Results of open reduction and internal fixation of proximal humerus fractures using a proximal humeral plating system with smooth pegs

Abstract

Background

The aim of this study was to evaluate clinical results and complication rate in patients with proximal humeral fracture treated with Anatomical Locking Plate System (A.L.P.S.) Proximal Humeral Plating System with smooth pegs.

Methods

Retrospective evaluation of 116 consecutive patients treated with A.L.P.S. Proximal Humeral Plating System. Shoulder function was evaluated by Constant score and disabilities of arm, shoulder and hand (DASH). Furthermore, European Quality of life 5 Dimensions (EQ-5D) and single shoulder value (SSV) were assessed. Fractures were classified according to Neer.

Results

Fifty-two patients were included with a mean follow-up of 49 months. Mean Constant score was 62.8. Mean DASH was 17, and the mean EQ-5D was 0.91. According to Neer classification 63% (n = 33) of the fractures were two-part, 31% (n = 16) three-part, and 6% (n = 3) four-part fractures. Peg penetration was registered in 13 cases (25%, n = 52). In total a reoperation rate of 23% (n = 12) and an overall complication rate of 37% (n = 19) were recorded.

Discussion/conclusion

Open reduction and plate fixation with the A.L.P.S. Proximal Humeral Plating System with smooth pegs provides a stable fixation with good clinical outcome and complication and reoperation rates in line with previous studies with plates using screws in the humeral head.

Introduction

Fractures of the proximal humerus are the third most observed osteoporosis-related fractures in the elderly. ¹ Patients > 60 years and women are overrepresented. ² The incidence of proximal humeral fractures has been rising, mainly because of population ageing, and studies have suggested that the incidence will continue to increase steadily over the next decades.^3,4 This trend is a public health concern because it may cause higher medical costs and loss of autonomy for the patients. ⁵

Although the majority of the proximal humeral fractures are treated conservatively, about 20% of the fractures are complex and treated surgically.^6,7 The operative options available are open reduction and internal fixation (ORIF) using locking plates or intramedullary locking nails, primary hemiarthroplasty and reversed shoulder replacement. ⁸ However, the optimal surgical treatment method of complex proximal humeral fractures is controversial. ORIF using locking plates have been used extensively, but several studies have shown high complication rates such as humeral head avascular necrosis (AVN), screw perforation, loss of reduction and subacromial impingement with plate.^9,10 A well-established implant with widely reported good outcome is the proximal humeral internal locking system, PHILOS (Synthes, Solothurn Switzerland).^11,12 Nevertheless, an overall complication rate of 35%, particularly due to screw perforation, has been reported. ⁹ To reduce the screw perforation and to provide optimum subchondral support in osteoporotic bone the screws were replaced by smooth pegs in the S3 Proximal Humerus Plate (Zimmer Biomet, Warsaw, Indiana, USA) and a retrospective study indicated a lower rate of complication, such as screw/peg perforation of the humeral head. ¹³ The successor Proximal Humerus Plating System based on the Anatomical Locking Plate System, A.L.P.S., (Zimmer Biomet, Warsaw, IN, USA), has slightly different designed features with more gracile smooth pegs, a multidirectional medial calcar locking screw, increased number of suture holes and the ability to choose between two designs for high or low placement of the plate. ¹⁴ Only a few studies have investigated the outcome of this novel locking plate system and the results have suggested lower total rate of complications compared to the well-established locking screw plates.^15,16 The aim of this study was to retrospectively evaluate the complication rate and the clinical functional outcome among patients with displaced proximal humeral fracture treated with the A.L.P.S. Proximal Humeral Plating System using smooth pegs.

Materials and methods

From 1^st of January 2017 until 22^nd of December 2018, 116 consecutive patients with displaced proximal humeral fractures were treated at our institution with ORIF using the A.L.P.S. Proximal Humeral Plating System. After ethic committee approval was obtained the patients were contacted by letter or telephone. Medical charts were reviewed, and these patients were offered to participate in the study. If the patient declined participation, medical charts and existing x-rays were reviewed and complications and fracture union rates were recorded.

Clinical evaluation was completed by a physiotherapist including questionnaires. Complications were identified by reviewing medical records and radiographic evaluation.

