Complex regional pain syndrome: working through the Budapest criteria properly

The wrist that should be settling by now

A distal radius fracture comes out of plaster. The bone has healed on the follow-up X-ray, but the pain hasn't settled the way anyone expected. The hand is stiff, maybe a little puffy, and the person mentions that it feels hot one day and cold the next, or that a sleeve brushing the skin is unbearable.

It's easy to file that under post-cast stiffness and get on with mobilising. Most of the time that's reasonable. Occasionally it's the early picture of complex regional pain syndrome (CRPS), and the Royal College of Physicians' UK guideline is blunt that, because the condition is uncommon and mimics many others, patients commonly experience a delay in diagnosis and in starting appropriate treatment.

CRPS is a clinical diagnosis. There's no scan or blood test that confirms it. That puts the job squarely in the examination room, and it means the quality of the diagnosis depends on how carefully someone works through a short set of criteria. This article goes through those criteria, how to examine for them, what has to be ruled out first, and what a likely CRPS should change in your plan.

What CRPS is, and what it isn't

The RCP guideline (Goebel and colleagues, 2018) describes CRPS as a debilitating, painful condition in a limb, associated with sensory, motor, autonomic, skin and bone abnormalities. It commonly follows an injury to that limb, but there's no relationship to the severity of the trauma, and in some cases (9% in the guideline's figures) there's no precipitating trauma at all.

A few older ideas are worth dropping. The guideline says the concepts that CRPS is predominantly a sympathetic problem, and that it runs through stereotyped stages, are now obsolete. It also states that CRPS is not associated with a history of pain-preceding psychological problems, somatisation or malingering, and notes that patients still report suffering from interactions with health professionals who don't believe their condition is real. If you trained when the condition was called reflex sympathetic dystrophy, Sudeck's atrophy, algodystrophy or shoulder–hand syndrome, those are all earlier names for the same group of conditions.

It's uncommon, but not rare enough to ignore. In a Dutch population-based study using primary care records, de Mos and colleagues (Pain 2007) estimated an incidence of 26.2 per 100,000 person-years. Women were affected at least three times more often than men, the highest incidence was in women aged 61 to 70, the upper limb was affected more often than the lower limb, and a fracture was the most common precipitating event (44%).

CRPS is divided into type 1, without a major nerve lesion and much more common, and type 2, with one. The RCP guideline notes that the distinction currently has no relevance for management outside some surgical situations.

The Budapest criteria, line by line

The criteria now used internationally are the IASP diagnostic criteria, usually called the Budapest criteria. All four parts must apply:

• A. Continuing pain that is disproportionate to any inciting event.
• B. At least one symptom reported in three or more of the four categories below.
• C. At least one sign present on examination in two or more of the categories.
• D. No other diagnosis better explains the signs and symptoms.

The four categories are:

1. Sensory: allodynia (to light touch, temperature, deep pressure or joint movement) and/or hyperalgesia to pinprick. A reported increase in skin sensitivity (hyperaesthesia) counts as a symptom.
2. Vasomotor: temperature asymmetry, skin colour changes or skin colour asymmetry.
3. Sudomotor/oedema: oedema, sweating changes or sweating asymmetry.
4. Motor/trophic: decreased range of motion, motor dysfunction (weakness, tremor, dystonia) or trophic changes in hair, nails or skin.

The distinction between symptoms and signs is the part most often blurred in practice. A symptom is something the person reports. A sign is something you can see or feel on examination at the time of evaluation. Someone can describe their hand going blotchy and sweaty in the evening, which counts towards the symptom count, but that only counts as a sign if you can see or feel it yourself.

Why the criteria changed

The original 1994 IASP criteria were sensitive but over-inclusive. Harden and colleagues' validation study (Pain 2010) compared structured evaluations of 113 people with CRPS type 1 and 47 people with non-CRPS neuropathic pain. The 1994 criteria had a sensitivity of 1.00 but a specificity of only 0.41. The Budapest clinical criteria kept nearly all of that sensitivity (0.99) and raised specificity to 0.68. A stricter research version, intended to reduce false positives in research samples, reached a specificity of 0.79, at the cost of the lowest sensitivity of the criteria sets tested. The analysis suggested that requiring features across four distinct components was what improved specificity.

How much the choice of criteria matters is clearest in a prospective Dutch cohort of 596 people with a single fracture of the wrist, scaphoid, ankle or fifth metatarsal (Beerthuizen and colleagues, Pain 2012). Using the 1994 IASP criteria, 48.5% met the definition of CRPS type 1. Using Veldman's criteria, 21.3% did. Using the stricter Harden and Bruehl criteria, 7.0% did. Same patients, same fractures. The authors' conclusion was that the incidence of CRPS after a fracture depends to a large extent on the criteria used.

