Rotator Cuff Treatment.

Last updated Wednesday, January 26, 2005

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Figure 17

Figure 18

Figure 19

Figure 20

Figure 21

Figure 21

Figure 23

Figure 24

Figure 25

Figure 26

Figure 27

Figure 28

Figure 29

Figure 30

Figure 31

Figure 32

Figure 33

Figure 34

Figure 35

Figure 36

Figure 37

Figure 38

Figure 39

Figure 40

Figure 41

Figure 42

Figure 43

Figure 44

Figure 45

Figure 46

Figure 47

Figure 48

Figure 49

Figure 50

Figure 51

Figure 52

Figure 53

Figure 54

Figure 55

Figure 56

Figure 57

Figure 58

Figure 59

Figure 60

Figure 61

Figure 62

Figure 63

Figure 64

Figure 65

Figure 66

Figure 67

Figure 68

Figure 69

Figure 70

Figure 71

Figure 72

Figure 73

Figure 74

Figure 75

Figure 76

Figure 77

Figure 78

Figure 79

Figure 80

Figure 81

Figure 82

Figure 83

Figure 84

Introduction

Eight clinical entities

Minimal symptoms

Slightly frozen shoulder

Posterior capsular tightness

Nonoperative treatment

Patients in whom the primary complaint is symptomatic subacromial crepitance (see figure 7) will usually benefit from reassurance and a home program of gentle stretching and strengthening exercises. Various nonoperative rotator cuff programs have been described for the general population and for athletes, including throwers. (Albright, Jokl, 1978, Atwater, 1979, Berry, Fernandes, 1980, Binder, Parr, 1984, Cofield, 1985, Cofield and Simonet, 1984, Fowler, 1979, Hawkins and Kennedy, 1980, Jobe and Moynes, 1982, Kerlan, Jobe, 1975, Neer, 1983, Neviaser, Neviaser, 1982, Pappas, Zawacki, 1985a, Pappas, Zawacki, 1985b, Richardson, Jobe, 1980, Rocks, 1979, Scheib, 1990) Exercises must address specifically any shoulder stiffness, which may cause obligate translation and loss of concentricity on shoulder movement (see figure 8). The effectiveness of nonoperative treatment was recognized many years ago by Neer who, in his initial article on anterior acromioplasty, pointed out that "Many patients . . . were suspected of having impingement, but responded well to conservative treatment." (Neer, 1972) Furthermore, he stated that patients were advised not to have an acromioplasty until the stiffness of the shoulder had disappeared and the disability had persisted for at least nine months. As a result of these conservative surgical indications, during the period covered by his report, this most active shoulder surgeon operated on an average of only ten shoulders a year with this diagnosis: the effectiveness of nonoperative management is worthy of emphasis!

The low success rate in returning athletes to competition after acromioplasty (Tibone, Jobe, 1985) reinforces the importance of nonoperative management in this population. Similar principles apply to workers who are required to use their shoulders in positions aggravating subacromial abrasion.

Subacromial injections of corticosteroids have been reported by some to produce symptomatic relief. (Hollingworth, Ellis, 1983) However, Withrington and coworkers (Withrington, Girgis, 1985) reported a double-blind trial of steroid injections and found no evidence of the efficacy of such treatment. Valtonen (Valtonen, 1978) found no difference between subacromial and gluteal injections of steroids. Berry and colleagues (Berry, Fernandes, 1980) compared acupuncture, physiotherapy, steroid injections, and anti-inflammatory medications and found no difference among these treatments.

Steroid injections in or near the cuff and biceps tendons may produce tendon atrophy or may reduce the ability of damaged tendon to repair itself. Such changes have been well documented in other tissues. (Lund, Donde, 1979, Rostron, Orth, 1979, Uitto, Teir, 1972) Uitto and colleagues, (Uitto, Teir, 1972) demonstrated corticosteroid-induced inhibition of the biosynthesis of collagen in human skin. The harmful effects of repetitive intra-articular injection of steroids have been noted. (Behrens, Shepherd, 1975, Cruess, Blennerhassett, 1968, Mankin and Conger, 1966, Salter, Gross, 1967, Sweetnam, 1969)

Ford and DeBender (Ford and DeBender, 1979) reported 13 patients who developed 15 ruptured tendons subsequent to nearby injection of steroids. Other authors have reported spontaneous ruptures of the Achilles tendon and patellar tendon after steroid injection. (Bedi and Ellis, 1970, Ismail, Balakishnan, 1969, Lee, 1957, Melmed, 1965, Smaill, 1961) Although Matthews and colleagues (Matthews, Sonstegard, 1974) failed to find a deleterious effect of corticosteroid injections on rabbit patellar tendons, Kennedy and Willis (Kennedy and Willis, 1976) found a substantial effect in the rabbit Achilles tendon. They concluded that physiological doses of local steroid placed directly in a normal tendon weaken it significantly for up to 14 days following the injection.

Watson (Watson, 1985) reviewed the surgical findings in 89 patients with major ruptures of the cuff. He found that all 7 patients who had had no local steroid injections had strong residual cuff tissue. Thirteen of 62 patients having one to four steroid injections had soft cuff tissue that held suture poorly; 17 of the 20 patients having more than four steroid injections had very weak cuff tissue; these shoulders with weak cuff tissue had poorer results after surgical repair. In this light, one can appreciate the potential hazard of making a diagnosis of "bursitis" or "bicipital tendinitis" and treating the situation with repeated steroid injections until the reality of a major cuff tendon deterioration becomes inescapable. (Darlington and Coomes, 1977, Kennedy and Willis, 1976)

The patient with subacromial abrasion is informed that this condition can usually be resolved with nonoperative management directed toward the restoration of normal mobility, strength, coordination and fitness.

Authors' preferred method of nonoperative management of subacromial abrasion

In our approach to subacromial abrasion, we recognize the important interplay between cuff weakness, stiffness of the posterior capsule, and subacromial roughness. We use a program designed by Sarah Jackins, a physical therapist who has worked with the University of Washington Shoulder and Elbow Service since its inception in 1975. This treatment regimen is analogous to one that would be used for managing a tennis elbow or Achilles tendinitis and includes:

1. avoidance of repeated injury,
2. restoration of normal flexibility,
3. restoration of normal strength,
4. aerobic exercise, and
5. modification of work or sport.

The emphasis is on simple, low-tech exercises that the patient can perform unassisted.

The Jackins Program

Step 1: Avoidance of Repeated Injury

Although it seems obvious that an affected shoulder must be rested, we see patients each week who are trying to continue vigorous overhead work or swimming hundreds of miles per week in the presence of cuff symptoms. It is difficult to treat these symptoms when the affected area is repeatedly irritated; activities may need to be temporarily modified--light duty, reducing mileage, less throwing, using the kickboard for a major part of the workout rather than continuing to try to "swim through" the problem, or working on the forehand and footwork rather than beating away at the serve. Once symptoms have subsided, the activity is progressively resumed with an emphasis on proper technique and a paced resumption of normal levels of performance.

Step 2: Restoration of Normal Flexibility

The goal of Step 2 is to stretch out all directions of tightness. Shoulders with subacromial abrasion are frequently stiff, especially in the posterior capsule. As described above for posterior capsular tightness, the most effective program is one that taught by the surgeon or therapist, but is carried out by the patient. The goal of the flexibility program is to restore the range of motion to that of the unaffected shoulder. The recommended treatment consists of gentle stretches performed five times a day by the patient (see figures 9-14). Each stretch is performed to the point where the patient feels a pull against the shoulder tightness, but not to the point of pain. Each stretch is performed for one minute, so that the patient invests about 30 minutes per day in their shoulder. Obvious improvement commonly occurs within the first month, but three months may be required to completely eliminate the condition.

Step 3: Restoration of Normal Strength

When near-normal passive flexibility of the shoulder is restored, the patient's attention is directed toward regaining muscle strength. As is the case in managing tennis elbow, it is most effective to delay strengthening exercises until normal range of motion is achieved. As with the flexibility exercises, the patient is given the responsibility for strengthening the shoulder. Internal and external rotator-strengthening exercises are carried out with the arm at the side (see figures 15 and 16) to strengthen the anterior and posterior cuff muscles without the potential for subacromial grinding that exists with exercises in abduction and flexion. These exercises are most conveniently performed against the resistance of rubber tubing, sheet rubber, bike inner tubes, springs or weights. It is convenient if the resistance device can be carried in a pocket or purse for frequent use through the day. As strength increases, the patient is advanced to more resistance: thicker tubing, tougher rubber sheets, or more springs. Deltoid strengthening is added when it can be performed comfortably (see figure 17) as are exercises to strengthen the scapular motors (see figure 18). Athletes are not returned to full activity until the shoulder has full mobility and strength.

Step 4: Aerobic Exercise

If a patient has gotten out of shape as a result of the shoulder problem, it is important to emphasize the need to regain normal fitness. To get back in shape and to improve the sense of well-being, a half hour of "sweaty" exercise five days a week is recommended. Brisk walking may be the safest and most effective type of aerobic exercise, but other suitable forms include jogging, biking, stationary biking, and so on. Aerobic calisthenics as usually defined must be carefully reviewed to ensure that they do not require arm positions which aggravate the patient's symptoms.

