Risk factors for failing to achieve improvement after anatomic total shoulder arthroplasty for glenohumeral osteoarthritis

ARTICLE IN PRESS J Shoulder Elbow Surg (2018) ■■, ■■–■■

www.elsevier.com/locate/ymse

ORIGINAL ARTICLE

Risk factors for failing to achieve improvement after anatomic total shoulder arthroplasty for glenohumeral osteoarthritis Gregory T. Mahony, MSa, Brian C. Werner, MDa,b, Brenda Chang, MS, MPHc, Brian M. Grawe, MDa,d,*, Samuel A. Taylor, MDa, Edward V. Craig, MD, MPHa, Russell F. Warren, MDa, David M. Dines, MDa, Lawrence V. Gulotta, MDa a

Sports Medicine and Shoulder Group, Hospital for Special Surgery, New York, NY, USA Department of Orthopaedic Surgery, University of Virginia Health System, Charlottesville, VA, USA c Epidemiology and Biostatistics, Hospital for Special Surgery, New York, NY, USA d Department of Orthopaedics and Sports Medicine, University of Cincinnati Academic Health Center, Cincinnati, OH, USA b

Background: Although anatomic total shoulder arthroplasty (TSA) successfully improves pain and function, not all patients improve clinically. This study was conducted to determine patient-related factors for failure to achieve improvement after primary TSA for osteoarthritis at 2 years postoperatively. Methods: This prospective study reviewed an institutional shoulder registry for consecutive patients who underwent primary TSA for osteoarthritis from 2007 to 2013 with baseline and 2-year postoperative American Shoulder and Elbow Surgeons (ASES) Standardized Shoulder Assessment Form scores. A failed outcome was defined as (1) a failure to reach the ASES minimal clinically important difference of 16.1 points or (2) revision surgery within 2 years of the index procedure, or both. Univariate and multivariable analyses of clinical and demographic patient factors were performed using logistic regression. Results: Of 459 arthroplasties that met inclusion criteria, 411 were deemed successful by the aforementioned criteria, and 48 (10.5%) failed to achieve a desirable outcome. Clinical risk factors associated with failure included previous surgery to the shoulder (P = .047), presence of a torn rotator cuff (P = .025), and presence of diabetes (P = .036), after adjusting for age, sex, race, and body mass index. A higher preoperative ASES score at baseline was associated with failure (P < .001). Conclusion: Previous shoulder surgery, a rotator cuff tear requiring repair during TSA, presence of diabetes, surgery on the nondominant arm, and a higher baseline ASES score were associated with a higher risk of failing to achieve improvement after anatomic TSA. Level of evidence: Level II; Prospective Cohort Design; Treatment Study © 2017 Journal of Shoulder and Elbow Surgery Board of Trustees. All rights reserved. Keywords: Total shoulder arthoplasty; glenohumeral arthritis; risk factors; MCID; postoperative outcomes; satisfaction; poor improvement; ASES score

The Hospital for Special Surgery Institutional Review Board approved this study (IRB #2013-014). *Reprint requests: Brian M. Grawe, MD, University of Cincinnati Academic Health Center - Orthopaedics and Sports Medicine, PO Box 670212, Cincinnati, OH 45267, USA. E-mail address: [email protected] (B.M. Grawe). 1058-2746/$ - see front matter © 2017 Journal of Shoulder and Elbow Surgery Board of Trustees. All rights reserved. https://doi.org/10.1016/j.jse.2017.12.018

ARTICLE IN PRESS 2 Anatomic total shoulder arthroplasty (TSA) has demonstrated an excellent clinical track record, with improvements in patient-reported pain relief, shoulder function, range of motion, and quality of life indices.25,27,28,33 Most large shoulder arthroplasty series have focused on implant survival and included patients who underwent TSA or hemiarthroplasty for their shoulder condition.16,28 Despite generally good results, a subset of patients experience suboptimal results. Data regarding patient-specific or surgical elements that may portend poor clinical outcomes after TSA are limited.2,4,16 As the landscape of health care economics evolves to prioritize cost-efficient and effective care, providers are incentivized to select patients who are most likely to experience significant functional improvement. Furthermore, as the Centers for Medicare and Medicaid Services investigates shifting financial risk to providers via bundled payment and outcomes-based reimbursement models, care providers must be able to identify which patients pose the highest risk for costly complications, revisions, and adverse outcomes.6,29,36 Understanding such risk factors is also valuable preoperatively to effectively manage surgeon and patient expectations. The purpose of this study was to determine patient-related factors for failure to achieve improvement after primary TSA for osteoarthritis at a minimum of 2 years postoperatively.

Materials and methods Data collection Prospective data from 2007 to 2013 were collected as part of an institutional shoulder arthroplasty registry. The registry collects preoperative, intraoperative, and postoperative information regarding shoulder arthroplasties performed at the Hospital for Special Surgery. Inclusion criteria consisted of patients who underwent a primary anatomic shoulder replacement for a diagnosis of osteoarthritis and for whom presurgical baseline data and 2-year followup data were available. The analysis excluded TSAs for a diagnosis other than osteoarthritis. Data collection at both points of interest was standardized and accomplished by mailed questionnaires and a web-based system with an interface that allowed patients to enter information directly online.

