Correlation of findings in clinical and high resolution ultrasonography examinations of the painful shoulder

Abstract

Objective

High resolution ultrasonography is a non-painful and non-invasive imaging technique which is useful for the assessment of shoulder pain causes, as clinical examination often does not allow an exact diagnosis. The aim of this study was to compare the findings of clinical examination and high resolution ultrasonography in patients presenting with painful shoulder.

Methods

Non-interventional observational study of 100 adult patients suffering from unilateral shoulder pain. Exclusion criteria were shoulder fractures, prior shoulder joint surgery and shoulder injections in the past month. The physicians performing the most common clinical shoulder examinations were blinded to the results of the high resolution ultrasonography and vice versa.

Results

In order to detect pathology of the m. supraspinatus tendon, the Hawkins and Kennedy impingement test showed the highest sensitivity (0.86) whereas the Jobe supraspinatus test showed the highest specificity (0.55). To identify m. subscapularis tendon pathology the Gerber lift off test showed a sensitivity of 1, whereas the belly press test showed the higher specificity (0.72). The infraspinatus test showed a high sensitivity (0.90) and specificity (0.74). All AC tests (painful arc II^a, AC joint tenderness^b, cross body adduction stress test^c) showed high specificities (^a0.96, ^b0.99, ^c0.96). Evaluating the long biceps tendon, the palm up test showed the highest sensitivity (0.47) and the Yergason test the highest specificity (0.88).

Conclusion

Knowledge of sensitivity and specificity of various clinical tests is important for the interpretation of clinical examination test results. High resolution ultrasonography is needed in most cases to establish a clear diagnosis.

Introduction

The prevalence of shoulder problems is high⁽¹⁾ and increases with age⁽²⁾. Shoulder disorders constitute the third most common musculoskeletal presentation in the general practice⁽³⁾ and they can lead to absenteeism from work, inability to perform social activities and serious economic hardship for affected individuals and their families. In 1995, 4.2 million days of sick leave were attributed to disorders of the upper limb and neck area in the UK⁽⁴⁾.

Because shoulder disorders can develop into chronic conditions^(5–7), it is essential to identify the exact pathology to select the most appropriate treatment.

Shoulder pain can originate in various structures of the shoulder joint or in affected periarticular structures⁽⁸⁾. Comprehensive history taking and clinical examination is therefore of great importance. However, with clinical examination alone often an exact diagnosis cannot be made. The high resolution ultrasonography (HRUS) provides a non-painful, noninvasive, cost-efficient and fast imaging technique which is increasingly used to evaluate patients with musculoskeletal disorders.

Up to now, there are only few studies addressing the correlation between clinical and HRUS examination findings of the painful shoulder^{(9, 10)}, although the application of HRUS is well established in the examination of the musculoskeletal system⁽¹¹⁾. In our study sensitivity and specificity of most common clinical tests for the assessment of a painful shoulder are determined using HRUS as the gold standard.

Patients and methods

Between August and December 2012, 100 adult patients referred for evaluation of unilateral shoulder pain were included in the study, either in the rheumatology department of the University hospital of Zurich or the Sonography Institute Bethanien in Zurich. Exclusion criteria were shoulder fractures, prior shoulder joint surgery and prior shoulder injection (local anesthetics or steroids) in the past month. All patients went through a routine clinical and HRUS examination.

Approval of the local ethics review board Zurich was obtained and all patients had to sign an informed consent. The study was performed according to good clinical practice and carried out in compliance with the Helsinki Declaration.

The HRUS examination was conducted according to the guidelines of the Swiss Society of Ultrasound Medicine (SGUM, musculoskeletal section)⁽¹²⁾ which are compatible with the Musculoskeletal Ultrasound Technical Guidelines from the European Society of Musculoskeletal Radiology (ESSR)⁽¹³⁾ and the guidelines of the European League Against Rheumatism (EULAR)⁽¹⁴⁾. The diagnostic criteria for HRUS findings are listed in Table 1 and Figure 1 shows several examples of HRUS pathologies of the shoulder.

Tab. 1.

