CARBOHYDRATE-DEFICIENT TRANSFERRIN IS ELEVATED IN CATABOLIC FEMALE PATIENTS

Alcohol & Alcoholism Vol. 36, No. 6, pp. 603–607, 2001 CARBOHYDRATE-DEFICIENT TRANSFERRIN IS ELEVATED IN CATABOLIC FEMALE PATIENTS ANDREAS REIF*, HEIKE KELLER, MARC SCHNEIDER, STEPHAN KAMOLZ, ARMIN SCHMIDTKE and ANDREAS J. FALLGATTER Department of Psychiatry and Psychotherapy, Julius-Maximilians-University Würzburg, Füchsleinstr. 15, D-97080 Würzburg, Germany (Received 14 March 2001; in revised form 21 May 2001; accepted 15 June 2001) Abstract — Serum carbohydrate-deficient transferrin (CDT) is currently widely used as a biochemical marker of alcohol misuse. However, various recent studies have questioned the diagnostic value of this parameter and reported low levels of both specificity and sensitivity, especially in women. Thus, we sought to identify sub-groups of female individuals in which CDT is elevated independently of alcohol consumption. Significantly increased CDT levels were found in catabolic disease states due to psychiatric disorders distinct from alcoholism. None of those patients reported frequent alcohol consumption. CDT therefore appears also to be increased by metabolic processes distinct from alcohol degradation. Possible biochemical mechanisms of this phenomenon are discussed. As a consequence of these findings, the measurement of CDT alone is not suitable to screen for alcohol misuse in catabolic subjects. INTRODUCTION and in a recent study CDT levels of peri-menopausal women were reported to have a sensitivity of 30% (van Pelt et al., 2000). Although some other studies found better values, it was consistently noted that CDT alone is not suitable as a biomarker of alcohol intake in women (Nystrom et al., 1992; Gronbaek et al., 1995; Yeastedt et al., 1998; Allen et al., 2000; Brathen et al., 2000). However, the cause for this genderspecific effect is not yet known and further investigations are needed to clarify this issue. As the determination of CDT is more than twice as expensive as that of γ-GT or MCV, and, even more importantly, a false-positive test result might have important consequences for the patient, we sought to identify conditions in which CDT is elevated independently of alcohol intake. Most noteworthy, several variables distinct from gender have been shown to have an influence on the CDT level, e.g. age, smoking status, obesity (Sillanaukee et al., 1998; Whitfield et al., 1998), hypertension (Fagerberg et al., 1994a), serum iron (De Feo et al., 1999) and insulin levels (Fagerberg et al., 1994b). However, most of these do not increase the CDT concentration above the cut-off value in teetotallers (Whitfield et al., 1998). As we had the clinical impression that serum CDT concentration is increased in women reporting recent weight loss, we screened all female patients who were in a catabolic (negative metabolic) state when attending our department. Our hypothesis was that catabolism of various aetiologies might result in elevated CDT levels in female (psychiatric) patients. Serum carbohydrate-deficient transferrin (CDT) has over the past 20 years been described as a biochemical marker for alcohol consumption (van Eijk et al., 1983; Stibler et al., 1988; Stibler, 1991). Human transferrin comprises at least six different isoforms, with respect to the number of sialic acid side chains: penta-, tetra-, tri-, di-, mono- and asialotransferrin (Wong and Regoeczi, 1977). In subjects with high levels of alcohol consumption, isoforms with 0–3 sialo residues are increased, whereas the 4 and 5 sialo forms are decreased (Stibler et al., 1979). The former are designated as CDT. The exact mechanism by which alcohol intake elevates CDT is not yet exactly known and seems to be a multi-step process (Sillanaukee et al., 2001): enzymes which glycosylate transferrin are inhibited by ethanol metabolites (Xin et al., 1995), enhanced loss of sialic acid groups occurs (Ghosh et al., 1993), and receptor-mediated CDT uptake might be inhibited (Petren and Vesterberg, 1988). To date, several commercial tests are available for clinical use, and the measurement of CDT is commonly applied in clinical and forensic medicine (Wetterling and Kanitz, 1997). Initially, very high levels of both sensitivity and specificity have been reported, so that CDT was considered to be the best biomarker of alcoholism available (Stibler et al., 1986; Gjerde et al., 1988; Kapur et al., 1989; Kwoh-Gain et al., 1990; Lesch et al., 1996; Burke et al., 1998; Reynaud et al., 1998). However, follow-up studies failed to reproduce these promising results; a large number of studies demonstrated that CDT is not superior to gamma-glutamyltranspeptidase (γ-GT) and mean corpuscular volume (MCV), which are not correlated to CDT (Helander et al., 1996), in the identification of alcohol misuse (Nilssen et al., 1992; Gronbaek et al., 1995; Aithal et al., 1998; Schmitt et al., 1998; Sillanaukee et al., 1998; Limin et al., 1999). This has been confirmed by two recent literature reviews (Salaspuro, 1999; Scouller et al., 2000). It also appeared that CDT has a worse predictive power in women compared to men; Schmitt et al. (1998) reported a sensitivity of 0% (at a specificity level of 95%) for females, SUBJECTS AND METHODS Subjects Female in-patients presenting at the admission ward of our department have been screened for elevated CDT levels if they reported a recent history of weight loss or showed clear clinical signs thereof. All of them suffered from psychiatric disorders distinct from alcoholism; only patients reliably denying any alcohol intake above one drink per week were included in the screening programme. Higher alcohol consumption was an exclusion criterion, as was a score of ≥1 in the CAGE questionnaire (Mayfield et al., 1974; Ewing, 1984). *Author to whom correspondence should be addressed. 603 © 2001 Medical Council on Alcohol 604 A. REIF et al. Alcohol abstinence was confirmed by relatives or professional staff, when available; three patients did not have any opportunity to obtain alcoholic beverages, due to immobilization or placement in a closed ward. Serum CDT and other parameters were determined within the first 4 days after admission. Eleven patients matched all the inclusion criteria described above; they were between 20 and 86 years old with a mean age of 45 years. The recruitment interval was 5 months. The causes of weight loss were varied, as were the psychiatric diagnoses (cf. Table 2). Patient #1 could not care for adequate food intake due to long-lasting dementia, whereas patient #2 refused to cook and eat, because of a paranoid schizophrenia resulting in a delusion of impoverishment. Patient #3 suffered from chronic malnutrition and neglect due to paranoid schizophrenia, and #4 reported that she refused to eat because of erotomania. Patient #5 suffered from anorexia nervosa. Patient #6 suffered from a personality disorder exaggerated by an acute polymorph psychotic disorder; additionally, she had juvenile diabetes. Due to both psychiatric diseases, she had very low co (...truncated)