Risk of schizophrenia-spectrum disorders among individuals with eating disorders: a nation-wide cohort study

Individuals with eating disorders (EDs) hold an increased risk of developing schizophrenia-spectrum disorders (SSDs). It is however unclear whether this applies to all ED subgroups. We conducted a nationwide register-based cohort study including all individuals diagnosed with EDs in Denmark between 01jan1994 and 31dec2015. We used Cox regression models to compare the risk of SSDs between individuals with and without EDs and how this risk was influenced by risk factors and time since diagnosis of ED. The cohort included 20,045 individuals with EDs and 79,720 without ED (median age [interquartile range]:19 [16-24] years; 93% female). During 1,123,372 person-years of follow-up, 1,335 (6.7%) individuals with EDs were diagnosed with SSDs, compared to 982 (1.2%) individuals without EDs. This reflected a 4-fold higher risk of any SSD among individuals with EDs (adjusted hazard ratio [aHR] 4.43; 95% confidence interval [95%CI] 4.04-4.84). Individuals with anorexia nervosa (aHR 4.56 [3.99-5.22]) and other EDs (aHR 4.99 [4.27-5.83]) had a higher risk of any SSD diagnosis compared to individuals with bulimia nervosa (aHR 3.46 [2.83-4.22]). History of previous psychiatric hospitalisation, substance abuse, alcohol abuse, male sex, and onset of ED <19 years were associated with increased risk of SSDs in individuals with EDs. The risk of any SSD diagnosis decreased over time, but persisted ≥11 years after ED diagnosis (aHR 2.23 [1.71-2.92]). In conclusion, individuals with EDs have a four-fold higher risk of developing SSDs, compared with individuals without EDs. Anorexia nervosa and other EDs entails a higher risk of a subsequent SSD diagnosis compared with bulimia nervosa.