Results of surgery using a locking plate for proximal humerus fractures in patients older than 80 years: A retrospective case series

Background: To evaluate the results of surgical treatment using a locking plate for proximal humeral fractures in patients aged >80 years. Methods: Between September of 2013 and March of 2016, there were 22 patients who received locking plate fixation from proximal humeral fractures over 80 years-old. Among the 22 cases, Clinical, radiological results were analyzed for 19 patients who were able to follow up more than one year. We analyzed bone union, neck-shaft angle, UCLA score, range of motion compared to opposite side and complication. Clinical, radiological results were investigated for medial comminuted fracture or not. Results: All the patients achieved bone union. The mean bone union time was 13.7 weeks, and the mean neck-shaft angle was 126.4. The mean University of California, Los Angeles, shoulder score was 22.4, and score was <28 point in 12 patients. The mean forward flexion, abduction, external rotation, and internal rotation angles were 129.2°, 112.3°, 44.2°, and L2. All motions were significantly different from the normal shoulder motion. A significant difference was found in the loss of neck-shaft angle according to the medial comminuted fracture. Conclusion: In the surgical treatment of proximal humeral fractures in patients aged >80 years, use of a locking plate attained bone union with relatively satisfactory results. However, we considered that prevention of and training for postoperative stiffness are necessary. Other surgical methods should be considered for patients with complex displaced fractures, especially those with medial comminuted fractures.


Background
Proximal humeral fracture is the third most common fracture and occurs in approximately 6% of total fractures, with increasing incidence recently [21]. Conservative treatment is performed in older patients with low expectations due to medical problems, and good results can be expected in the absence of displacement or in simple two-part fractures [12,15]. However, in patients with displaced, comminuted, or angulated fractures proximal humeral fractures, surgical treatment should be performed to restore shoulder function with firm fixation and early mobilization. In these cases we can perform several methods of surgery, percutaneous techniques, intramedullary nailing, plating, or arthroplasty. [3,27] Some authors recommend percutaneous pinning or screw fixation for 2-part fractures in the elderly patient. [5] [27]Intramedullary nails have the benefit of limited soft tissue dissection and blood loss for 2-par fracture, but there many complications like distraction, screw backout and nonunion. [3] Open reduction and locking plate fixation is generally chosen [16,24]. However, in elderly patients, the frequency of complications such as varus deformity and screw penetration due to osteoporosis is high. Especially in combined medial comminuted fractures, loss of fracture reduction occurs more frequently [17,20]. Therefore, the number of reverse total shoulder arthroplasty cases is increasing in elderly patients [16]. Some authors said that the preference for RSA over HA was especially high among patients aged 80 years or older, because they have a higher prevalence rotator cuff disease. [13,30] Furthermore, Patients over 80 years of age are considered a high risk group of fractures. [2,8] There are reports of surgical outcomes in older patients with proximal humeral fractures, but very few reports have been reported on locking plates in very elderly patients. In this regard, we would like to know how poor surgical results in these patients is and how many complications they have. Therefore, the purpose of this study was to analyze the results of surgical treatment using a locking plate for proximal humeral fractures in patients aged >80 years. Furthermore, we examined the radiological and clinical differences among elderly patients with medial comminuted fractures and complications of plate fixation for proximal humeral fractures.