Defining failure to improve Functional outcome after TSA was assessed using the American Shoulder and Elbow Surgeons (ASES) Standardized Shoulder Assessment Form score.35 This patient-reported outcome is graded on a scale of 0 (worst) to 100 (best) and consists of questions concerning pain, function, and stability in regards to the affected shoulder. The ASES has demonstrated adequate responsiveness for patients undergoing TSA and has shown excellent validity and reliability, with minimal administrator and responder burden.1,38 ASES improvement has been shown to plateau at 2 years postoperatively in the TSA population, indicating that at our 2-year assessment, patients will have achieved close to maximum functional improvement.34

G.T. Mahony et al. To assess poor postoperative improvement after TSA, a definition of “failure to improve” was established. First, any patient who self-reported a revision arthroplasty or an additional surgical intervention to the joint within the 2-year follow-up period was defined as a patient who failed to improve after the index procedure.45 Our second definition of failure to improve was modeled after the minimal clinically important difference (MCID) of the ASES score reported by Werner et al45 for the anatomic TSA population. Werner et al45 used a method that anchored the patients’ 2-year change in ASES scores to their postoperative satisfaction responses. We selected from their study the reported MCID of 16.1-point improvement as our cutoff. We therefore defined failure to improve as those patients who underwent a subsequent operation within 2 years or failed to achieve an MCID of a 16.1-point improvement on the ASES questionnaire, or both.

Patient satisfaction As a secondary outcome, patient satisfaction with the surgery was assessed against the definition of failure to improve. Patient overall satisfaction with surgery, with pain relief, with ability to return to work, and with ability to return to recreational activities were assessed by a 5-point Likert scale that varied from 1 (very satisfied) to 5 (very dissatisfied). Satisfaction with the operation’s ability to improve quality of life was assessed by a 6-point Likert scale that ranged from 1 (more improvement than ever dreamed possible) to 6 (quality of life is worse). To convert the Likert scales to binary variables, patients who rated the procedure as a 1 or a 2 for a given satisfaction metric were considered to be “satisfied,” and patients who rated the procedure lower than 2 were considered to be “dissatisfied.” Satisfaction between patients who improved vs. those who failed to improve was then assessed using χ2 tests.

Patient-specific factors Multiple patient-specific factors were assessed against the outcome of interest, including demographic variables, social history variables, and medical history elements. Sex, age at the time of surgery, body mass index (BMI), race, education level, and living arrangement composed the primary demographic and social variables. The medical comorbidities considered were the presence of heart disease, hypertension, lung disease, diabetes, ulcer/stomach disease, kidney disease, liver disease, anemia or other blood diseases, cancer, depression, osteoarthritis (in a joint other than the one being operated on), back pain, rheumatoid arthritis, and an open write-in field that was assessed by a surgeon. The sum of medical comorbidities for a given patient was also included to assess the effect of the total comorbidity burden on outcome. Also assessed were the American Society of Anesthesiologists (ASA) Physical Status Classification, previous use of bisphosphonates, use of narcotic pain medication preoperatively, whether the operation was performed on the patient’s dominant arm, and history of surgery to the affected joint. The preoperative clinical assessments analyzed included the Mental Component Score of the 12-Item Short Form Health Survey questionnaire, a visual analog scale (VAS) pain score, a VAS instability score, and the patient’s score on the Marx Activity Scale.5

ARTICLE IN PRESS Failure to improve after TSA

Surgical technique and surgical factors The most commonly used implant system in this series was the Biomet Comprehensive (Warsaw, IN, USA). A standard deltopectoral approach was used for access to the humeral head and glenoid in all procedures. Surgical factors assessed for this study included intraoperative complications, management of biceps tendon, subscapularis approach and subscapularis repair, rotator cuff status (intact, attenuated, or torn), whether a torn rotator cuff was repaired, and the Walch classification.43

Statistical analysis Univariate analyses of patient-related factors, including demographic and clinical characteristics, were performed using logistic regression with the single risk factor of interest. Patient-related factors that were univariately associated with the outcome at the 0.10 significance level were considered for inclusion when modeling the multivariable model. The multivariable logistic regression of patient-related factors that influence failure was fitted with clinical factors with a significance level of 0.05 included in the final model, adjusting for age, sex, BMI, and race. Patient satisfaction with surgery was compared between groups using χ2 or Fisher exact tests. Analyses were conducted using SAS 9.3 software (SAS Institute Inc., Cary, NC, USA). A power analysis for a multivariable logistic regression was completed using RStudio 0.99.484 software (Boston, MA, USA). With a small effect size of .05 and 10 total variables to include in the model, the study would require 405 patients to reach a power of 80%.

