Adhesive Capsulitis in Physical Medicine and Rehabilitation

Updated: Feb 08, 2023
  • Author: André Roy, MD, FRCPC; Chief Editor: Stephen Kishner, MD, MHA  more...
  • Print

Practice Essentials

Adhesive capsulitis, most commonly referred to as frozen shoulder (FS), is an idiopathic disease with 2 principal characteristics: pain and contracture.


Shoulder pain associated with FS is progressive and initially felt mostly at night or when the shoulder is moved close to the end of its range of motion (ROM). It can be caused by certain combined movements of the shoulder, such as abduction and external rotation (eg, grooming one's hair, reaching for a seatbelt overhead) or extension and internal rotation (eg, reaching for a back pocket or bra strap). The pain usually progresses to constant pain at rest that is aggravated by all movements of the shoulder and that may be worsened by repetitive movements of the involved upper extremity, psychological stress, exposure to cold or vibration, and changes in the weather. In approximately 90% of patients with FS, this pain usually lasts 1-2 years before subsiding. [1]


The second principal characteristic of FS is progressive loss of passive ROM (PROM) and active ROM (AROM) of the glenohumeral joint in a capsular pattern. That is, the movements are usually restricted to a characteristic pattern, with proportionally greater passive loss of external rotation than of abduction and internal rotation.

In 1934, Codman stated, "This entity [FS] is difficult to define, difficult to treat, and difficult to explain from the point of view of pathology." Codman's statement continues to hold true today.

In 1992, the American Shoulder and Elbow Surgeons Society agreed on the following definition of FS by consensus: a condition of uncertain etiology that is characterized by clinically significant restriction of active and passive shoulder motion that occurs in the absence of a known intrinsic shoulder disorder.

Workup in adhesive capsulitis

In general, idiopathic FS is considered a clinical diagnosis that does not require confirmation with radiologic imaging. Currently, the principal utility of these tests is in ruling out concomitant conditions that may influence the treatment of an individual patient.

All patients presenting with FS should undergo plain radiography of the shoulder, with the acquisition of soft-tissue views of the rotator cuff to rule out a septic or metastatic process.

A plain radiograph may also show evidence of a large calcification of the rotator cuff in the painful resorptive phase, an avascular necrosis of the humeral head (that is, Milwaukee shoulder), or a Charcot joint.

Management of adhesive capsulitis

Physical therapy

Physical therapy alone has not been shown to improve function in FS. However, physical therapy associated with an intra-articular injection of corticosteroid improves function and ROM more rapidly than does intra-articular corticosteroid injection alone. [2, 3]

Although many therapeutic exercises are described, few have been evaluated in the treatment of FS. Therapeutic exercises that have been studied include articular stretching and pulley therapy. [4, 5, 6] Passive articular stretching exercises improve ROM. The superiority of supervised versus home exercise programs has yet to be demonstrated. [7]

Surgical intervention

Considering the favorable prognosis for patients with idiopathic FS, surgical intervention should probably be reserved for rare patients whose condition does not respond to maximal conservative modalities implemented over a sufficient period of time. [8]


Intra-articular injections of steroid derivatives are the second most common medical intervention for treating painful joint conditions. (The most common intervention is the administration of nonsteroidal anti-inflammatory drugs [NSAIDs]).

Many authors have shown that the pathology of FS is in the extra-articular structures, such as the coracohumeral ligament, the interval of the rotator cuff, the subacromial space, and the articular capsule. [9, 10, 11] From a pathophysiologic point of view, a rationale supports the use of corticosteroid injection in the subacromiodeltoidian bursa to treat FS.

A comparison of intra-articular and intrabursal injections with lidocaine alone or with lidocaine and a corticosteroid showed similar decreases in pain and increases in ROM in all groups. [6] However, a significant short-term decrease in pain was observed in the lidocaine-plus-corticosteroid group compared with the lidocaine-only group.



Neurologic, surgical, and histologic findings

The pathophysiology of FS continues to be largely mysterious. However, certain consistent neurologic, surgical, and histologic findings in soft-tissue specimens of patients with FS have been identified and appear to be specific to the pathology of FS.

Evaluation of anatomic, histologic, and surgical specimens from subjects affected by idiopathic FS demonstrates that the glenohumeral joint synovial capsule is often involved in this disease process. However, most of the notable loss of ROM is caused by disease in structures outside the synovial capsule glenohumeral joint, such as the coracohumeral ligament, soft tissues in the rotator interval, the subscapularis muscle, and the subacromial bursae.

Most authors do not describe clinically significant capsular adhesions as a predominant finding in the chronic phase of this condition. Instead, pathologic data confirm an active process of hyperplastic fibroplasia and excessive type III collagen secretion that lead to soft-tissue contractures of the aforementioned structures (ie, the coracohumeral ligament, soft tissues of rotator interval, the subscapularis muscle, the subacromial bursae). However, these findings were observed in surgical patients who had severe and late-phase disease and cannot be applied to early phases of the disease.

