PubMed Health. A service of the National Library of Medicine, National Institutes of Health.

Maund E, Craig D, Suekarran S, et al. Management of Frozen Shoulder: A Systematic Review and Cost-Effectiveness Analysis. Southampton (UK): NIHR Journals Library; 2012 Mar. (Health Technology Assessment, No. 16.11.)


The decision problem

The commissioning brief requested an evidence synthesis comparing the clinical effectiveness and cost-effectiveness of different treatments for frozen shoulder to determine the most appropriate management strategy. The aims of the project were to:

  1. evaluate, by way of a systematic review, the clinical effectiveness (including adverse effects) of strategies currently used in the NHS for the management of frozen shoulder and identify the most appropriate intervention by stage of condition, specifically physical therapies, steroid and other shoulder injections, manipulation under anaesthesia (MUA), arthrographic distension, capsular release, watchful waiting and combinations of these interventions
  2. collate, by way of a systematic review, patients' views and experiences of the interventions being assessed in the review
  3. identify, by way of a systematic review, the cost-effectiveness of the different interventions in order to inform the development of a decision model
  4. develop a decision-analytic model to estimate the cost-effectiveness of alternative treatment options for frozen shoulder
  5. make recommendations for clinical practice
  6. identify any gaps in the evidence, undertake value of information analysis to assess the potential value of future research on interventions for frozen shoulder and make specific recommendations for further research.

Frozen shoulder

Frozen shoulder is a painful condition in which movement of the shoulder becomes severely restricted. The condition can vary in severity from mild to severe pain and/or from some to severe restriction in movement. It was first described in 1875 by the French pathologist Duplay, who named it péri-arthrite scapula-humérale. The most well-known definition and name for the condition was provided in 1934 by an American surgeon EA Codman. He defined it as coming on slowly ‘with pain usually felt near the insertion of the deltoid; inability to sleep on the affected side; painful and incomplete elevation and external rotation; restriction of both spasmodic and mildly adherent type; atrophy of the spinati; little local tenderness; [and] X-rays negative except for bone atrophy’ and named it ‘frozen shoulder’.1 However, there is an acknowledged absence of a specific definition of the condition2,3 and of a diagnostic label,3 with additional names for frozen shoulder including adhesive capsulitis, painful stiff shoulder, retractile capsulitis, Checkrein shoulder, monoarticular arthritis and steroid-sensitive arthritis.4 Throughout this report the term ‘frozen shoulder’ will be used.

Frozen shoulder is typically characterised as having three overlapping phases:5

  1. phase 1, in which there is progressive stiffening and loss of motion in the shoulder with increasing pain on movement, which may be worse at night (months 2–9), usually referred to as the painful phase
  2. phase 2, in which there is a gradual decrease in pain but stiffness remains and there is considerable restriction in the range of movement (months 4–12), usually referred to as the stiffening or ‘freezing’ phase
  3. phase 3, in which there is an improvement in range of movement (months 12–42), usually referred to as the resolution phase.

There are different views about the underlying fundamental process: inflammation, scarring and scarring produced in reaction to inflammation.6 Bunker describes a pathology of fibrous contracture of the rotator interval and coracohumeral ligament of the shoulder joint.7 In addition to the capsular contracture, there is often a reduced joint volume: 3–4 ml compared with the normal volume of 10–15 ml.8 The formation of new blood vessels in the synovial membrane, which is most marked in the rotator interval area, is a feature of the early stage of the condition. In the stiff phase this declines and thick white scar tissue can be seen and palpated within the capsule at arthroscopic and open surgery. Histology also shows thickening and contracture of the capsule. A significant increase in fibroblasts, which lay down scar tissue and myofibroblasts, which contract scar tissue, has been identified.8 In addition, pathological studies have identified the presence of inflammatory cells (mast cells, T cells, B cells and macrophages), suggesting a process of inflammation leading to scarring.9

The cumulative incidence of frozen shoulder is estimated at 2.4 per 1000 population per year based on a Dutch general practice sample.10 A large UK-based primary care study found that frozen shoulder affected 8.2% of men and 10.1% of women of working age.11 In contrast, frozen shoulder was estimated to affect only 0.75% of the UK population based on a specialist shoulder surgeon's hospital care experience.8 This discrepancy in estimated prevalence may be explained by the fact that only the most resistant cases of frozen shoulder are seen in a hospital setting.6 Frozen shoulder most commonly occurs in people in their mid-50s and is thought to be slightly more common in women than in men.

