In proximal humerus reconstructions, osteoarticular massive allograft (OA) allows a biologic reattachment of tendineous structures of the host providing stability and functional recovery. Allograft-prosthesis composites (APC) are more technically demanding, but functional advantages with new prosthetic design are expected with lower fracture and failure rate with time. We compared the outcome of patients treated with OA reconstruction (group A) and patients treated with an APC (group B) after humerus resection for bone tumours. Thirty six patients treated between 1996 and 2009 were included in the study (18 group A; 18 group B). The diagnosis was a malignant bone tumour in 28 cases, an aggressive benign bone tumour in 6 and a bone metastasis in 2. Three total humerus resections were performed (two in group A and one in group B). Three different type of prosthesis were used (conventional, reverse and resurfacing). In group A, at an average follow-up of 66 months (2-116), 16 patients were continuously disease free. Ten patients (56%) had no complication; in 6 cases (33%) we observed a fracture of the graft and in 3 cases (17%) a pseudoarthrosis. In group B, at an average follow up of 44 months (3-164), 16 patients were continuously disease free. Thirteen patients (72%) had no complication, in 2 cases (11%) stem loosening, in 2 cases (11%) dislocation, in 1 case pseudoarthrosis and in 1 case resorption and detachment of allograft's great tuberosity was observed. Functional outcome was evaluated using MSTS scoring system: a mean score of 24 (13-30) in group A and 23 (15-27) in group B was observed. In our experience, proximal humerus APC showed functional results comparable to OA allograft with lower complication rate. In conclusion, in biologic reconstruction of the proximal humerus, APC seems to provide a more reliable solution than OA allograft at long term