Clinical and arthroscopic findings in recreationally active patients
© Fowler et al; licensee BioMed Central Ltd. 2010
Received: 3 February 2009
Accepted: 15 January 2010
Published: 15 January 2010
To examine the diagnostic accuracy of standard clinical tests for the shoulder in recreational athletes with activity related pain.
Cohort study with index test of clinical examination and reference standard of arthroscopy.
Sports Medicine clinic in Sheffield, U.K.
101 recreational athletes (82 male, 19 female; mean age 40.8 ± 14.6 years) over a six year period.
Bilateral evaluation of movements of the shoulder followed by standardized shoulder tests, formulation of clinical diagnosis and shoulder arthroscopy conducted by the same surgeon.
Main Outcome Measurements
Sensitivity, specificity, likelihood ratio for a positive test and over-all accuracy of clinical examination was examined retrospectively and compared with arthroscopy.
Isolated pathology was rare, most patients (72%) having more than one injury recorded. O'Brien's clinical test had a mediocre sensitivity (64%) and over-all accuracy (54%) for diagnosing SLAP lesions. Hawkins test and Jobe's test had the highest but still not impressive over-all accuracy (67%) and sensitivity (67%) for rotator cuff pathology respectively. External and internal impingement tests showed similar levels of accuracy. When a positive test was observed in one of a combination of shoulder tests used for diagnosing SLAP lesions or rotator cuff disease, sensitivity increased substantially whilst specificity decreased.
The diagnostic accuracy of isolated standard shoulder tests in recreational athletes was over-all very poor, potentially due to the majority of athletes (71%) having concomitant shoulder injuries. Most likely, this means that many of these injuries are missed in general practice and treatment is therefore delayed. Clinical examination of the shoulder should involve a combination of clinical tests in order to identify likely intra articular pathology which may warrant referral to specialist for surgery.
Athletic injuries to the shoulder are both common and unique, being both difficult to diagnose and manage . The majority of shoulder injuries occur from direct or indirect trauma or as a result of repetitive use  with the most frequently presenting shoulder conditions at primary care being rotator cuff pathology, glenohumeral and acromioclavicular joint disorders and pain referral from the neck . The most commonly reported shoulder pathology in older athletes is rotator cuff pathology [1, 2] with 85% of patients at primary care level being diagnosed with this pathology . Obtaining a specific diagnosis, in addition to determining the prognosis and appropriate treatment strategy for a patient is essential  particularly considering the resolution of shoulder conditions is extremely poor with 41% of primary care patients having persistent pain 12 months post initial consultation . Fundamental to shoulder disorder diagnosis is the lack of accepted diagnostic criteria . Whilst numerous clinical examination tests for the shoulder exists, some of which were specifically designed to identifying specific disorders, such as O'Brien's test for labral lesions , difficulty exists in diagnosing and differentiating shoulder disorders by one single test .
Whilst the diagnostic accuracy of various clinical tests in identifying specific shoulder pathologies has been widely discussed in previous research [8–12], some of the research only selected and analysed patients with a specific condition in mind . Consequently the probability of a test diagnosing a patient who has the pathology is extremely high, resulting in the test being assessed in an already biased setting. Whilst this ensures the clinical tests assessed are specific to identifying a particular pathology when that is known, it does not necessarily mean that the test is accurate to demonstrate this injury in a clinical setting when people clinically present with undiagnosed shoulder pain. For the non specialist clinician, the main aim of such clinical examination, utilizing specific shoulder tests, is to try to establish a working diagnosis and thereby distinguish between cases requiring surgical referrals and cases where conservative treatment may be sufficient. It is obvious that the more structures that are injured in the shoulder the more problematic this examination will be to interpret. This article draws attention to the fact that upon patent presentation it is difficult to surmise how many patients will present with "isolated and specific injuries" or have more complex scenarios.
To our knowledge no study has examined clinical tests for the shoulder undertaken on consecutive recreational athletes within a routine clinical setting, without any exclusion criteria. The definition of a recreationally active patient for inclusion in this study was; a person who took part in competitive sport, of a non-professional nature, on three or more occasions a week. The aim of this paper was to assess the diagnostic accuracy of routine clinical tests for the shoulder in consecutive recreationally active patients, presenting in a sports medicine clinic.