Methods
This was a retrospective study, and final approval of informed consent exemption by the institutional review board was obtained (IRB NO. 05-2019-013).
Patients who underwent locking plate fixation for the treatment of proximal humeral fracture by the single author at one institution between September 2013 and March of 2016 were included. We performed reverse total shoulder arthroplasty for proximal humeral fractures in patients aged >80 years who had cuff arthropathy or glenohumeral arthritis as diagnosed by radiography and computed tomography (CT). Twenty-two patients aged >80 years received locking plate fixation for proximal humeral fractures. Among the 22 patients, 3 were excluded, including 2 who died within 1 year after surgery for unrelated causes and one who had a follow-up period of only 8 months. Finally, we retrospectively evaluated 19 patients (4 men and 15 women) who underwent locking plate fixation with proximal humeral fractures within the same period and were followed up for at least 1 year. The mean age at the time of surgery was 82.2±3.08 years (range, 80-91 years), and the mean follow-up duration was 15.4±5.09 months (range, 12-39 months). By using dual-energy X-ray absorptiometry, we measured the bone density of the patients' hip and spine, which according to the lowest T-score, showed a mean bone mineral density of −2.8±0.86 (range, −1.3 to −5.9). According to the Neer classification system, 9 patients had two-part fractures, 8 had three-part fractures, and 2 had fourpart fractures (Table 1).
Radiological evaluation was performed immediately and at 2 weeks, 4 weeks, 3 months, 6 months, 9 months, and 12 months after surgery. We defined successful bone union when the fracture site was completely adjoined through callus formation. Humeral neck-shaft angles (NSAs) were measured in the anteroposterior plane. We defined humeral NSA as the angle formed between the line perpendicular to the articular surface and the line following the humeral shaft. The humeral NSA was

Overall clinical results
We achieved successful bone union in all the patients, and the mean time to bone union was 13.7±1.21 weeks (range, 12-16 weeks). (Figure 2.) The humeral NSA at final follow-up was 126.4±14.23. The mean UCLA score was 22.4±6.49 points and was excellent in 2 patients, good in 5, 5 and poor in 12. Significant differences of 129.2±19.40°, 112.3±14.75°and 44.218.50° were found in the mean forward elevation, abduction, and external rotation, as compared with the opposite arm.
Significant differences were also noted in the internal rotation of L2-L3(S-T12), as compared with the opposite arm ( Table 2).
We found a statistically significant difference between the preoperative NSA and the NSA at final follow-up. In the group with medial comminuted fractures, the mean range of motion was 108.5°± 10.69° for forward elevation, 99.3°±6.07° for abduction, and 26.4°±12.48° for the internal rotation of L2. In the group without medial comminuted fracture, statistically significant differences were noted in forward elevation of 139.4°±11.64°, external rotation of 120.1°±12.79°, internal rotation of 54.5°±12.51°, and internal rotation of L4 (S-T12). The UCLA shoulder score showed significant mean differences of 18.3±1.71 and 24.7±5.63 in the groups with and without medial comminuted fracture, respectively ( Table 3).

Comparison of outcomes associated with malreduction of tuberosity
Five patients had malreduction of tuberosity after surgery.   (Table 5).
Complications were osteonecrosis in 1 patient and screw penetration in 2 patients. One patient with osteonecrosis was only observed because the pain was not severe. Penetrating screw removal was performed in 1 patient. Other patient did not want to remove the screw, So We decided to conservation for that because it did not affect joint destruction.

