REF-068 · Stellingwerff, 2016 · International Journal of Sport Nutrition & Exercise Metabolism, · Paper
Competition Nutrition Practices of Elite Ultramarathon Runners (three 100-mile specialists)
Participants
n=3 veteran male ultrarunners (age 35 ± 2 years; body mass 59.5 ± 1.7 kg; 100-mile best times 16.7 ± 2.5 h). Throughout 2014 they competed in 16 ultramarathons with a total of 8 wins, including the Western States Endurance Run 100-miler (14.9 h).
Protocol
Competition-specific nutrition intake questionnaire for 100-mile races, completed with verbal and visual portion-size instruction to confirm dietary intake. Background: anecdotal claims that ultramarathoners increasingly train on low-carbohydrate ketogenic diets and race with very low carbohydrate intake; very little empirical evidence on what elite ultrarunners actually do.
Key Findings
Pre-race breakfast: 70 ± 16 g carbohydrate, 29 ± 20 g protein, 21 ± 8 g fat.
Carbohydrate during the race: 1,162 ± 250 g in total, i.e. 71 ± 20 g/h, with minor fat and protein intakes.
Energy: 5,530 ± 1,673 kcal in total (333 ± 105 kcal/h); 93% of calories from commercial products.
Caffeine: 912 ± 322 mg over the whole race.
Sodium: 6.9 ± 2.4 g over the whole race (a race total; hourly rates are not in the abstract — on the mean 16.7 h best time it is roughly 400 mg/h by simple arithmetic).
Despite limited professional nutrition input, all three practised fuelling that maximises carbohydrate intake and is congruent with contemporary evidence-based recommendations — the opposite of the low-carbohydrate racing anecdote.
Practical Application
Three elite male 100-mile runners (16.7 h best times, one a Western States winner) fuelled with 71 ± 20 g carbohydrate/h for the whole race (1,162 g total), 912 mg caffeine and 6.9 g sodium over the event, 93% of calories from commercial products — elite ultrarunners race high-carbohydrate, not low-carbohydrate, in line with evidence-based recommendations ([REF-066]).
Treat caffeine and sodium as race-long cumulative totals in this population (912 ± 322 mg caffeine; 6.9 ± 2.4 g sodium), not as hourly spikes; the paper does not report how they were spread across the race.
The pre-race breakfast was moderate (70 g carbohydrate); the race intake did the work.
Not from this paper; the base's own practice: digestive tolerance is individual and is the real constraint — rehearse the exact products and rate in long sessions at race intensity, never for the first time on race day.
Limitations
n=3, one sex, one race format (100 miles); a description of what three winners did, not a test of what works.
Self-reported questionnaire intake (portion-size aided) — under- or over-reporting cannot be excluded.
Race-total caffeine and sodium only; hourly patterns and individual tolerance are not reported.
Elite 100-mile specialists at 59.5 kg body mass; absolute intakes will not transfer to heavier or slower athletes without scaling.
How this entry was checked and changed
Why this level
regraded 2026-09-27 from B+: elite case series, n=3, self-reported intake
Content correction 2026-09-27
"Sodium 600-1000 mg/h", "600-800 mg/h", "CHO:sodium ~100:1" -> abstract: 6.9 ± 2.4 g sodium over the whole race (about 400 mg/h on the mean best time); removed the invented hourly range and ratio.
"Ultra-distance trail races, 24-100 h" -> 100-mile races, 16.7 ± 2.5 h best times; "70-75 g/h" -> 71 ± 20 g/h; "~800-1000 mg caffeine" -> 912 ± 322 mg.
"Deliberately below the ≤90 g/h laboratory ceiling", "pushing past ~90 g/h risks GI distress" -> not in the abstract; removed. Added n=3, breakfast, energy, 93% commercial products and the study's anti-low-carb conclusion.
Type "field practices review" -> descriptive case series; Level B+ -> B (elite case series, n=3); byline year 2014 -> 2016; Sport adds trail running.