You are on page 1of 7

J Shoulder Elbow Surg (2012) -, 1-7

Efcacies of corticosteroid injection at different sites of the shoulder for the treatment of adhesive capsulitis
Sang-Jin Shin, MD, PhD*, Seung-Yup Lee, MD
Department of Orthopaedic Surgery, School of Medicine, Ewha Womans University, Seoul, South Korea
Background: A corticosteroid injection in the glenohumeral joint conducted blindly is technically demanding with a low rate of accuracy despite satisfactory clinical outcomes in the treatment for adhesive capsulitis. This study prospectively compared the clinical outcomes of patients with idiopathic adhesive capsulitis treated by a single corticosteroid injection in different locations of the shoulder. Materials and methods: We randomly assigned 191 patients with adhesive capsulitis to 1 of 4 groups based on corticosteroid injection location: group I, subacromial; group II, intra-articular; group III, intra-articular combined with subacromial space; and group IV, medication. Pain relief and patient satisfaction were assessed with a visual analog scale and functional outcomes were evaluated with the American Shoulder and Elbow Surgeons score up to 24 weeks after treatment. Results: Patients treated with corticosteroids achieved faster pain relief and had greater satisfaction levels than patients in group IV during the 16 weeks after treatment. However, no signicant difference in pain scores was observed among the 4 groups at 24-week follow-up visits (P .670). Shoulder motion and function improved in all groups at nal follow-up. However, shoulder motion in the injection groups recovered faster than that in group IV. At 24 weeks after treatment, no signicant differences in shoulder motion or functional outcomes were found among the 4 groups (P .117). Conclusions: The efcacy of a single corticosteroid injection was not found to be related to the site of injection. However, a single corticosteroid injection provided faster pain relief, a higher level of patient satisfaction, and an earlier improvement in shoulder motion and function than medication in patients with adhesive capsulitis. Level of evidence: Level II, Randomized Controlled Trial, Treatment Study. 2012 Journal of Shoulder and Elbow Surgery Board of Trustees. Keywords: Corticosteroid injection; adhesive capsulitis; subacromial space; intra-articular

Adhesive capsulitis traditionally has been regarded as a self-limiting condition with spontaneous recovery within 2 years.16 However, long-term follow-up studies have

Institutional review board approval: This article was approved by Ewha Womans University Institutional Board Review (No. ECT 12-04B-41). *Reprint requests: Sang-Jin Shin, MD, PhD, Department of Orthopaedic Surgery, School of Medicine, Ewha Womans University, 1071 Anyangcheon-ro, Yangcheon-Ku, 158-710 Seoul, South Korea. E-mail address: (S.-J. Shin).

shown that 40% of patients have persistent mild pain and shoulder motion limitation and that 11% of patients have permanent functional disability of the shoulder joint.8,19 Various treatment modalities have been introduced to reduce pain and time to healing and to enable an early return to daily activities. A single corticosteroid injection is viewed as a conservative treatment modality that achieves good clinical outcomes.2,11,15 Corticosteroid injection therapy has been advised in cases of adhesive capsulitis because it is believed that inammation plays an important

1058-2746/$ - see front matter 2012 Journal of Shoulder and Elbow Surgery Board of Trustees.

2 role in its pathogenesis. Furthermore, the majority of studies on corticosteroid injection therapy as a treatment for adhesive capsulitis have concluded that it is effective with regard to shoulder function recovery up to 12 to 16 weeks after injection compared with other treatments, including physiotherapy or placebo distension.9,10,21 When a corticosteroid injection is applied for adhesive capsulitis treatment, most physicians favor an intra-articular injection. However, without the assistance of radiologic guidance, an intra-articular injection is technically demanding. Studies on the accuracy of blind glenohumeral joint injection have reported accuracy rates of only 10% to 42%.12,22 Moreover, an inappropriate injection into the shoulder muscles or tendons can cause serious complications and nullify the therapeutic effects of the corticosteroid. On the other hand, an injection into the subacromial space is widely used to treat rotator cuff disorders and is known to be relatively straightforward with a higher injection accuracy rate than an intraarticular injection. Therefore, if similar clinical outcomes can be obtained with subacromial corticosteroid injection and with intra-articular injection, subacromial injection can be substituted for the technically demanding intra-articular injection in the treatment of adhesive capsulitis. The hypothesis of this study was that corticosteroid injection in the subacromial bursa has the same therapeutic effect and achieves the same functional outcomes as an intra-articular injection. The purposes of this randomized, controlled, prospective clinical study were to analyze the efcacy of a single corticosteroid injection administered at different sites of the shoulder in patients with primary adhesive capsulitis and to compare the clinical outcomes achieved with those of medication.

