Diagnostic usefulness of carbohydrate-deficient transferrin for detecting alcohol-related problems in hospitalized patients

A. Gómez*,, A. Conde1,, J. A. Aguiar, J. M. Santana, A. Jorrín and P. Betancor1,

Family Practice, San Gregorio Health Care Centre and
1 Internal Medicine, Doctor Negrín General Hospital, Las Palmas, Spain

Received 24 March 2000; in revised form 23 August 2000; accepted 29 December 2000


    ABSTRACT
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 PATIENTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
We evaluated the diagnostic usefulness of carbohydrate-deficient transferrin (CDT) for detecting alcohol-related problems (ARP) in hospitalized patients, assessed potential differences according to gender and age, and compared this value to the other screening measures conventionally used, namely the CAGE questionnaire and standard biological markers MCV (mean corpuscular volume) and GGT ({gamma}-glutamyltransferase). CDT sensitivity for detecting ARP was 70%, specificity 75% and area under the receiver operating characteristic curve 0.76%. Its sensitivity was shown to be lower both in the female group (65 vs 71%) as well as in the younger-age group whereas its specificity was lower in the liver-damaged group (55%). CDT diagnostic value did not exceed that of CAGE (sensitivity 78% and specificity 99%), MCV (sensitivity and specificity 74%) or GGT (sensitivity 82% and specificity 53%) but was better than those of other biochemical measures. We conclude that CDT is an effective biological marker for the detection of ARP among hospitalized patients, but that its diagnostic usefulness does not exceed that of some other alternative measures commonly used.


    INTRODUCTION
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 PATIENTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
Alcohol-related problems (ARP) represent one of the greatest influences affecting public health in many countries. It is estimated that 15–20% of primary care visits and 20–30% of hospitalizations are related either directly or indirectly to excessive alcohol consumption (Bruguera et al., 1994Go). To improve the global treatment of this important problem, it is essential to have access to accurate diagnostic tools. Accurate testing will aid both in avoiding further complex, expensive and at times invasive, diagnostic tests as well as establishing an earlier treatment.

Blood markers, in contrast to other diagnostic tests, could provide objective criteria for alcohol misuse because they can be applied in patients who may be unwittingly underestimating their own alcohol intake or in patients unable to respond, in some cases secondarily to injury.

An ideal marker must possess certain characteristics. It should possess a high sensitivity and specificity and not be modified by certain traits of individual patients. Its detection should not be dependent on temporal sampling and it should correlate well with the amount of alcohol consumed. The ideal marker should be easily measured with a rather simple and low-cost technique (Rosman and Lieber, 1992Go). To date, no marker is known to fulfil all these requirements, although the recently studied marker carbohydrate-deficient transferrin (CDT) has raised expectations.

The aim of this project was to assess the diagnostic usefulness of CDT for detecting ARP in the hospital setting, and also to evaluate the influence that gender and age may have on it and to compare this value to that of other screening measures, such as other biological markers and the CAGE questionnaire.


    PATIENTS AND METHODS
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 PATIENTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
This is a descriptive cross-sectional study conducted to evaluate diagnostic tests carried out between February 1997 and August 1998 at a university hospital (El Sabinal Hospital, Las Palmas, Spain).

For an estimated prevalence of 20% (Bruguera et al., 1994Go), the sample size was calculated to achieve a sensitivity of 85%, at a precision level of 7% and 95% confidence interval. Thus 99 eligible patients were sampled consecutively. All these patients fulfilled ICD-10 criteria for harmful alcohol consumption or alcoholic dependence (World Health Organization, 1992Go) and would have been consuming alcohol during the month prior to the admission so as to be considered hazardous drinkers by The Royal College of Physicians of London (1987) (i.e. 168 g per week of alcohol for males and 120 for females). A total of 80 control subjects were randomly selected and matched to the study group for age and gender. They met none of the above criteria. Ex-drinkers were not excluded from the control group. Exclusion criteria were patients under 15 years of age and those suffering from any mental disorder that precluded the completion of our questionnaire.

