Adhesive capsulitis is a common (occurring in 2% of the general population), but poorly understood condition. Consequently, it is frequently underdiagnosed. Codman was the first to publish an extensive review of the subject in 1934, which he referred to as “frozen shoulder.” In 1945, the Nevaiser was the first to introduce the term “adhesive capsulitis” and identified the pathology of this condition as “a thick capsule, tight glenohumeral joint and obliteration of the axillary fold.” It is important to recognize adhesive capsulitis as a distinct diagnosis, not to be confused with other conditions that cause stiffness and pain in the shoulder, such as proximal humerus post-fracture stiffness, glenohumeral osteoarthritis and calcific tendinitis. Accurate diagnosis will indicate appropriate treatment and result in the best outcome for patients.
90% of patients with adhesive capsulitis are women, age 40-60 years, typically presenting with a chief complaint of the insidious onset of pain and stiffness in the shoulder without a history of significant trauma. Often patients may describe a trivial everyday event (such as reaching overhead or sudden movement of the arm) insufficient to cause mechanical bone or soft tissue injury, but may trigger the onset of progressive pain and stiffness.
The most common and disabling symptom may be night pain that interrupts sleep. Other symptoms include compromised activities of daily living that require full range of motion of the shoulder, such as donning a coat, reaching overhead or fastening a bra. Patients with severe cases of adhesive capsulitis may even complain of pain at rest. However, the majority of patients with adhesive capsulitis will not complain of pain during activities performed within their limited range of shoulder motion.
Past medical history should include inquiry regarding other conditions associated with adhesive capsulitis such as diabetes mellitus, thyroid disease, and regional surgical procedures such as recent breast surgery or hand surgery. How these conditions contribute to the development of adhesive capsulitis remains obscure. However, such history will alert the clinician to the proper diagnosis.
The most important aspect of the physical examination and the key to establishing the diagnosis is global limitation of passive range of motion of the shoulder. Range of motion of the symptomatic shoulder should be compared to the normal contralateral side. Up to 15% of patients subsequently develop adhesive capsulitis in the contralateral shoulder but presentation of bilateral adhesive capsulitis simultaneously is extremely rare. The basic shoulder exam includes three ranges of motion: forward elevation in the scapular plane (best assessed with the patient supine on the exam table), external rotation with the arm at the side, and internal rotation noted by which posterior vertebral spinous process can be reached by the tip of the thumb.
Typically, forward elevation is limited to 90-120 degrees (normal 170 degrees), external rotation 0-30 degrees (normal 70 degrees), and internal rotation to the lumbosacral junction (normal T7). There is a firm and painful end point to this limited motion. Limited range of motion distinguishes adhesive capsulitis from impingement syndrome in which full or nearly full passive range of motion is usually preserved.
With careful and gentle passive rotation of the glenohumeral joint by the examiner, the patient will have no pain within the limited range of motion but complain of sharp pain when the arm is moved to the extremes of the limited range. This physical finding distinguishes adhesive capsulitis from glenohumeral osteoarthritis in which there is painful crepitus within the limited range of motion.
Typically shoulder strength is normal and there is no focal tenderness in the shoulder, distinguishing adhesive capsulitis from symptomatic acromioclavicular joint injuries and rotator cuff disease. Range of motion of the cervical spine and the neurological examination of the involved upper extremity are also normal thereby ruling out pathology of the cervical spine as a cause of shoulder pain.
Shoulder radiographs including scapular AP, scapular lateral and an axillary view should be obtained. Scapular plane images are essential in order to evaluate the glenohumeral joint (Figure 1) and are best to rule out other conditions that may cause pain and limited motion of the shoulder in this patient population such as proximal humerus fracture, glenohumeral osteoarthritis and calcific tendinitis.
Magnetic resonance imaging of the shoulder is rarely required to establish the diagnosis but if obtained, will show the characteristic findings of a thickened inferior capsule, obliteration of the inferior axillary joint space, well-preserved glenohumeral cartilage, and an intact rotator cuff (Figure 2).
