Study design
The present study was a hospital-based, prospective cohort trial and was approved by the Medical Ethics Committee for Clinical Research of Zhongda hospital.
Inclusion criteria: (1) age ≥ 60 years; (2) 2-part or 3-part proximal humeral surgical neck fracture, according to the Neer system [20]; (3) fresh fracture ≤ 3 weeks old; and (4) systematic functional rehabilitation exercises performed after surgery.
Exclusion criteria: (1) pathological fracture; (2) humeral anatomical neck fracture; (3) concomitant humeral head fracture; (4) isolated tuberosity fracture; (5) fracture dislocation; (6) open fracture; (7) multiple fractures of the ipsilateral upper extremity; (8) previous or active infection in the shoulder; (9) frozen shoulder, irreparable rotator cuff tear, or plexus or nerve injury; (10) surgery history in ipsilateral shoulder; and (11) loss to follow-up.
Participants
From June 2014 to November 2016, 216 patients with a proximal humeral fracture were treated with surgery in our hospital, and 63 patients who were willing to participate and met the inclusion and exclusion criteria were divided into the nail group (TriGen®, Smith & Nephew, London, UK) and the plate group (PHILOS®, DePuySynthes, Oberdorf, Switzerland) at the discretion of the surgeon and finally of the patient.
Four patients were lost to follow-up at 12 months postoperatively. At 24 months postoperatively, 4 patients were lost to follow-up, and 2 patients died of a cause unrelated to the surgery. The flowchart is shown in Fig. 1.
Preoperative evaluation
We obtained a detailed medical history from every patient. All patients underwent an X-ray examination including an anteroposterior view of the shoulder and a lateral view of the scapular plane, as well as CT for 3D reconstruction to evaluate the fracture according to the Neer classification system.
Interventions and surgical procedures
All surgical procedures were performed by senior trauma surgeons with more than 15 years of clinical experience. All patients underwent general anesthesia combined with an interscalene brachial plexus block in the beach chair position. All patients received prophylactic intravenous antibiotics within 30 minutes before the surgery.
Nail group
We used a syringe needle to confirm the entry point preliminarily under C-arm guidance, and an approximately 4.5-cm-long incision was made parallel to Langer’s line, lateral to the lower acromion. The deltoid was split between the anterior and middle thirds. We fixed and stretched the deltoid muscles to protect the axillary nerve. If necessary, the “joystick” technique [21] and Kirschner wires could be used to fix the fracture temporarily. Care was taken not to block the path of the nail. A 1-cm-long incision was made in the supraspinatus muscle close to the tendon in line with its muscle fiber orientation. The incision was optimized according to the position of the rotator cuff tear. Under radioscopic imaging, the entry point was located posterolateral to the long head of the biceps tendon in the sulcus between the greater tuberosity and the humeral head margin. The guidewire was introduced after the medullary canal was opened with an awl, and after the fracture was fully reduced, a medullary canal was reamed. The nail was inserted into the medullary cavity along with the guidewire until the nail was placed 4 mm below the cartilage. Then, at least 3 proximal locking screws and 2 distal locking screws were inserted. Additional tension band sutures were tied around the proximal screw heads to fix the rotator cuff if a comminuted tuberosity was observed, based on the so-called “rope-over-bitt” principle [22]. Finally, an end cap was screwed on, fracture reduction and accurate screw positioning were examined by radioscopic imaging, and a drainage tube was placed before wound closure.
Plate group
The fracture was exposed via the standard deltopectoral approach. A small bundle of deltoid muscles was first split and stretched to the internal side to wrap and protect the cephalic vein. The hematoma and embedded soft tissue were removed before fracture reduction. Kirschner wires were used to fix the fragments temporarily. Then, a locking plate was positioned on the proximal part of the humerus along the lateral side to the bicipital groove. To avoid subacromial impingement, the plate was placed approximately 5 mm inferior to the tip of the greater tuberosity and 4 mm lateral to the bicipital groove. Next, a cortical screw was used for distal fixation, and the plate position was examined by radioscopic imaging. A minimum of 5 proximal locking screws and 3 distal locking screws were inserted. Here, a calcar screw could be used to provide medial column support. Radioscopic imaging was used to evaluate fracture reduction and implant position. Finally, the rotator cuff tear was carefully repaired with nonabsorbable tension band sutures through the proximal plate holes, and a drainage tube was placed before wound closure.
Postoperative procedures
Antibiotics were used continuously for 24 hours after surgery. Postoperative analgesia consisted of Celebrex at 400 mg the first time, followed by 200 mg twice daily for a minimum of 5 days. Narcotic analgesics could be used if necessary. The drainage tube was removed after 48 hours or if the drainage fluid volume was less than 50 ml within 24 hours after surgery.
Postoperative rehabilitation
Slings were used for 4 weeks. On the operative day and the first postoperative day, active exercises of the fingers, wrist and elbow were performed. Passive forward elevation and abduction were performed on the second postoperative day, and passive external and internal rotation were introduced gradually. Active-assisted exercises were started after 6 weeks, and active antigravity exercises were introduced when callus formation was observed on radioscopic imaging. Improved muscle-strengthening and active-resistance exercises were started after 12 weeks. Detailed rehabilitation protocol is given in a table [see Additional file 1].
Demographic characteristics
We recorded the sex, age, injured shoulder, Neer classification, coronal deviation, interval between injury and surgery, number of medial cortical comminutions and follow-up period.
Intraoperative parameters
The intraoperative parameters included the operative duration, incision length, blood loss, and hemoglobin loss at 24 hours postoperatively.
Postoperative radiographic parameters
Radiographs were obtained for the first time after the drainage tube was removed and at 1.5, 3, 6, 12 and 24 months postoperatively. The parameters included the neck-shaft angle (NSA) immediately postoperatively and at 24 months postoperatively, NSA loss, and time to callus formation and bony union. Bony union was defined upon the formation of callus bridging the main fracture fragments in 80% of the cortex.
Functional outcomes
Functional results were evaluated using the American Shoulder and Elbow Surgeons (ASES) score, Constant–Murley score, and visual analog scale (VAS) score at 3, 6, 12 and 24 months postoperatively. In addition, the VAS score was recorded on the first postoperative day. Forward elevation, abduction, and external rotation were measured by a goniometer and are expressed as an angle (°), and internal rotation was measured according to the height of the back touched by the thumb, which was ranked using a continuous number.
Complications included tuberosity resorption, loss of reduction of the greater tuberosity (≥ 0.5 cm), varusmalunion (NSA ≤ 120°), screw penetration, subacromial impingement, humeral head avascular osteonecrosis, refracture, infection, iatrogenic nerve injury, iatrogenic rotator cuff tear and nonunion.
Statistical analysis
All data were analyzed using SPSS Statistics 20.0 (SPSS, IBM, New York, USA). Continuous variables are presented as the mean and standard deviation and were analyzed using Student’s t-test when normally distributed or the Mann–Whitney U test when non-normally distributed. Categorical variables are presented as absolute numbers and were analyzed using the chi-squared test. In addition, internal rotation and the VAS scores are presented as the median and range and were analyzed using the Mann–Whitney U test. P< 0.05 was considered statistically significant.