Development of a liquid chromatography–mass spectrometry method for the high-accuracy determination of creatinine in serum

An LC–MS method for the high-accuracy determination of creatinine in serum has been developed and used to provide results for an international measurement evaluation programme (IMEP) and the Comité Consultatif pour la Quantité de Matière (CCQM) international inter-laboratory studies. An assessment of different sample preparation methods, including ion-exchange chromatography, solid-phase extraction, plasma ultrafiltration and ethanol protein precipitation, revealed that no bias or reduced precision was associated with the quicker less extensive clean-up methods, when using liquid chromatography–isotope dilution mass spectrometry (LC–IDMS) for quantitation. A number of different calibration regimes were also investigated. External calibration was shown to provide adequate calibration for most routine analysis with a relative associated expanded uncertainty (k=2) of 6% at the 95% confidence level. The use of a non-isotopically labelled internal standard was shown to improve the relative expanded uncertainty (k=2) to 4%. However, the difference in retention time between the internal standard and the creatinine was such that a matrix interferent produced an observed bias of over 16%. The use of an isotopically labelled internal standard was shown to reduce any bias to less than 0.2% with an expanded uncertainty (k=2) of less than 0.3%. The developed method was then used, in a blind trial organised jointly by IMEP and CCQM, to determine the amount of creatinine in human serum. The method performed well against the established reference method of ion-exchange chromatography followed by derivatisation gas chromatography (GC)–IDMS. The observed difference between the values determined by LC–IDMS and the key comparison reference value (average of all the submitted results) was less than 0.3%. The biggest advantage of the described method is in the speed of analysis. With a chromatographic run time of less than 10 min and sample preparation consisting of a simple protein precipitation, without the need for a derivatisation stage, the analysis is vastly simpler then the conventional GC–IDMS reference method.