Multidirectional Glenohumeral Instability Treatment & Management

Indications for surgical treatment of multidirectional instability (MDI) include the presence of persistent symptoms to a disabling degree and failure of conservative management, including a supervised rehabilitation program and a trial of activity modification or restriction. A reasonable trial of conservative treatment is 6-9 months. (See Nonoperative Management.)  Any patient for whom conservative management has failed may be counseled regarding the option of surgical treatment. The following points must be considered:

According to Neer and Foster, ^[2] contraindications for surgical management of MDI include the following:

• Willful, habitual, or voluntary shoulder instability
• Collagen connective tissue disorders (eg, Ehlers-Danlos syndrome, Marfan syndrome)
• Lack of a trial of, or noncompliance with, a supervised rehabilitation program

Future and controversies

Areas of future development in the treatment of MDI likely will parallel the further development of operative shoulder arthroscopy. The trend has been to drift away from open surgery, and it is very likely that this will continue. This trend is driven by patient and surgeon perceptions of less surgical invasiveness and reduced morbidity associated with arthroscopic techniques.

It remains to be seen, however, whether the effectiveness of arthroscopic management approaches that of the conventional open capsular shift technique, particularly after long-term follow-up. Open techniques include a capsular reconstruction with allograft to regain stability. 

Experience with thermal stabilization has mirrored that of thermal chondroplasty. These procedures have been mostly abandoned. Significant complications (eg, chondrolysis, capsular necrosis, tendon rupture, adhesive capsulitis, and axillary nerve injury) have been reported. The use of this technique grew rapidly, perhaps because it was relatively easy to perform and also because it avoided arthroscopic knot-tying. Although basic science studies underscore the apparent effectiveness of thermal stabilization in decreasing capsular laxity, clinical studies have not shown satisfactory results.

With regard to arthroscopic suture techniques, increasing basic science data are available that favorably compare the effectiveness of this evolving stabilization technique with the customary open shift. There are many questions, including where to place sutures, how many sutures to use, whether to use absorbable versus nonabsorbable sutures, how tight the plication should be, and what rehabilitation modifications are necessary.

Whether arthroscopic suture techniques can mimic an open shift repair over the long term remains unknown, but early clinical results have appeared promising. The possibility of overconstraining the joint must also be considered; this can lead to increased joint reactive force with arthritis. Again, more work is needed to continue to support the general use of the arthroscopic approach over its open counterpart.

Neer and Foster ^[2] stressed the importance of conservative management before surgery. In their original case series, the patients they selected for surgery had symptoms and disability for 1 year and had also undergone a trial of rotator-cuff and deltoid rehabilitation that failed. Furthermore, Neer and Foster carefully excluded patients with emotional problems (now referred to as intentional, habitual, and willful voluntary dislocators). ^[2]

Many patients respond positively to a supervised vigorous shoulder-conditioning program. Supervision is essential, at least initially, to ensure both compliance and effective instruction in the proper execution of an exercise program. Exercises should address all portions of the rotator cuff. A low-resistance, high-repetition, subimpingement-range, isotonic cuff-strengthening program works well with use of stretch cords or hand weights. These exercises are most beneficial when performed three to five times per week.

Isokinetic equipment is useful but not essential; it is helpful for interval testing to assess progress. The most important factor is to teach the patient the value of a persistent, ongoing effort at shoulder-girdle strengthening (ie, strengthening for life). Many patients initially do well with their exercise program but then lose discipline when symptoms subside—with recurrences subsequently following. Most patients learn to return to conditioning exercises when symptoms return.

In addition to a strengthening program, activity modification is necessary to eliminate, or at least reduce, pain and instability symptoms. Avoidance of unnecessary overhead motions (eg, throwing, racquet sports, or swimming), side carrying and lifting, and pushing and pulling may be required. Work modifications also may be considered.

Anti-inflammatory medications or analgesics may help during exacerbations of MDI. Steroid injections have a limited place in the nonoperative treatment of MDI but may be helpful in the treatment of secondary impingement syndrome. Narcotic medications have no place in the management of MDI.

A reasonable duration for conservative treatment is 6-9 months; any patient in whom conservative treatment has failed may be counseled regarding the option of surgical treatment. The following points must be considered:

• Does the patient have sufficient disability to make surgery a worthwhile endeavor?
• Has the patient shown satisfactory effort and dedication to the preoperative rehabilitation program?
• Is the patient willing to continue postoperative rehabilitation for at least 6 months, and does he/she understand the importance of shoulder rehabilitation and muscle strengthening to the stability of the shoulder?
• Is the patient willing to comply with postoperative limitations such as immobilization, activity limits, and work and sports restrictions?
• Is the patient willing to accept the possibility of lost range of motion (ROM)?
• Is the patient willing to accept the possibility of recurrent symptoms (ie, failure of the procedure)?