S.-J. Shin, S.-Y. Lee

to undergo one of the following treatment methods: corticosteroid injection into the subacromial space (group I), glenohumeral joint (group II), or glenohumeral joint combined with subacromial space (group III) or administration of oral NSAID medication (group IV). Randomization was implemented by use of a computer-generated random sequence by an independent researcher, and group assignment codes were shown to a physician at the time of intended treatment. Group randomization assignments were permuted after 4 assignments to balance group demographics. A sample size in each group was determined beforehand by use of power and sample size calculation. Power analysis indicated that a total number of 36 patients per group would provide a power of 80% (b .20, a .05) to detect signicant differences in American Shoulder and Elbow Surgeons (ASES) shoulder scores, assuming a mean difference of 10 points and an SD of 15 points. These assumptions were based on means and standard deviations of ASES shoulder scores observed in a pilot study of 20 patients. To account for possible loss to follow-up of 30%, we enrolled at least 47 patients per group. Of the 191 patients, 18 (9%) were excluded because they were lost to follow-up during the 24 weeks after treatment, and 15 patients were excluded because of a treatment method change during the study period (Fig. 1). After a diagnostic shoulder ultrasound examination performed to evaluate concomitant pathologic lesions, patients were administered a single glenohumeral joint and/or subacromial space injection by 1 physician. All injections were performed through a posterior approach under ultrasonographic guidance. The corticosteroid injection mixture was composed of 4 mL of 2% lidocaine and 40 mg of triamcinolone (1 mL). In group III, this dosage was equally divided between the glenohumeral joint and subacromial bursa. Patients in group IV were administered oral aceclofenac (100 mg) twice daily for 6 weeks. All patients were followed up using the same standardized rehabilitation protocol by a physical therapist weekly for the rst month and then were asked to follow a simple home exercise program. Passive pendulum and gentle rangeofshoulder motion exercises with the assistance of a bar were started from 7 days after treatment. Active-assisted rangeofshoulder motion exercises began at 6 weeks in patients receiving instructions not to exceed their pain thresholds, and resisted shoulder motion exercises were started at 3 months. All patients were assessed with a shoulder function assessment form based on the ASES scoring system before treatment and at 2, 4, 8, 16, and 24 weeks after treatment. In addition, pain integrity and patient satisfaction with treatment were measured with visual analog scales. Active range of shoulder motion, including forward exion and external rotation at the side, was assessed with a goniometer, and internal rotation was assessed by use of the spinal levels for the affected and nonaffected shoulders. Patients who regained range of shoulder motion to within 15 of the contralateral side for both forward exion and external rotation at the side and internal rotation to within 2 spinal levels of the contralateral side were considered to have recovered. Shoulder range of motion was measured to the point of pain with a goniometer. Data regarding clinical evaluations and functional outcomes were collected blindly by a physicians assistant not otherwise involved in the study. A repeated-measures analysis of variance with the Tukey post hoc test and c2 test was used to identify associations between initial and outcome parameters by use of SPSS software (SPSS, Chicago, IL, USA). P < .05 was considered signicant.

Materials and methods

From April 2007 to July 2009, 208 consecutive patients with a diagnosis of primary adhesive capsulitis were enrolled in this study. Our institutional review board approved the study protocol, and informed consent was obtained from all participants. The inclusion criteria applied were as follows: age of 18 years or older, shoulder pain with limitation of both active and passive shoulder movement in at least 2 directions (forward exion <120 or 50% restriction of contralateral external rotation and internal rotation), duration of shoulder pain greater than 3 months, and the availability of nonspecic radiographic and ultrasound ndings of the affected shoulder. This study excluded patients found to have a shoulder disorder in the subacromial space or the glenohumeral joint, those with bilateral adhesive capsulitis, those with a history of shoulder injury or surgery, and those with an arthritic change of the glenohumeral joint. Thirteen patients were excluded because of partial- or fullthickness rotator cuff tears found during ultrasound examination, and four patients chose not to participate. The remaining 191 patients with primary adhesive capsulitis were treated by a corticosteroid injection or with nonsteroidal anti-inammatory drugs (NSAIDs). These 191 patients were randomly divided into 4 groups

Efcacies of corticosteroid injection

Figure 1 Flow diagram of patient recruitment based on CONSORT (Consolidated Standards of Reporting Trials) criteria. IA, intraarticular; SA, subacromial.

