The Painful Anterior Apprehension Test – an Indication of Occult Shoulder Instability

Introduction: To evaluate the clinical relevance of the painful anterior apprehension test in shoulder instability. Materials and methods: We performed a retrospective study of 155 patients that underwent arthroscopic anterior Bankart repair between 2014–2016. Exclusion criteria were previous ipsilateral shoulder surgery, bony Bankart lesions, glenohumeral osteoarthritis and concomitant surgery involving rotator cuff tears, biceps tendon pathology and superior labrum from anterior to posterior (SLAP) lesions. The study cohort was divided into three groups: apprehension test with apprehension only, apprehension test with pain only, and apprehension test with both apprehension and pain. Patient demographics, clinical characteristics, radiological imaging, arthroscopy findings and surgical outcomes (Constant, American Shoulder and Elbow Surgeons (ASES), SF-36 scores) were evaluated. Results: A total of 115 (74.2%) had apprehension only, 26 (16.8%) had pain only and 14 (9.0%) had pain and apprehension with the apprehension test. Univariate analysis showed significant differences between the groups in patients with traumatic shoulder dislocation (p=0.028), patients presenting with pain (p=0.014) and patients presenting with recurrent dislocations (p=0.046). Patients with a purely painful apprehension test were more likely to have a traumatic shoulder dislocation, more likely to present only with pain, and less likely to present with recurrent shoulder dislocations. Multivariate analysis showed that none of these factors alone were significant as single predictors for shoulder instability. All three groups were otherwise similar in patient profile, MRI and arthroscopic assessments, and clinical outcomes of surgery. Excellent clinical outcomes were achieved in all groups with no difference in pre-operative and post-operative scores across all groups at all time points. Conclusion: The painful apprehension test may suggest underlying shoulder instability.


INTRODUCTION
The anterior apprehension test is routinely used to detect anterior shoulder instability with several studies showing the diagnostic value of this simple clinical test 1,2 . A level 1 prospective cohort study showed that the overall accuracy of the apprehension test was 81.7% 2 , while a separate metaanalysis also showed a strong association between a positive apprehension test and anterior shoulder instability with a high positive likelihood ratio 3 . In most of these previous studies the hallmark of a positive apprehension test was apprehension or a sense of instability rather than pain.
The painful apprehension test had been previously described as a possible indication of an underlying subtle anterior shoulder instability in the overhand or throwing athlete presenting with pain [4][5][6] . The unstable painful shoulder (UPS) and multidirectional instability (MDI) are two forms of shoulder instability that can present with pain that are often missed 6 . UPS can occur without an apparent history of glenohumeral dislocation or subluxation 7 and multidirectional instability (MDI) often do not have a history of trauma 6 , making instability as a cause of shoulder pain particularly difficult to diagnose.
Farber et a1 1 reported that the specificity and overall accuracy of the anterior apprehension test decreased significantly when pain rather than apprehension was the diagnostic criteria used in the clinical test. The co-existence of a variety of other pathologies such as a partial rotator cuff tear, rotator cuff tendinitis, biceps tendinitis, posterior labral tear or superior labral anterior posterior (SLAP) tear can similarly cause a painful apprehension test which may be a cause for the reduced specificity of this clinical test 8 . A dedicated study on this subject of patients with a painful apprehension test has never been carried out to the authors' knowledge.
The objectives of this study were to (i) evaluate the clinical relevance of the painful anterior apprehension test, (ii) correlate the painful apprehension test with MRI and arthroscopic findings and (iii) understand the underlying patho-mechanism of the painful apprehension test.

MATERIALS AND METHODS
This is a retrospective review of all patients that underwent arthroscopic anterior Bankart repair between 2014-2016 at a tertiary hospital. All surgeries were performed by two fellowship-trained shoulder surgeons during the study timeframe. Ethical approval by the hospital's ethics committee was secured for this study.
All patients with surgical stabilisation of an anterior Bankart lesion confirmed on arthroscopy were included. Patients with concomitant surgery involving rotator cuff tears, biceps tendon pathology and SLAP lesions were excluded as these were confounding factors with pain generators addressed in the same setting affecting outcome correlation. Patients with bony Bankart lesions, glenohumeral osteoarthritis and previous ipsilateral shoulder surgery were also excluded as they were potential pain generators that could confound results of the study.
All patients were interviewed and baseline demographic data collected. These included age, gender, arm dominance, smoking history, sporting activity and occupation. Presenting complaints of pain, number of dislocations and severity of the trauma incurred at the first dislocation were recorded.
Besides a general shoulder examination, the Beighton score for joint laxity was also recorded. Patients with Beighton score ≥4 were deemed to have generalised joint laxity or hypermobility. The anterior apprehension test was then conducted and apprehension or pain during the manoeuvre was recorded. The study cohort was divided into three groups according to the test results: (i) apprehension test with apprehension only, (ii) apprehension test with pain only, and (iii) apprehension test with both apprehension and pain. The groups were then compared.
All patients had pre-operative plain radiographs with anterior/posterior and axillary views as well as an MRI of the affected shoulder. The presence of Bankart lesions, Hill-Sach's lesions, SLAP tears, biceps tendon lesions and rotator cuff pathology were recorded. Pre-operative Constant, ASES and SF-36 scores were documented for all patients.
All patients underwent arthroscopic assessment of the shoulder with the patient in the beach chair position. An arthroscopic Bankart repair was carried out using standard portals. A mean of 4 (range 3-5) suture anchors were used.
Post-operatively the arm was immobilised for three weeks in an arm sling. Pendulum exercises were commenced at three weeks, followed by a standardised physiotherapy protocol involving passive followed by active range of motion exercises. External rotation beyond neutral was not allowed until six weeks after surgery. Strengthening exercises were commenced at eight weeks post-surgery.
Post-operative Constant, ASES and SF-36 scores were documented at 6 months, 12 months, and 24 months after surgery.
Data was compiled and analysed using SPSS version 21 [SPSS Inc., Chicago, Illinois, USA]. The Chi-Square test was used for analysis of categorical variables. Continuous variables were analysed for normality using the Shapiro-Wilk test; normally distributed variables were analysed using parametric tests, otherwise non-parametric tests were used. One-way ANOVA and Tukey post hoc tests were used for univariate analysis of continuous variables. A multinomial logistics regression model was used to identify risk factors for shoulder instability in those with the painful apprehension test. Only significant covariates after univariate analysis are included in the multivariate modelling. Statistical significance was set at p <0.05.

