Frozen shoulder
BMJ 2022; 377 doi: https://doi.org/10.1136/bmj-2021-068547 (Published 21 April 2022) Cite this as: BMJ 2022;377:e068547
- Marta Karbowiak, trauma and orthopaedics core trainee1,
- Thomas Holme, trauma and orthopaedics registrar2,
- Maisum Mirza, general practitioner3,
- Nashat Siddiqui, consultant orthopaedic and upper limb surgeon2
- 1Royal Hampshire County Hospital, Winchester, UK
- 2Kingston Hospital NHS Foundation Trust, Kingston upon Thames, UK
- 3Warlingham Green Medical Practice, Warlingham, UK
- Correspondence to M Karbowiak mkarbowiak{at}doctors.org.uk
What you need to know
Patients with diabetes are at higher risk of developing frozen shoulder and having bilateral symptoms than the general population
Recovery times vary, but can be years, and some patients are left with residual pain or functional impairment
Physiotherapy is the most commonly used intervention and can be supplemented by intra-articular steroid injections
Treatments offered in secondary care include joint manipulation under anaesthesia, arthroscopic capsular release, and hydrodilatation
The UK FROST trial compared manipulation under anaesthetic, arthroscopic capsular release, and early structured physiotherapy with intra-articular corticosteroid injections, and found that none of the interventions were clinically superior
Frozen shoulder is a common and often debilitating condition that lacks a clear consensus on management, partly owing to a lack of high quality evidence on the various treatments options. In this clinical update, we offer an overview of the latest evidence on management of frozen shoulder, incorporating the clinical implications of recently published research, including the UK FROST study—the largest randomised controlled trial in this field to date, which compares surgical treatments with early structured physiotherapy and intra-articular corticosteroid injections.
What is frozen shoulder?
Frozen shoulder is a condition that results in development of thickened, fibrosed joint capsule, contraction of the joint, and reduced intra-articular volume.1 The exact cause of these changes is unknown, with several possible processes suggested in the literature.1 Over the years, uncertainty has surrounded the definition and classification of this condition, leading to inconsistencies in both clinical practice and scientific studies.2 This is partially owing to the wide spectrum of clinical presentations, with patients experiencing different levels and combinations of symptoms. This also means their lives can be affected in many different ways, depending on the severity of the condition and their daily activities.
Who gets it?
The age of onset is usually in the fifth decade of life, with peak incidence between the ages of 40 and 60.3 Women are more commonly affected than men, with one study reporting the incidence as 3.38 and 2.36 per 1000 person years, respectively.4
Patients with diabetes have a 10% to 20% lifetime risk of developing frozen shoulder,56 and are more likely to have bilateral shoulder involvement than the general population.7 Frozen shoulder has been linked to conditions such as hypothyroidism, hypercholesterolaemia, and heart disease, although evidence is insufficient to determine whether these associations are independent.8
How is it diagnosed?
Frozen shoulder is primarily a clinical diagnosis (box 1). Patients can present with a range of symptoms related to the shoulder, although pain is often the initial trigger for presentation. Three distinct phases are commonly described,11 with each phase typically lasting several months:
Diagnosis of frozen shoulder9
History—insidious onset of shoulder pain, often anterolateral initially; pain at night; sometimes minimal trauma associated around time of onset
Examination—painful movement restriction, passive external rotation less than 30°, passive elevation less than 100°; cases where the disease affects the posterior capsule more than the anterior can present with reduction in internal rotation
Investigations—plain radiographs are useful to check for arthritic changes in the glenohumeral joint and are recommended by the British Elbow and Shoulder Society10; ultrasound and magnetic resonance imaging may be considered depending on the clinical features and differential diagnoses
Freezing/proliferative phase—stiffness with progressively worsening pain (usually constant but exacerbated by movement)
Frozen/adhesive phase—ongoing stiffness with improved pain levels, reduction in range of motion, in particular on external rotation
Thawing/resolution phase—gradual improvement in range of motion.
The clinical course can be variable2 and not all people with frozen shoulder will experience all three of these stages. Pain at night is a common feature, often causing considerable disruption to sleep. Patients can also experience sudden jerking movements associated with pain.2 Diagnostic pointers for frozen shoulder are summarised in box 2, and differential diagnoses are listed in table 1.
