Abstract
Background
Both anatomic total shoulder arthroplasty (aTSA) and reverse total shoulder arthroplasty (RTSA) are being increasingly performed. In the event of a complication necessitating revision, RTSA is more commonly performed in both scenarios. The purpose of this study was to compare clinical outcomes between patients undergoing revision RTSA for failed primary anatomic versus reverse total shoulder arthroplasty.
Methods
We performed a retrospective review of a prospective single-institution shoulder arthroplasty database. All revision RTSAs performed between 2007 and 2019 with a minimum 2-year clinical follow-up were initially included. After excluding patients with a preoperative diagnosis of infection, an oncologic indication, or incomplete outcomes assessment, we included 45 revision RTSAs performed for failed primary aTSA and 15 for failed primary RTSA. Demographics, surgical characteristics, active range of motion (external rotation [ER], internal rotation, forward elevation [FE], abduction), outcome scores (American Shoulder and Elbow Surgeons score, Constant Score, Shoulder Pain and Disability Index, Simple Shoulder Test, and University of California, Los Angeles score), and the incidence of postoperative complications was compared between groups.
Results
Primary aTSA was most often indicated for degenerative joint disease (82%), whereas primary RTSA was more often indicated for rotator cuff arthropathy (60%). On bivariate analysis, no statistically significant differences in any range of motion or clinical outcome measure were found between revision RTSA performed for failed aTSA vs. RTSA. On multivariate linear regression analysis, revision RTSA performed for failed aTSA vs. RTSA was not found to significantly influence any outcome measure. Humeral loosening as an indication for revision surgery was associated with more favorable outcomes for all four range of motion measures and all five outcome scores assessed. In contrast, an indication for revision of peri-prosthetic fracture was associated with poorer outcomes for three of four range of motion measures (ER, FE, abduction) and four of five outcome scores (Constant, Shoulder Pain and Disability Index, Simple Shoulder Test, University of California, Los Angeles). A preoperative diagnosis of fracture was associated with a poorer postoperative range of motion in ER, FE, and abduction, but was not found to significantly influence any outcome score. However, only two patients in our cohort had this indication. Complication and re-revision rates after revision RTSA for failed primary aTSA and RTSA were 27% and 9% vs. 20% and 14% (P = .487 and P = .515), respectively.
Conclusion
Clinical outcomes of patients undergoing revision RTSA for failed primary shoulder arthroplasty did not significantly differ based on whether aTSA or RTSA was initially performed. However, larger studies are needed to definitively ascertain the influence of the primary construct on the outcomes of revision RTSA.
Keywords: Baseplate loosening, Revision surgery, Outcome scores, Reverse shoulder arthroplasty, Shoulder replacement
As the indications for reverse total shoulder arthroplasty (RTSA) continue to expand, the decision to perform a primary anatomic total shoulder arthroplasty (aTSA) vs. an RTSA can require consideration of factors beyond preoperative patient characteristics and suspected shoulder pathology. One such consideration is the need for subsequent revision procedures given the nontrivial failure rates of both primary aTSA and primary RTSA. The revision rate after primary aTSA or RTSA has been reported up to 19% depending upon the patient population and follow-up period.15,19,20,24,26,31,34,35 Of the options for revision, RTSA is most utilized because of reduced reliance on a functional rotator cuff, stronger glenoid fixation, and improved glenohumeral stability.16
Previous studies have demonstrated that revision RTSA for failed shoulder arthroplasty can improve patient function and decrease pain.2,5,17,25,28,32 However, studies comparing clinical outcomes of revision RTSA performed for failed primary aTSA vs. RTSA are lacking. Shields et al compared patients undergoing revision RTSA for failed aTSA with a mechanical etiology (ie, rotator cuff failure or component loosening) to a matched cohort of primary RTSAs and found that patients undergoing revision surgery had worse subjective shoulder value scores, lower satisfaction, and a greater incidence of complications.33 These findings suggest that even in the best case-scenario, revision RTSA after failed primary aTSA is not equivalent to primary RTSA. Given the increasing need for revision and increasing overlap of indications between aTSA and RTSA, a difference in clinical outcomes after revision RTSA may influence surgeon decision-making at the time of primary arthroplasty.11
The purpose of this study was to compare range of motion (ROM), outcome scores, and the incidence of postoperative complications between patients undergoing revision RTSA for failed primary aTSA versus RTSA at minimum 2-year follow-up. We hypothesized that clinical outcomes of revision RTSA would not differ when stratified based on primary shoulder arthroplasty procedure.
Methods
We conducted a retrospective review of all revision RTSAs performed at a single institution between October 2007 and May 2019. Medical records and our institutional prospectively collected shoulder arthroplasty database were reviewed for demographic, operative, and clinical information. All patients between the ages of 40-90 years old at the time of revision surgery with a minimum 2-year clinical follow-up were reviewed. Patients were excluded if the primary arthroplasty was performed for an oncologic diagnosis. Shoulders were also excluded if they were being revised for infection or had previously undergone revision arthroplasty surgery given poorer outcomes documented in these cohorts.7,21,33 Surgeries were performed by one of four fellowship-trained surgeons.
Surgical procedure
Revision implants included multiple manufacturers depending on the previously placed implant. If portions of the primary arthroplasty were salvageable, efforts were made to utilize these implants to prevent further bone loss. All patients completed a rehabilitation protocol consisting of a physical therapist-directed home exercise program. Patients were placed in a sling for 6 weeks postoperatively and passive ROM was performed starting at 2 weeks. Active ROM was initiated after sling use was discontinued. Patients were advised to avoid weight-bearing activities until strengthening exercises were introduced three months postoperatively.
