REF-142 · Maunder et al., 2018 · Sports Medicine, Vol. · Paper

Substrate Metabolism During Ironman Triathlon: Different Horses on the Same Courses

C+thin evidenceMeasured on recreational to world-class athletesWritten before October 2026, being re-checked

Key Points (abstract)

Ironman is an ultra-endurance event of extreme duration; performance level varies dramatically — elite competitors finish in ~8:00:00, lower-performing amateurs in ~14-15:00:00.
Applying published swimming, cycling and running economies to these times shows that the absolute energy cost is high and the rate of energy expenditure increases with the athlete's competitive level (a calculation, not a measurement).
Because endogenous carbohydrate storage is finite, the authors argue that minimising the endogenous carbohydrate cost of exercise at competitive intensity should be a goal of Ironman preparation.
Strategies that "may help", "including, but not limited to": a low-carbohydrate diet, exogenous carbohydrate supplementation, and periodised training with low carbohydrate availability.
The authors propose that performance level be considered when choosing these strategies ("different horses on the same courses"):
Elite and top-amateur triathletes, with very high rates of energy expenditure: periodised training with low carbohydrate availability combined with exogenous carbohydrate during competition "might be most appropriate".
Some lower-performing amateurs (>12 h): a low-carbohydrate or ketogenic diet "might be appropriate", because the high fat oxidation rates associated with that diet "may be almost completely sufficient" to match the energy demand.

Key Findings — what the paper models and argues

1

Same course, different rate (modelled): the modelled rate of energy expenditure falls from 20.1 kcal/min (elite, ~8 h) to 16.6 (top amateur, ~9 h) and 13.3 (lower amateur, ~13 h), and the modelled carbohydrate oxidation rate falls with it. The total cost does not fall in step with the rate: 9626 kcal (elite), 9062 (top amateur) and 10,425 (~13 h, the highest); the ~13 h model's total carbohydrate cost range (875-1999 g) overlaps the elite's (985-1675 g).

2

Fat oxidation is an input, not a result: the model enters wide lower and upper average fat oxidation rates, set lower for slower athletes (0.3-0.9 g/min for the lower amateur vs 0.6-1.2 for the elite). On those inputs the lower amateur still oxidises 1.12-2.56 g/min of carbohydrate.

Across those inputs fat supplies roughly 22-66% of the lower amateur's modelled energy vs 29-58% of the elite's (calculated from Table 1 at 9.75 kcal/g), so the model does not establish a larger fat share for the slower athlete; near-complete fat coverage is proposed only after dietary fat adaptation (Fig. 1, 'Optimised low-amateur').

3

Endogenous carbohydrate is likely limiting: the elite's estimated carbohydrate requirement (985-1675 g) exceeds whole-body carbohydrate stores (<740 g), and heat is expected to make it worse; the authors call fatigue from carbohydrate depletion "of particular importance" to elites.

4

Elite and top amateur (~8-9 h) — a "cross-over approach" is recommended (opinion): raise fat-oxidation capacity through training (planned low-carbohydrate-availability sessions, possibly with gut-training sessions), maximise pre-race glycogen, and take carbohydrate in the race from multiple transportable sources. The reason given: even the highest fat oxidation achievable through diet is suggested to be insufficient at ~20 kcal/min.

5

Lower amateur (~13 h; abstract >12 h) — a low-carbohydrate or even ketogenic diet is called "plausible" at ~13.3 kcal/min, with minimal exogenous carbohydrate. The paper says it "may be favourable" for those with gut discomfort on race-day carbohydrate who are focused primarily on simply completing the event, and that its performance effects for Ironman "are not known".

The same paper reports critics' point that these diets lack shown benefit on ecologically valid performance measures and may blunt fast carbohydrate use for brief hard efforts.

Inference, not tested here: the paper is a model-and-argument piece; it tests neither strategy in a trial, and no trial compares them head-to-head.