Diagnostic criteria for HRUS findings used in the study

Shoulder structure	HRUS finding	Tear-Grade	Diagnostic criteria
Bursa	Bursitis		Abnormal hypoechoic (relative to subdermal fat, but it sometimes may be isoechoic or hyperechoic) tissue that is nondisplaceable and poorly compressible. May exhibit Doppler signal
	Effusion		Abnormal hypoechoic (relative to subdermal fat, but it sometimes may be isoechoic or hyperechoic) material, that is displaceable and compressible. Does not exhibit Doppler signal
Rotator cuff tendons	Tendon calcification		Hyperechoic echotexture with or without an acoustic shadow (depending on the amount of calcification) within the rotator cuff
	Tendinosis		Thick swollen tendon with hypoechoic echotexture. Partial interruption may occur inside the tendon. Irregularity of fibrillar pattern, fragmentation, and focal hypoechoic or hyperechoic areas
	Tendon tear	Partial thickness	Hypoechoic zone or focus within the rotator cuff. Focal cuff thinning on the articular or deltoid located side
		Intramural	Hypoechoic zone or focus within the rotator cuff. Focal cuff thinning neither on the articular nor on the deltoid located side
		Full thickness	Complete loss of tendon substance with visualization of the cuff margins. Naked tuberosity, nonvisualization of the rotator cuff with approximation of the deltoid muscle to the surface of the humeral head
AC joint	Joint space narrowing		Narrowing of the space of the two articular forming bones (comparison with the contralateral side)
	Osteophytes		A step-up bony prominence at the end of the normal bone contour, or at the margin of the joint seen in two perpendicular planes, with or without acoustic shadow
	Joint effusion		Extension of the joint capsule filled with anechoic or inhomogeneously hypoechoic fluid (comparison with the contralateral side)
Long biceps tendon	Tendon sheath effusion		Extension of the joint capsule filled with anechoic or inhomogeneously hypoechoic fluid (comparison with the contralateral side)
	Subluxation		Visualization of a subluxated long biceps tendon
	Tendon tear		Complete loss of tendon substance with visualization of a empty tendon sheath

Fig. 1.

Examples of HRUS pathologies of the shoulder ^{(43, 44)} A. Shoulder lateral longitudinal view. No pathology. M. supraspinatus tendon (star), humerus (triangle). B. Shoulder anterior transverse view. Bursitis subacromialis. Effusion (star), synovial proliferation (arrow). C. Shoulder lateral longitudinal view. Tendon calcification. M. supraspinatus tendon (star), calcification with distal ultrasound cancellation (arrow). D. Shoulder lateral transversal view. Intramural tear. M. supraspinatus tendon (star), intramural tear (arrow). E. Schoulder lateral longitudinal view. Partial thickness tear. M. supraspinatus tendon (star), partial thickness tear (triangle). F. Shoulder lateral longitudinal view. Full thickness tear. M. supraspinatus tendon (star), zone of tear (triangle). G. Shoulder frontal view. AC joint osteoarthritis. Clavicula (star), acromion (plus sign), osteophyte (triangle). H. Shoulder anterior transversal view. Long biceps tendon sheath effusion. Long biceps tendon (star), effusion (arrow)

The HRUS examination was performed by two expert sonographers, both medical doctors with several years of experience performing HRUS examination of the musculoskeletal system.

For HRUS examination an Esaote MyLab50XVision with a linear 12 MHz transducer, respectively a Philips HD 11 with a linear 12–15 MHz transducer and a GE Voluson E6 with a 5–14 MHz linear array transducer were used.

Clinical examination was carried out by one trained clinician (rheumatology fellow). The tests included in the clinical examination were:

• bursitis sign;

• Jobe supraspinatus test;

• painful arc I and II;

• drop arm test;

• Hawkins and Kennedy impingement test;

• Gerber lift off test;

• belly press (abdominal compression) test;

• infraspinatus test;

• acromioclavicular (AC) joint tenderness;

• cross body adduction stress test;

• Abbott-Saunders test;

• palm up test (= speed bicepsor straight arm test);

• Yergason test;

• Hueter sign (Table 2).

