Carbohydrate-deficient transferrin (CDT) has been widely investigated as a biological marker of heavy alcohol consumption. In general, it has been found to be at least as sensitive and probably more specific than gamma-glutamyltransferase (GGT). Because the two analytes are not highly correlated, their use in parallel enhances the sensitivity of detection of heavy alcohol consumption, especially in clinical populations. Women as a group produce more CDT under natural conditions and may produce less CDT in response to heavy drinking. Carbohydrate-deficient transferrin has found a place in the monitoring of alcoholics during treatment. Changes in CDT levels from individualized baseline values seem to be more sensitive to lower level relapse drinking than is the use of raw cut-off values. There are some conditions such as severe liver disease in which higher than normal levels of CDT are produced, thereby reducing the specificity of this marker for detecting heavy drinking under certain conditions. Future directions for the use of CDT include standardization and automation of measurement technology, evaluation of how to use it wisely in myriad medical and institutional settings, understanding more thoroughly the gender issues in its production, and greater evaluation of its performance as a monitoring tool during treatment and follow-up situations. How to combine CDT with both verbal tools of alcohol assessment and newer biological markers will also need more extensive evaluation.