This author currently uses a traditional open surgical approach for severe MDI; uses an arthroscopic plication technique for mild-to-moderate MDI, as well as for anteroinferior-predominant and posteroinferior-predominant MDI; and does not use or advocate thermal treatment, in light of the reported complications, especially with the risk of axillary nerve injury, the potential for capsular necrosis (see Complications), and the reports of poor results. ^[27]

Key elements of the various options are discussed below.

Open surgical management

The landmark paper on MDI was that of Neer and Foster, published in 1980. ^[2] Not only did this classic paper facilitate widespread recognition of the MDI problem that was previously believed to be rare relative to unidirectional instability, but it also described a capsulorrhaphy to address the pathologic capsular laxity associated with MDI.

Of the 32 patients in this study, 31 (97%) had satisfactory results following the inferior capsular shift procedure with no recurrent instability, no significant postoperative pain, full strength, and full return to activity. ^[2] However, this was a preliminary study; by today's standards, ROM was questionable, with satisfactory defined as ROM within 10° of elevation and 40° of rotation of the contralateral side.

A biomechanical study by Wang et al ^[28] of joint reactive forces and glenohumeral translations in a cadaveric model reported that the inferior capsular shift was superior to a unidirectional anterior capsular repair in reproducing more normal forces, kinematics, and mechanics.

In a follow-up report by Neer and Foster, ^[29] the longevity of the results of the inferior capsular shift procedure for MDI was documented in larger numbers of patients with longer follow-up.

Other reports of the use of the inferior capsular shift procedure for MDI were published early in the learning curve. Recurrence rates consistently were reported at 10% or lower. ^{[15, 30, 31, 16]} However, average loss of motion has remained variable, with the best cited as only 6° of elevation loss and 3° of external rotation loss in a series of active-duty naval personnel at 28 months, ^[16] and only 5° and 4° of external rotation loss at 0° and 90° of abduction, respectively, at 36 months. ^[31]

Postoperative ROM obviously varies from patient to patient and surgeon to surgeon, and it is likely to also be a function of the rehabilitation program.

The classic Neer-type inferior capsular shift has been modified by making the shift in the glenoid side, rather than the humeral side, and by applying it to patients with a predominantly anterior-inferior instability pattern. In 42 shoulders, of which 90% had a concomitant Bankart repair and 50% had generalized ligamentous laxity, four (9.5%) had recurrent instability. ^[31] Interestingly, three of these four (7.1% of 42) had recurrent posterior instability. However, in the successful category, motion loss was relatively small, averaging only 5° of external rotation, with no more than 5° of elevation loss.

The classic humeral-side inferior capsular shift procedure has been applied to lesser degrees of MDI as well. The procedure has also held up well in patients who are very active.

Bigliani et al ^[32] reviewed 63 patients, in whom a combined total of 68 inferior capsular shifts had been performed. Most were performed primarily for anterior-inferior instability, excluding combined anterior and posterior patholaxity and associated glenoid fractures. These patients were athletic, including 31 throwers, and 21 (30.9%) of these surgeries also included a Bankart procedure.

Results were good to excellent in 94% of the cases, with 92% of patients able to return to their previous sports. ^[32] However, only 75% of the patients were able to return at their previous level of play (only five of 10 were elite throwers). Motion loss was less than in Neer and Foster's series, averaging only 7° of lost external rotation. Two patients (3.2%) had redislocations resulting from falls, one (1.6%) had musculocutaneous nerve palsy that resolved, and 10% of patients had persistent minor clicking.

Arthroscopic management

It is only comparatively recently that arthroscopy in the setting of MDI made the transition from the diagnostic to the operative realm. In the first (1994) edition of his textbook Shoulder Arthroscopy, Snyder made no mention of arthroscopic management techniques for MDI, but he described the diagnostic findings at arthroscopy. ^[33] Snyder stated, "If surgery is required for atraumatic laxity, most often an open capsular shift procedure is used." He referred the reader to Neer and Foster's classic 1980 paper. ^[2]

However, various options have continued to evolve for arthroscopic management of MDI. The same basic principles that apply to open surgery apply to arthroscopic surgery as well. The goals are as follows ^[34] :

• Reduction of overall capsular patholaxity anteriorly, posteriorly, and inferiorly
• Closure of the rotator-cuff interval
• Minimization of morbidity and risk of complications

Thermal capsulorrhaphy

Thermal or radiofrequency (RF) "capsular shrinkage" was advocated in the past because of its technical ease and simplicity relative to traditional open capsulorrhaphy or arthroscopic suture techniques.

Abundant basic science research has shown that the application of RF energy to collagen tissues results in shortening proportional to temperature and duration of contact, ^{[35, 36, 37, 38, 39, 40]} as well as ultrastructural changes ^[41] , in energy and time-dependent fashions. ^{[42, 43]}  Technique-dependent ^[44] temporary decreases in strength and stiffness of treated capsular tissue ^[45]  have been demonstrated and are likely the basis for observed mixed clinical results, with high failure rates (up to 60%) reported in some cases. ^{[46, 27, 18, 19]}

This technique has largely been abandoned.

