A randomised trial comparing two rehabilitation approaches following reverse total shoulder arthroplasty

Abstract

Background

Rehabilitation contributes to post-operative success following reverse total shoulder arthroplasty; however, randomised trials comparing the effectiveness of rehabilitation following reverse total shoulder arthroplasty are lacking. This study sought to determine if early, active mobilisation targeting the deltoid and the external rotator muscles, would exhibit greater improvements in post-operative outcomes compared to a delayed and deltoid-focused mobilisation programme.

Methods

Patients scheduled for reverse total shoulder arthroplasty were randomly assigned to either an early active or delayed active rehabilitation group. Patient-reported outcomes for pain and function were assessed pre-surgery and at 3, 6 and 12 months post-surgery. Objective measures (Constant Score, range of motion, isometric strength) were assessed at 3, 6 and 12 months post-surgery.

Results

Sixty-one patients (63 shoulders) underwent reverse total shoulder arthroplasty. There were no significant interaction effects or between-group differences for any patient-reported outcomes or objective measures at 3, 6 or 12 months post-surgery. However, significantly better (p = 0.019) active arm flexion was observed in the early active group at three months post-surgery. Significantly more patients in the early active group reported improvement in patient-reported function that reached minimal clinically important difference from three to six months post-surgery (p = 0.016).

Conclusion

Early, active rehabilitation after reverse total shoulder arthroplasty is safe and effective, and may have early clinical benefits over a conservative, delayed mobilisation programme.

Level of evidence

Therapy, level 1b. Trial registered 15 June 2016 at www.anzctr.org.au (ACTRN12616000779471).

Introduction

Reverse total shoulder arthroplasty (RTSA) is a common treatment offered to patients with massive rotator cuff tears, rotator cuff arthropathy and osteoarthritis of the glenohumeral joint. The number of RTSA procedures has increased substantially in the last decade, now representing 69% of all shoulder arthroplasties performed in Australia, ¹ 51% in the United Kingdom ² and 33% in the United States. ³ The initial reverse design pioneered by Grammont and Lelaurain in 1987 optimises the length–tension relationship and moment arm of the deltoid muscle by distalising and medialising the centre of rotation. ⁴ With the improved biomechanical efficiency of the deltoid, RTSA has enabled good recovery of active overhead shoulder range of motion (ROM) and functional outcomes in pseudoparalytic shoulders.^5–7

Post-operative rehabilitation is recommended to restore and enhance post-operative strength following RTSA, and is considered an important factor contributing to post-operative success. ⁸ However, rehabilitation guidelines following RTSA have been largely based on low-quality evidence and expert opinion,^9,10 which typically advocate for early joint protection, gentle passive ROM, and active ROM and strengthening that prioritises the deltoid.^11,12 Furthermore, these protocols are still largely based on Grammont-style RTSA designs, which poorly utilise any remaining cuff for recovery of external rotation,^4,8 and concomitant subscapularis repair which requires prolonged soft tissue protection, delaying rehabilitation.

Advances in RTSA prosthetic design have led to prostheses with a medial glenoid, lateralised humerus (MGLH) which optimise not only the moment arms and length–tension relationship of the deltoid, but also the preserved posterior cuff muscles and posterior deltoid. Furthermore, equivalent clinical outcomes have been demonstrated in RTSA performed with and without subscapularis repair. ¹³ This may support a rationale for accelerated exercise rehabilitation, including early ROM and deltoid conditioning, as well as emphasising rotation exercises due to the improved potential for abductor and rotator muscle recruitment, potentially improving patient outcomes. ¹⁴

To date, no randomised controlled trial (RCT) has been published comparing the effectiveness of rehabilitation programmes following RTSA and to our knowledge, none specific to more recent implants using the MGLH design. The current study sought to determine if a rehabilitation programme focused on early deltoid and active rotator exercise (EA) would exhibit significantly greater improvements in post-operative shoulder strength and ROM after RTSA, compared to patients receiving a delayed active (DA), and deltoid-only rehabilitation.

Methods

Trial design

This multicentre RCT was prospectively registered in a clinical trials database (ACTRN12616000779471; Australian New Zealand Clinical Trials Registry) and the flow of participants through the trial is outlined in Figure 1. This trial was approved both by Hospital Human Research Ethics Committee, and the University of Western Australia Human Research Ethics Committee. Participants were provided with a detailed information sheet and signed a written consent form, then were randomly allocated to one of two post-operative exercise rehabilitation programmes.

Figure 1.

Patient randomisation and assessment throughout the trial for the delayed (DA) and early (EA) rehabilitation interventions after reverse total shoulder arthroplasty (RTSA).

