Association between per capita sugar consumption and diabetes prevalence mediated by the body mass index: results of a global mediation analysis

PURPOSE: The aim of this study was to examine the mediation of body mass index (BMI) on the association between per capita sugar consumption and diabetes prevalence using country-related data. RESEARCH DESIGN AND METHODS: In this ecological study, based on 192 countries, data on per capita sugar consumption were obtained from the Food and Agriculture Organization of the United Nations (FAO), on BMI from the World Health Organization and on diabetes prevalence from the International Diabetes Federation. Data on demography and economic factors were obtained from the Central Intelligence Agency, the United Nations and the FAO. Multiple linear regression analysis was performed to investigate the association between per capita sugar consumption and diabetes prevalence, and mediation analysis to detect the mediated percentage of BMI on this association. RESULTS: Each increase of 100 kcal/day per capita sugar consumption was associated with a 1.62% higher diabetes prevalence [adjusted β-estimator (95% CI): 1.62 (0.71, 2.53)]. Mediation analysis using BMI as the mediator demonstrated an adjusted direct association of 0.55 (95% CI: − 0.22, 1.32) and an adjusted indirect association of 1.07 (95% CI: 0.54, 1.68). Accordingly, the BMI explained 66% (95% CI: 34%, 100%) of the association between per capita sugar consumption on diabetes prevalence. CONCLUSIONS: These findings indicate that the association between dietary sugar intake and the occurrence of diabetes is mediated by BMI to a large proportion. However, it seems that other mechanisms may explain the association between sugar consumption and development of type 2 diabetes. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (10.1007/s00394-020-02401-2) contains supplementary material, which is available to authorized users.