A total of 101 recreationally active patients were examined by a senior orthopaedic consultant and after examination underwent arthroscopic surgery between January 2001 and December 2006. All data was subsequently entered into a standardized database and retrospectively analysed as a consecutive subjects design. Prior to assessment, consent was obtained in writing from each patient and ethical approval was obtained from the University of Sheffield, U.K. The same orthopaedic surgeon conducted all physical examinations and arthroscopies.
Bilateral evaluation of all active, passive and resisted movements of the shoulder was a pre-requisite to the physical assessment. A battery of routine shoulder tests were incorporated into the examination in all patients; these were O'Brien's test , Jobe's test , Hawkins-Kennedy test , Palm-up test , Apprehension-relocation test  and Gerber's lift-off test .
Following subjective and physical examination findings, a clinical diagnosis was formulated. Where deemed appropriate, further investigation was prescribed involving either an x-ray or MRI scan. An arthroscopy was undertaken later by the same consultant, whereby all procedures were recorded onto DVD. The operative findings were routinely documented, citing all intra articular structural pathology and outcomes of functional intra-operative dynamic assessments referring to internal and external impingement. Shoulder arthroscopy was undertaken as a day case, with no complications reported in any patient. All patients were referred post-operatively for physiotherapy, within the orthopaedic surgeon's supportive team.
All patient documentation from initial consultation to subsequent discharge, was systematically audited and analysed, retrospectively, by an unbiased observer, uninvolved in any stage of the treatment given to patients. The data was analysed using the software, Statistical Package for the Social Sciences (SPSS; version 15.0) where assessment of diagnostic test accuracy for all tests in various pathologies was calculated using a 2 × 2 table. Test sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), likelihood ratio for a positive test and overall accuracy were calculated .
There was 101 recreational athletes with shoulder pathology (82 male, 19 female), mean age 40.8 (+/- 14.6 years) with 58% (n = 59) presenting with right shoulder pain. The patients, who were all keen participants in sport, competing in their chosen activity three or more times per week in addition to being employed full-time, were regarded as being recreationally active. A total of 92% of the cohort complained of activity-related pain whilst pain on overhead activities was reported in 89% (n = 90) of patients.
The average time from injury to initial orthopaedic consultation was 35.6 weeks with 75% of patients attending physiotherapy prior to the assessment. All 101 recreational athletes underwent arthroscopy, waiting an average of 3.6 weeks from orthopaedic consultation to surgery. Of the entire recreational cohort, 71% (n = 72) of patients had more than one shoulder pathology present, as verified by arthroscopy. The most common operative finding was a grade II or III SLAP lesion, evident in 50.4% (n = 51) of patients with only 7 patients having an isolated SLAP lesion. Rotator cuff pathology, defined as partial under surface tears or complete tears, with or without macroscopic degenerative changes, was detected in a total number of 35 patients (35%) with only three patients presenting an isolated rotator cuff injury. Thirty-one cases (30%) of internal impingement (characterised by pinching of flap tears from the SLAP and/or under surface rotator cuff and/or subscapularis between the humeral head and the glenoid on passive overhead simulation-patients in beach chair position) were reported with none occurring in isolation. External impingement (characterised by compression of the rotator cuff tendons against the acromion on passive abduction and internal rotation) was present in 28% (n = 28) of patients and was the most common operative finding occurring in isolation with 11 reports of this pathology. No normal arthroscopies were reported.
Diagnostic Accuracy of Clinical Tests
Over-all diagnostic values of clinical tests for shoulder pathology in recreational athletes.
Clinical preoperative test
Over-all Accuracy (%)
Rotator cuff pathology
The sensitivity and specificity of tests utilised to diagnose SLAP lesions, within this study, ranged from 12% to 64% and 36% to 87% respectively (Table 1). O'Brien's test had mediocre sensitivity (64%), PPV (56.1%) and over-all accuracy (54.3%).
Gerber's lift off test reported highest specificity (86.7%), whilst Palm-up test had highest NPV (55.6%) for SLAP lesions. The highest likelihood ratio for a positive test in patients with a SLAP lesion was 1.13 for O'Brien's test (Table 1).
Diagnostic accuracy when one of the combinations of shoulder clinical examination tests utilised to diagnose SLAP lesions as verified by arthroscopy, is positive during clinical examination, in recreational athletes.
