| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
1 Laboratorium voor Analytische Chemie and
3
Laboratoria voor Medische Biochemie en Klinische Analyse, Faculteit Farmaceutische Wetenschappen, Universiteit Gent, Harelbekestraat 72, B-9000 Gent, Belgium
2 Klinisch Laboratorium, AZ Middelheim, Lindendreef 1, 2020 Antwerpen, Belgium
aauthor for correspondence: fax 32-9-264-8198, e-mail linda.thienpont{at}rug.ac.be
In the past, standardization of measurements of diagnostically important polypeptides and proteins was hampered by the noncommutability of primary standards (1). One way to overcome this problem is to establish a method comparison with a reference measurement procedure. Until now, reference measurement procedures, such as isotope-dilution mass spectrometry (ID-MS), have been scarce. Recent developments in the MS field, however, have made the technique easily applicable to the analysis of polypeptides and proteins (kinetic studies, sequence analysis, and determination of molecular mass and posttranslational modifications). To the best of our knowledge, only two groups have used ID-MS for the quantitative determination of a specific polypeptide/protein. One of these groups described the measurement of apolipoprotein A-I after enzymatic digestion (2), the other described offline ID-liquid chromatography (LC)-MS assays for serum proinsulin, insulin, and C-peptide (3)(4). Neither group, however, examined the potential of ID-MS for standardization of the respective routine test systems (usually, immunoassays).
Here we report on the use of an ID-MS measurement procedure for standardization/recalibration of C-peptide measurements in urine by use of a method-comparison study with split-sample measurements. In view of the model character of the study for future applications, we chose urinary C-peptide over the clinically more important serum C-peptide because of the ease of sample collection and MS measurement. The measurement procedure applies online ID-LC-electrospray tandem MS (ID-LC-MS/MS) and is described in detail elsewhere (5). For calibration, it makes use of a commercial C-peptide preparation with a peptide content of 89% and a purity by HPLC of >99% (according to the manufacturer’s information). This purity was taken into account for calculation of the C-peptide content in the calibrators. The C-peptide preparation was delivered in a vial containing 250 µg of freeze-dried material; a calibration solution was prepared by carefully weighing the added volume (
1 mL) of a solution containing, per liter, 10 g of protease-free bovine serum albumin (BSA). The exact dissolution volume of the vial contents could be derived from the gravimetric data and the density of the BSA solution. From this stock solution, we prepared a series of 1:5 gravimetrically diluted solutions in the same BSA solution to obtain working solutions of respectively 50 and 10 mg/L. Immediately after preparation, the working solutions were divided in 100-µL portions in plastic vials and frozen at -20 °C until the day of analysis. Each day of analysis, an aliquot was thawed and gravimetrically diluted 1:10 with the 10 g/L BSA solution to obtain working solutions of 1 mg/L. Once diluted, the aliquots were never reused.
The C-peptide immunoassays evaluated in the study (see the acknowledgements) were all performed in the application laboratories of the respective manufacturers/distributors for Belgium with the exception of one manual test, which was performed in the routine clinical laboratory of one of the authors. All of the manufacturers/laboratories strictly followed the prescribed assay protocols, with particular care for adequate internal quality control. The results are therefore considered representative for the respective assays.
Random urine specimens were collected during 4 weeks from 45 apparently healthy male and female volunteers between 15 and 65 years of age, according to the standards of the Committee for Medical Ethics of the Ghent University. The C-peptide concentrations in the collected samples were spread across a wide range. With respect to the collection conditions, we had to adhere to the assay instructions, which do not foresee special sample treatment except for storage at -20 °C if the analysis is not performed within 24 h. Precautions were taken to ensure that all samples had been handled uniformly. Directly after collection, the samples were stored in bulk at -20 °C and kept at that temperature until analysis, which started 4 weeks after the panel had been assembled.
Just before the start of the measurements, the samples were thawed (and cleared by centrifugation, if necessary), aliquoted in portions of 1 mL, and stored again at -20 °C. They were then shipped frozen (on dry ice) to the measurement sites. They were thawed on the day of analysis and never reused. All measurements were performed within 2 weeks. Thus, storage time differences at the time of analysis were, at maximum, 2 weeks at -20 °C. Although reports on the stability of C-peptide under these conditions are conflicting (6), our experience with ID-MS gave no indication that samples deteriorated in that short period. Nevertheless, deterioration of specific samples can never be definitively excluded. From the data below, however, we conclude that if slight sample deterioration had taken place, it had no influence on the results of the study. The samples were measured by the immunoassays in one run, in duplicate. For the ID-LC-MS/MS measurement procedure, however, measurements were done in triplicate, spread over 6 measurement days.