The Neer injection test, consisting of injection of 10cc of lidocaine into the subacromial space, may be considered if concomitant impingement syndrome is suspected, but usually provides neither relief of pain nor improved range of motion since the pathology is intra-articular and capsular. However, some patients with adhesive capsulitis may have associated subacromial bursitis and may obtain some but incomplete relief of pain following a subacromial injection. A poor response to the injection test, i.e. no pain relief, indicates that adhesive capsulitis is the primary diagnosis.
Adhesive capsulitis is a self-limiting disease that resolves in 90% of patients. Hence, one could argue that “benign neglect” is a reasonable course of nonoperative treatment. However, the natural history of the disease may last up to 2 years and some degree of residual stiffness may persist in up to 50% of patients after pain resolves. Consequently, few patients choose to ignore the disease, awaiting an uncertain and prolonged resolution.
The mainstay of nonoperative treatment is gentle, assisted range of motion exercises that are performed for brief periods (2-3 minutes) and frequently (3-5 times daily). Ideally, these exercises are physician directed, and the patient is instructed to perform these exercises independently. Instruction can be provided in the physician’s office or by a physical therapist, but the essential aspect of the program is that the patient understands the exercises and performs them consistently. Simple diagrams with instructions of the exercise program provided by the physician’s office are extremely useful (Figure 3).
Many patients request and benefit from supervised physical therapy which often is very effective. However, it is extremely important that the stretching program provided by the therapist not be overly aggressive causing increased pain and inflammation and thereby exacerbating the condition and prolonging the recovery. Physical therapy described by patients as “torture” and “I feel worse after therapy” indicates an overly aggressive program. The concept of “no pain no gain” should be discouraged.
The actual stretching should be gentle but the patient should feel the stretch at terminal range of motion which may be uncomfortable but tolerable. Any pain associated with the stretch should abate within a few minutes after the stretch. Application of moist heat to the shoulder prior to the stretching program often enhances the amount of stretching that can be achieved. Thus, such an appropriate stretching program could be described as “work, not torture.”
The initial medical management of adhesive capsulitis includes Tylenol or NSAIDs to help decrease pain. These medications taken 30-60 minutes prior to therapy sessions may improve the amount of stretching achieved during the session. Since the most common complaint is pain at night that interrupts sleep, these medications should be taken just prior to bedtime in an effort to permit a good night’s sleep. However, night pain is often so severe that narcotic medication may be required as well.
Intra-articular injection of cortisone may make sense since the pathology is intra-articular capsular inflammation. However, there is little evidence that intra-articular cortisone injections alter the natural history of the disease. Intra-articular brisement, the injection of saline under pressure which ruptures capsular tissue, has been described but lacks convincing evidence in well controlled studies.
The most important aspect of the nonoperative program is to educate the patient and emphasize that brief and frequent stretching sessions, though tedious, are by far the most effective. The patient should be advised that the stretching program may take up to 3-4 months or longer before substantial resolution of pain and stiffness is achieved. A follow-up examination should be scheduled within 6 weeks into the program to ensure that range of motion is improving and the patient is improving.
Indications for Surgery
Disabling pain and stiffness that compromise sleep and activities of daily living, unresponsive to a minimum of 3-6 months of a supervised stretching program are the indications for operative treatment which is necessary in less than 10% of patients.
Patients who have demonstrated consistent improvement in range of motion and pain after 3-4 months of exercises but who have not yet achieved complete resolution of their symptoms should be encouraged to continue the nonoperative protocol with the expectation that improved shoulder function will be achieved within 3 months.
Operative management of adhesive capsulitis begins with examination under anesthesia, followed by manipulation of the shoulder and arthroscopic capsular release. The procedure is performed under a regional scalene block anesthesia and intravenous sedation. Long-acting anesthetic agents, usually 0.25% Marcaine, supplemented with 4 mg of Decadron, consistently afford 18 to 24 hours of pain relief postoperatively. Patients should be advised preoperatively not to be alarmed by prolonged numbness and weakness of the arm and advised to keep the upper extremity in a sling until the morning following surgery by which time motor control has returned to the arm. General anesthesia may be considered as well but will not provide the desired prolonged analgesia post-operatively that a scalene block affords.
