This Article Extract Full Text (PDF) Letters to the Editor: Submit a response Alert me when this article is cited Alert me when Letters to the Editor are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to My File Cabinet Download to citation manager Reprints and Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by BUDOFF, J. E. Articles by GUIDI, E. J. Search for Related Content PubMed PubMed Citation Articles by BUDOFF, J. E. Articles by GUIDI, E. J. Social Bookmarking             What's this?

Current Concepts Review - Débridement of Partial-Thickness Tears of the Rotator Cuff without Acromioplasty. Long-Term Follow-up and Review of the Literature*

JEFFREY E. BUDOFF, M.D., ROBERT P. NIRSCHL, M.D. and ERIC J. GUIDI, M.D., ARLINGTON, VIRGINIA

Investigation performed at Nirschl Orthopedic and Sportsmedicine Clinic, Arlington

	Introduction

Top Introduction Review of the Literature Analysis of the Present... Glenohumeral Instability... Evaluation of Patients Return to Activities of... Overview References

 
The rotator cuff, a musculotendinous unit that acts in combination with the deltoid to allow elevation of the shoulder, is a frequent source of pain and disability. In addition, the rotator cuff maintains the humeral head centered on the glenoid and opposes the superior translatory and shearing force of the deltoid by compressing the humeral head in the glenoid concavity as well as by imparting an inferiorly directed force vector to the proximal aspect of the humerus.

The term impingement syndrome has been used to describe symptoms related to the rotator cuff in the absence of a full-thickness tear of the cuff. The abnormality of the rotator cuff may range in severity from an acute strain or tendinitis to frank tearing. In all but the most acute situations, the lesion is one of chronic tendinopathy or angiofibroblastic hyperplasia rather than tendinitis. We recommend the use of the term tendinosis in lieu of the histologically inaccurate term tendinitis, since histopathological studies as well as analyses of cadaveric specimens have repeatedly shown that inflammatory cells are not part of the abnormality. Disorders of the rotator cuff may be due to an intrinsic factor, such as intrasubstance degenerative tearing or tendinosis caused by avascularity, aging, or overuse. Alternatively, they may be due to an extrinsic factor, such as outlet stenosis or glenohumeral instability¹⁹.

The most commonly performed procedure to treat symptoms of impingement in the absence of a full-thickness tear of the rotator cuff is acromioplasty. This treatment is based on the theory that primary abnormal acromial morphology (an extrinsic cause), popularized by Neer³⁶ in 1972, is the initiating factor leading to dysfunction of the rotator cuff and eventual tearing. Therefore, subacromial decompression, which is now often performed arthroscopically, is done in an attempt to alter presumed aberrant acromial morphology, thereby eliminating a theorized impingement on the rotator cuff.

However, newer evidence suggests that, in most patients who have an abnormality of the rotator cuff, the primary problem is intrinsic. The supraspinatus, a small and relatively weak muscle, is in a key position and is therefore susceptible to overuse and injury. When eccentric tensile overload occurs at a rate that is greater than the ability of the rotator cuff to repair itself, injury occurs, resulting in weakness of the musculotendinous rotator-cuff unit. Trauma to the shoulder may initiate the same process, and a weak, fatigued, or injured rotator cuff is unable to oppose the superior pull of the deltoid effectively and to keep the humeral head centered on the glenoid during elevation of the arm. This leads to inappropriate superior migration of the humeral head with active elevation of the arm, which functionally narrows the subacromial space. Continued dysfunction of the rotator cuff and further superior migration of the humeral head cause the greater tuberosity and the rotator cuff to abut against the undersurface of the acromion and the coracoacromial ligament, leading to signs of secondary impingement.

The secondary impingement between the greater tuberosity and the acromion may lead to reactive and degenerative osseous changes, such as osteophytic spurring of these structures. The already injured and weakened rotator cuff is then damaged further by this process of secondary impingement, especially by osteophytes on the undersurface of the acromion. Subacromial impingement does occur, therefore, but in most circumstances it is a secondary, rather than a primary, phenomenon. Superimposed trauma may then cause the weakened cuff to rupture. Thus, in most instances impingement is a dynamic process secondary to intrinsic failure of the tendon that, with time, results in reactive osseous changes, causing the classic radiographic changes seen with impingement syndrome. Erosion or exostosis of the greater tuberosity has been reported to occur in nineteen (20 per cent) of ninety-six cases, probably as a result of the same mechanism⁴⁰.

The coracoacromial ligament, which stabilizes the rotator cuff to prevent uncontrolled superior migration of the humeral head⁴⁰, may also undergo degenerative changes within its substance⁵⁶ or a traction spur may form at its insertion into the anteromedial corner of the acromion, or both phenomena may occur. The traction spur, first described by Neer³⁶ in 1972, may easily be mistaken for an abnormal acromial hook, or type-3 acromion⁶, on a radiograph of the supraspinatus outlet. This reactive spur has often been thought, erroneously we believe, to be the initiating factor in dysfunction of the rotator cuff. This etiological theory of primary extrinsic impingement led to the currently popular operative approach of subacromial decompression, which we also believe is erroneous.

We recommend that decisions involving operative treatment be based on demonstrated pathoanatomy, not theoretical concepts. Because the coracoacromial ligament and the undersurface of the acromion both function as important passive stabilizers against superior migration of the humeral head, we believe that these structures should not be sacrificed without evidence of pathological changes. Both of these structures are usually normal. Neer, in his original series of partial-thickness tears³⁶, found reactive acromial changes in only eight of nineteen patients, a prevalence similar to that which we found in our patients (twenty of seventy-nine shoulders). Neer did not mention any changes in the twenty patients who had a full-thickness tear. Although spurs may result in secondary damage to the cuff in only a few patients, acromioplasty was recomended by Neer 100 per cent of the time.

Acromioplasty fails to address the primary problem, which is intratendinous degeneration or tendinosis, in most symptomatic rotator cuffs. This was demonstrated by Gerber¹¹, who found that an injection into the subacromial space does not relieve pain caused by resisted abduction in patients who have a partial-thickness tear. It is our contention that at least some of the pain in a patient who has a partial tear is due to the damaged supraspinatus tendon. Failure of operative treatment, such as acromioplasty, likely occurs when the pathoanatomy of the tendinosis is not addressed. In addition, acromioplasty disrupts the periosteum and cortical bone of the acromion, exposing a large surface of raw cancellous bone that may predispose to extensive scar formation. Subsequent fibrosis of the coracoacromial arch may result in decreased motion of the shoulder.

Since 1988, the senior one of us (R. P. N.) has based his treatment of symptoms of impingement on demonstrated pathoanatomy^38,39. The degenerative tissue of the rotator cuff is identified and arthroscopically debrided, thus removing the primary pathoanatomy^40,51. Clearly identified excrescences underneath the acromion are also resected; however, such excrescences are the exception rather than the rule. A complete acromioplasty is not performed and the coracoacromial ligament is not resected, although small (less than 25 per cent) segments may be removed when occasional osseous subacromial excrescences are addressed.