Arthroscopic suture plication (suture capsulorrhaphy)

Results of arthroscopic suture management of MDI are premature, and long-term follow-up is pending. Snyder ^[14] reported on 23 of 24 patients who had a capsular plication for glenohumeral instability in the absence of a Bankart lesion. He noted that MDI was present in eight of the 23 seen for follow-up at an average of more than 24 months. American Shoulder and Elbow Surgeons (ASES) scores improved on average, and Rowe scores were good or excellent in 78% of cases. Snyder termed this technique a promising alternative to open techniques. For further information, see Shafer et al. ^[47]

Wolf and Durkin ^[13] presented their results of suture plication in 20 of 26 patients with average 34-month follow-up (minimum follow-up, 24 months). With regard to pain, strength, activity, ROM, stability, and overall satisfaction, results were good to excellent in 75% of cases. Workers' compensation claims correlated with unsatisfactory outcome, and five patients with recurrent instability required a total of eight additional surgical procedures.

Treacy et al ^[17] reported their results of an arthroscopic capsular shift for MDI in 26 patients and noted 88% satisfactory results in 24 patients available for review at an average of 52 months (range, 28-72 mo). Three patients had postoperative instability. All but one of the 24 patients had regained full symmetrical ROM. The authors felt that these results were comparable to the results of open surgery.

Two clinical studies reported 2- to 5-year results. Gartsman et al ^[48] evaluated 47 patients with MDI managed with arthroscopic plication with or without interval closure, reporting 94% good-to-excellent Rowe scores accompanied by a significant increase in UCLA shoulder score, with 86% return to sports. Baker et al ^[49] also demonstrated similar functional improvements and high rate of return to athletic activities after arthroscopic management of MDI.

A review article by Caprise and Sekiya ^[50] is worthy of mention and comment. The authors noted that although arthroscopic suture techniques for the treatment of MDI are new and evolving, these methods "have comparable results to open techniques when the multifactorial nature of the disease is recognized and the multiple techniques are used in combination to fully treat all pathology.... The advantages of a less invasive procedure make arthroscopic capsular plication attractive, but it is associated with increased technical difficulty and a steep learning curve."

Caprise and Sekiya stressed that further research and follow-up are needed, and they emphasized that the goal of any surgery for MDI, whether open or arthroscopic, is "addressing the capsular laxity and redundancy to restore anatomic capsuloligamentous tension without overconstraining the shoulder." ^[50]

Cadaveric studies have shed light on the effects of capsular plication on capsular tightness. ^[47] Gerber et al ^[51] studied the effects of selective capsular plications on the ROM of the shoulder in eight cadaveric specimens and found predictable patterns of motion loss. Medial-to-lateral 1-cm plications resulted in significant losses of motion, as follows:

• Anterosuperior plication, 30.1° loss of external rotation in adduction
• Anteroinferior plication,19.4° loss of abduction and 20.6° loss of external rotation
• Posterosuperior plication, 16.1° loss of internal rotation in adduction

Furthermore, total anterior and total posterior plication resulted in loss of flexion of 20° and abduction of more than 15°, whereas total anterior plication had a greater than 30° loss of external rotation and total posterior plication had a greater than 20° loss of internal rotation. ^[51] Finally, total inferior plication resulted in loss of abduction of 27.7°.

Alberta et al ^[52] found that in six cadaveric specimens that underwent "stretching" of the anteroinferior capsule that resulted in increased external rotation of 23.2° without increased glenohumeral translation, selective anteroinferior plication reduced this external rotation by more than 12°. The center of rotation of the glenohumeral joint (GHJ) was posteriorly and inferiorly shifted, with  49-61% loss of anterior translation at 15N and 20N loads, respectively. Finally, the capsulolabral "bumper" height more than doubled, from 2.9 mm to 6.4 mm. Such observed restrictions of motion may improve clinical instability but may have consequences for the long-term function of the GHJ.

Flanigan et al ^[53] found losses of cadaveric capsular volumes of 16.2% and 33.7% with 5- and 10-mm plications. Sekiya et al ^[54] found cadaveric capsular volume decreases greater than open techniques when "multiple pleats" were taken, mirroring arthroscopic techniques.

There has been evidence that aggressive tightening, especially with interval closures, may limit ROM, especially external rotation. ^{[55, 56]}

Thus, the application of these techniques involves a degree of subjective judgment, and more research is necessary to delineate guidelines for appropriate tightening.

This discussion of the open surgical technique focuses first on the classic treatment for MDI, followed by modifications. Useful and recommended modifications are in parentheses.