Step 5: Modification of Work or Sport

Obviously, the purpose of the program is to return the patient to the comfortable pursuit of normal activities. Not infrequently this requires some analysis of working and recreational techniques. Occasionally this is as simple as having the short grocery clerk stand on a platform at work. The technique of swimmers is reviewed to ensure, for example, adequate roll on the freestyle stroke. Throwers are taught the importance of body position and rotator cuff strength. Adequate knee bend and lumbar extension is reinforced in the execution of the tennis serve. If the patient has an occupation that requires vigorous or repeated use of the shoulder in painful positions, vocational rehabilitation to a different job may be required.

Subsequent Steps

It may take six weeks before substantial benefit is realized. As long as the patient is making progress, we continue this program. If improvement is not forthcoming, the program is reviewed to be sure it is being conducted in an ideal way. The shoulder and the patient are also reevaluated to make sure there are no other factors that may be interfering with recovery. If repeat clinical evaluation indicates positive tendon signs (see figure 19) or other evidence of cuff fiber failure, tendon imaging studies may be considered if their results would change the patient's management. If a well-motivated patient continues to have symptomatic subacromial abrasion after six months of a well-conducted program, subacromial smoothing may be discussed as an alternative to continued non-operative management. Poor compliance with an exercise program may foretell an equally poor result from surgical treatment.

Operative treatment

Open acromioplasty

In his classic description of acromioplasty, Neer (Neer, 1972) described approaching the shoulder through a 9 cm incision made in Langer's lines from the anterior edge of the acromion to a point just lateral to the coracoid. The deltoid is split for 5 cm distal to the acromioclavicular joint in the direction of its fibers. It is then dissected from the front of the acromion and the acromioclavicular joint capsule. The stump of the deltoid's tendinous origin is elevated upward and preserved for the deltoid repair. Using an osteotome, a wedge-shaped piece of bone .09 cm by 2.0 cm is resected from the anterior undersurface of the acromion, along with the entire attachment of the coracoacromial ligament. If acromioclavicular osteophytes are present, the distal 2.5 cm of the clavicle are also excised along with the prominences on the acromial side of the joint. After the procedure, the deltoid is carefully repaired to the acromioclavicular joint capsule, the trapezius, and its tendon of origin.

Many reports regarding the results of open acromioplasty have been published. (Armstrong, 1949, Diamond, 1964, Ha'eri, Orth, 1982, Hammond, 1962, Hammond, 1971, Hawkins and Kennedy, 1980, Jackson, 1976, Kessel and Watson, 1977, McLaughlin, 1944, Michelsson and Bakalim, 1977, Moseley, 1969, Neer, 1972, Neer, 1983, Neviaser, Neviaser, 1982, Penny and Welsh, 1981, Pujadas, 1970, Skoff, 1995, Smith-Petersen, Aufranc, 1943, Thorling, Bjerneld, 1985, Watson-Jones, 1960) However, the interpretation of these reports is made difficult by the admixture of patients with intact cuffs, partial thickness cuff lesions, and full thickness cuff tears as well as by the inclusion of a wide range of additional elements to the surgery. Stuart et al (Stuart, Azevedo, 1990) reported a series which included acromioplasty with or without cuff repair, distal clavicle excision, and biceps tenodesis; 23% were still painful. Rockwood and Lyons (Rockwood and Lyons, 1993) reported on a series of 71 patients who had a modified acromioplasty with or without cuff repair and concluded that cuff repair did not influence the percentage of excellent results. Bosley (Bosley, 1991) reported on 35 patients with total acromionectomy, including patients with and without long standing massive cuff tears; most failures were attributed to either the underlying pathology or to failure of deltoid reattachment. Bjorkenheim et al (Bjorkenheim, Paavolainen, 1990) reported a failure rate of over 25%, attributing the failures to "associated bony as well as soft-tissue subacromial lesions". Oglvie-Harris et al (Ogilvie-Harris, Wiley, 1990) evaluated 67 shoulders in 65 patients who had pain and dysfunction for more than two years after an initial acromioplasty for impingement syndrome without a rotator cuff tear. In almost half of the cases there were "diagnostic errors" and even in those where there was a correct diagnosis and no operative errors, the failure rate was almost 20%.

Arthroscopic acromioplasty

Ellman, (Ellman, 1987) in 1987, published the first large series of fifty patients (average age 50 years) with mixed shoulder pathology who under went arthroscopic acromioplasty; ten had full thickness tears. At an average follow-up of seventeen months, 88% had good or excellent results. These results persisted at 2.5 year followup of the same treatment group. (Ellman and Kay, 1989)

Since then others have reported results of arthroscopic acromioplasty. (Ellman, 1987, Esch, Ozerkis, 1988) Gartsman, (Gartsman, 1990) Speer et al, (Speer, Lohnes, 1991) Altchek et al, (Altchek, Warren, 1990) and Roye et al (Roye, Grana, 1995) reported series of arthroscopic acromioplasties on shoulders without cuff tears; each finding 83-94% of the results were satisfactory. Approximately 75% of the patients were able to return to sports activity. Recovery times in these series ranged from 2 to 4 months. Most authors describe the procedure as technically demanding. The control of bleeding and the determination of the amount of bone to resect are two commonly technical difficulties in performing arthroscopic subacromial decompression. Many describe a learning curve associated with this technique and have recommended this procedure be performed on cadaver shoulders before it is used clinically.

In early years, after the introduction of the arthroscopic technique of acromioplasty, controversy arose as to whether a subacromial decompression performed arthroscopically was technically equivalent to that performed open. Gartsman et al. (Gartsman, 1990, Gartsman, Blair, 1988) in a cadaver study, was able to perform arthroscopic bony resection with release of the coracoacromial ligament equivalent to the open technique described by Neer. He suggested criteria for the technical adequacy of the acromioplasty: (1) the entire anterior acromial protuberance is resected, (2) the undersurface of the acromion is flattened, (3) the deltoid fibers are visible from the acromioclavicular joint to the lateral edge of the acromion, (4) the inferior aspect of the acromioclavicular joint is debrided to remove any downward protrusion, (5) the coracoacromial ligament is completely released from the anterior portion of the acromion and the acromioclavicular joint, (6) a portion of the ligament is resected, (7) an adequate subacromial bursectomy is performed to allow complete inspection of the bursal surface of the rotator cuff, and (8) finally, no subacromial abrasion is observed when the arm is taken through a range of motion at the completion of the procedure.

Most authors state that the indications for arthroscopic acromioplasty should be identical to those for the open procedure described by Neer in 1972. However, compared to the rate with which Neer used open acromioplasty in his practice, it is apparent that arthroscopic acromioplasty is performed much more commonly with broader indications. Although overall "satisfactory" results were obtained in the majority of reports, some authors were uncertain whether relief was obtained from the modifications of the acromial shape or other aspects of the treatment.

To keep things in perspective, Brox et al (Brox, Staff, 1993) compared the effectiveness of arthroscopic acromioplasty, an exercise program and a placebo in a randomized clinical trial. The study group consisted of 125 patients aged 18-66 who had had rotator cuff disease for at least three months and whose condition was resistant to treatment. The authors concluded that surgery or a supervised exercise regimen significantly, and equally, improved rotator cuff disease compared with placebo; however the surgical treatment was substantially more costly.

Comparison of open and arthroscopic acromioplasty

In 1994, Sachs et al (Sachs, Stone, 1994) reported on a series of 44 patients with stage II impingement prospectively randomized into open (22 patients/average age 49) and arthroscopic treatment groups (19 patients/average age 51). In both groups, full recovery took at least 1 year for the majority of patients. In both groups over 90% of patients achieved a satisfactory result (good or excellent). Final analysis showed that the main benefits of arthroscopic acromioplasty were evident in the first 3 months postoperatively with the arthroscopic patients regaining flexion and strength more rapidly than did patients treated with open decompression.

Furthermore, the arthroscopic treatment group had shorter hospitalizations, used less narcotics, and returned more quickly to both work and activities of daily living, leading the authors to suggest arthroscopic acromioplasty may have significant economic advantages.

In 1992, Van Holsbeeck et al (van Holsbeeck, DeRycke, 1992) compared their results of 53 patients treated by arthroscopic acromioplasty and 53 patients treated by an open acromioplasty. Based on the UCLA rating scale, good or excellent results were identical for both groups at a 2 year followup. The authors suggested arthroscopic acromioplasty was associated with a shorter recovery time, however, in the long term there was no difference in strength of forward flexion between the open and arthroscopic groups.

Hawkins at al (Hawkins, Saddamis, 1992) reported 40% satisfactory results with arthroscopic subacromial decompression, while they reported 87% satisfactory results with a concurrent series of open acromioplasty.

Roye et al (Roye, Grana, 1995) reported a series of 90 arthroscopic acromioplasties and found that the most of the patients who were not throwing athletes obtained satisfactory results and that the presence or absence of a cuff tear did not affect the result.

Lindh et al, in 1993, reported on a series of 20 patients who were randomly selected for either open or arthroscopic acromioplasty (10 patients in each group). The average duration of symptoms before surgery was over 5 years. Functional results in both the arthroscopic and open surgery groups were good and similar. Patients in the arthroscopic group were observed to demonstrate earlier restoration of full range of motion and reduction in time away from work.

Proponents of arthroscopic acromioplasty have argued this procedure requires less surgical dissection and produces less scarring and less post operative morbidity. In most instances the procedure can be performed on an outpatient basis. Post operative discomfort is moderate, and can usually be controlled by oral analgesics. Additionally, cosmesis is good, and patient acceptance is high.

