REF-1308 · Bradbury et al., 2020 · Journal of Strength and Conditioning Research · Paper · PubMed

Should I build volume before intensity for a 5 km?

In 30 recreational runners on 3 sessions a week, both orders improved 5000 m (−77 s vs −113 s) with no detectable difference; 10 per group. B.

Main caveat: Small and not clearly randomised: 10 per group, allocated on pre-test results; randomisation, blinding of testers and dropouts are not reported in the abstract.
limited evidenceMeasured on recreational people

Bradbury and colleagues compared two orders of the same training in recreational runners. Thirty runners were assigned on their pre-intervention test results to a linear periodization group (high-volume block first, then a higher-intensity, lower-volume block), a reverse linear periodization group (the intensity block first) or a control group, ten in each.

The training groups did three sessions a week (two supervised, one unsupervised) in two 6-week blocks, with total volume and intensity equated between them. Body mass, skinfolds, treadmill VO2max and running economy, and a 5000-m time trial on a 400-m grass track were tested at weeks 0, 7 and 14.

The 5000-m time improved by 76.8 ± 55.8 s with volume first and 112.8 ± 83.4 s with intensity first, against 3.6 ± 59 s in the control group; the difference between the two orders was not significant (p = 0.321, d = 0.51). VO2peak and running economy at 9 and 11 km/h improved over time with no group differences.

The authors conclude that neither order is superior, but that periodized training improved performance more than non-periodized training. Our reading: no evidence of a difference between the orders, from a trial too small to detect one — the 36-s gap in favour of intensity first has a 95% CI of roughly −31 to +103 s (our arithmetic from the abstract's means and SDs).

It is the one independent endurance trial with volume and intensity equated inside REF-948.

Protocol

Who: n = 30 recreational runners (abstract: 25.2 ± 7.4 years, 175.4 ± 8.1 cm, 69.0 ± 9.8 kg). Sex split, training history and baseline performance are not in the abstract.
Allocation: three groups of 10 "based on preintervention test results" (abstract); randomisation is not mentioned. REF-948's table lists 11 traditional, 11 reverse and 13 control runners (35).
Training: linear (LPG): a 6-week high-volume block, then a 6-week higher-intensity, lower-volume block, with the midpoint tests at week 7 between them; reverse (RPG): the same blocks in the opposite order; 3 sessions a week, 2 supervised and 1 unsupervised; volume and intensity equated between LPG and RPG. Session content, zones and weekly volume are not in the abstract.
Control (CG): not described in the abstract; the conclusion calls its training "nonperiodized".
Tests: weeks 0, 7 and 14 — body mass and sum of 8 skinfolds; treadmill VO2max and running economy at 9 and 11 km/h; 5000-m time trial on a 400-m grass track.
Funding and interests: not given in the abstract (full text not read); the first affiliation is the University of Western Australia.

Key Results

OutcomeLinear (volume first), n = 10Reverse (intensity first), n = 10Control, n = 10Between orders
5000-m time trial, improvement week 0 → 14 (mean ± SD of change)76.8 ± 55.8 s (p = 0.009, d = 1.27)112.8 ± 83.4 s (p = 0.002, d = 1.51)3.6 ± 59 sp = 0.321, d = 0.51
VO2peakimproved over time (p = 0.010)no group difference (p = 0.955)
Running economy at 9 km/himproved over time (p < 0.001)no group difference (p = 0.329)
Running economy at 11 km/himproved over time (p = 0.004)no group difference (p = 0.558)
What the p values next to each training group test: not stated, but the d values reproduce the difference from control with the pooled SD of change (73.2 s, d = 1.27; 109.2 s, d = 1.51; our arithmetic), so they are most likely comparisons with control, not within-group changes.
The order comparison in numbers (our arithmetic from the abstract's means and SDs): a 36-s larger improvement with intensity first, 95% CI about −31 to +103 s (Welch approximation); a trial of 10 per group had about a one-in-five chance (18%) of detecting a true effect of d = 0.5 at p < 0.05.
Individual response: the SD of change was 55.8 s with volume first, 83.4 s with intensity first and 59 s in the control group. The volume-first spread does not exceed the control's, so it is consistent with measurement noise and day-to-day variation rather than with responders and non-responders; the intensity-first spread is larger, but with 10 runners per group that is no evidence of a wider true response.
Count mismatch: 30 runners in the abstract against 35 (11/11/13) in the table of REF-948; perhaps enrolled vs analysed, not checkable without the full text.

Mechanisms

Measured: VO2peak and running economy improved across the training groups without a group difference, so the 5000-m gains are not attributable to one order through either.
Authors' interpretation: neither linear nor reverse linear periodization is superior; planned training structure elicited greater improvements than non-periodized training.
Inference: the control group's training is not described in the abstract, so the trial cannot separate "periodized" from "supervised, more, or structured" training; the conclusion about periodization per se is stronger than the design.
Inference: with equal total load, the two orders delivered the same sessions, only shuffled; that both ended in a similar place fits the base's null on sequencing at equal load (SYNTH-90). Whether either order changes what an unstable athlete can tolerate and keep doing was not the question here (see REF-1307, REF-1290).

Practical Application

Should I build volume before intensity for a 5 km? In 30 recreational runners on 3 sessions a week, both orders improved 5000 m (−77 s vs −113 s) with no detectable difference; 10 per group. B.
For a recreational runner with a steady three-session week and a short goal (5-10 km), either order is defensible (B, one small trial; REF-948, B+, found no evidence of a difference on endurance tests, but it includes this trial, so it is not independent support): choose the order the athlete will execute and progress each block. Not from this paper; the base's own practice: the default sequence and block lengths are in SYNTH-90.
Not tested for long events: for a marathon the base's evidence points to volume and the long run as the gap (SYNTH-64; REF-1314 for a volume floor and injury) — this trial does not speak to it.
Not tested in an unstable base: the runners could already complete the programme; for true beginners and returners the order question is secondary to regularity and graded load (SYNTH-90 row "Tier 1-2, or a patchy log"; SYNTH-75).
Do not promise a bigger gain from intensity first: the 36-s difference is within noise for this sample (95% CI about −31 to +103 s, our arithmetic).

Limitations

Small and not clearly randomised: 10 per group, allocated on pre-test results; randomisation, blinding of testers and dropouts are not reported in the abstract.
Read from the abstract: session content, weekly volume, intensity zones, the control group's training, sex split, baseline times, adherence and funding were not available.
A count that does not match: 30 in the abstract vs 35 (11/11/13) in the review that cites it (REF-948).
Short and specific: 12 weeks, a 5000-m time trial on grass, recreational runners already able to do the sessions; no race, no injury or adherence outcome.
Funding and interests: not in the abstract.
How this entry was checked and changed
Why this level

(from the abstract) — a parallel trial of 30 sits in the rubric's B+ band by size; held at B because the order comparison rests on 10 vs 10 runners, the groups were "assigned … based on preintervention test results" (randomisation is not stated), only the abstract was read (the full text is not open), and the review that cites it counts 11, 11 and 13 runners (REF-948), which the abstract's 10 per group does not match