- Email: [email protected]
Reasons for failure after surgical repair of anterior shoulder instability
Reasons for failure after surgical repair of anterior shoulder instability Mark Tauber, MD,a Herbert Resch, MD,a Rosemarie Forstner, MD,b Michael Raffl, MD,a and Josef Schauer, MD,a Salzburg, Austria
A total of 41 patients presenting with recurrent anterior instability of the shoulder after surgical repair were followed up after a mean period of 49 months (range, 24-81 months). The failed procedures were arthroscopic Bankart repair in 25 cases, open Bankart repair in 6 cases, Eden-Hybinette procedure in 4 cases, rotational osteotomy in 2 cases, capsular T- shift operation in 1 case, Bristow-Latarjet in one case, and a J-bone graft procedure in one case. In one case the index procedure was unknown. At revision surgery, the findings were a defect of the anterior bony glenoid rim in 23 patients (56%), a large capsule in 9 (22%), and a laterally torn capsule in 2 (5%). In 7 patients (17%) a typical Bankart lesion with good capsule quality was found. At revision surgery, these lesions were addressed by a bone graft procedure in 21 cases and fixation of the rim fragment with screws in 2 cases. In the 9 patients with a large capsule, a T-shift operation was performed in 6 and a Bankart repair with capsulorrhaphy was performed in the remaining 3. In the 7 patients with a typical Bankart lesion, a Bankart repair was performed, and in the 2 patients with a laterally torn capsule, an open suturing technique was used. At follow-up, none of the patients had had further redislocation or subluxation. The Rowe score was excellent in 81% of the cases and good in 19%. In 19 patients (46%) no increase in arthritic change was detected on radiographic assessment at follow-up, whereas 13 (32%) showed an increase of 1 degree and 4 (10%) showed an increase of 2 degrees. The results show that good and very good outcomes can be achieved with surgical repair provided that the basic pathology of the unstable shoulder is taken into account. (J Shoulder Elbow Surg 2004;13:279 – 85.) From the Departments of Traumatology,a and Deparment of Radiology,b General Hospital Salzburg. Reprint requests: Herbert Resch, MD, Landeskliniken Salzburg, Mu¨llner Hauptstrasse 48, 5020 Salzburg, Austria (E-mail: [email protected]). Copyright © 2004 by Journal of Shoulder and Elbow Surgery Board of Trustees. 1058-2746/2004/$30.00 doi:10.1016/j.jse.2004.01.008
S
everal methods have been described for the surgical treatment of recurrent anterior instability of the shoulder. There is general acceptance of those methods that target surgical repair of the Bankart lesion (ie, capsulolabral detachment from the anterior glenoid rim). In addition to the traditional methods of open repair,1,10,11,15,17,25,27 in recent years more emphasis has been placed on arthroscopic techniques.5,8,12,13,16,18,19,22,29 Recurrence rates of 2% to 8% can be expected with open surgical stabilization techniques.7, 15,17,23,27 The recurrence rate is higher for arthroscopic stabilization depending on the procedure used.8,9,13,18,22 The fact that the recurrence rate is so low, especially for open surgery, makes it all the more interesting to investigate the causes of failure of surgical stabilization. After surgery, redislocation can occur as a result of the application of high levels of external force. However, cases in which recurrence occurs without the involvement of any significant external force are more deserving of attention. In such cases one must consider the quality of the original clinical diagnosis of instability, as well as the misinterpretation or failure to appreciate lesions in the humeral head, glenoid cavity, or soft tissues of the shoulder joint. Burkhart and De Beer3 recently drew attention to the significance of glenoid rim fractures. The question that still remains to be answered, however, is how big the bony defect of the glenoid rim must be to be taken into account for surgical repair. A similar question relates to the role of the size of a Hill-Sachs lesion. A number of authors are of the opinion that major lesions must be taken into account at surgery.3,28 There is consensus on the importance of an intact capsule for the stability of the shoulder joint. According to Rowe et al26 and experimental studies performed by Bigliani et al,2 the capsule is always overstretched in shoulder instability, and capsular redundancy must be addressed at surgery. In cases of capsular laxity Neer and Foster17 and other authors26 have recommended a capsular shift combined with closure of the rotator interval to reduce capsular volume. The purpose of this study was to answer the question of the extent to which the clinical diagnosis and
279
280
Tauber et al
J Shoulder Elbow Surg May/June 2004
Table I Procedures at initial surgery Procedure Arthroscopic Bankart Open Bankart Eden-Hybinette Weber rotational osteotomy J-bone graft Bristow-Latarjet Capsular T-shift Unknown
Table II Procedures at second surgery No. 25 6 4 2 1 1 1 1
Procedure Open Bankart J-bone graft Capsular T-shift Arthroscopic Bankart Weber rotational osteotomy Eden-Hybinette Open reduction and fragment fixation
No. 14 13 7 4 1 1 1
Table III Procedures at third surgery, (8 cases)
the shape, size, and extension of these lesions played a role in the development of failure after surgical repair. The investigation focused in particular on those cases in which postoperative redislocation occurred without any significant trauma. That made it necessary to distinguish between traumatic and non traumatic groups. Determining the exact cause of postoperative recurrence is important not only for the stabilization of the shoulder involved but also for higher success rates at initial surgery. MATERIAL AND METHODS From 1993 to 1999, a total of 47 patients underwent surgery at our institution for failed surgical repair of recurrent anterior instability of the shoulder. Of the 47 patients, 41 were followed up for a mean of 49 months (range, 24-81 months), with personal clinical and radiologic evaluation in 36 patients; the remaining 5 were interviewed by telephone. The follow-up rate was 87%. Of these 41 patients, 32 (78%) were male and 9 (22%) were female. In 23 cases (56%) the right shoulder was involved, and in 18 cases (44%) the left shoulder was involved. In 24 patients (59%) the dominant shoulder was affected. Before the patients underwent reoperation, we performed a detailed evaluation of their history with special regard to the number and type of previous operative procedures.
Number and type of preceding surgery There were a mean of 12 dislocations (range, 3-50 dislocations) before the initial surgical repair was undertaken. The method of surgical repair is shown in Table I. The mean age of the patients at this time was 26.1 years (range, 17-52 years). The surgical procedures used for revision are shown in Table II. In 8 patients a further redislocation occurred after the second operation, and operative stabilization was required for a third time. The procedures used in those cases are listed in Table III. In one patient, the third operation was followed again by redislocation that had no significant traumatic cause, and a fourth repair was performed. In summary, before we performed revision surgery, 32 patients had undergone one previous surgical intervention, 8 patients had undergone two, and 1 patient had undergone three.
Procedure
No.
J-bone graft Eden-Hybinette and Weber rotational osteotomy Capsular T-shift Open Bankart
5 1 1 1
Cause and number of dislocations In 35 patients (85%) the initial dislocation was caused by a traumatic event, with some form of sport involved in 23 cases (snowboarding, 6, skiing, 4, soccer, 4, cycling, 2, rock climbing, 2, basketball, 2, ice skating, 1, judo, 1, tennis, 1). In one patient the initial dislocation was the result of a grand mal episode. In the remaining 6 patients (15%), the initial dislocation had no significant traumatic cause, these were diagnosed as unidirectional recurrent anterior instability. The mean age of all patients at revision surgery was 30.7 years (range, 19-54 years). The mean period of time between initial surgery and revision surgery was 10.6 years. Redislocation of the shoulder occurred a mean of 31.8 months (range, 2-180 months) after initial surgery. In 24 cases (59%) redislocation after initial surgery was caused by a traumatic event, with sports injuries again being the main factor involved (skiing, 6, soccer, 5, snowboarding, 2, judo, 1, boxing, 1, water skiing, 1, surfing, 1). In the case of the patient whose initial dislocation was caused by a grand mal episode, the postoperative recurrence had the same etiology. In 19 cases (46%) postoperative redislocation occurred with no significant traumatic cause. Most of these patients said that a clumsy movement had caused the dislocation.
