Biological markers of alcohol consumption have been used in both clinical and research settings to aid in the identification of relapse drinking. Although carbohydrate-deficient transferrin (CDT) has been shown to be a sensitive and specific marker for the identification of heavy drinkers, little data are available as to its utility as a marker for relapse drinking during treatment, particularly in comparison with the more widely used serum gamma-glutamyltransferase (GGT). CDT and GGT were measured in 35 male alcoholics before entering, and every 4 weeks during, a 12-week outpatient treatment trial combining pharmacotherapy and cognitive behavioral therapy. CDT and GGT were again measured 14 weeks after completion of treatment. During the 12-week treatment period, CDT showed a significant difference in those individuals who abstained from drinking (30% decrease), compared with those who relapsed (10% increase). GGT decreased on average in all individuals, and the change from treatment entry did not differ significantly across the drinking outcome groups. The change in CDT, but not GGT, from study entry to termination, significantly correlated with total alcohol consumption during the trial. At the 14-week posttreatment, follow-up evaluation CDT showed about a 60% elevation and GGT showed a 30% elevation, on average, from study entry values in those individuals who had relapse drinking by self and/or collateral report. The change in both markers differed between those individuals who remained abstinent or relapsed during the poststudy period. In general, the change in CDT from pretreatment levels seemed more sensitive to drinking status during treatment and follow-up than GGT. This indicates that CDT may be more sensitive marker for evaluating drinking status during both clinical and research treatment trials.