Type 2 diabetes is linked to higher risk of many non-metabolic conditions (e.g. cancers, neuropsychiatric and musculoskeletal disease) which all share a common risk factor: obesity. The causal pathways from excess weight to disease are unknown. We aimed to use specific genetic variants to separately test the causal roles of higher adiposity with and without its adverse metabolic effects. Using metabolic biomarkers from 442,278 UK Biobank individuals, we identified two clusters of body fat % increasing alleles: a) 36 “favourable adiposity” (FA) alleles associated with a favourable metabolic profile (higher HDL and SHBG and lower triglyceride and liver enzymes); b) 38 “unfavourable adiposity” (UFA) alleles associated with an unfavourable profile (lower HDL and SHBG and higher triglyceride and liver enzymes). We used two-sample Mendelian randomization with 43 diseases from non-UK Biobank GWASs. We identified two sets of diseases. 1) Those where the metabolic effect of higher adiposity is the likely primary cause of the disease. Here the FA and UFA clusters had an opposing effect on the risk of disease: type 2 diabetes (FA: 0.11 OR per 1 standard deviation higher body fat % [95%CI: 0.08, 0.16] vs. UFA: 5.50 [4.29, 7.05]), stroke (0.65 [0.52, 0.83] vs. 1.43 [1.23, 1.67]), gout (0.44 [0.29, 0.68] vs. 2.49 [1.88, 3.29]), breast cancer (1.12 [0.64, 1.94] vs. 0.52 [0.39, 0.68]) and colorectal cancer (0.67 [0.52, 0.85] vs. 1.20 [1, 1.50]). 2) Those where the mechanical excess weight (or other non-metabolic effects) is likely a cause of the disease. Here the FA cluster, despite having lower metabolic risk, and the UFA cluster were both associated with higher disease risk: venous thromboembolism (2.52 [1.82, 3.47] vs. 1.63 [1.25, 2.13]), osteoarthritis (1.45 [1.19, 1.76] vs. 2.20 [1.64, 2.95]), depression (2.85 [1.38, 5.89] vs. 2.59 [1.60, 4.21]). Our results assist in understanding the consequences of higher adiposity uncoupled from its adverse metabolic effects in non-metabolic diseases pathophysiological mechanisms.


S. Martin: None. J. D. Bell: None. T. M. Frayling: None. H. Yaghootkar: None. J. Tyrrell: None. E. L. Thomas: None. M. Bown: None. J. T. Elder: None. C. Kabrhel: Research Support; Self; Diagnostica Stago, Grifols. N. Papadimitriou: None. M. Gunter: None. E. E. Vincent: None.


Diabetes UK (17/0005594)

Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at http://www.diabetesjournals.org/content/license.