Intraoperative findings Intraoperative findings during revision surgery relating to the size, shape, and extension of the Bankart lesion and the presence of a bony fragment were noted. The bony fragments were graded on the basis of size (thickness): less than 2 mm, 2 to 5 mm, and more than 5 mm. Loss of convexity distal to the glenoid notch was considered a major defect with a thickness of more than 5 mm. In addition, the bony glenoid rim was classified as eroded or non-eroded. In the case of Bankart lesions, a distinction was made between fraying of a labrum that was still attached, a detached labrum, absence of a labrum, or sacular detachment of the capsule from the scapular neck.
Tauber et al
J Shoulder Elbow Surg Volume 13, Number 3
281
Clinical follow-up Clinical follow-up involved assessment of the stability of the glenohumeral joint on the basis of the apprehension and relocation tests. Load-and-shift tests were also performed on all patients to identify any laxity.6 Range-of-motion testing was based on flexion, abduction, internal rotation, and external rotation, with external rotation measured at both 0° and 90° abduction. Strength testing comprised (1) internal rotation strength with the arm at maximum external rotation and adduction and (2) elevation strength of the extended arm in 90° abduction in the scapular plane, with the force measured at the wrist by use of a spring balance. All tests and clinical examinations were also performed on the contralateral side. The clinical results were evaluated with the Rowe score.24 Subjective satisfaction was assessed on a scale ranging from 0 (dissatisfied) to 5 (very satisfied). Pain was assessed with a visual analog scale ranging from 0 (severe pain) to 10 (zero pain). Information was also elicited about return to work and sports. Depending on the degree of risk involved, a distinction was made between overhead sports (tennis, handball, basketball, volleyball, squash), sports with high shoulder involvement (skiing, snowboarding, swimming, ice hockey, soccer, golf), and sports with low shoulder involvement (running, cycling). A separate classification was used for contact sports such as soccer, handball, basketball, and ice hockey. Three categories were used for the frequency of the sporting activity involved: infrequent (once a week), frequent (2-4 times a week), and competitive level (⬎4 times a week). With regard to employment, a distinction was made between physical and nonphysical work. For statistical analysis, the paired t test was used.
Figure 1 J-bone graft procedure: A 10-mm-deep slot is chiseled 5 mm away from the cartilage; the J-bone graft harvested from the iliac crest is inserted by means of a pointed impactor.
1. Normal: no narrowing of the joint space, no osteophytes, and no sclerosis 2. Slight: minor irregularity of the joint cavity, narrowing of the joint space by less than 1 mm, and slight sclerosis 3. Moderate: narrowing of the joint space by less than 2 mm, minor osteophyte formation, and moderate sclerosis 4. Severe: clear narrowing of the joint space, pronounced presence of osteophytes, severe sclerosis, and formation of cysts.
shape. Attention is paid to the fact that the bone block does not protrude onto the curvature of the glenoid in order to avoid the possibility of later arthritic change. Through an anterior skin incision, the subscapularis tendon is approached via the deltopectoral interval. The tendon is split in a longitudinal direction on the transition zone between the middle and inferior third of the tendon. The capsule is exposed and detached from the undersurface of the tendon as far as possible, and a T-shaped capsular incision is made. The capsule together with the destroyed labrum and the rest of the fragment is detached from the bony rim and the neighboring scapular neck. The area of the defect is abraded with a burr. Parallel to the cartilage-bone border and about 5 mm away from it, a slot approximately 20 mm in length is chiseled in the craniocaudal direction. The chisel is directed posteromedially away from the plane of the glenoid surface to avoid any danger of fracture (Figure 1). The depth of the slot is 10 to 12 mm. A corticocancellous bone block about 15 to 15 mm in size is harvested from the ipsilateral iliac crest. The bone block consists of the outer cortex and the cranial cortex of the iliac crest. With the oscillating saw, the bone block is now modeled into a J shape, with a thin part consisting of the cortical bone only and a thicker part, with cancellous bone beneath the cortex. The bone block with the thinner part, which is sharpened at the end, is hammered into the slot by means of an impactor until the cancellous bone of the second part is pressed firmly against the rim. With a high- speed burr, the block is adapted precisely to the cartilage (Figure 2). The bone block should be precisely confluent with the curvature of the cartilage and should not protrude. The capsule is closed above the inserted bone by end-to-end suturing without capsulorrhaphy and without attaching it to the graft.
Technique of J-bone graft procedure
RESULTS
As the technique of the J-bone graft procedure has only been published in the German literature,20,21 it is described here in brief. The only indication for this technique is a chronic defect of the anterior glenoid rim caused by fracture. The goal is to reconstruct the glenoid rim in its size and
In 21 patients (51%) a bony Bankart lesion that was more than 2 mm thick was found. In 2 other patients (5%) the bony anterior glenoid rim was eroded, although no fragment was present (having
Radiologic follow-up In all patients who attended the follow-up in person (36 cases), anterior-posterior and axillary radiographs of the shoulder joint were obtained. The radiographs were assessed for arthritic change and compared with the original radiographs taken before the index stabilization procedure. The radiographs were evaluated by an independent radiologist (R.F.). Arthritic change was assessed on the basis of the classification proposed by Rosenberg et al23:
282
Tauber et al
Figure 2 Once the graft is inserted, it is modeled by use of a high-speed burr to continue precisely the curvature of the cartilage of the glenoid; the capsule is closed end to end without attachment to the graft.
