Causal Association of Obesity and Dyslipidemia with Type 2 Diabetes: A Two-Sample Mendelian Randomization Study

Abstract

Recent studies have suggested an association between obesity and dyslipidemia in the development of type 2 diabetes (T2D). The purpose of this study was to explore the causal effects of obesity and dyslipidemia on T2D risk in Asians. Two-sample Mendelian randomization (MR) analyses were performed to assess genetically predicted obesity using body mass index (BMI) and dyslipidemia using high-density lipoprotein cholesterol (HDL), low-density lipoprotein cholesterol (LDL), total cholesterol (TCHL), and triglycerides (TG) versus T2D susceptibility using genome-wide association study (GWAS) results derived from the summary statistics of Biobank Japan (n = 179,000) and DIAbetes Meta-ANalysis of Trans-Ethnic association studies (n = 50,533). The MR analysis demonstrated evidence of a causal effect of higher BMI on the risk of T2D (odds ratio (OR) > 1.0, p < 0.05). In addition, TG showed a protective effect on the risk of T2D (ORs 0.68-0.85). However, HDL, LDL, and TCHL showed little genetic evidence supporting a causal association between dyslipidemia and T2D. We found strong genetic evidence supporting a causal association of BMI with T2D. Although HDL, LDL, and TCHL did not show a causal association with T2D, TG had a causal relationship with the decrease of T2D. Although it was predicted that TG would be linked to a higher risk of T2D, it actually exhibited a paradoxical protective effect against T2D, which requires further investigation.

Keywords: Mendelian randomization; dyslipidemia; obesity; single-nucleotide polymorphisms; type 2 diabetes.

Figure 1

Schematic of the analytical study design. SNP: single-nucleotide polymorphism; HDL: high-density lipoprotein cholesterol; LDL: low-density lipoprotein cholesterol; TCHL: total cholesterol; TG: triglycerides; BMI: body mass index.

Figure 2

Forest plot of causal associations of BMI and dyslipidemia with T2D. BMI: body mass index; T2D: type 2 diabetes; SNP: single nucleotide polymorphism; OR: odds ratio; CI: confidence interval; IVW: inverse-variance weight; MR: Mendelian randomization; SIMEX: simulation extrapolation; PRESSO: polyhedral sum of residuals and outliers; HDL: high-density lipoprotein cholesterol; LDL: low-density lipoprotein cholesterol; TCHL: total cholesterol; TG: triglycerides.

Figure 3

Scatter plots of MR tests assessing the effect of BMI on T2D. The dots represent the effect size (β) of each SNP on BMI (x-axis) and T2D (y-axis), and the grey lines show their standard errors. Regression slopes show the estimated causal effect of BMI on T2D. The light blue, dark blue, light green, and dark green regression lines represent the IVW, MR-Egger, MR-Egger (SIMEX), and weighted median estimate, respectively. MR: Mendelian randomization; BMI: body mass index; T2D: type 2 diabetes; SNP: single nucleotide polymorphism; IVW: inverse-variance weight; SIMEX: simulation extrapolation.

Figure 4

Scatter plots of MR tests assessing the effect of dyslipidemia on T2D. (A) Causal effect of HDL on T2D; (B) causal effect of LDL on T2D; (C) causal effect of TCHL on T2D; and (D) causal effect of TG on T2D. The dots represent the effect sizes (β) of each SNP on dyslipidemia (x-axis) and T2D (y-axis), and the grey lines show their standard errors. Regression slopes show the estimated causal effect of dyslipidemia on T2D. The light blue, dark blue, light green, and dark green regression lines represent the IVW, MR-Egger, MR-Egger (SIMEX), and weighted median estimate, respectively. MR: Mendelian randomization; T2D: type 2 diabetes; HDL: high-density lipoprotein cholesterol; LDL: low-density lipoprotein cholesterol; TCHL: total cholesterol; TG: triglycerides; SNP: single nucleotide polymorphism; IVW: inverse-variance weight; SIMEX: simulation extrapolation.