OBJECTIVE: To assess the diagnostic performance of the CAGE and AUDIT (Alcohol Use Disorder Identification Test) and its derivatives, and laboratory tests for screening alcohol abuse or dependence in a male medical hospital population. DESIGN: A diagnostic cross-sectional prevalence study. SETTING: Three general hospitals and one university hospital. PATIENTS: All male patients older than 18 years admitted to the hospitals, during a period of 6 weeks, were consecutively included in the study (N = 233). MEASUREMENTS: Calculation of diagnostic measurements with 95% CI and ROC curves for different scores of CAGE, AUDIT and derivatives, laboratory tests and % Carbohydrate Deficient Transferrin (CDT), using DSM-III-R as the reference standard, derived from the CIDI. RESULTS: A current diagnosis of alcohol abuse or dependence was found in 29 medical male inpatients, representing 12.4% (95% CI: 8.6-17.5). Ten of these (4.2%) fulfilled criteria of alcohol abuse and 19 (8.2%) the criteria of alcohol dependence. Laboratory tests are useless as screening tools with sensitivities between 10% (%CDT) and 52% (GammaGT). Only the Fiveshot questionnaire seems to yield reasonable diagnostic parameters at the recommended cutpoint of > or = 2.5 with a sensitivity of 79.3% and a specificity of 87.7%. CONCLUSIONS: With a prevalence of 12.4%, our results are similar with other published studies for alcohol abuse and dependence according DSM criteria. The AUDIT as well as the Fiveshot seems to have the best diagnostic properties in this male medical inpatient population, and regarding to brief interventions, these questionnaires can be used as screening instruments as well as opportunities to talk about alcohol problems with admitted patients.