Results A total of 699 TSAs were performed during the study window that met inclusion criteria; of these, 459 (441 patients) responded to follow-up at an average of 2.1 years (range 1.62.8 years). The average age at the time of surgery was 66.8 ± 9.3 (standard deviation) years (range, 25-9 years). There was a relatively equal distribution of men and women, at 55.3% (n = 254) and 44.7% (n = 205), respectively. The right shoulder was more likely to be replaced, with only 41.4% (n = 190) undergoing left-sided surgery. According to arm dominance, 51% (n = 234) of the TSAs were to the patient’s dominant arm, 37.9% (n = 174) were to the nondominant arm, and arm dominance was not reported for 11% (n = 51). Most patients were moderately healthy: 368 patients (80.2%) carried an ASA status of I or II, and the average BMI for this cohort was 28.6 ± 5.6 kg/m2 (range, 17.554.5 kg/m2). By race, 96.1% of patients (n = 441) were white, non-Hispanic. Additional baseline demographics and clinical characteristics of the study population are summarized in Table I. The operation was classified as a failure in 41 patients (8.9%) who failed to achieve an MCID of a 16.1-point improvement on the ASES questionnaire. Of the 41 patients who failed to achieve a MCID, 8 (19.5%) also self-reported at follow-up that they required additional surgery to the joint of interest within the 2-year follow-up window. In addition, 7 patients who achieved the MCID at 2 years were also

3 Table I Demographic and clinical characteristics of study population Characteristics

No.

%

Patients Demographic characteristics Sex Male Female Age, yr Body mass index, kg/m2 Race White, non-Hispanic Nonwhite Missing Preoperative clinical characteristics ASES score Mental Component Score Marx Shoulder Activity score VAS pain score VAS instability score Intraoperative clinical characteristics ASA status I II III Missing Surgery on dominant arm No Yes Missing Revisions Primary

459

100.0

254 205 459 458

55.3 44.7 100.0 99.8

441 14 4

96.1 3.1 0.9

459 315 442 459 422

100.0 68.6 96.3 100.0 91.9

16 352 74 17

3.5 76.7 16.1 3.7

174 234 51

37.9 51.0 11.1

459

100.0

Mean

SD

66.8 28.6

9.3 5.6

38.06 52.31 7.49 5.42 3.73

18.08 11.75 4.92 2.64 3.11

SD, standard deviation; ASES, American Shoulder and Elbow Surgeons Standardized Shoulder Assessment; VAS, visual analog scale; ASA, American Society of Anesthesiologists Physical Status Classification.

classified as failures because they self-reported an additional surgery to the joint of interest within the follow-up window. Therefore, 48 arthroplasties (10.5%) were assigned to the failure-to-improve group for this study, and 411 (89.5%) were categorized as having successful outcomes. After the initial univariate analysis (Table II), multivariate logistic regression revealed that the following clinical risk factors were associated with failure to improve: previous surgery to shoulder (P = .047), presence of a torn rotator cuff at time of arthroplasty (P = .025), and the presence of diabetes (P = .036). Surgery on the dominant arm was associated with a lower risk of failing to improve (P = .007). A higher preoperative ASES score at baseline was associated with a greater risk of failing to improve (P < .001; Table III). None of the other patient-specific and surgeon-specific factors affected the outcome of interest in a statistically significant fashion. For the group that failed to improve, the average VAS pain scores were 4.19 ± 2.75 preoperatively and 4.20 ± 3.10

ARTICLE IN PRESS 4

G.T. Mahony et al. Table II Univariate analyses of patient-related factors associated with failure to achieve a minimal clinically important difference or require later revision within 2 years Variable

Outcome: failure* Odds ratio

Preoperative scores ASES Pain visual analog scale Previous surgery vs. none Rotator cuff status (ref: intact) Attenuated Torn Rotator cuff repair performed vs. not Comorbidities Diabetes Back pain Surgery on dominant arm vs. not Biceps (ref: not performed because already torn or still intact) Tenotomy Tenodesis

95% CI

P value

Lower

Upper

1.04 0.82 3.34

1.02 0.73 1.33

1.05 0.92 8.37

<.001 .001 .010

1.86 4.66 3.69

0.40 1.35 1.11

8.78 16.11 12.25

.432 .015 .033

3.27 2.09 0.52

1.23 1.12 0.27

8.74 3.88 0.99

.018 .020 .048

0.40 0.32

0.17 0.13

0.93 0.79

.004 .014

CI, confidence interval; ASES, American Shoulder and Elbow Surgeons Standardized Shoulder Assessment. * Defined as failure to achieve a minimal clinically important difference or require revision within 2 years.

Table III

Multivariable logistic regression model of patient-related factors on failure

Variable

Outcome: failure* Odds ratio

Age at surgery Female vs. male Body mass index Nonwhite vs. white Preoperative ASES score Surgery on dominant arm vs. not Previous surgery vs. none Rotator cuff status (ref: intact) Attenuated Torn Comorbidities Diabetes vs. no diabetes Back pain vs. no back pain

95% CI

P value

Lower

Upper

1.00 1.42 1.01 2.55 1.05 0.33 3.47

0.95 0.61 0.93 0.42 1.03 0.14 0.96

1.04 3.32 1.08 11.62 1.08 0.73 11.44

.890 .418 .887 .256 <.001 .007 .047

2.02 6.06

0.23 1.14

12.01 28.62

.471 .025

4.15 1.97

1.00 0.90

15.06 4.34

.036 .089

CI, confidence interval; ASES, American Shoulder and Elbow Surgeons Standardized Shoulder Assessment. * Defined as failure to achieve a minimal clinically important difference or require revision within 2 years.

at 2 years on a maximum pain level of 10. Average VAS pain scores for the group that improved after arthroplasty were 5.56 ± 2.59 at baseline and 0.66 ± 1.27 at 2 years. The group that failed to improve postoperatively also reported greater dissatisfaction with the operation across all satisfaction metrics assessed (P < .001 for all metrics; Table IV). For the failure group, the 15 patients who underwent a subsequent surgery differed little from the 34 who did not require subsequent surgery. No significant differences were found in age, sex, BMI, presence of diabetes, presence of back pain,

ASA status, previous operations, or rotator cuff status. The group that underwent subsequent surgery had a lower baseline ASES score (Table V).