probably been removed at earlier surgery or absorbed). In all 23 patients the glenoid had shown reduced or zero convexity at the anteroinferior margin below the glenoid notch, representing a significant reduction in diameter in the transverse plane. The detached fragment could only be used for screw fixation in 2 patients. In 19 patients a J-bone graft procedure was used, and in 2 cases the graft was fixed with two screws. In one patient, in whom a rotational osteotomy28 had been performed at earlier surgery, an osteotomy of the humeral head was done, and it was derotated back to normal. All other patients had soft-tissue defects only. In 9 patients the capsule was described as large. In 5 of these, the clinical examination had revealed a positive sulcus sign and the shoulder was described as lax. In the remaining 4 cases, the capsule looked overstretched but there was no clinical sign of laxity. In 6 of the 9 patients a capsular T-shift operation according to Neer and Foster17 was performed, and in 3 cases open Bankart repair with additional capsulorraphy was performed. In 7 cases the capsulolabral complex was detached from the anterior glenoid rim. The capsule was of good quality, and a Bankart repair was therefore performed. In 3 patients this was performed arthroscopically. In 2 other cases the capsule was torn on the lateral side. This was recognized during arthroscopic examination. The capsule was repaired by an open technique. After a mean follow-up of 49 months (range, 24-81 months), none of the patients had had a recurrent dislocation or subluxation. In 5 patients (12%) the apprehension test was positive. However, only 1 of these patients remained symptomatic with a transient
J Shoulder Elbow Surg May/June 2004
feeling of instability when performing overhead activities. The Rowe score was obtained for 37 patients. The mean score was 95.5 points (range, 75-100 points). Of 37 patients, 30 (81%) had an excellent result and 7 (19%) had a good result. None of the patients had a moderate or poor result. The sulcus sign as a clinical indicator of a large joint capsule was positive preoperatively in 5 cases. In all 5 cases it was present in both shoulders. Intraoperatively, all 5 patients were found to have a large joint capsule, and in all of these cases a capsular T-shift operation was carried out (see above). Compared with the contralateral shoulder, the average loss of motion for the operated shoulder was 5.4° of flexion (0°-25°) and 2.1° of abduction (0°20°). External rotation at 0° abduction was on average 5° (0°-25°) less than that in the contralateral shoulder. External rotation at 90° abduction was on average 5.5° (0°-30°) less than that in the contralateral shoulder. Internal rotation was not restricted in 18 patients (47%)—that is, they could move the back of the hand to the interscapular position; 16 patients (45%) could reach the thoracic spine, and 3 patients (8%) could reach the lumbar spine. The mean loss of internal rotation strength in comparison with the contralateral shoulder was 0.55 kg (range, 0-4 kg). In 10 patients (24%), internal rotation strength was higher on the operated side. The mean loss of strength at 90° abduction was 0.6 kg (range, 0-4 kg). Four patients (ten percent) had more strength on the operated side. The 8 patients who had undergone three surgical repairs for shoulder instability had a mean loss of internal rotation strength of 0.63 kg with the arm in adduction and 1.1 kg at 90° abduction. There was no significant difference between operated and nonoperated shoulder (P ⬎ .05). The mean time to resumption of sporting activity was 3.6 months (range, 1-6 months). Of the 41 patients, 31 (76%) returned to the performance level that they had reached before surgery. Two patients (five percent) were actually able to improve further following revision surgery. Eight patients (twenty percent) were unable to achieve their preoperative levels. Four of these explained the problem in terms of psychological barriers. With one exception, all 8 patients who had performed some form of sport at a competitive level preoperatively (ie, trained at least 5 times a week) were able to regain their preoperative level of performance. Of the 41 patients, 21 had originally performed nonphysical work and 20 had performed physical work. The patients with nonphysical duties returned to work after a mean of 1.3 months (range, 0.25-2 months), and those employed in physical work returned after a mean of 2.4 months (range, 1-4
Tauber et al
J Shoulder Elbow Surg Volume 13, Number 3
283
Table IV Procedures at initial surgery and frequency of subsequent traumatic and atraumatic redislocations
Total No.
Traumatic redislocations
Atraumatic redislocations
Atraumatic redislocations (%)
25 6 4 2 1 1 1 1
15 4 2 2 1 0 0 0
10 2 2 0 0 1 1 1
40 33.3 50 0 0 100 100 100
Arthroscopic Bankart Open Bankart Eden-Hybinette Weber J-bone graft Capsular shift Bristow-Latarjet Unknown
months). None of the patients had to change his or her job. The mean score for reported pain on the 10-point scale was 9.5 points (range, 5-10 points). Of the 41 patients, 33 (81%) reported complete pain relief, 5 complained of an occasional slight hypersensitivity to changes in the weather, and 3 reported pain during continuous overhead activities. Mean patient satisfaction on the 5-point scale was 4.7 points (range, 1- 5 points). Of the 41 patients, 33 (81%) were very satisfied, 7 (17%) were satisfied, and 1 (2%) was fairly satisfied. Analysis of atraumatic failures
In 19 patients (46%) the first redislocation after surgical repair occurred without significant trauma (atraumatic failures). Six had previously undergone surgery on two occasions. In the majority of cases a Bankart repair had been performed, with an arthroscopic procedure selected in 7 cases and an open procedure in 6. Bone grafting procedures were performed in two patients, a capsule shift operation in one, a rotational osteotomy in one, and Bristow-Latarjet in one (Table IV). In the remaining case the applied technique of the index procedure was unknown. Intraoperatively, at revision surgery, a large bony Bankart lesion was found in 8 patients (42%) and a severely eroded glenoid rim in 1 (5%)—that is, bony deficiencies were present in approximately half of the cases. Four patients (twenty-one percent) had a large joint capsule. In the other cases typical Bankart lesions or extensive capsular detachment from the glenoid rim was found. Radiographic examination
At revision surgery, only 14 patients (34%) had no signs of arthritis. At follow-up, only 5 (12%) remained with no arthritic changes on radiographic analysis. In 19 patients no increase in arthritis was seen, whereas the arthritic change progressed by one degree in 13 and by two degrees in 4.
Of the 19 patients without an increase in arthritic change, 8 underwent a bone graft procedure and 11 underwent a soft-tissue repair. Of the 13 patients whose arthritis changed by one degree, 6 had been stabilized by J-bone graft and 7 by a soft-tissue repair. There was no significant difference between bone graft procedures and soft-tissue repairs in the progression of arthritis. Postoperative complications
Postoperative complications occurred in 3 patients. Out of a total of 23 patients in whom bone graft was harvested from the iliac crest, hypesthesia was reported postoperatively in the area of the iliac crest in 2 patients. One patient had a postoperative infection. All of these patients had a successful resolution of the complication. DISCUSSION Knowledge of the causes of redislocation after surgery for shoulder stabilization is important not only for the success of revision surgery but also for improved results at initial surgery. According to Levine et al,14 the most frequent causes of unsuccessful shoulder stabilization relate to the failure to correct an excessively large anterior-inferior capsular pouch and detached capsulolabral complex. They found the best results in revision surgery in those cases where redislocation was due to a traumatic event after just one surgical repair and in the absence of voluntary instability. Burkhart and De Beer3 stress the importance of reconstructing bony defects during arthroscopic procedures. According to their findings, the arthroscopic Bankart repair is capable of delivering the same results as an open Bankart procedure in the absence of significant bony defects. Even though these findings relate to arthroscopic procedures for initial surgery, great importance is attached to an unchanged bony glenoid rim for the stability of the glenohumeral joint. Half of all our patients requiring revision surgery
284
Tauber et al
