A dual nationwide epidemiological study investigating the genetic etiology of eating disorders in denmark and sweden

Eating disorders (ED) are mental disorders associated with high levels of morbidity, mortality, and social, psychological and physical impairment. Currently, eating disorders are more prevalent in women, yet they affect people from all sectors of society (i.e., sex, ancestry, sexual orientation, age etc). Lifetime prevalence estimates of ED vary dramatically between studies. This high level of variability is likely due to sample size variations, differences in diagnostic practices, and ascertainment biases (e.g., clinical vs population wide sampling). These biases influence downstream estimates of the genetic contribution to the liability of EDs. Consequently, while family, twin, and adoption studies have shown that genetic factors contribute to the risk of developing an eating disorder, the variability in estimates is substantial. This highlights the need for large population representative studies. Here, we estimated the heritability of three eating disorders (anorexia, bulimia, and ED NOS) and the genetic correlations between them and seven other mental disorders using information from nationally collected civil- and medical registers in Denmark (>6 million) and Sweden (>13 million). These registers contain information about all individuals in the two Nordic countries during roughly the last 100 years, their extended genealogy, and all hospital records. Both datasets reflect a total population, uniformly diagnosed across the country in a nationally run public healthcare system. Therefore, the effect of different kinds of biases associated with cohorts ascertained from (psychiatric) clinics or voluntary participants is reduced. Using Danish data, we showed that all eating disorders were heritable (h2: 0.28–0.35) and presented substantial positive genetic correlations with other mental disorders (rG: 0.23–0.63 (anorexia); 0.24–0.46 (bulimia); 0.42–0.73 (ED NOS)). While there is some variation, all genetic correlations presented overlapping confidence intervals. OCD showed the largest genetic correlation with both Anorexia and ED NOS. A similar association was observed using genotype-based estimates of genetic correlations. Eating disorders, especially anorexia, are commonly observed comorbid with OCD and anxiety phenotypes. The observed large genetic correlations with eating disorders point to a substantial relevance of shared genetic effects as a partial explanation for those clinical observations. The same pattern of results was observed using the larger Swedish data, indicating a high degree of similarity between countries. From the sex-stratified analyses, we observed that the female estimates were nearly identical to the main results, while the male estimates could not be considered reliable. These results suggest that due to the substantial difference in ED prevalence between males and females, and despite using the complete data of two Nordic countries, the estimations we observed by using males and females together are driven by the female samples and that we are underpowered to estimate heritability and genetic correlations in males. Here, using unascertained population-wide register data of two Nordic countries representing 20 million people, we have shown that eating disorders are complex heritable outcomes with substantial evidence for partially shared genetic risk across many mental disorders.