Outcome of Revision Shoulder Arthroplasty After Resurfacing Hemiarthroplasty in Patients with Glenohumeral Osteoarthritis

Rasmussen, Jeppe V. MD, PhD; Olsen, Bo S. MD, PhD; Al-Hamdani, Ali MD; Brorson, Stig MD, PhD, DMSc

Journal of Bone & Joint Surgery - American Volume:
doi: 10.2106/JBJS.15.00934
Scientific Articles
Abstract

Background: Patients are often treated with a resurfacing hemiarthroplasty in the expectation that the bone-preserving design facilitates revision should the need for a revision arthroplasty arise. The aim of this study was to report the outcome of patients with glenohumeral osteoarthritis who underwent revision shoulder arthroplasty after resurfacing hemiarthroplasty.

Methods: We reviewed all patients with osteoarthritis reported to the Danish Shoulder Arthroplasty Registry from 2006 to 2013. There were 1,210 primary resurfacing hemiarthroplasties, of which 107 cases (9%) required a revision surgical procedure, defined as the removal or exchange of the humeral component or the addition of a glenoid component. The Western Ontario Osteoarthritis of the Shoulder (WOOS) index was used to evaluate outcome at 1 year.

Results: The median WOOS of revision arthroplasty after failed resurfacing hemiarthroplasty was 62 points (interquartile range, 40 to 88 points). Of the 80 cases that had follow-up, 33 (41%) had an unacceptable outcome, defined as a WOOS of ≤50 points. Of the 107 cases that required a revision surgical procedure, 11 arthroplasties (10%) required a further revision surgical procedure. The resurfacing hemiarthroplasty was revised to a stemmed hemiarthroplasty (n = 39), anatomic total shoulder arthroplasty (n = 31), or reverse shoulder arthroplasty (n = 30). In 7 cases, the revision arthroplasty design was unknown. The median WOOS of patients who underwent revision stemmed hemiarthroplasty (48 points) was significantly inferior (p = 0.002) to that of patients who underwent primary stemmed hemiarthroplasty (75 points); the median WOOS of patients who underwent revision anatomic total shoulder arthroplasty (74 points) was also significantly inferior (p = 0.007) to that of patients who underwent primary anatomic total shoulder arthroplasty (93 points). However, the median WOOS of patients who underwent revision reverse shoulder arthroplasty (68 points) was not significantly different (p = 0.66) from that of patients who underwent primary reverse shoulder arthroplasty (77 points) used in the treatment of osteoarthritis.

Conclusions: The outcome of revision shoulder arthroplasty after failed resurfacing hemiarthroplasty was variable and, in many cases, disappointing. It is important that resurfacing hemiarthroplasty is used for the correct indications and with adequate technique and skill. When resurfacing hemiarthroplasty is used in the treatment of osteoarthritis, revision cannot be counted upon as a fallback.

Level of Evidence: Therapeutic Level II. See Instructions for Authors for a complete description of levels of evidence.

Author Information

1Department of Orthopaedic Surgery, Herlev University Hospital, Herlev, Denmark

E-mail address for J.V. Rasmussen: jevera01@heh.regionh.dk

Article Outline

The first resurfacing shoulder arthroplasty design to be routinely used was the Scan (Scandinavian) Shoulder (MITAB) with cemented fixation used between 1981 and 1985 in the treatment of rheumatoid arthritis1. A few years later, a similar design was used in the treatment of glenohumeral osteoarthritis2. In the following years, there were advances in osseous ingrowth technology, and the first modern hydroxyapatite-coated resurfacing arthroplasty with a central grooved impact-fit taper peg for cementless fixation was introduced in 19932. Since then, this arthroplasty design has been widely used, although, to our knowledge, there have been limited documentation of efficacy and no comparative studies. In a recent publication, patients who underwent resurfacing hemiarthroplasty were reported to have a functional outcome comparable with that of patients who underwent stemmed hemiarthroplasty, but inferior to that of patients who underwent anatomic total shoulder arthroplasty3.

Younger patients, with a long life expectancy, are often treated with a resurfacing hemiarthroplasty to avoid problems with glenoid loosening and with the expectation that the bone-preserving design facilitates later revision arthroplasty4-6. However, the rationale for this thought is not well validated.

