Clinical Outcomes of Reverse Total Shoulder Arthroplasty
Clinical Outcomes of Reverse Total Shoulder Arthroplasty
Ryan W. Simovitch, MD
Richard J. Friedman, MD
Reverse total shoulder arthroplasty (RTSA) designs came into clinical use in the early 1970s. However, the early designs were fraught with many complications, and most failed within 2 to 3 years, as the glenoid component often loosened due to the high forces present, especially in cases where the rotator cuff was absent. As a result of this, combined with the early success of Neer’s anatomic total shoulder arthroplasty (ATSA), RTSA fell out of favor.1
While ATSA provided excellent pain relief and restoration of function in patients with an intact rotator cuff, those with a cuff-deficient shoulder did not do as well. Patients with glenohumeral arthritis and massive rotator tearing developed instability in a superior or anterior-superior direction because the absent rotator cuff could not provide a fixed fulcrum for the humeral head against the glenoid. Patients with glenohumeral arthritis and massive rotator cuff tears, referred to as cuff tear arthropathy (CTA), exhibited shoulder weakness, incongruous joints surfaces, instability, and bone loss. The promise of RTSA was that it could restore glenohumeral joint stability, provide smooth articulating surfaces, replace bone loss, and optimize the remaining rotator cuff muscles and deltoid to improve strength and function.
It was not until the middle 1980s when Paul Grammont in France developed and articulated his principles of RTSA that clinical interest in the prosthesis emerged.2 Grammont proposed moving the center of rotation (COR) medially and distally, thereby protecting the glenoid component from early failure and providing stability for the prosthetic construct. The glenoid had a convex weight-bearing portion matching up with a concave humeral cup, and the center of the glenoid sphere (glenosphere) was at or within the glenoid neck. His original prosthesis, first implanted in 1985, underwent further development and design changes in 1991 and 1994, and each redesign showed clinical promise and avoided the failures of earlier designs.3 This was the catalyst that led to the development of the modern RTSA.
Initial reports on the clinical outcomes of RTSA were from Europe, given the fact that the current design originated there and was not released for use in the United States until 2003. Published studies showed significant clinical improvements in a very difficult patient population, but the complication rates were significantly higher than those reported for ATSA. These included hematoma formation,4 infection,4,5 scapular notching,6,7 instability,8,9 acromial insufficiency,4,10 and glenoid component failures.11,12 However, prosthetic designs have evolved, surgical techniques have improved, and surgical experience has increased, and as a result, the complication rate for RTSA is now similar to and, in fact, may be less than ATSA.13
The original indication for RTSA when it was first approved by the United States Food and Drug Administration in December 2003 was CTA. The indications have expanded significantly since that time, as the clinical outcomes and survivorship have been favorable and remained so over time. Current indications for RTSA include CTA, massive irreparable cuff tear with osteoarthritis, irreparable massive rotator cuff tear, primary osteoarthritis in older patients with an intact rotator cuff, rheumatoid arthritis and other inflammatory arthropathies, proximal humerus fractures, sequelae of proximal humerus fractures such as nonunions and malunions, revision total shoulder arthroplasty, chronic dislocations, periprosthetic fractures, tumors, infection sequelae, severe glenoid deformity with or without an intact rotator cuff, and severe posterior subluxation with posterior glenoid erosion.
Given the expanding indications for RTSA, one could argue that there is one indication for ATSA and all other patients would be candidates for an RTSA. The indication for an ATSA is a patient with osteoarthritis who is relatively young, has no severe systemic illnesses, has no significant glenoid deformity, has an intact functional rotator cuff, and has not had any previous open shoulder surgery that violated the subscapularis. This makes for a very homogeneous group with consistent and reliable outcomes. RTSA is done for many different indications, and therefore the clinical outcome and survivorship vary based upon the underlying pathology. One must be careful when comparing the results of ATSA and RTSA to make sure the comparison is between similar groups with similar indications.
FIGURE 27.1 Shoulder arthroplasty usage from 2009 to 2019 in the United States, demonstrating the increase in reverse total shoulder arthroplasty (RTSA) and decrease in hemiarthroplasty over that time. Note that 2014 was the year that RTSA outnumbered anatomic total shoulder arthroplasty (ATSA).
With the expanding indications for RTSA, the clinical volume has grown exponentially over the past 15 years, far surpassing the growth of ATSA.14 There has been a 17-fold increase in the use of RTSA over the past 15 years, compared to a 3-fold increase in ATSA (FIGURE 27.1). In 2014, the use of RTSA surpassed that of ATSA in the United States for the first time and has continued to do so.15 It is projected that RTSA utilization will increase 122% over the next 5 years.
