Since Codman’s initial treatise on the surgical treatment of rotator cuff tears in 1911,⁴⁵ improved operative techniques have been responsible for a high success rate, with enduring patient satisfaction.⁴⁷,⁵⁷,⁷⁵,¹⁰⁷,¹⁰⁸,¹²⁰,¹³⁷,²¹⁴,²⁴¹ However, recurrent or persistent rotator cuff defects have been reported to occur in 20% to 90% of the cases,²⁰,³⁸,⁷³,⁷⁷,⁹⁰,¹⁰²,¹³³,¹⁴¹,¹⁴³,¹⁶⁰,¹⁶¹,¹⁷⁹,²²⁷,²³²,²³³,²⁴³ with the risk of recurrence increasing relative to the size of the initial tear. The failure rates between arthroscopic and open repairs appear to be equal when the tear is small and involves minimal retraction of the musculotendinous unit. The incidence of failure is highest among elderly patients with chronic and retracted tears involving two or three tendons. Arthroscopic repairs under these conditions represent the highest failure rate in both in vitro and clinical studies,⁴⁰,⁴⁴,⁷³,⁷⁵,⁸³,¹⁶⁰,²¹⁹,²³² although the gap appears to be closing as arthroscopic techniques continue to improve with failure rates anywhere from 30% to 50% for the large to massive tears.¹³⁷ Recently, Millar and colleagues showed improvement in outcomes and re-tear rates using arthroscopic methods versus open methods.¹⁶⁰ However, Bishop and colleagues showed an almost threefold increase in re-tear rates for tears greater than 3 cm done by arthroscopic techniques.²⁰ Clearly, controversy continues to exist for the best method of treating large to massive rotator cuff tears.

Persistent defects are not necessarily the sine qua non for failure, since the presence of a persistent rotator cuff defect is compatible with a good postoperative result following rotator cuff repair.³⁸,⁵⁷,⁷³,⁷⁷,¹³³,¹⁴³,²³² This process of converting a symptomatic tear into an asymptomatic re-tear is not entirely clear, although it may involve adequate subacromial decompression (SAD), debridement, biceps tenotomy, partial healing of the rotator cuff, and adequate postoperative rehabilitation.³⁰,³¹,¹⁸⁸,²¹⁰ The quality of functional results, however, depends on the size of the persistent defect, associated atrophy of the rotator cuff muscles, the integrity of the deltoid and the coracoacromial arch, and the age and functional demands of the patient.⁵⁷,⁶⁵,¹⁰²,¹⁰³,¹²²,¹⁴⁴,¹⁷¹,¹⁸⁵,²⁵² Patients with persistent rotator cuff defects will be capable of overhead function when the deltoid is intact and the anterior and posterior portions of the rotator cuff are intact and balanced.²⁹,³¹,³²,⁵⁷,²³² However, they will generally complain of fatigue with overhead activities, and limitation in activities that require vigorous or sustained overhead strength, as compared with patients with an intact rotator cuff.⁷⁷,¹⁰²,¹⁷⁹ Therefore, the goal of rotator cuff repair is long-term restoration of a functional, healed musculotendinous unit. While this may not always be attainable in primary rotator cuff repair, the development of recurrent rotator cuff tears may be minimized through a combination of careful preoperative patient selection, meticulous surgical technique, and attention to appropriate postoperative protection and rehabilitation.

