PURPOSE OF THE STUDY: Core decompression of the femoral head is a conservative surgical treatment with controversial efficacy. We studied retrospectively a series of 32 cases of femoral head osteonecrosis treated by core decompression between 1988 and 2000 in 25 patients. We examined the epidemiological and clinical features as well as the laboratory findings, comparing cases requiring secondary hip replacement and those who had a favorable outcome. We search for prognostic factors. MATERIAL AND METHODS: The series included 32 hips, one case was lost to follow-up. Mean age at decompression was 41.3 years (22-55). In eight hips, osteonecrosis was favored by corticosteroid treatment, in three by chronic alcoholism, and in one by hypertriglyceridemia. No favoring factors were present for 20 hips. According to the ARCO classification there were 15 stage I hips, 13 stage II, 3 stage III, and one stage IV. Core decompression was centered in 24 hips and mean time to decompression was 6.4 months (14 days to 40 months). We reviewed hips without a total prosthesis using the Postel-Merle-d'Aubigne function score and for the radiological assessment the ARCO stage and the Koo index. RESULTS: Favorable outcome was noted in 12 hips. Total hip arthroplasty was required for 19, one hip was lost to follow-up. Mean follow-up in the success group was 82 months (26-176) and mean "time of participation" in the failure group was 11 months (1-38). Mean survival after core decompression was 14 months. Time between onset of symptoms and decompression did not influence outcome. Lesions which remained asymptomatic before decompression remained stable. The stage I hips did not have more favorable outcome than the stage II hips (p < 0.05). Stage III or IV hips had unfavorable outcome. Hips with a Koo index > 40 had a poor outcome (p < 0.05). DISCUSSION: Epidemiological factors which can worsen outcome after core decompression for osteonecrosis are controversial in the literature. Early stage disease (I or II) is considered as an ideal indication for decompression, but is insufficient alone to guarantee success. As other authors, we consider that ARCO stage III and IV and a Koo index > 40 are contraindications for decompression. Improved outcome after core decompression can only be achieved by limiting indications.