RESUMEN
BACKGROUND: The significance of erythrocyte membrane fatty acids (EMFAs) and their ratios to predict hyperglycemia and incident type 2 diabetes is unclear. OBJECTIVE: We investigated EMFAs as predictors of the worsening of hyperglycemia and incident type 2 diabetes in a 5-y follow-up of a population-based study. DESIGN: We measured EMFAs in 1346 Finnish men aged 45-73 y at baseline [mean ± SD age: 55 ± 6 y; body mass index (in kg/m(2)): 26.5 ± 3.5]. Our prospective follow-up study included only men who were nondiabetic at baseline and who had data available at the 5-y follow-up visit (n = 735). RESULTS: Our study showed that, after adjustment for confounding factors, palmitoleic acid (16:1n-7; P = 2.8 × 10(-7)), dihomo-γ-linolenic acid (20:3n-6; P = 2.3 × 10(-4)), the ratio of 16:1n-7 to 16:0 (P = 1.6 × 10(-8)) as a marker of stearoyl coenzyme A desaturase 1 activity, and the ratio of 20:3n-6 to 18:2n-6 (P = 9.4 × 10(-7)) as a marker of Δ(6)-desaturase activity significantly predicted the worsening of hyperglycemia (glucose area under the curve in an oral-glucose-tolerance test). In contrast, linoleic acid (18:2n-6; P = 0.0015) and the ratio of 18:1n-7 to 16:1n-7 (P = 1.5 × 10(-9)) as a marker of elongase activity had opposite associations. Statistical significance persisted even after adjustment for baseline insulin sensitivity, insulin secretion, and glycemia. Palmitoleic acid (P = 0.010) and the ratio of 16:1n-7 to 16:0 (P = 0.004) nominally predicted incident type 2 diabetes, whereas linoleic acid had an opposite association (P = 0.004), and n-3 polyunsaturated fatty acids did not show any associations. CONCLUSION: EMFAs and their ratios are associated longitudinally with changes in glycemia and the risk type 2 diabetes.