Double Crush Syndrome in the Lower Limb

What Works

Graded honestly, and the top of the scale is empty.

There is no Tier 1 evidence for this condition, in any limb. No randomised trial of double crush syndrome has ever been run, and there is no clinical practice guideline, no Cochrane review, no systematic review and no meta-analysis. Verified directly against PubMed rather than inferred: a title search returns 110 records while every evidence-tier filter returns zero, and three of the four records tagged as randomised trials are coronary artery stenting studies that share the word "crush".

1. Test hip abductor power MODERATE

The highest-yield single measurement here, and the one most often skipped, because a foot problem does not prompt anyone to test the hip. Gluteus medius is supplied by L5 above the point where the peroneal nerve branches off, so it is weak in a back-origin problem and spared in a knee-origin one.Sn: 85.7% | Sp: 96.4%

Used inversely here: normal hip abductor power in a patient already carried as an L5 problem is the cue to go looking further down the leg.

Evidence: MODERATE. Prospective diagnostic study, 61 consecutive patients with foot drop. Hip abductor weakness in 85.7% of lumbar radiculopathy against 3.6% of peroneal neuropathy; PPV 94.7%, NPV 90%. Single study, soft reference standard, not externally validated (Jeon 2013).

2. Nerve conduction studies and needle EMG MODERATE

The only test that can actually demonstrate two lesion sites on one nerve, and the established method for excluding the entrapment neuropathies and whole-nerve diseases that mimic radiating leg pain. Nothing about the treatment target should be finalised before this.

Evidence: MODERATE for radiculopathy (modest sensitivity, high specificity). Accuracy for the two-site question specifically: NO EVIDENCE

3. Decompression of the lower lesion where one is found LOW–MODERATE

Reported success across three uncontrolled series: 93% with pain improved in every patient and dorsiflexion in 78% (n=14); improvement across dorsiflexion, big toe extension and eversion in all 10 patients of a second series, with numbness falling from 90% to 40%; and 88% weakness improvement even with the back lesion left untreated.

Evidence: MODERATE for direction, LOW for magnitude. No randomisation, no blinding, follow-up as short as 6.8 months.

Tier 3 and the options that sound stronger than they are

4. Treating both sites, together or in stages NO EVIDENCE

Recommended by one 32-patient upper-limb study on the basis of improved conduction velocities and disability scores, and a second upper-limb series found simultaneous surgery shortened operating time and hospital stay compared with staging it. Neither study compared treating both sites against treating one, so neither can support the recommendation it makes. One of the two reports 79 patients in its abstract while its groups sum to 68.

5. Rehabilitation borrowed from the responsible lesion EXTRAPOLATION

Once the responsible lesion is identified, use the established protocol for that lesion. This is borrowed, it is labelled as borrowed, and the patient should be told it is borrowed.

Exercise Prescription

There is no exercise program for this condition, and writing one would be inventing it. No study has ever tested exercise, stretching, nerve gliding, orthotics or manual therapy in patients with two compression points on one nerve, in any limb. Filling this slot with a plausible-looking protocol would be the single worst thing this page could do, because it is the section most likely to be copied into a patient email.

What replaces it: the rehabilitation belongs to whichever lesion is actually responsible, and those protocols are well established. The job of this page is to make sure the right one gets chosen. In the meantime, four things are genuinely supported.

Take the pressure off the outside of the knee. Stop sitting with legs crossed; stop resting that side of the knee against a hard edge, a chair leg or a car door. The nerve sits just under the skin there.
Outline the numb area with a pen and photograph it. The exact border is one of the most useful pieces of information for working out which nerve is involved, and it is almost never recorded.
Write down whether the foot symptoms started before, with, or after the back problem. In the largest dataset 60.6% of the leg lesions were diagnosed before the spine surgery, so the ordering is genuinely diagnostic and almost nobody is asked.
Keep moving within comfort. There is no evidence that rest helps this, and good evidence that rest harms recovery in the individual conditions it is made of.