Data reporting weekly alcohol intake, as that referred to the 2 weeks prior to hospital admission, were collected from the history using the ‘Standard Drink Unit’ (SDU) system for quantification purposes by applying the equivalence proposed by The Royal College of Physicians of London (1981) (i.e. 1 SDU = 8 g of alcohol). Afterwards CAGE was administered, since it is the most widely used questionnaire (Ewing, 1984Go), marking the cut-off value at two positive answers. Also the eventual presence of hepatic damage was considered, being that defined by clinical findings suggesting portal hypertension, biochemistry (i.e. liver function tests and serology) or ultrasonography data. Finally, a blood sample was obtained within the 48 h following hospital admission to determine CDT, {gamma}-glutamyltransferase (GGT), mean corpuscular volume (MCV), aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase, platelets, triglycerides and uric acid levels. CDT was determined with a Cobas-MiraPlus instrument based on column chromatography through an immunological technique using Axis biological reactants from Norway. The test accorded with internal as well as external consistency. Positive values were considered beyond 5%.

Diagnostic usefulness for detecting ARP was determined for CDT, conventional biological markers and CAGE, as well as the influence that gender and age could have on it. As for the latter, three age categories were established (<=40, 41–59 and >=60 years old). We also evaluated the relationship between biochemical markers and CAGE with the amount of alcohol consumed weekly, as well as that of classical biological measures with CDT. Statistical analysis was performed using ‘Statistical Package for Social Sciences’ (SPSS) 8.0 version for Windows and ‘Tabulated Data Epidemiological Analysis’ (EPIDAT) 2.1 version for Windows. Sensitivity, specificity, overall value, positive and negative predictive values and likelihood ratios and the area under the receiver operating characteristic (ROC) curve were calculated to assess the diagnostic usefulness. {chi}2-Test (for quantitative variables), Student's t-test and Mann–Whitney test (for qualitative and quantitative variables of normal and non-normal distribution respectively), Pearson and Spearman rank correlation coefficients (quantitative variables of normal and non-normal distribution respectively) were used for univariant analysis. Level of statistical significance was set at 5%.


    RESULTS
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 PATIENTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
Descriptive study
The sample comprised a population of 99 in-patients considered to have ARP: 82 (83%) were men and 17 (17%) women. The control group consisted of 80 patients, 64 (80%) were males and 16 (20%) females. The mean age for both groups was 49 years (± SD of 12 and 15 respectively).

Among those patients with alcohol disorders, 66 (67%) had concurrent liver disease (94% alcoholic type and 6% viral type), whereas in the control group only 20 (25%) exhibited some form of liver disease (30% viral type, 25% alcoholic type, 10% fatty liver and 35% other types).

No patient refused to collaborate in data collection. Average weekly SDU consumption by patients with ARP was 102.6 ± 107.3 whereas for the control group it was 4.3 ± 6.5. Measures of central tendency and dispersion for gender and age are as shown in Table 1Go. A positive result for CAGE was found in 77 (78%) patients with ARP (95% CI: 68.1–85.3) and in one control patient (1%).


View this table:
[in this window]
[in a new window]
 
Table 1. Measures of central tendency and dispersion for average weekly standard drink units intake
 
Diagnostic utility
The sensitivity shown by CDT for the detection of ARP was 70%, specificity 75%, overall value 72% and area under ROC curve 0.76 (95% CI: 0.68–0.83) (Table 2Go). In the male group, the sensitivity was 71% and specifity 83%, whereas in the female group these were 65% and 44% respectively (Table 3Go). According to age, in the younger group the sensitivity was 62% and specificity 67%, in the mid-age group it was 72% and 77% respectively and for the older group 73 and 80% respectively (Table 3Go). Sensitivity was similar for the two groups based on the liver disease status, whereas specificity showed a decrease from 83% to 55% for the hepatic dysfunction group. Half of false positive cases detected were carriers of a non-alcoholic liver condition (data not shown).


View this table:
[in this window]
[in a new window]
 
Table 2. Diagnostic usefulness of CAGE and biological markers for detecting alcohol-related problems
 

View this table:
[in this window]
[in a new window]
 
Table 3. Diagnostic usefulness of carbohydrate-deficient transferrin for detecting alcohol-related problems according to gender and age
 
Among biological markers, GGT and MCV yielded sensitivities of 83% and 74% and specificities of 53% and 74% respectively (Table 2Go). CAGE used for the same diagnostic purposes had a sensitivity of 70% and a specificity of 75% (Table 2Go).

