REF-1393 · Alexander et al., 2022 · British Journal of Sports Medicine · Paper · PubMed

What actually prevents knee injuries in runners?

Of 18 prevention trials, only soft-landing retraining cut risk (one trial of 320, RR 0.32, low certainty). Shoes, exercise plans, gentler build-ups, education: no clear effect.

Main caveat: Read from the abstract only. Per-trial effects, confidence intervals of the null comparisons, outcome definitions, comparators and follow-up could not be…
fairly strong evidenceMeasured on non-athletes to trained athletes

Alexander and colleagues set out to evaluate interventions that prevent or manage running-related knee injuries, searching five databases to May 2022 for randomised controlled trials whose primary aim was knee injury prevention or management. They included 30 RCTs, 18 on prevention and 12 on management, in novice and recreational runners.

For prevention, low-certainty evidence from one trial of 320 participants showed that running technique retraining to land softer reduced knee injury risk against control treadmill running (RR 0.32, 95% CI 0.16 to 0.63).

Very low to low certainty evidence from the other 17 prevention trials (24 to 3,287 participants each) indicated that various footwear options, multicomponent exercise therapy, graduated running programmes and online and in-person injury-prevention education did not influence knee injury risk (RR range 0.55 to 1.06).

For runners with patellofemoral pain, very low to low certainty evidence indicated that running technique retraining, medial-wedged foot orthoses, multicomponent exercise therapy and osteopathic manipulation can reduce knee pain in the short term (SMD range −4.96 to −0.90).

The authors conclude that retraining to land softer may reduce knee injury risk by two-thirds and that running-related patellofemoral pain may be managed by a variety of active and passive interventions.

Our reading: the single positive prevention trial matches the base's REF-1119 (320 novice runners, impact-feedback gait retraining against treadmill running without feedback; 16% vs 38% injured over 12 months for all running injuries) — the same n and comparison, though the abstract does not name it — so "land softer" rests on one trial, not on a body of evidence; and everything sold for knee prevention besides it (shoes, generic exercise programmes, gentler build-ups, education) showed no effect on knee injuries in runners.

Protocol

Design: systematic review and meta-analysis of RCTs, registered (PROSPERO CRD42020150630); certainty of evidence graded (very low to low across findings).
Sources: MEDLINE, EMBASE, CINAHL, Web of Science and SPORTDiscus up to May 2022.
Eligibility: RCTs with a primary aim of evaluating intervention(s) to prevent or manage running-related knee injury.
Included: n = 30 RCTs (18 prevention, 12 management) in novice and recreational runners; prevention trials ranged from 24 to 3,287 participants. The total number of participants, the outcome definitions (what counted as a knee injury) and follow-up lengths are not in the abstract.
Funding and interests: funding not available from the abstract or the PubMed record; Crossref lists no funder. Competing interests as printed in the PubMed record: "AGC and CJB are Associate Editors of BJSM." No device maker, federation or data platform is named.

Key Results

QuestionIntervention(s)Result (abstract)Certainty
Preventionrunning technique retraining to land softer vs control treadmill running (1 trial, 320)RR 0.32 (95% CI 0.16 to 0.63) for knee injurylow
Preventionfootwear options; multicomponent exercise therapy; graduated running programmes; online and in-person injury-prevention education (17 trials, 24 to 3,287 participants each)did not influence knee injury risk; RR range 0.55 to 1.06very low to low
Management of patellofemoral painrunning technique retraining; medial-wedged foot orthoses; multicomponent exercise therapy; osteopathic manipulationshort-term knee pain reduced; SMD range −4.96 to −0.90very low to low
Authors' conclusion: low-certainty evidence that retraining to land softer may reduce knee injury risk by two-thirds; very low to low certainty that running-related patellofemoral pain may be managed by active (running technique retraining, multicomponent exercise therapy) and passive (foot orthoses, osteopathic manipulation) interventions.
Not in the abstract: the confidence interval of each null comparison (only the range of point estimates is given), which footwear options were tested, comparators and follow-up for the management trials, and any long-term pain or return-to-running outcome.

