Running injuries
Red flags and bone-like pain: no running, a clinician, a symptom-free return; a sore tendon: swap the provoking load, then rebuild under a pain limit over months; front-of-knee and heel pain: a clinician's exercise, taping and education first; every return gated by symptoms; in runners only supervised hip-and-core work and soft-landing retraining prevented injury; no load rule is safe; never name the injury
The question it answers
A runner reports pain — Achilles, shin, knee, heel, hip, buttock, thigh, calf or foot — or asks how not to get hurt. Do they stop and see someone, keep running with one change, or swap running for something that does not hurt while a clinician leads the loading; when and how does running come back; and what is worth doing to prevent the next injury?
Short answer: red flags and bone-like pain go to a clinician with no running; otherwise find and take out what changed, apply the rule for the tissue (bone symptom-free, tendon within a pain scale, knee and heel with a clinician's first-line package), keep fitness on what does not hurt, and come back on symptoms, never on a date.
The assistant repeats what the athlete reports ("pain along the inner edge of the shin"), never names a condition, and uses a clinician's diagnosis as given (SYNTH-73).
What this guide owns and what it leaves
| Owns (decided here) | Leaves to |
|---|---|
| The runner's first questions, the rule by tissue (bone, tendon, knee and heel, muscle), what to keep doing and the clinician's usual first line by injury | The cross-sport triage, the never-name-a-diagnosis rule, the two-week rule and HYROX niggles → SYNTH-73; pain on the bike → SYNTH-92 |
| The runner's bone-stress sites, added to SYNTH-73's red flags | Layoffs, illness and the return after a break → SYNTH-36; the ramp and the deload → SYNTH-29; the long run → SYNTH-84 |
| Return to running after an injury: which gates are written down and which are practice | Foot strike → SYNTH-86; step rate → SYNTH-55; strength for runners → SYNTH-44; new runners → SYNTH-75 |
| What prevents running injuries and what does not, and the honest status of load rules | The REDs screen → SYNTH-54; isometric holds → SYNTH-88; soreness after new work, which is not injury → SYNTH-23 |
Default — the decision, in order (what to prescribe first)
Not from a study; the base's own practice for the order of questions, the ≤3/10 ceiling and the two-week rule; the sources say what each step rests on:
Repeat what they report, then ask: exactly where; when it hurts (first steps in the morning, the start of a run, building through the run, afterwards, at rest or at night); since when; what changed in the weeks before (a run much longer than recent ones, more hills, speed work or racing, new shoes or a drop change, a new surface); whether it has happened before; how they are eating for the training and, for women, whether periods have changed.
Why these: a previous injury is the best-supported warning sign — a previous running injury in runners over about 20 km a week (moderate quality) and a previous non-running injury in lower-volume, mostly novice runners (high quality), with no risk model able to predict who gets hurt (REF-1387, A-); in 2,378 event runners an injury in the previous year went with about twice the odds of a new one (OR 2.21, 95% CI 1.84-2.65; REF-1385, A); bone stress injuries were about twice as common in women (REF-1388, A-) and accumulated with triad-related factors (REF-1392, B).
The questions themselves are practice.
Bone-like pain → no running. Walking only if it is pain-free; a clinician grades the site and the risk; the return is symptom-free (REF-1383, C+; Return section below).
Tendon-like pain — the Achilles, the sit bone, the outer hip (the guideline below is for the Achilles; at the sit bone and outer hip it is a transfer, and the outer hip has its own trial, see the table): swap the load that provokes it for one that does not, bring it back gradually under a pain scale, and start progressive loading for at least 12 weeks with a clinician, reassessed at 3 months (REF-1380, A — by the guideline's own account the loading advice rests on clinical expertise and the pain numbers on expert opinion; the 12 weeks is its working group's choice, not a graded conclusion).
Running that stays within the pain limit and is no worse the next morning is compatible with it (REF-628, B+: one trial of 38, not a non-inferiority test). The ceiling the assistant gives is ≤3/10 — the base's own practice (SYNTH-73), more conservative than the 5/10 in that trial and the guideline's 0-5/10.
Front-of-knee pain: cut the load that provoked it; once a clinician has diagnosed it, knee-targeted exercise therapy with hip exercise as needed and education underneath is the first line (REF-1381, A-), and in the one trial of runners education on symptoms and load alone did as well (REF-917, B+) — give both.
Heel pain on the first steps: taping, plantar-fascia stretching and education on shoes and load, together for about 4-6 weeks, then the clinician's next step (REF-1382, A-).
Keep fitness on a pain-free substitute and keep some intensity in it (healthy-runner data, a transfer; "Keeping fitness" below). Offer it rather than wait to be asked (practice).