Tab. 2.

Overview of the clinical examinations according to Buckup et al.⁽¹⁵⁾

Test	Procedure	Assessment
Bursitis sign	The examiner palpates the anterolateral subacromial region with his or her index and middle fingers	Localized tenderness to palpation in the subacromial space suggests irritation of the subacromial bursa
Jobe supraspinatus test	With the elbow extended, the patients arm is held at 90° of abduction, 30° of horizontal flexion, and in internal-neutral and external rotation. The examiner exerts pressure on the upper arm during th e abduction and horizontal flexion motion	Where the test elicits pain and the patient is unable to abduct the arm 90° and hold it against gravity, this indicates a tear or pathology of the supraspinatus tendon, or muscle
Painful arc I	The arm is passively and actively abducted from the rest position alongside the trunk	Pain occurring in abduction between 70° and 120° is a sign of a lesion of the supraspinatus tendon, which becomes impinged between the greater tubercle of the humerus and the acromion in this phase of the motion. Patients are usually free of pain above 120°
Drop arm sign	The Patient is seated and the extended arm passively abducted 90°. The Patient is instructed to hold the arm in this position without support and then slowly lower it	Weakness in maintaining the position of the arm, with or without pain, or sudden droppping of the arm suggests a m. supraspinatus lesion
Hawkins and Kennedy impingement tests	The examiner immobilizes the scapula with one hand while the other hand adducts the patient's 90°-forward-flexed and internally rotated arm (moving it toward the contralateral side of the body)	Pain indicates a positive test for supraspinatus pathology
Gerber lift off test	The patient places the dorsum of the hand on his or her back with the arm in internal rotation. The patient then lifts the hand away from the back, the examiner should apply a load, pushing the hand toward the back to test the strength of the subscapularis and to test how the scapula acts under dynamic loading	Where a tendon rupture or insufficiency of the subscapularis is present, the patient will be unable to lift the hand off the back against the examiner's resistance. Where pain renders maximum internal rotation impossible, the belly press test may be performed
Belly press test	The patient's forearm lies along the abdomen with the elbow flexed. The patient attempts to continue forcefully pressing arm against abdomen	A tear in the supraspinatus tendon results in loss of the internal rotation component. The elbow deviates laterally and posteriorly under the influence of the latissimus dorsi and teres major. Flexion also occurs in the wrist
M. infraspinatus test	The patient's arms should hang relaxed with the elbows flexed 90° but not quite touching the trunk. The examiner places his or her palms on the dorsum of each of the patient's hands and then asks the patient to externally rotate both forearms against the resistance of the examiners hands	Pain or weakness in external rotation indicates a disorder of the infraspinatus (external rotator). As infraspinatus tears are usually painless, weakness in rotation strongly suggests a tear in this muscle
Painful arc II	The patient's arm is passively and actively abducted from the rest position alongside the trunk	Pain in the acromioclavicular joint occurs between 140° and 180° of abduction
AC Joint tenderness	The examiner palpates the acromioclavicular joint	Localized tenderness in the acromioclavicular joint indicates pathology
Cross body action	The 90° abducted arm on the affected side is forcible adducted across the chest toward the normal side.	Pain in the acromioclavicular joint suggests joint pathology
Abbott–Saunders test	The patient's arm is externally rotated and abducted about 120° with progressive internal rotation. The examiner slowly lowers the arm from this position. The examiner guides this motion of the patients arm with one hand while resting the other on the patients shoulder and palpating the bicipital groove with the index and middle finger	Pain in the region of the bicipital groove or a palpable or audible snap suggest a disorder of the biceps tendon (subluxation sign)
Palm up test	The patient's arm is extended in supination at 90° of abduction and 30° of horizontal flexion. The patient attempts to either maintain this position or continue to abduct and pronate the arm against the downward pressure of the examiners hand	A positive test elicits increased tenderness in the bicipital groove especially with the arm supinated and is indicative of bicipital tendinitis or tendinosis
Yergason test	The patient's arm is alongside the trunk and flexed 90° at the elbow. One of the examiners hands rests on the patients shoulder and palpates the bicipital groove with the index finger while the other hand grasps the patients forearm. The patient is asked to supinate the forearm against the examiners resistance. This places isolated tension on the long head of the biceps tendon	Pain in the bicipital groove is a sign of a lesion of the biceps tendon, its tendon sheath or its ligamentous connection via the transverse ligament. The typical provoked pain can be increased by pressing on the tendon in the bicipital groove
Hueter sign	The patient is seated with the arm extended at the elbow and the forearm in supination. The examiner grasps the posterior aspect of the patient's forearm. The patient is then asked to flex the elbow against the resistance of the examiners hand	In a rupture of the long head of the biceps tendon, the distally displaced muscle belly can be observed as a “ball” directly proximal to the elbow

