Approach to management

• Frozen shoulder is common, and is usually managed in primary care with a combination of analgesia, injections, and physiotherapy [Dias et al, 2005].
• There are a lack of well-designed trials on the common treatments for chronic shoulder disorders, but most people can be treated conservatively with some combination of activity modification, physical therapy, medication, and corticosteroid injection (if necessary); this approach produces satisfactory results in the majority of people [Burbank et al, 2008].

Advice

• A detailed explanation of the natural history is important so that people do not develop unrealistic expectations [NZGG, 2004]. The natural history of frozen shoulder is for resolution (complete relief of pain and partial improvement of range of movement) over 18–24 months, regardless of treatment [Dias et al, 2005; NHS Lothian, 2007a]. Vigorous stretching in the early, painful phase of a frozen shoulder will exacerbate the pain [NZGG, 2004].

Analgesia

• A New Zealand guideline on the diagnosis and management of soft tissue shoulder injuries and related disorders [NZGG, 2004], two reviews on the management of shoulder pain in primary care [Hazleman, 2005; Mitchell, 2008], and a review of the treatment of chronic shoulder pain [Burbank et al, 2008] recommend analgesia for the pain of frozen shoulder.
• Pain relief is particularly important in the initial painful stage [Dias et al, 2005] and may be useful in helping compliance with physical therapy [Moskowitz and Blaine, 2005].
• Moderately-strong analgesics are required, but rarely adequately control the pain at night [NZGG, 2004].
• CKS found no evidence on the use of paracetamol, nonsteroidal anti-inflammatory drugs (NSAIDs), or codeine for the treatment of shoulder pain caused by frozen shoulder, therefore the decision should be based on clinical judgement, taking into account the severity of the person's symptoms and comorbidities.
• For more information on the adverse effect profile and contraindications to prescribing NSAIDs, see the CKS topic on NSAIDs - prescribing issues.
• CKS has not recommended the use of stronger opioids in primary care as long-term treatment should be undertaken with caution because of the potential for dependence [Iannotti and Kwon, 2005; Moskowitz and Blaine, 2005].

Physiotherapy

• The optimal timing of interventions is uncertain, but CKS recommends referral for physiotherapy only if the person can tolerate movement, as exercise in the acutely painful phase will exacerbate the pain; it should be initiated as the acute pain settles. A supervised exercise programme has been found to give a faster improvement in range of movement than a simple home programme of gentle exercise [NZGG, 2004].
• CKS recommends ensuring adequate analgesia is provided so that the person is able to tolerate the movements.
• CKS found no evidence that physiotherapy alone is beneficial for people with frozen shoulder, although a Cochrane systematic review found limited evidence that a supervised exercise regimen is beneficial for mixed shoulder disorders in the short and long term [Green et al, 2003]. Compared with intra-articular corticosteroid injections, there is limited evidence that physiotherapy appears to be less effective in the short term, but not the long term [van der Windt et al, 1998; Speed, 2008].

Corticosteroid injections

• A New Zealand guideline on the diagnosis and management of soft tissue shoulder injuries and related disorders [NZGG, 2004], two articles on the management of shoulder pain in primary care [Hazleman, 2005; Mitchell, 2008], and reviews of the treatment of chronic shoulder pain [Moskowitz and Blaine, 2005; Burbank et al, 2008] recommend corticosteroid injections for frozen shoulder.
• Efficacy: many cases of capsulitis settle spontaneously but if no, or poor, improvement is seen after some weeks with initial treatment, early corticosteroid injections (together with analgesia and physiotherapy) are usually therapeutic [Hazleman, 2005; Burbank et al, 2008]. CKS found limited evidence suggesting that intra-articular steroid injections for people with frozen shoulder may be more effective than placebo at relieving pain and improving disability and function in the short term, but not the long term [Speed, 2008].
• Timing: early treatment with a steroid injection into the glenohumeral joint may reduce synovitis and pain and increase the range of motion in the short term. Steroid injections should therefore be given early on (during the active inflammatory stage that lasts 2–3 months) [NZGG, 2004; Dias et al, 2005; Hazleman, 2005]. The interval between injections should be at least 3–4 weeks [Silver, 2002]. Steroid injections are not recommended in the adhesive phase as the inflammatory stage is over [Dias et al, 2005].
• CKS found no evidence on when to give a corticosteroid injection in relation to physiotherapy. Reviews suggested that people with adhesive capsulitis have been shown to respond to intra-articular injections with decreased pain and increased function, especially in combination with physical therapy for stretching and that giving a corticosteroid injection may increase the benefits of physical therapy in people with adhesive capsulitis [Iannotti and Kwon, 2005; Moskowitz and Blaine, 2005; Burbank et al, 2008].
• Choice of corticosteroid and use of lidocaine:
• CKS found no guidelines informing the choice of corticosteroid. A number of corticosteroid preparations are listed in the British National Formulary [BNF 55, 2008], but feedback from expert reviewers suggested that triamcinolone or methylprednisolone are the most commonly used.
• Feedback from expert reviewers also suggested that local anaesthetic is commonly used in conjunction with a corticosteroid. Reasons given for this included that local anaesthetic can give immediate pain relief, can help confirm the diagnosis, and may decrease the risk of steroid atrophy by diluting the corticosteroid.
• CKS found evidence from one randomized controlled trial (RCT) that intra-articular corticosteroid injections plus lidocaine may be no more effective at 4–24 weeks than lidocaine or placebo at relieving pain in people with frozen shoulder, and at 24 weeks, intra-articular corticosteroid injections plus lidocaine may be no more effective than lidocaine in terms of improving range of motion [Speed, 2008].
• Number of injections: local injection with corticosteroids may be useful for people with persistent shoulder pain, but should not be given if minimal or no benefit was obtained from previous injection [Iannotti and Kwon, 2005]. Evidence from a systematic review found that up to three corticosteroid injections were beneficial for frozen shoulder until 16 weeks from the date of the first injection [Shah and Lewis, 2007].
• Adverse effects: CKS advises discussion of adverse effects when considering corticosteroid injections. These include infection (rare if sterile technique used), tendon rupture, and hyperglycaemia in people with diabetes [NZGG, 2004].
• A New Zealand guideline advises monitoring blood glucose in people with diabetes after corticosteroid injection as this may be elevated for 24–48 hours [NZGG, 2004].

• A New Zealand guideline on the diagnosis and management of soft tissue shoulder injuries and related disorders [NZGG, 2004] and an article on the management of shoulder pain in primary care [Mitchell, 2008] recommend referral of frozen shoulder if there is diagnostic uncertainty, or red flag features.
• The need for surgical intervention (e.g. manipulation under anaesthesia [MUA] or arthroscopic capsular release) is rare and referral is not indicated until the person has failed 6 months of non-operative treatment, or if the diagnosis is in question [Burbank et al, 2008].
• MUA may be effective, but subsequent mobilization through physiotherapy is essential. Fractured humerus is a recognized complication of MUA, however [Hazleman, 2005]. A review advises referral for MUA if the person cannot tolerate the pain and disability of frozen shoulder [Dias et al, 2005].