Physiological Response to Cycling With Variable Versus Constant Power Output
Participants
n=15 elite competitive cyclists. Baseline day established lactate threshold (LT) and VO2max.
Protocol
Three lab days: Day 1 baseline (lactate profile + VO2max). Days 2–3: four 28-min bouts comparing:
CP (constant power): steady wattage at target intensity
VP (variable power): 15% fluctuation of the average power output every second minute, same mean power
Two intensity levels:
Low: 70% of lactate threshold (LT)
High: 95% of LT
Measurements: VO2, RER, HR, blood lactate, RPE, cadence, power.
Key Results
| Metric | Constant Power vs Variable Power |
|---|---|
| VO2 (low and high intensity) | significantly higher with VP |
| HR (low and high intensity) | significantly higher with VP in the abstract; in the Results no main effect of power condition on HR (p > 0.24; 138 vs 139 bpm low, 167 vs 168 bpm high) (corrected 2026-10-12, full text) |
| Blood lactate (low and high intensity) | significantly higher with VP; at high intensity VP also produced a greater rise in lactate across the bout |
| RPE (high intensity) | whole-bout RPE similar between conditions; the rise in RPE across the bout was smaller with VP than with CP |
Authors' reading: despite considerable differences in demand during the bouts, the perceptual and physiological differences directly after them were minor; training with VP is a viable alternative to CP, at least at high intensity.
Mechanism
Inference, not tested: variable power introduces repeated surges above and below the mean; because the VO2-power relationship is non-linear the surges cost more than the troughs save, and above-threshold surges accumulate lactate faster than below-threshold troughs clear it. The measured result is higher VO2 and lactate at the same mean power (HR: higher in the abstract, no main effect in the Results (corrected 2026-10-12, full text)).
Practical Application
Variable power costs more than constant power at the same mean wattage: in 15 elite cyclists, 28-min bouts with a 15% power fluctuation every 2 min at 70% and 95% of LT produced significantly higher VO2 and blood lactate (1.2 vs 1.4 mmol/l at 70% LT, 2.8 vs 3.9 mmol/l at 95% LT) than constant power — HR higher in the abstract but with no main effect in the Results — while whole-bout RPE at 95% LT was similar (at 70% LT it was higher with VP, 12.3 vs 11.3) — at high intensity the rider does not feel the extra metabolic cost (corrected 2026-10-12, full text).
Inference, not tested here (the study had no ERG condition and no training period): ERG mode holds constant power, and in this study constant power cost less VO2 and lactate than the same mean power ridden with surges (heart rate did not differ in the Results (corrected 2026-10-12, full text)) — so a session ridden in ERG probably understates the metabolic cost of a variable-power race.
Variable-power work is a viable alternative to constant power, at least at high intensity (the authors' practical implication) — race-simulation and group-ride-style sessions can be trained with surges without losing the training effect of the mean power.
Inference, not tested: use ERG for precise sweet-spot/threshold stimulus at a target wattage; ride race-specific and long endurance sessions with ERG OFF to expose the athlete to the variable-power metabolic cost.
Limitations
Only submaximal intensities tested (70% and 95% LT), no supra-threshold / VO2max intervals
15% variability every second minute may not represent full outdoor variability spectrum
Acute responses only — no chronic adaptation data
Elite cyclists only — may differ in recreational populations
The abstract reports directions and significance, not the lactate or VO2 values themselves; the full text gives them (Table 2) (corrected 2026-10-12, full text)
How this entry was checked and changed
Why this level
regraded 2026-09-26 from A: crossover, n=15 elite
Content correction 2026-09-27
Lactate values "2.8 -> 3.9 mmol/L (+39%)" and "1.2 -> 1.4 mmol/L (+17%)" -> deleted; not in the abstract, which reports only "significantly higher LA with VP". Superseded 2026-10-12: the full text reports 1.2 vs 1.4 and 2.8 vs 3.9 mmol/l (Table 2), so the original values were right (corrected 2026-10-12, full text).
"HR similar (NS)" -> corrected: the abstract reports significantly higher VO2, HR and LA with VP at both intensities. Superseded 2026-10-12: the Results report no main effect of power condition on HR (p > 0.24), so the original line matched the Results (corrected 2026-10-12, full text).
"RPE (low intensity) higher with VP, P < 0.05" -> deleted; the abstract reports whole-bout RPE similar at high intensity and a smaller rise in RPE with VP at high intensity, nothing on low-intensity RPE. Superseded 2026-10-12: the full text reports low-intensity whole-bout RPE 11.3 (CP) vs 12.3 (VP), significantly higher with VP (corrected 2026-10-12, full text).
"male", "of 19 recruited, 4 excluded" -> deleted; not in the abstract. Superseded 2026-10-12: the full text reports 19 elite male cyclists recruited, 4 excluded (corrected 2026-10-12, full text). Added the authors' conclusion (differences minor; VP a viable alternative to CP at high intensity).
Cross-reference descriptions checked against the cited entries' current content; wrong years, wrong entries and claims the cited entry does not make were corrected.
2026-09-29: the "ERG mode understates race demands" bullet read as a finding; the study compared constant vs variable power acutely, with no ERG condition and no training — relabelled as inference.
Full text (checked 2026-10-12)
19 elite competitive male cyclists were recruited and 4 were excluded (three for measurement errors, one could not complete the protocol); the 15 included had VO2max 72.9 ± 5.1 ml/kg/min and LT (power at 4 mmol/l) 310.5 ± 21.7 W (full text, Table 1).
Each 28-min bout was a 5 min incremental ramp, a 20 min work-matched CP or VP main period and a final 3 min at 95% of LT, ridden seated at freely chosen cadence; the 15% fluctuation every second minute let the highest intensity reach 110% of LT (full text).
Mean power in the 20-min main period did not differ: LV 218 ± 15 W vs LC 217 ± 15 W and HV 294 ± 20 W vs HC 294 ± 20 W, and cadence was not different across the four bouts (full text).
Blood lactate over the main period was 1.2 ± 0.3 (LC) vs 1.4 ± 0.4 mmol/l (LV) and 2.8 ± 0.8 (HC) vs 3.9 ± 0.7 mmol/l (HV), both significantly higher with VP (full text, Table 2).
VO2 was 44.9 ± 3.4 (LC) vs 46.0 ± 3.7 ml/kg/min (LV, significant) and 59.2 ± 4.3 (HC) vs 59.9 ± 4.1 ml/kg/min (HV) (full text, Table 2).
Whole-bout RPE was 11.3 ± 0.7 (LC) vs 12.3 ± 1.6 (LV, significantly higher) and 14.9 ± 1.3 (HC) vs 14.9 ± 1.4 (HV); from PRE to POST, RPE rose 3.1 ± 2.0 after HC vs 1.9 ± 2.1 after HV, while lactate rose 0.8 ± 1.0 after HC vs 1.7 ± 0.8 mmol/l after HV (full text, Tables 2 and 3).
Correction: the entry says HR was "significantly higher with VP" at low and high intensity; the full text reports HR 138 vs 139 bpm (LC vs LV) and 167 vs 168 bpm (HC vs HV) and states "There was no main effect of power condition for HR or RER (both p > 0.24)" (full text, Table 2).
Correction: the entry's Content correction deleted the lactate values, low-intensity RPE result and "male" as not in the paper; the full text reports them (2.8 vs 3.9 and 1.2 vs 1.4 mmol/l; RPE 11.3 vs 12.3 at low intensity; 19 male cyclists recruited, 4 excluded) (full text).