Complex Regional Pain Syndrome (CRPS)

The VerdictMODERATE CONVICTION

A limb hurting far more than the injury should, changing color and swelling?

If you have just broken your wrist, ask your doctor about vitamin C, 500 mg a day for about 7 weeks (50 days), starting now. In studies it roughly halved the chance of developing this pain problem. It only works as prevention. It does nothing once CRPS has started.

What this actually is: a pain response that goes into overdrive after an injury like a broken wrist, so the limb hurts, swells and changes color far more than it should.
The one thing that makes it worse: resting and protecting the limb. Immobilizing it feeds the problem.
Start here: if you already have it, keep the limb gently moving and retrain it with mirror therapy, and get a pain team involved early.

After an injury, the alarm system in that limb gets stuck on. The nerves and the brain's map of the limb become over-sensitive, so ordinary touch and movement fire the alarm, and the limb changes temperature, color and swelling. Resting it lets the brain's map of the limb fade further, which turns the alarm up, not down. Gently using the limb and retraining that map is how you turn the alarm back down.

What Works

Cinematic anatomy of a limb in rehabilitation

The honest headline: the strongest single action here is prevention, not treatment. And every number below comes from a field of 40 reviews built on small, old trials, so treat them as direction, not promise.

Tier 1 — Strongest available evidence MODERATE

Vitamin C after a wrist fracture (PREVENTION).
500 mg/day · for 50 days · starting the day of injury
Roughly halves the risk of developing CRPS (RR 0.54; OR 0.33). It does not treat CRPS once it exists.

Graded motor imagery / motor imagery for established CRPS.
3 stages: left-right recognition → imagined movement → mirror therapy · daily · slow progression
Improves pain and disability (moderate-quality evidence). Exact weeks and reps aren't well defined, so follow your therapist's pace.

Exercise Prescription

Left / right recognition: quickly decide left or right from pictures of hands or feet. 10–15 min · several short sessions daily · no pain flare

Imagined movement: picture the sore limb moving normally without moving it. 5–10 min · daily · mental practice, no pain

Mirror therapy: reflect the good limb over the sore one, move the good limb, watch the reflection. 5–10 min · daily · stop if it strongly flares

Graded normal use + desensitization: reintroduce everyday tasks and textures with the limb. little and often · daily · mild discomfort OK, sharp escalation not

See Tier 2 and the specialist tier

Tier 2 LOW–MODERATE — Mirror therapy and graded/aerobic activity added to conventional rehab (especially the poststroke shoulder and hand). Keep the limb in graded normal use to avoid the disuse spiral.

Specialist tier (a pain team decides, not a physical therapist) — bisphosphonates and short courses of oral steroids carry the best drug-tier signal. Spinal cord stimulation helps a small refractory minority. IV ketamine gives short-term relief that fades. These are for refractory cases and are referred onward.

What Doesn't Work

Resting / immobilizing the limb. The single most common and most damaging mistake.
Vitamin C as a treatment for CRPS you already have. It is prevention only.
Treating CRPS as "in your head." The low mood and anxiety are consequences, not the cause.
Trusting a normal bone scan to rule it out. A negative scan does not exclude CRPS.

Return to Training

CRPS confirmed and a multidisciplinary pain plan in place
Color, temperature and swelling settling, and no longer flaring with normal use
Full or near-full pain-free active range restored in the limb
Can tolerate graded loading of the limb without a 24-hour symptom flare
Stable or improving function on a tracked measure across consecutive reviews

Red Flags — When to Refer

Cinematic anatomy of an inflamed distal limb

Any suspected CRPS → refer early to a multidisciplinary pain team. Early recognition is the single biggest lever on the outcome.
Hot, red, spreading limb with fever or feeling unwell → possible infection. Urgent medical review.
A leg that suddenly swells and hurts → rule out a blood clot (DVT) first. Urgent medical review.
A defined area of numbness/weakness in one nerve's territory → possible nerve injury. Refer for neurological assessment.
CRPS that keeps worsening despite rehab → back to the pain team for the medication/procedure tier.

Refer to: GP + multidisciplinary pain service for suspected CRPS; A&E / urgent medical review for suspected infection or clot; Neurology for a suspected nerve lesion.

MODERATE

The whole field is roughly 40 reviews built on small, old, mixed trials, with no large stand-alone randomized trial to anchor it. The direction of the evidence is consistent and the harm of the physical-therapy tools is near-zero, but the confidence is not high.

What would change this

A large (N≥200), multi-center, double-blind trial of graded motor imagery with a fully specified dose, in early CRPS, with a 12-month pain and function endpoint, showing a durable effect, would upgrade the rehab conviction to HIGH and finally give the field a dosing schedule.

Go Deeper

Don't want to guess what actually works the next time an injury turns into something stranger? Join The Verdict for free, evidence-scored protocols every week.

