Eating around training
Set carbohydrate by what the day holds, keep protein near 1.8 g/kg spread through the day (not lower on rest days), fuel every key session, and use "train low" only on easy sessions inside an adequately fed week; meal timing is a smaller lever than the daily totals
The question it answers
Default (what to prescribe first)
Energy availability is the floor. The UCI statement (REF-748, A) sets a habitual 40-45 kcal/kg fat-free mass per day (lower end for riders with little activity outside training) and accepts short, planned periods of low availability only if recovery follows.
REDs affects men and women (REF-341, A); that low energy availability can look like overtraining is the entry's own application line. In 12 trained-to-elite men, 14 days at ~22 kcal/kg FFM cut explosive power and haemoglobin before hormones changed (REF-545, B, uncontrolled pre-post).
Carbohydrate by the day's load, in g/kg, not as % of energy (REF-748; REF-070, B pending verification, same unit). UCI rows: light endurance 5-7 g/kg/day, moderate endurance (<3 h/day) 6-10, intensified (>4 h/day) >8, extreme racing 10-25; the entry's own application, not a line of the statement, is a 90-min session = a 5-7 g/kg day and a 4-h ride = a >8 g/kg day (REF-748). Fat ~20-35% of energy, moving inversely to carbohydrate on the day (REF-748).
Protein ~1.8 g/kg/day, spread over the day. The endurance review advocates ~1.8 g/kg/day and notes it may exceed 2.0 on rest days and in carbohydrate-restricted training per its abstract; its Table 1 targets 2.0 on recovery days and 1.95 on low-carbohydrate training days (corrected 2026-10-12) (REF-151, A, narrative); UCI gives 1.6-2.1 with the upper end in low-carbohydrate or energy-restricted periods (REF-748).
By the IAAO method: recommended 1.83 g/kg/day in 6 trained men (REF-156, B), 1.81 in men and 1.89 in women runners (REF-725, B+). The ISSN stand doses 0.25 g/kg (20-40 g) evenly every 3-4 h (REF-578, A).
Not from a study; the base's own practice: four to five protein feedings of roughly 0.4 g/kg (arithmetic on 1.8 g/kg/day), one within a couple of hours after the session. REF-151's Table 1 recommends 0.5 g/kg in immediate post-exercise recovery (preliminary, single study) (corrected 2026-10-12).
Fuel the key sessions; practise train-low only on easy ones (REF-543, B+, framework). Details below.
Decision (who → what)
| The day | Carbohydrate | Protein and timing | Source and how sure |
|---|---|---|---|
| Rest day | No rest-day row in the base. Not from a study; the base's own practice: eat to appetite, at or a little below the light-day row, and do not cut to "earn" food | Hold protein at or above the training-day total | REF-151 A (may exceed 2.0 g/kg on rest days); REF-152 B (endurance-trained men at a fixed 0.93 g/kg: the rest day had the lowest protein synthesis and net balance) |
| Easy day (~1-1.5 h easy) | 5-7 g/kg | ~1.8 g/kg; the day a fasted or protein-only start fits | REF-748 A; REF-543 B+ (train-low on base sessions); REF-157 B+ (fasted start acutely costless in one short session) |
| Hard / interval day | 6-10 g/kg (UCI's <3 h/day row; mapping an interval day to it is the base's own practice) | Normal distribution; a carbohydrate-rich meal before the session is practice | REF-748 A; REF-543 B+: intensity must not be compromised |
| Long day (over ~4 h) | >8 g/kg; in-ride grams → SYNTH-02 | Breakfast before; calcium-rich if long or sweaty (practice; the trial was one 90-min ride) | REF-748 A; REF-544 B+ (female cyclists, surrogate markers) |
| Two sessions, <~6 h apart (the trial tested a 4-h gap; ~6 h is practice) | Recovery dose 1-1.2 g/kg/h starting soon after the first; fast (high-GI) and part fructose | 0.3-0.4 g/kg protein with it | REF-748 A; REF-734 B+; REF-057 B; REF-056 B; the whole decision for close sessions: SYNTH-85 |
| Race eve and morning, event >90 min | 10-12 g/kg/day for 1-2 days before a single-day event; 1-4 g/kg in the 1-4 h before the start | Rehearse the menu in training (practice) | REF-748 A; REF-819 B+ (loading pays only >90 min) |
| Train-low session | Reduced availability for that one easy session; the day and week still cover the work | Protein up: 1.95 g/kg in REF-151's Table 1 (its abstract: may exceed 2.0) (corrected 2026-10-12); train-low raised the requirement without quantifying it (REF-726) | REF-543 B+; REF-726 B; REF-003 B+; REF-954 C+ (ten trained men, crossover: after a glycogen-depleting session, peak fat oxidation in a graded test the next day rose further with a low-carbohydrate diet than with a high-carbohydrate one; mitochondrial coupling control and efficiency unchanged; no performance outcome) |
Before a session. In 17 trained male cyclists, a session of submaximal stages plus 6 × 3-min intervals gave the same interval power (367 ± 51 W), RPE and hunger after an overnight fast, a 1 g/kg carbohydrate meal, or a protein-fat meal (0.45 g/kg protein + 0.24 g/kg fat) eaten 30 min before; fasting raised fat oxidation, and the protein meal kept it similarly high (REF-157, B+).
