REF-1290 · Ramskov et al., 2018 · BMJ Open Sport & Exercise Medicine · Paper · PubMed · Free full text

Does adding hard running raise injury risk more than adding volume?

In 447 recreational runners on 3 runs a week, progressing hard running injured no more than progressing easy volume over 16 weeks (16% vs 19%); wide CI. A-.

Main caveat: Underpowered and wide. 447 entered the compared phase against a planned 620.
fairly strong evidenceMeasured on recreational people

Run Clever asked whether progressing running intensity injures more runners than progressing running volume. Healthy recreational runners were randomised to an intensity schedule (Sch-I) or a volume schedule (Sch-V); all ran three times a week, in 4-week blocks in which weekly volume rose 23% in the first week and fell 10% in the last.

For the first 8 weeks everyone followed the same plan; then for 16 weeks Sch-I increased the weekly amount of running at or above 88% of estimated VO2max (hard) and Sch-V increased total volume at or below 80% (easy). GPS (phone app or Garmin watch) recorded every run, and the app told runners in real time to speed up or slow down.

The outcome was a running-related injury that reduced training for at least 7 days, confirmed by blinded physiotherapists. Of 839 randomised, 392 were censored during preconditioning (47% and 46%), so 447 entered the compared phase.

Over 16 weeks 36 of 221 (16%) in Sch-I and 44 of 226 (19%) in Sch-V were injured; the risk difference was +2.9 points at week 2, +1.8 at week 4, −4.7 at week 8 and −14.0 at week 16 (95% CI −36.9 to 8.9), with no time point significant.

The companion paper looked at specific injuries (Achilles tendinopathy, calf injuries and plantar fasciitis as "intensity" injuries; patellofemoral pain, iliotibial band syndrome and patellar tendinopathy as "volume" injuries) and found no difference either.

The authors caution that the hard running rose by only 0.8 km a week in total, and that more Sch-I runners were injured early, when their 3 km a week of moderate running became 3 km of hard running — a 100% jump in intensity.

Our reading: the only randomised trial of intensity vs volume progression and running injury; it does not show that progressing a small amount of hard running in a three-run week is riskier than progressing easy volume, but its interval is too wide to show equivalence, and it says nothing about two hard sessions in a short week.

Protocol

Design: randomised parallel-group trial, 24-week follow-up: 8-week preconditioning (same schedule for all) then 16-week specific-focus period. Block randomisation in blocks of 10 by a secure back-end system at inclusion; physiotherapists who examined injuries were blinded. Registered at ClinicalTrials.gov on 23 January 2015 (NCT02349373); protocol submitted to BMC Musculoskeletal Disorders; run April 2015 to March 2016. Ethics: Northern Denmark Region (N-20140069).
Participants: 1535 answered the eligibility questionnaire; 839 included and randomised (419 Sch-I, 420 Sch-V); 447 entered the compared phase (221 Sch-I, 226 Sch-V). Eligible: 18-65 y, 1-3 runs a week in the past 6 months, a smartphone; excluded: injured in the past 6 months, pregnant, vigorous activity contraindicated.Compared-phase sample: 62% women, age 39.6 ± 10.3 and 39.4 ± 9.6 y, BMI 24.0 and 24.3, previous injury 53% (timing not given; injury in the past 6 months excluded), running experience median 7 (IQR 3-15) and 6 (3-12) y.
Schedules: three runs a week throughout; 4-week blocks (+23% volume in week 1, −10% in week 4). Sch-I: progression of weekly running at an intensity at or above 88% of estimated VO2max. Sch-V: progression of total weekly running at or below 80% of estimated VO2max. Session details are in the protocol paper, not in this report.
Intensity prescription: VO2max estimated by Daniels' algorithm from a self-administered run (1500 m, Cooper test or 5000 m) through the app at baseline and every 8 weeks; pace guidance in real time by vibration and audio.
Outcome: time-loss running-related injury (a reduction in distance, intensity or frequency for at least 7 days), weekly email questionnaire, phone check, then physiotherapist examination.
Analysis: primary instrumental-variable analysis (randomisation as the instrument; adherence < 80% counted as non-compliant), ITT and per-protocol secondary (in the supplement, not read); cumulative risk difference by the pseudo-observation method at 2, 4, 8 and 16 weeks, also by days, kilometres and minutes; non-running injuries as competing risks; censoring for pregnancy, illness, accidents, lack of motivation or more than 10% manually uploaded running.Planned sample 620 (80% power to detect a 15-point difference) plus 15%.
Funding and interests: "DR was funded by Aarhus University, the Graduate School of Health. The Danish Rheumatism Association supported the study with a grant (grant number A3419)." Competing interests: none declared. Help2Run (the app provider) is thanked for assistance; no funding from device makers is stated.

