The Association Between Changes in Weekly Running Distance and Running-Related Injury: Preparing for a Half Marathon
Participants
n=261 healthy (non-injured) runners followed prospectively for 14 weeks while preparing for a half marathon. Running data collected daily and objectively (GPS watch or smartphone); injuries via weekly e-mailed questionnaires.
Protocol
Primary exposure: change in weekly running distance (categories compared: <20% vs 20-60% increase)
Outcome: running-related injury (RRI); time-to-event models producing cumulative risk difference
Effect-measure modifier tested: the running schedule followed before the change
Key Findings
56 of 261 runners (21.5%) sustained an RRI during the 14 weeks.
At 21 days: significantly more runners were injured among those who had increased weekly distance by 20-60% than among those who increased by <20% — risk difference 22.6% (95% CI 0.9-44.3%; P = 0.041).
At 56 and 98 days: no significant difference between the increase categories.
No effect-measure modification by running schedule — the prior schedule did not change how much increase a runner tolerated.
Read together with REF-565 (Damsted 2018, "10% rule"): the penalty for a large jump is an early one; over the whole build the categories converge, which is why a single fixed "safe %" is weak as a season-long rule but a 20-60% jump is a real short-term risk.
Practical Application
Prospective cohort, 261 healthy recreational runners in a 14-week half-marathon build: 21.5% got injured; at 21 days, runners who had raised weekly distance by 20-60% had ~23 percentage points more injuries than those who raised it by <20% (95% CI 0.9-44.3), with no difference left at 8 and 14 weeks. A big weekly jump costs you in the next three weeks; the same total build spread more gently does not.
When adding distance, keep week-to-week increases under 20% — that is the comparison the paper tested. The wide CI means the exact size of the penalty is uncertain; its direction is not.
Nothing here was modified by the runner's previous schedule, so "I have a good base" is not, on this evidence, licence for a 20-60% jump.
Population is healthy, non-injured runners building for a race. This is not a return-to-run study: it says nothing about how to rebuild after injury or layoff.
Limitations
Observational cohort in healthy runners (not a post-injury rehab trial); cannot be used for return-to-run timelines, and bone stress injury needs clinician-led, site-specific rules.
The 20-60% category is broad and the 95% CI (0.9-44.3%) is wide; the effect is significant but its size is imprecise.
Injuries were self-reported by weekly questionnaire; recreational half-marathon runners only.
How this entry was checked and changed
Content correction 2026-09-27
Title "Return-to-running / graded loading: rigid percentage rules fail — progress by symptoms" -> the paper's title; the study is on healthy runners in a race build, not return-to-run.
"Larger increases were not consistently associated with injury" -> at 21 days a 20-60% weekly-distance increase produced significantly more injuries than <20% (risk difference 22.6%, 95% CI 0.9-44.3); no difference at 56 and 98 days.
Return-to-run protocol and Silbernagel pain-monitoring rules (walk-run, ≤3-4/10 pain, volume before intensity) -> not in this paper; moved under "Not from this paper".
Added n=261, 56 injured (21.5%), no effect modification by running schedule.