Discussion Most patients (89.5%) achieve a successful outcome at 2 years postoperatively after primary anatomic TSA for osteoarthritis. We found that risk factors for failure (revision surgery or failure to achieve the ASES MCID) were surgery on the nondominant

ARTICLE IN PRESS Failure to improve after TSA

5

Table IV Results of satisfaction survey at 2 years postoperatively Variable Overall satisfaction with results of surgery Very satisfied Somewhat satisfied Neither satisfied nor dissatisfied Somewhat dissatisfied Very dissatisfied Satisfied* Not satisfied Pain Very satisfied Somewhat satisfied Neither satisfied nor dissatisfied Somewhat dissatisfied Very dissatisfied Satisfied* Not satisfied Work Very satisfied Somewhat satisfied Neither satisfied nor dissatisfied Somewhat dissatisfied Very dissatisfied Satisfied* Not satisfied Activities Very satisfied Somewhat satisfied Neither satisfied nor dissatisfied Somewhat dissatisfied Very dissatisfied Satisfied* Not satisfied Quality of life More improvement than I ever dreamed possible A great improvement A moderate improvement A little improvement No improvement at all The quality of my life is worse Satisfied* Not satisfied

Failure Nonfailure P value (No.) (No.)

13 13 1

370 30 1

7 13 26 21

2 3 400 6

18 9 1

383 19 0

7 12 47 27

3 1 402 4

13 12 3

357 32 12

8 10 25 21

5 1 389 18

10 14 3

325 52 18

9 10 24 22

7 4 377 29

3

147

14 10 8 4 7 17 22

237 20 4 1 0 384 25

<.001

<.001

<.001

<.001

<.001

<.001

<.001

<.001

<.001

<.001

* Satisfied includes “very satisfied” and “somewhat satisfied,” and “not satisfied” is all others.

arm, previous shoulder surgery, presence of a torn rotator cuff, diabetes, and a higher baseline ASES score. Overall, shoulder arthroplasty has consistently provided clinical success for a variety of degenerative conditions of

the glenohumeral joint. Excellent pain relief, good survival, and symptomatic objective patient improvement have all been noted in multiple series.3,10,16,19,28,33,42 Previous studies have focused on implant survival, included a significant number of humeral head replacement operations in the overall series, or used nonvalidated objective clinical scores.10,16,28,42 Demonstrating value in surgical care for pathologic conditions is becoming increasingly important, and this study confirms that 89.5% of patients undergoing primary TSA for osteoarthritis will experience a clinically significant gain in subjective functional outcome without the need for additional surgery. Our results corroborate previous studies that investigated outcomes after total shoulder replacement. Raiss et al33 recently demonstrated an average improvement of 34 points, as measured by the Constant score, in 46 patients. In separate studies, Tammachote et al42 and Deshmukh et al10 reported the longterm results of cemented TSA. Tammachote et al demonstrated 78% excellent-to-satisfactory results by the rating criteria of Neer, whereas Deshmukh et al showed a mean final Disabilities of the Arm. Shoulder and Hand score of 49 points.10,42 Neither of these studies examined preoperative patient scores and, therefore, were unable to demonstrate an overall improvement. Furthermore, the respective scoring systems used in these investigations remain nonvalidated (Neer), or nonjoint specific (Disabilities of the Arm, Shoulder and Hand). A growing body of literature confirms that certain patientspecific factors affect the overall outcome after shoulder arthroplasty. Recent studies performed by Ponce et al32 and Bot et al4 demonstrated that diabetes and the presence of a psychiatric comorbidity are associated with a worse shortterm result after TSA; however, neither study assessed clinical outcomes. Our findings are in line with these results, because diabetes was also adversely related to outcome in our population. However, mental health status (as assessed by the Mental Component Score of the 12-Item Short Form Health Survey and by the self-reported presence of depression at baseline) failed to statistically affect outcome. Preoperative rotator cuff pathology adversely affects outcomes after TSA.37 Typically, patients with irreparable rotator cuff tears are indicated for reverse TSA rather than TSA.44,47 Although the presence of full-thickness rotator cuff tears is a contraindication for TSA, a repairable rotator cuff tear does not necessarily exclude TSA as a treatment option, because TSR can be justified if the quality of the cuff is sufficient to maintain movement and stabilize the humeral implant.13 In our study, 4 patients (8.3%) of the failure group had a rotator cuff tear to the supraspinatus at the time of arthroplasty. Comparatively, 8 patients (2.0%) of the nonfailure group had rotator cuff tears; of whom 2 patients had tears to both the infraspinatus and supraspinatus, 2 had tears to both the subscapularis and supraspinatus, and the 4 remaining patients had tears to only the supraspinatus. All tears identified during arthroplasty were repaired with suture anchor fixation. Norris et al28 previously reported that 6.0% of patients undergoing TSA had repairable tears to their supraspinatus. Similarly, a systematic review of medical literature by Levy et al20 found a 6.4%

