A thorough physical examination was performed by using following tests:

Speed’s test: This test is most sensitive for provoking biceps tendon pain. It was performed by asking the subject to place the shoulder in 90° forward flexion position with the elbow fully extended and forearm supinated. The subject was then instructed to resist as the examiner attempted to push the patient’s arm downward. The subject with biceps tendonitis tends to complain of pain with this maneuver and exhibits difficulty resisting the examiner’s downward pressure.⁶

Resisted abduction (Jobe test): Supraspinatus was evaluated using resisted abduction (supraspinatus isolation or Jobe test). To perform the test, the examiner stood in front of the subject, who was asked to abduct the arms to 90° with the elbows fully extended. The subject’s arms were then brought forward to a position 30° anterior to the true coronal plane and maximally rotated so that the thumbs would point downwards. The subject was then asked to push toward the ceiling while the examiner provided resistance. Weakness noticed during this test signified muscle inhibition due to pain or true muscle dysfunction.⁶

Resisted external rotation for the infraspinatus tendon: To test the infraspinatus, the subject was asked to place the arms tightly at the sides with the elbows flexed to 90°. The subject was then asked to externally rotate the arms while the examiner provided resistance.⁶

Resisted internal rotation for the subscapularis tendon: The subject was asked to stand with the elbow flexed to 90° and the hand on the abdomen. The subject was instructed to press the hand down against the abdomen while the examiner attempted to lift the subject’s hand away from the trunk. This method was particularly helpful in assessing the strength of the subscapularis in patients with restricted internal rotation.⁶

Adduction stress test: In this test, the elbow and shoulder of the subject were extended and then passively adducted behind the back. This maneuver produced pain in case of acromioclavicular joint pathology. Passive cross-chest adduction was also used as a test for acromioclavicular joint symptoms. The examiner stood facing the subject and passively brought the patient’s arm into maximal cross-chest adduction. This maneuver reproduced pain if acromioclavicular joint injury or arthritis was present.⁶

Charles Neer proposed the concept of the impingement syndrome, to detect this he described Neer impingement sign. To elicit this, the examiner passively flexes the patient’s shoulder to the position of maximal forward flexion while stabilizing the patient’s scapula with the other hand. Reproduction of the patient’s symptomatic pain at maximal forward flexion is designated a positive impingement sign and is considered evidence of impingement syndrome.

All patients and controls underwent ultrasonographic examination within 1 week of physical examination with a real time equipment ( PHILIPS HDI-ATL-5000 Bothell, Washington) using 7.5MHz Linear phased array transducer. Transverse and longitudinal planes from the Biceps tendon groove, Rotator cuff and transverse planes from posterior Glenohumeral recess, Glenoid labrum and acromioclavicular joint were scanned. In all patients, comparable images of the opposite shoulder were obtained in order to facilitate detection of subtle abnormalities.

Features like tendon thickness, homogeneity of the fibrillar pattern and regularity of the margins, presence of effusion and calcification were studied.

Positioning of patient: The biceps tendon groove, the subscapularis tendon, and the acromioclavicular joint were examined with the arm held in neutral position, the elbow flexed to 90°, and the forearm in a supinated position on top of the thigh.

On the anterior aspect of the shoulder, the long head of biceps tendon was examined in both planes as a fibrillar hyperechoic structure into the humeral groove, surrounded by a 1-2 mm thick hypoechoic halo of fluid within the synovial sheath.

Medial to the biceps tendon, the hyperechoic subscapularis tendon was identified, inserting on the lesser tuberosity, with some fibers continuing across the bicipital groove to form the transverse humeral ligament, better seen when the shoulder of the patient is moved into external rotation.⁷

Next, the transducer was moved laterally to scan the rotator cuff. The supraspinatus and infraspinatus tendon were examined with the patient’s shoulder in hyperextension and internal rotation in order to expose the supraspinatus from underneath the acromion. These tendons appeared as a hyperechoic homogenous fibrillar layer, convex shaped on transverse images and curved triangular shaped on longitudinal views, deep to the deltoid muscle covering the humeral head.

The subacromial-subdeltoid bursa was imaged as a hypoechoic line, 1-2 mm thick with a variable amount of peribursal echogenic fat, between the deltoid muscle and the supraspinatus and infraspinatus tendons.

The humeral articular cartilage was seen as a thin hypoechoic layer between the supraspinatus and infraspinatus tendons and the humeral head. The posterior infraspinatus and teres minor tendons were evaluated from a posterior view with the arm in neutral position and the elbow flexed at 90°. A normal small amount of fluid was seen in the glenohumeral joint. The cartilaginous posterior labrum was viewed as a hyperechoic triangle separating the infraspinatus and teres minor tendons from the glenoid.