Patients
303 revision RTSAs in 273 patients were initially identified (Fig. 1). We excluded 106 revision RTSAs in 92 patients due to infection, oncologic indication, or multiple previous arthroplasty surgeries on the involved shoulder. Subsequently, 63 revision RTSAs in 63 patients were removed due to the primary arthroplasty procedure being a hemiarthroplasty or resurfacing arthroplasty. Finally, 50 revision RTSAs in 50 patients were excluded due to a lack of 2-year follow-up outcome data (n = 36) or incomplete postoperative outcome scores or ROM data (n = 14). In total, 60 revision RTSAs (45 primary aTSA, 15 primary RTSA) were included for analysis.
Figure 1.
Flowchart of patient inclusion and exclusion. aTSA, anatomic total shoulder arthroplasty; RTSA, reverse total shoulder arthroplasty.
Study outcomes
ROM and outcome scores were collected preoperatively and postoperatively at all follow-up appointments. ROM measurements were assessed using a goniometer and included active forward elevation (FE), active abduction, and active external rotation (ER). Additionally, active internal rotation was measured by vertebral segments and scored as follows: 0; hip, 1; buttocks, 2; sacrum, 3; L5 to L4, 4; L3 to L1, 5; T12 to T8, 6; and T7 or higher, 7.12 Several clinical outcome scores were collected: the American Shoulder and Elbow Surgeons (ASES) score,29 the Constant Score,8 the Shoulder Pain and Disability Index score,30 the Simple Shoulder Test score,23 and the University of California, Los Angeles score.1 The incidence of complications and re-revision at the time of this study was assessed from all 110 patients that underwent revision RTSA for failed primary aTSA (n = 75) or primary RTSA (n = 35) regardless of whether 2-year follow-up data and complete scores were available to mitigate follow-up bias.
Statistical analysis
The study sample size was fixed thus a priori power analysis was not performed. We estimated an effect size of d = 0.85 for 80% statistical power given N1 = 45, N2 = 15, two-tailed testing, and α = 0.05. Continuous data were analyzed using two-tailed unpaired t-tests. Fisher’s exact test was used to analyze categorical data. Multivariate regression analysis was performed to determine whether the primary procedure (aTSA or RTSA) was predictive of improved outcome scores, ROM, or incidence of complications, independent of patient demographics (age at surgery, sex, body mass index), and surgical indications (reason for revision and preoperative diagnosis of the primary procedure). Backward stepwise selection was used for variable selection in all multivariable regression models to reduce the influence of multicollinearity on effect size estimates.37 The implementation of this method has been previously described in the shoulder arthroplasty literature.13,22 All statistical analyses were performed using R Software (version 3.6.3; R Foundation for Statistical Computing, Vienna, Austria) with P < .05.
Results
Age at surgery and the proportion of female patients were comparable between cohorts undergoing revision RTSA for failed primary aTSA and RTSA (Table I). The mean follow-up was 54 months after conversion to RTSA from aTSA and 44 months from RTSA (P = .619). The most common indication for primary aTSA was degenerative joint disease (82%), whereas primary RTSA was more often indicated for rotator cuff arthropathy (60%). Bone graft was used behind the baseplate in 29% (N = 13) of revised aTSAs and 13% (N = 2) of revised RTSAs (P = .313). The stem was revised in all cases of humeral loosening.
Table I.
Demographics of included patients.
| Characteristic | Primary aTSA revised to RTSA (N = 45) | Primary RTSA revised to RTSA (N = 15) | P value |
|---|---|---|---|
| Age at surgery (years) | 67.4 ± 7.9 | 70.1 ± 13 | .286 |
| Female sex | 53.3% (24) | 53.3% (8) | 1 |
| Body mass index (kg/m2) | 29.8 ± 5.7 | 29.5 ± 4.7 | .975 |
| Follow-up (months) | 53.9 ± 30.3 | 44.3 ± 27.5 | .619 |
| Comorbidities | |||
| Inflammatory arthritis | 8.9% (4) | 6.7% (1) | 1 |
| Hypertension | 57.8% (26) | 60.0% (9) | 1 |
| Heart disease | 13.3% (6) | 13.3% (2) | 1 |
| Diabetes mellitus | 15.6% (7) | 6.7% (1) | .666 |
| Tobacco use | 8.9% (4) | 0.0% (0) | .564 |
| Reason for revision | - | ||
| Humeral loosening | 11.1% (5) | 20.0% (3) | .400 |
| Glenoid loosening | 33.3% (15) | 26.7% (4) | .755 |
| Rotator cuff failure | 33.3% (15) | 6.7% (1) | .050 |
| Dislocation | 2.2% (1) | 13.3% (2) | .151 |
| 15.6% (7) | 13.3% (2) | 1 | |
| Unexplained pain | 11.1% (5) | 6.7% (1) | 1 |
| Periprosthetic fracture | 4.4% (2) | 13.3% (2) | .258 |
| Implant wear | 6.7% (3) | 0.0% (0) | .566 |
| Instability | 2.2% (1) | 0.0% (0) | 1 |
| Polyethylene dissociation | 2.2% (1) | 0.0% (0) | 1 |
| Preoperative diagnosis of primary shoulder arthroplasty | - | ||
| DJD | 82.2% (37) | 40.0% (6) | .003 |
| 2.2% (1) | 6.7% (1) | .441 | |
| Rotator cuff arthropathy | 0% (0) | 60.0% (9) | <.001 |
| Instability arthropathy | 4.4% (2) | 6.7% (1) | 1 |
| Avascular necrosis | 2.2% (1) | 0.0% (0) | 1 |
| Rheumatoid arthritis | 2.2% (1) | 0.0% (0) | 1 |
| Unknown | 6.7% (3) | 6.7% (1) | 1 |
aTSA, anatomic total shoulder arthroplasty; DJD, degenerative joint disease; RTSA, reverse total shoulder arthroplasty.