With the understanding that the correlation between postoperative subjective and functional results and anatomic results (i.e., rotator cuff integrity) is variable,⁷⁷,⁸⁵,¹⁰²,¹⁴³ outcome studies have begun to focus on patient satisfaction in terms of patient-derived subjective assessments of symptoms and function.¹¹⁷,¹⁵⁹,¹⁸²,²¹²,²³²,²⁴³ Preoperative and surgical variables that are associated with poorer patient satisfaction include older patients, patients with chronic or irreparable subscapularis tears, large to massive supraspinatus and infraspinatus tears, tear retraction, atrophy of the rotator cuff muscles, especially the infraspinatus, and fatty infiltration of the muscle bellies.⁴⁷,⁸⁰,⁸²,⁸⁵,⁸⁸,¹⁰⁷,¹¹⁷,¹⁴⁹,¹⁸²,¹⁸⁵,²⁰⁴,²¹⁴,²²⁹,²⁴³,²⁵⁰ Objective postoperative variables that are associated with poorer patient satisfaction include diminished and weakened forward elevation, impingement signs, and acromioclavicular joint pain and tenderness. Subjective variables associated with poorer patient satisfaction include persistent pain, functional impairment, and work disability.¹⁸²,²⁵⁰ Numerous studies have shown that fat infiltration and muscle atrophy are not reversible after surgery and are positive predictors of re-tear and decreased functional outcomes.⁸¹,⁸³,⁸⁵,⁹⁰,¹⁴¹,¹⁶¹,²⁴³ In a 13-year follow-up study on the results of rotator cuff repair in patients under 50 years of age, Sperling and colleagues have shown that while patients experience significant long-term pain relief following repair, the functional results are inferior to those seen in a mixed-age population.²²⁹ A prospective longitudinal analysis of rotator cuff repairs by Galatz and colleagues indicated that as patients advance in age, their functional requirements decrease, and therefore satisfaction does not decline.⁷⁵ This was also found by Oh and colleagues, who showed that functional outcomes were improved in older patients after rotator cuff repair despite decreased rotator cuff integrity when compared with a younger population.¹⁸⁵

Associated risk factors for recurrent tears include advanced age, tear size, chronicity and atrophy, poor tendon quality, fatty infiltration of the muscle belly, poor bone quality, inappropriate rehabilitation, inadequate SAD, smoking, steroid injections, and diabetes.⁵⁹,⁷⁷,⁹⁰,¹⁰²,¹⁷⁵,¹⁸⁴,¹⁸⁵,²⁰⁴,²³³

Preoperative variables exist which will have a bearing on the ability to obtain long-term tendon-to-bone healing. In the presence of an acute rotator cuff tear, the biologic potential for healing appears greater when the repair is performed within 3 weeks of injury.¹²,¹⁰¹ In long-standing tears or delayed repairs, muscle atrophy and fatty infiltration may develop, which do not appear to be reversible after rotator cuff healing, and have a negative effect on the outcome following rotator cuff repair.⁸⁵,¹⁴¹,²²⁰ These changes may be graded using computed tomographic scanning or magnetic resonance imaging, and increase with elapsed time from the tendon rupture.⁷²,¹⁰¹,¹⁶¹,²³³,²⁵⁹ Additionally, chronic retraction and scarring of the musculotendinous unit may preclude the surgeon from obtaining an adequate tendon-to-bone repair. Therefore, in the presence of an acute or acute-on-chronic rotator cuff tear with retraction of the tendon, early repair is more likely to result in long-term tendon-to-bone healing as compared with late repair. This potential advantage should be considered in the context of appropriate patient selection criteria such as age, physical demands, comorbidities (diabetes and smoking), and motivation (willingness to comply with rehabilitation), prior to recommending surgical intervention.

The surgical principles that most likely reduce postoperative recurrent tears or persistent defects include adequate mobilization of the tendon to the greater tuberosity, preparation of the tendon and bone interfaces, and secure fixation of
the tendon to bone. These principles hold true for all repair methods. The superficial surface should be free from the overlying deltoid, acromion, subdeltoid bursa, scapular spine, coracoid, and the trapezius muscle. Capsular releases are frequently required to release the tenodesis effect of the underlying capsule.²⁶¹ Interval releases may also be required to allow full excursion of the contracted tendon.¹⁶,¹⁷¹,¹⁷² and ¹⁷³ A thorough understanding of the various tear configurations will enable the surgeon to perform the proper releases, repair longitudinal tears in a convergent manner, and repair the tendon to bone with minimal tension. A repair that has been overly tensioned will eventually fail.³² Inability to adequately mobilize the tendon may be due to substantial intramuscular scarring and the development of fatty atrophy. Tendon convergence and advancement methods as well as the management of irreparable rotator cuff tears have been discussed in previous chapters.

