Saphenous Nerve Entrapment

Stand on one leg on the painful side for 10 seconds. Then push up onto that toe, once.

If you can do both, this nerve is not causing weakness — because it doesn't control a single muscle anywhere in your leg. That's the fork this whole condition turns on: burning and numbness with normal strength points at the nerve, and any genuine weakness points somewhere else entirely.

TAKES 15 SECONDS

Don't wait on these. Real weakness, the leg giving way, catching your toe, a growing lump in the inner thigh, or new electric-shock sensations in your calf after recent knee surgery all mean see a doctor now, not a protocol.

The Verdict

A nerve that only carries feeling gets squashed near your knee — so it burns, but nothing is weak.

Think of it as the doorbell wire, not the door hinge. The wire only carries a signal — it doesn't hold anything up and it doesn't move anything. When a surgeon's cut or a tight strap crushes that wire, the signal arrives scrambled: burning, tingling, or nothing at all. That's why the knee itself still works perfectly. And it's why the repair is slow — a cut wire regrows from its severed end far slower than a bruise fades, which is why the honest timeline here is months rather than weeks, and why for about a third of people the signal never fully comes back.

What this actually is: a skin-sensation nerve that runs across the inside of your knee, and the common knee operations cross straight over its path.
What most people get wrong: numbness after knee surgery gets treated as damage that needs fixing, when for most people it's expected, harmless, and no reason to hold back training.
Start here: check whether anything is actually weak — this nerve works no muscle at all, so real weakness means something else is going on.

Best for

Burning, numbness or tingling around the inside or front of the knee — especially after an ACL reconstruction, arthroscopy, knee replacement or vein surgery — where the knee itself feels mechanically fine.

Skip if

You have weakness, the leg gives way, you catch your toe, or there's a growing lump in your inner thigh. Those need a doctor, not a protocol.

Want the full evidence? Keep scrolling

What Works

Cinematic anatomical rendering of the medial knee and adductor canal region

Read the grading before the list. This condition's treatment hierarchy is unusual, and flattening it would be dishonest. The strongest evidence is preventive and belongs to the surgeon. The physical therapy tier is graded NO EVIDENCE — not because it fails, but because it has never been tested.

Tier 1 — Strong evidence

Explanation and mapping for an isolated numb patch HIGH

The intervention most people actually need, and the one most often skipped. Sensory loss follows 39.7% to 88% of hamstring graft harvests. Only 22.6% resolve, at a mean of 11.1 months, and 32.1% persist beyond two years.

Evidence: STRONG for the underlying facts — Hardy 2017 (36 articles), Egerci 2024 (n=159, 2-year follow-up). Timeline: immediate reassurance, honest 6–12 month horizon for any recovery that happens.

Oblique rather than vertical incision at surgery MODERATE

Risk ratio 0.64 (95% CI 0.49–0.84) across 8 randomised trials and 626 patients, with 29 cm² less area of sensory loss. This is a surgeon's decision made before a physical therapist meets the patient — its value here is pre-operative counselling, not treatment.

Evidence: STRONG design, MODERATE certainty — Yuwarungsikul 2026; the authors themselves rate certainty low on heterogeneity (I²=73.6%). And it reduces risk rather than removing it: the sartorial branch is injured as often as the infrapatellar, via tendon-stripper proximity the incision can't affect (Sanders 2007).

Tier 2 — Moderate evidence

Ultrasound-guided diagnostic anaesthetic block MODERATE

Diagnostic first, therapeutic second, and the gateway to every nerve-directed treatment. Responders carried an entrapment or neuroma on ultrasound in 97% of cases against 6% of non-responders.

Evidence: MODERATE — Yang 2022, retrospective diagnostic cohort, n=55.

Reassess for joint instability when nerve treatment fails MODERATE

A sequencing rule rather than a treatment. Non-responders to nerve-directed therapy had knee joint instability at 25% against 4% — a competing pain generator hiding behind a nerve diagnosis.

Evidence: MODERATE — Yang 2022. Directly actionable and cheap to apply.

Tier 3 — Emerging and case-level only

Ultrasound-guided hydrodissection LOW

Mechanistically coherent, direct evidence weak. For this nerve the evidence is single case reports (Watanabe 2020, Eliot Hu 2022). The only randomised trial is in a different nerve — dextrose beat corticosteroid at 4 and 6 months in meralgia paresthetica, with 0 adverse effects against 6 (Shi 2024, n=56). That is an analogy, not saphenous evidence.

