Aim: Treatment of unrecons tructible comminuted fractures of the radial head remains a therapeutic challenge. There is limited information on the outcome of management of these injuries with metal radial head implant. Before choosing one of them, two groups of patients were rewieved in a retrospective study comparing resection and Swanson implant.

Material and methods: 39 patients sustained unreconstructible radial head fracture between 1969 and 1992. Two groups of 24 patients were reviewed clinically and radiologically by a surgeon not involved in treatment. Functionnal outcomes of the elbow (morrey scoring – SOO scoring system, instability, cubitus valgus) and involvement of the wrist (pain, grasp, RUD instability) were evaluated with a mean follow up of 15 years (6–27,7). Elbow or wrist arthritis, ulnar variance, and evolution of Swanson implant’s were evaluated on standard Xray. 1/2 patient was a man and 1/3 cases consisted in a professional injury. Resec tion group: 16 patients with a mean age of 39 (19–65), treated with a mean preoperative delay of 29,6 days (0–150) were rewieved with a mean follow up of 18,9 years (6–27,7). There were 75% Mason type-III injuries associated with dislocation of the elbow in 18% cases. Swanson group: 8 patients with a mean age of 36 (21–57), treated with a mean preoperative delay of 4,1 days (0–15) were rewieved with a mean follow up of 12 years (8,1–20). There were 88% Mason type-III injuries with no dislocation in that group.

Results: The following criteria did not show any significant differences between both groups: Morrey scoring 77/100, SOO scoring 7,4/11, mean flexion was 130°, mean deficient extension was 18°, mean pronation 60°, mean supination 67°, grasp reached 90% of the contralateral side. Arthritis was pointed in same frequency at the elbow level (87%), and wrist level (66%) in each group. 94% of pa tients in resection group and 89% in Swanson group were satisfied. Excellent and good results are reported most frequently in Swanson group (37% resection group, 51% Swanson group). In resection group following complications were significantly more frequent: ulnar nerve irritation (2x), ulnar head dislocation (2x), ulnar head instability (3x), paraarticular ossification (5x), ulnar variance positive in all cases (mean value 3,20 mm). In the Swanson group only one implant was destroyed at the follow up.