Neer and Foster ^[2] emphasized the basic principle of capsular detachment from the humeral neck on the predominant side of instability and then shifting the capsule to the opposite side of the calcar. The goals are to reduce the capsular redundancies on the more-involved approach side and also on the opposite side, while additionally obliterating the axillary pouch. Neer and Foster also emphasized the creation of thickenings in the repaired shifted capsule by folding and overlapping capsular flaps. Permanent nonabsorbable sutures are used.

After a thorough examination under anesthesia (EUA) to elucidate the dominant direction of instability, the patient is placed in a beach-chair position for a planned anterior approach. If arthroscopy is performed before a planned open anterior approach, it is useful to perform the arthroscopic segment in the beach-chair position as well. If a posterior approach is planned, arthroscopy is best performed in a lateral position, with open surgery continuing via a posterior approach in the same position.

In other words, the surgical position should be clear following EUA, with patient position maintained through both the arthroscopic and open portions of the procedure. In this author's experience, arthroscopic findings rarely change the plan following EUA.

For severe MDI, both anterior and posterior aspects of the shoulder are exposed, and a deltopectoral approach is planned. A 9-cm incision is made from the axillary crease distally to the coracoid proximally. (This is a long incision and is rarely necessary. In fact, an axillary incision measuring more than 4 cm is sufficient in a patient who is relatively slim. See the images below.)

The interval between the deltoid and pectoralis major muscles is identified, as well as the cephalic vein, which then is retracted medially. (Take the vein in any direction it wants to follow.) Retract the conjoined tendon medially. (Respect the musculocutaneous nerve, which typically penetrates the conjoined tendon 3-5 cm below the coracoid process.)

The subscapularis tendon is identified. Cauterize the vessels at their lower margins, superficial to the subscapularis. Neer and Foster recommended next dividing the superficial one-half thickness of the subscapularis tendon 1 cm medial to the long head of the biceps, whereas the deeper one-half thickness is left on the capsule for reinforcement. (It is wise to leave slightly more [1.5 cm] lateral subscapularis tendon.)

The medial subscapularis flap is tagged, dissected from its deeper component, and retracted medially. (The lower third of the subscapularis tendon is mostly muscle, with minimal tendon fibers.) The underlying capsule is inspected, and the interval between superior and middle glenohumeral ligaments is closed with nonabsorbable sutures (see the image above). (This is a constant finding, and the rotator-cuff interval closure is essential to reduce inferior glenohumeral translation, but it may be difficult to reach with a very small incision in the axilla.)

Next, a T-shaped incision is made in the capsule, with the stem of the T aimed at the glenoid, traversing the interval between the middle and inferior glenohumeral ligaments, and the top of the T parallel to the humeral anatomic neck (1 cm from its lateral insertion). Two capsular flaps thereby are created, one superomedial (SM) and the other inferomedial (IM) (see the images below). During exposure and repair, it is very helpful to tag each flap corner with a long suture to control the flaps.

The proximal-distal incision is gradually extended distally, and the IM flap is divided from the inferior humeral neck around to the posterior portion of the neck, while the axillary nerve is carefully protected with a flat retractor. To assist in control of the flap and to aid in visualization, it is helpful to place nonabsorbable sutures at 1-cm intervals in the lateral margin of the IM flap while progressing distally; the sutures will be used in the repair on the way out.

Once the dissection reaches the posterior portion of the capsule and humeral neck, it is useful to apply traction to the IM flap in order to test the effect of capsular shortening on posterior capsular tension and to estimate and adjust necessary capsular repair tension.

With a gauge or curette (a rongeur works well), a shallow groove is then fashioned in the anterior-inferior portion of the humeral neck adjacent to the capsular reflection. Then, the IM flap is advanced in a proximal direction to eliminate the inferior pouch and increase posterior tension. The lateral edge of the IM flap of the capsule is sutured to the remnant lateral capsular tag or adjacent subscapularis stump (by using the aforementioned sutures). Once this has been accomplished with the IM flap, any redundant superior portion may be reflected inferiorly to thicken the anterior capsule.

Finally, the SM flap is advanced distally and inferiorly and similarly sutured to the superior and anterior lateral capsular remnant and subscapularis stump.

An overlap develops as the SM flap is advanced inferiorly. This overlap serves to further reinforce the anterior tissues (see the image below). Neer and Foster recommended securing the capsule with the arm in slight forward flexion and at about 10° of external rotation. To avoid excessive tension, the author secures the repair sutures with the arm in at least 45° of abduction and 45° of external rotation.

The subscapularis tendon is closed at its normal location, with care taken to avoid shortening anatomically. Matsen et al ^[11] showed that a shortening of 1 cm can theoretically limit rotation by 20°. After the remaining closure is finished, the arm is immobilized at the side in a splint or a sling with a chest pad on neutral flexion and 20° of internal rotation.