Debridement of the greater tuberosity, along with abrasion of the cortical surface at the proposed site of tendon attachment, may enhance tendon healing. While open rotator cuff techniques often include the creation of a shallow cancellous trough, most arthroscopic techniques repair the tendon to the cortical surface of the greater tuberosity. The rates of tendon healing between a shallow cancellous groove or trough, and a cortical bone surface appear to be equal.²³⁰ Therefore, complete removal of the cortical surface of the humeral attachment site is probably not necessary.¹⁵⁸ While excessive debridement of the tendon edges is not necessary or advisable, excision of the friable, necrotic edges will reveal a healthy tendon edge for suture placement.⁸⁷ The majority of rotator cuff repair failures occur when the suture pulls through the tendon.⁵⁴ Tendon grasping suture techniques such as the Mason-Allen technique provide excellent pullout strength and presumably a more durable repair during open rotator cuff surgery, as compared with simple or mattress sutures.⁴⁰,⁸³,²²² The holding strength of arthroscopically tied horizontal mattress sutures appears to be higher than that of the Mason-Allen suture. The massive cuff stitch, which includes a horizontal mattress stitch that is reinforced with a vertical loop, may act in a manner similar to that of the Mason-Allen suture technique.¹⁴⁶ While the holding strength that is required for successful tendon-to-bone healing is unknown, data exist supporting the superior holding strength of open transosseous techniques as well as for arthroscopic fixation using bone anchors.⁴⁰,⁴⁴,¹⁶⁰,¹⁶¹,²⁰⁵ As a result, some authors continue to prefer open repair methods in order to reduce the risk of re-tear, especially in the case of large to massive rotator cuff tears.

The advent of improved suture anchoring devices has led to their widespread acceptance and use in rotator cuff repair procedures. The advantages of suture anchors include the ease of use, decreased operating time, and decreased surgical exposure and morbidity. While most rotator cuff repair failures occur at the tendon— suture interface, anchors may subside without actually pulling out of the bone, leading to gap formation, poor tendon-to-bone healing, and rotator cuff repair failure.³²,³³,³⁷,⁴⁰,⁴⁴,¹⁴⁶,¹⁵⁵,¹⁵⁸,¹⁶¹,²²²,²³⁵ Anchor placement depends on the type of bone in which they rely for fixation, and pullout strength varies according to the bone quality of the humeral head. Tingart and colleagues have shown that the higher bone mineral density in the proximal—medial and proximal anterior regions of the greater tuberosity (i.e., closer to the articular surface) is associated with increased pullout strength of suture anchors.²³⁵ While suture anchors may represent an attractive option in young patients with high bone density, like any fixation device in orthopedic surgery, bone stock and strength remain major limitations and caution should be exercised among patients with poor bone quality such as postmenopausal women, smokers, and elderly patients, as these devices may not provide adequate pullout strength for rotator cuff repair. Under these circumstances, passing sutures through bone tunnels and tying the sutures over a lateral bone bridge may provide superior strength. This may be augmented with a thin plate or button to provide additional resistance to pullout of the sutures.³⁷ Recently, a novel arthroscopic technique has been developed that incorporates the advantages of both a true transosseous technique and an arthroscopic approach to the shoulder. Short- and long-term studies are necessary to fully evaluate this technique.

Proper postoperative protection and rehabilitation play an important role in preventing postoperative recurrent or persistent rotator cuff defects. Early postoperative mobilization has been a subject of increased debate and controversy. Although many surgeons have advocated for early passive range of motion after rotator cuff surgery in the past,¹²⁶,¹³¹,¹⁷¹,²¹¹ recent in vitro studies by Peltz and colleagues¹⁹⁴,¹⁹⁵ showed that early passive range of motion in a rat model was associated with increased joint stiffness. The same group has also shown that while there is an increase in stiffness at early time points after immobilization, this was transient with a resolution of range of motion in the immobilization group. A recent in vitro study has shown an improved tendon biomechanics when activity and exercise was limited in a rat model.¹⁹⁵ The delicate balance between tendon healing and postoperative stiffness was recently addressed in a study by Parsons and colleagues who found that 6 weeks of immobilization had no effect on range of motion at 1 year.¹⁸⁹ Immobilization typically lasts 4 to 6 weeks with patients who return to their postoperative visits exhibiting increased stiffness undergoing initiation of simple passive exercises of external rotation and pendulum exercises. Larger tears that have been repaired will typically be managed with increased immobilization and a greater tolerance for immediate postoperative stiffness. Immobilization is usually in a simple sling; however, in larger posterior-superior tears, the protective orthosis should place the arm at the side in a small amount of abduction in order to allow the tendon to heal under a minimum amount of tension.¹¹¹,²⁰⁶ Chronic or massive tears may require the use of a larger abduction pillow or orthosis, depending on the amount of tension that is observed at the time of surgery. Passive motion exercises should be instituted at 4 to 6 weeks with active motion exercises initiated between 8 and 10 weeks, avoiding resisted strengthening and isokinetic exercises until 3 months following repair. Increases in resistance are continued until 6 months at which time the patient can be allowed to return to full activity. The initiation of early active motion and the use of weights in the early postoperative period have been associated with failure of the repair.¹⁷⁵

