Background: We attempted to determine whether including trisialo-Fe2-transferrin in carbohydrate-deficient transferrin (CDT) affects the diagnostic accuracy of CDT as a marker of chronic excessive alcohol intake.
Material/methods: The criterion standard tests for the diagnosis of alcoholism and alcohol intake were the Composite International Diagnostic Interview (CIDI) and the Timeline-Followback (TLFB). The study groups (alcohol intake in each of the last 4 weeks before blood sampling) were comprised of 56 controls (< or = 280 g/week, no alcoholism), 54 hazardous drinkers (>280 g/week, no alcoholism), 63 alcoholics (>280 g/week, alcoholism diagnosis). CDT analysis was performed with %CDTri-TIA, which includes about 50% of trisialo-Fe2-transferrin in CDT, and ChronAlcoI.D, which excludes this transferrin isoform from CDT.
Results: Depending on the cut-offs for the CDT/transferrin ratio (upper or lower limit of the test-specific borderlines) and on the patient group, the diagnostic sensitivity was 28.1%-72.3% for %CDTri-TIA, as opposed to 50.0%-82.5% for ChronAlcoI.D. The diagnostic accuracy was 62.8%-78.5% for %CDTri-TIA and 71.8%-86.6% for ChronAlcoI.D. The latter test consistently showed higher diagnostic sensitivity and accuracy than %CDTri-TIA. The diagnostic specificity was 85.7%-98.2% for %CDTri-TIA and 91.1%-92.2% for ChronAlcoI.D. The areas under the ROC curve were 0.810%-0.885 for %CDTri-TIA and 0.867%-0.896 for ChronAlcoI.D.
Conclusions: The present study and data from the literature indicate that including parts of trisialo-Fe2-transferrin by the %CDTri-TIA test significantly reduces the diagnostic sensitivity and thus accuracy of CDT as a marker of chronic excessive alcohol use.