Chapter 9: Arthroscopic Management of Traumatic Shoulder Dislocations
Background
Shoulder dislocation can be classified into one of two types:
- TUBS: Traumatic, Unidirectional, Bankart, Surgery
- AMBRI: Atraumatic, Multidirectional, Bilateral, Rehabilitation, Shift1
The great majority of TUBS are anterior dislocations (98 per cent in Rowe's2 series of 500), in whom a Bankart lesion is present in 85 per cent. The treatment of recurrence in this group is surgery.
AMBRI is a smaller group, usually initiated without trauma, often multidirectional (anterior, inferior and posterior), occurring in patients with generalized joint laxity, the opposite shoulder usually being loose and demonstrating a sulcus sign. These patients should not be operated upon. Treatment initially consists of a programme of supervised shoulder strengthening exercises. If the shoulder does not respond to such a programme, the patient should be referred on to a shoulder specialist to consider a capsular shift operation. There is no place for arthroscopic repair in these patients.
Two large studies have been performed on the incidence of recurrent dislocation following traumatic anterior dislocation (Figure 9.1). We have already seen (Chapter 6)
Hovelius
Figure 9.1 Incidence of recurrence following dislocation according to age at first dislocation. Data from Rowe and Hovelius.2,3
that there is a spectrum of both Bankart lesions (Figure 9.2) and Hill-Sachs lesions (Figure 9.3), and it is likely that the severity of damage to the anterior capsular structures is related to the number of recurrent dislocations.
Figure 9.2 Bankart lesion of the glenoid labrum IGHL complex.
Figure 9.3 Hill-Sachs osteochondral lesion in the back of the humeral head.
A concept of anterior traumatic dislocation is where the aim of surgery is to identify precisely the pathology responsible for dislocation and to perform a selective repair: 'Not all shoulder instabilities are created equal, nor are they treated the same' (Johnson[4]).
Recurrent Anterior Dislocation
The classic paper on stabilizing mechanisms preventing anterior dislocation of the glenohumeral joint by Turkel et al[5] should be read by all orthopaedic surgeons with an interest in shoulder dislocation. Turkel performed studies on 46 cadaveric shoulders with selective sectioning of the anterior structures to identify the relative contribution of each to shoulder instability, in various positions of the shoulder. The shoulder joints were opened posteriorly and wire sutures were inserted to mark the superior border of the middle glenohumeral ligament (MGHL), the superior and inferior margins of the thickening of the anterior inferior glenohumeral ligament (AIGHL) and the superior and inferior margins of the subscapularis muscle (marked from in front). Radiographs were then taken with the shoulder at 0, 45 and 90 degrees of abduction in external rotation (Figure 9.4). They concluded that at 0 degrees of abduction, anterior dislocation is prevented largely by subscapularis but, as the shoulder reaches 90 degrees abduction in external rotation (the position of apprehension), then the subscapularis rolls over the top of the humeral head and its inferior margin rides up exposing the lower half of the anterior surface of the head. The AIGHL was tight in this position and was the only structure covering the vulnerable anteroinferior portion of the humeral head.
Figure 9.4 Position of the subscapularis (lines D and E), middle glenohumeral ligament (top margin is marked A) and inferior glenohumeral ligament (lines B and C), with the arm at the side, at 45 degrees of abduction and at 90 degrees of abduction and external rotation (the position of apprehension). (From Turkel et al.5)
In the selective cutting experiments, section of subscapularis alone did not allow subluxation or dislocation. When all the structures except the AIGHL had been sectioned, dislocation did not occur. However, as soon as the superior band of the AIGHL was sectioned, subluxation or dislocation did occur, and this increased as the axillary fold of the AIGHL was sectioned. A further selective cutting experiment division of the AIGHL alone, with all other structures intact, led to dislocation. These studies show the vital role of the AIGHL complex.