Participants and recruitment

All consecutive patients scheduled for RTSA between March 2017 and July 2018 were screened for eligibility by three orthopaedic surgeons experienced in shoulder arthroplasty and were recruited providing they consented and met the inclusion criteria. Eligible participants were aged 55–85 years, had symptomatic massive rotator cuff tears or glenohumeral joint osteoarthritis, with or without rotator cuff dysfunction, and were subsequently recommended for RTSA after consultation with an orthopaedic surgeon experienced in shoulder arthroplasty. Diagnosis and patient selection for RTSA was made via clinical examination and confirmed on imaging using X-ray or computed tomography with ultrasound, or magnetic resonance imaging performed by an independent radiologist not associated with this study. Patients presenting with acute fractures or dislocations were deemed ineligible, as were patients undergoing revision RTSA. Patients with pre-existing conditions associated with upper extremity pain, including ongoing infection, peripheral nerve compression syndrome, cervical neuropathy or other nerve pathology were also excluded, as were patients with inflammatory arthritis. Patients with prior rotator cuff repairs were included, since studies have found no difference in outcomes after primary RTSA, with or without previous rotator cuff surgery. ¹⁵

Randomisation procedure

Patients were randomised by a senior research supervisor using a random number generator via Excel (Microsoft, Redmond, WA). While patients were recruited and consented pre-operatively, randomisation was performed two weeks after surgery, though before the onset of shoulder rehabilitation. This specified whether the patient was assigned to either the EA rehabilitation group, or to the DA group. A concealed allocation procedure was used, whereby only the study coordinator had access to the randomisation list.

RTSA surgery

All patients received RTSA under general anaesthesia in a semi ‘beach chair’ position with routine antibiotic prophylaxis. A deltopectoral approach was used in all cases, with the subscapularis tendon tagged and mobilised. A limited tenotomy of the superior edge of the pectoralis major tendon was performed for mobilisation of the proximal humerus and to improve exposure of the glenoid. All RTSAs were performed using a MGLH design (Equinoxe Reverse Shoulder Design, Exactech, Inc; Gainesville, FL) and in all cases, the subscapularis was not repaired. The long head of biceps, if present, was treated by tenotomy at the glenoid margin and tenodesed in the distal bicipital groove. Implant fixation was stable in all cases, and the prosthetic joint reduction was stable at the end of the surgical procedure.

Post-operative rehabilitation protocol

The post-operative rehabilitation protocols for both groups (EA and DA) are described in Table 1. All patients were asked to wear an immobilisation sling for six weeks and received the same immediate post-operative rehabilitation for the first two weeks. Equipment for the relevant rehabilitation programme was supplied and all participants, irrespective of their group assignment were educated regarding the nature of their surgical procedure, the rationale for exercise treatment, and the importance of compliance to their programme by an Accredited Exercise Physiologist. All patients received a ‘post-surgery home exercise guide’ detailing their exercise rehabilitation programme, including exercises, sets, repetitions and frequency.

Table 1.

Post-operative rehabilitation programmes for the delayed active, deltoid-only (DA) and early active deltoid and rotator exercise (EA) groups.

	DA group			EA group
Phase	Aims/Goals	Exercises	Dose (Sets × Repetitions, Frequency)/Load	Aims/Goals	Exercises	Dose (Sets × Repetitions, Frequency)/Load
Phase 1 (weeks 0 to 2)	Shoulder protection (immobilisation)	Initial hospital-based exercises: Wrist and hand AROM	Dosage: 2–3 × 10–15, 2× daily	Shoulder protection (immobilisation)	Initial hospital-based exercises: Wrist and hand AROM	Dosage: 2–3 × 10–15, 2× daily
Phase 2 (weeks 2 to 6)	Continued shoulder protection (immobilisation) Commencement of PROM	PROM • Self-supported FF to 90° • Self-supported pendulum • Self-assisted ER	Dosage: 2–3 × 10–15, 2× daily Rep tempo: 6 s (3 s concentric, 3 s eccentric)	Continued shoulder protection (immobilisation) Commencement of PROM and AAROM	Submaximal isometric contractions • Deltoid (anterior, middle, posterior) • Internal and external rotatorsPROM • Self-supported FF to 90° • Self-supported pendulum • Self-assisted ER • Wall-assisted ERAAROM • FF sliding on table • Supine ‘self-assisted’ elevation → supine assisted press-up with washcloth • Seated self-assisted elevation	Isometrics • 10 × 30 s, 20% MVC, 3×/dayPROM & AAROM: • Dosage: 2–3 × 10–15, 2× daily • Rep tempo: 6 s (3 s concentric, 3 s eccentric)
Phase 3 (weeks 6 to 12)	Discontinued shoulder immobilisation Progression of PROM into AAROM	AAROM • Pulley-assisted elevation • Assisted upright wall slide	Dosage: 2–3 × 10–15, 2× daily. Rep tempo: 6 s (3 s concentric, 3 s eccentric).	Discontinued shoulder immobilisation Progression of AAROM to AROM	AAROM • Pulley-assisted elevation • Assisted upright wall slideAROM • Active FF from supine → incline → upright (Figure 2).Active ER strengthening • Seated, supported active ER in 45–90° abduction (Figure 3). • Seated ER in 0 to 45° abduction.	AAROM & AROM • Dosage: 2–3 × 10–15, 2× daily. Rep tempo: 6 s (3 s concentric, 3 s eccentric).ER strengthening • Dosage: 2–3 × 10–15, 2× daily. Rep tempo: 6 s (3 s concentric, 3 s eccentric). • Load: yellow TB; DB (0.5–1 kg) • Intensity: 3 to 5 RPE
Phase 4 (weeks 12 to 20)	Activity as tolerated	Activity as tolerated	NA	Active shoulder strengthening	Anterior deltoid strengthening • Supine → seated→ standing active FF with weight (0.5 kg)ER / IR strengthening • Supine → standing external rotation in 0 to 45° abduction using TB • Seated, supported external rotation in 45–90° abduction using DB • Standing IR in 0–45° abduction with TB • Seated → standing lat pulldown with TBScapula strengthening • Seated scapula setting / isometric retractions → seated rows → standing rows (0–45° → 90° abduction) • Supine scapula protractions → gentle wall press.	Deltoid strengthening: • Dosage: 2–3 × 10–15, 2× day. Rep tempo: 6 s (3 s concentric, 3 s eccentric) • Load: 0.5 kg • Intensity: 3–5 RPEER/IR strengthening: • Dosage: 2–3 × 10–15, 2× day. Rep tempo: 6 s (3 s concentric, 3 s eccentric) • Load: TB (yellow/red/green/blue); DB (0.5 kg – 1 kg) • Intensity: 3–5 RPEScapula strengthening: • Dosage: 2–3 × 10–15, 2× day. Rep tempo: 6 s (3 s concentric, 3 s eccentric). • Load: TB (yellow/red/green/blue); DB (1–5 kg) • Intensity: 3 to 5 RPE