ARTICLE IN PRESS 6

G.T. Mahony et al. Table V

Comparison of failure group: those requiring additional surgery vs. no additional surgery

Variables

Surgery group

Nonsurgery group

Continuous Age at surgery, yr Body mass index, kg/m2 Baseline ASES score Categoric Female Surgery on dominant arm ASA of I or II Previous surgery Torn rotator cuff Diabetes Back pain vs. no back pain

Mean ± SD 61.3 ± 12.4 28.0 ± 4.1 40.79 ± 22.3 No. (%) 5 (33.3) 6 (40.0) 13 (92.9) 2 (13.3) 1 (6.7) 2 (13.3) 8 (53.3)

Mean ± SD 66.9 ± 11.7 28.8 ± 5.6 51.22 ± 15.4 No. (%) 15 (45.5) 12 (36.4) 22 (73.3) 5 (15.2) 3 (9.1) 8 (24.2) 21 (63.6)

P value .119 .992 .021* .299 .286 .213 .551 .328 .692 .980

SD, standard deviation; ASES, American Shoulder and Elbow Surgeons Standardized Shoulder Assessment; ASA, American Society of Anesthesiologists Physical Status Classification. * Significant at P < .05.

rate of supraspinatus tears identified at the time of TSA. Both rates are comparable to our findings. Consideration of the rotator cuff at the time of arthroplasty is important, primarily because the prosthetic joint relies on the rotator cuff for both stability and strength.26,44 Furthermore, rotator cuff tears complicate the TSA procedure and also require modifications and limitations to the postoperative rehabilitation protocol.7,44 Previous research by Iannotti et al16 showed that in the osteoarthritic population, repairable rotator cuff tears that were isolated to the supraspinatus tendon did not affect the postoperative ASES score. However, that study included hemiarthroplasty and TSA patients.16 Additional research is needed to further understand the effect of preoperative rotator cuff tears on TSA outcomes. Particular consideration should be given to the quality of the rotator cuff tendons at the time of TSA. In this study, 14.5% (n = 7) of patients who failed to improve after their TSA had undergone a surgical procedure to the joint of interest before the arthroplasty. Comparatively, 4.8% (n = 20) of patients that had successful outcomes underwent prior surgery (P = .047). Among the patients who failed to improve after TSA, 3 patients had arthroscopic rotator cuff repairs and 1 underwent a previous open rotator cuff repair. One patient had previous open shoulder stabilization surgery, and the subscapularis in this patient was intact at the time of index procedure. One had open reduction and internal fixation of a fracture (of which there was no malunion at time of index TSA procedure), and 1 patient underwent removal of a humeral lipoma. Interestingly, none of the patients in the failure-to-improve group who underwent rotator cuff repair surgery before the arthroplasty had significant rotator cuff pathology at the time of the TSA. Previous surgery to the joint of interest has been implicated in poor outcomes after arthroplasty.2,11,15,23,26,39 Because soft tissue and rotator cuff integrity are important stabilizers of the prosthetic humeral head, prior operations to the joint of interest may be linked to poor arthroplasty outcomes.12,28

Although this study indicates that hand dominance could affect outcome after TSA, the data on this topic are inconsistent. Cvetanovich et al9 analyzed postoperative TSA patients and found that at a minimum follow-up of 1 year, patients who underwent surgery on their dominant extremity had greater range of motion. They also reported that hand dominance did not affect validated outcome scores at the same time point, but their analysis was underpowered.9 In 63 patients at a follow-up of 15 to 20 years, Raiss et al33 found that hand dominance was not associated with clinical outcomes in patients who had not undergone surgical revision. Some radiographic analyses have shown hand dominance increases the presence of radiolucent lines at follow-up, whereas others have shown that, at least with regards to radiographic outcomes, hand dominance is inconsequential.8,31,40,41 We also note that a high baseline ASES score correlated with poor postoperative outcome in this study. Past studies corroborate that higher function at baseline can predict less dramatic improvements postoperatively.18,21,46,48 To further evaluate patient outcomes, we also calculated the percentage of the maximal possible improvement a patient attained on the ASES. This methodology was previously reported by Matsen et al23,30 for the shoulder arthroplasty population. Three of the 41 patients who failed to reach the MCID of the ASES questionnaire did attain a high percentage of maximal possible improvement on their ASES scores. With high starting ASES scores of 82.50, 81.33, and 78.33 (of a maximum score of 100), these patients ultimately achieved 89.6%, 81.6%, and 68.1% of maximal score improvement without the need for an additional operation and were satisfied with the outcome of their arthroplasty. Patients with a high baseline ASES score experience a ceiling affect: they have a smaller window for potential improvement, making it more difficult to achieve an MCID. Our study methodology ultimately classified these 3 operations as failures, but they could have been classified as successes using alternative, previously reported, methods. Using the MCID is often considered the norm for reporting