Mechanisms

Measured (in the one positive prevention trial, as summarised): retraining that aimed at softer landings was followed by fewer knee injuries than treadmill running without it; the base's REF-1119 reports that the 2-week visual-feedback retraining also lowered vertical loading rates.
Authors' interpretation: softer landing may cut knee injury risk by about two-thirds; several active and passive interventions can reduce short-term patellofemoral pain in runners.
Inference: lowering impact loading per step is a plausible route for knee injury prevention, but the base holds no injury trial of step rate itself; a higher step rate lowers knee energy absorption in healthy runners (REF-620, a load measure, not an injury outcome). The abstract does not say whether the retraining trials in patellofemoral pain used softer landing, step rate or another cue.

Practical Application

What actually prevents knee injuries in runners? Of 18 prevention trials, only soft-landing retraining cut risk (one trial of 320, RR 0.32, low certainty). Shoes, exercise plans, gentler build-ups, education: no clear effect.
Do not sell a shoe, a generic exercise routine, a slower build-up or advice as knee-injury prevention: in 17 trials none influenced knee injury risk (RR range 0.55 to 1.06, very low to low certainty; A-, from the abstract).This agrees with REF-716 (no shoe type reliably reduces running injuries) and REF-925 (a graded programme did not lower injuries in novices), and with REF-1244 for light prevention routines in runners. A supervised hip-and-core programme is a separate question (REF-1384).
Soft-landing retraining is the one prevention option with a positive trial, and it is one trial: the base's REF-1119 (B+, 320 novice runners, 2 weeks of real-time visual feedback on impact) is very likely the trial pooled here. Not from this paper; the base's own practice: offer it as something to try with a qualified coach or clinician, not as a proven shield.
Front-of-knee pain in a runner: several things helped in the short term: running technique retraining, medial-wedged foot orthoses, multicomponent exercise therapy and osteopathic manipulation reduced short-term knee pain (very low to low certainty).In the one runner RCT the base holds, education alone did as well as education plus exercises or gait retraining (REF-917, B+); the best-practice guide in REF-1381 puts exercise therapy with education first. Which option suits a runner is the clinician's choice (SYNTH-73).
What it is good for as a source: the runner-specific answer to "what prevents knee injury" (one method, one trial) and the list of popular strategies that showed no effect on knee injuries. What it does not show: that any strategy prevents knee injury with more than low certainty; long-term effects on patellofemoral pain; which retraining cue, orthosis or exercise works best; effects in trained or elite runners; anything about injuries other than the knee.

Limitations

Read from the abstract only. Per-trial effects, confidence intervals of the null comparisons, outcome definitions, comparators and follow-up could not be checked.
Low certainty throughout. Every finding is graded very low to low; the positive prevention finding is a single trial.
Ranges, not estimates. The "no influence" finding is given as a range of point estimates (0.55 to 1.06) across different interventions, not as one pooled result; a point estimate of 0.55 with a wide interval is compatible with a benefit that the trials could not show (our reading).
Short-term pain only for patellofemoral pain management; an SMD of −4.96 is implausibly large for a pain outcome and suggests a small trial with little spread (inference).
Interests: competing interests as printed in PubMed: two authors (AGC, CJB) are Associate Editors of the journal. Funding not available from the abstract or the PubMed record; Crossref lists no funder.
How this entry was checked and changed
Why this level

(from the abstract) — a systematic review of RCTs (rubric A) held at A- because every conclusion is very low or low certainty, the one positive prevention finding rests on a single trial, and the management findings come from small trials (an SMD as large as −4.96 suggests at least one very small or imprecise trial; inference). Read from the abstract (closed access; Europe PMC holds no full text and OpenAlex lists no open copy)