In this chapter, we will explore the various indications for RTSA, the clinical outcomes achieved, and the factors that can influence outcomes and survivorship.
INFLUENCE OF BIOMECHANICS ON OUTCOMES
Initial early success with the use of a traditional Grammont-style RTSA spawned the development of multiple reverse shoulder implant designs and implantation techniques that are utilized today. Variations in implant design and implantation techniques have been shown to impact clinical and radiographic results.
Implant variability includes both glenoid- and humeral-sided considerations. Different RTSA designs utilize glenospheres of varying spherical radii and thickness, which together influence the COR.16 Altering the COR can influence rotator cuff and soft-tissue tension, thus impacting range of motion (ROM), strength, and stability. Furthermore, the COR influences the deltoid moment arm, whereby a medialized COR results in a greater moment arm in turn providing the deltoid with greater efficiency in elevating the arm during abduction. The converse is true in the case of a glenosphere with a lateralized COR.16,17,18,19,20
Humeral-sided design variability includes neck-shaft angle, inlay versus onlay design, stem shape, and humeral tray offset, all of which result in a specific amount of humeral-sided lateralization. The degree of humeral-sided lateralization influences soft-tissue tensioning and deltoid wrapping, which can influence both strength and stability. Other design characteristics that can influence clinical and radiographic outcomes after RTSA are polyethylene depth and constraint as well as stem length. A combination of glenoid- and humeral-sided design attributes has been shown to influence the rate of scapula notching, impingement-free arc of motion, active ROM, acromial fracture rate, and stability.21,22,23,24,25,26,27,28,29,30,31,32,33
Routman et al16 first described a classification system that categorized various RTSA prosthesis designs based on their glenoid- and humeral-sided design attributes. Specifically, this classification stratified prostheses into those with a glenosphere with a COR of 5 mm or less lateral to the glenoid face labeled as a medialized glenoid (MG) or with a COR greater than 5 mm lateral to the glenoid labeled as a lateralized glenoid (LG). Similarly, they classified humeral prostheses according to the degree of offset, which was defined based on the horizontal distance between the humeral stem axis and the center of the humeral liner. It was stipulated that a humeral implant with an offset of 15 mm or less was a medialized humerus (MH) and a humeral implant with an offset greater than 15 mm was a lateralized humerus (LH). Different implant systems combine a specific glenoid design with a specific humerus design, thus yielding three general design combinations: MG/MH, MG/LH, and LG/MH. Each design has various advantages and disadvantages specific to its biomechanical design.
Werthel et al34 attempted to stratify RTSA prosthesis designs by considering combined global implant lateralization. The measurement of humeral lateralized offset was similar to the technique employed by Routman et al16; however, the contribution from glenoid lateralization included the sum of the offset contributed by the COR offset in addition to the perceived radius of the glenosphere. The quantitative global implant lateralization was then fit into five categories relative to the Delta III prosthesis (Grammont design): medialized reverse shoulder arthropathy (RSA), minimally lateralized RSA, lateralized RSA, highly lateralized RTSA, and very highly lateralized RSA.
The classification systems of both Routman et al16 and Werthel et al34 help to lay the framework for understanding the biomechanical differences in various RTSA implant designs. Understanding the various biomechanical attributes of each implant system can help predict their strengths and weaknesses and can also help to avoid potential pitfalls. While specific implants have defined biomechanical properties, the technique of implantation and hence surgeon choices can alter the biomechanical principles and expected outcomes. For example, it has been demonstrated that the position of glenosphere implantation matters greatly. Based on previous studies, a position biased inferiorly6,35 and posteriorly36 is associated with reduced rates of scapula notching. Thus, the increased rates of scapular notching in Grammont-style prostheses (MG/MH) can be mitigated, but not eliminated, by lowering the position of the glenoid baseplate6 or augmenting the glenoid bone with autograft such as in the BIO-RSA.37 Other implants reduce the rate of scapula notching through the implant design attributes alone or in combination such as the MG/LH and LG/MH designs. Another example of a technique variation that impacts outcome is the decision to repair the subscapularis after RTSA. Repair of this tendon may result in certain limitations and certain benefits depending on the implant utilized.38 This confirms the connected interplay between implant choice and surgical technique. This complex interplay should be appreciated when reviewing the literature and the results of RTSA. Implant designs are vividly different, and thus comparing results among different cohorts in the literature can be deceiving or, at best, confusing.
