Operative skills have advanced sufficiently so that. Arthroscopic Treatment of Multidirectional Glenohumeral Instability: 2- to 5-Year Follow-up

Transcription

1 Arthroscopic Treatment of Multidirectional Glenohumeral Instability: 2- to 5-Year Follow-up Gary M. Gartsman, M.D., Toni S. Roddey, Ph.D., P.T., O.C.S., and Steven M. Hammerman, M.D. Purpose: We present the results of a prospective study evaluating arthroscopic technique in repairing multidirectional glenohumeral instability. Type of Study: Case series. Methods: The 47 patient study group consisted of 26 men and 21 women. Inclusion criteria were multidirectional glenohumeral instability diagnosed on physical examination and at arthroscopy. Exclusion criteria were unidirectional anterior or posterior instability and prior instability operation. Four patients declined to participate in the study. The average age at the time of operation was 30 years (range, 15 to 56 years). The average interval from operation to final evaluation was 35 months (range 26 to 67 months). The American Shoulder and Elbow Surgeons (ASES) Shoulder Index, Constant, Rowe, and UCLA scores were recorded preoperatively and at final evaluation. Results: Preoperatively, no patients were rated overall as good to excellent according to the Rowe scale; at final follow-up 94% (44 of 47 patients) were rated as good to excellent. One patient was considered a failure of the index operation due to persistent instability and underwent a second operative procedure. One patient noted a loss of strength during sports, and 2 patients had pain that limited their throwing ability. The ASES Shoulder Index improved to 94.7 from 45.4 (P.001). The absolute Constant score improved to 91.7 from 60 (P.001). The Rowe score improved to 93.7 from 14.2 (P.001). The UCLA total score improved to 33.1 from 17.4 (P.001). Average passive external rotation at 90 abduction measured Twenty-two of 26 patients (85%) returned to their desired levels of sports following the operations. Conclusions: Patients with multidirectional glenohumeral instability have multiple lesions within the shoulder and the surgeon must individualize the operative treatment. Arthroscopic surgery produced successful results in 44 of 47 patients. Key Words: Arthroscopic multidirectional glenohumeral instability. Operative skills have advanced sufficiently so that surgeons can repair glenohumeral instability using arthroscopic techniques. The proposed advantages of arthroscopic stabilization include the following: smaller skin incisions, more complete glenohumeral joint inspection, ability to treat intra-articular lesions, access to all areas of the glenohumeral joint for repair, less soft-tissue dissection, and maximal preservation of motion. 1-3 METHODS From Texas Orthopedic Hospital, Fondren Orthopedic Group (G.M.G., S.M.H.); The Department of Orthopaedic Surgery, University of Texas Houston Health Science Center (G.M.G.); and Texas Woman s University (T.S.R.), Houston, Texas, U.S.A. T.S.R. was supported in part by a grant from Columbia/HCA and Texas Orthopedic Hospital, Houston, Texas. Address correspondence and reprint requests to Gary M. Gartsman, M.D., Texas Orthopedic Hospital, 7401 South Main St, Houston, TX 77030, U.S.A. garyg@onramp.net 2001 by the Arthroscopy Association of North America /01/ $35.00/0 doi: /jars We report the results of a prospective study for patients treated between January 1994 and November 1997 with arthroscopic repair of multidirectional glenohumeral instability. The study was prospective as all patients possessed a single clinical diagnosis and were evaluated preoperatively and at final examination with shoulder scales and standardized range of motion and stability testing on physical examination as described below. 236 Arthroscopy: The Journal of Arthroscopic and Related Surgery, Vol 17, No 3 (March), 2001: pp

2 MULTIDIRECTIONAL GLENOHUMERAL INSTABILITY 237 Study Hypotheses Because the purpose of this study was to evaluate the final result of patients undergoing arthroscopic repair of multidirectional glenohumeral instability, we decided, based on a literature review, 2,4-10 that a final mean Rowe score of greater than 85 was a satisfactory result. We made this decision because either functional limitation (in work or sports) or discomfort when the arm is examined for apprehension (patient apprehension when the shoulder is placed in 90 abduction, externally rotated and an anterior stress is applied) will deduct 15 points and could result in a total score of 85. We believed that the presence of either functional limitation or discomfort on apprehension testing was unacceptable. Therefore the null (primary) hypothesis (H o ) of this study was that arthroscopic repair of multidirectional glenohumeral instability results in a final mean Rowe score of 85 (H o : 85). The alternative hypothesis (H a ) was that the final mean Rowe score would be greater than 85 (H a : 85). The alpha level was set at The effect size was selected based upon a literature review 1-3 and our prior experience. We estimated that a mean score greater than 10 points from 85 would result in an unfavorable outcome. Based on this, an effect size of 0.60 was estimated. A power analysis was then performed to quantify the number of patients required to reject our null hypothesis if in fact it should be rejected. Using an effect size of 0.6, with a 1-tailed or unidirectional analysis at a desired power of 0.95, it was determined that approximately 35 subjects were required. 