Removing a sustained compression source LOW

Where one genuinely exists — equipment, seating, prolonged cross-legged sitting. Narrative and case-level only (Obana 2024, whose prevention plan is explicitly unvalidated).

Manual therapy and home exercise NO EVIDENCE

One case report, n=1. Settergren 2013 resolved a competitive ultra-marathon runner in two treatments — and stated in the same paper that no documented conservative interventions existed in the literature at all. Thirteen years on, that has changed only by the addition of further single cases.

Surgical neurolysis, neurectomy or translocation LOW

Effective in correctly selected refractory cases — complete relief in 4 knees after translocation (House 1977), 5 of 5 cured after failed conservative care (Lippitt 1993). Selection is the entire problem, and the diagnostic block is how selection is done.

Exercise Prescription

An honest note before the table. No exercise programme has ever been tested for this condition. What follows keeps the nerve moving and the surrounding muscles working. It is reasonable and low-risk, and nobody has proven it speeds anything up. Saying so is the point — a protocol presented with false confidence here would be inventing evidence that doesn't exist.

Exercise	How	Dose	Pain guide
Saphenous nerve glide	Lie on your back. Straighten the affected leg, let it roll slightly outward, then slowly point and flex the ankle with the knee straight. You're flossing the nerve, not stretching it	2 × 10 slow reps, twice daily	Mild pulling only. If burning or tingling rises and stays up, reduce the range
Adductor isometric squeeze	Sit or lie with knees bent and a pillow between them. Squeeze gently and hold	3 × 20 seconds, daily	Effort in the inner thigh, no sharp pain
Straight-leg raise	Lie on your back, other knee bent. Lift the straight leg to the height of the opposite knee, lower slowly	3 × 10, daily	Effort in the front of the thigh; stop if sharp
Sit-to-stand	From a normal chair, stand and sit under control, no hands if possible	3 × 10, daily	Some discomfort fine; should settle within an hour

The review point matters more than the exercises. If nothing has meaningfully changed by six weeks, that is the moment to refer for imaging and a diagnostic block — not the moment to add more exercises. Draw around the numb or painful area at the start so you can actually tell whether it's shrinking.

What doesn't work

Treating the numb patch. The commonest presentation is an expected post-surgical finding that needs explaining, not treating. Treating it turns a normal finding into a problem and starts a course of care with no endpoint.
Repeating a diagnostic block that didn't work. Non-response is information pointing at another pain generator, usually joint instability.
Modifying the ACL rehabilitation programme because of sensory loss. Nerve injury here isn't related to anterior knee pain or range-of-motion limitation.
Borrowing the surgical incidence figures for a case with no surgery. The meta-analyses say nothing about a runner with adductor-canal pain.
Corticosteroid as the default injectate where a choice exists — 6 of 28 patients reported an adverse effect against 0 of 28 on dextrose in the comparable-nerve trial.

Red Flags

Stop and refer — do not work through these

Dark cinematic anatomical rendering of the lower limb neurovascular structures

Any motor weakness, foot drop, or progressive neurological deficit. This nerve works no muscle at all, so weakness means a different and possibly more serious diagnosis — L3/L4 radiculopathy, femoral neuropathy or common peroneal palsy. Refer: GP or Neurology.
New electric-shock sensations in the inner calf on deep knee bend after a recent meniscal repair. Possible entrapment of a nerve branch by suture material — described with immediate resolution on release. Refer: the operating surgeon, promptly.
A palpable, enlarging or progressive lump in the inner thigh. Saphenous neuropathy has been caused by a large cyst within the adductor muscle. Refer: GP for urgent imaging.
Calf swelling, colour change, absent pulses, or a claudication pattern. This condition has been documented simulating a vascular disorder. Refer: A&E or same-day GP depending on severity.
Sensory loss spreading beyond the nerve's territory, or progressing rapidly. The diagnosis is wrong. Refer: GP or Neurology.
Systemic features — fever, weight loss, night pain unrelieved by position change. Refer: GP.

Refer to: GP for workup and imaging in most cases; Orthopaedics where the cause is surgical or a block is being considered; Neurology where a motor or spreading sensory deficit appears; A&E for suspected vascular emergency.

Return to Training

For an isolated numb patch the answer is simpler than the checklist below suggests: change nothing. There's no tissue-healing timeline and no structure at risk. The criteria below apply to genuine neuropathic pain.

