| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Enzymes and Protein Markers |
1
EP Central Hospital Laboratory, FIN-60220 Seinäjoki, Finland, and
2
Departments of Clinical Chemistry, University of Helsinki and Oulu, Fin-90220 Oulu, Finland.
a Author for correspondence. Fax 358-6-415-4924.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
x = 1.51, r = 0.744). CDTect
results correlated positively with serum transferrin
(r = 0.224, P <0.001), whereas the %CDT
TIA results showed a slight inverse correlation with serum transferrin
(r = -0.132, P = 0.07). The data
suggest that CDTect is more sensitive than %CDT TIA in detecting
drinking problems. However, the %CDT TIA method yields more
specificity when analyzing samples from patients with high serum
transferrin concentrations. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CDT consists of several subfractions with different amounts of sialylation (1)(3). To date, the most widely used method for measuring CDT in Europe has been based on ion-exchange microcolumn separation of the desialylated fraction from transferrin, followed by a transferrin radioimmunoassay for quantification (CDTect, Pharmacia & Upjohn). Transferrin isoforms have also been quantified by isoelectric focusing/immunoblotting and by HPLC (4)(5)(6), which was found recently to provide clinical sensitivity and specificity similar to those of the CDTect (7).
Recently, new approaches for CDT test kits have been introduced (8). In such modifications, the amount of CDT is reported as a relative amount to total transferrin (%CDT). Here, we report the first comparisons on the analytical characteristics and clinical value of the new semiautomated Axis %CDT TIA and the conventional CDTect method. Unlike in the CDTect method, the %CDT TIA determines both the desialylated fraction and total transferrin, using an automated turbidimetric procedure. These assays were shown to be markedly different with respect to the clinical value as alcohol markers.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The healthy controls were 42 volunteers (22 women and 20 men), who did not drink or who were social drinkers (<30 g of ethanol/day on any occasion). As additional controls, we analyzed 72 hospitalized nondrinking individuals (56 women and 16 men) who were patients with non-alcohol-related liver disease (n = 15), patients with abnormalities in their iron balance (n = 32), or pregnant women (n = 25). The nondrinking subjects with expected abnormalities in their serum transferrin concentrations were chosen for this study to obtain information on the effect of serum transferrin concentration on the specificity of the assays. One patient in the group with non-alcohol-related liver disease had increased serum transferrin. Among the patients with an abnormal iron balance, there were 20 patients with increased and 12 patients with decreased concentrations of serum transferrin. Among the pregnant women, there were 14 individuals with increased serum transferrin.
All serum samples were stored at -70 °C until analysis. All participants of the study gave their informed consent and the study was carried out according to the provisions of the Declaration of Helsinki.
cdt analyses
CDT was analyzed by two different methods. In the first method,
CDT was measured by anion-exchange chromatography followed by
radioimmunoassay using a commercially available assay kit (CDTect,
Pharmacia & Upjohn) according to the instructions of the manufacturer.
In this procedure, serum transferrin isoforms are separated by a
microcolumn, and the eluted transferrin fraction, which is deficient in
its carbohydrate moieties, is subsequently quantified by a
radioimmunoassay in which the CDT in the eluate competes with
I-labeled transferrin for antibody binding sites.
The reference range in this assay is 0–20 U/L for men and 0–26 U/L
for women.
In the second method, CDT was analyzed by the Axis %CDT turbidimetric immunoassay (%CDT TIA, Axis Biochemicals AS), in which serum transferrin is first saturated with Fe3 before the low sialic acid transferrin (the CDT) is separated by an ion-exchange chromatography minicolumn (8). In this procedure, the CDT content of the eluate and the total transferrin content of the Fe3-saturated serum sample are measured separately by turbidimetric measurement, using the same anti-transferrin antibodies. The measurements are evaluated using a calibration curve, and the %CDT value is calculated. According to the manufacturer, amounts exceeding 6% are considered increased. In the present study, a Kone Optima Analyzer (Kone Instruments) was used for the measurements. To allow comparisons with previously published material (9), measurements were also performed by the %CDT radioimmunoassay (%CDT RIA, Axis Biochemicals AS). The upper reference range limit for the %CDT RIA is 2.5%.
