DEFINING THE PROBLEM

The shoulder is a versatile complex of joints with a large functional range of motion. However, stability is sacrificed for this freedom of movement. The glenohumeral joint is a loosely opposed “ball-and-socket” type joint, and as such is the most commonly dislocated large joint in the body. The lifetime incidence of anterior shoulder dislocation has been reported to be 1% to 2% in the general population.^1,2 Recurrent instability after primary dislocation is a common problem.

After initial reduction of the joint, the traditional treatment for primary shoulder dislocations has been immobilization in a sling, with the arm in a position of adduction and internal rotation. The length of the immobilization period is controversial. A common recommendation is for 3 to 6 weeks in a sling to allow for early capsular healing followed by several months of rehabilitation including range-of-motion and strengthening exercises. Some authors have recommended abbreviated periods of immobilization, as short as 1 week, followed by early institution of therapy to promote restoration of a full range of motion.

The clinical course of patients after nonsurgical treatment has been investigated extensively. Of particular interest is the relatively high rate of recurrent instability in young patients. Estimates of the recurrence rate in this group have been reported to be anywhere between 17% and 96%.^2–10

The anatomic abnormality associated with the majority of traumatic shoulder dislocations is the Bankart lesion—an avulsion of the anterior glenoid labrum and shoulder joint capsule at the insertion of the inferior glenohumeral ligament. The anterior capsulolabral complex is thought to be the major stabilizer of the shoulder to anterior subluxation. The frequency of recurrence, therefore, is postulated to be related to failure of this complex to heal in an anatomic position. Proponents of acute arthroscopic Bankart repair have emphasized that it is the ability to directly oppose and secure the lesion to its anatomic position at surgery that imparts immediate stability versus nonoperative methods. Early series of arthroscopic treatment of shoulder instability have shown it to be an effective modality in decreasing recurrence.^11,12

Despite evidence in favor of arthroscopic stabilization, uncertainty still exists among the orthopedic community as to optimal timing of surgical intervention.^13,14 The cost, risk, and anxiety associated with surgery have caused some surgeons to adopt a strategy of “watchful neglect,” allowing patients to demonstrate recurrent episodes of instability or dislocation before proceeding with an operative solution. Furthermore, although early surgery has the potential to result in a number of patients avoiding redislocation in the short term, the long-term functional benefits over a conservative approach are less well known.

This chapter attempts, via an evidence-based approach, to address whether early surgical intervention improves outcome after a primary shoulder dislocation. For the most part, this is a narrative overview of published investigations that describe results after conservative or arthroscopic treatments, or both, for first-time anterior shoulder dislocations. For the purposes of this review, the studies were divided into three categories: those reports describing primarily nonsurgical treatments, those describing arthroscopic treatments, and those designed as prospective comparative cohorts and/or randomized trials utilizing both surgical and nonsurgical treatment arms.

The specific patient population of interest for the review was young patients (<40 years old) who had sustained first-time, traumatic, anterior dislocations of the glenohumeral joint. These patients represent a group at high risk for recurrent dislocation and in whom treatment decisions are controversial. Studies pertaining to the methods of and issues regarding the reduction of shoulder dislocations were excluded. Also excluded were those investigations that included (or whose primary focus was) recurrent shoulder dislocations/instability, atraumatic dislocations, posterior dislocations, and multidirectionally unstable patients.

The concept of symptomatic recurrent instability merits further clarification. For the purposes of this review, patients were considered to have recurrent instability if the patient had sustained a subsequent (documented) dislocation or if the patient sought surgical intervention for symptoms of instability and/or subluxation, (but not necessarily dislocation) which he or she felt was intolerable. This definition excludes those patients who may have demonstrated clinical findings consistent with instability (most notably a positive apprehension test) but who did not wish to proceed with further intervention.

There were 43 potentially relevant study titles. More studies were identified secondarily on the basis of bibliographical reviews and hand searches of proceedings. Application of the study eligibility criteria eliminated 11 of the original titles, with 32 articles remaining for review: 13 describe primarily nonsurgical management, 9 describe arthroscopic interventions, 4 are observational studies that include the results for both surgically and nonsurgically treated patients, and 6 are randomized trials.

