Proper Identification

One potential complication of subscapularis tears that is very common, and often underreported, is failure to identify these tears in the first place. One cannot treat what one does not diagnose. The diagnostic process begins with a comprehensive history and physical exam. There are several provocative physical examination maneuvers that have been identified which may be helpful in making this diagnosis. In 1996 Greis et al. used electromyography to document the importance of both the upper and lower subscapularis when performing the “lift-off test”: placing the shoulder in internal rotation with the dorsum of the hand placed against the lumbar spine, and then “lifting off.” The authors found that the subscapularis muscle activity was approximately 70% of its maximal contraction during this test, which was significantly more than any other muscle around the shoulder. In 2003, Tokish et al. performed a similar study evaluating the “belly press test” (the palm of the hand is placed on the abdomen during shoulder internal rotation, and held there under resistance). The study found that, similar to the “lift-off test,” the subscapularis (both upper and lower) is activated more than any other muscle about the shoulder, leading the authors to conclude that this was also a good test for evaluation the subscapularis. Additionally, electromyography (EMG) data from the “belly press test” was compared with the lift-off test in 16 patients. They showed that the “belly press” activates the upper subscapularis more, whereas the lift-off test activates the lower subscapularis more, making it important to perform both tests, especially if a partial tear is suspected.

In 2008, Barth et al. described a new physical examination maneuver, known as the “bear hug test,” and assessed the sensitivity and specificity of this test compared with the lift-off and belly press maneuvers. The bear hug test is described as placing the shoulder in internal rotation while the palm is held on the opposite shoulder with the elbow held in a position of maximal anterior translation. The authors examined 68 patients who were scheduled to undergo arthroscopic shoulder surgery, and in all of them performed the lift-off test, the belly press test, and the bear hug test. At the time of surgery, a subscapularis tear was found in 29.4% of patients, with 40% of the tears not suspected by any of the preoperative tests. With regard to the physical examination maneuvers, the bear hug was most sensitive (60%), followed by belly press (40%), and the lift-off test (17.6%) ( Fig. 32.1 ). All of the tests had a high specificity: bear hug (91.7%), belly press (97.9%), and lift-off (100%). The authors concluded that the bear hug test was the best for detecting upper subscapularis tears, and also should be used routinely because it gave the greatest chance at correctly diagnosing a tear. The authors also recommended using all three tests, as multiple positive tests could signify a larger tear. Overall, these studies have shown that, although we do have specific physical examination maneuvers for diagnosing subscapularis tears, these tears may still be missed when relying solely on physical examination (40% according to Barth et al.). For this reason, the first step in preventing complications caused by these injuries is to diagnose them to treat them sufficiently. The examiner should have an appropriate clinical suspicion when evaluating patients and their imaging, and ultimately confirming the presence or absence of pathology during arthroscopic surgery.

Detection of subscapularis dysfunction by the (A) belly press test, (B) bear hug test, and (C) lift-off test.

Technical Pitfalls

Exposure. Subscapularis tears can be treated be both arthroscopically and by open surgery, with the majority being repaired arthroscopically. Repair of the subscapularis can be technically demanding and requires a thorough knowledge of the arthroscopic anatomy of the shoulder. A majority of suscapularis repairs will require concomitant tenodesis of the biceps, as the continuation of the upper subscapularis, superior glenohumeral ligament, and coracohumeral ligament (which are routinely injured with a subscapularis tear) form the medial aspect of the biceps sling and prevent medial subluxation of the long head of the biceps. Note that failure to address any associated biceps pathology will likely lead to continued symptoms because the biceps typically subluxates, or dislocates medially in the case of subscapularis tears. Biceps tenodesis will also improve exposure, as the biceps tendon is often subluxated onto or medial to the lesser tuberosity ( Fig. 32.2 ).

T2 axial magnetic resonance image demonstrating subscapularis tear with medially subluxated biceps tendon.

Additional aspects of exposure for arthroscopic repair of subscapularis tears is the requirement of the surgeon to work in the subcoracoid space, which is subject to soft tissue swelling from fluid extravasation. The best way to avoid excess fluid extravasation is by limiting damage to the subscapularis fascia during debridement and by being expeditious during this portion of the case. More chronic tears may be difficult to identify, and can be visualized by finding the torn medial biceps sling, composed of the coracohumeral ligament, superior glenohumeral ligament, and superior lateral corner of the subscapularis; this has been referred to as the “comma sign” ( Fig. 32.3 ). Depending on the chronicity and degree of retraction of the tendon, release of all attachments to the coracoid base may be required to gain adequate mobilization before repair ( Fig. 32.4 ). The tendon should be released on three sides. Posterior to the tendon, release is performed past the level of the glenoid rim, releasing the anterior capsule. Anteriorly, there may be adhesions to the coracoid which need to be released as well, taking care not to injure the innervating subscapular nerves. Superiorly, release of the rotator interval, superior glenohumeral ligament, and coracohumeral ligament can also increase mobilization ( Fig. 32.5 ). As previously mentioned, retracted tears which require this type of “three-sided” release of the tendon require careful attention to crucial anatomic structures, such as branches of the thoracoacromial artery, specifically when performing dissection beneath the coracoid neck and the subscapular nerves. Release anterior to the subscapularis tendon is largely done under poor visualization, making knowledge of the anatomy extremely important during this step of the procedure. Exposure on the inferior aspect of the tendon is avoided secondary to the proximity of the axillary nerve and posterior circumflex humeral artery.