PROM: passive ROM; AAROM: active-assisted range of motion; AROM: active ROM; DB: dumbbell; ER: external rotation; IR: internal rotation; FF: arm flexion; MVC: maximal voluntary contraction; RPE: rating of perceived exertion; TB: theraband.

Patients randomised to the DA group were provided with a programme in line with their orthopaedic surgeon’s standard practice and similar to what has been reported previously.^7,11,16 This programme was home-based, and incorporated passive ROM exercises from two weeks, followed by active-assisted ROM exercises from 6 to 12 weeks (Table 1). No additional exercise rehabilitation was prescribed beyond 12 weeks post-surgery, though patients were encouraged to continue with their prescribed programme and return to their usual daily activities as tolerated.

Patients randomised to the EA group received a rehabilitation programme that comprised of early mobilisation and active strengthening of the deltoid and external rotators. These patients embarked on a home-based rehabilitation programme incorporating passive and selected active-assisted ROM exercises, as well as submaximal isometric deltoid strengthening exercises from two weeks post-surgery. These early exercises were selected based upon known levels of deltoid activation during early stage shoulder exercises from electromyographical studies.^17,18

From six weeks, exercises were progressed into active-assisted ROM exercises, graduated active ROM exercises and early external rotation strengthening exercises (Table 1). AROM exercises were commenced by starting with gravity-minimised exercises and progressing to inclined- and upright-assisted elevation exercises in an effort to regain active elevation (Figure 2).^17,19 This advanced set of external rotator exercises were tailored to the MGLH design of RTSA that has been shown to permit greater external rotator recruitment.^14,20–22 Specifically, exercises were selected based upon cadaveric and electromyographical studies which have suggested that greater external rotation, via the teres minor and remaining infraspinatus, may be achieved post-operatively when the humerus is positioned in abduction. ²² As such, external rotator strengthening exercises were prescribed to patients to be performed with the affected shoulder in at least 45° abduction (Figure 3).

Figure 2.

Progressive active forward flexion exercises, starting with (a) supine ‘pray and lift’; (b) supine forward flexion; (c) incline forward flexion; (d) upright forward flexion.

Figure 3.

Active external rotation strengthening with the arm positioned in 45° to 90° of abduction.

From 12 weeks, active strengthening of the deltoid internal and external rotators, and scapulothoracic muscles were prescribed and supervised at a local rehabilitation centre once per week over an eight-week period, alongside a daily home exercise programme (Table 1). All exercises were performed with sets, repetitions and intensity in keeping with strength and conditioning principles for muscular endurance made up of low load and high repetitions,^23,24 to mitigate any risk of premature glenoid failure or loosening. ²⁵ These muscular endurance dose parameters are also consistent with recommendations for resistance training in older adults. ²⁶ A range was provided (i.e. minimum and maximum sets and repetitions) for each exercise to provide scope for progression of each exercises, within each phase. This detail has also been provided in Table 1.

Clinical outcome measures

Patient-reported outcome measures (PROMs) were assessed prior to surgery (baseline time point) and at 3, 6 and 12 months post-surgery. The American Shoulder and Elbow Surgeons (ASES) score was the primary outcome variable for this study, for which validity and reliability has been demonstrated in both non-surgical and surgical patients alike, including those following shoulder arthroplasty. ²⁷ In RTSA, the ASES has previously shown to have a minimal clinically important difference (MCID) of 13.6 points ²⁸ and a difference defining a substantial clinical benefit (SCB) of 25.6 points. ²⁹ Secondary outcome variables included the Visual Analog Scale (VAS) for pain, where 0 indicates no pain and 10 indicates the highest possible level of pain, which has previously shown to have an MCID of 1.1 points ²⁸ and an SCB of 2.6 points. ²⁹ Global Shoulder Function (GSF) was also assessed, whereby 0 indicates the poorest level of function and 10 indicates the highest possible level of function. In RTSA, the GSF has previously shown to have an MCID of 1 point ²⁸ and an SCB of 2.4 points. ²⁹