Each test had the option of being positive or negative⁽¹⁵⁾. Table 3 shows the “assumed corresponding shoulder structure” for each clinical test. These shoulder structures were directly evaluated in the HRUS examination.

Tab. 3.

Assumed corresponding shoulder structure of the clinical tests

Clinical test	Assumed corresponding shoulder structure
Bursitis sign	Bursa subacromialis
Jobe supraspinatus test	M. supraspinatus
Painful arc I
Drop arm test
Hawkins and Kennedy impingement test
Gerber lift off test	M. subscapularis
Belly press test
Infraspinatus test	M. infraspinatus
AC joint tenderness	AC joint
Painful arc II
Cross body adduction stress test
Abbott–Saunders test	Long biceps tendon (luxated)
Palm up test	Long biceps tendon
Yergason test
Hueter sign

Neither the clinical examiner nor the HRUS examiner had information about the history of the patients and both were blinded to the results of each other's examinations.

Statistical calculations were done with IBM SPSS Statistics 21 and Microsoft Excel 2010.

Using HRUS findings as the gold standard, the following diagnostic values were calculated: sensitivity, specificity, efficiency, positive predictive value, negative predictive value, positive likelihood coefficient, negative likelihood coefficient, Youden index, empirical Pearson's correlation coefficient and uniweighted Cohen's κ coefficient.

Empirical Pearson's correlation coefficient was used as an indicator of linear dependence between the two variables (clinical examination test results and HRUS examination findings respectively). In order to measure the correlation in terms of inter-rater agreement, the uniweighted Cohen's κ coefficient was used.

Results

The median age of the patients included in the study was 54, female was the predominant gender and the right shoulder was more affected than the left one (cf. Table 4).

Tab. 4.

Baseline characteristics of the subjects (n = 100)

Characteristic	Patients (n = 100)
Age (years)	mean 53.5 ± 14.3 (range 20–84)
	median (50% percentile) 54
Sex (male/female)	41/59
Affected shoulder site (r/l)	57/43

HRUS examination revealed a variety of pathologic findings (cf. Table 5). Pathologies of the subacromial bursa were found in 87%, of the m. supraspinatus tendon in 67%, of the AC joint in 24%, of the long biceps tendon in 20%, of the m. subscapularis tendon in 11%, of the m. infraspinatus tendon in 10% and of the m. pectoralis major tendon in 1% of the cases. All pathologic subacromial bursae showed signs of inflammation. The most frequent pathology of the rotator cuff (m. supraspinatus^d, m. subscapularis^e and m. infraspinatus^f) was tendon calcification (^d51%, ^e72%, and ^f60% of each tendon pathologies), besides tendon tear and tendinosis. Tendon calcification occurred more frequent in females (male/female = 0.32) and were found more frequently in patients older than 30 and generally correlated positively with age. Eighty-five percent of all tendon calcifications were located in the m. supraspinatus. The pathologically affected AC joints mostly showed joint space narrowing and osteophytes. Tendon sheath effusion was the predominant finding in the HRUS examination of the long biceps tendon. Patients older than 45 years showed more long biceps tendon pathologies than the younger and isolated (no other pathologies at the shoulder except bursitis and AC joint pathologies) long biceps tendon pathologies was more often seen in males (male/female = 1.50).

Tab. 5.

