Diagnostic Accuracy Study of Direct and Formula Based Method for Estimation of LDL Cholesterol

Abstract

Introduction: Increased LDL-C is a well-known risk factor for cardiovascular disease. Also, LDL-C is used for risk stratification and in planning appropriate treatment. This requires the accurate measurement of LDL-C. Though β quantification is the reference method for LDL-C estimation it can’t be used routinely. Even though homogenous assays are available, some clinical laboratories use the Friedewald formula, which has its disadvantages. Aim & Objectives: To identify which LDL-C formula best correlates with direct LDL-C among various levels of TG. Study Design: A retrospective diagnostic accuracy study. Materials and Methods: Complete lipid data of 15,094 was obtained from the hospital information system. Based on TG level data, it was grouped into TG <100 mg/dl (N=6022) and TG 100-200 mg/dl (9072). Direct homogenous assay was used to measure LDL-C. LDL-C was calculated using seven formulae (Friedewald, Hattori, de Cordova, Vujovic, Anandaraja, Chen, and Sampson). Results: Correlation between direct and formula-based LDL-C by ICC showed maximum correlation by Vujovic (0.978). Bland Altman plot between Vujovic and the direct method shows a bias of -0.38 and it may overestimate or underestimate by 23 mg/dl. Conclusion: LDL-C calculated using Vujovic showed maximum correlation with direct homogenous assay when TG <200 mg/dl.