Energy availability (EA) is defined as the amount of dietary energy available to sustain physiological function after subtracting the energetic cost of exercise. Insufficient EA due to increased exercise, reduced energy intake, or a combination of both, is a potent disruptor of the endocrine milieu. As such, EA is conceived as a key etiological factor underlying a plethora of physiological dysregulations described in the female athlete triad, its male counterpart and the Relative Energy Deficiency in Sport models. Originally developed upon female-specific physiological responses, this concept has recently been extended to males, where experimental evidence is limited. The majority of data for all these models are from cross-sectional or observational studies where hypothesized chronic low energy availability (LEA) is linked to physiological maladaptation. However, the body of evidence determining causal effects of LEA on endocrine, and physiological function through prospective studies manipulating EA is comparatively small, with interventions typically lasting ≤ 5 days. Extending laboratory-based findings to the field requires recognition of the strengths and limitations of current knowledge. To aid this, this review will: (1) provide a brief historical overview of the origin of the concept in mammalian ecology through its evolution of algebraic calculations used in humans today, (2) Outline key differences from the ‘energy balance’ concept, (3) summarise and critically evaluate the effects of LEA on tissues/systems for which we now have evidence, namely: hormonal milieu, reproductive system endocrinology, bone metabolism and skeletal muscle; and finally (4) provide perspectives and suggestions for research upon identified knowledge gaps.

The physiological effects of low energy availability (EA) have been studied using a homogenous daily EA pattern in laboratory settings. However, whether this daily EA pattern represents those of free-living athletes and is therefore ecologically valid is unknown. To investigate this, we assessed daily exercise energy expenditure, energy intake and EA in 10 free-living elite male road cyclists (20 min Mean Maximal Power: 5.27 ± 0.25 W · kg⁻¹) during 7 consecutive days of late pre-season training. Energy intake was measured using the remote-food photography method and exercise energy expenditure estimated from cycling crank-based power-metres. Seven-day mean ± SD energy intake and exercise energy expenditure was 57.9 ± 10.4 and 38.4 ± 8.6 kcal · kg FFM⁻¹ · day⁻¹, respectively. EA was 19.5 ± 9.1 kcal · kg FFM⁻¹ · day⁻¹. Within-participants correlation between daily energy intake and exercise energy expenditure was .62 (95% CI: .43 – .75; P < .001), and .60 (95% CI: .41 – .74; P < .001) between carbohydrate intake and exercise energy expenditure. However, energy intake only partially compensated for exercise energy expenditure, increasing 210 kcal · day⁻¹ per 1000 kcal · day⁻¹ increase in expenditure. EA patterns displayed marked day-to-day fluctuation (range: −22 to 76 kcal · kg FFM⁻¹ · day⁻¹). The validity of research using homogenous low EA patterns therefore requires further investigation.

The carbohydrate deficit induced by exercise is thought to play a key role in increased post-exercise insulin action. However, the effects of replacing carbohydrate utilized during exercise on postprandial glycaemia and insulin sensitivity are yet to be determined. This study therefore isolated the extent to which the insulin-sensitizing effects of exercise are dependent on the carbohydrate deficit induced by exercise, relative to other exercise-mediated mechanisms. Fourteen healthy adults performed a 90-min run at 70% $\dot{\mathrm{V}}{\mathrm{O}}_2$ max starting at 1600–1700 h before ingesting either a non-caloric artificially-sweetened placebo solution (CHO-DEFICIT) or a 15% carbohydrate solution (CHO-REPLACE; 221.4 ± 59.3 g maltodextrin) to precisely replace the measured quantity of carbohydrate oxidized during exercise. The alternate treatment was then applied one week later in a randomized, placebo-controlled, and double-blinded crossover design. A standardized low-carbohydrate evening meal was consumed in both trials before overnight recovery ahead of a two-hour oral glucose tolerance test (OGTT) the following morning to assess glycemic and insulinemic responses to feeding. Compared to the CHO-DEFICIT condition, CHO-REPLACE increased the incremental area under the plasma glucose curve by a mean difference of 68 mmol·L⁻¹ (95% CI: 4 to 132 mmol·L⁻¹; p = 0.040) and decreased the Matsuda insulin sensitivity index by a mean difference of −2 au (95% CI: −1 to −3 au; p = 0.001). This is the first study to demonstrate that post-exercise feeding to replaceme the carbohydrate expended during exercise can attenuate glucose tolerance and insulin sensitivity the following morning. The mechanism through which exercise improves insulin sensitivity is therefore (at least in part) dependent on carbohydrate availability and so the day-to-day metabolic health benefits of exercise might be best attained by maintaining a carbohydrate deficit overnight.