11 Inclusion/Exclusion Criteria The inclusion criterion was a diagnosis of multidirectional glenohumeral instability on preoperative examination that we confirmed at the time of arthroscopic operation. This diagnosis was made on a combination of the following signs and symptoms: (1) The patient s description of shoulder dislocation or sensation of looseness and slipping. (2) Physical examination that demonstrated pain and/or apprehension with excessive anterior, inferior and posterior translation. (3) Findings at arthroscopic operation that documented multidirectional glenohumeral instability. Exclusion criteria were unidirectional anterior or posterior instability and instability in only 2 directions (anterior-inferior or posterior-inferior). Patients with prior instability operations were excluded. We also excluded Workers Compensation patients because they have various issues that adversely affect their outcome. Misamore and others have documented inferior results following shoulder operations in this population. Fifty-six patients met the criteria described above. Four patients declined to participate in the study because they did not think they could attend the number of required clinic visits for follow-up evaluation. Five patients failed to return for final evaluation. Therefore, this report evaluates 47 patients with 47 shoulder operations. Study Group Population We recruited the patients from the private practice of a single orthopaedic surgeon who performed all operations at a private hospital. We gathered no information about annual income or ethnicity but all patients had private health insurance. The authors were not designated physicians for any high school, college, or professional sports team, and the majority of our athletic patients participated at a recreational level. Six patients participated in high school sports (1 in football, 2 in basketball, and 1 in softball) and 1 participated in college athletics (football defensive lineman). Twenty-one patients assessed their own level of sports activity as recreational. Nineteen patients did not participate in sports. No professional or semiprofessional athletes were included in this study. The group consisted of 26 men and 21 women; the average age at the time of operation was 30 years (range, 15 to 56 years); 23 dominant and 24 nondominant shoulders were involved. Preoperative Assessment To allow comparison of this report with others in the literature, we collected sufficient data to rate patients according to the American Shoulder and Elbow Surgeons (ASES) Shoulder Index, the Constant scoring system, the scoring system of Rowe, and the University of California at Los Angeles (UCLA) Shoulder Scale. 9,15-17 Before surgery, all patients completed self-assessment questionnaires to document their levels of pain, satisfaction, and function. Patients who had sustained a specific traumatic event were questioned about the mechanism of injury and asked to recall, if possible, the position of the arm at the time of the trauma. All patients were questioned about the position of the arm or activity that reproduced their symptoms. We recorded the sports participation, if any, of each patient. The level of sports participation was categorized as

3 238 G. M. GARTSMAN ET AL. high school team sport, college team sport, or recreational. We measured active ranges according to the Constant rating system, which included forward flexion, abduction, external rotation in abduction, and behind the back internal rotation. Passive elevation and external rotation (with the arm adducted), as well as external rotation and internal rotation with the arm abducted 90 were measured by the examiner using a hand-held goniometer and recorded to the nearest 5. Elevation strength was measured with the arm elevated in the scapular plane 90 and internally rotated using a dynamometer with the result recorded in pounds. We did not use anesthetic injections before examining the shoulder for strength, so we are unable to quantify how much of the strength loss was the result of pain. The stability examination was performed on both shoulders. We compressed (loaded) the humeral head into the glenoid during all maneuvers. Anterior translation was assessed with an anterior force applied to the shoulder with the arm in 90 of abduction. The Rowe test 18 was used to assess inferior-anterior translation. To perform this examination, the patient stands and flexes the trunk from the hips approximately 30. In this relaxed position, the shoulders are effectively elevated 30 and the examiner applies a distraction force. Inferior translation was assessed with an inferior force applied with the shoulder at 0 of abduction (Sulcus test). Posterior translation was examined with the arm elevated 90, adducted slightly, and rotated internally. The amount of internal rotation was altered from 90 to 30 as a posterior force was applied and then the shoulder was extended. We recorded the presence or absence of pain and apprehension for each instability maneuver. We graded the amount of humeral head translation on the glenoid surface as 0 (stable or trace laxity), 1 (up to 50%), 2 (more than 50% but not dislocatable) and 3 (dislocatable). 