So the criteria aren't a formality. Loose application turns ordinary post-injury pain and swelling into a diagnosis that will follow someone around. Skipping them altogether means people with genuine CRPS wait.

Examining for it properly

In 2019 an IASP special interest group met in Valencia to address ambiguities in how the Budapest criteria were being applied. The resulting consensus (Goebel and colleagues, Pain 2021) deliberately left the criteria wording unchanged, but clarified the assessment. Several points translate directly into clinic.

Ask about every symptom category, every time. CRPS signs and symptoms fluctuate, so the consensus recommends asking systematically about all the listed symptoms at each formal diagnostic evaluation, even ones the person hasn't mentioned before.

Compare with the other side. For a unilateral presentation, asymmetry means the affected side compared with the unaffected side, and it's based on clinical judgement rather than a prespecified threshold.

Test hyperalgesia with a single pinprick. Apply one pinprick in the centre of the most affected region and an identical pinprick at the corresponding spot on the other limb. The test is positive if the pain is more intense or lasts longer on the affected side.

Test allodynia with a stimulus that shouldn't hurt. Light touch or a brush, vibration from a tuning fork over a bony prominence, cool or warm temperature, or pressure. For pressure, the consensus describes pressing with just enough force to make your own fingernail bed blanch, which is far less than the pressure used for tender point examination. Only one positive stimulus is needed.

Feel for temperature asymmetry with the back of your hand. It should be obvious to the touch of the dorsum of the examiner's hand when compared with the corresponding area on the other limb.

Describe colour changes specifically. The consensus asks for obvious regional colour asymmetry (hand, foot, knee or larger), and for the nature of the change to be recorded: red, blue, pale or mottled.

Ask the motor question that separates pain from something else. The RCP checklist points out that in CRPS, reduced movement or weakness in the motor/trophic category isn't due to pain, nerve damage or a joint or skin problem. It suggests asking: "If I had a magic wand to take your pain away, could you then move your fingers?" Many people with CRPS answer no.

The Valencia group also tidied up the labels for people who don't meet every criterion. "CRPS with remission of some features" is for someone who previously met full criteria but no longer does. "CRPS not otherwise specified" is now reserved for people who have never been documented as meeting full criteria but have some features that nothing else explains. That's a change from the RCP's 2018 wording, which applied NOS to both groups. Both documents make the same clinical point: losing some limb signs isn't the same as recovery. The RCP guideline notes that swelling, sweating and colour or temperature changes usually reduce with time even where pain and motor symptoms persist, and that where pain persists the condition is best considered active.

Criterion D is a clinical safety job

It's tempting to tick A, B and C and treat D as a footnote. The RCP guideline calls CRPS a diagnosis of exclusion and lists differentials that include:

• infection of bone, soft tissue, joint or skin
• orthopaedic mal-fixation and joint instability
• arthritis or arthrosis
• bone or soft tissue injury, including stress fracture or ligament damage
• compartment syndrome
• peripheral nerve injury, compression or entrapment, spinal or central lesions, or neuropathy
• thoracic outlet syndrome
• arterial insufficiency, Raynaud's, and lymphatic or venous obstruction
• brachial neuritis (Parsonage–Turner syndrome)
• self-harm

The list isn't exhaustive, and several of those are urgent. The surgical section adds that "scalding" pain in the distribution of a peripheral nerve after injury or surgery should be reviewed urgently by the surgeon because of the possibility of nerve damage.

There are no specific investigations that confirm CRPS, but investigations may be needed to rule out other causes. That's where your referral route matters. For someone who has already been discharged by the trauma or surgical team, the guideline suggests the GP consider re-referral to allow ongoing pathology to be definitely excluded. Where there was no preceding trauma, or only minor trauma, it recommends referral to secondary care, such as rheumatology, to exclude or address specific pathology, and advises that an isolated referral to physiotherapy or occupational therapy is arranged with caution and only when the referrer is certain there's no identifiable underlying cause.

The guideline also says that, although experienced therapists may make a diagnosis of CRPS, it should generally be confirmed by a doctor, such as the patient's consultant or GP.

Timing: too early to call, too late to miss

Early after an injury, the RCP guideline says a CRPS diagnosis is difficult to secure, because the signs and symptoms are often indistinguishable from the normal response to injury or surgery. Pain out of proportion to the injury is common at that stage, and the vast majority of those responses resolve. The emergency medicine section recommends that CRPS isn't diagnosed until 2 to 4 weeks after the triggering trauma, and that people with enhanced pain early after injury are reassured that it settles in most cases. The guideline also notes that the onset of symptoms for the majority occurs within a month of the trauma or immobilisation of the limb.