Data are presented as mean ± standard deviation or % (N).
Bold values indicate statistical significance.
Outcome scores
Outcome scores assessed preoperatively, postoperatively at the latest follow-up, and the improvement between preoperative and postoperative time points did not statistically differ based on the primary procedure (Table II). On multivariate analysis, the primary procedure did not influence postoperative outcomes (Table III). Favorable postoperative outcome scores were positively associated with revision RTSA performed for humeral loosening without an associated fracture, while less favorable postoperative outcome scores were associated with revision RTSA performed for periprosthetic fracture.
Table II.
Comparison of range of motion and clinical outcome scores between patients undergoing revision RTSA for failed primary aTSA versus RTSA.
| Outcome measure | Primary aTSA revised to RTSA (N = 45) | Primary RTSA revised to RTSA (N = 15) | P value |
|---|---|---|---|
| Preoperative | |||
| SPADI score | 64.0 ± 21.4 | 54.2 ± 23.8 | .236 |
| SST score | 4.3 ± 3.2 | 5.2 ± 2.6 | .370 |
| ASES score | 42.7 ± 18.5 | 52.9 ± 19.9 | .151 |
| UCLA score | 14.7 ± 5.5 | 18.2 ± 7.6 | .228 |
| Constant Score | 40.7 ± 18.7 | 48.1 ± 17.8 | .271 |
| Active ER (°) | 29 ± 27 | 19 ± 26 | .343 |
| Active FE (°) | 70 ± 36 | 84 ± 36 | .294 |
| Active abduction (°) | 66 ± 35 | 81 ± 35 | .212 |
| Active IR score | 4.0 ± 2.0 | 3.4 ± 2.2 | .391 |
| Postoperative | |||
| SPADI score | 40.8 ± 21.1 | 38.2 ± 23.3 | .698 |
| SST score | 6.9 ± 3.4 | 7.0 ± 3.6 | .950 |
| ASES score | 61.3 ± 21.2 | 61.9 ± 24.1 | .929 |
| UCLA score | 23.3 ± 7.5 | 24.4 ± 8.6 | .650 |
| Constant Score | 58.3 ± 18.4 | 58.5 ± 27.8 | .971 |
| Active ER (°) | 29 ± 20 | 19 ± 26 | .220 |
| Active FE (°) | 105 ± 35 | 109 ± 48 | .748 |
| Active abduction (°) | 98 ± 32 | 102 ± 46 | .721 |
| Active IR score | 4.0 ± 1.6 | 3.8 ± 2.0 | .761 |
| Improvement | |||
| SPADI score | −23.8 ± 26.0 | −17.3 ± 30.1 | .526 |
| SST score | 3.1 ± 3.7 | 1.9 ± 3.6 | .366 |
| ASES score | 18.9 ± 23.3 | 7.3 ± 22.9 | .159 |
| UCLA score | 9.2 ± 8.3 | 4.6 ± 9.6 | .214 |
| Constant Score | 18.3 ± 20.0 | 10.2 ± 24.3 | .356 |
| Active ER (°) | 2 ± 27 | 5 ± 20 | .768 |
| Active FE (°) | 40 ± 34 | 21 ± 41 | .185 |
| Active abduction (°) | 34 ± 34 | 19 ± 38 | .252 |
| Active IR score | −0.2 ± 2.0 | 0.5 ± 2.3 | .438 |
ASES, American Shoulder and Elbow Surgeons; ER, external rotation; FE, forward elevation; IR, internal rotation; SPADI, Shoulder Pain and Disability Index; SST, Simple Shoulder Test; UCLA, University of California, Los Angeles; aTSA, anatomic total shoulder arthroplasty; RTSA, reverse total shoulder arthroplasty.
Data presented as mean ± standard deviation.
Statistically significant comparisons are denoted in bold.
Table III.
Multivariate linear regression assessing the influence of primary shoulder arthroplasty (aTSA vs. RTSA) on postoperative outcome scores after revision RTSA.
| Preoperative predictor | Outcome score/regression coefficient and P value |
||||
|---|---|---|---|---|---|
| SPADI | SST | ASES | UCLA | Constant | |
| Intercept | 36.4 | 7.0 | 67.7 | 25.7 | 66.2 |
| Revision RTSA from RTSA vs. aTSA | - | - | - | - | - |
| Age at surgery (years) | - | - | - | - | - |
| Male sex | - | 1.2, P = .118 | - | - | - |
| Body mass index (kg/m2) | - | - | - | - | - |
| Comorbidities | |||||
| Inflammatory arthritis | - | - | - | - | - |
| Heart disease | - | - | - | 3.9, P = .130 | 10.1, P = .143 |
| Diabetes mellitus | 11.2, P = .110 | - | - | - | −9.3, P = .162 |
| Tobacco use | - | - | - | - | - |
| Reason for revision | |||||
| Humeral loosening | −27.6, P < .001 | 3.8, P = .001 | 26.3, P = .001 | 8.5, P = .001 | 22.9, P = .001 |
| Glenoid loosening | - | - | - | - | - |
| Rotator cuff failure | - | - | - | −4.5, P = .034 | −11.4, P = .040 |
| Instability, dislocation, or subluxation | −8.1, P = .164 | - | - | - | - |
| Periprosthetic fracture | 32.2, P = .002 | −5.0, P = .003 | −26.4, P = .014 | −11.5, P = .002 | −37.3, P < .001 |
| DJD/implant wear | - | - | - | - | - |
| Preoperative diagnosis of primary shoulder arthroplasty | |||||
| DJD | 10.1, P = .047 | −1.4, P = .087 | −13.0, P = .017 | −3.3, P = .083 | −9.3, P = .059 |
| Fracture | - | - | - | - | - |
| Rotator cuff arthropathy | - | - | - | - | - |
| Instability arthropathy | - | - | - | - | - |
| Avascular necrosis | −32.5, P = .081 | 5.0, P = .107 | - | - | - |
ASES, American Shoulder and Elbow Surgeons; aTSA, anatomic total shoulder arthroplasty; DJD, degenerative joint disease; RTSA, reverse total shoulder arthroplasty; SPADI, Shoulder Pain and Disability Index; SST, Simple Shoulder Test; UCLA, University of California, Los Angeles.
