Which interval type raises VO2max most in elite athletes?
Long aerobic reps beat speed-endurance sprints by 2.6% (between studies); 30/15-type short-short sets were not tested against long reps. Time trial: no clear type effect. A-
Main caveat: The interval-type effect is a linear trend across studies. Type was one number from 1 to 6, its effect mostly a comparison between studies with different…
Wiesinger, Stöggl and colleagues pooled every controlled or uncontrolled HIIT intervention they could find in highly trained endurance athletes and elite athletes of other sports (34 studies, searched February 2021) and estimated, by meta-regression, the effect of HIIT against usual (control) training on seven measures: sprint, repeated-sprint ability, time trial, peak speed or power, aerobic or anaerobic threshold, VO2max and economy.
They sorted HIIT into six types on a scale from most aerobic to most anaerobic — 1 traditional long intervals (2-10 min, submaximal), 2 intermittent short intervals (15-60 s, submaximal), 3 speed-endurance maintenance and 4 speed-endurance production (10-75 s at maximal effort, production with about twice the recovery), 5 repeated sprints and 6 sprint intervals (2-10 s) — and entered the type as one linear number.
For VO2max (22 studies), five weeks of long-interval HIIT replacing part of usual training in season was predicted to add 7.7% (90% CI 4.9 to 10.6) over control in male endurance athletes, 7.4% in female endurance athletes and 8.6-11.5% in other athletes; the higher the starting VO2max, the smaller the gain.
Moving from type 1 to type 4 cost 2.6% (−4.1 to −0.7) of VO2max gain, "very good evidence"; on the same linear scale one step, from long to intermittent short intervals, would be about 0.9% (our arithmetic), under the authors' 1% threshold.
For time-trial power the type made no clear difference (+0.4%, −1.8 to 2.6), and the authors write that it would be "premature to recommend a particular type of HIIT for endurance athletes". Their conclusion for VO2max: athletes who lack VO2max should use "the most aerobic type of HIIT".
Our reading: good evidence that interval training still adds VO2max in trained athletes and that long aerobic intervals beat sprint and speed-endurance formats for VO2max; it does not test, and cannot settle, short-short aerobic sets such as 30/15 against long reps.
Protocol
Design: systematic review with meta-regression, PRISMA, pre-registered (PROSPERO CRD42021236384). Search 18 February 2021 in PubMed, Scopus, SPORTDiscus and Web of Science (title, abstract, keywords); 547 papers passed title and abstract screening, plus 7 from reference lists; 34 included. English or German, peer reviewed.
n: n = 34 studies; for VO2max n = 22 studies with 394 HIIT and 132 control participants.
Who: highly trained endurance athletes with mean baseline VO2max at least 60 (men) or 55 (women) mL/kg/min, or "other elite athletes" (first league, national team or international level).VO2max analysis: 22 studies (15 endurance, 7 other), 34 HIIT and 17 control estimates, 394 HIIT and 132 control participants, 10.3 ± 8.2 per estimate, maleness 0.87 (HIIT) and 0.80 (control); baseline VO2max 62.8 (endurance men, 28 estimates), 65.9 (endurance women, 7), 50.4 (other men, 10) and 44.6 (other women, 6) mL/kg/min.
HIIT eligibility: at least two sessions a week for at least one week and at least five sessions; intensity at least 90% HRmax, at least 90% VO2max, or all-out; no supplements, altered environment or garments. Studies whose training and testing were unrelated to the sport were excluded.
Interval types (Figure 1, after Stöggl et al. 2024): 1 aerobic traditional long intervals, 2-10 min; 2 aerobic intermittent short intervals, 15-60 s; 3 speed-endurance maintenance; 4 speed-endurance production (speed endurance 10-75 s, production with about twice the recovery of maintenance); 5 repeated-sprint intervals; 6 sprint intervals (2-10 s).Coded 1-6 as a numeric linear moderator, a progressive programme as the average of two adjacent codes (e.g. 1.5). In the VO2max analysis the HIIT arms averaged type 2.1 ± 1.4 (range 1-5).
How the types were compared: not matched for total work, work intensity, session duration or energy. Each HIIT arm entered with its type code; the type effect is the slope across arms and studies, reported as the predicted difference between type 4 and type 1. Within-study contrasts of types existed in a few studies only (e.g. Laursen 2002, types 1 vs 4; Clark 2014, types 4 vs 5); none contrasted type 2 with type 1 for VO2max.
Other moderators (VO2max model): type of athlete, maleness, intervention duration (5.7 ± 3.1 weeks, range 0.86-12; 14.4 ± 6.2 sessions), type of HIIT, log of baseline VO2max, HIIT added to or replacing usual training, and phase of season. All except extra training and type interacted with treatment.
