Phenome-wide screening of GWAS data reveals the complex causal architecture of obesity.

Abstract

OBJECTIVE\nIn the present study, we sought to identify causal relationships between obesity and other complex traits and conditions using a data-driven hypothesis-free approach that uses genetic data to infer causal associations.\n\n\nMETHODS\nWe leveraged available summary-based genetic data from genome-wide association studies on 1498 phenotypes and applied the latent causal variable method (LCV) between obesity and all traits.\n\n\nRESULTS\nWe identified 110 traits causally associated with obesity. Of those, 109 were causal outcomes of obesity, while only leg pain in calves was a causal determinant of obesity. Causal outcomes of obesity included 26 phenotypes associated with cardiovascular diseases, 22 anthropometric measurements, nine with the musculoskeletal system, nine with behavioural or lifestyle factors including loneliness or isolation, six with respiratory diseases, five with body bioelectric impedances, four with psychiatric phenotypes, four related to the nervous system, four with disabilities or long-standing illness, three with the gastrointestinal system, three with use of analgesics, two with metabolic diseases, one with inflammatory response and one with the neurodevelopmental disorder ADHD, among others. In particular, some causal outcomes of obesity included hypertension, stroke, ever having a period of extreme irritability, low forced vital capacity and forced expiratory volume, diseases of the musculoskeletal system, diabetes, carpal tunnel syndrome, loneliness or isolation, high leukocyte count, and ADHD.\n\n\nCONCLUSIONS\nOur results indicate that obesity causally affects a wide range of traits and comorbid diseases, thus providing an overview of the metabolic, physiological, and neuropsychiatric impact of obesity on human health.