Over-all accuracy (%)
Jobe's + Palm-up + O'Brien's
Jobe's + O'Brien's
Palm-up + O'Brien's
Palm-up + Jobe's
Rotator Cuff Pathology
Rotator cuff pathology, as verified by arthroscopy, was not diagnosed accurately over-all by Hawkins test (67%) (Table 1), whilst the likelihood ratio was the highest of all tests at 2.06. This clinical test also showed a PPV of 53%, NPV of 76% and sensitivity of 58%; the latter of which was preceded slightly by Jobe's test (67%). Specificity ranged from 35% to 90% for the shoulder tests, the latter of whish corresponded to Gerber's lift-off test.
Diagnostic accuracy when one of the combinations of shoulder clinical examination tests utilised to diagnose rotator cuff pathology, as verified by arthroscopy, is positive during clinical examination, in recreational athletes.
Over-all accuracy (%)
Hawkins + Jobe's + O'Brien's
Hawkins + Jobe's + Palm-up
Hawkins + O'Brien's + Palm-up
Jobe's + Palm-up + Obrien's
Hawkins + O'Brien's
Jobe's + O'Brien's
Hawkins + Jobe's
Palm-up + Jobe's
Hawkins + Palm-up
Palm-up + Obrien's
Diagnostic accuracy when one of the combinations of shoulder clinical examination tests utilised to diagnose labral tears, as verified by arthroscopy, is positive during clinical examination, in recreational athletes.
Over-all accuracy (%)
Jobe's + O'Brien's
Jobes + Palm-up + O'Brien's
Palm-up + O'Brien's
Palm-up + Jobes
Over-all diagnostic values of clinical tests for internal and external impingement in recreational athletes.
Clinical preoperative test
Over-all accuracy (%)
Jobe's test was the most sensitive test, with the highest NPV at diagnosing external impingement in a pathological shoulder (76% and 83% respectively), being 16% more sensitive than Hawkins test; the latter of which had the highest specificity, PPV and likelihood ratio (70%, 42% and 2.0 respectively) (Table 5).
In the entire recreational cohort, the Apprehension test had high sensitivity and specificity values for diagnosing Bankart lesions (79% and 87% respectively) and Hill Sachs lesions (81.5% and 80.7% respectively), in addition to having high likelihood ratios for both conditions (6.16 and 4.23 respectively) (Table 1).
Clinical examination tests aim to screen, diagnose, grade and monitor the progression of a disease . The accuracy of numerous clinical tests for the shoulder with respect to arthroscopic findings is extremely variable in both this study and previous literature [8–10, 12]. The primary focus of this paper was to determine the diagnostic accuracy of standard clinical tests for shoulder pathology in consecutive recreationally active patients with shoulder pain in relation to arthroscopic findings. This patient category is common in most general practices around the UK and their injuries are, in our experience, often missed and not referred for orthopaedic specialist evaluation. The fact that most patients suffered from more than one pathology may blur the perception and clinical diagnosis, which may lead to wrong conclusions for further treatment; in particular requirements for surgery.
A SLAP lesion is a relatively unique condition, with disagreement amongst authors as to the most provocative test to identify this pathology [10–12]. Whilst numerous authors have previously discussed the diagnostic accuracy of O'Brien's clinical test for the shoulder [10–12], it must be noted that the substantial diagnostic accuracy reported by O'Brien et al  for this test was based on its ability to assess labral abnormalities. Parentis et al  analysed O'Brien's test specific to SLAP lesions and reported adequate sensitivity (62.5%) as did this current study (64%). An observation of O'Brien's test being more sensitive than specific is also evident between this investigation and the aforementioned study, a finding contradictory to the findings of McFarland et al (sensitivity 47%, specificity 55%). However, McFarland et al had the same over-all accuracy value for O'Brien's test diagnosing SLAP lesions (54%) as that reported here. Whilst this study cited findings similar to previous research for assessing O'Brien's accuracy in diagnosing SLAP lesions, several discrepancies exist between these studies.
McFarland et al reported the diagnostic tests for SLAP lesion identification (compression/rotation test, O'Brien's and anterior slide test) were positive at almost the same rate in the control and SLAP lesion groups utilised in their study, consequently querying the validity of O'Brien's test as a diagnostic aid for this pathology. However, McFarland et al allocated patients with Type I SLAP lesions into the control group of the study, as the authors regarded this condition as being degenerative and un-related to mechanical symptoms. Considering Kim et al reported 74% of 139 patients suffering from SLAP lesions were Type I, McFarland et al appear to have excluded a substantial proportion of the population, although neither study restricted their subject population to recreational athletes. Parentis et al utilised a less extreme exclusion criteria omitting patients with adhesive capsulitis and those who were unable to complete a physical examination. This investigation however, analysed all patients irrespective of severity and type of shoulder pathology, including all individuals with SLAP lesions.