Univariant analysis
Analysis of the relationship between alcohol intake, biological markers and CAGE revealed a statistically significant positive correlation between average SDU weekly intake and CDT values (P < 0.001), GGT (P < 0.001), MCV (P < 0.001), AST (P < 0.001), ALT (P < 0.017), CAGE (P < 0.001) and a negative correlation with platelet count (P < 0.015) (Table 4Go). Analysis of this relationship in the group of patients with alcohol disorders only showed positive and significant correlations between SDU weekly intake and CDT levels (P = 0.044), AST (P = 0.048) and CAGE (P < 0.001).


View this table:
[in this window]
[in a new window]
 
Table 4. Relationship between average weekly standard drink units intake, levels of biological markers and CAGE
 
Analysis of the relationship between CDT, conventional biological markers and CAGE revealed a statistically significant positive correlation between CDT levels and those of GGT (P < 0.001), MCV (P < 0.001), AST (P < 0.004), ALT (P < 0.019) and triglycerides (P < 0.001) (CAGE, P < 0.001) and a negative correlation with that of platelet count (P < 0.001) (Table 5Go). Analysis of this relationship in the group of the patients with alcohol disorders only showed positive and significant correlations between CDT levels and MCV (P = 0.033), AST (P < 0.001), ALT (P = 0.018) and triglycerides (P < 0.001) and a negative correlation with that of platelet count (P = 0.07).


View this table:
[in this window]
[in a new window]
 
Table 5. Relationship between levels of carbohydrate-deficient transferrin (CDT), biological markers and CAGE
 

    DISCUSSION
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 PATIENTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
The interest that CDT as a biological marker has fostered recently is especially due to its high specificity as compared with some of the other commonly used biochemical measures. Stibler (1991), in a review of all the works conducted in the 1980s, found an overall sensitivity and specificity for CDT of 82% and 97% respectively. Since then, the majority of published studies, including our own, have shown that CDT sensitivity, as well as specificity, are somewhat lower than formerly described. While comparing studies, sample characteristics should be taken into consideration (specifically gender, age, average alcohol intake and prevalence of liver disease), as well as the specific aim and context with which the study has been conducted. Therefore, the aim of some studies, such as this one, is to detect ARP. Huseby et al. (1997) demonstrated that variation in CDT diagnostic value depends on the setting. Thus, a higher sensitivity and lower specificity can be obtained among detoxification unit patients as compared with those of surgery wards (sensitivities of 77% vs 51% and specificities of 81% vs 86% respectively), showing that the higher the average alcohol intake the higher the sensitivity, and the higher the potential false positive cases the lower the specificity. The results in the present study were intermediate, probably due to it being conducted in a Medicine ward with moderate levels of alcohol consumption, where the overall prevalence of liver disease was lower than it would have been in a detoxification unit. Perhaps this could be the reason why the studies conducted to date in primary care settings provide worse results (Nilsson et al., 1992; Yersin et al., 1995Go; Aithal et al., 1998Go). Other studies tried to detect consumers after a given amount of alcohol intake, e.g. attempts to detect an average of 50 or 60 g daily intake among in-patients gave sensitivities of 69% and 70% (Bell et al., 1994Go; Stauer et al., 1995Go). These figures, at least theoretically, should be lower than ours, as they relate to lower alcohol intake.

Our results agree with most published studies indicating that the sensitivity of CDT determination decreases in women (Anton an Moak, 1994; Gronbaek et al., 1995Go; Yersin et al., 1995Go; Huseby et al., 1997Go). This difference has been attributed to gender-specific consumption patterns, as well as lower physiological testosterone level and/or the relative increase in oestrogens in this group (Yersin et al., 1995Go). Studies evaluating the age effect showed no coherent results (Yersin et al., 1995Go; Huseby et al., 1997Go).