Dark cinematic anatomical study of nerve decompression at the fibular head

What Doesn't Work

Using the label to explain non-recovery. "It must be a double crush" absorbs any unexplained failure without requiring evidence that two lesions exist, and it displaces the differential that actually needs excluding. A review in JAAOS puts it bluntly: the diagnosis may be an illness construction that does more harm than good by offering a physical explanation for symptoms that may be driven by something else entirely.
Treating the spine because the patient has leg pain. Back pain was present in only 49% of patients whose leg pain turned out to be a trapped nerve further down, and 36 patients in one 4,000-patient referral series had that as the sole cause. This is the expensive error on record, and it runs in the opposite direction to the one people expect.
Waiting for typical foot symptoms before examining the foot. Only a third of confirmed second lesions produced typical symptoms. A symptom-triggered search finds a minority of them.
Screening every back patient for a second lesion. At a base rate near 5% nothing supports indiscriminate screening, and the patients most likely to acquire a second label are the ones in whom a second structural label is most likely to be wrong.

Red Flags

Progressive weakness is time-critical. None of the assessment below substitutes for having these looked at.

Any change in bladder or bowel control, or numbness in the saddle area. Possible cauda equina. Emergency, same day.
Symptoms starting in the other leg. Bilateral involvement changes the differential entirely.
Weakness that is getting worse, or a foot you cannot lift (below MRC grade 3). May indicate an expanding lesion or a root needing urgent decompression. Time-critical for recovery.
A lump you can feel at the outside of the knee or behind the ankle bone. Cyst, ganglion or tumour on the nerve. Surgical, not conservative.
Numbness in a symmetrical stocking pattern across both feet. Points to a whole-nerve disease rather than two pinch points; needs blood tests, not decompression.
Painless, progressively worsening foot drop. The spine gets ruled out by default here and should not be. Central causes also sit in this differential.
Night pain, unexplained weight loss, fever, or a cancer history. Malignancy and infection belong in this differential.

Refer to neurology or clinical neurophysiology first, because the two-site question cannot be answered clinically. Refer to spinal surgery where a structural lesion is shown or a deficit is progressing. Go to A&E for cauda equina signs.

Dark cinematic study of the lower limb nerve pathway

Return to Training

Governed by the lesion actually being treated, not by this label. These are the gates that apply regardless.

The responsible lesion has been identified, or a clear recorded decision exists about which one is being treated and why
Hip abduction, dorsiflexion, big toe extension and eversion all at full strength, tested and documented separately
No foot drop or slapping gait at walking pace and at speed
Full foot and ankle control through the intended movement, unloaded first, then at half your previous working weight
No increase in numbness or sensory change in the 24 hours after a full session
Symptom-free, or at a stable documented baseline, for two consecutive weeks
Red flags absent at reassessment, specifically no progressive deficit and no palpable lump

Regression trigger: any increase in numbness, any new or worsening weakness, or symptoms spreading to new territory. These are not the acceptable soreness of a tendon problem. A nerve getting worse under load is a reason to stop, not to push through.

Lie on your side and lift your top leg straight up while someone presses down on it. Compare both sides.

If that hip muscle is weak on the bad side, your foot problem is coming from your back. If it is strong and your foot is still weak, the problem is further down your leg, near the outside of your knee. That is the part almost nobody checks, and it is usually the more fixable of the two.

Do not test, do not wait: if your foot is getting weaker, if you are catching your toe when you walk, or if there is any change in bladder or bowel control, book an urgent appointment today.

The Verdict

Treating your back may not fix your foot. Weakness that keeps getting worse needs a doctor now.

Think of the nerve as a long extension cable running from your lower back to your foot, threaded through several tight doorways on the way. Everyone checks the first doorway, the one in your spine, because that is where the pain usually starts. But the cable also squeezes past the outside of your knee and behind your ankle bone, and a pinch at one of those later doorways produces exactly the same dead foot as a pinch at the first one. Fixing the door in your spine does nothing at all if the cable is trapped at the knee.

What this actually is: two separate squeeze points on one nerve, one in your lower back and one further down the leg, where the second one is usually the one nobody looked for.
What most people get wrong: the assumption is that the back problem caused the leg problem, and in people that has never been shown; what has been shown is that both turn up in the same patients, and the second one is often the more fixable.
What to watch for: a foot that is getting weaker, catching your toe when you walk, or numbness spreading to the other leg, because those need assessing quickly rather than exercising.