Arthroscopic image of a retracted subscapularis tear with an arthroscopic grasper on the lower “comma” tissue.

Reduced subscapularis following arthroscopic repair.

Arthroscopic image showing releases around the subscapularis tendon ( SSc ) with release anteriorly and superiorly from the coracoid ( C ) and rotator interval.

Inadequate repair. The difficult technical nature of arthroscopic subscapularis repair can lead to one of the more common complications of this procedure, namely failure of the repair. Successful repair of the subscapularis tendon relies first on adequate visualization of the tendon and the lesser tuberosity footprint. Visualization can be enhanced by use of a 70-degree arthroscope if needed. Additionally, the surgeon can have his or her assistant perform the “posterior lever push,” levering the proximal humerus from anterior to posterior and the distal aspect posterior to anterior to better visualize the front of the shoulder and the subdeltoid space. ^, Internal rotation and shoulder flexion can also aid visualization. Additionally, coracoplasty can and should be performed if there is concern for subcoracoid impingement. Denard et al. recommend performing coracoplasty if the subcoracoid coracohumeral distance is less than 7 mm. During coracoplasty, care should be taken to not dissect medial to the coracoid to avoid injury to neurovascular structures such as branches of the thoracoacromial artery, or potentially the axillary or musculocutaneous nerve.

Anchor placement. Proper anchor position is also essential to adequate repair. To achieve the proper anchor orientation of 30 to 45 degrees of inclination into the lesser tuberosity, the surgeon needs to place his or her hand medially, sometimes getting very close to the patient’s face (the so called “hand on jaw” position). ^, Failure to place the anchor in the proper orientation can result in a weak repair or intraarticular anchor penetration. When performed properly, good outcomes have been reported following isolated subscapularis repair. Lafosse et al. reported on the outcomes of 17 patients following subscapularis repair and found that the structural integrity of the repair was intact in 15 patients at an average 29 months’ follow-up. ^,

Anchor pullout. Anchor pullout is a mode of failure described historically in arthroscopic rotator cuff repair in general, but with newer anchor design is likely declining. ^{, ,} Bensen et al. retrospectively reviewed 269 patients who had 550 metallic suture anchors placed, with an anchor pullout rate of 2.4%. The authors reported an increased risk of pullout for tears larger than 3 cm. Of note, none of the 61 patients who had a subscapularis repair had any reported failures. Barber et al. performed a biomechanical study on the failure mechanisms of newer anchors in 2013. Anchors tested included ReelX (Stryker Endoscopy, San Jose, CA); Footprint Ultra PK (4.5 and 5.5 mm [Smith & Nephew, Andover, MA]); TwinFix (4.5, 5.5, and 6.5 mm made from polyether ether ketone [PEEK], hydroxyapatite, and titanium [Ti] [Smith & Nephew Endoscopy, Andover, MA]); Morphix (2.5 and 5.5 mm [MedShape Solutions, Atlanta, GA]); CrossFT BC (ConMed-Linvatec, Largo, FL); JuggerKnot (1.5 and 2.8 mm [Biomet Sports Medicine, Warsaw, IN]); Healicoil (Smith & Nephew Endoscopy, Andover, MA); Quattro (X, Link, and GL) (Cayenne Medical, Scottsdale, AZ); Healix (Biocryl Rapide [BR], PEEK, and Ti [DePuy Mitek, Raynham, MA]); Twin Loop (3.5 mm, PEEK [Stryker Endoscopy, San Jose, CA]); PressFT (2.1 and 2.6 mm [ConMed Linvatec, Largo FL]); Y-Knot (ConMed Linvatec, Largo FL); Gryphon (BR and PEEK [DePuy Mitek, Raynham, MA]); and Iconix (1, 2, and 3 [Stryker Endoscopy, San Jose, CA]). Their study showed that these newer anchors used for rotator cuff fixation more frequently fail at the eyelet as opposed to pullout from the bone.