Anatomically we have seen that the anterior labrum is an extension of the AIGHL complex, and the Bankart lesion is an avulsion of the AIGHL complex from its origin on the scapula. O'Brien and colleagues6 have drawn attention to the role of the posterior band of the IGHL, and have shown how the IGHL can be likened to a hammock upon which the humeral head rests (see Figure 5.26). The Bankart lesion can be likened to section of one of the ropes supporting the hammock, and thus the head falls forward (Figure 9.5). Oveison and Nielson7 have shown that a large lesion of the posterior capsule is needed for anterior dislocation and puts forward the concept of the capsule as a ring, the anterior part of which can only break if the posterior section comes forward as well. In the 10-15 per cent of patients who are not shown to have a Bankart lesion, two factors must be considered. The first is that the diagnosis is incorrect, these patients may in fact be AMBRI with a lax AIGHL, in which case they must not undergo arthroscopic repair (Figure 9.6). The second is that the diagnosis of recurrent TUBS is correct but there has been plastic deformation, or stretching of the AIGHL rather than avulsion from its origin and these patients will need an inferior capsular shift.
Figure 9.5 O'Brien et al6 have described the inferior glenohumeral ligament/labral complex as a hammock, upon which the head lies. When the front supporting rope of the hammock is cut then the humeral head falls 'out of bed'.
Figure 9.6 Multidirectional instability. Note the inferior shift of the humeral head so that its equator has almost slipped over the inferior glenoid into the infraglenoid recess.
Concerning the relevance of the Hill-Sachs lesion to recurrent TUBS, it should be recalled that the Hill-Sachs lesion is a result of the dislocation and not a cause of it. It is an impaction fracture of the dislocated head against the anterior glenoid. The only relevance of the Hill-Sachs lesion to the arthroscopic surgeon is that, if the defect is massive, then consideration should be given to a Connolly procedure (placing the insertion of infraspinatus within the defect), or a rotational osteotomy, rather than arthroscopic repair.
Fracture of the glenoid rim is a further contraindication to arthroscopic repair. The defect should be selectively repaired, which may mean a bony operation to restore the anterior glenoid. However, it should be said at this point that such a massive defect is rarely encountered. Minor crevassing of the anterior glenoid rim (Figure 9.7) is not a contraindication to arthroscopic repair.
Figure 9.7 Crevassing of the anterior glenoid rim.
Figure 9.8 The labrum has been traumatized so much that it has disappeared from a long segment of the anterior glenoid.
Unfortunately, by the time patients are referred for surgery, they have often had multiple episodes of recurrent dislocation, and what may have started as a simple avulsion of the AIGHL complex (Bankart lesion) undergoes repeated trauma, the Bankart lesion extends, the labrum fibrillates and tears within its substance and eventually disintegrates so that there is no trace left of it or the middle or inferior glenohumeral ligaments (Figure 9.8). Arthroscopic repair of the Bankart lesion is now impossible, and although Caspari's group have performed arthroscopic augmentations, the less experienced arthroscopic surgeon should abort the arthroscopy and move on to selective open repair. To abort an arthroscopic procedure is a sign of wisdom on the part of the surgeon, not an admission of failure.
Arthroscopic suture repair
On 14 September 1982, Johnson performed the first arthroscopic repair for anterior dislocation of the shoulder at Ingham Medical Center, Lansing, Michigan (Johnson LL, Detrisac DA, personal communication, 1983). The patient had sustained a traumatic anterior dislocation 6 months previously, with 100 subsequent dislocations, four in the week prior to surgery. After arthroscopic preparation of the anterior glenoid rim, the Bankart lesion was repaired using a specially devised staple, which was introduced through the anterior arthroscopic portal. By 1987, 195 repairs had been performed at Ingham Medical.
Although Perthes first used staples to repair the detached anterior capsule in 1906, the technique was popularized by Du Toit and Roux from Johannesburg.[8] They reported on an open staple repair which they had been performing since 1932. They credited the procedure to Fouche and Allen whose surgical staples were fashioned from bicycle spokes. The anatomical basis is to reattach the avulsed AIGHL labrum complex to the glenoid neck with the staple. The method was modified by Boyd and Hunt [9] who introduced a barbed staple. The simple nature of this open operation led to great popularity, but loosening of the staples and degenerative lesions of the glenoid rim have led to its discontinuation as an operation.
