Sensitivity and specificity of carbohydrate-deficient transferrin as a marker of alcohol abuse are significantly influenced by alterations in serum transferrin: comparison of two methods.

Abstract

Despite a number of investigations suggesting the value of carbohydrate-deficient transferrin (CDT) as a marker of alcohol abuse, a variety of issues on the applicability of CDT measurements in clinical settings have remained unexplored. Earlier studies in this field have focused on the relationship of CDT and the amount of alcohol consumption or presence of liver disease, whereas the influence of alterations in serum transferrin concentrations on CDT has received less attention. In this study, we compared two different methods for measuring CDT (CDTect and %CDT) and total transferrin concentrations in a sample of 83 alcohol abusers (20 patients with alcoholic liver disease and 63 heavy drinkers who were devoid of liver disease, despite excessive alcohol consumption) and 89 controls, who were social drinkers or abstainers. The control population included 53 hospitalized patients with expected abnormalities in serum transferrin concentrations caused by conditions such as negative iron balance, pregnancy, or nonalcoholic liver disease. Both methods gave significantly higher values in alcohol abusers than in controls (p < 0.01), but the overall sensitivity for detecting alcohol abuse was clearly higher for CDTect (59%) than for %CDT (34%). The correlation between the results obtained by the two methods (r = 0.629) significantly improved, when the CDTect values were replaced by the ratio of CDTect/total transferrin (r = 0.770) (p < 0.05). There was a positive correlation between the CDTect and serum transferrin (r = 0.201, p < 0.01), which was significant both in the alcoholics (r = 0.240, p < 0.05), and especially in the controls (r = 0.727, p < 0.001). A significant inverse correlation emerged between %CDT and total transferrin (r = -0.302, p < 0.01). The sensitivities of CDTect and %CDT for correctly classifying alcohol abusers in the subgroup of alcoholic liver disease patients were 90% and 70% and in the subgroup of heavy drinkers without liver disease (49% and 22%), respectively. Specificities for CDTect and %CDT in this sample were 81% and 100%, respectively. However, in the subgroup of hospitalized control patients with abnormal serum transferrin, the specificity of CDTect was only 48%. According to present data, CDTect seems to be more sensitive than %CDT for detecting alcohol abuse. However, any alteration in serum total transferrin concentration markedly decreases the assay specificity. This should be considered when interpreting the assay results in patients with elevated serum transferrin, such as iron deficiency, pregnancy, or liver diseases.