➢ With the advent of magnetic resonance imaging (MRI) and arthroscopy, the diagnosis and treatment of subscapularis tendon tears have been increasingly reported.
➢ The patient history and physical examination findings are important to heighten the suspicion of a subscapularis tendon tear and to distinguish acute traumatic tears from degenerative tears.
➢ While nonoperative treatment is the primary option for subscapularis tendon tears, surgery may be considered for acute traumatic tears and tears for which nonoperative treatment has failed.
➢ Both open and arthroscopic subscapularis tendon repairs have provided good clinical outcomes.
➢ For arthroscopic repair, maximizing visualization is paramount to recognizing the subscapularis tendon tear and allowing anatomical repair.
The published literature on tears of the subscapularis tendon is relatively sparse when compared with that on tears of the supraspinatus or infraspinatus tendons. Nevertheless, subscapularis tendon tears are in fact extremely common. Bennett, in a retrospective study, reported that the overall prevalence of subscapularis tears among patients undergoing shoulder arthroscopy was 28% (46 of 165) and was even higher in the subset of patients who had rotator cuff tears (35%; 46 of 130)1. Other authors have reported similar or even higher rates of up to 49%2-4.
Historically, the subscapularis tendon tear was first identified by Smith in 18345,6. However, in modern orthopaedics, Gerber et al. highlighted the presence of subscapularis tendon tears in 2 reports involving 16 isolated tears7,8. In many reports, subscapularis tendon tears have been reported as isolated lesions7-11, likely because of their relative rarity and apparently unique anatomy, with the subscapularis being “separate” from the posterosuperior part of the rotator cuff (comprising the supraspinatus, infraspinatus, and teres minor). In addition, many reports were based on open surgical approaches, in which the prevalence of partial tears was likely under-reported12.
However, with the advent of shoulder arthroscopy and imaging modalities, the identification and recognition of subscapularis tendon tears has substantially improved, and there is a general consensus that, in fact, subscapularis tendon tears are relatively common13. In what we believe to be the largest reported series of subscapularis tendon tears, Toussaint et al. reported on 208 patients with subscapularis tendon tears who were included in a multicenter study by the French Arthroscopy Society14. That report (as well as the report by Warner et al.15) demonstrated that subscapularis tendon tears are common but most often occur in association with supraspinatus and/or infraspinatus tendon tears (i.e., anterosuperior tears). Isolated tears of the subscapularis still remain relatively uncommon and represent approximately 10% to 25% of subscapularis tendon lesions in most series3,4,14,16.
Interestingly, subscapularis tendon tears appear to be more prevalent in patients with a traumatic onset of symptoms17. For example, among 259 consecutive patients with acute traumatic shoulder pain, Aagaard et al. found that 38 (63.3%) of 60 patients with a full-thickness rotator cuff tendon tear had a tear involving the subscapularis18.
Whether isolated or combined, subscapularis tendon tears are now recognized as an important component of rotator cuff tendon tears. Furthermore, ignoring a subscapularis tendon tear may cause substantial dysfunction even after the repair of a tear involving the posterosuperior part of the rotator cuff15.
Surgical Anatomy and Function of the Subscapularis Tendon
Recently, there has been a renewed interest in the anatomy of the rotator cuff and its insertions. Classically, the subscapularis has been described as originating from the ventral surface of the scapular body, with both muscular and tendinous insertions onto the lesser tuberosity and neck of the humerus. The morphology of the subscapularis tendon footprint is broad proximally (i.e., tendinous insertion) and tapered distally (i.e., muscular insertion), with previous authors having described the shape of this footprint in a variety of ways (e.g., comma19,20, human ear21, state of Nevada22, pear23). Gross anatomic dissections of cadavers have demonstrated that the insertion of the subscapularis has a transverse width ranging from 16 to 20 mm (with the broadest region being located superiorly) and a longitudinal length ranging from 25 to 40 mm19-22.
