A 26-Year-Old Woman With Shoulder Pain

DR PARKER: Mrs B is a 26-year-old teacher whose activities are severely limited by chronic left shoulder pain. She has been treated with nonsteroidal anti-inflammatory drugs (NSAIDs), physical therapy, and cortisone injections without adequate improvement. Mrs B lives in a suburb of Boston and has managed care insurance.

During adolescence, Mrs B experienced back pain and was diagnosed as having kyphosis and lordosis, for which she wore a brace. Despite this, she remained active in sports. Four years ago, pain began in both shoulders without any preceding trauma or other joint symptoms. Her internist, and subsequently several orthopedists, evaluated her. The initial diagnoses included subdeltoid bursitis, rotator cuff tendonitis, and impingement.

Mrs B notes that shoulder pain interrupts her sleep every night. She cannot brush her hair, brush her teeth, or lift her arms without pain. She enjoys swimming but cannot swim the crawl or backstroke because of pain. She describes "cracking" and "popping" of the shoulder with activity. Initially, she did not have numbness or tingling in either arm, but she subsequently developed radiating pain and numbness in her fingers. Mrs B notes that her shoulders seem to improve in warm weather and become painful again with the onset of cold weather.

During the past 3 years, Mrs B received 2 cortisone injections in the right shoulder and 3 in the left. She described significant relief of the pain with improved range of motion, but the symptoms recurred within weeks. Physical therapy offered some relief, especially when using modalities of ultrasound, ice, and massage. She took ibuprofen, which afforded only minimal relief, and she was not interested in taking medication to solve the problem.

Her past medical history is notable only for occasional migraine headaches. Her family history is unremarkable. She does not smoke or abuse alcohol.

Physical examination revealed asymmetry of the left scapular motion, described as dysfunctional scapulothoracic motion, accompanied by palpable and audible crepitus. Range of motion in the shoulder revealed active forward flexion to only 90° and active abduction to 90° with significant pain limiting motion at that point. She had approximately 150° of passive forward flexion. Sensation, strength, and reflexes were normal. She had a markedly positive impingement test, with no pain with direct palpation of her acromioclavicular joint or sternoclavicular joint. She had a lot of pain with resisted external and internal rotation.

Findings from plain films of the shoulder were unremarkable, as were the results of an electromyogram. A magnetic resonance imaging (MRI)–arthrogram from May 1999 showed a small, high-signal focus within the supraspinatus tendon (<4 mm), consistent with degeneration vs tear; fluid within the subacromial-subdeltoid bursa but no definite extension of contrast to confirm a small focal tear; no large rotator cuff tear; and labrum blunted but otherwise intact.

Mrs B wonders whether the pain will persist or get worse. She wishes to know whether surgery would afford her a long-term solution and how difficult the recovery would be, considering her profession.

I'm a health and physical education teacher and also a coach, so I'm always involved in extracurricular activities. Approximately 4 years ago I started experiencing shoulder pain. Originally, the pain was in my right shoulder, but as it progressed, it began to develop in both shoulders. Now the pain is more severe in my left shoulder, to the point where I haven't had full range of motion for 2 years.

I received a series of cortisone shots in my shoulders. The cortisone injection itself was very painful. The last one was the most severe. Other than shots, physical therapy seemed to work the best. I had physical therapy for about 2 months. I felt the strength was better and felt a greater range of motion. The therapist treated me with ultrasound, ice, massage, and exercises, and I thought that was beneficial. It was convenient for me at that time. But to fit in physical therapy, as beneficial as it is, is very time-consuming.

The shoulder pain has affected my personal fitness and activities. I've really had to eliminate weight training, which I enjoy even more than cardiovascular exercise. I'm also concerned about when my husband and I are ready to have children: will this affect my ability to carry a child? I'm struggling with that.

I've had serious difficulty sleeping because of the pain, so I've used medication for sleeping. My sleep is still very interrupted every night, with shifting into different positions. But I don't like taking medications. I'd rather try to figure out something else to do.

It's a very personal decision whether to have surgery. I'm wondering what the pros and cons of surgery are. Obviously, you hear that the pain would subside, but how long will it subside for? Would the cracking in my shoulder subside? After the surgery, how long would the recovery process be? Would it inhibit my work with young children?