The Single Assessment Numerical Evaluation (SANE) involves patients rating their shoulder from 0% to 100%, with 100% as normal. The SANE has demonstrated good agreement with other commonly used PROMs in RTSA ³⁰ and has an MCID of 28.8, and an SCB of 50.2. ³⁰ The four-dimension version of the Assessment of Quality of Life (AQOL-4D) was also used to measure health-related quality of life, which includes items describing both physical and psycho-social health, and which has shown to be reliable and valid. ³¹ Shoulder Activity Level (SAL) was used to evaluate a patient’s overall shoulder activity level based on the frequency with which he or she completes five common activities of the shoulder, by summing individual item scores, ranging from a minimum of 0 points (if a patient answers ‘never or less than once a month’ for all five items) to a maximum of 20 points (if the patient answers ‘daily’ for all five items). ³² A patient satisfaction questionnaire was also administered at 12 months after surgery to investigate patients’ level of satisfaction with the RTSA surgery overall, as well as their satisfaction with the rehabilitation programme they received. A Likert response scale was employed with the following descriptors: (1) very satisfied, (2) satisfied, (3) neither satisfied, nor dissatisfied, (4) dissatisfied, and (5) very dissatisfied.

Objective outcome evaluations were undertaken at 3, 6 and 12 months post-surgery. This included the Constant score, a 100-point scoring system that includes subjective domains of pain (0–15 points), activities of daily living (0–20 points), ROM free of pain (0–40 points), and an objective measure of strength (0–25 points). A standard goniometer was used to measure arm flexion, abduction and external rotation ROM with patients lying supine. Internal rotation ROM was assessed as the highest spinal level that the patient’s thumb was able to reach behind the back (based on the scoring system used in the Constant Score). A total of 2 points was allocated if the end of thumb could reach the greater trochanter, 4 points to the sacrum, 6 points to L3–L1, 8 points at T12–T8 and 10 points from T7 to T1. ⁵ Peak isometric abduction strength, specifically for the strength component of the Constant Score, was measured using the IDO isometer (Innovative Design Orthopaedics, Redditch, UK). Peak isometric shoulder strength was also measured for arm flexion, abduction, internal and external rotation using a hand-held digital dynamometer (Commander Powertrack II, JTech, USA), with patients positioned in supine lying and adopting standardised shoulder positions. ³³ Two trials measuring efforts involving 5-s maximal contractions were completed for both the operated and contralateral side for all measures. ³³

Blinding

Although best attempts at patient blinding were made, due to the ethical nature of this patient-informed trial, patients were made aware of the two rehabilitation pathways and, therefore, were able to determine their allocated rehabilitation group.

Sample size

An a priori power calculation was performed using G-Power (Dusseldorf, Germany) for the primary outcome variable (ASES) which demonstrated that 62 shoulders (31 per group) were required to detect a MCID of 13.6 points on the ASES score, ²⁸ at a p < 0.05 significance level and with 80% power.

Statistical analysis

Descriptive statistics were performed for patient characteristics and baseline outcomes. Continuous data on both the primary and secondary outcome measures were analysed using a two-factor, mixed-models analysis of variance, whereby rehabilitation group (EA and DA) and time (pre-surgery, 3, 6 and 12 months) were entered to measure the effects of two different rehabilitation protocols on outcomes throughout the post-operative timeline. Further comparisons were performed which included analysing within-group treatment effects at each follow-up time point (3, 6 and 12 months) and the mean differences in improvement between groups, with 95% confidence intervals (CIs) using post-hoc t tests. Participant data were analysed based on the intention-to-treat principle, meaning that patients were analysed in the group to which they were randomised, regardless of whether participants adhered to the group rehabilitation protocol. Intention-to-treat analysis was performed by using a multiple-imputation approach to handle missing data. Little’s test showed that the data were missing at random (x²= 144.5, p = 0.167).

Between-group effect sizes (mean differences between groups divided by the pooled standard deviation) for all patient-reported outcomes were measured at all post-operative time point. An effect size greater than 0.8 was considered large, around 0.5 moderate, and less than 0.2 small. ³⁴ Clinically important effects were calculated for the primary outcome variable (ASES), and the secondary outcomes (VAS Pain and GSF) to aid in determining the clinical importance of the differences. Improvement was assessed based on whether participants improved from baseline or the previous time-point by more than the MCID for ASES (13.6 points), VAS Pain (1.1 points) and GSF (1.0 points). ²⁸ All analyses were performed using SPSS software (Version 26.0, IBM, Chicago, IL) and all tests were two-tailed with alpha set at p < 0.05.

Results

Seventy patients who were scheduled for RTSA between March 2017 and July 2018 were screened by their orthopaedic surgeon for eligibility to participate in this clinical trial. Sixty-one patients who met the eligibility criteria agreed to participate and signed informed consent. Two participants underwent RTSA on both shoulders separately during the data collection period. These two patients were initially randomised into one rehabilitation pathway for their first shoulder, and subsequently allocated to the alternative pathway after surgery for their contralateral shoulder. Therefore 63 shoulders were included in the intention-to-treat analysis. Figure 1 shows a flow diagram of patient recruitment and retention to 12 months post-surgery. Both rehabilitation groups were well matched at baseline with respect to demographics, pain and patient-reported function (Table 2). Eight patients were lost to follow-up at 12 months (Figure 1). Furthermore, 10 patients failed to attend clinical follow-ups for objective assessment, mainly due to transport and logistical limitations in attending follow-ups. Therefore, 51 patients (53 shoulders) were assessed for objective outcomes of ROM and strength. All missing data were accounted for in the intention-to-treat analysis.

