REF-1703 · Zhang et al., 2026 · Frontiers in Physiology, · Paper · PubMed

In trained adolescent athletes, mostly in soccer and other team sports, 4-12 weeks of interval training (1-4 sessions a week, 8-32 sessions) raised VO2max moderately against non-interval training (SMD 0.65, 95% CI 0.46-0.83, I² 0%; 16 studies, 522 athletes; GRADE moderate) and improved sprint, change-of-direction, jump and intermittent running tests by small to moderate amounts, in a meta-analysis of 35 controlled studies with 988 athletes, none rated low risk of bias overall

Main caveat: The authors': wide differences in participants, protocols, tests and comparators; comparator-stratified analysis not possible.
fairly strong evidenceMeasured on trained to highly trained athletesRead in full

Zhang and colleagues (Kyiv and Taian) pooled controlled trials that tested interval training in adolescent athletes classified as McKay Tier 2 or above, asking what it does to aerobic fitness, jumping, sprinting, change of direction and repeated sprints, and what modifies the effect.

Five databases searched to February 2026 gave 35 studies with 988 athletes aged on average 11 to 18, 17 of the studies in soccer.

Against groups that did not do interval training, interval programmes of 4-12 weeks raised VO2max by a moderate amount (SMD 0.65) with no heterogeneity, and improved intermittent field tests (Yo-Yo and 20-m shuttle, SMD 0.65, very heterogeneous; 30-15 test final speed, 1.13), countermovement jump (0.44), sprints of 10 m or less (effect −0.79, negative meaning faster), sprints of 20 m or more (−0.28) and change of direction (−0.54).

Every repeated-sprint study used repeated sprint training, and it improved repeated-sprint ability (−0.70). Only two subgroup splits differed: the intermittent field-test gain appeared in athletes 16 or older and not in younger ones, and the longer-sprint gain in Tier 3 and not Tier 2 athletes; the authors call both exploratory.

Risk of bias was "some concerns" in 31 studies and high in 4, mainly from unclear randomisation and unblinded training, and every outcome was downgraded for it in GRADE.

The authors conclude that interval training is a useful part of a broader programme for aerobic fitness, intermittent running and short acceleration, but not a stand-alone tool for jump, longer sprint or change of direction.

Participants

35 studies, 988 trained adolescent athletes; 21 studies male, 10 female, 4 mixed; mean ages about 11.0 to 18.4 years.
Sports: soccer 17 studies; basketball 4; handball 4; taekwondo 3; tennis 2; futsal, karate, cross-country skiing, aerobic gymnastics and volleyball 1 each.
Level (authors' McKay classification): Tier 2 in 19 studies, Tier 3 in 16.
Group sizes in the study table: 5 to 29 per group (one karate study had 5 and 5).

Protocol

Search: PubMed, Scopus, Embase, Web of Science, Cochrane Library to 26 February 2026; English, peer-reviewed; reference lists screened. 3,635 records, 1,281 screened, 74 full texts, 35 included.
Inclusion: trained adolescent athletes (10-19 years, McKay Tier 2 or above); a structured interval programme; a control group doing no interval training (usual sport training, coordination, moderate continuous, strength or other); randomised or non-randomised controlled trials.
Interventions: repeated sprint training 11 studies, short intervals 10, sprint interval training 5, long intervals 4, mixed short/long 4 (one study had two arms); running-based in 24, sport-specific in 9, both in 2; 4-12 weeks, 1-4 sessions a week, 8-32 sessions in all.
Analysis: Stata 15; change scores with a change-score SD imputed at r = 0.5 where not reported; Hedges' g, random effects; I²; subgroups by sex, age, duration, frequency, total sessions, competitive level, interval type and mode (median splits for continuous variables); exploratory univariable meta-regression for jump; leave-one-out; Egger's test with 10 or more studies.
Risk of bias and certainty: RoB 2 by two reviewers; GRADE per outcome.
Funding and interests: no financial support received; no conflicts declared.