Best for: people with a confirmed nerve root problem in the lower back whose leg or foot symptoms have stalled, persisted, or come back after the back was treated. Skip if: your symptoms are steadily improving on your current treatment, or you have not yet had a nerve root problem confirmed. This page is about the second lesion, not the first.

Want the full evidence? Keep scrolling⌄

Conviction

LOW overall, and a single score would bury the half of this that is worth acting on. Thirteen claims were scored separately. The two that should change what a clinician does on Monday are both MODERATE; the mechanism the condition is named after is not supported in humans at all.

Second lesions co-occur with back nerve problems more often than chance	MODERATE
Those second lesions are usually atypical or silent, so symptoms are a poor trigger to look	MODERATE
Hip abductor power separates a back-origin from a knee-origin foot drop	MODERATE
A trapped nerve lower down can be the sole cause of apparent sciatica	MODERATE
Treating the lower lesion works even with the back lesion untreated	LOW–MODERATE
Two-site patients do worse overall than single-lesion patients	MODERATE, upper limb only, confounded by case mix
The back lesion CAUSES the leg lesion (the actual hypothesis)	NOT SUPPORTED in humans
A missed back lesion is why the lower surgery underperforms	NOT SUPPORTED
Treating both sites beats treating one	NO EVIDENCE
Any exercise, manual therapy or conservative treatment	NO EVIDENCE
How common this is outside surgical and nerve-testing clinics	NO EVIDENCE
Accuracy of any test for confirming BOTH lesions	NO EVIDENCE

LOW and NO EVIDENCE are different, and the difference is load-bearing. LOW means somebody looked and found little. NO EVIDENCE means nobody looked.

What would change my mind on the mechanism

A prospective cohort of at least 1,000 consecutive patients with confirmed lumbosacral radiculopathy, screened at the fibular head and tarsal tunnel at baseline, followed 24 months with the second lesion's incidence date recorded relative to the first, analysed against controls matched on healthcare encounter frequency as well as age, sex and comorbidity. Encounter-frequency matching is the specific missing control: every existing dataset confounds having more nerve lesions with having more nerve appointments.

What would change my mind on treating both sites

A randomised trial of decompressing both sites versus the lower site alone, in confirmed two-site lower-limb cases, N at least 120, primary endpoints dorsiflexion strength and pain at 12 months. Nothing in the current literature can answer this, and the two studies that recommend treating both never compared it against treating one.

The Full Picture — Anatomy, Diagnosis & Evidence

What's Actually Going On

In 1973 Upton and McComas proposed that a squeezed nerve root starves the rest of that nerve of the material it needs to stay healthy, so a second squeeze further down, one that would otherwise cause no trouble, becomes symptomatic. The claim is specific and directional: the first lesion has to sit on the same nerve fibres as the second.

In animals it holds. In people it does not. A canine study found that two compressions produced a deficit larger than the sum of the two separate ones, and that releasing only one of them gave incomplete recovery. That is the strongest mechanistic evidence in existence, and it is a dog.

Three human findings point the other way. First, when the hypothesis's own two requirements were applied to 12,736 limbs, that the root must anatomically supply the nerve below and the lower lesion must show actual nerve-fibre loss, only 69 limbs (0.5%) qualified. Second, the one study that tested the idea electrically exploited a neat dissociation, that one median nerve response comes from C6 and C7 and another from C8, so a genuine double crush makes a specific falsifiable prediction. Every prediction failed, and the authors reported no evidence for a neurophysiological explanation. That same study found the co-occurrence in 22.1% of cases, and offers it as the reason clinicians believe the idea anyway.

Third, and cleanest: the same association turns up between nerves that cannot possibly share an axon. Meralgia paresthetica, a nerve to the outer thigh, is associated with carpal tunnel syndrome at an odds ratio of 7.7 across 173,375 patient-years of general practice records. An association that survives after you remove the mechanism's one anatomical requirement is not evidence for that mechanism.

What the field's own experts now think: a Delphi study of 17 international nerve specialists produced 14 candidate mechanisms, nine of which had never previously been linked to double crush, with impaired transport along the nerve only one of four rated highly plausible. The single-mechanism story in the name is a 1973 hypothesis the field has already moved past.

