MULTICENTRE VALIDATION STUDY OF INSTRUMENT APPLICATIONS FOR %CDT, AN IMMUNOASSAY FOR QUANTIFICATION OF CARBOHYDRATE-DEFICIENT TRANSFERRIN IN SERUM{dagger}

Anders Helander,*

Department of Clinical Neuroscience, Karolinska Institutet, and Division of Clinical Chemistry, Karolinska Hospital, Stockholm, Sweden

Received 23 November 2001; in revised form 7 February 2002; accepted 11 February 2002


    ABSTRACT
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
— This multicentre trial assessed the inter-laboratory transferability and agreement of results for five instrument applications (Immage, BN A, BN II, Cobas Mira and Microtiter) of the Axis–Shield %CDT kit, a new version immunoassay for quantification of the alcohol marker carbohydrate-deficient transferrin (CDT) in serum. Two %CDT kit controls and three authentic serum samples were compared by 14 laboratories in six European countries, and each application was evaluated at three study sites. The %CDT results showed an overall good agreement both within and between sites, although it was also demonstrated that the analysis might be biased due to site performance. The data indicate that transferability of the %CDT assay is high, and that the instrument applications may be used interchangeably in routine quantification of %CDT.


    INTRODUCTION
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
Carbohydrate-deficient transferrin (CDT) refers to an abnormal microheterogeneity of serum transferrin and is used for identifying subjects engaged in heavy consumption of alcohol and for monitoring abstinence (Stibler, 1991Go; Allen et al., 1994Go). Serum levels of CDT are elevated by continuous high alcohol intake and persist for 2–4 weeks or longer after cessation of drinking (Helander and Carlsson, 1996Go). The major advantage of CDT compared with traditional laboratory tests for chronic alcohol exposure, such as the liver function test {gamma}-glutamyltransferase, is its higher specificity for alcohol (Meerkerk et al., 1998Go). Reported causes for incorrect determination of CDT (i.e. false positives/negatives) in the identification of alcohol misuse include rare congenital disorders of glycosylation, genetic transferrin variants, body iron, and advanced non-alcoholic liver disease (Stibler, 1991Go; De Feo et al., 1999Go; DiMartini et al., 2001Go), although it should be noted that this risk is partly dependent on the specific assay used (Helander et al., 2001aGo). Several different methodologies are currently available for the quantification of CDT in serum, but the transferrin isoforms covered by each test may vary considerably (Arndt, 2001Go; Helander et al., 2001aGo). Ongoing activities for the standardization of CDT quantification are thus important to improve further the value of CDT as a sensitive and specific alcohol marker.

Immunological assays for CDT are very convenient and time-efficient in routine laboratories with a high specimen throughput. The immunoassays utilise an initial separation of the CDT isoforms on anion-exchange chromatography minicolumns. Most studies published to date have used the CDTect® radioimmunoassay (Axis-Shield ASA), which measures the sum of asialo, monosialo and a minor part of disialo transferrin in units, or mg, per litre of serum. However, having an abnormally high or low serum total transferrin concentration might render falsely high and falsely low CDT results, respectively, in identification of heavy drinking when expressing the CDT content as an absolute amount, but rarely in percentage of total transferrin (Helander, 1999Go). A new version of the Axis–Shield %CDT immunoassay for quantification of CDT normalized to the total transferrin concentration was recently introduced (Helander et al., 2001bGo). The new assay measures primarily the asialo, monosialo and disialo transferrin isoforms (i.e. the ‘classical' CDT isoforms), and was found to perform favourably compared with the CDTect and previous %CDT immunoassays (Anton et al., 2001Go; Helander et al., 2001bGo).

This multicentre trial was designed to document the inter-laboratory transferability and agreement of results for five instrument applications of the Axis–Shield %CDT new version turbidimetric immunoassay.


    MATERIALS AND METHODS
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
Instrument applications of the new %CDT kit for five widespread chemistry analyser systems (Beckman Coulter Immage®, Dade Behring BN® II and BN A, Roche Diagnostics Cobas Mira®, and a Microtiter plate version) were evaluated by 14 European hospital and university laboratories. Each site was responsible for one instrument application, except for one laboratory that performed two, and each application was evaluated at three study sites.