Since the risk of recurrent rotator cuff tears is highest among elderly patients with chronic tears involving two or more tendons,¹⁰² the index of suspicion for this complication should be high when these patients complain of persistent pain and functional impairment. Physical findings vary according to the size and location of the tear, the integrity of the coracoacromial arch, and the presence of soft tissue contracture. Lack of both active and passive elevation is more likely to represent capsular contracture rather than residual rotator cuff insufficiency. It is extremely difficult, if not impossible, to determine the presence and clinical relevance of recurrent rotator cuff tears
in a stiff shoulder. Once the soft tissue contracture has been excluded or corrected, the physical findings associated with a recurrent rotator cuff tear will become more apparent.

Symptomatic patients will have subacromial crepitance, weakness on isolated muscle testing, and will often demonstrate a positive “lag sign” on physical examination.¹¹² When pain appears to be a limiting factor during strength testing, a subacromial lidocaine injection will usually alleviate a significant amount of associated pain, and increase the reliability of the examination. Recurrent defects in the posterior-superior portion of the rotator cuff (supraspinatus and upper infraspinatus tendons) generally demonstrate weakness of arm abduction and weakness of external rotation with the arm at the side. This external rotation weakness generally improves when the arm is brought into 90 degrees of elevation in the scapular plane. Conversely, a tear involving the posterior-inferior portion of the rotator cuff (infraspinatus and teres minor tendons) will usually demonstrate external rotation weakness with the arm at the side as well as in 90 degrees of elevation in the scapular plane. External rotation lag signs refer to the inability of the patient to actively maintain maximal external rotation of the arm when it has been passively placed in this position by the examiner. Positive external rotation lag signs with the arm at the side, or with the arm at 90 degrees of elevation in the scapular plane, indicate recurrent defects in the posterior-superior and posterior-inferior rotator cuff, respectively. External rotation lag signs at any level of abduction are not particularly reliable in the detection of weakness associated with an isolated supraspinatus tear.¹²¹ Abduction of the arm in the scapular plane, with the elbow extended and the humerus internally rotated, is generally accepted to represent supraspinatus function.¹¹²,¹²⁶ Therefore, weakness in this position may indicate a recurrent or persistent tear of the supraspinatus.

Subscapularis insufficiency results in increased passive external rotation of the shoulder as well as weakness of terminal internal rotation. Detection of a recurrent defect requires careful isolation of the muscle from other internal rotators of the shoulder girdle.⁹³ The internal rotation lag sign represents the inability of the patient to maintain the dorsum of the hand away from the mid-lumbar spine after it has been passively placed in this position of maximal internal rotation by the examiner. Similarly, the lift-off test is sensitive for detecting subscapularis insufficiency and describes the inability of the patient to actively lift the dorsum of the hand away from the lumbar spine.⁸²,¹¹² These tests require full passive internal rotation in order to place the arm in the appropriate position. In the presence of posterior capsular contracture, the abdominal compression test appears to be as specific as the lift-off test in assessing subscapularis insufficiency.²³⁶ The test is positive when the patient is unable to maintain the flexed elbow anterior to the coronal plane of the body while simultaneously maintaining the palm of the hand compressed against the abdomen. Scapular protraction is often difficult to control and will interfere with the performance of this test. When this occurs, it may be easier to control scapular protraction with the patient in the supine position.

