REF-096 · Sylta et al., 2016 · Medicine & Science in Sports & Exercise, Vol. · Paper

The Effect of Different High-Intensity Periodization Models on Endurance Adaptations

A-fairly strong evidenceMeasured on trained to highly trained athletesWritten before October 2026, being re-checked

Sylta et al. compared three ways of ORDERING the same high-intensity interval sessions across a 12-week programme, with total load balanced and a high volume of low-intensity training underneath.

This is NOT a polarized-vs-threshold training-distribution trial; all three groups did the same 24 interval sessions and the same low-intensity volume, differing only in which interval format came in which 4-week mesocycle.

All groups improved 5-10% in 40-min all-out power, peak power output and VO2peak; no adaptation differences emerged among the groups, and the likelihood of large, moderate or non-response was similar in each.

Participants

n=63 cyclists (VO2peak 61.3 ± 5.8 mL·kg·min), randomized to three groups: increasing HIT (n=23), decreasing HIT (n=20), mixed HIT (n=20).

Protocol

12 weeks, 24 interval sessions per group plus a high volume of low-intensity training and laboratory testing.
Increasing HIT: 4 × 16 min in weeks 1-4, 4 × 8 min in weeks 5-8, 4 × 4 min in weeks 9-12.
Decreasing HIT: the same mesocycles in the opposite order (4 × 4 first, 4 × 16 last).
Mixed HIT: all three interval formats mixed within every mesocycle.
Intervals were prescribed as maximal session efforts; mean blood lactate was 4.7, 9.2 and 12.7 mmol·L in the 4 × 16-, 4 × 8- and 4 × 4-min sessions respectively (P < 0.001).
Tests pre and post: mean power in a 40-min all-out trial, peak power output in an incremental test to exhaustion, VO2peak, power at 4 mmol·L lactate.

Key Results

All groups improved 5-10% in 40-min mean power, peak power output and VO2peak (P < 0.05).
No adaptation differences among the three groups (P > 0.05).
Individual-response analysis: similar likelihood of large, moderate or non-response in each group (P > 0.05).
Session lactate: 4.7 (4 × 16 min), 9.2 (4 × 8 min), 12.7 mmol·L (4 × 4 min).

Practical Application

In 63 trained cyclists, 12 weeks of the same 24 maximal-effort interval sessions produced the same 5-10% gains in 40-min power, peak power and VO2peak whether the interval formats were ordered long-to-short, short-to-long or mixed — the ORDER of interval work within a training block has little or no effect when total load is equal.
Do not expect a specific mesocycle sequence (e.g. "long intervals first, short last") to add adaptation by itself; choose the order that fits the athlete's calendar, motivation and race demands.
Mixing 4 × 16, 4 × 8 and 4 × 4 within each block is as effective as blocking them, so variety within a week is a legitimate option for athletes who find monotonous blocks hard to sustain.
A high volume of low-intensity training sat under every arm; the trial says nothing about replacing that base with more intensity.
Correction to earlier versions of this entry: this trial did NOT compare polarized, intensified and pyramidal distributions and did NOT find polarized superior. For polarized-vs-other distribution evidence use REF-303 (trial) and REF-079 (narrative review).

Limitations

Trained cyclists, 12 weeks; interval sessions prescribed as maximal efforts, so the arms differed in lactate/intensity by format but not in total load; no measure of season-long or race outcomes.
How this entry was checked and changed
Why this level

regraded 2026-09-27 from A: single randomized trial, n=63, three arms of 20-23; no group difference

Content correction 2026-09-27
Entry described a polarized (75/10-15/10-15) vs intensified (50/30/20) vs pyramidal (60/30/10) comparison with polarized winning (+6.2% VO2max, +7.1% TT vs +3.8-4.5%) -> the paper compares increasing, decreasing and mixed ORDER of the same HIT sessions; all groups improved 5-10% and no group differences emerged. Title, summary, key figures and lead rewritten from the abstract.
"Expected 6-7% VO2max / 7-8% TT gains from a 14-week polarized plan" and "empirical evidence definitively favors polarized" -> deleted; not from this paper and contradicted by it.
Level A -> A- (single RCT, n=63, three arms of 20-23). Tier text updated to the abstract's VO2peak; sample sizes per group added.
REF-079 was an empty stub and is now filled: it is a C+ narrative review (Hydren & Cohen 2015), so pointers for polarized-vs-other evidence now lead with the trial REF-303.