Calculation of LDL-cholesterol from total cholesterol, triglyceride and HDL-cholesterol: a comparison of methods in the Jerusalem Lipid Research Clinic Prevalence Study.

Methods for estimating the cholesterol content of the plasma very low-density lipoprotein (VLDL-C) and low-density lipoprotein (LDL-C) fractions were examined in an ethnically diverse population sample from the Jerusalem Lipid Research Clinic Prevalence Study. None of these methods required the use of preparative ultracentrifuge. The often-used method for estimating VLDL-C by multiplying the triglyceride (TG) value by 0.20 was compared with two other models in which linear regression functions were computed for VLDL-C (determined by ultracentrifugation) as the dependent variable and TG as the independent variable. Subsequently, the intercept and the slope coefficients were used for estimation of VLDL-C and LDL-C. In the first regression model, overall age and sex regression coefficients were applied, whereas in the second regression model, the coefficients were calculated within different TG ranges. All three models were similar in their predictive accuracy when applied to the total population sample. However, in subjects with high TG levels, the second regression model was preferable.