Table 2.

Patient demographics patient-reported outcome measures for the early deltoid and active rotator exercise (EA) and the delayed active and deltoid-only (DA) rehabilitation interventions at baseline.

Variable	EA group a	DA group a
No. of patients, n	29	32
No. of shoulders, n	30	33
Dominant arm, n (%)	22 (73)	18 (55)
Sex
Female, n (%)	18 (60)	21 (64)
Age, y	75.1 (6.1)	73.6 (6.1)
Height, m	1.65 (0.1)	1.64 (0.1)
Weight, kg	79.9 (23.1)	77.9 (13.5)
BMI	32.6 (12.8)	28.8 (4.5)
ASES	30.5 (16.9)	34.7 (15.9)
VAS pain	6.7 (1.9)	6.2 (1.9)
GSF	3.5 (2.2)	4.0 (2.0)
SANE	35.8 (22.5)	43.0 (18.2)
SAS	4.2 (4.6)	4.8 (5.3)
AQOL-4D	55.6 (19.6)	61.2 (20.4)

BMI: body mass index; ASES: American Shoulder and Elbow Surgeons score; VAS: Visual Analog Scale; GSF: global shoulder function; SANE: Single-Assessment Numerical Evaluation; SAS: Shoulder Activity Scale; AQOL-4D: Assessment of Quality of Life instrument;

^aValues are mean ± SD unless otherwise indicated.

There were no significant interaction effects or between-group differences for the primary outcome variable (ASES) or any other patient-reported outcomes (p > 0.05). Within-group time effects from baseline to 3, 6 and 12 months were observed for both groups (p < 0.001; Table 3). Furthermore, no significant between-group differences for improvement in patient-reported outcomes from baseline to 3, 6, or 12 months, for any outcome variable, was observed (p > 0.05; Table 3). Small effect sizes were generally seen throughout all patient-reported functional outcomes at all post-operative time points, except for moderate effects in GSF score at six months (Table 3). Moderate effect sizes were also observed for quality of life (AQOL-4D) scores at all time points, and activity (SAS) at 12 months (Table 3). Overall, 76% of patients in the EA group met the MCID in the GSF (>1.0 point) from three to six months, compared to 43% of patients in the DA group (p = 0.016). No other differences in the proportion of patients achieving MCIDs between groups were observed (Table 4). All patients at 12 month follow-up (n = 53) were either ‘very satisfied’ or ‘satisfied’ with the outcome of their surgery, bar one patient in the DA group who was ‘dissatisfied’ with the rehabilitation they received.

Table 3.

Effects of early deltoid and active rotator exercise versus delayed active and deltoid-only rehabilitation on patient-reported outcomes from baseline to 3, 6 and 12 months post-surgery.

	Mean (SD)		Within-group change (95% CI)		Between-group differences in change scores (95% CI)	Between-group effect size (95% CI)	p value a
Variable	EA	DA	EA	DA
ASES
3 months	70.0 (11.8)	70.9 (14.3)	39.5 (33.0 to 45.9)	36.2 (30.0 to 42.3)	3.3 (−5.6 to 12.2)	−0.07 (−0.6 to 0.5)	0.466
6 months	80.8 (11.7)	80.7 (12.6)	50.3 (42.8 to 57.8)	46.0 (38.6 to 53.4)	4.3 (−6.3 to 14.8)	0.01 (−0.5 to 0.5)	0.428
12 months	87.4 (9.5)	87.0 (9.3)	57.0 (50.4 to 63.7)	52.5 (46.8 to 58.1)	4.5 (−4.2 to 13.3)	0.04 (−0.5 to 0.6)	0.306
VAS-P
3 months	1.7 (1.6)	1.7 (1.8)	−5.0 (−5.9 to −4.2)	−4.4 (−5.1 to −3.7)	−0.6 (−1.6 to 0.5)	0 (−0.5 to 0.5)	0.282
6 months	1.1 (1.4)	1.4 (1.7)	−5.6 (−6.5 to −4.8)	−4.7 (−5.7 to −3.8)	−0.9 (−2.2 to 0.4)	−0.2 (−0.7 to 0.3)	0.161
12 months	0.5 (0.8)	0.8 (1.2)	−6.2 (−7.0 to −5.5)	−5.4 (−6.1 to −4.7)	−0.9 (−1.9 to 0.1)	−0.3 (−0.8 to 0.2)	0.080
GSF
3 months	7.3 (1.3)	7.3 (1.7)	3.8 (3.0 to 4.6)	3.2 (2.3 to 4.2)	0.6 (−0.7 to 1.8)	0 (−0.5 to 0.5)	0.384
6 months	8.5 (1.2)	7.8 (1.5)	4.9 (4.1 to 5.7)	3.8 (2.9 to 4.7)	1.1 (−0.1 to 2.2)	0.5 (0 to 1.0)	0.075
12 months	9.0 (1.7)	8.9 (1.0)	5.5 (4.4 to 6.5)	4.8 (4.0 to 5.6)	0.6 (−0.7 to 2.0)	0.07 (−0.5 to 0.6)	0.362
SANE
3 months	72.6 (15.6)	74.7 (16.2)	36.7 (26.0 to 47.4)	31.4 (22.9 to 39.9)	5.2 (−8.3 to 18.7)	−0.13 (−0.7 to 0.4)	0.448
6 months	85.2 (10.2)	83.7 (12.5)	49.4 (40.7 to 58.1)	40.8 (32.8 to 48.7)	8.5 (−3.3 to 20.2)	0.13 (−0.4 to 0.7)	0.158
12 months	91.7 (9.0)	89.9 (8.8)	55.6 (47.0 to 64.2)	47.1 (39.5 to 54.7)	8.8 (−2.5 to 20.2)	0.2 (−0.3 to 0.7)	0.128
AQOL-4D
3 months	65.7 (19.3)	71.6 (20.4)	10.0 (3.0 to 17.0)	9.6 (3.1 to 16.1)	−0.3 (−8.3 to 8.9)	−0.3 (−0.8 to 0.2)	0.942
6 months	69.5 (24.3)	78.3 (14.8)	13.3 (3.5 to 23.2)	16.1 (5.0 to 27.3)	−1.0 (−13.3 to 15.3)	−0.4 (−1 to 0.1)	0.886
12 months	68.6 (18.0)	77.7 (17.7)	12.9 (4.9 to 20.9)	14.9 (6.2 to 23.6)	−3.3 (−18.2 to 11.7)	−0.5 (−1 to 0)	0.653
SAS
12 months	9.3 (3.7)	11.0 (3.4)	5.1 (2.3 to 7.9)	6.3 (2.7 to 9.8)	−1.2 (−5.7 to 3.4)	−0.5 (−1 to 0.1)	0.605