The study was carried out with two different production lots (A and B) of the Axis–Shield %CDT assay manufactured by Axis–Shield ASA (Oslo, Norway). Quantification was performed following the manufacturer's instructions in the package insert. The samples compared were two lyophilized %CDT kit controls (low and high %CDT level; reconstituted in deionized water) and three pools of authentic sera. Serum sample 1 was selected to have a low %CDT value, serum 2 to have a value close to the cut-off of the method, and serum 3 to have an elevated %CDT value. Tentative target %CDT values for the controls and serum samples set by a high-performance liquid chromatography (HPLC) method (Bean et al., 1998Go) were as follows (data from Axis–Shield ASA): low control = 2.2%, high control = 5.2%, serum 1 = 1.7%, serum 2 = 2.5%, and serum 3 = 4.0%. The reconstituted controls were stored at 2–8°C and the serum samples at –18°C or below until testing.

Calibration of the instruments was done for each run using the %CDT Calibrator set. Prior to the main assay precision study, the laboratory personnel carrying out the assays underwent a familiarization training where the kit controls of lots A and B were tested in six replicates per day on 3 days within a 10-day period. The acceptable range for %CDT results was defined as the target value ± 20%, and the total coefficient of variation (CV) should be <10%. Based on these results, individual laboratory reference ranges for the low and high kit controls were established as the mean values ± 3 SD.

Once the familiarization training was successfully completed, the assay precision study was started in which controls and serum samples were run with both kit lots in four replicates per day on 4 days (separate aliquots of the sera were used) within a 10-day period. Finally, the reproducibility of the instruments was determined by measuring one low and one high kit control in nine replicates in one run. Statistical calculations were carried out using MedCalc software.


    RESULTS AND DISCUSSION
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
The individual results of the main assay precision study with the low and high controls and three serum samples are given in Table 1Go. There was mostly good agreement of %CDT results both within and between the different instrument models, and also between the two kit lots. However, systematic deviations were observed for some laboratories. One laboratory evaluating the Microtiter application (site 1) constantly reported very high %CDT values, and the results for serum samples 1 and 2 were identified as outliers by statistical analysis (box-and-whisker plots). Another laboratory (site 1 for the BN II application) obtained very low %CDT results for the low control with both kit lots. Two laboratories (one Immage and one BN II application) were excluded from the calculations, because the assay precision study was started before the training results met the acceptance criteria of a total CV <10%. Here it should be pointed out that their results from the main study did not differ markedly from the other sites (Table 1Go).


View this table:
[in this window]
[in a new window]
 
Table 1. Results of the multicentre instrument application study for two production lots (A and B) of the Axis–Shield %CDT immunoassay
 
The results of the instrument reproducibility study mostly yielded CV values well below 10% (data not shown), with the Cobas Mira application showing the lowest inter-laboratory variation. However, the BN A application gave striking differences in reproducibility between study sites, the instrument CV range being 2.4–12.1% in one laboratory compared with 1.2–2.7% in another.

The %CDT mean values obtained with the different instrument applications of the immunoassay differed from the tentative target values set by an HPLC method. For example, the %CDT mean values for all low and high controls (lots A and B combined) were 2.1 and 4.9%, respectively, compared with target values of 1.7 and 5.2%, and for the serum samples the immunoassay consistently produced higher values (Table 1Go). However, that %CDT results obtained by HPLC may deviate from the corresponding immunoassay values is well known (Bean et al., 1998Go; Helander et al., 2001aGo,bGo). This may result from differences in the separation of transferrin isoforms between the HPLC analytical columns and the immunoassay minicolumns, and also the integration mode (baseline or valley–valley) used in the HPLC chromatograms.

The relative imprecision (CV) of the %CDT results between all study sites and instrument applications ranged from 4.7 to 15.1%. As expected, the highest CV values were obtained at low %CDT values (i.e. low kit control and serum sample 1) (Table 1Go). When the outliers were omitted from these calculations, the CV values were typically <10% at all %CDT levels. In absolute terms, %CDT values of 2.4–3.3% (outliers omitted) were obtained for serum sample 2 which was selected to have a value close to the cut-off limit of the method. Accordingly, irrespective of using a cut-off of <2.6% (Axis–Shield %CDT Instruction manual; Anton et al., 2001Go) or <3.0% (Helander et al., 2001bGo), %CDT values at, below, as well as above, the threshold limit were obtained for this sample. In doubtful cases of a high %CDT result with the immunoassays, it is therefore recommended that the result be verified by HPLC, or possibly capillary electrophoresis or isoelectric focusing, to rule out any interference by genetic variants and isoform types of transferrin (Helander et al., 2001aGo). This is especially important whenever a positive test result could lead to serious consequences for the individual.

In summary, this standardized multicentre investigation evaluated the transferability and agreement of results for five instrument applications (Immage, BN A, BN II, Cobas Mira, and Microtiter) of the Axis–Shield %CDT new version immunoassay. The test results showed an overall good agreement both within and between different instrument models, although it was also demonstrated that the analysis might be biased due to site performance. Ongoing activities for the use of certified reference materials and proficiency testing, and development of an HPLC reference method, will help to improve further the quality of intra- and inter-laboratory test results and, hence, the value of CDT as an alcohol marker. To conclude, the evaluated %CDT applications for some of the most widespread chemistry analyser systems showed adequate performance, and may thus be used interchangeably in routine quantification of %CDT in serum.


