Carbohydrate deficient transferrin in detecting relapse in alcohol dependence.

Patients reports together with findings at clinical examination and information from an informant such as a relative were used to categorise patients as relapsed or not relapsed during a 6 month period of out-patient treatment at an alcohol problems clinic. At each fortnightly visit, blood was taken for measurement of serum gamma-glutamyl transferase and carbohydrate deficient transferrin (Pharmacia method). A total of 53 patients attended for at least one follow-up visit. Mean CDT differentiated relapsers from non-relapsers at seven of the 11 visits (P < 0.05), but at no visit did mean GGT differentiate. CDT tended to become elevated after a relapse more quickly than GGT. However, whether using upper limit of normal (ULN), or defining a 'positive test' as > last test and either > 20% above lowest previous test or > ULN, specificity (averaged over the 11 visits) was greater for GGT than CDT. Some of the false positive results for CDT were in patients who, shortly after having a positive test, relapsed, suggesting that a rising CDT can herald a relapse admitted by the patient. This could not be shown for false positive GGT results. Inspection of individual trajectories of alcohol consumption and blood test results shows that for some patients GGT is the more effective marker of relapse, whilst for others CDT operates better.