REF-992 · Skovereng et al., 2018 · Frontiers in Physiology, Vol. 9, article · Paper · PubMed

Effects of Initial Performance, Gross Efficiency and VO2peak Characteristics on Subsequent Adaptations to Endurance Training in Competitive Cyclists

Blimited evidenceMeasured on highly trained athletesWritten before October 2026, being re-checked

The study examined whether initial levels of cycling performance, VO2peak and gross efficiency (GE) affect the subsequent adaptations of those variables after high-intensity training (HIT) designed to increase VO2peak.

Sixty competitive cyclists (VO2peak = 61 ± 6 mL·kg-1·min-1) were assigned a 12-week programme of twenty-four supervised high-intensity interval sessions plus ad libitum low-intensity training (1.5 ± 0.3 and 8.3 ± 2.7 h·week-1). For all athletes pooled, VO2peak and 40-min time-trial power increased and GE decreased (all p < 0.05), with no change in body mass.

Initial VO2peak correlated moderately and negatively with its change (r = -0.32), initial 40-min power weakly with its change (r = -0.27 and -0.28) and initial GE moderately with its change (r = -0.44). Four groups formed from low/high initial VO2peak and GE did not differ in their response in VO2peak, GE or 40-min power (all p > 0.12).

The authors conclude that the effects of initial level are statistically significant but small to moderate, indicating limited influence on training practice.

Participants

Sixty competitive cyclists, VO2peak = 61 ± 6 mL·kg-1·min-1. Sex and age are not given in the abstract.

Protocol

Training: 12 weeks; twenty-four supervised high-intensity interval training sessions and ad libitum low-intensity training. Average volume 1.5 ± 0.3 h·week-1 high-intensity and 8.3 ± 2.7 h·week-1 low-intensity.
Measures: VO2peak; GE calculated at 125, 175 and 225 W; performance as mean power during a 40-min time trial.
Analysis 1 (between athletes): correlation between each variable's initial level and its pre-to-post change.
Analysis 2 (between profile groups): changes compared between four groups categorised by low and/or high initial VO2peak and GE.
No non-training control group is described.

Key Results

All athletes pooled: a significant increase in absolute and body-mass-normalized VO2peak and in 40-min power (p < 0.05), and a significant decrease in GE (p < 0.05). No change in body mass.
Initial VO2peak vs change in VO2peak: moderate negative correlation (r = -0.32; p < 0.05).
Initial 40-min power vs its change: small negative correlation, absolute and per body mass (r = -0.27 and -0.28; both p < 0.05).
Initial GE vs change in GE: moderate negative correlation (r = -0.44; p < 0.01).
No differences between the four profile groups (low/high initial VO2peak and GE) in the response of VO2peak, GE or 40-min power (all p > 0.12).
Authors' conclusion: statistically significant effects of initial level, but the small and moderate effects indicate limited influence on training practice.

Practical Application

"Does my profile — big engine but poor efficiency, or the reverse — tell me how I will respond to a VO2max block?" In 60 competitive cyclists (VO2peak 61 ± 6) who all did the same 12-week programme with twenty-four supervised high-intensity sessions (single cohort; between-athlete correlations and a four-group comparison), riders grouped by low or high initial VO2peak and gross efficiency did not differ in their gains in VO2peak, efficiency or 40-min power (all p > 0.12); the starting level correlated only weakly to moderately with the change (r = -0.27 to -0.44). The authors' own verdict: limited influence on training practice.
The negative correlations mean riders who started higher gained somewhat less. Inference, not tested here: part of that is regression to the mean, which an uncontrolled design cannot separate from a true ceiling effect.
Group-level trade-off: VO2peak and 40-min power went up while gross efficiency went DOWN. A block aimed at the engine did not leave efficiency untouched in this cohort (compare REF-444).
What was not tested: whether a low-efficiency rider would do better on a different kind of block. Everyone did high-intensity training, so the study cannot match lever to profile — it shows only that the profile did not predict the response to this lever.

Limitations

No control group and no repeated block: individual change includes measurement error and regression to the mean.
Profile groups were formed by splitting on initial VO2peak and GE; group sizes are not given in the abstract, and a non-significant difference between four groups is not proof of equivalence.
One type of programme only; low-intensity training was ad libitum.
Sex, age and the interval prescriptions are not given in the abstract.
The size of the changes in VO2peak, 40-min power and GE is not given in the abstract.
Written from the abstract and record metadata only (full text not read).
How this entry was checked and changed
Why this level

uncontrolled pre-post and correlational analysis; n = 60 is large for a training study, but there is no control group and the profile groups were formed after the fact