Impingement syndrome is evaluated by dynamic examination. ⁸ A dynamic view of the supraspinatus tendon is obtained by moving the patient’s arm from a neutral position to 90° abduction in order to detect encroachment of the acromion into the rotator cuff.

Table 1 lists the US diagnostic criteria for shoulder abnormalities, which are based on those widely described in the literature.^{7, 9, 10, 11, 12, 13}

Figure 1

Table 1- Ultrasonographic diagnostic criteria of shoulder abnormalities

US findings from the clinically evaluated painful shoulders, asymptomatic opposite shoulders and both shoulders of controls were recorded.

Statistical analysis of data was done after compiling and tabulation. Data were analyzed with the Statistical Package for the Social Sciences (SPSS, SPSS Inc software, Chicago, IL, USA). Statistical analysis was performed using the χ² test. A p value < .05 was considered to be statistically significant. The sensitivity, specificity, positive predictive value and negative predictive value of physical examination with respect to sonography were calculated and compared using a 2×2 table.

The chief complaints of patients were pain and limitation of shoulder movements. All forty cases with painful shoulders and twenty controls underwent Physical examination. Physical examination was positive in 35 painful shoulders; negative in all forty asymptomatic shoulders of cases and twenty controls.

Figure 2

Table 2: No. of +ve painful shoulders on physical examination

The speed’s test was positive in 14(35%) painful shoulders, resisted abduction test in 31 (77.5%), resisted external rotation in 1(2.5%), resisted internal rotation in 4(10%) and adduction stress test in 9(22.5%) painful shoulders.

Sonographic examination of all cases and controls was performed within 1 week of PE and the various anatomic structures were systematically reviewed. Sonography showed the presence of alteration in 33(82.5%) painful shoulders. Observations on individual structures were recorded.

Ultrasound showed supraspinatus lesions in 21 (52.50%) of painful shoulders, 2 (5%) asymptomatic shoulders and in 1 (2.50%) control. Full-thickness tear (Fig.1) was detected in 4 (10%) of painful shoulders. Partial-thickness tear (Fig.2) was seen in 10 (25%) of painful shoulders, 1 (2.50%) asymptomatic shoulder and in 1 (2.50%) control.

Tendonitis was detected in 5 (12.50%) of painful shoulders and 1(2.50%) asymptomatic shoulder. Tendonitis was not seen in controls. Calcific tendonitis and thinning of supraspinatus tendon each was seen in 1 (2.50%) of painful shoulders. Full-thickness tear, calcific tendonitis and thinning of tendon were not seen in asymptomatic shoulders and controls.

Alteration in long head of biceps tendon were found in14 (35%) painful shoulders, 2 (5%) asymptomatic shoulders and in 2 (5%) controls. The presence of fluid in tendon sheath (Fig.3) was seen in 7 (17.50%) painful shoulders and 1 (2.50%) asymptomatic shoulder. No control showed fluid in tendon sheath. Partial subluxation was seen in 3 (7.50%) of painful shoulders, 1 (2.50%) asymptomatic shoulder and 2 (5%) controls. Partial-thickness tear and tendonitis of long head of biceps, each was seen only in 1 (2.50%) painful shoulder. No control and asymptomatic shoulder showed tear and tendonitis.

Subscapularis showed alteration in 4 painful shoulders, 2 (5%) tendonitis and 2 (5%) partial–thickness tear. Subscapularis tendon was normal in all asymptomatic shoulders and controls.

Only 2 (5%) of painful shoulders and 1 (2.50%) asymptomatic shoulder showed tendonitis of infraspinatus. Infraspinatus tendon was normal in all controls.

Teres minor was normal in all study groups.

Acromioclavicular joint abnormality was found in 14 (35%) of painful shoulders, 1 (2.5%) asymptomatic shoulder and 1 (2.5%) control. Degenerative changes (osteophytes and capsular hypertrophy) alone were seen in 4 (10%) of painful shoulders, 1 (2.50%) asymptomatic shoulder and 1 (2.5%) control. Presence of fluid in joint alone was seen only in 2 (5%) painful shoulders. Degenerative changes and presence of fluid in joint both (Fig.4) were seen in 8 (20%) of painful shoulders only. Asymptomatic shoulders and controls did not show fluid alone and in combination with degenerative changes.

Abnormality in the form of fluid in the joint cavity was detected in1 (2.5%) of painful shoulder only. Fluid in the joint cavity was associated with rotator cuff tear. Glenohumeral joint did not show any abnormality in asymptomatic shoulders and controls.

Irregularity of greater tuberosity was identified in 5 (12.50%) painful shoulders and 1 (2.50%) asymptomatic shoulder. No abnormality detected in controls. Rotator cuff tear was associated with irregularity of greater tuberosity in 3 (7.50%) of painful shoulders.