had a bony Bankart defect extending to the anteriorinferior portion straight downward from the glenoid notch. In all of these cases the bony defect was reconstructed by bone grafting, and all shoulders remained stable. This would seem to confirm the importance of an uncompromised or reconstructed anterior glenoid rim for the stability of the shoulder joint and good function. At revision surgery, we found that the size of the remaining fragment did not correlate with the size of the glenoid defect. The absence of loading on the fragment may lead to atrophy and partial absorption. Therefore, the fragment could only be used to repair the glenoid in a few cases. The incidence of bony glenoid lesions in our series was much higher than that reported by other authors.3,24 However, this series presents a selected group of patients after failed stabilization surgery. Another reason for the higher incidence might be the fact that a careful intraoperative examination with special regard to the condition of the bony rim was performed at the time of revision surgery. An incidence of approximately 50% of bony glenoid rim defects in our revision cases demonstrates that a nonreconstructed bony glenoid rim plays a major role in the development of a postoperative failure. On the basis of this outcome, we believe that a defect of the bony glenoid rim thicker than just the cortex should be reconstructed by bone grafting at surgery. The size of the joint capsule and its importance for the stability of the shoulder joint have been stressed by several authors.4,17,25,26,30 They are of the opinion that the capsule is overstretched by the dislocations and, therefore, has to be shortened at the time of surgery.17,25 This has to be taken into consideration in case of revision surgery. Particular attention has to be given to the lax shoulder. In these cases the capsule was voluminous even before dislocation had occurred and was quite often of poor quality. In our series 5 patients have shown laxity at clinical examination. In none of them was this taken into account at the primary surgery. Only the capsule T-shift operation that we have used in these patients has resulted in permanent stability. In this study the most frequent cause of redislocation after initial surgery was a further traumatic episode, with sports injuries as the most frequent cause. Our study focused in particular, however, on patients who had redislocation without a traumatic cause (i.e, where surgical repair seemed to have failed to achieve stability). In such cases it is reasonable to assume that the repair failed because of poor technique, an inappropriate choice of procedure, failure to take the presence of lesions into account, or simply an incorrect diagnosis. Of the 19 patients who had redislocation without any external force, 8 (42%) had a glenoid rim frag-
J Shoulder Elbow Surg May/June 2004
ment of more than 2 mm that had been ignored at the index procedure. Another 5 patients (26%) from this group had a clearly overstretched capsule that had not been taken into account at previous surgery. In summary, in 13 of 19 cases (68%) a clear cause for the failure of the previous surgery could be found. A large proportion of the revisions we performed related to arthroscopic stabilization procedures. Of 25 patients treated arthroscopically, there were 10 in whom the labrum had not healed onto the glenoid rim despite the absence of bony defects. With such findings, redislocations are due primarily to problems of technique rather than a failure to address lesions. The Hill-Sachs lesion did not play a major role in the preoperative planning of the revision surgery in this series. On the basis of our good results in terms of both stability and function, we would conclude that the importance of the Hill-Sachs lesion is minor compared with the Bankart lesion and that the Hill-Sachs lesion may play a role only in the rare case of a huge defect. As the radiographic evaluations show, the progression of arthritic changes cannot be stopped in all cases despite successful stabilization; at best, it can only be slowed down. Before the last revision surgery, 34% of the patients had no radiographic signs of arthritis, whereas at follow-up 4 years later, only 5% were without arthritic changes. There was no difference between bone graft procedures and soft-tissue repairs. In this respect it has to be emphasized that the bone graft addressed only the size of the defect and did not alter the congruity of the glenoid surface. Given the very good functional results, a restricted range of motion as a factor in arthritic progression presumably can be discounted. It would seem to be a process that has already been triggered and cannot really be stopped by surgical repair procedures, at least not in all cases. These results show that surgical repair that takes full account of the clinical and intraoperative findings leads to very good results in terms of both stability and function. Had due attention been paid to these aspects at initial surgery, this series of revisions would not have been necessary. REFERENCES
1. Altchek DW, Warren RF, Skyhar MJ, et al. T-plasty modification of the Bankart procedure for multidirectional instability of the anterior and inferior types. J Bone Joint Surg Am 1991;73:10512. 2. Bigliani LU, Pollock RG, Soslowsky LJ, et al. Tensile properties of the inferior glenohumeral ligament. J Orthop Res 1992;10:18797. 3. Burkhart SS, De Beer JF. Traumatic glenohumeral bone defects and their relationship to failure of arthroscopic Bankart repairs: significance of the inverted-pear glenoid and the humeral engaging Hill-Sachs lesion. Arthroscopy 2000;16:677-94. 4. Burkhead WZ, Richie MF. Revision of failed shoulder reconstruction. Contemp Orthop 1992;24:126-33.
Tauber et al
J Shoulder Elbow Surg Volume 13, Number 3
5. Caspari RB. Arthroscopic reconstruction for anterior shoulder instability. Tech Orthop 1988;3:61-6. 6. Gerber C, Ganz R. Clinical assessment of instability of the shoulder. J Bone Joint Surg Br 1984;66:551-6. 7. Gill TJ, Micheli LJ, Gebhard F, et al. Bankart repair for anterior instability of the shoulder. J Bone Joint Surg Am 1997;79:850-7. 8. Hawkins RB. Arthroscopic stapling repair for shoulder instability: a retrospective study of 50 cases. Arthroscopy 1989;5:122-8. 9. Hayashida K, Yoneda M, Nakagawa S, et al. Arthroscopic Bankart suture repair for traumatic anterior shoulder instability: analysis of the causes of a redislocation. Arthroscopy 1998;14: 295-301. 10. Hovelius L, Akermark C, Albrektsson C, et al. Bristow-Latarjet procedure for recurrent anterior dislocation of the shoulder. Acta Orthop Scand 1983;54:284-90. 11. Jobe FW, Giangarra CE, Kvitne RS, et al. Anterior capsulo-labral reconstruction of the shoulder in athletes in overhead sports. Am J Sports Med 1991;19:428-34. 12. Johnson LL. Diagnostic and surgical arthroscopy of the shoulder. St Louis: Mosby; 1993. p. 276-352. 13. Landsiedl F. Arthroscopic therapy of recurrent anterior luxation of the shoulder by capsular repair. Arthroscopy 1992;8:296-304. 14. Levine WN, Arroyo JS, Pollock RG, et al. Open revision stabilization surgery for recurrent anterior glenohumeral instability. Am J Sports Med 2000;28:156-60. 15. Levine WN, Richmond JC, Donaldson WR. Use of the suture anchor in open Bankart reconstruction: a follow-up report. Am J Sports Med 1994;22:723-6. 16. Morgan CD, Bodenstab AB. Arthroscopic Bankart suture repair: technique and early results. Arthroscopy 1987;3:111-22. 17. Neer CS, Foster CR. Inferior capsular shift for involuntary inferior and multidirectional instability. J Bone Joint Surg Am 1980;62: 897-908. 18. Pagnani MJ, Warren RF, Altchek DW, et al. Arthroscopic shoul-
19. 20. 21. 22. 23. 24. 25. 26. 27.
28.
29. 30.
285
der stabilization using transglenoid sutures. A four-year minimal follow-up. Am J Sports Med 1996;24:459-67. Peterson CA, Newton SE. Arthroscopic Bankart repair. Am J Arthroscopy 1991;1:17-9. Resch H. The Bankart procedure. A modification using transosseus sutures. Orthop Traumatol 1993;2:19-28. Resch H. J-Span Plastik. In: Kremer K, Lierse W, Platzer W, et al, editors. Chirurgische Operationslehre, Schultergu¨rtel und obere Extremita¨t. Stuttgart: Thieme; 1995. p. 102-3. Resch H, Povacz P, Wambacher M, et al. Arthroscopic extraarticular Bankart repair for the treatment of recurrent anterior shoulder dislocation. Arthroscopy 1997;13:188-200. Rosenberg BN, Richmond JC, Levine WN. Long-term followup of Bankart reconstruction: incidence of late degenerative glenohumeral arthrosis. Am J Sports Med 1995;23:538-44. Rowe CR, Patel D, Southmayd WW. The Bankart procedure: a long-term end-result study. J Bone Joint Surg Am 1978;60:1-16. Rowe CR, Zarins B. Recurrent transient subluxation of the shoulder. J Bone Joint Surg Am 1981;63:863-72. Rowe CR, Zarins B, Ciullo JV. Recurrent anterior dislocation of the shoulder after surgical repair. J Bone Joint Surg Am 1984;66: 159-68. Thomas SC, Matsen FA. An approach to the repair of avulsion of the glenohumeral ligaments in the management of traumatic anterior glenohumeral instability. J Bone Joint Surg Am 1989;71: 506-13. Weber BG, Simpson LA, Hardegger F. Rotational humeral osteotomy for recurrent anterior dislocation of the shoulder associated with a large Hill-Sachs lesion. J Bone Joint Surg Am 1984; 66:1443-6. Wolf EM, Wilke RM, Richmond JC. Arthroscopic Bankart repair using suture anchors [abstract]. Oper Tech Orthop 1991;1:184. Zabinski ST, Callaway GH, Cohen S, et al. Revision shoulder stabilization: 2- to 10-year results. J Shoulder Elbow Surg 1999; 8:58-65.
