Outcomes of reverse shoulder arthroplasty in patients under 55 years old: Results from the New Zealand joint registry

Abstract

Background

The appropriate age at which to perform reverse shoulder arthroplasty is controversial. The aim of this study was to compare the outcomes of reverse shoulder arthroplasty between younger and older patients.

Methods

Patients who underwent primary reverse shoulder arthroplasty between January 2000 and December 2019 were identified from New Zealand Joint Registry records. Patients were stratified into two cohorts according to age at the time of surgery: < 55 years and ≥ 55 years. These two groups were then compared with regard to baseline characteristics, indications for surgery, revision rates, and patient-reported outcomes using the Oxford Shoulder Score and American Shoulder and Elbow Score (ASES).

Results

A total of 5518 primary reverse shoulder arthroplasty cases were identified, with 75 patients < 55 years at the time of surgery (range: 34–54 years). The mean duration of follow-up was 2.36 years (range: 0.11–13.37 years) in the younger cohort and 3.10 years (range: 0.01–16.22 years) in the older patient cohort. Indications for surgery differed significantly between the two groups, with younger patients having higher rates of inflammatory arthritis (p < 0.001), posttraumatic arthritis (p < 0.001), and avascular necrosis (p = 0.049). The younger cohort had an inferior 6-month postoperative Oxford Shoulder Score compared to the older cohort (mean: 28.5 [younger cohort] vs. 35.7 [older cohort]; p < 0.001). There was no significant difference in revision rate between the younger and older patient cohorts during the study period (1.56 [<55 years] vs. 0.74 [≥55 years] revisions per 100 component-years; p = 0.332).

Conclusion

Our early results suggest that younger patients undergoing reverse shoulder arthroplasty demonstrate high implant retention rates, comparable to older patients. Longer-term patient-reported outcomes in younger patients are required to guide appropriate patient selection for reverse shoulder arthroplasty.

Level of evidence

Level III, retrospective case-control study.

Introduction

Reverse shoulder arthroplasty (RSA) was developed as a treatment option for elderly low-demand patients with cuff tear arthropathy (CTA) or glenohumeral osteoarthritis with significant rotator cuff dysfunction.^1–3 Among the elderly population, favorable long-term outcomes can be expected^1,4 and recently indications for RSA have expanded to include a range of glenohumeral pathologies, such as proximal humeral fracture and malunion,^5,6 posttumor reconstruction, ⁷ and inflammatory arthropathies.

While favorable outcomes have consistently been reported following reverse shoulder arthroplasty, the majority of published literature focuses on elderly patients with lower functional demand.^1,2,8,9 On the other hand, there remains a paucity of data on the long-term outcomes of RSA performed in younger patients, which presents unique challenges regarding functional outcomes, implant longevity, and the requirement for future revision surgery. Furthermore, younger patients tend to present with more complex glenohumeral pathology, with higher rates of inflammatory arthropathies, capsulorrhaphy arthropathy, trauma, and posttraumatic sequelae.^10,11

Recent studies investigating the outcomes following RSA in patients younger than 60 years of age have yielded encouraging results.^12–17 However, only one study to date has specifically evaluated the outcomes in patients younger than 55 years of age, ¹³ and long-term clinical outcome studies are currently lacking. Therefore, controversy remains regarding the expected implant longevity or the exact age and indication for RSA in younger patients.

The purpose of this study was to assess outcomes in patients younger than age 55 years of age undergoing RSA and compare the functional outcomes and implant retention rate with older patients undergoing the same procedure. The authors hypothesized that younger patients had higher revision rates and lower functional outcomes when compared to older patients.

Materials and methods

After a National Health and Disability Ethics Committee approval was granted, data was collected from the New Zealand (NZ) Joint Registry for all patients undergoing primary RSA between 1 January 2000, and 31 December 2019. Patients were stratified into two groups according to age at the time of surgery: < 55 years and ≥ 55 years. The two patient groups were then compared with regard to demographic data, operative characteristics (indications for surgery, surgical approach, prosthesis used, surgeon case volume, and experience), and postoperative outcomes (Oxford Shoulder Score [OSS], revision rate, and morbidity at 6 months and 5 years). Revision rates were further stratified into early (< 2 years) and late revisions (> 2 years), to further assess for any differences between each cohort. In the younger cohort of patients (< 55 years old), those with a minimum of one-year follow-up underwent repeat functional assessment using the OSS and American Shoulder and Elbow Score (ASES). Functional outcome assessments were performed using a combination of hardcopy, electronic, and telephone questionnaires which were administered according to patient preference.

