Complication rates and outcomes stratified by treatment modalities in proximal humeral fractures: a systematic literature review from 1970-2009.

Background: The optimal treatment of complex, displaced proximal humeral fractures is controversial. A systematic literature review of the time period from 1970 to 2009 was conducted. The purpose was to evaluate the clinical success and complications of the available treatment modalities to determine specific treatment recommendations for the different fracture patterns.
Methods: The databases (PubMed/EMBASE) were searched for the time period (01/1970-09/2009). Study quality, treatment modalities, classification, outcome scores and complications of 200 publications including 9377 patients were analyzed. Interventions were compared by analysis of variance with subsequent Tukey's-test. Complication rates among methods were compared by using Pearson's-chi-square-test and pairwise comparisons using Fisher's-two-tailed-exact-test.
Results: Hemiarthroplasty, angle-stable plate and non-operative treatment were used for 63% of the follow-up-patients. For 3- and 4-part fractures, patients with hemiarthroplasty [3-Part: 56.4 (lower/upper 95% confidence interval (CI): 43.3-68.7); 4-Part: 49.4 (CI: 42.2-56.7)] received a lower score than different surgical head-preserving methods such as ORIF [3-Part: 82.4 (CI: 76.6-86.9); 4-Part: 83.0 (CI:78.7-86.6)], intramedullary nailing [3-Part: 79.1 (CI:74.0-83.4)] or angle-stable plates [4-Part: 66.4 (CI: 59.7-72.4)].The overall complication rate was 56%. The most common complications were fracture-displacement, malunion, humeral head necrosis and malreduction. The highest complication rates were documented for conventional plate and hemiarthroplasty and for AO-C, AO-A, for 3- and 4-part fractures. Only 25% of the data were reported with detailed classification results and the corresponding outcome scores.
Discussion: Despite the large amount of patients included, it is difficult to determine adequate recommendations for the treatment of proximal humeral fractures because a relevant lack of follow-up data impaired subsequent analysis. For displaced 3- and 4-part fractures head-preserving therapy received better outcome scores than hemiarthroplasty. However, a higher number of complications occurred in more complex fractures and when hemiarthroplasty or conventional plate osteosynthesis was performed. Thus, when informing the patient for consent, both the clinical results and the possibly expected complications with a chosen treatment modality should be addressed.