From August 1992 through November 2007, any patient undergoing shoulder surgery for any reason at our institution was prospectively entered into a database. We retrospectively reviewed that database and found, of the total of 1854 patients it had recorded, 1331 had undergone diagnostic arthroscopy with or without a subsequent open procedure; 681 of the 1331 patients had a positive Neer sign preoperatively. Of the 681 patients, we excluded 283: 263 because, during arthroscopy, observation of the cuff-glenoid contact was obscured by synovitis or was absent secondary to a large massive rotator cuff tear; and 20 with a diagnosis of frozen shoulders because the contact under anesthesia was evaluated after the manipulation. The remaining 398 patients (191 males, 207 females) had an average age of 44.2 years (range, 13–86 years). This study was approved by our Institutional Review Board.
All 398 patients underwent a thorough preoperative history, completed subjective questionnaires within 4 weeks before surgery, and had a comprehensive physical examination performed by the senior author (EGM) or under his supervision. This examination included 646 data points for motion, strength testing, and provocative maneuvers, as previously reported [6
The Neer sign was performed in the office with the patient standing. Using the technique described by Neer [28
], the examiner stabilized the affected arm by placing one hand on the shoulder to limit scapular rotation and used the second hand to elevate the patient’s arm passively in flexion with the arm in internal rotation until the patient felt pain in the anterior or lateral deltoid only. If the patient had pain elsewhere, it was not considered a positive Neer sign. For the sign to be positive, the pain had to be in the anterior or lateral portion of the shoulder near the acromion or in the deltoid region. The sign was considered negative if there was pain in the acromioclavicular joint, trapezius muscle, posterior shoulder joint line, or shoulder blade region. Once pain in the anterior or lateral deltoid was reported, the examiner stabilized the patient’s arm in that position, and a handheld goniometer then was used to determine the degree of arm flexion (Fig. ).
A goniometer is held by the examiner to determine the degrees of arm flexion where pain occurred.
All patients subsequently underwent arthroscopy of the shoulder under general anesthesia with or without a scalene block. After the induction of anesthesia, the patient was placed in a lateral decubitus position with a bean bag and was secured with towels and tape to prevent thorax motion. The arm was held in a commercially available arm holder with 10 pounds of traction. The arthroscope was placed in a standard posterior portal, and the joint was insufflated with saline via a pressurized pump set at 80 mm Hg. The senior author performed a thorough diagnostic evaluation of the glenohumeral joint, including inspection for rotator cuff and labrum abnormalities. Rotator cuff tears of the supraspinatus were stratified by severity by noting whether they were partial or full thickness [8
]. The rotator cuff tears were not débrided until the site had been inspected for cuff-glenoid contact.
The patient’s arm then was removed from the arm holder, and an operative assistant on the opposite side of the surgeon moved the arm into flexion as for a Neer sign [24
]. An assistant held the scapula with one hand, a technique similar to that used in the office examination and as described by Neer [28
]. With the arthroscope held at the posterior joint line as described previously [18
], the arm was gradually flexed until cuff-glenoid contact (between the rotator cuff and the superior labrum anterior to the biceps attachment) was noted. With the arm held in this position, the degree of elevation of the shoulder was measured with a handheld goniometer. This position was measured to the nearest 5° and represented the degrees at which the rotator cuff tendon first made contact with the superior glenoid. Next, maximum arm flexion was measured with the goniometer as the assistant fully elevated the arm in flexion [18
The final diagnosis for all patients was determined by history, physical examination, and arthroscopic findings. Radiographic studies were not used as the source of the final diagnosis in this study because all patients had undergone diagnostic arthroscopy.
We used a paired-samples Student’s t test to compare the degrees of flexion noted during the preoperative positive Neer sign with those obtained during shoulder arthroscopy. A two-tailed Pearson correlation was used to test correlation between the degrees of flexion noted during the preoperative positive Neer sign and those obtained during shoulder arthroscopy. We used a standard statistical program (SPSS® 16.0; SPSS, Inc, Chicago, IL, USA) for our analyses.