Mrs B is very healthy except for her shoulders. The problem is that she's a physical education teacher. That's good and bad. The good thing is she's very knowledgeable about range of motion exercises. She's religious in doing these, and she could educate other people about doing the exercises. The bad thing is that she needs to use her shoulders a lot when teaching classes. Plus, it is uncomfortable for her when she has a flare-up of her symptoms. It interferes with her sleep.

Despite treatment in the past, she's still symptomatic. Both her life and her work are affected by her shoulder discomfort. I referred her to an orthopedist, who, after a couple of visits, presented Mrs B with 2 options: surgery or observation and continued conservative treatment.

Mrs B had multiple steroid injections to both shoulders. Are there any adverse effects from repeated injections? What are safe intervals between consecutive injections? When does she need more than empirical treatment?

What is the prevalence of shoulder pain and dysfunction in the general population? What is the correct history and physical examination for a patient presenting with a painful shoulder? What is the role of plain films and MRI in the routine evaluation? What are the common syndromes of shoulder pain and dysfunction? What are the roles of physical therapy, NSAIDs, other medications, and injected steroids in the treatment of the painful shoulder? When is arthroscopic intervention indicated? What are the risks and benefits? When is shoulder replacement indicated? What do you recommend for Mrs B?

DR BURKHART: Mrs B is a 26-year-old physical education teacher with bilateral shoulder pain that is more severe on the left side. She has a history of scoliosis in adolescence, which was treated with a brace. She has pain with overhead motion as well as painful popping of the left shoulder that she can volitionally reproduce with active scapular motion. She has winging of the left scapula as she lowers the left arm. Electromyographic findings are normal. Magnetic resonance imaging results suggest a possible partial articular-surface rotator cuff tear. Physical therapy and NSAIDs have failed to give Mrs B adequate relief, and subacromial corticosteroid injections have provided only temporary relief.

Shoulder pain, largely related to the rotator cuff, ranks second only to back and neck pain as a cause of dysfunction in the working population.¹ The prevalence of persistent shoulder pain in those involved in static postures, heavy work, repetitive work, or overhead work ranges between 9% and 20% compared with a prevalence of 2% to 4% in age-matched controls.^{2 - 4}

Rotator cuff tear is the most common pathologic entity in the shoulder. Although cadaver studies of individuals older than 40 years have generally shown a prevalence of full-thickness rotator cuff tears between 5% and 20%,^{5 - 7} the prevalence of partial-thickness rotator cuff tears has been shown to be in the 30% to 40% range in adult cadavers.^{6 - 7} Furthermore, the incidence of rotator cuff tear increases with age, with approximately 50% of individuals older than 55 years demonstrating an arthrographically detectable rotator cuff tear.⁸ Other shoulder diagnoses are less common than rotator cuff tears. Adhesive capsulitis (frozen shoulder) occurs in approximately 2% of the adult general population⁹ and in 11% of the adult diabetic population.¹⁰ The prevalence of calcific tendinitis has been reported to be between 3%¹¹ and 7%.¹²

The shoulder has the greatest range of motion of any joint, with a vast array of muscles and tendons to power the glenohumeral and scapulothoracic articulations (Figure 1). The complex kinematics probably account for the fact that strains and sprains may remain symptomatic for much longer than in other joints. The glenohumeral joint accounts for approximately two thirds of shoulder motion, and the scapulothoracic joint provides the other one third.¹³

The main function of the shoulder is to position the hand in space so that the hand can perform its fine motor functions. Inability to position the hand results in profound impairment of function of the entire upper extremity. In the case of Mrs B, the inability to position the hand overhead has caused her to have a significant amount of dysfunction.

The shoulder as a functional unit comprises a unique combination of mobility and stability. Compromise of either the mobility or stability results in dysfunction. Mobility depends on a healthy articular surface, intact muscle-tendon units, and supple capsuloligamentous restraints. Stability requires intact capsuloligamentous structures and integrity of osseous articular structures (both the glenoid articular platform and the humeral articular arc¹⁴ ).

