The Lequesne algofunctional indices for the knee and the hip require a physician interview. We adapted the Lequesne indices for the German language and tested the metric properties, test-retest reliability and validity of a self-report questionnaire-format in 51 patients with hip or knee OA. The symptom and function components as well as the global indices had a high test-retest reliability (intraclass correlation coefficient knee: 0.87/0.92 and 0.86, hip: 0.96/0.85 and 0.94). The symptom component showed an insufficient internal consistency (Cronbach's coefficient alpha knee: 0.55; hip: 0.63). Also, the symptom component was not or only weakly associated with radiological OA-severity and the limitations of range-of-motion. The most likely explanation is the grading of the Lequesne symptom questions which addresses such different concepts as pain presence, pain on movement, duration of pain and appearance of pain after a certain period of time. Because of the insufficient internal consistency and questionable validity the tested self-report German-version of the Lequesne algofunctional indices may not be recommended for use in clinical research and practice.