Multivariate linear regression performed with backward stepwise selection.
Statistically significant comparisons are denoted in bold.
Range of motion
ROM assessed preoperatively, postoperatively at the latest follow-up, and the improvement between preoperative and postoperative time points did not differ based on the primary procedure (Table II). On multivariate analysis, the primary procedure did not influence postoperative ROM (Table IV). Superior postoperative ROM was positively associated with revision RTSA performed for humeral loosening, while poorer postoperative ROM was associated with revision RTSA performed for a periprosthetic fracture and a primary shoulder arthroplasty preoperative diagnosis of fracture.
Table IV.
Multivariate linear regression assessing the influence of primary shoulder arthroplasty (aTSA vs. RTSA) on postoperative range of motion after revision RTSA.
| Preoperative predictor | Active ROM measure/regression coefficient and P value |
|||
|---|---|---|---|---|
| ER (°) | FE (°) | Abduction (°) | IR score | |
| Intercept | 54.4 | 127.6 | 123.1 | 6.1 |
| Revision RTSA from RTSA vs. aTSA | - | - | - | - |
| Age at surgery (years) | - | - | - | - |
| Male sex | - | - | - | - |
| Body mass index (kg/m2) | −0.9, P = .074 | - | - | −0.1, P = .042 |
| Comorbidities | ||||
| Inflammatory arthritis | - | −24.1, P = .077 | −26.3, P = .041 | - |
| Heart disease | 13.7, P = .086 | 30.1, P = .006 | 32.3, P = .003 | - |
| Diabetes mellitus | - | - | - | - |
| Tobacco use | - | - | - | - |
| Reason for revision | ||||
| Humeral loosening | 17.5, P = .020 | 32.1, P = .004 | 30.9, P = .003 | 1.6, P = .008 |
| Glenoid loosening | - | 12.1, P = .200 | - | - |
| Rotator cuff failure | - | −21.0, P = .025 | −17.4, P = .033 | - |
| Instability, dislocation, or subluxation | - | 15.3, P = .127 | ||
| Periprosthetic fracture | −25.9, P = .022 | −80.6, P < .001 | −73.9, P < .001 | - |
| DJD/implant wear | - | - | - | - |
| Preoperative diagnosis of primary shoulder arthroplasty | ||||
| DJD | - | −28.6, P = .004 | −23.9, P = .009 | - |
| Fracture | −32.9, P = .025 | −50.1, P = .025 | −56.7, P = .008 | - |
| Rotator cuff arthropathy | −10.2, P = .095 | −19.0, P = .077 | −17.1, P = .086 | - |
| Instability arthropathy | - | - | - | - |
| Avascular necrosis | - | - | - | - |
ROM, range of motion; aTSA, anatomic total shoulder arthroplasty; DJD, degenerative joint disease; ER, external rotation; FE, forward elevation; IR, internal rotation; RTSA, reverse total shoulder arthroplasty.
Multivariate linear regression performed with backward stepwise selection.
Statistically significant comparisons are denoted in bold.
Complications
The rates of patients with a complication from revision RTSA for failed primary aTSA and RTSA were 27% (20 out of 75) and 20% (7 out of 35), respectively (P = .487) (Table V). Baseplate loosening was the most common complication of revision RTSA for failed aTSA (12%), whereas baseplate loosening and dislocation occurred at equal rates in the patient cohort revised from primary RTSA (6% for both). Seven patients (9%) that underwent revision RTSA for failed primary aTSA required re-revision compared to five patients (14%) after revision RTSA for failed primary RTSA (P = .515). Patients that required re-revision after revision RTSA for failed primary aTSA included: infection requiring component explantation and revision to an antibiotic spacer, humeral loosening revised to a long stemmed humerus, baseplate loosening secondary to graft resorption that was revised to hemiarthroplasty, polyethylene wear secondary to coracoid impingement that was revised to RTSA, glenosphere loosening that was revised to an RTSA with the stem retained, unexplained pain that was managed with revision RTSA at another institution, and a periprosthetic fracture that was revised to a hemiarthroplasty. Patients that required re-revision after revision RTSA for failed primary RTSA included: dislocation that was revised to RTSA with the stem retained, periprosthetic fracture that was revised to a long stem hemiarthroplasty, two cases of baseplate loosening that were revised to hemiarthroplasty with the stem retained, and humeral loosening that was revised to RTSA with a new stem.
Table V.
