One day of low carbohydrate cost nineteen watts at the moderate to heavy transition in female cyclists

What the moderate to heavy transition is

Exercise intensity is usually divided into domains, and the first boundary is the point where breathing starts to climb faster than the workload, called the first ventilatory threshold. It is the ceiling on the pace an endurance athlete can hold comfortably for a long time, which is why it matters more to a long event than a maximum does.

Drawn from background physiology, not from this paper.

Why they ran it

Carbohydrate availability falls during prolonged exercise, and the authors set out to test whether that fall is part of why power at the intensity domain transitions drifts down over a long ride. Their stated route is that glycogen depletion may impair contractile function, force recruitment of higher threshold motor units, and reduce energetic efficiency.

Drawn from the paper's introduction.

Nine well trained female cyclists completed a randomised, counterbalanced crossover. In each condition they performed glycogen depleting exercise, then over roughly the next 24 hours ate either at least 9 grams of carbohydrate per kilogram of body mass or no more than 1 gram per kilogram, before an incremental cycling test and a three minute all out test.

Power at the first ventilatory threshold was lower after the low carbohydrate day, 133 plus or minus 24 watts against 152 plus or minus 28, a difference of 19 plus or minus 14 watts at P equals 0.011.

Power at the lactate threshold did not differ between the two conditions, and neither did end test power in the three minute all out test.

Gross cycling efficiency during submaximal cycling was reduced after the low carbohydrate day, at P equals 0.003.

Electromyographic median power frequency differed significantly between conditions in both the vastus lateralis, at P equals 0.025, and the vastus medialis, at P equals 0.007. Electromyographic amplitude did not differ. The direction of the median power frequency difference is stated inconsistently within the paper and is set out under what this does not show rather than reported here as a result.

The numbers

High carbohydrate condition	at least 9 g/kg body mass over about 24 hours
Low carbohydrate condition	no more than 1 g/kg body mass over about 24 hours
Power at the first ventilatory threshold, low carbohydrate	133 plus or minus 24 W
Power at the first ventilatory threshold, high carbohydrate	152 plus or minus 28 W
Difference at the first ventilatory threshold	19 plus or minus 14 W lower, P equals 0.011
Power at the lactate threshold	no difference between conditions
Gross cycling efficiency during submaximal cycling	reduced after the low carbohydrate day, P equals 0.003
End test power, three minute all out test	no difference between conditions
Electromyographic median power frequency	differed between conditions, vastus lateralis P equals 0.025, vastus medialis P equals 0.007
Electromyographic amplitude	no difference between conditions

Why this might happen

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

Their account is that with glycogen depleted from the low threshold type I fibres, the muscle compensates by recruiting higher threshold type II fibres to produce the same power, and those fibres are less energetically efficient than the ones they are standing in for.

That single move is meant to carry both findings at once. It costs efficiency, and it lowers the power that can be held at the moderate to heavy transition.

Drawn from the paper's discussion.

What this does not show

Two of the four performance measures did not move. Power at the lactate threshold was unchanged, and so was end test power in the three minute all out test. Whatever low carbohydrate availability did here, it did not do it everywhere on the intensity range.
The paper states the muscle activation result in two opposite directions. The title and the discussion say median power frequency was raised by low carbohydrate availability, which is what the authors' motor unit recruitment explanation requires. The results sentence in the abstract says it was higher in the high carbohydrate condition. Both cannot be right, so the direction is not repeated here as a finding. The two P values are unambiguous and the amplitude null is unambiguous.
It does not show that a long ride does this. The carbohydrate was removed by diet over about a day, not by hours of riding. The authors offer the result as relevant to durability over a long event, and that is an extrapolation from a controlled feeding design, not something this study measured.
Nine cyclists. A small crossover. It is the only study in this issue conducted in women, which makes it valuable and does not make it large.

Where this leaves us

The cost of low carbohydrate availability now has a number attached at the intensity that decides long events, and it has one in women. This adds to rather than confirms the adaptation studies in this issue, because no adaptation was involved: a single day of eating was enough to move the ceiling on sustainable pace by nineteen watts.

Nine well trained female cyclists, tested about 24 hours after glycogen depleting exercise and a controlled day of eating.

The same measurements taken during a long ride rather than after a feeding manipulation, which is the setting the authors say the result speaks to and the one they did not test.

Caveats worth holding

The paper reports the direction of the median power frequency difference inconsistently between its title and discussion on one hand and its abstract results on the other. The significance is not in doubt, the direction is.
A correction was published on 13 July 2026, PMID 42440090. It replaced a duplicated figure. According to the correction notice no numerical values, statistical findings or conclusions were altered.
Nine participants in a crossover, with trials 5 to 14 days apart.
This is acute carbohydrate availability over about a day, not adaptation to a diet over weeks. It sits alongside the retooling studies in this issue rather than inside them.
The two conditions differ enormously in muscle glycogen by construction, so this is not a fine grained dose response.
Well trained female cyclists only.