ASES: American Shoulder and Elbow Surgeons score; VAS: Visual Analog Scale; GSF: global shoulder function; SANE: Single-Assessment Numerical Evaluation; AQOL-4D: Assessment of Quality of Life instrument; SAS: Shoulder Activity Scale; SD: standard deviation.

^aAnalyses of between-group differences for improvement in patient-reported outcomes from baseline to 3, 6, or 12 months.

Table 4.

Proportion (n, %) of patients meeting Minimal Clinically Important Differences (MCIDs) between time points in ASES, VAS pain and GSF in both rehabilitation groups.

	Minimal Clinically Important Difference (n, %)
	ASES		VAS pain		GSF
	EA	DA	EA	DA	EA	DA
Pre-surgery to
3 months	28/29 (97%)	27/29 (93%)	26/29 (90%)	26/29 (90%)	27/29 (93%)	24/29 (83%)
6 months	28/29 (97%)	26/28 (93%)	27/29 (93%)	24/28 (86%)	28/29 (97%)	25/28 (89%)
12 months	26/27 (96%)	28/28 (100%)	26/27 (96%)	26/28 (93%)	27/27 (100%)	28/28 (100%)
3 months to
6 months	13/29 (45%)	12/28 (43%)	8/29 (28%)	3/28 (11%)	22/29 (76%)*	12/28 (43%)
12 months	22/27 (82%)	18/28 (68%)	10/27 (37%)	8/28 (29%)	23/27 (85%)	20/28 (71%)
6 months to
12 months	9/27 (33%)	11/28 (39%)	5/27 (19%)	5/28 (18%)	14/27 (52%)	16/28 (57%)

ASES: American Shoulder and Elbow Surgeons; GSF: Global Shoulder Function; VAS-P: Visual Analog Scale pain.

^*p<0.05. Bolded values denote between group differences with p < 0.05.

No interaction or group differences were observed in the Constant Scores, neither were there any at each individual time point (Figure 4). No significant interaction effects or between-group differences were observed in arm flexion, abduction, external rotation and internal rotation ROM (p > 0.05; Figure 5). Post-hoc t tests revealed significantly better (p = 0.019) arm flexion ROM at three months post-surgery for the EA group (Figure 5). Within-group time effects from 3 to 6 and 12 months post-surgery were also observed for both groups (p < 0.001). Furthermore, no between-group differences for improvement in any ROM variable from 3 to 6 month, or 6 to 12 month time points were observed (p > 0.05).

Figure 4.

Mean (SE) Constant Scores at 3, 6 and 12 months post-surgery for the DA (n = 33) and EA (n = 30) rehabilitation interventions.

Figure 5.

Mean (SE) range of motion (ROM) for (a) arm flexion, (b) arm abduction, (c) arm external rotation and (d) arm internal rotation at 3, 6 and 12 months post-surgery for the DA (n = 25) and EA (n = 26) rehabilitation interventions.

Similarly, no significant interaction or between-group differences were observed in arm flexion, abduction, external rotation and internal rotation peak isometric strength at 3, 6 or 12 months post-surgery (p > 0.05; Figure 6). Within-group time effects for both groups from 3 to 6 and 12 months post-surgery were observed (p < 0.001). A significant between-group difference for improvement existed for arm flexion strength at the three- to six-month interval (mean difference, −0.8; 95% CI: −1.5 to −0.4; p = 0.038) favouring the DA group.

