Association between glycemic index, glycemic load, and fructose with insulin resistance: the CDC of the Canary Islands study

The involvement of carbohydrates in triggering insulin resistance (IR) remains a source of controversy. To study the relation between glycemic index (GI), glycemic load (GL), and fructose with insulin resistance in a predominantly rural population in the Canary Islands. Cross-sectional study carried out in 668 nondiabetic people aged 18-75. IR was estimated with serum glucose and C-peptide (HOMA2-IR). Nutrient intakes were obtained from a validated food frequency questionnaire. ANOVA was used to analyze nutrient distribution across quartiles of HOMA2-IR. Four multivariate nutrient density models (dependent variable: log-transformed HOMA2-IR) which differed only in the kinds of carbohydrates included were tested (Model 1: carbohydrates; Model 2: GI and then GL; Model 3: free fructose, other simple sugars and starch; Model 4: total fructose, remaining sugars and starch). There was no association between GI and IR. There was a direct association between GL (P < 0.001), fructose (free [P = 0.001], total [P = 0.013]), energy intake (P < 0.001), fruit fiber (< 0.001), and glucose (P = 0.003) with IR. There was an inverse association between cereal (P = 0.008) and vegetable fiber (P < 0.001) and IR. Multivariate models corroborated the association of carbohydrates, GL, fructose, vegetable fiber, and energy intake with IR. The association between GL and IR disappeared when Model 2 was adjusted by total fructose intake. There was a direct association between fructose intake and IR. There was no relationship between GI and IR. Although a direct association of GL with IR was detected, it was attributable to the consumption of fructose.