Anatomic or reverse shoulder arthroplasty for cuff intact glenohumeral osteoarthritis

Abstract

Glenohumeral osteoarthritis in the presence of an intact rotator cuff is a debilitating condition that often leads to surgical intervention in the form of joint arthroplasty. Historically, anatomic total shoulder replacement (ATSR) had been considered the treatment of choice. Over the past decade, however, there has been a shift in treatment patterns to the use of reverse shoulder replacement to treat this condition due to concerns of subsequent rotator cuff failure and glenoid component loosening with ATSR. With continued innovation of implants and surgical techniques, it is unclear which procedure provides the best outcomes for patients. The Reverse or Anatomical replacement for Painful Shoulder Osteoarthritis, Differences between Interventions trial aims to definitively answer this question and evaluate the cost effectiveness of both procedures.

Introduction

The incidence of shoulder arthroplasty continues to increase worldwide.^1–3 Anatomic total shoulder arthroplasty has long been the treatment of choice for advanced glenohumeral osteoarthritis (GHOA). Its use requires an intact rotator cuff to preserve glenohumeral kinematics and avoid early loosening of the glenoid component. Reverse shoulder arthroplasty was subsequently developed by Paul Grammont as a treatment for patients with rotator cuff deficiency and has been in use since 1991.^4,5 The indications for reverse shoulder replacement (RTSR) then expanded to its use in proximal humeral fractures, non-union or malunion, chronic shoulder dislocations, inflammatory arthropathy, revision arthroplasty and tumour.^4,6–15 Over the last decade, its use in primary GHOA has increased. In the United States, RTSR is used to treat approximately 33% of surgically managed primary GHOA. ¹ This increasing role for RTSR in GHOA is potentially due to the historical rates of subsequent anatomic total shoulder replacement (ATSR) glenoid component as well as rotator cuff failure. To date, there is no clear consensus for the optimal arthroplasty option in patients with GHOA with an intact rotator cuff.

Anatomic total shoulder arthroplasty

Anatomic total shoulder arthroplasty has been considered the standard surgical treatment for GHOA.^16,17 It has undergone some substantial changes and improvements in surgical technique and implant design since its first reported use in 1893. ¹⁸ In long-term studies, ATSR has been shown to reliably improve pain and functional outcomes in patients with GHOA.^19–21 Survival rate of ATSR for various indications using revision surgery as the end point, at 10 years has been reported to be approximately 93%.^19,20 A Nordic-based registry study found the 10 year survivorship of ATSR, when completed for GHOA, is also approximately 93%. ²² Another multicenter study that included 601 ATSR for GHOA found the 10 year survivorship to be approximately 70%. ²³ This notable difference was found to be attributable to the use of metal-backed glenoid components. ²⁴ Haines et al. demonstrated 90% survival when ATSR was indicated for GHOA while osteonecrosis and posttraumatic arthritis had survival rates of 75% and 33%, respectively. ²⁵ These studies highlight examples of how prior long-term studies might have been confounded by implant design, inferior materials, surgical technique, and patient selection considerations.

Studies have investigated the relationship between patient age and ATSR survivorship. There has been long standing caution to implanting ATSR in young patients due to worry of overactivity resulting in glenoid component failure. Saltzman et al. evaluated a cohort of patients who had undergone ATSR between 1990 and 2008 and found a discrepancy for the surgical indication based on age. They noted that only 21% of those undergoing shoulder arthroplasty under the age of 50 years had ATSR for GHOA compared to 66% of patients ≥ 50. ²⁶ This suggested that most young patients undergoing ATSR have more complex pathologies such as capsulorrhaphy arthropathy, inflammatory or posttraumatic arthritis as well as more severe glenoid deformity that may influence the outcomes of these procedures. ²⁶ Similarly, within the Australian Registry, patients with GHOA treated with stemmed ATSR aged ≥75 had the lowest 10-year cumulative revision risk (8%) while those <55 years old had the highest risk (12.9%). ²⁷ The increased demand and stress on the glenoid component and rotator cuff have been suggested as reasons for these findings.

