Depletion beat the taper, but only by day seven

Why they ran it

Several loading protocols were already known to produce supercompensated glycogen, so rather than test one again the authors compared two modified versions head to head, one that opens with a glycogen depleting ride and one that opens with a taper.

They also wanted to know whether the light training an athlete actually does during a loading week costs any of what is being stored.

Drawn from the stated aims of the paper.

On the morning of day five, muscle glycogen had risen to 1.45 times baseline in the depletion group and 1.24 times baseline in the taper group. Both increases were significant. The difference between the groups was not.

By day seven the depletion group was significantly ahead, at 130 against 104 millimoles per litre. The advantage of the depletion protocol is real, and it takes a full week to appear.

Twenty minutes of daily cycling at 65 percent of peak oxygen uptake cost about 10 to 14 millimoles per litre of muscle glycogen, and the muscle was back to or above its pre exercise value within 24 hours. Light training through the loading week did not interfere with loading.

The numbers

Day 5, depletion group	1.45 times baseline, p<0.001
Day 5, non depletion group	1.24 times baseline, p<0.001
Day 5, difference between groups	not significant
Day 7, depletion vs non depletion	130 +/- 7 vs 104 +/- 5 mmol/l, p<0.01
Cost of the daily 20 minute ride	10 +/- 2 and 14 +/- 5 mmol/l, recovered within 24 h
Performance outcomes measured	none

Why this might happen

Not available The full discussion is not openly available and the abstract offers no mechanism, so we do not know what the authors think caused this.

Drawn from Abstract only. The full discussion is not openly available..

What this does not show

Nobody rode a time trial. Every outcome here is a glycogen concentration. The trial cannot say whether the day seven advantage would appear anywhere a stopwatch is running, and it does not claim to.
Day five and day seven answer different questions. On day five, which is where most real world loading sits, the two protocols could not be told apart. The depletion advantage only exists if the athlete keeps loading for a full week.
The groups were unequal in size and not crossed over. Fifteen men in one arm and ten in the other, each person doing one protocol only. Person to person differences in how much glycogen someone stores sit inside the comparison rather than cancelling out of it.
The depleting bout was severe, and its cost was not measured. Two hours of cycling followed by one minute sprints to exhaustion is a hard session to do in race week. Nothing here tracked fatigue, soreness or readiness, so the price of that session is not in the result.

Where this leaves us

This is the cleanest direct comparison of a depletion protocol against a taper protocol that exists, and the depletion phase does add glycogen. But it adds it slowly. At day five the two protocols were indistinguishable, and only by day seven had they separated. Alongside the finding that light daily training costs nothing, this refines the depletion question rather than closing it: the phase is not useless, it is slow, and no performance number was attached to it.

Twenty five men in a metabolic ward, cycling, with glycogen measured by magnetic resonance rather than biopsy.

The same two protocols with a performance test on day five and again on day seven, which would show whether the glycogen gap that appears late corresponds to anything an athlete can do.

Caveats worth holding

Measured by carbon 13 magnetic resonance spectroscopy and reported in millimoles per litre, which is not the same unit as the millimoles per kilogram of dry or wet muscle used in the biopsy literature. The numbers are not directly comparable across this issue.
A parallel group design with unequal group sizes, and the paper does not describe the allocation as randomised.
Men only.
Both groups ate the same diet and did the same daily ride, so the comparison isolates the opening session and nothing else.