Analysis: SAS Proc Mixed; dependent variable 100 × ln(1 + percent change/100), weighted by inverse squared standard errors; random effects for study and estimate within study, separate for HIIT and control; missing standard errors imputed from typical errors of measurement (7 HIIT and 6 control estimates for VO2max); 90% CIs; Bayesian probabilities of substantial or trivial effects with a minimally informative prior; magnitude thresholds 1.0, 3.0, 5.5, 8.6 and 14% (from competition variability of top athletes).Publication bias examined by plots and simulation.
Quality: Galna-based scale, 9.7 ± 0.9 (range 6.7-10.9) out of 13; no study excluded on quality.
Funding and interests (as printed): no financial support; "Authors TS and AS were employed by the Red Bull Athlete Performance Center" (in the supplementary appendix Thomas Leonard Stöggl and Anna Schmuttermair carry that affiliation). Our observation: review authors wrote included studies (Stöggl and Sperlich 2014, REF-303; Stepto et al. 1999 with Hopkins, REF-1002).
Key Results
| Outcome (setting of the prediction) | HIIT minus control, % (90% CI) | Type of HIIT, 4 vs 1, % (90% CI) | Other moderators |
|---|---|---|---|
| VO2max (5 weeks of long-interval HIIT, on season, replacing usual training) | endurance men 7.7 (4.9 to 10.6); endurance women 7.4 (0.2 to 15.1); other men 8.6 (5.8 to 11.5); other women 11.5 (7.4 to 15.8) | −2.6 (−4.1 to −0.7), "very likely" smaller with more anaerobic types | baseline VO2max +20%: −1.8 (HIIT arm), −1.4 (−3.3 to 0.6) against control; on season +4.3 (1.4 to 7.2); HIIT added to training +2.3 (−0.4 to 5.2); heterogeneity SD 1.1 |
| Time-trial power (male endurance athletes, intermittent short intervals, type 2) | 25-min test: 2.8 (0.7 to 5.0) at 5 weeks, −0.3 at 9 weeks; 45-min test: 5.3 (2.2 to 8.6) at 5 weeks, 2.1 at 9 weeks | +0.4 (−1.8 to 2.6), trivial, unclear | longer test +4.9 (2.1 to 7.8); control groups "catching up" over longer studies |
| Peak speed or power, incremental (male endurance athletes, speed-endurance maintenance, type 3) | 3.1 (1.1 to 5.1) | −0.3 (−1.5 to 1.0), likely trivial | |
| Aerobic or anaerobic threshold (male endurance athletes, long intervals) | 4.8 (1.3 to 8.3) | −2.7 (−6.9 to 1.8), unclear | women 12.6 (2.6 to 23.5) |
| Economy | −0.3 (−3.9 to 3.3), unclear | not estimable |
What the VO2max type effect is made of (Supplementary Table 8 and Figure 8): 19 of the 34 VO2max HIIT estimates are type 1 (our count; 17 of them in endurance athletes), including both arms of Sandbakk 2013 (REF-1052: its "short" 2-4-min group is type 1 by this scale) and the polarized arms of Stöggl and Sperlich 2014; type 2 has two arms, Stenqvist 2020 (+2.4%, endurance athletes) and Yang 2017 (+6.5%, canoeists); the six-day block of daily 30/15 sessions in elite cyclists, Rønnestad and Vikmoen 2019 (REF-1198, +3.8% in 0.9 weeks), is coded type 3; with it, 13 arms are speed-endurance or sprint training (types 3-5), 7 in endurance athletes and 6 in canoe, badminton and soccer players (our count).
The step from long to intermittent short intervals (our arithmetic): the model's type effect is linear, so −2.6% over three steps is about −0.9% per step (90% CI roughly −1.4 to −0.2%), under the 1% smallest important change. This is a model projection; no included study compared type 2 with type 1 for VO2max.
Authors' wording: "There was very good evidence of a negative effect of more anaerobic types of HIIT on VO2max (−2.6%)"; "greater improvements were seen with longer intervals and lower intensities"; "It would therefore be premature to recommend a particular type of HIIT for endurance athletes"; conclusion: "the strategy for improving athletes who are deficient in VO2max is the most aerobic type of HIIT".
Training status: all participants are highly trained or elite, so the type effect is measured in trained athletes; within that range a higher baseline VO2max went with a smaller gain (−1.8% per +20% of baseline, HIIT arms).
A number that does not add up (our reading of Figure 8): Johansen 2021's HIIT arm is listed as −3.8% (−7.1 to −0.4) in Supplementary Table 8 but plotted near +6% in Figure 8.
Mechanisms
Measured: VO2max gains were larger with more aerobic HIIT types and smaller with higher baseline VO2max; the type made no clear difference to time-trial gains (unclear) and a likely trivial one to peak power.
Authors' interpretation: different HIIT types stress different systems along an aerobic-anaerobic continuum (their Figure 1, after Stöggl et al. 2024); the most aerobic types load the oxygen-transport system most and so move VO2max most, while sprint types act more on anaerobic and neuromuscular qualities. They add that a combination of types may suit endurance athletes (from their Part II analysis of how changes in measures relate).
