A novel interaction between dietary composition and insulin secretion: effects on weight gain in the Quebec Family Study

Background: Clinical trials of low-fat diets characteristically produce small mean long-term weight loss but a large interindividual variation in response. This variation has been attributed to psychological and behavioral factors, although biological differences may also play a role. Objective: The objective was to determine whether physiologic differences in insulin secretion explain differences in weight gain among individuals consuming low- and high-fat diets. Design: Of 276 individuals followed in the Quebec Family Study for a mean of 6 y, we compared those in the lowest with those in the highest dietary fat tertiles. We performed oral-glucose-tolerance tests at baseline and examined the insulin concentration at 30 min (insulin-30) as a proxy measure of insulin secretion. Six-year changes in body weight and waist circumference were the primary endpoints. We determined the associations between insulin-30 and the primary endpoints by linear regression analysis, with adjustment for potentially confounding factors. Results: Mean changes in body weight and waist circumference did not differ significantly between the lowest- and highest-fat diet groups. However, these endpoints were strongly associated with insulin-30, especially among individuals consuming the lowest-fat diet. Insulin-30 at baseline was significantly associated with 6-y weight gain (r = 0.5 1, P < 0.0001) and change in waist circumference (r = 0.55, P < 0.0001) in the lowest diet fat, group (r = 0.18, P = 0.086), but not in the highest diet fat group (r = 0.20, P = 0.058). Individuals in the highest insulin-30 and lowest dietary fat group gained 1.8 kg more than did those in the highest insulin-30 and highest dietary fat group (51%; P = 0.034); they gained 4.5 kg more than did those in the lowest insulin-30 and lowest dietary fat group (6.5-fold; P = 0.0026). Conclusion: A proxy measure of insulin secretion strongly predicts changes in body weight and waist circumference over 6 y in adults, especially among those consuming lower-fat diets, which demonstrates the existence of a novel diet-phenotype interaction.