Despite O'Brien's test having on average, the highest sensitivity, likelihood ratio, PPV and over-all accuracy for diagnosing SLAP lesions in this study, these values are mediocre and consequently this study is unable to validate this test, or any other, as a single accurate, reliable tool for diagnosing SLAP lesions. Considering 50% of patients had SLAP lesions in this study, with 86% having concomitant injuries with the SLAP lesion, it is feasible that it was going to be unachievable to diagnose SLAP lesions utilising only O'Brien's test. No one test utilised in the clinical examination in this study, attained high values in all diagnostic categories thereby suggesting that no single test is both sensitive and specific in diagnosing SLAP lesions, a conclusion supported by Parentis et al . Oh et al  recommended combining two relatively sensitive clinical tests and one relatively specific clinical test as a means of increasing the diagnostic accuracy for detecting SLAP lesions.
With Hawkins clinical test being considered as one of the most commonly used tests to assess for external shoulder impingement , of which rotator cuff pathology is a potential contributor to this condition , this test is subsequently discussed in relation to rotator cuff disease. Clinical tests utilised to diagnose rotator cuff pathology are generally more sensitive than specific  a fact evident in the findings of MacDonald et al  whereby sensitivity exceeded specificity by 45% for Hawkins test, whilst Bin Park et al  reported a similar trend (sensitivity 72%, specificity 66%). This particular investigation had contradictory trends with specificity being 15% higher than sensitivity. Remarkably similar over-all accuracy results for Hawkins test and rotator cuff disease are evident however between this study and that of Bin Park et al  (67% and 69.7% respectively), and a likelihood ratio of 1.85. However, Bin Park et al  utilised a patient population that had the diagnosis of subacromial syndrome; a condition defined in the study as patients with partial and full rotator cuff tears, bursitis, or rotator cuff tendonitis. Patients suffering from impingement with concomitant injuries such as SLAP lesions or instability were excluded. Whilst this ensures the diagnostic accuracy of shoulder tests are specific to rotator cuff pathology, or bursitis, it nonetheless excludes a vast majority of patients from a study. Murrell and Walton  reported several clinical tests for the shoulder as being unable to distinguish between patients with rotator cuff tears and alternate shoulder pathologies. Over-all, this study cannot promote the use of one single clinical test to specifically diagnose rotator cuff pathology.
A result to possibly note is the high specificity and NPV associated between Hawkins test and internal impingement (66% and 73% respectively). Considering specificity and NPV relate to a test being able to predict the absence of pathology in patients without the disease, it is unsurprising Hawkins attained high scores in these categories. However, as Pappas et al  reported that internal impingement can be elicited by both Neers and Hawkins tests, it may explain why the latter test obtained 60% over-all accuracy for internal impingement of the shoulder. The role of Hawkins test in diagnosing internal impingement is unclear and warrants further investigation. It may simply be due to the proximity between the SLAP and under surface of the rotator cuff compressing an impinging flap during the test. This finding may also be of importance in respect to the relatively common situation whereby a suspected subacromial impingement is treated arthroscopically but not successfully, due to a missed SLAP tear impingement in the glenohumeral joint.
Although a number of clinical tests used for the diagnosis of painful shoulder are considered accurate in determining the location of the periarticular lesions, these entities may be difficult to differentiate by physical examination . The inability of isolated clinical tests for the shoulder, to assess for specific shoulder disorders is evident in this study, with no one test being significantly outstanding for detecting any condition. This study further demonstrated that a positive response gained in one of a combination of clinical tests caused test sensitivity to increase substantially in all pathological conditions, with specificity subsequently plummeting.