Statistical analysis

Statistical comparisons of the demographic and surgical outcome measures were undertaken using independent t-tests and χ² tests as appropriate. The outcome measures were compared using independent t-tests for the OSS and log-rank tests for the revision rates. An ANOVA was undertaken to determine the independent effects of the diagnosis group and age cohort on the 6-month OSS. A two-tailed p-value <0.05 was taken to indicate statistical significance.

Methodologic considerations

Established in 1998, the NZ Joint Registry became a nationwide operational registry by 1999, and recent data among public hospitals suggest registry compliance exceeding 95%. ¹⁸ The registry collects prospective data for all patients who undergo arthroplasty surgery within our country, including demographic data, operative characteristics, and patient-reported postsurgical outcomes. In patients undergoing shoulder arthroplasty, the self-assessed OSS ¹⁹ is used as the patient-reported outcome measure, with patients completing the questionnaire at 6 months and 5 years after surgery. Our registry uses the method described by Kalairajah et al. ²⁰ to classify the OSS, with a score > 41 being excellent, 34–41 being good, 27–33 being fair, and < 27 being poor.

Results

Between January 2000 and December 2019, 5518 patients who underwent primary RSA were identified from NZ Joint Registry. Of these patients, 75 were under the age of 55 (range: 34–54 years), and 5443 were ≥ 55 at the time of surgery (range: 55–96 years). While there was a higher proportion of female patients in both groups there was no statistically significant difference between each group (percentage female, 54.7% [< 55] vs. 63.0% [≥ 55]; p = 0.136). The mean duration of follow-up was 2.36 years (range: 0.11–13.37 years) in the younger cohort and 3.10 years (range: 0.01–16.22 years) in the older patient cohort.

The indications for surgery differed significantly according to age (Table 1). The three most common indications for RSA in the younger patient cohort (< 55 years) were CTA, inflammatory arthritis, and posttraumatic. Comparatively, CTA, osteoarthritis, and acute proximal humerus fractures were the most common indications in the older patient cohort (≥ 55 years).

Table 1.

Indications for surgery according to age.

Indication	< 55 years	≥ 55 years	p-value
Cuff tear arthropathy	31%	46%	0.010
Inflammatory arthritis	17%	5%	<0.001
Posttraumatic	13%	6%	0.008
Osteoarthritis	12%	36%	<0.001
Avascular necrosis	5%	2%	0.049
Postdislocation	5%	1%	0.008
Acute fractures	7%	10%	0.317

Operative characteristics

A deltopectoral approach was the preferred approach during the study period (rate, 88.0% [< 55 years] vs. 85.5% [≥ 55 years]; p = 0.547). Surgeon experience was similar in both groups, with the majority being performed by high-volume surgeons (rate, 76.0% [< 55 years] vs. 73.0% [≥ 55 years]; p = 0.567), as defined by those completing ≥ 10 cases per year.

The most frequently used prostheses during the study period were the SMR Modular Shoulder (Limacorporate S.p.a., Udine, Italy; used in 2615 cases) and Delta Xtend Reverse Shoulder System (DePuy, Warsaw, IN, USA; used in 1810 cases). Overall rates of specific prosthesis utilization were similar between the younger and older patient cohorts.

Postoperative outcomes

6-month follow-up

The first functional outcome score was performed at 6-months postoperatively using an OSS, with patients in the younger cohort scoring significantly poorer compared to the older cohort (mean, 28.5 [< 55] vs. 35.7 [≥ 55]; p < 0.001).

A multivariate analysis was conducted to assess whether lower functional outcome scores in the younger cohort were due to age alone or the pathology experienced in each group, with higher rates of trauma and inflammatory arthropathy in younger patients. Using this method, mean OSS remained significantly lower in the younger cohort of patients (28.5; 95% confidence interval (CI), 25.2–31.9) than for the older cohort, (34.6; 95% CI 34.0–35.2), p < 0.001

Final follow-up

After a mean follow-up of 54.9 months (range: 13–166 months) for 31 patients in the younger cohort with a minimum of one-year follow-up, the mean ASES and OSS were 71.1 and 23.5, respectively.