For the posterior-inferior predominant instability pattern, a posterior approach may be chosen. One trend has been to perform all shifts from the front; the reasoning behind this is that the rotator-cuff interval cannot be closed from the back, and reasonable posterior tightening can be obtained from the front.

For a posterior approach, according to Neer and Foster, a 10-cm incision is made either horizontally or vertically over the posterior-lateral scapular spine and posterior glenohumeral joint. They recommended detachment of the deltoid from the posterior acromion and scapular spine, followed by a vertical 2- to 3-cm split to expose the underlying external rotators. However, detachment of the deltoid can usually be avoided, in that it can be more simply retracted upward.

As with the dissection of the subscapularis anteriorly as described above, the infraspinatus is divided near its insertion and peeled medially, leaving some of its fibers on the posterior capsule for reinforcement. The posterior capsule normally is very thin posteriorly; hence, this step is important.

Again, a T-shaped capsulotomy is made, creating SM and IM flaps. The IM flap is dissected and released progressively around the inferior humeral neck in an anterior direction while the axillary nerve is carefully protected throughout. A trough is prepared on the posterior-inferior humeral neck. Eventually, the IM flap is advanced and repaired gradually in a superior direction, eliminating both anterior patholaxity and the axillary recess. Then, the SM flap is advanced over the top of the IM flap to reinforce and add bulk to the middle posterior capsule.

Afterward, the infraspinatus is repaired in an anatomic fashion, followed by the posterior deltoid.

The arm is immobilized in neutral flexion-extension and 10° of external rotation for 6 weeks. This author generally immobilizes the arm in 30-45° of abduction and 30° of external rotation for 4 weeks, followed by a 1-week transition to the Neer and Foster position, followed by a sling.

As recently as 1994, the role of the arthroscope in the evaluation and management of MDI was limited to a diagnostic function. ^[33] Since then, however, as with other shoulder applications, operative arthroscopy for MDI has been developing and evolving rapidly.

Operative arthroscopy for MDI can be used for either primary or adjunctive functions. Open surgery—namely, the open capsular shift—is predictable, safe, and successful, with a proven track record. Thus, any new arthroscopic approaches must be evaluated against open surgery with regard to efficacy and safety.

The general principles of open surgical treatment must be addressed via arthroscopic means. Furthermore, should arthroscopic approaches prove easier, as well as effective and safe, they must not displace or preempt a routine trial of conservative management before consideration of surgical treatment.

Arthroscopic stabilization of the MDI shoulder can be performed in either the beach-chair or the lateral position. A thorough EUA must precede diagnostic arthroscopy.

Routine utilitarian portals are established. The posterior portal is made 1.5 cm distal and medial to the posterior-lateral corner of the acromion; the anterior portal is made1.5 cm medial and proximal to the coracoid process, between the long head of the biceps and the upper edge of the subscapularis intra-articularly. If the surgeon plans on suturing the posterior capsule, or if there is known posterior labral pathology, then the posterior portal should “cheat” more laterally and superiorly to facilitate placement of sutures, anchors, or both.

Arthroscopic suture repair techniques were developed by Snyder, ^[14] who called this approach "capsular pinch-tuck," or "plication surgery." The arm position is lateral, at 70° abduction and 10° flexion. Two anterior portals in the rotator-cuff interval are created, as well as a posterior superior portal. The synovial surfaces are excoriated on the capsule and adjacent areas of the labrum.

While the surgeon views from the posterior portal and uses a suture hook through one of the anterior portals, a pinch-tuck of capsular tissue is taken 1 cm lateral to the labrum, and the needle and tissue are approximated to the edge of the labrum. ^[14] The needle is then passed through the labrum (it has now captured both the capsule and labrum). First, a suture relay is passed through the suture hook; then a suture is passed via the relay in the opposite direction out the original cannula. This thus leaves a suture crossing the labrum and also through the capsular fold.

The process is repeated at 1-cm intervals along the labrum in an inferior direction; each suture is tied with an arthroscopic knot pusher by a sliding knot technique (Snyder recommended the Tennessee slider knot). ^[14] The number of tucks and the extent of anterior, inferior, and posterior tightening are left to the individual surgeon's judgment. Snyder warned that the axillary nerve is at risk with a deep pass through the inferior capsule.

The images below show posterior plication and are representative of posterior capsular laxity (see the image below). The view is of the right shoulder of a patient in the beach-chair position from the anterosuperior portal, just anterior to the biceps long head, aimed in a posteroinferior direction. The working portal is the typical posterior portal, which is cheated slightly higher and more laterally.

First, a suture passer device (Suturelasso; Arthrex, Naples, FL) is placed through the working cannula; next, it is initially passed through a pinch of posterior capsule 1 cm from the labrum and then through the posterior labrum itself (see the image below).