From the chromosomal, cytochemical, and histologic points of view, the soft-tissue contractures are identical to those seen in a Dupuytren contracture of the hand. These contractures result in the classic progressive loss of ROM of the glenohumeral joint, which affects external rotation and abduction, then flexion, adduction, and extension (in descending order of severity). Despite these histopathologic similarities, the favorable and regressive outcome of adhesive capsulitis differs from the unfavorable and progressive outcome of Dupuytren disease.

Genetic abnormalities

Specific genetic abnormalities have been identified with this condition. In particular, frequent trisomy 7 and trisomy 8 in the fibroblasts were confirmed in cultures of tissue samples obtained from glenohumeral joint capsules resected at the time of surgery for FS. These pathologic findings were observed in studies of tissue specimens that usually were obtained from subjects with chronic FS that did not respond to typical conservative modalities. Little is known about the tissue characteristics of the acute phase of this condition.

Pathophysiologic model and complex regional pain syndrome type 1 versus FS

To date, no robust pathophysiologic model explains the relationship between the principal characteristics of this disease (ie, pain, contracture). Neurologic factors seem to be the principle mediators of the pain, whereas a process resembling fibromatosis causes the contracture, as described in the above text. Why the pain precedes the contracture and why it resolves before the contracture does in most subjects remain unclear. The fact that the pain of FS often precedes stiffening of the joint tends to support the notion that the initial pain probably does not stem from altered mechanics of the glenohumeral joint.

Clinical research suggests the presence of autonomic sympathetic dysfunction in the upper extremities in many patients with isolated idiopathic FS. This dysfunction is not detectable during usual clinical examination. However, various examination techniques, such as 3-phase nuclear technetium-99m (99m Tc) bone thermography, laser Doppler flowmetry, and transcutaneous measurement of oxygen tension, have been used to detect the anomalies found in more than 80% of subjects examined in various studies. The subjects did not have any of the dramatic manifestations of full-blown complex regional pain syndrome type 1 (CRPS1). Studies have shown similar histologic findings in joints affected by CRPS1 and FS.

A broad look at the similarities between FS and CRPS1 with regard to risk factors and the usual clinical course of the disease shows many parallels between the conditions. For example, risk factors common to both conditions are trauma, diabetes, thyroid disease, and dyslipidemia.

The initial, painful phase, without any inflammatory, exudative cellular joint process, lasts for several weeks or months. This phase usually progresses to include the gradual development of connective-tissue contracture. In most patients, the pain and contracture gradually resolve, although in a few individuals, 1 or both persist.

Neurologic mechanisms contributing to the generation and perpetuation of pain in CRPS1 probably also play a role in the pain and disability associated with FS. These mechanisms include peripheral alpha-adrenoreceptor hyperresponsiveness, dorsal-root reflexes (DRRs), central nervous system (CNS) factors, myxoid globular degeneration, and sympathetic autonomic hyperactivity.

Sympathetic dysfunction in patients with FS is believed to mediate hyperresponsiveness of peripheral alpha adrenoreceptors in the nerve endings of somatosensory neurons, including various joint nociceptive and proprioceptive fibers of the shoulder. This hyperresponsiveness probably contributes to the pain (allodynia) produced with gentle, passive mobilization of the shoulders observed in patients with FS.

DRRs, involved in previously unknown efferent neuronal activity in the somatic sensory nerves, are said to contribute to the pain of certain neurologically mediated painful conditions. These conditions are related to the release of substance P and of other local-tissue pain-and-inflammation – causing neurotransmitters into the regional tissues.

CNS factors, such as sensitization of wide-dynamic-range interneurons (WDRIs) in the dorsal horn of the spinal cord, may contribute to the allodynia affecting patients with FS.

In healthy subjects, myxoid globular degeneration of the terminal sensory nerve fibers of the glenohumeral joint capsule was particularly important in the fifth and sixth decades of life (when as many as 40% of terminal sensory nerve fibers of the shoulder are in a degenerated state). The importance of this phenomenon in the genesis and perpetuation of FS remains unknown. However, of note, FS is most common in the fifth and sixth decades of life, when such degenerative changes in the terminal nerve fibers of the shoulder joint become prevalent. Destruction of terminal sensory nerve fibers as a result of myxoid globular degeneration may play a role in the apparent dysfunction of the nervous system in FS.

Wohlgethan implicated sympathetic autonomic hyperactivity in the genesis and perpetuation of painful FS in patients with hyperthyroidism. [12] Several authors reported an apparently dramatic improvement in the pain and disability of FS in patients whose hyperthyroidism was adequately treated.

Despite the aforementioned similarities, FS should not be considered CRPS1. First, idiopathic FS does not fit the diagnostic criteria for CRPS1 that the International Association for the Study of Pain (IASP) proposed in 1996. Second, the diagnosis of CRPS1 is relatively nonspecific because of the heterogeneous pathologies it encompasses. Further expansion of the diagnosis of CRPS1 to include FS dilutes its meaning and falsely implies that any painful condition associated with anomalies of the autonomic nervous system must be categorized as CRPS. A preferable option is to include neuromodulatory concepts from CRPS research in a pathoetiologic model in which FS is a distinct clinical entity.