Nor does unanimity exist when assessing the diagnostic usefulness of CDT in the presence of liver disease (Ouyahya et al., 1995Go; Radosavljevic et al., 1995Go; Nalpas et al., 1997Go) or, particularly, the alcoholic type (Meregalli et al., 1995Go; Rublo et al., 1997Go; Henriksen et al., 1997Go). Niemela et al. (1995) showed that CDT increases during the first stages of hepatic damage, and a negative correlation was seen between CDT values and the morphologic severity index. In our present study, we observed significantly higher levels in the liver dysfunction group of patients, perhaps due to the fact that most of them were carriers of a mild to moderate degree of hepatic disease. Given that ex-drinkers with current non-hazardous consumption were not excluded from the control group, among this latter group, some cases of alcoholic liver disease could eventually be found. Concerning these and steatosis cases, we cannot have absolute certainty about the absence of current harmful consumption. Although it might have been interesting, we did not assess the differences among each of the liver disease groups, in view of the small numbers of subjects. As expected, we found a positive correlation between CDT values and the amount of alcohol intake, as much in the whole group of patients as in the group of patients with alcohol disorders solely, showing a coefficient similar to those reported by Niemela et al. (1995).

In the present study, the diagnostic usefulness of CDT did not exceed that of some classical markers such as GGT or MCV. The results reported in the literature are still controversial (Bell et al., 1994Go; Stauer et al., 1995Go; Yersin et al., 1995Go; Huseby et al., 1997Go). While investigating CDT in combination with GGT we did not find, contrary to some authors (Randell et al., 1998Go), any greater effectiveness, while sensitivity increased up to 93% and specificity decreased to 39%. Therefore biological markers would constitute a useful diagnostic tool for patients denying or underestimating their alcohol consumption and for those not available for interview. Thus, we can conclude that CDT is an effective diagnostic measure to detect ARP, its effectiveness not exceeding that of conventional biological markers.


    ACKNOWLEDGEMENTS
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 PATIENTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
We would like to thank The Higher College of Physicians of Las Palmas for the trust given to this project, expressed in one of the annual grants offered for its completion. The authors also wish to express their grateful appreciation to Dr Vicente H. Gracias, Assistant Professor from the University of Pennsylvania, for his generous contribution in the revision of the typescript.


    FOOTNOTES
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 PATIENTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
* Author to whom correspondence should be addressed at: C/Mary Sánchez n° 21, apartamento 410, 35009 Las Palmas De Gran Canaria, Spain. Back


    REFERENCES
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 PATIENTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
Aithal, G. P., Thornes, H., Dwarakanath, A. D. and Tanner, A. R. (1998) Measurement of carbohydrate-deficient transferrin (CDT) in a general medical clinic: is this test useful in assessing alcohol consumption? Alcohol and Alcoholism 33, 304–309.[Abstract]

Anton, R. F. and Moak, D. H. (1994) Carbohydrate-deficient transferrin and gamma-glutamyltransferase as markers of heavy alcohol consumption: gender differences. Alcoholism: Clinical Experimental Research 18, 747–754.[ISI][Medline]

Bell, H., Tallaksen, C. M., Try, K. and Haug, E. (1994) Carbohydrate-deficient transferrin and other markers of high alcohol consumption: a study of 502 patients admitted consecutively to a medical department. Alcoholism: Clinical Experimental Research 18, 1103–1108.[ISI][Medline]

Bruguera, M., Gual, A., Salleras, L. and Rodés, J. (1994) Screening for excessive alcohol consumption. Medicina Clínica (Barcelona) 102 (Suppl. 1), 85–92.