HRUS examination findings (n = 100)

Shoulder structure	Pathology	Tear grade	Patients (n = 100)
Bursa subacromialis			87
	Bursitis		87
M. supraspinatus			69
	Tendon calcification		35
	Tendinosis		23
	Tendon calcification + tendinosis		7
	Tendon tear		29
		Partial thickness	20
		Intramural	2
		Full thickness	7
M. subscapularis			11
	Tendon calcification		8
	Tendinosis		2
	Tendon calcification + Tendinosis		2
	Tendon tear		2
		Partial thickness	1
		Intramural	0
		Full thickness	1
M. infraspinatus			10
	Tendon calcification		6
	Tendinosis		2
	Tendon tear		2
		Partial thickness	1
		Intramural	0
		Full thickness	1
AC joint			24
	Joint space narrowing		17
	Osteophytes		15
	Joint effusion		2
Long biceps tendon			20
	Tendon sheath effusion		17
	Tendinosis		3
	Subluxation		1
	Tendon tear		1
M. pectoralis major			1
	Tendon tear		1

Table 6 gives an overview of the diagnostic values of clinical examinations compared with HRUS examination findings (any pathology).

Tab. 6.

Diagnostic values of clinical examinations compared with HRUS examination findings

Shoulder structure	Clinical examination	Sensitivity	Specificity	Positive predictive value	Negative predictive value	Empirical Pearson's corr. coeff.	Uniweighted Cohen's kcoeff.
Bursa subacromialis	Bursitis sign	0,09	1	1	0,14	0,11	0,03
M. supraspinatus	Jobe supraspinatus test	0,81	0,55	0,8	0,57	0,36	0,36
	Painful arc 1	0,83	0,35	0,74	0,48	0,2	0,19
	Drop arm test	0,12	1	1	0,34	0,2	0,08
	Hawkins and Kennedy impingement test	0,86	0,45	0,78	0,58	0,33	0,33
M. subscapularis	Gerber lift off test	1	0,55	0,22	1	0,34	0,21
	Belly press test	0,73	0,72	0,24	0,96	0,3	0,24
M. infraspinatus	Infraspinatus test	0,9	0,74	0,28	0,99	0,41	0,33
AC joint	Painful arc II	0,25	0,96	0,67	0,8	0,31	0,27
	AC joint tenderness	0,38	0,99	0,9	0,83	0,52	0,45
	Cross body adduction stress test	0,38	0,96	0,75	0,83	0,44	0,4
Long biceps tendon (luxated)	Abbott-Saunders test	1	0,99	0,5	1	0,7	0,66
Long biceps tendon	Palm up test	0,47	0,75	0,31	0,86	0,2	0,19
	Yergason test	0,32	0,88	0,38	0,85	0,21	0,2
	Hueter sign	0,05	1	1	0,81	0,2	0,04

corr. coeff. = correlation coefficient

Bursitis sign has a very low sensitivity (0.09) and a specificity of 1. The uniweighted Cohen's κ coefficient shows a very low correlation (0.03).

The m. supraspinatus tests (Jobe supraspinatus test^g, painful arc I^h and Hawkins and Kennedy impingement testⁱ) show high sensitivities (^g0.81, ^h0.83, and ⁱ0.86) and low specificities (^g0.55, ^h0.35, and ⁱ0.45) for any pathology. As opposed to this, the drop arm test^j has a specificity of 1 and a low sensitivity (0.12) for any pathology of the rotator cuff. All m. supraspinatus tests have only moderate correlation as measured by the values of the empirical Pearson's correlation coefficient (^g0.36, ^h0.20, ⁱ0.33, and ^j0.20) and only weak correlation as measured by the uniweighted Cohen's κ coefficients (^g0.36, ^h0.19, ⁱ0.33, and ^j0.08). Comparing the sensitivity of the m. supraspinatus tests (Jobe supraspinatus test, painful arc I, drop arm test, and Hawkins and Kennedy impingement test) versus the grade of tendon tear [partial thickness (+), intramural (++), and full thickness (+++)], it was observed, that the higher the tendon tear grade, the higher the sensitivity was. Only in the case of the Hawkins and Kennedy impingement test the sensitivity of detecting partial thickness tendon tears was higher (0.80) than the sensitivity of detecting full thickness tendon tears (0.71). The drop arm test showed a high specificity (0.98) and sensitivity (0.86) for detecting full thickness supraspinatus tendon rupture. The Hawkins and Kennedy Impingement test showed a high sensitivity (0.86) and a specificity of 1 to detect a supraspinatus tendon rupture.

