Abstract
About 20% of individuals with anorexia nervosa (AN) remain chronically ill. Therefore, early identification of poor outcome could improve care. Genetic research has identified regions of the genome associated with AN. Patients with anorexia nervosa were identified via the Swedish eating disorder quality registers Stepwise and Riksät and invited to participate in the Anorexia Nervosa Genetics Initiative. First, we associated genetic information longitudinally with eating disorder severity indexed by scores on the Clinical Impairment Assessment (CIA) in 2843 patients with lifetime AN with or without diagnostic migration to other forms of eating disorders followed for up to 16 years (mean = 5.3 years). Second, we indexed the development of a severe and enduring eating disorder (SEED) by a high CIA score plus a follow-up time ≥5 years. We associated individual polygenic scores (PGSs) indexing polygenic liability for AN, schizophrenia, and body mass index (BMI) with severity and SEED. After multiple testing correction, only the BMI PGS when calculated with traditional clumping and p value thresholding was robustly associated with disorder severity (βPGS = 1.30; 95% CI: 0.72, 1.88; p = 1.2 × 10–5) across all p value thresholds at which we generated the PGS. However, using the alternative PGS calculation method PRS-CS yielded inconsistent results for all PGS. The positive association stands in contrast to the negative genetic correlation between BMI and AN. Larger discovery GWASs to calculate PGS will increase power, and it is essential to increase sample sizes of the AN GWASs to generate clinically meaningful PGS as adjunct risk prediction variables. Nevertheless, this study provides the first evidence of potential clinical utility of PGSs for eating disorders.
Highlights
Eating disorders are complex psychiatric conditions that arise from a combination of genetic and environmental factors [1]
In bulimia nervosa (BN), findings are mixed with a high frequency of compensatory behaviours [16] and comorbid psychiatric diagnoses associated with poor outcome [17,18,19], whereas perfectionism, obsessionality, anxiety, and genetic factors may increase the likelihood of developing a severe and enduring eating disorder (SEED) [20]
Participants were recruited for genotyping into Anorexia Nervosa Genetics Initative (ANGI) via Swedish treatment centres and two national registers: (1) Riksät, the national quality register established in 1999, including patients treated for Anorexia nervosa (AN), BN, or eating disorders not otherwise specified (EDNOS) [40] and (2) Stepwise, the internet-based clinical quality assurance database for specialised eating disorder care, established in Ethics All participants provided written informed consent for participation in ANGI
Summary
Eating disorders are complex psychiatric conditions that arise from a combination of genetic and environmental factors [1]. The largest genome-wide association study (GWAS) published as of 2020 associated eight risk loci with AN and indicated a genetic sharing between AN and BMI (rg = −0.32) and between AN and schizophrenia (SCZ; rg = 0.25), corroborating observed comorbidity and familiality [23] in clinical and epidemiological studies [24,25,26] These findings implicate both anthropometric and psychiatric factors in the origin of AN [27]. Given the observed negative genetic correlation between AN and BMI [27], and the association between low phenotypic BMI and poor outcome [36,37,38], we hypothesised that a BMI PGS would separately be associated with severity and the development of a SEED. Most patients were first registered in the region of Stockholm (32.3%), followed by Västra
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