The results of the method comparison study are presented in Fig. 1A
. Because some methods (Eurogenetics and Biosource C-PEP-RIA New) visually did not correlate with the ID-LC-MS/MS measurement procedure in a linear way, we performed second-order regression analysis. The regression equations and correlation coefficients are presented in Table 1
. As can be seen in Fig. 1A
, there were considerable systematic differences between the results obtained with the immunoassays and the ID-LC-MS/MS measurement procedure and, more importantly, among the immunoassays themselves. Indeed, with the ID-LC-MS/MS measurement procedure, the mean C-peptide concentration was 40.5 µg/L, whereas the respective immunoassays gave values of 32.6 µg/L (Biosource C-PEPsp-RIA), 61.9 µg/L (DPC), 71.8 µg/L (Eurogenetics), 76.1 µg/L (Diasorin), and 98.0 µg/L (Biosource C-PEP-RIA New test). These data clearly indicate the existence of severe problems in the standardization process of C-peptide immunoassays.
|
|
The intermethod differences appear striking as all manufacturers used the same primary calibrator (IRP 84/510) for assigning values to the secondary calibrators in their immunoassays. The IRP 84/510 certification report, however, had already described systematic differences among immunoassay results (7), irrespective of the use of a local standard or the IRP for calibration. Consequently, the present study shows once again that standardization and/or recalibration of immunoassays in general is difficult to achieve with a common calibrator alone but needs split-sample measurements with a reference measurement procedure. This stems from the well-known fact that immunoassays may suffer from differences in behavior related to differences between the matrices of calibrators and those of real samples (1).
Standardization/recalibration makes sense only when sufficient correlation of the investigated immunoassay is achieved with the comparison measurement procedure (or sufficient specificity). As shown in Table 1
, this prerequisite was fulfilled in the present study because all assays had an excellent second-order correlation with the ID-LC-MS/MS measurement procedure (0.980 < r < 0.992). The outcome of the recalibration of the test systems, which was done by use of the respective second-order regression equations (Table 1
), is shown in Fig. 1B
. As can be seen from Fig. 1B
, all test systems were successfully recalibrated by this approach.
Standardization cannot solve specificity/interference problems in immunoassays. Indeed, in this study, three samples showed considerable sample-related effects in the immunoassays because of properties of the matrix. Investigation of the reasons to which the phenomenon can be attributed was beyond the scope of this study.
The present ID-LC-MS/MS measurement procedure does not yet have the status of a reference measurement procedure. It is not part of a complete reference system comprising an international primary calibrator such as the IRP 84/510. As described, the ID-LC-MS/MS measurement procedure used here was calibrated with a commercial C-peptide preparation. This was done, on the one hand, for economic reasons; on the other hand, it was also done because we did not consider the current C-peptide IRP superior to the commercial calibrator. The IRP, being a recombinant product, contains an impurity of 10%, and the ampoule content was only nominally assigned the value of 10 µg by comparison with several commercial C-peptide preparations. Ideally, a primary calibrator should be certified in terms of mass units by techniques such as amino acid analysis and MS.
Another reason that the ID-LC-MS/MS measurement procedure used here cannot be claimed as a reference measurement procedure is that, as part of the validation process, it should be assessed in a round-robin trial for its ability to satisfy predefined performance specifications in terms of trueness/accuracy and precision. Such an assessment should preferably be endorsed by an authoritative organization (8)(9).
In conclusion, the present study demonstrates that a method comparison with a real reference measurement procedure would be the ideal basis for standardization/recalibration of test systems. Alternatively, it can guide manufacturers in judging whether their assays are sufficiently specific and need recalibration. If this is done, one would also naturally measure the calibrators to support the industry in their standardization process and to get information about assay commutability. For the latter, a deviation from the reference measurement procedure values in a defined mathematical way is required. The ID-LC-MS/MS for quantification of urinary C-peptide appears useful for recalibration of immunoassay test systems. It could thus provide an excellent basis for a reference measurement procedure.
Acknowledgments
We acknowledge the much appreciated collaboration of the following manufacturers of C-peptide routine test systems: Byk-Sangtec Diagnostica (Dietzenbach, Germany), with the C-peptide immunoassay performed automatically on a Liaison analyzer (Diasorin, Saluggia, Italy); DPC (Los Angeles, CA), with the assay performed on the Immulite 2000 automated analyzer; Tosoh (Minato-ku Tokyo, Japan), with the AIA-Pack C-peptide assay dedicated for performance on the AIA-21 automatic test system; and Biosource (Nivelles, Belgium), with two manual RIAs, the C-PEPsp-RIA and C-PEP-RIA New. Funding for the ID-LC-MS/MS work was provided by the Research Fund of the University of Ghent (Grant BOF 01102096) and the National Fund for Scientific Research (Grant 3G0001096).
References
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  A. E. Buyken, Y. Kellerhoff, S. Hahn, A. Kroke, and T. Remer Urinary C-Peptide Excretion in Free-Living Healthy Children Is Related to Dietary Carbohydrate Intake but Not to the Dietary Glycemic Index J. Nutr., July 1, 2006; 136(7): 1828 - 1833. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Anderson Candidate-based proteomics in the search for biomarkers of cardiovascular disease J. Physiol., February 15, 2005; 563(1): 23 - 60. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Groschl, M. Uhr, and T. Kraus Evaluation of the Comparability of Commercial Ghrelin Assays Clin. Chem., February 1, 2004; 50(2): 457 - 458. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
, Diasorin; +, DPC; 
, Eurogenetics; 
, Biosource C-PEP-RIA New; 
, Biosource C-PEPsp-RIA. Dotted line, y = x.