A new accurate, simple formula for LDL-cholesterol estimation based on directly measured blood lipids from a large cohort

Background: Low-density lipoprotein cholesterol (LDL-c) is the major measured parameter for cardiovascular risk assessment. The generally accepted formula (LDL-F) for estimating LDL-c developed by Friedewald and colleagues in 1972 using data from 448 individuals suffers from known inaccuracies at extremes of triglyceride (TG) and total cholesterol (TC) values. Methods: We generated new formulas based on a large Brazilian database containing directly measured lipid values from 10,664 fasted individuals. This database LDL-c was measured by the LDL-C Select FS (DiaSys) system, a homogeneous method without centrifugation. The formulas were generated using linear and non-linear approaches, and the formula with the highest accuracy and simplicity for general clinical use was selected. Results: The simple formula LDL-c = 3/4 (TC - HDL-c) provided an accurate estimate of LDL-c, a higher correlation with directly measured LDL (r = 0.93) compared with LDL-F (r = 0.87), and also a higher accuracy. Conclusions: The new formula outperformed several other LDL-c formulas over a wide range of TC, HDL-c and TG values. The validation and application of this formula in other populations is warranted.