Suitability of various control sera for use in testing the accuracy of cholesterol determination with enzymic kits.

We used 20 different commercial enzymic kits to measure cholesterol in 19 commercial control sera and in a pooled specimen of human serum and compared the relative biases, with the Abell -Kendall procedure as reference. Our purpose was to select those sera in various enzymic kits showing behavior similar to that of human serum and which thus can be used to measure the accuracy of these kits. The overall mean relative biases obtained with each of the 20 kits for the pooled human serum on the one hand and for a given control serum on the other generally correlated significantly. On the basis of the correlation coefficients and regression equations, we could select the control sera best suited for measurement of accuracy. They were all human-serum based, with cholesterol concentrations greater than 5 mmol/L. Animal sera with above-normal lipid values produced by feeding special diets appeared to be less reliable control sera in this respect.