A countervailing advantage of open acromioplasty is the advantage of being able to observe directly the subacromial space during motions which preoperatively caused the patient's symptomatic subacromial crepitance and the ability to assure that the crepitance is resolved before the procedure is concluded.

Deltoid retraction can be a significant problem after open procedures which require detachment and subsequent reattachment of the deltoid to the anterior acromion. (Bigliani, Cordasco, 1988) Arthroscopic acromioplasty has the theoretical advantage of leaving the deltoid origin almost totally undisturbed. However, in a recent report Torpey et al (Torpey, McFarland, 1996) it was pointed out that much of the deltoid arose from the anterior acromion. Their analysis indicated that a 4 mm anterior acromioplasty would detach approximately half of the deltoid fibers, whereas a 6 mm anterior acromioplasty would detach approximately 75% of the fibers. They conclude that neither an open nor an arthroscopic acromioplasty can be performed without substantial compromise of the anterior deltoid origin.

Arthroscopy offers the ability to directly inspect the glenohumeral joint as well as subacromial space. During an open acromioplasty, the deep surface of the cuff (where most cuff lesions begin) is not visible. By contrast at arthroscopy, partial or complete thickness tears of either surface of the rotator cuff as well as other findings can be identified by the experienced observer. However, even with arthroscopy the common intratendinous lesions remain inaccessible. Paulos and Franklin (Paulos, Harner, 1988) in their series of 80 arthroscopic acromioplasties reported a high number of unsuspected diagnoses that were made during arthroscopy. These included twenty-six (26) partial rotator cuff tears, twelve (12) labral tears; eight (8) instances of humeral chondrosis; four (4) cases of biceps tendon fraying; and two (2) loose bodies in the glenohumeral joint. They reported that for most of these shoulders these findings would have been missed with the open technique.

Altchek et al (Altchek, Warren, 1990) in their series of forty-four (44) patients treated by arthroscopic acromioplasty, eleven (11) patients had lesions of the glenoid labrum. Preoperatively, these patients had no evidence of instability, either by history or physical exam in action. Five (5) of these patients had a tear involving the inferior part of the labrum and failed to recover completely after the acromioplasty and were unable to return to full participation in sports. The authors felt that undetected slight instability may have played a role in the production of these patients' symptoms. The authors argued that arthroscopic inspection of the glenohumeral joint makes it possible to detect such problems providing information that is important for prognosis. Others have reported a higher than anticipated percentage of unsuspected associated lesions in those shoulder being treated arthroscopically for impingement symptoms. (Jobe and Kvitne, 1989) Burns and Turba (Burns and Turba, 1992) reported on their findings in 29 patients treated with arthroscopic acromioplasty which included anterior glenoid labrum tear (15), undersurface rotator cuff tear (8), chondromalacia of the humerus (3), biceps rupture (1), posterior glenoid labrum tear (1), and acromioclavicular arthritis (1).

These results indicate that the preoperative diagnosis of "impingement syndrome" has been associated with a wide range of shoulder pathologies. They leave unanswered the question of the prevalence of these same findings in asymptomatic shoulders and the role played by each of the findings in producing clinical symptoms. Hopefully in the future methodical clinical-pathological correlation will lead to improved accuracy in preoperative diagnosis and greater specificity in treatment.

The primary difficulty in interpreting these studies on open and arthroscopic acromioplasty is that, although the outcome of the procedure is characterized in terms such as of "good" or "excellent", the effectiveness of the procedure is often undetermined because the preoperative status, or ingo was not characterized in the same way. Ideally, a "good" result from surgery would indicate the change in the patient's condition as a result of the procedure, rather than the status of the shoulder post operatively.

The definition of the indications for and the effectiveness of acromioplasty must await multipractice studies which define accurately the pretreatment clinical findings and shoulder functional status, the nature of and compliance with a non-operative program, the nature of the surgery, and the change in the shoulder function realized after the procedure using the same parameters of comfort and function before and after surgery. The effectiveness of a treatment is the difference between the outcome and the ingo.

Authors' preferred method for subacromial smoothing

In our experience, the results of subacromial smoothing are likely to be best in the following circumstances:

1. a well-motivated patient over 40,
2. absence of posterior capsular stiffness,
3. presence of symptomatic subacromial crepitus (see figure 20) which the patient agrees is the dominant clinical problem,
4. absence of tendon signs (see figure 21) and other shoulder pathology and
5. symptoms which are not associated with a work-related injury.

Poor prognostic signs include:

1. age less than 40,
2. stiffness,
3. absence of subacromial crepitus,
4. presence of tendon signs or evidence of other shoulder pathology,
5. attribution of problem by the patient to his or her occupation,
6. concomitant evidence of glenohumeral instability, and
7. neurogenic cuff muscle weakness.

We use an open approach to subacromial smoothing. The patient is positioned with the head up at 30 and the arm draped free. Before making the incision, we note the positions and motions in which subacromial crepitus can be palpated through the acromion. The shoulder is approached through an incision in the skin lines over the anterolateral corner of the acromion (see figure 22). The acromion is exposed striving to maintain the continuity of the deltoid fascia, the acromial periosteum and the trapezius fascia. The deltoid tendon is split in line with its fibers along the strong tendon of origin that divides the anterior and middle deltoid. This allows two strong "handles" on the deltoid for repair. This split is deepened under direct vision until the bursa is entered. Rotating the humerus provides easy differentiation between the deltoid (which does not move with humeral rotation) and the superficial surface of the cuff (which does).

On entering the subacromial aspect of the humeroscapular motion interface (see figures 23 and 24), the subacromial space is observed while the preoperatively identified crepitus-producing movements are carried out. This step reveals the cause of the crepitance, which is usually some combination of roughness on the undersurface of the acromion, hypertrophic bursa, adhesions between the cuff and acromion, roughness of the superior surface of the rotator cuff or a prominent tuberosity. By gently rotating the arm, most of the cuff can be brought to the incision as pointed out by Codman. The rotator cuff is thoroughly explored and palpated for evidence of superior surface blisters, partial tears, thinning, or full-thickness defects. Although deep surface cuff fiber failure cannot be seen through this approach, it is also true that such fiber disruption cannot be causing the subacromial crepitance. The methylene blue "dye test" of Hiro Fukuda (Fukuda, Mikasa, 1983) or more recently the "Fukuda-lite" test with saline is used to evaluate shoulders with suspicious cuff integrity. In this test fluid is injected to distend the glenohumeral joint to further explore suspected thinning or small cuff defects.

Hypertrophic bursa is resected. Superior surface cuff defects are smoothed by either resection of their protruding aspects or occasionally by reattaching a superior surface cuff flap. Prominences of the tuberosity are smoothed so that the tuberosity passes easily beneath the coracoacromial arch. The undersurface of the coracoacromial arch is palpated to identify areas of roughness or prominence. These are smoothed with a "pine cone" burr, although an osteotome or rongeur may be useful for larger lesions. No attempt is made to resect the coracoacromial ligament unless it can be demonstrated to be the cause of the subacromial roughness.

If a substantial amount of bone needs to be removed, a thin-bladed osteotome is used (see figure 25). The osteotomy is oriented in line with the extrapolated undersurface of the posterior acromion (identified by palpation and direct vision). Care is taken that the osteotomy does not continue into the posterior acromion or scapular spine. The undersurface of the acromion is then smoothed using a "pine cone" power bur, taking care that no spurs are left laterally in the deltoid origin or medially at the acromioclavicular joint. The shoulder is thoroughly irrigated to remove all bone fragments.

Additional surgery is avoided unless clearly indicated. Inferiorly directed acromioclavicular osteophytes are resected if they scrape on the cuff. The biceps is left undisturbed unless it appears to be seriously inflamed, obviously unstable, or doomed to imminent rupture, in which case we perform a tenodesis to the proximal humerus.

The shoulder is gently manipulated through a complete range of motion to assure the absence of stiffness or additional adhesions. The entire humeroscapular motion interface (see figures 23 and 24) is inspected to assure absence of adhesions and for other pathology. Before the procedure is concluded, the upper surface of the cuff and tuberosities and the undersurface of the coracoacromial arch are carefully palpated to assure the absence of residual roughness. The entire range of passive shoulder motion must be free of subacromial crepitance.

On closure, a secure deltoid reconstitution is top priority so that early postoperative motion may be instituted. The deltoid is repaired by side-to-side closure of the medial and lateral aspects of the tendon split using No. 2 nonabsorbable suture (see figure 26). The tendon is secured to the acromion, using suture to bone as necessary. Suture from the medial hole is passed through the lateral part of the deltoid tendon and suture from the lateral hole is passed through the medial part of the deltoid tendon to effect a crisscross closure. This avoids the "telltale V" defect that reveals a poor deltoid closure. All knots are placed on the superficial aspect to avoid recreating subacromial roughness.

Postoperative program

After any type of subacromial surgery, there is a great potential for adhesions between the cuff and the arch. In cases of failed acromioplasty, such scarring seems to be a dominant feature and appears to be often related to delay in the institution of motion following the surgery. To avoid such problems, we begin motion as soon as possible, preferably with continuous passive motion in the recovery room (figure 27). CPM is set to move the arm slowly through an arc of 0 to 90 degrees of elevation and from 50 to 0 degrees of internal rotation. CPM is applied whenever the patient is in bed during their hospitalization, but is not continued after discharge. On the day of surgery, the patient is instructed in the "140/40 passive program" in which the opposite hand is used to assist the operated shoulder in achieving 140 degrees of elevation (see figure 9) and 40 degrees of external rotation (see figure 11). Emphasis is also placed on posterior capsular stretching, including cross body adduction (see figure 14), reaching up the back (see figure 13) and internal rotation of the abducted arm.