Training with low carbohydrate availability (LCA) has been proposed as an independent determinant of physiological perturbations commonly attributed to low energy availability (LEA) and to increase skeletal muscle oxidative machinery, yet the effects of LCA in isolation from LEA remain unclear. We examined whether short‐term carbohydrate restriction under energy balance alters endocrine and metabolic markers associated with LEA and skeletal muscle proteomic response. In a randomized crossover design, eight trained males completed 4 days of either a low‐carbohydrate high‐fat diet (LOW; 12% carbohydrate, 69% fat, 19% protein) or a normal‐carbohydrate diet (NORM; 62% carbohydrate, 19% fat, 19% protein), while undertaking daily cycloergometer exercise (15 kcal kg FFM ⁻¹ day ⁻¹ ) and maintaining energy availability at 45 kcal kg FFM ⁻¹ day ⁻¹ . LOW induced a clear metabolic shift consistent with LCA, evidenced by elevated circulating free fatty acids, glycerol and β‐hydroxybutyrate, in fasting conditions and fat oxidation at rest and during exercise, alongside reduced exercise glucose concentrations. Despite these responses, LOW did not alter insulin, testosterone, triiodothyronine, leptin, hepcidin, or P1NP. In contrast, β‐CTX increased and IGF‐1 decreased relative to NORM. Muscle glycogen concentration decreased only in LOW (40% ± 14%). Proteomic analysis identified 671 proteins; 57 differentially expressed in LOW relative to NORM were limited to fatty acid metabolism pathways and suppression of ribosomal, sarcomeric, and extracellular matrix proteins. These findings indicate that isolated LCA exerts limited endocrine disruption but may selectively compromise bone turnover and muscle anabolic response, suggesting that without acute LEA, LCA has limited influence on muscle oxidative phenotype.

Energy deficit is a potent physiological stressor that has shaped human evolution and can improve lifespan and healthspan in a wide range of species. Preserving locomotive capacity was likely essential for survival during the human hunter-gatherer period but surprisingly little is known about the molecular effects of energy deficit on human skeletal muscle, which is a key tissue for locomotion and metabolic health. Here we show that a 5-day 78% reduction in energy availability with concomitant aerobic exercise in healthy men leads to a profound modulation of skeletal muscle phenotype alongside increases in fat oxidation at rest and during exercise and a 2.1 ±0.8 kg loss of fat free mass and 0.8 ±0.6 kg of fat mass. We used stable isotope (D ₂ O) labelling and peptide mass spectrometry to investigate the abundance and turnover rates of individual proteins. Abundance (1469 proteins) and synthesis rate (736 proteins) data discovered a shift toward a more oxidative phenotype and reorganisation of cytoskeleton and extracellular matrix structure during energy deficit. Mitochondrial components: TCA, electron transport chain and beta-oxidation, were prominently represented amongst proteins that increased in abundance and synthesis rate, as well as proteins related to mitochondrial proteostasis, remodelling and quality-control such as BDH1 and LONP1. Changes in muscle metabolic pathways occurred alongside a reduction in extracellular matrix proteins, which may counteract the age-related muscle fibrosis. Our results suggest that muscle metabolic pathways are not only preserved but positively affected during periods of concomitant low energy availability and exercise.