A positive response in one or more of Jobe's, O'Brien's or Palm-up tests, when assessed together, was the most sensitive for diagnosing intra articular pathology such as both SLAP lesions (94%) and labral tears (95%). Jobe's and O'Brien's test elicited the same sensitivity result when diagnosing labral tears (95%) in addition to being only 2% lower in sensitivity for SLAP lesions. Whilst sensitivity values are extremely impressive when the clinician reports one or more of a battery of tests to be positive for various pathologies, specificity is subsequently extremely poor with Jobe's and O'Brien's tests being only 45% and 18% specific for labral and SLAP lesions respectively. Even though the exact diagnosis is vague, surely a finding like this would render a referral to orthopaedic surgeon with the query of arthroscopic intervention.
Diagnosing rotator cuff pathology using a positive response in one or more of either Jobe's, O'Brien's or Palm-up tests resulted in sensitivity of 91%. The substitution of O'Brien's test with Hawkins clinical test resulted in a 3% and 11% increase in sensitivity and over-all accuracy respectively. The high sensitivity values were accompanied by extremely poor specificity results. With minimal differences observed between the aforementioned results for sensitivity, this highlights not only the ability of these tests in detecting the presence of shoulder pathology, but more importantly, emphasises their ineffectiveness at differentiating different shoulder pathologies.
The majority of previous research analysing the diagnostic accuracy of clinical tests for the shoulder, implemented varying degrees of exclusion criteria for patients. Previous research eliminated patients with concomitant shoulder injuries [8, 10] which is potentially unwise considering the wide spectrum of shoulder disease which may present in a patient, as evident in this study. Having a large variation in subject pathology in this study gives a good over-view as to how the clinical tests for the shoulder may perform in other professional settings .
Despite clinical tests being an integral part of the over-all examination procedure  utilising these tests in isolation or combination, in recreational athletes, must be cautioned, considering the poor diagnostic accuracy presented in this study. Considering that clinical diagnostic tests for the shoulder are most often undertaken in general practice and can be positive in the presence of other conditions , as evident in this investigation, the clinician should consider the results of the examination on the basis of the clinical presentation and subjective history of the patient. It is recommended that a systematic evaluation, including a combination of tests be utilised to establish an accurate working diagnosis for shoulder injuries  of which positive findings within these tests may warrant further investigations by a specialist.
The present study involved a retrospective review of the information collected by one surgeon. Although the surgeon did not have access to the results of the clinical tests at the time of the surgery, lack of independence of clinical tests and surgical findings, introduces a potential bias related to blindness. This study involved only surgical patients. Patients with potential pathology who had negative tests and did not have surgery were not included in analysis. This may indicate a verification or work-up bias, causing inflation of sensitivity and underestimating specificity.
No funding was received for this study.
- Owens S, Itamura J: Differential diagnosis of shoulder injuries in sports. Oper techn sport med. 2000, 8: 253-257. 10.1053/otsm.2000.17773.View ArticleGoogle Scholar
- Mitchell C, Adebajo A, Hay E, Carr A: Shoulder pain: diagnosis and management in primary care. BMJ. 2005, 331: 1124-1128. 10.1136/bmj.331.7525.1124.View ArticlePubMedPubMed CentralGoogle Scholar
- Oster A, Richards C, Prevost A, Speed C, Hazelman B: Diagnosis and relation to general health of shoulder disorders presenting to primary care. Rheumatology. 2005, 44: 800-805. 10.1093/rheumatology/keh598.View ArticleGoogle Scholar
- van der Heiijden G: Shoulder disorders: a state of the art review. Balliers clin rhem-. 1999, 13: 287-309.View ArticleGoogle Scholar