NONSURGICAL STUDIES

The natural history—or clinical course in the absence of active treatment (whether it be surgical or rehabilitative)—after anterior dislocation of the shoulder has been investigated extensively. Reports dating to the early part of the twentieth century highlight the recurrent nature of the problem.

In those reports predating 1990, retrospective (Level II, III, and IV) studies predominate.^4–6 These series are subject to the limitations of bias and confounding inherent in any investigation that relies on retrospectively collected data. Furthermore, most of these series include both primary and recurrent dislocators in the study sample. Perhaps the largest early series is that of Rowe and colleagues,⁴ which reports on 488 patients with shoulder dislocations seen at Massachusetts General Hospital during the 20-year period from 1934 to 1954. Of the entire group, 398 (82%) were patients with primary dislocations with the remainder being recurrent. No information is provided as to immobilization or rehabilitative protocols prescribed for these patients. Although the inclusion of patients with recurrent instability makes it difficult to draw specific inferences on those patients with primary instability, Rowe was the first to document a relatively high recurrence rate (38%) after primary dislocation in a large series and to make conjectures about possible prognostic factors for recurrence including young age at initial dislocation and traumatic cause. Age was identified as the strongest single prognostic factor, with 83% of those patients younger than 20 years going on to further instability episodes.

In more recent investigations, there has been conflicting information as to the rate of recurrent instability and the optimum nonsurgical management. Henry and Genung⁷ reported on a group of 121 first-time dislocators that was also reviewed retrospectively (Level II). A rate of recurrence of 85% to 90% was noted in these patients within 18 months of the injury; this was significantly greater than the previous reports of Rowe and others.⁴ All patients in that study were younger than 32. Half of the group was immobilized in a sling for a variable period. Although randomization was not performed, equivalent rates of recurrent instability in immobilized and nonimmobilized patients led the authors to conclude that immobilization was of little benefit in reducing recurrence. Another Level III retrospective series of 124 patients by Simonet and Cofield⁸ reports a lower overall rate of recurrence (33%), but again stresses the importance of age as a risk factor with more than 50% of those patients younger than 40 years experiencing further instability.

Controversy as to the optimal rehabilitation program after first-time dislocation was introduced by both Aronen and Regan,⁹ as well as Yoneda and coworkers,¹⁰ who reported seemingly lower recurrence rates of 25% and 17.3%, respectively, with aggressive physical therapy. The former of these investigations was prospectively conducted on a small cohort of young male military cadets (n = 20)—a recognized “high-demand” population (Level IV). The authors used a rigorous rehabilitation protocol. Despite the fact that this protocol was described in detail, subsequent studies utilizing it on a similar population have failed to replicate the results.¹¹

A prospective, multicenter cohort study was initiated in 1978 in a Swedish population (Level II). The results of this cohort have been published at 2-, 5-, and 10-year follow-up.^15–17 The group of 247 primary dislocations in patients between the ages of 12 and 40 years were quasi-randomized (on the basis of even or odd date of injury) to receive one of three methods of conservative treatment: sling and swathe for 3 weeks, sling for 1 week followed by restricted range of motion, or a “mixed” treatment for patients believed to be unfit for either of the first two arms. No significant differences were found in the rate of recurrent dislocations at 10 years after injury among the three groups. Overall, approximately 50% of the patients had experienced symptoms of instability at 10-year follow-up. Young age was again put forth as the single strongest predictor for recurrence. The presence of an associated fracture of the greater tuberosity was identified as protective. No association between recurrent instability and sex, handedness, or side of dislocation was identified. This Level I prognostic study represents the highest current level of evidence available regarding the natural history of anterior shoulder dislocations with long-term, prospectively collected data. This study has an impressively low loss to follow-up (3%) over the 10-year course.

A recently published study has suggested that immobilizing acutely dislocated shoulders in a position of relative external rotation reduced subsequent dislocation rates.¹⁸ In a quasi-randomized prospective trial (Level II), Itoi and coauthors¹⁸ demonstrate a 0% recurrence rate in 20 patients immobilized in external rotation versus a 30% recurrence rate in 20 patients immobilized in a sling at a mean follow-up of 15.5 months. Several randomized clinical trials are ongoing in Canada, the United States, and Australia to test this concept in a young at-risk population. The characteristics of the nonoperative studies are outlined in Table 98-1.