Long-term studies (32-49 wk) of the turnover of plasma cholesterol were conducted in 24 subjects. Eight subjects were normilipidemic, six had hypercholesterolemia, eight had hypercholesterolemia and hypertriglyceridemia, and two had hypertriglyceridemia alone. 10 of the hyperlipidemic patients had a definite familial disorder. In all subjects (except one for whom complete data were not available), the same three-pool model previously described gave the best fit for the data. The parameters of the three-pool model observed in the normal subjects were compared with the model parameters found in the patients with the different kinds of hyperlipidemia. In addition, single and multiple regression analyses were conducted to explore the relationships between the model parameters and various physiological variables, including age, body size, and serum lipid concentrations. Using this approach, significant differences between groups, or correlations with serum lipid levels were seen for several parameters of the three-pool model: the production rate (PR); the size of the rapidly exchanging pool 1 (M1); all estimates of the size of the most slowly equilibrating pool 3 (M3); and the rate constant k21. The PR in normal subjects (1.14 +/- 0.19 g/day, mean +/- SD) was not significantly different from that found in patients with hypercholesterolemia, with or without hypertriglyceridemia. The major determinant of cholesterol PR was overall body size, expressed either as total body weight or as surface area. The correlations between PR and indices of adiposity (percent ideal weight and excess weight), although statistically significant, were much weaker in this nonobese population. After adjustment for body size variation, cholesterol PR was not correlated with the serum cholesterol concentration but was probably (P less than 0.05) correlated with the triglyceride concentration. When the two patients with very high triglyceride concentrations were excluded, however, no correlation was observed between adjusted PR and triglyceride level. It is probable that hypertriglyceridemic patients represent a heterogeneous population, in which the majority do not show increased cholesterol PR. M1 was correlated with all body size variables, but most strongly with excess weight. After adjusting for the effects of body size, M1 was also correlated and triglyceride. Major differences were found in the relationships between the physiological variables and the sizes of pools 2 and 3. M2 was correlated neither with any of the indices of body size or adiposity, nor with the serum levels of either cholesterol or triglyceride. In contrast, all estimates of M3 were correlated with indices of adiposity (but not of overall body size) and with the serum cholesterol concentration. Thus, the amount of cholesterol in slowly equilibrating tissue sites appears to particularly increase with elevations of the serum cholesterol level. The results also confirm previous data that adipose tissue cholesterol is an important part of pool 3.