Inclusion of the aforementioned neuromodulatory concepts of joint pain into the authors' proposed pathoetiologic model of FS leaves room for the exploration of therapeutic modalities, such as bupivacaine suprascapular nerve blocks, stellate-ganglion blocks, and intra-articular morphine injections, to treat the pain associated with FS.

Ischemia of the soft tissues may link the apparently distinct and separate pathophysiologic entities observed in FS, namely, neurologically mediated pain and fibromatosis-like contracture. The histologic and cytochemical profiles of connective-tissue specimens in patients with FS are identical to those of patients with Dupuytren contracture, a condition whose pathoetiology has been linked to microangiopathy and local tissue ischemia. Pre-existing microvascular disease, often related to hypertriglyceridemia or diabetes in patients with FS, may predispose them to ischemia of the connective tissues of the glenohumeral joint capsule, which itself is subject to the ischemic stresses because of mechanical stress and tension. Furthermore, hypersensitization of peripheral vasomotor alpha-adrenergic receptors and a pain-induced increase in local vasomotor tone also can contribute to ischemia of the local connective tissues of the shoulder.

Resultant regional ischemia of the soft tissues of the shoulder can lead to the local release of free radicals and a platelet-derived growth factor that can initiate a cycle of fibroblastic hyperplasia and excessive deposition of collagen and glycosaminoglycans. The whole process ends with the active development of a tough, thick, fibrous contracture of the connective tissues of the shoulder. In his 1990 review on FS, Hazelman noted that microvascular disease might be the common thread linking FS to diabetes and dyslipidemias. [13]

In cases of FS occurring after fractures of the humeral head or following rotator cuff tears, the capsular, ligamentous, and tendinous circulation of the local soft tissue may be compromised. This effect may contribute to the aforementioned ischemic mechanisms leading to FS.

Care should be taken to avoid overusing the term FS secondary to shoulder tendonitis. In 1934, Codman described signs of inflammation of the tendon of the long portion of the biceps in cadaveric specimens of FS, giving rise to the still-popular theory that FS is a consequence of an inflammatory process of the tendons of the shoulder. Data from subsequent studies have largely refuted this as a pathophysiologic mechanism for FS, but the idea remains difficult to erase from the minds of clinical practitioners. Inflammation seen in the tendon of the long portion of the biceps in Codman's cadaveric specimens was most likely a late sequela of FS itself, which causes narrowing of the bicipital groove secondary to a retraction of the overlying soft tissues.




United States

Shoulder pain is the third most common cause of musculoskeletal disability after low back pain (LBP) and neck pain. The prevalence of FS in the general population is reported to be 2%, with an 11% prevalence in unselected individuals with diabetes. For patients with type I diabetes, the risk of developing FS in their lifetime is approximately 40%.

FS may affect both shoulders, either simultaneously or sequentially, in as many as 16% of patients. The frequency of bilateral FS is higher in subjects with diabetes than in those without diabetes. In 14% of patients, while FS still is active in the initial shoulder, the contralateral shoulder also becomes affected. Contralateral FS usually occurs within 5 years of disease onset. A relapse of FS in the same shoulder is unusual.

FS most frequently occurs in subjects with hyperthyroidism and hypertriglyceridemia. Although various authors report that heart disease, tuberculosis, and many other medical conditions are associated with FS, these associations are largely anecdotal and are not supported in proper, case-controlled studies.

Most survivors of cerebral vascular accidents (CVAs) that cause hemiplegia develop painful stiffening of their shoulders. However, the painful hemiplegic shoulder has distinct characteristics, which are not discussed in this article.

FS will undoubtedly become increasingly common as the baby-boom generation ages, because this condition most frequently occurs in the fifth and sixth decades of life. Patients who present with an idiopathic FS when they are younger than 40 years should definitely be examined to rule out occult diabetes, hyperthyroidism, hypertriglyceridemia, or concomitant neurologic or systemic rheumatologic disorder affecting the upper extremity.


The data reported above are derived from the international literature. No regional variations are recognized in the frequency of occurrence of FS.


Shoulder pain is the third most common cause of musculoskeletal disability in the workplace after LBP and neck pain. For more information, see the Musculoskeletal Disorders and Workplace Factors, published by the National Institute for Occupational Safety and Health, on the Centers for Disease Control and Prevention Web site.

The degree of pain and disability caused by idiopathic FS is highly variable and depends on the stage of the disease. Retrospective data tend to show that patients cope well with a painless glenohumeral contracture. Therefore, most disabilities occur during the initial, painful phase and the subsequent, freezing phase.

See also Patient Education.


Although there has been a lack of racial variation cited in the literature, a study by Kingston et al of 2190 patients with adhesive capsulitis found the rate of African American and Hispanic individuals in the cohort to be higher than that of white patients. [14]


FS affects women more frequently than men, with a female-to-male ratio of about 1.4:1. Menopause is often reported as a cause of FS in women, although Lundberg seems to have ruled out this hypothesis by demonstrating that age is the principal predictor. [15] He specifically demonstrated that women with early menopause did not have FS any earlier than their counterparts who undergo late menopause.


To date, the best data available seem to show that FS affects women somewhat earlier than it does men. The mean ages of onset are 52 years for women and 55 years for men, with a standard deviation of 7.