Dark cinematic anatomical rendering of a peripheral nerve pathway

How to Identify It

The condition only applies once a nerve root problem is established. The question this section answers is whether there is a second lesion, and the honest answer is that you cannot settle it clinically.

Hip abductor powerSn: 85.7% | Sp: 96.4% Weak means the problem is above the knee; strong means below it. PPV 94.7%, NPV 90%.
Nerve conduction studies and needle EMGSn: modest | Sp: high For radiculopathy. The only test that can show two sites. Accuracy for the two-site question itself is unmeasured.
Tinel sign at the fibular head or behind the ankle boneno controlled data Positive in 15 of 27 confirmed concurrent cases and 6 of 10 in another series. Those are hit rates, not a sensitivity, because there was no control group.
Ultrasound at the lower siteno accuracy data Finds a cyst, ganglion or tumour. Changed management in reported cases.

The finding that most changes practice: in 561 patients with confirmed lumbosacral radiculopathy, 27 (4.8%) had a concurrent tarsal tunnel syndrome, and only 33.3% of them had typical symptoms. Another 51.9% had merely suspicious ones. Waiting for a classic presentation misses two thirds of these.

A second finding that quietly undermines the mechanism: concurrent tarsal tunnel frequency did not differ across single-level L4, L5 or S1 radiculopathy, or between single and multi-level. If the proposed mechanism worked, the lower lesion should track the root that actually supplies that nerve. It does not.

Dark cinematic study of clinical lower limb assessment anatomy

The Debate

Association versus mechanism

The largest supporting study queried a database of 90.7 million patient records and found back nerve problems raise the odds of a peripheral nerve problem, and the reverse, both highly significant, and reported this as support for the hypothesis. The objection: a hypothesis whose whole content is directional cannot be confirmed by a symmetric association. A bidirectional odds ratio is exactly what a shared cause produces.

What the risk factors are actually describing

In the largest lower-limb dataset (650,562 surgical patients, 5.1% with a second lesion, odds ratio 3.10 against matched controls), the strongest predictors were complex regional pain syndrome (3.33), fibromyalgia (1.73), osteoarthritis (1.61) and high overall comorbidity (1.50). That is a description of a patient who accumulates diagnoses and appointments, not a description of nerve biology. The matching controlled for demographics and comorbidity but not for how often someone is examined.

Do two-site patients do worse? Two answers, and they are answering different questions

A matched case-control of 80 patients found the two-site group significantly more disabled at 18 months (QuickDASH 29 versus 13). But it compared people who had both operations against people who needed one, so the comparison contains case severity. The only lower-limb study that compared what a clinician actually faces, lower-site surgery with an untreated back lesion versus without one, found no significant difference (88% versus 100% weakness improvement, p=0.62). Both results stand. Set expectations down; do not conclude a second operation is therefore indicated.

Honest Limitations

The whole literature is surgical, and the question usually is not

Every management study here reports operative outcomes from neurosurgical, orthopaedic or podiatric services. Zero conservative or exercise interventions have ever been tested. The decision most clinicians actually face, whether to send someone for nerve testing, is addressed by no study at all. Anything on this page about rehabilitation is extrapolation and is labelled as such.

The prevalence figures come from enriched populations

5.1% among patients who reached lumbar surgery. 4.8% among patients with electrically confirmed radiculopathy. 5.3% among mostly post-traumatic back pain referred for nerve testing. The convergence near 5% across three independent datasets is the striking part, but all three are populations that had already reached a surgeon or a nerve lab. The rate in a community clinic is unknown. Treat 1 in 20 as a ceiling, not a base rate.

Nobody has separated more nerve lesions from more nerve appointments

The headline odds ratio of 3.10 comes from propensity matching on demographics and comorbidity, not on how often a patient is seen. Someone inside a spine surgical pathway is examined by more clinicians, more often, with more nerve testing, than a matched control outside it. Part of that odds ratio is a measure of who gets looked at. It should not be quoted to a patient as their risk of developing a second problem.

The Nuance

The premise that commissioned this page turned out to be backwards, and that is the most useful thing on it. The expectation was the patient whose lower-limb decompression underperforms because the back lesion was never addressed. Every study that can speak to it runs the other way.