Whereas the %CDT TIA method measures asialylated, monosialylated, disialylated, and 50% of the trisialylated serum transferrin isoforms (8), the %CDT RIA measures transferrin variants with 0–2 sialic acid residues. In the CDTect assay, serum transferrin isoforms with pI values higher than 5.7 (mono- and asialotransferrins) and minor amounts of isotransferrin with pI values of 5.7 (disialotransferrin) are detected (10).
transferrin analyses
Serum total transferrin concentrations were measured by the
Array® Protein System (Beckman Instruments), which
measures nephelometrically the rate of light-scatter formation produced
by an immunoprecipitation reaction with the protein. The reference
range for transferrin is 1.7–3.4 g/L. The method is not affected by
the degree of transferrin desialylation.
statistical methods
Values were expressed as mean ± SD. The raw data of the CDT
methods were subjected to a logarithmic transformation to yield
gaussian nonskewed distributions, and ANOVAs were performed on the
transformed values followed by the Bonferroni's multiple comparisons
procedure to test for statistical differences among pairs of groups.
The differences were considered statistically significant at
P <0.05.
Linear regression analysis was used for the comparison studies. Additionally, the Bland-Altman plot (11) was used to monitor the agreement between the methods. To allow such comparisons between %CDT TIA and CDTect, the ratio CDTect/serum transferrin was first computed and transformed to the scale of %CDT TIA [referred to as (CDTect/Transferrin)'].
The 95% confidence intervals for observed indices were estimated
(p ± 1.96 x (SE), where SE =
, where n
>30, and p is the specificity or sensitivity) or the exact confidence
ranges were calculated according to Armitage and Berry (12),
when appropriate. The ROC plot areas ± SEs and the differences
between the areas were calculated according to Hanley and McNeil
(13)(14).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The values from the different subgroups of alcoholic and nonalcoholic
patients are shown in Fig. 1
. The mean %CDT TIA values in the alcohol abusers and in the
healthy controls were 5.4 ± 2.5% and 2.6 ± 0.8%
(mean ± SD), respectively. For the CDTect, the corresponding
values were 27.5 ± 13.8 U/L and 11.5 ± 3.6 U/L,
respectively. The differences were significant in both of the above
comparisons. There were no gender differences in the %CDT TIA or in
the CDTect values among alcohol abusers, whereas among the healthy
controls, the CDTect values were significantly higher (P
<0.01) in women. In the group of hospitalized nondrinking patients,
the mean values for the %CDT TIA and CDTect were 3.0 ± 0.9% and
19.9 ± 8.9 U/L, respectively, which were both significantly
higher than those of the healthy controls.
|
The slope and the intercept for linear regression between CDTect and
%CDT TIA results (with 95% confidence limits, n = 192) were 0.13
(0.12–0.15) and 1.16 (0.73–1.59), respectively. The
Syx was 1.51, and the correlation coefficient was
0.744 (Fig. 2
). A considerable disagreement between the %CDT TIA and the
CDTect results was also noted when they were examined by plotting the
%CDT TIA data and the CDTect/Transferrin' ratios according to the
method of Bland and Altman (11) (Fig. 3
).