ARTICLE IN PRESS Failure to improve after TSA surgical outcomes; however, this discrepancy underscores the need for further research into the best way of categorizing successful arthroplasties. To further evaluate successful outcomes, we investigated the relationship between failure to improve and patient satisfaction. Evaluating postoperative satisfaction with the surgery in tandem with other outcome measures is becoming increasingly important because satisfaction is often included as a factor in bundled payment and outcome-based reimbursement models. Jacobs et al18 showed that lower ASES score improvement is correlated with patient dissatisfaction. Petri et al30 reported that demographic, anatomic, and surgical variables do not affect satisfaction but also showed that failure to improve after surgery was significantly associated with lower satisfaction. In our study, patients in the failure group were more likely to report overall dissatisfaction with their index procedure and dissatisfaction with the surgery’s ability to improve their pain, their ability to return to work, their ability to participate in recreational activities, and ability to improve overall quality of life. Interestingly, the current study did not find any association between a number of factors that have previously been shown to affect shoulder arthroplasty outcomes. Studies have indicated that demographic variables, such as advanced age, female sex, and race, are all associated with poor outcomes; however, these factors were not significant in our population.14,22 Other studies have shown that the medical comorbidity burden is implicated in postoperative complications; however, we found that the total number of complications in a patient was not predictive of a poor outcome.17,46 Use of opioids has been shown in other studies to be a significant predictor of outcome but was insignificant in our population.18,24 None of the surgical technique variables assessed—biceps management, subscapularis approach, subscapularis repair, and occurrence of intraoperative complications—were associated with adverse results at 2 years. Finally, neither of the social factors considered (living arrangement and education level) achieved significance in our study. The data presented here are not without weaknesses. Our follow-up does not include long-term data, which might be expected to affect outcomes and need for subsequent surgery. Furthermore, we only present minimal objective data; range of motion, radiographic, and strength values were not assessed. Although our study population is large, the results may be significantly affected by omitted variable bias or confounding factors, or both. As such, the correlations presented in this study should not be interpreted as causative.

Conclusion Primary TSA confers excellent short-term gains in subjective patient-reported shoulder function for degenerative osteoarthritis of the glenohumeral joint. Factors such as

7 high preoperative shoulder function, diabetes, prior surgical history to the joint of interest, arm dominance, and rotator cuff pathology are associated with failure to improve after TSA.

Disclaimer Lawrence V. Gulotta receives compensation in the form of consulting and speaking payments from Biomet Inc., which is related to the subject of this work. Edward V. Craig, Russell F. Warren, and David M. Dines received royalties for intellectual property from Biomet Inc., which is related to the subject of this work. 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.

References 1. Angst F, Goldhahn J, Drerup S, Aeschlimann A, Schwyzer HK, Simmen BR. Responsiveness of six outcome assessment instruments in total shoulder arthroplasty. Arthritis Rheum 2008;59:391-8. http://dx.doi.org/ 10.1002/art.23318 2. Bigliani LU, Weinstein DM, Glasgow MT, Pollock RG, Flatow EL. Glenohumeral arthroplasty for arthritis after instability surgery. J Shoulder Elbow Surg 1995;4:87-94. 3. Boileau P, Avidor C, Krishnan SG, Walch G, Kempf JF, Molé D. Cemented polyethylene versus uncemented metal-backed glenoid components in total shoulder arthroplasty: a prospective, doubleblind, randomized study. J Shoulder Elbow Surg 2002;11:351-9. http://dx.doi.org/10.1067/mse.2002.125807 4. Bot AG, Menendez ME, Neuhaus V, Ring D. The influence of psychiatric comorbidity on perioperative outcomes after shoulder arthroplasty. J Shoulder Elbow Surg 2014;23:519-27. http://dx.doi.org/ 10.1016/j.jse.2013.12.006 5. Brophy RH, Beauvais RL, Jones EC, Cordasco FA, Marx RG. Measurement of shoulder activity level. Clin Orthop Relat Res 2005;439:101-8. 6. Burwell SM. Setting value-based payment goals–HHS efforts to improve U.S. health care. N Engl J Med 2015;372:897-9. http://dx.doi.org/ 10.1056/NEJMp1500445 7. Cofield RH. Shoulder arthrodesis and resection arthroplasty. Instr Course Lect 1985;34:268-77. 8. Collin P, Tay AK, Melis B, Boileau P, Walch G. A ten-year radiologic comparison of two-all polyethylene glenoid component designs: a prospective trial. J Shoulder Elbow Surg 2011;20:1217-23. http://dx.doi.org/10.1016/j.jse.2011.06.012 9. Cvetanovich GL, Chalmers PN, Streit JJ, Romeo AA, Nicholson GP. Patients undergoing total shoulder arthroplasty on the dominant extremity attain greater postoperative ROM. Clin Orthop Relat Res 2015;473:3221-5. http://dx.doi.org/10.1007/s11999-015-4400-0 10. Deshmukh AV, Koris M, Zurakowski D, Thornhill TS. Total shoulder arthroplasty: long-term survivorship, functional outcome, and quality of life. J Shoulder Elbow Surg 2005;14:471-9. http://dx.doi.org/ 10.1016/j.jse.2005.02.009 11. Donigan JA, Frisella WA, Haase D, Dolan L, Wolf B. Pre-operative and intra-operative factors related to shoulder arthroplasty outcomes. Iowa Orthop J 2009;29:60-6. 12. Edwards TB, Kadakia NR, Boulahia A, Kempf JF, Boileau P, Némoz C, et al. A comparison of hemiarthroplasty and total shoulder arthroplasty

ARTICLE IN PRESS 8

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25. 26. 27. 28.

