In 43% of energy-availability studies, menstrual status is reported too vaguely to grade

Energy availability is the arithmetic underneath relative energy deficiency in sport: energy intake minus exercise energy expenditure, divided by fat-free mass [s1]. Three quantities, one division. An audit published in Sports Medicine on August 26 asked how the field actually measures them, and the answer is that the measurement is far less standardised than the concept.

What was audited

The authors applied the Smith et al. 2020 standardised protocol to human studies that directly assessed or manipulated energy availability, searching PubMed in June 2024 and October 2025 [s1]. Studies using indirect measures such as questionnaires or surveys, studies of energy balance only, studies of diseased populations, and non-English reports were excluded [s1]. Included publications were assessed on population, athletic calibre, menstrual classification, study design and methodology [s1].

That left 203 publications [s1].

Who is in this literature

Most included females — 70% — and half studied females alone [s1]. High-performance athletes were well represented at 45% (tiers 3–5), particularly among studies that assessed rather than manipulated energy availability [s1].

Both of those are, in context, good news. The authors note that high-performance female athletes are well represented in energy-availability research compared with other areas of sports science and medicine research [s1] — a field that has spent decades being criticised for the opposite.

The reporting problem

The audit's grading system for menstrual-status assessment found that 57% of publications used techniques graded gold, silver or bronze [s1]. The remaining 43% provided insufficient terminology or detail to classify female participants' menstrual status at all, and were recorded as ungraded [s1].

Among those that did report, 54% included naturally menstruating females, 28% included athletes with menstrual irregularities, and 18% included athletes using hormonal contraceptives [s1].

The significance of the 43% is specific rather than general. Menstrual dysfunction is one of the main outcomes that low energy availability is thought to produce; it is also a variable that changes how the body responds to energy restriction. A study that cannot say whether its participants were menstruating normally, irregularly, or on hormonal contraception cannot cleanly be pooled with one that can — and the authors say as much, noting that despite the well-known association between energy availability and menstrual dysfunction, many studies failed to use high-quality methodology to describe or account for menstrual characteristics [s1].

How the numbers get made

The methodological detail explains why estimates diverge. Among the 146 observational studies, food records were the dominant method for energy intake, at 82%, and metabolic equivalent scores were used for exercise energy expenditure in 40% [s1]. Self-reported food records are a known source of under-reporting, and METs are a table lookup rather than a measurement.

The 57 intervention studies were better instrumented: 77% provided food to participants, and 70% determined exercise energy expenditure by physiologically derived methods such as indirect calorimetry or heart rate [s1].

So the studies that describe how common low energy availability is tend to use the weakest methods, and the studies that use the strongest methods tend to be small experiments. That is the inverse of what a reader of prevalence figures would assume.

What the downstream evidence looks like

A meta-analysis published on August 19 shows what gets built on this base. It searched seven databases for studies from January 2007 to July 2023 comparing body composition and hormonal status between athletes with low and normal energy availability, and ultimately included 11 studies covering 1,006 athletes — 471 female (49.4%) and 535 male (50.6%) — of whom 509 (50.6%) were classified as having low energy availability [s2].

Athletes with low energy availability tended to have lower weight, body mass index and total body bone mineral density Z-score [s2]. Several physiological markers were lower, cortisol was higher, and leptin did not differ significantly [s2].

Eleven studies is a thin base for a question this prominent, and a cross-sectional design cannot say which came first. The authors' conclusion is correspondingly hedged: low energy availability is associated with adverse health outcomes in athletes [s2].

What the audit recommends

A standardised calculation of energy availability and best-practice protocols for its assessment may improve study quality — but the authors argue the bigger need is a shift toward intervention studies rather than further assessment of energy availability in free-living athletes, if the effects of low energy availability on health and performance are to be understood [s1]. They also call for better characterisation of menstrual cycles, and for investigation of whether the causes of and responses to low energy availability differ in male athletes [s1].

What to watch

Whether the next wave of consensus documents makes menstrual-status grading a reporting requirement rather than a recommendation. Until it is one, the ungraded 43% will keep propagating into every meta-analysis built on this literature.

Sources

• Methods of Calculating Energy Availability in Current Research: A Standardised Audit — Sports Medicine, 2026-08-26
• A study for markers of low energy availability in athletes: a meta-analysis of cross-sectional studies — European Journal of Clinical Nutrition, 2026-08-19