The association between running injuries and training parameters: a systematic review

Key points

1. 0126.2% of runners across 36 studies sustained a running-related injury, with rates of 14.9% in novice, 26.1% in recreational, and 62.6% in competitive runners
2. 02Evidence on whether distance, duration, frequency, or intensity causes injury is conflicting across high-quality studies
3. 03The knee (25.8%), foot/ankle (24.4%), and lower leg (20.9%) are the most frequently injured body parts
4. 04No consistent link was found between sudden or gradual changes in training load and injury onset
5. 05The widely recommended '10% rule' for increasing weekly distance is not justified by the current evidence

How it was conducted

Design

Systematic review of prospective cohort studies and RCTs (PRISMA 2009); registered in PROSPERO (CRD42018112913)

Databases searched

MEDLINE/Ovid, CINAHL, Embase, SPORTDiscus from inception to July 7, 2020

Included studies

36 studies (33 prospective cohort, 3 RCTs)

Participants

23,047 runners (44% female), aged 17 years and older

Quality assessment

QualSyst tool; median quality score 86.4% (IQR 19.3%); ICC for interrater reliability 0.97 (95% CI 0.95, 0.99)

Primary outcomes

Incidence of running-related injuries and their association with training distance, duration, frequency, intensity, and recent changes in these parameters

What they found

• Overall RRI incidence: 6,043 of 23,047 runners injured (26.2%); incidence range 8.8% to 91.3% across studies
• Novice runner incidence: 14.9% (703/4720; range 9.4%-94.9%)
• Recreational runner incidence: 26.1% (2057/7888; range 17.9%-79.3%)
• Competitive runner incidence: 62.6% (259/414; range 52.6%-91.3%)
• Most injured sites: knee 25.8%, foot/ankle 24.4%, lower leg 20.9%
• Distance and injury risk: weekly distance >30 km associated with increased RRI risk in one HQ study (HR=3.28; 95% CI 1.23, 8.75; P=.02); distance >64 km associated with adjusted relative risk of 2.88 in another HQ study; however other HQ studies found no association or a protective effect (e.g., relative risk 0.901, 95% CI 0.820, 0.991 for knee injury with greater distance)
• Duration and injury risk: session duration >45 min associated with higher incidence in novice runners (20%, 24%, and 54% for 15-, 30-, and 45-min groups respectively); running >60 min/week was protective in one HQ study (HR=0.41, 95% CI 0.20, 0.86); conflicting findings across studies
• Frequency and injury risk: running 7 days/week versus 0-2 days/week associated with increased risk (relative risk males 5.92, 95% CI 2.49, 12.75; relative risk females 5.50, 95% CI 1.44, 17.39); running only 1 day/week was a risk factor in females (OR=3.6, 95% CI 1.1, 12.3)
• Intensity and injury risk: higher intensity associated with more injuries in novice runners (HR=1.28, 95% CI 1.18, 1.40); tempo runs in first 6 weeks increased RRI odds (OR=3.96, 95% CI 1.35, 11.61); regular interval training protective against knee injuries (OR=0.49, 95% CI 0.26, 0.93)
• Changes in training: injured runners had 86% greater week-before-injury distance progression than other weeks (95% CI 12.9%, 159.9%; P=.026); graded 10% weekly increase showed no benefit over faster progression (OR=0.8, 95% CI 0.6, 1.3; P=.90); increasing distance 20%-60% in one week associated with more injuries at 21 days (RD=22.6%, 95% CI 0.9%, 44.3%) but not at 56 or 98 days

Limitations

• High heterogeneity in RRI definitions across studies, with fewer than 40% using the consensus definition, preventing meta-analysis
• Most studies relied on self-reported training data; only 5 of 36 studies (14%) used GPS-based objective training load measurement
• Confounding factors such as biomechanics, footwear, psychosocial variables, hormonal factors, sleep, and recovery were rarely measured or controlled for
• Wide variation in follow-up periods (1 month to 60 months) and runner classifications makes cross-study comparisons unreliable

Why it matters

For patients

Runners cannot rely on simple rules like the '10% rule' to prevent injury, and should work with a health professional to account for their individual fitness, recovery, and lifestyle factors.

For clinicians

No single training parameter reliably predicts running injury risk, so prescribing or restricting training based on distance, duration, or intensity thresholds alone is not evidence-based; individualized, multifactorial assessment is warranted.

For readers

This review exposes a persistent gap in running injury research: without standardized definitions and objective load monitoring, the field cannot yet deliver clear training guidelines.