29.

30.

G.T. Mahony et al. in the treatment of primary glenohumeral osteoarthritis: results of a multicenter study. J Shoulder Elbow Surg 2003;12:207-13. http://dx.doi.org/10.1016/S1058-2746(02)86804-5 Eichinger JK, Galvin JW. Management of complications after total shoulder arthroplasty. Curr Rev Musculoskelet Med 2015;8:83-91. http://dx.doi.org/10.1007/s12178-014-9251-x Griffin JW, Hadeed MM, Novicoff WM, Browne JA, Brockmeier SF. Patient age is a factor in early outcomes after shoulder arthroplasty. J Shoulder Elbow Surg 2014;23:1867-71. http://dx.doi.org/10.1016/ j.jse.2014.04.004 Hasan SS, Leith JM, Campbell B, Kapil R, Smith KL, Matsen FA 3rd. Characteristics of unsatisfactory shoulder arthroplasties. J Shoulder Elbow Surg 2002;11:431-41. http://dx.doi.org/10.1067/mse.2002 .125806 Iannotti JP, Norris TR. Influence of preoperative factors on outcome of shoulder arthroplasty for glenohumeral osteoarthritis. J Bone Joint Surg Am 2003;85-A:251-8. Issa K, Pierce CM, Pierce TP, Boylan MR, Zikria BA, Naziri Q, et al. Total shoulder arthroplasty demographics, incidence, and complications-A Nationwide inpatient sample database study. Surg Technol Int 2016; 29:240-6. Jacobs CA, Morris BJ, Sciascia AD, Edwards TB. Comparison of satisfied and dissatisfied patients 2 to 5 years after anatomic total shoulder arthroplasty. J Shoulder Elbow Surg 2016;25:1128-32. http://dx.doi.org/ 10.1016/j.jse.2015.12.001 Khan A, Bunker TD, Kitson JB. Clinical and radiological follow-up of the Aequalis third-generation cemented total shoulder replacement: a minimum ten-year study. J Bone Joint Surg Br 2009;91:1594-600. http://dx.doi.org/10.1302/0301-620X.91B12.22139 Levy DM, Abrams GD, Harris JD, Bach BR Jr, Nicholson GP, Romeo AA. Rotator cuff tears after total shoulder arthroplasty in primary osteoarthritis: a systematic review. Int J Shoulder Surg 2016;10:78-84. http://dx.doi.org/10.4103/0973-6042.180720 Maier MW, Lauer S, Wolf SI, Dreher T, Klotz MC, Zeifang F, et al. Low preoperative Constant score is a negative predictive factor for postoperative proprioception after total shoulder arthroplasty in osteoarthritis. Arch Orthop Trauma Surg 2015;135:171-7. http://dx.doi.org/10.1007/s00402-014-2148-8 Matsen FA 3rd, Li N, Gao H, Yuan S, Russ SM, Sampson PD. Factors affecting length of stay, readmission, and revision after shoulder arthroplasty: a population-based study. J Bone Joint Surg Am 2015;97:1255-63. http://dx.doi.org/10.2106/JBJS.N.01107 Matsen FA 3rd, Russ SM, Vu PT, Lucas RM, Comstock BA. What factors are predictive of patient-reported outcomes? A prospective study of 337 shoulder arthroplasties. Clin Orthop Relat Res 2016;474:2496-510. http://dx.doi.org/10.1007/s11999-016-4990-1 Morris BJ, Sciascia AD, Jacobs CA, Edwards TB. Preoperative opioid use associated with worse outcomes after anatomic shoulder arthroplasty. J Shoulder Elbow Surg 2016;25:619-23. http://dx.doi.org/10.1016/ j.jse.2015.09.017 Neer CS 2nd. Articular replacement for the humeral head. J Bone Joint Surg Am 1955;37-a(2):215-28. Neer CS 2nd, Kirby RM. Revision of humeral head and total shoulder arthroplasties. Clin Orthop Relat Res 1982;(170):189-95. Neer CS 2nd, Watson KC, Stanton FJ. Recent experience in total shoulder replacement. J Bone Joint Surg Am 1982;64:319-37. Norris TR, Iannotti JP. Functional outcome after shoulder arthroplasty for primary osteoarthritis: a multicenter study. J Shoulder Elbow Surg 2002;11:130-5. http://dx.doi.org/10.1067/mse.2002.121146 Oberlander J, Laugesen MJ. Leap of faith–Medicare’s new physician payment system. N Engl J Med 2015;373:1185-7. http://dx.doi.org/ 10.1056/NEJMp1509154 Petri M, Euler SA, Dornan GJ, Greenspoon JA, Horan MP, Katthagen JC, et al. Predictors for satisfaction after anatomic total shoulder arthroplasty for idiopathic glenohumeral osteoarthritis. Arch Orthop Trauma Surg 2016;136:755-62. http://dx.doi.org/10.1007/s00402 -016-2452-6