The aim of this study was to report the outcome of patients with glenohumeral osteoarthritis who underwent revision shoulder arthroplasty after resurfacing hemiarthroplasty. We hypothesized that the results of patients who underwent revision arthroplasty after failed resurfacing hemiarthroplasty would be similar to the outcome of patients who underwent primary arthroplasty and that age would not be a risk factor for inferior outcomes.

Back to Top | Article Outline

Materials and Methods

The Danish Shoulder Arthroplasty Registry (DSR) was established in 20047. All Danish public hospitals and private clinics provide data. The registry is independent of commercial interests. Data are electronically reported by the surgeon at the time of the surgical procedure. During the first few years of the registry, reporting was voluntarily and the compliance of reporting was unacceptably low. Reporting became mandatory in 2006, and, since 2007, in each year, the registry has obtained data on 92% to 96% of all shoulder arthroplasty cases.

Patient-reported outcome is assessed by the registry 1 year after the operation using the Western Ontario Osteoarthritis of the Shoulder (WOOS) index8. The Danish version of WOOS has been translated and has undergone cross-cultural adaption as per Guillemin et al.9 and was validated with use of the classical test theory10. WOOS is a disease-specific questionnaire for measuring the quality of life of patients with glenohumeral osteoarthritis. There are 19 questions about physical symptoms and their effect on sport, work, lifestyle, and emotions. Each question is answered on a visual analog scale ranging from 0 to 100. The total score ranges from a raw score of 0 to 1,900, with a raw score of 1,900 being the worst. For simplicity of presentation, the total score is converted to a percentage of the maximum score, with 100 points being the best. A score of ≤50 points is regarded as unsatisfactory by the DSR. The minimal clinically important difference on the WOOS has never been validated, to our knowledge. In the present study, it is defined as a raw score of 190 (e.g., 10% of a maximum score), which is an extrapolation from studies validating the minimal clinically important difference of other shoulder-specific measures11,12 (e.g., the Constant-Murley score and the Oxford Shoulder Score).

The WOOS is sent to the patients by a secretary located at the DSR facilities at the Department of Clinical Epidemiology, Aarhus University Hospital, Aarhus, Denmark, without involving orthopaedic surgeons. For primary and revision procedures, the questionnaire is only sent at 1 year without any formal long-term follow-up or preoperative assessment. A single reminder is sent to nonresponders and to those returning an incomplete questionnaire. In case of revision, death, or immigration within the first year, the WOOS score cannot be obtained.

A revision is defined as the removal or exchange of the humeral (resurfacing) component or the addition of a glenoid component. A revision is reported to the registry by the surgeon at the time of the revision procedure. It contains the same variables as a primary procedure with additional information about the revision arthroplasty, date, and indication for revision. The indication and decision for revision are based on a clinical examination and radiographs, with additional diagnostic procedures such as computed tomography (CT), magnetic resonance imaging (MRI), ultrasound, bone scintigraphy, and microbiological culturing performed if necessary. The available reasons for revision include subluxation, loosening, glenoid arthrosis, periprosthetic fracture, infection, technical failure including overstuffing and malposition of the implant, and rotator cuff problems. The surgeons can also report the reason for revision as “other” and then can describe the reason. Pain as the isolated basis for revision is not an option. However, in the present study, pain is registered as the reason for revision if it is reported and is listed as “other” and if no other reason is reported. A revision is linked to the primary procedure using the patient-specific civil registration number given at the time of birth.

We reviewed all patients with glenohumeral osteoarthritis who underwent a primary resurfacing hemiarthroplasty reported to the registry between January 2006 and December 2013, and 1,210 primary resurfacing hemiarthroplasties in 1,109 patients were eligible. For these primary arthroplasties, we collected data on the annual number, the WOOS 1 year after the operation, survival rates, and indications for revision. For the revisions, we collected data on the revision arthroplasty design, the WOOS 1 year after the revision, and the need for re-revision. The revised resurfacing hemiarthroplasties were separated into 3 groups on the basis of the arthroplasty design used in the revision procedure: stemmed hemiarthroplasty, anatomic stemmed total shoulder arthroplasty, or reverse total shoulder arthroplasty. The results were then compared with those of primary stemmed hemiarthroplasty, primary anatomic stemmed total shoulder arthroplasty, and primary reverse total shoulder arthroplasty used in the treatment of osteoarthritis and reported to the registry in the same time period.