Table I

Clinical characteristics of 4 groups at baseline I (SA) II (IA) 42 55.1 4.6 16/26 7.4 3.4 30/12 III (SA IA) 39 56.3 5.8 14/25 7.0 2.6 29/10 IV (PO) 36 57.3 6.4 13/23 6.8 2.7 26/10 P value .183 .986 .580 .857 41 53.9 4.1 14/27 7.7 3.3 27/14

No. of patients Age (mean SD) (y) Sex (M/F) Duration of symptoms (mean SD) (mo) Side (dominant/nondominant)

F, female; IA, intra-articular; M, male; PO, oral medication; SA, subacromial.

The patients demographic and clinical characteristics at baseline are summarized in Table I. Mean age, gender distribution, and dominant shoulder involvement were similar in the 4 groups. No signicant demographic differences were observed among the 4 groups. The visual analog scale score for pain was signicantly improved after treatment in all groups at nal follow-up (Fig. 2). However, pain relief was achieved signicantly faster after a corticosteroid injection (groups I, II, and III) than after starting oral medication (group IV), and this difference was maintained for up to 16 weeks: 1.4 0.5 in group I, 1.4 0.4 in group II, 1.2 0.8 in group III, and 3.1 0.5 in group IV (P < .05). Among the 3 corticosteroid injection groups, group III showed the fastest pain relief, followed by group I and then group II; however, the differences were not signicant. At 24 weeks follow-up, pain was gradually improved in group IV and no signicant intergroup differences were found (P .670). The corticosteroid injection groups showed signicantly better patient satisfaction for up to 16 weeks (P .022) (Fig. 3). High patient satisfaction in these 3 groups was achieved

within 2 weeks of treatment. Although patients treated with corticosteroids showed better early satisfaction with treatment, no signicant differences were observed among the 3 corticosteroid groups and group IV after 24 weeks followup (P .07). At nal follow-up, mean range of shoulder motion improved in all 4 groups. Forward exion of patients in groups I, II, and II was signicantly more rapidly restored than that of patients in group IV regardless of injection site for up to 16 weeks after treatment (P < .05) (Fig. 4). Forward exion in patients who received a corticosteroid injection increased by a mean of 12.4 3.1 before physiotherapy started. However, no signicant difference was observed in ranges of forward exion among the 4 groups at nal follow-up evaluations (P .117). External rotation and internal rotation improvements followed a similar pattern in the 4 groups throughout the 24-week follow-up (Figs. 5 and 6). No signicant intergroup differences were observed in baseline ASES scores (Table II). All shoulder functional outcomes recovered signicantly in the 4 groups at nal follow-up (P < .05). Although shoulder scores were higher in the corticosteroid injection groups than in group IV at each visit up to 16 weeks

S.-J. Shin, S.-Y. Lee

Figure 2 Visual analog scale (VAS)assessed pain improved signicantly faster in the corticosteroid injection groups (groups I, II, and III) than in group IV and remained lower for up to 16 weeks after injection (asterisk, P < .05). No signicant difference was observed among the 4 study groups at nal follow-up. IA, intraarticular; PO, oral medication; SA, subacromial.

Figure 4 Forward exion was restored signicantly faster in the corticosteroid injection groups (groups I, II, and III) than in group IV, and this difference was maintained for up to 16 weeks (asterisk, P < .05). No signicant differences were observed among the 4 groups at nal follow-up. IA, intra-articular; PO, oral medication; SA, subacromial.

Figure 3 Patients were signicantly more satised from 2 weeks after corticosteroid injection than with NSAIDs, and this difference was maintained for up to 16 weeks (asterisk, P < .05). No signicant differences were observed among the 4 groups at nal follow-up. IA, intra-articular; PO, oral medication; SA, subacromial; VAS, visual analog scale.

Figure 5 External rotation improved regardless of treatment method at nal follow-up evaluations. No signicant differences were observed among the 4 groups during serial follow-up evaluations. IA, intra-articular; PO, oral medication; SA, subacromial.