Practical Application

Opinion piece with a modelled energy budget: the rate of energy expenditure in an Ironman rises with competitive level (modelled 20.1 kcal/min at ~8 h vs 13.3 at ~13 h), so the authors propose matching the fuelling strategy to performance level — for elite and top amateurs, planned low-carbohydrate-availability training plus carbohydrate from multiple transportable sources in the race; for some amateurs beyond 12 h, a low-carbohydrate or ketogenic diet "might" suffice, mainly for those with gut trouble on race-day carbohydrate who aim simply to finish, with performance effects unknown. Place yourself by expected finishing time before copying a fuelling plan; a slower finish lowers the modelled rate, while the modelled total stays in the same range (9062-10,425 kcal across the three models).
For a fast athlete the race-day question is exogenous carbohydrate; the paper cites a guideline of ~0.8:1 fructose-glucose at 1.5-1.8 g/min and gut training. For intake targets see SYNTH-02; for gut-training detail see REF-008.
The paper's rationale for planned low-carbohydrate-availability sessions is to drive adaptations in fat metabolism; it calls the fat-oxidation results "promising" and flags concern that impaired high-intensity power in those sessions might detrain. It does not show a race-intensity fuel shift in triathletes.
The paper does not make the low-carbohydrate or ketogenic option a default for slow amateurs; the base's own handling (do not prescribe it, keep a carbohydrate plan) is in SYNTH-57.
Not from this paper; the base's own practice: whatever the strategy, rehearse it in long sessions before race day and expect the athlete's own response, not the averaged one, to decide.

Limitations

Current Opinion piece; energy costs and substrate use come from a theoretical model (published economies applied to race times, fat oxidation rates entered as wide assumptions), not from measurements in a race, though the totals agree with one measured study of eight men.

The model is male-based and partly built on an author's own race; one author coaches elite Ironman triathletes. No trial compares the proposed strategies. The low-carbohydrate or ketogenic suggestion for slow amateurs is explicitly conditional ("plausible", "might", for some, mainly gut-sensitive athletes aiming to finish) and its performance effects are stated as unknown.

How this entry was checked and changed
Why this level

regraded 2026-10-14 from B+: an opinion piece without a systematic search; its numbers come from a model the authors call a "postulated estimate" (published economies applied to two real race times and one constructed ~13 h time, with fat oxidation rates entered as assumptions), checked against one small measured study; the recommendations for slower amateurs are explicitly untested, and one author coaches elite Ironman triathletes and supplied his own race as the top-amateur input