The patient is placed in the beach chair position at which time an examination under anesthesia is performed to confirm that the limited passive range of motion recorded preoperatively is mechanical and not limited by pain or apprehension while the patient was awake during preoperative examination in the office. Next, a manipulation under anesthesia is performed.
The manipulation begins with slow and steady forward elevation in the scapular plane. The arm should be supported along the distal humerus to minimize a long lever arm that may place the humeral shaft at risk for fracture. With progressive pressure and forward elevation, there is often audible crepitus which indicates rupture of the contracted inferior capsular tissue. Patients with severe stiffness preoperatively, less than 90 degrees forward elevation, or with a history of diabetes, frequently have capsular thickening so severe that manipulation cannot rupture the pathologic tissue. Hence, in such cases, if a hard endpoint during manipulation is encountered, excessive force should be avoided and, if progressive forward elevation is not possible, the manipulation should be terminated in order to avoid humeral shaft fracture.
Once full elevation is achieved, the arm is placed at the side and external rotation is performed to rupture of the anterior capsule. Again, excessive force should be avoided to minimize the risk of humeral shaft fracture. Next, the arm is placed in 90° of abduction and then internally rotated to release the contracted posterior capsule. Finally, with the arm in 90°abduction, the arm is adducted across the chest to complete division of the posterior capsular tissue.
If the surgeon prefers lateral decubitus position for the arthroscopy, manipulation under anesthesia is performed with the patient in the supine position on the operating table prior to positioning and the same steps followed as described above.
Slight modification of the routine posterior arthroscopic portal will enhance the surgeon’s ability to enter the glenohumeral joint. Owing to the contracted posterior capsule that is thicker than normal and, consequently narrows the glenohumeral joint space, the posterior portal placed 1 cm more superior than a standard posterior portal will allow entry into the glenohumeral joint space slightly above the equator and permit easier entry into the joint where there is more potential space along the superior curve of the humeral head. Entry into the posterior joint is often difficult due to the thickened posterior capsule and requires more force than usual. Care should be taken to avoid inadvertent penetration of the humeral head and for this reason a blunt trocar should always be used.
Following a manipulation under anesthesia with rupture of the highly vascular capsular tissue, a significant hemarthrosis is always encountered upon entering the glenohumeral joint. For this reason, copious irrigation through the arthroscopic cannula is required to provide adequate visualization for placement of the anterior portal.
A slight modification of the routine anterior portal 1 cm medial to the usual entry point will enhance the surgeon’s ability to release the most inferior capsular tissue. Placement of the anterior portal is best assessed using a spinal needle and confirming under direct visualization that the tip of the spinal needle can access the inferior capsule without being obstructed by the humeral head. Once proper placement is confirmed, the anterior portal is established and a synovial resector is used to provide high flow irrigation of the joint to improve visualization.
Once the portals are established, inspection of the joint begins. The hyperemic and thickened anterior capsule is always present (Figure 4) but may be ruptured following a successful manipulation. Typically the articular surfaces of the glenoid and humeral head are normal as are the rotator cuff and biceps tendons.
Capsular resection and synovectomy can usually be performed with just two arthroscopic instruments: electrocautery and a synovial resector. If the surgeon encounters difficulty accessing the inferior capsule, less common when the patient is placed in the lateral decubitus position, a 30° arthroscopic biter and angled soft tissue elevator may be helpful to release the inferior capsule.
The procedure begins with electrocautery through the anterior portal and a complete synovectomy using the coagulation setting is performed. There is often an extensive synovitis in the superior capsular fold above the labrum and along the undersurface of the rotator cuff and biceps anchor. Use of the coagulation setting avoids any inadvertent resection of the tendons. Tissue in the rotator interval, the space between the superior border of the subscapularis tendon and the anterior edge of the supraspinatus tendon, is consistently inflamed and contains thickened and contracted tissue which is resected using electrocautery on the ablation setting.
The capsular tissue is then addressed. The capsule should not simply be divided but actually a portion of it resected to produce at least a 1 cm gap in the tissue that may prevent subsequent healing in the previously contracted state. Laterally, the capsule and subscapularis tendon are confluent but separate at the level of the musculocutaneous junction medially where the plane between the anterior capsule and subscapularis tissue is most easily identified. This plane can be developed by blunt dissection with a probe and the electrocautery device on the ablation setting is used to resect the capsule.
