J Clin Endocrinol Metab. 2022 Dec 31:dgac758. doi: 10.1210/clinem/dgac758. Online ahead of print.
ABSTRACT
AIMS/HYPOTHESIS: Hyperinsulinemia and adiposity are associated with one another, but the directionality of this relation is debated. Here, we tested the direction of the causal effects of fasting insulin (FI) levels, body fat accumulation/distribution using two-sample bidirectional Mendelian randomization (MR).
METHODS: We included summary statistics from large-scale genome-wide association studies for body mass index (BMI, n=806,834), waist-to-hip ratio adjusted for BMI (WHRadjBMI, n=694,649), abdominal subcutaneous, visceral, and gluteofemoral adipose tissue (n=38,965), FI levels (n=98,210), pancreatic islets gene expression (n=420) and hypothalamus gene expression (n=155). We used inverse variance-weighted and robust MR methods that relied on statistically and biologically driven genetic instruments.
RESULTS: Both BMI and WHRadjBMI were positively associated with FI. Results were consistent across all robust MR methods and when variants mapped to the hypothalamus (presumably associated with food behaviour) were included. In multivariable MR analyses, when waist circumference and BMI were mutually adjusted, the direct effect of waist circumference on FI was 2.43 times larger than the effect of BMI on FI. FI was not associated with adiposity. By contrast, using genetic instruments mapped to gene expression in pancreatic islets (presumably more specific to insulin secretion), insulin was positively associated with BMI and abdominal subcutaneous and gluteofemoral adipose tissue, but not with visceral adipose tissue.
CONCLUSIONS/INTERPRETATION: Although these results will need to be supported by experimental investigations, results of this MR study suggest that abdominal adiposity may be a key determinant of circulating insulin levels. Alternatively, insulin secretion may promote peripheral adipose tissue accumulation.
PMID:36585897 | DOI:10.1210/clinem/dgac758