More recent studies have highlighted the complexity of the subscapularis tendon insertion and have identified a number of anatomic insertions that are separate from the main tendinous attachment24. These include the medial capsuloligamentous insertion20,23, the inferior musculocapsular insertion23, the “lateral hood” (intact lateral tissue commonly observed in association with upper subscapularis tendon tears)25, and the superior “tendinous slip,” which is attached to the fovea capitis of the humerus and contributes to biceps stability (along with the superior glenohumeral ligament [SGHL] and coracohumeral ligament [CHL])26. Furthermore, Mochizuki et al. also reported variations of the supraspinatus tendon insertion in which the leading edge of the supraspinatus tendon extends over the biceps and inserts onto the superiormost region of the lesser tuberosity27. These findings have collectively improved our understanding of the anatomy and relationship of the subscapularis and biceps tendons.
Functionally, the subscapularis is the largest and most powerful rotator cuff muscle28. In isolation, the subscapularis muscle predominantly functions as an internal rotator of the humerus; however, it is also known to contribute to dynamic stability of the glenohumeral joint in combination with the posterior rotator cuff muscles by creating a transverse-plane force couple28,29. Therefore, disruption of the subscapularis tendon can cause not only pain but also dysfunction of the shoulder.
The majority of tears of the subscapularis tendon begin at the articular surface of the superior insertion of the tendon and extend inferiorly4,25. At that insertion, there is an intimate relationship between the upper border of the subscapularis tendon and the medial sling of the biceps tendon30. Disruption of the subscapularis tendon usually leads to concomitant incompetence of the medial sling of the biceps (i.e., the SGHL and CHL) and/or the transverse humeral ligament, with subsequent biceps tendon instability1,2,26,31-33. In the clinical study by Arai et al., there were no cases in which the subscapularis tendon was intact when the biceps tendon was unstable2. Therefore, biceps tendon instability should raise the suspicion of a concomitant subscapularis tendon tear and should be addressed when subscapularis tendon repair is performed.
Concomitant disruption of the medial sling of the biceps tendon as a part of a subscapularis tendon tear results in a pathoanatomical landmark (the “comma sign”) as seen arthroscopically. The comma sign is considered to be present when the medial sling of the biceps tendon remains in continuity with the upper lateral border of the torn subscapularis tendon34,35 (Fig. 1). Therefore, identifying this landmark can be extremely useful when reducing the subscapularis tendon tear, especially when the tendon is medially retracted and adhesively associated with the anterior structures (e.g., inner deltoid fascia, conjoint tendon).
Furthermore, because the leading edge of the torn supraspinatus tendon is often attached to this comma tissue in patients with combined subscapularis and supraspinatus tears27, anatomical reduction of the comma sign and subscapularis tendon repair will assist in the reduction of the leading edge of the supraspinatus tendon, simplifying repair of the superior rotator cuff13.
The etiological factors associated with subscapularis tendon tears are similar to those associated with supraspinatus tendon tears. While most tears are multifactorial in nature, chronic tears in older individuals involve at least some degree of degenerative change, whereas younger patients experience more acute, traumatic tears36-38.
The relationship between subscapularis tendon abnormality and the presence of subcoracoid impingement has long been investigated39-46. The concept of subcoracoid (or coracoid) impingement originally was introduced by Gerber et al. in 198540, followed by Dines et al. in 199039. More recently, Lo et al. defined subcoracoid impingement as mechanical contact between the coracoid and the lesser tuberosity and defined associated subcoracoid stenosis as narrowing of the coracohumeral space, emphasizing their relationship with anterosuperior rotator cuff tears41.
Although Richards et al.43 demonstrated a correlation between a narrowed coracohumeral distance and subscapularis tendon tears, it is still not clear whether so-called subcoracoid impingement is truly an etiological factor or the consequence of a large rotator cuff tear involving the subscapularis tendon that results in anterior and superior translation of the humeral head42,45. Furthermore, very limited evidence is available with regard to the clinical outcome after coracoplasty44,46,47. Therefore, further studies with higher levels of evidence are necessary.
In contrast, acute traumatic subscapularis tendon tears are differentiated from degenerative anterosuperior rotator cuff tears and typically are reported separately as isolated ruptures of the subscapularis tendon7,10,17,48-51. In general, acute subscapularis tendon tears are associated with a traumatic episode with the arm forced into an abducted, extended, and externally rotated position17. These tears may be associated with anterior shoulder instability in older patients or lesser tuberosity avulsion in younger patients52,53. Although sometimes difficult to distinguish from degenerative or acute-on-chronic tears, acute tears may benefit from early surgical intervention17,49,50.