	Review of the Literature

Top Introduction Review of the Literature Analysis of the Present... Glenohumeral Instability... Evaluation of Patients Return to Activities of... Overview References

 
We believe that 90 to 95 per cent of abnormalities of the rotator cuff are secondary to tension overload, overuse, and traumatic injury. There is no objective evidence that primary extrinsic factors are involved in most disorders of the rotator cuff³⁹, as changes within the rotator cuff often occur without accompanying changes on the acromion^19,42. Ozaki et al.⁴² examined 200 cadaveric shoulders and found that, while a lesion on the anterior one-third of the undersurface of the acromion was always associated with a tear of the cuff, the reverse was not always true. In fact, in all specimens that had a partial-thickness tear on the articular side, the undersurface of the acromion was intact. Ozaki et al., like one of us (R. P. N.)^38,40, concluded that "the pathogenesis of most tears of the rotator cuff is a degenerative process" that predates the formation of osteophytes and that the acromial changes, when present, are reactive osseous changes secondary to impingement from superior humeral migration rather than primary acromial variants. Ozaki et al. also noted that the severity of the pathological changes on the undersurface of the acromion were commensurate with the severity of the damage to the rotator cuff and that a cycle subsequently develops, with the irregularity of the undersurface of the acromion abrading the cuff and vice versa. The prevalence of incomplete tears increased with increasing age.

Gerber¹⁰ stated that if a repair of the rotator cuff and an acromioplasty are performed together and the repair fails, a type-3 acromion will recur. The change in the morphology of the acromion following an acromioplasty cannot cause the rupture, but the rupture causes new deformation of the acromion¹⁰.

Neer³⁶, in 1972, and Bigliani et al.⁶, in 1986, popularized the theory of extrinsic subacromial impingement. However, it cannot be concluded from either the clinical series described by Neer or the cadaver series described by Bigliani et al. that acromial variations are primary (due to heredity) as opposed to secondary changes. Neer frequently noted, in 100 dissected scapulae, a characteristic ridge of proliferative spurs and excrescences on the undersurface of the anterior process of the acromion that were apparently caused by repeated impingement of the rotator cuff on the humeral head, with traction of the coracoacromial ligament. He also noted that many radiographs showed corresponding areas of proliferation at the anterior edge of the acromion. One of us (R. P. N.)^38-40 observed similar findings and concluded that Neer misinterpreted the etiology of these pathoanatomical abnormalities.

Morrison¹⁰ has stated that the basis of rotator-cuff disorders is a muscle imbalance between the elevators and depressors of the humeral head—that is, between the rotator cuff and the deltoid. As a natural part of the aging process, the deltoid retains its strength longer than the relatively diminutive rotator cuff, resulting in a loss of the depressor effect of the rotator cuff on the humeral head during elevation and leading to subsequent impingement. Morrison also thought that acromial morphology is not the cause of tears of the rotator cuff, although he believed that acromial variants may predispose some people to tears of the rotator cuff once muscle imbalance occurs.

Deutsch et al.¹⁵, in a radiographic study, noted that patients who had stage-II or III impingement, with or without a full-thickness tear, had superior migration of the humeral head with elevation of the arm. The humeral head stayed centered within the glenoid cavity in normal volunteers under normal conditions. However, when the volunteers were fatigued, they also had migration of the humeral head. This may explain the belief that occupational and athletic activities may play a greater role in degeneration of the tendon than does age¹¹. Young patients frequently overload the rotator cuff during intense athletic activities, specifically swimming, racquet sports, or throwing sports, or because of occupational demands. This hypothesis may also explain the beneficial effects of rest and rehabilitative exercises in restoring strength to the rotator cuff. According to Gerber¹⁰, "the main problem is degeneration and weakness in the musculotendinous unit and the tendinous insertion on the bone."

Uhthoff¹⁰ believed that most changes on the acromion result from intrinsic tendinopathies. Tears of the cuff, if they are caused by attrition or impingement, should originate at the bursal aspect of the cuff¹⁰; however, autopsy evaluations^42,56 as well as clinical observations have revealed a higher prevalence of partial tears on the articular side than on the bursal side. Gerber¹⁰ concurred, reporting that iatrogenic impingement exclusively caused lesions on the superior surface of the cuff in rats. He believed that the confusion regarding the etiology of dysfunction of the rotator cuff stems from the fact that the pain experienced by the patient is at least partly caused by secondary impingement.

Burkhead¹¹ found that the hook of a type-3 acromion represents spurring on the medial aspect of the acromion. We agree that this hook, when present, is a traction spur of the coracoacromial ligament rather than a radiographically misinterpreted primary acromial variant.

Uhthoff et al.⁵⁶ confirmed the presence of degenerative changes in the coracoacromial ligament in patients who had impingement syndrome. We believe that this degeneration is probably the result of repetitive overloading of the ligament as it works to function as a secondary, passive superior stabilizer of the humeral head once the rotator cuff is too weak or dysfunctional to dynamically maintain the humeral head centered in the glenoid cavity.

Osteoarthrosis of the acromioclavicular joint is present in association with many symptomatic problems of the rotator cuff that necessitate operative intervention. While inferior acromioclavicular osteophytes may contribute to subacromial stenosis and subsequent impingement and may need to be removed, osteoarthrosis of the acromioclavicular joint itself appears to be independent of injury of the rotator cuff⁴⁰.

Ozaki et al.⁴² found that, in contrast to partial-thickness tears on the articular side, each partial-thickness tear on the bursal side was associated with an attritional lesion on the coracoacromial ligament as well as on the anterior one-third of the undersurface of the acromion. Those authors observed the same kind of attritional changes in each specimen that had a full-thickness tear.

One of us (R. P. N.)^40,51 also noted attritional subacromial changes in association with full-thickness tears, but only in 46 per cent (forty-four) of ninety-six tears.

Several mechanisms may be responsible for partial-thickness tears on either the articular side or the bursal side. If the articular side degenerates simultaneously with weakening of the cuff, a tear occurs on the articular side and secondary impingement may then lead to a partial-thickness tear on the bursal side, as well as to osteophytes on the underside of the acromion. If weakness occurs without a tear on the articular side, an isolated partial-thickness tear occurs on the bursal side. In either case, further damage from continued overload, overuse, secondary impingement, or traumatic injury may then lead to a full-thickness tear. We believe that a primarily intrinsic etiology is responsible for most tears on the bursal side, but a few tears may be due to extrinsic causes, such as acromial variants.

Pathophysiology
Both Uhthoff et al.⁵⁶ and Ozaki et al.⁴² performed anatomical studies on cadaveric shoulders and found that most partial-thickness tears were on the articular side. Payne et al.⁴⁴ found that 91 per cent of the partial-thickness tears in forty-three young athletes were on the articular side. Ellman¹⁷ noted that, when partial-thickness tears were openly explored, "a surprising degree of lamination can be encountered, with the most inferior [articular surface] fibers demonstrating the greatest degree of retraction." Loehr and Uhthoff³⁴ studied 306 cadaveric rotator cuffs and concluded that most degenerative tears originated on the articular side of the supraspinatus tendon, near the insertion, and that such tears are primarily intrinsic tendinopathies and are not secondary to anatomical variations or wear and tear resulting from contiguous structures. This finding may be due to the poor blood supply of the articular side of the supraspinatus insertion. Lohr and Uhthoff³⁵ confirmed the presence of the hypovascular critical zone just proximal to the insertion of the supraspinatus tendon that had first been noted by Codman¹² in 1934. Lohr and Uhthoff³⁵ reported that the articular side of the supraspinatus insertion had only a sparse vascular supply, with almost no vessels. However, the bursal side of the supraspinatus insertion was well vascularized. Those authors also noted that this critical zone of ischemia on the articular side seemed to correspond well with the common site of rupture of the tendon. The relative ischemia of this area may be one reason why the cuff is unable to repair itself, leading to degeneration and weakness.

