The results for 103 consecutive hips (eighty-nine patients) that had been treated with free vascularized fibular grafting because of symptomatic osteonecrosis of the femoral head were reviewed in a prospective study. The disease was associated with consumption of alcohol in 30 percent of the hips, use of steroids in 17 percent, trauma in 13 percent, and Perthes disease in 3 percent; in the remaining 38 percent, the condition was idiopathic. All patients, except for one who died of unrelated causes 4.5 years after the operation, were followed for at least five years. By the time of the most recent follow-up evaluation, a total arthroplasty had been performed in thirty-one hips: two of the nineteen that were in stage II, according to the criteria of Marcus et al., at the time of the operation; five (23 percent) of the twenty-two that were in stage III; seventeen (43 percent) of the forty that were in stage IV; and seven (32 percent) of the twenty-two that were in stage V. Kaplan-Meier survivorship analyses demonstrated that the probability of conversion to a total hip arthroplasty within five years after free vascularized fibular grafting was 11 percent for the stage-II hips, 23 percent for the stage-III hips, 29 percent for the stage-IV hips, and 27 percent for the stage-V hips. There was a trend toward a lower rate of conversion to a total hip arthroplasty in patients who were less than thirty years old, but this difference did not reach significance (p = 0.06). No association was found between a causative factor and the probability of conversion to a total hip arthroplasty. The average Harris hip scores had improved at the latest follow-up evaluation, compared with the preoperative values (p < 0.001). For the stage-II hips, the average score improved from 56 to 80 points; for the stage-III hips, from 52 to 85 points; for the stage-IV hips, from 41 to 76 points; and for the stage-V hips, from 36 to 75 points. An outcome questionnaire, completed for 73 percent of the hips, revealed that 59 per cent of the hips that had not been subsequently treated with an arthroplasty did not limit or only slightly limited the patient's ability to carry out daily activities, and 62 percent did not limit or only slightly limited the patient's ability to work.(ABSTRACT TRUNCATED AT 400 WORDS)