No Dose Response Relationship in the Effects of Commonly Consumed Sugars on Risk Factors for Diabetes across a Range of Typical Human Consumption Levels

Abstract

Questions have been raised as to whether dietary carbohydrate intake is directly related to the development of type 2 diabetes. Of particular importance, fructose-induced insulin resistance has been previously shown in animals. However, the implications of such findings for humans are unclear as these models typically use very high doses of sugars and from sources not commonly consumed. Little is known about how the typical consumption of sugar in humans affects risk factors for diabetes. 355 weight-stable (weight change < 3% in previous 30 days) individuals aged 20 - 60 years old drank sugar-sweetened low fat milk every day for 10 weeks as part of their usual diet. Added sugar was provided in the milk as either high fructose corn syrup or sucrose at 8%, 18% or 30% of the calories required to maintain body weight. Insulin resistance was measured using the Homeostasis Model Assessment (HOMA IR) on fasting measures and a standard Oral Glucose Tolerance Test (OGTT) was used to measure insulin and glucose areas under the curve resistance (AUC30 g * AUC30 I) and whole body insulin sensitivity and hepatic insulin resistance using the Matsuda Composite Insulin Sensitivity Index (ISI). There was a small increase in weight in the entire cohort (169.1 ± 30.6 vs 171.6 ± 31.8 lbs, p < 0.001), which was greater in the 30% level than in the 8% or 18% levels (p < 0.05). Glucose, insulin, HOMA, glucose AUC, insulin AUC, Matsuda insulin sensitivity index, and hepatic insulin resistance did not vary by sugar level (p > 0.05) nor by sugar type (p > 0.05). In the entire cohort insulin sensitivity decreased as evidenced by an increase in HOMA IR (1.8 ± 1.3 vs 2.3 ± 3.4, p < 0.01) and a decrease in the Matsuda ISI (13.1 ± 21.3 vs 11.6 ± 16.1, p < 0.05). Hepatic insulin resistance was unchanged (2.4 ± 1.7 vs 2.4 ± 1.7 p > 0.05). Neither sugar level nor sugar type had any effect on any of these three measures (interaction p > 0.05). These data show that risk factors for diabetes do not vary between the 8% (25 th percentile), and the 30% group (95 th percentile) although insulin sensitivity may be affected by sugar consumption across a wide range of typical consumption levels. Importantly, the type of sugar (HFCS versus sucrose) had no effect on any response.