This was a retrospective study including patients who underwent osteosynthesis for a fracture of the greater tuberosity at three municipal general hospitals between 2008 and 2021. We included adult patients who underwent surgery for isolated greater tuberosity fracture diagnosed by plain radiographs and computed tomography (CT). Several previous studies have recommended that a superior displacement of the greater tuberosity of ≥ 5 mm is an indication for surgery of the greater tuberosity fracture [8, 9] as it is thought to cause abnormal shoulder mechanics during shoulder abduction [10] and subacromial impingement [8, 11]. Therefore, the indications for surgery for greater tuberosity fractures at the institution where this study was performed were patients who could undergo general anesthesia and had a superior displacement of greater than 5 mm. We determined a superior displacement when the superior margin of the greater tuberosity fragment was ≥ 5 mm superior to the superior margin of the articular fragment of the humeral head on the anteroposterior view of plain radiography or the coronal view of CT, regardless of the fracture pattern. We excluded patients with other fractures complicating the affected upper extremity, history of surgery involving the affected upper extremity, and paralysis of the affected upper extremity due to cerebral infarction or other causes, and patients who underwent osteosynthesis using intramedullary nail.

Eleven orthopedic surgeons performed surgery. In all cases, surgery was performed in the beach-chair position under general anesthesia. Osteosynthesis was performed using the delto-pectoral or deltoid split approaches using plates in 24 patients, cannulated cancellous screw (CCS) in 16 patients, transosseous wiring or suture in 15 patients, suture-bridge technique in 8 patients, and tension band wiring (TBW) in 5 patients, at the discretion of the surgeon. The implant used for plate fixation was the PHILOS® plate (Depuy Synthes, Oberdorf, Switzerland), LCP® plate (Depuy Synthes, Oberdorf, Switzerland), or MODE® plate (MDM, Tokyo, Japan). The implant used for CCS fixation was the ACE® (Zimmer Biomet, Warsaw, IN, USA) or Asnis® III cannulated screw system (Stryker, Kalamazoo, MI, USA). For transosseous wiring or suture, a surgical wire or FiberWire® (Arthrex, Naples, FL, USA) was fastened through the rotator cuff and the bone hole created distally in the humeral fragment. For the suture-bridge procedure, suture anchors were inserted proximally and distally to the fracture site and the bone fragments were reduced and fixed. Healix Advance™ BR anchor (Mitek, Raynham, MA, USA), JuggerKnot® anchor (Zimmer Biomet, Warsaw, IN, USA), and Quattro® Link Knotless Anchor (Zimmer Biomet, Warsaw, IN, USA) were used. TBW was performed using Kirschner wires and surgical wires, AI-Wiring system (Aimedic MMT, Tokyo, Japan) or RING PIN system (Nakashima Medical, Okayama, Japan). In this study, the mean operative time and blood loss for osteosynthesis performed in this study were 100.5 ± 32.5 min and 65.5 ± 98.1 g, respectively. In eight cases, drainage was performed by inserting an SB VAC™ (Sumitomo Bakelite, Tokyo, Japan) into the fracture site for 2 days after osteosynthesis. Immobilization of the arm was achieved by sling fixation postoperatively for 1–3 weeks after which passive range of motion (ROM) exercises were initiated. Active ROM exercises were initiated 4–6 weeks after surgery. Patients did not undergo additional fixation periods even when inferior subluxation occurred postoperatively.

Various methods of humeral head inferior subluxation have been previously reported [3, 12, 13]. Carbone et al. define inferior subluxation as a distance of ≥ 1 cm between the humeral anatomical neck level and the glenoid inferior edge level [3]. We adopted this method for this study because a good intra- and excellent inter-rater reliability was reported [3]. Based on the previous studies [3, 13], plain radiographs of the shoulder in the upright position taken before surgery and at 1 week, 1 month, 3 months, and 6 months postoperatively were evaluated by one examiner. Inferior subluxation on the plain radiograph taken 1 week postoperatively was defined as inferior subluxation immediately after osteosynthesis, as previously described [2] (Fig. 1).

Radiographic assessment of inferior subluxation of the humeral head. A distance of ≥ 1 cm between the humeral anatomical neck level and the glenoid inferior edge level was defined as the presence of humeral head inferior subluxation. Postoperative radiograph images after surgery using a cannulated cancellous screw (A) or tension band wiring (B)

The patients were further divided into two groups according to the presence of inferior subluxation at 1 week postoperatively: patients with inferior subluxation (+ IS group) and patients without inferior subluxation (−  IS group).

Multivariate analysis was performed to clarify the factors affecting postoperative subluxation, and inferior subluxation on plain radiograph at 1 week postoperatively was used as the dependent variable. The explanatory variables were age, sex, affected side of the shoulder, body mass index (BMI), history of smoking, local osteoporosis, time period from injury to surgery, preoperative axillary nerve injury, fracture dislocation, preoperative inferior subluxation, surgical approach (delto-pectoral/deltoid split approach), surgical method (plate, CCS, transosseous wiring, suture-bridge technique, TBW), operative time, amount of blood loss, and drainage after surgery. Local osteoporosis was assessed by measuring the average cortical thickness at two points of the humerus, and an average proximal humeral cortical thickness of ˂ 6 mm was defined as local osteoporosis as previously reported [14]. Axillary nerve injury was assessed using clinical notes on numbness of the axillary nerve region.

Postoperative outcomes were postoperative complication rate (delayed bone union, nonunion, infection, screw perforation into the joint, fixation failure) and ROM (elevation and external rotation [ER] at side) at 6 months after surgery. A single evaluator, who was blinded on the results of postoperative inferior subluxation, investigated the postoperative complications based on the clinical notes and plain radiographic images. Delayed union was defined as a lack of bone bridging at 6 months postoperatively. We defined fixation failure as a residual displacement of the greater tuberosity fragment of ≥ 5 mm. Postoperative ROM was assessed by the surgeon who performed osteosynthesis or the occupational therapist. We compared the postoperative outcomes between the + IS and −  IS groups during a follow-up period ≥ 6 months.

All statistical analysis was conducted using SPSS software (version 27.0*, IBM, Armonk, NY, USA). In univariate analyses, we used the Mann–Whitney U test to compare the average of continuous values (age, BMI, time from injury to surgery, operative time, and blood loss). We used Fischer’s exact test to compare the proportions (sex, side of injury, smoking, local osteoporosis, preoperative axillary nerve injury, fracture dislocation, preoperative inferior subluxation, surgical approach, surgical method, and postoperative drainage). Baseline variables with P < 0.05 in univariate analysis were included in the logistic regression analysis to clarify the independent predictive factors of inferior subluxation. The regression model fit was estimated by the Hosmer–Lemeshow goodness-of-fit test. On the examination of the effect of subluxation on surgical outcomes, the Mann–Whitney U test was used to compare the average of ROM and Fischer’s exact test was used to compare the complication rate. P < 0.05 was considered statistically significant.