Revision rate

There was a total of 21,399 observed component-years during the study period. There was no significant difference in the revision rate between the younger and older patient cohorts (1.56 [< 55 years] vs. 0.74 [≥ 55 years] revisions per 100 component-years; p = 0.332). The overall revision rate during the study period was 0.74 revisions per 100 component-years (Table 2). There was no significant difference in early revision (< 2 years, 1.84 [< 55 years] vs. 1.15 [≥ 55 years] revisions per 100 component-years; p = 0.568) or late revision rates (> 2 years, 1.20 [< 55 years] vs. 0.43 [≥ 55 years] revisions per 100 component-years; p = 0.292) between both cohorts.

Table 2.

Revision rates.

Variable	< 55 years	≥ 55 years	p-value
Number of cases	75	5443		0.332
Observed component years	192.2	21206.7
Number of revisions	3	156
Revision rate (percentage)	4.0	2.9
Revision/100 component-years	1.56	0.74

In the younger patient cohort, three required revision during the study period, two due to glenoid component loosening and one due to humeral component loosening. In the older patient cohort, 156 patients required revision during the study period, with the three most common indications for revision including dislocation (25.6%), glenoid component loosening (20.5%), and infection (21.2%).

Mortality

There was no significant difference between the younger and older patient cohorts in mortality rates at 6 months (2.9% [< 55 years] vs. 0.8% [≥ 55 years]) and 1 year (2.9% [< 55 years] vs. 1.6% [≥ 55 years]) postoperatively (p = 0.717).

Discussion

To our knowledge, this is the largest comparative study investigating the outcomes of primary RSA in patients < 55 years of age. In these patients, RSA was associated with low revision rates, comparable to older patients receiving the same procedure. Although postoperative functional outcomes were lower in the younger cohort of patients, this may reflect the more complex glenohumeral pathology experienced by the younger cohort, with a significantly higher proportion of patients receiving RSA for trauma or inflammatory arthropathy.

Due to concerns over implant survivorship and the lack of revision options, RSA has traditionally been cautioned against in younger patients. Increased implant constraint has led to concerns of early loosening in younger patients with higher functional demands, even when compared to anatomical shoulder prostheses.^21,22 At present, there is a paucity of robust long-term outcome studies for RSA performed in younger patients. Previous studies have focused predominantly on elderly populations and are limited by evolving surgical techniques and implants, as well as heterogeneous patient populations with varying surgical indications. However, as indications for RSA have expanded its use has increased dramatically over the past 20 years, ¹¹ leading many surgeons to consider its use in younger patients with the advanced glenohumeral disease or as a salvage procedure for failed anatomic total shoulder arthroplasty (TSA) or hemiarthroplasty (HA). Cuff et al.^1,23 reported their outcomes on a cohort of patients with a mean age of 78 years. Outcomes based on ASES were 77 and 74 at 5 years and 10 years, respectively. Compared to results achieved in the present study, although surgical indications were dissimilar, the difference in outcomes would be below what is considered the minimal clinically important difference in ASES. ²⁴ Favard et al. ⁹ reported a 10-year implant survival of 89% and a complication rate of 18% in their cohort of 489 patients undergoing RSA with a mean age of 73 years. They also demonstrated significantly lower constant scores in patients with > 9-year follow-up as compared with patients with shorter follow-up periods leading authors to caution against the use of RSA in younger patients. However, the contribution of age-related decline in function was not considered and makes it challenging to extrapolate these results to younger patients. In contrast, recent reviews by Ernstbrunner et al. ¹² and Ek et al. ¹⁴ have demonstrated maintained functional outcome scores beyond 10-year follow-up in separate cohorts of < 60-years old and < 65 years old, respectively.