The simplicity of this open procedure led to its reintroduction by Johnson as an arthroscopic technique in 1982.4 A special 4 mm staple was devised for arthroscopic delivery. The method according to Johnson is described below.
Method
Examination under anaesthesia
Both shoulders should be examined using the tests for instability (see Chapter 3). Johnson has found that external rotation is limited in the dislocating shoulder preoperatively by 12 degrees (with the elbow at the side) compared to the unaffected shoulder
The patient is positioned in the lateral position for shoulder arthroscopy, and a single shoulder holder is used to suspend the arm. If the arm rolls into external rotation and starts to sublux in the suspension apparatus, then the procedure may become more than usually difficult. No more than 15lb of traction should be used, as further traction can cause damage to the brachial plexus.
Diagnostic arthroscopy
The shoulder must always be arthroscoped prior to any surgical procedure for instability. Firstly, the joint can be examined far more thoroughly than at arthrotomy, allowing precise identification of the pathology, which in turn allows the surgeon to plan the selective repair required. For instance, a simple Bankart lesion needs a Bankart repair, either arthroscopic or open. A stretched AIGHL requires an open inferior capsular shift, and a patient with no labrum or glenohumeral ligaments needs an extra-articular augmentation, such as a Putti-Platt or Magnusson-Stack procedure.
Secondly, a large proportion of patients will have second pathology (Table 9.1) which can be assessed or corrected. Finally, the surgeon is given a legitimate reason for shoulder arthroscopy. The dislocating shoulder is ideal for training in shoulder arthroscopy as the joint is usually of large capacity, entry is easy, and there is a large amount of abnormal pathology on view.
Table 9.1 Associated lesions found during staple anterior repair (Johnson4).
| n = 195 |
| 13% Cuff tear |
| 90% Hill-Sachs |
| 4% Anterior glenoid fracture |
| 14% Loose body |
| 10% Posterior labral tear |
Exposure
The key to any surgical procedure is good exposure and this is as true of arthroscopy as of any open technique. Correct scope positioning, atraumatic single insertion, and joint distension are important for getting a good view. Control of bleeding is achieved by increasing intra-articular pressure to above systolic blood pressure. This can be done either by elevating the head of fluid, or by the use of a pump, increasing flow to wash out the joint, or the use of f ml 1/100000 adrenaline solution in 3000 ml saline for irrigation. If the exposure is not adequate, traction must not be increased. The procedure should be terminated and open repair undertaken.
Arthroscopic stapling method
Preparation of the bed for reattachment of the AIGHL labrum complex is vital to the success of the procedure. The Bankart lesion should be extended interiorly so as to allow the AIGHL to be shifted upwards and medially later in the procedure. The extension of the lesion must be performed with the greatest of care, firstly in order that the AIGHL complex is not damaged further; secondly because the axillary nerve lies just inferior to the axillary fold of the IGHL. Extension of the Bankart lesion also allows better access to the front of the glenoid neck which can be prepared by using a curette, the motorized shaving system or drill holes (Figure 9.9). The idea is to create a bleeding bed, for the bed must be vascularized for healing to occur. The staple is then inserted through the anterior portal, and the superior band of the AIGHL is impaled between the tines of the staple (Figure 9.10). It is now vital that the staple, with the impaled AIGHL, is shifted both superiorly up the anterior glenoid rim, and then medially in front of the glenoid rim to tighten up the inferior glenohumeral ligament complex and place it on the previously prepared vascularized bed (Figure 9.11). The staple is then hammered home. Loose staples are usually a result of the staple not being firmly inserted at operation, and it is sometimes difficult to judge if placement is secure, as the staple lies in front of the glenoid and can only be seen correctly with the arthroscope in the anterior portal.
