Reliability and correlation study of a new homocysteine assay.

BACKGROUND: Hyperhomocysteinemia is a risk factor for cardiovascular, cerebrovascular diseases and is also associated with congenital birth defects, pregnancy complications and cancer. Many homocysteine assays have been developed since the 1960s and most companies have commercially available kits for homocysteine determination. In this paper, we report the performance characteristics of the new Architect homocysteine assay and the correlation with HPLC and other routine analytical methods. METHODS: Recovery, limit of detection (LOD), total imprecision and interferences were evaluated; furthermore, the method was compared with HPLC and with two routine methods. RESULTS: Recovery was higher than 95%, LOD was lower than 1.0 micromol/L, the range for total imprecision was 1.7%-3.8%. No significant degree of interference was observed up to 5 g/L hemoglobin, 20 mg/dL (342.5 micromol/L) bilirubin and 60 g/L lipids. For the comparison study, this new method shows a lightly positive bias with HPLC (mean+/-SD: 0.03+/-1.7 micromol/L) with R(2)=0.96, a positive bias with the immunonephelometric method (mean+/-SD: 1.7+/-1.2 micromol/L) with R(2)=0.98, and a negative bias with the enzymatic method (mean+/-SD: -1.8+/-2.2 micromol/L) with R(2)=0.94. CONCLUSIONS: This new method from Abbott can be included among the methods suitable for homocysteine routine analysis.