Pre-exercise carbohydrate or protein ingestion influences substrate oxidation but not performance or hunger compared with cycling in the fasted state
Participants
n=17 trained male cyclists, VO2peak 62.2 ± 5.8 ml/kg/min, age 31.2 ± 12.4 years, mass 74.8 ± 9.6 kg
Protocol
3 conditions (randomized, counter-balanced crossover), meal 30 min before exercise:
1. CARB: carbohydrate-rich meal, 1 g/kg CHO
2. PROTEIN: protein-rich meal, 0.45 g/kg protein + 0.24 g/kg fat
3. FASTED: water only (overnight fasted)
Exercise: 20 min submaximal cycling (4 × 5 min at 60%, 80%, 100% of VT and at VT + 20% of the VT-to-peak difference) → 3 × 3 min at 80% peak aerobic power → 3 × 3 min maximal effort
Measurements: fat and CHO oxidation, average power in the high-intensity intervals, RPE, hunger, oxidative stress (F2-isoprostanes)
Key Results
| Parameter | CARB | PROTEIN | FASTED | Significance |
|---|---|---|---|---|
| Fat oxidation (at and below VT) | Lower | ≈ FASTED | High | CARB < FASTED at and below VT; CARB < PROTEIN at 60% VT only |
| Average power, high-intensity intervals | 367 ± 51 W (all trials) | — | — | ns (p = 0.516) |
| RPE | no difference | no difference | no difference | ns |
| Hunger | no difference | no difference | no difference | ns |
| F2-isoprostanes (oxidative stress) | no difference | no difference | no difference | ns |
Fat oxidation: overnight-fasted exercise raised fat oxidation during submaximal work compared with the CHO-rich breakfast; pre-exercise protein allowed similarly high fat oxidation (differed from CARB only at 60% VT)
Performance identical: average power in the high-intensity intervals 367 ± 51 W with no difference between trials
Hunger, RPE and oxidative stress identical: fasted training did not increase hunger or perceived exertion, and pre-exercise nutrition had no influence on exercise-induced oxidative stress (novel finding)
Mechanisms
Inference, not measured here: CHO pre-exercise → insulin rise → lipolysis inhibition → lower fat oxidation; fasted or protein-rich → low insulin response → fat oxidation preserved. Insulin was not reported in the abstract.
Performance did not suffer in this session (20 min of submaximal stages, then 3 × 3 min at 80% peak power and 3 × 3 min all-out): an acute substrate shift without an acute power cost
Practical Application
In 17 trained male cyclists (VO2peak 62), 20 min of submaximal stages followed by 3 × 3 min at 80% peak power and 3 × 3 min all-out produced the same interval power (367 ± 51 W), RPE, hunger and oxidative stress whether preceded by nothing, by 1 g/kg carbohydrate or by a 0.45 g/kg protein + 0.24 g/kg fat meal 30 min before; fat oxidation during the submaximal work was higher fasted than after the carbohydrate meal, and the protein meal preserved similarly high fat oxidation — so a fasted or protein-only start acutely raises fat use at no performance cost for a session of this length; whether that produces a training adaptation was not measured.
Protein option: the tested dose was 0.45 g/kg protein + 0.24 g/kg fat 30 min before (≈ 34 g protein for a 75-kg athlete) — if fully fasted feels too hard, a protein-rich, low-carbohydrate pre-session meal keeps the fat-oxidation effect.
Not from this paper: the "fasted morning Z2 → post-training 30 g whey + CHO" routine, protein for muscle protection on train-low days (REF-151), and the rule to take a CHO breakfast before races — house practice built on other entries.
Limitations
n = 17, crossover in trained cyclists; acute substrate and performance measures only, no training-block outcome.
One pre-exercise dose and one session format (20 min submaximal, 4 × 5-min stages, + 6 × 3-min intervals (corrected 2026-10-12, full text)); the fasted-versus-fed question changes with session length and with what follows it.
