EFFECTS OF DIETARY FIBER, CARBOHYDRATE, LIPID AND PROTEIN LEVELS ON SERUM AND LIVER LIPIDS IN THE RAT (CHOLESTEROL, TRIGLYCERIDES)

Response surface methodology was used to study the effects of selected combinations of four diet components (fiber, carbohydrate, lipid and protein) in the diet of male weanling Sprague-Dawley rats. Fiber (X) in the form of wheat bran, carbohydrate (X) from corn starch and sucrose, lipid (X) from beef tallow, corn oil, and safflower oil, and protein (X) from casein. Minimum serum and liver cholesterol and triglyceride levels (Y, Y, Y, Y) and maximum serum HDL cholesterol levels (Y) and feed efficiency (Y) were considered.;Because the dietary components (X s) were not statistically independent those variables (X s) were studied in combinations of three or less. Of the four dietary components, there were four possible combinations of three variables at a time, (3-X), and six possible combinations of two variables at a time, (2-X). All ten possibilities were studied.;The three-variable (3-X) combinations were the most useful in locating the desired maximum or minimum dependent variable responses (Y s) in terms of the proportions of the diet components. Response surface regression analyses indicated that of the dietary variables (X s), carbohydrate (X), lipid (X) and protein (X) were better than fiber (X) for predicting the dependent responses (Y s).;Surface response contour and three dimensional plots were developed for each of the dependent variables (Y s) except for serum triglycerides (Y), which was not predictable based on the proportion of dietary components. The contours were used to determine combinations of the diet components that would produce an overall diet that would generate desired responses of the dependent variables (Y s) in the desired maximum or minimum region. Generally, the statistical analyses represented by the response surface contour and three dimensional plots indicated that desired responses for the dependent variables (Y s) could be obtained with an overall diet low in bran NDF fiber (4%) and lipid (6-10%) and high in carbohydrate (54-55%) and protein (26-30%).;Serum triglycerides (Y) were unpredictable based on the proportions of the dietary components, as indicated by the low multiple correlation coefficient (maximum R = 0.078004). Therefore, data for serum triglycerides were analyzed further.