Despite the lower revision risk of ATSR in elderly patients, ²⁸ concern regarding the risk of subsequent rotator cuff failure with increasing age, has led some surgeons to be cautious in the use of ATSR in older patients. ²⁹ How ‘secondary rotator cuff failure’ post ATSR is defined however can influence its reported rates. Early studies reported superior migration of the humeral head to be a marker for rotator cuff failure. Young et al. reviewed 518 patients with ATSR and reported a 16.8% (87 patients) rate of secondary rotator cuff failure based on superior migration of the humeral head on post-operative radiographs. ³⁰ Despite that alarming rate, only one of these patients required revision surgery for rotator cuff failure. ³⁰ Jensen et al. evaluated 377 patients age ≥70 who underwent ATSR for GHOA and found 98.9% survivorship at five years. ³¹ They found five patients who had secondary rotator cuff failure that was confirmed through imaging and only two of these patients required a reoperation. ³¹ They noted that despite the presence of radiographic superior humeral head migration on postoperative radiographs, this did not correlate with decreased ASES and VAS scores. ³¹ The relative lower risk of failure in this age group may of course be related to their lower level of activity.

Within the United Kingdom, National Joint Registry (NJR) and Australian Registry, rotator cuff insufficiency is the leading cause for revision of ATSR.^27,32 Other causes of failure included glenoid component loosening and infection. Recent studies have demonstrated more favourable outcomes for ATSR in GHOA and this is likely due to improved surgical techniques and implants that allow for correction of glenoid deformities that were not corrected in prior studies. In a study evaluating the outcomes of shoulder arthroplasty over time, based on year of operation, ASES scores were found to increase each year despite controlling for patient sex, American Society of Anaesthesiologists classification, indication, rotator cuff condition and Walch Classification. ³³ This likely demonstrates that improvement in patient selection over the years has also resulted in improved outcomes for shoulder arthroplasty.

In summary, there is no doubt that the clinical outcome after ATSR can provide excellent pain relief, movement, and function. Survivorship figures of 90% and above over a 10-year period are reported. Generally, this applies to cemented polyethylene glenoid components and cobalt chrome humeral heads fitted to a stemmed humeral component. This figure does diminish, particularly when metal-backed glenoid components have been used and in younger patients. Regarding the indications for revision, these include failure of the rotator cuff and glenoid component loosening.

Reverse shoulder arthroplasty for GHOA

The use of RTSR has been successful in improving pain and functional outcome scores in patients with rotator cuff tear arthropathy. A recent systematic review evaluated 5824 shoulders that had undergone primary reverse shoulder arthroplasty for various indications with a minimum of two year follow up. They demonstrated a mean clinically significant improvement of 42 in ASES score and 37 in the Constant score. ³⁴ Mizuno et al. were the first to describe the outcomes of RTSR for GHOA in patients over 70 years of age whilst preserving all of the rotator cuff in 2013. ³⁵ This revelation for the use of RTSR stemmed from the failure rate of previously designed ATSR components in the setting of a biconcave glenoid. ³⁶ They reviewed 27 patients who had underwent RTSR from 1998 to 2009 for GHOA with a biconcave glenoid. At mean follow up of 34 months, there was a significant improvement in the Constant score, forward flexion, external (ER) and internal (IR) rotation. ³⁵ Complications included scapular notching (37%), glenoid component loosening (3.7%) and neurological issues (11%). Since then, there has been an increase in the use of RTSR for primary GHOA due to considerations such as severe glenoid bone loss, age and concern regarding possible future rotator cuff tears.