Next, with nonabsorbable suture employed in an all-arthroscopic knot-tying technique (sliding knot first, backed up by an alternating post, alternating half-hitch technique), knots are placed from inferior to superior, plicating the capsular pinch to the labrum. Capsular pinches or tucks may vary at the surgeon's discretion, and the number of sutures and the spacing between them also may vary (1-cm spacing is typical). Caution is advised in passing sutures in the inferior regions anteriorly and posteriorly, given the proximity of the axillary nerve to the inferior capsule. (See the images below.)

An arthroscopic interval closure is also typically added to reduce inferior laxity; this may be done last, after completion of plication sutures (see the images below).

Sometimes, the labrum may be deficient, hypoplastic, abraded, or torn and thus insufficient for use via direct suture passage technique. In such a situation, suture anchors may be helpful (see the images below). The anchor is placed on the margin of the articular surface, and the attached suture is then passed through the capsule to achieve a standard "tuck," with or without the labrum if possible, to achieve a "caposulolabral reconstruction" and a "bumper-stop" configuration, to enhance stability.

General complications of instability repairs may arise with any technique of MDI stabilization, whether open or arthroscopic. Failed repairs can result from a number of causes, including (but not limited to) the following:

• Errors in diagnosis
• Failure to address specific pathology (eg, omitting a Bankart repair in favor of a capsulorrhaphy when a labral detachment is present)
• Loose repair
• Problems with healing (including failure to recognize collagen diseases such as Ehlers-Danlos and Marfan syndromes)
• Postoperative noncompliance
• Reinjury

Norris ^[57] stressed that errors in diagnosis can include treating impingement as primary with decompression, missing secondary impingement caused by instability. This problem is especially common in throwers, swimmers, and other athletes who use overhead arm motions. A high index of suspicion for secondary impingement is required. Overtensioning the tighter side of a multidirectionally loose shoulder does not appear to be a common problem, but it is possible that a shift performed from the anterior side of a posterior predominant MDI pattern may worsen the posterior component, despite correcting the inferior component.

The inferior capsular laxity must be addressed in the MDI shoulder, and this has been discussed above. Failure to satisfactorily correct the inferior capsular laxity, failure to tighten the rotator-cuff interval, or failure to adequately support the shoulder postoperatively may lead to recurrent instability. Furthermore, for revision surgery after initial MDI surgery has failed, it is important to be sure that the interval has been repaired, that there are no labral detachments, and that the capsular flaps are firmly secured to the glenoid.

The axillary nerve is at particular risk during the inferior dissection and during development of the inferior flap in both anterior and posterior open approaches. ^[58] The relation of the nerve to the inferior capsule must also be kept in mind with use of arthroscopic thermal and suture techniques.

Some authors advocate exposure and isolation of the axillary nerve during this portion of the procedure. However, dissecting around the axillary nerve merely for the purpose of identifying it may paradoxically cause injury. It may be enough to maintain an elevator immediately beneath the inferior capsule while working on the inferior flap. (This author has not seen any axillary nerve injuries in hundreds of repairs and has yet to make a specific effort to identify and isolate the nerve.)

In an excellent cadaveric study, Price et al ^[59] examined the relation of the axillary nerve to the inferior capsule as the nerve passes through the quadrangular space, in order to define the risk to this structure from an arthroscopic perspective, specifically in the axillary nerve's relationship to the glenoid rim and the inferior glenohumeral ligament (IGHL). The authors used a simulated lateral decubitus position, akin to typical arthroscopic positioning, with the arm in 5-lb (~2.25 kg) traction and in 45° abduction and 20° flexion.

The following findings from this study are most relevant. ^[59] The axillary nerve branches from anterior to posterior. The branch to the teres minor was closest to the rim of the glenoid, and the branch to the anterior deltoid was the farthest, with the branch to the posterior deltoid and the superior lateral cutaneous branch both intermediate in position, the latter closer to the glenoid than the posterior deltoid motor branch.

The axillary nerve was closest to the glenoid rim at the 6 o'clock meridian, averaging a distance of 12.4 mm (11.6-13.2 mm at the 95% confidence interval [CI]) at this site. ^[59] At 10 mm anterior and posterior to the 6 o'clock meridian, the axillary nerve averaged a distance of 14.5 mm and 13.9 mm, respectively. Furthermore, the nerve was a mere 2.3 mm (1.7-2.9 mm at the 95% CI) from the IGHL at the 6 o'clock meridian, and averaged 2.8 mm at 10 mm anterior and posterior to the 6 o'clock meridian.

The limitations of this study were discussed by the authors. ^[59] They did not replicate capsular abnormalities that may be associated with unstable shoulders (ie, a loose capsule) or the effects of arthroscopic distention, both of which may alter the "normal" relations defined above. Clinical applications of this work may explain the predominance of sensory deficits with arthroscopic axillary nerve injury. The teres minor must be carefully evaluated because injury to the teres minor branch of the axillary nerve may be difficult to discern.