When RTSA was introduced in the United States, many surgeons recommended avoiding the procedure in patients younger than 70 years because survivorship was unknown. Over time, however, it has become clear that the survivorship is similar to that reported following ATSA. Given the fact that the glenoid component in RTSA relies on bone ingrowth and not methyl methacrylate as with an ATSA for long-term fixation, the survivorship of RTSA could surpass that of an ATSA as we acquire longer term follow-up.
Earlier studies utilizing a first-generation Grammont-style prosthesis reported survivorship of 89% at 10 years, but the clinical outcomes deteriorated over that time period.39 This was likely related to the high incidence of scapular notching in these patients, which is much less frequent now with more contemporary designs and compensatory surgical techniques. While previous studies suggested that scapular notching did not affect the clinical outcome, more recent articles have shown that the clinical results deteriorate over time if scapular notching is present when compared to those without scapular notching.22 Another early study looking at the 10-year results of RTSA performed in patients younger than 65 years for massive irreparable rotator cuff tears showed that the significant clinical improvements were maintained out to 10 years, but there was a high complication rate.40 A large multicenter study utilizing the same first-generation RTSA prosthesis with a minimum follow-up of 10 years showed that survivorship depended on the underlying diagnosis.41 The overall survivorship was 92% at a minimum of 10 years. RTSAs performed for osteoarthritis with or without rotator cuff disease had a 97% survivorship, while those performed for revision of a previously placed prosthesis, such as conversion of a hemiarthroplasty (HA) to an RTSA or revision of an ATSA to an RTSA, had a survivorship of 88%.
Time and experience have demonstrated that this procedure has excellent outcomes and survivorship even in younger patients using contemporary prosthetic designs. Otto et al reported on 67 patients younger than 55 years undergoing RTSA for a variety of diagnoses and reported 91% survivorship at a mean follow-up of 5 years, with the range being from 2 to 12 years. Patients undergoing a primary RTSA had significantly better outcomes compared to those who had a failed previous arthroplasty revised to RTSA.42
Two early studies looking at the long-term survivorship of RTSA for osteoarthritis and/or rotator cuff disease both found survival rates of 95% at 8 and 12 years, respectively.5,9 Bacle et al found an implant survival rate of 93% at a minimum of 10 years for multiple indications using what would be considered second-generation implant designs.43 Another more recent study looking at the 10-year minimum survivorship of a contemporary design performed only for the treatment of rotator cuff deficiency showed 91% survivorship.24
These survivorship results, for both first-generation designs and more contemporary implants, are excellent and compare favorably with those for ATSA. This is true for both primary RTSA done for arthritis and/or rotator cuff disease and for multiple other indications, including revision arthroplasty. Torchia et al reviewed the results of 113 first-generation Neer ATSAs performed for a variety of indications, such as osteoarthritis, rheumatoid arthritis, and posttraumatic arthritis, and reported 93% survivorship at 10 years and 87% survivorship at 15 years.44 Another study of survivorship of a Neer total shoulder arthroplasty in patients with both osteoarthritis and rheumatoid arthritis showed similar results, with survivorship of 93% at 10 years, 88% at 15 years, and 85% at 20 years.45 This is the standard to which RTSA needs to be compared and, at this time, appears to be similar.
National registry databases from countries where all arthroplasties are entered into a national registry also provide valuable information on survivorship. The 2019 Annual Report from the Australian Orthopaedic Association National Joint Replacement Registry documents a survivorship of approximately 91% for ATSA and 94% for RTSA at 10 years of follow-up for all patients, and this difference is statistically significant (P < 0.001).46 Survivorship data for ATSA from the 2019 Swedish Shoulder and Elbow Registry Annual Report showed that for the 1999-2003 cohort, the survivorship was 92%, and for the 2003-2008 cohort, it was 96% at 15 years.47 For RTSA, the 15-year survivorship was 89% for the 1999-2003 cohort and 93% for the 2003-2008 cohort, showing, once again, a favorable comparison.
A variety of shoulder outcome metrics exist within the literature and are utilized to evaluate outcomes following RTSA. Some of these metrics rely only on patient reporting. These are called patient-reported outcome measures (PROMs). Other outcome metrics involve patient reporting and objective measures quantified by an examiner. These may have a PROM component but are not purely patient reported. For brevity, only the most common measures will be detailed.