Key Results

Outcome (interval training vs non-interval control)Studies, athletesSMD (95% CI)I²GRADE
VO2max16, 5220.65 (0.46 to 0.83)0%moderate
Intermittent field tests (Yo-Yo IR1, IR2, 20-m shuttle)9, 2880.65 (0.07 to 1.23)79.4%very low
30-15 Intermittent Fitness Test final speed5, 1531.13 (0.63 to 1.63)42.5%low
Countermovement jump20, 5430.44 (0.13 to 0.74)62.7%low
Sprint ≤10 m (negative = faster)6, 165−0.79 (−1.26 to −0.32)50.1%low
Sprint ≥20 m18, 423−0.28 (−0.47 to −0.08)0%moderate
Change of direction20, 550−0.54 (−0.72 to −0.37)0%moderate
Repeated-sprint ability (all repeated sprint training)7, 208−0.70 (−1.02 to −0.39)12.7%low
Subgroups: intermittent field tests, age ≥16 SMD 1.07 (0.43 to 1.70) vs <16 −0.09 (−0.95 to 0.76), between-group p = 0.033; sprint ≥20 m, Tier 3 −0.47 (−0.75 to −0.20) vs Tier 2 −0.04 (−0.33 to 0.24), p = 0.034. No other subgroup difference for VO2max, jump, change of direction or repeated sprints, including training frequency, total sessions and interval type (numbers in the supplement, not read).
Sensitivity: leaving out one study at a time did not change direction or significance for most outcomes; dropping one study each cut I² for jump (62.7% to 24.7%), the 30-15 test (42.5% to 8.8%) and sprint ≤10 m (50.1% to 0%).
Publication bias: Egger p = 0.707 (VO2max), 0.194 (jump), 0.785 (sprint ≥20 m), 0.691 (change of direction); not tested for outcomes with fewer than 10 studies.
Risk of bias: overall 31 some concerns, 4 high (Sandbakk 2011, the cross-country skiing study, among them); randomisation 6 low, 25 some concerns, 4 high; deviations from intended interventions some concerns in all 35; missing data low in 34; outcome measurement and reporting low in all.

Practical Application

Interval training works in trained teenagers: across 35 controlled studies it raised VO2max by a moderate amount (SMD 0.65, moderate certainty), made sprints of 10 m or less moderately faster (low certainty) and longer sprints and turns slightly faster, compared with training that had no intervals. The athletes were mostly teenage team-sport players, the programmes ran 4-12 weeks at 1-4 sessions a week, and the review measured no injury, burnout or other harm.
Where it meets [SYNTH-81]: the guide keeps intervals optional and at most weekly for a 12-14-year-old cyclist and allows up to two a week for an older junior in season.This review supports that intervals improve fitness in trained adolescents; it does not show they are needed at 12-14, it found no difference by training frequency or total sessions (exploratory subgroups), and it tested no harms, so it gives no reason to raise the guide's dose.Inference, not tested here: one or two interval sessions a week for an older junior who already trains regularly fits the range these studies used.
Match the session to the target. Repeated-sprint work was the only format tested for repeated-sprint ability; the authors say jump height, longer sprints and change of direction need specific strength, plyometric or sprint work rather than intervals alone.
Age may matter for intermittent running: the field-test gain appeared at 16 and older and not below (an exploratory split of 9 studies across different tests, very low certainty).
Not settled by this design: anything for teenage cyclists or endurance runners (one skiing study is the only endurance sport, rated high risk of bias); which interval format is best; long-term development or results; the effect of maturity rather than age.

Limitations

The authors': wide differences in participants, protocols, tests and comparators; comparator-stratified analysis not possible; few studies for some outcomes and subgroups; repeated-sprint results come only from repeated sprint training; English only; change-score SDs imputed in some studies; weaknesses in randomisation, allocation concealment and blinding; soccer dominates.
Certainty: no study low risk of bias overall; GRADE moderate for three outcomes, low for four and very low for the intermittent field tests, which pool three different tests.
Comparators: the control groups ranged from usual sport training to continuous running and strength training, so the effect is interval training against whatever the controls did, not interval training added on top of an unchanged programme.
Design: randomised and non-randomised trials pooled; group sizes mostly under 20; a control group shared by several arms had its size split, which keeps the same controls in more than one comparison.
Not given: adverse events or injuries; the subgroup and meta-regression numbers beyond the two significant splits (supplement not read).
How this entry was checked and changed
Why this level

risk of bias and certainty read in the full text; held at A- because no study was low risk overall (31 some concerns, 4 high), most trials are small (5 to 29 per group in the study table, mostly under 20), and certainty was moderate for VO2max, sprint over 20 m and change of direction but low or very low for the other outcomes