In general, rotator cuff tears are commonly classified by size (e.g., small, medium, large, massive)54,55 or by the number of tendons involved56. However, a number of authors have developed classification systems based specifically on the subscapularis tendon tear. Lafosse et al. proposed a unique classification system that incorporated both the pattern of tear progression as well as the surgical strategy to repair4 (Table I). In this classification system, subscapularis tendon tears were divided into 5 types, with type-I tears being partial tears of the superior one-third and types II through V being complete full-thickness tears with progressive involvement of the entire length of the footprint. Type-V tears were considered to be the most severe, with static anterior subluxation of the humeral head and stage-3 (or greater) fatty degeneration of the subscapularis muscle belly.
Recently, Yoo et al. proposed a new classification system based on the 3-dimensional insertional anatomy of the subscapularis tendon25. Through both cadaveric dissection and arthroscopic evaluation, they demonstrated that the subscapularis tendon had a 4-facet insertional anatomy along the curvature of the anterior part of the humeral head. This classification system divided subscapularis tears into 5 categories based on the involvement of the 4 facets of the subscapularis tendon insertion.
While these classification systems have been useful for understanding the anatomy and surgical strategy related to subscapularis tendon tearing, neither system has been validated when evaluated under strict scientific rigor57.
Recently, there has been an explosion of interest in the natural history and nonoperative treatment of rotator cuff tendon tears58,59. While the natural history of rotator cuff tears has improved our understanding of the role of nonoperative treatment and the indications for surgical repair, there is very little information specifically related to tears of the subscapularis tendon. For example, Maman et al. evaluated nonoperatively treated symptomatic rotator cuff tears with magnetic resonance imaging (MRI) scans (made >6 months apart) and reported that age (>60 years), a full-thickness tear, and fatty infiltration were factors that were associated with tear progression60. Those authors evaluated a wide range of tears, including some subscapularis tendon tears, although the number of subscapularis tendon tears was relatively small. While it is not clear whether this information can be applied to subscapularis tendon tears in general, it is reasonable to suggest that subscapularis tendon tears likely increase in size with time and that these factors may also accelerate tear progression.
The clinical history of patients with subscapularis tendon tears is difficult to differentiate from that of patients with posterosuperior rotator cuff tears. However, an acute history of trauma, particularly with the arm forced into abduction, extension, and external rotation, should raise the suspicion of a subscapularis tendon tear. While classically subscapularis abnormalities often cause anterior pain, such pain is not specific for subscapularis tears and may be the result of a concomitant biceps lesion or other associated abnormality. Patients also may complain of weakness of internal rotation (e.g., while tucking in a shirt), which may be more specific to subscapularis dysfunction.
A number of physical examination tests have been described to evaluate the function of the subscapularis tendon. Importantly, all of these tests are designed in part to isolate the function of the subscapularis muscle from that of the pectoralis major muscle. The classic lift-off test7 and belly-press test8 were introduced by Gerber et al. and are commonly used together as many patients cannot perform the lift-off test because of restricted range of motion or pain. The belly-press test is known to activate the upper subscapularis and therefore is thought to be more sensitive for the detection of upper subscapularis tendon tears7,61. However, the belly-press test may be negative even if the upper 50% of the subscapularis tendon is torn3.
The bear-hug test, described by Barth et al.3, was proposed to be more sensitive for upper subscapularis tendon tears. The test is considered to be positive if the patient demonstrates pain or weakness.
In addition, the presence of increased external rotation in the injured shoulder as compared with the contralateral shoulder should be assessed8 and biceps provocative tests (e.g., Speed test62, Yergason test63) should be performed.
While a number of these tests have been evaluated for accuracy and have demonstrated reasonable utility (with Barth et al.3 reporting an accuracy of 77.8% for the lift-off test, 80.9% for the belly-press test, and 82.4% for the bear-hug test), not all studies have demonstrated uniform accuracy. In general, while the specificity of these tests has been reported to be relatively high (range, 79% to 100%), the sensitivity has been lower and more variable (range, 12% to 60%)3,64,65.