A total of 41 patients presenting with recurrent anterior instability of the shoulder after surgical repair were followed up after a mean period of 49 months (range, 24-81 months). The failed procedures were arthroscopic Bankart repair in 25 cases, open Bankart repair in 6 cases, Eden-Hybinette procedure in 4 cases, rotational osteotomy in 2 cases, capsular T- shift operation in 1 case, Bristow-Latarjet in one case, and a J-bone graft procedure in one case. In one case the index procedure was unknown. At revision surgery, the findings were a defect of the anterior bony glenoid rim in 23 patients (56%), a large capsule in 9 (22%), and a laterally torn capsule in 2 (5%). In 7 patients (17%) a typical Bankart lesion with good capsule quality was found. At revision surgery, these lesions were addressed by a bone graft procedure in 21 cases and fixation of the rim fragment with screws in 2 cases. In the 9 patients with a large capsule, a T-shift operation was performed in 6 and a Bankart repair with capsulorrhaphy was performed in the remaining 3. In the 7 patients with a typical Bankart lesion, a Bankart repair was performed, and in the 2 patients with a laterally torn capsule, an open suturing technique was used. At follow-up, none of the patients had had further redislocation or subluxation. The Rowe score was excellent in 81% of the cases and good in 19%. In 19 patients (46%) no increase in arthritic change was detected on radiographic assessment at follow-up, whereas 13 (32%) showed an increase of 1 degree and 4 (10%) showed an increase of 2 degrees. The results show that good and very good outcomes can be achieved with surgical repair provided that the basic pathology of the unstable shoulder is taken into account. (J Shoulder Elbow Surg 2004;13:279 – 85.) From the Departments of Traumatology,a and Deparment of Radiology,b General Hospital Salzburg. Reprint requests: Herbert Resch, MD, Landeskliniken Salzburg, Mu¨llner Hauptstrasse 48, 5020 Salzburg, Austria (E-mail: [email protected]). Copyright © 2004 by Journal of Shoulder and Elbow Surgery Board of Trustees. 1058-2746/2004/$30.00 doi:10.1016/j.jse.2004.01.008
S
everal methods have been described for the surgical treatment of recurrent anterior instability of the shoulder. There is general acceptance of those methods that target surgical repair of the Bankart lesion (ie, capsulolabral detachment from the anterior glenoid rim). In addition to the traditional methods of open repair,1,10,11,15,17,25,27 in recent years more emphasis has been placed on arthroscopic techniques.5,8,12,13,16,18,19,22,29 Recurrence rates of 2% to 8% can be expected with open surgical stabilization techniques.7, 15,17,23,27 The recurrence rate is higher for arthroscopic stabilization depending on the procedure used.8,9,13,18,22 The fact that the recurrence rate is so low, especially for open surgery, makes it all the more interesting to investigate the causes of failure of surgical stabilization. After surgery, redislocation can occur as a result of the application of high levels of external force. However, cases in which recurrence occurs without the involvement of any significant external force are more deserving of attention. In such cases one must consider the quality of the original clinical diagnosis of instability, as well as the misinterpretation or failure to appreciate lesions in the humeral head, glenoid cavity, or soft tissues of the shoulder joint. Burkhart and De Beer3 recently drew attention to the significance of glenoid rim fractures. The question that still remains to be answered, however, is how big the bony defect of the glenoid rim must be to be taken into account for surgical repair. A similar question relates to the role of the size of a Hill-Sachs lesion. A number of authors are of the opinion that major lesions must be taken into account at surgery.3,28 There is consensus on the importance of an intact capsule for the stability of the shoulder joint. According to Rowe et al26 and experimental studies performed by Bigliani et al,2 the capsule is always overstretched in shoulder instability, and capsular redundancy must be addressed at surgery. In cases of capsular laxity Neer and Foster17 and other authors26 have recommended a capsular shift combined with closure of the rotator interval to reduce capsular volume. The purpose of this study was to answer the question of the extent to which the clinical diagnosis and
279
280
Tauber et al
J Shoulder Elbow Surg May/June 2004
Table I Procedures at initial surgery Procedure Arthroscopic Bankart Open Bankart Eden-Hybinette Weber rotational osteotomy J-bone graft Bristow-Latarjet Capsular T-shift Unknown
Table II Procedures at second surgery No. 25 6 4 2 1 1 1 1
Procedure Open Bankart J-bone graft Capsular T-shift Arthroscopic Bankart Weber rotational osteotomy Eden-Hybinette Open reduction and fragment fixation
No. 14 13 7 4 1 1 1
Table III Procedures at third surgery, (8 cases)
the shape, size, and extension of these lesions played a role in the development of failure after surgical repair. The investigation focused in particular on those cases in which postoperative redislocation occurred without any significant trauma. That made it necessary to distinguish between traumatic and non traumatic groups. Determining the exact cause of postoperative recurrence is important not only for the stabilization of the shoulder involved but also for higher success rates at initial surgery. MATERIAL AND METHODS From 1993 to 1999, a total of 47 patients underwent surgery at our institution for failed surgical repair of recurrent anterior instability of the shoulder. Of the 47 patients, 41 were followed up for a mean of 49 months (range, 24-81 months), with personal clinical and radiologic evaluation in 36 patients; the remaining 5 were interviewed by telephone. The follow-up rate was 87%. Of these 41 patients, 32 (78%) were male and 9 (22%) were female. In 23 cases (56%) the right shoulder was involved, and in 18 cases (44%) the left shoulder was involved. In 24 patients (59%) the dominant shoulder was affected. Before the patients underwent reoperation, we performed a detailed evaluation of their history with special regard to the number and type of previous operative procedures.
Number and type of preceding surgery There were a mean of 12 dislocations (range, 3-50 dislocations) before the initial surgical repair was undertaken. The method of surgical repair is shown in Table I. The mean age of the patients at this time was 26.1 years (range, 17-52 years). The surgical procedures used for revision are shown in Table II. In 8 patients a further redislocation occurred after the second operation, and operative stabilization was required for a third time. The procedures used in those cases are listed in Table III. In one patient, the third operation was followed again by redislocation that had no significant traumatic cause, and a fourth repair was performed. In summary, before we performed revision surgery, 32 patients had undergone one previous surgical intervention, 8 patients had undergone two, and 1 patient had undergone three.
Procedure
No.
J-bone graft Eden-Hybinette and Weber rotational osteotomy Capsular T-shift Open Bankart
5 1 1 1
Cause and number of dislocations In 35 patients (85%) the initial dislocation was caused by a traumatic event, with some form of sport involved in 23 cases (snowboarding, 6, skiing, 4, soccer, 4, cycling, 2, rock climbing, 2, basketball, 2, ice skating, 1, judo, 1, tennis, 1). In one patient the initial dislocation was the result of a grand mal episode. In the remaining 6 patients (15%), the initial dislocation had no significant traumatic cause, these were diagnosed as unidirectional recurrent anterior instability. The mean age of all patients at revision surgery was 30.7 years (range, 19-54 years). The mean period of time between initial surgery and revision surgery was 10.6 years. Redislocation of the shoulder occurred a mean of 31.8 months (range, 2-180 months) after initial surgery. In 24 cases (59%) redislocation after initial surgery was caused by a traumatic event, with sports injuries again being the main factor involved (skiing, 6, soccer, 5, snowboarding, 2, judo, 1, boxing, 1, water skiing, 1, surfing, 1). In the case of the patient whose initial dislocation was caused by a grand mal episode, the postoperative recurrence had the same etiology. In 19 cases (46%) postoperative redislocation occurred with no significant traumatic cause. Most of these patients said that a clumsy movement had caused the dislocation.
