Functional outcomes of reverse shoulder arthroplasty compared with open reduction and internal fixation for acute proximal humeral fractures

Abstract

Introduction

Reverse shoulder arthroplasty (RSA) was traditionally used in treatment of massive rotator cuff arthropathy. In recent years, this surgical modality was adopted to be an alternative in the treatment of acute complex fracture of proximal humerus for geriatric patient and this management is gaining its popularity. This is a local study that aims at comparing the functional outcomes among the two surgical options in treatment of acute complex fracture of proximal humerus (3-parts, 4-parts, and fracture dislocation), i.e. open reduction and internal fixation (ORIF), and RSA.

Methods

A retrospective review was conducted on patients aged ≥60 years who underwent RSA or ORIF for acute proximal humeral fractures between 2015 and 2023 at a single local center. Data were extracted from the Clinical Management System. Outcomes included Oxford shoulder score (OSS), range of motion (ROM), revision rates, and intra-operative parameters. This study incorporates computed tomography-based stratification of Neer classification to optimize fracture characterization and ensure appropriate treatment allocation. Statistical analysis employed Chi-square test for categorical comparison, and ANOVA for group comparisons.

Results

60 patients were included (29 ORIF, 31 RSA). RSA patients were older (mean age 77 vs. 67 years, p < 0.001) with higher Charlson Comorbidity Index (CCI) scores (p = 0.040). For 4-part fractures, RSA showed superior OSS (41.2 vs. 30.2, p = 0.006) despite older age/poorer CCI. For 3-part fractures, no significant differences were observed in OSS or ROM. Revision rates were higher in ORIF (6.9% vs. 0%, p = 0.137).

Conclusion

RSA is preferable for 4-part fractures due to better functional outcomes, even in older patients with poorer comorbidities. ORIF remains appropriate for 3-part fractures and dislocations, offering comparable outcomes with less operative time and cost. Still, careful case selection, pre-operative planning, good surgical technique, and rehabilitation are the keystones for successful surgery.

Keywords

RSA ORIF proximal humeral fracture functional outcome RSR reverse shoulder arthroplasty

Introduction

Acute proximal humeral fracture is the third most common fragility fracture, after hip fracture and distal radius fracture in Hong Kong and its incidence is continually rising due to ageing of population.¹ In order to manage an acute fracture of proximal humerus, orthopedic surgeons have to consider patient's factors such as their premorbid functional status, comorbidities, hand dominance, expected functional demands, and disease factors such as the fracture patterns and classification.

Neer classification is the most commonly used system for classification of the patterns of fracture proximal humerus.^2,3 By this classification, the more displaced parts the fracture has, the more complex the fracture is and it is also believed to be associated with a higher risk of avascular necrosis of humeral head^2–5 and poorer functional outcome scores after open reduction and internal fixation (ORIF).⁶ Besides Neer classification, ones can also predict the risk of AVN by using the Hertel criteria.⁷ If the risk of AVN is high, surgeons, therefore, may reasonably prefer shoulder arthroplasty to internal fixation.⁷

Non-displaced or minimally displaced fracture of proximal humerus can be treated with conservative treatment with satisfactory outcomes while for Surgical management is typically recommended for patients with acute proximal humeral fracture involving Neer type 3 or 4, fracture dislocation, or head-splitting pattern as it can result in improved functional outcomes and quality of life.^8–10 The options of surgical management included in our study are ORIF and reverse shoulder arthroplasty (RSA).

This study aims at comparing the functional outcome between these surgical treatments of acute complex fracture of proximal humerus.

Statistical analysis

All the parameters were compared between two groups using Chi-square test and ANOVA tests, as appropriate. Chi-square test was used to compared the categorical parameters including the gender, revision rate, associated neurovascular injuries, injury involving dominant hand or not, postoperatively rehabilitation allowing free mobilization or not between the two groups while one-way ANOVA test was used for comparing the numerical data including the age, premorbid Charlson Comorbidities Index (CCI), intra-operative parameters(operative time and blood loss), postoperative functional score (Oxford Shoulder Score (OSS))¹¹, and range of motion(ROM) (Forward flexion and abduction). ANOVA test was also used in subgroup analyses based on fracture type (3-part, 4-part, and fracture-dislocation patterns) to evaluate outcome differences between ORIF and RSA groups. A p-value < 0.05 was taken to indicate statistical significance.