31. Pfahler M, Jena F, Neyton L, Sirveaux F, Molé D. Hemiarthroplasty versus total shoulder prosthesis: results of cemented glenoid components. J Shoulder Elbow Surg 2006;15:154-63. http://dx.doi.org/10.1016/ j.jse.2005.07.007 32. Ponce BA, Menendez ME, Oladeji LO, Soldado F. Diabetes as a risk factor for poorer early postoperative outcomes after shoulder arthroplasty. J Shoulder Elbow Surg 2014;23:671-8. http://dx.doi.org/10.1016/ j.jse.2014.01.046 33. Raiss P, Bruckner T, Rickert M, Walch G. Longitudinal observational study of total shoulder replacements with cement: fifteen to twenty-year follow-up. J Bone Joint Surg Am 2014;96:198-205. http://dx.doi.org/ 10.2106/JBJS.M.00079 34. Razmjou H, Stratford P, Kennedy D, Holtby R. Pattern of recovery following total shoulder arthroplasty and humeral head replacement. BMC Musculoskelet Disord 2014;15:306. http://dx.doi.org/10.1186/ 1471-2474-15-306 35. Richards RR, An KN, Bigliani LU, Friedman RJ, Gartsman GM, Gristina AG, et al. A standardized method for the assessment of shoulder function. J Shoulder Elbow Surg 1994;3:347-52. 36. Rosenthal MB. Physician payment after the SGR–the new meritocracy. N Engl J Med 2015;373:1187-9. http://dx.doi.org/10.1056/NEJMp1507757 37. Rozing PM, Nagels J, Rozing MP. Prognostic factors in arthroplasty in the rheumatoid shoulder. HSS J 2011;7:29-36. http://dx.doi.org/ 10.1007/s11420-010-9172-1 38. Schmidt S, Ferrer M, González M, González N, Valderas JM, Alonso J. Evaluation of shoulder-specific patient-reported outcome measures: a systematic and standardized comparison of available evidence. J Shoulder Elbow Surg 2014;23:434-44. http://dx.doi.org/10.1016/ j.jse.2013.09.029 39. Simmen BR, Bachmann LM, Drerup S, Schwyzer HK, Burkhart A, Goldhahn J. Development of a predictive model for estimating the probability of treatment success one year after total shoulder replacement—cohort study. Osteoarthritis Cartilage 2008;16:631-4. http://dx.doi.org/10.1016/j.joca.2007.10.010 40. Sperling JW, Cofield RH, O’Driscoll SW, Torchia ME, Rowland CM. Radiographic assessment of ingrowth total shoulder arthroplasty. J Shoulder Elbow Surg 2000;9:507-13. 41. Szabo I, Buscayret F, Edwards TB, Nemoz C, Boileau P, Walch G. Radiographic comparison of flat-back and convex-back glenoid components in total shoulder arthroplasty. J Shoulder Elbow Surg 2005;14:636-42. http://dx.doi.org/10.1016/j.jse.2005.05.004 42. Tammachote N, Sperling JW, Vathana T, Cofield RH, Harmsen WS, Schleck CD. Long-term results of cemented metal-backed glenoid components for osteoarthritis of the shoulder. J Bone Joint Surg Am 2009;91:160-6. http://dx.doi.org/10.2106/JBJS.F.01613 43. Walch G, Badet R, Boulahia A, Khoury A. Morphologic study of the glenoid in primary glenohumeral osteoarthritis. J Arthroplasty 1999;14:756-60. 44. Waldman BJ, Figgie MP. Indications, technique, and results of total shoulder arthroplasty in rheumatoid arthritis. Orthop Clin North Am 1998;29:435-44. 45. Werner BC, Chang B, Nguyen JT, Dines DM, Gulotta LV. What change in American Shoulder and Elbow Surgeons score represents a clinically important change after shoulder arthroplasty? Clin Orthop Relat Res 2016;474:2672-81. http://dx.doi.org/10.1007/s11999-016-4968-z 46. Werner BC, Wong AC, Mahony GT, Craig EV, Dines DM, Warren RF, et al. Causes of poor postoperative improvement after reverse total shoulder arthroplasty. J Shoulder Elbow Surg 2016;25:e217-22. http://dx.doi.org/10.1016/j.jse.2016.01.002 47. Westermann RW, Pugely AJ, Martin CT, Gao Y, Wolf BR, Hettrich CM. Reverse shoulder arthroplasty in the United States: a comparison of national volume, patient demographics, complications, and surgical indications. Iowa Orthop J 2015;35:1-7. 48. Wong SE, Zhang AL, Berliner JL, Ma CB, Feeley BT. Preoperative patient-reported scores can predict postoperative outcomes after shoulder arthroplasty. J Shoulder Elbow Surg 2016;25:913-9. http://dx.doi.org/ 10.1016/j.jse.2016.01.029