Variations of cholesterol and total lipid concentrations in sera of healthy young men. Differentiating analytic error from biologic variability.

Three major sources of variation affecting serum cholesterol and serum total lipid concentration values were studied in a group of healthy men. The first source of variation, analytic error, was separated into pre-instrumental and instrumental components. The pre-instrumental component, which consisted of all uncertainties occurring from the instant of venipuncture to the entrance of the sample into the instrument, was statistically significant for total lipids. The second major source of variation-preparation of the subject-was evaluated for four major stresses: diet, exercise, posture prior to venipuncture, and duration of tourniquet application. Diet affected total lipids, while the specifics of posture and tourniquet application were statistically significant for both serum cholesterol and total lipids. The third source of variation considered was the within-day and the day-to-day changes. Within-day changes were divided into within-hour and hour-to-hour effects. Day-to-day changes occurred over a two-week period. The magnitude and statistical significance of the changes were evaluated using an ANOVA model. The C.V.'s of day-to-day changes for cholesterol and total lipids (biologic) were 5.3% and 10.2%, respectively while the hour-to-hour within-day C.V.'s were 3.8% for cholesterol and 10.4% for serum total lipids. In all cases the biologic variation was in far excess of the analytic variation. The significance of these factors affecting intraindividual variation in healthy subject is discussed.