(P .036), no intergroup difference was observed at 6 months follow-up (P .651). No infections related to corticosteroid injection occurred around the shoulder joint. Three patients showed temporary skin color changes around the injection site, and a steroid are reaction developed in seven patients.

In this study, the use of corticosteroid injection led to more rapid pain relief, better functional outcomes, and higher patient satisfaction for up to 16 weeks after injection than

oral medication. However, patients treated with corticosteroid injection in the glenohumeral joint and the subacromial space achieved similar clinical outcomes at each visit. Although adhesive capsulitis is a common disorder, its optimum treatment remains controversial. In the majority of cases, a stiff shoulder responds well to conservative treatments, which include oral NSAIDs, oral corticosteroids, glenohumeral intra-articular corticosteroid injections, and physical therapy.2,3,7,12,14,15 Of these treatments, intraarticular corticosteroid injection achieves better clinical outcomes than the other treatments. Lorbach et al15 concluded that an intra-articular injection of corticosteroid showed superior shoulder motion restoration compared with oral corticosteroid. Patients treated with intra-articular corticosteroid injection with physical therapy achieved faster symptom relief than those who underwent physical

Efcacies of corticosteroid injection

5 However, the effect of corticosteroids on pain and satisfaction was more marked than those on shoulder range of motion, especially rotation. External rotation and internal rotation improved at a lower rate than forward exion regardless of the treatment method used and required continuous stretching exercise. In this study, patients injected in the subacromial space showed similar functional recoveries to patients injected in the glenohumeral joint. Previous studies compared only 2 different injection sites and found that there were no signicant differences in clinical outcomes between an intrabursal and intra-articular injection.17,20 We added another group of patients who were treated with both intraarticular and subacromial corticosteroid injections to compare clinical outcomes of patients who were treated with separate injections. However, we could not nd any better benecial effects after simultaneous intra-articular and subacromial injection. It is interesting that shoulder function restoration is similar after a subacromial or intraarticular injection because adhesive capsulitis mainly provokes the glenohumeral capsular inammatory changes that trigger brosis and proliferative myelobrosis.5,8 Although adhesive capsulitis primarily affects glenohumeral capsular tissue, histologic ndings proved that contractures of the coracohumeral ligament and rotator interval are also the main lesions in chronic adhesive capsulitis.18 In addition, release of the rotator interval appears to be an essential part of the operative treatment of adhesive capsulitis in terms of obtaining successful restoration of shoulder motion.6,24 The coracohumeral ligament, which reinforces the rotator interval, lies supercial to the superior glenohumeral ligament and may be encountered in the subacromial space. Therefore, the use of a subacromial injection is considered to have a positive effect on adhesive capsulitis. Andrieu et al1 found that the subacromial space is almost invariably involved in adhesive capsulitis and suggested an adjuvant subacromial corticosteroid injection in patients who do not respond to an intra-articular injection. However, in our study, no additional benets were found after combined intra-articular and subacromial injection compared with an isolated injection in either space. A subacromial injection is technically easier than an intra-articular injection when conducted blindly. However, a subacromial injection of corticosteroid has potential side effects, especially when injected into muscle or tendon, because collagen bers are weakened and late rotator cuff ruptures may occur when corticosteroids are inadvertently injected directly into the rotator cuff tendon. NSAIDs have potent anti-inammatory properties and are used to treat tendinitis of the rotator cuff and adhesive capsulitis.13,14 Although literature on the effect of oral NSAIDs for adhesive capsulitis treatment is sparse, the use of NSAIDs has never been shown to improve pain or function as compared with a placebo.10 However, in this study, we found that an oral aceclofenac dose of 100 mg daily combined with physical therapy restored shoulder

Figure 6 Internal rotation improved regardless of treatment method at nal follow-up evaluations. No signicant differences were observed among the 4 groups during serial follow-up evaluations. IA, intra-articular; PO, oral medication; SA, subacromial.