To detect m. subscapularis tendon pathology the Gerber lift off test has a sensitivity of 1, the belly press test a high specificity (0.73).

The infraspinatus test showed a high sensitivity (0.90) and specificity (0.74). Forty-one hundredths is the empirical Pearson's correlation coefficient and 0.33 the uniweighted Cohen's κ coefficient.

All m. supraspinatus tests (Jobe supraspinatus test^g, painful arc I^h, Hawkins and Kennedy impingement testⁱ, drop arm test^j) provided high sensitivities (^g0.91, ^h0.86, ⁱ1, ^j0.86) and low specificities (^g0.43, ^h0.17, ⁱ0.15, ^j0.17), in order to detect tendon calcification. The other rotator cuff tests (Gerber lift off test^m, belly press testⁿ and infraspinatus test^o) show low sensitivities (^m0.41, ⁿ0.39, and ^o0.46) and high specificities (^m1, ⁿ0.91, and ^o0.96).

Tendon tears were diagnosed across all age groups. Patients aged 55 to 59 years have significantly more often a partial thickness rupture than other patient groups and regarding people older than 65 years there is a significant accumulation of full thickness rupture.

All three AC joint tests (painful arc II^a, AC joint tenderness^b, cross body adduction stress test^c) show low sensitivities (^a0.25, ^b0.38 and ^c0.38) and high specificities (^a0.96, ^b0.99 and ^c0.96). AC osteophytes are in 9 out of total 13 cases (69%) associated with pathologies in the rotator cuff, either tendon tear or tendinosis or both.

To examine the long biceps tendon both tests (palm up test^k and Yergason test^l) have a high specificity (^k0.75 and ^l0.88) and a low sensitivity (^k0.47 and ^l0.32). The Abbott-Saunders test shows a sensitivity of 1 and a high specificity (0.99) in order to detect a long biceps tendon subluxation. The Hueter Sign has a low sensitivity (0.05) and a specificity of 1 to identify a long biceps tendon tear. In 18 out of total 20 patients (90%) with at least one long biceps tendon pathology (tendon sheath effusion, tendinosis, tendon subluxation or tendon rupture) a concomitant tendon tear of one or multiple rotator cuff tendons was noticed.

Discussion

The study of Naredo et al.⁽⁹⁾ showed that with the clinical tests performed in the study an exact diagnosis of the underlying shoulder pathology could not be made. Iagnocco et al. found in his study⁽¹⁰⁾ that the sonographic examination has a higher sensitivity and specificity than the clinical examination. Therefore, in our study the primary goal was to find out if the clinical tests are useful for detection of the location of the shoulder pathology and in a second step to examine whether some of the tests are specific for distinct pathologies. Furthermore in this study the clinical examination was more detailed and included a larger number of patients than in the study of Naredo et al.⁽⁹⁾

HRUS is an inexpensive, non-painful, non-invasive and fast method for a real-time and dynamic examination of the shoulder. The accuracy is comparable to MRI examination to evaluate superficial tendons and soft tissues with a high degree of resolution, including the rotator cuff and muscles, and concomitant pathology of the long biceps tendon, AC joint, and bursa subacromialis⁽¹⁶⁾. Furthermore HR US is not limited by MRI contraindications, such as patient's body habitus, claustrophobia, inability to lie flat, or implanted devices not compatible with the magnetized environment in a MRI. For this reasons HRUS was chosen as the gold standard. However, there exist occasions in which HRUS is inferior to MRI examination of the shoulder. For instance if labral and ligamentous injuries are suspected, or when underlying osseous pathology is a concern^{(17, 18)}.

