Prediction of plasma lipids in a cross-sectional sample of young women

We studied 60 premenopausal women to identify the best predictors of plasma lipid levels. Eligible subjects gave one venous blood sample during menstruation, kept a 3-day food record, completed a demographic questionnaire, had their height and weight measured, and participated in a standardized physical activity interview. Dietary constituents, anthropometric measures, serum progesterone and estradiol concentrations, and leisure-time activity level were then examined for relationships to plasma lipids. Pearson correlations revealed associations among diet, adiposity, alcohol intake, and plasma lipids. Saturated fatty acid (SFA) intake was positively related to plasma total cholesterol, low-density-lipoprotein (LDL) cholesterol, and high-density-lipoprotein (HDL) cholesterol. Height and weight were significantly related to total triglycerides. Furthermore, in the multivariate analyses, dietary constituents and height or body weight explained more of the variance in plasma lipids than did physical activity level or sex hormones. The final models for each lipid are as follows: for plasma total cholesterol and LDL cholesterol, SFA intake was the only significant predictor; for HDL cholesterol, refined carbohydrate and SFA intake; and for total plasma triglycerides, body weight and crude fiber intake.