Not improving within about two weeks of sensible changes, or recurring → a clinician (SYNTH-73's rule, practice). A clinician treating an Achilles reassesses at 3 months (REF-1380) and one treating front-of-knee pain after at least 6 weeks (REF-1381); the assistant's two weeks is the point to involve one, not a verdict on treatment.
Promise nothing — no date, no prevention, no safe load. The assistant's own class of intervention, risk-factor advice, did not prevent injuries in a large trial (REF-1385, A).
When to stop running and see a clinician
SYNTH-73's red flags apply unchanged: night pain, rapid swelling, a "pop" or giving way, fever or feeling unwell with the pain, pain not tied to load, numbness or weakness that persists or returns, pain that changes the movement or climbs during the session, and bone pain with missed periods, low libido or a previous bone injury.
No paper tested a red-flag list as a screen; the runner-specific signs below are practice unless a source is given, each meaning no running and a clinician, never a diagnosis:
Decision table by injury
"(practice)" marks a call that is not from a study; the base's own practice. Every first-line cell is what a clinician usually starts with once they have made the diagnosis; the assistant offers it as such, never as its own treatment.
| What the athlete reports | Clinician now if | May keep doing while it settles | First line, with a clinician — what the base shows (refId, level, key number) | Return to running |
|---|---|---|---|---|
| Pain at one spot on a bone, building through the run, on walking or at night (front of the shin, top of the foot, outer forefoot, groin, buttock) | Always; no running | Pain-free cycling or deep-water running; walking only if pain-free; cross-training to keep fitness is the leading group's advice (REF-1383), the modality is practice | Load reduced first (partial or non-weight bearing) until symptoms settle; painkillers only briefly and only for pain at rest or at night (REF-1383, C+, abstract, low-risk tibial and metatarsal shaft injuries); the clinician grades site and risk — the complication-prone sites are the femoral neck, tarsal navicular, anterior tibial shaft and fifth metatarsal (REF-1379, A-); fuelling and periods checked (REF-1392, B; SYNTH-54) | After 5 consecutive pain-free days of daily activity; volume before speed; no symptoms during, after or the next day (REF-1383, C+, untested); for the tibia, five criteria first and alternate-day walk-run (REF-3498, C+, level IV); pooled study means 44 days for the posteromedial tibia (95% CI 27-61), 56 for the fibula (13-100), 107 for the femoral neck (79-135) and 127 for the navicular (102-151), with more than 90% returning (REF-1379, A-) — averages across athletes, not a date |
| Achilles stiff and sore in the morning or after sitting, sore at the start of runs | A sudden kick or pop, cannot rise on the toes, rapid swelling, night pain (practice) | A swap first — swimming, cycling, a cross-trainer or walking for pain above the heel, swimming or cycling for pain where the tendon meets the heel bone (REF-1380, expert opinion); running comes back gradually and may continue only within the pain limit (≤3/10, practice), no worse the next morning and not rising week to week (REF-628) | Education, a temporary swap of the provoking load, a gradual return steered by a pain scale and progressive calf strengthening for at least 12 weeks, reassessed at 3 months; any active treatment beat wait-and-see at 3 months, very low certainty (REF-1380, A guideline, GRADE low to very low, the loading advice expert opinion); heavy slow resistance matched eccentric training at 12 and 52 weeks with compliance 92% vs 78% (REF-626, A-, 58 patients); continued running up to 5/10 showed no evidence of a difference from rest in 38 patients (not a non-inferiority test), with rest numerically ahead at 12 months (REF-628, B+); isometric holds are not a same-day painkiller (REF-1148, B; SYNTH-88) | Months, not weeks: full symptom-free loading only after a few months of active treatment, and a return within days went with more recurrence; the working group rebuilds towards 5 km over 6-8 weeks once daily activity is within the limits, and suggests keeping single-leg calf work about three times a week afterwards; 23-37% still have symptoms years later and 85% of athletes return to sport (REF-1380, expert opinion and very low to low evidence) |
| Front of the knee: stairs, squats, sitting long, downhill | Swelling, locking, giving way (practice; swelling and giving way are SYNTH-73's) | Shorter, flatter runs within the pain limit; cycling if pain-free (practice) | Knee-targeted exercise therapy (short-term pain SMD 1.16, 95% CI 0.66-1.66, high certainty), hip exercise added as needed, education underneath everything; prefabricated foot orthoses (OR 4.31, low certainty), taping and running retraining (never adequately tested, graded "low") as add-ons by need (REF-1381, A-, trial populations not described, not runner-specific, no trial past 3 months); in 69 runners education on symptoms and training load alone did as well over 20 weeks as education plus exercise or plus gait retraining (REF-917, B+); in runner trials retraining, medial-wedged orthoses, exercise and osteopathic manipulation cut short-term pain at very low to low certainty (REF-1393, A-, abstract) | By symptoms (practice); no source gives return criteria |