Back to Top | Article Outline
Statistical Analysis

Although it violates the assumption of independency, patients with bilateral replacements were included in the statistical analyses as if they were independent. The Kaplan-Meier method was used to illustrate the unadjusted cumulative survival rates. A Cox regression model was used to estimate the relative risk of revision, and a general linear model was used to analyze differences in WOOS scores when primary resurfacing hemiarthroplasty brands were compared. Age, sex, year of surgical procedure, and type of osteoarthritis were included in the models. A general linear model was also used to analyze differences in the WOOS when patients ≤55 years of age and older patients (those who were >55 years of age) were compared. Sex, year of the surgical procedure, primary resurfacing hemiarthroplasty brand, and type of osteoarthritis were included in the model. The assumptions of a general linear model, including the assumption of normality, are less important when large subpopulations are compared13. The results of revision arthroplasty designs were analyzed using nonparametric (Mann-Whitney) testing. The analyses were performed using SPSS version 21.0 (IBM). The level of significance was set at p < 0.05 and p values were two-tailed.

Back to Top | Article Outline

Results

The Distribution of the Shoulder Implant Designs in the Study Period

From 2006 to 2013, 2,293 primary shoulder replacements in patients with osteoarthritis were reported to the registry. Although the annual number of anatomic total shoulder arthroplasties has increased, resurfacing hemiarthroplasty was the most frequently used arthroplasty design in the study period (Fig. 1).

Back to Top | Article Outline
Primary Resurfacing Hemiarthroplasty

There were 1,210 primary resurfacing hemiarthroplasties reported: 1,070 performed for primary osteoarthritis, 118 performed for secondary osteoarthritis, and 22 performed for osteoarthritis of an unreported type. The mean patient age (and standard deviation) was 65 ± 11 years. In 663 cases (55%), the patients were female. There were 677 (56%) Copeland resurfacing arthroplasties (Biomed), 481 (40%) Global CAP resurfacing arthroplasties (DePuy), 50 other brand arthroplasties (4%), and 2 arthroplasties (0.2%) in which the brand was not reported. A completed WOOS was returned by patients undergoing 964 arthroplasties (80%). The mean WOOS was 69 ± 26 points, and the median WOOS was 76 points (interquartile range, 50 to 92 points). In 241 cases (25%), patients had a WOOS of ≤50 points. The Copeland resurfacing hemiarthroplasty had a similar outcome (p = 0.06) compared with that of the Global CAP, with a mean difference in WOOS of 3 points (95% confidence interval [95% CI], 0 to 7 points). There were 107 arthroplasties (9%) that required revision, with an 8-year cumulative survival rate of 87% (Fig. 2). The most common reported reasons for revision were glenoid arthrosis (n = 29), a rotator cuff problem (n = 24), and technical failure including overstuffing or malposition of the implant (n = 21) (Table I). The relative risk of revision after the Copeland hemiarthroplasty compared with the Global CAP arthroplasty was 1.1 (95% CI, 0.8 to 1.9; p = 0.31). There was no significant difference (p = 0.16) in the unadjusted cumulative survival rates between the two designs (Fig. 3).

Back to Top | Article Outline
Revised Resurfacing Hemiarthroplasties

There were 107 revisions, of which 51 (48%) were performed in women. The mean patient age at the time of revision was 63 ± 11 years. The mean time to revision was 27 ± 19 months. A final WOOS was unavailable in 8 cases because of a re-revision, death, or immigration within 1 year. Of those who were available, 80 patients (81%) returned a complete questionnaire. One year after failed resurfacing hemiarthroplasties, the mean WOOS was 61 ± 27 points and the median WOOS was 62 points (interquartile range, 40 to 88 points). Thirty-three (41%) of these patients had an unacceptable outcome with a WOOS of ≤50 points. Eleven arthroplasties (10%) required further revision surgical procedures, with deep infection as the reason in 6 cases. The relative risk of revision in case of a WOOS of ≤50 points at 1 year was 3.3 (95% CI, 2.1 to 5.0; p < 0.001) compared with cases with a WOOS score of >50 points.

Back to Top | Article Outline
Outcome of Revision Shoulder Arthroplasty

Thirty-nine arthroplasties were revised to a stemmed hemiarthroplasty. The most frequent reason for revision was technical failure including overstuffing and malposition of the implant (Table II). The median WOOS was significantly inferior to that of primary stemmed hemiarthroplasty (p = 0.002) (Table III).

