Proximal humeral fractures account for about 4%-5% of all fractures in the whole body, and are only next to the distal radial fractures in the upper limb fractures. Thereinto, approximately 80% of proximal humeral fractures have slight displacement and can be treated conservatively, while the other 20% of them have obvious displacement which requires surgical interventions. The varus impacted proximal humeral fractures account for 10% of all proximal humeral fractures, and the efficacy of conservative treatment is unsatisfactory as simple application of locking plate is easy to produce a series of postoperative complications such as loss of reduction, varus impaction, plate fixation failure, screw piercing, fracture nonunion, humeral head necrosis, etc. The implantation of allogeneic fibula into proximal humeral intramedullary provides effective medial support, and combined with locking plate, it has been widely used in the treatment of comminuted metaphyseal fractures of proximal humerus, which has achieved satisfactory curative effect. However, there are still no reports about the clinical efficacy of locking plate combined with allogeneic fibula in the treatment of varus impacted proximal humeral fractures.

(1) General information.From June 2013 to June 2015, a total of 35 patients with varus impacted proximal humeral fractures treated in our hospital were included in this study. Thereinto, 11 cases were treated with locking plate and allogeneic fibula fixation, while the other 24 cases were treated with simple locking plate fixation. The group of locking plate combined with allogeneic fibula included 6 males and 5 females with an average age of (60.27±12.32) years. The group of locking plate fixation included 10 males and 14 females with an average age of (50.21±17.60) years. There was no statistical difference between the two groups at the baseline level (P>0.05) . (2) Inclusive and exclusive criteria.Inclusive criteria:①18 years of age or older;②acute varus impacted proximal humeral fractures (within 3 weeks and neck-shaft angle <130°) treated by locking plate fixation with or without allogeneic fibular graft;③follow-up time≥1 year with complete data. Exclusive criteria: ①pathological fracture, open fractures or multiple fractures; ②combination of neurovascular injuries; ③medical history that affects shoulder function, such as previous trauma history of ipsilateral limb, chronic arthritis, etc. (3) Treatment methods.All the enrolled patients completed the relevant examinations after admission, and were treated with open reduction and internal fixation. The patients in the experimental group were treated by locking plate fixation combined with allogeneic fibular graft, while the patients in the control group were simply treated with locking plate.After brachial plexus block or general anesthesia, the patient was in beach chair position during the operation. Through the deltoid-pectoralis approach, the soft tissue was softly separated to minimize periosteal dissection. After the exposure of fracture ends, the greater and lesser tuberosities were lifted by pulling the rotator cuff insertions with sutures and the humeral head was reduced. Then the allogeneic fibula was reversely inserted into the medullary cavity of humeral shaft. With the confirmation and adjustment of the fibular length (usually 4-6 cm) , the proximal fibula was under the cartilage of humeral head. The greater and lesser tuberosities were reduced with sutures and temporarily fixed with Kirschner wires. As the fracture reduction was satisfactory under fluoroscopy, the locking plate was used for fixation with sutures knotted through the suture holes on the plate. With satisfactory position of fracture fragments, plate and screws under fluoroscopy, the wound was irrigated and closed after hemostasis. The allogeneic fibular graft was saved in the control group, and all the locking plates used in the two groups were PHILOS plates. All the procedures were performed by 3 attending surgeons or higher ranks.All the patients followed the same rehabilitation programme: the affected limb was suspended for 4 weeks before starting passive movements, and active movements were allowed 8 weeks later. All the patients were followed up in the 1st, 3rd, 6th and 12th months with anteroposterior and lateral X-ray films of shoulder joint taken every time. The data of all patients were collected, including general information, preoperative X-ray films, CT scans, X-ray films on the 1st postoperative day, X-ray films taken in each follow up and functional scores (Constant-Murley scores and DASH scores) . The neck-shaft angles were measured on the shoulder radiographs of anteroposterior view to calculate the difference between values between the 1st postoperative day and the 12th month after operation. (4) Statistical analysis.The SPSS 20.0 software was adopted for statistical analysis. The ages, neck-shaft angels and functional scores were considered as quantitative data, while the gender, fracture types, smoking or not, dominant hand or not and comminution of medial column or not were considered as qualitative data. All the quantitative data were tested by the normality test and described as mean value±standard deviation, and the independent sample t test was used to analyze whether the difference was statistically significant. The qualitative data was described as constituent ratio, and chi square test was used to analyze whether the difference was statistically significant. When P <0.05, the difference was considered statistically significant.

The fractures healed in two groups after 1 year. The change of neck-shaft angles was (-1.36±2.58) ° in the group of locking plate combined with allogeneic fibula while the angle was (-7.21±8.06) ° in the simple plate group, and the difference was statistically significant (P=0.003) . The Constant-Murley score was (76.82±6.11) points in the group of locking plate combined with allogeneic fibula while it was (64.29±9.15) points in the simple plate group, and the difference was statistically significant (P=0.0002) . The DASH score was (15.55±2.98) points in the group of locking plate combined with allogeneic fibula while it was (25.96±9.35) points in the simple plate group, and the difference was statistically significant (P=0.001) .In the respect of postoperative complications, 1 case in the group of locking plate combined with allogeneic fibula had wound infection 2 weeks after operation, and the patient achieved wound healing with positive disinfection and wound redressing. There were 3 cases of screw piercing in the simple locking plate group. Thereinto, 1 patient without obvious symptoms was given no treatment while the other 2 patients received secondary procedures to replace locking screws. Ischemic necrosis of humeral head occurred in 1 patient, which was type II of Cruss classification, and the patient was given conservative treatment.

Regarding the treatment of varus impacted proximal humeral head fractures, locking plate with allogeneic fibula can effectively increase the stability of fracture fragments, decrease the changes of neck-shaft angle, reduce the postoperative complication rate and improve the shoulder joint function. In addition, prospective randomized controlled trials are expected to further validate the relevant results.