Serum phospholipid fatty acids, dietary patterns and type 2 diabetes among urban Ghanaians

Franziska Jannasch, George Bedu-Addo, Matthias B. Schulze, Frank P. Mockenhaupt, Ina Danquah

Previously, a "purchase" pattern (rich in vegetable oil, manufactured foods, red meat and poultry, fruits, and vegetables) was identified among adults in urban Ghana and was inversely associated with T2D, while a "traditional" pattern (rich in fish, palm oil, plantain, green-leafy vegetables, beans, garden egg, fermented maize products,) increased the odds of T2D. To investigate, if specific fatty acids (FAs), partly reflecting the intakes of certain food groups and cooking methods, might explain the observed diet-disease relationships, serum phospholipid fatty acid profiles were characterized and their relationships with blood lipids that are common risk factors for T2D were analyzed. The relative proportions of 28 FAs (%) in 653 Ghanaians without T2D were measured by gas chromatography. In a cross-sectional analysis, the associations of FAs with dietary patterns and with serum lipids that are likely involved in T2D development were investigated. The FAs distributions across dietary pattern scores were examined. Standardized beta coefficients (β) were calculated for the associations of dietary pattern scores (per 1 standard deviation (SD) increase) with FAs. Across the tertiles of selected diet-related FAs, adjusted means of serum triglycerides, cholesterol, HDL-cholesterol and LDL-cholesterol were calculated. In this mainly female (76%), middle-aged (mean age: 46.4, SD: 15.3 years) and predominately overweight study population (mean body mass index: 25.8, SD: 5.4 kg/m(2)), saturated FAs (SFAs) contributed 52% to total serum FAs, n-6 polyunsaturated FAs (PUFAs) 27%, monounsaturated FAs 12%, n-3 PUFAs 9% and trans FAs (TFAs) <1%. The "purchase" pattern was related to lower proportions of n-3 PUFAs (β per 1 score SD: -0.25, p < 0.0001), but higher proportions of linoleic acid (LA) (β per 1 score SD: 0.24, p < 0.0001). The "traditional" pattern was characterized by lower proportions of arachidic acid (β per 1 score SD: -0.10, p = 0.001). LA was inversely associated with triglycerides, but positively with HDL-cholesterol and LDL-cholesterol. In this Ghanaian population, serum FA profiles reflected the intake of key components of dietary patterns, such as fish and vegetable oil. FAs from manufactured foods (SFAs) and deep-fried meals (TFAs) did not contribute to the observed associations between dietary patterns and T2D. Still, LA might partly explain the health-beneficial effect of the "purchase" pattern.