Rothman and Parke⁴⁷ noted that areas of hypovascularity coincided with the common sites of degeneration in the supraspinatus and infraspinatus tendons. Those authors concluded that this strongly suggests that a poor blood supply may be important in the pathogenesis of degeneration of the rotator cuff. They found no effect of aging on this vasculature. We also noted that the area of degenerative tendinosis occurs most often in the hypovascular zone of the articular side of the supraspinatus insertion and the anterior infraspinatus insertion. This area corresponds well with the crescent described by Burkhart^8,10. In fact, just outside this well defined area of degeneration one can often see the cable originally described by Burkhart⁸. We have never found this cable to be degenerated, and all attempts to debride it with a motorized shaver have been futile; this further proves that the tissue is intrinsically healthy. We agree with Burkhart⁸ that loss of the hypovascular crescentic area of rotator cuff tissue within the cable may be a normal part of aging, similar to hair loss. Elevation of the arm is still possible because of stress transmission through the cable. This may explain why approximately 72 per cent of all tears of the rotator cuff are asymptomatic¹¹: they are a normal part of aging. The approximately 28 per cent of tears that cause pain may do so because they involve more than just the crescentic area of the insertion or are associated with increased tendinopathy.

This degenerative tissue, identified histologically as angiofibroblastic hyperplasia^39,40, appears to be painful in and of itself, although we do not know why this is so. Goldie²⁵ noted free nerve endings in this tissue, but we have been unable to duplicate his findings. We suggest that it is painful because it is ischemic or chemically irritated. If degeneration of the crescentic area is a normal part of aging, however, then it should be painless. This is probably true in most people, in whom the tissue degenerates at a rate at which the body can absorb it, leaving an asymptomatic non-retracted full-thickness tear of the rotator cuff with the cable intact. Symptoms may develop when the rate of tissue breakdown, or the formation of tendinosis, exceeds the body's ability to absorb the tissue, as may occur with overuse or traumatic injury even in the absence of a full-thickness tear.

Because of the high prevalence of asymptomatic shoulders with a full-thickness tear of the rotator cuff as well as the high prevalence of very painful shoulders without a full-thickness tear, we do not routinely perform preoperative magnetic resonance imaging or arthrography. Once a diagnosis of tendinosis and dysfunction of the rotator cuff is made on the basis of history, physical examination, and routine radiographs and after appropriate non-operative treatment has failed, decisions regarding operative treatment are made on the basis of the symptoms and wishes of the patient, not the presence or absence of a hole in the cuff. Once diagnostic arthroscopy is performed, it becomes clear whether the tear is full thickness.

Arthroscopic Subacromial Decompression
Most reports regarding arthroscopic treatment of abnormalities of the rotator cuff have presented only short-term results and have focused on arthroscopic subacromial decompression. While this procedure has been performed since 1985, no long-term results (after a minimum of five years of follow-up) have been published, to our knowledge. Our objections to subacromial decompression and the destruction of the normal static superior stabilizers of the humeral head concern the long-term effects of iatrogenic removal of the coracoacromial arch. Long-term follow-up is essential in the evaluation of any procedure. Many procedures that had good short-term results have now been condemned on the basis of the long-term data. The orthopaedic experience with total medial meniscectomy and various total joint arthroplasty techniques and prostheses has borne this out.

The rate of short-term (average, 25.9 months; range, 16.7 to 48.0 months) success after arthroscopic subacromial decompression has been reported to be between 46 per cent²⁹ (forty-six of 100) and 100 per cent⁵⁸ (fifty of fifty)^{1,3,16-18,21,22,24,30,33,43,44,48-50,53,57}. These so-called successes probably occur because subacromial decompression relieves the pain caused by secondary impingement, which, even though it is a secondary manifestation of primary dysfunction of the rotator cuff, is the cause of the clinical symptoms. This situation is analogous to the subtle anterior instability and resultant dysfunction of the rotator cuff in a thrower; the primary pathoanatomy (glenohumeral instability) is subtle and is easily missed if one focuses exclusively on the immediate source of pain. Harryman¹⁰ stated: "Before focusing on the coracoacromial arch and the shape of the acromion, we must understand why fiber failure occurs." Gerber¹⁰ agreed that acromioplasty is effective in relieving pain, but he did not think that impingement causes the tear of the cuff. Ryu⁴⁹ reviewed the literature and concluded that the improvement noted after subacromial decompression in patients who have a partial-thickness tear on the articular side of the rotator cuff may be a result of postoperative rest rather than an intrinsic benefit of the operation itself. Another potential source of pain relief could be the denervation of subacromial soft tissue.

We are concerned that, with the passage of time after an acromioplasty, the lack of the passive stabilizing influence of the coracoacromial arch in the presence of a marginally functional rotator cuff may compromise the ability to elevate the arm. This point is best illustrated by patients who have a full-thickness tear of the rotator cuff with marginal function that is dependent on an intact coracoacromial arch. Acromioplasty may cause a torn rotator cuff that is functional to become completely non-functional, with the patient no longer able to elevate the arm secondary to uncontrolled superior humeral migration¹¹. Since it is impossible to predict which torn rotator cuffs may become marginally functional in the distant future, prudence suggests retaining the secondary stabilizers if at all possible.

Ozaki et al.⁴², in support of this point of view, stated that "histological evidence ... suggests that the anterior one-third of the undersurface of the acromion plays a major role in the mechanism of movement of the shoulder, because it works as a subacromial joint." Paulos and Franklin⁴³ believed that excessive resection of acromial bone or removal of the coracoacromial ligament may not be advisable in patients who have a massive irreparable full-thickness tear of the rotator cuff because these structures provide restraint to anterior and anterosuperior migration of the humeral head. Harryman¹¹ believed that the function of the shoulder may become fully dependent on the deltoid if the rotator cuff fails and remains irreparable. In such instances, the coracoacromial arch functions as a secondary constraint to provide humeroscapular stability. When a major portion of the acromion is resected, resection of the coracoacromial ligament should be avoided or severe anterosuperior instability may occur; if such instability does occur, no reconstruction will be satisfactory¹¹. Harryman also thought that the shoulder may eventually become dependent on the arch for functional stability. Williams¹¹ agreed, stating that patients who have a marginally functional torn cuff need the coracoacromial arch. If the arch is removed, the cuff may become completely non-functional. In fact, Williams stated that he was "appalled to find that some patients have a shoulder that is no longer functional following what was considered to be reasonable surgery." He also noted that a marginally functional torn rotator cuff can be difficult to identify preoperatively.

We believe that acromioplasty is infrequently indicated for disorders of the rotator cuff. Primary aberrant acromial morphology is probably the cause of less than 5 to 10 per cent of impingement syndromes. In addition, Burkhart⁹ described congenital subacromial stenosis, which also may warrant a traditional acromioplasty. However, both of these conditions are uncommon.

Adolfsson and Lysholm¹ reported an overall 67 per cent rate of satisfactory results, according to a modification of the system described by Neer, following seventy-nine arthroscopic subacromial decompressions performed for various stages of impingement, including full-thickness tears. Of the twenty-nine patients who had a partial-thickness tear, 90 per cent had a satisfactory result; however, of the eleven patients who had clinical signs of impingement alone without tendinosis or a tear, only 19 per cent had a satisfactory result. The average duration of follow-up was seventeen months (range, nine to twenty-four months).