Standard radiographs may be utilized to assess for proximal migration of the humeral head as an indicator of a massive rotator cuff tear or cuff tear arthropathy66-69. However, this finding is neither specific nor sensitive for a subscapularis tendon tear. While static anterior subluxation may indicate a torn subscapularis tendon, it may not lead to narrowing of the acromiohumeral interval70.
Ultrasound recently has gained acceptance for the detection of full-thickness tears of the rotator cuff. Although the technique is highly observer-dependent, when performed by an experienced physician, it can be as accurate as MRI for the detection of supraspinatus and infraspinatus tears71,72. However, its utility for the detection of subscapularis tendon tears has not been thoroughly evaluated. Most studies have been restricted to small subsets of patients73,74, and, even when used by an expert, ultrasound appears to be inferior to MRI for the detection of subscapularis tendon tears, particularly partial-thickness or upper subscapularis tendon tears74.
MRI or magnetic resonance arthrography (MRA) are the diagnostic imaging modalities of choice and have been reported to have sensitivities and specificities of 36% to 94.7% and 82.4% to 100%, respectively75-79. In particular, MRI scans should be evaluated for 4 criteria; specifically, axial images should be evaluated for (1) tearing of the subscapularis tendon from the lesser tuberosity and (2) subluxation or dislocation of the long head of the biceps tendon, and sagittal oblique images should be evaluated for (3) atrophy of the subscapularis muscle and (4) tearing of the subscapularis tendon from the lesser tuberosity76.
The gold standard for the diagnosis of a subscapularis tendon tear remains arthroscopic evaluation57,83. Diagnostic arthroscopy may be performed with use of only a 30° arthroscope through a posterior portal; however, this limited technique has been associated with fair to moderate interobserver and intraobserver reliability57. To improve visualization of the subscapularis tendon, placing the arm in forward flexion and internal rotation in addition to applying a posteriorly directed force can draw the fibers of the subscapularis away from the humeral head, revealing the footprint (Fig. 2-A). Some authors have also proposed the use of a 70° arthroscope (Fig. 2-B) or a lateral viewing portal to further improve visualization, particularly in cases of complete subscapularis tendon tears that extend inferiorly.
Arthroscopic evaluation should include the assessment of the footprint, the medial sling, and the position of the biceps tendon. A spectrum of abnormality may be present, including partial-thickness tears, full-thickness upper subscapularis tendon tears (Fig. 3-A), full-thickness complete subscapularis tendon tears with minimal retraction (Fig. 3-B), and full-thickness complete tears with retraction (Fig. 3-C).
In either case, the identification of the “comma sign” (Fig. 1) can be useful for identifying the presence of a tear and determining the borders of the subscapularis tendon. This “comma sign” is particularly helpful when evaluating retracted or adhesively associated tears that are scarred to the inner deltoid fascia. In this scenario, extensive releases may be required to obtain sufficient excursion of the subscapularis tendon for repair to bone.
The nonoperative treatment of a subscapularis tendon tear is similar to that of a standard posterosuperior rotator cuff tear. In general, conservative treatment may include cold and/or heat therapy, pain medications, nonsteroidal anti-inflammatory medications, injections (e.g., subacromial or glenohumeral), and physical therapy. To our knowledge, no protocol has been specifically designed to address the subscapularis tendon or to determine the results of nonoperative treatment.
In general, conservative treatment of a chronic rotator cuff tendon tear has been reported to be successful in 70% to 75% of patients, even after >10 years of follow-up84,85. Interestingly, in the study by Moosmayer et al., while early surgical intervention led to significantly greater improvement in outcomes than physiotherapy treatment, the differences were small and may be below clinical importance86. However, more than one-third of the patients who had been managed with physiotherapy demonstrated an increasing tear size, with inferior outcomes86. As there is minimal evidence that nonoperative treatment delays tear progression, the risk of tear progression persists even in patients for whom conservative treatment is clinically successful60,87. However, the majority of this information was obtained from studies of patients with posterosuperior rotator cuff tears88,89, and therefore it is unclear how these findings relate to patients with concomitant or isolated subscapularis tendon tears.