| "Will a knee brace help in my marathon?" | — | — | Braces lowered new kneecap pain in two small trials of recruits and young athletes in 6-8 weeks of very heavy training (RR 0.40, 95% CI 0.22-0.73, low certainty), and the authors advise against widespread bracing outside such settings; no trial tested a brace for existing pain or in a race (REF-1386, A-) | Nothing new on race day: a brace a clinician suggested is worn in training first (practice) |
| Inner edge of the shin, diffuse, tender along a stretch rather than at one spot | One tender spot on the bone, front-of-shin pain, pain at rest or at night (→ bone row) | Reduced running if mild and within the pain limit; otherwise a substitute (practice) | After a clinician's diagnosis, a graded running programme; adding calf stretching and strengthening or a sports compression stocking did not change the time to complete it in 74 athletes (REF-975, B+, no untreated control) | The clinician's graded programme; 18.9% did not complete it (REF-975) |
| Heel pain on the first steps in the morning or after sitting | Numbness or tingling, night pain, swelling, fever, pain after a fall (practice) | Reduced running within the pain limit; break up long standing (the guide's expert advice, REF-1382) | Taping (first-step pain SMD 0.47 at 1 week), plantar-fascia stretching (SMD 1.21 against radial shockwave, short term) and individualised education — comfortable, supportive shoes with a heel-to-toe drop, no barefoot walking or flat unsupportive shoes until it has gone, no rapid training changes, pain-monitoring — together for about 4-6 weeks; then shockwave, then custom orthoses (REF-1382, A-: one moderate trial each for taping and stretching, the education untested, no runner subgroup, a condition the authors describe as mainly of sedentary adults); calf stretching showed no effect (REF-1382); heel raises were ahead of stretching at 3 months only, in 48 patients who were not runners (REF-920, A-) — untested at the guide's bar, not refuted | Load-managed; no tested return (practice) |
| Outer knee, starting at a set distance, worse downhill | Swelling, locking, giving way | Shorter, flatter runs within the pain limit (practice) | The clinician's choice: no cause, weak hip abductors included, and no treatment is established (REF-919, B+, a 2012 review); hip strengthening is a common clinical choice, not an established treatment (not from the base) | By symptoms (practice) |
| Deep buttock or sit-bone pain; sitting hurts; worse uphill or with speed | Sudden pain with bruising; pain or numbness down the leg (SYNTH-73); buttock pain with impact (→ bone row) | Flat, easy running within the pain limit; no hills, sprints, lunges or deep hip-flexion stretches early on (practice; that deep hip flexion presses the tendon against the sit bone is the review authors' proposal, REF-1389) | Nothing proven: 12 studies, 2 small RCTs, every outcome very low certainty; shockwave beat a 6-week multimodal programme in one trial of 40 professional athletes (REF-1389, B+); the authors' pragmatic advice is to try first what works in other tendinopathies; in the base that is progressive tendon loading under a clinician (our reading, a transfer from REF-1380 and REF-626) | By symptoms; no study (practice) |
| Outer hip; lying on that side hurts | Groin pain or pain with impact (→ bone row); night pain | Running within the pain limit; avoid positions that press the outer hip, such as lying on that side (practice; the trial's education covered compression, and its list of positions sits in an appendix not read, REF-1390) | Education on tendon load plus targeted exercise, 14 sessions over 8 weeks: global success 77% at 8 weeks against 58% after one corticosteroid injection and 29% with wait-and-see (number needed to treat 2.0 against wait-and-see), and 78% vs 57% vs 52% at 52 weeks (REF-1390, A; 204 adults aged 35-70, 82% women, not runners — a transfer) | By symptoms (practice); the trial says nothing about running load |
| A sudden grab at the back of the thigh while running fast | Always, for the diagnosis and the rehabilitation plan; bruising, a gap or a limp → promptly (practice) | Pain-free cycling or walking (practice) | Clinician-led rehabilitation from about day 3 — hamstring strengthening including eccentric work, and graded running drills — progressed with pain up to 4/10 allowed: clearance a median 17 vs 15 days with pain-free progression (P = .37), 15% more knee-flexor strength regained, 2 reinjuries in each group by 6 months (REF-1391, A-, 43 men, team sports); afterwards, programmes including the Nordic hamstring exercise halved hamstring injuries across several sports (REF-766, A) | Clearance required no pain on palpation, on stretch, in a maximal contraction and in a full sprint (REF-1391); the 4/10 applied to rehabilitation exercise, not to running through pain |