Methodology

Data collection and comorbidity assessment

The premorbid CCI¹² was calculated using International Classification of Disease diagnosis codes retrieved from our hospital's Clinical Management System (CMS, Electronic Health Records). This index categorizes comorbidities based on patient age and the severity of 19 predefined medical conditions, with higher scores indicating poorer comorbidity status and prognosis. Additional baseline data (age, sex, fracture patterns, hand dominance) and intra-operative parameters (blood loss and operative time) were also extracted from CMS. To ensure secure data handling, all information was transferred and stored using encrypted USB devices.

Fracture classification

All patients underwent pre-operative computed tomography (CT) with multiplanar reconstructions alongside standard radiographs. CT imaging allowed detailed assessment of fracture patterns, fragment displacement, and intra-articular involvement, enabling precise Neer classification. This approach minimized diagnostic uncertainty and improved stratification of 3-part, 4-part, and fracture-dislocation patterns for surgical planning.

Surgical protocol

All surgeries were performed for acute proximal humeral fractures by a dedicated group of surgeons under the supervision of a single senior surgeon (to standardize techniques). Patients were positioned in the beach-chair position, and a deltopectoral approach was used. The ORIF group received fixation using the DePuy Synthes PHILOS plating system, while the RSA group received the DePuy Synthes Delta XTEND Reverse Shoulder System. Intra-operative time and blood loss were recorded in CMS.

Outcome measures

OSS: Functional outcomes were assessed at 1 year postoperatively using the OSS, a validated 12-item questionnaire (total score: 48) evaluating pain and shoulder function. A single-trained interviewer conducted all OSS assessments via structured phone interviews after obtaining verbal consent. To minimize bias, the interviewer followed a standardized script, though blinding to treatment allocation was not feasible due to the nature of the study.

ROM: Postoperative ROM (forward flexion, abduction, external rotation) was measured at 1-year follow-up in our Specialist Outpatient Clinic using a standardized goniometric protocol. Measurements were performed by the treating orthopedic surgeon during routine clinical consultation, with the patient seated and the scapula manually stabilized. To enhance reliability, reference was also taken from independent measurements performed by physiotherapists (not involved in the surgical team) during separate rehabilitation sessions using the same protocol. The surgeon's measurements served as the primary reference for analysis, while physiotherapist-recorded values provided supplementary data for reliability assessment.

Rehabilitation

Both ORIF and RSA groups share similar supervised rehabilitation by our hospital's physiotherapy team including early mobilization, gradual range-of-motion exercises, and strengthening exercises.

Results

During the study period, there were 31 patients underwent RSA, 29 underwent ORIF in our hospital.

Patient demographics

The mean age of the patients underwent RSA (76.3, p = 0.000) was significantly higher compared with ORIF group (Table 1). More female patients were noted in RSA group (p = 0.009). In terms of fracture patterns, more 4-part fractures were treated by RSA (Figure 1) while more 3-part fractures were treated by ORIF (Figure 2) though it is not significant (Table 2). Moreover, RSA group (4.3, p = 0.04) has a significantly higher premorbid CCI than ORIF group(3.4). There is no significant difference between the two groups in terms of whether the injury involved the dominant hand, and any associated neurovascular injury.

Table 1.

Patient demographics.

	RSA	ORIF	P-value
Age(years)	77.4(63–90)	66.9(60–83)	0.000^a
Female sex	29(93.5)	19(65.5)	0.007^a
CCI	4.3(2–9)	3.4(2–9)	0.04
Hand dominance (%)	61.3	41.3	0.196
Associated NV injuries (%)	0	6.90	0.229
Fracture patterns			0.000^a
3-parts	4(16.7)	20(83.3)
4-parts	21(80.8)	5(19.2)
Fracture dislocation	6(60)	4(40)

Age and CCI are recorded as mean (range) and female sex as number of patients (percentage of group of patients).

Hand dominance and NV injuries are recorded as percentage.

Fracture patterns are recorded as number of patients (percent).

Statistically significant.

Abbreviations: RSA: reverse shoulder arthroplasty; ORIF: open reduction and internal fixation; CCI: Charlson Comorbidity Index.

Table 2.

Fracture patterns (numbers of patients).