Fluid in subacromial-subdeltoid bursa (Fig.1) was seen in 3 (7.50%) painful shoulders in association with rotator cuff lesion. SASD bursa was normal in asymptomatic shoulders and controls.

One painful shoulder showed buckling of supraspinatus tendon in association with SASD bursal fluid. Impingement was not seen in asymptomatic shoulders and controls.

Figure 3

Fig.1: Full-thickness tear (non visualization of supraspinatus tendon) with fluid in subacromian-subdeltoid bursa.

Figure 5

Fig.3: Transverse section of biceps tendon in the bicipital groove shows fluid in tendon sheath

Figure 6

Fig.4: Degenerative changes with fluid in the acromioclavicular joint.

Figure 7

Table 3: Sonographic findings with the spectrum of the specific alterations among the various anatomic structures of the shoulder.

Figure 8

Table 4: Sonographic findings in cases (painful shoulders) and asymptomatic shoulders of cases

Figure 9

Table 5: Sonographic findings in patients (painful shoulders) and controls.

Figure 10

Table 6: Comparison of Physical examination (PE) and Ultrasonography (US)

Figure 11

Table 7: Comparison of physical examination and ultrasonographic findings in painful shoulders.

Figure 12

Table8: Comparative analysis of findings on Physical examination by various authors

Table 7 shows the sensitivity, specificity, positive predictive value and negative predictive value of PE with respect to US in detection of above lesions. Sensitivity was good in the clinical diagnosis of supraspinatus lesion and low in other shoulder lesions, especially the infraspinatus and the acromioclavicular joint. Specificity was high for the lesion of infraspinatus, subscapularis and the acromioclavicular joint. However, it was fairly good for the biceps tendon pathology and very low for the lesion of the supraspinatus tendon.

PE was unable to differentiate rotator cuff tendonitis from tear, and partial-thickness tears from full-thickness tear. Moreover, it missed 3 SASD bursal fluids, 5 greater tuberosity irregularities, 1 rotator cuff impingement and 1 glenohumeral joint effusion.

Almost whole of the spectrum of pathologies was found in our study except for adhesive capsulitis, glenohumeral instability and laberal pathology and soft tissue tumors. Adhesive capsulitis has non-specific findings on ultrasound.¹⁷ Soft tissue tumors are usually large and extensive by the time they reach tertiary set-up and it is uncommon for them to present solely as pain as a symptom. Glenohumeral instability and labral pathologies are usually seen in patients with past history of trauma, which is one of the exclusion criteria in present study.

Rotator cuff was the most common structure to show abnormality on ultrasound in our study. 27 painful shoulders had rotator cuff abnormality out of the 40 painful shoulders i.e. 67.5%. This is in concordance with the previous studies. In study done by Hawkins et al, 1989.¹⁸ 60% of the shoulder abnormalities had been attributed to rotator cuff abnormalities. Most rotator cuff lesions involved the “critical zone” in the anterior aspect of tendons as in previous study.¹⁹

In our study, among the tendons of rotator cuff, Supraspinatus tendon most commonly showed abnormality on ultrasound (52.5%). Subscapularis was the next common tendon to show abnormality (10%). Only in 5% rotator cuff lesions Infraspinatus tendon was involved. Teres minor was normal in all the patients.

As in our study, the studies done by Iagnocco et al, 2003 and Naredo et al, 2002 also showed supraspinatus to be the most commonly involved tendon and teres minor least commonly involved. However, in our study subscapularis was more commonly showed abnormality as compared to infraspinatus. While in the studies of Iagnocco et al and Naredo et al, infraspinatus was more frequently involved. This variability could be due to different in sample size.

The reason for frequent involvement of supraspinatus tendon is because of the characteristic anatomical location of it between two bones (head of humerus and acromion).It gets compressed during forward flexion of shoulder joint. This leads to ischemia and later degeneration of the tendon.

Partial thickness tears were more common than full thickness tears in the present study. Partial thickness tear is seen in younger age group of 40-60 years than full thickness tears. The reason of finding partial thickness tear more commonly than full thickness tears in our study is due to large number o f the patents included in our study belonging to the 40-60 age group.

Brennke et al, 1992²⁰ and Hawkins et al, 1989¹⁸ studies also showed partial thickness tears to be commoner of the two. Articular surface partial-thickness tears was most common type of partial-thickness tear in our study in accordance with previous study by Holsbeeck et al, 1992.²¹

Statistically significant findings observed in partial thickness tears as compared to control shoulders were focal hypoechoic areas, focal discontinuity of fibers, focal thinning of tendon as was seen in previous studies.^{14, 18, 20,, 21}

Statistically significant findings observed in full thickness tears as compared to control shoulders were non-visualization of tendon, herniation of deltoid and subacromial-subdeltoid bursa in the cuff as was seen in previous studies.^{14,15,22,23,24}

Alterations of the long head of the biceps were found in 14 of the 40 painful shoulders (35%) which are statistically significant as compared to controls. Alterations observed were tendonitis, fluid around the tendon, partial thickness tear and partial subluxation. Rotator cuff tears were observed in 6 of the 14 patients with biceps tendon alterations. Most commonly observed pathology of biceps tendon in association with the rotator cuff tears was tendon sheath effusion. These findings are in accordance with Iagnocco et al, 2003 and Naredo et al, 2002.

