Comparison of three homogeneous methods for measuring high-density lipoprotein cholesterol with a precipitation method in diabetic and non-diabetic subjects.

Three homogenous methods for the automated measurement of high-density lipoprotein-cholesterol (HDL-C) are studied and compared with a precipitation method in diabetic and non-diabetic subjects. All three methods meet the precision criteria of the National Cholesterol Education Programme (NCEP). Triglycerides did not cause significant interference up to 10 mmol/L in the immuno-inhibition method (Wako) and up to 40 mmol/L in the Randox and Roche methods. Haemoglobin, up to a concentration of 5 g/L, had only negligible effect on the performance of all three homogeneous methods. Bilirubin caused an increasing positive bias in all methods above a concentration of 50 micromol/L. In comparison with the precipitation method, the new homogeneous methods agreed for type 1 diabetic patients but showed a positive bias for the control subjects and patients with type 2 diabetes. The bias of HDL-C levels in type 2 diabetes may be sufficient to affect the calculation of cardiovascular risk, and may therefore influence the decision to prescribe lipid-lowering medication.