Indications for Operative Intervention
The general indications for surgical intervention for the treatment of a subscapularis tear are essentially the same as those for any rotator cuff repair. Patients who fail to improve after a trial of conservative management including physical therapy, activity modification, and oral anti-inflammatory medications are reasonable candidates for surgical intervention. However, while we are not aware of any critical evidence, acute traumatic subscapularis tendon tears should likely be treated urgently to prevent further retraction and muscle atrophy, similar to acute tears of the posterosuperior part of the rotator cuff.
Because of limited evidence, the contraindications for subscapularis tendon repair follow the contraindications for posterosuperior rotator cuff repair. Although there are no clear guidelines, primary repair may not be achievable or the risk of postoperative retearing may be too high if both the length and width of the tear exceed 40 mm90 or if there is radiographic evidence of a narrowed acromiohumeral interval69 and/or severe fatty infiltration (Goutallier stage 3 or 4)81. However, as reported by Burkhart et al., even patients with Goutallier stage-3 or 4 fatty infiltration can have clinical improvement after arthroscopic repair91. Furthermore, attempting to repair the subscapularis in a shoulder with a massive tear sometimes may be necessary in order to balance the so-called force couples about the glenohumeral joint29,92.
Debridement Versus Repair
While repair of subscapularis tendon tears can restore anatomy, debridement of subscapularis tendon tears has been reported11,93,94. In 2006, Edwards et al.11 reported on 11 carefully selected patients with isolated subscapularis tendon tears. Following debridement with or without biceps tenotomy, 9 (82%) of the 11 patients were satisfied with the outcome and the mean Constant score improved from 49 to 80. Similarly, in a limited prospective randomized trial, Randelli et al. compared arthroscopic repair with debridement for the treatment of the small partial-thickness upper subscapularis component of an anterosuperior tear94. The authors reported no significant difference in clinical outcomes among the 20 patients who were available for follow-up.
Importantly, while it appears that debridement of a subscapularis tendon tear can improve outcomes, in many studies11,93,94, concomitant procedures such as biceps tenotomy or tenodesis and repair of the posterosuperior part of the rotator cuff were simultaneously performed, complicating interpretation of the data. While the evidence is limited, debridement appears to be a viable option when a repair is not achievable or may not be necessary. In any patient with a subscapularis tendon tear, it is important to determine preoperatively if the tear is symptomatic and to address any concomitant abnormalities, if present.
Open Versus Arthroscopic Repair
While the majority of recent studies have investigated arthroscopic subscapularis tendon repair, a number of articles have shown equivalent clinical outcomes following open subscapularis tendon repair7,8,10,50,95-98. Some authors have suggested that open surgical repair may be preferred for the treatment of isolated lesions associated with severe tendon retraction or tears isolated to or involving a major portion of the lower subscapularis tendon insertion99. However, particularly in cases involving concomitant large posterosuperior tears, a combined open approach (i.e., a combined deltopectoral and transdeltoid approach) may be required for sufficient visualization15. Furthermore, many partial tears or upper subscapularis tendon tears may be missed during an open approach; therefore, routine investigation of the bicipital groove and subscapularis may be required to identify such “hidden lesions.”12
On the other hand, visualization and instrumentation can be challenging when performing arthroscopic subscapularis tendon repair. Some authors have recommended the use of a 70° arthroscope in combination with a standard 30° arthroscope to maximize visualization of the entire subscapularis footprint, particularly its medial and inferior portions30,100. Because of its relative complexities, arthroscopic subscapularis repair should only be attempted when the arthroscopist is comfortable with standard rotator cuff repair techniques.
Postoperative Protocol After Repair
Most authors have recommended sling immobilization for 4 to 6 weeks postoperatively7,49,94,96,101. During this period, hand, wrist, and elbow motions are encouraged. Passive external rotation is permitted but is limited to 0° or as determined intraoperatively. Forward elevation exercises are started once immobilization ends, with progression from active-assisted to active range of motion. Rotator cuff and deltoid strengthening are started 12 weeks postoperatively, with progressive return to functional activities and sports.