The capsule is divided at its junction with the labrum which should remain attached to the rim of the glenoid. As the release proceeds inferiorly, the capsule thickens and care must be taken to ensure complete division of the capsule. As the release of the anterior capsule proceeds, the surgical assistant can progressively increase external rotation and with slight abduction of the arm to further enhance visualization and access to the inferior capsule. The more medial placement of the routine anterior portal allows easier access to the inferior aspect of the glenohumeral joint and allows for the complete release of the most inferior capsular tissue. In this inferior region, it is important that the electrocautery device remain close to the rim of the glenoid in an effort to avoid injury to the axillary nerve.
If access to the inferior capsule is not possible, then an angled arthroscopic biter or soft tissue elevator may provide better access to complete the inferior capsular resection. In the beach chair position, capsular release can routinely be performed to the level of the inferior posterior capsule. Visualization of the inferior and posterior capsule is usually easier when the patient is placed in the lateral decubitus position which is why some surgeons prefer this position to the beach chair position.
Once a complete inferior capsular release is accomplished, the arthroscope is placed in the anterior portal and the electrocautery device is placed in the posterior portal to complete the posterior capsulectomy and synovectomy. Complete inspection of the joint is then performed and meticulous hemostasis obtained to minimize postoperative hemarthrosis that may further stimulate the inflammatory response and contribute to excessive scar formation and subsequent recurrent contracture.
The subacromial space is then inspected via the posterior portal. If the subacromial space is normal, the arthroscope is removed and the portals are closed. However, if the subacromial tissue is inflamed, which is not unusual, a bursectomy is performed via a routine lateral portal using a synovial resector and electrocautery. A coracoacromial ligament release and acromioplasty are not performed for two reasons: (1). the bursitis is believed to be simply associated with the inflammation in the glenohumeral joint and not due to an impingement lesion and (2). to minimize bleeding that may contribute to subsequent adhesions and scar formation.
Prior to skin closure, range of motion of the shoulder is recorded. A further gentle manipulation of the shoulder is performed in the direction of any residual limitation of motion. Following application of dressings, the arm is placed in a sling with a waist strap to fully secure and protect the partially paralyzed arm as the long-acting scalene anesthetic block resolves over the ensuing 18-24 hours.
Pearls and Pitfalls of Technique
surgeon preference most important
lateral decubitus position permits easier access to inferior and posterior capsule
beach chair position allows easier motion of the shoulder to assess adequacy of release
Manipulation under Anesthesia
proximal holding of the humerus to avoid long lever arm
slow and steady pressure until full motion achieved often with audible lysis of adhesions
if a firm endpoint met, do not apply excessive force, stop manipulation to avoid humeral shaft fracture, and proceed with arthroscopic capsular release under direct visualization
posterior portal should be 1 cm superior to the usual placement to enhance access to the glenohumeral joint
anterior portal should be approximately 1 cm medial to the standard anterior portal to enhance access to the inferior capsule
use of a spinal needle to determine best placement of the anterior portal that will permit access to the inferior capsule
Technique to help enter the glenohumeral joint
use blunt, never a sharp trocar to localize the humeral head
surgical assistant rotates the arm to help identify the curve of the humeral head
palpate the curvature of the superior humeral head with the blunt trocar and enter the glenohumeral joint superior to the equator
Identify anterior capsule
develop the plane separating the anterior capsule from the subscapularis tendon at the musculo-tendinous junction medially
divide the anterior capsule with electrocautery at its junction with the labrum
resect 1 cm of anterior capsule to produce a gap that will minimize the chance for recurrent contracture
Access to inferior capsule
superior and medial placement of the anterior portal
advance arthroscope between humeral head and inferior glenoid to enhance direct visualization of inferior capsule
surgical assistant abducts humerus with shoulder in neutral or slight external rotation to improve exposure of inferior capsule
Avoid injury to axillary nerve
resect capsule near glenoid rim at attachment to the labrum
shoulder abduction in neutral rotation places axillary nerve further away from inferior capsule
The most common complication following arthroscopic release for adhesive capsulitis is recurrent contracture and stiffness which occurs in up to 10% of operative cases. The most important aspect to avoid recurrent contracture is no synovectomy and capsular release during the index procedure and early range of motion of the shoulder postoperatively. Humeral shaft fracture following shoulder manipulation under anesthesia is a well-appreciated risk but there are very few reports of this complication. While there is understandable concern for injury to the axillary nerve during capsular release, this also is only rarely reported. Risk of infection following shoulder arthroscopy is well under 1%.