RESULTS
There were no significant differences between all 3 groups in term of gender, mean age at presentation, handedness, smoking history, involvement in high risk occupations, participation in competitive sports or symptom duration, and severity of trauma incurred at the first dislocation (Table I).
Univariate analysis showed significant differences between the 3 groups in patients with a history of traumatic shoulder dislocation (p=0.028), patients with pain as the presenting complaint (p=0.014) and patients with recurrent dislocations as the presenting complaint (p=0.046) ( Table I). Patients with the painful apprehension test were more likely to have a history of traumatic dislocation, more likely to present with pain, and less likely to present with recurrent shoulder dislocations (Table I). However, none of these factors were shown to be significant as single predictors for shoulder instability on multivariate analysis (Table II).
Similar rates of detecting Bankart lesions, Hill-Sachs lesions, rotator cuff pathology and SLAP lesions were found across all groups for both MRI and arthroscopic assessments. (Table III). On clinical examination except for the findings of the apprehension test, no other differences were noted, including the severity of joint laxity.  In this group of patients, pain was likely the protective mechanism limiting their shoulder range of motion before the shoulder could exceed its arc of stability. As such, rather than instability, they would more likely present with persistent shoulder pain and activity limitation instead. The abduction, external rotation (ABER) manoeuvre translates the humeral head antero-inferiorly causing it to "roll over" the Bankart lesion, which may not be extensive enough to destabilise the shoulder but adequately disrupted to cause pain. The universal arthroscopic finding in our series in patients with a painful shoulder was focal capsular synovitis around the frayed torn edges of the labrum, which was the likely pain generator. Repetitive ABER motion might potentially perpetuate this "roll-over synovitis" leading to the predominant symptom of pain which may distract patients from a sensation of instability. We found no significant difference in the incidence of Bankart lesions, cuff pathology or SLAP lesions between patients with and without pain on the apprehension test, making these other pathologies less likely causes of pain. Interestingly, the rates of Hill-Sachs lesions found in patients with apprehension only (49.6%) was higher than patients with who had pain (30.8%). This provided circumstantial evidence supporting our postulation that pain during the ABER manoeuvre might be protective against frank shoulder dislocations resulting in impaction fractures of the humeral head.
Furthermore, the proportion of patients with a preceding traumatic dislocating event in the group with a painful apprehension test was higher than the group with a purely apprehensive test (80.8% vs 69.9%), and the latter group was also nearly twice as likely to have ligamentous laxity than the former (14.8% vs 7.7%). This suggested a greater role of trauma, rather than ligamentous laxity, as the underlying aetiology for patients with a painful "roll-over" lesion, with predominant symptoms of pain rather than instability. These associations were not statistically significantly, and this is likely due to the underpowered nature of our study but they were certainly clinically relevant.
Boileau et al emphasised the importance of recognising UPS early in the young, hyperlax athlete complaining of deep, anterior shoulder pain as soft tissue or bony lesions indicative of instability may be found in the absence of an apparent history of shoulder subluxation or dislocation which were amenable to surgery with excellent outcomes 7 . Twenty patients were identified over a 5-year period with data collected prospectively and followed-up for a minimum of 2 years (mean 38 months, range 24-76 months Instability of the shoulder can exist in the purely painful form, with or without a history of traumatic shoulder dislocation, recurrent dislocation or ligamentous laxity. Our study shows that the painful apprehension test is common in shoulder instability, particularly in patients presenting with pain as their main symptom despite a history of traumatic shoulder dislocation. These patients tend to have lower rates of recurrent dislocation but can achieve predictable, satisfactory outcomes with arthroscopic stabilisation of their Bankart lesions. The strengths of our study are: (i) having 2 fellowshiptrained shoulder surgeons who work closely with each other ensures uniformity in clinical assessments, patient selection and quality of surgeries performed, thereby reducing the number of variables from poor standardisation, and (ii) utilisation of multiple outcome measures provides a more holistic assessment. Our study limitations include: (i) the retrospective design of the study limiting collection of more comprehensive data, and (ii) relatively small sample size.

CONCLUSION
The painful apprehension test should not be dismissed as a negative finding for shoulder instability. The underlying pain generator in this group of patients is likely the focal capsular synovitis around the torn labrum and persistent "rolling over" of the humeral head over the torn labrum perpetuates the painful synovitis. These patients are likely to present with shoulder pain and fewer episodes of dislocation. Clinicians should have a lower threshold for advanced imaging to identify these patients early.

CONFLICT OF INTEREST
The authors declare no potential conflicts of interest.