REF-1375 · Wen et al., 2019 · Journal of Science and Medicine in Sport · Paper · PubMed

Do long intervals raise VO2max more than short ones?

In 53 RCTs of healthy, overweight and athletic adults, trials with ≥2-min reps gained more than ≤30-s ones (subgroups; no head-to-head shown); short reps still worked. Abstract, B+

Main caveat: Read from the abstract only; the included trials, their protocols, heterogeneity, risk of bias and the subgroup tests could not be checked.
moderate evidenceMeasured on non-athletes to highly trained athletes

Wen and colleagues pooled 53 randomised controlled trials, searched to April 2018, in which healthy, overweight or obese, or athletic adults did HIIT and were compared with a control group or with moderate-intensity continuous training (MICT), with VO2max or VO2peak measured before and after.

They classified HIIT by work-interval length (short, 30 s or less; long, 2 min or more), session volume (low, 5 min or less of work; high, 15 min or more) and programme length (short-term, 4 weeks or less; moderate to long-term, 4-12 weeks or more). The VO2max gain differed by population (SMD 0.41-1.81).

Against controls, even short-interval, low-volume and short-term HIIT raised VO2max clearly (SMD 0.79-1.65), but long-interval, high-volume and longer programmes "displayed significantly larger effects" (SMD 0.50-2.48). Against MICT, only the long-interval, high-volume and longer programmes showed a benefit (SMD 0.65-1.07).

The authors conclude that short, low-volume and short-term HIIT are "effective and time-efficient strategies for developing VO2max, especially for the general population", and that "to maximize the training effects on VO2max, long-interval (≥2min), high-volume (≥15min) and moderate to long-term (≥4-12weeks) HIIT are recommended." Our reading: in mixed adult populations, trials using long reps and more work gained more VO2max than trials using short reps; the abstract does not separate interval length from volume, does not say how short intervals were done (sprints or 30/15-type sets), and gives only overlapping ranges, so it cannot speak to 30/15 against long reps in trained cyclists.

Protocol

Design: systematic review and meta-analysis; PubMed, Scopus, Medline and Web of Science searched up to April 2018.
Inclusion (abstract): randomised controlled trials; healthy, overweight or obese, or athletic adults; VO2max or VO2peak measured before and after training; HIIT compared with a control group or with MICT.
n: n = 53 studies (participant numbers not in the abstract).
Categories (abstract): work intervals short (≤30 s) or long (≥2 min); session volume low (≤5 min) or high (≥15 min); programme short-term (≤4 weeks) or moderate to long-term (≥4-12 weeks). An intermediate interval category, if any, is not named in the abstract (REF-1363 adopted a >30 s to <2 min category "after Wen et al. 2019").
What was matched: not stated. Long-interval and short-interval effects are subgroups of HIIT arms, each against its own control or MICT arm, and so (our inference) come mostly from different trials; interval length, volume and programme length are separate classifications, and the abstract does not say whether they were analysed together.
Not in the abstract: participant numbers, the number of trials per subgroup, confidence intervals, heterogeneity, the test of subgroup differences, which SMD belongs to which population, whether short intervals were sprints or submaximal short-short sets, work intensities, risk of bias, registration.
Funding and interests: Crossref and Europe PMC list the Key Laboratory of Physical Motor Function Assessment, General Administration of Sport of China (ZGD1607235) — a state sports administration body; conflicts of interest are not in the abstract.

Key Results

Comparison (abstract)SMDNote
HIIT effect by population0.41-1.81, p < 0.05which population is which is not given
Against control: short intervals (≤30 s), low volume (≤5 min), short-term (≤4 weeks)0.79-1.65, p < 0.05"clear beneficial effects"
Against control: long intervals (≥2 min), high volume (≥15 min), moderate to long-term (≥4-12 weeks)0.50-2.48, p < 0.05"significantly larger effects"
Against MICT: long intervals, high volume, moderate to long-term0.65-1.07, p < 0.05"only" these showed a benefit
What "favours long intervals" rests on: subgroup comparisons (each HIIT arm against its own control or MICT arm, so mostly across trials, our inference), reported as ranges that overlap (0.79-1.65 for the short, low-volume, short-term subgroups; 0.50-2.48 for the long, high-volume, longer subgroups). No within-trial comparison of long and short intervals is reported in the abstract.
Authors' conclusion: short, low-volume and short-term HIIT "represent effective and time-efficient strategies for developing VO2max, especially for the general population. To maximize the training effects on VO2max, long-interval (≥2min), high-volume (≥15min) and moderate to long-term (≥4-12weeks) HIIT are recommended."

Mechanisms

Measured: larger pooled VO2max gains in trials with long intervals, more work per session and longer programmes than in trials with short intervals, little work and short programmes.
Authors' interpretation: not given in the abstract beyond the recommendation.
Inference: interval length and work per session go together in practice (a few 30-s sprints add up to a few minutes, 4 × 4 min to 16), so the long-interval subgroups may be showing the dose of work as much as the length of the rep; the abstract gives no analysis that separates them.
Inference: short-interval HIIT in general-population trials is often sprint work with long recoveries; a 30/15-type set at sub-sprint power with 15-s recoveries keeps VO2 high in trained cyclists (REF-1363) and is a different session, whose place in this review's categories cannot be seen from the abstract.

Practical Application

Do long intervals raise VO2max more than short ones? In 53 RCTs of healthy, overweight and athletic adults, trials with ≥2-min reps gained more than ≤30-s ones (subgroups; no head-to-head shown); short reps still worked. Abstract, B+
More work, more weeks: high-volume (≥15 min of work) and longer (4-12 weeks or more) HIIT gained more than low-volume or short programmes, and only these beat continuous training (subgroups from the abstract; between trials is our inference).
Short HIIT is not wasted: short-interval, low-volume and short-term HIIT still raised VO2max clearly against controls, which the authors call "effective and time-efficient", especially for the general population.
Not from this paper; the base's own practice: for trained cyclists SYNTH-33 keeps 30/15 as the first VO2max format on head-to-head trials (REF-045, REF-1050, B); this review compares long reps with short reps of unstated type in mixed adult populations, in subgroups (across trials, our inference), and does not test 30/15-type sets against long reps head to head.

Limitations

Read from the abstract only; the included trials, their protocols, heterogeneity, risk of bias and the subgroup tests could not be checked.
Our inference, not stated in the abstract: the long-versus-short difference is mostly between trials, and interval length may be confounded with session volume and programme length, which the abstract classifies separately and does not disentangle.
Mixed populations (healthy, overweight or obese, athletic adults); the abstract does not report the long-versus-short contrast for athletes separately.
Overlapping SMD ranges; "significantly larger" cannot be checked from the abstract.
How short intervals were performed (sprints or submaximal short-short sets with brief recoveries) is not in the abstract.
Funding from a state sports administration laboratory per the Crossref record; conflicts of interest not in the abstract.
How this entry was checked and changed
Why this level

a meta-analysis of randomised trials would sit at A, or A- if its trials are small (participant numbers are not in the abstract); held lower because the paper's headline — long beats short — is a subgroup comparison, by our inference largely between trials since each trial needed its own control or MICT arm, which the abstract does not separate from session volume and programme length, and because, read from the abstract, its heterogeneity, risk of bias, registration and subgroup tests cannot be checked