Accuracy and reproducibility of cholesterol assay in the western Cape

The accuracy and precision of cholesterol assay in the western Cape region is reported. The survey was carried out over 15 weeks utilising three human EDTA plasma pools with normal, borderline high and high cholesterol levels respectively. All 11 laboratories in the region providing a service to academic, provincial or military hospitals or to the private medical sector were included in the study. Ten of the 11 laboratories utilised automated enzymatic methods of cholesterol assay whereas 1 used a manual procedure based on the Liebermann-Burchard reaction. Methods were standardised by means of a variety of commercial calibrator material in all except 1 laboratory which used reference sera from the Centers for Disease Control, Atlanta. The performance of the 4 best laboratories met the standard of precision recommended for cholesterol assay, viz. total coefficient of variation of less than or equal to 2.5%. However, only 2 of the 11 laboratories achieved the optimum objective of an overall bias of less than 2.0% together with precision of less than or equal to 2.5%. Rational use of cholesterol assay for diagnosis and management will therefore require standardisation of cholesterol assay on a common reference material and greater attention to analytical factors influencing the reproducibility of results. Intrinsic biological variation also contributes uncertainty to the interpretation of a single value. Thus important clinical decisions must be based on two or more assays carried out using appropriate methodology.