SYNTH-73 · Guide
Pain and overuse injury in runners and cyclists
Red flags go to a clinician, load-related Achilles tendon pain can be trained under a pain-monitoring rule (one trial; any other site is transfer), the load that hurts is swapped rather than dropped, and the return is gated by symptoms, not dates; the assistant never names a diagnosis
The question it answers
"My knee / Achilles / shin / back / hands hurt — do I train today, change the session, stop, or see someone?" It does not diagnose: the assistant repeats what the athlete reports ("pain at the front of the knee when pedalling"), never names a condition, and uses a clinician's diagnosis as given.
Default (what to prescribe first)
Not from a study; the base's own practice — the order of the questions:
1
Red flags first (list below). Any one → stop the painful activity, see a clinician, no "easy version".
2
No red flag, pain tied to load: keep training under the pain-monitoring rule.
Its only tested source is REF-628 (B+): 38 patients with Achilles tendinopathy, all on the same tendon-loading rehabilitation, randomised to keep running and jumping under a pain-monitoring model or to rest from tendon-loading activity for 6 weeks — no significant difference in the rate of improvement over 12 months.
The rule in that trial (full text; not checked against the abstract): pain during activity up to 5 on a 0-10 scale, not increasing week to week, settled by the next morning; otherwise reduce load. The 5/10 is the trial's model, not a tested cut-off. The ceiling the assistant gives is ≤3/10 — the base's own, more conservative practice, the same number as SYNTH-36.
3
At any site other than the Achilles the rule is transfer. Inference: no entry tests it for knee, shin, hip, heel, back or any cycling complaint.
4
Find the load change that preceded the pain and take that out, not all training (REF-624, C+: a conceptual tendon model read as load modification; no protocol; using it for pain that is not in a tendon is inference).
5
Swap, do not drop: replace the painful load with one that does not provoke it (below).
6
Not improving within about two weeks of sensible changes, or recurring → clinician. Not from a study; the base's own practice.
When NOT to use it / red flags
Not from a study; the base's own practice (the list is carried as practice in the REF-622, REF-624, REF-3497 and REF-641 entries; no paper tested it as a screen). Any one means stop and see a clinician; the pain-monitoring rule does not apply:
Possible bone stress: localised bony tenderness; impact pain that worsens through a run; pain at rest or at night; groin, front-of-shin or forefoot pain in a runner. An international Delphi panel of 41 reached consensus on 41 of 58 statements on bone stress injuries; this base does not hold their grading, imaging or return criteria, so do not improvise them (REF-3497, A for the consensus, not for the base's coverage of it).
Night pain, rapid swelling, a "pop" or giving way, fever or feeling unwell with the pain, pain not tied to load.
Neurological signs: numbness, tingling or weakness that persists after the session or keeps returning; back pain that travels down a leg.
Pain that changes the movement — a limp, or an altered pedal stroke — or that climbs during the session.
Missed periods, low libido or a previous bone injury alongside bone pain: in 59 national/world-class distance athletes, bone injuries were about 4.5-fold more prevalent in the amenorrhoeic and low-testosterone groups (REF-343, B, cross-sectional, elite runners and walkers); REDs is physician-diagnosed (REF-341, A). Screening questions: SYNTH-54.
Never mask pain to train. Not from a study; the base's own practice. The base's only painkiller trial with a clinical outcome is about kidneys: in 89 ultramarathon runners, acute kidney injury followed ibuprofen in 52% vs 34% on placebo, a difference whose confidence interval includes zero (REF-892, A-).