The early implementation of passive motion is facilitated if the procedure is performed under brachial plexus block (Tetzlaff, Yoon, 1994), which lasts from 12 to 18 hours. The post operative exercises are already familiar to the patient, having been performed as part of the preoperative trial of the Jackins program. The patient is allowed active use of the shoulder within the realm of comfort unless there is concern for the strength of the deltoid reattachment. Internal (see figure 15) and external rotation (see figure 16) strengthening exercises are also begun immediately. Deltoid strengthening is initiated at 6 weeks after the repair is secure (see figure 17). As soon as they can be performed comfortably, exercises to strengthen the scapular motors are added (see figures 18 and 18). Athletics are not allowed for three months after surgery and until normal motion and strength are regained.

Failure to achieve relief of symptoms

Such results occur in every series of acromioplasty, even if the technique of the procedure seems appropriate. The incidence of these failures ranges from 3 to 11 per cent. (Neer, 1972, Post and Cohen, 1985, Raggio, Warren, 1985, Thorling, Bjerneld, 1985) In Post and Cohen's series, 11 per cent continued to have significant pain after surgery. (Post and Cohen, 1985, Post and Cohen, 1986) Fifty-six per cent of those with weakness before surgery still had weakness postoperatively; 29 per cent of those with preoperative limitations of motion still had limitation of motion after surgery. The rate of return to high-level athletics or challenging occupations is lower. Tibone and colleagues (Tibone, Jobe, 1985) found that of 35 athletes having impingement syndrome treated by anterior acromioplasty, 20 per cent still had moderate to severe pain, and 9 per cent had pain at rest and with activities of daily living. Only 43 per cent returned to their preinjury level of competitive athletics, and only 4 of 18 returned to competitive throwing. Hawkins and coworkers (Hawkins, Chris, 1987) have shown that it is difficult for patients injured on the job to return to their original occupations following acromioplasty.

Why is this? Failure to achieve complete relief of symptoms through acromioplasty may indicate:

1. pathology other than subacromial roughness,
2. failure to achieve subacromial smoothness,
3. failure of deltoid reattachment,
4. excessive acromial resection,
5. postoperative complications such as dense scarring between the cuff and the acromion, or
6. failure of rehabilitation.

Many of these problems can leave the patient more symptomatic than before the surgery.

Acromioclavicular joint problems were thought to be responsible for five failures in Post's series, a "frequent cause of failure of surgical treatment" in the series of Penny and Welsh, (Penny and Welsh, 1981) and the cause of the only unsatisfactory result in Neer's series. In their series of patients having persisting problems after acromioplasty, Hawkins and colleagues (Hawkins, Chris, 1987) reported that forty-five per cent of the patients had a diagnosis other than continuing impingement, including acromioclavicular joint problems, cervical spondylosis, reflex sympathetic dystrophy, rotator cuff tear, thoracic outlet syndrome, glenohumeral osteoarthritis, and glenohumeral instability. Thirty-three per cent were thought to have continuing subacromial abrasion. The striking finding in this series was the relative lack of improvement in patients on workmen's compensation after revision acromioplasty. Even in these authors' series of primary acromioplasties, twenty-two per cent of the workmen's compensation cases had an unsatisfactory result, compared with eight per cent failure rate with non--workmen's compensation cases. (Hawkins and Brock, 1979) Post and Cohen (Post and Cohen, 1986) also observed that worse results were obtained from surgery performed for work-related impingement syndrome. This inability to return to work may be due to partial-thickness cuff tears, residual tendon scarring, and residual weakness. Post and Cohen emphasized the need for recovery of muscle strength before the laborer is returned to work; otherwise, recurrence can be anticipated. The difficulty of returning workers to their jobs after acromioplasty is reminiscent of the problems described by Tibone and coworkers (Tibone, Jobe, 1985) in returning athletes to a competitive level of function.

Bosley (Bosley, 1991) reported that most failures were attributed to either the underlying pathology or to failure of deltoid reattachment. Bjorkenheim et al (Bjorkenheim, Paavolainen, 1990) reported a failure rate of over twenty-five per cent, attributing the failures to "associated bony as well as soft-tissue subacromial lesions". Oglvie-Harris et al ( Ogilvie-Harris, Wiley, 1990) evaluated 67 shoulders in 65 patients who had pain and dysfunction for more than two years after an initial acromioplasty for impingement syndrome without a rotator cuff tear. In almost half of the cases there were "diagnostic errors" and even in those where there was a correct diagnosis and no operative errors, the failure rate was almost twenty per cent.

Radical acromionectomy may worsen a patient's comfort and function. This procedure removes the origin of the deltoid muscle and facilitates scar formation between the deltoid muscle and the rotator cuff. Neer and Marberry have pointed out that a radical acromionectomy may seriously compromise shoulder function without achieving subacromial smoothness. (Neer and Marberry, 1981) In their series of thirty patients, all had marked shoulder weakness and almost all had persistent pain. In the twenty shoulders reoperated upon, all had a retracted and scarred middle deltoid that was adherent to the cuff and humerus. Fifteen of the patients had residual cuff tears. Attempts to reconstruct these severely damaged shoulders were disappointing. The effects of loss of the deltoid attachment and the permanent contracture could not be reversed. In addition, these authors observed a high incidence of wound problems and infections following the radical acromionectomy, which further complicated their attempts at revision.

To help understand some of the other causes of unsuccessful acromioplasty, Flugstad and coworkers (Flugstad, Matsen III, 1986) reviewed nineteen patients referred to the University of Washington Shoulder and Elbow Service because of persistent pain and stiffness after open acromioplasty performed elsewhere. The average age was 42; 16 were male. Eleven patients had a traumatic onset of their shoulder problem; eight of these were work related. The average time of postoperative immobilization was four weeks. At the time of presentation, the patients complained of pain and stiffness. Physical examination revealed an average of 126 degrees of forward flexion and 36 degrees of external rotation and internal rotation so that the thumb could touch T12. In thirteen of these patients revision surgery was performedafter an exercise program failed to improve their symptoms. The average interval between the initial surgery and revision surgery was fifteen months. At the revision surgery, ten patients had roughness of the undersurface of the acromion. Five patients had distinct spurs protruding from the lateral or medial acromion; eight patients had large amounts of subacromial scarring in which heavy bands of cicatrix connected the undersurface of the acromion to the rotator cuff. Three patients had acromioclavicular joint spurs, one had a large ununited acromial fragment, and another had an os acromial. Although no patient had a full-thickness cuff tear, the incidence of partial-thickness deep surface or midsubstance cuff tears is unknown. The revision surgical procedure included excision of scar tissue, revision of the acromioplasty to assure adequate resection of the anterior and inferior acromion, resection of acromioclavicular spurs, inspection of the rotator cuff, and careful deltoid repair. Immediately after surgery, gentle range-of-motion exercises were initiated to minimize restriction from postoperative scar. Follow-up at an average of ten months postoperatively revealed substantial although incomplete improvement in comfort, range of motion, and ability to work.

This report emphasizes the importance of accurate diagnosis and effective subacromial smoothing. However, the key lesson was the importance of rapid restoration of full joint motion before restricting adhesions have the opportunity to form: the average patient in this series had a one-month delay between surgery and the implementation of motion.

Authors' preferred method for the management of failed acromioplasty

Patients who have had previous acromioplasty with unsatisfactory results need to be carefully reevaluated to determine presence of stiffness, weakness, instability or persisting roughness. The social and vocational context of the shoulder problem must be reevaluated as well.

The Jackins non-operative program is instituted, even if the patient has already "had therapy"; since surgery has failed once already, there is plenty of time for conservatism and a period of observation.

Patients with positive tendon signs (see figure 29) may be considered for cuff imaging studies if these signs are refractory to rehabilitation. Vocational rehabilitation may be essential; if one procedure has not gotten the patient back to their job, the odds would seem not much better the second go-round.

Reoperation is considered in well-motivated patients with evidence of residual subacromial roughness or stiffness that is attributable to postoperative scarring in the humeroscapular motion interface (see figures 30 and 31). In contrast to primary acromioplasty, we are willing to reoperate on patients with refractory shoulder stiffness, because this stiffness may be due to dense scarring between the cuff and the acromion which cannot be managed nonoperatively. Our revision procedure is identical with the primary subacromial smoothing described earlier.

Signs and symptoms

Intrasubstance and articular surface

Judging from the cadaver studies reviewed earlier in this chapter, intrasubstance and articular surface partial thickness cuff tears represent the commonest forms of cuff involvement. These lesions usually involve the supraspinatus tendon near its anterior insertion, but may also involve the infraspinatus and subscapularis. Clinical observation of patients with documented partial thickness cuff lesions suggests that they produce symptoms analogous to other partial thickness tendon lesions, such as a partial Achilles tear, a partial tear of the patellar tendon, or a partial tear of the tendon of origin of the extensor carpi radialis brevis ("tennis elbow"). These partial tendon lesion symptoms include stiffness of the joint on passive motion in a direction that stretches the tendon and tendon signs, i.e. pain or weakness on isometric contraction of the tendon's muscle (see figure 32). These partial tendon lesions are often much more painful than full thickness tears. This is because, in contrast to full thickness tears, partial thickness defects of the cuff give rise to stiffness and unphysiologic tension on the remaining fibers.

