A Multilaboratory Commutability Evaluation of Proficiency Testing Material for Carbamazepine and Valproic Acid: A Study Within the Framework of the Dutch Calibration 2000 Project

K. Robijns*, N.W. Boone, A.W.H.M. Kuypers, R.T.P. Jansen, C. Neef, D.J. Touw

*Corresponding author for this work

Research output: Contribution to journalArticleAcademicpeer-review

Abstract

BACKGROUND:: Medical laboratories are required to participate in interlaboratory comparisons of the analyses they perform. The materials used in these comparisons need to be of sufficient quality so the comparison provides a picture of the performances. One of the main characteristics of the testing material is commutability, which is the ability of a material to yield the same numerical relationships between results of measurements as those relationships obtained when the same procedures are applied to patient samples. The aim of this study was to assess the commutability of three different matrices for the preparation of proficiency testing material (PTM) for the analysis of carbamazepine and valproic acid. METHODS:: Patient samples and PTM containing various concentrations of carbamazepine and valproic acid were collected and prepared and shipped to different laboratories for analysis. Reported results for patient samples from each laboratory were plotted against results for patient samples of each of the other laboratories and the corresponding regression line was calculated. The distance of results from PTM to the regression line is a measure for commutability. The distance is expressed as a multiple of the SDwl (average within-laboratory SD as calculated from EQAS results) and referred to as relative residual. A commutability decision limit of 2 SDwl was set. RESULTS:: For carbamazepine and valproic acid a total of 78 and 105 laboratory couples respectively could be formed. The number of relative residuals for liquid human serum outside the commutability decision limit were 1, 4 and 0 for low, medium and high concentrations of carbamazepine respectively and 3, 1 and 0 for low, medium and high concentrations of valproic acid respectively. In both liquid and lyophilized bovine serum, the number of relative residuals outside the commutability decision limit were between 2 and 15 and between 6 and 21 for carbamazepine and valproic acid respectively. CONCLUSIONS:: Even though not all results for PTM with carbamazepine and valproic acid are within the commutability decision limits, a preference for human serum can be seen.
Original languageEnglish
Pages (from-to)445-450
Number of pages6
JournalTherapeutic Drug Monitoring
Volume37
Issue number4
DOIs
Publication statusPublished - Aug 2015

Keywords

  • LINEAR-REGRESSION PROCEDURES
  • CLINICAL-CHEMISTRY
  • CROSS-REACTIVITY
  • IMMUNOASSAYS
  • CARBAMAZEPINE-10,11-EPOXIDE
  • PETINIA
  • DRUGS

Cite this