REF-871 · Vanhatalo et al., 2008 · Medicine & Science in Sports & Exercise, Vol. · Paper · PubMed

A 3-min all-out cycling test is sensitive to a change in critical power

Blimited evidenceMeasured on recreational peopleWritten before October 2026, being re-checked

Nine habitually active subjects did a ramp test, two 3-min all-out cycling tests (end power, EP, and work done above end power, WEP) and three predicting trials to establish critical power (CP) and W′ with the work-time model, before and after 12 supervised high-intensity interval sessions over 4 weeks.

CP rose in every subject and EP rose with it; the two estimates did not differ and were highly correlated before and after training, and the change in EP tracked and did not differ from the change in CP. W′ and WEP did not change significantly. The authors conclude the 3-min all-out test closely estimates CP and is sensitive to training-induced changes in CP.

Participants

n=9 habitually active subjects (abstract: "Nine habitually active subjects"). Sex, age and fitness values are not given in the abstract beyond CP.

Protocol

Before and after training: ramp test; two 3-min all-out tests (end power, EP; work done above EP, WEP); three predicting trials for CP and W′ from the work-time model (W = CPt + W′).
Training: 12 supervised high-intensity interval training sessions over 4 wk (interval format not given in the abstract).

Key Results

MeasurePre-trainingPost-trainingTest
CP (predicting trials)230 ± 53 W255 ± 50 Wt8 = 7.47, P < 0.001 (rose in all subjects)
3-min test EP225 ± 52 W248 ± 46 Wt8 = 6.26, P < 0.001
W′17.2 ± 4.2 kJ15.5 ± 3.8 kJt8 = 2.03, P = 0.08 (not significant)
WEP——no change (t8 = 1.89, P = 0.10)
EP and CP were not different and were highly correlated before (r = 0.96, P < 0.001) and after training (r = 0.95, P < 0.001).
The increase in EP correlated with (r = 0.77, P = 0.016) and did not differ from (t8 = 0.60, P = 0.57) the increase in CP.

Practical Application

In nine habitually active people, 12 interval sessions over 4 weeks raised critical power in everyone (230 to 255 W) and the end power of a 3-min all-out test rose with it (225 to 248 W); the change in end power tracked the change in CP (r = 0.77) and did not differ from it, while W′ did not change significantly. To "can I retest with a single 3-min all-out effort to see whether my CP went up after a block?": yes — in this group the 3-min test picked up the training change as well as the full set of predicting trials did.
The test replaces three or more separate predicting trials with one effort, which makes retesting after a block practical.
Inference, not tested here: these were habitually active people with a large training response; in trained cyclists, whose changes are smaller, whether the test can separate a real change from day-to-day noise is not shown here.
The test depends on a truly all-out start with no pacing; the conditions are described with the original validation REF-044.
Not from this paper; the base's own practice: which threshold test to choose and how often to retest is in SYNTH-53.

Limitations

n=9 habitually active subjects; no control group, so test-retest learning is not separated from training.
The interval training format is not given in the abstract.
Sensitivity to the smaller changes typical of trained cyclists was not tested.
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, n=9