Altchek et al.³ reported a 73 per cent rate of good and excellent results at an average of seventeen months (range, one to three years) after forty arthroscopic subacromial decompressions performed for various stages of impingement. The twenty-four patients who did not have a partial or full-thickness tear had the highest rate of success (83 per cent). Seventy-six per cent of the thirty-three patients who had participated in sports activity had returned to sports activity at the time of the most recent follow-up.

In a series described by Ellman¹⁶, 88 per cent of forty-nine arthroscopic subacromial decompressions performed for various stages of impingement yielded a good or excellent result. The rate of success for the thirty-nine patients who did not have a full-thickness tear was 90 per cent. The average duration of follow-up was seventeen months (range, twelve to thirty-six months). In a later study¹⁷, Ellman reported a satisfactory short-term result in fifteen (75 per cent) of twenty patients who had been managed with arthroscopic subacromial decompression for a partial-thickness tear.

Esch et al.¹⁸ reported a 78 per cent rate of good and excellent results, according to the rating scale of the University of California at Los Angeles, after seventy-one arthroscopic subacromial decompressions performed for impingement syndrome and partial-thickness tears. The average duration of follow-up was nineteen months (range, twelve to thirty-six months).

Gartsman²¹ noted improvement, at an average of twenty-eight months, in eighty-one (88 per cent) of ninety-two patients who had had arthroscopic subacromial decompression for stage-II impingement. In another report, Gartsman²² reported that seventy-eight (88 per cent) of eighty-nine shoulders that had been treated with arthroscopic subacromial decompression for stage-II impingement without a partial-thickness tear had a satisfactory result at an average of thirty-one months (minimum, two years). In addition, thirty-three (83 per cent) of forty shoulders had a satisfactory result at an average of twenty-nine months after arthroscopic subacromial decompression combined with débridement of the frayed edges of a partial-thickness tear. Gartsman and Milne²⁴ later reported an 88 per cent rate of subjective satisfaction after arthroscopic subacromial decompression, combined with débridement of the rotator cuff when appropriate, in 111 shoulders with a partial-thickness tear on the articular side of the rotator cuff. The average duration of follow-up was thirty-two months (range, twenty-six to eighty-four months). Objective scoring systems were not used to rate the outcome of the operation.

In the study by Hawkins et al.³⁰, there was a satisfactory result (indicating no or minimum pain, normal use, within 20 degrees of a full range of motion, normal strength, and a negative impingement sign) after 46 per cent (fifty-one) of 110 arthroscopic subacromial decompressions performed for stage-II impingement. When patients who were involved in a Workers' Compensation claim were excluded, the rate was 56 per cent (thirty-seven of sixty-six shoulders). The minimum duration of follow-up was two years (maximum, fifty-two months).

Van Holsbeeck et al.⁵⁷, in a study of fifty-three arthroscopic subacromial decompressions performed for advanced stage-II or early stage-III disease (full-thickness tears of less than one centimeter), reported an 83 per cent rate of good and excellent results, according to the rating scale of the University of California at Los Angeles. The average duration of follow-up was twenty months (range, twelve to twenty-six months).

Lazarus et al.³³ reported a good or excellent result (on the basis of the University of Pennsylvania score) after 57 per cent of forty-six arthroscopic subacromial decompressions performed for impingement syndrome without a full-thickness tear. The average duration of follow-up was twenty-five months (minimum, twelve months). Those authors stated that patients who had a more hooked acromion postoperatively did better. Patients who had made a Workers' Compensation claim had the worst results. The results following twenty-four open subacromial decompressions were comparable.

In a study of forty-two arthroscopic subacromial decompressions and twenty-four so-called simple decompressions, which involved resection of the bursa and the coracoacromial ligament without acromioplasty, Paulos and Franklin⁴³ noted improvement after 85 per cent of the arthroscopic subacromial decompressions, no change after 10 per cent, and worsening after 5 per cent. They also noted improvement after 83 per cent of the simple decompressions and no change after 17 per cent. The average duration of follow-up was thirty-two months (range, twelve to fifty-four months). Patients who did not have an acromioplasty did slightly better, especially with regard to pain at night and persistence of the impingement sign. These patients also tended to be younger, which could mean that they had less extensive disease but probably also meant that they had higher functional demands.

Payne et al.⁴⁴ reported a satisfactory result, as determined with the rating system of the American Shoulder and Elbow Surgeons, for 72 per cent of forty-three athletes who had a partial-thickness tear treated with arthroscopic subacromial decompression or débridement; 51 per cent of the athletes returned to sports. However, the results of the two procedures were not compared. The duration of follow-up averaged forty-eight months (range, twenty-four to 120 months). Patients who had subtle instability had the worst results.

Roye et al.⁴⁸ reported a good or excellent result, according to the rating scale of the University of California at Los Angeles, after 94 per cent of ninety arthroscopic subacromial decompressions performed for various stages of impingement, including full-thickness tears of less than one centimeter. Sixty-one per cent (thirty-four) of the fifty-six athletes overall and 53 per cent (eighteen) of the thirty-four who participated in a sport that required throwing returned to sports activity. Only 79 per cent of the thirty-four athletes whose sport involved throwing and eight of the twelve competitive baseball or softball pitchers had a satisfactory result. The average duration of follow-up was forty-one months (range, twenty-four to eighty-two months).

At an average of twenty-three months (range, twelve to fifty months) after fifty-three arthroscopic subacromial decompressions performed for various stages of impingement, Ryu⁴⁹ reported an 81 per cent rate of good and excellent results, according to the rating scale of the University of California at Los Angeles. The result was good or excellent after 86 per cent of the thirty-five procedures that were performed for a partial-thickness tear.

Sampson et al.⁵⁰ reported a good or excellent result, according to the rating scale of the University of California at Los Angeles, after 90 per cent of ninety-one arthroscopic subacromial decompressions. The minimum duration of follow-up was one year; the average duration was not reported.

In a study of twenty-five arthroscopic subacromial decompressions performed for impingement without a full-thickness tear, Speer et al.⁵³ reported an 88 per cent rate of good and excellent results, on the basis of the rating scale of the University of California at Los Angeles. The patients were able to return to sports after 76 per cent of the procedures. The average duration of follow-up was twenty months (range, fourteen to thirty-two months). Only four of the twelve patients who were involved in a sport that required throwing were able to return to the sport without discomfort, even though examination under anesthesia was negative for glenohumeral instability.

Warner et al.⁵⁸ reported a successful result after 94 per cent of seventy arthroscopic subacromial decompressions at an average of twenty-four months (range, twenty-two to twenty-eight months) postoperatively. The modified technique that was used by those authors led to complete resolution of the symptoms at an average of twenty-two months (range, eighteen to twenty-four months) postoperatively in all fifty patients who had impingement syndrome. No objective measures were used to evaluate the success of the procedure.

Bigliani et al.⁷ reported an 81 per cent rate of good and excellent results following twenty-six arthroscopic subacromial decompressions performed on patients who were less than forty years old. The average duration of follow-up was thirty-three months (range, twelve to eighty months). Seven of the ten patients who had been recreational athletes were able to return to their recreational activities.