Discussion
Although various methods such as percutaneous K-wire fixation, plate fixation, and arthroplasty have been proposed for the treatment of proximal humeral fracture with osteoporosis, a definite treatment method has not yet been established [23]. In the case of complications such as infection and loss of reduction due to weakening of the fixation, pin fixation should be used when bone quality is good and in the absence of a comminuted fracture [9]. According to Boudard [4], the study did not show that locking plates resulted in better functional outcome, radiographic results or reduction of postoperative complications than intramedullary nail. However, many other literatures have reported 7 various complications for intramedullary nail surgery. It causes various complications such as distraction, nonunion, screw back-out, and rotator cuff tear. [3,7,27] In the case of arthroplasty, the incidence rate of complications such as dislocation, nerve damage, and infection ranges from 5% to 40% [31]. Furthermore, Patients older than 70 years and patients with chronic debilitating disease showed significantly poorer clinical outcome after hemiarthroplasty for proximal humeral fracture. [22] Reverse total shoulder arthroplasty for complex proximal humeral fractures is still controversial but RSA has been increasingly used as an alternative to hemiarthroplasty in the recent years. [29] However, RSA cause unique complications, including scapular notching, acromial fracture, and a higher incidence of implant instability. We need to long-term outcome studies to define RSA prosthesis longevity and possible late implant failure in elderly patients for proximal humeral fractures. [25,29] When we cannot achieve adequate reduction due to comminuted fracture at the surgical neck including medial calcar and severe bone loss, we can't expect a medial buttress even after reduction.
In that cases, varus deformity and loss of fracture were reported. The reason for this is that medial comminuted fracture is associated with blood flow injury, poor bone quality, and high energy damage [19]. The authors also reported that varus deformity was more severe in the medial comminuted fractures. After operation, we achieved restoration of the NSA but found no statistically significant difference at final follow-up. Clinical results also showed decreases in the range of motion and clinical performance. For this reason, Gardner et al. [12] proposed a fibular strut allograft augmentation and reported good results when the medial support was difficult.
It is often accompanied by shoulder stiffness and scapular dyskinesia after proximal humeral fracture.
Thus, various rehabilitation methods have been introduced to prevent this [12,28]. Aggressive rehabilitation can reduce the many discomforts associated with stiffness. The authors presented various exercise methods to patients, including scapular, stretching, and strengthening exercises.
However, in elderly patients, compliance was poor and aggressive rehabilitation was difficult because of pain, and these factors are also considered to contribute to the poor clinical outcome.
Open plate fixation was generally good, but complication and reoperation rates of 15% and 12.7% 8 were reported, respectively [11]. Various complications such as malunion, loss of fracture reduction, metal failure, nonunion, and osteonecrosis have been reported. In 8% of patients, screw penetration was the most common cause of reoperation [18]. Two patients had screw penetration, one of whom had reoperation to remove the screw. To prevent this, during the operation, the length of the screw should examined using C-arm imaging. Furthermore, the far-cortex should be examined with a depth gage, based on which the screw length can be decided.
Furthermore, we need to understand the screw configurations according to the plate type and proceed with the surgery. Of course, preoperatively we have to check evaluation of bone quality of the patient, accurate analysis of the fracture type, plate selection.
The incidence rate of osteonecrosis after surgical treatment of proximal humeral fractures is reported to be 3-35% [26]. Especially in the proximal humeral head, after 48 hours, the frequency of this is high. In these patients, the frequency of shifting to arthroplasty was high.(80%) The authors also performed surgical treatment in 1 patient on postoperative day 4, which resulted in osteonecrosis.
However, considering the age and activity of this patient, we decided to perform conservative treatment.
Recently, the number of reverse total shoulder arthroplasty cases has been increasing among complex humeral fractures. The results are excellent as compared with those of hemiarthroplasty, and the 5-and 10-year survival rates are 94% and 91%, respectively. However, the technique of reverse total shoulder arthroplasty for fractures was found to be more difficult than the procedure for rotator cuff disease, and the functional score decreased from 6 years after the procedure, so that the Constant-Murley score was maintained at ≥30 points in 60% of patients only [14]. Generally, the complication rate of reverse total shoulder arthroplasty is reported to be 19-68%. Furthermore, complication rate of reverse total shoulder arthroplasty for fracture vary among studies and range from 5% to 40% such as periprothetic fracture, infection, dislocation and nerve paly. [14] Cheung et al reported that treatment of the these complications is difficult. [6] Therefore, reverse total shoulder arthroplasty is not an easy choice for treating proximal humeral fractures even in the elderly [10].
One of the study's limitations is that, as a retrospective study, the findings may have been influenced 9 by selection bias. Although we considered Neer classification and medial comminution, we did not make an analysis of the severity of fractures exactly. Second, the total number of cases are only 19, therefore it is difficult to anticipate a statistically significant difference. Third, because the study design consisted of a relatively short follow-up, long-term complication, such as osteonecrosis, arthritis, cannot be out-ruled. Lastly, It has the disadvantage of not comparing with other surgical methods.

Conclusions
In the surgical treatment of proximal humeral fractures in patients aged >80 years, use of a locking plate attained bone union with relatively satisfactory results. However, postoperative shoulder motion was significantly reduced compared to normal. Therefore, it seems that active physical therapy is    Data are presented as mean ± SD.