Arthroscopic capsular release for adhesive capsulitis is performed under a regional scalene block using a long acting anesthetic on an ambulatory basis. The patient is discharged with the arm in a sling with a strap around the waist to provide adequate protection during the period of prolonged analgesia which may last up to 18 -24 hours.
The sling is removed the morning following surgery and is not replaced. Active use of the operated shoulder below the horizontal for activities of daily living is encouraged immediately as comfort permits. The patient is instructed to place a pillow between the arm and chest while seated and by the second day postoperative is instructed to perform home pendulum exercises.
Since the patient had undergone an extensive course of self-directed assisted range of motion exercises preoperatively, the same exercise routine, already familiar to the patient, is started on POD #2, and is expanded as comfort permits.
The first postoperative visit is typically 10 days following surgery at which time sutures are removed and range of motion of the shoulder assessed. At this time, the patient should have achieved, at a minimum, the range of motion that was present preoperatively. Further instructions regarding assisted range of motion exercises are shared with the patient at this time and, if it is apparent that persistent pain precludes an adequate home exercise program, supervised physical therapy for-assisted range of motion exercises is recommended at this time.
The patient is reexamined at 6 weeks following surgery at which time a minimum of 140° forward elevation and 40° external rotation is achieved. Additional stretching exercises, focusing on terminal range of motion, are instituted at this time with the expectation that full range of motion will be achieved by the next visit at 3 months following the procedure.
The patient should be advised preoperatively that improvement in range of motion can continue up to 6 months postoperatively but that the majority of the motion will be regained by 3 months. Progressive improvement in range of motion is the most important aspect to assess during the postoperative recovery, even if full range of motion is not achieved at 3 months.
If range of motion of the shoulder is not improving by 3 months postoperatively, a repeat manipulation of the shoulder under a light general anesthesia should be considered. However, a repeat manipulation at this time is only rarely required if the patient has complied with the postoperative exercise routine.
Outcomes/Evidence in the Literature
Neviaser, AS, Hannafin, JA. “Adhesive capulitis: A review of current treatment”. Am J Sports Med. vol. 38. 2010. pp. 2346-2356. (Full review of literature and results of major trials.)
Shaffer, B, Tibone, JE, Kerlan, RK. “Frozen Shoulder. A long-term follow-up”. J Bone and Joint Surg. vol. 74-A. 1992. pp. 738-746. (62 patients treated non-operatively for adhesive capsulitis. Average 7-year follow up 50% of patients had remaining pain or stiffness with objective loss of ROM. Only 11% of patients reported mild functional limitation.)
Miller, MD, Wirth, MA, Rockwood, CA. “Thawing the frozen shoulder: The "patient" patient”. Orthopedics. vol. 19. 1996. pp. 849-853. (Retrospective review of 50 patients with adhesive capsulitis treated with physician taught home exercises, moist heat and ibuprofen. All patients achieved "functional motion" by 36 months.)
Griggs, SM, Ahn, A, Green, A. “Idiopathic Adhesive Capsulitis: A prospective functional outcome study of nonoperative treatment”. J Bone and Joint Surg. vol. 82-A. 2000. pp. 1398-1407. (Prospective study of 75 patients with adhesive capsulitis placed into directed four direction shoulder stretching PT. Both objective and subjective criteria analyzed. 90% of patients with satisfactory outcome. Objective measures demonstrated significant improvement though pain and function did not equal that of contralateral side. Male gender and diabetes associated with poorer outcomes.)