Heifner et al. completed a systematic review that included 460 patients who underwent RTSR for GHOA and demonstrated a mean satisfaction rate of 92% and significant improvement in the Constant score. ³⁷ The overall major complication rate was 3.9% which included deep infection (7), periprosthetic fracture (2), instability (1) and one case of baseplate failure. The overall revision rate was 3.1%. ³⁷ A recently published study evaluated 230 patients who underwent RTSR for GHOA with an intact rotator cuff and had a mean follow up of 33.4 months. ³⁸ In 97.4% the change in ASES score exceeded the minimal clinically important difference (MCID = 10.3) ³⁹ and 90.1% exceeded the substantial clinical benefit (SCB = 25.9) ⁴⁰ for the ASES score. ³⁸ After multivariate analysis, preoperative ASES score or Walch B3 classification were predictors of excellent postoperative ASES scores. Meanwhile, postoperative complications or preoperative opioid use were predictors of poor postoperative ASES scores. ³⁸

In a longer-term (minimum five years) follow up study including seven centres, 49 shoulders had underwent RTSR for cuff intact GHOA with Walch types B1, B2, B3 or C glenoid. ⁴¹ Sixteen cases required glenoid bone grafting and eight of those resulted in partial lysis at the inferior aspect of graft at latest follow up. Patients had an average increase of 38 points with respect to the Constant score. Major complications included two infection cases and one transient radial nerve palsy. There was a 43% rate of scapular notching and survivorship was reported to be 98% at five years. ⁴¹

With regard to survivorship, we are aware of only one publication with 10-year outcomes. Favard et al. reported a figure of 89% at 10 years with removal of implant being the end point. ⁴² However, if the Constant-Murley score of less than 30 was used as the end point, this survivorship was reduced to 72% with a marked reduction at eight years.

The NRJ reports instability/dislocation to be the leading cause for revision of RTSR. ³² Through a multicenter analysis of 6621 patients who underwent RTSR for all indications, the overall dislocation rate was 2.1%. When surgical indications were taken into account, patients with GHOA and intact rotator cuff had the lowest rate of dislocation at 0.8%. ⁴³ A similar multicenter group evaluated 4764 patients and found the overall rate or acromial stress fracture to be 4.1%. ⁴⁴ Those with the primary diagnosis of GHOA had the lowest incidence of 2.1% compared to those with cuff tear arthropathy or massive rotator cuff tears (5.2%). ⁴⁴

To summarise therefore the clinical outcome of RTSR is generally very good in that it provides good pain relief, movement and function for the indications described above. Whilst long term survival data remains limited from what is available it does seem to be comparable to ATSR (89% at 10 years). Complication rates are also similar although the complications themselves are different. Many surgeons however remain concerned about recommending RTSR in patients under 65 principally due to the lack of appropriate revision options.

Anatomic versus reverse for GHOA

Polisetty et al. completed a value-based matched analysis to compare the results of ATSR and RTSR (101 patients in each group) alongside multiple patient reported outcome measures in the setting of GHOA. ⁴⁵ At mean follow up of 39 months, they found the implant cost to be higher for the RTSR but operating room, anaesthetic and cements costs to be higher for ATSR. ⁴⁵ While there were no revision cases in the RTSR cohort, there was a 2.4% revision rate within the ATSR cohort. Following inclusion of revision costs and outcome scores, ATSR and RTSR demonstrated similar outcomes and value when used to managed GHOA. ⁴⁶ Nonetheless, this study had a smaller cohort of patients within the RTSR group and shorter term follow up. Therefore, future studies are needed to assess the long-term value analysis of these treatment options.

A propensity score-matched analysis was completed on patients who had undergone ATSR or RTSR for cuff intact GHOA with a mean follow up of 30 months by a single surgeon. ⁴⁷ Matching was based on age, sex, body mass index, preoperative ASES score, preoperative forward elevation and Walch classification. There were no differences found between the ATSR and RTSR groups in terms of postoperative outcome measure scores and overall complication rates. The ATSR group had significantly better postoperative forward elevation (7°), ER (6°) and IR but these were of questionable clinical significance. A single ATSR case was revised due to a secondary rotator cuff tear and no revisions were completed within the RTSR group. ⁴⁷ Of note, there was a variety of ATSR implants included while a single RTSR design was used. Furthermore, no augmented anatomic glenoid components were used.