The suspected diagnosis of a recurrent rotator cuff tear may be confirmed with ultrasonography, arthrography, or magnetic resonance imaging.³⁸,⁷⁷,⁹⁶,¹⁴¹,¹⁴⁸,¹⁶¹,¹⁸⁶,²⁴³ The presence of postsurgical artifact will interfere with the interpretation of imaging studies, and the diagnostic criteria for a full-thickness recurrent tear are more stringent than that for a shoulder that has not undergone surgery. Magnetic resonance arthrography (MRA) or ultrasonography can be used. However, our preferred imaging modality is MRA because it provides information regarding tear size, muscle atrophy and fatty infiltration, and concomitant biceps and labral pathology. While the clinical relevance of minor tendon signal abnormalities is uncertain,¹⁶⁵ a well-defined tendon gap that is traversed by fluid is a reliable indicator of a persistent or recurrent rotator cuff defect (Fig. 1-1).⁹⁶,¹⁷⁹,²³³

Normal shoulder kinematics in the presence of an intact rotator cuff, functioning deltoid, and intact coracoacromial arch, will maintain the geometric center of the humeral head within 2 to 3 mm of the center of the glenoid during active elevation of the arm.⁷¹,¹²⁷,¹⁹⁸ Although the greater tuberosity may contact the anterior acromion and coracoacromial ligament with an elevation above 60 degrees, the compressive force of the normal rotator cuff will actively center the humeral head into the glenoid fossa during deltoid muscle contraction, and serve as the primary restraint to anterosuperior translation of the humeral head.⁶⁷,⁷¹ Under normal loading conditions, the static coracoacromial arch plays a limited role as a secondary passive restraint to anterosuperior subluxation, but assumes a crucial role in the presence of a dysfunctional rotator cuff.⁶⁷,²⁵¹ In the presence of small rotator cuff tears, there are usually sufficient forces anteriorly and posteriorly to compensate for the loss of function that accompanies the tear. Although glenohumeral mechanics are altered, the remaining rotator cuff muscles continue to serve as the primary restraint to superior translation, but the coracoacromial arch probably plays a more important role as a passive restraint.²⁵⁷ When rotator cuff tears become too large, or are “unbalanced,” the remaining rotator cuff muscles are unable to compensate for the loss of joint compressive forces.²⁸ In these “uncompensated” rotator cuff tears, the humeral joint reaction force is not directed into the glenoid. Rather, the muscle imbalance between a dysfunctional rotator cuff and a strong deltoid directs the humeral head anterosuperiorly, and the intact coracoacromial arch becomes the only restraint to further anterosuperior subluxation
(Fig. 1-2). In the setting of uncompensated rotator cuff dysfunction, coracoacromial insufficiency from prior surgery will severely compromise overhead function.²⁵² The incidence of this complication following rotator cuff repair is unknown, but when it occurs, the results are devastating and have historically been unsalvageable by tendon repair, augmentation, tendon transfer, coracoacromial arch reconstruction, or any combination of these techniques.⁷⁰,²³¹,²⁵² In selected patients, reverse total shoulder replacement provides a reliable method to restore some function in the shoulder. Patients who have poor preoperative function will often indicate that they are worse following surgery that results in anterosuperior humeral head subluxation. Their pain and dysfunction is often intolerable, even with activity modification and analgesics.

Prevention of anterosuperior humeral head subluxation involves preservation of the coracoacromial arch during acromioplasty and rotator cuff repair, particularly in the presence of a large rotator cuff tear, or a decreased preoperative acromiohumeral interval. While the indications for coracoacromial ligament preservation and repair continue to evolve,²⁷,⁷⁰,²⁵¹ this generally includes patients with two and three tendon tears, in whom the risk of a recurrent rotator cuff defect is sufficiently high to justify restoration of the coracoacromial arch. To facilitate repair of the ligament to bone, the anterior length of the acromion should be preserved while smoothing the undersurface of the acromion.²⁸,³⁴,¹⁷¹,¹⁷³ Some authors have advocated rotator cuff repair without performing an acromioplasty, in order to avoid disrupting the integrity of the coracoacromial arch.²⁷,⁷⁶,⁸⁶ Denard and colleagues examined three different modes of SAD in a cadaveric model: a “smooth and move,” a limited acromioplasty with coracoacromial ligament preservation, and a coracoacromial ligament resection were performed on six cadaveric specimens with intact rotator cuffs.⁵⁸ They concluded that a limited acromioplasty with coracoacromial ligament preservation may best provide decompression of the rotator cuff while avoiding potential anterosuperior glenohumeral translation. Clearly, partial excision of the coracoacromial ligament during acromioplasty has met with good results among patients with small rotator cuff tears,¹⁰²,¹⁰⁷,¹⁰⁸,¹⁷¹,²¹¹ suggesting that repair of the coracoacromial ligament is not required in all rotator cuff repairs. This was confirmed by Wellmann and colleagues who showed that coracoacromial ligament resection had limited effect on anterior and superior translation of the humeral head with an intact rotator cuff.²⁵¹ Although there may be a growing trend toward coracoacromial ligament preservation during both open and arthroscopic acromioplasty, little data exist in the literature to support routine repair of the ligament. However, Fagelman and colleagues examined the biomechanical consequences of coracoacromial arch alteration and subsequent reconstruction in cadaveric specimens concluding that reconstruction of the coracoacromial ligament in patients with massive rotator cuff tears may provide the necessary stabilization force needed to prevent anterosuperior escape.⁶⁷

