Outcome after reconstruction of the proximal humerus for tumor resection: a systematic review

Abstract

Background: Tumors of the appendicular skeleton commonly affect the proximal humerus, but there is no consensus regarding the best reconstructive technique after proximal humerus resection for tumors of the shoulder.

Questions/purposes: We wished to perform a systematic review to determine which surgical reconstruction offers the (1) best functional outcome as measured by the Musculoskeletal Tumor Society (MSTS) score, (2) longest construct survival, and (3) lowest complication rate after proximal humerus resection for malignant or aggressive benign tumors of the shoulder.

Methods: We searched the literature up to June 1, 2013, from MEDLINE, EMBASE, and the Cochrane Library. Only studies reporting results in English, Dutch, or German and with followups of 80% or more of the patients at a minimum of 2 years were included. Twenty-nine studies with 693 patients met our criteria, seven studies (24%) were level of evidence III and the remainder were level IV. Studies reported on reconstruction with prostheses (n = 17), osteoarticular allografts (n = 10), and allograft-prosthesis composites (n = 11). Owing to substantial heterogeneity and bias, we narratively report our results.

Results: Functional scores in prosthesis studies ranged from 61% to 77% (10 studies, 141 patients), from 50% to 78% (eight studies, 84 patients) in osteoarticular graft studies, and from 57% to 91% (10 studies, 141 patients) in allograft-prosthesis composite studies. Implant survival ranged from 0.38 to 1.0 in the prosthesis group (341 patients), 0.33 to 1.0 in the osteoarticular allograft group (143 patients), and 0.33 to 1.0 in allograft-prosthesis group (132 patients). Overall complications per patient varied between 0.045 and 0.85 in the prosthesis group, 0 and 1.5 in the osteoarticular graft group, and 0.19 and 0.79 in the prosthesis-composite graft group. We observed a higher fracture rate for osteoarticular allografts, but other specific complication rates were similar.

Conclusions: Owing to the limitations of our systematic review, we found that allograft-prosthesis composites and prostheses seem to have similar functional outcome and survival rates, and both seem to avoid fractures that are observed with osteoarticular allografts. Further collaboration in the field of surgical oncology, using randomized controlled trials, is required to establish the superiority of any particular treatment.

Fig. 1

The flowchart shows our literature search and selection of relevant articles. The last search was performed on June 1, 2013.

Fig. 2

Low reporting of patient and complication selection and outcome assessment increase the risk for selection and outcome bias.

Fig. 3A–C

The funnel plots show MSTS scores and standard errors for (A) prosthesis studies, (B) osteoarticular allograft studies, and (C) allograft-prosthesis composite studies. Asymmetry seems to reflect clinical and methodologic heterogeneity rather than publication bias.