|
|
The %CDT TIA method showed a significantly higher (P
<0.05) correlation with CDTect than the earlier %CDT RIA method
(r = 0.629, n = 112). Serum transferrin was found
to correlate with the %CDT TIA results in the subgroups of alcohol
abusers (r = -0.248, n = 90, P
<0.05), hospitalized nondrinking controls (r = 0.274,
n = 60, P <0.05), and healthy controls
(r = -0.297, n = 42, P <0.05),
although not in the total study group (r = -0.132,
P = 0.07; Table 1
). The correlation of %CDT RIA with serum transferrin
(r = -0.302, n = 112, P <0.01) was
higher (P = 0.07) than that of %CDT TIA. CDTect
results, in turn, were found to correlate significantly with serum
transferrin in the total study group (r = 0.239, n
= 192, P <0.001). The correlation was significantly higher
(P <0.05) in women (r = 0.425, n =
104, P <0.001) than in men (r = 0.098,
n = 100, not significant). The correlation between serum
transferrin and CDTect results was particularly strong in the subgroups
of hospitalized nondrinking patients and healthy controls
(r = 0.774, n = 72, P <0.001 and
r = 0.546, n = 42, P <0.001,
respectively).
|
The CDTect and the %CDT TIA results were also compared by ROC
analysis. The results for both genders were analyzed separately. For
men, the area under the curve (mean ± SE) was significantly
(P <0.05) higher for CDTect (0.990 ± 0.009) than for
%CDT TIA (0.941 ± 0.025), whereas for women, no significant
differences (0.923 ± 0.040 and 0.901 ± 0.045, respectively)
were found on the basis of the results obtained from the healthy
controls and alcohol abusers (Fig. 4
). The area under the ROC curve for CDTect results for men was
significantly higher (P = 0.05) than that for women,
whereas for %CDT TIA, no significant gender differences were found.
For the total study population of women, including the hospitalized
nondrinkers with high serum transferrin concentrations, the area under
the curve (mean ± SE) was significantly (P <0.05)
higher for %CDT TIA (0.861 ± 0.049) than that for CDTect
(0.740 ± 0.061). For the total study population of men, including
the hospitalized nondrinkers, no significant differences (0.921
± 0.027 and 0.899 ± 0.031, respectively) were found.
|
The sensitivities of the methods for detecting alcohol abuse, based on the cutoff limits recommended by the manufacturers of the present tests, were 29 ± 9% for the %CDT TIA and 59 ± 10% for the CDTect, respectively, for 100% (92–100%) specificity in the healthy controls. The %CDT RIA method yielded sensitivities of 34 ± 10%, for 100% (88–100%) specificity. For the hospitalized nondrinking patients, the %CDT TIA gave a specificity of 100% (94–100%), whereas that for the CDTect was 71 ± 10%. In the 35 nondrinking patients with increased serum transferrin, there were 21 (60%) false-positive values for the CDTect and none for the %CDT TIA. In the patients with non-alcohol-related liver disease, there was one (7%) false positive for the CDTect and none for the %CDT TIA. There were nine alcohol abusers with increased serum transferrin. In these subjects, %CDT TIA gave six (66%) and CDTect three (33%) false-negative results when the cutoff limits given by the manufacturers were used. When the cutoff limits based on the present healthy control group (mean 2 SD) were used, the sensitivities of the methods were 61 ± 10% for %CDT TIA and 86 ± 7% for the CDTect, respectively. The specificities were 98% (87–100%) and 95% (77–100%), respectively. However, with these cutoff limits, the specificities of the methods toward the hospitalized nondrinkers decreased to 88 ± 8% for the %CDT TIA and 53 ± 12% for CDTect.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Many recent reports have indicated that 15–30% of all hospital admissions in general hospitals are related to alcohol abuse (3)(8)(15)(16)(17). Because of the high prevalence of alcohol-related problems and because these problems are associated with serious health and social consequences, screening for alcohol-related problems is most important. However, the heterogeneity of alcohol disorders complicates the development of a "gold standard" that can be used to determine the predictive validity of screening tests. Studies have shown that alcohol-related health problems arise at a consumption rate of 60 grams (men) or 40 grams (women) of alcohol per day. Therefore, it is crucial to detect excessive drinking as the underlying cause of morbidity particularly in patients who are not obvious alcoholics. In the present study, we examined patients with a wide variety of alcohol-related problems to obtain a representative sample of consecutive admissions of patients with alcohol-related problems in general hospitals. This may also explain the present finding of lower sensitivities than those seen in many previous studies contrasting more extreme groups (3)(18)(19)(20).