Open Subacromial Decompression
Hawkins et al.²⁹ reported a satisfactory result (no or minimum pain, normal use, within 20 degrees of a full range of motion, normal strength, and a negative impingement sign) for 87 per cent of 108 patients who had been managed with an open subacromial decompression because of impingement without a full-thickness tear. When the thirty-five patients who had made a Workers' Compensation claim were excluded, the rate of satisfactory results was 92 per cent. The long-term follow-up evaluation was performed at an average of 5.2 years (range, 2.6 to 8.1 years).

Van Holsbeeck et al.⁵⁷ reported a good or excellent result, on the basis of the rating scale of the University of California at Los Angeles, after 81 per cent of fifty-three open subacromial decompressions performed for stage-II and early stage-III disease. The average duration of follow-up was twenty-seven months (range, twelve to forty-seven months).

In a study by Neer³⁶, the result was satisfactory for fifteen of sixteen shoulders at an average of two and a half years after an open subacromial decompression.

Neviaser et al.³⁷ reported that all patients had a subjective decrease in pain, compared with the preoperative status, following a four-in-one arthroplasty (subacromial decompression, excision of the lateral end of the clavicle and the coracoacromial ligament, and biceps tenodesis). The average duration of follow-up was four years (range, two to eight years). Many authors^11,40 currently condemn this so-called shotgun operative approach, and it should be noted that Neviaser et al. did not document any objective criteria for success.

In a study of seventy-two open subacromial decompressions performed for impingement syndrome without a full-thickness tear, Post and Cohen⁴⁶ reported an 80 per cent rate of good and excellent results at an average of twenty-three months (range, five to forty-eight months) postoperatively.

Stuart et al.⁵⁴ reported a successful result after 73 per cent of forty open subacromial decompressions without resection of the lateral end of the clavicle. The average duration of follow-up was eight years (range, three to thirteen years).

In the series described by Thorling et al.⁵⁵, the result was good or excellent for twenty-six (65 per cent) of forty patients at an average of twenty-one months (range, six to forty-two months) after an open subacromial decompression performed for impingement without a full-thickness tear.

Open Division of the Coracoacromial Ligament
The short-term results of open division of the coracoacromial ligament without subacromial decompression are at least as good as, if not better than, those of open or arthroscopic subacromial decompression.

Nineteen (90 per cent) of twenty-one patients in a study by Ha'eri and Wiley²⁷ reported that they were satisfied after an open resection of the coracoacromial ligament, which was combined with open subacromial decompression (four patients) and open resection of the lateral end of the clavicle (six patients) when appropriate. The minimum duration of follow-up was one year, but no average was noted.

Johansson and Barrington³² reported a 95 per cent rate of satisfactory results at an average of thirty-six months (range, eight to seventy-six months) after forty-one open resections of the coracoacromial ligament in the absence of substantial acromial osteophytes.

In the study by Penny and Welsh⁴⁵, twelve of fourteen patients who had signs of impingement without a full-thickness tear of the rotator cuff returned to sports activity. There was no difference in the results between ten patients who had a simple excision of a one-to-two-centimeter segment of the coracoacromial ligament and ten who had a complete open subacromial decompression. The average duration of follow-up was three and a half years (minimum, six months).

Débridement of the Rotator Cuff
The reported results of open or arthroscopic débridement of the rotator cuff have been good; however, in many series, the procedure was combined with a subacromial decompression. Fu et al.¹⁹ stated that "the emphasis of treatment is shifting from that of decompression to restoring the health of the rotator cuff."

Altchek and Carson² reported an 80 per cent rate of favorable results after arthroscopic débridement of a symptomatic rotator cuff in fifty athletes who engaged in a sport that required throwing. Five of the fifty also had an arthroscopic subacromial decompression for a narrowed subacromial space that had been noted either on radiographs or intraoperatively. Those authors believed that the 20 per cent rate of failure was due to excessive capsular laxity that could not be controlled by strengthening of the shoulder.

Andrews et al.⁴ reported a good or excellent result for 85 per cent of thirty-four young athletes (average age, twenty-two years) who had an arthroscopic débridement of the rotator cuff; 64 per cent of the patients were pitchers. The average duration of follow-up was thirteen months.

In a study of thirty-nine open subacromial decompressions combined with open débridement of the rotator cuff performed for partial-thickness tears, Fukuda et al.²⁰ found a 92 per cent rate of satisfactory results; no objective criteria were given. The average duration of follow-up was thirty-four months; no range was reported.

Snyder et al.⁵² reported a good or excellent result, on the basis of the rating scale of the University of California at Los Angeles, for 93 per cent of thirty-one shoulders in which a partial-thickness tear was treated with débridement; eighteen of the shoulders were also treated with an arthroscopic subacromial decompression. The average duration of follow-up was twenty-three months (range, ten to forty-three months).

Ogilvie-Harris and Wiley⁴¹ reported a successful result for approximately two-thirds of twenty patients who had an arthroscopic débridement of the rotator cuff for tendinitis and for approximately 50 per cent of sixty-five patients who had that procedure for a partial-thickness tear. The minimum duration of follow-up was one year.

	Analysis of the Present Study

Top Introduction Review of the Literature Analysis of the Present... Glenohumeral Instability... Evaluation of Patients Return to Activities of... Overview References

 

Materials and Methods
Ninety-eight shoulders (ninety-five patients) were treated with arthroscopic débridement because of tendinosis of the rotator cuff in the absence of a full-thickness tear. Seventy-nine of these shoulders (seventy-six patients) were available for follow-up. Forty-two shoulders were examined by one of us (J. E. B.), and thirty-seven were evaluated with use of a telephone interview. Fifty shoulders were in male patients, and twenty-nine were in female patients. Fifty-three of the shoulders were dominant, and twenty-six were non-dominant. Sixty-five of the shoulders were in patients who were involved in athletic activity on at least a recreational level. The average age of the patients was forty-five years (range, seventeen to seventy-seven years). Fifty-two shoulders were in patients who were forty years old or more, and twenty-seven shoulders were in patients who were less than forty years old.

The overall duration of follow-up was fifty-three months (range, twenty-five to ninety-three months). Two subgroups were examined: thirty-two shoulders (thirty-one patients) that were followed for five years or more (average, seventy-five months; range, sixty to ninety-three months) and forty-seven shoulders (forty-five patients) that were followed for two to less than five years (average, thirty-nine months; range, twenty-five to fifty-six months).

The main preoperative symptom was pain in all patients. In addition, forty-five patients had weakness and thirty-eight had stiffness or loss of functional motion. The average duration of the preoperative symptoms was thirty-four months (range, two months to twenty-three years).

The preoperative diagnosis of impingement syndrome or abnormality of the rotator cuff was made on the basis of the history and the physical examination; radiographs were made for all patients. All of the patients had failed to respond to non-operative therapy that included an intensive program of rehabilitative exercises designed to restore strength, flexibility, and endurance to the rotator cuff, deltoid, scapulothoracic stabilizers, and thoracic muscles. All of the rehabilitation programs were supervised by a physical therapist.

