4.7 Instability

Stability of any articulation can be defined as asymptomatic normal mechanical behaviour at rest and in motion. It depends on structural integrity and intact neural control systems (afferent, efferent, and neuromuscular connections). For the glenohumeral joint (GHJ) stability is therefore the product of a functionally (not necessarily anatomically) intact rotator cuff, competent capsular and labral structures, a sufficient surface arc (or surface area) of contact between humeral head and glenoid, and an intact neuromuscular system comprising central and peripheral connections. Motion about the centroid of rotation and ‘containment’ of the humeral head on the size-mismatched glenoid is the result of concavity compression. This is generated by the centralizing joint reaction force (CJRF) of the rotator cuff (including the tendon of the long head of biceps, LHB), and is facilitated by the conformity of the glenoid/labral surface. Rotator cuff competence is a function of centrally-driven activation, and the strength and inertia of the cuff muscles. Shoulder joint position sense and motion are the product of right parietal cortical programming (the parietal ‘reach region’), and visual, vestibular, and cervical afferent inputs. Mechanoreceptors in the capsulolabral structures, tendon, and muscle, and the gliding planes (acromio–coraco–deltoid, and scapulothoracic) maintain this activity.

Instability is usefully defined as the condition of symptomatic abnormal motion of the joint. This universal definition combines the mechanistic view and the clinical perspective, and distinguishes the pathological state from the constitutional state of laxity, in which anatomical anomalies may exist, but which does not necessarily cause impairment of function. Since it is a symptom, the term implies recognition (a cerebral cortical event) of the abnormal motion (a difference in position, motion direction, or motion velocity has been identified). Thus ‘instability’ cannot be determined by examining the shoulder under anaesthetic (EUA). Laxity (looseness) of the GHJ can be determined by EUA either by comparison of the perceived abnormal with the perceived normal side or by reference to what is considered a normal range of laxity for the population.

The relationship between the dominant factors in creating stability in the GHJ can be described by an equation (see Box 4.7.1).

Each factor varies in relevance to the individual patient, but several points are applicable:

The clinical syndrome of instability is a disturbance of one or more of these factors in isolation or together, and the relative importance of each factor to the syndrome can change over time. The pathologies causing instability comprise structural (RC, SAC, CLPM) and non-structural (NS) elements (Box 4.7.2 and Figure 4.7.1). The structural elements may be damaged by extrinsic force (traumatic structural); acquire microtraumatic lesions over time (atraumatic structural); or be congenitally abnormal, or comprised of abnormal collagen. The non-structural elements can be congenitally abnormal or can be acquired over time as pertubations of neuromuscular control, particularly at periods of skeletal growth.

The combined concepts that structural (traumatic and atraumatic) and neurological system disturbances can cause instability, the observation that more than one pathology can be present in the shoulder at the time of diagnosis, and that pathology can change over time led Bayley to the classification of instability as a continuum of pathologies, which can be graphically displayed as a triangle (Figure 4.7.2). The polar pathologies are labelled types I (traumatic instability), type II (atraumatic instability), and type III (neurological dysfunctional or muscle patterning). Polar groups I and II and the axis I–II representing the spectrum between the two poles correspond to the TUBS-AMBRI classification which also allows for a spectrum of structural shoulder pathology but which does not admit those shoulders in which there is no structural (traumatic or atraumatic) cause for the instability. The characteristics for each polar group are given in Box 4.7.3. Jaggi has further subdivided polar group III into peripheral, central, positional, protective, and combination subtypes (personal communication). The interpolar spectrum describes dual pathologies in which traumatic, atraumatic and muscle-patterning factors play a variable role in the emergence of instability.

The traumatic structural group includes the TUBS (Traumatic, Unilateral, Bankart lesion present [in the anterior instability type], Surgery usually indicated) group but also includes the posterior traumatic instability characterized by the McLaughlin or Reversed Hill–Sachs lesion and posterior labral damage or locked posterior dislocation. Instability can thus be recurrent (acute or chronic) or persistent (acute or chronic). Luxatio erecta is a version of anterior instability, usually of the more elderly, and may be associated with infraclavicular plexopathy, and extensive bruising indicative of rupture of the circumflex vessels. In general shoulder clinics, acute recurrent anterior GHJ instability is the commonest form of instability. Assessment is based on the history of the initiating event, the examination is discussed in Chapter 4.1 and investigation is best undertaken by magnetic resonance arthrography (MRA) or EUA and arthroscopy.

