The Stages of Complex Regional Pain Syndrome (CRPS/RSD)

Complex regional pain syndrome (CRPS), still widely called reflex sympathetic dystrophy or RSD, is traditionally described in three stages: an acute inflammatory stage, a dystrophic stage, and an atrophic stage. The staging language is useful shorthand for what a limb can look like, but it is not a timeline — the largest studies that looked for three consecutive phases did not find them, and how long a person has had CRPS does not predict which “stage” they resemble.¹²

This page explains what the three stages describe, why sequential staging was abandoned, and what clinicians use in its place: the Budapest diagnostic criteria, and a simpler description of the affected limb as warm or cold.

What the classic three-stage model describes

The three-stage model dates to clinical descriptions written long before anyone had a mechanism for the condition. It grouped findings that tend to appear together. The time ranges attached to each stage differ between accounts, and no prospective study has followed patients through them in order.¹ Treat the ranges below as the way the model is taught, not as a schedule any individual follows.

Stage 1 — the acute or “warm” picture

Burning or aching pain out of proportion to the original injury, swelling, a limb that is warmer and redder than the other side, increased sweating, faster nail and hair growth, and pain provoked by light touch (allodynia) or by ordinarily mild pressure (hyperalgesia). Usually described as roughly the first one to three months.

Stage 2 — the dystrophic picture

Swelling becomes firmer, the skin may turn cool and mottled or bluish, nails thicken and become brittle, hair growth slows, joints stiffen, and muscle around the joint weakens. Bone density on the affected side may fall. Usually described as roughly months three to nine.

Stage 3 — the atrophic picture

The limb is cool, pale or glossy, the skin and subcutaneous tissue thin, contractures limit motion, and function is lost. Pain may spread beyond the original territory. Usually described as beyond a year.

Why the three stages are not a sequence

Two studies are usually cited against the staging model, and both used large patient samples.

A prospective Dutch study of 829 patients with reflex sympathetic dystrophy examined signs and symptoms directly and reported that it found no evidence consistent with three consecutive phases of the disease. It also found that the early picture is dominated by regional inflammation rather than by sympathetic overactivity — hyperhidrosis, the classic “sympathetic” sign, was infrequent enough to have no diagnostic value, while pain was present in 93 percent, hyperpathy in 75 percent, and reduced sensation in 69 percent.¹

A second study took 113 patients meeting the older International Association for the Study of Pain criteria and let a clustering algorithm sort them by their actual pattern of signs and symptoms across four domains: pain and sensory abnormality, vasomotor change, edema and sweating, and motor and trophic change. Three statistically distinct groups emerged — but they did not differ meaningfully in how long the patients had been in pain, which is exactly what a staging model predicts they should. The group that looked most like “classic RSD,” with the heaviest motor and trophic findings, had directionally the shortest pain duration of the three.²

So the clusters are real. The ordering is not. A person can arrive with a cold, stiff, atrophic-looking limb at four months, or with a hot swollen limb at three years.

What clinicians use instead

Warm CRPS and cold CRPS

The distinction that has held up better is temperature. Reviews of the mechanisms describe a shift from “warm” CRPS, in which peripheral inflammation predominates, toward “cold” CRPS, in which central sensitization, altered sympathetic and catecholaminergic signaling, and reorganized somatosensory maps in the brain matter more. The relative contribution of these mechanisms differs between patients and within the same patient over time.³ This is a description of which mechanism is currently loudest, not a stage a patient graduates from.

The Budapest criteria

CRPS has no confirmatory blood test or scan. It is diagnosed clinically, using the Budapest criteria: continuing pain disproportionate to the inciting event, at least one reported symptom in three of the four domains (sensory, vasomotor, sudomotor/edema, motor/trophic), at least one sign observed at the time of examination in two of those domains, and no other diagnosis that better explains the findings.⁴

The validation study compared these criteria against the older IASP criteria in 113 CRPS-I patients and 47 patients with other neuropathic pain. The old criteria were almost perfectly sensitive (1.00) but specific only 0.41 — they labeled a large share of non-CRPS patients as CRPS. The Budapest clinical criteria kept sensitivity at 0.99 and raised specificity to 0.68; the stricter research version reached 0.79.⁴ A later systematic review of every screening and diagnostic tool proposed for CRPS concluded that for adults the Budapest criteria are the preferred tool, and that no validated screening instrument exists at all.⁵ A fuller walkthrough is on our Budapest criteria page.