A substantial number of other patients had arthroscopic procedures during this time-period, but they were excluded from the present study because they had a full-thickness tear of the rotator cuff, glenohumeral instability (noted on examination under anesthesia or on glenohumeral arthroscopy, or both), glenohumeral chondromalacia that was grade III (fissuring to bone) or grade IV (subchondral bone exposed), osteoarthrosis of the acromioclavicular joint that necessitated full resection of the lateral end of the clavicle, arthroscopically confirmed adhesive capsulitis, previous repair of the rotator cuff, previous subacromial decompression, rheumatic disease, calcific tendinitis, or osteochondromatosis. In addition, one patient had concomitant chronic neck pain with radiculopathy and could not distinguish that pain from the pain in the shoulder.

Preoperative and postoperative pain was assessed according to whether it occurred at night (that is, whether it was present while the patient was trying to fall asleep or it awoke the patient), at rest, with activities of daily living, or during athletic activities.

Operative Technique
The procedure was performed with the patient under general anesthesia, and antibiotic prophylaxis was provided. If needed, the shoulder was manipulated to regain a full passive range of motion. The patient was placed in the lateral decubitus position with ten to fifteen pounds (4.5 to 6.8 kilograms) of traction placed on the arm. Marcaine (bupivacaine hydrochloride) with epinephrine was injected into the subacromial space to aid in hemostasis of the coracoacromial arch. An arthroscopic pump (Davol, Cranston, Rhode Island) was utilized for the operative interventions in both the glenohumeral joint and the coracoacromial arch. A standard posterior arthroscopic portal was made. The rotator interval between the biceps tendon and the leading edge of the subscapularis was identified, and an anterior portal was established in this space through an inside-out technique with use of a Wissinger rod. Diagnostic arthroscopy was then performed in the glenohumeral joint.

Areas of labral degeneration were debrided with a motorized shaver. SLAP lesions (lesions of the superior portion of the labrum, anterior and posterior, or a bucket-handle tear of the labrum) were repaired or debrided as appropriate. The superior aspect of the glenoid as well as the underside of the labrum were burred to enhance biological healing. The biceps tendon was debrided if it was degenerated. The supraspinatus and infraspinatus tendons were evaluated for any degenerative tendinotic changes, which were usually identified by a change in appearance. (A normal rotator-cuff tendon appears firm, taut, shiny, and yellow-white, whereas tendinosis gives the tendon a softened texture and it appears gray, dull, sometimes edematous, and frayed.)

The degenerative portions of the affected tendons are arthroscopically debrided with straight or curved motorized shavers moved in a sweeping motion across the involved area. This is the arthroscopic equivalent of the Nirschl scratch test developed for open débridement of degenerative tendinosis about the elbow. Healthy tendon tissue is not substantially affected by a motorized shaver that is operated without undue force. However, analogous to peeling paint, friable, degenerative intratendinous tissue is readily removed, without harming healthy tissue. The abnormal tissue was most frequently located in the critical zone of the supraspinatus tendon, especially in its insertions, and in the more anterior aspect of the infraspinatus. If the abnormality extended inferiorly into the infraspinatus, the portals were reversed to debride the abnormal areas through the posterior portal.

We believe that many partial-thickness tears of the rotator cuff are missed by both experienced and inexperienced shoulder surgeons because the synovial tissue covering the articular side of the rotator cuff gives it a deceptively normal appearance. The underside of the rotator cuff is therefore routinely tested with a motorized shaver. If the diagnosis is correct, a cuff that had a normal appearance will be found to be abnormal and to have a partial-thickness lesion. On the basis of our findings, we believe that most patients who are diagnosed as having an impingement syndrome have, instead, a partial-thickness intrasubstance tear of the rotator cuff or tendinosis.

The same anterior and posterior skin portals were used to enter the subacromial space, which was first explored with use of a blunt trocar to disrupt any adhesions and to feel the undersurface of the acromion and the lateral aspect of the clavicle for any inferior osteophytes.

A bursectomy was performed as needed to visualize the rotator cuff, especially at its insertion onto the greater tuberosity. The bursal side of the rotator cuff was tested with a motorized shaver, and tendinotic changes were identified and debrided in the same manner as the intrasubstance changes on the articular side. Any thickened bursa or subacromial or subclavicular osteophytes contributing to stenosis of the coracoacromial arch were arthroscopically debrided with motorized instruments. No formal acromioplasties were performed, and the coracoacromial ligament was preserved in all shoulders.

Early in the study, one patient had an open bursectomy and débridement of osteophytes because technical difficulties were encountered during subacromial arthroscopy.

The two portals were routinely closed. Postoperative rehabilitation was begun as soon as tolerated, usually within two days. The patients progressed with active and active-assisted exercise as tolerated. All of the patients followed a postoperative program of active therapeutic exercise that was similar to the preoperative program.

Assessment of the Patients
The subjective result was rated satisfactory when the patient thought that the result was good or excellent and that he or she would have the same procedure performed again under similar circumstances. Otherwise, the rating was unsatisfactory.

Three objective scoring systems were used: the rating scale of the University of California at Los Angeles¹⁶ (Table I), the functional rating score of Constant and Murley¹³ (Table II), and a new rotator-cuff score designed by two of us (R. P. N. and J. E. B.) (Table III). Only the rating scale of the University of California at Los Angeles, which has a maximum possible score of 35 points, was used to determine the objective result. A score of 34 or 35 points was considered excellent; a score of 28 to 33 points, good; a score of 21 to 27 points, fair; and a score of 0 to 20 points, poor. A good or excellent result was considered objectively satisfactory, and a fair or poor result was considered a failure. Any patient who had a second procedure on the shoulder was considered to have had an objective failure of the index operation, regardless of the score.

Statistical Analysis
Statistical analysis was performed with the chi-square test of independence with the Yates correction. We analyzed the objective result with respect to the age of the patient, side of involvement, tendons involved, performance of exostectomy, duration of follow-up, and presence of a Workers' Compensation claim.

Results
Sixty-one (77 per cent) of the seventy-nine shoulders had a labral abnormality. Eleven shoulders had a SLAP-type lesion (a lesion of the superior portion of the labrum, anterior and posterior, or a bucket-handle tear of the labrum that necessitated excision of the loose fragment). Also, most of the shoulders had labral degeneration in addition to a partial-thickness tear, often in more than one location: fifty-four had degeneration of the anterosuperior portion of the labrum; twenty-five, of the posterosuperior portion; five, of the superior portion; two, of the mid-posterior portion; and two, of the anteroinferior portion.

Grade-I (softening of the cartilage) or grade-II (fibrillation) chondromalacia (excluding the normal spot of chondromalacia often found in the center of the glenoid) was noted in nineteen glenoids and seven humeral heads. Subacromial bursitis or thickening was noted in fifty-four shoulders (68 per cent). Seven shoulders (9 per cent) had degeneration of the coracoacromial ligament, and eleven (14 per cent) had thickening of the ligament.

The partial-thickness tear of the rotator cuff was on the articular surface in fifty-one shoulders, on the bursal surface in one, and on both surfaces in twenty-seven. Thirty-eight of the fifty-one tears of the articular surface were of the so-called critical area of the supraspinatus insertion, ten were of the supraspinatus and infraspinatus, two were of the infraspinatus alone, and one was of the supraspinatus and the subscapularis. The tear on the bursal surface was of the supraspinatus insertion. Sixteen of the twenty-seven tears that involved both surfaces were of the supraspinatus insertion alone, and eleven were of the supraspinatus and the infraspinatus. No significant association was detected, with the numbers available, between either the side of the partial-thickness tear (p = 0.64) or the tendon or tendons involved (p = 0.75) and the success or failure of the procedure.

