REF-958 · Thurber et al., 2019 · Science Advances, Vol. · Paper · PubMed

Extreme events reveal an alimentary limit on sustained maximal human energy expenditure

Blimited evidenceMeasured on a mixed groupWritten before October 2026, being re-checked

The authors pooled measurements of basal metabolic rate (BMR) and total energy expenditure (TEE) recorded during human endurance events, plus new data from adults who ran ~250 km per week over 20 weeks of a transcontinental race.

Maximum sustained metabolic scope (SusMS), their measure of sustained expenditure, depended on event duration: across events of 0.5 to 250+ days it fell along a curve as duration grew and levelled off below 3× BMR. Shorter events such as marathons follow a different relationship.

Bringing in overfeeding studies, the authors see evidence that human alimentary energy supply is limited to ~2.5× BMR, with any spending beyond that drawn from the body's energy stores. Data from the transcontinental race suggest people can partly cut TEE over long events and so extend endurance.

Participants

Not given in the abstract beyond: people in the compiled endurance events, and adults who ran ~250 km per week over 20 weeks of a transcontinental race (the new data). The number of participants, sex, age and training level are not in the abstract (n not stated).

Protocol

Compiled measurements of TEE and BMR from human endurance events lasting 0.5 to 250+ days.
New data: BMR and TEE measured in a transcontinental race, in adults who ran ~250 km per week over its 20 weeks.
Outcome: maximum sustained metabolic scope (SusMS, TEE as a multiple of BMR) as a function of event duration.
Data from overfeeding studies were added to estimate the limit of energy supply.

Key Results

SusMS is a function of event duration: for events lasting 0.5 to 250+ days it decreases curvilinearly with duration, plateauing below 3× BMR.
The relationship differs from that of shorter events (e.g., marathons).
Adding overfeeding data, the authors see evidence that alimentary energy supply tops out at ~2.5× BMR; spending beyond that has to come from the body's energy stores.
The data from the transcontinental race suggest people can partly cut TEE over long events as a way to extend endurance.
The abstract gives no values for individual events, no sample size and no statistics.

Practical Application

Across compiled endurance events lasting from half a day to more than 250 days (observational compilation plus new data from a 20-week transcontinental running race), the highest sustainable energy expenditure fell curvilinearly with event duration and flattened below 3× BMR, and with overfeeding data added the authors put the limit of energy supply through eating at about 2.5× BMR — above that, the body's energy stores pay the difference. To "can I just eat more to keep a very high load going for months?": on this evidence the long-run limit sits on the supply side, so a load above it is paid from the body.
The ceiling is duration-dependent: what can be held for a day or a week is far above what can be held for months. Do not apply the long-run figure to a single hard week.
Inference, not tested here: weight loss during very long events fits this model (expenditure above what can be taken in is covered by body stores); body mass change is not reported in the abstract.
Not from this paper: the same model was tested in ultra-endurance athletes (REF-959) and against a year of professional cycling (REF-960); how the base uses it for recovery weeks and annual load is in SYNTH-24.
Never quote 2.5× BMR to an individual athlete as a target or a personal limit; it is a population estimate.

Limitations

The abstract gives no sample size, no participant description and no statistics; the strength of the fitted curve cannot be judged from it.
A compilation of events measured by different groups with different methods; the alimentary limit is an inference from adding overfeeding data, and energy intake and absorption during the events are not described in the abstract.
The new cohort is runners in one transcontinental race.
A population relationship, not an individual prescription.
Written from the abstract and record metadata only (full text not read).
How this entry was checked and changed
Why this level

cross-sectional / correlational compilation with a small new cohort; the number of people and events is not given in the abstract