Figure 9.9 The glenoid neck is prepared using a bur
Figure 9.10 The superior band of the inferior glenohumeral ligament is impaled between the tines of the staple.
Figure 9.11 The staple is shifted upwards and medially to tighten the inferior glenohumeral ligament complex.
The staple is removable if there is a failure to secure it to the bone, misplacement, unexplained pain, repeat surgery or if in the throwing athlete. A biodegradable staple is currently being tested by Johnson.
Postoperative regime
The patient is immobilized in a sling in internal rotation for a full 3-week period. During weeks 4-6, pendulum exercises are encouraged. At week 7, overhead exercises are encouraged. When the patient has regained his preoperative range of motion, shoulder-strengthening exercises are started. When deltoid and cuff power have returned to preoperative levels then sports are allowed, but no throwing is allowed for 6 months.
Results
By the end of 1987, 195 cases had been performed and 147 had a follow-up of over 2 years. The following associated pathology was found: 13 per cent had rotator cuff tears, 90 per cent had a Hill-Sachs lesion, 4 per cent had an anterior glenoid fracture (usually small), 14 per cent had loose bodies, and 10 per cent had a posterior labral tear. All the patients had some form of Bankart lesion. In 75 per cent, this consisted of detachment of the AIGHL alone, 16 per cent had detachment of both the MGHL and AIGHL, and 9 per cent had complete absence of ligaments.
Throughout the series there was a shift from using multiple staples to using a single staple. Preparation also developed from burring to drilling the anterior glenoid neck. The following technical problems were encountered. In 9 per cent, the ligaments were inadequate, a subscapulars tenodesis was attempted but 5 out of 15 redislocated, so this procedure has been abandoned. If patients have no glenohumeral ligaments, they require open augmentation. Inadequate glenoid decortication was solved by changing the scope portal to the front or using a 90 degree arthroscope. Staple problems were encountered in the form of bending, breaking, missing the bone or loosening.
In the first 5 years, 6 months, there was no case of infection and no neurovascular injury, 15 staples had to be removed: 8 because of pain, 3 in athletes, 2 for loosening and 2 for poor position. The redislocation rate was 21 per cent. If the table of results is examined closely (Table 9.2), it can be seen that there was an unusually high dislocation rate in 1984. An analysis of the redislocations showed that those of 1984 were partly due to a change of postoperative policy to immobilization for only 2 weeks. This has been reinstated to 3 weeks. Two patients were AMBRI. Twenty-two patients redislocated while undertaking activities such as prizefighting, barfighting, basketball, football, baseball, waterskiing and diving.
Table 9.2 Results of staple anterior repair by year of surgery (Johnson4).
|
Year |
Cases |
Lost |
Radiolocation |
| 1982 | 2 | --- | 0 |
| 1983 | 16 | --- | 3 |
| 1984 | 55 | 1 | 16 |
| 1985 | 36 | 2 | 3 |
| 1986 | 42 | 1 | 8 |
| 1987 | 44 | --- | 1 |
|
Total |
195 |
4 |
31 (15%) |
External rotation was not limited subsequent to the procedure. Johnson, as mentioned, has made the interesting observation that external rotation is limited preoperatively, and this brings into question historical reports of normal rotation with various open procedures.
In summary the advantages of staple repair are: accurate diagnosis, microdebridement, lesion assessment and selective repair, which can be performed as a day-case procedure. The disadvantages are that it is a difficult technique, with a learning curve, and it can only be performed if the ligaments are present. The redislocation rate is 21 per cent but if short immobilization and collision sports are excluded, the redislocation rate is only 6 per cent. No extra-articular reinforcement is possible and there are staple problems. A biodegradable staple is being tested.
Arthroscopic Suture Repair: Morgan technique
In 1959 Viek and Ben[10] described a technique for reattaching the Bankart lesion using pullout sutures. They acknowledged Luckey with the concept. Three 0.3 cm (1/8 inch) Steinmann pins were drilled through the glenoid from the front and then out of the back of the patient, passing through infraspinatus and skin. Pullout sutures were taken through these transglenoid tunnels and tied over on the skin. The sutures were removed at 3 weeks. Morgan and Bodenstab[11] took this concept and performed the operation arthroscopically. The advantages of this technique are its simplicity, the lack of damage to the glenoid neck, and the fact that no metallic hardware is left in the joint.
