COMPARISON OF 2 HOMOGENEOUS HIGH-DENSITY-LIPOPROTEIN CHOLESTEROL ASSAYS

Background.-High-density lipoprotein cholesterol (HDL-C) is an indepen dent inverse risk factor for coronary artery disease. Current methodol ogies for measurement of HDL-C in most clinical laboratories involve c hemical precipitation-based methods. However, these methods are time-c onsuming, affected by high triglycerides, are not suitable for complet e automation, and require a large sample size. New direct homogeneous methods are now available that do not have these constraints. Design.- We evaluated the performance of 2 direct homogeneous methods, Liquid N -geneous HDL-C assay (LN-HDL) and Boehringer Mannheim HDL Cholesterol assay (BM-HDL), and compared these methods against a modified Centers for Disease Control and Prevention reference method (MR-HDL) in 126 pa tients with normotriglyceridemia (triglycerides < 4.5 mmol/L, range 0. 6-4.3 mmol/L) and 50 patients with hypertriglyceridemia (triglycerides greater than or equal to 4.5 mmol/L, range 4.5-18.8 mmol/L). Results. -Excellent precision profiles were exhibited by both homogeneous metho ds. Both LN-HDL and BM-HDL correlated well with MR-HDL in normotriglyc eridemia (r = 0.98, slope = 0.93 and r = 0.97, slope 1.0, respectively ). However, compared with the modified reference method, the LN-HDL co rrelated better than the BM-HDL in hypertriglyceridemic samples (r = 0 .97, slope = 1.0 and r = 0.91, slope = 0.9, respectively). The 1998 Na tional Cholesterol Education Program guidelines for accuracy (bias < /-5%) were met by LN-HDL in both normotriglyceridemic and hypertriglyc eridemic samples (bias 1.3% and 3.3%, respectively); however, BM-HDL f ailed to meet the National Cholesterol Education Program accuracy crit eria in both triglyceride subgroups (bias = 8.2% and 11.3%, respective ly). In addition, the total error for LN-HDL in both normotriglyceride mia (6.6%) and hypertriglyceridemia (8.6%) was well within the Nationa l Cholesterol Education Program guidelines for total error (less than or equal to 13%); BM-HDL exhibited a higher total error than LN-HDL in normotriglyceridemia (11.9%) and failed to meet the National Choleste rol Education Program guidelines in hypertriglyceridemia (15.0%). Conc lusion.-Although both homogeneous methods are precise, the LN-HDL assa y is superior in accuracy to the BM-HDL assay when compared with the m odified reference method.