A minimal model for human whole body cholesterol metabolism

Important work by others has shown that human whole body cholesterol metabolism can be described by a three-compartment model computed from plasma cholesterol specific activity after an intravenous infusion of labeled cholesterol. However, some parameters of that model cannot be estimated precisely [coefficient of variation (CV) 15-19% after 40 wk of follow-up], making its use in routine clinical investigation difficult. On the other hand, a simpler two-compartment model can be calculated with excellent precision from only 10 wk of data (CV 2-8%), but its parameters are inaccurate (for example, the size of the central pool is overestimated by 20%, and the rate constant for fractional excretion of cholesterol from the body is underestimated by 15%). Thus both three-compartment and two- compartment models of cholesterol turnover have important limitations. An alternative is provided by a minimal model that takes advantage of the increased precision expected in the solution of models with fewer parameters. A three-compartment structure is used, but only four (rather than 6 or more) parameters are calculated: the mass of the rapidly mixing central cholesterol compartment, the fractional rate of cholesterol elimination from the body, and the average forward and reverse rate constants for cholesterol transfer between the rapid compartment and both slower compartments. Each of these parameters can be determined unambiguously (without the need to use a minimum or maximum estimate), accurately (mean values within 2% of theory), and with precision (CV 3-13%). The minimal model is applicable to conventional cholesterol turnover studies lasting 40 wk, but most of the metabolic information available can be obtained in only 10 wk; such studies in 10 normal subjects gave expected parameter values. These results emphasize that both precision and accuracy are important considerations when choosing a working model of cholesterol turnover.