Impingement syndrome¹⁵ is the term given to the pain syndrome in which a tight subacromial space (eg, one in which the space is compromised by an anterior acromial osteophyte) causes abrasion of the rotator cuff against the acromion with overhead motion. The pain is secondary to the rotator cuff tendinitis and subacromial bursitis caused by the abrasion. Impingement-like symptoms may occur in patients with instability (instability-impingement overlap). Impringement syndrome is uncommon in individuals younger than 30 years.¹⁵

Rotator cuff tears can be caused by chronic impingement or by an acute traumatic injury with tensile overload of the rotator cuff.¹⁶ Weakness of some degree always accompanies rotator cuff tears. The amount of weakness is directly related to the size of the tear, and in small tears the weakness may not be clinically detectable. Sudden profound weakness with inability to raise the arm overhead signifies a massive tear. Acute massive tears demand prompt evaluation for surgical repair because they cannot be expected to improve with nonoperative management. Partial tears, such as the one suggested by Mrs B's MRI findings, require arthroscopy for definitive diagnosis.

Calcific tendinitis can be acute or chronic.¹⁷ The etiology of calcific tendinitis is unknown. In general, the condition is found more frequently in women than men.¹¹ Hsu et al reported a higher incidence of calcific tendinitis among Chinese patients than among other groups.¹⁸ A relationship to occupation must be suspected given the high incidence among clerical workers observed by Uhthoff and Sarkar.¹⁹ Sengar et al²⁰ found that HLA-A1 was found in 50% of patients with calcifications vs 27% of controls. The extreme pain associated with acute calcific tendinitis frequently resolves with corticosteroid injection but may require arthroscopic surgical excision of the calcific deposits. Chronic calcific tendinitis generally presents with impingement symptoms of pain with overhead motion.

Spinoglenoid ganglion cysts can cause an impingement-like pain syndrome by virtue of compression of the suprascapular nerve and can only be diagnosed by MRI.²¹

Instability of the shoulder comes in a variety of forms. Acute traumatic instability with dislocation of the shoulder is the most dramatic variety, usually requiring manipulative reduction by a physician. In younger age groups (about 25 years and younger), the chance of recurrent anterior dislocation after the initial event is greater than 95%.²² Chronic recurrent dislocations of the shoulder can lead to degenerative arthritis. Some lesser traumatic injuries can cause subluxation of the shoulder to such a degree that recurrent subluxation rather than dislocation becomes a source of dysfunction.²³

Atraumatic instability occurs in loose-jointed individuals.²⁴ Such individuals may have pain due to inability to control their laxity by means of their muscles, and they may develop enough superiorly directed instability that they present with impingement-like symptoms (instability-impingement overlap).²⁵

Athletes performing overhead movements, particularly baseball pitchers, may develop a "dead arm" syndrome²⁶ in which they have a painful shoulder with throwing and can no longer throw a baseball with their preinjury velocity. The root problem is usually a tear of the superior labrum, the so-called superior labrum anterior-posterior (SLAP) lesion. About 30% of these pitchers will have associated rotator cuff tears. Jobe²⁷ also proposed an "internal impingement" mechanism for the cuff tears with the arm in the cocked position of abduction and external rotation, a position in which the rotator cuff may be pinched between the posterosuperior glenoid and the greater tuberosity of the humerus. Nonathletes may also sustain traumatic SLAP lesions,²⁸ usually in falls or motor vehicle crashes (eg, drivers who have their hands on the wheel and sustain a rear-end impact).

Scapular dyskinesis can cause impingement-like symptoms due to protraction of the superior scapula.²⁶ This scapular protraction may be secondary to weakness of the lower trapezius, lower rhomboids, or serratus anterior; nerve palsy (eg, long thoracic nerve); or structural thoracic abnormalities due to congenital or developmental factors. Mrs B potentially could have scapular dyskinesis with scapular protraction due to developmental thoracic malformation secondary to her scoliosis.

Shoulder pain has a broad spectrum of patterns and characteristics. Boring, unrelenting, aching pain, even at rest, is characteristic of adhesive capsulitis (frozen shoulder).²⁹ Pain with overhead activities, as Mrs B has, is typical of subacromial disease, encompassing subacromial bursitis,³⁰ impingement syndrome,¹⁵ rotator cuff tendinitis,^{15 - 16} and rotator cuff tear.⁴ Pain with all motions, both at the side and overhead, is characteristic of osteoarthritis. Periscapular pain is often associated with local muscle strain or is referred from the cervical spine.