    ACKNOWLEDGEMENTS
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
Participating laboratories: France: Hospital of Nancy; Hospital of Clermont-Ferrand; Hôpital Fernand Vidal, Paris. Germany: Diagnostisches Zentrum, Kaiserslauten; University of Greifswald; Universitätsklinik Homburg; Klinikum der FSU Jena; Klinikum Chemnitz GmbH. Italy: Laboratoria Analisi, Alessandria. Norway: Fylkesykehuset, Molde. The Netherlands: Diagnostisch Centrum, Delft; Ziekenhuis Eemland, Amersfoort. Sweden: Karolinska Hospital, Stockholm (main investigator); Länssjukhuset, Kalmar.

This study was supported in part by Axis-Shield ASA. The author has no other financial involvement with Axis-Shield related to this work, and have complete independence in the interpretation of data and writing of the report.


    FOOTNOTES
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
{dagger} See acknowledgements for a full list of the participating laboratories. Back

* Alcohol Laboratory, L7:03, Karolinska Hospital, SE-171 76 Stockholm, Sweden. Back


    REFERENCES
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
Allen, J. P., Litten, R. Z., Anton, R. F. and Cross, G. M. (1994) Carbohydrate-deficient transferrin as a measure of immoderate drinking: remaining issues. Alcoholism: Clinical and Experimental Research 18, 799–812.[ISI][Medline]

Anton, R. F., Dominick, C., Bigelow, M. and Westby, C. (2001) Comparison of Bio-Rad %CDT TIA and CDTect as laboratory markers of heavy alcohol use and their relationships with {gamma}-glutamyltransferase. Clinical Chemistry 47, 1769–1775.[Abstract/Free Full Text]

Arndt, T. (2001) Carbohydrate-deficient transferrin as a marker of chronic alcohol abuse: a critical review of preanalysis, analysis, and interpretation. Clinical Chemistry 47, 13–27.[Abstract/Free Full Text]

Bean, P., Husa, A., Liegmann, K. and Sundrehagen, E. (1998) Semi-automated carbohydrate-deficient transferrin in primary biliary cirrhosis: a pilot study. Alcohol and Alcoholism 33, 657–660.[Abstract]

De Feo, T. M., Fargion, S., Duca, L., Mattioli, M., Cappellini, M. D., Sampietro, M., Cesana, B. M. and Fiorelli, G. (1999) Carbohydrate-deficient transferrin, a sensitive marker of chronic alcohol abuse, is highly influenced by body iron. Hepatology 29, 658–663.[ISI][Medline]

DiMartini, A., Day, N., Lane, T., Beisler, A. T., Dew, M. A. and Anton, R. (2001) Carbohydrate deficient transferrin in abstaining patients with end-stage liver disease. Alcoholism: Clinical and Experimental Research 25, 1729–1733.[ISI][Medline]

Helander, A. (1999) Absolute or relative measurement of carbohydrate-deficient transferrin in serum? Experiences with three immunological assays. Clinical Chemistry 45, 131–135.[Free Full Text]

Helander, A. and Carlsson, S. (1996) Carbohydrate-deficient transferrin and gamma-glutamyl transferase levels during disulfiram therapy. Alcoholism: Clinical and Experimental Research 20, 1202–1205.[ISI][Medline]

Helander, A., Eriksson, G., Stibler, H. and Jeppsson, J. O. (2001a) Interference of transferrin isoform types with carbohydrate-deficient transferrin quantification in the identification of alcohol abuse. Clinical Chemistry 47, 1225–1233.

Helander, A., Fors, M. and Zakrisson, B. (2001b) Study of Axis–Shield new %CDT immunoassay for quantification of carbohydrate-deficient transferrin (CDT) in serum. Alcohol and Alcoholism 36, 406–412.

Meerkerk, G. J., Njoo, K. H., Bongers, I. M., Trienekens, P. and van Oers, J. A. (1998) The specificity of the CDT assay in general practice: the influence of common chronic diseases and medication on the serum CDT concentration. Alcoholism: Clinical and Experimental Research 22, 908–913.[ISI][Medline]

Stibler, H. (1991) Carbohydrate-deficient transferrin in serum: a new marker of potentially harmful alcohol consumption reviewed. Clinical Chemistry 37, 2029–2037.[Abstract/Free Full Text]