When to refer to secondary care
British Elbow and Shoulder Society guidelines10 advise to refer:
Cases of atypical presentation or marked functional limitation
Persistence of pain despite primary care interventions beyond three months
American Family Physician guidelines recommend referral to a shoulder specialist if no improvement is seen with physiotherapy and corticosteroid injections after three months12
RETURN TO TEXTDifferential diagnoses
What is the clinical course of frozen shoulder?
Frozen shoulder is often described in literature as a “self-limiting” condition, and patients typically experience resolution of symptoms without or regardless of any treatment.13 Most people with the condition make a full recovery, although recovery time tends to be slow—between one and three years.1415 Some experience residual symptoms: the original prospective study on frozen shoulder from 1975 found that half of patients had residual clinical restriction in range of movement after 5-10 years, and 7% had ongoing functional limitation.11 Similar rates were reported in more recent literature, with one study6 of patients under the care of a specialist shoulder clinic followed up at average 52 months finding that 41% reported residual symptoms. Recurrence of primary frozen shoulder after the initial resolution of symptoms is poorly reported in literature, but in our experience is rare. Up to 20% of patients can develop the condition on the opposite side.5 Patients with diabetes generally have poorer response to treatment and, with interventions such as manipulation under anaesthetic, are at higher risk of requiring further procedures.16
How is frozen shoulder managed?
After establishing a clinical diagnosis of frozen shoulder, explain the typical progression of the condition. Discuss the range of available management options and the risks associated with each intervention (table 2). An individual approach involving exploring the extent of functional limitation and establishing treatment goals can aid in deciding the appropriate treatment.
Summary of treatment options
Advise patients to continue to use the arm as pain allows.9 Over-the-counter or prescription painkillers can help to alleviate pain, which is often the most debilitating symptom experienced in the early stages and can limit engagement with physiotherapy. Sleeping on the unaffected side or using pillows for support in bed can help with night time pain. Heat or ice packs over the affected area can be used for additional pain relief. Shoulder stiffness can lead to other musculoskeletal symptoms, most commonly neck and lower back pain, which can also be targeted with physiotherapy. In the early stages, we recommend patients try simple home exercises such as placing things higher up to encourage reaching, gentle stretching, and pendulum exercises.17
Physiotherapy
The main role of physiotherapy is in the frozen/adhesive phase (when the initial symptoms of pain have subsided) with stretches and strengthening exercises. This should be sustained with additional resistance based exercises in the thawing/resolution phase.18 Structured approaches include group or individual physiotherapy, with formal range of motion exercises, soft tissue massage, and trigger point release.19 The recommended initial treatment course is six to 12 weeks.5
Corticosteroid injections
Corticosteroid injections can help in reducing pain and improving range of movement, particularly in the early stages of the condition and when combined with physiotherapy. A 2020 systematic review and meta-analysis on the management of frozen shoulder assessed the effectiveness of available treatment strategies across 65 studies with 4097 patients.20 The authors found that intra-articular (IA) corticosteroid injections were associated with short term improvement in external rotation and pain compared with no treatment or placebo. IA corticosteroid injections with physiotherapy were found to be superior to IA corticosteroid injections alone for early range of motion only. Physiotherapy with IA corticosteroid injections was found to be superior to physiotherapy alone for short term outcomes using several symptom and functional scoring systems, but not for range of motion or medium term function.
IA steroid injections are associated with better pain relief compared with subacromial injections.21 Interestingly, a 2021 clinical trial found that, even though ultrasound guided IA injections were associated with greater accuracy than blind IA injections, no difference was seen between them in pain and functional outcome scores.22
Corticosteroid injections are generally considered safe and are associated with mild side effects only. In one study, three of 58 patients (5.2%) reported mild self-limiting nausea and dizziness.23 Another reported that, of 133 participants, one patient (0.7%) experienced prolonged pain at the injection site, and three patients (2.3%) developed transient facial flushing.24 Steroid injections can affect blood glucose control in patients with diabetes, particularly in the first day after the intervention.25
Surgical options—arthroscopic capsular release and manipulation under anaesthetic
Arthroscopic capsular release (ACR) is a surgical procedure carried out under general or regional anaesthesia. The shoulder capsule is divided using arthroscopic instruments and the shoulder is re-examined to confirm optimal release. Manipulation under anaesthetic(MUA) is a procedure where the shoulder is manipulated by the surgeon to stretch and tear the joint capsule. It is carried out under general or regional anaesthesia.