No motor weakness anywhere in the limb, confirmed on testing — a safety re-check, since this should have been true throughout
Neuropathic symptoms (burning, electric, tingling) stable or reducing across 2 consecutive weeks
Mapped area of sensory disturbance stable or reducing — drawn and dated at baseline, not estimated from memory
Can complete the sport- or lift-specific movement that previously provoked symptoms without a lasting increase (settles within an hour, no next-day escalation)
For lifters: 3 sets of 10 at previous working weight with no symptom increase persisting past 24 hours
Compressive source identified and removed, or confirmed absent
Patient understands that a residual numb patch is not a reason to delay return to training

Conviction

MODERATE

Endpoint-stratified, because a single label would be misleading here. This topic is evidence-rich about a surgical complication and evidence-empty about a treatment, and those two halves must not borrow each other's confidence.

Iatrogenic injury is common after hamstring harvest HIGH
The nerve is purely sensory; weakness means another diagnosis HIGH
Most sensory deficit doesn't impair knee function MODERATE-HIGH
Oblique beats vertical incision MODERATE
About a third of deficits persist beyond 2 years MODERATE
A diagnostic block identifies who benefits from nerve-directed treatment MODERATE
Persistent injury associated with ACL re-rupture LOW
Hydrodissection for this nerve specifically LOW
Any physical therapy intervention for established entrapment NO EVIDENCE

What would change this: a randomised trial of structured conservative care against watchful waiting in block-confirmed saphenous entrapment. It has never been run in any form.

What would change my mind on "no physical therapy evidence"

Adults with medial knee or leg pain, no motor deficit, symptom relief on an ultrasound-guided diagnostic block as the entry criterion. N≥120 for a 20-point pain difference. Neurodynamic sliders plus adductor-canal soft-tissue work plus activity modification, against advice alone. 12 weeks, 6-month follow-up, proportion achieving ≥50% pain reduction. That single trial would move this row off NO EVIDENCE in either direction.

What would change my mind on the re-rupture association

A prospective cohort of ≥400 hamstring-graft reconstructions with sensory mapping and re-rupture as a pre-specified endpoint, adjusted for incision type, surgeon and return-to-sport testing. If real, it changes return-to-sport gating for a large population. If confounded, it should be retired from the literature before it propagates further. Right now it rests on one retrospective cohort with no proposed mechanism, which is why nothing on this page acts on it.

The Full Picture — Anatomy, Diagnosis & Evidence

What's Actually Going On

Cinematic anatomical rendering of the saphenous nerve course through the adductor canal

The saphenous nerve is the terminal sensory branch of the femoral nerve, and it is purely sensory along its entire course. That single anatomical fact does more diagnostic work than any test on this page.

It runs down the thigh inside the adductor canal beneath sartorius, then pierces a sheet of connective tissue roughly a hand's breadth above the inner knee bone. Then it divides:

The infrapatellar branch crosses the front of the knee sideways, supplying skin over the patellar tendon and the inner joint line. That transverse course is exactly why a vertical surgical cut severs it and an oblique or horizontal one can pass between branches — the entire prevention literature rests on this geometry.
The sartorial branch continues to the inner ankle and the arch of the foot. It is not the minor player the name order implies: when both branches are surveyed for, 74% of patients have some post-harvest sensory disturbance, with isolated sartorial injury (23%) slightly more common than isolated infrapatellar injury (19%). The nerve runs against the gracilis tendon for 4.6 cm in the lower thigh, so part of the damage comes from the tendon stripper rather than the skin cut (Sanders 2007).

Three mechanisms produce symptoms: iatrogenic transection or traction (by far the commonest), compressive entrapment at the adductor canal or behind the sartorius tendon, and secondary compression from an adjacent lesion — pes anserine bursitis has compressed this nerve and presented as a tibial stress fracture, resolving when the bursitis was treated (Hemler 1991).

How to Identify It

Dark cinematic rendering of medial knee anatomy for clinical assessment

No bedside test for this condition carries a published sensitivity and specificity anywhere in the 106 papers reviewed. That is stated rather than filled in, because converting the numbers that do exist into accuracy figures would mean inventing measurements no study made.

Tinel's sign over the adductor canal or medial knee Sn: DATA UNAVAILABLE · Sp: DATA UNAVAILABLE — the one paper linking a bedside sign to a reference standard found a positive Tinel predicted response to a diagnostic block at p=0.002, and Tinel plus absence of joint instability predicted response in 90% (26/29). Those are conditional probabilities, not accuracy figures (Yang 2022).
Sensory mapping against the nerve's territory Sn: DATA UNAVAILABLE · Sp: DATA UNAVAILABLE — light touch and pinprick, drawn and compared to the other limb and to the L3/L4 dermatomes. The distribution is the diagnosis.
Saphenous nerve tension test Sn: DATA UNAVAILABLE · Sp: DATA UNAVAILABLE — named in the surfing-injury literature with no published accuracy data of any kind (Obana 2024).
Hagert's tetrad (pain, weakness, scratch collapse test, orthogonal taping) Sn: DATA UNAVAILABLE · Sp: DATA UNAVAILABLE — proposed for dynamic compression after knee replacement, and its own authors state it needs validation (Ferembach 2025).
Ultrasound-guided diagnostic block — the reference standard, and not a bedside test. Responders carried an entrapment or neuroma in 97% of cases against 6% of non-responders (Yang 2022, n=55).

