Cyclists do not need to incorporate off-bike resistance training to increase strength, muscle-tendon structure, and pedaling performance: Exploring a high-intensity on-bike method
Participants
n=37 well-trained cyclists (male, per the companion MSS paper; not in the abstract of this paper):
OFF-bike RT (n=12): full squat
ON-bike RT (n=12): high-intensity all-out pedaling efforts
CON (n=13): continued the cycling routine with no additional RT stimulus
Protocol
Duration: 10 weeks of RT added to the cycling routine
The only difference between RT groups was the exercise: full squat vs all-out pedaling efforts. RT intensity (% maximal dynamic force, MDF), volume, sets and rest were identical between groups.
Companion MSS paper (not in the abstract of this paper): 2 RT sessions per week, relative load 70% of 1RM, on-bike efforts against very high resistance and thus very low cadence.
Cycling volume at each intensity zone matched between the three groups.
Outcomes in the abstract: maximal aerobic power, power at the respiratory compensation point (RCP) and at the ventilatory threshold, off-bike MDF and pedaling (on-bike) MDF, quadriceps size, patellar tendon size, time-to-exhaustion, injury-related symptoms.
Key Findings
| Parameter | OFF-bike | ON-bike | CON |
|---|---|---|---|
| Maximal aerobic power | ↑ (ES ≥ 0.37) | ↑ (ES ≥ 0.37) | — |
| Power at RCP | ↑ (ES ≥ 0.20) | ↑ (ES ≥ 0.20) | — |
| Power at ventilatory threshold | — | significant ↑ (ES = 0.24) | — |
| Off-bike MDF | meaningful ↑ (ES ≥ 0.16) | meaningful ↑ (ES ≥ 0.16) | significant ↓ (ES ≤ −0.40) |
| Pedaling (on-bike) MDF | — | significant ↑ (ES = 0.67) | significant ↓ (ES ≤ −0.40) |
| Quadriceps size | significant ↑ (ES = 0.22) | tendency ↑ (ES = 0.15) | significant ↓ (ES = −0.26) |
| Patellar tendon size | — | tendency ↑ (ES = 0.35) | — |
| Time-to-exhaustion | ↑ (ES ≥ 0.30), considerable but not significant | ↑ (ES ≥ 0.30), considerable but not significant | — |
| Injury-related symptoms | tendency ↑ (ES ≥ 0.33) | — | — |
ON-bike vs OFF-bike: no significant differences between the two RT methods in any parameter.
Both RT groups improved maximal aerobic power and RCP power and increased off-bike MDF; the on-bike group additionally increased pedaling MDF and power at the ventilatory threshold significantly.
CON declined: without an RT stimulus the control group significantly lost off- and on-bike MDF (ES ≤ −0.40) and quadriceps size (ES = −0.26) over the 10 weeks.
Safety: the off-bike (squat) group tended toward more injury-related symptoms (ES ≥ 0.33); nothing of the kind is reported for the on-bike group.
Companion MSS paper (not in the abstract of this paper): no within- or between-group effect on VO2max or body composition; squat and pedaling strength rose 2.6-5.8% (off-bike) and 4.5-7.3% (on-bike) vs −5.8% to −3.9% in controls (p < 0.05); Wingate +4.1% and +4.3% vs −4.9% (p ≤ 0.018); muscle cross-sectional area +2.5% and +2.2% vs −2.3% (p ≤ 0.008).
Mechanisms
Inference, not tested here: all-out pedaling against very high resistance loads the same musculature as a squat while keeping the movement pattern of the sport, which fits the larger pedaling-specific MDF gain in the on-bike group (ES = 0.67).
The control group's loss of strength and quadriceps size while cycling training continued suggests that endurance cycling alone does not maintain these qualities over 10 weeks (the abstract's finding; the mechanism is not measured).
Practical Application
In n=37 well-trained cyclists, 10 weeks of all-out high-resistance pedaling efforts (on-bike RT, n=12) matched full-squat training (off-bike RT, n=12) with no significant difference in any parameter: both improved maximal aerobic power (ES ≥ 0.37), RCP power (ES ≥ 0.20) and off-bike strength (ES ≥ 0.16), while the no-RT control (n=13) lost off- and on-bike strength (ES ≤ −0.40) and quadriceps size (ES = −0.26). For an athlete without gym access or time, resistance work done on the bike is a real option, not a compromise.
The on-bike group was the one with the significant pedaling-specific strength gain (ES = 0.67) and the significant ventilatory-threshold power gain (ES = 0.24); the squat group was the one that tended toward more injury-related symptoms (ES ≥ 0.33).
Programming detail that made the on-bike work "resistance training": efforts all-out against very high resistance at very low cadence, 2 sessions per week, matched to 70% 1RM (companion MSS paper; not in the abstract of this paper).
Not training strength has a cost: the control group, cycling as usual, lost strength and quadriceps size in 10 weeks. Inference, not tested: some strength stimulus should stay in every phase; the minimum dose was not studied here.
Not from this paper: Barranco-Gil 2024 REF-119 on why normal pedaling loads are too low to act as strength training.
Limitations
Well-trained, not elite, cyclists; 10 weeks; three groups of 12-13.
Several effects are reported as effect sizes with "tendency" or "considerable but not significant" (time-to-exhaustion, tendon size, injury symptoms) — they are directions, not established differences.
The Biology of Sport abstract carries no VO2max or body-composition result; those come from the companion MSS paper and are labelled as such.
12, 12 and 13 well-trained men; 10 weeks.
"No need for off-bike training" means no difference was detected between the two strength groups, not that equivalence was proven.
Time-to-exhaustion gains were not significant, and quadriceps growth was significant only in the off-bike group.
How this entry was checked and changed
Why this level
regraded 2026-09-27 from A: parallel RCT, n=37 in three arms
Content correction 2026-09-27
"On-bike MDF ↑ in both RT groups", "quadriceps CSA stable/↑ in both", "patellar tendon ↑ (ES 0.35) in both" -> abstract: pedaling MDF significant only in the on-bike group (ES = 0.67); quadriceps size significant in off-bike (ES = 0.22), tendency in on-bike (ES = 0.15); tendon tendency (ES = 0.35) in on-bike only.
"~70% on-bike MDF", "cadence adjustment ±5 rpm (~2.5% MDF)", "squat 1RM", "Wingate", "DXA", "VO2max no change" -> not in this abstract; cadence/MDF numbers removed, VO2max/Wingate/% changes kept only as labelled companion-MSS-paper data (70% of 1RM, not 70% MDF).
"Gym remains preferred; maintenance dose at least 1×/week; never abandon strength" -> house advice not from the paper; removed or reduced to labelled inference.
Tier 4-5 -> Tier 2-3 (well-trained); Level A -> B+ (parallel RCT, n=37) per KB-LEVEL-RUBRIC. "Also published as" -> companion paper on the same trial (different title and DOI).