Biceps tendon is subject to mechanical forces that contribute to cuff impingement because of it anterior location in the impingement area. Also, the synovial sheath of the biceps is an extension of the glenohumeral synovial membrane. Hence the frequent association of cuff tears with biceps tendon abnormalities is observed.

SASD bursal fluid was found in 3(7.5%) painful shoulder in our study which is not significant statistically as compared to controls. Out of three, two were associated with the rotator cuff tear and one with impingement of supraspinatus tendon which is in accordance with the study done by Hollister et al, 1995 (specificity 96%)²⁵ and in the study by Van Holsbeeck MT et al, 1993 (specificity 90%)²⁶ for the diagnosis of rotator cuff tears.

Fluid within both bursa and joint in the presence of a full-thickness tear of the rotator cuff is due to direct communication between the two compartments through the defect, akin to the positive findings during a single-contrast arthrogram. Fluid in a partial-thickness tear is more likely due to direct mechanical irritation of the bursa by impingement or fenestrations within the partially torn cuff, allowing the communication. In full-thickness tears, the volume of fluid within the bursa is far greater than that seen in partial-thickness tears or impingement.

Acromioclavicular joint abnormalities were found in 14 (35%) of painful shoulders in present study which is statistically significant as compared to controls. Degenerative changes (osteophytes and capsular hypertrophy) alone were seen in 4 (10%) of painful shoulders.

Presence of fluid in joint alone was seen only in 2 (5%) painful shoulders. Degenerative changes and presence of fluid in joint both were seen in 8 (20%) of painful shoulders only. The findings in our study was in accordance with the study by Iagnocco et al,2003 but ACCL joint abnormality as a cause of painful shoulder was observed less frequently as compared to the study by Naredo et al, 2002.

In our study irregularity of greater tuberosity was identified in 5 (12.50%) painful shoulders which is statistically significant as compared to controls. Rotator cuff tear was associated with irregularity of greater tuberosity in 3 (7.50%) of painful shoulders. Cortical irregularity of greater tuberosity was considered to be one of the secondary sonographic signs of full-thickness tear by Middleton WD et al, 1989²³ and Dondelinger RF et al, 1995²⁴ in their studies. In 2 painful shoulders there was no association between rotator cuff tear and irregularity of greater tuberosity in our study, the changes may be due to aging process, as 1 asymptomatic shoulder also showed irregularity of greater tuberosity.

Calcific tendonitis was seen in one (2.5%) painful shoulder in our study. Hyper-reflective lines were seen without acoustic shadowing. Calcium deposits were seen in supraspinatus tendon. MRI of this painful shoulder was normal. Absence of shadowing seen in this case may be due to soft form of stage 3 calcific tendonitis. Calcific tendonitis seen in present study occurred less commonly than the study by Naredo et al, 2003 this may be due to less mean duration of symptoms (2.8 months) in our study cohort as compared to study by Naredo et al, 2003 (mean duration of symptoms group I-3.3 months and group II-8.6 months).⁹²

In our study only one (2.5%) painful shoulder showed abnormality in the form of fluid in the joint cavity which is not significant statistically as compared to controls. Fluid in the joint cavity was associated with rotator cuff tear. This finding is less common than the studies done by Iagnocco et al, 2003(11.2%) and Naredo et al, 2003 (group I -12%, group II-7%). This may be due to different study population and different sample size and limited numbers of subjects with synovitis.

Impingement is rarely an isolated finding. There are usually secondary changes within the supraspinatus tendon or the subacromial bursa. The presence of a fluid collection alone in the subacromial-subdeltoid bursa has been used as a further criterion for the presence of early-stage impingement syndrome, i.e., Neer grade 1.

One (2.5%) painful shoulder showed buckling of supraspinatus tendon in association with SASD bursal fluid in our study which is not significant statistically as compared to controls. This finding is less common than the study done by Naredo et al, 2002. This may be due to different study population and different in age range.

From above discussion we observed that high frequency ultrasonography is an accurate, non-invasive and cheap imaging technique available to assess the shoulder lesions with high sensitivity and specificity as concluded from study by Teefy SA, et al, 2000.²⁷

In conclusion, the high sensitivity and specificity, non-invasiveness and low costs of High Resolution Sonography justify its routine utilization for the assessment of painful shoulder.