Isolated rupture of the tendon of the subscapularis muscle was reported by Gerber and Krushell in 19911. Due to the relative rarity of this lesion, clinical results after repair have only been reported in a few studies, most of which reported short-term results or included a limited numbers of patients2-7. Studies of open repair have reported satisfactory outcomes2,3,6,8. Despite these results, postoperative subscapularis testing has demonstrated an incomplete functional correction2,6. It has been recognized that probable causes for this incomplete functional improvement could be the initial larger tendon tears and preoperative severe fatty infiltration of the subscapularis muscle6,9.
More recently, the popularity of arthroscopic repair of isolated subscapularis tendon lesions has increased, with clinical results reported to be equivalent to those of open repair4,5,7,10. Referring to the supraspinatus repair, many studies that have attempted to correlate the structural results to functional outcomes have shown controversial conclusions. Retears were not synonymous with failed clinical outcomes, and similar results were demonstrated when retears were compared with intact repairs11-16. On the other hand, better postoperative strength and function were observed when the tendon healed15-20. Structural outcomes of subscapularis tendon repairs were reported by Fuchs et al.8 and Lafosse et al.7 after open and arthroscopic repairs, respectively. Both types of repair demonstrated a good rate of tendon-healing, but the changes observed postoperatively with regard to fatty infiltration of the muscle after open repair8 were not confirmed after arthroscopic repair7.
The purpose of this study was to evaluate subscapularis tendon-healing after arthroscopic repair and to correlate the structural results with the clinical outcomes. We hypothesized that structural outcomes have an influence on subjective clinical and functional results and also the results seen on clinical testing (belly-press test and lift-off test) of the subscapularis muscle.
Between January 2004 and October 2007, 883 rotator cuff repairs were performed at our institution by the senior author (L.N.-J.). Forty-seven patients (5.3%) had an isolated subscapularis tear: the tears in twenty-two of them were repaired with use of an open approach, and the tears in twenty-five of them were repaired arthroscopically. In 2004, most of the procedures were done with use of an open approach, whereas, in 2007, all procedures were done arthroscopically.
The inclusion criteria of this retrospective study were any patient who had an isolated tear of the subscapularis tendon and underwent either open or arthroscopic repair with a minimum of a two-year clinical follow-up, had availability of preoperative and postoperative magnetic resonance imaging (MRI) or computed tomography arthrography (CTA) data, and had no previous shoulder surgery. In the arthroscopic group, one patient refused a postoperative MRI or CTA examination and was excluded from the study. Two patients underwent additional surgery on the same shoulder within two years after the index repair (repair of a traumatic subscapularis retear in one patient and of a supraspinatus full-thickness tear associated with a disinsertion of the upper part of the repaired subscapularis tendon in the other) and were excluded from the study. This resulted in twenty-two arthroscopic cases with complete clinical and MRI or CTA data. In the open surgery group, seven patients were lost to follow-up or did not agree to complete clinical or MRI evaluations, and two patients underwent reoperation as a result of retear that occurred before the completion of two years of follow-up (repair for a nontraumatic subscapularis retear in one patient and for an interstitial lesion of the supraspinatus tendon in the other). This resulted in thirteen open surgery cases with complete clinical and MRI or CTA data. This open group was used for comparison with the arthroscopic group. The study was approved by our institutional review committee.
In all patients except those who had a traumatic large tear, the indication for surgical repair was a traumatic or degenerative tear of the subscapularis tendon in patients who did not respond to nonoperative treatment. The predominant symptoms reported were pain and weakness, while only internal rotation motion was partially limited. Nonoperative treatment consisted of at least three months of selective rest, activity modification, physiotherapy, and administration of nonsteroidal anti-inflammatory medications and corticosteroid injections.
The arthroscopic group included twenty men and two women with an average age of 54.7 years (range, forty-six to seventy-four years) at the time of surgery. There were fourteen right shoulders, and the dominant side was involved in fourteen patients (64%). In thirteen patients (59%), the onset was traumatic followed by functional impairment of the affected arm. The open group included thirteen men with an average age of 49.5 years (range, twenty-two to sixty-two years) at the time of surgery. There were nine right shoulders, and the dominant side was involved in nine patients (69%). In eight patients (62%), the onset was traumatic, followed by functional impairment of the affected arm.
