Interlaboratory proficiency survey of cholesterol and high-density lipoprotein cholesterol measurement.

We conducted a proficiency survey of cholesterol and high-density lipoprotein (HDL) cholesterol analysis in local clinical laboratories to determine whether increased national emphasis on cholesterol measurement had resulted in changes in performance from previous surveys. Sets of frozen aliquots of plasma and HDL supernate pools were sent to nine laboratories for analysis; results were compared with Northwest Lipid Research Center values, and relationships were determined by linear regression. Of all the cholesterol measurements, 81% were considered acceptable (i.e., within 9% of the NWLRC value), and 61% of the HDL cholesterol measurements were considered acceptable (within 50 mg/L of NWLRC values). These data represented no improvement over previous surveys. Workload had increased significantly: 79% for cholesterol and 284% for HDL cholesterol. On a bias plot, six of the laboratories demonstrated inaccuracy greater than the +/- 3% recommendation within the critical range of 2000-2400 mg/L for total cholesterol. For HDL cholesterol, two laboratories demonstrated bias greater than 10% at the critical point of 350 mg/L, with three additional laboratories displaying strong bias outside the decision point. The survey results indicate that apparently further improvements must be made for laboratories to achieve acceptable performance in cholesterol analysis.