Get free weekly protocols

The Full Picture — Anatomy, Diagnosis & Evidence

What's Actually Going On

Cinematic anatomy of nerves and inflammation in a limb

CRPS is a disproportionate regional pain response that usually follows a clear event: a fracture (classically the wrist), a crush, a sprain, surgery, or a stroke. It is not one lesion. At least four systems act at once: inflammation in the limb and the nervous system, a link carrying that inflammation into the spinal cord and brain, a distortion of the brain's sensory map of the limb, and autonomic changes that alter temperature, color, sweating and swelling.

That distorted brain map is the key to treatment. When the brain's picture of the limb blurs, the pain system over-reacts to ordinary signals. Graded motor imagery and mirror therapy retrain that map, which is why they help where stretching and resting don't.

CRPS type I has no confirmed nerve injury (the old name was reflex sympathetic dystrophy). Type II has a confirmed nerve lesion (old name causalgia).

How to Identify It

Cinematic anatomy of a clinician examining a limb

There is no single test that confirms CRPS. It is a clinical-pattern diagnosis (the Budapest criteria): continuing pain out of proportion to the injury, PLUS signs across sensory, temperature/color, swelling/sweating, and movement/skin changes, with no better explanation.

Budapest clinical criteria — the reference standard no gold-standard test
Three-phase bone scintigraphy — specialist imaging that supports the diagnosis Sn 0.55 · Sp 0.94 vs Budapest
A negative bone scan does NOT rule CRPS out — go on the clinical picture, not the scan

What it is not: plain post-fracture stiffness lacks the temperature/color/swelling cluster; infection brings fever and spreading redness; a clot (DVT) swells a leg without the sensory and skin changes; a defined nerve-territory deficit points to a nerve injury instead.

Cinematic anatomy contrasting limb conditions

The Debate

Rest it vs move it

Older instinct: protect and rest the painful limb.

Recent evidence: keeping the limb in graded use, with graded motor imagery and mirror therapy, improves pain and disability (PMID 38265184, 36650605). Immobilizing it feeds the disuse and remapping spiral.

Follow: keep the limb in graded, pain-contingent use.

Is CRPS psychological?

Old framing: a psychogenic or "personality" condition.

Recent evidence: no support for a causal CRPS personality or pre-existing psychopathology; the depression and anxiety are consequences of a severe, poorly-understood pain condition (PMID 22961122).

Follow: treat the pain and the person; psychology is support, not the cause.

Honest Limitations

Rehab works in trials, but the trials are thin

The main rehab review rated its evidence low or very low certainty, with most trials small, short and at high risk of bias, and dosing rarely reported. Real-world delivery is improvised on clinical judgment.

Prevention gets mistaken for treatment

Vitamin C's evidence is entirely about preventing CRPS after a fracture. It does nothing once CRPS exists, yet it is the most-quoted "CRPS supplement." That conflation is the trap.

The diagnosis is moving under the treatment evidence

New chronic-stage criteria and objective tests are emerging, so cohorts diagnosed different ways aren't perfectly comparable, which is part of why the treatment literature is so noisy.

The Nuance

Cinematic anatomy representing a treatment decision pathway

There is no curative surgery for CRPS. For a refractory minority who fail rehab and the medication tier, a pain service may consider spinal cord stimulation, which reduced pain versus usual care in a small trial base (MD −1.17 points). Ketamine infusions can give short-term relief (immediate 30%+ relief in about 69% of patients) that does not last.

One methodological point sharpens all of this: in long-standing CRPS the placebo response is nearly absent (PMID 25478803). That means a genuine improvement is more believable than usual, but it also removes the easy excuse for weak trials. Most drug and procedure options still lack durable, blinded, adequately-powered support, which is why they stay specialist-tier and referred, not physical-therapy prescriptions.

Sources

Aïm F, et al., 2017 — meta-analysis, 3 RCTs, N=875: vitamin C 500 mg/day × 50 days after wrist fracture, CRPS-I RR 0.54 (95% CI 0.33–0.91). (PMID 28274883)
Chen S, et al., 2022 — SR & meta-analysis, 8 studies: perioperative vitamin C, CRPS-I OR 0.33; no effect on pain/function once present. (PMID 34961681)
Motor imagery SR & meta-analysis, 2024 — 6 RCTs: pain SMD −1.07 (95% CI −1.53 to −0.60), moderate-quality. (PMID 38265184)
Rehabilitation SR & meta-analysis — 33 RCTs, GRADE low: mirror therapy + graded motor imagery improve poststroke CRPS-1. (PMID 36650605)
Bone scintigraphy SR & Bayesian meta-analysis — vs Budapest criteria Sn 0.55 / Sp 0.94; a negative scan does not exclude CRPS. (PMID 28301606)
Duong S, et al., 2018 — updated SR, 35 RCTs: continued support for bisphosphonates and short oral steroid courses. (PMID 29492826)
Placebo-response SR & meta-analysis — near-absent placebo response in long-standing CRPS. (PMID 25478803)
Psychosocial-factors SR — CRPS distress is a consequence; no support for a causal CRPS personality. (PMID 22961122)