In view of the present ROC analysis based on comparisons of alcohol abusers and healthy controls, the overall diagnostic performance of %CDT TIA is weaker than that of CDTect in detecting alcohol abuse in men, whereas for women, there is no significant difference. It should also be noted that, for CDTect, the ROC plot area is significantly higher for men than that for women, which is in line with previous observations on CDTect assays, which indicated higher sensitivities for men (7)(21)(22). However, the ability of CDTect to make diagnostic distinctions decreases markedly when comparisons are made between alcohol abusers and controls with increased serum transferrin concentrations. In nondrinking patients who have high serum transferrin concentration, CDTect results are frequently increased. This should be noted particularly in studies concerning women with a high prevalence of iron deficiency. This phenomenon may also account for the higher mean CDTect values in women, for the need of higher cutoff values, and for the apparent lack of diagnostic accuracy for alcohol-consuming women (21)(23)(24)(25)(26). In contrast to the CDTect method, high serum transferrin may sometimes be associated with false-negative results in the %CDT TIA, such as in patients with early phases of alcohol-induced liver disease, when transferrin synthesis is active (23). Interestingly, the group of nondrinking patients with high serum transferrin concentrations showed higher %CDT TIA values than those of healthy controls, suggesting differences in the sialylation process or glycoprotein uptake in these populations.
In contrast to the present data, a recent study by Stowell et al. (20), who used %CDT RIA for %CDT measurements, achieved (at the cutoff values given by the manufacturers) sensitivities of 78–94% for the %CDT RIA and 83–88% for CDTect, which were not different from each other. This may be because Stowell et al. (20) reported findings on alcoholics who had been actively drinking for 2 weeks before sampling in amounts ranging from 120 to 342 g of ethanol per day, whereas our patients represent problem drinkers with a mean continual alcohol consumption of 35–143 g per day and 0–7 day abstinence before sampling. However, when the patients with the most severe alcohol dependence were analyzed separately, the sensitivity of both assays in the present study material were also found to be equal and markedly higher (64–73%) than in the total population, indicating that the different CDT assays may be equally effective in detecting an advanced stage of alcoholism (data not shown). Apparently, the number of carbohydrate moieties attached to serum transferrin may also change as a function of the amount of alcohol consumed and/or as a function of severity of liver disease (1)(23). Thus, the assays may be more different from each other in the detection of early-phase drinking problems or binge-drinking than in detecting more severe alcoholism.
A summary of the practical characteristics of the %CDT TIA and CDTect
assays is given in Table 2
. The precision of both assays appears to suffer from the
multistep procedure. In fact, the performances of %CDT TIA and the
CDTect method were comparable despite the fact that the former
procedure is semiautomated. The instability of the eluted samples in
the %CDT TIA procedure may complicate routine work, whereas
corresponding samples in the CDTect procedure are stable for >1 week
after ion- exchange elution, when stored refrigerated. The %CDT and
the CDTect methods appear, however, to differ with respect to several
analytical characteristics and, therefore, can not readily replace each
other in routine laboratory work.
|
|
Acknowledgments |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-glutamyltransferase as markers of heavy alcohol consumption: gender differences. Alcohol Clin Exp Res 1994;18:747-754.