Method
Set up
The patient is placed in the lateral position and the arm is suspended, using two shoulder holders, in a position of internal rotation.
Diagnostic arthroscopy
This is performed for the same reasons as in the staple repair. Bleeding is controlled by adding 1 ml 1:1000 adrenaline solution to each 31 bag of saline.
Procedure
An anterior portal is made through which a 7mm utility arthroscopic cannula is passed. The Bankart lesion is probed and assessed. The anterior glenoid neck is prepared exactly as for staple repair.
The arthroscopic suture is performed using a specially designed stainless steel suture pin, 2 mm in diameter and 30 cm in length. This pin has a specially sharp pointed tip which is necessary to 'pick up' the labrum. The trailing end of the suture pin has a recessed eye to carry the suture. Two passes of the suture pin are made, the lower one first. The suture pin is introduced through the 7 mm cannula into the front of the joint (Figure 9.12) and the sharp tip is used to spear the IGHL labral complex at their junction. The sharp tip of the pin can now be seen between the labrum and the prepared anterior glenoid rim (Figure 9.13) and the pin is now used to shift the speared soft tissue both superiorly and medially on the glenoid neck in order to tighten up the IGHL labrum complex. The pin is knocked into the prepared bone with a tap from a mallet, in a position about 3 mm medial to the glenoid neck.
Figure 9.12 The suture pin is inserted through a 7 mm cannula into the front of the joint.
Figure 9.13 The suture pin passes through the labrum.
Figure 9.14 The suture pin should pass downward (caudad) 15 degrees, and either parallel to or 15 degrees medial to the glenoid articular surface to avoid damage to the joint surface and the suprascapular nerve.
It is important at this stage to angle the pin for a safe passage through the glenoid. In order to miss the suprascapular nerve on exiting the glenoid, the pin must be angled 15 degrees downward (caudad) to a line drawn perpendicular to the long axis of the glenoid. The pin must also run parallel to the articular surface of the glenoid or, at the most, 15 degrees angled in to the glenoid. Angled any less, the pin may break through into the joint and, angled any more, it might hit the suprascapular nerve (Figure 9.14). Rose has developed a guide, similar to an anterior cruciate ligament (ACL) guide, which can be used to make passage of the pins more accurate. The results are yet to be published.
The pin is then driven through the bone using a power drill and a 5 mm stab incision made over the skin at the point where the skin is tented up by the pin. The pin now passes from the front of the shoulder, through the cannula, the Bankart lesion, the glenoid and out of the back of the shoulder (Figure 9.15). A length of 1 PDS suture is passed through the eye of the suture pin and a haemostat attached to the two ends of the thread. A pair of pliers is now used to grasp the sharp end of the pin and pull it through the shoulder from the back (Figure 9.16) so that it is now replaced by a double thickness length of 1 PDS suture.
Figure 9.15 The pin is drilled through the glenoid.
9.16 The assistant grasps the suture pin and pulls the suture out through infraspinatus and the skin.
Figure A second passage of the pin is made through the cannula piercing the Bankart lesion about 1.5 cm above (cephalad) to the first passage. The pin is angled parallel to the first passage and drilled through the bone. It is possible to pick up the first suture on the rotating suture pin which makes a mess of tangled suture within the joint. This can be avoided either by being aware of this possibility or by using a reciprocating drill. The second pin should exit the skin about 2 cm above the first pin. Once again a length of 1 PDS is passed through the eye of the pin, and a haemostat attached to the ends. The pliers are then used to bring the second suture through the glenoid.
