Power output at the moderate-to-heavy intensity transition decreases in a non-linear fashion during prolonged exercise
In 12 trained cyclists VT1 power had fallen 5% after 139 ± 78 min, task failure came at 234 ± 66 min, and both tracked rested fat-oxidation rate
Participants
n=12 trained cyclists, age 40 ± 8 y, VO2peak 52.3 ± 5.2 ml/kg/min.
Protocol
An exhaustive cycling protocol alternating (a) incremental exercise tests that re-determined power at the moderate-to-heavy transition via the first ventilatory threshold (VT1) with (b) 30-min bouts at 90% of the VT1 power estimated in the rested state.
The individual time course of VT1 power was fitted with linear and second-order polynomial functions; the time to a 5% drop in VT1 (Δ5%VT1) was estimated from the best-fitting model.
Fat-oxidation rates (absolute and relative) were measured at specific intensities during the rested incremental test.
Key Results
Power at VT1 decreased along a second-order polynomial (non-linear, accelerating) curve in 11 of 12 participants.
Time to task failure: 234 ± 66 min. Time to a 5% drop in VT1 power: 139 ± 78 min.
The two were correlated (rs = 0.676, p = 0.016), and both were strongly correlated with absolute and relative fat-oxidation rates measured at rest.
Authors' conclusions: (i) the moderate-to-heavy transition falls non-linearly during prolonged exercise; (ii) its durability matters for prolonged exercise capacity; (iii) that durability is related to fat-oxidation rates.
Mechanisms
The paper links durability of VT1 to fat-oxidation capacity (measured correlation), not to a tested mechanism.
Inference, not tested here: progressive glycogen depletion and a rising O2 cost of the same power are the usual explanations for a moving VT1; this study did not measure glycogen or substrate kinetics during the ride.
Practical Application
In 12 trained cyclists (40 y, VO2peak 52), power at VT1 fell along a non-linear, accelerating curve during repeated 30-min bouts at 90% of rested VT1: on average VT1 had dropped 5% after 139 ± 78 min and task failure came at 234 ± 66 min; the two were correlated (rs = 0.676), and both tracked the rider's rested fat-oxidation rate. So the "easy" boundary is not fixed — it moves down with time in the saddle, faster the longer you ride, and riders who oxidise more fat hold it longer.
Inference, not tested: for a multi-hour event, a power that sits just under VT1 at the start will be above it later; a margin below rested VT1 — and a long-ride check of how HR and breathing drift at fixed power — is the practical answer. How big that margin should be was not tested here.
Limitations
n=12, trained (not elite) cohort, single group, no control condition; the effect size in a WorldTour peloton is not known from this study.
The protocol (repeated 30-min bouts at 90% of rested VT1, alternated with incremental tests) is not a race; surging may change the time course.
No direct measurement of glycogen or substrate kinetics during the ride — the fat-oxidation link is a correlation with rested-state measurements.
The repeated incremental tests themselves add fatigue beyond a steady ride.
Sex of participants and nutritional state during the protocol are not reported in the abstract.
How this entry was checked and changed
Why this level
regraded 2026-09-27 from A: uncontrolled single-group correlational study, n=12
Content correction 2026-09-27
"Tier 4-5 (Elite/International)" -> Tier 2: 12 trained cyclists, age 40 ± 8, VO2peak 52.3.
Table "GET 100% → ~93% at 60 min → ~82% at 120 min", "2 hours at 90% GET with ramps every 30 min", "p < 0.05 at all time points after 30 min", "higher R²" -> not in the abstract; replaced with the abstract's numbers: Δ5%VT1 at 139 ± 78 min, task failure at 234 ± 66 min, rs = 0.676, polynomial fit in 11 of 12.
House advice ("start 10-15 W below LT1, ~190 W instead of 205 W", "HR rises >10% over 3 h", "80-100 g/h carbohydrate", "weeks 5-6 long rides") -> not from the paper; removed. Mechanisms (glycogen depletion, fibre recruitment, slow component) -> labelled inference.
Level A -> B: uncontrolled single-group correlational study, n=12.
Descriptions of REF-144 repeated its pre-2026-09-26 version (muscle biopsies, glycogen/fibre-type depletion, fat oxidation doubling, power "preserved/maintained" at 100 g/h). The corrected entry took no biopsies and reports 6-min power FALLING ~10% after 4 h at 100 g/h; descriptions updated.
Full text (checked 2026-10-12)
The twelve trained cyclists were 10 male and 2 female, with a weekly training volume of 10.3 ± 3.4 h·week−1 (full text).
On the protocol day, participants ate a breakfast with ~ 2 g kg−1 of carbohydrate and drank ~ 800 mL of water 1 h before, and drank 600 mL of water with an electrolyte mix during each 30 min bout (full text).
After a 3-min warm-up at 100 W, each incremental test was 6 × 4 min at 70%, 80%, 90%, 100%, 110% and 120% of the VT1 power from visit one; the protocol was capped at 6 h, and task failure was volitional or an inability to keep cadence ≥ 50 rpm (full text).
VT1 power was 184 ± 35 W rested, 181 ± 37 W after an hour (p = 0.290) and 165 ± 34 W in the test before task failure (p < 0.001) (full text).
The individual second-order polynomial fits had R2 = 0.92 ± 0.09 (range: 0.72–1.00), and Δ5%VT1 ranged from 21 to 279 min between participants (full text).
Compared with the rested state, energy expenditure at VT1 was 7.3 ± 6.2% lower (range: − 0.4 to − 19.3%) and gross efficiency 4.1 ± 5.2% lower in the test before task failure (full text).
Cadence in the 90% VT1 steps did not change significantly (p = 0.528): 82 ± 9 rpm rested, 80 ± 9 rpm after an hour and 78 ± 9 rpm before task failure (full text).