Genetic and environmental influences on eating disorders and associated adversities and comorbidities

Abstract: Eating disorders (EDs), including anorexia nervosa (AN) and bulimia nervosa (BN), are severe psychiatric disorders. Adversities (including suicide) and comorbidities have been illustrated in clinical observations with varying sample sizes, but evidence from large epidemiological research is still lacking. Further, the mechanisms underlying the observed associations remain largely unclear. Taking advantage of the unique Swedish national registries, this thesis aims to examine the associations between EDs and between EDs and potential adversities and comorbidities at population level and deepen the understanding of the mechanisms underlying these associations using genetically informative study designs. Study I applied quantitative genetic modeling to estimate genetic and environmental effects on AN and BN and their overlap. This study used registry data in siblings and half-siblings, which significantly increased the sample size and extended the literature from self-reported behavioral measures to clinical diagnosis. Consistent with twin studies, moderate heritability was found for both AN and BN. Furthermore, moderate genetic and environmental correlations were found between clinically diagnosed AN and BN, suggesting partially overlapped etiologies between the two EDs in the general population. Study II focused on the associations between EDs and suicide attempts and death by suicide. At population level, significantly increased risks of both suicide attempts and death by suicide were found in individuals with EDs (over 5 times the risk) compared to in individuals without EDs. Individuals with full-sibling or cousins with EDs were also at increased risks of suicide attempts. The familial co-aggregation pattern suggested that EDs and suicide attempts might share familial liabilities, which could include genetic and/or environmental risk factors shared by family members. Study III assessed the risks of committing theft and other crimes in EDs in a nationwide female cohort. Firstly, significantly higher risks of both theft and other crimes were found in exposed females (i.e., had been diagnosed with AN or BN) than in unexposed females; theft was more common than other crimes altogether in exposed groups; and both the absolute and relative risks were higher in BN than in AN. Next, sibling comparison design, where the risks were compared between differentially exposed full-sisters, was applied to account for potential confounding effects of familial factors shared between sisters. The relative risk of theft decreased but remained statistically significant in BN and did not decrease in AN. The finding suggests that familial confounders (e.g., genetic and/or familial environmental confounders) were likely to explain part of the association between BN and theft but not the association between AN and theft, potentially reflecting different etiologies of the two EDs. Study IV examined the genetic associations between EDs and attention-deficit/hyperactivity disorder (ADHD) using multiple approaches, namely assessing familial co-aggregation, quantitative genetic modeling, and analysis of polygenic risk scores (PRS, a measure of genetic risk of a disorder). 1) Increased risks of being diagnosed with AN and non-AN EDs (including BN) were found in individuals diagnosed with ADHD and their full- and maternal half-siblings and cousins, compared to individuals without ADHD and their relatives, suggesting familial liabilities shared between ADHD and the EDs. 2) Moderate genetic correlations were found between non-AN EDs and ADHD and between BN and ADHD, and mild genetic correlation was found between AN and ADHD. 3) ADHD PRS significantly predicted ED symptoms including drive for thinness and body dissatisfaction in a large genotyped population sample, indicating that the polygenic risk of ADHD influenced some ED symptoms. The findings of the three approaches converged and together illustrated significant genetic correlations between EDs, especially non-AN EDs, and ADHD at both diagnostic and symptomatic levels. Both ADHD and theft behaviors (in Study III) might reflect multi-impulsive forms of EDs which, as suggested by previous studies, may be associated with relatively poorer treatment response. Taken together, this thesis highlighted the seriousness of EDs by revealing their associations with adversities (suicide and crime) and comorbidity (ADHD) at population level. Further, it revealed the genetic and/or environmental influences on these associations and the associations among EDs. The findings suggest that EDs are correlated yet different disorders and provide insights on the etiologies underlying these important associations, encouraging future research to identify specific risk factors that target the shared etiologies. Clinical implications include the identification of subgroups in individuals with EDs who display high impulsivity and high risk of suicide as well as vigilance of forensic issues that could complicate treatment and recovery. The findings also highlighted increased risks of EDs, adversities, and comorbidity in family members of individuals with EDs, calling for clinical attention to the psychological robustness of the relatives especially when they serve as the caregivers of ED patients and are expected to engage intensively in treatment.