Poster Topical Area: Nutritional Epidemiology
Location: Hall D
Poster Board Number: 761
Objective: To evaluate if dietary patterns (DP) are associated with glucose and lipid biomarkers in near-normal range healthy women from the NutriHS baseline.
Methods: This cross-sectional analysis included 190 women aged ≤45 yrs without diabetes or cancer. Diet was assessed by food frequency questionnaire. Total cholesterol (TC), HDL-c and triglycerides (TG) were measured by enzymatic colorimetric method and LDL-c was calculated. Fasting plasma glucose and insulin (immunoenzymatic assay) were used to estimate HOMA-IR. Factor analysis by principal components was used to obtain DP, analyzed in tertiles (T) taking T1 as reference. DP were exposures in multiple logistic regression models performed to estimate odds ratio (OR) to be classified in T3 of biomarkers [vs. T1+T2 (ref)]. Model 1 was adjusted for age, education, physical activity and smoking. Model 2 further included energy intake.
Results: Median age was 22 yrs and mean values of TC and glucose were 173.2±34.6 and 81.8±8.9 mg/dL, respectively. Processed, Prudent, Brazilian and Vegetarian DP were identified, which explained 27% of total variance. Women classified in T2 of Processed DP had higher chance to be classified in the higher T of glucose compared to those who were less adherent (T2: OR 2.95, P=0.013; T3: OR 2.16; P=0.13), and those in T3 showed higher chance to have higher TC (T2: OR 1.76, P=0.26; T3: OR 2.96; P=0.034) and LDL-c (T2: OR 1.83; P=0.19; T3: OR 5.34, P=0.002). Prudent DP was inversely associated with TG (T2: OR 0.35, P=0.027; T3: OR 1.22, P=0.69) and LDL-c (T2: OR 0.55; P=0.20; T3: OR 0.28; P=0.022) in all models and with LDL-c/HDL-c ratio only in model 2 (T2: OR 0.52; P=0.15; T3: OR 0.32; P=0.031). In contrast, the association with glucose verified in model 1 (T2: OR 0.37; P=0.025; T3: OR 0.35; P=0.016) was lost after adjustment for energy. Brazilian DP was inversely associated with LDL-c (T2: OR 0.26; P=0.003; T3: OR 0.53; P=0.22). Vegetarian DP was directly associated with LDL-c/HDL-c ratio in model 2 (T2: OR 1.40; P=0.45; T3: OR 2.73; P=0.029). DP were not associated with insulin or HOMA-IR.
Conclusion: Harmful associations of Processed DP with glucose, TC and LDL-c and favorable of Prudent and Brazilian DP with lipid biomarkers in healthy women reinforce the impact of dietary habits in metabolic parameters even when they are still within near-normal range.
School of Public Health/ São Paulo University
São Paulo, Sao Paulo, Brazil