Clinical Characteristics of Carbohydrate-Deficient Transferrin (%Disialotransferrin) Measured by HPLC: Sensitivity, Specificity, Gender Effects, and Relationship with other Alcohol Biomarkers

Alcohol & Alcoholism Vol. 43, No. 4, pp. 436–441, 2008 Advance Access publication 14 April 2008 doi: 10.1093/alcalc/agn017 ASSESSMENT AND DETECTION Clinical Characteristics of Carbohydrate-Deficient Transferrin (%Disialotransferrin) Measured by HPLC: Sensitivity, Specificity, Gender Effects, and Relationship with other Alcohol Biomarkers Jonas P. Bergström and Anders Helander∗,1 Department of Clinical Neuroscience, Karolinska Institutet, Stockholm, Sweden ∗ Author to whom correspondence should be addressed: Anders Helander, Alcohol Laboratory, L7:03, Karolinska University Hospital Solna, SE-171 76 Stockholm, Sweden. Tel.: +46-8-51771531; Fax: +46-8-51771532; E-mail: 1 On Behalf of the WHO/ISBRA Study on State and Trait Markers of Alcohol use and Dependence Investigators. (Received 19 October 2007; first review notified 30 November 2007; in revised form 23 December 2007; accepted 7 February 2008; advance access publication 14 April 2008) Abstract — Aims: The sensitivity and specificity of the relative disialotransferrin amount (%DST), considered the primary single target for measurement of the alcohol biomarker carbohydrate-deficient transferrin (CDT), were compared with the absolute CDT amount determined by the CDTect assay and with GGT and AST. Methods: Serum samples (n = 1387) were collected within the WHO/ISBRA Study on State and Trait Markers of Alcohol Use and Dependence. The subjects had been classified as “nondrinkers” (26%), “light/moderate drinkers” (50%), or “heavy drinkers” (24%) by use of the WHO/ISBRA Interview Schedule. An HPLC candidate reference method for CDT was used to quantify individual transferrin glycoforms. Results: No gender difference in %DST was noted for nondrinkers, but light/moderate and heavy drinking males had significantly higher levels than females. Of the alcohol biomarkers examined, %DST showed the strongest correlation with self-reported alcohol intake, except for female heavy drinkers. The area under the %DST ROC curve for male (0.83) and female (0.82) heavy drinkers was significantly higher compared with CDT by CDTect (0.68) and GGT (0.69). At the 40, 60, or 80 g ethanol/day thresholds, %DST showed lower test sensitivity in women but there was no significant gender difference in overall accuracy according to ROC curve analysis. Conclusions: %DST measured by HPLC showed overall higher sensitivity for “heavy drinking” and better correlation with recent high alcohol intake, compared with the absolute CDT amount, and GGT and AST. The observation that several “light/moderate drinkers” had elevated %DST levels and some also a measurable asialotransferrin indicated misclassification with the WHO/ISBRA Interview Schedule and emphasize the limitations of self-reports of drinking. INTRODUCTION The iron-transportglycoprotein transferrin exists in different genetic variants owing to variations in the amino acid sequence (Kamboh and Ferrell, 1987; de Jong et al., 1990). Several glycoforms differing in the number of and/or structure of the maximum two N-linked oligosaccharide chains (N-glycans) are always present (Arndt et al., 2001; Helander et al., 2001). The major glycoform under normal conditions is tetrasialotransferrin that contains two disialylated biantennary glycans (i.e., four terminal sialic acids). Tetrasialotransferrin typically makes up ∼75–80% of serum transferrin (Helander et al., 2003). Other less abundant glycoforms normally found in blood are disialo-, trisialo-, pentasialo-, and hexasialotransferrin. Abnormal transferrin profiles are seen and used for preliminary diagnosis of rare congenital disorders of glycosylation (CDG) (Helander et al., 2004; Jaeken and Matthijs, 2007). A more common cause of an abnormal serum transferrin profile is sustained heavy alcohol consumption. Excessive drinking causes a transient change in the glycoform pattern by increasing the relative amounts of disialo- and asialotransferrin at the expense of tetrasialotransferrin (Bergström and Helander, 2008). This change is named carbohydrate-deficient transferrin (CDT; originally defined as the sum of asialo-, monosialo-, and disialotransferrin) and is used as an alcohol biomarker for heavy drinking (Stibler, 1991). When drinking is discontinued, the transferrin glycoform pattern normalizes with a half-life of ∼1.5 weeks (Jeppsson et al., 1993), and reaching the baseline level usually takes 3–5 weeks (Helander and Carlsson, 1996). The alcohol-induced change in transferrin glycosylation was identified more than 30 years ago (Stibler, 1991). Today CDT is routinely applied in many countries mainly for detection and follow-up of alcohol-related problems in health care and forensic (traffic medicine) settings (Helander, 2003). Over the years, several analytical techniques have been used for the measurement of CDT (Arndt, 2001; Bortolotti et al., 2006). In the large international WHO/ISBRA Study on State and Trait Markers of Alcohol Use and Dependence (Tabakoff et al., 2001; Conigrave et al., 2002), CDT quantification was achieved using the CDTectTM immunoassay (Pharmacia, Sweden), which was the standard routine method at that time but is no longer in use. A number of limitations of CDTect have hampered the interpretation and generalization of the CDT data arising from the WHO/ISBRA Study. With this assay, a CDT fraction consisting of asialo-, monosialo-, a minor part of disialo-, and traces of trisialotransferrin (Arndt et al., 1998) was separated from the non-CDT glycoforms using small ion-exchange columns followed by a transferrin immunoassay (Stibler et al., 1991), but the individual glycoforms were not recognized. Furthermore, the CDT content was reported in an absolute amount (units/L), and not normalized to the total transferrin concentration (%CDT) which is the standard way today (Jeppsson et al., 2007). Because of this, gender-specific reference intervals had to be applied and the CDTect result was influenced by many causes of an elevated or a reduced total serum transferrin concentration (e.g., anemia, estrogens, pregnancy, and inflammation). Lastly, genetic transferrin variants were not detected but represented a cause of incorrect CDT values with the CDTect assay (Helander et al., 2001). These method-dependent shortcomings are likely to have underrated the diagnostic accuracy  C The Author 2008. Published by Oxford University Press on behalf of the Medical Council on Alcohol. All rights reserved Sensitivity and Specificity of %DST Measured by HPLC and clinical value of the CDT biomarker (Salaspuro, 1999; Scouller et al., 2000; Koch et al., 2004; Fleming et al., 2004). The present study used a high-performance liquid chromatography (HPLC) method, recently proposed as a candidate CDT reference method (Jeppsson et al., 2007), to identify and quantify individual transferrin glycoforms in serum samples collected within the WHO/ISBRA Study. The sensitivity and specificity of the relative disialotransferrin amount to total transferrin, (...truncated)