Durability as an independent parameter of endurance performance in cycling
5-min and 20-min power after a 1000 kJ fatigue protocol in amateur cyclists
Participants
n=20 well-trained amateur road cyclists (abstract gives no age, sex or W/kg)
Protocol
Home-based testing on two occasions
Fresh: standardized warm-up, then a 5-min and a 20-min all-out cycling test
Fatigued: warm-up, then riding at 80% of the fresh 20-min power until 1000 kJ of work was completed, followed by the same 5-min and 20-min all-out tests
Durability = % decline of 5-min and 20-min power from fresh to fatigued; correlated with VO2max and relative FTP
Key Findings
1
Power fell after the 1000 kJ fatigue protocol: −10.1% in the 20-min test and −10.8% in the 5-min test (both significant vs fresh); the SDs are inconsistent within the paper: 20-min ± 6.5 and 5-min ± 7.8 in the abstract, 5-min ± 6.5 and 20-min ± 7.8 in Table 1 (corrected 2026-10-12, full text)
2
No significant correlation between durability and VO2max or relative FTP — the amateurs with the best fresh numbers were not the ones who faded least
3
The authors' conclusion: durability is a parameter independent of traditional physiological measures of cycling performance
4
The protocol is home-based and scaled to the rider's own 20-min power, so it is usable outside a laboratory and outside the professional cohorts of earlier durability studies
Mechanism Implications
Inference, not tested: fresh VO2max and FTP describe capacity, not how that capacity holds up under accumulated work; the wide spread of declines (SD 6.5-7.8 percentage points) means fading is an individual trait worth measuring in its own right
Which physiological factors explain the between-rider spread was not analysed in the abstract
Practical Application
In 20 well-trained amateur road cyclists, riding at 80% of fresh 20-min power until 1000 kJ of work cut subsequent 20-min power by 10.1% and 5-min power by 10.8% (SDs ± 6.5 and ± 7.8 in the abstract, swapped between the two tests in Table 1; corrected 2026-10-12, full text), and the size of that drop was not correlated with VO2max or relative FTP — so an amateur's fresh FTP or VO2max does not predict how much they fade, and durability has to be tested separately.
A usable field protocol: 5-min + 20-min fresh; another day, 80% of fresh 20-min power to 1000 kJ, then repeat both tests; a drop around 10% is the cohort average, clearly more than that flags durability as the limiter.
Limitations
Amateur cohort only; abstract does not report age, sex or W/kg
n=20 — moderate
Single fatiguing protocol (1000 kJ at 80% of fresh 20-min power); a fixed work amount takes stronger riders less time
Only VO2max and relative FTP as fresh-state correlates (LT1, LT2, GE not analysed)
Home-based tests — equipment and pacing not standardised as in a laboratory
Cross-sectional
How this entry was checked and changed
Why this level
regraded 2026-09-26 from B+: test-retest cohort, 20 amateurs
Content correction 2026-09-27
"1-min and 20-min power" -> the paper tested 5-min and 20-min power; "~3 h steady-state" -> 80% of fresh 20-min power until 1000 kJ of work.
Declines "~5-9% (1-min) / ~3-6% (20-min)" -> −10.8 ± 7.8% (5-min) and −10.1 ± 6.5% (20-min) per the abstract; Table 1 gives 10.8 ± 6.5 (5-min) and 10.1 ± 7.8 (20-min) (corrected 2026-10-12, full text).
"1-min vs 20-min durability only moderately correlated (r ~0.5), durability is duration-specific" and "FTP 3.7-4.4 W/kg" -> not in the abstract; removed. Correlate list now VO2max AND relative FTP (both non-significant).
Added Tier 2 and Sport cycling (lines were missing).
Cross-reference descriptions checked against the cited entries' current content; wrong years, wrong entries and claims the cited entry does not make were corrected.
Removed "consistent with Passfield 2000, who reached the same conclusion": REF-423 ran no predictor analysis.
2026-10-01: cross-reference repointed from a merged framework to the one that replaced it (SYNTH-01→31, 05→34, 09→70, 12/15→32, 19→26).
2026-10-01: citation of the superseded pointer SYNTH-12 repointed to SYNTH-32, which carries the cited content.
Full text (checked 2026-10-12)
All twenty participants were male (age 34,6 ± 7,3 years; body weight 74,4 ± 8,7 kg; 3,4 ± 1,06 years in endurance sport; 8,21 ± 2 h·week-1 training) (full text).
Relative power fell from 4.3 ± 0.8 to 3.8 ± 0.9 W⋅kg−1 in the 5-min test and from 3.5 ± 0.6 to 3.2 ± 0.8 W⋅kg−1 in the 20-min test (mean differences 35,2 W and 24,9 W, both P < 0.001) (full text, Table 1).
The fatigue protocol was 20 min at 70% of fresh 20-min power as warm-up (counted toward the 1000 kJ), then 80% until 1000 kJ, an immediate 5-min all-out test, 40 min of active recovery at RPE ≤ 2 out of 10, then the 20-min all-out test; the fresh test used the same 40-min recovery between efforts and a cadence of 80 to 100 rpm (full text).
Carbohydrate intake during both tests was standardized to 60 g·h − 1 (1:1 glucose to fructose), after a breakfast of ~2 g·kg − 1 carbohydrate and ~800 mL water one hour before; tests were on two consecutive weekends on the riders' own bikes on direct-drive trainers at 15 °C to 20 °C (full text).
The regression predictors were training volume, VO2max, years in sport and age; neither model was significant (5-min drop R2 0.127, p = 0.704; 20-min drop R2 0.392, p = 0.095), and W/kg was removed because it correlated with VO2max at r = 0.95 (full text, Table 2).
Correction: the entry (following the abstract) gives "−10.1 ± 6.5% in the 20-min test and −10.8 ± 7.8% in the 5-min test", whereas Table 1 lists the drop as 10.8 ± 6.5 (5 min) and 10.1 ± 7.8 (20 min), so the paper is internally inconsistent on the SDs (full text, Table 1).