In its less common form involving the bursal aspect of the cuff tendon, partial thickness cuff lesions may be associated with subacromial abrasion, yielding subacromial crepitance on passive joint motion.

There is not a lot of published information regarding the results of operative treatment for partial thickness cuff lesions. Fukuda and colleagues (Fukuda, Mikasa, 1983, Fukuda, Mikasa, 1987) described the management of six patients with partial-thickness bursal-side tears by acromioplasty and/or wedge resection with tendon repair to bone. They used an intraoperative "color test" in which dye was injected into the shoulder joint to indicate the extent of joint side tears. The results were satisfactory in 90 per cent of cases. Itoi and Tabata (Itoi and Tabata, 1992b) reported their results in managing 38 shoulders with partial thickness cuff lesions. The average follow-up period of 4.9 years and the average age at operation was 52.2 years. Three types of lesions were identified: superficial (12 shoulders), intratendinous (three), and deep surface tears (23). The authors performed full-thickness resection of the cuff including the lesion and repaired the defect with side-to-side suture (13 shoulders), side-to-bone suture (eight), fascial patch grafting (16), or side-to-bone suture with fascial patch grafting (one). The overall results were satisfactory in 31 shoulders (82%). The results were not affected by the tear types, operative methods, or follow-up period.

Arthroscopic treatment

Andrews et al (Andrews, Broussard, 1985) presented thirty-six patients with partial thickness tears of the supraspinatus portion of the cuff treated with arthroscopic debridement of the rotator cuff defect. No acromioplasty was performed. The average age was 22.5 and 64% of the patients were baseball pitchers. Of the 34 patients available for follow-up, 85% had an excellent (26 patients) or good (3 patients) result and were able to return to sports. The authors suggested that the debridement may initiate a healing response. Arthroscopy revealed a tear of some part of the glenoid labrum in all patients. Six had partial tears of the long head of the biceps tendon. These observations point to the difficult of deciding which surgical findings are responsible for the patient's symptoms.

Ogilvie-Harris and Wiley (Ogilvie-Harris and Wiley, 1986) reported on arthroscopic treatment of 57 incomplete tears of the rotator cuff with symptoms of impingement. These tears were debrided and no acromioplasty was performed. Half of the patients improved.

Wiley (Wiley, 1985) reported on thirty-three patients treated arthroscopically for partial tears of the rotator cuff. Only three patients achieved a satisfactory result.

Ellman reported good results from arthroscopic acromioplasty performed in conjunction with arthroscopic debridement of partial thickness tears of the rotator cuff. (Ellman and Kay, 1991)

Esch et al, (Esch, Ozerkis, 1988) in 1988, reported on thirty-four patients with stage II rotator cuff disease and partial thickness rotator cuff tears treated with arthroscopic acromioplasty and tear debridement. Twenty-eight patients were satisfied with their results; 16 patients were rated excellent, 10 were good, 6 were fair, and 2 were rated as poor.

Gartsman (Gartsman, 1990) presented forty patients with partial thickness rotator cuff tears in a group of 125 patients treated with arthroscopic acromioplasty. Of these partial thickness tears, thirty-two involved the articular surface of the supraspinatus tendon and four tears involved the bursal side. Four infraspinatus tears were identified, three of which involved the articular surface. Notably, in these forty patients, there were twenty-seven cases of labral fraying with six instances of biceps/labral complex detachment, again indicating the difficulty of relating symptoms to surgical findings. Of the forty patients, thirty-three (83% satisfactory results) had major improvement in their ratings for pain, activities of daily living, work and sports, at an average of 28.9 months post-arthroscopic debridement. Two patients, who had an unsatisfactory result, had a second operation: one, open acromioplasty and the other, repair of the rotator cuff with satisfactory results. Of the thirty patients in this group engaged in sports preoperatively, ten patients returned to those sports at the same level of performance as before the symptoms had started.

Altchek et al (Altchek, Warren, 1990) reported four of six good or excellent results in patients with partial thickness rotator cuff tears treated with arthroscopic acromioplasty and debridement of the rotator cuff defect.

Roye et al (Roye, Grana, 1995) presented thirty-eight patients with partial thickness rotator cuff tears (thirty-two involving the supraspinatus) treated with arthroscopic acromioplasty. A satisfactory result was achieved in 95%.

As part of a larger series, Ryu (Ryu, 1992), reported on thirty-five patients with partial thickness rotator cuff tears treated with arthroscopic acromioplasty. Thirty of thirty-five patients (86%) were rated with excellent or good results (5 fair, no poor) at a minimum follow-up of 12 months. Of the group with partial tears, four were found to involve only the articular surface. Three of these four were considered failures

In 1994, Olsewki and Depew (Olsewski and Depew, 1994) reported on their experience with 61 consecutive patients treated with arthroscopic acromioplasty and debridement of the rotator cuff defect (17 of 21 patients (81%) with a partial thickness rotator tear rated a satisfactory result (UCLA rating scale). This was identical to the result achieved in 27 patients treated with arthroscopic acromioplasty for rotator cuff "tendonitis" with an intact rotator cuff. As was the case with the series of Roye et al (Roye, Grana, 1995), the extent of the tear did not statistically affect the result.

From this group of reports it is difficult to define (1) the indications for surgery, (2) which aspects of the patients' pathologies were responsible for their symptoms, (3) why from 15-50% of patients failed to achieve a satisfactory result, and (4) which aspect of the surgery (acromioplasty or debridement) was responsible for improvement after surgery. It seems likely that those patients benefiting from this procedure were able to heal their tendon debridement in a way that stabilized the insertional mechanism, distributing the loads from muscle to bone in a way that prevented disproportionately large loads from being concentrated on the neighboring intact tendon fibers.

Authors' preferred method of treating partial thickness cuff lesions

Nonoperative treatment

The nonoperative management of partial thickness cuff tears is similar to that for subacromial abrasion described earlier in this chapter.Just as with partial lesions of the Achilles, patellar or extensor radialis brevis tendons, the program must emphasize stretching against all directions of tightness, including internal rotation (see figure 33), cross-body adduction (see figure 34), elevation (see figures 35 and 36) and occasionally external rotation (see figure 37). As in a tennis elbow rehabilitation program, when a comfortable normal range of passive motion is reestablished, gentle progressive muscle strengthening is instituted (see figures 38 and 39) An emphasis is always placed on gentle and comfortable progress of this rehabilitation program. The goal of this program is to assure that the scar collagen which forms in the defect will become as supple as normal tendon; otherwise scar contracture will tend to concentrate the loads of the cuff on the lesion leading to recurrence and propagation of injury.

Operative treatment - Open surgery

Just as is the case for partial Achilles, patellar and extensor carpi radialis brevis tendon lesions, there is no surgical treatment which reliably restores the tendon to its normal condition. Preoperatively, it is important to determine whether the patient's primary problem is due to stiffness or to difficulties upon active muscle contraction so that the procedure can be biased accordingly. On the one hand, sectioning of the fibers which remain intact (as in a tennis elbow release) may worsen the problem of weakness; although this may be the basis of the arthroscopic "debridement" advocated by some surgeons for this lesion. On the other hand, excision of the defect and repair would worsen the problem of stiffness. (Zuckerman, Leblanc, 1991) Furthermore, such surgical tightening of involved part of the cuff would cause the area of damage and repair to bear the majority of the load when the cuff muscles contract (reminiscent of the "quadregia" phenomenon in hand surgery). Thus excision and repair of partial thickness cuff lesions should include efforts to assure that the tendon load is distributed evenly at the insertion, by carrying out a repair that is isometric, allowing uniform load distribution and by carrying out a release of the capsule tightened in the repair. (Harryman, Matsen, 1996)

The surgical exposure to the partial thickness cuff lesion is identical to that described for the management of subacromial roughness (see figure 40). If symptoms are related to subacromial abrasion (i.e. symptomatic subacromial crepitance), subacromial smoothing is performed as described previously in this chapter.

The decision to convert a partial thickness cuff defect to a full thickness defect and then to repair it (see figure 41) is based on the patients preoperative evaluation and surgical findings. The thickness of the cuff can be determined at surgery by inspection, palpation and the Fukuda test described above. A depth gauge or calibrated nerve hook inserted in the area of the lesion may help determine the percentage of the tendon that remains intact. If the decision is made to perform an open repair, a tenotomy is performed in the most suspicious area along the line of the tendon fibers to explore the full thickness of the tissue. If, as is usually the case, the defect is within the substance of the tendon or on its deep surface near the anterior insertion of the supraspinatus, a longitudinal tenotomy and capsulotomy are performed along the anterior aspect of the supraspinatus near the rotator interval. This cut is then extended at right angles posteriorly through the partially detached cuff at its insertion to the greater tuberosity, turning back the flap of cuff until normal tendon of full thickness is encountered. Next an attempt is made to retrieve and consolidate any split laminations of cuff which may have retracted medially (see figure 41). These are usually on the deep articular surface where the cuff lesion begins and may have retracted up to 1 cm. Release of the coracohumeral ligament and the rotator interval capsule from the base of the coracoid (see figures 42 and 43) as well as release of the capsule from the glenoid lip (see figure 44) will minimize tension on the repair. Then the full thickness defect is repaired (see figures 45 and 46) with care to render the cuff insertion isometric with respect to all it fibers and smooth on its superior surface. Finally, with the anterior undersurface of the acromion in full view, the shoulder is put through a full range of motion to verify the elimination of any subacromial abrasion (see figure 47) and to assure that the repair has not restricted shoulder motion.