therapy only.23 In our study, patients treated with corticosteroid injection showed faster pain relief and shoulder functional recovery than patients treated with oral medication. However, this study showed that the benets of corticosteroid were not maintained beyond 16 weeks after injection. Similar to the results of our study, the majority of studies have concluded that injected corticosteroid is probably of limited short-term benet in adhesive capsulitis cases. Bulgen et al4 showed that the initial response to treatment was most marked in patients treated with corticosteroids but found no signicant long-term outcome difference versus patients treated with physical therapy or benign neglect. Another study found that a patient treated with an intra-articular corticosteroid had a satisfactory functional outcome at 6 weeks after injection compared with patients treated with physiotherapy only; however, this difference disappeared at 16 weeks after treatment.21 Early treatment with corticosteroids may chemically ablate synovitis and thus limit the subsequent development of brosis and shorten the natural history of the disease.9 Despite the short-term benets of corticosteroid injection, we considered it worthwhile to use corticosteroid for the treatment of adhesive capsulitis. One of the benecial effects of corticosteroid injection is pain relief, and reduced pain facilitates further treatment. Patients are satised with the fast effects of the injection treatment, because most have had shoulder pain for a long time. In our study, patient satisfaction after a corticosteroid injection was higher even though pain and shoulder function were no different from those achieved while patients were receiving oral medication. Range of shoulder motion can be increased without physical therapy when pain is reduced after injection; therefore, patients have more condence in shoulder rangeof-motion exercises because they can exercise more aggressively with little pain. In this study, shoulder motion, especially forward exion, spontaneously increased about 12 before physiotherapy started because of reduced pain.

Table II Group Group Group Group

S.-J. Shin, S.-Y. Lee

Differences in ASES shoulder scores from baseline to 24 weeks after treatment according to corticosteroid injection site Initial I (SA) II (IA) III (IA SA ) IV (PO) 38.8 42.6 39.5 37.7 3.6 3.1 2.6 2.9 2 wk 69.4 73.9 72.9 49.4 2.7 2.6 4.6 3.6) 4 wk 76.3 85.1 85.6 55.9 3.4 3.1 1.6 3.1) 8 wk 81.9 86.4 86.5 70.8 3.7 2.1 1.9 3.8) 16 wk 87.1 88.4 90.7 73.1 3.2 2.9 2.8 2.0) 24 wk 89.4 91.1 90.7 84.1 1.9y 1.3y 1.6y 2.3y

IA, intra-articular; PO, oral medication; SA, subacromial. Data are given as mean SEM. ) Signicant difference between corticosteroid groups (groups I, II, and III) and group IV (P < .05). y Signicant difference between pretreatment and post-treatment scores in each group (P < .05).

function at 24 weeks despite a slower rate than corticosteroid injection. This study has several limitations. First, a subacromial corticosteroid injection is not commonly used to treat adhesive capsulitis, and it is not fully known how it affects shoulder function recovery, despite the good clinical results obtained. Second, the numbers of patients in the study groups were relatively small, and long-term outcomes were difcult to assess because of the high dropout rate. The overall participation rate was 77%. The most common reason for dropping out was an unwillingness to continue because of the failure of medication treatment. Furthermore, some adhesive capsulitis cases may have comprised rotator cuff tendinopathy, which showed similar clinical characteristics to adhesive capsulitis. It was very difcult to make a differential diagnosis between these 2 diseases using limited diagnostic tests, such as physical examination and ultrasonography used in this study. Finally, patients with different phases of adhesive capsulitis were present in each group, and this was not adjusted for, which may have inuenced clinical outcomes and effects of corticosteroids.

