OBJECTIVE We report our experience with vascularized fibular transplant using the technique described by Urbaniak for the treatment of avascular necrosis of the femoral head. METHODS Sixteen patients, mean age 39 years 4 months at surgery, were evaluated at mean 3 years 6 months follow-up (minimum 15 months). Mean pain score (analog visual scale) was 6.25/10. Mean Postel-Merle-d'Aubigné (PMA) function score was 12. The Ficat classification was grade 2A (n=9) or grade 2B (n=7). RESULTS There were no cases of infection or migration of the fibula. The only complication at the donor site was one case of paresia of the long extensor of the hallux which resolved spontaneously. There were five failures which required a hip prosthesis (31%). Three failures occurred early during the first 18 months. Two were late. These patients had grade 2B (n=4) or grade 2A (n=1) necrosis. At last follow-up, eleven hips had not required prosthetic revision (69%). The pain score improved in 81% of patients: mean score=2.3. Nine patients were satisfied or very satisfied (81%). Among these eleven hips, function was good or excellent in eight (72%). Walking distance was improved in nine patients, unchanged in one and worse in one. Joint motion improved in nine patients, was unchanged in one and decreased in one. Nine patients resumed their full-time job. Mean sick leave was 13.3 months. Radiographically, only two of the initial Ficat grade 2A hips remained at this stage. Likewise only one of the three hips initially rated grade 2B remained at this stage. Overall radiographic results were: failure (n=5), worsening (n=9), stabilization and no improvement (n=2). Among the nine 2A hips, one was converted to a total hip prosthesis and five femoral heads remained spheric. Among the seven 2B hips, four were converted to a total hip prosthesis and two femoral heads collapsed. CONCLUSIONS At mean follow-up (3.5 years), 69% of the hips had not required total hip arthroplasty. The grade 2A hips did better but the small sample size did not allow statistical comparison. There was a clear discordance between the clinical presentation and the radiographic findings among the hips which were not treated with a prosthesis: good function score, maintained occupational and daily life activities contrasting with unfavorable radiographic evolution. Results could be improved by detailed quantification of the lesion using 3D imaging and by associating intraoperative navigation.