Preparing for a road ultra (50 km, 100 km, 24 hours)
Start slower than feels right, fuel at a rehearsed rate, drink to thirst, and treat volume as the injury risk it is; the training structure is practice, not evidence
The question it answers
"I have a flat road or loop ultra — 50 km, 100 km or 24 h. How do I train, pace, fuel, drink, and what can go wrong?" Mountain and trail ultras are SYNTH-61.
Default (what to prescribe first)
For a healthy Tier 1-3 runner who already runs a marathon comfortably:
Build volume slowly and lengthen the long run. Economy held better over 90 min in runners doing regular long runs, and its durability correlated with the length of the weekly longest run (r = -0.67) (REF-160, B, cross-sectional, 26 well-trained men; transfer to ultra is inference).
In 169 male 100-km runners, mean training speed (r = −.56) and weekly running kilometres (r = −.43) were, with age, the best predictors of race time (r² = .43); weekly hours correlated more weakly (r = −.21) and body fat, though correlated (r = .45), dropped out of the regression (REF-893, B+, cross-sectional — faster runners may simply train faster).
A narrative review ranks previous experience (ultra finishes, marathon best) as the strongest predictor (REF-826, C+). Not from a study; the base's own practice: over 12-20 weeks, a weekly long run growing toward ~4-5 h for 100 km, some weekends a second long run the next day (back-to-back), and ultra distance built over several races.
Pace the first hours below the day's average. In a 24-h race, better performers started at lower speeds relative to their own 24-h average (r = -.58 for the first 2 h; r = .03 for the last 2 h) (REF-778, B+, 398 men and 103 women, one race over 5 editions, observational).
Fuel at a rehearsed rate, with calorie-dense mixed food and some protein (REF-528, A, ISSN stand for single-stage ultras).
Taper two weeks, keep the quality (SYNTH-34, B+).
Decision (who → what)
| Situation | What changes | Source and how sure |
|---|---|---|
| 50 km (~4-6 h for most runners, the base's estimate) | Fuel like a long marathon: gut-trained multiple-transportable carbohydrate, the 3-6 h row of SYNTH-02 | SYNTH-02 B (its matrix is practice); REF-663 B+ (~90 g/h supported for ≥3 h) |
| 100 km or 24 h (>6 h) | Start fuelling lower (60-80 g/h, practice), food-based, protein 5-10 g/h; a speed ceiling for the first hours (practice) | SYNTH-02 B; REF-528 A; REF-778 B+ observational |
| Runner new to ultras, or younger | Slower volume build; this group reported more injury | REF-779 B+, self-reported, association; the slower build is that entry's inference |
| Woman with focal foot bone pain | Stop running, refer — suspected stress injury | REF-779 B+ (stress fractures more common in women, 44.5% in the foot); referral rule is practice (REF-622, C+) |
| Salty sweater planning to drink back most losses over ~160 km | Sweat test and targeted sodium may be worth it | REF-223 C (model, no participants) |
| Hot race | Fluid needs rise: plan bottles per hour from a weighed sweat rate, never above the measured loss; thirst overrides the plan downward, never upward (SYNTH-68, the same rule as SYNTH-64) | REF-225 C (editorial: planned drinking favoured in heat for efforts over 60 min); REF-528 A (needs rise in heat); REF-776 A- (never beyond losses; thirst cues persist in heat); the plan itself is practice |
| Wants pain-killers for the race | No ibuprofen (NSAIDs) during the race | REF-892 A-: 89 runners, 80 km desert stage, 400 mg every 4 h: acute kidney injury 52% vs 34% on placebo (number needed to harm 5.5; CI −4% to 41%, harm not proven, safety not shown); AKI 44% overall and more with greater weight loss |
Training volume and back-to-back long runs
Pacing
Fuelling and GI — the ultra carbohydrate conflict, stated
The base holds numbers that do not agree:
How far the entries resolve it: duration splits part of it — a 50 km inside ~6 h sits under REF-663's ≥3 h statement and SYNTH-02's 3-6 h row; 100 km and 24 h sit where SYNTH-02 itself steps back. For >6 h the A-level stand (30-50 g/h) and the elite practice (71 g/h) differ, and no entry in the base tests carbohydrate rates in events longer than ~6 h.
Not from a study; the base's own practice: start at 60-80 g/h if rehearsed, fall back toward the stand's 30-50 g/h when the gut refuses, never try a rate on race day that training has not tolerated. Expect weight loss in ultras and treat it as absorption, not discipline (SYNTH-24, B).
Caffeine: the stand supports strategic use in the latter stages, particularly when sleep deprivation may compromise safety (REF-528).
Drinking and sodium
Injury risk
Among 1,212 ultrarunners, 64.6% reported an injury costing training days in the past year (median 14 days), knee most common, 5.5% had a stress fracture (44.5% in the foot), stress fractures were more common in women, and injured runners were younger and less experienced (REF-779, B+, retrospective, self-reported; race distances and terrain not given).
Sleep in a 24-h race
No entry studies sleep in a 24-h run. Transfer: in 23 riders of a multi-day ultracycling race, those who slept more per 24 h finished higher and less sleep went with worse cognition (REF-796, B, observational; multi-day cycling, not a one-night run). At UTMB, sleep loss plus effort impaired cognition up to visual hallucinations (REF-665, B, n=17, mountain).
Before the race: ≤6 h of sleep cut performance ~7.6% on average (REF-204, A-, meta-analysis, 89% men). Not from a study; the base's own practice: protect the two nights before, expect a sleepiness trough before dawn, place caffeine there (REF-528 supports late-race use).
Taper and race week
A ≤21-day taper cutting volume 41-60% with intensity and frequency kept improved time-trial performance (REF-3492, A, 14 studies, endurance athletes); two weeks is the default, longer for very long racing only as house reasoning (SYNTH-34, B+). No ultra taper entry exists.
Carbohydrate loading: a moderate regimen matched the depletion regimen for stored glycogen but did not speed a 20.9 km run (REF-774, B); the entry's own practice line, not the study, puts a 2-3 day high-carbohydrate phase before a marathon or longer.
Recovery
How to progress / periodise
Not from a study; the base's own practice: base (volume, strength, easy running) → build (longer long runs, back-to-backs every 2-3 weeks, fuelling at race rate) → specific (race pace late in long runs, a night run for 24 h, loop or road rehearsal) → 2-week taper. Raise one load at a time; keep weekly steps under 20% (REF-623, half-marathon cohort).
When NOT to use it / red flags
What the evidence does and does not show
Shows: better 24-h performers started slower relative to their own average (association); faster 100 km men trained faster and ran more kilometres (association); ibuprofen went with more kidney injury in an ultra RCT (52% vs 34%, CI crossing zero); flat 24-h strength loss is mostly central; ultrarunners report injury often, knee most, stress fractures more in women; overdrinking causes hyponatraemia and thirst-guided drinking is the recommended prevention; the gut is trainable.
Does not show: an ultra training plan, back-to-backs, walk breaks in an ultra (only a marathon RCT), a tested carbohydrate rate beyond ~6 h, an ultra taper, 24-h sleep strategy, recovery after a flat ultra.
Evidence gaps
How this entry was checked and changed
(fuelling and fluid safety rest on an A stand and an A- consensus; pacing on one B+ observational race; the NSAID kidney signal on an A- RCT whose difference was not significant; training predictors on one B+ cross-section; training plan, back-to-backs, taper and sleep are house practice)