Evaluation of five different high-density lipoprotein cholesterol assays: the most precise are not the most accurate.

We evaluated the accuracy and performance of four different test kits for the direct determination of high-density lipoprotein (HDL)-cholesterol and compared them with the phosphotungstic acid/MgCl2 assay. All four homogeneous assays were precise (within-run CV of < 2.0% and between-run CV of < 6.4%); both assays based on immuno-inhibition had the lowest CVs (within-run 1.3% and 0.9%; between-run 2.3% and 2.2%). Interference from haemolysis was negligible, but triglyceride concentrations gave a negative interference. The effects of conjugated and unconjugated bilirubin were opposite; conjugated bilirubin showed a negative interference of up to 40%; unconjugated bilirubin interfered positively up to 50%. Using the recently validated indirect phosphotungstic acid/MgCl2 method as a comparison, all four homogeneous assays did not fulfil the National Cholesterol Education Program total error standard, mainly due to the positive biases of 12 to 42%, apparently associated with improper calibrators. Both assays involving immuno-inhibition showed a concentration-dependent bias.