Scoring metrics can generally be divided into those that are purely patient reported and hence PROMs and those that have an examiner component. The Simple Shoulder Test (SST) is a PROM composed of 12 questions that assess pain and function of the shoulder. The maximum score is 12 and pain is weighted 20% while function is weighted 80%.48 The American Shoulder and Elbow Surgeons (ASES) score is a PROM that combines questions related to function and pain, each weighted 50%, to generate a maximum score of 100.49 The Shoulder Pain and Disability Index (SPADI) is a PROM that evaluates pain (38%) and disability (62%). The maximum score for SPADI is 130.48,50 Unlike most other metrics, a low score denotes a better clinical state. The Oxford Shoulder Score (OSS) is also a PROM that utilizes 12 multiple choice questions to evaluate pain (33%) and function (67%). The maximum score is 60. In 2009, a change was made to the scoring methodology of the OSS. Prior to that date, a higher score denoted a worse disease state; following that date, a higher score denoted a reduced disease state. The rank of the questions was inverted though the content remained the same.51 A visual analog scale (VAS) for pain is a PROM that asks a patient to rate their pain numerically between 0 and 10.52 Similarly, a Single Assessment Numeric Evaluation (SANE) metric is a PROM that requires a patient to evaluate the shoulder as a percent of normal between 0 and 100. A score of 100 is considered a “normal” shoulder.53 Similar to the SANE, the subjective shoulder value (SSV) is a PROM that asks a patient to rate the shoulder as a percent of normal between 0% and 100%.
The following scoring metrics combine patient-reported measures and examiner contributions, typically with an assessment of ROM and/or strength. The University of California at Los Angeles (UCLA) Shoulder metric assimilates a score based on a combination of ROM, strength, function, satisfaction, and pain. The maximum score is 35.54,55 The Constant metric assesses strength (25%), pain (15%), sleeping comfort, and activity level (20%) as well as ROM (40%).56
There is a group of data collection tools called health-related quality of life (HRQoL) surveys. These are increasingly utilized in the shoulder arthroplasty literature allowing the comparison of results of different cohorts as well as to follow the effect of the intervention (shoulder arthroplasty) on a patient’s general quality of health. These surveys include but are not limited to Patient-Reported Outcomes Measurement Information System (PROMIS) 10, 36-Item Short Form (SF-36), 12-Item Short Form (SF-12), and Verterans Rand 12 (VR-12). The PROMIS-10 allows the measurement of symptoms, function, and HRQoL for chronic diseases and conditions such as arthritis using 10 questions. It can be separated into physical (global physical health) and mental (global mental health) components. The SF-36 relies on 36 questions to evaluate physical function, role limitations due to physical health, role limitations due to emotional problems, energy and fatigue, emotional well-being, social functioning, general health, and pain. The SF-12 and SF-8 are shorter but similar surveys. The VR-12 examines health domains similar to the SF-36 but is summarized into a Physical Component Score (PCS) and a Mental Component Score.
Radiographic evaluation of RTSA centers on the evaluation of inferior scapula notching. The degree of inferior notching is measured according to the method of Sirveaux and Nerot. grade 0 notching denotes the absence of a notch, grade 1 denotes involvement of the inferior pillar, grade 2 describes a defect that reaches the inferior screw, grade 3 describes a defect that rises above the inferior screw, and grade 4 denotes a notch extending underneath the baseplate.
MCID AND SCB
Historically, the clinical results of outcome studies were reported by showing differences in means and recording statistical significance, generally represented as a P-value. Statistical significance denotes that a difference is not due to random chance. However, statistical significance does not equate to meaningful clinical change. Otherwise stated, statistical significance does not mean a change from preoperative to postoperative would be considered meaningful by a patient. Toward that end, the concepts of the minimal clinically important difference (MCID) and substantial clinical benefit (SCB) are becoming more popular. There are a variety of methods that can be utilized to calculate MCID and SCB, although a comprehensive discussion of these techniques is outside the scope of this chapter.57 MCID has been defined as “the smallest difference in score in the domain of interest which patients perceive as beneficial and which would mandate, in the absence of troublesome side-effects and excessive cost, a change in the patient’s management.”58 SCB has been defined as the smallest difference in score in the domain of interest that exceeds the minimum threshold of improvement—a value that surgeons would choose to aim for.59 The difference between statistical significance and reaching MCID is illustrated in a study by Torrens et al,60 which studied 60 patients undergoing RTSA for rotator cuff insufficiency and found that improvement in the lateral rotation and strength components of the Constant score reached statistical significance but failed to reach the MCID, indicating that despite reaching a level of statistical significance, these values were not clinically relevant to patients.
Values for MCID and SCB of common clinical metrics have previously been reported for RTSA (TABLE 27.1). MCID and SCB values after RTSA have been demonstrated to be lower than the values after ATSA. Furthermore, these values appear to be impacted by gender, length of follow-up, age at time of surgery, and preoperative function.57,59,61 These reported MCID and SCB values can be used as an additional lens by which to evaluate clinical improvement after RTSA as opposed to relying solely on statistical significance that can be influenced by a myriad of factors including sample size.