Postoperative management is the same as for the repair of full thickness defects with a particular emphasis on continuous passive motion (see figure 48) and on the early restitution of a full range of motion to prevent stiffness and adhesions (see figures 34, 35, 36, 37, 49, and 50).

Diagnosis

A full thickness defect of one or more of the cuff tendons can be demonstrated on ultrasonography, arthrography, MRI, arthroscopy or open surgery.

While the diagnosis is not difficult, a number of key factors must be considered in selecting the appropriate treatment for cuff defects.Some defects cannot be repaired, because as McLauglin pointed out they offer only "rotten cloth to sew" (McLaughlin, 1944, McLaughlin, 1962, McLaughlin, 1963, McLaughlin and Asherman, 1951) The recognition that full thickness cuff tears may exist without clinical symptoms (Harryman, Mack, 1991, Matsen, Lippitt, 1994, Milgrom, Schaffler, 1995, Pettersson, 1942, Sher, Uribe, 1995) cautions that cuff defects need not be repaired just because they are there.

Nonoperative treatment

Substantial data are available on the results of nonoperative treatment for full thickness cuff defects. The programs generally include some combination of "compound tincture of time" along with physical therapy, administration of nonsteroidal anti-inflammatory medications, rest, avoidance of precipitating activities, and steroid injections.

Improvement with nonoperative management was noted to be 33 per cent in Wolfgang's series (Wolfgang, 1978), 44 percent in Takagishi's series (Takagishi, 1978), 59 percent in Samilson and Binder's series (Samilson and Binder, 1975) and 90 percent in Brown's series. (Brown, 1949)

Steroid injections do not seem to be a major enhancement to the nonoperative management program. Although Weiss (Weiss, 1981) presented some evidence that patients with arthrographically proven cuff tears are symptomatically improved by intra-articular injections, there is little evidence for a protracted benefit from this method. Other observers found that steroid injections offered no benefit to patients with cuff tears. Coomes and Darlington, (Coomes and Darlington, 1976, Darlington and Coomes, 1977) Lee and colleagues, (Lee, Lee, 1974) and Connolly (Connolly, 1972) compared steroid and local anesthetic injections in patients with tendinitis and tendon tears. They found a small subjective benefit in relief of pain but no effect on function in the steroid-treated group.

There has been a recent resurgence of reports confirming the value of nonoperative management for chronic cuff tears. Bartolozzi et al (Bartolozzi, Andreychik, 1994) studied the factors predictive of outcome in 136 patients with cuff disease who were treated nonoperatively. Mean followup was 20 months (range, 6-41 months). The authors found 66-75% good or excellent results with indication that the clinical result improved significantly as followup duration increased. Prognostic factors that were associated with an unfavorable clinical outcome included a rotator cuff tear over one square centimeter, a history of pretreatment clinical symptoms for over one year duration, and significant functional impairment at initial presentation.

Hawkins and Dulap (Hawkins and Dunlop, 1995)found that over half of patients with full thickness cuff tears treated with a supervised nonoperative program of rotator strengthening exercises obtained satisfactory results at an average of four years followup. Bokor et al (Bokor, Hawkins, 1993) managed 53 patients (average age 62 years) with full thickness cuff tears documented arthroscopically using a program of non-steroidal medications, stretching, strengthening, and occasional steroid injections. At an average of 7.6 years later, thirty-nine of the 53 patients (74%) had only slight or no shoulder discomfort. Of the 28 shoulders presenting within three months of injury, 24 (86%) were rated as satisfactory at the time of latest evaluation. Of the 16 patients who initially had had shoulder pain for over 6 months, only nine (56%) were rated as satisfactory. Most patients showed improvement with regard to their ability to perform activities of daily living. Average active total elevation was 149 degrees compared with 121 degrees at initial presentation. Thirty-two of the 34 patients examined (94%) had evidence of weakness on muscle testing and 19 (56%) had demonstrable muscle atrophy.

Itoi and Tabata (Itoi and Tabata, 1992a) followed 62 shoulders with complete rotator cuff tears which were treated conservatively from 1980 until 1989. The follow-up period averaged 3.4 years. Fifty-one shoulders (82%) rated satisfactory. The overall scores of pain, motion, and function improved significantly. The authors concluded that conservative treatment affords satisfactory results when it is given to the patients with well-preserved motion and strength, although in some cases function may deteriorate with time.

In our own practice we have followup data on 56 patients (23 female, 33 male) with full thickness cuff tears managed nonoperatively. The average age was 61 ± 10 years (range 45-84) and the mean followup time was 25 months. The initial and final responses to the questions of the Simple Shoulder Test are shown in table 1, below.

Table 1

Taken together these results clearly offer encouragement for a trial of nonoperative management for chronic full thickness cuff tears, particularly in cases where the prospect of achieving a durable cuff repair appears doubtful.

Operative treatment

Cuff repair

Patient Selection

Substantial information bearing on the potential reparability of a rotator cuff defect can be obtained from the history along with the physical examination and plain radiographs (see table 2).

Table 2

Acute tears in younger, healthy individuals without prior shoulder disease are more likely to be repairable. Long-standing tears associated with major weakness in older patients carry a poorer prognosis. The prognosis for a durable repair is even worse if the history reveals the administration of local or systemic steroids, smoking, or difficulties in healing previous injuries or surgeries.

These guidelines are derived from our experience, but also are supported by the literature. Postacchini et al (Postacchini, Perugia, 1992) found in a study of 73 cuff repairs, that while seventy-three percent of the cases had satisfactory results, rotator cuff repair is almost always successful in patients with more than 60 degrees of active arm flexion and either small or medium-size tears. Less than two-thirds of the patients with major tears and less than 60 degrees of motion achieved satisfactory results, particularly if there was cuff muscle atrophy.

Watson (Watson, 1985) reviewed the surgical findings in 89 patients with major ruptures of the cuff. He found that all seven patients who had had no local steroid injections had strong residual cuff tissue. Thirteen of 62 patients having one to four steroid injections had soft cuff tissue that held suture poorly; 17 of the 20 patients having more than four steroid injections had very weak cuff tissue; these shoulders with weak cuff tissue had poorer results after surgical repair.

Misamore et al (Misamore, Ziegler, 1995) evaluated 107 consecutive cuff repairs, including 24 patients on Workers' Compensation and 79 who were not. Although other factors such as the age and sex of the patients, the size of the tear of the rotator cuff, and the preoperative strength, pain, and active range of motion of the shoulder were comparable, only 54 per cent of the shoulders covered by Workers' Compensation were rated good or excellent, compared with 92 per cent who were not. Forty-two per cent of the patients on Workers' Compensation returned to full activity, compared with 94 per cent who were not.

Samilson and Binder listed the following most reasonable indications for operative repair of non-acute cuff tears: (Samilson and Binder, 1975)

1. a patient "physiologically" younger than 60 years,
2. clinically or arthrographically demonstrable full-thickness cuff tear,
3. failure of patient to improve under nonoperative management for a period not less than six weeks,
4. patient's need to use the involved shoulder in overhead elevation in his or her vocation or avocation,
5. full passive range of shoulder motion,
6. patient's willingness to exchange decreased pain and increased external rotator
strength for some loss of active abduction, and
7. ability and willingness of the patient to cooperate.

Grana et al (Grana, Teague, 1994) reviewed their experience with 54 patients having open repair of chronic cuff tears. They concluded that pre repair arthroscopic evaluation did not affect the functional outcome, but did increase the cost by about $2000 per patient.

Laboratory studies on repair techniques

Gerber et al (Gerber, Schneeberger, 1994) studied the mechanical properties of several techniques of tendon-to-bone suture employed in rotator cuff repair in cadavers. Two simple stitches failed at 184 N; four simple stitches failed at 208 N. Two Mason-Allen stitches failed at 359 N. These results indicate that in addition to the quality of the bone and the quality of the cuff tissue, the number of sutures and the suture technique affect the load to failure.

Technique of cuff repair

A most important recent study bearing on the technique of cuff repair was published by Zuckerman et al. (Zuckerman, Leblanc, 1991) These authors used a cadaver model to determine the effect of arm position and capsular release on the tension in the repaired tendon as reflected by strain gauges on the greater tuberosity. They found that with repair of supraspinatus-only defects, tension in the repair increased significantly as the arm was lowered from 30 to 15 degrees of abduction. Release of the capsule from the glenoid rim (see figures 51 and 52) significantly reduced the tension at 15 and 0 degrees of abduction. For tears involving the supraspinatus and infraspinatus, abduction of at least 30 degrees was required to reduce tension in the repair. Release of the capsule from the glenoid (figure 52) resulted in a 30% reduction in repair tension when the arm was adducted.

