Objectives: Examine the reproducibility of plasma lipid and lipoprotein measurements in the HERITAGE Family Study.
Design and methods: In a sample of 379 subjects (191 men and 188 women), reproducibility was determined for lipids, lipoproteins (done on two occasions) and post-heparin lipase assays using an Intracenter Quality Control study by generating split samples from an additional 60 subjects (35 men and 25 women), which were assayed in a blind fashion by the lipid core laboratory. Reproducibility was estimated using intraclass correlation coefficients (ICC) for the selected variables. Analytical error (ANER) and coefficient of variation (CV) were also calculated. Day-to-day variation for 10 variables including plasma cholesterol and triglycerides (TG), HDL-cholesterol and its subfractions HDL2-cholesterol and HDL3-cholesterol, LDL-cholesterol and VLDL-cholesterol, as well as apoprotein (apo) A-I, apo B, and LDL-apo B were assessed.
Results: In the HERITAGE study, all lipid and lipoprotein variables had ICC above 0.79. Plasma VLDL-cholesterol (31 %) and TG (23%) levels, which are well known to be highly variable from one day to another, had CVs greater than 20%. Other variables had CVs lower than 10% except for HDL2-cholesterol which reached 16%. In the intracenter reliability sub-study, the measurement errors were found to be low except for HDL2-cholesterol. For the lipases, the reproducibility of repeated samples was very high, with ICC over 0.95. The within-assay CV corresponded to 2.1 and 5.3% for hepatic lipase (HL) and lipoprotein lipase (LPL), respectively, whereas the between-assay CV reached 8-12% for HL and about 15% for LPL. Due to the complexity of these two assays, the results are considered to be quite satisfactory.
Conclusions: The reproducibility of plasma lipid and lipoprotein measurements, as well as of post-heparin lipase activities, is good in the multicenter HERITAGE Family Study. In addition, the well-documented gender difference in the plasma lipoprotein profile was confirmed in the present study, women having lower fasting triglyceride and LDL-cholesterol levels than men as well as reduced cholesterol/HDL-cholesterol and increased HDL2-cholesterol/ HDL3-cholesterol ratios compared to men. Results of the present study support the notion that the higher LPL and low HL activities found in women compared to men are important factors contributing to explain gender difference in the lipoprotein profile. However, additional factors not examined in the present study are involved beyond the contribution of post-heparin lipase to the sex dimorphism in plasma lipoprotein levels.