1. Andrieu V, Dromer C, Fourcade D, Zabraniecki L, Ginesty E, Marc V, et al. Adhesive capsulitis of the shoulder: therapeutic contribution of subacromial bursography. Rev Rhum Engl Ed 1998;65:771-7. 2. Arslan S, Celiker R. Comparison of the efcacy of local corticosteroid injection and physical therapy for the treatment of adhesive capsulitis. Rheumatol Int 2001;21:20-3. 3. Binder A, Hazleman BL, Parr G, Roberts S. A controlled study of oral prednisolone in frozen shoulder. Br J Rheumatol 1986;25:288-92. 4. Bulgen DY, Binder AI, Hazleman BL, Dutton J, Roberts S. Frozen shoulder: prospective clinical study with an evaluation of three treatment regimens. Ann Rheum Dis 1984;43:353-60. 5. Bunker TD, Anthony PP. The pathology of frozen shoulder. A Dupuytren-like disease. J Bone Joint Surg Br 1995;77:677-83. 6. Gerber C, Espinosa N, Perren TG. Arthroscopic treatment of shoulder stiffness. Clin Orthop Relat Res 2001;390:119-28. 7. Griggs SM, Ahn A, Green A. Idiopathic adhesive capsulitis. A prospective functional outcome study of nonoperative treatment. J Bone Joint Surg Am 2000;82:1398-407. 8. Hand GC, Athanasou NA, Matthews T, Carr AJ. The pathology of frozen shoulder. J Bone Joint Surg Br 2007;89:928-32. http://dx.doi. org/10.1302/0301-620X.89B7.19097 9. Hannan JA, Chiaia TA. Adhesive capsulitis. A treatment approach. Clin Orthop Relat Res 2000;372:95-109. 10. Hsu JE, Anakwenze OA, Warrender WJ, Abboud JA. Current review of adhesive capsulitis. J Shoulder Elbow Surg 2011;20:502-14. http:// 11. Jacobs LG, Smith MG, Khan SA, Smith K, Joshi M. Manipulation or intra-articular steroids in the management of adhesive capsulitis of the shoulder? A prospective randomized trial. J Shoulder Elbow Surg 2009;18:348-53. 12. Jones A, Regan M, Ledingham J, Pattrick M, Manhire A, Doherty M. Importance of placement of intra-articular steroid injections. BMJ 1993;307:1329-30. 13. Karthikeyan S, Kwong HT, Upadhyay PK, Parsons N, Drew SJ, Grifn D. A double-blind randomised controlled study comparing subacromial injection of tenoxicam or methylprednisolone in patients with subacromial impingement. J Bone Joint Surg Br 2010;92:77-82. 14. Levine WN, Kashyap CP, Bak SF, Ahmad CS, Blaine TA, Bigliani LU. Nonoperative management of idiopathic adhesive capsulitis. J Shoulder Elbow Surg 2007;16:569-73. 15. Lorbach O, Anagnostakos K, Scherf C, Seil R, Kohn D, Pape D. Nonoperative management of adhesive capsulitis of the shoulder: oral cortisone application versus intra-articular cortisone injections. J Shoulder Elbow Surg 2010;19:172-9. 16. Miller MD, Wirth MA, Rockwood CA Jr. Thawing the frozen shoulder: the patient patient. Orthopedics 1996;19:849-53.

Different corticosteroid injection sites of the shoulder were found to have similar clinical outcomes in adhesive capsulitis cases. However, a single corticosteroid injection provided faster pain relief, a higher level of patient satisfaction, and earlier improvements in range of motion and shoulder function than oral NSAIDs at up to 16 weeks after treatment.

The authors, their immediate families, and any research foundations with which they are afliated have not received any nancial payments or other benets from any commercial entity related to the subject of this article.

Efcacies of corticosteroid injection

17. Oh JH, Oh CH, Choi JA, Kim SH, Kim JH, Yoon JP. Comparison of glenohumeral and subacromial steroid injection in primary frozen shoulder: a prospective, randomized short-term comparison study. J Shoulder Elbow Surg 2011;20:1034-40. jse.2011.04.029 18. Ozaki J, Nakagawa Y, Sakurai G, Tamai S. Recalcitrant chronic adhesive capsulitis of the shoulder. Role of contracture of the coracohumeral ligament and rotator interval in pathogenesis and treatment. J Bone Joint Surg Am 1989;71:1511-5. 19. Reeves B. The natural history of the frozen shoulder syndrome. Scand J Rheumatol 1975;4:193-6. 20. Rizk TE, Pinals RS, Talaiver AS. Corticosteroid injections in adhesive capsulitis: investigation of their value and site. Arch Phys Med Rehabil 1991;72:20-2.

21. Ryans I, Montgomery A, Galway R, Kernohan WG, McKane R. A randomized controlled trial of intra-articular triamcinolone and/or physiotherapy in shoulder capsulitis. Rheumatology 2005;44:529-35. 22. Sethi PM, Kingston S, Elattrache N. Accuracy of anterior intraarticular injection of the glenohumeral joint. Arthroscopy 2005;21: 77-80. 23. van der Windt DA, Koes BW, Deville W, Boeke AJ, de Jong BA, Bouter LM. Effectiveness of corticosteroid injections versus physiotherapy for treatment of painful stiff shoulder in primary care: randomised trial. BMJ 1998;317:1292-6. 24. Warner JJ, Allen A, Marks PH, Wong P. Arthroscopic release for chronic, refractory adhesive capsulitis of the shoulder. J Bone Joint Surg Am 1996;78:1808-16.