The French Society of Clinical Biochemistry conducted this study to compare the accuracy and performances of the best creatinine enzymatic assays and the compensated Jaffe methods from the same manufacturers. Creatinine was measured in 3 serum pools with creatinine levels of 35.9+/-0.9 mumol/L, 74.4+/-1.4 mumol/L, and 97.9+/-1.7 mumol/L (IDMS determination). The performances of the assays (total error that includes the contribution of bias and imprecision) were evaluated using Monte-Carlo simulations and compared against desirable NKDEP criteria. The enzymatic assays always fell within the desirable total Error of 7.6%. By contrast, this requirement was never obtained for the compensated Jaffe methods at the critical level of 74.4+/-1.4 mumol/L. Only the compensated Jaffe creatinine on Olympus analyzer reached this specification at 35.9+/-0.9 and 97.9+/-1.7 mumol/L levels. This study demonstrates that, despite substantial improvement regarding traceability to the IDMS reference method and precision, compensated Jaffe creatinine methods, by contrast to enzymatic ones, do not reach the desirable specifications of NKDEP at normal levels of creatinine.