In forty-six (58 per cent) of the seventy-nine shoulders, no osseous procedure was performed in the subacromial space. Acromial osteophytes were debrided in twenty shoulders (25 per cent), both acromial and clavicular osteophytes were debrided in twelve (15 per cent), and clavicular osteophytes were debrided in one. No significant association was detected between the performance of an exostectomy or its location and the success or failure of the procedure (p = 0.40).

All sixty-five patients who were involved in athletic activities had pain with these activities preoperatively, and twenty of these patients (31 per cent) continued to have such pain postoperatively. Preoperatively, sixty-five shoulders (82 per cent) overall had pain with activities of daily living, forty-five (57 per cent) had pain at rest, and fifty-two (66 per cent) had pain at night. Postoperatively, seven shoulders (9 per cent) had pain with activities of daily living, three (4 per cent) had pain at rest, and fifteen (19 per cent) had pain at night. The severity of the pain was almost always markedly diminished postoperatively.

The average time until the patients returned to usual activities of daily living was twenty-five days (range, one day to five months), the average time until they returned to full activities including sports was four months (range, two weeks to one year), and the average time to maximum improvement was eight months (range, two weeks to three years). Of the sixty-five patients who were involved in athletic activities preoperatively, fifty-two (80 per cent) were able to return to the same level of competition postoperatively. Most of these patients were recreational athletes, although some were engaged in competitive tennis.

Two shoulders had a decreased range of active and passive motion postoperatively. One was manipulated with the patient under anesthesia at seven months postoperatively, and the other was manipulated with the patient under anesthesia and was treated with an arthroscopic débridement of adhesive capsulitis at seven months postoperatively. At the most recent follow-up examination, both shoulders had excellent motion without sequelae. The result, according to the rating scale of the University of California at Los Angeles¹⁶, was excellent for one shoulder and good for the other.

At the most recent follow-up evaluation, the result for sixty-nine (87 per cent) of the seventy-nine shoulders was considered satisfactory by the patient. With use of the rating scale of the University of California at Los Angeles¹⁶, there were forty-three (54 per cent) excellent, twenty-five (32 per cent) good, six (8 per cent) fair, and five (6 per cent) poor results. According to the functional score of Constant and Murley¹³, twenty-seven shoulders (34 per cent) had the maximum score of 100 points, thirty-one (39 per cent) had a score of 90 to 99 points, ten (13 per cent) had a score of 80 to 89 points, six (8 per cent) had a score of 70 to 79 points, three (4 per cent) had a score of 69 points or less, and two (3 per cent) did not have sufficient data to be rated with this system.

Forty-two shoulders were evaluated with the new rotator-cuff score developed by two of us (maximum score, 100 points), and the score was 90 to 100 points (excellent) for thirty-one, 80 to 89 points (good) for three, 70 to 79 points (fair) for four, and 69 points or less (poor) for four.

The success or failure of the arthroscopic débridement of the rotator cuff was found to be independent of the age of the patient (p = 0.85) and the duration of follow-up (p = 0.49), with the numbers available.

Eight patients had made a Workers' Compensation claim. According to the rating scale of the University of California at Los Angeles¹⁶, the result of the procedure was excellent for three of them, good for two, fair for one, and poor for two. With these eight patients excluded from the analysis, sixty-four (90 per cent) of the remaining seventy-one shoulders had a result that was considered satisfactory by the patient and sixty-three shoulders (89 per cent) had a good or excellent result according to the rating scale of the University of California at Los Angeles.

One patient in the long-term follow-up group (followed for at least five years) had a second operation. The result after the second operation was good, according to the rating scale of the University of California at Los Angeles¹⁶. The first operation was considered an objective failure, even though the patient rated the result of both operations as excellent. Another patient had, after the index operation, an open débridement of the rotator cuff with an exostectomy of the greater tuberosity, and the result was considered excellent.

A full-thickness tear developed in two patients (both of whom had made a Workers' Compensation claim), and the procedure was considered a failure. The tear was repaired in one patient and debrided (because the tear was not retracted) in the other. Both of the subsequent operations were performed forty-two months after the initial procedure, thereby emphasizing the need for long-term follow-up. Of the other patients in whom the procedure was considered a failure, one had avascular necrosis of the humeral head without collapse, and arthroscopic débridement of the rotator cuff was performed because it was believed that the symptoms originated from the rotator cuff. This may have been a failure in diagnosis. One patient who had bilateral involvement may have had rheumatic disease. In addition to bilateral pain related to the rotator cuff, she had bilateral tennis elbow, bilateral generalized pain involving multiple joints in the hand, and an elevated erythrocyte sedimentation rate. Tests for rheumatoid factor and antinuclear antibodies were negative. The patient refused to see a rheumatologist. As no definitive diagnosis could be made, the result for both shoulders was considered a failure.

Discussion
The short-term (two to less than five-year) results of arthroscopic débridement of the rotator cuff in our series were as good as those reported after open or arthroscopic subacromial decompression; the overall rate of good and excellent results was 89 per cent, and the rate when the patients who had made a Workers' Compensation claim were excluded was 93 per cent. It should be noted that our short-term follow-up is still longer than the follow-up in most reported studies of arthroscopic subacromial decompression, except for those by Payne et al.⁴⁴ and Roye et al.⁴⁸, and is longer than the follow-up in most reported studies of open subacromial decompression, except for those by Hawkins et al.²⁹, Neviaser et al.³⁷, and Stuart et al.⁵⁴.

Our long-term follow-up (minimum, five years) is the longest of which we are aware in a study of the arthroscopic treatment of abnormalities of the rotator cuff. The long-term results of arthroscopic débridement of the rotator cuff in the present study (an 81 per cent rate of good and excellent results overall and an 83 per cent rate when the patients who had made a Workers' Compensation claim were excluded) are similar to the short-term results reported for open subacromial decompression; no comparable follow-up data on open débridement are available.

It is possible that the short-term results in our study were better than the long-term results for two reasons: first, our operative technique became more refined as we gained experience with the procedure and, second, the results may deteriorate with time.

Arthroscopic débridement of the rotator cuff is probably not as technically demanding as arthroscopic subacromial decompression. It is, however, more intellectually demanding in that the surgeon must be able to identify the abnormal areas of the rotator cuff accurately. Mere visualization of the undersurface of the cuff is not enough; the surgeon must routinely test the suspicious areas with a motorized shaver. We do this in all arthroscopic procedures on the shoulder. A motorized shaver with a five-millimeter meniscal blade will not damage healthy tissue on either the articular or the bursal side; however, tissue that appears superficially normal or areas where the abnormal tissue is covered by synovial tissue will be débrided away. Once healthy tissue is exposed, the shaver is no longer effective and the débridement is complete. The shaver will not create a full-thickness tear through healthy, functional tissue.

It is our belief that impingement syndrome in the absence of a full-thickness tear of the rotator cuff is most often secondary to a partial-thickness tear. That tear usually becomes apparent after the initial débridement but may be missed if the surgeon does not perform a meticulous dissection. We have found a sweeping motion of the shaver to be most effective for débridement of the rotator cuff. An angled shaver makes it easier to reach the more posterior infraspinatus insertion in the humeral head; internal rotation also makes this area accessible. If the more posterior structures need to be debrided, it may be necessary to reverse the portals (by placing the arthroscope anteriorly and the instrumentation posteriorly). We performed all arthroscopic débridements of the rotator cuff with the patient in the lateral decubitus position with ten to fifteen pounds (4.5 to 6.8 kilograms) of traction on the arm. Of technical note, we have found it extremely difficult to fully inspect and debride the posterior insertion of the cuff with the patient in the beach-chair position.