ASES, American Shoulder and Elbow Surgeons; ER, external rotation; MCID, minimal clinically important difference; SCB, substantial clinical benefit; SPADI, Shoulder Pain and Disability Index; SST, Simple Shoulder Test; UCLA, University of California at Los Angeles; VAS, visual analog scale.
Rotator CTA can be defined as severe arthritis of the glenohumeral joint with a high-riding humeral head resulting in acetabularization of the undersurface of the acromion and femoralization of the humeral head, glenoid bone loss, massive tearing of the rotator cuff tendons, and muscle changes including atrophy and fatty infiltration.62 Prior to RTSA, standard surgical treatment included an attempt at repair of part of the rotator cuff along with a standard head size HA or a large head size HA. While patients experienced pain relief and improvement in function, these did not approximate the improvement reported after ATSA and for glenohumeral arthritis with an intact rotator cuff.63,64
With the introduction of RTSA in the United States in 2003, CTA was the sole indication for this procedure. Since then, as the indications have expanded, it has revolutionized the ability to manage complex problems about the shoulder, including CTA. Studies have shown that RTSA far outperforms HA for CTA.65,66,67,68 ROM, pain relief, and functional outcomes documented by PROM were all significantly better in patients who underwent RTSA compared to those who underwent an HA. Other studies have documented the pain relief, superior functional outcomes, and improved ROM in patients with CTA treated with RTSA, which has become the procedure of choice in these patients.66,68,69,70,71,72,73 Studies have shown that these outcomes can be expected in both younger and older patients.74
Inflammatory arthropathies, the most common being rheumatoid arthritis, frequently affect the periarticular soft tissues. For the shoulder, this represents a high incidence of rotator cuff disease including attenuation, dysfunction, or full-thickness tearing. While the incidence of rotator cuff involvement has decreased with the use of biologics over the past 20 years, it is still an issue that has to be addressed in the face of disabling arthritic changes about the glenohumeral joint. ATSA used to be the procedure of choice, but patients would often develop cuff failure after arthroplasty, leading to poor function, increased pain, and early failure. With the availability of RTSA, the condition or subsequent viability of the rotator cuff is no longer an issue that will affect the patient’s outcome.
Studies have shown similar results for RTSA performed in the setting of inflammatory arthropathy compared to RTSA performed for other diagnoses. Patients experience predictable, consistent, and significant pain relief and improvements in ROM and function.72,75,76,77 Complication rates are similar to RTSA performed for other indications, and recent studies have shown complication rates of RTSA to be similar to those reported for ATSA.38,78
Osteoarthritis With Rotator Cuff Tear
RTSA has also been utilized in patients with osteoarthritis and rotator cuff tears of varying sizes without CTA. Significant improvements in overhead function and improvements in pain, ROM, and PROM have been documented with follow-up of 20 years without significant deterioration over time.24,79
The use of RTSA in patients with an intact rotator cuff and severe glenoid bone loss has been reported and is growing in frequency. The surgeon has more reconstruction options in patients with a B2-, B3-, or C-type glenoid when using a reverse prosthesis, with or without bone graft. Glenoid bone grafting is also a much easier procedure with an RTSA baseplate and screw fixation compared to an ATSA with screw fixation and the use of methyl methacrylate cement for implant fixation. Mizuno et al reported on deformities with a mean retroversion of 32° and mean posterior subluxation of 87% with excellent clinical outcomes and no recurrences of posterior subluxation with follow-up of 5 years.80 Other subsequent studies have also shown excellent clinical results in these patients with or without bone grafting.81,82
More recently, RTSA has been used in older patients with severe osteoarthritis and an intact rotator cuff, as our comfort level with the procedure has grown and the longer term results have shown favorable survivorship compared to ATSA. In fact, in some countries such as Australia and South Korea, 78% to 90% of all total shoulder arthroplasties are performed using a reverse prosthesis regardless of the indication or condition of the rotator cuff.46 Studies have shown similar outcomes when comparing RTSA performed with an intact rotator cuff compared to ATSA performed for osteoarthritis with an intact cuff.83,84 A retrospective cohort study looked at 135 patients who underwent either ATSA or RTSA, and all had an intact rotator cuff confirmed with advanced imaging.83 There were no significant differences in pain relief, ROM including internal rotation, or patient-reported outcomes. Satisfaction was very high in both groups, and complications and revision rates were similar.