One physiotherapist carried out all the clinical follow-ups. The physiotherapist evaluated shoulder function by measuring Constant score of the affected and contralateral arm. Shoulder strength in 90° abduction was measured with a spring balance (Isobex 4.0, Medical Device solutions AG, Bern, Switzerland). Painful test of strength or inability to reach abduction of 90° resulted in a strength score of 0. This is according to recommendations. ¹⁷ In those cases where the patient was unable to come to the clinic and meet the physiotherapist a patient reported Constant score was used. ¹⁸ Disabilities of arm, shoulder and hand (DASH), European Quality of life 5 Dimensions (EQ-5D), single shoulder value (SSV) was also used for evaluation. SSV was assessed by the physiotherapist where the patients were asked to evaluate their shoulder function on a scale of 1% to 100% of a normal shoulder. EQ-5D was calculated and interpreted with Swedish reference values. ¹⁹

Radiographic evaluation

A follow-up x-ray including a frontal, an axial and a subscapular view were performed. If considered necessary, a frontal view tangential to the plate was added. The images were independently interpreted by a 4^th-year radiology resident, a 4^th-year orthopedic resident and a senior orthopedic surgeon specialized in shoulder surgery. Furthermore, the acute x-ray and CT scan (if available), postop images and x-rays taken before follow-up were also retrospectively reviewed, and the fractures were classified according to Neer classification. ²⁰

If a smooth peg in the humeral head was protruding through the cortical surface of caput humeri it was recorded as a penetration. AVN was defined as collapsed humeral head with loss of sphericity. By comparing postop x-rays with follow-up x-rays, a loss of reduction could be identified and recorded.

Surgical technique

All surgeries were performed by orthopedic surgeons subspecialized in shoulder surgery. Patients were positioned in a beach chair position and underwent general anesthesia with or without ultrasound-guided interscalene nerve block, which was decided on a case-by-case basis.

A prophylactic single dose of 600 mg Clindamycin (Pfizer, New York, USA) was intravenously administered 30 minutes before surgery to all patients. A standard deltopectoral incision was used. All patients underwent intraoperative fluoroscopy to determine reduction of the fracture, plate position and peg/screw placement. The final result was documented either with intraoperatively taken fluoroscopy images or postoperative conventional x-rays (Figure 1).

Figure 1.

Displaced proximal humeral fracture treated with open reduction and internal fixation with the A.L.P.S. Proximal Humeral Plating System. (A) Preoperative image. (B) Postoperative radiograph.

Postoperative treatment

The patients were given a sling for 7 to 10 days—but were allowed to use the arm for simple ADL as soon as possible. After 2 weeks, physiotherapy was started with assisted active motions, and after 4 to 6 weeks active range of motion was allowed. A routine follow-up with radiographs was performed after 8 to 12 weeks.

Data collection and statistical analysis

Data was gathered through patient journals, follow-up x-rays, patient reported questionnaires and the clinical follow-ups performed by the physiotherapist. All data was registered in a database in Microsoft Excel. The complications risk was estimated as the proportion complications in all patients with an exact binominal 95% confidence interval.

Fisher exact test was used when the different fracture types were compared to determine any difference between two-part and three-part fractures. Odds ratio was calculated with a p-value calculated according to Sheskin.

In the analyses, where complications according to age were compared, a risk ratio between patients < 70 and > 70 years was calculated. Odds ratio was calculated with the group < 60 years as reference and a p-value calculated according to Sheskin.

Results

Patient demographics

Out of the 116 identified eligible patients, 52 accepted the invitation to the study, 36 women and 16 men. Of the remaining 64 patients, 7 patients had no follow-up with x-rays and were considered lost to follow-up. Furthermore, 13 patients were deceased, leaving 44 patients declining to participate in the study.

The main reason not to participate was morbidity, advanced age, and the ongoing COVID-19 pandemic. All of the 52 subjects, with a mean age of 69 (range 49–83, SD 8), were followed-up with x-ray during November 2021 to July 2022. Out of these 52 patients, 4 did not meet with the physiotherapist. Two patients could not be evaluated because of health issues, 2 did not participate due to difficulties traveling to the hospital. One patient met the physiotherapist, but was unable to cooperate and the result therefore was not included. Four of these 5 patients did, however, answer a patient reported Constant score which is included in the result.

Before the final decision of surgery, 19 out of 52 patients were examined with CT scan in addition to the acute standard x-ray. The mean time to surgery after the trauma was 8 days. In total, six different orthopedic surgeons, all subspecialized in shoulder surgery, were involved in the treatment. In 83% (n = 43) of the cases, osteosutures or sutures through rotator cuff tendon fixed to the plate, was performed to further stabilize fracture fragments. At the time of surgery five patients reported being smokers. In 46% (n = 23) the treated fracture was in the dominant arm.

Clinical results

The mean Constant score of the affected arm was 62.8 (range 23.5–87.7, SD 15). Among the five patients that responded with the patient reported Constant score the mean score was 57.2 (range 42.5–100) and none of these reported pain in the contralateral arm/shoulder. The mean relative Constant score when compared to the contralateral shoulder was 78.9% (range 34–106%, SD 15.7%). Among the patients reporting no problems with the contralateral side the mean relative constant score was 77.5% (range 34–100%, SD 15.6%).