Decision (who → what)
| What the athlete reports | What to do | Source and how sure |
|---|---|---|
| Runner, Achilles pain, load-related, no red flag | Keep running under the pain-monitoring rule; progressive calf loading for about 12 weeks, set by a clinician; isometric calf holds can be one form of that loading (the base's practice), not a same-day painkiller: no relief in the 9 of 16 patients with pain to test, and the most painful got worse (REF-1148, B; SYNTH-88) | REF-628 B+, n=38 patients with diagnosed Achilles tendinopathy, all on a tendon-loading rehabilitation programme — for Achilles pain with no diagnosis and no such programme the rule is transfer; heavy slow resistance equalled eccentric training at 12 and 52 weeks, with compliance 92% vs 78%, in 58 patients with chronic mid-portion tendinopathy (REF-626, A-) |
| Runner, front-of-knee or hip pain at easy pace, no red flag (groin pain is a red flag above, not this row) | Cut the recent load jump; optionally raise step rate 5-10% at the same pace | REF-620 B: 45 healthy recreational runners, less energy absorbed at the knee at +5% and +10% and at the hip at +10% only; no injury outcome measured — load redistribution, not treatment. Diagnosed patellofemoral pain: in 69 runners, education on symptoms and training modifications did as well at 4, 8 and 20 weeks as education plus exercises or gait retraining (REF-917, B+, RCT, no untreated control) |
| Runner, suspected bone stress or shin pain with bony tenderness | Stop running, clinician. Return is theirs: walk-run, distance before speed, gated by symptoms | REF-3498 C+: scoping review, level IV evidence, no study compared return approaches; REF-622 C+ narrative |
| Cyclist, knee pain | Check saddle height as well as load: a lower saddle seems to raise patellofemoral compressive force; the authors recommend 25-30° of knee flexion; changes under 4% of trochanteric length seem not to matter | REF-882 B+: review of mechanics, not injury counts — a candidate cause, not a proven one. In 109 professional road cyclists the knee was 23% of overuse injuries but 57% of the time-loss ones (REF-921, B+, 12-month recall, Tier 4-5 — transfer) |
| Cyclist, hand numbness | Vary hand position, pad contact points, check weight on the hands (practice). Numbness or weakness that persists or keeps returning → clinician | REF-890 B: 14 riders, 6-day 420-mile tour, slower conduction in the deep motor branch of the ulnar nerve afterwards; no control group |
| Cyclist, lower-back pain | Ask about previous episodes; shorten or raise the position (practice). Persistent, worsening or leg-radiating → clinician | The lower back was 45% of overuse injuries in 109 professional road cyclists (REF-921; no causes). Transfer from rowing: in 2,082 rowers previous low back pain was the one significant risk factor (OR 2.65), training volume was not (OR 1.26, 0.83-1.91) (REF-641, B+) |
| Cyclist, neck or saddle-area pain or numbness | Say that the base has no entry; apply only the red flags, the load question and the two-week rule | No entry — practice only |
| Runner, shin pain that is not at the front of the shin and has no bony tenderness (front-of-shin pain is a red flag above) | Red flags, load question, two-week rule (practice). Once a clinician has diagnosed medial tibial stress syndrome: their graded return to running; calf exercises or a compression stocking are optional | REF-975 B+: in 74 athletes the time to complete a graded running programme did not differ with added calf stretching and strengthening or a sports compression stocking; 14 (18.9%) did not complete; no untreated control |
| Runner, outer-knee pain a clinician has called ITB syndrome | Red flags, load question, two-week rule; no cause (weak hip abductors included) or treatment presented as established | REF-919 B+: systematic review — cause limited or conflicting, treatment studies inconsistent |
| Plantar heel pain diagnosed by a clinician | Loading is the clinician's; heel raises are worth asking about | REF-920 A-: 48 patients, not runners; heel raises every second day vs daily stretching, both with shoe inserts — foot function index 29 points lower at 3 months (95% CI 6-52), no difference at 1, 6 or 12 months; no loads given |
| HYROX or hybrid athlete, "niggle" | Ask about pain and reduced training, not days off; manage the running load first | REF-904 B: 89 athletes, 12 weeks, 73% of injuries overuse, knee 22% of episodes, only half of problems involved time loss |
What to substitute (keeping fitness)
Runner who cannot run pain-free: deep-water running or cycling, whichever is pain-free. Inference: no entry tests cycling as a substitute for injured runners; the nearest measurement is muscle damage, not injury — after 3 days × 2.5 h at 70% VO2max, 13 trained runners had 133% higher creatine kinase than 22 trained cyclists (REF-235, B+, not randomised).Deep-water running: in 16 trained male runners split into matched treadmill and water groups on the same hard-and-long programme (6 wk, 5 d per week), VO2max, ventilatory threshold and economy were unchanged in both (REF-922, B); 11 competitive runners who ran only in deep water for 4 wk held 5-km treadmill time (1142.7 vs 1149.8 s; REF-923, B, no control).Alternating run and cycle sessions at 85-90% maximum heart rate for 5 wk improved 5000-m time as much as running only (N = 6 and N = 5; REF-953, B). All healthy runners — transfer to the injured, at the same duration and intensity.
Keep some intensity in the substitute. In general populations endurance was held for up to 15 weeks on 2 sessions a week or 33-66% less volume, provided intensity was maintained (REF-587, B+, narrative; data in athletes insufficient). Inference: keep hard sessions on the pain-free modality.
How to progress / periodise (the return)
Gate by symptoms, not by a calendar. After a tibial bone stress injury a scoping review names five things to address before running — bony tenderness, pain-free walking, radiological healing if high-risk, strength and loading tests, contributing factors (REF-3498, C+, level IV). The clinician owns those gates.