Preoperative Evaluation
Preoperatively, all of the patients were evaluated with the Constant-Murley score21. This consisted of individual scores for pain (15 points maximum, with 0 indicating severe pain and 15 indicating no pain), activity (20 points maximum, with 0 indicating a poorer score and 20 indicating a better score), range of motion (40 points maximum, with 0 indicating a poorer score and 40 indicating a better score), and strength (i.e., pounds of abduction force with the shoulder at 90° of elevation in the plane of the scapula, measured once per patient with a handheld mechanical dynamometer) (25 points maximum, with 0 indicating a poorer score and 25 indicating a better score).
Subscapularis muscle-testing included the lift-off test and the belly-press test, as described by Gerber et al.1,2, for each patient. For both tests, the results were considered to be intermediary or asymmetric when they were not positive but were weaker or incomplete in comparison with the opposite side1,2. In the arthroscopic group, preoperatively, the lift-off test was positive in four patients and asymmetric in two patients but was not performed on sixteen patients due to restricted painful internal rotation. In the open-surgery group, preoperatively, the lift-off test was negative in one patient, positive in two patients, and asymmetric in three patients but was not performed on seven patients due to the same reason. The preoperative belly-press test, in the arthroscopic group, was positive in twelve patients, asymmetric in four, and negative in six. In the open group, this test was positive in six patients, asymmetric in four, and negative in three.
All shoulders were assessed with the use of standard radiographs, and none of them showed evidence of glenohumeral osteoarthritis. Nineteen patients had preoperative CTA (eleven in the arthroscopic group and eight in the open group) and nineteen patients had MRI (twelve in the arthroscopic group and seven in the open group) to diagnose the subscapularis lesions (three patients had both examinations). The size of the tear was determined on the preoperative imaging study and confirmed by intraoperative assessment. According to Lafosse et al.7 and Edwards et al.6, subscapularis lesions are classified into three groups by dividing the tendon into thirds from superior to inferior (Table I). In the arthroscopic group, four tears (18%) were limited to the superior one-third of the subscapularis tendon, six tears (27%) involved the superior two-thirds of the subscapularis tendon, and twelve tears (55%) involved the complete height of the subscapularis tendon. In the open group, three tears (23%) were limited to the superior one-third of the subscapularis tendon, six tears (46%) involved the superior two-thirds of the subscapularis tendon, and four tears (31%) involved the complete height of the subscapularis tendon. The long head of the biceps tendon was subluxed or dislocated in eighteen patients in the arthroscopic group and in twelve patients in the open group. There was no supraspinatus or infraspinatus tendon tear in either group on the basis of the preoperative imaging and as confirmed intraoperatively.
For each muscular component, the degree of fatty infiltration was graded according to the classification system of Goutallier et al.22. The classification system of Goutallier et al. (modified by Fuchs et al.23 for MRI) was used to evaluate both the CTAs and the T1-weighted sequences of the MRIs23 (Table I). In the arthroscopic group, nineteen shoulders (86%) showed stage-0 or stage-1 fatty infiltration of the subscapularis muscle (which is considered to be a normal condition22,23) and three shoulders (14%) demonstrated stage-2 fatty infiltration. In the open-surgery group, eleven shoulders (85%) showed stage-0 or stage-1 fatty infiltration22,23 and two shoulders (15%) demonstrated stage-2 fatty infiltration.
The two groups were comparable in terms of sex, age, preoperative Constant-Murley score, preoperative fatty infiltration, and subscapularis tendon testing (Tables I and II).
Surgical Technique
All patients received an interscalene block and general anesthesia. For arthroscopic repair, patients were placed in the beach-chair position. In fourteen cases of retracted subscapularis tendon tears, a soft-tissue release was performed as traction sutures were placed to mobilize the tendon3. The subscapularis tendon reinsertion was done from inferior to superior. Placement of anchors (nonabsorbable in two cases and absorbable in twenty cases; Mitek Worldwide, a division of Ethicon, Westwood, Massachusetts), with two number-2 nonabsorbable sutures, was followed by suture passage with use of a clever hook (T.A.G. Medical Products, Kibbutz Gaaton, Israel) through the torn tendon from the anterior portal. Horizontal mattress suturing was carried out. Depending on the size of the lesion, one anchor was used for the limited superior tears and two or more for the larger lesions. Supraspinatus and infraspinatus tendons were inspected and palpated from articular and bursal sides, and they showed no lesions.