Expect normal range of motion and normal strength throughout. Any deficit reframes the whole presentation.

The Debate

No clinical practice guideline exists for this condition as of 2026-08-15. No NICE, APTA, BOA, EULAR or JOSPT guideline addresses it. That's an honest gap rather than a search failure — the high-grade evidence lives inside the ACL reconstruction surgical literature rather than under this condition's own name. So the disagreements below are between studies, not between a guideline and a study.

Which incision actually wins?

Papastergiou 2006 — n=230 knees

Horizontal incision: 14.9% injury against 39.7% for vertical, P<0.001. The strongest single-study result in the field.

VS

Yuwarungsikul 2026, Grassi 2018 — meta-analyses

Oblique recommended: RR 0.64 (95% CI 0.49–0.84) across 8 RCTs and 626 patients.

Resolved rather than picked: this is an evidence-availability artefact, not an anatomical finding. Oblique wins the meta-analyses because oblique is what most trials randomised. The defensible verdict is that anything other than vertical is better, and the oblique-versus-horizontal question is genuinely open.

Alarming injury rates, or a non-problem?

Hardy 2017, Pękala 2017, Egerci 2024

39.7–88% affected; 32.1% still affected beyond two years.

VS

Grassi 2018, Yuwarungsikul 2026, Du 2025

No difference in Lysholm, IKDC or pain scores between groups with very different injury rates.

Both are true, and the resolution is the most useful thing on this page: sensory loss is common and mostly not disabling. But the instruments reporting "no difference" are knee-function scores, and numbness is not knee function. Where satisfaction was measured directly, it did differ (p=0.02, Mahmood 2020). Don't mistake a null on the wrong instrument for a null.

Honest Limitations

The best evidence answers a question the therapist can't act on

The finding: every Level I result here concerns which skin incision the surgeon makes.

The gap: by the time the patient reaches physical therapy, the incision has been made. This evidence cannot be converted into a treatment.

The adjustment: its value moves upstream and downstream — pre-operative counselling and post-operative expectation setting. Nothing else.

The outcome instruments can't see the outcome

The finding: "no difference in functional outcomes between incision groups."

The gap: Lysholm and IKDC measure knee function. Sensory loss isn't knee function. These scores will read normal in a patient with a large numb area and an unchanged complaint.

The adjustment: measure the mapped sensory area and ask the satisfaction question directly. Neither has an established minimum clinically important difference, and saying so is better than borrowing one.

The conservative literature is an archive of successes

The finding: case reports describe conservative or surgical resolution.

The gap: case reports get published when the diagnosis was made and something worked. The denominator — how often conservative care fails — is unknown and unknowable from this literature.

The adjustment: never quote case-report success as a prognosis. Agree a review point rather than an open-ended course of treatment.

The Nuance

Cinematic anatomical rendering contrasting medial knee structures

Surgery versus conservative care: there are no comparative data at all. This is the one section where the evidence genuinely won't support a recommendation, and pretending otherwise would be the worst thing this page could do. No trial has ever compared conservative management against surgery, or either against doing nothing, for this condition.

Conservative success rate: DATA UNAVAILABLE. No cohort has followed an unselected group through conservative care. Single successful cases can't generate a rate.
Surgical success rate: no controlled data. Case series report high success in selected patients — complete relief in 4 knees after translocation (House 1977), 5 of 5 cured after failed conservative care (Lippitt 1993).

What the evidence does support is a sequence rather than a choice. Confirm the pain is nerve-mediated with a block before anyone operates — because the patients who did well surgically are the ones in whom the nerve was demonstrably the problem, and the patients who did badly turned out to have an unstable knee (Yang 2022). Most people with a saphenous problem after knee surgery have a numb patch that will never hurt them, and roughly a third will still have it in two years. A much smaller group has genuine neuropathic pain, and for them the pathway runs through a diagnostic block rather than through an escalating series of treatments no study has shown to work.

