INTRODUCTION

Frozen shoulder is a condition of uncertain aetiology, characterised by substantial restriction of both active and passive movement in the shoulder without any intrinsic disorder.1 Some of its pathophysiology have been reported.2 Physiotherapy, analgesics, steroid injection, manipulation, and surgical release are all effective treatments. Adhesion of the long head of the biceps (LHB) tendon to the rotator interval was identified in patients with frozen shoulder using arthroscopy. For those unresponsive to conservative treatment, outcomes of arthroscopic capsular release targeting the LHB tendon were assessed.

MATERIALS AND METHODS

From 2003 to 2005, 87 consecutive patients aged 36 to 77 (mean, 54) years underwent arthroscopic release of 88 frozen shoulders by a single surgeon. Preoperative treatments included rehabilitation, steroid and/or hyaluronic acid injections. Pain was managed by non-steroidal anti-inflammatory drugs (NSAIDs). The inclusion criteria were severe night pain, no improvement of flexion (90°) and external rotation (0°), and poor response to rehabilitation for at least 6 months. Exclusion criteria were complete rotator cuff tear, acromioclavicular subluxation, and biceps tendon rupture noted on magnetic resonance images (MRIs).

Shoulders were divided into 3 types; types A/ B/C indicated slight/moderate/severe degree of synovitis and adhesion of the LHB tendon to the rotator interval (Fig. 1). 23 shoulders were type A, 26 type B, and 39 type C. 18 of the 39 type-C shoulders were controls with release of the capsule alone but not the LHB tendon.

The patient was placed in the beach-chair position under general anaesthesia. The range of movement in flexion, external rotation, and internal rotation were measured. A 4-mm diagnostic arthroscopy was inserted through a standard posterior portal. An anterior portal was created just superior to the subscapularis tendon using the outside-in technique to facilitate manoeuvres of shavers and a radiofrequency instrument. The severity of the adhesion of the LHB tendon to the rotator interval was assessed. A scope was moved to the subacromial space via a lateral and anterolateral portal. The synovium in the subacromial bursa was shaved and the rotator cuff checked. The scope was then moved to the glenohumeral joint to precede the capsular release. First, the adhesion of the LHB tendon to the rotator interval was released using a radiofrequency instrument and rasp. Then the joint capsule (anterior, antero-inferior, superior, and superoposterior) next to the labrum was released and moved into the glenoid neck without compromising nerves. The range of movement of the LHB from maximum internal to external rotation and recovery of sliding of the LHB tendon to the rotator interval were assessed. The shoulder was manipulated in external rotation, internal rotation, and flexion in the scapular plane. The range of movement after manipulation was measured.

Passive, active-assistive, and stooping exercises were commenced for forward flexion and external rotation one day after surgery with the aid of a physical therapist. Active exercises to strengthen the rotator cuff and scapular stabilisers were started one week later. Patients were able to return to work 4 weeks later, without any limitations to daily activity. Rehabilitation was continued for 3 months to regain complete muscle strength. No patient received steroid treatment.

Patients were followed up for a mean of 21 (range, 12-35) months. Changes in the American Shoulder and Elbow Surgeons (ASES) score, range of movement, and muscle strength (flexion and external rotation) among types A, B, C, and controls were compared using the Mann-Whitney U test. The ASES score is for patients to self-evaluate shoulder function for the activities of daily living using a visual analogue scale and a questionnaire.3 One-way analysis of variance of pain, duration of symptoms, restriction of movement, age, and blood sugar were performed. A p value of <0.05 was considered significant.

RESULTS

In type-A adhesion, the LHB tendon tended to lie downward toward a 3-o'clock position in the internal rotation at 0° of abduction, and was directed upward slightly toward a 12-o'clock position in external rotation at 0° of abduction (Fig. 2a). In type-C adhesion, this sliding movement was prevented even in external rotation (Fig. 2b). After arthroscopic capsular release, the sliding movement recovered (Fig. 2c).

The severity of the adhesion of the LHB tendon to the rotator interval was related to the ASES score (Table). Postoperatively, the differences in ASES score between control and types A, B, and C were significant (p=0.001, p=0.02, and p=0.04, respectively), as were those between types C and A/B (p=0.001/p=0.027). In all adhesion types, the range of movement in flexion, external rotation, and internal rotation improved postoperatively (Table). Flexion of types A, B and C improved significantly after surgery and remained so for 24 months compared with controls (p=0.001, p=0.023, and p=0.045, respectively). The muscle strength of types A, B, and C increased significantly 6 months after surgery compared to controls (flexion: p=0.001, p=0.034, and p=0.047; external rotation: p=0.001, p=0.021, and p=0.038, respectively; Table).

DISCUSSION

Shoulder functions depend partly on the sliding movement of the LHB tendon, especially for external rotation. MRI findings in frozen shoulder typically reveal a thickening of the coracohumeral ligament (CHL).4 CHL thickness was evident in all 3 types of LHB tendon adhesion, but MRI was of limited use in differentiating its severity, except by showing high signal intensity around the LHB tendon in T2- weighted images.

In a study on open procedures for recalcitrant frozen shoulder, all patients regained normal strength, and 94% had pain relief and complete range of movement.5 Another study used the same procedure for 25 patients, 20 of whom achieved goodto- excellent outcomes and in 3 outcomes were fair. The 2 failures were associated with severe diabetes in a patient and Dupuytren's disease in another.6 In our study, the ASES scores for the diabetic and idiopathic patients were not significantly different.

Antero-inferior release has also been reported. In a study on release of the inferior and middle glenohumeral ligaments with a follow-up of 13.5 months in 24 patients (26 shoulders), 88% were satisfied, and 76% had normal function.7 The shoulders of these patients were placed in full abduction during the procedure, and steroids were injected into the joint and subacromial space at the end of the procedure. Another study in 40 patients compared manipulation with arthroscopic capsular release (removal of synovium from the rotator interval, release of the anterior glenohumeral ligament and the intra-articular portion of the subscapularis tendon, and division of the anterior half of the inferior capsule).8 Patients having arthroscopic capsular release achieved better outcomes. In 23 patients who underwent antero-inferior release combined with manipulation and subacromial decompression, followed by physiotherapy for 48 hours under interscalene local anaesthetic blockade, the mean Constant score improved (from 13 to 77), as was the range of movement.9 In 35 patients with recalcitrant frozen shoulder who underwent arthroscopic capsular release (of the antero-inferior capsule, intra-articular portion of the subscapularis tendon, superior and middle glenohumeral ligaments, and coracohumeral ligament), 83% had normal or mildly symptomatic shoulders after a mean follow-up of 22 months. These patients received a tapered 21-day course of oral prednisolone.10 In 73 patients with arthroscopic selective capsulotomy,11 it took a mean of 2.2 weeks for pain to subside and 5.5 weeks for the range of movement to improve to within 10% of the contralateral limb. One month postoperatively is the most important period for obtaining better results by rehabilitation. 37% of the patients received steroid injections and 11% had recurrent pain. In 45 patients with arthroscopic 360° release of adhesion of the LHB tendon to the rotator interval, several patients still complained of slight night pain without NSAIDs.12 In our study, the anterior, antero-inferior, superior, and superoposterior capsules were released in addition to eliminating the LHB adhesion to the interval. Arthroscopy of the LHB adhesion revealed the mechanism responsible for the decreased shoulder function associated with frozen shoulder. Further analysis of rotator cuff dysfunction (including incomplete tears of the supraspinatus) is needed to clarify the mechanism responsible for night pain in patients with frozen shoulder.

© 2008 Western Pacific Orthopaedic Association Provided by ProQuest LLC. All Rights Reserved.