Biomechanical and musculoskeletal measurements as risk factors for running-related injury in non-elite runners: a systematic review and meta-analysis of prospective studies

Key points

1. 0123 of 25 meta-analyses found no significant difference in biomechanical or musculoskeletal measures between runners who did and did not develop injury
2. 02Knee extension strength was slightly lower in injured runners (SMD -0.19, 95% CI -0.36 to -0.02, p=0.03), but the effect size is trivial
3. 03Hip adduction velocity during stance was lower in injured runners (MD -12.80 degrees/s, 95% CI -25.22 to -0.38, p=0.04), but this analysis included only two studies
4. 04No significant pooled associations were found for joint range of motion, static foot posture, running kinetics, or running kinematics
5. 05Current evidence does not support biomechanical screening as a standalone injury prevention tool for recreational runners

How it was conducted

Design

Systematic review and meta-analysis of prospective studies (PRISMA-compliant, PROSPERO registered CRD42018089392)

Databases searched

MEDLINE, EMBASE, CINAHL, SPORTDiscus, AMED, Cochrane Library, grey literature; search to January 13, 2021

Participants

3,404 non-elite adult runners across 30 included studies (66.6% female); recreational, novice, and collegiate non-NCAA-D1/D2 runners

Risk factors tested

Over 100 discrete biomechanical and musculoskeletal variables including muscle strength, joint ROM, static alignment, plantar pressure, running kinematics, and running kinetics

Pooled analyses

25 separate meta-analyses performed; 19 studies eligible for pooling

Effect size interpretation

Hopkins modified Cohen scale: <0.20 trivial, 0.20-0.59 small, 0.60-1.19 moderate, >1.20 large

What they found

• 23 of 25 meta-analyses found no statistically significant association between the biomechanical or musculoskeletal variable and running-related injury
• Knee extension strength meta-analysis (4 studies, 594 runners, 299 injured): significantly less strength in injured runners (SMD -0.19, 95% CI -0.36 to -0.02, p=0.03, I2=0%); effect size classified as trivial
• Hip adduction velocity meta-analysis (2 studies, 253 runners, 89 injured): significantly lower velocity in injured runners (MD -12.80 degrees/s, 95% CI -25.22 to -0.38, p=0.04, I2=0%); manually calculated Cohen's d estimated as small (0.32)
• No significant pooled associations for hip abduction strength, hip external rotation strength, ankle dorsiflexion ROM, hip rotation ROM, vertical loading rates, or static foot posture measures
• Individual study findings not included in meta-analysis: peak braking force in highest tertile associated with injury at HR 5.08 (95% CI 1.71-15.03, p=0.003) compared to middle tertile (Napier et al.); knee joint stiffness increase of 6.89 Nm/deg associated with 18% increase in RRI odds (OR 1.184, 95% CI 1.021-1.374, p=0.03) (Messier et al.)
• Navicular drop >10 mm associated with 6.6 times greater likelihood of medial exercise-related lower leg pain in one non-pooled study (OR 6.6, 95% CI 1.2-38.0, p=0.03) (Bennet et al., unconfirmed eligibility)

Limitations

• Many biomechanical variables could not be pooled due to heterogeneous measurement methods and statistical reporting across studies, potentially missing important associations
• Lack of a consensus definition for 'elite' and 'sub-elite' runners required exclusion of three studies from meta-analysis, limiting generalisability
• Most included studies did not control for training load, which is itself a likely contributor to injury risk, potentially confounding results
• Publication bias cannot be excluded; studies with null findings may not have been published, which could inflate the apparent importance of significant individual findings

Why it matters

For patients

Recreational runners should not expect that passing or failing a biomechanical screening will reliably tell them whether they will get injured; injury prevention likely depends on training load management as much as physical measurements.

For clinicians

Current evidence does not justify routine pre-participation biomechanical or musculoskeletal screening for injury prediction in non-elite runners; clinical decisions should not rely on these measurements in isolation.

For readers

This is the most comprehensive prospective-only meta-analysis on this topic to date, and its null findings challenge the widespread clinical assumption that gait and strength measures are reliable predictors of running injury.