It is unknown whether the association between body mass index (BMI) and chronic kidney disease (CKD) is mediated by genetic confounding or obesity-associated diabetes. We investigated the association between BMI and incident CKD in 29,136 Swedish twins with no history of CKD or diabetes, first using traditional Cox regression in a cohort design, and second controlling for shared genetic factors within twin pairs. Hazard ratios (HR) per unit increase in BMI were calculated and adjusted for age, sex, comorbidities, and lifestyle factors. We contrasted CKD risk between twins with discordant BMI and adjusted for diabetes as a time-varying covariate. During an average follow-up of 12.8 years, 1,113 (3.8%) incident CKD and 2,282 (7.8%) diabetes cases were observed. BMI was associated with CKD risk in the traditional cohort analysis (HR 1.13; 95% CI, 1.11-1.14). Effect sizes were similar in analyses clustered by twin pairs to adjust for factors shared by twins (HR 1.16; 95 CI 1.09-1.22 in same-sex fraternal twins and HR 1.17; 95% CI 1.06-1.29 in identical twins). Heavier twins had higher incidence rates of both CKD and diabetes than their leaner siblings, particularly when BMI differed by >2 kg/m2. The proportion of CKD cases attributable to BMI ≥25 kg/m2 was estimated to be 32%, with approximately 19% of those cases explained by diabetes. We conclude that a higher BMI, irrespective of genetic confounding, is associated with increased CKD risk. A large proportion of CKD cases might be prevented if the population maintained a normal BMI.
Keywords: BMI; CKD; dizygotic; monozygotic; twins.
Copyright © 2019 International Society of Nephrology. Published by Elsevier Inc. All rights reserved.