Another point to be considered is the interobserver variability of the HRUS examination. Scheel et al. evaluated in the study Interobserver reliability of rheumatologists performing musculoskeletal ultrasonography: results from a EULAR ‘Train the trainers’ course⁽¹⁹⁾ an interobserve r agreement. Using modified κ index for majority agreement⁽²⁰⁾ the agreement was estimated to be 0.76 for the shoulder which was described as good. Middleton et al.⁽²¹⁾ found a high agre ement (92%) between the observers in the diagnosis of rotator cuff partial and full thickness tear. On the opposite Naredo et al.⁽²²⁾ found a low inter observer agreement (κ = 0.5) for the diagnosis of shoulder tendon lesions.

It is believed that the interobserver variabilities are low due to lacking standardization of US scanning techniques and definitions of different pathological US lesions. In this study the HRUS examinators had clear definitions of both.

Looking at previous studies, the diagnostic values of the m. supraspinatus tests for determining the location of the rotator cuff lesion are similar to this study. Leroux et al.⁽²³⁾ used surgery as th e gold standard; the sensitivity and specificity of Jobe supraspinatus test for detecting tendon lesions was 0.86 and 0.50, respectively. The lacking specificity of the Jobe supraspinatus test may be explained by the fact that a present pathology of the long biceps tendon, the adjacent rotator cuff tendons or a neuropathy of the suprascapular nerve result in a false-positive test result.

The sensitivity and specificity of painful arc I calculated by Park et al.⁽²⁴⁾ was 0.71 and 0.47 al so using surgery as the gold standard. A positive painful arc can be the result of complete or incomplete rotator cuff tears, swelling and inflammation as a result of bursitis and abnormality of the margin of the acromion⁽¹⁵⁾.

For detecting subacromial impingement the drop arm test showed a sensitivity of 0.04 and a specificity of 1, calculated by Caliş et al.⁽²⁵⁾ using surgery as the gold standard. In the study of MacDonald et al.⁽²⁶⁾ the sensitivity and s pecificity of Hawkins and Kennedy impingement test were 0.88 and 0.43 again using surgery as the gold standard. These two tests are not suitable for any m. supraspinatus pathologies but the drop arm test is an ideal test for detecting full thickness supraspinatus tendon rupture. In patients in which a tendon rupture is suspected, the Hawkins and Kennedy Impingement Tests are highly sensitive and specific for supraspinatus tendon tear.

Regarding the sensitivity of 1 but low specificity (0.55) of the Gerber lift off test and the high specificity (0.72) of the belly press test, the Gerber lift off test is suitable as screening test and the belly press test may serve as a confirmatory test for m. subscapularis tendon pathology.

Regarding other studies, Gerber and Krushell⁽²⁷⁾ calculated a sensitivi ty of 0.98 for the Gerber lift off test and Barth et al.⁽²⁸⁾ calculated a specifici ty of 0.98 for the belly press test. Both studies used surgery as the gold standard in their statistical analysis. These values suggest that the belly press test may be valuable as a confirmatory test for subscapularis muscle tear after a positive Gerber lift off test result.

The infraspinatus test showed a high sensitivity and specificity. In the study of Walch et al.⁽²⁹⁾ the diagnostic values of the external rotation lag sign, which is an alteration of the infraspinatus test, were calculated using surgery as the gold standard. A sensitivity and specificity of each 0.98 to detect m. infraspinatus muscle tear was shown. Therefore the alteration of the infraspinatus test may be diagnostic of an infraspinatus muscle tear.

In this study, it is shown that patients, aged 55 to 59 years, have significantly more often a partial thickness rupture than other patient groups. Regarding people older than 65 years, there is a significant accumulation of total thickness rupture. Similar results have been reported in the study of Sher et al.⁽³⁰⁾ in which the prevalenc e of partial-thickness tears increased significantly with age. Tendon ruptures in older individuals are more often full thickness tears, traumatic ruptures in younger individuals are mainly partialthickness tears⁽³¹⁾.