Figure 6.

Mean (SE) peak isometric strength for (a) arm flexion, (b) arm abduction, (c) arm external rotation and (d) arm internal rotation movements at 3, 6 and 12 months post-surgery for the DA (n = 5) and EA (n = 26) rehabilitation interventions.

Complications

Two patients reported adverse events within the first three months following RTSA, who were both randomised to the DA group. One patient suffered a traumatic fracture of the scapula, and the other an acromion stress fracture. Both patients were subsequently withdrawn from the study prior to three-month follow-up, and referred to their orthopaedic surgeon for suitable treatment.

Adherence

Of the 27 patients who were randomised to the EA programme, 13 patients, in conjunction with their home exercise programme, either attended the rehabilitation centre for a mean of 7.8 supervised therapy visits over eight weeks. Twelve patients opted to continue with their rehabilitation programme at home only and did not receive supervised rehabilitation. Two patients did not do any rehabilitation following the initial three months and were subsequently lost to follow-up. One patient, despite meeting all clinical follow-ups, did not return their completed home exercise diary and was considered ‘non-compliant’. Given the DA group did not undertake any prescribed rehabilitation from three months post-surgery, they were not tracked for compliance as to their ongoing rehabilitation routine or specific exercise habits.

Discussion

The main finding of the current study demonstrates that patients, irrespective of their allocated rehabilitation group, experience improvements in pain and function to 12 months following RTSA. Furthermore, within-group mean improvements in ASES, VAS Pain and GSF scores (as well the lower bound 95% CIs) in both groups exceed the known MCIDs at 3, 6 and 12 month follow-up. Importantly, no between-group effects, and none of large magnitude, were observed in the EA group over the DA group.

The results of this trial are similar to previous intervention studies for arthroplasty surgery of the shoulder.^35,36 Hagen et al. ³⁶ showed, like our study, that early- and delayed-rehabilitation groups result in significant improvements in ROM and PROMs. A key difference in our study was that both of our protocols were more accelerated than those used in the study by Hagen et al. ³⁶ In their study, the delayed group was restricted from any motion until the sixth post-operative week, whereas our delayed group was prescribed PROM exercise from the second post-operative week. Additionally, our EA group was prescribed submaximal isometric exercises from the second post-operative week and resisted external rotation from the sixth post-operative week, compared to Hagen et al. ³⁶ commencing these at 6 and 12 weeks respectively. Denard and Laderman ³⁵ investigated ‘immediate’ versus ‘delayed’ passive and active-assisted ROM exercises in an anatomic total shoulder arthroplasty cohort, with no differences observed for ROM and function at 3, 6 and 12 months post-surgery. ³⁵ Interestingly, at earlier time points to three months, patients in the ‘immediate’ exercise group demonstrated higher VAS, SANE and ASES scores, though at three months there were no further group differences. ³⁵ In the current study, the first post-operative evaluation was at three months, so it is possible that we did not observe early differences in functional outcomes and shoulder ROM that may have been present between the two rehabilitation groups. Future studies should look to measure outcomes at earlier time points.

Moderate effect sizes were observed for self-reported function via the GSF at six months post-surgery. Furthermore, the point estimate for the difference between mean group improvements from baseline to six months post-surgery was 1.1 points which slightly exceeds the MCID for the GSF (1 point). This may suggest that patients in the EA group demonstrate greater perceived improvement in function over the course of their rehabilitation programme compared to those in the DA group. Furthermore, this appears to be most significant between the 3rd and 6th month post-operatively, whereby a significantly greater proportion of patients in the EA group met the MCID. Furthermore, our results showed that SCB changes from pre-surgery in ASES, VAS pain and GSF were observed for both groups, which again, exceed the respective SCBs at 3, 6 and 12 month follow-up. However, whilst no between-group differences were observed and both groups met the MCID, only the EA group reported an SCB change in SANE score, defined as 50.2 point change, ³⁰ from pre-operative to 12 months post-surgery. This may be explained by EA programme’s progression into more strength-based exercises from three months post-surgery, in contrast to the DA group’s ‘activities as tolerated’ whereby no additional exercises beyond what they received at six week post-surgery. This is consistent with previous work which has shown strength in internal rotation, external rotation and arm flexion to be associated with a higher functioning shoulder after RTSA. ¹⁶ However, it is also possible that the patients in the EA group were biased towards a greater perceived benefit, given that simply interacting with a physical therapist in a rehabilitation setting is linked with reduced pain, reduced disability, and higher treatment satisfaction. ³⁷

Significantly better arm flexion ROM was observed in the EA group at three months post-surgery, which may be explained by the inclusion of early deltoid isometric exercises and progressive active-assisted and active anterior deltoid exercises within the first three months following surgery. The primary goal after RTSA is to gradually restore overhead motion with many studies having demonstrated the association between deltoid strength and active arm flexion.^38,39 It is well known that the deltoid, in particular the anterior head, is an important rehabilitation target after RTSA due to its primary role in arm flexion. ⁴⁰ While many rehabilitation protocols acknowledge the importance of deltoid strengthening following RTSA, many are conservative in their prescription due to concerns around risk of scapular notching, acromial stress fractures and implant loosening. ⁹ In a systematic review of rehabilitation protocols, Bullock et al. ⁹ reported that only three of the six protocols recommended early deltoid isometrics within the first six weeks following RTSA, and that active motion should be achieved by 12 weeks posts-surgery.