Frozen shoulder can be described as either primary (idiopathic), whereby the aetiology is unknown, or secondary, when it can be attributed to another cause. Secondary frozen shoulder has been defined as that associated with diabetes (although some classify this as primary frozen shoulder), trauma, cardiovascular disease and hemiparesis. The incidence of frozen shoulder is reported to be 10–36% amongst people with diabetes, who tend not to respond as well to treatment as those without diabetes.5 The proportion of frozen shoulder attributed to trauma varies (9–33%) and the trauma is often not particularly severe. There is also a discrepancy between the extent of trauma and severity of subsequent frozen shoulder.4 These uncertainties mean that in practice it can be difficult to differentiate between primary and secondary frozen shoulder.

Although for most people frozen shoulder is a self-limiting condition of approximately 1–3 years' duration, it can be extremely painful and debilitating; people with the condition may struggle with basic daily activities and be worn down by sleep disturbance as a result of the pain.12 As well as interfering with domestic and social activities, it can affect the ability to work. There may not be a complete resolution for all patients and there is variation across case series in the proportion of patients who do not regain full shoulder motion,5 possibly a reflection of variation in how outcome was assessed. Based on the largest series of patients with a mean follow-up of 4.4 years from onset of symptoms, 59% had normal or near normal shoulders, 35% had mild to moderate symptoms with pain being the most common complaint and 6% had severe symptoms at follow-up.13 Recurrence is unusual although it is estimated that the other shoulder becomes affected in 6–17% of patients within 5 years.

Diagnosis and management

Diagnosis, in both primary and secondary settings, is based on clinical examination and medical history. A key alerting feature is restriction of shoulder movement in all directions – passive and active range of movement.14 Blood tests, radiography and ultrasound are usually normal and not routinely required unless history or physical examination suggests the need to rule out other pathologies, for example if rheumatoid arthritis or osteoarthritis is suspected. Frozen shoulder is commonly managed in the primary care setting. In a UK study of patterns of referral and diagnosis of shoulder conditions it was estimated that 22% of patients were referred to secondary care, up to 3 years following initial presentation, although most referrals occurred within 3 months.15 There is little evidence available on referral patterns in relation to frozen shoulder specifically.12

There are a number of management options, both surgical and non-surgical, but there is no consensus about management. For the purpose of the report we have classified the interventions as conservative and invasive (Table 1).

TABLE 1. Classification of interventions for frozen shoulder.


Classification of interventions for frozen shoulder.

The aims of treatment, depending on stage of condition, are pain relief, increasing arm movement, reducing the duration of symptoms and return to normal activities for the patient. Treatment options include:

  • Watchful waiting or ‘supervised neglect’, which involves explaining the condition to the patient, and education and advice about mobilisation within pain limits and use of pain relief.
  • Oral medications such as non-steroidal anti-inflammatory drugs (NSAIDs) and oral steroids. Although the use of oral steroids is described in the literature they are not a commonly used intervention in the UK.
  • Gentle exercise supervised by a physiotherapist or as part of a home exercise programme.
  • Physical therapies to help regain range of movement and prevent further stiffness. These encompass the wide range of techniques used by physiotherapists as well as acupuncture, and chiropractic and osteopathic techniques. Several different regimes have been described in the literature including supervised exercises, mobilisation, physiotherapy and use of electrotherapeutic interventions such as laser therapy and ultrasound. Mobilisation is therapist-applied passive movement of joints or other structures performed in such a way that they are always within the control of the patient. Electrotherapies include transcutaneous nerve stimulation (TENS), interferential therapy, short-wave diathermy (SWD) and pulsed short-wave diathermy (PSWD), and ultrasound. TENS and interferential therapy consist of electric pulses or currents and have an analgesic effect. SWD and PSWD use radio frequency energy to generate heat in tissues, which has an analgesic effect and reduces muscle spasm and joint stiffness. Ultrasound, mechanical vibration at very high speed, also generates heat and has the same therapeutic effect as SWD and PSWD.6 Low-level laser therapy is non-thermal and is believed to reduce pain and inflammation, although the exact mechanism of its effect is unknown.16 Acupuncture is a form of ancient Chinese medicine in which fine needles are inserted into the skin at certain points on the body. Chiropractic uses a range of manual therapies, with an emphasis on manipulation of the spine, whereas osteopathy uses gentle stretching, massage and manipulation of muscles and joints.17
  • Intra-articular corticosteroid injections to reduce inflammation and provide pain relief. A range of different doses and number of injections is described in the literature. This intervention is usually delivered in the primary care setting but also in the secondary care setting, depending on how services are organised in a particular region.
  • Arthrographic distension (also called hydrodilatation), which involves controlled dilatation of the joint capsule with sterile saline or other solution such as local anaesthetic or steroid, guided by radiological imaging (arthrography). The procedure is performed under local anaesthetic.
  • MUA, in which the shoulder is freed by rotation while the patient is under short general anaesthesia. This can be undertaken as a day procedure.
  • Arthroscopic capsular release, a surgical procedure conducted under general or regional anaesthesia during which the contracted tissue is released. Open capsular release is another surgical option, usually recommended in those resistant to arthroscopic intervention. Both can be undertaken as a day procedure.