Key Results

Risk difference, Sch-I vs Sch-V (reference)Week 2Week 4Week 8Week 16
Reference cumulative risk (Sch-V)4.6%8.2%13.2%28.0%
Risk difference (95% CI), points2.9 (−5.7 to 11.6)1.8 (−9.1 to 12.8)−4.7 (−17.5 to 8.1)−14.0 (−36.9 to 8.9)
p0.510.740.470.23
Counts: 80 injured in the compared phase — 36 of 221 (16%) Sch-I and 44 of 226 (19%) Sch-V.
Other time scales (no difference anywhere): day 112: −8.9 (−26.3 to 8.5); 400 km: −3.5 (−23.8 to 16.7); 3000 min: −6.1 (−41.6 to 29.5).
Adherence: preconditioning 87 ± 16% (Sch-I) and 89 ± 19% (Sch-V) of scheduled sessions; compared phase 75 ± 23% and 76 ± 22% of 48 scheduled runs.
Dose actually progressed: the weekly amount of hard running rose by 0.8 km in total over the 16 weeks in Sch-I; weekly volume rose by 6 km in total in Sch-V. At week 16 the cumulative injury proportion was 12.1-18.8% in Sch-I and 25.5-30.6% in Sch-V across analyses.
The switch: entering the compared phase, Sch-I runners turned about 3 km a week of moderate running into 3 km of hard running (a 100% jump in intensity), and more Sch-I than Sch-V runners were injured in the first weeks (week-2 risk difference +2.9 points, not significant).
Companion paper (JOSPT, from the abstract): "intensity" injuries in Sch-I, risk difference at 16 weeks −5.1% (−16.5% to 6.3%); "volume" injuries in Sch-V −3.4% (printed as "(13.2%, 6.2%)", the lower bound's minus sign missing); no difference at 2, 4 or 8 weeks either.
Loss to follow-up: 198 of 419 (47%) Sch-I and 194 of 420 (46%) Sch-V censored during preconditioning; 173 eligible volunteers declined because of the rigid schedule or GPS recording.

Mechanisms

Measured: injuries were no more frequent with intensity progression than with volume progression; the only numerical excess in the intensity arm was in the first weeks after the switch from moderate to hard pace.
Authors' interpretation: because progression was relative to the previous week, the intensity arm's absolute increase in hard running was small and left more recovery between sessions; a sudden change (3 km moderate to 3 km hard) is the likelier cause of the early excess. Changing the structure (number of sessions, weekly frequency, absolute rather than relative progression) would probably change the result.
Inference: hard running loads tendons and the plantar fascia more per kilometre, but at 3-4 km a week of it the cumulative load in Sch-I stayed small; the trial cannot say what happens at 6-10 km of hard running a week or with two hard sessions in three runs.

Practical Application

Does adding hard running raise injury risk more than adding volume? In 447 recreational runners on 3 runs a week, progressing hard running injured no more than progressing easy volume over 16 weeks (16% vs 19%); wide CI. A-.
Progress the hard running gradually and avoid a jump at the start (A-, authors' interpretation of a non-significant early excess): the intensity arm's early injuries followed a switch of the same 3 km from moderate to hard pace in one step. Inference, not a guide line: bring hard running in as a small share and grow it over weeks rather than switching a whole block of running from moderate to hard pace at once.
The spike the base has measured is a single long run (REF-1102, B+: a single run more than 10% longer than the longest of the past 30 days raised overuse injury); that cohort did not examine spikes in intensity, so it does not clear the hard session. Run Clever adds that a small, progressed dose of hard running inside three runs a week did not show up as extra risk.
What it does not license: two hard sessions in a three- or four-run week, or more than a few kilometres of hard running a week, were not tested.Not from this paper; the base's own practice: SYNTH-47 allows one quality session at most below three runs a week; for recreational runners on three runs its default is one quality day + the long run + one easy run, and on 4-5 runs both quality days if recovery holds, the count following SYNTH-26's table (a recovery-budget judgement, not a measured optimum).
Injury is common either way: about one in six to one in five runners had a time-loss injury in 16 weeks on a structured, app-guided plan with a 28.0% reference cumulative risk by week 16.

Limitations

Underpowered and wide. 447 entered the compared phase against a planned 620; the week-16 interval runs from a 36.9-point benefit to an 8.9-point harm, so equivalence is not shown.
Loss and randomisation timing. 47% were censored during the 8-week preconditioning; randomisation happened before preconditioning, not after it as the protocol stated, which the authors say probably let imbalance in unmeasured factors into the compared groups.
Small, relative progression. Hard running grew by 0.8 km a week over 16 weeks; the result does not extend to bigger absolute increases (the authors' own caution).
Estimated intensity. VO2max and the 88% / 80% bands came from a self-administered field test and Daniels' equations, not a lab test; adherence in the compared phase was about 75%, and the instrumental-variable primary analysis counts non-compliant Sch-I runners with the reference group.
Session structure not reported here. How the hard running was spread across the three weekly runs is in the protocol paper, which was not read; "one hard session a week" is not a description this report supports.
Companion paper read as an abstract only; its printed interval for volume injuries at 16 weeks lacks a minus sign.
Who. Recreational runners, 62% women, about 39 years old, not injured in the past 6 months, willing to follow a rigid app-guided schedule (173 declined for that reason).
Funding and interests: Aarhus University Graduate School of Health (salary of the first author) and the Danish Rheumatism Association (grant A3419); no competing interests declared.
How this entry was checked and changed
Why this level

large registered RCT, n = 447 in the compared phase of 839 randomised, read from the full text (open access, CC BY-NC); the companion JOSPT paper from the abstract. By size the rubric would allow A; held at A- because the trial fell short of its planned 620, 47% were censored during preconditioning, randomisation came before that loss (not after it, as the protocol stated), and the week-16 confidence interval spans a 37-point benefit to a 9-point harm

Full text (checked 2026-10-27)
In the first eight weeks all participants (n=839) followed the same running schedule; participants (n=447) not censored entered the 16-week training period (full text)
80 participants sustained an RRI: 36 of 221 runners (16%) in Sch-I and 44 of 226 (19%) in Sch-V (full text)
Risk difference at week 16: reference 28.0%, RD −14.0 (–36.9 to 8.9), P value 0.23 (full text)
During preconditioning, 198 (47% of 419) in Sch-I were censored and 194 (46% of 420) in Sch-V were censored (full text)
Sch-I increased the weekly volume of running at an intensity equal to or above a VO2max of 88%; Sch-V increased total running volume equal to or below a VO2max of 80% (full text)
The total increase in the volume of running intensity during the 16-week specific focus training period was 0.8 km (full text)
A sample size of 620 participants was required; the randomisation was introduced prior to preconditioning and not prior to the 16-week intervention period, as stated in the protocol article (full text)
Funding: Aarhus University, the Graduate School of Health; The Danish Rheumatism Association grant (grant number A3419); Competing interests: None declared (full text)