19 The grading of instability was subjective, as we made no attempt to measure the degree of translation with fluoroscopic observation. We recorded the presence of laxity in the contralateral shoulder and elbows, and the patient s ability to bring the thumb to the forearm, but did not use any formal grading system for the degree of generalized ligamentous laxity. Ligamentous laxity was categorized as present or absent on the basis of this examination. We excluded other sources of shoulder pain (rotator cuff lesions, acromioclavicular joint arthritis, thoracic outlet syndrome, brachial plexus lesions, and glenohumeral arthritis) through patient history, physical examination, and radiographic analysis. Routine radiographs included anteroposterior glenoid, axillary, and supraspinatus outlet views. 20 Radiographic imaging (magnetic resonance, computed tomography, and arthrography) was not routinely obtained. We did not require the presence of radiographic abnormalities for inclusion in this study because the majority of patients with multidirectional instability have normal plain radiographs. The primary indication for surgery was persistent shoulder pain caused by multidirectional glenohumeral instability that had not responded to a minimum 6-month nonoperative program that consisted of avoidance of painful activities, nonsteroidal anti-inflammatory medication, and a home physical therapy program designed to maintain or improve shoulder girdle strength. Our goal was to improve the strength of those muscles responsible for glenohumeral stability. Therefore, patients performed resistive exercises of the deltoid, internal rotators, external rotators, biceps, triceps, and the scapular muscles with surgical tubing and light weights (maximum 5 lb). 21 We evaluated patients at 5 postoperative intervals within the first year: 2 weeks, 6 weeks, 3 months, 6 months, and 1 year. Thereafter, they were evaluated yearly. Patient Classification All patients showed excessive translation in the anterior, inferior, and posterior directions. To increase diagnostic precision, we classified each shoulder in regard to chronicity, degree, and traumatic onset. We documented (according to the patient s description) whether the instability was chronic or acute (less than 6 weeks), and also classified the instability degree as either recurrent dislocation or recurrent subluxation. We recorded whether patients developed instability after a traumatic event of sufficient magnitude to damage the glenohumeral ligaments (traumatic or atraumatic). Operative Technique Before administration of general anesthesia, all patients received an interscalene block to diminish postoperative pain. The anesthesiologist administered 1 g of cephalosporin intravenously. We placed patients in the sitting position and examined both shoulders as described above. The shoulder joint was entered with a cannula and blunt trocar through a posterior skin incision placed 1.5 cm inferior and 1.5 cm medial to the posterolateral border of the acromion. An anterior portal was identified with a spinal needle (outside-in technique) so that the cannula entered the shoulder

4 MULTIDIRECTIONAL GLENOHUMERAL INSTABILITY 239 joint immediately superior to the subscapularis tendon and 1 cm lateral to the glenoid. An inspection of the glenohumeral joint was performed. We re-examined the shoulder for translation while viewing the shoulder through the arthroscope. The arthroscope was removed and inserted through the anterior cannula to more completely inspect the posterior glenohumeral joint. The arthroscope was then returned to its posterior cannula. All structures within the glenohumeral joint were examined systematically and any evidence consistent with instability recorded. Such signs were variable and included partial-thickness tears of the rotator cuff, rotator interval, and biceps tendon. To evaluate the glenohumeral ligaments for midsubstance tear or plastic deformation, we also assessed them for laxity by directly observing and palpating them with an arthroscopic probe and applying translational stresses as we rotated the shoulder. We documented the location on the glenoid of labrum detachment. Labra that were frayed or had midsubstance tears were noted. The cartilage was inspected for damage to the glenoid and to the humeral head (Hill-Sachs lesion). Once the diagnosis of glenohumeral instability was confirmed, an anterosuperior portal was created and a second cannula placed through the rotator interval, 1 cm lateral to the glenoid. The goals of the procedure were to repair ligament or labrum detachments anatomically, and then to recreate adequate capsular tension. Since we believe that the labrum contributes to glenohumeral instability through the mechanism of concavity-compression, 22 we repaired labrum separations from the glenoid (if present) anatomically. The