Associations of erythrocyte fatty acid patterns with insulin resistance

Synergistic or additive effects or both on cardiometabolic risk may be missed by examining individual fatty acids (FAs). A pattern analysis may be a more useful approach. In addition, it remains unclear whether erythrocyte FA composition relates to insulin resistance among Hispanics/Latinos.

We derived erythrocyte FA patterns for a Puerto Rican cohort and examined their association with diet and insulin resistance in cross-sectional and prospective analyses.

At baseline, principal components analysis was used to derive factor patterns with the use of 24 erythrocyte FAs from 1157 participants of the Boston Puerto Rican Health Study (aged 45-75 y). Dietary intake was assessed with a validated semiquantitative food-frequency questionnaire. The homeostasis model assessment of insulin resistance (HOMA-IR) was calculated at baseline and at the 2-y follow-up. Relations between FA patterns and HOMA-IR were analyzed in a sample of 922 participants with available data.

Five FA patterns were derived, differentiated by 1) relatively high de novo lipogenesis (DNL) FAs and low n-6 (ω-6) FAs, 2) high very-long-chain saturated FAs, 3) high n-3 (ω-3) FAs, 4) high linoleic acid and low arachidonic acid, and 5) high trans FAs. The DNL pattern was positively correlated with sugar and inversely with n-6 and monounsaturated FA intakes. Only the DNL pattern was positively related to baseline HOMA-IR [adjusted geometric means (95% CIs) for quartiles 1 and 4: 1.72 (1.58, 1.87) and 2.20 (2.02, 2.39); P-trend < 0.0001]. Similar associations were observed at 2 y, after adjustment for baseline status [quartiles 1 and 4 means (95% CIs): 1.61 (1.48, 1.76) and 1.84 (1.69, 2.00); P-trend = 0.02]. These results remained consistent after the exclusion of participants with diabetes (n = 485).

Our findings suggest that upregulated DNL associated with a diet high in sugar and relatively low in unsaturated FAs may adversely affect insulin sensitivity in a Hispanic/Latino cohort.