In the only controlled lower-limb comparison, an untreated back lesion did not degrade the outcome of surgery further down (88% versus 100%, p=0.62). In a 14-patient series, half the patients arrived with symptoms that had persisted or returned after their back was treated, and were resolved by decompressing the lower site. In a second series, 100 patients were screened and only 10 had genuine two-site disease; all 10 improved after lower-site surgery, and only 3 had ever had back surgery.

And the most extreme version: of roughly 4,000 patients referred with suspected lumbar nerve root pain, 36 had a trapped nerve further down the leg as the sole cause, with back pain present in only 49% of them. Those authors concluded that peripheral lesions should be excluded before anyone considers spinal surgery.

Surgery versus conservative care: conservative success rate is NO EVIDENCE, because no study has ever reported it. Surgical series report 88% to 100% improvement, all uncontrolled. Surgery is indicated where a structural lesion is demonstrated, where a deficit is progressing, or where symptoms persist after the back lesion has been properly treated. The honest position on treating both sites is that the second operation has no comparative evidence behind it, not that it is wrong.

Dark cinematic anatomical study of the lower limb differential

Treated for the wrong lesion once already?

Every week, one condition taken apart to the level of what the evidence actually supports, including where it supports nothing.

Join The Verdict, free

Sources

Morgan G, Wilbourn AJ (1998). Neurology. PMID 9443461. 12,736 limbs; only 0.5% satisfied the hypothesis's own anatomic and pathophysiologic requirements.
Richardson JK, Forman GM, Riley B (1999). Muscle & Nerve. PMID 9883859. Direct electrophysiological test; every prediction failed. Co-occurrence 22.1%.
Karakash WJ, Avetisian H, Gallo MC, et al. (2026). Spine Surg Relat Res. PMID 41695903. N=650,562; 5.1% with a peripheral nerve lesion, OR 3.10 versus matched controls.
Zheng C, Zhu Y, Jiang J, et al. (2016). Eur Spine J. PMID 26407567. N=561; concurrent tarsal tunnel 4.8%, typical symptoms in only 33.3%.
Jeon CH, Chung NS, Lee YS, et al. (2013). Spine. PMID 22805342. N=61; hip abductor power Sn 85.7%, Sp 96.4%, PPV 94.7%, NPV 90%.
van Slobbe AM, Bohnen AM, Bernsen RM, et al. (2004). J Neurol. PMID 15015008. 173,375 patient-years; meralgia paresthetica with carpal tunnel OR 7.7 (95% CI 1.9–31.1).
Ishii M, Nishimura Y, Hara M, et al. (2023). Neurol Med Chir. PMID 37286483. The only controlled lower-limb comparison; 88% versus 100%, p=0.62.
den Boogert HF, Schuuring J, de Ruiter GCW (2025). J Clin Med. PMID 40725716. N=14; 93% success, 7 of 14 presented after the back had been treated.
Saal JA, Dillingham MF, Gamburd RS, et al. (1988). Spine. PMID 2847334. 36 of ~4,000 referrals had peripheral entrapment as the sole cause; back pain in 49%.
Santangelo G, Singh A, Catanzaro M, et al. (2025). World Neurosurg. PMID 39586462. 100 screened, 10 with genuine two-site disease, all improved after distal surgery.
Mills ES, Mertz K, Fresquez Z, et al. (2024). Global Spine J. PMID 36321208. 90.7 million records; association bidirectional and reported as supportive.
Wessel LE, Fufa DT, Canham RB, et al. (2017). Plast Reconstr Surg. PMID 27627055. Matched case-control, N=80; QuickDASH 29 versus 13. Upper limb.
Schmid AB, Coppieters MW (2011). Manual Therapy. PMID 21646036. Delphi, 17 experts; 14 mechanisms, 9 never previously linked to double crush.
Nemoto K, Matsumoto N, Tazaki K, et al. (1987). J Hand Surg. PMID 3611653. Canine model. Preclinical support for the mechanism, not clinical efficacy.
Upton AR, McComas AJ (1973). Lancet. PMID 4124532. Origin of the hypothesis.