Lorbach et al. evaluated the biomechanical strength of 25% and 50% tears of the subscapularis fixed with a single suture anchor in cadaver shoulders. Tears were created and fixed with a double-loaded suture anchor (5.5-mm Bio-Corkscrew with two No. 2 FiberWire [Arthrex, Naples, FL]) creating a double-mattress suture repair. There was no significant difference in anchor displacement during cyclical loading or ultimate failure load. In smaller, 25% tears, most failures were owing to lesser tuberosity fracture and anchor pullout, whereas larger tears typically failed at the tendon–suture interface. This led the authors to conclude that proper suture configuration became more important as the size of the tear increased. Once again, proper technique is paramount and can result in positive outcomes. Lanz et al. reported 2- to 4-year outcomes in 46 patients who underwent subscapularis repair. Patients had an increase in mean Constant score from 46.4 points to 79.9 points. Additionally, subscapularis strength was 92% and external rotation was 96% of the nonoperative shoulder. The overall rerupture rate was 11%.

Nerve Injury

Nerve injury is a potential risk during shoulder arthroscopy in general, typically because of traction on the arm, and is more common in patients who are placed in the lateral decubitus position. Nerve injury can also theoretically occur because of fluid extravasation around the shoulder. Obviously, direct nerve injury is a risk as well. Positioning in the lateral decubitus position typically involves placement of traction on the arm, and can lead to a neurapraxia of the brachial plexus. A retrospective review of 14,329 shoulder arthroscopy cases by the Arthroscopy Association of North America found the prevalence of neurological injuries to be 0.1%.

Portal placement, specifically around the coracoid, can also lead to iatrogenic nerve injury. Matthews et al. dissected 20 cadavers to determine the relationship of the anterior portal to pertinent neurovascular structures. The authors identified danger to the divisions of the brachial plexus medial and inferior to the coracoid. Slightly lateral and inferior to the coracoid lies the musculocutaneous and axillary nerve. The authors recommended placement of the anterior portal directly lateral and adjacent to, or slightly superior to, the coracoid to avoid inadvertently injuring these structures. These relationships are important to be aware of, especially if variation in the traditional portal placement is required to adequately debride the subscapularis footprint at the lesser tuberosity, and for the placement of any anterior accessory portals in order to pass suture for subsequent tendon repair.

Additionally, retracted tears of the subscapularis will require aggressive mobilization to properly reduce the tendon. Lafosse et al. classified subscapularis tears from grade 1 to 5 based on severity of tear and degree of retraction. Grade 1 is defined as a partial tear of the superior third of the tendon, grade 2 is a complete superior third tear, grade 3 is a complete superior two-thirds tear, grade 4 is a complete tear (often with retraction and with less than grade 3 fatty infiltration), and grade 5 is a complete tear with less than grade 3 fatty infiltration. In larger, retracted tears (grade 4 or 5), release of the subscapular nerves may be necessary to avoid traction injury during repair. Despite the risks of nerve injury, it is still very rare in isolated subscapularis repair. Lafosse et al. reported zero nerve injuries in a review of 74 patients who underwent arthroscopic subscapularis repair. Similarly, Saltzman et al. performed a systematic review of isolated subscapularis repairs which included eight studies with 115 total patients. There were no reported nerve injuries in any of the studies. Bradley et al. retrospectively reviewed all patients who presented to their brachial plexus clinic with iatrogenic nerve injury following shoulder surgery and identified 29 nerve injuries in 26 patients from 2000 to 2010. Of note, only one of the patients underwent arthroscopic subscapularis repair. This patient had a radial nerve injury which spontaneously resolved within 2 months. The axillary nerve courses inferior and medial to the subscapularis tendon, and dissection in this area should be avoided unless absolutely necessary, to limit iatrogenic injury. ^,

Injury to the subscapular nerves is more common during open subscapularis repairs. This complication is more commonly discussed in the shoulder arthroplasty literature because of the need to detach and repair the subscapularis in some manner as a part of the procedure. Armstrong et al. correlated post–total shoulder arthroplasty subscapularis function with clinical examination and EMG findings in a group of 30 patients. Patients were examined at an average of 1 year following total shoulder arthroplasty with subscapularis tenotomy and repair. On clinical examination six patients had a positive lift-off test, whereas the belly press test was negative in all patients. Ultrasound revealed a subscapularis rupture in two patients. EMG did not show any patients with active denervation; however, nine patients had abnormal findings in their subscapularis consistent with chronic denervation and reinervation. The authors concluded that a combination of surgical exposure, traction on the arm, and use of regional anesthesia may contribute to some of the chronic denervation changes seen about the shoulder, all of which should be considered when performing open surgery involving the subscapularis. Although there are no articles addressing EMG changes before and after arthroscopic subscapularis repair, this is a potential area of future study.

