Comparison of equations for the calculation of LDL-cholesterol in hospitalized patients

Background: The Friedewald equation is widely used to calculate LDL-C for cardiovascular risk prediction but is less accurate with comorbidities and extreme lipid values. Several novel formulae have been reported to outperform the Friedewald formula. Methods: We examined 14,219 lipid profiles and evaluated four formulae (Friedewald, Chen, de Cordova, Hattori) and compared these to direct measurement of LDL-C across various triglyceride (TG), total cholesterol (TC) and HDL-cholesterol (HDL-C) ranges using Beckman reagents and instruments. Linear regression and ROC analysis were performed. Results: The de Cordova formula showed a high correlation with directly measured LDL-C (r = 0.90, P < 0.001), comparable to the Friedewald calculated values for directly measured LDL-C (r = 0.95, P < 0.001). The de Cordova formula was favorable in some ranges of HDL, TC and the lowest TG range (r = 0.97, P < 0.001) but performed least well in comparison with the three other LDL-C calculations (AUC = 0.8331), demonstrating inconsistent bias. The Chen formula performed better than Friedewald (AUC = 0.9049). The Hattori formula outperformed all formulae including Friedewald over various ranges of lipid values (AUC = 0.9097). Conclusions: We observe favorable correlations of the de Cordova formula with Friedewald at low TG values. However, the Hattori formula appears to be best for application in hospitalized patients, even at extreme lipid values.