Additional key phrases: turbidimetry; nephelometry; protein reference standardsThere is increasing interest in the determination of serum apolipoproteins (apo) A-I and B, because of the evidence indicating that they may be more useful laboratory indices of coronary heart disease than quantification of serum lipids. The lack of standardized methods, common reference ranges and clinical decision values are delaying the widespread use of apo A-I and B measurements. The possibility of obtaining accurate and comparable apo A-I and B values and reference limits has been enhanced by the development of reference materials for these apclipoproteins.l? In previous work, we devised a turbidimetric method for apo A-I and apo B with a Cobas Bio analyser and calculated reference intervals in a sample of 494 apparently healthy subjects.' In this study we have converted these intervals into values that may be obtained using three commercial methods for apo A-I and B which are calibrated against the International Federation of Clinical ChemistryWorld Health Organization (I FCC-WHO) reference standards for apolipoproteins. MATERIALS AND METHODSPatient samples Specimens (220) were collected over a period of 10 days from samples sent to the laboratory for routine analysis. These samples had apo A-I and B concentrations below (44 samples), above (44) or evenly distributed throughout the laboratory reference range (132). Haemolysed, icteric or turbid samples were excluded. Measurements of apo A-I and B of the serum samples were performed on the same day they were drawn, in 10 runs on different days by the four methods, Correspondence: Josefina Eugui. and the instruments were calibrated before each run. MethodsIn-house method A turbidimetric procedure developed for the Cobas Bio analyser,' using reagents obtained from Boehringer Mannheim SA (Barcelona, Spain). The standards had assigned values based on Centers for Disease Control (CDC) reference materials. Briefly, 5 ilL (for apo A-I) or 30 ilL (for apo B) of 51-fold pre-diluted serum was mixed with 200 ilL of the respective antibody (30-fold diluted for apo A-I, 40-fold for apo B). The assays were conducted at 30°C, and absorbance changes (290 nm) were measured between O' 5 sand 15 min for apo A-I, and between O' 5 sand 5 min for apo B. Sample concentrations were calculated using the four-parameter logistic model. The interday coefficients of variation (CV) for this method (n = 10) were 6·0% for apo A-I at 1·37 giL and 5·0% for apo B at 0·69 g/L, Our reference limits were established with this procedure in a sample of 246 women and 248 men, ages 20--83 years. ' The other three methods employed standards which had been calibrated by the manufacturers using the IFCC-WHO reference preparations SPI and SP3. In all the assays, the fourparameter logistic model was used, and the reaction temperature was 37°C. In brief, the assay conditions for these methods were as follows:BN Jl nephelometer (Behring Diagnosticos Iberica SA, Barcelona, Spain) Ten microlitres of a 20-fold automatical...