At this point, there are two doubled threads of 1 PDS coming out of the cannula, each with a haemostat on the end. The two doubled threads are tied together with a double square knot and the excess trimmed off (Figure 9.17). The sutures are cut from the suture pins at the back and replaced with haemostats and by pulling on these haemostats the knotted PDS suture disappears down the cannula and comes to rest on the Bankart lesion. Further tension from behind pulls the Bankart lesion down on the prepared glenoid neck where it will be held until it has healed.
Figure 9.17 The sutures are tied together firmly and the excess cut off.
The only remaining problem is what to do with the suture at the back. A small stab incision is made between the two exiting pairs of sutures and each is retrieved subcutaneously using a haemostat and brought out through the central stab. The two pairs of threads are then tied using a double square knot over the fascia of infraspinatus. A hook probe is then inserted through the cannula and the repair is tested. If all is satisfactory, the cannula and arthroscope are withdrawn and the puncture sites are closed.
Postoperative regime
The shoulder is immobilized in internal rotation for 6 weeks. Approximately 10 days postoperatively, the skin sutures may be removed and the patient is allowed to remove the shoulder immobilizer three times daily for elbow extension exercises. The patient is allowed to shower, as long as the arm is kept in internal rotation. At 4 weeks, active assisted and Codman pendulum exercises are used under the physiotherapist's direction. At 6 weeks, TheraBand exercises are used to strengthen the rotator cuff. Vigorous exercises including contact sports and throwing are not allowed until 6 months after the repair.
Results
Morgan presented his 2-5 years results at the American Academy of Orthopaedic Surgeons' 1989 meeting.12 Of 60 patients, 55 had a Bankart lesion, all had a positive anterior apprehension sign, and none were lax jointed or had a sulcus sign. Follow-up was graded by the Rowe assessment at an average of 37 months. There were only two failures: one redislocated playing American football, fracturing his glenoid neck at the same time, and one subluxed. One patient had a neuropraxia of the medial antebrachial cutaneous nerve, but no evidence of injury to the suprascapular nerve and no problems with the suture over infraspinatus. Of the 96 per cent graded as excellent, which is comparable to the best open series, 53 had excellent results, including 46 who achieved a full range of shoulder movement, and 7 who lacked 5 degrees of movement.
Inferior Capsular Shift
Caspari and Rose have each devised a method of picking up the IGHL labrum complex with a suturing device. Caspari devised a suture punch which could be used to deliver a suture through the superior band of the IGHL. The problem with the suture punch is its size, in particular withdrawing it from the front of the joint with its jaws open. Rose's suture passer (Figures 9.18 and 9.19) is easier to insert into the joint but needs a separate instrument to grasp the free end of the suture to withdraw it from the joint (Figure 9.20). In both techniques, the sutures, once inserted into the Bankart lesion, are passed on a suture pin through the glenoid so as to shift and tighten the IGHL labrum complex and hold it down to the prepared glenoid neck (Figure 9.21).
Figures 9.18 and 9.19 Rose's suture passer is used to pass a suture through the glenoid labrum.
Figure 9.20 The suture placed through the labrum and exiting the cannula.
Figure 9.21 With several sutures placed, the suture pin is inserted through the prepared glenoid neck. The sutures are taken through the glenoid to be tied over the infraspinatus fascia.
Results
Caspari et al also presented their results at the American Academy of Orthopaedic Surgeons' 1989 meeting.[13] At that time they reported on 100 cases: 49 with a labral detachment had a suture repair, 51 with a midsubstance tear of the IGHL labrum complex had a semiclosed augmentation.
In the 49 suture repairs, 5-8 sutures were inserted in each patient, the maximum being 12 sutures. The patients wore an immobilizer for 3 months after surgery. The average number of preoperative dislocations was five in this group, 92 per cent had a satisfactory result, two patients had a further dislocation and two patients a further subluxation.
Of the 51 patients who had an augmentation, 80 per cent had a satisfactory result. These patients had an average of 19 preoperative episodes of dislocation, there were eight redislocations, seven in American football players, and one in a road traffic accident. This is an extremely difficult technique which is beyond the remit of this book.