- van der Windt D, Koes B, Boeke A, Deville W, de Jong B, Bouter L: Shoulder disorders in general practice: prognostic indicators of outcome. Br J Gen Pract. 1996, 46: 519-523.PubMedPubMed CentralGoogle Scholar
- Hawkins R, Bokor D: Clinical evaluation of shoulder problems. The Shoulder. Edited by: Rockwood C, Matsen F. 1998, Philadelphia: WB Saunders Company, 164-197. 2Google Scholar
- Brox J: Shoulder pain. Best pract res cl rh. 2003, 17: 33-56. 10.1016/S1521-6942(02)00101-8.View ArticleGoogle Scholar
- Bin Park H, Yokota A, Gill H, Rassi G, McFarland E: Diagnostic accuracy of clinical tests for the different degrees of subacromial impingement. J Bone Joint Surg Am. 2005, 87A: 1446-1455. 10.2106/JBJS.D.02335.View ArticleGoogle Scholar
- MacDonald R, Clar P, Sutherland K: An analysis of the diagnostic accuracy of the Hawkins and Neer subacromial impingement signs. J Shoulder Elbow Surg. 2000, 9: 299-301. 10.1067/mse.2000.106918.View ArticlePubMedGoogle Scholar
- McFarland E, Kim T, Savino R: Clinical assessment of three common tests for superior labral anterior-posterior lesions. Am J Sports Med. 2002, 30: 810-815.PubMedGoogle Scholar
- O'Brien S, Pagnani M, Fealy S, McGlynn S: The active compression test: a new and effective test for diagnosing labral tears and acromioclavicular joint abnormality. Am J Sports Med. 1998, 26: 610-613.PubMedGoogle Scholar
- Parentis M, Glousman R, Mohr K, Yocum L: An evaluation of the provocative tests for Superior Anterior Posterior lesions. Am J sports Med. 2006, 34: 265-268. 10.1177/0363546505279911.View ArticlePubMedGoogle Scholar
- Jobe F, Jobe C: Painful athletic injuries of the shoulder. Clin Orthop. 1983, 173.Google Scholar
- Hawkins R, Kennedy J: Impingement syndrome in athletes. Am J Sports Med. 1980, 8: 151-157. 10.1177/036354658000800302.View ArticlePubMedGoogle Scholar
- Gilcreest EL: Dislocation and elongation of the long head of the biceps brachii. An analysis of 6 cases. Ann Surg. 1936, 104: 118-138. 10.1097/00000658-193607000-00012.View ArticlePubMedPubMed CentralGoogle Scholar
- Gerber C, Krushell R: Isolated rupture of the subscapularis muscle. Clinical features in 16 cases. J Bone Jt Surg Br. 1991, 73: 389-394.Google Scholar
- Duffy M, Munro B, Jacobsen B: Statistical Methods for Health Care Research. 2005, Philadelphia: Lippincott, Williams and Wilkins, 5Google Scholar
- Deeks J: Systematic reviews of evaluations of diagnostic and screening tests. BMJ. 2001, 323: 157-162. 10.1136/bmj.323.7305.157.View ArticlePubMedPubMed CentralGoogle Scholar
- Kim T, Queale W, Cosgarea A, McFarland E: Clinical features of the different types of SLAP lesions: an analysis of one hundred and thirty-nine cases. J Bone Joint Surg Am. 2003, 85A: 66-71.Google Scholar
- Oh J, Kim Y, Kim W, HS Gong, Lee J: The evaluation of various physical examinations for the diagnosis of Type II superior labral anterior and posterior lesions. Am J Sports Med. 2008, 36: 353-359. 10.1177/0363546507308363.View ArticlePubMedGoogle Scholar
- Akuthota VCL, Drake DF, Nadler SF, Toledo SD: Sport and performing arts medicine. 2. Shoulder and elbow overuse injuries in sports. Arch Phys Med Rehabil. 2004, 85: S52-S58. 10.1053/j.apmr.2003.11.010.View ArticlePubMedGoogle Scholar
- Murrell G, Walton J: Diagnosis of rotator cuff tears. The Lancet. 2001, 357: 769-770. 10.1016/S0140-6736(00)04161-1.View ArticleGoogle Scholar
- Pappas G, Blemker S, Beaulieu C, McAdams T, Whalen S, Gold G: In vivo anatomy of the Neer and Hawkins sign positions for shoulder impingement. J Shoulder Elbow Surg. 2006, 15: 40-49. 10.1016/j.jse.2005.04.007.View ArticlePubMedGoogle Scholar
- Nelson M, Leather G, Nirschi R, Pettrone F, Freedmann M: Evaluation of the painful shoulder. J Bone Jt Surg Am. 1991, 73: 707-716.Google Scholar
- Holtby R, Razmjou H: Accuracy of the Speed's and Yergason's tests in detecting biceps pathology and SLAP lesions: Comparison with arthroscopy findings. Arthroscopy. 2004, 20: 231-236.View ArticlePubMedGoogle Scholar
- Calis M, Akgun K, Birtane M, Karacan I, Calis H, Tazum F: Diagnostic values of clinical diagnostic tests in subacromial impingement. Ann Rheum Dis. 2000, 59: 44-47. 10.1136/ard.59.1.44.View ArticlePubMedPubMed CentralGoogle Scholar