Tendon calcification was the most frequently detected pathology of the rotator cuff (42% of all cases). Female was the predominantly affected gender (male/female = 0.32), as has been reported previously⁽³²⁾. However, a further ex planation for this female predominance has not yet been worked out. With regard to location, the m. supraspinatus was involved in 85.4% of the cases. Age was found to be associated with the frequency of tendon calcifications. The m. supraspinatus tests have high sensitivities and the m. subscapularis and m. infraspinatus tests have high specificities to detect calcification in their corresponding tendon. Only 35% to 45% of the individuals with tendon calcification are symptomatic⁽³³⁾. An asymptomatic tendo n calcification hardly seems to provoke a positive test. In contrast, HRUS detects calcification in asymptomatic patients. A positive power Doppler signal was shown to correlate with symptomatic tendon calcifications⁽³⁴⁾. In our study Doppler signal was not used in the HRUS examination.

The HRUS examination findings of AC osteophytes with consequent rotator cuff pathologies as tendon tear and tendinosis are correlated with the according test results: only in 4 cases out of total 13 (30%) AC osteophytes occurred without tendinosis and tear. Therefore in cases with AC osteophytes rotator cuff tendon tears or tendinosis should be suspected as AC osteophytes can mechanically damage the rotator cuff⁽³⁵⁾.

As with AC osteophytes, a strong correlation between long biceps tendon pathology and rotator cuff tears was found. Eighteen out of 20 patients (90%) with biceps pathologies showed one or multiple rotator cuff tears. In cases of rotator cuff incompetence the bicipital tendon acting as a humeral head depressor has to compensate for the ruptured rotator cuff. Also, the presence of a complete rotator cuff tear exposes the intra-articular portion of the bicipital tendon to the overlying acromion and further impingement ⁽³⁶⁾. Therefore if a pathologic long biceps tendon is observed one should carefully look for a rotator cuff pathology.

All three AC test results (painful arc II^a, AC joint tenderness^b, cross body adduction stress test^c) showed low sensitivities (^a0.25, ^b0.38 and ^c0.38). This can be explained by the fact that the main pathologic findings in the AC joint detectable by HRUS; joint space narrowing and osteophytes can be asymptomatic. Especially the asymptomatic osteoarthritis of the AC joint occurs frequently with aging ^{(37, 38)}.

Holtby et al.⁽³⁹⁾ found for the Yergason test a sensitivity of 0.43 and a specificity of 0.79, for the palm up test a sensitivity of 0.32 and a specificity of 0.75 using surgery as the gold standard. These findings are in agreement with our findings.

It is very difficult to exclusively test the long biceps tendon. The palm up test includes the anterior part of the deltoid muscle as well as the m. supraspinatus, m. subscapularis and upper part of the major pectoralis muscle in its movement. The Yergason test involves a supination movement of the forearm which can also be performed by the m. supinator. Pathologies of the rotator cuff with concomitant subacromial bursitis often lead to a long biceps tendon sheath effusion, because the subacromial bursa communicates with the long biceps tendon sheath when a full thickness tear is present.

The fact that males (male/female = 1.50) and older patients showed more isolated long biceps tendon pathologies is in agreement with other studies. Long biceps tendon tears occur more often in patients 40–60 years of age and in male patients in general ^{(40, 41)}. In the elderly degenerative tendon changes can result in ruptures without any trauma⁽⁴²⁾.

Conclusions

This study shows that knowledge of sensitivity and specificity of various clinical tests is important for the interpretation of clinical examination results and the identification of cases, for which further imaging procedures are necessary to make distinct diagnosis.

Only few clinical tests are sensitive and specific enough to allow a diagnosis without further imaging. The drop arm sign is diagnostic for full thickness m. supraspinatus tendon ruptures. If a rotator cuff rupture is suspected, the Hawkins Kennedy impingement test is diagnostic for m. supraspinatus tendon rupture. Regarding m. subscapularis tendon pathology, the Gerber lift off test is suitable as screening test and the belly press test serve as a confirmatory test. A positive AC joint test is diagnostic for any AC Joint pathology.

Conflict of interest

The authors have no conflicts of interest to declare.