Other overuse pain: Not from a study; the base's own practice: restart at roughly half the prior volume of the painful activity, easy first, intensity last, each step held while the pain rule is met (SYNTH-36, C+, whose 50-70% first week is also house practice).
Do not rebuild with a jump. In 261 recreational runners building to a half marathon, those who raised weekly distance by 20-60% had 22.6 percentage points more injuries at 21 days than those who raised it by under 20% (95% CI 0.9-44.3); no difference at 56 and 98 days (REF-623, B+, healthy runners, not a return-to-run study).
A previously injured site gets the slowest ramp (inference from a risk factor; no ramp was tested): previous injury is a strong-evidence risk factor in long-distance runners (REF-564, A-), and a candidate marker in HYROX athletes (rate ratio 2.60, 95% CI 1.00-6.75, compatible with no association; REF-904, B).
Masters: Inference, not tested: progress tendon-loading work more slowly — ageing tendon loses modulus (REF-599, B–C, narrative).
Load errors that precede injury — and prevention
Running, jumps in load: 3 of 4 studies in a systematic review linked a sudden load increase to injury — including recent changes in velocity, distance or frequency (p = 0.037) and 86% more distance in the week before injury — but a 10% and a 24% weekly increase did not differ; the authors call the evidence very limited, and there is no support for a "10% rule" (REF-565, B+).In 874 novice runners a >30% two-week progression gave HR 1.59 (0.96-2.66, p = .07) for distance-related injuries, a non-significant trend (REF-619, B).
Do not steer by an acute:chronic ratio: a commentary finds no evidence supporting it for injury prevention (REF-365, C).
Cyclists: no entry links load jumps to injury in cyclists; the runner evidence transfers weakly (SYNTH-29, C+).
Strength as prevention: across 25 randomised trials (26,610 participants) in general sport, strength programmes cut injuries (RR 0.315, 0.207-0.480), overuse injuries were also reduced (RR 0.527) and stretching did nothing (RR 0.963) (REF-861, A-). Transfer: not endurance athletes — direction, not effect size.The one trial in runners is null: 720 first-time marathon runners, a self-directed 10-minute programme 3 times weekly, overuse injury resulting in noncompletion 7.1% vs 7.3% (risk ratio 0.97, 0.57-1.63; REF-924, A-; light, unsupervised — heavy or supervised lifting untested).Programmes including the Nordic hamstring exercise halved hamstring injuries (RR 0.49; 15 studies, 8,459 athletes, several sports; REF-766, A) — for athletes who sprint; not shown for endurance overuse pain.
What the evidence does and does not show
Shows, each in one population: pain-monitored running did not slow Achilles tendinopathy recovery (n=38 patients); a 20-60% jump in distance cost recreational runners injuries within three weeks; strength programmes prevent injuries in general sport; deep-water running held fitness for 4-6 weeks in healthy trained runners.
Does not show: that the pain rule is safe at any other site; that heavy isometric holds reliably relieve tendon pain (patellar: no better than dynamic exercise, under 1/10 of relief, in a pre-registered crossover of 21, REF-1147, B+; not superior to isotonic exercise in a review of 10 small RCTs, REF-1149, A-; Achilles: no relief in the 9 of 16 patients with pain to test, REF-1148, B); that any percentage ramp is safe; that changing saddle height resolves knee pain; that cross-training preserves fitness in injured athletes; that strength training prevents injury in runners; that any red-flag list works as a screen.
Every row is load management around a symptom; none is a diagnosis or a treatment plan.
Evidence gaps
Overuse injury in amateur cyclists, with risk factors (REF-921 is professionals).
Saddle-area numbness and saddle design; neck pain and reach; cleats and crank length versus knee pain — no entry.
Cross-training in injured (not healthy) runners; elliptical — no entry.
Supervised or heavy strength training against injury in distance runners — REF-924 tested a light self-directed programme.
Load change and injury in cyclists; a tested pain threshold (3 vs 5 of 10); a validated coach-level red-flag screen.
How this entry was checked and changed
Why this level
(continue-under-pain rests on one B+ RCT in Achilles tendinopathy; load errors on B/B+ observational runner cohorts; bone stress on an A consensus whose criteria the base does not hold plus C+ reviews; bike fit on a B+ mechanical review; prevention on A/A- general-sport meta-analyses and one null A- trial in first-time marathoners; the triage itself — red flags, substitution, return steps — is house practice)
Content correction 2026-09-27
Revision 2026-10-01: folded REF-917 (patellofemoral pain: education alone did as well as education plus exercises or gait retraining, 69 runners), REF-919 (ITB systematic review: nothing established) and REF-920 (plantar heel pain: heel raises ahead at 3 months only, 48 patients) — the 'no entry' row is split, ITB and plantar heel now have their own rows gated on a clinician's diagnosis, and the lede no longer lists them as uncovered.Non-bony shin pain, neck and saddle-area pain stay 'no entry'.