Specificity matters here because the old criteria inflated the apparent size of the problem. When one prospective study screened 306 consecutive patients with foot and ankle fractures using the Budapest criteria, 110 reported at least one CRPS symptom, but only one met the full symptom-and-sign combination — an incidence of 0.3 percent.⁶

CRPS type 1 and type 2 — the distinction that does exist

Type 1 (historically RSD) follows an injury with no identifiable nerve lesion. Type 2 (historically causalgia) follows a documented nerve injury. The presentation and the treatment approach are largely the same; the difference is whether a discrete nerve lesion can be demonstrated, which is where electrodiagnostic testing is useful. In the clustering study above, the subgroup dominated by neuropathic pain and sensory abnormality was the one whose EMG and nerve conduction findings suggested type 2.² Type is not a severity grade; it is a statement about whether a nerve lesion was found.

What the course actually looks like

A systematic review of 18 studies covering 3,991 patients found the honest answer is split. The three prospective studies showed that for many patients symptoms improve markedly within six to thirteen months of onset. The retrospective studies were far more heterogeneous and documented lasting impairment in many patients. Across studies, sweating and vasomotor findings were the most likely to resolve, while motor findings — stiffness and weakness — were the most likely to persist. The review’s authors noted that little is known about which patients fall into which group, and rated the overall quality of the evidence as poor.⁷

For scale: a population-based study drawing on 600,000 Dutch primary-care records estimated an overall incidence of 26.2 CRPS cases per 100,000 person-years (95% CI 23.0–29.7), with women affected at least three times as often as men, peak incidence in women aged 61–70, upper limb more often than lower, and fracture as the precipitating event in 44 percent of cases. Read that number with care: its authors point out it is more than four times higher than the only other population-based estimate then available, from Olmsted County in the United States — a gap they report rather than resolve, and a reminder that how often CRPS is counted depends heavily on how it is defined.⁸

Does the stage change the treatment?

Not in the way the model implies. There is no evidence-based protocol keyed to a stage number. Current practice guidelines are built around restoring function early — active movement, desensitization, and graded loading — with interventional and pharmacologic measures used to make that movement tolerable rather than as ends in themselves. The 5th edition guidelines are candid that most of the underlying research is low to medium quality with small samples, in part because CRPS is uncommon and because the diagnostic category is broad enough that a treatment effective in one subset can be diluted to non-significance across a whole trial.⁹

That caveat applies to specific treatments too. A double-blind randomized trial of a 4.2-day intravenous ketamine infusion in 60 CRPS-1 patients with long-standing severe pain found significantly lower pain scores than placebo across a 12-week window, with the largest separation at week 1 and no statistically significant difference remaining by week 12 — and no measurable functional improvement. Psychomimetic side effects occurred in 76 percent of the ketamine group versus 18 percent on placebo.¹⁰ That is a real effect worth knowing about and also a limited one; we describe the evidence by route on our ketamine therapy page.

The practical implication of dropping the staging model is this: nothing about a cool, stiff limb means the window has closed, and nothing about a warm swollen limb means it will inevitably progress. What guides treatment is the individual’s findings on examination, not a stage number.

If you are in the St. Louis area

Padda Institute evaluates and treats CRPS and RSD at 4477 Woodson Rd, Suite 100, St. Louis, MO 63134, and at 12174 Natural Bridge Road, Suite 100, St. Louis, MO 63044. Dr. Gurpreet Singh Padda, MD, MBA, MHP performs the Budapest-criteria examination and builds a plan around restoring use of the limb. Details are on the CRPS treatment page, or call (314) 481-5000.