labrum tear(s) were repaired beginning, if needed, with the posterior labrum and proceeding as necessary to the inferior and anterior and superior labrum. We did not make any attempt to shift the labrum. Drill holes for suture anchors were placed through the glenoid articular surface approximately 2 mm from the lateral glenoid margin. The detached labrum was sutured so that it was in contact with the scapular neck and extended on to the glenoid articular surface in order to establish the labrum as a bumper and to recreate optimal conditions for concavity-compression. If the labrum was intact but the glenohumeral ligament had torn from the labrum, and the labrum was of sufficient size to allow suture placement within its substance, the ligament was repaired directly to the labrum with monofilament suture. If the labrum was absent, the capsule was advanced onto the glenoid articular cartilage surface and repaired with suture anchors (as described above) to create a labrum bumper. We could restore capsular tension at the same time by placing the capsular suture further from the glenoid. Typically, 5 to 15 mm of lateral and superior ligament advancement was required. We used the same techniques if the capsule was lax but not torn from the glenoid or labrum. To assess the adequacy of the repair, we examined the shoulder for translation after we inserted each suture. After we inserted all posterior, inferior, and anterior sutures, if persistent excessive translation was present, a suture was placed in the rotator interval. A suture passer was used to place a monofilament suture through the capsular tissue immediately anterior to the anterior supraspinatus tendon and then through the capsule superior to the subscapularis tendon. If a greater degree of tightening was required then the superior capsular tissue was sutured to the middle glenohumeral ligament. A sliding knot was used to tie the suture extra-articularly. The details of this technique have been previously described. 23 Postoperative Management Postoperative management was similar for all patients. A soft pillow sling supported the arm in 15 of abduction. An ice pack wrap decreased postoperative shoulder swelling and pain. Patients went home either the same day or the morning after operation. Active range of motion exercises of the fingers, wrist, and elbow, as well as deltoid muscle isometric exercises were started the morning after the operation and continued at home for a period of 2 weeks. At 2 weeks, we obtained an anteroposterior radiograph to document the position of any metallic suture anchors. We instructed the patients to remove the sling 4 times a day and perform active range of motion exercises in elevation and external rotation. Elevation was limited to 90 (to protect any inferior and posterior repair) and external rotation was limited to 40 (to protect the anterior and rotator interval repair). Sling wear continued for 6 weeks, at which point the sling was removed and the patient began unrestricted active range of motion and strengthening exercises. Patients continued range of motion and strengthening exercises for 1 year. We used self-assessment forms to document shoulder pain, function, satisfaction, and level of sports activity. We measured active and passive ranges of motion as well as strength. No postoperative radiographic imaging studies (ultrasound, magnetic resonance, arthrography) were performed routinely.

5 240 G. M. GARTSMAN ET AL. Data Analysis Initial data screening was accomplished using scatterplots, histograms, and frequency tables for all variables. Further diagnostics were completed on any potential outliers using regression diagnostics and Studentized residuals. Violations of linearity, homoscedasticity, and independence were assessed on the scatterplots. 24 A -square goodness of fit test was used to evaluate our null hypothesis that the mean postoperative Rowe score would be 85 points. Our alpha level was set at Paired t tests were used to determine if there were differences between preoperative and postoperative scores. Within-subjects analyses of variance were used to evaluate whether there were any differences among items (frequency of preoperative dislocations) with 3 or more variables. Tukey post hoc testing was completed for all possible pair-wise comparisons, with the overall experimental alpha level maintained at Standard statistical software (SPSS, Chicago, IL) was used to analyze the data. RESULTS The average duration of symptoms prior to surgery was 16 months (range 6 to 120 months). The average time from operation to final evaluation was 35 months (range, 26 to 67 months). Twenty-seven patients had recurrent subluxation and 20 had recurrent dislocation. Twenty-one patients developed shoulder instability after a single traumatic event and 26 patients developed instability without trauma. We classified the instability as chronic in all 47 patients. Radiographic Analysis All patients underwent preoperative radiographic imaging with standard radiographs. The radiographs were normal in all 47 patients. One patient underwent computed tomography before evaluation in our clinic and the abnormality identified was a SLAP lesion. Seventeen