Significant axillary nerve risks and morbidity have been reported in both in-vivo and in-vitro studies of thermal procedures used to treat the inferior capsule. ^{[60, 61, 62, 63, 64]}

The relation of the axillary nerve to arthroscopically placed capsulolabral sutures was also studied by Eakin et al. ^[65] Ten cadaveric shoulders underwent suture placement that mimicked arthroscopic suture plication techniques. Sutures were placed in a simulated lateral decubitus position with the arm in 45° abduction and 20° flexion with 10-lb (~4.5 kg) traction, through the capsule 1 cm from the glenoid rim, and then through the labrum at anterior (3:00 or 9:00 o'clock), anteroinferior (4:30 or 7:30), posteroinferior (4:30 or 7:30), and posterior (3:00 or 9:00) positions.

The average distance of each suture position to the axillary nerve was 16.7 mm (13.7-19.7 mm at the 95% CI) for the anterior sutures, 12.5 mm (10.2-14.8) for the anteroinferior, 14.4 mm (10.9-17.9) for the inferior, 24.1 mm (19.7-28.5) for the posteroinferior, and 32.3 mm (28.4-36.4) for the posterior sutures. ^[65]

The authors noted a statistically significant trend for the axillary nerve to lie closest to the anteroinferior sutures and then gradually at farther distances from more posteriorly placed sutures. They concluded that a "safe zone" exists between the common locations of suture placement for arthroscopic plication and the axillary nerve, but they urged caution during anteroinferior and inferior suture placement. ^[65]

Axillary nerve injuries are not unique to arthroscopic management of shoulder instability. Neer and Foster reported three cases of axillary neurapraxia in their landmark presentation of open inferior capsular shift. ^[2]

Before the abandonment of thermal capsulorrhaphy of the shoulder, complications were increasingly being reported. Weber ^[66] reported on 15 patients referred to his practice for complications related to this treatment method, including recurrent instability (11 patients), axillary nerve injury (three), adhesive capsulitis (two), and capsular necrosis (two). He advocated salvage of recurrent instability with revision open capsular shift.

Weber noted the transient nature of the axillary nerve injuries, but painful neuralgia persisted in two cases. ^[66] The stiff shoulders required subsequent capsular release but failed to gain complete motion at final follow-up. Capsular necrosis is difficult to treat and may require autografting or allografting for salvage (Warner JP, personal communication, 2001). Weber stressed that these complications are serious but that true rates of complications are unknown and, therefore, the "RF technique" should be used with caution until more data are available.

Although there some successes were reported (Ceballos et al ^[46] reported high patient satisfaction, high return to preoperative activity levels (including athletes), and no neurologic problems), Karas et al ^[27] reported an overall failure rate of 26%; 50% of those failures occurred in individuals with posterior instability and 30% occurred in individuals with MDI.

D'Alessandro et al ^[18] and Hawkins et al ^[19] reported high failure rates, up to 60%, and a significant risk of chondrolysis and neurologic injury. Miniaci et al ^[67] noted a failure rate of 47% and reported four transient axillary nerve problems (three sensory and one motor, all resolving by 9 months) in 19 MDI patients followed for 2 years after monopolar thermal capsulorrhaphy. 

Wong and Williams ^[68] reported the results of a survey compiled by members of the American Shoulder and Elbow Surgeons, the Arthroscopy Association of North America, and the American Orthopaedic Society for Sports Medicine. The authors focused on recurrence of instability, axillary nerve injury, and the incidence of capsular necrosis following monopolar, bipolar, and laser thermal treatment for glenohumeral instability.

Thermal treatment was reported to be used in 14,277 (6%) of 236,015 cases, with most utilizing monopolar RF, where the rates of recurrent instability ranged from 7.1% and 8.4%. ^[68] Furthermore, between 18% and 33% of patients requiring revision surgery showed evidence of capsular attenuation (33% with laser treatment). The incidence of associated axillary nerve injury was 1.4% (least with laser), with 95% recovering between 2 and 4 months.

Proximal long head biceps tendon rupture has also been reported after thermal capsular capsulorrhaphy. ^{[69, 70]}

No axillary nerve injuries were reported in either of the suture plication series of Snyder ^{[33, 14]} and Wolf and Durkin. ^[13]

Neer and Foster recommended 6 weeks of postoperative immobilization, followed by heat and gentle assisted exercises. ^[29] Their goal was for ROM to be 20° less than the opposite shoulder. They advocated that patients perform isometric exercises at 8 weeks postoperatively and progressive resistive exercises beginning at 12 weeks postoperatively. Additionally, Neer and Foster restricted playing sports and lifting more than 20 lb (~9 kg) for 9 months and advised against swimming using back and butterfly strokes, heavy overhead use of the arm, and contact sports for 12 months after surgery.

