Context .- Cystatin C is becoming an increasingly popular biomarker for estimating glomerular filtration rate, and accurate measurements of cystatin C concentrations are necessary for accurate estimates of glomerular filtration rate. Objective .- To assess the accuracy of cystatin C concentration measurements in laboratories participating in the College of American Pathologists CYS Survey. Design .- Two fresh frozen serum pools, the first from apparently healthy donors and the second from patients with chronic kidney disease, were prepared and distributed to laboratories participating in the CYS Survey along with the 2 usual processed human plasma samples. Target values were established for each pool by using 2 immunoassays and ERM DA471/IFCC international reference material. Results .- For the normal fresh frozen pool (ERM-DA471/IFCC-traceable target of 0.960 mg/L), the all-method mean (SD, % coefficient of variation [CV]) reported by all of the 123 reporting laboratories was 0.894 mg/L (0.128 mg/L, 14.3%). For the chronic kidney disease pool (ERM-DA471/IFCC-traceable target of 2.37 mg/L), the all-method mean (SD, %CV) was 2.258 mg/L (0.288 mg/L, 12.8%). There were substantial method-specific biases (mean milligram per liter reported for the normal pool was 0.780 for Siemens, 0.870 for Gentian, 0.967 for Roche, 1.061 for Diazyme, and 0.970 for other/not specified reagents; and mean milligram per liter reported for the chronic kidney disease pool was 2.052 for Siemens, 2.312 for Gentian, 2.247 for Roche, 2.909 for Diazyme, and 2.413 for other/not specified reagents). Conclusions .- Manufacturers need to improve the accuracy of cystatin C measurement procedures if cystatin C is to achieve its full potential as a biomarker for estimating glomerular filtration rate.