except Quebec) between April 1, 2006, and March 31, 2021 (n = 1.3 million admissions). Cases of AN were identified using ICD-10 (F50.0 and F50.1) codes. First admissions during this interval for AN and other psychiatric conditions were linked to vital statistics data. Mortality was characterized through cross-tabulation, Cox proportional hazards models, and competing cause regression. Results: After adjustment for age and sex, there was no significant difference in mortality between AN and those with other psychiatric conditions (HR = 1.04; p = 0.644). Among AN admissions, 25% (95% CI 18.6-31.4) of deaths were attributed to psychiatric conditions (ICD-F codes), with 88% of these (comprising 22% of all deaths in the AN group) having AN itself identified as the underlying cause of death. In contrast, only 8% of deaths among non-AN admissions were attributed to a mental disorder. Discussion: Prevention of premature mortality in the general psychiatric population emphasizes modification of metabolic (e.g., hyperlipidemia) and lifestyle-related (e.g., sedentary behavior) risk factors. However, as AN itself makes a major contribution to mortality, specialized preventive strategies may be required.

Postdischarge Mortality in a Cohort Hospitalized With Anorexia Nervosa