Intraoperative findings Intraoperative findings during revision surgery relating to the size, shape, and extension of the Bankart lesion and the presence of a bony fragment were noted. The bony fragments were graded on the basis of size (thickness): less than 2 mm, 2 to 5 mm, and more than 5 mm. Loss of convexity distal to the glenoid notch was considered a major defect with a thickness of more than 5 mm. In addition, the bony glenoid rim was classified as eroded or non-eroded. In the case of Bankart lesions, a distinction was made between fraying of a labrum that was still attached, a detached labrum, absence of a labrum, or sacular detachment of the capsule from the scapular neck.
Tauber et al
J Shoulder Elbow Surg Volume 13, Number 3
281
Clinical follow-up Clinical follow-up involved assessment of the stability of the glenohumeral joint on the basis of the apprehension and relocation tests. Load-and-shift tests were also performed on all patients to identify any laxity.6 Range-of-motion testing was based on flexion, abduction, internal rotation, and external rotation, with external rotation measured at both 0° and 90° abduction. Strength testing comprised (1) internal rotation strength with the arm at maximum external rotation and adduction and (2) elevation strength of the extended arm in 90° abduction in the scapular plane, with the force measured at the wrist by use of a spring balance. All tests and clinical examinations were also performed on the contralateral side. The clinical results were evaluated with the Rowe score.24 Subjective satisfaction was assessed on a scale ranging from 0 (dissatisfied) to 5 (very satisfied). Pain was assessed with a visual analog scale ranging from 0 (severe pain) to 10 (zero pain). Information was also elicited about return to work and sports. Depending on the degree of risk involved, a distinction was made between overhead sports (tennis, handball, basketball, volleyball, squash), sports with high shoulder involvement (skiing, snowboarding, swimming, ice hockey, soccer, golf), and sports with low shoulder involvement (running, cycling). A separate classification was used for contact sports such as soccer, handball, basketball, and ice hockey. Three categories were used for the frequency of the sporting activity involved: infrequent (once a week), frequent (2-4 times a week), and competitive level (⬎4 times a week). With regard to employment, a distinction was made between physical and nonphysical work. For statistical analysis, the paired t test was used.
Figure 1 J-bone graft procedure: A 10-mm-deep slot is chiseled 5 mm away from the cartilage; the J-bone graft harvested from the iliac crest is inserted by means of a pointed impactor.
1. Normal: no narrowing of the joint space, no osteophytes, and no sclerosis 2. Slight: minor irregularity of the joint cavity, narrowing of the joint space by less than 1 mm, and slight sclerosis 3. Moderate: narrowing of the joint space by less than 2 mm, minor osteophyte formation, and moderate sclerosis 4. Severe: clear narrowing of the joint space, pronounced presence of osteophytes, severe sclerosis, and formation of cysts.
shape. Attention is paid to the fact that the bone block does not protrude onto the curvature of the glenoid in order to avoid the possibility of later arthritic change. Through an anterior skin incision, the subscapularis tendon is approached via the deltopectoral interval. The tendon is split in a longitudinal direction on the transition zone between the middle and inferior third of the tendon. The capsule is exposed and detached from the undersurface of the tendon as far as possible, and a T-shaped capsular incision is made. The capsule together with the destroyed labrum and the rest of the fragment is detached from the bony rim and the neighboring scapular neck. The area of the defect is abraded with a burr. Parallel to the cartilage-bone border and about 5 mm away from it, a slot approximately 20 mm in length is chiseled in the craniocaudal direction. The chisel is directed posteromedially away from the plane of the glenoid surface to avoid any danger of fracture (Figure 1). The depth of the slot is 10 to 12 mm. A corticocancellous bone block about 15 to 15 mm in size is harvested from the ipsilateral iliac crest. The bone block consists of the outer cortex and the cranial cortex of the iliac crest. With the oscillating saw, the bone block is now modeled into a J shape, with a thin part consisting of the cortical bone only and a thicker part, with cancellous bone beneath the cortex. The bone block with the thinner part, which is sharpened at the end, is hammered into the slot by means of an impactor until the cancellous bone of the second part is pressed firmly against the rim. With a high- speed burr, the block is adapted precisely to the cartilage (Figure 2). The bone block should be precisely confluent with the curvature of the cartilage and should not protrude. The capsule is closed above the inserted bone by end-to-end suturing without capsulorrhaphy and without attaching it to the graft.
Technique of J-bone graft procedure
RESULTS
As the technique of the J-bone graft procedure has only been published in the German literature,20,21 it is described here in brief. The only indication for this technique is a chronic defect of the anterior glenoid rim caused by fracture. The goal is to reconstruct the glenoid rim in its size and
In 21 patients (51%) a bony Bankart lesion that was more than 2 mm thick was found. In 2 other patients (5%) the bony anterior glenoid rim was eroded, although no fragment was present (having
Radiologic follow-up In all patients who attended the follow-up in person (36 cases), anterior-posterior and axillary radiographs of the shoulder joint were obtained. The radiographs were assessed for arthritic change and compared with the original radiographs taken before the index stabilization procedure. The radiographs were evaluated by an independent radiologist (R.F.). Arthritic change was assessed on the basis of the classification proposed by Rosenberg et al23:
282
Tauber et al
Figure 2 Once the graft is inserted, it is modeled by use of a high-speed burr to continue precisely the curvature of the cartilage of the glenoid; the capsule is closed end to end without attachment to the graft.