Modern protocols for repair of traumatic instability are more aggressive, with the philosophy having shifted in parallel with knee rehabilitation. The focus is on obtaining complete ROM, with earlier institution of rotator-cuff strengthening in order to protect the surgical repair. Whether this opinion applies to the MDI-reconstructed shoulder may be debatable. This author has used the same protocol for both types of surgery.

According to Norris, ^[57] the most common complication of rehabilitation is recurrent instability caused by early motion and return to activity before complete healing. The opposite consideration, slow motion, is of at least equal concern because of the consequence of permanent motion loss and, if severe, iatrogenic arthritis similar to failed Magnuson-Stack and Putti-Platt procedures (incidence reported at 43%). ^[57]

The author's current protocol for anterior capsular shift repairs is divided into four phases and is outlined below.

Phase I: immediate postoperative phase (weeks 1-6)

Weeks 3-4

Treatment consists of the following:

• Swelling and pain management
• Sling for 3 weeks; can remove when awake, but observe ROM restrictions
• Sleep in immobilizer for 6 weeks
• Elbow/hand ROM
• Passive/gentle active assisted ROM exercises, with motion to tolerance and comfort - By the end of 4 weeks: flexion and abduction 60-75º, external rotation in scapular plane 15-20º, internal rotation in scapular plane 30-40º
• Shoulder isometrics (submaximal/pain-free isometrics) for flexion, internal rotation and adduction
• Cryotherapy and modalities as indicated
• NO MOTION ABOVE SHOULDER HEIGHT
• NO ACTIVE EXTERNAL ROTATION, EXTENSION, OR ABDUCTION

Weeks 4-5

Treatment consists of the following:

• Discontinue sling during the day if indicated
• Sleep in immobilizer
• Elbow/hand ROM
• Passive/gentle active assisted ROM exercises, with motion to tolerance and comfort - By the end of 5 weeks: flexion 90º, abduction to 90º, external rotation in scapular plane 30º, internal rotation in scapular plane 30º
• Shoulder isometrics (submaximal/pain-free isometrics) for all planes
• Light tubing external/internal rotation with the arm at the side
• Upper-body ergometer (UBE) for ROM only
• Cryotherapy and modalities as indicated
• NO MOTION ABOVE SHOULDER HEIGHT

Weeks 5-6

Treatment consists of the following:

• Sleep in immobilizer until end of week 6
• Passive/gentle active assistive ROM exercises, with motion to tolerance and comfort - By the end of 6 weeks: flexion 135-140º, abduction to 140º, external rotation at 90º and abduction 45º, internal rotation in scapular plane 45º
• May initiate stretching exercises
• Initiate scapular stabilization exercises; emphasize posture
• May start active ROM for all shoulder motions
• Pool therapy (no swimming)
• Modalities as indicated

Weeks 6-7

Treatment consists of the following:

• Active warmup on UBE
• Passive/active assistive ROM exercises, with motion to tolerance and comfort - By the end of 7 weeks; flexion 170-180º, abduction to 170-180º, external rotation at 90º and abduction 60-75º, internal rotation at 90º and abduction 65-70º
• Initiate rotator cuff–strengthening exercises
• Modalities as indicated

Phase II: moderate protection phase (weeks 7-14)

Weeks 7-8

Treatment consists of the following:

• Active warmup on UBE
• Passive/active assistive ROM exercises, with motion to tolerance and comfort - By the end of 8 weeks: external rotation at 90º and abduction 90º, internal rotation at 90º and abduction 80º, horizontal adduction 45-50º
• Add resistance to rotator cuff program

Weeks 11-12

Treatment consists of the following:

• Active warm up on UBE
• Passive/active assistive ROM exercises, with motion to tolerance and comfort - By the end of 12 weeks: external rotation at 90º and abduction 115-125º if overhead thrower
• Continue with all stretches as above
• Progress to more aggressive strengthening
• Initiate golf-swing motion (week 12)
• Initiate light swimming (week 12)

Phase III: minimal protection phase (weeks 14-21)

Weeks 14-18

Treatment consists of the following:

• Active warmup on UBE
• Continue with all stretches and flexibility as above
• Full rotator-cuff program
• Proprioceptive neuromuscular facilitation (PNF) manual resistance - PNF rhythmic stabilization drills
• Endurance training
• Initiate plyometric drills
• Two-handed drills progressing to one-handed
• Initiate swinging of bat: hit off tee (week 16)

Weeks 17-21

Treatment consists of the following:

• Active warmup on UBE
• Continue with all stretches and flexibility as above
• Continue with all strengthening as above
• Initiate interval throwing program

Phase IV: return to activity phase (weeks 22-32)

Weeks 21-23

Treatment consists of the following:

• Continue all strengthening and stretching as above
• Progress interval throwing to throwing off of the mound
• Progress to unrestricted sports participation at 23 weeks; continue with rotator-cuff strengthening