Incidence of surgical complications and re-revision in all patients undergoing revision RTSA for failed primary aTSA versus RTSA.
| Complication | Primary aTSA revised to RTSA (N = 75) | Primary RTSA revised to RTSA (N = 35) | P value | ||
|---|---|---|---|---|---|
| All-cause complications | 20 | 26.7% (20) | 7 | 20.0% (7) | .487 |
| Humeral stem loosening | 2 | 2.7% (2) | 1 | 2.9% (1) | - |
| Baseplate loosening | 9 | 12.0% (9) | 2 | 5.7% (2) | - |
| Glenosphere loosening | 1 | 1.3% (1) | 1 | 2.9% (1) | - |
| Polyethylene wear | 1 | 1.3% (1) | 0 | 0.0% (0) | - |
| Dislocation | 1 | 1.3% (1) | 2 | 5.7% (2) | - |
| Proximal humerus fracture | 0 | 0.0% (0) | 0 | 0.0% (0) | - |
| Infection | 1 | 1.3% (1) | 0 | 0.0% (0) | - |
| Acromial fracture | 1 | 1.3% (1) | 0 | 0.0% (0) | - |
| Periprosthetic fracture | 2 | 2.7% (2) | 1 | 2.9% (1) | - |
| Unexplained pain | 1 | 1.3% (1) | 0 | 0.0% (0) | - |
| Intraoperative fracture: greater tuberosity | 1 | 1.3% (1) | 0 | 0.0% (0) | - |
| Intraoperative fracture: humeral shaft cortex | 0 | 0.0% (0) | 1 | 2.9% (1) | - |
| Re-revision rate | 7 | 9.3% (7) | 5 | 14.3% (5) | .515 |
aTSA, anatomic total shoulder arthroplasty; RTSA, reverse total shoulder arthroplasty.
Discussion
Since indications for shoulder arthroplasty have expanded and the prevalence of shoulder arthroplasty is rising, an increase in the incidence of revision shoulder arthroplasty is to be expected.7,11 In a prospective, case-control study with a minimum 2-year follow-up, Kiet et al reported revision rates of 11% and 9% after primary aTSA and RTSA, respectively. When necessary, conversion to RTSA is often the only option due to rotator cuff failure/degeneration or extensive glenoid bone loss. When the indication to perform a primary aTSA vs. RTSA is not clear, an important factor to consider is the possible outcome following revision surgery, if needed in the future. There is a growing body of literature evaluating outcomes after revision RTSA for failed shoulder arthroplasties.2,5,17,18,28,33 However, there are no studies to date that directly compare the outcomes of revision RTSA for primary aTSA vs. RTSA. Herein, we found no influence of the type of primary shoulder arthroplasty (aTSA vs. RTSA) performed on postoperative clinical outcomes after revision RTSA.
To our knowledge, our study is the first to compare outcomes of revision RTSA based on the primary procedure performed. Whether a primary aTSA or RTSA was initially performed did not influence the clinical outcomes of revision RTSA on either univariate or multivariate analysis (Table II, Table III, Table IV). Black et al compared patients under the age of 65 that were revised from an aTSA to RTSA to an age-matched cohort undergoing primary RTSA and found similar Simple Shoulder Test (58.8 ± 25.1 vs. 67.3 ± 26.1, P = .124) and ASES scores (69.7 ± 18.0 vs. 74.0 ± 23.8, P = .180), but worse Subjective Shoulder Values in the revision RTSA group (60 ± 29% vs. 76 ± 26%, P = .015).5 These results demonstrate that RTSA is a viable salvage procedure for failed primary aTSA; however, comparison to revision RTSA for failed primary RTSA was not performed. Our data demonstrated improvements in both outcome scores and ROM from preoperative to postoperative after conversion of aTSA to RTSA. However, there was no significant difference in postoperative outcomes or preoperative to postoperative improvement between the primary aTSA and RTSA cohorts on either univariate or multivariate analysis, suggesting that the primary procedure has a negligible influence on measured clinical outcomes for revision RTSA. Given the limited sample size of this study, future investigations are needed before definitive conclusions can be made.
In addition, our study also did not demonstrate a significant difference in complication rates (Table V) between primary procedure types (aTSA, 27%; RTSA, 20%). Our complication rates are similar to previously reported complication rates after revision RTSA for failed primary aTSA (24%-31%),3,6,33 but lower than reported rates for failed primary RTSA (35%-56%).4,6,27 However, we recognize that the low revision rate for RTSA in the present study may have been affected by the limited number of patients that were able to be included (n = 35).
This study evaluated 60 revision RTSAs, representing one of the largest cohort studies assessing clinical outcomes in this patient population.4,10,17, 18, 19,24,33,36 Both study groups had a similar mean age at surgery (67 vs. 70 years, P = .286) and proportion of females (53% for both). In our study, no association between sex and clinical outcomes was identified on multivariate analysis. Moreover, we did not find a disproportionate sex distribution between patients undergoing revision RTSA for primary aTSA vs. RTSA. Although prior studies have not assessed the influence of patient sex on revision RTSA outcomes, prior studies have assessed its influence on the risk of requiring revision shoulder arthroplasty. A previous observational study of 3795 primary aTSAs and 8878 RTSAs by Gill et al suggested that sex may be associated with an increased risk of revision.14 While they did not identify a significant cumulative risk for revision between sexes in a 4-year span, men that underwent RTSA had higher revision rates in the first 3 months (hazard ratio 4.0, 95% confidence interval 1.72-9.09) whereas women that underwent aTSA had a greater revision risk at 3 months and onward (hazard ratio 2.77, 95% confidence interval 1.55-4.92). A similar large British registry study found that younger males have a greater lifetime risk of revision after both elective primary aTSA and RTSA.9
Disparities in the reasons for conversion to RTSA from primary aTSA vs. RTSA on univariate analysis did not meet the threshold for statistical significance in this study. However, an indication of humeral loosening was associated with improved postoperative outcomes and periprosthetic fracture was associated with worse outcomes independent of whether revision RTSA was performed for failed primary aTSA or RTSA (Tables III and IV). The influence of an indication of humeral loosening or periprosthetic fracture for revision shoulder arthroplasty has not been previously studied. Otte et al found significant improvements in outcome scores and ROM after revision RTSA with no difference between indications for revision with the exception of more favorable postoperative active FE in patients undergoing revision surgery for a loose glenoid component.25 In contrast, glenoid loosening as an indication for revision RTSA was not found to influence any postoperative clinical outcomes on multivariate analysis. Poorer outcomes following revision shoulder arthroplasty for infection have also been well established, thus we excluded those patients from our study.21 Our results suggest that revision RTSA can be successful regardless of the indication for revision when aseptic.