	RSA	ORIF
3-parts	4	20
4-parts	21	5
Fracture dislocation	6	4

Abbreviations: RSA: reverse shoulder arthroplasty; ORIF: open reduction and internal fixation.

Two patients in the ORIF group (6.9%, 2 out of 29) presented with nerve injuries associated with their initial trauma. The first case involved a contralateral ulnar nerve palsy while the second case had a complete brachial plexus injury evident at initial presentation. Both nerve injuries were managed conservatively: the ulnar neuropathy resolved completely within 6 months, while the brachial plexus injury showed partial recovery but retained residual ulnar nerve deficits at final follow-up. These were monitored with serial clinical examinations and electrodiagnostic studies when indicated.

Figure 1.

Reverse shoulder arthroplasty done for 4-part fracture proximal humerus.

Intra-op parameters

The intra-operative parameters including the operative time and intra-operative blood loss between the two groups are compared. Generally, the operative time of the ORIF group (148.2 min) is shorter than RSA group (161.9 min, p = 0.132) (Table 3) and the intra-operative blood loss of ORIF group (354.5 ml) is less than the RSA group (404.8 ml, p = 0.32). Yet, both results of the intra-operative comparisons are not statistically significant.

Table 3.

Intra-operative parameters.

	RSA	ORIF	P-value
Operative time(mins)	161.9	148.2	0.132
Blood loss(ml)	404.8	354.5	0.32

Abbreviations: RSA: reverse shoulder arthroplasty; ORIF: open reduction and internal fixation.

Functional outcomes

We included the postoperative ROM (forward flexion and abduction), OSS, and revision rate at 1 year after the operations of the two groups for comparisons .

Figure 2.

Open reduction and internal fixation done for 3-part fracture proximal humerus.

In overall, the RSA group had higher range of forward flexion and abduction (130 deg, p = 0.329; 128 deg, p = 0.195) than the ORIF group (122 degrees, p = 0.329; 116 degrees, p = 0.195) but not statistically significant (Table 4).

Table 4.

Postoperative range of motion.

	RSA	ORIF	P-value
Forwards flexion (mean°)	130.6	122.8	0.329
Abduction (mean°)	128.4	116.7	0.195

Data are presented as mean ± standard deviation.

Abbreviations: RSA: reverse shoulder arthroplasty; ORIF: open reduction and internal fixation.

The RSA group had higher OSS (41.1 p = 0.091) than the ORIF group (37.6) but again not statistically significant (Table 5).

Table 5.

Postoperative outcome.

	RSA	ORIF	P-value
Oxford shoulder score	41.1	37.6	0.091
Revision rate (%)	0	6.9	0.229

OSS is presented as mean ± standard deviation.

Revision rate is presented as percentage of patient (number of patients).

Abbreviations: RSA: reverse shoulder arthroplasty; ORIF: open reduction and internal fixation; OSS: Oxford Shoulder Score.

There was no revision case in the RSA groups while ORIF group had 6.90% (2 out of 29) of revision rate within the study period. The reason for revision was malunion as well as implant failure, i.e. screws protruding into shoulder joint and the PHILOS plate was removed and converted to RSA (Figure 3).

Figure 3.

(a) ORIF with plating; (b) complicated with screws intra-articular protrusion; (c) revision surgery as reverse shoulder arthroplasty. ORIF: open reduction and internal fixation.

Subgroup analysis

3-part fracture type

No significant differences in terms of the intra-operative parameters and postoperative function were found except the RSA group was significantly older than the ORIF group (Table 6).

Table 6.

3-part fracture.

	RSA	ORIF	P-value
Age	80.75	66.65	0.001^a
CCI	4.0	3.6	0.713
Operative time(mins)	151.5	140.1	0.440
Blood loss(ml)	325	362.5	0.698
ROM(FF)	140°	123.5°	0.304
ROM(Abd)	127.5°	116.8°	0.602
Oxford shoulder score	38.75	39.4	0.885

Statistically significant.

Abbreviations: RSA: reverse shoulder arthroplasty; ORIF: open reduction and internal fixation; CCI: Charlson Comorbidity Index; ROM: range of motion.

4-part fracture type

RSA performed significantly better in terms of OSS despite RSA patients being significantly older and having higher CCI scores (Table 7).

