Abstract
AIMS/HYPOTHESIS: Fasting hyperinsulinaemia is a key feature of obesity and is implicated in diabetes progression. However, the following aspects of insulin secretion remain unclear: (1) which index of obesity is most important; (2) what is the shape of the dose-response curve between obesity and insulin secretion; (3) what physiological mechanisms sustain insulin hypersecretion; (4) what are the underlying causes; and (5) whether sex-related differences exist. METHODS: We analysed data from 1250 healthy participants (547 men, 703 women) of the EGIR-RISC cohort followed up for 3.5 years, with age 30-60 years and BMI 18.5-40.0 kg/m(2). Assessments included body composition, insulin secretion, beta cell function modelling from an OGTT and clamp-derived insulin sensitivity. Endogenous glucose production (EGP) was measured in a subset of 368 participants. Multivariable regression models and stratifications for BMI, body fat per cent, WHR and fat mass were applied to evaluate the effect of obesity on insulin secretion and beta cell function. RESULTS: The impact of obesity on fasting insulin secretion (FIS) was continuous across the full spectrum of BMI and WHR values and was greater in men than women. Among adiposity indices, fat mass (standardised β coefficient [Stβ] 0.27, p<0.0001) and waist circumference (Stβ 0.21, p<0.0001) were the strongest predictors of FIS. Insulin secretion increased 2.4-fold across BMI deciles, and adiposity-associated insulin hypersecretion appeared to be driven by the combination of hyperglycaemia and an increase in a specific beta cell function variable (insulin secretion rate at 5 mmol/l glucose [ISR@5]). In the follow-up cohort, weight gain (mean ± SD ∆ weight=+5.1 ± 3.8 kg) was associated with an increase in FIS and fasting glucose (+0.20 ± 0.63 mmol/l, p<0.03), whereas weight loss (-4.7 ± 2.8 kg) led to a reduction in FIS and fasting glucose (+0.06 ± 0.55 mmol/l, p<0.006). ISR@5 declined in both weight losers and those with stable weight (-0.17 ± 1.9 and -0.16 ± 1.0 U/h, respectively; p<0.002 for both) but not in weight gainers (-0.06 ± 1.1 U/h). Peripheral insulin resistance, plasma NEFA and leptin accounted for only part of obesity's effect on insulin secretion. Subset analysis of fasting and clamp EGP data suggested a rightwards shift in the dose-response curve across fat mass quintiles, indicating progressive hepatic glucose overproduction despite a preserved hepatic insulin response. CONCLUSIONS/INTERPRETATION: The effect of body mass on insulin secretion is continuous, more pronounced in men, driven by fat mass and waist, sustained by hyperglycaemia and by an upregulation of beta cell insulin secretion and is only partially explained by typical hormonal and metabolic consequences of obesity. We suggest that hepatic glucose overproduction contributes to the fasting hyperinsulinaemia observed in individuals with obesity.