The operating surgeon should exercise caution and even restraint when evaluating patients with marginal overhead function and complete obliteration of the acromiohumeral interval, as these patients are at high risk for deteriorating function following SAD and rotator cuff repair. If surgical intervention is contemplated in these patients, imaging studies such as magnetic resonance imaging and ultrasound will provide critical information regarding reparability of the rotator cuff. In the setting of an irreparable rotator cuff tear, a more conservative deltoid-sparing open approach or arthroscopic decompression may be considered. Arthroscopic SAD usually requires subperiosteal release of a portion of the coracoacromial ligament, although the most medial portion or the acromial attachment can often be preserved.¹⁰²,¹⁴⁰ Studies indicate that the coracoacromial ligament may heal back to the acromion provided that the anterior length of the acromion is preserved. Reversed arthroscopic decompression with biceps tenotomy may represent a reasonable alternative to conventional arthroscopic acromioplasty in patients with massive rotator cuff tears, and preserves the coracoacromial arch.²²¹

The coracoacromial ligament can be released and repaired in one of two ways during open SAD. First, the deltoid insertion and coracoacromial ligament can be released in a single layer and reattached to the acromion at the completion of the procedure. Alternatively, the interval between the deltoid attachment and the coracoacromial ligament can be dissected, releasing the deltoid from the anterior acromion (Fig. 1-3). The underlying ligament can then be subperiosteally released from the undersurface of the acromion in order to maximize the length of the ligament. The ligament is then immediately sutured to the deep deltoid fascia and ultimately repaired to the anterior acromion as a single layer. Tuberoplasty, as described by Fenlin and colleagues, represents an alternative to conventional open acromioplasty in patients with massive rotator cuff tears, and may assist in preserving the coracoacromial arch. This technique is typically performed by arthroscopic surgery.

The manifestations of anterosuperior humeral head subluxation are usually apparent on clinical examination. With the arm at the side, the humeral head rests within the glenoid cavity. However, when the patient attempts to actively elevate the arm, the humeral head will rise out of the glenoid into a subcutaneous position anterior to the acromion. Full passive forward elevation is usually possible unless the anterosuperior subluxation is chronic and fixed. Due to the massive rotator cuff tear, the patient will be unable to raise the arm overhead, and will demonstrate weakness in external rotation with the arm at the side, and with the arm at 90 degrees of elevation in the scapular plane. Although the subscapularis may be involved, some inferior fibers may remain intact, leading to negative abdominal compression and lift-off tests. However, internal rotation strength will often be weak and the internal rotation lag sign may be present.⁸²,¹¹² This complication is often also accompanied by deltoid insufficiency, consisting of thinning, detachment, or atrophy associated with denervation.⁷⁴,²⁵²

Radiographic studies are obtained to evaluate acromial bone loss and humeral head morphology. The standard radiographic views include anteroposterior views with the humerus internally and externally rotated, axillary, 30-degree caudaltilt, and supraspinatus outlet views. Acromial bone loss is best appreciated on the 30-degree caudal-tilt and axillary views, while the anteroposterior view will reveal humeral head articular changes associated with rotator cuff tear arthropathy, and static humeral anterosuperior subluxation. Magnetic resonance imaging is not required to confirm the diagnosis of anterior subluxation or recurrent rotator cuff tear, but will provide prognostic information regarding atrophy and fatty infiltration of the rotator cuff musculature. It will also facilitate treatment decisions regarding muscle transfer by indicating which muscles are still intact.⁸¹,⁸⁴