One finding is carried here and deliberately not acted on: persistent nerve injury tracked with a higher ACL re-rupture rate in a single retrospective cohort (Egerci 2024). There's no proposed mechanism and obvious uncontrolled confounders. It's recorded because a signal that would change practice if true deserves to be visible with its weakness attached — and it must not gate anyone's return to sport until it's replicated.

Sources

Pękala PA, et al. (2017). Risk of iatrogenic injury to the infrapatellar branch of the saphenous nerve during hamstring tendon harvesting: A meta-analysis. Muscle & Nerve. PMID 28120438. Meta-analysis, 11 studies, n=1,050. Vertical 51.4%, oblique 26.0%, horizontal 22.4%.
Yuwarungsikul C, et al. (2026). Oblique incision is associated with lower sensory loss compared with vertical incision for hamstring harvest in ACL reconstruction. Journal of ISAKOS. PMID 42229723. Meta-analysis of 8 RCTs, n=626. RR 0.64 (0.49–0.84); certainty rated low.
Grassi A, et al. (2018). Association between incision technique for hamstring tendon harvest and risk of injury to the infrapatellar branch of the saphenous nerve. KSSTA. PMID 29423546. Meta-analysis, 8 studies, n=675.
Hardy A, et al. (2017). Complications following harvesting of patellar tendon or hamstring tendon grafts for ACL reconstruction. OTSR. PMID 28888527. Systematic review, 36 articles. Sensory deficit 39.7–88%.
Egerci OF, et al. (2024). The prognosis of iatrogenic saphenous nerve injuries during hamstring tendon harvesting in ACL reconstruction. J Orthop Surg Res. PMID 39049103. Cohort, n=159, ≥2 years.
Sanders B, et al. (2007). Prevalence of saphenous nerve injury after autogenous hamstring harvest: an anatomic and clinical study of sartorial branch injury. Arthroscopy. PMID 17868834. n=164 survey plus 11 cadavers.
Yang SR, et al. (2022). Diagnostics of infrapatellar saphenous neuralgia — a reversible cause of chronic anteromedial pain following knee surgery. European Radiology. PMID 34342695. Diagnostic cohort, n=55.
Papastergiou SG, et al. (2006). Injuries to the infrapatellar branch(es) of the saphenous nerve in ACL reconstruction: vertical versus horizontal incision. KSSTA. PMID 16328463. n=230 knees. 39.7% vs 14.9%, P<0.001.
Mahmood A, et al. (2020). Clinical and electrophysiological assessment of injury to infrapatellar branch(es) of saphenous nerve during ACL reconstruction using oblique incision. The Knee. PMID 32563427. Prospective, n=50.
Patterson DC, et al. (2019). No safe zone: the anatomy of the saphenous nerve and its posteromedial branches. The Knee. PMID 30902515. 16 cadaveric knees.
Settergren R (2013). Conservative management of a saphenous nerve entrapment in a female ultra-marathon runner. J Bodyw Mov Ther. PMID 23768272. Case report — states no documented conservative interventions existed in the literature.
Shi X, et al. (2024). A randomized double-blind trial of 5% dextrose versus corticosteroid hydrodissection for meralgia paresthetica. Pain Physician. PMID 39621981. RCT, n=56 — a different nerve, carried as an analogy only.
House JH, Ahmed K (1977). Entrapment neuropathy of the infrapatellar branch of the saphenous nerve. Am J Sports Med. PMID 907036. Case series, 4 knees.
Lippitt AB (1993). Neuropathy of the saphenous nerve as a cause of knee pain. Bull Hosp Jt Dis. PMID 8443553. 5 cases.
Hemler DE, et al. (1991). Saphenous nerve entrapment caused by pes anserine bursitis mimicking stress fracture of the tibia. Arch Phys Med Rehabil. PMID 2009053.
Du T, et al. (2025). The effect of preoperative ultrasound localization on the incidence of infrapatellar branch of the saphenous nerve injury after hamstring tendon harvesting. J Orthop Surg Res. PMID 40140833. n=60.
Ferembach B, et al. (2025). Total knee arthroplasty and persistent pain: a neuropathic perspective on peroneal and saphenous nerve compression. International Orthopaedics. PMID 40009174.
Obana KK, et al. (2024). Surfer's neurapraxia — an uncommon surfing injury of the saphenous nerve. Phys Sportsmed. PMID 38975984.

Go Deeper

Most health advice never tells you how strong the evidence actually is, or where it runs out. The Verdict does, one topic a week, free.

Join The Verdict

A nerve that only carries feeling gets squashed near your knee, so it burns, but nothing is weak.