That doesn't mean early pain is uninformative. Moseley and colleagues (Journal of Pain 2014) followed 1,549 people with a wrist fracture managed nonsurgically, assessed within a week of the fracture and again four months later. CRPS developed in 3.8%, about one in 26. A model using four clinical assessments discriminated very well between those who did and didn't go on to develop CRPS, and a simple 0 to 10 pain rating did almost as well on its own. The authors suggest treating a pain score of 5 or more in the first week after fracture as a "red flag" for CRPS.

The RCP guideline adds that a tight plaster or dressing, a sense of tightness, and disproportionate pain in plaster or when it comes off may be early warning signs. Its emergency medicine section suggests that people with excessive, disproportionate pain early after injury who aren't being followed up elsewhere should be considered for review four to six weeks after the trauma, in clinic or through their GP, to establish a definite diagnosis.

In practice, high early pain is a reason to plan a review, not to apply a label. Persistent disproportionate pain beyond the first few weeks is a reason to run the full checklist rather than assume stiffness.

What a likely CRPS changes in your plan

The RCP guideline builds management around four pillars of equal importance: education, pain relief, physical rehabilitation and psychological intervention. For therapists, its recommendations are to recognise the clinical signs, know the Budapest criteria, start treatment as early as possible, educate the person about the condition, know where the nearest multidisciplinary pain clinic or CRPS specialist centre is, and initiate referral when moderate or severe symptoms aren't resolving.

Its rehabilitation algorithm gives practical referral triggers. An early referral to a multidisciplinary pain clinic or specialist unit is recommended when there are moderate to severe signs and symptoms (unless they're of very recent onset after trauma and resolving quickly), when dystonia is present, when there's no positive response to treatment within four weeks, or when the condition deteriorates or improvements aren't sustained. Treatment should continue while the person waits to be seen.

The guideline is also clear about immobilisation. Splinting is generally short term and limited to where it's clinically appropriate, avoiding prolonged periods of immobilisation or keeping the limb covered up. Early after injury, where it isn't contraindicated, people are encouraged to touch and stroke the skin of the painful area and to use the limb gently, even when that seems counterintuitive.

The evidence for specific physical treatments is weak, and it's worth saying so. The 2022 Cochrane review of physiotherapy for CRPS (Smart and colleagues) included 34 randomised trials with 1,339 participants. Most interventions were tested in small single trials, most trials were at high risk of bias, and the certainty of evidence was rated very low for every comparison and outcome where evidence was found. The authors concluded that the evidence is very uncertain about the effects of physiotherapy interventions on pain and disability in CRPS. Rehabilitation is still one of the four pillars in the UK guideline, so the uncertainty argues for describing it honestly and not promising that any particular protocol will fix the problem.

How you explain the diagnosis matters too. The RCP guideline notes that being told about CRPS can be distressing and should be done carefully and with reassurance. It reports that around 15% of people have unrelenting pain and physical impairment two years after onset, and that the majority get better. A clear diagnosis, clear information, realistic goals and encouragement to keep gently using the limb are a better start than either silence or a list of worst cases.

What changes on Monday morning

For the next person whose limb pain isn't settling after a fracture, surgery or a minor injury:

1. Rate pain early. In the first week after a wrist fracture, a pain score of 5 or more out of 10 is a reason to record it and book a follow-up review.
2. Don't label too early. Disproportionate pain in the first few weeks is common and usually resolves. CRPS shouldn't be diagnosed until 2 to 4 weeks after the trigger.
3. When pain stays out of proportion, run the full checklist. Ask about all four symptom categories, then examine for signs in each, comparing with the other side.
4. Record symptoms and signs separately. Note which categories were reported and which you found yourself, plus the pinprick, allodynia, temperature and colour findings.
5. Take criterion D seriously. Work through the differentials, and refer back to the trauma team, GP or secondary care when something else needs excluding. Urgent features need urgent review.
6. Get the diagnosis confirmed and know your referral route. Refer when there's dystonia, moderate to severe features, deterioration, or no positive response within four weeks, and keep treating while the person waits.
7. Keep the limb in use and explain it well. Avoid prolonged immobilisation, encourage gentle touch and use, and make it clear that you believe the pain is real.

None of this needs new equipment. It needs a pin, a brush, the back of your hand and a willingness to work through four short criteria properly, especially the last one.

Keeping assessment frameworks like this current is the kind of learning that is easy to put off. ManualCPD offers online, on-demand CPD for chiropractors, physiotherapists, osteopaths and sports therapists, broken into manageable sections you can work through between patients, at lunch or after clinic. Every ManualCPD course is accredited by The CPD Group (Provider #788442), and you receive a CPD certificate on completion.