| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Letters to the Editor |
1 Laboratorio Analisi Chimico Cliniche 1 Azienda Ospedaliera "Spedali Civili", Brescia, Italy
aAddress correspondence to this author at: Laboratorio Analisi Chimico Cliniche 1, Azienda Ospedaliera "Spedali Civili", 25125 Brescia, Italy. Fax 39-030-3995369; e-mail panteghi{at}bshosp.osp.unibs.it.
To the Editor:
In recent years, meaningful efforts toward standardization of cardiac markers have been initiated by several organizations (1). The final objective of these standardization projects should be the promotion of result traceability by means of a reference measurement system by providing reference measurement procedures and reference materials (2). The progress in standardization of cardiac protein immunoassays has, however, been slow. As an interim solution, some authors have proposed assay harmonization by recalibrating various assays to give the same results (3). In this case, results may, however, be biased in terms of trueness.
A major prerequisite for guaranteeing comparability of results among different methods is the availability of suitable reference materials, appropriately and thoroughly defined by a set of characteristics (2). Reference materials can be used for calibration of routine methods, but when reference materials are intended for direct value assignment to manufacturers’ calibrators, they should be extensively investigated for commutability (2). Commutability has been defined as the ability of a reference/control material for a given analyte to show interassay properties comparable to those of the same analyte in human serum (4). To directly calibrate manufacturers’ methods, matrix-based materials are desirable because they are more likely to behave in a similar fashion to test samples, but this does not eliminate a priori the matrix problem because different immunoassays may have different matrix problems (5).
Some years ago, a project of the AACC subcommittee on creatine kinase MB (CK-MB) mass assay standardization led to identification and characterization of a lyophilized CK-MB candidate reference material (in recombinant form) (6). Although this material did not have a certified concentration value nor any uncertainty statement, it appeared effective in obtaining assay harmonization, achieving a reduction in the systematic error among commercial methods from 40% to 13% during a pilot experiment (6). The material was not tested for commutability, but providers suggested that it be diluted in the manufacturer’s diluent to compensate for matrix effects. The European Community recently issued a new directive to be implemented by December 2003, requiring that calibration of all in vitro diagnostic assays be traceable to the highest available reference material or procedure (7). To fulfill this directive, some manufacturers recently decided to use the AACC CK-MB material (available from Seradyn, Indianapolis, IN) to recalibrate their CK-MB assays, considering it as the highest level CK-MB reference material currently available. In particular, Roche Diagnostics used the recombinant material starting from September 2002 to recalibrate the Elecsys CK-MB assay (cat. no. 11731432), and Beckman Coulter released in March 2003 a reformulated Access CK-MB assay (cat. no. 386371) standardized to the AACC material diluted in buffered bovine serum albumin-based matrix (8).
To evaluate the status of harmonization between the two assays after this realignment phase, we recently performed a comparison study between the Access and Elecsys CK-MB methods, using leftover sera from a group of patients with myocardial infarction for whom blood samples at peak CK-MB concentrations were obtained (9). A total of 45 samples, representing CK-MB concentrations in the calibration range of the assays (12–279 µg/L; Access values), were tested on each system in duplicate, subdivided on 2 different days (10). The analyzers were handled strictly according to the manufacturers’ instructions, and the manufacturers’ control samples were used to validate the analytical runs. The comparison data were analyzed by Deming regression and bias plots.
The obtained results are shown in Fig. 1
. Quite surprisingly, although good correlation was observed (r = 0.982; 95% confidence interval, 0.968–0.990; Fig. 1
, left-hand panel), the method comparison produced evidence of a highly significant proportional bias [slope (SE), 1.328 (0.087); P <0.0001]. A mean difference in percentage of 29% (95% confidence interval, 26–33%; bottom right-hand panel in Fig. 1
) confirmed the significant systematic bias in the correlation involving the two assays without relationship with the CK-MB concentration. The antibody specificities of the two assays are identical, and we observed no constant bias related to antibody specificity [intercept (SE), –0.6 (5.3) µg/L; P = 0.61].
|
Lacking a CK-MB reference measurement procedure, it is currently impossible to demonstrate that one of the two assays produces true CK-MB values. The problem is, however, that results coming from the two recalibrated assays are clearly not comparable. In spite of limited experimental data, a significant difference in the behavior of the AACC material when used as working calibrator in two CK-MB systems can be assumed, which would give a measurement relative error. Even if instability of the material or different handling by the two assay manufacturers cannot be excluded, apparently a major issue includes commutability of the material with biological samples in the evaluated assays. Two recently published studies showed the perverse effect of recalibrating immunoassays measuring cardiac or tumor markers with noncommutable materials and the consequent misinterpretation of patient results (11)(12). As demonstrated previously for another proposed CK-MB reference material (13), our results show that the AACC CK-MB standardization material cannot be used to directly calibrate commercial assays without rigorous verification of its commutability among procedures. In particular, convertibility of material results between two assays requires experimental demonstration of the same constant ratio between results obtained for the reference material and for a large number of patients’ samples (5).
Acknowledgments
We thank Beckman Coulter (Chaska, MN) for loan of the Access apparatus and the gift of Access CK-MB reagents. We gratefully acknowledge the skillful technical assistance of Francesca Stefini.
References
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  M. Panteghini Standardization of Cardiac Troponin I Measurements: The Way Forward? Clin. Chem., September 1, 2005; 51(9): 1594 - 1597. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