For all patients who had open repair, a deltopectoral surgical approach was used. When the subscapularis lesion was not apparent, an arthrotomy was performed to open the rotator interval and the bicipital groove. The torn subscapularis tendon required release and mobilization in seven patients. The subscapularis tendon reinsertion was done with nonabsorbable anchors (Mitek Worldwide) in all patients, and with two number-2 nonabsorbable sutures. Horizontal mattress suturing was carried out. Depending on the size of the lesion, one anchor was used for the limited superior tears and two or more anchors were used for the larger lesions.
In both groups, the decision to perform a biceps tenodesis (in nineteen cases in the arthroscopic group and thirteen cases in the open group) was based on the recommendation of Edwards et al.6, who found that biceps tenodesis improved significantly the results of subscapularis repair, independently of the preoperative condition of the biceps tendon. The decision to perform a tenotomy (in three patients in the arthroscopic group and none in the open group) was made when patients had major delamination of the biceps tendon6. In all patients in our study, the biceps tendons were either tenodesed or tenotomized.
After surgery, the arm was supported in a sling at 20° of abduction for six weeks. Passive-motion exercises were initiated on the first postoperative day, and, when possible, hydrotherapy was initiated after skin healing. For the security of the subscapularis repair, passive external rotation was limited to 0° for six weeks. Active shoulder motion was allowed after six weeks. Patients were not allowed to perform any strengthening or strenuous work for three months after the surgery. Light sports and demanding activities were allowed after six months.
Postoperative Evaluation
The follow-up evaluation was carried out at least twenty-four months later by an independent evaluator who did not participate in the surgical procedures. Patients were evaluated on the basis of the shoulder subjective value (SSV), the Constant-Murley score21, and the performance of the same subscapularis muscle testing that had been performed preoperatively.
At a minimum of six months after surgery, anatomic controls were performed by analyzing postoperative MRI in all patients of the open group and nineteen patients of the arthroscopic group. A CTA was performed in the remaining three patients of the arthroscopic group. Postoperative MRI protocol consisted of five sequences, including three fat-suppressed T2-weighted sequences in coronal, sagittal oblique, and transverse planes and two T1-weighted sequences in transverse and sagittal oblique planes. Postoperative imaging control was independently reviewed by a radiologist trained in musculoskeletal pathology (Y.C.). The tendon healing on the footprint, the thickness of the healed tendon, and the subscapularis fatty infiltration (according to the classification system of Goutallier et al.22) were the analyzed criteria. In accordance with the classification system of Sugaya et al.20 in relation to the healing of the repaired supraspinatus tendon, we considered the subscapularis tendon to be “thin” when its thickness was ≤3 mm.
Statistical Analysis
Statistical analysis was performed with specialized software. For each group, we compared continuous variables using the Student t test or linear regression. We compared categorical data using the chi-squared test. Multiple-regression analysis was used to evaluate the relative influence of clinical factors on the clinical outcome. For all tests, a probability value (p) of <0.05 was considered significant.
Source of Funding
There was no external funding source for this study.
Clinical Results
The mean duration of the clinical follow-up in the arthroscopic group was 35.7 months (range, twenty-five to forty-nine months) and twenty-three months (range, six to thirty-six months) for postoperative imaging. The mean duration was 47.8 months (range, thirty-six to fifty-seven months) for both clinical and imaging follow-up in the open group. The mean subjective shoulder value was 77.5% (range, fifty to ninety-five) in the arthroscopic group and 89.8% (range, sixty to 100) in the open group. The preoperative and postoperative ranges of shoulder motion and Constant-Murley scores are shown in Table III.
Postoperative Rotator Cuff Testing
Concerning the lift-off test, comparisons with preoperative data were not possible due to preoperative limitations in performing this test. Postoperatively, improvement in internal rotation enabled the performance of the lift-off test in all cases (Table IV). In the arthroscopic group, the percentage of patients who had a negative belly-press test improved from 28% preoperatively to 55% postoperatively and the percentage of patients who had an asymmetric belly-press test varied from 18% preoperatively to 32% postoperatively. In the open group, the percentage of patients who had a negative belly-press test improved from 23% preoperatively to 62% postoperatively and the percentage of patients who had an asymmetric belly-press test varied from 31% preoperatively to 23% postoperatively. The improvement in the belly-press test (Table II) was significant (p < 0.05) for both groups. In no patient did the lift-off test or the belly-press test became positive after surgery.