probably been removed at earlier surgery or absorbed). In all 23 patients the glenoid had shown reduced or zero convexity at the anteroinferior margin below the glenoid notch, representing a significant reduction in diameter in the transverse plane. The detached fragment could only be used for screw fixation in 2 patients. In 19 patients a J-bone graft procedure was used, and in 2 cases the graft was fixed with two screws. In one patient, in whom a rotational osteotomy28 had been performed at earlier surgery, an osteotomy of the humeral head was done, and it was derotated back to normal. All other patients had soft-tissue defects only. In 9 patients the capsule was described as large. In 5 of these, the clinical examination had revealed a positive sulcus sign and the shoulder was described as lax. In the remaining 4 cases, the capsule looked overstretched but there was no clinical sign of laxity. In 6 of the 9 patients a capsular T-shift operation according to Neer and Foster17 was performed, and in 3 cases open Bankart repair with additional capsulorraphy was performed. In 7 cases the capsulolabral complex was detached from the anterior glenoid rim. The capsule was of good quality, and a Bankart repair was therefore performed. In 3 patients this was performed arthroscopically. In 2 other cases the capsule was torn on the lateral side. This was recognized during arthroscopic examination. The capsule was repaired by an open technique. After a mean follow-up of 49 months (range, 24-81 months), none of the patients had had a recurrent dislocation or subluxation. In 5 patients (12%) the apprehension test was positive. However, only 1 of these patients remained symptomatic with a transient
J Shoulder Elbow Surg May/June 2004
feeling of instability when performing overhead activities. The Rowe score was obtained for 37 patients. The mean score was 95.5 points (range, 75-100 points). Of 37 patients, 30 (81%) had an excellent result and 7 (19%) had a good result. None of the patients had a moderate or poor result. The sulcus sign as a clinical indicator of a large joint capsule was positive preoperatively in 5 cases. In all 5 cases it was present in both shoulders. Intraoperatively, all 5 patients were found to have a large joint capsule, and in all of these cases a capsular T-shift operation was carried out (see above). Compared with the contralateral shoulder, the average loss of motion for the operated shoulder was 5.4° of flexion (0°-25°) and 2.1° of abduction (0°20°). External rotation at 0° abduction was on average 5° (0°-25°) less than that in the contralateral shoulder. External rotation at 90° abduction was on average 5.5° (0°-30°) less than that in the contralateral shoulder. Internal rotation was not restricted in 18 patients (47%)—that is, they could move the back of the hand to the interscapular position; 16 patients (45%) could reach the thoracic spine, and 3 patients (8%) could reach the lumbar spine. The mean loss of internal rotation strength in comparison with the contralateral shoulder was 0.55 kg (range, 0-4 kg). In 10 patients (24%), internal rotation strength was higher on the operated side. The mean loss of strength at 90° abduction was 0.6 kg (range, 0-4 kg). Four patients (ten percent) had more strength on the operated side. The 8 patients who had undergone three surgical repairs for shoulder instability had a mean loss of internal rotation strength of 0.63 kg with the arm in adduction and 1.1 kg at 90° abduction. There was no significant difference between operated and nonoperated shoulder (P ⬎ .05). The mean time to resumption of sporting activity was 3.6 months (range, 1-6 months). Of the 41 patients, 31 (76%) returned to the performance level that they had reached before surgery. Two patients (five percent) were actually able to improve further following revision surgery. Eight patients (twenty percent) were unable to achieve their preoperative levels. Four of these explained the problem in terms of psychological barriers. With one exception, all 8 patients who had performed some form of sport at a competitive level preoperatively (ie, trained at least 5 times a week) were able to regain their preoperative level of performance. Of the 41 patients, 21 had originally performed nonphysical work and 20 had performed physical work. The patients with nonphysical duties returned to work after a mean of 1.3 months (range, 0.25-2 months), and those employed in physical work returned after a mean of 2.4 months (range, 1-4
Tauber et al
J Shoulder Elbow Surg Volume 13, Number 3
283
Table IV Procedures at initial surgery and frequency of subsequent traumatic and atraumatic redislocations
Total No.
Traumatic redislocations
Atraumatic redislocations
Atraumatic redislocations (%)
25 6 4 2 1 1 1 1
15 4 2 2 1 0 0 0
10 2 2 0 0 1 1 1
40 33.3 50 0 0 100 100 100
Arthroscopic Bankart Open Bankart Eden-Hybinette Weber J-bone graft Capsular shift Bristow-Latarjet Unknown
months). None of the patients had to change his or her job. The mean score for reported pain on the 10-point scale was 9.5 points (range, 5-10 points). Of the 41 patients, 33 (81%) reported complete pain relief, 5 complained of an occasional slight hypersensitivity to changes in the weather, and 3 reported pain during continuous overhead activities. Mean patient satisfaction on the 5-point scale was 4.7 points (range, 1- 5 points). Of the 41 patients, 33 (81%) were very satisfied, 7 (17%) were satisfied, and 1 (2%) was fairly satisfied. Analysis of atraumatic failures
In 19 patients (46%) the first redislocation after surgical repair occurred without significant trauma (atraumatic failures). Six had previously undergone surgery on two occasions. In the majority of cases a Bankart repair had been performed, with an arthroscopic procedure selected in 7 cases and an open procedure in 6. Bone grafting procedures were performed in two patients, a capsule shift operation in one, a rotational osteotomy in one, and Bristow-Latarjet in one (Table IV). In the remaining case the applied technique of the index procedure was unknown. Intraoperatively, at revision surgery, a large bony Bankart lesion was found in 8 patients (42%) and a severely eroded glenoid rim in 1 (5%)—that is, bony deficiencies were present in approximately half of the cases. Four patients (twenty-one percent) had a large joint capsule. In the other cases typical Bankart lesions or extensive capsular detachment from the glenoid rim was found. Radiographic examination
At revision surgery, only 14 patients (34%) had no signs of arthritis. At follow-up, only 5 (12%) remained with no arthritic changes on radiographic analysis. In 19 patients no increase in arthritis was seen, whereas the arthritic change progressed by one degree in 13 and by two degrees in 4.
Of the 19 patients without an increase in arthritic change, 8 underwent a bone graft procedure and 11 underwent a soft-tissue repair. Of the 13 patients whose arthritis changed by one degree, 6 had been stabilized by J-bone graft and 7 by a soft-tissue repair. There was no significant difference between bone graft procedures and soft-tissue repairs in the progression of arthritis. Postoperative complications
Postoperative complications occurred in 3 patients. Out of a total of 23 patients in whom bone graft was harvested from the iliac crest, hypesthesia was reported postoperatively in the area of the iliac crest in 2 patients. One patient had a postoperative infection. All of these patients had a successful resolution of the complication. DISCUSSION Knowledge of the causes of redislocation after surgery for shoulder stabilization is important not only for the success of revision surgery but also for improved results at initial surgery. According to Levine et al,14 the most frequent causes of unsuccessful shoulder stabilization relate to the failure to correct an excessively large anterior-inferior capsular pouch and detached capsulolabral complex. They found the best results in revision surgery in those cases where redislocation was due to a traumatic event after just one surgical repair and in the absence of voluntary instability. Burkhart and De Beer3 stress the importance of reconstructing bony defects during arthroscopic procedures. According to their findings, the arthroscopic Bankart repair is capable of delivering the same results as an open Bankart procedure in the absence of significant bony defects. Even though these findings relate to arthroscopic procedures for initial surgery, great importance is attached to an unchanged bony glenoid rim for the stability of the glenohumeral joint. Half of all our patients requiring revision surgery
284
Tauber et al
