The effect of low- vs high-cadence interval training on the freely chosen cadence and performance in endurance-trained cyclists
Participants
n=16 male endurance-trained cyclists. Baseline: submaximal rides at 60% Wmax at 50, 70, 90 and 110 rpm and at freely chosen cadence (FCC) for gross efficiency (GE), RPE and crank torque profile.
Protocol
6 weeks, 18 interval-based training sessions, randomly assigned:
Low-cadence group (LC): intervals at 20% below FCC
High-cadence group (HC): intervals at 20% above FCC
Performance test: 15-min time trial, preloaded by a ride at 60% Wmax — distance and average power
Pre/post: GE at each cadence, crank torque profiles
Key Results
| Metric | Low-Cadence Group | High-Cadence Group |
|---|---|---|
| FCC change | unchanged (93 rpm) | 92 → 101 rpm (P = 0.01) |
| 15-min TT distance and power | improved — superior to HC | improved |
| GE | not reported as changed | 22.7 → 23.6% at 90 rpm (P = 0.05); 20.0 → 20.9% at 110 rpm (P = 0.05); 22.8 → 23.2% at FCC |
| Crank torque profile | minimal change | minimal change |
Both groups significantly increased TT distance and average power; the LC group recorded the superior performance measure
The authors: FCC can be altered with HC interval training; the determinants of optimal cadence are multifactorial and not fully understood; LC interval training may improve short-duration TT results via strength development or neuromuscular adaptation
Mechanism
Inference (the authors' suggestion, not measured): high-cadence training shifts the preferred cadence upward, and low-cadence training may improve short TT performance through strength development or neuromuscular adaptation. Type II recruitment, torque per stroke and cardiovascular cost are plausible routes but were not tested here.
Practical Application
RCT, 16 male endurance-trained cyclists, 18 interval sessions over 6 weeks at 20% below or above their preferred cadence: both groups improved a preloaded 15-min TT, the low-cadence group more; high-cadence training raised preferred cadence from 92 to 101 rpm and improved gross efficiency at 90 and 110 rpm, low-cadence training left preferred cadence unchanged. Low-cadence intervals bought more short-TT performance; high-cadence intervals bought a faster, more efficient spin.
Low-cadence intervals (about 20% under your natural cadence) are a defensible way to structure some interval work when the target is sustained power
High-cadence intervals raise the cadence you naturally settle at and make spinning at 90-110 rpm cheaper — useful when the event demands fast legs (criterium, track), less so for steady gravel efforts
Both approaches improved performance; 6 weeks was enough for a measurable cadence adaptation
Not from this paper: absolute cadence targets (65-75 rpm), pairing low-cadence work with LT1 or high-cadence work with VO₂max sessions — see REF-084 for low-cadence HIT.
Limitations
16 subjects, only male — small sample
6 weeks — no data on longer adaptation periods
Fixed % of FCC — absolute cadences varied between individuals
Single preloaded 15-min TT — untested at longer durations relevant to gravel; the abstract gives no numbers for the size of the LC advantage
How this entry was checked and changed
Content correction 2026-09-27
Performance test "15-min TT at 60% Pmax" -> 15-min maximal time trial preloaded by a ride at 60% Wmax.
Tier "mixed→Tier 2-3 (review ...)" -> Tier 2-3; this is an RCT, not a review.
Derived cadences (~74 / ~112 rpm), house targets (65-75 rpm, LT1/VO₂max pairing) -> removed or moved under "Not from this paper"; GE numbers added from the abstract; group sizes marked full text.
Cross-references to duplicate stubs repointed to their canonical entries (REF-014->REF-084); the stubs were emptied on 2026-09-27.
Cross-reference descriptions checked against the cited entries' current content; wrong years, wrong entries and claims the cited entry does not make were corrected.
2026-09-28: lines labelled "full text; not checked" read against the open-access full text (utoronto.scholaris.ca, accepted manuscript apnm-2015-0562.R2): 0 confirmed, 0 corrected, 1 removed. Removed: "Group sizes 8/8" -- the manuscript reports n=16 randomly assigned to HC/LC but never states the group sizes.
Full text (checked 2026-10-12)
Participants were aged 31.4 (6.2) yr, body mass 74.6 (6.1) kg, VO2max 4.85 (0.5) L·min-1, and all were training > 300 km·wk-1 (full text).
Each of the 18 supervised sessions (three per week, 45 - 60 mins) had a 10 minute warm up (5 minutes at 100 and 125 W), then 4 × 4 min at 70%Wmax with 2 min at 100 W between intervals in weeks 1-2, five bouts in weeks 3-4 and six bouts in weeks 5-6, and a 10 min warm down at 100 W (full text).
Training cadences ranged from 96 to 121 rpm in the HC group and 60 to 81 rpm in the LC group; outside the lab participants were not to do any interval training and had to pedal at their prescribed cadence (full text).
The performance test was a 15 min preload at 60%Wmax followed by a 15-min time trial in which participants received no feedback except time (full text).
TT distance improved 4% in HC (9.8 to 10.2 km) and 7% in LC (9.8 to 10.5 km); TT average power rose 8% in HC (269 W to 291 W) and 16% in LC (269 W to 312 W), with LC improving significantly more (F = 5.28, p = 0.04) (full text).
Wmax rose from 363 W to 383 W in HC and from 365 W to 388 W in LC, and VO2max rose in HC from 4.55 to 4.72 L·min-1 (full text).
The optimal metabolic cadence (highest GE) was 50 rpm in both groups before and after training, and the HC group's GE increase at FCC (22.8% to 23.2%) was not significant (full text).
FCC during the time trial did not change in either group (HC 100-102 rpm, LC 102 to 103 rpm) (full text).