[Web of Science][Medline]
[Order article via Infotrieve]
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  J. B. Whitfield, V. Dy, P. A.F. Madden, A. C. Heath, N. G. Martin, and G. W. Montgomery Measuring Carbohydrate-Deficient Transferrin by Direct Immunoassay: Factors Affecting Diagnostic Sensitivity for Excessive Alcohol Intake Clin. Chem., July 1, 2008; 54(7): 1158 - 1165. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. HIETALA, H. KOIVISTO, P. ANTTILA, and O. NIEMELA COMPARISON OF THE COMBINED MARKER GGT-CDT AND THE CONVENTIONAL LABORATORY MARKERS OF ALCOHOL ABUSE IN HEAVY DRINKERS, MODERATE DRINKERS AND ABSTAINERS Alcohol Alcohol., September 1, 2006; 41(5): 528 - 533. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Anttila, K. Jarvi, J. Latvala, and O. Niemela METHOD-DEPENDENT CHARACTERISTICS OF CARBOHYDRATE-DEFICIENT TRANSFERRIN MEASUREMENTS IN THE FOLLOW-UP OF ALCOHOLICS Alcohol Alcohol., January 1, 2004; 39(1): 59 - 63. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Anttila, K. Jarvi, J. Latvala, J. E. Blake, and O. Niemela DIAGNOSTIC CHARACTERISTICS OF DIFFERENT CARBOHYDRATE-DEFICIENT TRANSFERRIN METHODS IN THE DETECTION OF PROBLEM DRINKING: EFFECTS OF LIVER DISEASE AND ALCOHOL CONSUMPTION Alcohol Alcohol., September 1, 2003; 38(5): 415 - 420. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. J. Legros, V. Nuyens, M. Baudoux, K. Zouaoui Boudjeltia, J.-L. Ruelle, J. Colicis, F. Cantraine, and J.-P. Henry Use of Capillary Zone Electrophoresis for Differentiating Excessive from Moderate Alcohol Consumption Clin. Chem., March 1, 2003; 49(3): 440 - 449. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. J. Legros, V. Nuyens, E. Minet, P. Emonts, K. Z. Boudjeltia, A. Courbe, J.-L. Ruelle, J. Colicis, F. de L'Escaille, and J.-P. Henry Carbohydrate-deficient Transferrin Isoforms Measured by Capillary Zone Electrophoresis for Detection of Alcohol Abuse Clin. Chem., December 1, 2002; 48(12): 2177 - 2186. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Tagliaro, F. Bortolotti, R. M. Dorizzi, M. Marigo, J. R. Delanghe, B. Wuyts, and M. L. De Buyzere Caveats in Carbohydrate-deficient Transferrin Determination Drs. Delanghe, Wuyts, and De Buyzere respond: Clin. Chem., January 1, 2002; 48(1): 208 - 209. [Full Text] [PDF]
|
|
|
|
|
|
R. F. Anton, C. Dominick, M. Bigelow, and C. Westby Comparison of Bio-Rad %CDT TIA and CDTect as Laboratory Markers of Heavy Alcohol Use and Their Relationships with {gamma}-Glutamyltransferase Clin. Chem., October 1, 2001; 47(10): 1769 - 1775. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Helander, M. Fors, and B. Zakrisson STUDY OF AXIS-SHIELD NEW %CDT IMMUNOASSAY FOR QUANTIFICATION OF CARBOHYDRATE-DEFICIENT TRANSFERRIN (CDT) IN SERUM Alcohol Alcohol., September 1, 2001; 36(5): 406 - 412. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Berkowicz, S. Wallerstedt, K. Wall, and H. Denison Carbohydrate-deficient transferrin in vitreous humour: a marker of possible withdrawal-related death in alcoholics Alcohol Alcohol., May 1, 2001; 36(3): 231 - 234. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Arndt Carbohydrate-deficient Transferrin as a Marker of Chronic Alcohol Abuse: A Critical Review of Preanalysis, Analysis, and Interpretation Clin. Chem., January 1, 2001; 47(1): 13 - 27. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Helander Absolute or Relative Measurement of Carbohydrate-deficient Transferrin in Serum? Experiences with Three Immunological Assays Clin. Chem., January 1, 1999; 45(1): 131 - 135. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
)
indicate the means of the groups. (A) upper reference limits
of the CDTect values for women (····) and for men (- -);
(B) upper reference limit of the %CDT TIA (- -).
) and men (
) and (B) in
alcohol abusers (
), hospitalized nondrinkers (
), and healthy
controls (
).