Ewing, J. A. (1984) Detecting alcoholism: the CAGE questionnaire. Journal of the American Medical Association 252, 1905–1907.[Abstract]

Gronbaek, M., Henriksen, J. H. and Becker, U. (1995) Carbohydrate-deficient transferrin: a valid marker of alcoholism in population studies? Results from the Copenhagen city heart study. Alcoholism: Clinical Experimental Research 19, 457–461.[ISI][Medline]

Henriksen, J. H., Gronbaek, M., Moller, S., Bendsten, F. and Becker, U. (1997) Carbohydrate-deficient transferrin (CDT) in alcoholic cirrhosis: a kinetic study. Journal of Hepatology 26, 287–292.[ISI][Medline]

Huseby, N. E., Nilssen, O., Erfurth, A., Wetterling, T. and Kanitz, R. D. (1997) Carbohydrate-deficient transferrin and alcohol dependency: variation in response to alcohol intake among different groups of patients. Alcoholism: Clinical Experimental Research 21, 201–205.[ISI][Medline]

Meregalli, M., Giacomini, V., Lino, S., Marchetti, L., DeFeo, T., Cappellini, M. D. et al. (1995) Carbohydrate-deficient transferrin in alcohol and nonalcohol abusers with liver disease. Alcoholism: Clinical Experimental Research 19, 1525–1527.[ISI][Medline]

Nalpas, B., Hispard, E., Thepot, V., Pot, S., Dally, S. and Berthelot, P. (1997) A comparative study between carbohydrate-deficient transferrin and gamma-glutamyltransferase for the diagnosis of excessive drinking in a liver unit. Journal of Hepatology 27, 1003–1008.[ISI][Medline]

Niemela, O., Sorvajarvi, K., Blake, J. E. and Israel, Y. (1995) Carbohydrate-deficient transferrin as a marker of alcohol abuse: relationship to alcohol consumption, severity of liver disease and fibrogenesis. Alcoholism: Clinical and Experimental Research 19, 1203–1208.[ISI][Medline]

Nilssen, O., Huseby, N. E., Hoyer, G., Brenn, T., Schirmer, H. and Forde, O. H. (1992) New alcohol markers. How useful are they in population studies: The Svalbard study 1988–1989. Alcoholism: Clinical Experimental Research 16, 83–87.

Ouyahya, F., Bacq, Y., Schellenberg, F., Metman, E. H. and Weill, J. (1995) Carbohydrate-deficient transferrin and liver diseases. Study of 94 patients. Gastroenterologie Clinique Biologique 19, 698–702.

Radosavljevic, M., Temsch, E., Hammer, J., Pfeffel, F., Mayer, J., Renner, F. et al. (1995) Elevated levels of serum carbohydrate-deficient transferrin are not specific for alcohol abuse in patients with liver disease. Journal of Hepatology 23, 706–711.[ISI][Medline]

Randell, E., Diamandis, E. P. and Goldberg, D. M. (1998) Changes in serum carbohydrate-deficient transferrin and gamma-glutamyl transferase after moderate wine consumption in healthy males. Journal of Clinical Laboratory Analysis 12, 92–97.[ISI][Medline]

Rosman, A. and Lieber, C. S. (1992) Biological markers of alcoholism. In Medical and Nutritional Complications of Alcoholism: Mechanisms and Management, pp. 531–563. Plenum Publishing Corporation, New York.

Royal College of Physicians of London (1987) A Great and Growing Evil? The Medical Consequences of Alcohol Abuse. Tavistock, London.

Rublo, M., Caballería, J., Deulofeu, R., Caballería, L., Gasso, M., Pares, A. et al. (1997) Carbohydrate-deficient transferrin as a marker of alcohol consumption in male patients with liver disease. Alcoholism: Clinical and Experimental Research 21, 923–927.[ISI][Medline]

Stauer, R. E., Stepan, V., Trauner, M., Wilders-Truschnig, M., Leb, G. and Krejs, G. J. (1995) Evaluation of carbohydrate-deficient transferrin for detection of alcohol abuse in patients with liver dysfunction. Alcohol and Alcoholism 30, 171–176.[Abstract]

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]

World Health Organization (1992) Mental and Behavioral Disorders: Diagnostic Criteria for Research. International Statistical Classifications of Diseases, 10th revision (ICD-10). World Health Organization, Geneva.

Yersin, B., Nicolet, J. F., Decrecy, H., Burnier, M., Van Melle, G. and Pécoud, A. (1995) Screening for excessive alcohol drinking: comparative value of carbohydrate-deficient transferrin, gamma-glutamyltranspeptidase and mean corpuscular volume. Archives of Internal Medicine 155, 1907–1911.[Abstract]