The high prevalence of eating disorders in Arab countries indicates a need for an Arabic language screening tool. This study aimed to validate an Arabic version (A-SCOFF) of the British SCOFF questionnaire, a brief tool for the screening of eating disorders in primary health care. After translation and back-translation the A-SCOFF was given to 123 female patients [mean age 32 (SD 8.8) years] visiting primary health-care centres in Beirut. Each patient was evaluated by an eating disorders specialist blinded to A-SCOFF results. The validated Arabic version of the Mini International Neuropsychiatric Interview and the DSM-IV criteria for eating disorders were used as diagnostic references. The best diagnostic threshold for the A-SCOFF was found to be at 2 positive answers with a sensitivity of 80.0%, a specificity of 72.7% and an area under the curve of 80.0%. The A-SCOFF questionnaire is accurate and reliable for the early detection of eating disorders in this high-risk population.