Previous studies investigating the outcomes of RSA performed in younger patients have yielded mixed results. Sershon et al. ¹⁶ reviewed outcomes following RSA performed in patients < 60 years with a range of surgical indications. After a mean follow-up of 2.8 years, the authors reported a clinical success rate in 75% of patients as defined by ASES > 50. Mean functional improvements were 31.4–5.8 for ASES, improved forward elevation from 56° to 121°, and a mean postoperative constant score of 54.3. In the review by Ernstbrunner et al. ¹² of outcomes following RSA in under 60 years old with irreparable rotator cuff tears, 39% of patients experienced at least one complication, and 26% required at least one reoperation. Mean improvements in a relative constant score of 40% and subjective shoulder score (SSV) of 51% were identified. Despite this relatively high complication rate, these improved functional outcomes scores were reassuringly maintained beyond a 10-year follow-up. Leathers et al. ¹⁷ compared outcomes of patients ≤ 65 years with patients ≥ 70 years. After a mean follow-up of 3.25 years, equivalent complication rates and ASES were reported, however, the younger cohort was able to achieve a greater postoperative range of motion (forward flexion: 133° vs. 117°, abduction: 127° vs. 110°, and external rotation: 40° vs. 27°), and overall high satisfaction rate of 97%. Otto et al. ¹³ reported on their series of 67 patients < 55 years old undergoing RSA. Of these, 32 patients with a mean age of 48.9 years underwent primary RSA for a range of indications. After a mean follow-up of 5 years, there was a complication rate of 18.7% and implant retention rate of 87.5%. Despite this, the authors demonstrated significant improvements in range of motion (forward flexion 64.8°–113.2°; abduction 51.8°–107.8°) and functional outcome scores (ASES 28.1 preoperative to 58.6 postoperatively), with an overall high satisfaction rate of 96% reported. Muh et al. ¹⁵ reviewed their short-term outcomes on 66 patients undergoing RSA with a mean age of 52.2 years. Of these patients, 50% had undergone previous unsuccessful shoulder procedures. Significant improvements in ASES (40.0 preoperatively to 72.4 postoperatively) and range of motion (forward elevation 54.6°–134.0°) were identified. However, with an overall complication rate of 15%, only modest satisfaction rates of 81% were reported. Ek et al. ¹⁴ reviewed 40 RSAs performed for irreparable cuff tears in < 65 years old with a mean follow-up of 7.75 years. Significant improvements in functional outcomes (mean relative constant scores, 34–74%) and active forward elevation (72°–119°) were reported. Overall implant survivorship was 98% at 5 years and 88% at 10 years. Despite an overall complication rate of 37.5%, the authors concluded that functional outcomes can be maintained for up to 10 years. In the review by Black et al. ²⁵ outcomes were compared between a series of patients undergoing primary RSA or revision to RSA following failed HA or TSA. A total of 69 patients with a mean follow-up of 4.6 years were studied. Mean postoperative ASESs were 74 and 69.7 in the primary and revision groups, respectively (p = 0.180). Major complication rates were 18.8% in the revision group and 15.2% in the primary group (p = 0.751). Satisfaction rates were 78.1% in the revision group and 93.9% in the primary group (p = 0.082). Despite consistently high complication rates being reported in younger patients undergoing RSA, positive functional results can be achieved, which reflects the results obtained in the present study.

In a recent systematic review by Vancolen et al., ²⁶ patients aged ≤ 65 years undergoing primary RSA, a revision rate of 18% was reported and an overall satisfaction rate of 89%. The two most common complications identified were dislocation and glenoid loosening, which may reflect the increasing functional demand younger patients place upon their implants. In the present study, a comparatively lower revision rate of 5.6% was reported in patients ≤ 55 years. Shoulder surgeons also tend to permit more restrictive activities for patients undergoing RSA when compared to TSA or HA due to concerns of early implant failure. ²⁷ However, despite this limitation, the majority of patients achieve medium to high functional demands following RSA, with rates comparable to patients undergoing TSA or HA. ²⁸

There are several limitations to the present study. Firstly, this is a retrospective and non-randomized study. Secondly, while functional outcome scores were assessed, no preoperative assessment was performed as a baseline, which limits the ability to interpret any procedure-related improvement in functional status, especially with the younger cohort frequently demonstrating more complex glenohumeral pathology. Thirdly, functional outcome scores used were subjective, and objective clinical and radiographic data was not available due to the retrospective nature of this study. Fourthly, recent evidence suggests patient-reported outcome measures do not necessarily correlate with patient satisfaction following RSA. ²⁹ Inclusion of a patient satisfaction measure may have yielded valuable results in addition to the scoring systems utilized. Furthermore, not all patients included were available for final follow-up which may have created a bias in the outcome. Lastly, while the surgical indication varied between the younger and older cohorts, this is likely to reflect the different pathologies experienced in both age groups.

Conclusion

Our early results suggest that younger patients undergoing RSA demonstrate high implant retention rates, comparable to older patients. Longer-term patient-reported outcomes in younger patients are required in order to guide appropriate patient selection for RSA.