Full text (checked 2026-10-14)
Article type and method: a Current Opinion piece that builds theoretical models for elite (~8 h), top-amateur (~9 h) and lower-amateur (~13 h) finishers from published swim, bike and run economies; the elite model is Lionel Sanders' 2017 Kona race (8:04:07, 313 W bike), the top-amateur model is author DJP's 2015 Kona race (9:05:46, 225 W), and the lower-amateur model takes Sanders' phase times increased by 63% with equal cycling work; body mass of 75 kg is assumed for both amateur models.The authors call the model "a postulated estimate" and invite readers to adjust the values. (full text)
Modelled energy cost and rate by finish time: elite 9626 kcal at 20.1 kcal.min-1, top amateur 9062 kcal at 16.6 kcal.min-1, lower amateur 10,425 kcal at 13.3 kcal.min-1. As a cross-check the paper cites a measured 10,036 ± 931 kcal in eight males finishing in 10.9–13.0 h, and 8570 ± 1014 kcal in five females, so the male-based values might be scaled down for women. (full text)
Substrate assumptions: wide lower and upper average fat oxidation rates are entered (0.6–1.2 g.min-1 elite, 0.5 and 1.1 top amateur, 0.3 and 0.9 lower amateur), with 9.75 kcal per g fat, 4.07 kcal per g carbohydrate and negligible protein.Resulting carbohydrate oxidation: 2.05–3.49 g.min-1 (elite), 1.44–2.88 g.min-1 (top amateur), 1.12–2.56 g.min-1 (lower amateur); total carbohydrate cost 985–1675 g (elite); for the top amateur 1573 g at the lower and 788 g at the upper fat-oxidation bound; for the lower amateur 1999 g and 875 g. (full text)
Endogenous storage assumptions: about 80% of carbohydrate is stored in muscle and 10–15% in the liver, typically under 3000 kcal (740 g) in total, against about 68,250 kcal (7000 g) of fat in a 70-kg athlete with 10% body fat. The elite's estimated carbohydrate requirement exceeds whole-body carbohydrate content, likely more so in Kona heat (~30 °C, ~65% relative humidity). (full text)
Strategies reviewed: low-carbohydrate high-fat (under 25% of energy from carbohydrate) or ketogenic (under 20 g carbohydrate a day) diets consistently raise whole-body fat oxidation, but critics cite a lack of benefit on ecologically valid performance measures; train-low and sleep-low look "promising"; ingested carbohydrate is oxidised at up to ~1.8 g.min-1; ketone supplementation is "very much in its infancy" (one study found a 2 ± 1% slower 31-km time trial after a ketone diester).(full text)
Field data cited: 53 finishers at the 2009 Kona race ate a mean 62 ± 26 g.h-1 of carbohydrate; over 30% of Ironman triathletes had serious gut problems, yet faster finishes went with higher carbohydrate intake rates (r = - 0.55 to - 0.45). (full text)
Elite and top amateur (~8–9 h): a cross-over approach is "recommended" — planned low-carbohydrate-availability training to drive fat-metabolism adaptations, maximised pre-race glycogen, and race carbohydrate from multiple transportable sources (~0.8:1 fructose–glucose at 1.5–1.8 g.min-1), because energy expenditure of ~20.1 kcal.min-1 (and ~16.6 kcal.min-1 for a top-amateur male, taken as an elite female) likely exceeds what fat alone can supply, particularly in heat.(full text)
Lower amateur (~13 h): at ~13.3 kcal.min-1 it is "plausible" that a low-carbohydrate or even ketogenic diet "might be sufficient" with minimal exogenous carbohydrate; this "may be favourable" for those with gut discomfort on race-day carbohydrate who are focused on simply completing, and the specific performance effects of this approach for Ironman "are not known". (full text)
Limitations and conflicts stated or visible: one swim economy (national-level Italian swimmers) is used for all three models; the economies are best estimates from the literature; the model is male; the top-amateur input and its fat-oxidation check are an author's own data; DJP competes in Ironman triathlons and coaches elite Ironman triathletes; no funding. (full text)
Content correction 2026-10-14 (full text)
Type "Narrative review with energy-cost calculations (no new data)" | Level B+ -> Current Opinion built on a theoretical model | Level C+ (opinion, no systematic search, model the authors call a postulated estimate, recommendations for slow amateurs untested, author conflict disclosed).
Tier "mixed→Tier 1-4 (review spanning ~8 h elite finishers to 14-15 h amateurs)" -> mixed→Tier 1-5: the three modelled athletes are ~8 h (Kona runner-up), ~9 h and ~13 h; there are no participants.
"The performance level should decide the strategy" -> the authors "propose" that performance level be considered.
Elite/top-amateur recommendation stated without hedging -> abstract: "might be most appropriate"; the body "recommends" it as opinion.
Slow amateurs: "because their high rates of fat oxidation may be almost completely sufficient" -> the high fat oxidation is the one associated with the low-carbohydrate or ketogenic diet, not the athlete's current rate; added the paper's conditions (gut discomfort on race-day carbohydrate, focus on simply completing) and its statement that the performance effects are not known.
Key Finding "the slower athlete works at an intensity where fat can cover most of the demand" -> not shown: fat oxidation rates are model inputs (lower for slower athletes), and the ~13 h model still oxidises 1.12-2.56 g/min of carbohydrate; on those inputs fat supplies roughly 22-66% of its modelled energy vs 29-58% for the elite (calculated); fat sufficiency is proposed only after dietary fat adaptation.
Key Finding "larger carbohydrate demand" for the faster athlete -> true per minute only; total energy cost is similar across the models (9626, 9062, 10,425 kcal; the ~13 h model highest) and the ~13 h total carbohydrate range overlaps the elite's.
Practical "whose fat oxidation can nearly cover the demand" -> "might" suffice for some, conditional and untested.
Practical "Train-low sessions ... raise the share of fat in the energy mix at race intensity" -> the paper's rationale is to drive fat-metabolism adaptations; it calls results "promising" and flags concern over impaired high-intensity power; no race-intensity shift in triathletes is shown.
Practical "for the intake rate and gut training see REF-543 and REF-008 (not from this paper)" -> the paper itself cites ~0.8:1 fructose-glucose at 1.5-1.8 g/min and gut training; intake targets now point to SYNTH-02, gut-training detail to REF-008.
Cross-reference 'REF-543 — race-day carbohydrate intake targets' -> REF-543 is Impey 2018 carbohydrate periodisation; relabelled, and intake targets pointed to SYNTH-02.
Limitations "calculated, not measured in a race" -> kept, and added the measured cross-check it cites, the male-based model, the author's own race as input and the disclosed conflict.
Added the full-text section (model inputs, energy costs and rates, storage assumptions, strategies, conditions, limitations), the Pfeiffer 2012 REF-787 and SYNTH-57 cross-references, and the train-low and low-carbohydrate-diet tags; removed the durability tag (the paper does not address durability).