| A sudden "pop" or grab in the calf | Cannot rise on the toes, a large swelling, a kick at the back of the heel → promptly (practice) | Pain-free cross-training (practice) | The clinician's; no entry in the base | No return study (practice) |
| An ankle sprain, or ankle pain that keeps coming back | Cannot walk four steps, or bone tenderness at the specific points → an X-ray (Ottawa rules, 97.6% sensitivity, REF-1022, A) | Exercise started in the first week improved function at weeks 1 and 2 against rest, ice, compression and elevation, with no more pain, swelling or re-injury (REF-1024, A-) | A brace or tape plus an exercise programme, supervised exercise before passive treatments, care with anti-inflammatories (REF-1025, B+) | On criteria in five domains — pain, ankle impairments, the athlete's own perception, sensorimotor control, sport performance — not on days (REF-1023, C+) |
| Sore legs the day after a new, hilly or downhill run, easing over days | — | Whatever is comfortable | Soreness after unaccustomed, mostly eccentric work appears the day after and fades over days; it is not an injury (PHYS-23, established; running, above all downhill, damages muscle more than cycling, PHYS-22, established; SYNTH-23) | Once it has eased (practice) |
Return to running — which gates are written down and which are practice
Keeping fitness while running is out
Prevention — what has trial support, what did not, and the load rules
Key Findings — what the evidence shows
Which bone matters. Pooled return-to-sport times ran from 44 days (posteromedial tibia) to 127 (navicular), more than 90% returned, and complications clustered at the femoral neck, navicular, anterior tibial shaft and fifth metatarsal (REF-1379, A-, 76 studies, abstract).
The Achilles default is a swap, then a slow rebuild. The national guideline starts with education, a temporary swap of the provoking load, a gradual return under a pain scale and at least 12 weeks of calf strengthening, and gives months as the time frame; nearly all of it is its working group's expert opinion (REF-1380, A). Continued running within a pain model showed no evidence of a difference from rest in one trial of 38, not a non-inferiority test (REF-628, B+).
Runner's knee: exercise with education first. Knee-targeted exercise had high-certainty short-term evidence (SMD 1.16), not in runners specifically (REF-1381, A-); in runners, education alone did as well as education plus exercise or gait retraining (REF-917, B+), and several treatments cut short-term pain at low certainty (REF-1393, A-).
A hamstring strain can be rehabilitated through mild pain. Up to 4/10 during rehabilitation did not slow clearance and kept more strength in 43 men (REF-1391, A-).
Prevention is programme-specific. A supervised hip-and-core programme lowered injuries and an ankle-and-foot programme did not (REF-1384, A); soft-landing retraining had one positive trial (REF-1119, B+); light routines, online advice, shoes and a 10%-a-week programme did not prevent injury (REF-1244, REF-1385, REF-716, REF-925).
It does not show: a tested return-to-running protocol for any injury; a safe pain number during running outside one Achilles trial; that any load rule is safe; that an injured runner keeps fitness on a substitute; that the hip-and-core programme works unsupervised, in men or in experienced runners; anything on calf strains, inner-knee pain, chronic exertional compartment syndrome or running and knee arthritis; or which runner will get hurt.
Conflicts and how they are resolved
Practical Application — how the assistant says it
What NOT to do
Evidence gaps
Limitations
How this entry was checked and changed
(the direction of each first-line step rests on guideline-level sources and trials: the Achilles guideline REF-1380 (A, whose own GRADE conclusions are low to very low and whose loading advice, 12 weeks and pain numbers are its working group's expert opinion), the patellofemoral best-practice guide REF-1381 (A-, high certainty only for short-term pain with knee-targeted exercise), the plantar heel pain guide REF-1382 (A-, its education and load parts untested), the bone-site review REF-1379 (A-, abstract), the gluteal tendinopathy trial REF-1390 (A, not runners), the hamstring-strain trial REF-1391 (A-, 43 men), the one runner trial of pain-monitored running REF-628 (B+, 38 patients) and the prevention trials REF-1384 (A), REF-1385 (A), REF-1119 (B+) with the reviews REF-1393 REF-1386 REF-1244 (A-). Above C+ because that direction is what this evidence says; below B because the bone return rule is a C+ commentary REF-1383 and a C+ scoping review REF-3498, no criteria-based return-to-running protocol has been tested for any injury, several rows transfer from non-runners, and the triage order, the red-flag list as a screen, the ≤3/10 ceiling, the two-week escalation and every return step outside bone are practice)