Warner et al (Warner, Krushell, 1992) studied the relationships of the suprascapular nerve to the cuff muscles in thirty-one cadaveric shoulders. The suprascapular nerve ran an oblique course across the supraspinatus fossa, was relatively fixed on the floor of the fossa, and was tethered underneath the transverse scapular ligament. In eighty-four percent of the shoulders, there were no more than two motor branches to the supraspinatus muscle, and the first was always the larger of the two. In eighty-four per cent, the first motor branch originated underneath the transverse scapular ligament or very near it. In one shoulder (3 percent), the first motor branch passed over the ligament. The average distance from the origin of the long tendon of the biceps to the motor branches of the supraspinatus was three centimeters. In forty-eight percent, the infraspinatus muscle had three or four motor branches of the same size. The average distance from the posterior rim of the glenoid to the motor branches of the infraspinatus muscle was two centimeters. The motor branches to the supraspinatus muscle were fewer, usually smaller, and significantly shorter than those to the infraspinatus muscle. The standard anterosuperior approach allowed only one centimeter of lateral advancement of either tendon and limited the ability of the surgeon to dissect safely beyond the neurovascular pedicle. The advancement technique of Debeyre et al., or a modification of that technique, permitted lateral advancement of each muscle of as much as three centimeters and was limited by tension in the motor branches of the suprascapular nerve. In some situations, the safe limit of advancement may be even less. The authors concluded that lateral advancement of the rotator cuff is limited anatomically and may place the neurovascular structures at risk.

Surgical approaches

The surgical approaches to the complete cuff tear vary substantially. These include a saber cut, (Codman, 1911) an anterior approach through the acromioclavicular joint, (Bateman, 1963) a posterior approach, (Debeyre, Patte, 1965) and an "extensile" approach. (Ha'eri and Wiley, 1980) Many authors prefer the anterior acromioplasty approach, taking care to preserve the deltoid attachment and acromial lever arm. (Cofield, 1981, Cofield, 1985, Neer, 1972, Neer and Marberry, 1981) This technique provides excellent exposure of the common sites of lesions--the anterior cuff, biceps groove, undersurface of the acromion, and acromioclavicular joint.

Packer and coworkers, (Packer, Calvert, 1983) reporting on 63 cuff repairs followed for an average of 32.7 months, found that those performed with acromioplasty yielded more pain relief than cuff repair without acromioplasty. If greater access to the supraspinatus is needed, the acromioclavicular joint can be excised. (Neer, 1983) Debeyre and associates (Debeyre, Patte, 1965) described a posterior approach with acromial osteotomy. Ha'eri and Wiley described an approach that is extensile through the acromioclavicular joint to the supraspinous fossa. (Ha'eri and Wiley, 1980)

Repair methods

Operative techniques for repairing full-thickness cuff defects include tendon-to-tendon repair and tendon advancement to bone. McLaughlin (McLaughlin, 1944, McLaughlin, 1962, McLaughlin, 1963, McLaughlin, 1994, McLaughlin and Asherman, 1951)described his approaches to transverse ruptures (reinsertion into bone), longitudinal rents (side-to-side repair), and tears with retraction (side-to-side repair followed by reinsertion of the retracted portion of cuff into the head wherever it will reach with ease with the arm at the side). Although many of his principles are still applied today, most authors would not concur with his use of the transacromial approach or his belief that "distinct benefits are gained by excising and discarding the outer fragment of the divided acromion." (McLaughlin, 1944, McLaughlin, 1963) Hawkins and colleagues used side-to-side repair for small tears and tendon-to-bone repair for larger defects. (Hawkins, Misamore, 1985) Cofield has emphasized the identification of the tear pattern and the use of direct repair and flaps as indicated by the tear pattern. (Cofield, 1982, Cofield, 1985) Nobuhara et al (Nobuhara, Hata, 1994) reviewed, at an average of seven years, one hundred eighty-seven patients (189 shoulders) treated surgically for massive rotator cuff tears using either a tendon-to-tendon repair or the McLaughlin procedure. Ninety-five per cent of the patients were 45 years or older. Excellent or good functional results were attained in 93% of patients. Thirty-three percent of those who underwent tendon to tendon repair complained of pain after overuse compared with only 18% who had the McLaughlin Procedure.

A number of authors have described extensive tendon mobilization or advancement of major tendon flaps to repair large defects. Cofield recommended the transposition of the subscapularis for repair of large cuff defects. (Cofield, 1982) In this technique the subscapularis and the anterosuperior capsule are freed from the anteroinferior capsule, leaving the middle and inferior glenohumeral ligaments intact. The tendon is then transferred superiorly to the anterior aspect of the greater tuberosity. Most patients required postoperative protection in an abduction splint or cast for four to five weeks. These patients, who had severe symptoms of pain and limitation of function preoperatively, had less pain and slight improvement in active motion; 12 of 26 patients gained more than 30 degrees of active abduction, and 4 lost this amount of motion. Two patients disrupted their repair during the acute postoperative period. Of the twenty-six, twenty-five were satisfied with the procedure.

Karas and Giachello (Karas and Giachello, 1996) recently reported their results with twenty patients treated with acromioplasty and subscapularis transfer for massive (>5 cm) tears of the cuff in which direct tendon to bone reconstruction could not be achieved. At a mean of 30 months after surgery, seventeen patients were satisfied. Nine had weakness and discomfort with overhead activities and two had lost active elevation despite relief of pain. The authors found this procedure useful when "traditional" methods of repair were insufficient, but cautioned against its use when patients had full functional elevation preoperatively.

In less than five per cent of his cuff repairs, Neer (Neer, 1983) shifted the infraspinatus and upper half of the subscapularis superiorly to close a defect in the supraspinatus, leaving the lower half of the subscapularis, the teres minor, and the intervening capsule intact. He described the use of a second incision posteriorly for better mobilization of the infraspinatus toward the top of the greater tuberosity. Neviaser and Neviaser (Neviaser and Neviaser, 1982) described the transposition of both the subscapularis and the teres minor to close the defect. Debeyre and colleagues and others described the use of a supraspinatus muscle slide to help close major cuff defects. (Debeyre, Patte, 1965, Ha'eri and Wiley, 1980, Ha'eri and Wiley, 1981) Ha'eri and Wiley (Ha'eri and Wiley, 1981) used the supraspinatus advancement technique of Debeyre; most of their 18 patients achieved satisfactory results.

Latissimus transfers as described for Erb's palsy (Phipps and Hoffer, 1995) have been used to manage large cuff defects. Gerber (Gerber, 1992) reported on sixteen irreparable, massive rotator cuff tears treated with latissimus dorsi transfer and reviewed after an average of 33 months. Pain relief was satisfactory in 94% of the shoulders at rest and in 81% on exertion. Flexion was 83 degrees preoperatively and 135 degrees postoperatively. If the subscapularis was torn and could not be adequately repaired, latissimus dorsi transfer was of no value. In cases with good subscapularis function but irreparable defects in the external rotator tendons, restoration of approximately 80% of normal shoulder function was obtained.

A flap of deltoid has been used to cover cuff defects. Thur and Julke (Thur and Julke, 1995) analyzed the results of shoulder reconstruction using an anterolateral deltoid muscle flap plasty in 100 patients with rotator cuff lesions which were at least 5 x 5 cm in size. Ninety per cent of patients were satisfied. Shoulder function improved significantly, and 72% recovered their strength completely. Most of the patients were able to work after six months. The overall result was good to very good in 83%.

Dierickx and Vanhoof (Dierickx and Vanhoof, 1994) reviewed twenty patients with a painful massive, irreparable rotator cuff tear treated with an open partial acromionectomy and an anterior deltoid muscle inlay flap. After follow-up averaging 12 months, 17 out of 20 patients were satisfied, and the UCLA score improved from an average of 9.35 to an average of 25.7 Active forward flexion improved in 17, and strength of forward flexion improved in 15 patients.

As an alternative approach to surgery for massive tears, Burkhart et al (Burkhart, Nottage, 1994) repaired the margins of the tear to restore force transmission, believing that complete coverage of the defect was not essential. In fourteen patients this procedure led to improvement in active elevation from 59.6 degrees to 150.4 degrees. Strength improved an average of 2.3 grades on a 0-to-5-point scale. The average score on the UCLA Shoulder Rating Scale improved from a preoperative value of 9.8 to a postoperative value of 27.6. All but one patient was very satisfied

Some authors have used biological and prosthetic grafts to repair large cuff defects. Neviaser, (Neviaser, 1971) Bush, (Bush, 1975) and McLaughlin and Asherman (McLaughlin and Asherman, 1951) employed grafts from the long head tendon of the biceps to patch cuff defects. Ting and coworkers (Ting, Jobe, 1987) found that the electromyographic activity and size of the long head tendon of the biceps is significantly greater in patients with cuff tears compared with the uninjured shoulder. Their study suggests that the long head of the biceps may be a greater contributor to abduction and flexion in the shoulder with cuff tear than in the normal shoulder and that sacrificing the intracapsular portion of the tendon for grafting material may not be advisable. Heikel (Heikel, 1968) used fascia lata to close cuff defects, and both Heikel and Bateman described the use of the coracoacromial ligament. Freeze-dried rotator cuff has been used by Neviaser and coworkers (Neviaser, Neviaser, 1978) In this report, sixteen patients with massive tears had cadaver grafts, producing decrease in nocturnal pain in all sixteen. The change in shoulder function and strength was not reported. Post (Post, 1985) reported on preliminary results in five patients in whom a carbon fiber prosthesis was used to manage massive cuff deficiencies. Three had excellent to good results and two failed, one because of possible infection. The author states that these results are no better than with conventional repairs. Finally, synthetic cuff prostheses have been used by Ozaki and colleagues (Ozaki, Fujimoto, 1984) and Post. (Post, 1985) The former found that of 168 shoulders with cuff tears (almost all of which were "chronic" and "massive"), twenty-five could not be repaired by standard surgical techniques. Their defects were typically 6 * 5 cm. These patients had cuff reconstruction with Teflon fabric, Teflon felt, or Marlex mesh. This procedure was followed by a structured postoperative program, including the use of an abduction orthosis to keep the arm elevated in the plane of the scapula for two to three months and continued rehabilitation for three to six months. At an average of 2.1 years follow-up, 23 of 25 patients gained 120 to 160 degrees of abduction (the other 2 having had axillary nerve injury). Whereas twenty had reported continual or intolerable pain preoperatively, pain was absent in twenty-three patients at follow-up. The authors found that results were better with the thicker felt and now recommend the use of 3- to 5-mm-thick Teflon felt in their patients with massive defects.