Among the patients who experienced one or more complications the mean Constant score was 51.5 (n = 13, range 23.5–72.9, SD 15.5), five of the patients with complications did not meet with the physiotherapist, the reasons for this are explained in patient demographics. Patients with no complications had a mean Constant score of 67.2 (range 28.4–87.7, SD 12.4).

The mean DASH was 17 (SD 17.5, median 11.36, range 0–72.7, n = 50). 48 patients reported SSV with a mean of 80% (SD 15.5%, range 40–100%). Mean EQ-5D was 0.91 (SD 0.08, range 0.58–0.91, N = 51). 51 patients reported the EQ-VAS with a mean of 78.4 (SD 17.7, range 15–100).

At the follow-up with the physiotherapist, 12% (n = 6) of the patients reported problems with the contralateral arm. These problems were subacromial pain (n = 2), osteoarthritis (n = 1) of the shoulder, pain from COVID-19 vaccination (n = 1), neck-and shoulder pain (n = 1) and pain from prior surgery to the contralateral elbow (n = 1).

No deep or superficial infections were seen. No nerve injury was reported.

Radiological results

The fractures were classified on the acute radiographs and CT scans. According to Neer classification 63% (n = 33) of the fractures were assessed as two-part fractures, 31% (n = 16) as three-part fractures and 6% (n = 3) as four-part fractures (Table 1). Out of the patients with two-part fractures, one patient had a fracture with an anterior dislocation and one patient had a proximal diaphyseal/metaphyseal extension of the fracture. Half of the cases (n = 26) had a preoperative varus malalignment, 23% (n = 12) a valgus malalignment and 27% (n = 14) were neutral.

Table 1.

Fracture classification.

Fractures according to Neer classification	No. of patients
Two-part	33
Three-part	16
Four-part	3

The follow-up x-ray was performed 37 to 63 months after fracture (mean value 49 months, SD 7.43). All fractures of the included 52 patients were healed. In 34 cases no complications were detected. Peg perforation of the humeral head was registered in 13 cases (25%, n = 52). None of these were seen in the peroperative or direct postoperative x-ray, in other words no primary peg perforation was detected. Of the 13 patients with peg penetration four had AVN with secondary peg perforation. An additional two patients suffered AVN without peg perforation making it a total of six cases with AVN (12%, n = 52). Of these six patients with AVN, one had a four-part fracture (17%, n = 1), two had a three-part fracture (33%) and three had a two-part fracture (50%). Out of the six patients with AVN, three fractures had a preoperative varus malalignment, two fractures had a neutral alignment and one fracture showed valgus malalignment. One patient with neutral alignment and a two-part fracture had an anterior dislocation. The mean age among AVN patients was 68.7 years (n = 6, SD 12.2) compared to 68 years in the patients without AVN (n = 46, SD 12).

Loss of reduction with no peg perforation was registered in two cases (6%, n = 52). One patient developed a big osteophyte inferior to caput humeri. One patient with peg perforation also suffered a fracture to the coracoid process, which was treated conservatively, but since this was not a consequence of the surgery it was not included in the result (Table 2).

Table 2.

Complications and reoperations.

Complications	Quantity	Reoperation
Primary peg penetration	0	0
Secondary peg penetration	9	7
Avascular necrosis (AVN)	2	0
Secondary peg penetration and AVN	4	4
Loss of reduction/malunion	2	1
Symptomatic osteophyte	1	0
Rotator cuff rupture	1	0
Infection	0	0
Nerve injury	0	0
Total	19	12

Out of the 13 patients with secondary peg perforation, reoperation with change of pegs or removal of plate was performed in six cases. One of these patients first underwent a change of peg and subsequently plate and peg removal. In three patients who had secondary peg perforation and AVN the plate and pegs were removed and replaced by reverse total shoulder arthroplasty. The fourth patient who had AVN and secondary peg perforation was treated with removal of the hardware only. One patient with loss of reduction but no peg perforation was treated with removal of hardware and arthrolysis. Another patient was registered with loss of reduction and secondary osteoarthritis, but this patient was not interested in reoperation. In one case a massive rotator cuff rupture was diagnosed 8 months postoperatively. This patient was also offered reoperation with reversed arthroplasty but declined. The patient who developed a big osteophyte suffered from pain and limited range of motion and is therefore planned for reoperation. In total 12 reoperations in 11 different subjects were performed, thus a reoperation rate of 23% (n = 12) and an overall complication rate of 37% (n = 19) were recorded.

