Treatment of advanced glenohumeral osteoarthritis with joint replacement has been reported to lead to good to excellent outcomes in >90% of cases1. Complications, however, do occur and include aseptic loosening, glenohumeral instability, rotator cuff tears, periprosthetic fractures, implant failure, infection, nerve injury, and failure of the subscapularis repair following surgery2.
Failure of the subscapularis repair3,4 remains a largely under-recognized problem following total shoulder replacement. Compromise of subscapularis integrity and function may occur secondary to injury, poor-quality tissue, excessive tension due to oversizing of components, or nerve injury during mobilization4. Miller et al.4 reported on forty-one patients who had total shoulder replacement with an intrasubstance subscapularis tenotomy. The authors identified abnormal subscapularis function in more than two-thirds of the cohort, with an abnormal lift-off test in 67% of the patients and an abnormal belly-press in 68%.
Unfortunately, complete failure of the subscapularis repair may result in pain, weakness, and instability, and the outcomes of treatment are generally poor3. In an effort to minimize the incidence of subscapularis repair failure following total shoulder replacement, an alternative technique involving a lesser tuberosity osteotomy has been described5. The technique involves a tangential osteotomy of the lesser tuberosity in continuity with the insertion of the subscapularis tendon6. The osteotomy is then repaired with transosseous sutures, allowing healing across the two opposing bone surfaces, thought by some authors to be better than bone-to-tendon healing5,7,8. The selection of a subscapularis mobilization technique is controversial as each technique has its advantages, disadvantages, and surgeon advocates. A recent biomechanical study comparing osteotomy with tenotomy demonstrated similar results between the two techniques in terms of load to failure, with greater displacement under cyclic loading following the subscapularis osteotomy9. One clinical study compared tenotomy with osteotomy8, and a second study compared osteotomy with a complete elevation of the subscapularis from the lesser tuberosity (peel)7. Both studies demonstrated superior clinical outcomes for shoulders that were treated with lesser tuberosity osteotomy.
The primary objective of the present study was to compare lesser tuberosity osteotomy with subscapularis peel during shoulder arthroplasty. The primary outcome was subscapularis muscle strength as measured with an electronic handheld dynamometer in the belly-press position at two years postoperatively. Secondary outcomes that were assessed at two years of follow-up included disease-specific quality-of-life and function as measured with the Western Ontario Osteoarthritis of the Shoulder Index (WOOS) and the American Shoulder and Elbow Surgeons (ASES) score. Our hypothesis was that patients who were managed with a lesser tuberosity osteotomy would have higher subscapularis muscle strength scores than those who were managed with a subscapularis peel.
This double-blind, balanced randomization (1:1) study was conducted at two orthopaedic surgery facilities. From November 2006 to June 2009, patients were randomly assigned to one of two methods of mobilization of the subscapularis tendon during shoulder arthroplasty: (1) lesser tuberosity osteotomy or (2) subscapularis peel. The study received approval from the ethics boards at both institutions. Patients were recruited in outpatient specialty clinics by the study coordinator or research assistants. The target population included men and women of any age with advanced arthritis of the glenohumeral joint who were considered by the treating surgeon to be a candidate for shoulder replacement. The arthritis had to be amenable to treatment with either a humeral head replacement or a standard total shoulder replacement.
The inclusion criteria for the present study included unsuccessful standard nonoperative treatment of osteoarthritis or inflammatory arthritis. Failed medical treatment was defined as persistent pain and disability despite adequate nonoperative treatment for six months. Medical treatment was defined as (1) the use of medications, including analgesics and nonsteroidal anti-inflammatory drugs, (2) physiotherapy consisting of stretching, strengthening, and local modalities, and (3) activity modification. The exclusion criteria were active joint or systemic infection, rotator cuff arthropathy, muscle paralysis, neuropathic arthropathy, major medical illness (a life expectancy of less than one year or an unacceptably high operative risk), inability to speak or read English or French, psychiatric illness that precluded informed consent, and an inability to be followed for two years. Individuals who met the inclusion criteria and agreed to participate were contacted and enrolled by a research assistant.
This trial was registered with www.ClinicalTrials.gov and the registry number is NCT00508105.