Some authors recommend postoperative immobilization in an abduction splint, (Bakalim and Pasila, 1975, Bateman, 1963, Debeyre, Patte, 1965, Heikel, 1968) while others advise against this. (McLaughlin, 1963, Nixon and DiStefano, 1975)

Results of treatment

Neer and coworkers (Neer, Flatow, 1988) reported the results of 233 primary cuff repairs with an average follow-up of 4.6 years. Results were excellent (essentially normal), in 77 per cent, satisfactory in 14 per cent, and unsatisfactory in 9 per cent. The unsatisfactory ratings were usually due to lack of strength rather than pain and usually occurred in patients with long-standing, neglected tears. Hawkins and coworkers found that 86 per cent of their patients had relief of pain after repair. (Hawkins, Misamore, 1985) Recovery of strength was more common in patients with smaller tears. (Hawkins, Misamore, 1985) In other series pain relief was reported in 58 per cent, (Peterson, 1982) 60 per cent, (Heikel, 1968) 66 per cent, (Debeyre, Patte, 1965) 74 per cent, (Godsil and Linscheid, 1970) and 85 per cent. (Samilson and Binder, 1975)

Gore and associates (Gore, Murray, 1986) reviewed the results from 63 primary cuff repairs with an average of 5.5 years' follow-up. The shoulders without a traumatic onset were repaired an average of 32 months after the onset of symptoms, whereas those with a traumatic onset were repaired an average of 6 months after the traumatic episode. The surgical approach and technique varied somewhat but usually consisted of acromioplasty and tendon repair to bone or to adjacent tendon. Six shoulders had biceps tendon grafts. Most shoulders were immobilized at the side for four to six weeks, but twelve had immobilization in abduction. Subjective improvement was seen in 95 per cent of shoulders with repaired cuffs. Flexion averaged 126 degrees actively and 147 degrees passively. Most patients had marked relief of pain and minimal or no problems with activities of daily living. Patients with tears less than 2.5 cm long had better results than those with larger tears. The superior results with repair of smaller tears is consistent with the observations of Godsil and Linscheid (Godsil and Linscheid, 1970) and Post and coworkers. (Post, Silver, 1983) Watson (Watson, 1985) found that results were worse in patients with larger cuff defects, with multiple preoperative steroid injections, and with preoperative weakness of the deltoid. Ellman and colleagues (Ellman, Hanker, 1986) reported a 3.5-year follow-up of 50 patients having rotator cuff repair. Techniques of repair included tendon-to-tendon suture, reimplantation into bone, grafts, and tendon flaps. Comfort and function were usually improved by these procedures. Their report provides additional support for timely repair: patients with symptoms of longer standing had larger tears and more difficult repairs. Shoulders with Grade 3 or less strength of abduction before surgery had poorer results; those with an acromiohumeral interval of 7 mm or less also had poorer results. Arthrography was not consistently accurate in estimating the size of the tear.

Hawkins found that acromioplasty and cuff repair relieved the patients' pain and restored the ability to sleep on the affected side in most patients. Seventy-eight per cent were able to use the arm above shoulder level after surgery, whereas only sixteen per cent were able to do so before surgery. Hawkins and coworkers (Hawkins, Misamore, 1985) found that the results of cuff repair were worse in patients on Workmen's Compensation. Only two out of fourteen patients unable to work because of cuff tears could return to work after surgery, whereas eight of nine patients not on Workmen's Compensation did return to work after operation. Other series of cuff repairs include those of Codman, (Codman, 1934b) Moseley, (Moseley, 1952) Neviaser, (Neviaser, 1971) Wolfgang, (Wolfgang, 1978) Bakalim and Pasila, (Bakalim and Pasila, 1975) Bassett and Cofield, (Bassett and Cofield, 1983) Earnshaw and coworkers, (Earnshaw, Desjardins, 1982) Packer and associates, (Packer, Calvert, 1983) Post and colleagues, (Post, Silver, 1983) Samilson and Binder, (Samilson and Binder, 1975) and Weiner and Macnab. (Weiner and Macnab, 1970a) Cofield (Cofield, 1985) averaged the results of many reports in the literature and found that pain relief occurred in 87 per cent (range 71 to 100 per cent), and patient satisfaction averaged 77 per cent (range 72 to 82 per cent). The reader is encouraged to compare and contrast these results with those following non operative treatment which was described earlier in this chapter.

Some reports focus on the results of acute repairs. Bakalim and Pasila reviewed their series of 55 patients with arthrographically verified rupture of the cuff tendons treated surgically. (Bakalim and Pasila, 1975) Whereas only half of the workers were able to return to their previous work, all workers operated upon within one month of a traumatic rupture of the cuff were able to return to their jobs. Bassett and Cofield (Bassett and Cofield, 1983) presented a series of 37 patients having surgical repair within three months of cuff rupture. At an average follow-up of seven years, active abduction averaged 168 degrees for those having repair within 3 weeks and 129 degrees for those having repair within 6 to 12 weeks after injury. Patients with small tears averaged 148 degrees and those with large tears averaged 133 degrees of elevation. The authors concluded that surgical repair must be considered within 3 weeks of injury to obtain maximal return of shoulder function.

The importance of continued postoperative exercises is emphasized by the data of Walker and associates, (Walker, Couch, 1987) who measured the isokinetic strength of the shoulder after cuff repair. They found a significant increase in strength between 6 and 12 months after surgery. One year after operation, abduction was 80 per cent of normal and external rotation was 90 per cent of normal. Brems (Brems, 1987 Jan) found that the strength of external rotation after cuff repair averaged 20 per cent at three months, 38 per cent at six months, 57 per cent at nine months, and 71 per cent at one year.

Rokito et al (Rokito, Zuckerman, 1996) followed at 3 month intervals the isokinetic strength of 42 patients having repair of full thickness defects. The torques for the operated shoulder (as a percent of the opposite uninvolved shoulder) are shown in table 3).

Table 3

Recovery of strength correlated primarily with the size of the tear: for small and medium sized tears, the recovery of strength was almost complete during the first year. For large and massive tears, recovery was slower and less consistent. The authors concluded that at least a year is required to regain strength after a cuff repair.

Kirschenbaum et al (Kirschenbaum, Coyle, 1993) came up with very similar results in their evaluation of 25 shoulders tested isokinetically with a pain-relieving subacromial lidocaine injection before and after cuff repair (see table 4).

Table 4

The analysis of the results of cuff repair is hampered by lack of a uniform approach to the description of

1. the shoulder's preoperative functional status,
2. the magnitude and location of the cuff defect,
3. the quality of the tissue available for repair,
4. the anatomical integrity at followup, and
5. the post operative functional status.

The need for correlation of anatomical and functional outcomes is demonstrated by the surprisingly good results obtained with debridement for irreparable cuff tears. Neer, (Neer, 1972) Rockwood, (Rockwood, 1983 and 1987) and others have reported that in certain cases when the cuff cannot be repaired, comfort and function may be improved by debridement of the shreds of residual cuff and subacromial smoothing followed by muscle strengthening and range-of-motion exercises. The realization that patients may have good function and comfort in the presence of major cuff defects makes the definition of "success" after a cuff repair challenging.

Interestingly, there have been few follow-up studies of the relationship of cuff integrity to the quality of the result after cuff surgery. Lundberg (Lundberg, 1982) followed 21 cuff repairs with arthrography and found leakage in 7. The results in the leaking cuffs were not as good as in those with sealed cuffs. Calvert and associates (Calvert, Packer, 1986) performed double-contrast arthrograms in 20 patients at an average of 30 months after operative repair of a torn cuff. In 17 of 20 shoulders the contrast leaked into the bursa, indicating a cuff defect. These defects were estimated to be small in 8, medium in 8, and large in 2. However, 17 patients had complete relief of pain, 15 had a full range of shoulder elevation, and 10 felt that they had regained full function. The authors suggest that a complete closure of the cuff is not essential for a good functional result and that arthrography may not be helpful in the investigation of failure of repair.

Ultrasonography appears to offer a greater potential for evaluating postoperative cuff integrity. Mack and coworkers (Mack, Nuberg, 1987) investigated the accuracy of ultrasonography in this regard. In a group of symptomatic postoperative shoulders that were subsequently reoperated, ultrasonography accurately diagnosed recurrent cuff tears in 25 of 25 cases and correctly confirmed cuff integrity in 10 of 11. Using expert ultrasonography Harryman (Harryman, Mack, 1991) correlated the integrity of the cuff with functional status following 105 surgical repairs of chronic rotator cuff tears in 89 patients at an average of five years postoperatively. The patients'