Anatomical results
The subscapularis tendons in the arthroscopic repair group healed with normal thickness (Fig. 1) in eleven patients (50%) and with a thinner thickness (Fig. 2) in eight patients (36%). Three patients (14%) had a partial retear that was limited to the superior one-third of the subscapularis tendon. The subscapularis tendons in the open repair group healed with normal thickness (Fig. 1) in eleven patients (85%), and with a thinner thickness (Fig. 2) in one patient (8%). One patient (8%) had a retear of the subscapularis tendon.
Assessment of MRIs and CTAs showed progression of fatty infiltration of the subscapularis muscle in 55% of the arthroscopic group. The progression was from stage 0 to 1 in one patient, from stage 0 to 2 in two patients, from stage 1 to 2 in seven patients, from stage 1 to 3 in one patient, and from stage 2 to 3 in one patient. In the open group, progression of fatty infiltration was noted in 62% of the patients (Table III). The progression was from stage 0 to 1 in four patients and from stage 1 to 2 in four patients.
In ten patients (45%) of the arthroscopic group and six patients (46%) of the open group, a localized severe fatty infiltration (stages 3 and 4) was found at the superior and deep part of the subscapularis muscle and was not seen preoperatively (Figs. 3 and 4) (Table III).
Anatomo-Clinical Correlation
The quality of the tendon-healing had no significant influence on the subjective (p = 0.7) and functional outcomes (p = 0.5). This parameter had no influence on the lift-off test (p = 0.17) or the belly-press test (p = 0.06), despite the fact that a thinner tendon was present more frequently in patients who had a positive or asymmetric belly-press test.
The postoperative fatty infiltration of the subscapularis muscle had no significant influence on the subjective (p = 0.9) and functional outcomes (p = 0.4) or on the results of the subscapularis tests (belly-press test [p = 0.20] or lift-off test [p = 0.77]). The presence of a localized severe subscapularis fatty infiltration was more frequent in patients who had large-size lesions, especially complete ruptures of the subscapularis (p < 0.05). This specific aspect did not show any influence on the postoperative Constant-Murley score (p = 0.66). No complication or infection was identified.
Arthroscopic repair of an isolated tear of the subscapularis tendon provides reliable clinical and structural outcomes. We observed improvement of the subjective shoulder value and functional outcome as assessed with use of the Constant-Murley score. The result of subscapularis testing improves but remains impaired in some patients. From an anatomical point of view, the tendon generally heals with a normal or thinner thickness. In addition, progression of subscapularis fatty infiltration was observed and in half of the patients in both groups, a localized severe fatty atrophy of the upper part of the subscapularis was found. These results were comparable with those seen after open repair. However, the subjective shoulder value and the postoperative strength as measured with the Constant-Murley method of testing were different, in favor of the open group.
Since the report of Gerber et al.2 in 1996 on isolated subscapularis tears, good clinical outcomes have been obtained after open repair2,3,6,8. More recently, satisfactory results have been reported for arthroscopic repair of these tears4,5,7,10. We report similar subscapularis healing rates after arthroscopic and open repair, as compared with the results of previous studies7,8. Quality of the healing was different in favor of the open group; most (84%) of the tendons in that group healed with a normal thickness as compared with the healed tendons (50%) in the arthroscopic group. Healing with a thin tendon was observed after repair of the larger subscapularis tears, which are more frequent in the arthroscopic group. We believe that open repair allows for better release, mobilization, and fixation of the tendon, particularly in the case of a retracted tendon tear, and may account for the observed difference. Sugaya et al.20 demonstrated that the quality of the healing after supraspinatus tendon repair had an influence on the clinical outcome; however, no information is available concerning the subscapularis tendon. If we hypothesize that thinner tendons are more frequent after arthroscopic subscapularis repair and cause poorer clinical outcomes or poorer corrections of the results of subscapularis testing, further investigation to determine the optimal technique for repair is needed.