In addition to pain, shoulder dysfunction is often caused or exacerbated by loss of motion or weakness. Loss of active and passive motion is usually caused by adhesive capsulitis.²⁹ Loss of active motion with preservation of passive motion is caused by rotator cuff tear³⁰ or, rarely, suprascapular nerve injury.³¹ Shoulder weakness can be caused by rotator cuff tear or nerve injury (suprascapular, axillary, long thoracic, or thoracodorsal nerves, or cervical nerve root injury).³²

The physical examination of patients like Mrs B must include inspection for muscle atrophy, palpation for areas of tenderness and crepitation, muscle testing of all major muscles about the shoulder, measurement of active and passive range of motion, observation of symmetry of scapulothoracic motion, and specific testing for superior labral tear (Speed test, O'Brien test, Jobe relocation test²⁸ ) or instability (apprehension test²⁸ ) (Figure 2). The most reliable indication on physical examination for anterior instability is the apprehension test, in which the shoulder is brought into adduction and external rotation; patients with anterior instability will become apprehensive in that position, feeling that their shoulder is about to "slip out." Finally, the cervical spine must be palpated and taken through a range of motion to determine whether neck pathology is causing pain referred to the shoulder.

A comprehensive battery of plain film studies is essential in diagnosing shoulder pathology. Unfortunately, many physicians' offices and emergency departments still obtain views in only 1 plane (anterior-posterior view with the humerus in internal rotation and a second anterior-posterior view with the humerus in external rotation). This practice has been well documented to result in misdiagnosis, particularly missed posterior dislocation.³³ Therefore, in trauma cases, it is necessary to obtain a third view at right angles to the anterior-posterior views, either an axillary or a transscapular lateral view. For evaluation of the acromioclavicular joint, a 30° caudal tilt x-ray can be useful. For patients with subacromial symptoms, an outlet view,³⁴ which is a profile view of the acromion, can show the configuration of the acromion and identify whether a hooked osteophyte might be present on the anterior acromion. In traumatic anterior instability, the 2 pathognomonic pathologic lesions that occur are avulsion of the anterior-inferior labroligamentous complex from the glenoid (Bankart lesion³⁵ ), sometimes in association with a piece of glenoid bone (bony Bankart lesion), and an impression fracture of the posterolateral humeral neck (Hill-Sachs lesion³⁶ ) caused by contact with the glenoid rim that indents the humerus when the shoulder is dislocated. For patients with instability, the West Point view,³⁷ an obliquely oriented axillary view, will frequently show glenoid rim fractures (bony Bankart lesions) that are not visible on the standard views, and the Stryker notch view,³⁸ an axial oblique view with the humerus adducted, is useful for visualizing the Hill-Sachs lesion³⁶ in profile.

Bone scans are rarely used in the diagnosis of shoulder pain. Increased uptake in the distal clavicle³⁹ may confirm the clinical impression of osteolysis of the distal clavicle or degenerative arthritis of the acromioclavicular joint.³³ Osteomyelitis, although unusual about the shoulder, will generally show areas of increased uptake, as will most primary and metastatic bone tumors.

A computed tomographic (CT) scan is useful as a preoperative planning aid in some cases. Axial views accurately assess the amount of posterior glenoid erosion in patients with osteoarthritis, which can be indispensable in determining which total shoulder arthroplasty patients will require glenoid bone grafting. In addition, the evaluation of intra-articular fractures is often enhanced by axial CT scan views.

An MRI scan is invaluable in evaluating posttraumatic shoulder pain or shoulder pain that is refractory to conservative management. In the case of traumatic dislocation of the shoulder, MRI is very accurate in demonstrating Bankart lesions and Hill-Sachs lesions as well as associated rotator cuff tears.⁴⁰ Magnetic resonance arthrography has been advocated as a means of enhancing the appearance of labral avulsions.⁴¹ An MRI scan can generally indicate the approximate size of a rotator cuff tear and can also show whether the critically important subscapularis tendon is torn. Such information is very important to the surgeon in planning the location of arthroscopic portals and surgical incisions. In my experience, MRI is not very reliable in the diagnosis of partial-thickness rotator cuff tears. Therefore, the significance of the abnormalities on Mrs B's MRI scan suggesting possible partial-thickness rotator cuff tear is indeterminate.

The MRI scan is the only way to demonstrate the extra-articular spinoglenoid ganglion cysts²¹ that occur most commonly in active individuals in their 30s and 40s. Since these cysts are extra-articular in location, they cannot be visualized arthroscopically, and the diagnosis will be missed if an MRI scan is not obtained. In general, I prefer to obtain an MRI scan of the shoulder in all patients with traumatic shoulder pain, as well as in all patients whose shoulder pain fails to improve after 6 weeks of conservative management.

