Carbohydrate timing and the actual session

Updated 10 Oct 2026

What would change today’s food?

The reported routine does not establish a need to buy a carbohydrate drink or add another rice pouch. It includes brief running intervals and one hard working set per lifting exercise, with warm-ups. Hard effort matters, but effort alone does not reveal how much stored carbohydrate was used. The carbohydrate guide covers the food estimate; this page examines timing.

There are three different questions: whether eating beforehand makes a session feel and perform better, whether carbohydrate during exercise sustains output, and whether earlier food improves recovery before another demanding session. A positive answer to one does not settle the other two. These controlled studies also do not establish personal energy needs or explain the reported shoulder pain.

Before lifting: carbohydrate versus eating something

In a 2025 crossover trial, 16 trained adults arrived after an overnight fast and had one of three liquid breakfasts two hours before lifting. High- and low-carbohydrate breakfasts contained similar energy and protein; carbohydrate was largely exchanged for fat. The third was a low-calorie, similarly textured placebo. Across twelve sets to fatigue in a roughly 93-minute session, total repetitions did not differ significantly. The trial was small and not prospectively registered; there was no water-only condition. It does not prove that skipping all food works equally well, or exclude a small benefit. Original meal trial.

A useful contrast is a 2020 trial in 22 trained men who normally ate breakfast, after an overnight fast. A carbohydrate breakfast and a nearly energy-free, thickened placebo both improved total squat repetitions over water; neither improved total bench-press repetitions significantly. The two breakfasts performed similarly, and both reduced hunger. Participants were told both contained energy, although some identified them afterward. This supports a role for meal expectation or appetite, without separating their contributions or proving a specific mechanism. Pairwise comparisons used Holm–Bonferroni adjustment. University funds supported the study; an author also disclosed food and beverage industry relationships. Original breakfast trial.

For this routine, a familiar banana or an ordinary meal before training is a reasonable option when hungry, leaving enough time to feel comfortable. That is a practical inference, not evidence that a banana reproduces either laboratory breakfast. Extra food should solve an observed problem—hunger, discomfort or repeated loss of output—rather than satisfy an assumed mandatory carbohydrate window.

During intervals: a drink can change blood tests without improving sprints

A 2024 crossover study tested carbohydrate drinks against a noncaloric placebo in 11 trained men who had eaten a carbohydrate-containing breakfast. The exercise was substantial: three blocks of ten 45-second cycling bouts, long rests between bouts and 20-minute intermissions. Carbohydrate raised blood glucose and insulin but did not significantly preserve whole-muscle glycogen, repeated-sprint performance or perceived effort. Its confidence intervals still allowed a potentially useful small performance effect. Two electrically evoked force measures favored carbohydrate at nominal significance, with missing measurements and biopsy/order effects complicating interpretation. The trial was not registered and did not specify a broad multiplicity correction. Original interval trial.

This was far more work than the recorded 13 ten-second efforts, and cycling does not exactly reproduce running. The result nevertheless weakens the assumption that any intense interval session requires a carbohydrate drink. It says little about starting depleted, adding long endurance work or extending exercise until exhaustion. The relevant comparison includes prior food and total work, not just the word “intense.”

After a demanding session: recovery can matter even when glycogen looks similar

Another 2024 crossover trial compared immediate with delayed carbohydrate after ten two-minute cycling intervals. Nine men completed both conditions; eight contributed the next-day exercise-capacity comparison. The prescribed recovery diet matched total energy, carbohydrate and protein. Both received a meal three hours afterward, but a substantial carbohydrate drink moved from early recovery to eight hours afterward in the delayed condition. Participants completed five fewer intervals on average the next day with delayed intake, despite similar measured whole-muscle glycogen and molecular responses. Recovery-variable analyses used false-discovery-rate adjustment; capacity used a paired comparison. The study was not preregistered and was powered around a molecular outcome. Original recovery trial.

This favors earlier refueling over the study’s substantial delay when next-day capacity matters. It does not locate a precise 30-minute deadline, prove that an ordinary meal is insufficient, or prescribe its research drink for brief sprints. Nor does similar glycogen establish identical functional recovery. If repeated substantial sessions become part of the routine, a prompt, tolerable carbohydrate-containing meal becomes more relevant; rice with chicken is already one available option.

A molecular response is a different outcome

A 2024 study moved the same daily glucose amount between exercise and evening, while providing protein around training in both conditions. Thirteen participants completed a twelve-day alternating-leg protocol. Exercise-time glucose did not significantly increase the primary total-RNA outcome, a marker of muscle ribosomal content. Most strength-recovery comparisons were also null, with one nominal favorable contrast. This was neither a months-long muscle-growth trial nor evidence that carbohydrate has no role in training. Original timing study, registered primary and secondary outcomes.

There is no demonstrated adaptation benefit here that warrants forcing another shake or meal beyond appetite. Equally, laboratory signaling is a poor reason to intentionally underfuel. Longer-term intake and recovery belong in the protein and energy discussion and energy-availability guide.

The useful next observation

The current log has no repeated, comparable performance series tied to meal timing. For a few similar sessions, record the last food and time, whether the session began hungry, the work actually completed and any stomach discomfort. Record sleep and unusually demanding preceding exercise alongside it. These observations cannot diagnose depleted glycogen, but they can reveal whether moving some already planned food earlier helps. The decision can then be a small timing change in food already enjoyed, rather than another compulsory daily purchase.

Back to investigations