Inference: sustained submaximal work keeps VO2 near its maximum for longer than all-out sprints with long rests, which fits the acute data (REF-1363); a short-short set with brief recoveries (type 2) also keeps VO2 high, and nothing in this analysis isolates the length of the work bout from its intensity or the length of the recovery.
Inference: the type effect is a slope across studies, so it can carry whatever else differed between them — sport, season, programme length, recovery between reps; 17 of the 19 type-1 arms are endurance athletes against 7 of the 13 type 3-5 arms (our count), and type of athlete entered the model as a single term.
Practical Application
Which interval type raises VO2max most in elite athletes? Long aerobic reps beat speed-endurance sprints by 2.6% (between studies); 30/15-type short-short sets were not tested against long reps. Time trial: no clear type effect. A-
Interval training still pays in trained athletes: about 5 weeks of long-interval HIIT replacing part of usual training was predicted to add about 7-8% VO2max over usual training in highly trained endurance athletes in the competition phase (about 4% less off-season) and less at a higher starting VO2max (A-, model predictions, mostly men).
For a VO2max block, keep the work aerobic: submaximal long reps or short-short aerobic sets rather than all-out sprints with long rests; sprint and speed-endurance formats gave about 2.6% less VO2max gain than long intervals.Between 30/15-type sets and long reps the model's linear projection is about 0.9% less gain for the short-short type (90% CI 0.2 to 1.4; our arithmetic from only two type-2 arms, none compared with long intervals in the same study), trivial to small and an extrapolation, so this paper does not overturn the head-to-head trials that favoured 30/15 in cyclists (REF-045, REF-1050, B).
Do not expect the type to change race performance much: the HIIT type made no clear difference to time-trial power (+0.4%, 90% CI −1.8 to 2.6), and control groups tended to catch up over 9 weeks (mostly unclear); the authors call any single-type recommendation premature.
Not from this paper; the base's own practice: rotate formats in a block (SYNTH-33); this paper's own authors suggest a combination of types for endurance athletes.
Limitations
The interval-type effect is a linear trend across studies. Type was one number from 1 to 6, its effect mostly a comparison between studies with different athletes, sports, durations and seasons; it was clear for VO2max only, unclear for threshold, trivial or unclear for time trial and peak power.
Small studies, borrowed controls. About 10 athletes per estimate; uncontrolled HIIT arms rely on the control arms of other studies; 13 of 51 VO2max standard errors imputed.
Who. Highly trained endurance athletes and elite athletes of other sports, mostly men; women's estimates rest on few studies; "elite" includes national-level and junior athletes.
Search date. Searched February 2021; the authors name four later studies they judge consistent with their results.
Reporting. Johansen 2021's VO2max arm differs between Supplementary Table 8 (−3.8%) and Figure 8 (about +6%, our reading).
Interests. No funding; two authors employed by the Red Bull Athlete Performance Center; review authors wrote two included studies (our observation).
How this entry was checked and changed
Why this level
a systematic review and meta-analysis of controlled trials would sit at A; held one step lower because the studies are small (about 10 athletes per estimate), 10 of the 34 VO2max HIIT estimates come from arms without a control group in their own study (our count), and several moderators rest on a few studies. The interval-type finding is weaker than the level: type entered as a single linear score from long aerobic intervals (1) to sprints (6), its effect is mostly a comparison between studies, and it was clear for VO2max only. Read from the full text and the supplementary appendix (open access, CC BY 4.0)
Full text (checked 2026-10-28)
On 18 February 2021, PubMed, Scopus, SPORTDiscus and Web of Science were scanned; 34 articles met the inclusion criteria; PROSPERO CRD42021236384 (full text)
Traditional long intervals with durations of 2–10 min and intermittent short intervals with durations of 15–60 s; for SET, intervals last between 10 and 75 s; for SIT, sprints last between 2 and 10 s (full text)
The type of HIIT was included as a numeric linear moderator (coded 1–6) (full text)
VO2max, endurance male athletes (HIIT–control): 7.7; 4.9–10.6; endurance female athletes 7.4; 0.2–15.1 (full text)
Type of HIIT (4.0 vs. 1.0) for VO2max: −2.6; −4.1–−0.7; pre-test VO2max +20% HIIT: −1.8; −3.1–−0.5 (full text)
Time-trial speed/power, type of HIIT (4.0 vs. 1.0): 0.4; −1.8–2.6; 45-min test, 5-week duration, HIIT–control 5.3; 2.2–8.6 (full text)
"It would therefore be premature to recommend a particular type of HIIT for endurance athletes" (full text)
Rønnestad et al. [7] 2019, Endurance, Exp, HIIT, type 3, 9 athletes, 76.8 ± 6.9, 3.8; 2.1 to 5.5; Stenqvist et al. [13] 2020, type 2, 2.4; 1.1 to 3.7 (full text)
The study quality score was 9.7 ± 0.9; no financial support was received; Authors TS and AS were employed by the Red Bull Athlete Performance Center (full text)