Patients with three- or four-part fractures had a tendency for higher risk for complications, although not statistically significant (Table 3). There was also a tendency towards higher risk of complications with a higher age. However, this was not statistically significant (Table 4). Time to surgery after diagnosis was 7.8 days (range 1–30, SD 5.6) in the group without complications and 7.5 days (range 1–17, SD 4.5) in the group with complications.

Table 3.

Complications per fracture type according to Neer classification.

Fracture type	N	N complications	Risk (%)	Exact 95% CI	Fisher P	Odds ratio (P)
2	33	8	24%	11–42		Reference
3	16	7	44%	20–70	0.20	2.4 (0.17)
4	3	1	33%	1–91		1.56 (0.73)

Table 4.

Risk for postoperative complications according to age.

Age	N	N complications	Risk (%)	Exact 95% CI	Risk ratio (P)	Odds ratio (P)
<=60	9	2	22.2	2.8–60		Reference
61–70	19	7	36.8	16.2–61.6		2.04 (0.44)
71–80	22	9	40.9	20.7–63.6	1.29 (0.57)	2.42 (0.33)
>80	2	1	50	1.3–98.7

The fractures of the remaining 57 patients that were not participating in the study, but had routine follow-up with x-rays 3–4 months postoperatively, were regarded as healed. Two patients had x-rays showing a slight peg penetration, approximately 1 mm, but did not need surgery. Four patients had undergone revision surgery, one had the plate, screws and pegs removed due to painful stiffness, and three were revised to RSA—one because of malunion and two because of AVN.

Discussion

Treatment of proximal humeral fractures is controversial and the benefit from surgical intervention is discussed. In this study using A.L.P.S. Proximal Humeral Plating System with smooth pegs in the humeral head, the clinical results and complications was comparable with results reported with conventional locking plates with screws.

The mean Constant score in our study was 62.8, which is comparable to the mean value of 64.4 reported by Hansson et al. using the S3 plate. ¹³ Olerud et al. reported a Constant score of 61 and a mean DASH of 32 using the PHILOS plate. ²¹ In the present study the mean DASH was 17. Using the PHILOS plate Brunner et al. showed a higher mean Constant score, but a comparable DASH score. ⁹ The difference in Constant score in these studies could be contributed to differences in patient selection, evaluation/interpretation and the way Constant score is measured.

The relative Constant score (Constant compared to uninjured arm) was 79%, which is lower than booth Hansson et al. (87%) and Brunner et al. (87%).^9,13

In our study 13 (25%) of the patients experienced peg perforation, although four of these were secondary to AVN and humeral head collapse. If we exclude these four patients, nine (17%) patients experienced peg perforation.

The fact that no primary peg perforations were seen could be due to surgeon's experience. In our clinic, all involved surgeons were experienced with this technique since another plate with smooth pegs was used prior to switching to the A.L.P.S. Proximal Humeral Plating System.

The rate of peg perforation is slightly higher than that shown by Hansson et al. when using the S3 plate, which also uses smooth pegs proximally. ¹³ In their study, a peg perforation rate of 13.6% was seen. This study was also conducted at our clinic, with the same surgeons and routines. The increase in peg perforation seen in our study could be due to the small differences in plate design, including a smaller diameter of the pegs. With smaller diameter pegs it is easier to perforate the articular surface when drilling the canal for the pegs. Fluoroscopy is therefore recommended to avoid drilling too deep. An increased risk for peg penetration can be seen when the tips of the pegs are close to the cortex on the direct postoperative x-rays, in combination with a slight compression of the head fragment during healing of the fracture.

Previous studies with locking plates using screws have shown an incidence of cut out ranging from 14 to 22%.^9,21,22 Chen et al. compared the A.L.P.S. Proximal Humeral Plating System and PHILOS plate retrospectively, although there was a trend of fewer cut outs using the A.L.P.S. Proximal Humeral Plating System there was no statistical significance. ¹⁶ Our study shows a secondary peg perforation which is comparable to other studies using the locking plate with screws. Our result is consistent with Bønes et al. supporting the hypothesis that pegs do not reduce the frequency of perforations to the glenohumeral joint when compared to screws. ²²

When examining 4700 proximal humeral fractures between 2011 and 2017 using the Swedish Fracture Register, Bergdahl et al. concluded plate fixation had the highest reoperation rate with 21%. ²³ The register does not specify if locking screws or smooth pegs were used but in our study the reoperation rate was comparable suggesting that implant choice might not be the determining factor since a different plate was widely used in Sweden during this period.

Conclusions

This study shows that open reduction and plate fixation with the A.L.P.S. Proximal Humeral Plating System provides a stable fixation with good clinical outcome and complication and reoperation rates in line with previous studies with plates using smooth pegs or screws in the humeral head. To find any significant differences a RCT would probably be needed.