How to implement traceability of creatinine results: a manufacturer's experience.

Abstract

In 1997, Roche Diagnostics introduced an enzymatic creatinine test based on the enzymatic conversion of creatinine via creatine to sarcosine and on the quantification of sarcosine by a Trinder method involving sarcosine oxidase, peroxidase and aminoantipyrine. This method is a priori free of the well-known unspecific reactions of the classical ‘‘Creatinine Jaffe’’ method. Owing to the use of established techniques, it does not give a colour reaction with endogenous creatine and shows little or no interference from substances known to interfere with the Trinder reaction. As a consequence, a method comparison performed with 95 human sera gave a very good correlation with negligible intercept to the reference method isotope-dilution mass spectrometry (ID-MS; Figure 1). A comparable correlation could not be obtained in the past with the Roche/Hitachi Jaffe method (Figure 2). The discrepancy between the Jaffe and the ID-MS creatinine results is caused by the reaction of the alkaline picrate reagent with a variety of endogenous sample components in addition to that with creatinine. This results in an intercept in the method comparison that cannot be corrected by routine calibration procedures. However, the difference is fairly constant at approximately 26 mmol/L (0.3 mg/ dL) for serum or plasma samples over the entire measuring range. This finding led Roche Diagnostics to introduce the ‘‘Creatinine Jaffe compensated’’ method that automatically corrects each result by 224 mmol/L (20.27 mg/dL). Using this compensation in combination with a new standardization, an acceptable correlation with the reference method results is obtained with the Roche/Hitachi Jaffe method as well (Figure 3). While, with implementation of the Jaffe compensated method, the Creatinine routine methods showed satisfactory agreement with and the desired traceability to the reference method ID-MS, a new problem arose with estimation of the glomerular filtration rate (eGFR). Estimation of GFR is used in screening for chronic kidney disease. According to the recommendations of the National Kidney Disease Education Program (NKDEP), eGFR is calculated by the ‘‘four-variable MDRD Study equation’’ from the serum creatinine concentration and the other three variables: age, gender, race ([2], for review see [3]). A GFR of v60 mL/min per 1.73 m is considered to be highly indicative of impaired kidney function. However, on switching from the Roche/Hitachi Jaffe to the ID-MS traceable methods Creatinine enzymatic or Creatinine Jaffe compensated, the GFR calculated with the original MDRD Study equation increased considerably, i.e. from 58 to 65 mL/(min61.73 m) for an 18-year-old Caucasian female with a Creatinine concentration of 102 mmol/L (1.15 mg/dL) as determined with an ID-MS traceable method (Figure 4). The problem of a shift in GFR values upon switching to an ID-MS traceable Creatinine method was recognized and the so-called ‘‘re-expressed 4variable MDRD Study equation’’ was developed [4] and is presently recommended for use with ID-MS traceable Creatinine methods by the NKDEP. The re-expressed formula is based on a re-assay of the MDRD Study samples with the same kinetic alkaline picrate method but calibrated to the ID-MS traceable Roche/Hitachi Creatinine Plus enzymatic assay [4]. The re-expressed equation gives GFR values slightly lower than the old formula (e.g. 61 mL/ (min61.73 m) versus 65 mL/(min61.73 m) in the a.m. example), but still significantly higher than those that had been obtained before with the Creatinine Jaffe values (Figure 4). For many customers it is still hard to accept that, according to ID-MS traceable CREA results and to the re-expressed MDRD formula, a result of ‘‘GFR w60 mL/(min61.73 m)’’ has to be reported for a sample that would have been ‘‘eGFR558 mL/ (min61.73 m)’’ with the classical Jaffe method. In summary, despite the undeniable benefits of having a standardized method, the pain of the transition is fading only slowly today, 10 years later.