Treatment comprises, in principle, surgical intervention for most cases in which capsulolabral or rotator cuff disruption is evident. There remains controversy regarding the choice of open versus arthroscopic surgical techniques.

In this group the incidence of ligamentous anomalies is the same as the population of traumatic instabilities. Within this group is the population defined as AMBRI (Atraumatic, often Multidirectional, often Bilateral, treated by Rehabilitation/physiotherapy)—in the original definition of AMBRI, surgical treatment by Inferior capsular shift is indicated if rehabilitation fails. This concept may be flawed: some patients will fall into the category of subgroup II/III and so will possess some neuromuscular control aberration. These patients may fail surgery because of the muscle-patterning behaviour, and may have a substantially greater probability of developing instability or postcapsulorrhaphy arthropathy if surgery is undertaken. The decision to operate should not be made on the basis of failed physiotherapy.

This group also contains a subgroup of individuals with benign hypermobility syndrome (of which Ehlers–Danlos syndrome is a variant), characterized by a Beighton score of greater than five out of a total possible nine points. Hypermobility may be limited to the upper limbs. These patients may present with pain rather than overtly abnormal displacements.

Anterior atraumatic instability is less common than posterior types. Posterior structural anomalies (including posterior glenoid dysplasia, excessive glenoid retroversion, glenoid hypoplasia, and medialized posterior capsular attachment) create the environment in which obligatory positional posterior displacement may occur during elevation of the arm in the sagittal plane. If abnormal clinically obvious muscle activation appears at the onset of this motion, then the diagnosis is modified to acknowledge the muscle patterning: the case is group II/III (peripheral subtype). If the muscular activation occurs during the motion, and appears to be provoked by the position of the arm at the shoulder, then the diagnosis is still II/III but with the suffix (protective subtype). If there is no clinically or electrophysiologically-proven aberrant muscle activation then the condition is labelled II (positional subtype).

Scapular dyskinesis is common, and reflects the attempt by the scapular postural muscles to maintain scapulohumeral homeostasis. Serratus anterior can be suppressed in some cases: the aberrant muscle activity now affects the scapular primarily. This represents a scapulothoracic (STh) type III condition. The total diagnosis is therefore STh(III)/GHJ(II).

Assessment of the structural diagnosis is by MRA and/or diagnostic arthroscopy. These are complementary, not alternatives; MRA is useful to look for capsular detachments, bony defects, and the bulk and quality of rotator cuff muscles (particularly subscapularis, for the anterior, and infraspinatus, for the posterior types), while arthroscopy is useful to look for the subtleties of internal lesions of occult instability: bicipital and deep surface cuff lesions, soft tissue Broca defects, internal impingement lesions, and external impingement lesions.

Treatment comprises rehabilitation, avoidance of activities promoting the instability, pain management, and surgical intervention in selected cases. The variants of medially- or laterally-based inferior capsular shift with augmentation of dysplastic labral tissue are complemented in some cases of glenoid insufficiency by glenoid augmentation using autologous bone blocks (anteriorly and/or posteriorly), or glenoplasty (the Scott posterior glenoplasty).

Muscle-patterning instability (MPI) comprises aberrant activation of large muscles identified by dynamic electromyography (latissimus dorsi, pectoralis major, and anterior deltoid have been characterized thus far) and simultaneous suppression of the rotator cuff (infraspinatus has been characterized thus far: whether suppression of infraspinatus reflects whole rotator cuff suppression or specific suppression of the infraspinatus is not yet clear). MPI can be clinically obvious, in polar group III, but may be occult, requiring dynamic electromyography (DEMG) for confirmation of the suspicion of the existence of MPI in the presentation (as in types II/III and III/II instabilities).

Patients with abnormal muscular activation leading to ‘pure’ MPI appear to present in a trimodal distribution with peaks at 6, 14, and 20 years. The mean age at onset of symptoms was 14 years and the mean duration of symptoms before presentation was 8 years. In the under 10 years age group there were more females (71% vs 47%), greater laxity (estimated with the Beighton score, 63% vs 29%), and bilaterality (54% vs 42%), with fewer presenting with pain (17% vs 50%). As age increased laxity decreased and pain increased. Bilaterality did not appear to be associated with gender, laxity, or pain. Laxity was associated with gender but not pain or bilaterality. These observations suggest different aetiologies for the MPI in the three cohorts.

