Manfred Maximilian Fichter1,2, Norbert Quadflieg1. 1. Department of Psychiatry and Psychotherapy, Ludwig-Maximilians-University (LMU), Munich, 80336, Germany. 2. Schön Klinik Roseneck Affiliated with the Medical Faculty of the University of Munich (LMU), Prien, 83209, Germany.
Abstract
OBJECTIVE: To report on long-term mortality in anorexia nervosa (AN), bulimia nervosa (BN), binge eating disorder (BED), and eating disorder not otherwise specified (ED-NOS), causes of death, and predictors of early death. METHOD: A large sample of consecutively admitted inpatients (N = 5,839) was followed-up on vital status through the German civil registry office. Of these patients 1,639 were treated for AN, 1,930 for BN, 363 for BED, and 1,907 for ED-NOS. Data from the main inpatient hospital treatment were applied to bivariate and multivariate Cox regression analyses on survival time from onset of eating disorder to death or end of observation. Standardized mortality ratios (SMR) were computed matched for age, gender, and person-years. RESULTS: SMR were 5.35 for AN, 1.49 for BN, 1.50 for BED, 2.39 for narrowly defined ED-NOS, and 1.70 for widely defined ED-NOS. Patients with AN died earlier than patients with BN, BED, or ED-NOS who did not differ. A diagnosis of AN, chronicity, later age of onset, not living in a relationship, and an irregular type of discharge from index inpatient treatment were major predictors of a shorter time to death. Suicidality was a univariate predictor of a shorter time to death in BN only. AN patients mostly died from natural causes related to their eating disorder. DISCUSSION: Mortality in AN is excessive and considerably higher than in BN, BED, and ED-NOS.
OBJECTIVE: To report on long-term mortality in anorexia nervosa (AN), bulimia nervosa (BN), binge eating disorder (BED), and eating disorder not otherwise specified (ED-NOS), causes of death, and predictors of early death. METHOD: A large sample of consecutively admitted inpatients (N = 5,839) was followed-up on vital status through the German civil registry office. Of these patients 1,639 were treated for AN, 1,930 for BN, 363 for BED, and 1,907 for ED-NOS. Data from the main inpatient hospital treatment were applied to bivariate and multivariate Cox regression analyses on survival time from onset of eating disorder to death or end of observation. Standardized mortality ratios (SMR) were computed matched for age, gender, and person-years. RESULTS: SMR were 5.35 for AN, 1.49 for BN, 1.50 for BED, 2.39 for narrowly defined ED-NOS, and 1.70 for widely defined ED-NOS. Patients with AN died earlier than patients with BN, BED, or ED-NOS who did not differ. A diagnosis of AN, chronicity, later age of onset, not living in a relationship, and an irregular type of discharge from index inpatient treatment were major predictors of a shorter time to death. Suicidality was a univariate predictor of a shorter time to death in BN only. AN patients mostly died from natural causes related to their eating disorder. DISCUSSION: Mortality in AN is excessive and considerably higher than in BN, BED, and ED-NOS.