In a longer-term study with a minimum of seven years follow up, Cuff et al. evaluated 242 cases of cuff intact GHOA with Walch B or C glenoid (eccentric wear) that underwent either ATSR with high side ream or RTSR. ⁴⁸ At the two-year follow up, the ATSR group had significantly better postoperative ASES scores, forward elevation, ER and IR. Patient satisfaction at this time point was 95% for the ATSR group and 93% for the RTSR group. Interestingly, at the final follow up, patient satisfaction was 82% and 95% for the ATSR and RTSR cohorts, respectively. ⁴⁸ While the ATSR cohort preserved improved IR and ER, the RTSR cohort has similar forward elevation and a better ASES score. ⁴⁸ The overall revision rate for the ATSR cohort was 8% compared to 2% of the RTSR group. ⁴⁸ This study suggested greater early satisfaction and outcomes with ATSR for eccentric GHOA but these outcomes were not as durable in the long-term compared to those who underwent RTSR. An important limitation that was mentioned within this study was the lack of augmented anatomic glenoid components used. Therefore, a minimum of 90% backside seating was targeted rather than the conventional standard today of 100% with the aid of augmented components.

Another recent matched-cohort analysis (101 patients in each cohort) where again no augmented anatomic components were used demonstrated a slightly lower VAS pain score and worse IR in the RTSR group compared to the ATSR group. ⁴⁵ At mean follow up of 39 months, there were no differences in patient reported outcome scores, complications or other motions measured. Although the overall complications rates were not statistically different, the complication profile did vary. There were two cases of baseplate loosening in the RTSR cohort and three cases of grossly loosened glenoid components in the ATSR cohort. There were four cases of acromial stress fractures in the RTSR group and one case of subscapularis failure in the ATSR group. ⁴⁵ The NJR reports a 10-year cumulative risk of revision for stemmed ATSR and stemmed RTSR to be 6.07 (95% CI [5.30–6.93]) and 4.12% (95% CI [3.60–4.72]), respectively. ³² While this difference is of concern, the extent to which indications for surgery and implant specific complications may confound these results is unknown.

Published data raises concerns regarding young age and revision for both ATSR and RTSR. In a retrospective study of 518 ATSR and 732 RTSR procedures, ATSR was found to be associated with a 3.4-fold increased risk of revision in patients younger than 65 years. ⁴⁹ Patients younger than 60 years who underwent RTSR were associated with a 4.8-fold increased risk of revision. ⁴⁹ Ardebol et al. retrospectively reviewed 104 patients ≥75 years with cuff intact GHOA who were treated with either ATSR (n = 67) or RTSR (n = 37). ⁵⁰ They demonstrated similar clinical outcome score improvement, complication rate, revision rate and satisfaction in both groups. Meanwhile, ATSR was found to have a significantly greater improvement in postoperative ER. Importantly, the two groups did vary preoperatively as RTSR was more commonly performed for Walch type B2 or B3 glenoids. ⁵⁰

Overall, there may be bias in the current literature when assessing the use of ATSR and RTSR in the setting of cuff intact GHOA. There have been significant innovations in implant design and surgical technique over the course of past 20 years. Earlier studies did not use augmented glenoid components and therefore increased high side reaming or retaining retroversion in eccentrically worn glenoids may have contributed to higher complication rates. As there has been a shift to the use of RTSR for these specific glenoid morphologies, this also creates a selection bias when evaluating retrospective studies comparing ATSR and RTSR.^51,52 Furthermore, there is significant variability in outcomes and complication profiles associated with these two arthroplasty techniques that may not present at similar time points in follow up.

To combat this deficiency in our knowledge, a multi-centre, pragmatic, parallel group, superiority randomised controlled trial is in progress within the National Health Service in the United Kingdom. The Reverse or Anatomical replacement for Painful Shoulder Osteoarthritis, Differences between Interventions trial aims to determine whether RTSR is superior to ATSR for the treatment of painful GHOA with an intact rotator cuff and suitable bone stock in patients aged ≥60 years as measured by patient-reported pain and function using the Shoulder Pain and Disability Index at 24 months. Secondary objectives will evaluate range of movement, shoulder strength, complications, re-operations and mortality. Furthermore, the cost effectiveness of these interventions will be compared from the NHS perspective. A parallel trial is also ongoing in Australia.