Mrs B has bilateral shoulder pain that is worse in the winter than in the summer. There is no history of trauma. Even though she has never dislocated either shoulder, the left shoulder feels as if it "pops out" when she raises it. Overhead motion is inhibited by pain. These symptoms are all nonspecific. Although the fact that she has pain with overhead motion might suggest an impingement syndrome, her description of possible instability symptoms raises the question of instability-impingement overlap,²⁵ a condition in which multidirectional instability permits enough superior migration of the humerus that impingement of the humerus under the acromion results. In this case, the impingement is not a primary phenomenon but is secondary to the instability.

The cause of shoulder pain may be trauma, inflammation, or degeneration. Traumatic shoulder pain, such as that caused by dislocation¹⁴ or rotator cuff tear,¹⁶ is well recognized. Repetitive trauma to the rotator cuff⁴² and glenoid labral tears^{28 ,43} are additional causes for shoulder pain, especially in athletes who perform overhead movements. Inflammation of the glenohumeral joint classically manifests as an adhesive capsulitis, or frozen shoulder.²⁹ Tendon degeneration has the histologic appearance of a hypercellular, disorganized structure with microfragmentation of normal collagenous architecture.⁴² However, tendon degeneration is not the only explanation for rotator cuff tears, particularly because cuff tears almost always initiate in the supraspinatus tendon.^{16 ,30} There is experimental evidence that degeneration must be accompanied by mechanical factors for cuff tears to occur. Soslowsky⁴⁴ simulated tendon degeneration in a rat model by injecting bacterial collagenase into the rat supraspinatus tendons and found lasting degradation of material properties only in those tendons to which a mechanical insult was added.

Most shoulder pain syndromes are appropriately treated initially by conservative means. Possible exceptions are acute traumatic shoulder dislocation and acute traumatic inability to raise the arm (acute massive rotator cuff tear). For all other shoulder pain syndromes, an initial minimum 6-week period of conservative empiric treatment is indicated. This regimen should include NSAIDs if tolerated, thermal modalities (ultrasound, ice, heat, or ice-heat contrast), and physical therapy. Physical therapy modalities generally involve the transfer of thermal energy. Heat therapy can decrease muscle guarding and thereby facilitate stretching.⁴⁵ Cold packs applied directly to the shoulder area are useful in decreasing pain.⁴⁶ Evidence for the role of massage and acupuncture is anecdotal.⁴⁷ In this cost-conscious era, patient-directed therapy with a portable home exercise kit for range of motion and strengthening exercises is becoming popular and is quite acceptable early after injury.

Formal physical therapy may be indicated in the high-performance athlete. One study demonstrated that a physical therapy program of rotator cuff strengthening could decrease rotator cuff and/or impingement symptoms to such a degree that surgery may be avoided.⁴⁸

Glucosamine and chondroitin sulfate theoretically serve as chondroprotective agents to treat or prevent arthritic pain in various joints, but studies of the drugs have limitations and few have evaluated the shoulder.⁴⁹ A meta-analysis has suggested some efficacy in hip and knee osteoarthritis pain.⁵⁰

Corticosteroid injection into the subacromial space may provide symptomatic relief in patients with impingement syndrome or supraspinatus tendinitis.⁵¹ Although concerns have been raised by animal studies that injection could weaken the tendon,⁵² only 1 case of rotator cuff rupture following steroid injection has been reported in the literature.⁵³

In my opinion, corticosteroid injections may be indicated in 3 situations. The first would be a patient with an impingement syndrome without rotator cuff tear (with MRI documentation) that is refractory to conservative management by the aforementioned approaches. The second situation would be an elderly, inactive individual with a rotator cuff tear who simply requires pain relief. I prefer to limit these individuals to 3 monthly corticosteroid injections, advising them that they must live with any residual pain that persists after the third injection or else consider surgical intervention. The third type of patient for whom corticosteroid injection has been shown to be helpful is the individual with adhesive capsulitis.⁵⁴ Patients with adhesive capsulitis have such a florid synovitis that any motion is painful. The goal of steroid injection is to reduce the synovitis to the point that the patient can comfortably stretch the shoulder to regain motion. In these patients, I typically give 3 corticosteroid injections at monthly intervals. If they have persistent pain and stiffness after 3 injections, they must decide whether they wish to continue conservative treatment without further corticosteroid injection or to pursue arthroscopic release and/or manipulation under anesthesia.