Frequently asked questions

How many stages of CRPS are there?

Three are traditionally described — acute, dystrophic, and atrophic — but studies that tested whether patients pass through them in order did not find a fixed sequence. The stages are best read as pictures a limb can present, not steps. Diagnosis rests on the examination instead, as described on our Budapest criteria page.

Is CRPS stage 3 permanent?

An atrophic-looking limb is not by itself proof of an irreversible condition, and it does not tell you how long the condition has been present. Motor findings such as stiffness and weakness are the ones most likely to persist, which is why treatment emphasizes movement early. Our page on whether CRPS can be cured covers what is and is not known about long-term outcomes.

What is the difference between CRPS type 1 and type 2?

Type 1 follows an injury without a demonstrable nerve lesion; type 2 follows a documented nerve injury. Both are diagnosed by the same clinical criteria. Whether a nerve lesion is present is assessed with nerve conduction and EMG studies — see what EMG and nerve conduction studies measure.

Is CRPS the same thing as RSD?

Largely, yes. Reflex sympathetic dystrophy was the older name for what is now called CRPS type 1, and much of the foundational research — including the 829-patient study that undercut the staging model — was published under the RSD label. Many clinicians, including ours, still use both terms. Our CRPS and RSD page uses them interchangeably.

Do sympathetic nerve blocks work for CRPS?

They help some patients and not others, and the response is used diagnostically as well as therapeutically. For an arm, that is usually a stellate ganglion block; for a leg, a lumbar sympathetic block. Whether and how often to repeat one depends on what the first one showed, not on a fixed schedule.

Does waiting make CRPS harder to treat?

The evidence on timing is weaker than the confident advice you will read online, but the findings most likely to persist are the motor ones — stiffness, weakness, and lost range — and those are the ones that respond to early, sustained use of the limb. We discuss the practical consequences on our page about the cost of delayed CRPS treatment.

Sources

1. Veldman PH, Reynen HM, Arntz IE, Goris RJ. Signs and symptoms of reflex sympathetic dystrophy: prospective study of 829 patients. Lancet. 1993. PMID 8105263
2. Bruehl S, Harden RN, Galer BS, Saltz S, Backonja M, Stanton-Hicks M. Complex regional pain syndrome: are there distinct subtypes and sequential stages of the syndrome? Pain. 2002. PMID 11790474
3. Bruehl S. An update on the pathophysiology of complex regional pain syndrome. Anesthesiology. 2010. PMID 20693883
4. Harden RN, Bruehl S, Perez RSGM, et al. Validation of proposed diagnostic criteria (the “Budapest Criteria”) for Complex Regional Pain Syndrome. Pain. 2010. PMID 20493633
5. Mesaroli G, Hundert A, Birnie KA, Campbell F, Stinson J. Screening and diagnostic tools for complex regional pain syndrome: a systematic review. Pain. 2021. PMID 33230004
6. Bullen M, Lang C, Tran P. Incidence of Complex Regional Pain Syndrome I following foot and ankle fractures using the Budapest criteria. Pain Medicine. 2016. PMID 28025369
7. Bean DJ, Johnson MH, Kydd RR. The outcome of complex regional pain syndrome type 1: a systematic review. The Journal of Pain. 2014. PMID 24530407
8. de Mos M, de Bruijn AGJ, Huygen FJPM, Dieleman JP, Stricker BHCh, Sturkenboom MCJM. The incidence of complex regional pain syndrome: a population-based study. Pain. 2007. PMID 17084977
9. Harden RN, McCabe CS, Goebel A, et al. Complex Regional Pain Syndrome: Practical Diagnostic and Treatment Guidelines, 5th Edition. Pain Medicine. 2022. PMID 35687369
10. Sigtermans MJ, van Hilten JJ, Bauer MCR, et al. Ketamine produces effective and long-term pain relief in patients with Complex Regional Pain Syndrome Type 1. Pain. 2009. PMID 19604642

Dr. Gurpreet Singh Padda, MD, MBA, MHP