Bayley (2005) used DEMG in latissimus dorsi (LD), pectoralis major (PM), anterior deltoid (AD), and infraspinatus (IS) to evaluate the patterns of muscular activation in patients who clearly demonstrated, visibly or by palpation, apparently abnormal muscular activation immediately before and during dislocation of the GHJ in the outpatient clinic. These muscles were chosen because they were reliably accessible. Early pilot studies were expanded into a cohort study of over 1000 cases of glenohumeral instability (GHI) seen at the Royal National Orthopaedic Hospital, a tertiary referral centre, over a 22-year period (1981–2003), from which DEMG data were extracted. The results are summarized in Box 4.7.4.

MPI was diagnosed, clinically and/or electrophysiologically, in 494 shoulders (45% of the entire cohort of 1097 shoulders) of 386 patients. In 44% the MPI affected both shoulders. The dominant arm was affected more often in right-handed patients, the non-dominant more often in left-handed patients. The central neurological basis for these differences remains obscure. There were 323 shoulders with ‘pure’ MPI; 161 shoulders presented a mixed picture of MPI in addition to either traumatic or atraumatic structural causes for the instability.

The value of recognizing this form of instability is given by the success of specialist physiotherapy using biofeedback techniques versus conventional therapy for retraining abnormal muscle activity: 76% of patients had either no change or a deterioration of their condition with conventional therapy (including conventional strengthening exercises) compared with 61% of patients overall achieving improvement in their condition with specialist therapy. The incidence of iatropathic arthropathy increases with inappropriate surgery in those with type III instability while no type III in which surgery was not performed developed arthritis.

Given the limitation that only four muscles have been studied in depth, the evidence strongly suggests that these, and perhaps other muscles, are implicated in many cases of instability. Inappropriate activity in PM, LD, and AD, with suppression of IS appear either singly or as couples in the generation of instability in shoulder with and without capsular collagen insufficiency. Whether suppression of IS reflects specific inappropriate inactivity in IS alone or reflects abnormal activation of the rotator cuff as a whole remains unknown.

DEMG provides additional information about abnormal muscle activation or suppression in subtle MPI in which clinical examination has a low specificity and sensitivity. DEMG was found to be most useful in patients presenting with pain, glenohumeral capsular insufficiency, and occult muscle patterning, categorized as the interpolar group II/III. Given that a chance of a successful outcome for rehabilitation of patients with MPI (symptoms abolished or controlled) was diminished fivefold for anterior and tenfold for posterior instability after inappropriate surgery in patients with MPI, it is clear that the identification of this specific cohort is important. These patients contribute to the failures of the AMBRI group after rehabilitation (if MPI has not been recognized). It is noteworthy that the pathology is not defined by direction but rather that the direction of instability is determined by the pathology.

In MPI rehabilitation is used as the first mode of treatment: there is usually a change in aberrant muscle activity within 6 months. Therapy has to go on for at least 2 years, possibly for cortical re-imprinting to be re-established There is a 25% recurrence rate within 2 years, needing ‘top-up’ therapy, including inpatient treatment in which a multidisciplinary approach is favoured. The value of specific suppression of specific muscles by injection of botulinum toxin is unclear in MPI.

Multidirectional instability (MDI) is not a diagnosis but a symptom. Multidirectional laxity is assessed by EUA. It may be difficult to determine when a lax shoulder is centralized during EUA. Fluoroscopy is invaluable in such cases. An individual shoulder may have the same pathology (i.e. group I, II, or III pathology) causing instability in each of more than one direction, or more than one pathology (e.g. group I in one direction, and group II/III in the other direction), thus distinguishing ‘simple’ MDI from ‘complex’ MDI. The underlying diagnoses of the different directions of instability are distinguishable by the history, clinical observations, arthroscopy, and/or MRA (Figure 4.7.3) supplemented by DEMG where necessary. Each pathology is treated separately, with rehabilitation of MPI elements taking priority.