The recently published UK FROST study compared ACR, MUA, and early structured physiotherapy with intra-articular corticosteroid injection.26 The study is the largest randomised controlled trial of these interventions to date, with 503 participants recruited across 35 UK hospitals. In UK FROST, short term outcomes of ACR at three months were overall worse compared with physiotherapy with corticosteroid injection or MUA. However, at 12 months ACR was found to be associated with better functional scores (Oxford Shoulder Score, OSS) compared with both MUA and physiotherapy (OSS difference 2.01 and 3.06, respectively), although this was less than the clinically important effect size of 4-5 OSS points. The study authors concluded that none of the three interventions was clinically superior, but that ACR carried higher risks (3.9% in this cohort had a serious adverse event compared with 1% of those who had MUA), and MUA was the most cost effective intervention.
One weakness of UK FROST is that it was unable to determine to what extent the improvements in outcomes were the result of the interventions rather than the natural course of the condition. A randomised controlled trial of 125 patients in Finland17 comparing MUA with supportive treatment (home exercises) found no difference in terms of pain levels or functional ability between the two groups at 12 month follow-up; minimal differences were noted in range of motion in favour of the MUA group. Another smaller RCT comparing ACR with supportive care did not find any significant differences between the two in functional outcome scores.27
MUA is generally associated with low rates of adverse events—UK FROST26 recorded two serious (1%) and 15 non-serious (7.5%) adverse events for 201 patients. These were largely minor reactions such as residual stiffness, nerve pain, and paraesthesia. One case (0.5%) of septic joint arthritis was recorded, and three patients (1.5%) experienced postoperative worsening of shoulder pain. MUA was found to be cost effective as functional improvement was seen sooner, meaning less need for prolonged physiotherapy and follow-up.
Hydrodilatation
Hydrodilatation involves injecting fluid into the shoulder joint to disrupt the capsular adhesions, and is usually performed in a clinic setting. Solutions used and volumes injected vary in literature, but most clinicians use normal saline with local anaesthetic and corticosteroid.28 Hydrodilatation was not included in the UK-FROST study as it was not widely available until recently. It has since become a common management option alongside MUA and ACR in many UK secondary care centres.
A 2008 Cochrane review29 noted hydrodilatation was associated with short term improvement in pain, function, and range of motion. However, a more recent 2018 systematic review and meta-analysis28 found that the procedure had an overall insignificant effect on clinical outcomes—the authors noted minimal improvement in pain and range of motion (number needed to treat of 12) with no significant improvement in disability. One small randomised controlled trial30 comparing MUA and hydrodilatation found that functional scores were significantly better in the hydrodilatation cohort with higher patient satisfaction rates, although the study was conducted in a small patient cohort (38 joints in total). At present, comparisons with surgical treatments are difficult owing to a lack of high quality evidence. A Delphi study31 backed by the British Elbow and Shoulder Society is under way to help inform future directions of research. Adverse events reported with the use of hydrodilatation include pain, flushing, syncopal episodes, and one case of glenohumeral joint infection.28
Education into practice
How might you explain the different stages of frozen shoulder to a patient first presenting with the condition?
What strategies could be used in the community to widen the access to physiotherapy for patients?
How patients were involved in the creation of this article
A patient with experience of the condition reviewed a draft of the manuscript. In response to their feedback, we developed the section on management to include more detail about analgesia and adjuncts to physiotherapy.
How this article was created
We conducted a Medline search using the terms “frozen shoulder” and “adhesive capsulitis” to identify the relevant references and review the latest published evidence. We also searched the Cochrane library and consulted the British Elbow and Shoulder Society website for current guidelines.
Information resources for patients
British Elbow & Shoulder Society Website: advice on common shoulder conditions, when to seek medical advice, and recommended self-care measures: https://bess.ac.uk
Patient.info: information about the condition and management: https://patient.info/bones-joints-muscles/frozen-shoulder-leaflet
American Academy of Orthopaedic Surgeons: resource outlining treatment options and home exercises: https://orthoinfo.aaos.org/en/diseases--conditions/frozen-shoulder/
The Chartered Society of Physiotherapy: information on managing shoulder pain and recommended exercises: https://www.csp.org.uk/publications/shoulder-pain-exercises
Footnotes
We have read and understood the BMJ policy on declaration of interests and declare the following interests: none.
Provenance and peer review: commissioned, based on an idea from the author; externally peer reviewed.
References
Log in
Log in using your username and password
Log in through your institution
Subscribe from £164 *
Subscribe and get access to all BMJ articles, and much more.
* For online subscription
Access this article for 1 day for:
£30 / $37 / €33 (excludes VAT)
You can download a PDF version for your personal record.