The subjects were selected during the period April - November, 2009, from those attending the Department of Rehabilitation Medicine, who complained of pain in the shoulder and upper extremities and were diagnosed as suffering from unilateral shoulder impingement syndrome on the basis of their history and a physical examination. Patients, who were experiencing pain in the shoulder and upper extremities through trauma, fracture, bone deformities, calcification, or shoulder joint instabilities, were excluded. All patients with a restriction in their range of motion during passive joint movement were also excluded. The patients selected as candidates for this study were those, who had positive results in all three of the following physical examinations: 1) the positive painful arc sign, which causes pain at a certain angle during flexion and abduction of the shoulder joint and then loss of the pain when the angle is increased, 2) the Neer and Hawkins-Kennedy tests, 3) pain induced in the resisted isometric external rotation, abduction and elbow extension test or pain in the Jobe test. A thorough examination of the contralateral shoulder was performed to rule out any patients with bilateral shoulder impingement syndrome.

Patients diagnosed with unilateral shoulder impingement syndrome, who had rheumatism, a systemic inflammatory disease, peripheral neuropathy, a history of stroke, prior surgery in the shoulder joint or cervical vertebrae, pain in the cervical vertebrae or restriction in the movement of the cervicothoracic vertebrae, spinal disease such as scoliosis, or adhesive capsulitis, were also rejected as study participants.

The following measurements were made on each of the 21 patients finally selected as study subjects. The shoulder gradients were obtained by having each subject stand on an angulometer with feet spread 10cm apart between both of the medial malleoli, and their heels were placed in a straight line. The ankles were fixed in a neutral position, and the knees completely extended. The relative heights of the shoulders were assessed by measuring the difference between the angles formed by both blades of the angulometer being placed on the upper, lateral side of the acromion of each shoulder (Fig. 1).

The acromiohumeral intervals were obtained from anteroposterior radiographs of both shoulders (Fig. 2). The rotation of the scapula was also obtained from the radiographs and involved measuring the angle formed by the vertical line and the line connecting the superior angle and the inferior angle. The angle was given a positive value if the scapula was rotated upwards (Fig. 3-A) and a negative value if the scapula was rotated downwards (Fig. 3-B). The average of three measurements was used.

The data were analyzed with SPSS for Windows (version 13.0). Differences between higher and lower shoulders and between shoulders affected or unaffected by impingement were compared in terms of acromiohumeral intervals and angles of scapular rotation using paired t-tests. Statistical significance was defined as a p-value≤0.05.

The 21 subjects (6 males and 15 females) were all right-handed, and their average age was 37.7 years (Table 1).

Thirteen subjects were affected with shoulder impingement syndrome in the right shoulder, and eight subjects were affected in the left. Whereas 16 subjects were affected in the relatively lower level shoulder, only five people were affected in the relatively higher level shoulder (Table 2). This observed relationship between the prevalence of impingement syndrome and the relative height of the affected shoulder was statistically significant (Fisher's exact probability test, p<0.05).

While the mean acromiohumeral interval in the relatively lower shoulder was 10.03±1.28 mm, the mean in the relatively higher shoulder was 10.46±1.50 mm, but the difference was not statistically significant. Similarly, while the mean Acromiohumeral interval of the affected shoulder was 9.98±1.29 mm, the Acromiohumeral interval of the unaffected shoulder was 10.51±1.48 mm, but again this difference failed to reach statistical significance (Table 3).

The mean rotational angle of -0.31±3.73 degrees in the relatively lower shoulder was significantly less (p<0.05) than the mean of 3.85±4.42 degrees in the relatively higher shoulder. Also, the shoulders affected by impingement syndrome had a mean angle of 0.00±3.72 degrees, which was significantly less (p<0.05) than the mean angle of unaffected shoulders (3.54±4.71 degrees) (Table 4).