Carbohydrate during intervals spared muscle glycogen but not liver glycogen

Nine male cyclists rode 8 intervals of 5 minutes with 1 minute recovery on three occasions: placebo, 60 g of maltodextrin, and the same maltodextrin plus caffeine, taurine, l theanine, l citrulline and citicoline. Liver and muscle glycogen were measured before and after with carbon 13 magnetic resonance spectroscopy, which means no biopsy.

Muscle glycogen use was around 40 percent lower with carbohydrate than placebo (p=0.006). With the multi ingredient version the sparing was not evident (p=0.073).

Liver glycogen use did not differ between any of the three trials (p=0.101), even though post exercise plasma glucagon was lower in both carbohydrate trials (p=0.001) and plasma glucose was comparable across trials (p=0.175).

Mean power output during the later intervals was 2.8 percent higher with the multi ingredient drink (p=0.046). That came with higher plasma lactate than carbohydrate alone (p=0.003), lower bicarbonate (p=0.005), a shift in base excess (p<0.001) and lower total carbon dioxide (p=0.004).

The numbers

Muscle glycogen use, carbohydrate vs placebo	around 40% lower, p=0.006
Muscle glycogen use, multi ingredient vs placebo	sparing not evident, p=0.073
Liver glycogen use across trials	no difference, p=0.101
Mean power output, late intervals, multi ingredient	+2.8%, p=0.046
Plasma lactate, multi ingredient vs carbohydrate	higher, p=0.003

Why this might happen

Proposed by the authors This is the explanation the authors offer in their discussion. This study did not test it.

The authors explain the disappearing sparing as a behavioural consequence rather than a metabolic one. Their words are that the effect is not evident with the caffeine containing blend, potentially due to an increased capacity to sustain higher power outputs resulting in greater glycogen utilisation.

In other words the fuel saving was spent rather than lost. The riders converted it into 2.8 percent more power in the later intervals.

The liver result is left unexplained. The authors note the dissociation plainly, that liver glycogen use did not differ despite lower post exercise plasma glucagon in both carbohydrate trials.

None of this was tested here. The paper's discussion is behind a paywall, so this is the mechanism as the authors stated it in their own abstract.

Drawn from Abstract and concluding statement, full text paywalled.

What this does not show

It does not show that the caffeine did it. The multi ingredient drink combined five actives at once. Caffeine cannot be separated from taurine, theanine, citrulline or citicoline in this design.
It does not show that liver glycogen is spared by carbohydrate in this setting. It shows a null with nine riders and p=0.101. That is weak evidence of no effect rather than evidence of no effect, and the paper does not claim the study was powered for it.
It does not show a race outcome. One interval session in a laboratory, with power output as the only performance measure. Nothing here follows the riders across consecutive days or into a competition.
It does not apply to women. Nine male cyclists.

Where this leaves us

Carbohydrate during high intensity intervals looks like a muscle glycogen story rather than a liver glycogen story, which refines the usual framing that in feeding carbohydrate you are protecting hepatic output.

Nine trained male cyclists doing 8 by 5 minute intervals after a standardised day of eating.

A design that gives caffeine on its own against the same carbohydrate, and holds power output fixed, would separate whether the blend removes the sparing or the rider does.

Caveats worth holding

Nine participants across three conditions.
The full text is paywalled, so the mechanistic reading here comes from the paper's own abstract rather than its discussion.
Volume, issue and page numbers are not yet assigned, so the citation is to the article number.
The performance gain and the loss of glycogen sparing are two readings of the same event and cannot be separated in this design.