had a bony Bankart defect extending to the anteriorinferior portion straight downward from the glenoid notch. In all of these cases the bony defect was reconstructed by bone grafting, and all shoulders remained stable. This would seem to confirm the importance of an uncompromised or reconstructed anterior glenoid rim for the stability of the shoulder joint and good function. At revision surgery, we found that the size of the remaining fragment did not correlate with the size of the glenoid defect. The absence of loading on the fragment may lead to atrophy and partial absorption. Therefore, the fragment could only be used to repair the glenoid in a few cases. The incidence of bony glenoid lesions in our series was much higher than that reported by other authors.3,24 However, this series presents a selected group of patients after failed stabilization surgery. Another reason for the higher incidence might be the fact that a careful intraoperative examination with special regard to the condition of the bony rim was performed at the time of revision surgery. An incidence of approximately 50% of bony glenoid rim defects in our revision cases demonstrates that a nonreconstructed bony glenoid rim plays a major role in the development of a postoperative failure. On the basis of this outcome, we believe that a defect of the bony glenoid rim thicker than just the cortex should be reconstructed by bone grafting at surgery. The size of the joint capsule and its importance for the stability of the shoulder joint have been stressed by several authors.4,17,25,26,30 They are of the opinion that the capsule is overstretched by the dislocations and, therefore, has to be shortened at the time of surgery.17,25 This has to be taken into consideration in case of revision surgery. Particular attention has to be given to the lax shoulder. In these cases the capsule was voluminous even before dislocation had occurred and was quite often of poor quality. In our series 5 patients have shown laxity at clinical examination. In none of them was this taken into account at the primary surgery. Only the capsule T-shift operation that we have used in these patients has resulted in permanent stability. In this study the most frequent cause of redislocation after initial surgery was a further traumatic episode, with sports injuries as the most frequent cause. Our study focused in particular, however, on patients who had redislocation without a traumatic cause (i.e, where surgical repair seemed to have failed to achieve stability). In such cases it is reasonable to assume that the repair failed because of poor technique, an inappropriate choice of procedure, failure to take the presence of lesions into account, or simply an incorrect diagnosis. Of the 19 patients who had redislocation without any external force, 8 (42%) had a glenoid rim frag-
J Shoulder Elbow Surg May/June 2004
ment of more than 2 mm that had been ignored at the index procedure. Another 5 patients (26%) from this group had a clearly overstretched capsule that had not been taken into account at previous surgery. In summary, in 13 of 19 cases (68%) a clear cause for the failure of the previous surgery could be found. A large proportion of the revisions we performed related to arthroscopic stabilization procedures. Of 25 patients treated arthroscopically, there were 10 in whom the labrum had not healed onto the glenoid rim despite the absence of bony defects. With such findings, redislocations are due primarily to problems of technique rather than a failure to address lesions. The Hill-Sachs lesion did not play a major role in the preoperative planning of the revision surgery in this series. On the basis of our good results in terms of both stability and function, we would conclude that the importance of the Hill-Sachs lesion is minor compared with the Bankart lesion and that the Hill-Sachs lesion may play a role only in the rare case of a huge defect. As the radiographic evaluations show, the progression of arthritic changes cannot be stopped in all cases despite successful stabilization; at best, it can only be slowed down. Before the last revision surgery, 34% of the patients had no radiographic signs of arthritis, whereas at follow-up 4 years later, only 5% were without arthritic changes. There was no difference between bone graft procedures and soft-tissue repairs. In this respect it has to be emphasized that the bone graft addressed only the size of the defect and did not alter the congruity of the glenoid surface. Given the very good functional results, a restricted range of motion as a factor in arthritic progression presumably can be discounted. It would seem to be a process that has already been triggered and cannot really be stopped by surgical repair procedures, at least not in all cases. These results show that surgical repair that takes full account of the clinical and intraoperative findings leads to very good results in terms of both stability and function. Had due attention been paid to these aspects at initial surgery, this series of revisions would not have been necessary. REFERENCES
1. Altchek DW, Warren RF, Skyhar MJ, et al. T-plasty modification of the Bankart procedure for multidirectional instability of the anterior and inferior types. J Bone Joint Surg Am 1991;73:10512. 2. Bigliani LU, Pollock RG, Soslowsky LJ, et al. Tensile properties of the inferior glenohumeral ligament. J Orthop Res 1992;10:18797. 3. Burkhart SS, De Beer JF. Traumatic glenohumeral bone defects and their relationship to failure of arthroscopic Bankart repairs: significance of the inverted-pear glenoid and the humeral engaging Hill-Sachs lesion. Arthroscopy 2000;16:677-94. 4. Burkhead WZ, Richie MF. Revision of failed shoulder reconstruction. Contemp Orthop 1992;24:126-33.
Tauber et al
J Shoulder Elbow Surg Volume 13, Number 3
5. Caspari RB. Arthroscopic reconstruction for anterior shoulder instability. Tech Orthop 1988;3:61-6. 6. Gerber C, Ganz R. Clinical assessment of instability of the shoulder. J Bone Joint Surg Br 1984;66:551-6. 7. Gill TJ, Micheli LJ, Gebhard F, et al. Bankart repair for anterior instability of the shoulder. J Bone Joint Surg Am 1997;79:850-7. 8. Hawkins RB. Arthroscopic stapling repair for shoulder instability: a retrospective study of 50 cases. Arthroscopy 1989;5:122-8. 9. Hayashida K, Yoneda M, Nakagawa S, et al. Arthroscopic Bankart suture repair for traumatic anterior shoulder instability: analysis of the causes of a redislocation. Arthroscopy 1998;14: 295-301. 10. Hovelius L, Akermark C, Albrektsson C, et al. Bristow-Latarjet procedure for recurrent anterior dislocation of the shoulder. Acta Orthop Scand 1983;54:284-90. 11. Jobe FW, Giangarra CE, Kvitne RS, et al. Anterior capsulo-labral reconstruction of the shoulder in athletes in overhead sports. Am J Sports Med 1991;19:428-34. 12. Johnson LL. Diagnostic and surgical arthroscopy of the shoulder. St Louis: Mosby; 1993. p. 276-352. 13. Landsiedl F. Arthroscopic therapy of recurrent anterior luxation of the shoulder by capsular repair. Arthroscopy 1992;8:296-304. 14. Levine WN, Arroyo JS, Pollock RG, et al. Open revision stabilization surgery for recurrent anterior glenohumeral instability. Am J Sports Med 2000;28:156-60. 15. Levine WN, Richmond JC, Donaldson WR. Use of the suture anchor in open Bankart reconstruction: a follow-up report. Am J Sports Med 1994;22:723-6. 16. Morgan CD, Bodenstab AB. Arthroscopic Bankart suture repair: technique and early results. Arthroscopy 1987;3:111-22. 17. Neer CS, Foster CR. Inferior capsular shift for involuntary inferior and multidirectional instability. J Bone Joint Surg Am 1980;62: 897-908. 18. Pagnani MJ, Warren RF, Altchek DW, et al. Arthroscopic shoul-
19. 20. 21. 22. 23. 24. 25. 26. 27.
28.
29. 30.
285
der stabilization using transglenoid sutures. A four-year minimal follow-up. Am J Sports Med 1996;24:459-67. Peterson CA, Newton SE. Arthroscopic Bankart repair. Am J Arthroscopy 1991;1:17-9. Resch H. The Bankart procedure. A modification using transosseus sutures. Orthop Traumatol 1993;2:19-28. Resch H. J-Span Plastik. In: Kremer K, Lierse W, Platzer W, et al, editors. Chirurgische Operationslehre, Schultergu¨rtel und obere Extremita¨t. Stuttgart: Thieme; 1995. p. 102-3. Resch H, Povacz P, Wambacher M, et al. Arthroscopic extraarticular Bankart repair for the treatment of recurrent anterior shoulder dislocation. Arthroscopy 1997;13:188-200. Rosenberg BN, Richmond JC, Levine WN. Long-term followup of Bankart reconstruction: incidence of late degenerative glenohumeral arthrosis. Am J Sports Med 1995;23:538-44. Rowe CR, Patel D, Southmayd WW. The Bankart procedure: a long-term end-result study. J Bone Joint Surg Am 1978;60:1-16. Rowe CR, Zarins B. Recurrent transient subluxation of the shoulder. J Bone Joint Surg Am 1981;63:863-72. Rowe CR, Zarins B, Ciullo JV. Recurrent anterior dislocation of the shoulder after surgical repair. J Bone Joint Surg Am 1984;66: 159-68. Thomas SC, Matsen FA. An approach to the repair of avulsion of the glenohumeral ligaments in the management of traumatic anterior glenohumeral instability. J Bone Joint Surg Am 1989;71: 506-13. Weber BG, Simpson LA, Hardegger F. Rotational humeral osteotomy for recurrent anterior dislocation of the shoulder associated with a large Hill-Sachs lesion. J Bone Joint Surg Am 1984; 66:1443-6. Wolf EM, Wilke RM, Richmond JC. Arthroscopic Bankart repair using suture anchors [abstract]. Oper Tech Orthop 1991;1:184. Zabinski ST, Callaway GH, Cohen S, et al. Revision shoulder stabilization: 2- to 10-year results. J Shoulder Elbow Surg 1999; 8:58-65.