Table 7.

4-part fracture.

	RSA	ORIF	P-value
Age	77	67.6	0.019^a
CCI	4.33	2.8	0.035^a
Operative time(mins)	160.1	159.8	0.986
Blood loss(ml)	421	420	0.987
ROM(FF)	129°	113°	0.369
ROM(Abd)	127.6°	108°	0.279
Oxford shoulder score	41.2	30.2	0.006^a

Statistically significant.

Abbreviations: RSA: reverse shoulder arthroplasty; ORIF: open reduction and internal fixation; CCI: Charlson Comorbidity Index; ROM: range of motion.

Fracture-dislocation type

All the comparisons have no statistical significance (Table 8).

Table 8.

Fracture dislocation.

	RSA	ORIF	P-value
Age	76.3	67.5	0.143
CCI	4.5	3.25	0.333
Operative time(mins)	174.8	174.3	0.983
Blood loss(ml)	400	232.5	0.402
ROM(FF)	131.25°	130°	0.949
ROM(Abd)	127.5°	131.7°	0.835
Oxford shoulder score	42	38	0.378

Statistically significant.

Abbreviations: RSA: reverse shoulder arthroplasty; ORIF: open reduction and internal fixation; CCI: Charlson Comorbidity Index; ROM: range of motion.

Discussion

In our study, the RSA groups had a larger proportion of female patients than that of the ORIF group. Generally female patients had a higher risk of developing osteoporosis and the fragility fractures.¹³ While the severity of fracture increases with the age of the population,¹⁴ we tend to perform joint replacement surgery to these patients in fear of the higher complication risks such as humeral head collapse, avascular necrosis, and nonunion, if ORIF was performed.^15–17

The RSA group of patients were significantly older (77.4, p = 0.00) and had poorer premorbid status in terms of higher Charlson Premorbid Index (4.3, p = 0.04)¹⁸^,¹⁹. Yet, it still performed better in terms of postoperative functional ROM (p = 0.329 & 0.195), and OSS (p = 0.091) although it is statistically insignificant. This result is in line with previous literature studies^20–23 The reason for choosing RSA as the surgical treatment in those patients with poorer premorbid status is due to the consideration of associated poorer outcome for ORIF in those patients, e.g. malunion, nonunion, AVN, etc.²⁴^,²⁵ Hardeman et al.²⁰ showed that there was worse clinical and radiological outcome when performing ORIF in older patients in terms of failure rate and revision rate. On the other hand, Younger patients with better bone quality were more likely to undergo ORIF, as there was a desire to preserve their bone stock due to their potentially higher risk of engaging in risky activities that could lead to complications and subsequent re-operations.^26,27

For the intra-operative characteristics, all of the surgeries performed made use of the same position (Beach chair) and deltopectoral approach. The difference in blood loss and operative time, therefore, lied on the technical demands, procedures and the familiarities of the surgeons to the operative steps.

Generally, the RSA group required a longer operative time and more blood loss than ORIF group although the result is not statistically significant. This difference could be because ORIF is a more familiar surgery to most surgeons and it usually requires less steps and procedures in terms of bone defect management and soft tissue tensioning. The occurrence of the new anatomical pre-contoured locking plates also gave ease to the reduction and stability of the fractures and the better bone quality in the younger patients also helps secure good screws purchase.

In terms of postoperative functional outcomes, the RSA group demonstrated better ROM (forward flexion and abduction), OSS, and revision rate than ORIF group. Particularly, the difference is statistically significant in comparing the OSS in 4-part fracture. The superior outcomes in the RSA group can be attributed to the procedure's better mechanical advantages and its ability to accommodate rotator cuff insufficiency.²⁸

Many literatures had supported the above statement. Klein et.al²⁹ suggested RSA use in comminuted fracture proximal humerus in elderly patients because it showed a good clinical outcome and short intervention time. Another local study also showed that RSA performed better in terms of functional scores and postoperative ROM (Forward flexion and abduction) postoperatively 1 year onwards.²² Delphi trial from JBJS 2020 suggested RSA performed superiorly in the complex proximal humeral fracture (displaced OTA/AO type-B2 and C2, equivalent to Neer 3-part and 4-part fracture) in geriatric population.²¹

There are several limitations in our study. Firstly, the premorbid characteristics of the patient groups are not equivalent. For example, the mean age of the RSA group is significantly higher and the premorbid status was significantly more complex. This could, therefore, underestimate the functional outcome of RSA group.

