Although mean corpuscular volume (MCV), aspartate aminotransferase (AST), alanine aminotransferases (ALT), and the AST/ALT ratio are sometimes used as markers of alcohol disorders, their utility has not been established in older persons. We determined the tests' performance characteristics for (1) at-risk drinking, (2) CAGE positivity, (3) at-risk drinking and CAGE positivity, and (4) a clinician-recorded diagnosis of alcohol abuse/dependence in a study of older male veterans receiving primary care. Participants (n = 587) included patients who had MCV, AST, and/or ALT data collected as part of routine care no more than 12 weeks before or after enrollment. MCV, AST, and ALT test results were obtained from the VA's database. At enrollment, the Timeline Followback and Alcohol Use Disorders Identification Test (AUDIT) were used to identify at-risk drinkers (> or = 15 drinks per week or AUDIT score > or = 8), and the CAGE questionnaire was administered to identify participants with a history abuse/dependent drinking (CAGE score > or = 2). Participants' medical records were reviewed to identify subjects with a clinician-recorded diagnosis of alcohol abuse/dependence. The prevalence of abnormal test results for MCV (threshold value = > 98), AST (> 41), ALT (> 41), and the AST/ALT ratio (> 2) was 11%, 4%, 4%, and 5%, respectively. The occurrence of at-risk drinking, CAGE positivity, at-risk drinking and CAGE positivity, and a clinician-recorded diagnosis of alcohol abuse/dependence was 11%, 25%, 5%, and 9%, respectively. Test sensitivity ranged from 3.9% to 25.4% and specificity from 88.5% to 97.1%, whereas positive likelihood ratios varied from 0.72 to 4.01 and negative likelihood ratios from 0.82 to 1.04. Areas under the receiver operating characteristic curve were similar (range = 0.50-0.58) across tests. In conclusion, MCV, AST, ALT, and the AST/ALT ratio are not useful markers of alcohol disorders in older male veterans.