A-214 Modified CDC Beta-Quantification Reference Method for HDL-C and LDL-C Determinations Using Isotope-Dilution Mass Spectrometry—Pilot Report

Abstract

Abstract Background It is known that there is average positive 1.6% bias of total cholesterol by modified Abell-Levy-Brodie-Kendall (AK) colorimetric method, which have been the traceable reference measurement procedure (RMP) used by most of clinical guidelines, comparing to those by isotope-dilution mass spectrometry (ID-MS), the most accurate RMP. The CDC beta-quantification (BQ) reference method is currently the only RMP registered JCTLM database for high-density lipoprotein cholesterol (HDL-C), and low-density lipoprotein cholesterol (LDL-C) using the AK colorimetric assay after ultracentrifugation. We compared the modified BQ method using ID-MS assay after ultracentrifugation with the original CDC BQ reference method of HDL-C or LDL-C. Methods For this study, 18 commutable serum pools according to the CLSI 37A and 6 external quality assessment survey materials operated by Cholesterol Reference Method Laboratory Network (CRMLN) were used. We used the same ultracentrifugation procedure as in the original CDC BQ reference method, and the bottom fraction cholesterol (BFC) and HDL-C from supernatant after apo B-containing lipoprotein precipitation using heparin/manganese were measured for each sample using both AK and ID-MS methods, then LDL-C was calculated as the difference between BFC and HDL-C. The comparability analysis between methods at multiple medical decision levels (MDL), and Passing-Bablok regression analysis for each analyte were performed. Results Each Passing-Bablok regression equation of modified CDC BQ reference method vs. original CDC BQ reference method for three fractions, BFC, HDL-C, and LDL-C, were as follows; BFC was y (IDMS) = 0.996x (AK) - 0.342 (R = 0.999), for HDL-C was y (IDMS) = 1.007x (AK)-1.037 (R = 0.995), and for LDL-C was y (IDMS) = 0.990x (AK) + 0.860 (R = 0.999). Both cholesterol results by two methods showed excellent correlation and agreement of the three fractions showing regression R larger than 0.995 and slope or intercept containing 1 and 0 in their 95% confidence interval. The mean differences of IDMS against AK method were −0.56% for BFC, −0.51 mg/dL for HDL-C, and −0.35% for LDL-C, respectively. When the regression prediction at MDLs, 40/50/60 mg/dL for HDL-C and 100/130/160 mg/dL for LDL-C were estimated, all the differences were confirmed to be within allowable level claimed by CRMLN member laboratory accreditation criteria, ±1mg/dL for HDL-C and ±2.0% for LDL-C, showing equivalent results against CDC BQ reference method, but average negative bias of −0.72%, −1.59%, and −0.28% for BFC, HDL-C, and LDL-C, respectively are still present. These results of the modified BQ method using ID-MS assay may be promising, because the tedious and less accurate AK-assay can be switched to the ID-MS assay, not sacrificing trueness LDL-C or changing clinical guidelines, but quite large negative bias of HDL-C for some samples should require further investigation. Conclusion We got comparable HDL-C and LDL-C results using modified BQ method using ID-MS assay with those by original CDC BQ reference method using AK colorimetric assay. The quite large negative bias of HDL-C for some samples require large-scale additional further multi-center study among CRMLN network laboratories to evaluate the results observed in our study.