Preoperative diagnoses related to the primary shoulder arthroplasty of fracture were associated with poorer postoperative ROM in external rotation, FE, and abduction independent of the primary shoulder arthroplasty procedure (Table IV), but the influence on outcome scores was minimal (Table III). While statistically significant, these findings should be interpreted with caution: only two patients had an initial diagnosis of fracture in this study cohort. Nonetheless, these findings generally corroborate previous studies that have found poorer outcomes associated with revision RTSA performed for failed shoulder arthroplasty with an initial fracture indication.18,28 For example, Holschen et al compared patients undergoing revision RTSA for failed primary shoulder arthroplasty and found that patients with an initial fracture indication had poorer ASES scores (58.9 vs. 71.3, P = .048), a greater rate of complications (24% vs. 9%), and a higher rate of scapular notching (33% vs. 14%) after revision RTSA.18 However, Constant Scores, ROM in FE and abduction, and pain scores did not differ based indication for initial arthroplasty.18
While the findings of this study are suggestive that the primary surgery, whether aTSA or RTSA, does not influence outcomes after a revision RTSA, there are certain limitations. First, its retrospective nature leads to potential selection bias. In particular, 36 patients had to be excluded for lack of a 2-year minimum follow-up and an additional 14 patients were excluded for incomplete patient data. Additionally, the small sample size may reflect an underpowered population to detect significant differences between outcomes between primary aTSA and RTSA. However, given the low prevalence of revision RTSA, this is not unexpected and our study of this patient population is one of the largest to date. Furthermore, these findings may be able to inform future study design when sufficient patient populations become available to assess outcomes of revision RTSA. Estimates of the influence of low prevalence variables included in our multivariate models should be interpreted with caution (eg, periprosthetic fracture as a reason for revision, a preoperative diagnosis of fracture related to the primary shoulder arthroplasty); thus, we were careful to only draw conclusions from the multivariate model results that demonstrated association across multiple outcome measures. The primary aim of our study was to assess the influence of the primary shoulder arthroplasty and thus we only aimed to control for the influence of other variables (eg, reason for revision, preoperative diagnosis of primary shoulder arthroplasty) in our multivariate analysis, we did not ensure adequate statistical power to detect other interactions. Finally, our study was performed at a single institution thus the generalizability of our conclusions to other hospitals and surgeons may be limited. Furthermore, the heterogeneity in the study population may limit the clinical insight that can be gained from this study. Future studies should focus on a multicenter prospective evaluation of these revision RTSA cases to best examine outcomes.
Conclusion
Clinical outcomes of patients who underwent revision RTSA for failed primary shoulder arthroplasty in our study suggest that whether aTSA or RTSA was initially performed has a negligible influence on outcomes of revision RTSA. Future studies with larger cohorts should be conducted to determine what other factors may influence outcomes including primary arthroplasty construct and primary arthroplasty indications.
Disclaimers
Funding: No funding was disclosed by the authors.
Conflict of interest: Dr. T. Wright is a consultant and receives royalties from Exactech, Inc. Dr. King is a consultant for Exactech, Inc and LinkBio Corp. Dr. Schoch is a paid consultant for Exactech. He receives royalties from Exactech, Innomed and Responsive Arthroscopy. Dr. Farmer is a consultant for Exactech, Inc and Arthrex Inc. The other authors, their immediate families, and any research foundations with which they are affiliated have not received any financial payments or other benefits from any commercial entity related to the subject of this article.
Footnotes
This study was approved by the University of Florida Institutional Review Board, IRB#: 201902620.