These interventions can be used individually or in combination, depending on the disease stage. There is currently not a consensus about the overall management of the condition and the sequence in which treatments should be offered to patients. A recent survey of 303 UK health-care professionals [general practitioner (GPs), including those with a special interest, physiotherapists, advanced scope physiotherapists and orthopaedic surgeons] found that the professional groups had different views on the most appropriate treatment pathway for frozen shoulder.18 However, there appears to be a fairly consistent view that the treatments used should depend on the phase of the disease and/or that a step-up approach should be adopted in terms of the degree of treatment invasiveness.5,12,18,19 There is a suggestion that aggressive mobilisation should be avoided in the early, severely painful phase.12,14,19 Surgical intervention is generally, although not exclusively, used when the condition is resistant to the other interventions,19 although there is no consensus as to what time point or level of pain should indicate surgical intervention.18

The most commonly used or recommended interventions for the painful phase, in the recent UK survey of health-care professionals, were conservative treatment (watchful waiting, education, oral pain relief) and physical therapy (mainly physiotherapy and mobilisation), each recommended in one-third of responses, and intra-articular steroid injection, recommended in 18% of responses. For patients in the resolution phase, surgery was the most preferred option (mainly MUA and arthroscopic capsular release), recommended in almost half of the responses, followed by conservative treatment (12%) and physical therapy (19%).

Previous systematic reviews

Several systematic reviews had previously been undertaken on interventions for frozen shoulder, some of which focused on shoulder pain in general and included a range of conditions. Those interventions evaluated were oral steroids,20 corticosteroid injections,2123 physiotherapy,2426 acupuncture,27 arthrographic distension28 and multiple interventions29 (Table 2). None of the literature searches was recent and the reviews required updating (Table 2). In addition, some of the reviews did not report results by type of shoulder complaint. A recent review of systematic reviews on frozen shoulder noted the tendency to focus on a single intervention and that updated reviews were required.30

TABLE 2. Previous systematic reviews.


Previous systematic reviews.

It is apparent from previous reviews that there is variation in how frozen shoulder is defined across studies. A review of 21 randomised controlled trials (RCTs) of interventions for frozen shoulder could not derive a consistent description of the condition from the trials investigating this patient group.2 Although the RCTs required that participants had restricted shoulder movement, there was inconsistency across trials in the number of degrees of restriction, the type of restriction (active or passive) and the direction of the restriction (abduction or external rotation).2 This highlights the difficulty of applying a strict definition for frozen shoulder within the context of a systematic review.

Focus of the synthesis

We undertook a systematic review of a range of NHS-relevant interventions for the treatment of primary frozen shoulder with the aim of informing a decision-analytic model. Previous reviews have tended to assess single treatments. The main focus was on comparing the main treatment options for frozen shoulder and identifying the most appropriate treatment by stage of frozen shoulder. Although variations within treatments were included in the synthesis, the main interest was in questions such as whether there is additional benefit from providing physiotherapy following steroid injection. Given the range of interventions being considered and the possibility that all treatment options would not have been compared in head-to-head in trials we planned to undertake a mixed-treatment comparison (MTC) in addition to pair-wise meta-analysis. MTC is an extension of traditional meta-analysis in which trials comparing the same intervention and the same comparator are pooled to estimate an overall treatment effect. A MTC overcomes the limitations of standard meta-analysis when there are no or few head-to-head comparisons or when the decision problem requires the comparison of several interventions.31,32 A ranking of interventions, based on the probability that each treatment is best,33 can be produced. This is of particular value when several treatment options are under consideration.

© 2012, Crown Copyright.

Included under terms of UK Non-commercial Government License.

Cover of Management of Frozen Shoulder: A Systematic Review and Cost-Effectiveness Analysis
Management of Frozen Shoulder: A Systematic Review and Cost-Effectiveness Analysis.
Health Technology Assessment, No. 16.11.
Maund E, Craig D, Suekarran S, et al.
Southampton (UK): NIHR Journals Library; 2012 Mar.

PubMed Health Blog...

read all...

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...