Knee arthroscopy was first performed in North America by Robert Jackson in the mid-1960s.⁵⁵ Early shoulder arthroscopists tried to adapt arthroscopic knee instruments to shoulder procedures, but in most cases, these instruments were woefully inadequate. After Ellman⁵⁶ introduced arthroscopic acromioplasty for impingement syndrome in the mid-1980s, excisional procedures in the shoulder rapidly developed. In the 1990s, as arthroscopic knot-tying techniques improved and suture anchors enabled surgeons to achieve secure fixation of soft tissues to bone arthroscopically, techniques for arthroscopic soft tissue repair steadily evolved.

By the end of the decade, some arthroscopic surgeons were able to perform arthroscopic repairs of massive rotator cuff tears¹⁶ and complex instabilities,¹⁴ but the degree of difficulty was much greater than for most other arthroscopic procedures. For that reason, arthroscopic instability repairs and rotator cuff repairs are currently "boutique procedures" routinely performed by only a handful of surgeons. Results reported by experienced arthroscopic surgeons are comparable to those of open surgical techniques,^{14 ,57} and the less invasive nature of these procedures greatly expedites rehabilitation. In addition, the arthroscopic procedures can routinely be done on an outpatient basis, achieving significant cost savings compared with open surgical procedures.⁵⁸ Patient demand for arthroscopic shoulder surgery is high, and I expect it to become the standard of care for shoulder surgery over the next decade.

Arthroscopic acromioplasty is widely used to treat refractory impingement syndrome. In general, a patient whose symtoms have not improved after 4 to 6 months of conservative treatment for impingement syndrome is a candidate for arthroscopic acromioplasty. Reported results of arthroscopic acromioplasty are generally equal or superior to those of open acromioplasty, with ability to return to preinjury activity levels in approximately 80% of patients.^{56 ,59}

Arthroscopic rotator cuff repairs have been reported to give excellent results by some investigators,^{16 ,60 - 61} but most surgeons still perform open surgical rotator cuff repairs, often in conjunction with an arthroscopic acromioplasty. In general, results of repair of the smaller rotator cuff tears are better than the results for repair of large tears. However, in a retrospective review of my patients, when I compared results of massive (>5-cm diameter) rotator cuff tears repaired arthroscopically with those repaired by open techniques (20 shoulders in each group), I found that patients in the arthroscopic group had a faster return of function and greater range of overhead motion than those in the open repair group (author's unpublished data, January 2000). Partial thickness rotator cuff tears, such as that suspected in Mrs B, may be amenable to debridement or repair.

Arthroscopic excision of calcific deposits can relieve the pain of the acute process. For chronic, established calcific deposits, arthroscopic excision may leave a large defect in the rotator cuff that requires concurrent arthroscopic rotator cuff repair.

Arthroscopic instability repair techniques have evolved from metallic staple fixation to transglenoid suture fixation to the current suture anchor fixation techniques. Reports of high recurrence rates,⁶² particularly with transglenoid sutures, led critics to the conclusion that the arthroscopic techniques were inferior to open reconstructions. However, arthroscopic suture anchor repairs performed by experienced arthroscopic surgeons now routinely yield results equal to those of open surgery, with recurrence rates of less than 10%.^{14 ,57 ,63 - 65} Patients with unstable shoulders and bone deficiency (bony Bankart lesion or large Hill-Sachs lesion) are at particular risk for recurrence after simple arthroscopic repair, and bone grafting of the defect may be indicated.¹⁴

Arthroscopic repair of superior labral lesions in athletes performing overhead movements is quite successful compared with open surgery techniques.^{26 ,28} Morgan et al²⁸ reported an 84% return to the preinjury level of performance in 44 consecutive baseball pitchers. The best reported results by open reconstructions (open capsulolabral reconstruction) in baseball pitchers was 50% in a much smaller group (12 patients).⁶⁶ Postoperative rehabilitation is an important part of the treatment of these athletes.