Thirty-one arthroplasties were revised to an anatomic total shoulder arthroplasty, of which 20 were revised because of glenoid arthrosis (Table II). The median WOOS was significantly inferior to that of primary anatomic total shoulder arthroplasty (p = 0.007) (Table III).

Thirty arthroplasties were revised to a reverse shoulder arthroplasty. The most frequent reason for revision was a rotator cuff problem (Table II). The median WOOS was not significantly different from that of primary reverse shoulder arthroplasty used in the treatment of osteoarthritis (p = 0.66) (Table III).

Back to Top | Article Outline
Resurfacing Hemiarthroplasty in Young Patients

Of the primary resurfacing hemiarthroplasties, 241 (20%) were used in patients who were ≤55 years of age. The mean time to revision was 30 ± 19 months. At 1 year, for this subpopulation, the mean WOOS was 56 ± 26 points, and the median WOOS was 56 points (interquartile range, 34 to 78 points). Of the 200 cases that had follow-up, 92 (46%) had a WOOS of ≤50 points. The WOOS score was significantly inferior (p < 0.001) to the results of older patients, with a mean difference of 18 points (95% CI, 14 to 22 points). For the young patients, 36 arthroplasties (15%) required revision, with an 8-year cumulative survival rate of 80% (Fig. 4). The reasons for revision were mainly glenoid arthrosis (n = 13) and technical failure including overstuffing and malposition of the implant (n = 9) (Table I). For revised resurfacing hemiarthroplasty cases at 1 year, the mean WOOS was 52 ± 22 points, the median WOOS was 47 points (interquartile range, 37 to 66 points), and 62% of the cases had a WOOS of ≤50 points. There were 16 revision stemmed hemiarthroplasty cases with a mean WOOS of 53 ± 27 points and a median WOOS of 47 points (interquartile range, 31 to 79 points), 13 revision anatomic total shoulder arthroplasty cases with a mean WOOS of 53 ± 17 points and a median WOOS of 59 points (interquartile range, 36 to 66 points), and four revision reverse shoulder arthroplasty cases. The type of revision arthroplasty was not reported in 3 cases.

Back to Top | Article Outline

Discussion

Over the past decades, resurfacing hemiarthroplasty has been frequently used in the expectation that the bone-preserving design facilitates revision should the need for a revision arthroplasty arise. However, we found disappointing outcomes when anatomic total shoulder arthroplasty and stemmed hemiarthroplasty were used to treat failed resurfacing hemiarthroplasties. There was no significant difference between primary reverse shoulder arthroplasty and reverse shoulder arthroplasty used to treat failed resurfacing hemiarthroplasties. A key study finding was the poor outcomes of primary resurfacing hemiarthroplasty and revision arthroplasty in young patients.

We were able to identify a few small case series reporting outcomes of revision shoulder arthroplasty in patients with glenohumeral osteoarthritis, but none of them showed outcomes after failed resurfacing hemiarthroplasty14-16.

During the last decade, Danish surgeons have preferred resurfacing hemiarthroplasty in the treatment of end-stage osteoarthritis. The reason for the high utilization of resurfacing hemiarthroplasty in Denmark compared with other countries, including the United States, is unclear but worth considering. The avoidance of anatomic total shoulder arthroplasty may have been an attempt to avoid replacement of the glenoid because of the risk of loosening.

Preoperative imaging is not included in the registry, and thus, we have no data on abnormal anatomy including glenoid arthrosis and do not know whether the initial indication for resurfacing hemiarthroplasty was appropriate. Glenoid arthrosis might have been present in some patients at the time of the primary operation.

The limited information about the mode of failure is worth considering. If an abnormality such as glenoid arthrosis or rotator cuff arthropathy was inadequately addressed at the primary operation, a disappointing outcome or the need of revision arthroplasty would be related to incorrect treatment selection rather than to the resurfacing arthroplasty design. Infection is also a recognized cause of pain associated with failed shoulder arthroplasty17, and, if inadequately addressed or not recognized before, the revision procedure might lead to a similarly poor outcome.

The results of patients who underwent a stemmed hemiarthroplasty after failed resurfacing hemiarthroplasty were disappointing, with a median WOOS of ≤50 points. The reason for using stemmed hemiarthroplasty to replace resurfacing hemiarthroplasty is unknown, but the use of stemmed hemiarthroplasty may have been an attempt to avoid replacement of the glenoid because of the risk of loosening.