References
- 1.Amstutz H.C., Sew Hoy A.L., Clarke I.C. UCLA anatomic total shoulder arthroplasty. Clin Orthop Relat Res. 1981;155:7–20. [PubMed] [Google Scholar]
- 2.Austin L., Zmistowski B., Chang E.S., Williams G.R. Is reverse shoulder arthroplasty a reasonable alternative for revision arthroplasty? Clin Orthop Relat Res. 2011;469:2531–2537. doi: 10.1007/s11999-010-1685-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Bartels D.W., Marigi E., Sperling J.W., Sanchez-Sotelo J. Revision reverse shoulder arthroplasty for anatomical glenoid component loosening was not universally successful: a detailed analysis of 127 consecutive shoulders. J Bone Joint Surg Am. 2021;103:879–886. doi: 10.2106/JBJS.20.00555. [DOI] [PubMed] [Google Scholar]
- 4.Black E.M., Roberts S.M., Siegel E., Yannopoulos P., Higgins L.D., Warner J.J.P. Failure after reverse total shoulder arthroplasty: what is the success of component revision? J Shoulder Elbow Surg. 2015;24:1908–1914. doi: 10.1016/j.jse.2015.05.029. [DOI] [PubMed] [Google Scholar]
- 5.Black E.M., Roberts S.M., Siegel E., Yannopoulos P., Higgins L.D., Warner J.J.P. Reverse shoulder arthroplasty as salvage for failed prior arthroplasty in patients 65 years of age or younger. J Shoulder Elbow Surg. 2014;23:1036–1042. doi: 10.1016/j.jse.2014.02.019. [DOI] [PubMed] [Google Scholar]
- 6.Bois A.J., Knight P., Alhojailan K., Bohsali K.I. Clinical outcomes and complications of reverse shoulder arthroplasty used for failed prior shoulder surgery: a systematic review and meta-analysis. JSES Int. 2020;4:156–168. doi: 10.1016/j.jses.2019.10.108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Chalmers P.N., Boileau P., Romeo A.A., Tashjian R.Z. Revision reverse shoulder arthroplasty. J Am Acad Orthop Surg. 2019;27:426–436. doi: 10.5435/JAAOS-D-17-00535. [DOI] [PubMed] [Google Scholar]
- 8.Constant C.R., Murley A.H. A clinical method of functional assessment of the shoulder. Clin Orthop Relat Res. 1987;214:160–164. [PubMed] [Google Scholar]
- 9.Craig R.S., Lane J.C.E., Carr A.J., Furniss D., Collins G.S., Rees J.L. Serious adverse events and lifetime risk of reoperation after elective shoulder replacement: population based cohort study using hospital episode statistics for England. BMJ. 2019;364:l298. doi: 10.1136/bmj.l298. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Dezfuli B., King J.J., Farmer K.W., Struk A.M., Wright T.W. Outcomes of reverse total shoulder arthroplasty as primary versus revision procedure for proximal humerus fractures. J Shoulder Elbow Surg. 2016;25:1133–1137. doi: 10.1016/j.jse.2015.12.002. [DOI] [PubMed] [Google Scholar]
- 11.Farley K.X., Wilson J.M., Kumar A., Gottschalk M.B., Daly C., Sanchez-Sotelo J., et al. Prevalence of shoulder arthroplasty in the United States and the increasing Burden of revision shoulder arthroplasty. JB JS Open Access. 2021;6 doi: 10.2106/JBJS.OA.20.00156. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Flurin P.H., Marczuk Y., Janout M., Wright T., Zuckerman J., Roche C. Comparison of outcomes using anatomic and reverse total shoulder arthroplasty. Bull Hosp Jt Dis (2013) 2013;71:101–107. [PubMed] [Google Scholar]
- 13.Friedman R.J., Eichinger J., Schoch B., Wright T., Zuckerman J., Flurin P.-H., et al. Preoperative parameters that predict postoperative patient-reported outcome measures and range of motion with anatomic and reverse total shoulder arthroplasty. JSES Open Access. 2019;3:266–272. doi: 10.1016/j.jses.2019.09.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Gill D.R.J., Page BMedSci R.S., Graves S.E., Rainbird S., Hatton A. A comparison of revision rates for osteoarthritis of primary reverse total shoulder arthroplasty to primary anatomic shoulder arthroplasty with a cemented all-polyethylene glenoid: analysis from the Australian Orthopaedic Association National Joint Replacement Registry. Clin Orthop Relat Res. 2021;479:2216–2224. doi: 10.1097/CORR.0000000000001869. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Goldenberg B.T., Samuelsen B.T., Spratt J.D., Dornan G.J., Millett P.J. Complications and implant survivorship following primary reverse total shoulder arthroplasty in patients younger than 65 years: a systematic review. J Shoulder Elbow Surg. 2020;29:1703–1711. doi: 10.1016/j.jse.2020.02.004. [DOI] [PubMed] [Google Scholar]
- 16.Hernandez N.M., Chalmers B.P., Wagner E.R., Sperling J.W., Cofield R.H., Sanchez-Sotelo J. Revision to reverse total shoulder arthroplasty restores stability for patients with unstable shoulder prostheses. Clin Orthop Relat Res. 2017;475:2716–2722. doi: 10.1007/s11999-017-5429-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Holcomb J.O., Cuff D., Petersen S.A., Pupello D.R., Frankle M.A. Revision reverse shoulder arthroplasty for glenoid baseplate failure after primary reverse shoulder arthroplasty. J Shoulder Elbow Surg. 2009;18:717–723. doi: 10.1016/j.jse.2008.11.017. [DOI] [PubMed] [Google Scholar]
- 18.Holschen M., Franetzki B., Witt K.-A., Liem D., Steinbeck J. Conversions from anatomic shoulder replacements to reverse total shoulder arthroplasty: do the indications for initial surgery influence the clinical outcome after revision surgery? Arch Orthop Trauma Surg. 2017;137:167–172. doi: 10.1007/s00402-016-2595-5. [DOI] [PubMed] [Google Scholar]