Transfer: one short session in trained men; it does not show that a long or race-intensity session can be done unfed. Before a long effort, a ~2 g/kg breakfast with about a third from fructose extended time to exhaustion by ~7 min (137 vs 130 min) in 8 trained men (REF-065, B; p = .046, a tweak).
A dairy meal with ~1350 mg calcium before a 90-min ride blunted the rise in bone-resorption markers in 32 well-trained female cyclists (REF-544, B+; markers, not bone density). Not from a study; the base's own practice: a larger meal 3-4 h before a hard session, or a small carbohydrate snack 30-60 min before if the session comes early.
After a key session — the "window". Timing matters when the next hard session is close.
In 12 male cyclists, 2 g/kg carbohydrate taken immediately after a 70-min ride stored glycogen at 7.7 vs 2.5 μmol/g/h over 2 h compared with waiting 2 h, and once fed the late group's rate (4.1) stayed below the early group's first-2-h rate (REF-056, B); the entry's own inference is that with 24 h and adequate intake glycogen catches up.
Over 24 h at the same 10 g/kg, high-GI carbohydrate restored more glycogen than low-GI (106 vs 71.5 mmol/kg, 5 cyclists; REF-057, B) — the entry applies it when the next hard session is within a day. With 4 h between two runs, fructose + maltodextrin recovery at the same 90 g/h gave a longer second run (81.4 vs 61.4 min, n=8; REF-734, B+; direction firm, magnitude not).
There is no compelling evidence that protein added to adequate carbohydrate improves performance or glycogen resynthesis (REF-151, A). Inference: post-session protein serves the day's total, not the glycogen. Pre-sleep 20-40 g casein raised overnight protein synthesis in young men after resistance exercise and in older men; there are no endurance-athlete data (REF-155, A).
Masters. Per-meal protein saturated MPS at ~0.40 g/kg in healthy men aged ~71 vs ~0.24 g/kg at ~22 (REF-348, B; p = .055 per kg body mass; untrained men — transfer). In resistance training, the lean-mass benefit of protein fell with age (REF-579, A; lifting, not endurance). UCI gives masters riders the upper protein band (REF-748).
The REF-348 entry's own application: size each feeding for roughly 0.4 g/kg. The REF-543 entry adds, as not from that paper, that masters athletes should start train-low at a lower frequency.
Women. Protein requirement did not differ by sex in endurance runners (1.60 vs 1.61 g/kg/day breakpoint, n=15; REF-725, B+); a narrative review found it premature to set sex-specific carbohydrate or protein guidelines provided energy needs are met (REF-582, C+).
Female endurance athletes are less likely than men to meet carbohydrate guidelines, because of energy restriction (REF-070, B). A lengthening or missing cycle during a hard block is a fuelling question first (REF-748; SYNTH-37 for the cycle itself).
Vegetarian and vegan athletes. Per gram, soy and wheat protein raised muscle protein synthesis less than animal protein — lower essential amino-acid content, often short in lysine or methionine, and slower digestion — and eating more, blending sources or fortifying can compensate (REF-912, B+, narrative review of acute studies; no gram target).
With total protein matched at 1.6 g/kg/day (soy for vegans, whey for omnivores), 19 habitual vegan and 19 omnivorous untrained young men gained the same leg lean mass (1.2 kg each) and strength over 12 weeks of resistance training (REF-911, B+, non-randomised; lifting, not endurance — transfer).
Inference: a plant-based endurance athlete keeps the same ~1.8 g/kg/day target, never a lower one, from combined sources; whether it should be higher is untested (REF-912's authors ask).
How to progress / periodise
When NOT to use it / red flags
What the evidence does and does not show
Evidence gaps
How this entry was checked and changed
(daily targets rest on A consensus written largely for trained or elite athletes, transfer to amateurs named as open; timing, recovery and train-low rows on B/B+ crossovers of 5-21 mostly trained men; example days are practice)