Substrate oxidation shifted but performance and hunger did not — an outcome easy to over-read in either direction. It shows no acute cost, not a proven adaptive benefit.
Male trained sample; women, masters and athletes with a history of low energy availability are not represented, and fasted work interacts badly with the latter (REF-341).
The claim that fasted training enhances fat oxidation over 4 h in elites is imported from other work, not measured here.
How this entry was checked and changed
Why this level
regraded 2026-09-27 from A: controlled crossover, n=17
Content correction 2026-09-27
"PROTEIN intermediate (FASTED > PROTEIN > CARB)" -> pre-exercise protein allowed similarly high fat oxidation as fasted; CARB was lower than FASTED at and below VT and lower than PROTEIN only at 60% VT.
Lead rewritten: acute fat-oxidation gain with no performance cost (367 ± 51 W, p = 0.516), no adaptation measured; tested protein dose (0.45 g/kg + 0.24 g/kg fat) given instead of "20-25 g / 0.3 g/kg".
"Reduces muscle damage risk" -> not in the abstract; removed. F2-isoprostane (oxidative stress) null result added. Level A -> B+ (crossover, n=17).
Second pass 2026-09-27 (number-by-number check against the abstract): the session was 20 min of submaximal stages (4 × 5 min; "40 min" here was wrong (corrected 2026-10-12, full text)) PLUS two sets of 3 × 3-min intervals, not a ~40-min session.
Descriptions of REF-144 repeated its pre-2026-09-26 version (muscle biopsies, glycogen/fibre-type depletion, fat oxidation doubling, power "preserved/maintained" at 100 g/h). The corrected entry took no biopsies and reports 6-min power FALLING ~10% after 4 h at 100 g/h; descriptions updated.
Cross-reference descriptions checked against the cited entries' current content; wrong years, wrong entries and claims the cited entry does not make were corrected.
2026-09-30: "40 min of submaximal stages" in the lead and Practical Application contradicted the Protocol; the abstract says twenty minutes (4 × 5-min stages). Corrected to 20 min.
Full text (checked 2026-10-12)
The seventeen participants were trained cyclists and triathletes with peak aerobic power 425 ± 55 W (5.7 ± 0.6 W·kg−1) and weekly training of 13.6 ± 3.1 h; inclusion required at least 8 h per week for the previous two years and VO2peak > 55 mL/kg/min, and one participant completed only two of the three trials because of an unrelated injury (full text).
For a 70-kg person, CARB was 51 g white bread with 19 g raspberry jam and 500 mL of a 7% CHO-electrolyte drink, and PROTEIN was 25 g whey protein isolate with 33 g peanut butter and 500 mL water; both meals provided 4 kcal per kg body mass (299 ± 38 kcal) (full text).
The four submaximal stages corresponded to 40.8 ± 4.4, 50.6 ± 5.9, 60.6 ± 7.2 and 71.8 ± 6.3%VO2peak; HR was lower in FASTED than in both CARB and PROTEIN, and gross efficiency was higher in FASTED than in PROTEIN at each intensity (full text).
The HIIT block was 6 × 3-min intervals with 3 min of active recovery at 100 W; intervals 1–3 were at 80%Wmax (340 ± 44 W) and intervals 4–6 were maximal efforts (full text).
During the intervals there were no differences in RPE or lactate, but HR was higher in PROTEIN than in FASTED at each interval (all p = 0.004) (full text).
Gut discomfort after exercise was higher with PROTEIN than with FASTED (p = 0.032) and CARB (p = 0.012), rising from ~12 out of 100 before exercise to ~21 out of 100 afterwards in PROTEIN only (full text).
Overall session RPE did not differ between trials (p = 0.076), with a trend for CARB (7.9) to be lower than FASTED (8.3) (full text).
Correction: the entry's Limitations describe the session as "40 min submaximal + 6 × 3-min intervals"; the paper's submaximal portion was "4 × 5-min stages", i.e. twenty minutes (full text).