In our cohort of patients, across both groups, we observed improvements in external rotation ROM and a delay in internal rotation ROM from three months post-surgery. This is likely explained by the surgical procedure used in this study whereby patients were treated with RTSA with a MGLH, and without subscapularis repair. This surgical method uses an implant which optimises external rotator recruitment to enhance external rotation ROM. While this outcome has been reported in other studies utilising the MGLH design,^6,7 our study’s findings differ from that of Hagen et al., ³⁶ who reported no improvement in external rotation in either the immediate or delayed rehabilitation groups. While it has been reported that patients with or without repair of the subscapularis show no differences in complication rates or clinical outcomes,^13,41 an intact subscapularis may provide improved shoulder internal rotation ROM, ⁴¹ which is critical for patients to effectively manage toileting, washing of the back and, for women especially, applying and adjusting a bra strap. In the current study, by 12 months post-surgery the mean score found in both groups was below six points (reaching to the level of the 3rd lumbar vertebrae – L1). This recovery of internal rotation ROM is in line with previous research which reported a slower recovery compared to other outcome measures, including arm flexion ROM. ⁴² Triplet et al. ⁴³ also found that whilst 67% of anatomic shoulder arthroplasty patients achieved internal rotation ROM behind the back to T12, only 32% of patients in the RTSA group were able to achieve this ROM level.

No differences in isometric strength for arm flexion, abduction, internal and external rotation were observed in the EA group over DA. In fact, we observed a significant between-group improvement in isometric arm flexion strength in favour of the DA group from 6 to 12 month post-surgery, which was surprising. This may reflect the DA group, given their limited rehabilitation, took longer to attain their maximal improvement. It has been shown that patients undergoing RTSA can expect significant reductions in pain and the majority of their functional gains to occur within six months from surgery, and maximal improvement by 12 months. ⁴² Nonetheless, our findings show that the eight-week strengthening programme had no effect on strength, post-operatively, which is similar to a previous cohort study by Uschok et al., ⁴⁴ who found that a six-week physical therapy programme following RTSA, had no effect on pain and strength at a mean time of 62 months post-surgery.

The results of this study may have significant economic implications. Overall, we observed similar outcomes between the two groups, with lower costs for the group who did not undertake supervised rehabilitation from three months. In our study, mean number of post-operative therapy visits for the EA group was eight sessions. In the context of the Australian Medicare system, the total cost for patients in the self-directed pathway equated to $253, versus a mean cost of $759 for the supervised rehabilitation group. This total cost also neglects to account for the lost time and cost of patients traveling to rehabilitation visits. Given that this study did show an improvement in active forward flexion at three months, and more patients achieved the minimum MCID for GSF at 3–6 months, the early functional benefits of an EA programme may justify the increased costs for employed patients and those retirees who value an active lifestyle. It is possible that the EA programme in this study can be prescribed as a more focused home-based (unsupervised) programme to allow for these benefits, whilst incurring no significant outpatient therapy costs.

We acknowledge several limitations with this study. The primary limitations of this study include a lack of blinding and a small sample size, due to patients lost to follow-up. A larger blinded, RCT is needed to reduce any potential bias and determine if the results observed in this study are consistent in a larger population of patients. Secondly, due to logistical issues and patient convenience, pre-operative patient data were only collected for PROMs, with no objective pre-operative shoulder strength or ROM assessments undertaken. Post-operative clinical outcomes have previously been associated with pre-operative arm flexion and deltoid strength ³⁹ and, therefore, the influence these pre-operative objective measures may have had on post-operative outcome could not be assessed in this study. Thirdly, whilst imaging was used to diagnose and schedule patients for RTSA in this study, no imaging was used to quantify the muscle profile of the posterior cuff post-surgery. Advanced post-operative fatty infiltration of the deltoid and teres minor muscles have been shown to negatively correlate with clinical outcomes.^45,46 Therefore, it is not known to what degree post-operative strength and ROM may be influenced by degenerative muscle properties.

Fourthly, only patients in the EA group received supervised exercise therapy beyond three months post-surgery, and as such, could not be blinded for the exercise therapy and may be biased for positive outcome (placebo effect). However, only half of the patients randomised to the EA group (48%) complied with the programme by attending supervised visits from three months post-surgery. Nonetheless, future studies should include a comparison group that ensures equal time is spent with individuals in each group to remove any potential confounding factors. Finally, we did not formally measure compliance to the prescribed programme within the DA group. Therefore it is not known whether patients within this group maintained their exercises beyond this time point, discontinued with exercises beyond this time point, or whether they sought external physiotherapy or rehabilitation services. Consequently, this group could have undertaken more, less, or equivalent exercise rehabilitation to the EA group.

Conclusions

The results of this trial show no additional benefits of one rehabilitation strategy over another. However, these results also suggest that an early, guided rehabilitation programme is safe and well tolerated by patients and may result in greater perceived benefit in function. This is the first RCT that provides higher-quality evidence on which practitioners can base their treatment choices for patient management following RTSA using a MGLH design.