Revision 2026-10-01: folded REF-921 (109 professional road cyclists, recall survey) into the cyclist knee and lower-back rows; the back row no longer says 'No cycling entry' but 'causes unstudied'. Folded REF-922 and REF-923 (deep-water running, healthy trained runners) and REF-953 (alternating run/cycle sessions) into the substitution section, replacing 'deep-water running: no entry'; all three labelled transfer to injured runners.
Revision 2026-10-01: REF-924 (720 first-time marathoners, self-directed 10-minute strength programme, no fewer injuries) added next to the general-sport prevention meta-analysis: the line no longer reads as if strength prevents injury in runners; the general-sport figure stays, the runner trial is stated as null and limited to a light unsupervised programme. Level unchanged at B-: the triage is still house practice.
Verification 2026-10-01: all 33 cited entries opened in the KB.md snapshot; every number in the body (38, 12 months, 5/10, 41 of 58, 59, 4.5-fold, 52% vs 34%, 58/92%/78%, 45, 69, 25-30°, 4%, 23%/57%, 45%, 2,082/OR 2.65/1.26, 14/420 miles, 48/29 points/6-52, 89/73%/22%, 133%, 16/6 wk, 1142.7/1149.8, N 6 and 5, 15 weeks/33-66%, 261/22.6/0.9-44.3, 2.60/1.00-6.75, 3 of 4/p=0.037/86%, 874/HR 1.59, 25 trials/26,610/0.315/0.527/0.963, 720/7.1 vs 7.3/0.97, 0.49/15/8,459) matches its entry.
Verification 2026-10-01: no clinical claim was found that is neither cited nor labelled practice: the red-flag list, substitutions, two-week rule, hand/back position advice and return steps all carry a practice label in the same line or section header. The safety problems are two row/red-flag overlaps (hip vs groin; shin) and the Achilles row/title dropping the trial's conditions (diagnosed patients, concurrent rehab programme).
Verification 2026-10-01: the framework does not state which pain ceiling the assistant gives (5/10 trial model or ≤3/10 house practice), and SYNTH-36 carries ≤~3-4/10; owner decision, no pair written.
Verification 2026-10-01: REF-564 also reports that an increase in weekly distance was protective for knee injuries; SYNTH-73 omits it (not a misstatement). No SOURCE flags: every cited entry has a DOI or PMID; REF-3497 is abstract-only and the SYNTH says so.
2026-10-01: pain ceiling made explicit: the assistant gives ≤3/10 (base practice, same as SYNTH-36); 5/10 is quoted only as what the supervised Achilles trial (REF-628) allowed.
2026-10-01: level tags next to cited frameworks updated to their current levels after the independent audit.
Revision 2026-10-01: REF-975 (Moen 2012, B+, 74 athletes with clinician-diagnosed MTSS) added as the shin row: calf exercises or a compression stocking added nothing measurable to a graded running programme; 18.9% did not complete; no untreated control.The row is split from the cyclist neck/saddle row; the red flag for front-of-shin pain and bony tenderness and the clinician gate are unchanged, and for undiagnosed shin pain the action is still red flags, load question and the two-week rule (practice). Level C+ unchanged.
Revision 2026-10-01: 'The base's only painkiller trial is about kidneys' narrowed to 'only painkiller trial with a clinical outcome' — the base now also holds REF-983 (ibuprofen and gut-injury markers); REF-983 is not cited here and the kidney claim is not strengthened.
Revision 2026-10-01: to hold the length — the clause 'and is quoted only as what that supervised rehabilitation trial allowed' (redundant with 'the trial's model, not a tested cut-off') and the duplicate 'Elliptical: no entry.' (still in Evidence gaps) were removed; the Moen candidate gap line became the remaining gap (no untreated comparison).
Revision 2026-10-16: added to 'Does not show' that heavy isometric holds reliably relieve tendon pain (REF-1147 patellar crossover, REF-1149 meta-analysis, REF-1148 Achilles); cross-references extended. Level unchanged.
Revision 2026-10-16: the Achilles row now says isometric calf holds are one tolerable form of the progressive loading, not a same-day painkiller (REF-1148, B: no relief in the 9 of 16 patients with pain to test, the most painful worse; "one form of the loading" labelled practice), with a pointer to SYNTH-88; cross-reference added. Level unchanged.