- 19.Kiet T.K., Feeley B.T., Naimark M., Gajiu T., Hall S.L., Chung T.T., et al. Outcomes after shoulder replacement: comparison between reverse and anatomic total shoulder arthroplasty. J Shoulder Elbow Surg. 2015;24:179–185. doi: 10.1016/j.jse.2014.06.039. [DOI] [PubMed] [Google Scholar]
- 20.Kim D.M., Aldeghaither M., Alabdullatif F., Shin M.J., Kholinne E., Kim H., et al. Loosening and revision rates after total shoulder arthroplasty: a systematic review of cemented all-polyethylene glenoid and three modern designs of metal-backed glenoid. BMC Musculoskelet Disord. 2020;21:114. doi: 10.1186/s12891-020-3135-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Kim K., Elbuluk A., Jia N., Osmani F., Levieddin J., Zuckerman J., et al. Revision shoulder arthroplasty: patient-reported outcomes vary according to the etiology of revision. J Orthop. 2018;15:922–926. doi: 10.1016/j.jor.2018.08.030. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Matsen F.A., Russ S.M., Vu P.T., Hsu J.E., Lucas R.M., Comstock B.A. What factors are predictive of patient-reported outcomes? A prospective study of 337 shoulder arthroplasties. Clin Orthop Relat Res. 2016;474:2496–2510. doi: 10.1007/s11999-016-4990-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Matsen F.A., Ziegler D.W., DeBartolo S.E. Patient self-assessment of health status and function in glenohumeral degenerative joint disease. J Shoulder Elbow Surg. 1995;4:345–351. doi: 10.1016/s1058-2746(95)80018-2. [DOI] [PubMed] [Google Scholar]
- 24.Monir J.G., Abeyewardene D., King J.J., Wright T.W., Schoch B.S. Reverse shoulder arthroplasty in patients younger than 65 years, minimum 5-year follow-up. J Shoulder Elbow Surg. 2020;29:e215–e221. doi: 10.1016/j.jse.2019.10.028. [DOI] [PubMed] [Google Scholar]
- 25.Otte R.S., Naylor A.J., Blanchard K.N., Cancienne J.M., Chan W., Romeo A.A., et al. Salvage reverse total shoulder arthroplasty for failed anatomic total shoulder arthroplasty: a cohort analysis. J Shoulder Elbow Surg. 2020;29:S134–S138. doi: 10.1016/j.jse.2020.04.013. [DOI] [PubMed] [Google Scholar]
- 26.Parada S.A., Flurin P.-H., Wright T.W., Zuckerman J.D., Elwell J.A., Roche C.P., et al. Comparison of complication types and rates associated with anatomic and reverse total shoulder arthroplasty. J Shoulder Elbow Surg. 2021;30:811–818. doi: 10.1016/j.jse.2020.07.028. [DOI] [PubMed] [Google Scholar]
- 27.Patel M., Johnson P., Shishani Y., Ahmed S., Gobezie R., Lederman E. Revision of failed reverse shoulder arthroplasty to revision reverse shoulder arthroplasty leads to significant clinical and functional improvement. Semin Arthroplasty. 2021;31:325–329. doi: 10.1053/j.sart.2021.01.007. [DOI] [Google Scholar]
- 28.Patel D.N., Young B., Onyekwelu I., Zuckerman J.D., Kwon Y.W. Reverse total shoulder arthroplasty for failed shoulder arthroplasty. J Shoulder Elbow Surg. 2012;21:1478–1483. doi: 10.1016/j.jse.2011.11.004. [DOI] [PubMed] [Google Scholar]
- 29.Richards R.R., An K.N., Bigliani L.U., Friedman R.J., Gartsman G.M., Gristina A.G., et al. A standardized method for the assessment of shoulder function. J Shoulder Elbow Surg. 1994;3:347–352. doi: 10.1016/S1058-2746(09)80019-0. [DOI] [PubMed] [Google Scholar]
- 30.Roach K.E., Budiman-Mak E., Songsiridej N., Lertratanakul Y. Development of a shoulder pain and disability index. Arthritis Care Res. 1991;4:143–149. [PubMed] [Google Scholar]
- 31.Shah S.S., Gaal B.T., Roche A.M., Namdari S., Grawe B.M., Lawler M., et al. The modern reverse shoulder arthroplasty and an updated systematic review for each complication: part I. JSES Int. 2020;4:929–943. doi: 10.1016/j.jseint.2020.07.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Sheth M.M., Sholder D., Getz C.L., Williams G.R., Namdari S. Revision of failed hemiarthroplasty and anatomic total shoulder arthroplasty to reverse total shoulder arthroplasty. J Shoulder Elbow Surg. 2019;28:1074–1081. doi: 10.1016/j.jse.2018.10.026. [DOI] [PubMed] [Google Scholar]
- 33.Shields E., Wiater J.M. Patient outcomes after revision of anatomic total shoulder arthroplasty to reverse shoulder arthroplasty for rotator cuff failure or component loosening: a matched cohort study. J Am Acad Orthop Surg. 2019;27:e193. doi: 10.5435/JAAOS-D-17-00350. [DOI] [PubMed] [Google Scholar]
- 34.Singh J.A., Sperling J.W., Cofield R.H. Revision surgery following total shoulder arthroplasty. J Bone Joint Surg Br. 2011;93:1513–1517. doi: 10.1302/0301-620X.93B11.26938. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Villacis D., Sivasundaram L., Pannell W.C., Heckmann N., Omid R., Rick Hatch G.F., 3rd Complication rate and implant survival for reverse shoulder arthroplasty versus total shoulder arthroplasty: results during the initial 2 years. J Shoulder Elbow Surg. 2016;25:927–935. doi: 10.1016/j.jse.2015.10.012. [DOI] [PubMed] [Google Scholar]
- 36.Wierks C., Skolasky R.L., Ji J.H., McFarland E.G. Reverse total shoulder replacement: intraoperative and early postoperative complications. Clin Orthop Relat Res. 2009;467:225–234. doi: 10.1007/s11999-008-0406-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Yoo W., Mayberry R., Bae S., Singh K., Peter He Q., Lillard J.W. A study of effects of multicollinearity in the multivariable analysis. Int J Appl Sci Technol. 2014;4:9–19. [PMC free article] [PubMed] [Google Scholar]