	Glenohumeral Instability Associated with Abnormality of the Rotator Cuff

Top Introduction Review of the Literature Analysis of the Present... Glenohumeral Instability... Evaluation of Patients Return to Activities of... Overview References

 
Glenohumeral instability, especially anterior instability, subjects the rotator cuff to increased eccentric tension loads, predisposing to overuse, injury, and dysfunction³⁸. This overuse leads to weakness, which allows secondary impingement and subsequent development of symptoms. The difficulty with diagnosing instability when it is associated with disease of the rotator cuff is that the more dramatic signs and symptoms related to the cuff often obscure the subtle signs of instability⁵. Therefore, a high index of suspicion of glenohumeral instability must be maintained during examinations of patients who have symptoms related to the rotator cuff, especially young patients who are involved in activities such as swimming, throwing, racquet sports, and other athletic activities that require overhead use of the upper limb. The underlying instability may be subtle, and often careful physical and diagnostic arthroscopic examinations are required to make the diagnosis. These patients often need a glenohumeral stabilization procedure in order for treatment to be effective. Our results with arthroscopic débridement of the rotator cuff without glenohumeral stabilization for patients who have had symptoms related to the rotator cuff and underlying glenohumeral instability have been uniformly poor.

Many authors have commented on the high prevalence of posterosuperior labral fraying in combination with lesions on the articular side of the rotator cuff. Altchek and Carson² noted a 40 per cent prevalence in fifty throwing athletes who had pain in the anterior aspect of the shoulder that was refractory to non-operative treatment. These lesions are often attributed to internal impingement. As described by Davidson et al.¹⁴, internal impingement occurs with the arm cocked in 90 degrees of abduction and full external rotation. In this position, the posterior portion of the supraspinatus can be pinched between the humeral head and the posterosuperior part of the glenoid rim¹⁴. It has been noted that the increased anterior humeral translation associated with anterior glenohumeral instability intensified internal impingement with the arm in a position of abduction and external rotation^14,31. However, we have found superior labral fraying during most operations for partial-thickness tears on the articular side of the rotator cuff, regardless of whether or not the patient engaged in a sport that involved throwing. Most patients who have this combination of lesions do not participate in athletic activities that require overhead movement and do not often assume the position of abduction and external rotation; therefore, they are unlikely to be subjected to substantial internal impingement forces. A more plausible hypothesis is that, in most patients, the superior labral degeneration is due to the superior shear forces imparted by the relatively unopposed deltoid on the humeral head, not by impingement or impaction forces. Because of this force-couple imbalance, the humeral head is repetitively sheared across the superior portion of the labrum during active elevation of the arm. The superior labral degeneration caused by the subtle superior instability associated with a lesion of the rotator cuff is analogous to the anteroinferior fraying of the labrum seen in patients who have chronic anteroinferior glenohumeral instability.

Halbrecht²⁸ found that contact between the posterosuperior aspect of the glenoid and the rotator cuff may be physiological with the arm in maximum abduction and external rotation. Anterior glenohumeral instability would then appear to decrease internal impingement. In addition, most shoulders with both a tear on the articular side of the rotator cuff and posterior labral fraying are not lax or unstable. Davidson et al.¹⁴ as well as Halbrecht reported that patients who had this combined abnormality without glenohumeral instability were managed successfully with arthroscopic débridement of the rotator cuff combined with débridement of the areas of labral degeneration and postoperative rehabilitation.

Interestingly, it has been our experience that, in the presence of dysfunction of the rotator cuff, the anterosuperior portion of the labrum degenerates more often than the posterosuperior portion does. The reason may be that the resultant instability caused by the dysfunction tends to be anterosuperior in direction^11,43. In the present series, sixty-one (77 per cent) of seventy-nine shoulders had an abnormality of the labrum; fifty-four (68 per cent) had anterosuperior labral degeneration, whereas only twenty-five (32 per cent) had posterosuperior labral degeneration. We refer to these lesions as supraspinatus labral instability patterns, or SLIP lesions²⁶. Some shoulders had no labral lesions.

Before the present study, we observed SLIP lesions in 90 per cent of patients who had a partial-thickness tear of the rotator cuff; 66 per cent of the lesions occurred anterosuperiorly, and only 11 per cent occurred posterosuperiorly²⁶.

	Evaluation of Patients

Top Introduction Review of the Literature Analysis of the Present... Glenohumeral Instability... Evaluation of Patients Return to Activities of... Overview References

 
It is unfortunate that our geographical location is characterized by a transient population, as this made it difficult for us to personally examine some of our patients for the follow-up data. However, while a personal examination is obviously more desirable, we do not believe that our inability to perform one for some of our patients affected our results. Patients convey their sense of satisfaction, or lack thereof, as well over the telephone as they do in person.

The rating scale of the University of California at Los Angeles¹⁶, which was used as a more objective indicator of success in our study as well as in several others^{16,18,48,49,52,53,57}, contains five items (Table I). The information on patient satisfaction is purely subjective, and the data on pain and function are gathered from the patient's history. Active flexion is rated on a scale of 5 points, but the maximum score is given as long as the patient can actively flex more than 150 degrees. (This is also the case with the functional score of Constant and Murley¹³.) In other words, a lack of less than 30 degrees of flexion is allowed before the score is affected. We believe that a lack of 30 degrees produces a substantial functional deficit. Strength of flexion is also rated on a scale of 5 points. When the grade was determined on the basis of a telephone interview, any suspected weakness was given a grade of 4 points and major weakness was given a grade of 3 points. No patient who was interviewed or examined was unable to raise the arm against gravity (which would indicate a grade of 0, 1, or 2 points).

The cutoff for a successful result, according to the rating scale of the University of California at Los Angeles¹⁶, is 28 points; a score of 27 points or less is considered a failure. No patient who had a score of 28 points had a score of 5 points for strength; thus, it was not possible that, because of a subtle undiagnosed weakness, a patient who was thought to have a successful result on the basis of a telephone interview actually had a failure. All patients who had a borderline successful score denied that they had extreme or substantial weakness. It is unlikely that a patient would not notice or admit to a strength deficit that would warrant a score of 3 points (indicating the ability to raise the arm against gravity but not against resistance). Therefore, we are confident that the nature of the result (satisfactory or unsatisfactory) was determined accurately for the patients who were interviewed by telephone with use of the scale of the University of California at Los Angeles.

Despite our confidence in the rating scale of the University of California at Los Angeles¹⁶ for the present study, we have noted certain deficiencies. As such, we developed an assessment scale that is specific to the rotator cuff (Table III). We plan to use this scale for future outcome studies of patients who have rotator cuff disease.

	Return to Activities of Daily Living

Top Introduction Review of the Literature Analysis of the Present... Glenohumeral Instability... Evaluation of Patients Return to Activities of... Overview References

 
The average time until patients have returned to activity after arthroscopic subacromial decompression has varied. Roye et al.⁴⁸ reported a return to activities of daily living within two weeks, whereas Speer et al.⁵³ noted that the average was three months (range, three weeks to seven months). Altchek et al.³ reported an average of four months un