patients underwent preoperative magnetic resonance imaging. The results were normal in 8 patients. Abnormal findings included 5 anterior labrum tears, 1 posterior labrum tear, 1 rotator interval tear, 3 large anterior capsules, and 3 partial-thickness rotator cuff tears. Operative Findings The essential finding in these shoulders was a large capsular volume and excessive humeral head translation anteriorly, inferiorly, and posteriorly. Our ability to assess capsular volume was subjective and was based largely on our observation of the failure of the capsule to possess adequate tension to contain the humeral head as we rotated and translated the shoulder in all 3 directions. No patient required repair of a partial-thickness rotator cuff tear. The operative findings responsible for the excessive translation were variable and all patients had multiple lesions. The findings are summarized in Table 1. Operative Repair In no patient was labrum repair alone sufficient to restore glenohumeral stability and all patients required capsular tightening. The findings are summarized in Table 2. We inserted suture anchors in 27 patients. The average number of bone/soft tissue anchors used was 2 (range, 0 to 4). TABLE 1. Operative Findings Rotator cuff Partial-thickness grade 1 2 Partial-thickness grade 2 5 Biceps Longitudinal partial tear 1 Rotator interval, abnormal 28 SGHL, abnormal 25 MGHL, abnormal 47 AIGHL, abnormal 47 Inferior capsule, abnormal 47 Posterior capsule, abnormal 47 Superior labrum Present 47 SLAP lesion 10 Type 2 9 Type 3 0 Type 4 1 Anterior labrum Present 38 Absent 9 Bankart lesion 10 Inferior labrum Present 30 Absent 17 Frayed 3 Separated 2 Posterior labrum Present 23 Absent 24 Frayed 2 Separated 6 Hill-Sachs lesions 5 Loose bodies 1 Abbreviations: SGHL, superior glenohumeral ligament; MGHL, middle glenohumeral ligament; AIGHL, anterior inferior glenohumeral ligament.

6 MULTIDIRECTIONAL GLENOHUMERAL INSTABILITY 241 Postoperative Scores and Shoulder Rating Systems The postoperative results are shown in Table 3, and represent the condition of the patients at final follow-up examination. The 4 rating systems reflected an improvement in shoulder status. Comparing the scores before surgery with those at final follow-up, paired t tests showed significant increases in the total and subscales scores for the ASES Shoulder Index, Constant, Rowe, and UCLA scoring systems (P.0001). Primary Hypothesis Testing The final mean Rowe score for all patients was To evaluate the results of our null hypothesis (that the mean postoperative Rowe score is 85 points), a -square goodness of fit statistic was calculated. The value for t was calculated and, with a df of 37, was P.0005 was determined to be significant. Therefore, we could reject the null hypothesis and accept the alternative that arthroscopic instability operation produced a mean Rowe score of greater than 85 points. Satisfaction TABLE 2. Surgical Repair Labrum repair Superior 10 Anterior 10 Inferior 2 Posterior 6 Capsular suture imbrication Anterior 47 Middle 47 Inferior 47 Posterior 47 Rotator interval repair 28 Patients used the UCLA scale to rate their levels of satisfaction. Preoperatively, no patient rated satisfaction as good to excellent (4-5, out of a possible 5) on the UCLA scale. Postoperatively, 45 of 47 patients (96%) rated their satisfaction as good to excellent (4-5), and 2 patients as fair to poor (0-3). Function Patients answered a shoulder function self-assessment questionnaire using the 4 shoulder scoring systems. Preoperatively, no patient rated function as good to excellent (Rowe function subscale, out of 50 points). Postoperatively, 44 (94%) patients rated their function as good to excellent, and 5 as fair to poor. All 10 ASES activities of daily living function items showed significant improvement (P.0001). The total ASES function score improved to 48.3 (out of a possible 50 points) from 25.9, and this was statistically significant (P.0001). Range of Motion No patient lost more than 5 of elevation. External rotation at 90 of abduction averaged 88. Two patients had external rotation of 85 at 90 of abduction. One patient had 80 of external rotation at 90 abduction. Return to Sports Participation Twenty-eight patients had participated actively in sports before the onset of their shoulder problems. At final follow-up evaluation, 2 patients did not participate in sports due to issues unrelated to their shoulders and 4 patients had decreased their level of participation. One patient had persistent instability, 1 patient with a stable shoulder on examination had pain while throwing, and 2 patients experienced loss of strength. None of the following parameters showed a statistically significant difference: recurrent dislocation and recurrent subluxation (P.112), traumatic or atraumatic etiology (P.269), the number of preoperative dislocations (P.315), age at operation (P.445), arm dominance (P.379), gender (P.266), or ligamentous laxity (P.785). TABLE 3. Scoring Systems ASES Constant Rowe UCLA Pre Post Pre Post Pre Post Pre Post Score SD NOTE. All postoperative scores significant, P Abbreviations: Pre, preoperative; Post, postoperative; SD, standard deviation.