Carette, S, Moffet, H, Tardif, J. “Intraarticular corticosteroids, supervised physiotherapy, or a combination of the two in the treatment of adhesive capsulitis of the shoulder: A placebo controlled trial”. Arthritis Rheum. vol. 48. 2003. pp. 829-838. (Level I Study Comparing the efficacy of a single intra-articular corticosteroid injection, a supervised physiotherapy program, a combination of two, and placebo. Early pain scores better in group that received fluoroscopically guided corticosteroid and directed PT. At one year all group demonstrated no statistically significant difference.)
Dodenhoff, RM, Levy, O, Wilson, A, Copeland, SA. “Manipulation under anesthesia for primary frozen shoulder: Effect on early recovery and return to activity”. J Shoulder Elbow Surg. vol. 9. 2000. pp. 23-26. (Prospective assessment of 39 shoulders diagnosed with adhesive capsulitis. All patients underwent isolated manipulation under anesthesia. Significant increases seen in constant score that was maintained 1 year after manipulation. 94% patient satisfaction with only five shoulders having continued severe disability.)
Ogilvie-Harris, DJ, Biggs, DJ, Fitsialos, DP, MacKay, M. “The resistant frozen shoulder: Manipulation versus arthroscopic release”. Clin Orthop Relat Res. 1995. pp. 238-248. (Prospective study of 40 patients comparing MUA to arthroscopic release. MUA and release equivalent for restoration of ROM. Arthroscopic release patients had significantly better pain scores and function.)
Le Lievre, HMJ, Murrell, GAC. “Long term outcomes after arthroscopic capsular release for idiopathic adhesive capsulitis”. J Bobe Joint Surg. vol. 94. 2012. pp. 1208-16. (Retrospective study of 43 patients treated with circumferential arthroscopic release. Follow up for 7.5 years. Significant improvements in patient reported pain, shoulder function, stiffness, and ability to complete activities. Improvements sustained over long follow up.)
Chen, J, Chen, S, Li, Y, Li, H. “Is the extended release of the inferior glenohumeral ligament necessary for frozen shoulder?”. Arthroscopy. vol. 26. 2010. pp. 529-535. (Level I study of 74 patients randomly assigned to anterior release versus full circumferential release. All patients demonstrated significant improvements over preoperative status. No statistical difference in pain scores between the groups. No statistical difference in ROM between the two groups at final follow up. ROM better at 3 months in circumferential group.)
Levine, WN, Kashyap, CP, Bak, SF, Ahmad, CS, Blaine, TA, Bigliani, LU. “Nonoperative management of idiopathic adhesive capsulitis”. J Shoulder Elbow Surg. vol. 16. 2007. pp. 569-573. (105 shoulders followed to resolution of symptoms or surgical treatment. 89.5% successfully treated non-operatively. Average duration of successful non-operative treatment 3.8 +/- 3.6 months. Average duration of failed non-operative treatment for patients requiring surgery 12.4 +/- 12.1 months. The vast majority, 89%, of patients are successfully treated non-operatively in 3-6 months.)
Adhesive capsulitis Is a common condition occurring in up to 2% of the population. Women, age 40-60 years, are most commonly affected. The most important aspects of history and physical exam include the atraumatic onset of pain, limited passive range of motion of the shoulder, the absence of both glenohumeral crepitus and shoulder tenderness and normal radiographs. Establishing the correct diagnosis is essential to recommending appropriate treatment and predicting prognosis. Brief (1-2 minutes) though frequent (3-5 times daily) self-assisted range of motion exercises provide resolution of symptoms in 90% of patients but may require 3-6 months of treatment. Aggressive physical therapy causing increased pain and “torture” should be avoided. The 10% of patients who fail non-operative treatment for a minimum of 3-6 months are candidates for shoulder manipulation under anesthesia and arthroscopic capsular release which is successful in 90% of cases. Compliance with the same stretching program post-operatively as was followed pre-operatively is very important to avoid recurrent capsular contracture and subsequent shoulder stiffness and pain.
Copyright © 2017, 2013 Decision Support in Medicine, LLC. All rights reserved.
No sponsor or advertiser has participated in, approved or paid for the content provided by Decision Support in Medicine LLC. The Licensed Content is the property of and copyrighted by DSM.