Having diagnosed the specific pathology (-ies), treatment is directed at each element in the instability syndrome. The pathology of instability can be expressed as a useful equation acting as a checklist to define the management strategy (Box 4.7.5). The aim of treatment is the restoration of (asymptomatic) stable motion by restoration of the CJRF, optimization of the angular range of stability, and so restoration of concavity compression. Management follows simple principles (Figure 4.7.4): surgery should be undertaken within the context of a well-considered rehabilitation programme largely centred around optimizing rotator cuff function. In general, disrupted anatomy is restored, preferably by anatomic operations with predictably good outcomes. Dysplastic anatomy is augmented, often by non-anatomic operations with less predictable outcomes. Revision stabilizations are generally non-anatomic, and have higher failure rates.

The surgical treatment of structural dislocation depends upon the pathology detected on plain radiographs, MRA, and arthroscopy. Of the structural elements the capsulolabral pathology (CLPM) is the most common. Structural SAC lesions are becoming more frequent due to the increasing professionalism of contact sports leading to bone damage to the glenoid. Finally rotator cuff (RC) structural damage is common in dislocation in middle and old age, and is more commonly recognized these days.

The labrum is often torn during traumatic dislocation. This can be the classic Bankart lesion, where the labrum tears from the bone along the anteroinferior quarter of the glenoid. A variety of Bankart variants have been described such as the ALPSA (anterior labrum periosteal sleeve avulsion) lesion and the Perthes lesion. The Bankart tear may extend into a SLAP lesion, or extend around the inferior glenoid, or even involve the entire circumference of the socket. These lesions will need to be released (as they often heal in an abnormal position), reduced to their proper position on the glenoid rim and re-attached using suture anchors to a freshened bed so that they can heal strongly in an effective position. Labral repairs for patients at the Polar I position can be done either arthroscopically or by classic open surgery through a deltopectoral approach.

The glenohumeral ligaments may stretch or tear, or avulse from the humeral surface as a HAGL (humeral avulsion of the glenohumeral ligaments) lesion. The ligaments may be thin, stretched, or elastic if the patient lies on the I/II line, as the majority do. If the ligaments are loose and baggy then they will need to be repaired at the time of surgery. The capsule is often globally loose and may need to be tightened South to North and often East to West as well. At open surgery this is easy to do. The capsule is opened using a vertical capsulotomy and three sutures are placed to close the rotator interval. This will usually effect the necessary inferior capsular shift. As the capsulotomy is closed it can be overlapped East to West making sure that it is not overtightened. If more of a shift is required then a Neer T-shaped capsulotomy is performed. The rotator interval is closed and then a South to North double-breasting is performed to create a full and proper Neer inferior capsular shift. The vertical capsulotomy is then closed with as much double breasting East to West as required.

This is where things get difficult at arthroscopic surgery. All capsular shifts depend upon closing the rotator interval. If the interval is not securely closed then any South to North plication will merely open the rotator interval more, negating the effect of the capsular shift. Arthroscopic rotator interval closure can be performed, but is not easy, because the anterior working cannulae themselves take up most of the interval. To place more than one suture in the interval is more difficult still. To place three sutures in the interval arthroscopically is just showing off, very few surgeons have the requisite skill to do this. Double breasting is just not possible arthroscopically. This means that the further down the Polar I/II line towards the Polar II end becomes more and more difficult to perform arthroscopically.

If there is bone loss to the anteroinferior glenoid then this bone loss will need to be made good. This can not be done using a ligamentous repair and will need bone to be transferred to make up the SAC loss. The normal way to do this is a Bristow–Latarjet transfer of the coracoid process. This operation is performed using a deltopectoral approach. The musculocutaneous nerve must be exposed and protected throughout the procedure. The bone surfaces of the coracoid and the glenoid neck must be decorticated so that the coracoid will unite to the scapula. The coracoid is attached with one or two AO lag screws.

If the defect is larger than usual, too large for a coracoid transfer, then a tri-cortical bone block may be needed from the pelvic rim to make up the SAC defect.

In the middle aged to elderly patient the dislocation may be complicated by tearing of the rotator cuff itself. In these patients recurrent dislocation is not the long-term problem, it is the cuff tear itself, and this will need to be repaired.