One assumption is that the results reported by national arthroplasty registries can be generalized to the average shoulder surgeon. However, the proportion of anatomic total shoulder arthroplasty in Denmark is low, and it can be hypothesized that the excellent outcome of primary anatomic total shoulder arthroplasty found in this study may reflect the results of a few highly experienced shoulder surgeons. However, the results of patients who underwent primary stemmed hemiarthroplasty and, in particular, primary anatomic total shoulder arthroplasty are similar to the WOOS scores reported by Lo et al. in a randomized controlled trial comparing the 2 arthroplasty designs18.

The definition of a young patient is debatable. For this study, we used the threshold of 55 years, as this has been used in previous studies evaluating the outcome of primary resurfacing shoulder arthroplasty in young patients19,20. In the present study, the outcome of primary resurfacing hemiarthroplasty and revision arthroplasty after failed resurfacing hemiarthroplasty was variable and, in many cases, disappointing. The 8-year cumulative survival rate of primary resurfacing hemiarthroplasty implants was 80%. In comparison, a previous study showed a 5-year survival rate of 98% but a 10-year survival rate of 63% after primary stemmed total shoulder arthroplasty in young patients21. Thus, the treatment of young patients with end-stage glenohumeral osteoarthritis remains controversial and challenging.

This study had limitations. There was no accounting for surgeon experience or the specific indications for the primary resurfacing hemiarthroplasty, and some of the disappointing outcomes may have been the result of inferior surgical technique and skill or incorrect treatment selection. The poor results of patients who underwent revision arthroplasty after failed resurfacing hemiarthroplasty might have been the result of similar limitations. There was no preoperative WOOS for either patients who underwent primary resurfacing hemiarthroplasty or those who underwent revision arthroplasty. Thus, improvement in shoulder function might have been more relevant to report than the postoperative shoulder function alone, especially for patients who underwent revision arthroplasties. Finally, incorrect reporting may diminish the accuracy and reliability of the data.

In conclusion, the outcome of patients who underwent revision shoulder arthroplasty after failed resurfacing hemiarthroplasty was variable and, in many cases, disappointing. A key study finding was the high proportion of disappointing outcomes of young patients who underwent revision arthroplasty after failed primary resurfacing hemiarthroplasty. It is important that resurfacing hemiarthroplasty is used for the correct indications and with adequate technique and skill. When resurfacing hemiarthroplasty is used in the treatment of osteoarthritis, revision cannot be counted upon as a fallback.

NOTE: The authors thank the orthopaedic surgeons in Denmark for data reporting.

Investigation performed at the Department of Orthopaedic Surgery, Herlev University Hospital, Herlev, Denmark

Disclosure: There was no external funding for this study. The Disclosure of Potential Conflicts of Interest forms are provided with the online version of the article.

Back to Top | Article Outline

References

1. Jónsson E, Egund N, Kelly I, Rydholm U, Lidgren L. Cup arthroplasty of the rheumatoid shoulder. Acta Orthop Scand. 1986 ;57(6):542–6.
2. Copeland S. The continuing development of shoulder replacement: “reaching the surface”. J Bone Joint Surg Am. 2006 ;88(4):900–5.
3. Rasmussen JV, Polk A, Brorson S, Sørensen AK, Olsen BS. Patient-reported outcome and risk of revision after shoulder replacement for osteoarthritis. 1,209 cases from the Danish Shoulder Arthroplasty Registry, 2006-2010. Acta Orthop. 2014 ;85(2):117–22.
4. Levy O, Copeland SA. Cementless surface replacement arthroplasty (Copeland CSRA) for osteoarthritis of the shoulder. J Shoulder Elbow Surg. 2004 ;13(3):266–71.
5. Levy O, Tsvieli O, Merchant J, Young L, Trimarchi A, Dattani R, Abraham R, Copeland SA, Narvani A, Atoun E. Surface replacement arthroplasty for glenohumeral arthropathy in patients aged younger than fifty years: results after a minimum ten-year follow-up. J Shoulder Elbow Surg. 2015 ;24(7):1049–60. Epub 2015 Jan 16.
6. Thomas SR, Wilson AJ, Chambler A, Harding I, Thomas M. Outcome of Copeland surface replacement shoulder arthroplasty. J Shoulder Elbow Surg. 2005 ;14(5):485–91.
7. Rasmussen JV, Jakobsen J, Brorson S, Olsen BS. The Danish Shoulder Arthroplasty Registry: clinical outcome and short-term survival of 2,137 primary shoulder replacements. Acta Orthop. 2012 ;83(2):171–3. Epub 2012 Feb 13.
8. Lo IK, Griffin S, Kirkley A. The development of a disease-specific quality of life measurement tool for osteoarthritis of the shoulder: The Western Ontario Osteoarthritis of the Shoulder (WOOS) index. Osteoarthritis Cartilage. 2001 ;9(8):771–8.
9. Guillemin F, Bombardier C, Beaton D. Cross-cultural adaptation of health-related quality of life measures: literature review and proposed guidelines. J Clin Epidemiol. 1993 ;46(12):1417–32.
10. Rasmussen JV, Jakobsen J, Olsen BS, Brorson S. Translation and validation of the Western Ontario Osteoarthritis of the Shoulder (WOOS) index - the Danish version. Patient Relat Outcome Meas. 2013 ;4:49–54.
11. Kukkonen J, Kauko T, Vahlberg T, Joukainen A, Aärimaa V. Investigating minimal clinically important difference for Constant score in patients undergoing rotator cuff surgery. J Shoulder Elbow Surg. 2013 ;22(12):1650–5. Epub 2013 Jul 12.
12. Christiansen DH, Frost P, Falla D, Haahr JP, Frich LH, Svendsen SW. Responsiveness and minimal clinically important change: a comparison between 2 shoulder outcome measures. J Orthop Sports Phys Ther. 2015 ;45(8):620–5. Epub 2015 Jun 25.
13. Lumley T, Diehr P, Emerson S, Chen L. The importance of the normality assumption in large public health data sets. Annu Rev Public Health. 2002;23:151–69. Epub 2001 Oct 25.
14. Sassoon AA, Rhee PC, Schleck CD, Harmsen WS, Sperling JW, Cofield RH. Revision total shoulder arthroplasty for painful glenoid arthrosis after humeral head replacement: the nontraumatic shoulder. J Shoulder Elbow Surg. 2012 ;21(11):1484–91. Epub 2012 Mar 23.
15. Levy JC, Virani N, Pupello D, Frankle M. Use of the reverse shoulder prosthesis for the treatment of failed hemiarthroplasty in patients with glenohumeral arthritis and rotator cuff deficiency. J Bone Joint Surg Br. 2007 ;89(2):189–95.
16. Cheung EV, Sperling JW, Cofield RH. Revision shoulder arthroplasty for glenoid component loosening. J Shoulder Elbow Surg. 2008 ;17(3):371–5. Epub 2008 Feb 20.
17. Pottinger P, Butler-Wu S, Neradilek MB, Merritt A, Bertelsen A, Jette JL, Warme WJ, Matsen FA 3rd. Prognostic factors for bacterial cultures positive for Propionibacterium acnes and other organisms in a large series of revision shoulder arthroplasties performed for stiffness, pain, or loosening. J Bone Joint Surg Am. 2012 ;94(22):2075–83.
18. Lo IK, Litchfield RB, Griffin S, Faber K, Patterson SD, Kirkley A. Quality-of-life outcome following hemiarthroplasty or total shoulder arthroplasty in patients with osteoarthritis. A prospective, randomized trial. J Bone Joint Surg Am. 2005 ;87(10):2178–85.
19. Bailie DS, Llinas PJ, Ellenbecker TS. Cementless humeral resurfacing arthroplasty in active patients less than fifty-five years of age. J Bone Joint Surg Am. 2008 ;90(1):110–7.
20. Raiss P, Pape G, Becker S, Rickert M, Loew M. [Cementless humeral surface replacement arthroplasty in patients less than 55 years of age]. Orthopade. 2010 ;39(2):201–8. German.
21. Denard PJ, Raiss P, Sowa B, Walch G. Mid- to long-term follow-up of total shoulder arthroplasty using a keeled glenoid in young adults with primary glenohumeral arthritis. J Shoulder Elbow Surg. 2013 ;22(7):894–900. Epub 2013 Jan 9.
Copyright 2016 by The Journal of Bone and Joint Surgery, Incorporated