Analytical performance of a direct assay for LDL-cholesterol

We evaluated a direct assay for the determination of LDL-cholesterol (LDL-C) L-Type assay, Wako Pure Chemicals in two laboratories. This assay is applicable to most random access clinical chemistry analyzers, allowing full automation. Between-run coefficient of variation (NCCLS EP5) varied between 1.29% and 3.13% and thus met the National Cholesterol Education Program (NCEP) goal. The assay was considered linear over a physiologically relevant range of LDL-C, 2.22 to 7.04 mmol/l (NCCLS EP6). Method comparison yielded identical results at both evaluation sites for LDL-C when assayed with the direct method. LDL-C results obtained with the homogeneous method under investigation (y) differed significantly from values from density-gradient ultracentrifugation (x) according to Chung (y = 0.87x + 0.43 mmol/l, s(yx) = 0.38 mmol/l, r = 0.91). With the latter method as a reference method, mean bias was 3.16% meeting the NCEP criteria. Diagnostic performance was excellent at a clinically relevant cut-off level of 3.37 mmol/l. Results of the direct method (y) and the commonly used Friedewald formula (x) were highly correlated (s(yx) = 0.22 mmol/l, r = 0.97), but both slope and intercept differed significantly from one and zero respectively (y = 0.90x + 0.37 mmol/l). Bilirubin, hemolysis and ascorbate did not interfere; triglycerides did not cause clinically relevant interference below 11.3 mmol/l. The direct method we investigated is user-friendly and provides an improvement in the determination of LDL-C in routine laboratories.