Clinical Outcomes Following Subscapularis Repair
Improvement of Clinical Measures
In general, the majority of patients have demonstrated reasonable improvement in clinical measures regardless of surgical approach (i.e., open or arthroscopic). The results from the literature are summarized in Table II4,7,9,10,14-16,49,82,92-96,98,101-117.
Evaluations of tendon integrity with use of computed tomographic (CT) arthrography, MRI, or ultrasound have demonstrated healing rates of 45.5% to 100% following subscapularis tendon repair (Table II). While a number of factors are associated with healing (e.g., repair technique, size of tear), subscapularis healing rates appear to be equivalent to healing rates for the posterosuperior part of the rotator cuff (20% to 90%)118. While the subscapularis tendon may heal following repair, the recovery of muscle atrophy and fatty infiltration has been more variable. While some studies82 have demonstrated improvement in terms of fatty infiltration and muscle atrophy in the subscapularis muscle following arthroscopic repair, the majority of studies have demonstrated either no progression or worsening of these findings with longer-term follow-up4,14,96,101,108,110,113,117. Lafosse et al. demonstrated no progression of fatty infiltration on postoperative CT arthrograms after arthroscopic repair of isolated subscapularis tendon tears in a study of 17 consecutive patients with a mean duration of follow-up of 29 months4. However, with longer-term follow-up (range, 35 to 75 months), other studies have shown no preventive effect or even worsening of fatty infiltration after subscapularis repair, regardless of tendon healing and improved outcome measures (Table II)110,113,117.
Regardless of the degree of anatomical tendon healing and improvement in clinical outcome measures, incomplete return of subscapularis function following surgical repair has been a concern7,95,110. In several studies, although the majority of patients had improvement following surgery, many continued to demonstrate subscapularis dysfunction on clinical examination. For example, persistent positive belly-press tests were reported for 5 (45%) of 11 patients in the study by Gerber et al.7, in 15 (20%) of 76 shoulders in the study by Edwards et al.95, and in 45% of patients in the arthroscopic group and 38% of those in the open group in the study by Nové-Josserand et al.110. Similarly, in those same studies, persistent positive lift-off tests were reported for 5 (45%) of 11 patients7, 17 (28%) of 60 shoulders95, and 59% of patients in the arthroscopic group and 54% of those in the open group110. Collectively, the discrepancy between clinical outcomes and incomplete correction of the results of subscapularis functional tests may be due to irreversible fatty infiltration in the subscapularis muscle. However, this discrepancy is likely multifactorial as Nové-Josserand et al. demonstrated that there was no correlation between postoperative fatty infiltration and the results of subscapularis tests after surgical repair110.
Clinical Care Recommendations
Because of the lack of prospective randomized trials and higher levels of evidence related to subscapularis tendon repair, no Grade-A recommendations for either nonoperative treatment or operative treatment can be made (Table III). However, studies of arthroscopic subscapularis tendon repair, including a number of Level-III studies, have shown consistent clinical improvement. While surgical indications still require clarification, repair of subscapularis tendon tears appears to provide reasonable outcomes.
In conclusion, subscapularis tendon tears are more common than previously reported. With the advent of MRI and arthroscopy, a more dedicated approach is now warranted to assess and treat subscapularis abnormalities. Surgical repair can provide reasonable clinical improvement and tendon healing and therefore should be considered for the treatment of acute traumatic tears or chronic tears that have failed to respond to nonoperative treatment. However, insufficient evidence is currently available to provide definite treatment recommendations. Further investigations of better quality and larger size are necessary to acquire higher levels of evidence.
Investigation performed at the University of Calgary, Calgary, Alberta, Canada
Disclosure: No external funding sources were utilized for this study. On the Disclosure of Potential Conflicts of Interest forms, which are provided with the online version of the article, one or more of the authors checked “yes” to indicate that the author had a relevant financial relationship in the biomedical arena outside the submitted work.
- Copyright © 2017 by The Journal of Bone and Joint Surgery, Incorporated