Postoperative subscapularis testing is positive (indicating pathological or weak results) in about 15% to 20% of patients after an open repair procedure2,6 and after arthroscopic repair24,25. According to Edwards et al.6, the large sizes of subscapularis ruptures and the severe preoperative fatty infiltrations of the subscapularis muscle were related to the persistence of positive subscapularis testing. These conditions are well known to be prognostic factors of failure after rotator cuff tendon repair. In our series, severe preoperative fatty infiltration of the subscapularis muscle was not present, and complete retear was not observed, except after one open repair. Lafosse et al.7 reported an improvement in the internal rotation strength after arthroscopic repair, but they did not report details of postoperative subscapularis testing. Adams et al.25 reported a relationship between incomplete return of subscapularis function postoperatively and the preoperative extent of the subscapularis tendon tear. We could not demonstrate that the quality of healing influenced the belly-press test, despite the fact that the percentage of thinner tendon was higher in patients with asymmetric or positive subscapularis testing postoperatively.
Using MRI, Fuchs et al.8 reported a progression of muscle fatty infiltration after an open tendon lesion repair for supraspinatus and subscapularis muscles. However, for the subscapularis, progression of fatty infiltration was not significant from preoperative to postoperative observations. Using CTA images, Lafosse et al.7 did not observe fatty infiltration progression in any of the patients in their series after arthroscopic repair. In our study, there was a similar progression of subscapularis fatty infiltration in both groups. While the cause of the fatty infiltration progression was not related to the lack of tendon-healing, the cause remains unclear. Our results confirm those of Fuchs et al.8 who stated that tendon rupture and then tendon repair negatively influenced the fatty infiltration of the muscle with the tendon repair. Furthermore, our results demonstrated that there was a particular aspect of fatty infiltration, localized on a limited area of the subscapularis muscle, that did not exist preoperatively. This specific fatty infiltration was always severe, graded 3 or 4 according to the classification system of Goutallier et al.22, and was located at the level of the subscapularis tendon fiber, leaving intact the inferior part of the subscapularis, the humeral insertion of which is purely muscular26 (Fig. 4). Observed with similar frequency in the arthroscopic and open groups, this anatomical finding was correlated to the size of the preoperative subscapularis tear and did not influence the clinical outcome According to Gerber et al.27, who supposed that the traction of the tendon repair induced irreversible changes in the muscle concerned, we hypothesized that this specific postoperative fatty infiltration could be the consequence of high tension in the subscapularis tendon repair. Another possible cause is injury of the subscapularis nerve on the anterior muscular belly during tendon dissection and release.
Given that an isolated lesion of the subscapularis is rare and that large sample sizes are difficult to obtain, our study presented limitations responsible for the weakness of the statistical analysis. We did not statistically compare the two groups due to the patient inclusion design and sample sizes. The open group was used for comparison, despite the fact that a number of patients were lost during the follow-up. In addition, the time for the imaging control was variable in the arthroscopic group (from six to thirty-six months) and between the two groups (twenty-three versus forty-seven months) and represents another bias. Finally, the biceps tenodesis that was performed in conjunction with the subscapularis repair can be considered as a co-intervention, the effect of which cannot be determined.
In summary, we evaluated the outcome of arthroscopic subscapularis tendon repair and compared it with the outcome in a cohort of patients who were treated with open repair. Arthroscopic repair of isolated subscapularis tears improved shoulder function. Although the tendon-healing rate was high, clinical correction shown by the lift-off and belly-press tests remained incompletely corrected. Postoperatively, a moderate fatty infiltration progression in the subscapularis muscle was observed in more than half of the patients, but it did not influence the clinical outcomes or the results of subscapularis testing. A particular severe fatty infiltration of the subscapularis muscle, which was localized at the superior and deep part of the muscle and which did not exist preoperatively, was also observed, but it had no influence on clinical outcome or subscapularis testing. Although most clinical and structural outcomes after arthroscopic repair were generally comparable with those that occurred after open repair, the strength and subjective outcomes were better after open repair. Further study of patient selection and repair technique for isolated subscapularis tendon tears is warranted.
Note: The authors thank Jean François Safar, MD, for his assistance in performing the statistical analysis in this study.
Disclosure: None 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 any aspect of this work. One or more of the authors, or his or her institution, has had a financial relationship, in the thirty-six months prior to submission of this work, with an entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written in this work. 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.