Subcoracoid Impingement

Impingement of the subscapularis tendon between the lesser tuberosity and the coracoid is known as subcoracoid impingement. A failure to both properly identify and properly manage subcoracoid impingement can lead to complications with regard to subscapularis repair, albeit in different ways. First, the surgical management of subcoracoid impingement can lead to iatrogenic injury to anatomic structures such as the musculocutaneous nerve, which is just lateral and distal to the coracoid. Additionally, to debride the coracoid, a window needs to be made in the rotator interval. This can theoretically lead to shoulder instability; however, no instance of this has been reported in the literature. Second, failure to properly identify subcoracoid impingement can also have implications. First, inadequate decompression of subcoracoid impingement will lead to decreased working space in the anterior shoulder, making adequate subscapularis repair difficult. If impingement of the coracoid on the subscapularis tendon or lesser tuberosity is present with dynamic motion, more aggressive bony debridement of the coracoid is likely required. Care should be taken to not debride so much coracoid as to release the conjoint tendon from its insertion. Also, postoperatively, failure to treat the impingement can lead to continued symptoms and even failure of the repair. Patients with subcoracoid impingement have been shown to experience redundancy and folding of the subscapularis in the subcoracoid space, which can lead to pain and potential failure of the repair.

Infection

Repair of the subscapularis carries the same risk of superficial and deep infection as any type of arthroscopic rotator cuff repair. It is important to be aware of potential risk factors for infection to further decrease this rare complication. Vopat et al. investigated risk factors for infection following rotator cuff repair and found that the overall infection rate in 1822 arthroscopic rotator cuff repairs was 0.77%, with the mini open technique and male sex both being significant risk factors for infection. Randelli et al. reviewed overall complications associated with arthroscopic rotator cuff repair, and found a 0.001% rate of superficial infection in 2890 cases. The most common organisms found in both superficial and deep shoulder infections are Propionibacterium acnes, Staphylococcus epidermidis, Staphylococcus aureus , Corynebacterium species, and Pseudomonas aeruginosa . Chuang et al. showed that, despite very low rates of infection, there remains a high rate of P. acnes colonization following shoulder arthroscopy. The authors performed superficial skin cultures before administration of prophylactic antibiotics, as well as sterile preparation of the shoulder at the proposed arthroscopic portal sites, and then repeated a deep tissue culture at the conclusion of the procedure. There was a 72.5% superficial colonization rate (46.1% of female patients and 81.6% of male patients), with 19.6% of patients having positive deep cultures at the end of the procedure. Despite this, no patients showed any sign of deep infection at 3 months. This study shows that, despite contamination of deep tissue with P. acnes during shoulder arthroscopy, the rate of infection remains quite low in the setting of proper sterile technique and utilization of perioperative antibiotics.

Postoperative Stiffness

Also, similar to any type of surgically repaired rotator cuff tear, subscapularis repairs are subject to complications postoperatively such as stiffness. Huberty et al. reviewed 489 rotator cuff repairs for postoperative stiffness, which was defined subjectively as patients being unsatisfied with their postoperative range of motion (ROM). There were 24 (4.9%) patients who were unsatisfied, with risk factors for stiffness being worker’s compensation, age younger than 50 years, calcific tendonitis, preoperative adhesive capsulitis, partial articular-sided tendon avulsion–type tears, or concomitant labral repair. Interestingly, patients who had concomitant coracoplasty were less likely to develop stiffness. Overall, only eight of the patients who developed stiffness had subscapularis repairs.

A systematic review by Gallagher et al. examined rates of stiffness with early versus delayed rehabilitation. They found significantly increased functional scores and motion within the first 3 to 6 months with early rehabilitation compared with the delayed group. However, only one continued to show a difference at a final follow-up of 15 months. No study reported any significant difference in rates of rotator cuff retear. However, there is also some evidence against the need for early ROM. Kim et al. performed a level I prospective study evaluating the effectiveness of passive ROM in the immediate postoperative period. All 105 patients were placed in an abduction brace for 4 to 5 weeks after surgery, with 56 being placed in the early passive ROM group and 49 in the no motion group. Patients in the early group performed passive ROM exercises 3 to 4 times a day while wearing the brace, whereas the no motion group did not perform any passive exercises. All patients were allowed to begin active assist ROM once the sling was discontinued. There were no significant differences in ROM or visual analog scale pain score in either group at any of the time points. Additionally, the rate of rotator cuff detachment was reported as 12% in the early group and 18% in the no motion group. Studies like this show that early passive motion may not be completely necessary following repair, and that rehabilitation protocols should be tailored to the individual patient based on tear pattern, quality of repair, and expected compliance.