Secondly, the pre-operative OSS was not measured. This important information may be one of the determining factors that leads to the different outcomes among the groups. And the interviewer for OSS was not blinded to the treatment allocation which may introduce bias in subjective outcome reporting.

Thirdly, although postoperative ROM was assessed using a standardized goniometric protocol, measurements were performed by different clinicians in outpatient settings, which may introduce inter-observer differences.

Fourthly, the sample size of our study is insufficient to establish a fully scientific model that can lead to a definitive conclusion, which may introduce the possibility of a type II error.

Fifthly, the subgroup comparison (3-parts, 4-parts, fracture dislocation) were imbalanced in sample size between the two groups. This disparity may give rise to selection bias and hence reduce the statistical power to detect the true difference

Finally, during the study period, both the RSA and ORIF groups followed similar rehabilitation protocols, including early mobilization, gradual range-of-motion exercises, and strengthening exercises. However, unlike ORIF, RSA currently lacks a standardized rehabilitation protocol, which may introduce variability in postoperative recovery. Moving forward, establishing an evidence-based rehabilitation strategy for RSA will be a key focus of our research. Recent studies on accelerated rehabilitation protocols have shown promising outcomes,^30–33 suggesting a potential framework for future standardization.

Conclusion

RSA is a preferable option for 4-part fractures due to better functional outcomes, even in older patients with poorer comorbidities. On the other hand, ORIF remains appropriate for 3-part fractures and dislocations, offering comparable outcomes with less operative time and cost. Future studies with larger and comparable sample size are needed to validate these results. Careful case selection, pre-operative planning, good surgical technique, and rehabilitation are the keystones for successful surgery.

Footnotes

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Sui Kit Chan

Yuk Chuen Siu

Chun Man Ma

References

Wong

RMY

Wai

, et al. Fragility fractures and imminent fracture risk in Hong Kong: one of the cities with longest life expectancies. Arch Osteoporos 2019; 14: 104.

Neer

2nd.

Displaced proximal humeral fractures: I. Classification and evaluation. J Bone Joint Surg Am 1970; 52: 1077–1089.

Carofino

Leopold

. Classifications in brief: the Neer classification for proximal humerus fractures. Clin Orthop Relat Res 2013; 471: 39–43.

Mohan

Hintze

Morrissey

, et al. Incidence of avascular necrosis following biceps tenodesis during proximal humerus open reduction and internal fixation. Acta Orthop Belg 2021; 87: 339–346. PMID: 34529390.

Kilcoyne

, et al.

The Neer classification of displaced proximal humeral fractures: spectrum of findings on plain radiographs and CT scans.

AJR Am J Roentgenol 1990; 154: 1029–1033.

Sproul

Iyengar

Devcic

, et al. A systematic review of locking plate ﬁxation of proximal humerus fractures. Injury 2011; 42: 408–413.

Hertel

Hempfing

Stiehler

, et al. Predictors of humeral head ischemia after intracapsular fracture of the proximal humerus. J Shoulder Elbow Surg 2004; 13: 427–433.

Erpala

Tahta

Öztürk

, et al. Comparison of treatment options of three- and four-part humerus proximal fractures in patients over 50 years of age. Cureus 2021; 13: e17516.

Mao

Zhang

Peng

, et al. Comparison of surgical versus non-surgical treatment of displaced 3- and 4-part fractures of the proximal humerus: a meta-analysis. J Invest Surg 2015; 28: 215–224.

10.

Charlson

Pompei

Ales

, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987; 40: 373–383.

11.

Dawson

Rogers

Fitzpatrick

, et al. The Oxford shoulder score revisited. Arch Orthop Trauma Surg 2009; 129: 119–123.

12.

Dawson

Fitzpatrick

Carr

. The assessment of shoulder instability. The development and validation of a questionnaire. J Bone Joint Surg Br 1999; 81: 420–426.

13.

Cawthon

. Gender differences in osteoporosis and fractures. Clin Orthop Relat Res 2011; 469: 1900–1905.

14.

