17.10.2026 · 8 min read
An ordinary energy deficit describes the difference between all the energy taken in and all the energy expended. Energy availability answers a narrower question: how much energy is left for physiological functions after the cost of training has been subtracted. In studies it is often expressed as kilocalories per kilogram of fat-free mass per day:
dietary energy − exercise energy expenditure, divided by fat-free mass.
Two athletes can have a similar daily deficit but different energy availability: one cut food intake and kept a large training volume, the other cut both calories and training. For the body these are not necessarily the same situation.
The International Olympic Committee links prolonged, problematic low energy availability to relative energy deficiency in sport (REDs). It can affect bone, reproductive and immune function, metabolism, recovery and athletic performance. A brief, managed energy shortfall and a chronic problematic state are not synonyms, however: magnitude, duration, repetition and individual vulnerability all matter.
Controlled experiments had already shown that a few days of sharply reduced energy availability can change markers of bone turnover. In the work of Papageorgiou and colleagues, 11 women and 11 men went through five-day regimens with an availability of 45 or 15 kcal/kg of fat-free mass per day. With energy restriction, P1NP fell, which is consistent with suppressed bone-formation processes.
A pilot study by Murphy and colleagues tested low energy availability in seven men, with both normal and increased protein content. The high-protein diet did not prevent the suppression of bone-formation markers or the rise in resorption markers. However, seven participants is too small a sample to conclude with confidence that protein cannot change the response at all.
The broader literature and the IOC consensus underline one more detail: when calories are cut, it is not only total energy that changes. The availability of carbohydrate or other nutrients often falls at the same time. That is why the effect of "low energy" is hard to separate completely from the composition of the diet.
The randomised crossover experiment involved 19 healthy adults — 9 women and 10 men; 16 completed all the trials. Each participant went through three five-day regimens in turn, separated by at least nine days:
In the protein-preserved regimen, carbohydrate and fat had to be cut more sharply to reach the same low calorie level. This matters: the comparison was not "protein versus no protein" but two versions of very low energy availability with a different distribution of macronutrients.
On the first and fifth days participants did 45 minutes of cycle ergometer riding at 65% of peak power, followed by a Wingate test. Blood was drawn fasted, before and after exercise, and during a two-hour recovery. The main outcomes were P1NP and CTX-1. P1NP is linked to type I collagen synthesis and is used as a reference marker of bone formation; CTX-1 reflects collagen breakdown and bone resorption. Parathyroid hormone and inflammatory cytokines were also measured.
After both low-energy-availability regimens, fasting P1NP was lower than after the adequate-energy regimen. In the protein-preserved variant the decrease persisted during recovery from exercise as well. CTX-1 changed less consistently: it rose convincingly only in the recovery period after the protein-preserved regimen.
Several cytokines — including IL-6, IL-8, IL-10, IL-1RA and IL-1β — were higher at certain time points after both energy-deficient regimens. The signal for IL-6 was the most consistent. But the design does not allow the conclusion that the inflammatory response caused the changes in bone markers: both processes may have run in parallel.
Preserving protein did not weaken the observed responses. Moreover, some deviations looked more pronounced in that very regimen. The authors cautiously suggest that the cause may have been the sharper reduction in carbohydrate rather than protein itself. Direct comparisons between the two low-energy diets mostly remained inconclusive, so it cannot be claimed that preserved protein "worsened" bone status.
Bone is constantly remodelled. Osteoblasts take part in creating new tissue, osteoclasts in resorbing old tissue. P1NP and CTX-1 give a quick, indirect view of these processes, but they are not a snapshot of skeletal strength.
P1NP is not absolutely specific to bone: type I collagen is also present in tendons, ligaments and skin. A fall in P1NP is consistent with a reduction in osteoblast-driven synthesis of bone matrix, but part of the signal could theoretically originate from other tissues.
The main limitation is time. Five days is too short to measure a meaningful change in mineral density or fracture rate. It is biologically plausible that prolonged and repeated suppression of bone formation could damage the skeleton over time, but this experiment did not test that outcome.
Protein is essential for muscle and the bone matrix, and when body mass is being reduced, adequate intake helps preserve fat-free mass. But amino acids are only one resource. In a serious energy shortfall the body redistributes energy toward more urgent functions. Hormonal signals, carbohydrate availability, recovery and training capacity may change at the same time.
So the advice "keep protein high" is useful but incomplete. It does not turn an aggressive cut into a physiologically neutral process. Moreover, if calories are fixed at a very low level, preserving protein automatically leaves less room for carbohydrate and fat. For an athlete this can affect the quality of high-intensity work and other adaptations.
The strengths of the study are the randomised crossover design, controlled feeding, the same participants in all three conditions and analysis at several time points. The result is consistent with previous controlled experiments.
The limitations are substantial: the sample is small; exposure lasted five days; healthy adults were studied rather than athletes in a full training cycle; biomarkers were measured rather than bone density or fractures; and the study lacked the power for a reliable analysis of differences between the sexes. Finally, in the protein-preserved regimen the other macronutrients were cut more sharply, so the pure effect of protein cannot be separated out.
The new study does not set a universal "safe" rate of weight loss. It supports a more cautious principle:
This is not individual medical advice. Adolescents, athletes with a history of stress fractures and people with eating disorders should be especially cautious.
The work of Dolan et al. reinforces rather than overturns existing knowledge. With sharply reduced energy availability, preserving 1.5 g of protein per kilogram of body mass did not prevent the short-term suppression of a bone-formation marker or the amplification of part of the cytokine response to exercise. Protein remains important, but it does not offset the whole physiological cost of an energy shortage.
Tags: protein, sports nutrition, RED-S, bone health, energy availability
Research on training, nutrition and recovery: the key points worth knowing. One email a week with the new Sport in Luxembourg articles: title, short description and a link to read.
We use cookies for visitor analytics. You can accept or decline their use. Learn more in our Privacy Policy