Arthroscopic capsular release is indicated in patients with refractory adhesive capsulitis (frozen shoulder).⁶⁷ These patients generally have symptoms for at least 6 months prior to arthroscopic release. Adhesive capsulitis has been portrayed as a self-limited disease, with symptoms usually resolving spontaneously within 2 years.²⁹ However, a number of patients, particularly those with diabetes, have pain and stiffness that persist beyond 2 years.²⁹ Patients with diabetes are not only more prone to adhesive capsulitis, but the capsulitis also is more severe. One study showed that the prevalence of frozen shoulder in 800 diabetic patients was 10.8%, compared with 2.3% in 600 controls without diabetes.⁶⁸ In diabetic patients with persistent shoulder stiffness lasting longer than 1 year, the only means reported that restores full motion is arthroscopic capsular release.^{67 ,69} Even for patients with adhesive capsulitis who do not have diabetes, arthroscopic capsular release can dramatically shorten the disease process, with an almost immediate recovery of normal range of motion.^{67 ,69}

For degenerative arthritis of the acromioclavicular joint or osteolysis of the distal clavicle, arthroscopic excision of the distal clavicle generally provides good relief of pain.^{70 - 71}

Replacement of the shoulder joint may take the form of hemiarthroplasty (proximal humeral replacement) or total shoulder replacement (replacement of the humeral and glenoid articular surfaces). Total shoulder replacement provides significantly greater pain relief than hemiarthroplasty, with approximately 80% of patients reporting pain relief after hemiarthroplasty vs more than 90% after shoulder replacement .⁷² However, glenoid component loosening has been problematic, particularly in active, muscular patients. These patients do not usually regain full overhead motion after arthroplasty. Therefore pain, rather than decreased motion, is the usual indication for shoulder replacement.

In my experience, most patients with osteoarthritis of the shoulder do not seek surgical treatment until the pain is so severe that they cannot sleep at night. When patients seek surgical intervention, I recommend hemiarthroplasty for the high-demand active patient (eg, a laborer performing overhead work or recreational golfer) because of the concerns about glenoid loosening and total shoulder arthroplasty for the low-demand patient (eg, a sedentary patient older than 65 years) because of the superior pain relief afforded by glenoid resurfacing.⁷² Recent manufacturer refinements of total shoulder arthroplasty components to a more anatomical design should improve the range of motion and decrease the rate of loosening.

Mrs B has refractory impingement symptoms. She received temporary relief from corticosteroid injections, but all other nonoperative treatments have failed to provide ongoing relief. I recommend arthroscopy of the shoulder with treatment based on the findings. If she has a partial-thickness rotator cuff tear, as the MRI findings suggest, I would recommend arthroscopic subacromial decompression with debridement or repair of the cuff tear, depending on how much of the thickness of the cuff is involved. If her clinical examination suggests an acromioclavicular joint derangement, I would excise the left distal clavicle arthroscopically. As for her apparent scapular dyskinesis, she should perform a specific exercise regimen to strengthen the scapular retractors. Even with surgery, it will be important for her to maintain an exercise program for the scapular stabilizers, deltoid, and rotator cuff.

MRS B: I'm at the point where I need to know whether surgery would improve this problem. Then, if surgery does improve it, what type of physical therapy goes along with surgery?

DR BURKHART: Even if you have surgery, you have a lifelong task of keeping your muscles strong in order to avoid additional surgery in the future.

A PHYSICIAN: If Mrs B has residual limitation of motion of her shoulder, wouldn't she benefit from mobilization by either manipulation or some other means before entering an arthroscopic procedure?

DR BURKHART: Mrs B has tried most of the conservative approaches already. If she had some adhesive capsulitis, her motion would definitely need to be maximized. One thing often mistaken for impingement is limitation of motion from adhesive capsulitis in the early stages. However, her motion is normal.

A PHYSICIAN: For Mrs B, would it be possible to have arthroscopy done as soon as possible or over the summer, during a break?

DR BURKHART: There is no rush because her rotator cuff is intact. Theoretically, you should fix rotator cuff tears earlier, before they have a chance to undergo any fatty degeneration within the muscle, but her cuff is intact.

MRS B: What would be the length of time after the surgery for recovery?

DR BURKHART: If you had a standard acromioplasty, you would recover rapidly in terms of getting your motion back. We start people moving the shoulder right away. Patients generally regain motion after 3 or 4 days, sometimes more quickly than that. Strength recovery takes a little longer, maybe 6 to 12 weeks. As patients get more active at about 3 months, their pain may worsen, and then it tends to improve again. Most people require about 4 months before they are back to normal.