REF-993 · Collins et al., 2022 · Journal of Applied Physiology, Vol. · Paper · PubMed

Critical power and work-prime account for variability in endurance training adaptations not captured by V̇O2max

Blimited evidenceMeasured on a mixed groupWritten before October 2026, being re-checked

Do individual differences in critical power (CP, the maximal metabolic steady state) and W′ (the work that can be done above it) explain how endurance responds to training? 22 adults were tested for CP, W′, VO2max and other variables before and after 8 wk of either HIIT (high-intensity interval training) or MICT (moderate-intensity continuous training), each set at fixed percentages of PGXT, the power linked to VO2max in a graded exercise test.

HIIT raised CP more on average (HIIT 27.5 ± 4.3% vs MICT 15.7 ± 3.1%; P = 0.03), but within each group the gain ranged widely (HIIT: 4%-61%, MICT: 4%-36%). Most of the variance in CP change was explained by how hard the prescribed exercise was relative to CP before training, not relative to PGXT (R2 = 0.61-0.64; P < 0.01).

Before training CP and VO2max were related (R2 = 0.92, P < 0.01), but how much CP changed with training showed no significant relation to how much VO2max changed (R2 = 0.06, P = 0.26). The combined changes in W′ and CP explained the shift in time-to-failure at the original PGXT better (R2 = 0.77) than the change in VO2max did (R2 = 0.24).

Participants

22 adults. Age, sex, training status and group sizes are not given in the abstract.

Protocol

Design: two training groups, 8 wk: moderate-intensity continuous training (MICT) or high-intensity interval training (HIIT). How participants were allocated is not stated in the abstract.
Prescription: fixed percentages of PGXT (the power associated with VO2max during a graded exercise test).
Measures before and after: CP, W′, VO2max, time-to-failure at PGXT, and other variables.
Analysis: group comparison of the CP change; regressions ACROSS participants — (a) pre-training values against each other (between-person, one time point), (b) each participant's training-induced change in one variable against their change in another (between-person comparison of within-person changes).

Key Results

CP increased more after HIIT: MICT 15.7 ± 3.1% vs HIIT 27.5 ± 4.3% (P = 0.03). The magnitude varied widely within each group (MICT: 4%-36%, HIIT: 4%-61%).
The intensity of the prescribed exercise relative to pre-training CP, not PGXT, accounted for most of the variance in the change in CP in response to a given protocol (R2 = 0.61-0.64; P < 0.01).
CP and VO2max were related before training (R2 = 0.92, P < 0.01), but the training-induced change in CP was not significantly related to the change in VO2max (R2 = 0.06, P = 0.26).
Pre-training, W′ was related to time-to-failure at PGXT (R2 = 0.52; P < 0.01); VO2max was not (R2 = 0.13; P = 0.10).
How training shifted time-to-failure at the original PGXT was explained better by the combined W′ and CP changes (R2 = 0.77, P < 0.01) than by the VO2max change (R2 = 0.24; P = 0.02).
Authors' conclusion: exercise may be more homogeneously and effectively prescribed in relation to CP than to traditional factors like VO2max.

Practical Application

"If I raise my VO2max, will my threshold power come up with it?" In 22 adults (training status not stated) who did 8 wk of continuous or interval training, critical power and VO2max were tightly related across people before training (R2 = 0.92), yet a person's training-induced change in critical power was not significantly related to their change in VO2max (R2 = 0.06, P = 0.26). A relationship seen between people at one time point did not describe how the two changed within people who trained.
What predicted the gain in critical power: how hard the prescribed work was relative to the person's own pre-training CP (R2 = 0.61-0.64), not relative to graded-test power. Two people given the same percentage of test power were working at different intensities relative to CP.
Consequence for prescription: anchoring sessions to CP gives a more uniform stimulus than anchoring to VO2max or ramp-test power — the authors' own conclusion.
Scope: 22 adults, two groups, 8 wk; a non-significant R2 = 0.06 shows no detectable link in this sample, not that none exists. Inference, not tested here: transfer to trained endurance athletes, whose range of change is much narrower than the 4%-61% seen here.

Limitations

n = 22 split over two groups; allocation method, sex, age and training status are not given in the abstract.
The key results are correlations of change scores across participants, which carry measurement error from both tests; no repeated block to separate true individual response from noise.
"Not significantly related" at n = 22 leaves a wide range of true relationships possible.
CP changes of 4%-61% suggest participants with a lot of room to improve; whether the pattern holds in trained athletes was not tested.
How CP and W′ were determined 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

22 adults in two groups, allocation method not stated in the abstract; the headline findings are correlations across participants