7 242 G. M. GARTSMAN ET AL. Analysis of Unsatisfactory Results One patient was rated as fair to poor (UCLA score 29 or Rowe score 70) at final evaluation. The patient (with a preoperative diagnosis of recurrent subluxation) had recurrent subluxation after arthroscopic repair. The time to failure was 18 months. The patient underwent revision surgery where we noted that the inferior capsular repair was intact but there was a disruption of the superior portion of the Bankart repair. This was repaired with suture anchors. Thereafter, the patient had a stable shoulder with full return to sports activities. Complications No major intraoperative or perioperative complications (permanent nerve injuries, wound infections) occurred. One patient noted paresthesias in the musculocutaneous nerve distribution that resolved by the 6-week postoperative visit. We did not observe any complications from suture anchors. Analysis of Missing Data Adequate follow-up was not available for 4 patients who were therefore excluded. However, demographic data and operative findings showed these 4 patients were representative of the entire study group population. Two patients did not return for a 1-year evaluation. Two other patients were examined 1 year after surgery but were not available for the 2-year minimum. DISCUSSION In the past decade, orthopaedic surgeons have witnessed an increased interest in arthroscopic repair of glenohumeral instability. While some have reported satisfactory results, 1-3,5,25-28 others have presented a less favorable experience. 29,30 There are few articles dealing specifically with multidirectional instability. Neer and Foster 31 reported their experience with an open technique and, recently, McIntyre et al. 1 reported their results with arthroscopic treatment in 19 patients. Their patients scored 91 out of a possible 100 points at final evaluation. Eighteen patients returned to athletics and, in 12 of 16 patients examined at final evaluation, there was no loss of motion and 3 had less than 5 loss of external rotation. Neither the Constant nor the ASES scoring system specifically describes what ratings are excellent or poor; Ellman et al. 16 categorized UCLA scores of 29 to 35 as good to excellent results and those less than 29 as fair and poor results. Rowe and Zarins 9 rated scores of 90 to 100 as excellent and 75 to 89 as good. The present study reports overall good and excellent results in 47 of 48 (97%) shoulders according to both the Rowe and UCLA scores. One patient had persistent instability at final evaluation. Rotator interval repair is not emphasized in reports of arthroscopic treatment but both Neer and Foster 31 and Rowe and Zarins 10 described repair of the rotator interval in their descriptions of open surgery. We also found that repair of the rotator interval was an essential element of the operative treatment. A number of weaknesses exist in the present study. The investigation was prospective, but we used neither patient randomization nor investigator masking. Although each patient had the clinical diagnosis of multidirectional instability based on patient history, physical examination, examination under anesthesia, and observation at arthroscopy, the spectrum of glenohumeral lesions indicates that the etiology of multidirectional instability is both multifactorial and variable. Our data suggest that multidirectional instability may represent the clinical expression of a variety of anatomic lesions. With the numbers available in the present study, we could not detect correlation in regard to the anatomic findings and the patient history or physical examination. In addition, the follow-up period was relatively short and we continue to study these patients so that we can evaluate the results over a longer interval. We also believe that these results may deteriorate with time and may parallel the experience after open repair. 32 Although the patients in this study had the clinical diagnosis of multidirectional instability and a large capsule permitting excessive anterior, inferior, and posterior translation, the lesions responsible for the translation were variable. Anatomic repair of all labrum detachments and suture capsulorrhaphy to restore capsular tension produced successful results. REFERENCES 1. McIntyre LF, Caspari RB, Savoie FH. The arthroscopic treatment of posterior shoulder instability: Two-year results of multiple suture technique. Arthroscopy 1997;13: McIntyre LF, Caspari RB, Savoie FH. The arthroscopic treatment of multidirectional shoulder instability: Two-year results of a multiple suture technique. Arthroscopy 1997;13: Savoie FH, Miller CD, Field LD. Arthroscopic reconstruction of traumatic anterior instability of the shoulder: The Caspari technique. Arthroscopy 1997;13: Altchek DW, Warren RF, Skyhar MJ, Ortiz G. T-Plasty mod-