Outcomes
The primary outcome measure was an assessment of subscapularis muscle strength in the belly-press position as measured with an electronic handheld dynamometer (microFET2; Hoggan Health Industries, West Jordan, Utah). Subscapularis strength was measured with use of a variation of the “belly-press test” described by Gerber et al.10. The test was conducted with the dynamometer strapped to the patient’s hand. The hand was passively placed onto the lower part of the sternum, with the elbow in line with, but not posterior to, the hand. The patient was asked to press the dynamometer into the sternum with use of a maximum force for five seconds. The sensitivity of the belly-press test has been reported to be 100% for the detection of complete subscapularis tears11, and the handheld dynamometer has been validated for use in subscapularis evaluation12.
Secondary outcome measures included the WOOS and ASES scores13. The WOOS is a disease-specific evaluation that is an accurate and valid assessment of function after shoulder replacement14. Because it is specific for osteoarthritis of the shoulder, it is highly sensitive to small but clinically important changes in patient function14. The ASES score is a joint-specific assessment tool developed by the American Shoulder and Elbow Surgeons for the evaluation of all shoulder disorders13. It consists of a patient self-assessment component and a physician-assessment component. The patient self-assessment component is divided into two sections: (1) pain and (2) activities of daily living. Pain is recorded on a visual analog scale, and activities of daily living are recorded on a numeric scale. The overall score is an equal weight of the two components and is scored out of 100; the higher the score, the better the outcome. After consent and before surgery, the research assistants, who were not involved in the surgical procedure, collected the functional and quality-of-life outcome measures in the clinical setting. As the WOOS and ASES were self-administered, no specific patient coaching was required on behalf of the research assistants for this portion of the assessment.
Other secondary outcomes that were monitored during the postoperative course included complications and the rate of revision surgery. Patients completed all outcome measures preoperatively and at three, six, twelve, and twenty-four months postoperatively. Standard radiographs, consisting of a true anteroposterior view in neutral and external rotation and an axillary view, were made and reviewed at three months, one year, and two years of follow-up. Any implant-related complications such as evidence of loosening or instability were recorded. Evidence of loosening was determined on the basis of the presence of radiolucent lines around the humeral component according to the method described by Sanchez-Sotelo et al.15 as applied to uncemented, ingrowth humeral stems16,17. Radiolucent lines around the glenoid were determined with the method of Lazarus et al.18. Anterior subluxation of the humeral component was determined on the axillary view19.
Surgery
Final eligibility for the study was determined intraoperatively following visual inspection of the subscapularis tendon to ensure that it was intact. All surgical procedures were performed by one of two shoulder-fellowship-trained surgeons (P.L.C.L. and G.S.A.). Once eligibility was confirmed, the patient was randomized to either subscapularis peel or lesser tuberosity osteotomy. Allocation was carried out with use of computer-generated blocked randomization, with the surgeon masked to block size. Treatment allocation was printed on cards that were inserted into sealed, opaque envelopes. The envelopes were opened by the circulating nurse once patient eligibility was confirmed.
The surgical procedures were performed with the patient under general anesthesia and in the beach-chair position. Prophylactic antibiotics were administered. The deltopectoral approach was utilized to expose the shoulder. The long head of the biceps was tenodesed to the pectoralis major in all cases when present. For the peel group, the subscapularis tendon was peeled off the lesser tuberosity, beginning at the intertubercular groove. Implants consisted of press-fit humeral stems and keeled, cemented glenoid components in all cases (Aequalis; Tornier, Montbonnot, France). Following arthroplasty, the tendon was repaired with three non-absorbable mattress transosseous sutures (Hi-Fi; ConMed Linvatec, Largo, Florida) passed within the bicipital groove and tied over a mini-plate placed on the lateral aspect of the greater tuberosity (Figs. 1 and 2).
The lesser tuberosity osteotomy was conducted as described by Gerber et al.5. A fragment of the lesser tuberosity (measuring 5 to 10 mm in thickness and 3 cm in length) was elevated with an osteotome along with the subscapularis insertion. After arthroplasty, the osteotomy repair was performed in a manner identical to the peel repair. The osteotomy was secured with three non-absorbable mattress sutures placed in a transosseous fashion through the bicipital groove, with the sutures tied over a mini-plate on the lateral side of the greater tuberosity (Figs. 1 and 2).
Postoperative care and physiotherapy were identical in both groups. Patients received twenty-four hours of prophylactic antibiotic coverage. A shoulder sling was worn for the first six weeks. Self-assisted forward elevation was initiated on the first postoperative day to a maximum of 90°, and self-assisted external rotation was limited to neutral for the first six weeks. Patients were instructed to perform exercises every two hours with the contralateral arm by guiding the involved arm through the exercise while maintaining the involved arm in a completely relaxed state. Patients were instructed to perform all exercises in the supine position. Patients were generally discharged by the second day and continued the rehabilitation program as outpatients. Physiotherapy was initiated at six weeks postoperatively. Active range of motion followed at six weeks. At twelve weeks postoperatively, gentle strengthening exercises were initiated.
Because of the nature of the surgical trial, it was not possible to blind the surgeon to the surgical intervention. However, a trained research assistant performed the follow-up assessments and was blinded to the surgical procedure. The assessor did not have access to the patient chart prior to the evaluation. The patient was also blinded to the treatment assignment.
Statistical Methods
In the absence of an established minimal clinically important difference for the primary outcome measure, the sample size calculation was based on observing a large effect size between the two groups. This assumption was justified on the basis of a previous clinical trial in which two subscapularis surgical procedures were compared with use of a dynamometer strength test20. Although both groups were reported to have improved from baseline, the mean difference (and standard deviation) between the groups in terms of postoperative strength was 1.9 ± 2.4 kg (effect size, 0.79). This difference was interpreted as clinically meaningful; therefore, with an effect size of 0.79, a beta value of 0.1, and an alpha value of 0.05, thirty-four participants were required in each group. An additional eighteen patients (25%) were added to account for potential loss to follow-up. Thus, forty-three participants were to be enrolled in each group, for a sample size of eighty-six patients total.
The mean, standard deviation, and range were reported for the outcome variables at each time point in each treatment group and overall. A longitudinal analysis, mixed-effect model, using patients as random effect, was used to assess whether the strength, WOOS, and ASES scores improved significantly in each group over time, and two-sided t tests were used to compare these outcomes between the two treatment groups at baseline, six months, one year, and two years. A multivariable logistic regression was performed to control for the effect of age when comparing the effect of treatment on strength at twenty-four months.
Source of Funding
The Physicians’ Services Incorporated (PSI) Foundation provided grant funding for this study. PSI is a not-for-profit peer-reviewed granting agency. PSI did not play a role in the investigation beyond providing the grant for the study.
In total, 102 patients were enrolled in the study. Fifteen patients who were enrolled were never randomized as they either postponed or cancelled the procedure. Thus, eighty-seven patients were randomized in the study. After randomization and the surgical procedure, one patient refused to complete any of the outcome measures. One patient died after the three-month follow-up, and another died after the twelve-month follow-up; both patients died of causes that were unrelated to the shoulder surgery. In total, eight patients were lost to follow-up by twenty-four months. Seven patients were randomized but were excluded from the analysis; of these, five had rheumatoid arthritis and two had had previous open stabilization surgery. Eighty-four percent of patients were included in the final analysis. The flow of patients in the study is summarized in the CONSORT (Consolidated Standards of Reporting Trials) flow diagram21 (Fig. 3).
Baseline demographic data are summarized in the Appendix. Baseline characteristics, including sex, affected side, and arthroplasty type (hemiarthroplasty or total shoulder arthroplasty), did not differ significantly between the groups. The mean age of the patients in the subscapularis peel group was lower than that in the osteotomy group (65.3 compared with 70.4 years). Forty-three patients were allocated to lesser tuberosity osteotomy, and forty-four patients were allocated to subscapularis peel. The average age of the study participants was 67.8 years (range, thirty-four to ninety years). Thirty-nine percent of the patients in the study cohort were male.
Analysis of the primary outcome (subscapularis muscle strength) at twenty-four months revealed no significant difference between the lesser tuberosity osteotomy group and the subscapularis peel group (4.4 ± 2.9 kg [range, 0.0 to 12.5 kg] compared with 5.5 ± 2.6 kg [range, 1.0 to 13.5 kg]; p = 0.1309). Multivariable regression analysis, with strength as the dependent variable, revealed that this outcome did not change after controlling for age. At twenty-four months, the average subscapularis strength in the lesser tuberosity osteotomy group was 0.91 kg lower than that in the subscapularis peel group when controlling for age (p = 0.186). Mean subscapularis strength increased significantly from baseline to twenty-four months in both the lesser tuberosity osteotomy group (p = 0.0042) and the subscapularis peel group (p < 0.0001). The mean increase in strength from baseline to twenty-four months was 1.71 ± 2.16 kg (range, −3.0 to 7.0 kg) for the osteotomy group and 2.41 ± 2.39 kg (range, 1.5 to 7.0 kg) for the peel group; this difference was not significant (p = 0.239).
Comparison of the secondary outcomes, including the WOOS and ASES scores, demonstrated no significant differences between the groups at any time point. The WOOS and ASES scores improved significantly in each group when preoperative scores were compared with the scores at twenty-four months of follow-up (p < 0.0001). All outcome data are summarized in Tables I, II, and III.
Two patients had a nonunion at the osteotomy site as seen on postoperative radiographs. Neither of these patients reported a feeling of instability, and both were satisfied with the results of surgery. We did not observe any evidence of loosening or instability of the implants in either group. Two patients in the subscapularis peel group underwent further surgery. The first patient underwent two-stage revision because of implant-related infection. The second patient underwent revision surgery to a reverse total shoulder arthroplasty after sustaining a massive posterosuperior rotator cuff tear. There were no reoperations for the treatment of subscapularis failure in either group.
Historically, shoulder arthroplasty was conducted with a subscapularis tenotomy or peel. Recently, the lesser tuberosity osteotomy was introduced as an option with hypothesized improved healing and subsequent strength due to the theoretical advantages of bone-to-bone healing. Our goal was to examine the functional outcomes of these approaches in a high-level-of-evidence study. Our results indicate that there was no clinically important difference in strength (assessed with a modified belly-press maneuver and measured with an electronic handheld dynamometer) between patients who underwent a lesser tuberosity osteotomy and those who underwent a complete subscapularis peel at any time point during the study period. Both groups demonstrated a significant improvement in subscapularis strength from baseline to the latest follow-up, but no significant differences were observed between groups in the strength change scores from baseline to the latest follow-up. Furthermore, there were no clinically important differences between groups in the secondary measures (WOOS and ASES scores).
In the absence of a previously established minimum clinically important difference for the primary outcome measure, the sample size calculation in this study was based on observing a large effect size between the two groups. This assumption was justified on the basis of a previous clinical trial in which two subscapularis surgical procedures were compared with use of a dynamometer strength test as one of the outcome measures20. Although both groups were reported to have improved from baseline, the difference between groups in terms of postoperative mean strength was 1.9 ± 2.4 kg (effect size, 0.79). This difference was interpreted as clinically meaningful. Thus, the current study was powered to detect a difference in subscapularis muscle strength of 1.9 kg. At twenty-four months, the final difference between groups was 1.1 kg (p = 0.1309); thus, we conclude that no significant difference existed between the groups.
Gerber et al. described the results of the lesser tuberosity osteotomy during shoulder arthroplasty in a study of thirty-nine shoulders (thirty-six patients) after a mean duration of follow-up of thirty-nine months5. The authors developed a technique of lesser tuberosity osteotomy on the basis of their experience with subscapularis dysfunction following tenotomy. At the time of the latest follow-up, they reported that 89% of the patients had a normal belly-press test and 75% had a normal lift-off sign. The authors concluded that the lesser tuberosity osteotomy was a safe and reliable approach and stated that it provided better outcomes than those reported in the literature in association with subscapularis tenotomy.
Ponce et al.22, in a study of seventy-six patients who were managed with a lesser tuberosity osteotomy, reported that the belly-press and lift-off signs were normal for sixty-two patients (82%), abnormal for five (7%), and not documented for nine (12%). In the biomechanical portion of the study, the authors observed that the strength of the lesser tuberosity osteotomy repair was significantly higher than that of the tenotomy repair. Qureshi et al.23, in a report on their results with the lesser tuberosity osteotomy, noted that a high proportion (83.3%) of patients were able to tuck in a shirt behind the back and that the results of the belly-press examination were normal for eighteen (60%) of thirty patients.
In support of the traditional intrasubstance tenotomy, Caplan et al. reported good results in association with tendon-to-tendon repair, with forty-one of forty-five patients having a negative lift-off test and four patients being unable to reach behind themselves to conduct the test24. However, other clinical studies have demonstrated poor results following subscapularis tenotomy. Jackson et al.25 reported that seven of fifteen patients who were managed with tenotomy had a complete tear of the subscapularis tendon on ultrasound. The authors found that the lift-off and abdominal compression tests correlated poorly with the sonographic condition of the tendon. Interestingly, the bear-hug test with use of dynamometry did correlate with tendon healing and integrity. Scalise et al.7 reported on thirty-five shoulders that were treated with either subscapularis tenotomy or osteotomy. Postoperative Penn scores were higher in the osteotomy group. Jandhyala et al. compared subscapularis tenotomy with osteotomy in a recent study of thirty-six patients8. The authors reported superior strength on the graded belly-press test on clinical evaluation in patients managed with osteotomy.
In the current study, strength as assessed with an electronic dynamometer was used as a clinical surrogate for healing and tendon integrity. Our observation that no difference exists between groups contrasts with the aforementioned studies on the subject, in which patients managed with osteotomy had higher functional scores7 and higher strength on the graded belly-press test8. The reasons for this are not clear. A possible explanation is that our method of repair in both groups relied on bone tunnels in the intertubercular groove and involved the use of a mini-plate for repair reinforcement. The mini-plate was used to prevent suture cut-out through bone and may provide a more robust repair. The outcome measure used in the current study also differs from that used in some previous reports23; it is possible that the ability to tuck in a shirt may not correlate well with subscapularis tendon integrity, as demonstrated by Jackson et al.25.
The limitations of the present study include a 17% rate of loss to follow-up. The sample size calculation, however, was increased to account for a 25% rate of loss to follow-up, and one extra patient was randomized. Thus, this trial has the requisite number of patients needed to power the study to detect a clinically important and statistically significant difference between groups. The primary outcome was assessed with use of a modified belly-press test involving an electronic handheld dynamometer; strength was measured with the patient’s hand on the xiphoid process on the lower part of the sternum rather than on the belly. This modification was necessary to obtain strength measurements as the dynamometer did not produce readings on the belly. It is possible that this modification may not perfectly replicate the original belly-press test; however, strength measurement with the device correlates closely with subscapularis tendon integrity when used in the bear-hug position25, which requires a greater degree of elbow flexion than the method used in the current trial. Another possible limitation was related to the inclusion of both hemiarthroplasty and total shoulder arthroplasty. As total shoulder arthroplasty may provide better pain relief than hemiarthroplasty, it is possible that the subscapularis strength may have been affected, although the numbers of patients managed with hemiarthroplasty were similar between the groups. Patients with rheumatoid arthritis (five patients) and patients who had undergone previous surgical procedures (two patients with previous open stabilization procedures) were excluded from the final analysis as their disease or previous surgery may have compromised the status of the subscapularis tendon. Finally, during the process of randomization, a significant difference in the mean ages between cohorts occurred. Although not a weakness per se, significant differences in baseline demographic characteristic can occur between groups in randomized trials as a result of chance alone. This difference was addressed with post hoc analysis. A regression analysis was conducted to investigate the effect of age on strength, and no significant difference was identified (p = 0.186).
In summary, the results of this clinical trial did not identify any significant difference between the lesser tuberosity osteotomy and subscapularis peel techniques in terms of the primary outcome (subscapularis muscle strength as measured with a modified belly-press test and an electronic handheld dynamometer). Additionally, no differences in disease-specific quality-of-life or functional outcomes were identified between these two techniques.
Note: The authors would like to acknowledge Physicians’ Services Incorporated Foundation for providing grant funding for this study.
Disclosure: One or more of the authors received payments or services, either directly or indirectly (i.e., via his or her institution), from a third party in support of an aspect of this work. None of the authors, or their institution(s), have had any financial relationship, in the thirty-six months prior to submission of this work, with any entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written in this work. Also, no author has had any other relationships, or has engaged in any other activities, that could be perceived to influence or have the potential to influence what is written in this work. The complete Disclosures of Potential Conflicts of Interest submitted by authors are always provided with the online version of the article.