Roux

Decroocq

El Batti

, et al. Epidemiology of proximal humerus fractures managed in a trauma center. Orthop Traumatol Surg Res 2012; 98: 715–719. Epub 2012 Sep 19. PMID: 23000039.

15.

Dai

Chai

Wang

, et al. Meta-analysis comparing locking plate fixation with hemiarthroplasty for complex proximal humeral fractures. Eur J Orthop Surg Traumatol 2014; 24: 305–313.

16.

Gupta

Harris

Erickson

, et al. Surgical management of complex proximal humerus fractures—a systematic review of 92 studies including 4500 patients. J Orthop Trauma 2015; 29: 54–59.

17.

Misra

Kapur

Maffulli

. Complex proximal humeral fractures in adults—a systematic review of management. Injury 2001; 32: 363–372.

18.

Cappellari

Trovarelli

Andriolo

, et al. Reverse shoulder arthroplasty for treatment of proximal humerus complex fractures in elderly: a single institution experience. Injury 2022; 53: S2–S7.

19.

Greiwe

Kohrs

Callegari

, et al. Open reduction internal fixation vs. reverse shoulder arthroplasty for the treatment of acute displaced proximal humerus fractures. Seminars in Arthroplasty: JSES 2020; 30.

20.

Hardeman

Bollars

Donnelly

, et al. Predictive factors for functional outcome and failure in angular stable osteosynthesis of the proximal humerus. Injury 2012; 43: 153–158.

21.

Fraser

Bjørdal

Wagle

, et al. Reverse shoulder arthroplasty is superior to plate fixation at 2 years for displaced proximal humeral fractures in the elderly: a multicenter randomized controlled trial. J Bone Joint Surg Am 2020; 102: 477–485. Erratum in: J Bone Joint Surg Am. 2020 Jun 17;102(12):e63.

22.

Tong

Fang

. Rehabilitation progress following reverse total shoulder replacement and internal fixation for geriatric three and four-part proximal humerus fractures - a propensity score matched comparison. BMC Musculoskelet Disord 2023; 24: 566.

23.

Yahuaca

Simon

Christmas

, et al. Acute surgical management of proximal humerus fractures: ORIF vs. hemiarthroplasty vs. reverse shoulder arthroplasty. J Shoulder Elbow Surg 2020; 29: S32–S40.

24.

Barlow

. Complications after ORIF of proximal humerus fractures-how to deal with them. AOJ 2021; 6.

25.

Klug

Wincheringer

Harth

, et al. Complications after surgical treatment of proximal humerus fractures in the elderly-an analysis of complication patterns and risk factors for reverse shoulder arthroplasty and angular-stable plating. J Shoulder Elbow Surg 2019; 28: 1674–1684.

26.

Rosas

Law

Kurowicki

, et al. Trends in surgical management of proximal humeral fractures in the medicare population: a nationwide study of records from 2009 to 2012. J Shoulder Elb Surg 2016; 25: 608–613.

27.

Dixit

Cautela

Cooper

, et al. ORIF versus arthroplasty for open proximal humerus fractures: nationwide inpatient sample data between 1998 and 2013. J Orthop Traumatol 2018; 19: 12.

28.

Grammont

Trouilloud

Laffay

, et al. Concept study and realization of a new total shoulder prosthesis. Rhumatologie 1987; 39: 407–418.

29.

Klein

Juschka

Hinkenjann

, et al. Treatment of comminuted fractures of the proximal humerus in elderly patients with the Delta III reverse shoulder prosthesis. J Orthop Trauma 2008; 22: 698–704.

30.

Edwards

Ebert

Joss

, et al. A randomised trial comparing two rehabilitation approaches following reverse total shoulder arthroplasty. Shoulder Elbow 2021; 13: 557–572.

31.

Boardman

3rd Cofield

Bengtson

, et al. Rehabilitation after total shoulder arthroplasty. J Arthroplasty 2001; 16: 483–486.

32.

Boudreau

Higgins

, et al. Rehabilitation following reverse total shoulder arthroplasty. J Orthop Sports Phys Ther 2007; 37: 734–743.

33.

Bullock

Garrigues

Ledbetter

, et al. A systematic review of proposed rehabilitation guidelines following anatomic and reverse shoulder arthroplasty. J Orthop Sports Phys Ther 2019; 49: 337–346. Epub 2019 Apr 25. PMID: 31021690.