8 MULTIDIRECTIONAL GLENOHUMERAL INSTABILITY 243 ification of the Bankart procedure for multidirectional instability of the anterior and inferior types. J Bone Joint Surg Am 1991;73: Arciero RA. Arthroscopic bioabsorbable tack stabilization of initial anterior shoulder dislocations. Arthroscopy 1995;4: Bigliani LU, Kurzweil PR, Schwartzbach CC, Wolfe IN, Flatow EL. Inferior capsular shift procedure for anterior-inferior shoulder instability in athletes. Am J Sports Med 1994;22: Cofield RH, Kavanagh BF, Frassica FJ. Anterior shoulder instability. Instr Course Lect 1985;34: Hawkins RJ, Bell RM. Posterior instability of the shoulder. Instr Course Lect 1989;38: Rowe CR, Zarins B. The Bankart procedure: Long-term endresult study. J Bone Joint Surg Am 1978;60: Rowe CR, Zarins B. Recurrent transient subluxation of the shoulder. J Bone Joint Surg Am 1981;63: Cohen J, ed. Statistical power analysis for the behavioral sciences. Hillsdale, NJ: Lawrence Erlbaum, Gartsman GM. Arthroscopic acromioplasty for lesions of the rotator cuff. J Bone Joint Surg Am 1990;72: Hawkins RJ, Misamore GW, Hobeika PE. Surgery for fullthickness rotator-cuff tears. J Bone Joint Surg Am 1985;67: Misamore GW, Ziegler DW, Rushton JLN. Repair of the rotator cuff. A comparison of results in two populations of patients. J Bone Joint Surg Am 1995;77: Constant CR. A clinical method of functional assessment of the shoulder. Clin Orthop 1987;214: Ellman H, Hanker G, Bayer M. Repair of the rotator cuff: End-result study of factors influencing reconstruction. J Bone Joint Surg Am 1986;68: Richards RR, An K-N, Bigliani LU, et al. A standardized method for the assessment of shoulder function. J Shoulder Elbow Surg 1994;3: Ellman H, Gartsman GM, eds. Arthroscopic shoulder surgery and related procedures. Philadelphia: Lea & Febiger, Cooper NA, Brems JJ. The inferior capsular-shift procedure for multidirectional instability of the shoulder. J Bone Joint Surg Am 1992;74: Neer CS, Poppen NK. Supraspinatus outlet. Orthop Trans 1987;11: Burkhead WZ Jr, Rockwood CA Jr. Treatment of instability of the shoulder with an exercise program. J Bone Joint Surg Am 1992;74: Harryman DT, Sidles JA, Harris SL, Lippitt SB, Matsen FA. The effect of articular conformity and the size of the humeral head component on laxity and motion after glenohumeral arthroplasty. J Bone Joint Surg Am 1995;77: Gartsman GM, Taverna E, Hammerman SM. Arthroscopic rotator interval repair in glenohumeral instability: Description of an operative technique. Arthroscopy 1999;15: Huck SW, Cormier WH, eds. Reading Statistics and Research. New York: Harper Collins, Arciero RA, Wheeler JH, Ryan JB, McBride JT. Arthroscopic Bankart repair versus nonoperative treatment for acute, initial anterior shoulder dislocations. Am J Sports Med 1994; 22: Morgan CD, Bodenstab AB. Arthroscopic Bankart suture repair: Technique and early results. Arthroscopy 1987;3: Snyder SJ. Shoulder instability. In: Snyder SJ. Shoulder arthroscopy. New York: McGraw Hill, 1994; Wolf EM, Wilk RM, Richmond JC. Arthroscopic Bankart repair using suture anchors. Oper Tech Orthop 1991;1: Grana WH. Arthroscopic Bankart suture repair. Am J Sports Med 1993;21: Walch G, Boileau P, Levigne C, Mandrino A, Neyret P, Donell S. Arthroscopic stabilization for recurrent anterior shoulder dislocation: Results of 59 cases. Arthroscopy 1995;11: Neer CS, Foster CR. Inferior capsular shift for involuntary inferior and multidirectional instability of the shoulder. J Bone Joint Surg Am 1980;62: Morrey BF, Janes JM. Recurrent anterior dislocation of the shoulder. J Bone Joint Surg Am 1976;58: