REF-1291 · Lenk et al., 2025 · Physiological Reports · Paper · PubMed · Free full text

Is a third weekly interval run worth it over two?

In 26 active young adults (14 women) doing 4×4 min for 6 weeks, VO2max rose ~1% at one a week, ~8% at two, ~6% at three; the between-arm test was not significant. B-.

Main caveat: No frequency effect shown. Interactions were not significant for any outcome, and every within-group effect-size interval included zero.
limited, mixed evidenceMeasured on recreational people

Lenk et al. randomised recreationally active adults to one, two or three 4×4-min treadmill interval sessions a week for 6 weeks; 26 of 44 recruited completed (14 women; the rest withdrew for time constraints or illness such as colds or flu).

Intensity started at 85-90% of maximal treadmill speed and HRmax and rose by 0.2-0.5 km/h steps to about 95% by the last session, with 3 min at 4 km/h between reps. Attendance was 100%, 96% and 91% of the 6, 12 and 18 planned sessions.

VO2max went from 53.0 to 53.6 mL/kg/min at one session a week (p = 0.60, d 0.09), from 53.5 to 57.6 at two (p = 0.02, d 0.48) and from 48.5 to 51.2 at three (p = 0.01, d 0.61); time to exhaustion from 598 to 601 s, 601 to 667 s and 575 to 626 s.

The main effect of time was significant for both outcomes, but the group-by-time interaction was not (VO2max p = 0.167, time to exhaustion p = 0.080), and every within-group 95% CI for d included zero. Submaximal oxygen uptake, VT1, RER, heart rate and lactate at an easy speed did not change.

The authors conclude that two or three sessions a week may improve VO2max and time to exhaustion with no clear extra benefit from the third, and call the findings preliminary.

Our reading: the same direction as REF-1288 (one a week ≈ nothing in 6 weeks, three ≈ two) in a mixed-sex running sample, but with arms of about nine and every interval crossing zero it is a hint, not a measurement.

Protocol

Design: exploratory randomised trial, three parallel arms (one, two or three 4×4 sessions a week), 6 weeks; randomisation method and per-arm numbers not reported; no a priori power calculation (the authors say so); no blinding reported. Ethics: University of Würzburg (EV2024/4-1509).
Participants: 44 recreationally active people recruited; n = 26 completed (59.1%), 14 women; age 23 ± 3 y; VO2max 51.3 ± 7.1 mL/kg/min; BMI 24.0 ± 2.3. Withdrawals for time constraints or illness (colds, flu); how many from each arm is not reported.
Session: 4 × 4 min treadmill running; first two sessions at 85-90% of maximal treadmill speed (Vmax) and 85-90% HRmax, then +0.2-0.5 km/h steps to about 95% Vmax and HRmax by the final session; 3-min active recoveries at 4 km/h; Polar H10 at every session.
Tests (before and after): incremental treadmill test (5 min steady at 7 km/h women, 8 km/h men, then +1 km/h a minute to 16/17 km/h, then +1% incline a minute to exhaustion) with a verification bout at 110% of maximal speed; gas exchange (Cortex Metamax 3B); submaximal VO2 (running economy), HR, RER and lactate at 7/8 km/h; VT1.
Analysis: two-way mixed ANOVA (time × group); Bonferroni post hoc tests after a time effect; between-group post hoc comparisons described as descriptive only; paired t-tests and Cohen's d with 95% CI per group (the paper labels 0.20 trivial, 0.50 moderate, 0.80 large).
Funding and interests: "This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors"; no competing interests declared.

Key Results

Outcome, pre → post (mean ± SD)1 a week2 a week3 a weekGroup × time
VO2max, mL/kg/min53.0 ± 7.0 → 53.6 ± 6.7 (p 0.60; d 0.09, −0.68 to 0.86)53.5 ± 8.1 → 57.6 ± 9.2 (p 0.02; d 0.48, −0.28 to 1.24)48.5 ± 5.1 → 51.2 ± 3.6 (p 0.01; d 0.61, −0.41 to 1.63)p = 0.167 (time p = 0.004)
VO2max change (our arithmetic)+0.6 (+1.1%)+4.1 (+7.7%)+2.7 (+5.6%)—
Time to exhaustion, s598 ± 115 → 601 ± 132 (p 0.92; d 0.02)601 ± 136 → 667 ± 133 (p 0.00; d 0.49, −0.27 to 1.25)575 ± 72 → 626 ± 59 (p 0.02; d 0.77, −0.30 to 1.84)p = 0.080 (time p = 0.002)
Submaximal VO2, mL/kg/min29.1 → 28.130.5 → 31.025.9 → 26.0p = 0.237
Submaximal HR, beats/min139 → 135145 → 139144 → 140p = 0.848
Sessions attended100% of 696% of 1291% of 18—
Between arms (the comparison that tests frequency): no interaction for any outcome (VO2max 0.167, time to exhaustion 0.080, submaximal VO2 0.237, VT1 0.376, RER 0.694, HR 0.848, lactate 0.218); the authors present group-specific changes "descriptively".
Within arms: for VO2max and time to exhaustion the paired tests were significant at two and three a week and not at one a week; "all intervals included zero" for the effect sizes (the authors' words).
Baseline imbalance: the three-a-week arm started lower on VO2max (48.5 vs 53.0-53.5).
Individual response: individual data and SDs of change were not reported; the Discussion notes "some participants exhibiting limited or no response", which cannot be separated from test noise without a control arm.

Mechanisms

Measured: no change in submaximal oxygen cost, VT1, RER or lactate at an easy speed at any frequency; any gain sat in VO2max and the incremental test.
Authors' interpretation: central (stroke volume) and peripheral (mitochondrial) adaptations explain the VO2max gain; 6 weeks of training at speeds far above the test's easy speed was too short and too unspecific to change economy; the lower attendance at three a week reflects the practical difficulty of sustaining that frequency.
Inference: in this test the time-to-exhaustion change mostly mirrors VO2max, so the two outcomes are not independent evidence.

Practical Application

Is a third weekly interval run worth it over two? In 26 active young adults (14 women) doing 4×4 min for 6 weeks, VO2max rose ~1% at one a week, ~8% at two, ~6% at three; the between-arm test was not significant. B-.
Read it with [REF-1288], not alone: two small 6-week trials of the same session, in men and in a mixed-sex running sample, point the same way — one a week barely moved VO2max and three showed no extra gain over two. Neither showed a significant difference between frequencies.Together they are consistent with SYNTH-26's table (usually two quality sessions on four to seven sessions a week) and its warning not to add a third hard session to "catch up" as a sensible default in active adults (B, concordant small trials, means only).
This answers [SYNTH-47]'s gap only partly: it is the nearest running test of how many quality sessions an amateur needs, but over 6 weeks in young, active, non-competitive people, with VO2max as the outcome.
Not from this paper; the base's own practice: SYNTH-47 caps quality at one session for runners on fewer than three runs a week; on three runs its default is one quality day + the long run + one easy run, 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).

Limitations

No frequency effect shown. Interactions were not significant for any outcome, and every within-group effect-size interval included zero.
Losses and missing design details. 18 of 44 (41%) did not complete, with no intention-to-treat analysis; the paper does not report the randomisation method, per-arm numbers at the start or end, or which arms lost participants.
No power calculation; small arms. The authors call the study exploratory and the estimates imprecise.
Uncontrolled background. Variations in participants' background activity were not controlled (the authors' limitation); the paper does not report their usual running volume.
Baseline imbalance. The three-a-week arm started about 5 mL/kg/min lower on VO2max.
Reporting. The abstract says VO2max and time to exhaustion improved significantly "across all training frequencies (p < 0.05)" and calls effect sizes of 0.09-0.61 "moderate-to-large"; in Table 1 the once-weekly arm's own tests were p = 0.60 (VO2max) and p = 0.92 (time to exhaustion), and the paper's own scale calls d below 0.20 trivial.The Introduction cites the Declaration of Helsinki (World Medical Association, 2013) for a claim about two vs three sessions a week, a wrong reference.
Funding and interests: no specific grant; no competing interests declared.
How this entry was checked and changed
Why this level

n = 26 completers of 44 recruited, read from the full text (open access, CC BY). Three arms of roughly nine make each comparison a small parallel trial (B); held at B- for a serious limitation beyond the design: 41% lost with no intention-to-treat analysis, group sizes and the randomisation method not reported, no power calculation, background training not controlled, and an abstract that overstates the per-group results

Full text (checked 2026-10-27)
26 participants (14 female; age: 23 ± 3 years; VO2max: 51.3 ± 7.1 mL·kg−1·min−1) out of 44 initially recruited completed the intervention (59.1%) (full text)
VO2max: 53.0 ± 7.0 to 53.6 ± 6.7 (once weekly; p 0.60; d 0.09; −0.68 to 0.86), 53.5 ± 8.1 to 57.6 ± 9.2 (twice; d 0.48), 48.5 ± 5.1 to 51.2 ± 3.6 (thrice; d 0.61); p (pre/post × group) 0.167 (full text)
Time-to-exhaustion: 598 ± 115 to 601 ± 132, 601 ± 136 to 667 ± 133, 575 ± 72 to 626 ± 59 s; p (pre/post × group) 0.080 (full text)
Out of the targeted 6, 12, and 18 sessions, adherence rates of 100%, 96%, and 91% (full text)
All confidence intervals included zero; no a priori power calculation was performed (full text)
This research did not receive any specific grant from funding agencies (full text)