Bone health in combat athletes

1. The question the literature does not answer

Start with the shape of the gap, because everything downstream is an approximation to it.

To answer "what did eight cuts do to my skeleton," a study would have to recruit combat athletes, scan them at a standardised facility, follow them through multiple competitive cycles over years, record each cut's magnitude and method, and scan them again with the same machine and the same technique. It would need enough athletes to separate the effect of the cutting from the effect of the training, and it would need controls who train the same way without cutting.

That study does not exist. Not in MMA, not in boxing, not in Muay Thai, not in BJJ, not in wrestling, not in judo, not in kickboxing. The closest thing in the literature follows elite judoists through one weight cycle. Everything else is either a cross-sectional snapshot of what combat athletes' bones look like at a moment, or a controlled laboratory study of five days of restriction in runners and cyclists who are not combat athletes at all.

This is worth being blunt about, because the vacuum attracts confident answers. "Weight cutting causes osteoporosis in fighters" is not a finding. No study supports it. The only weight-cycling study with scans found density unchanged, and the only retired-combat-athlete dataset found better bone than in controls. Equally, "combat sports build bone so cutting doesn't matter" is not a finding either, because nobody has isolated the cutting from the training in anyone.

What follows is the evidence that exists, arranged by what it can and cannot support.

2. What "bone health" is actually measured as

Four instruments appear in this literature and they answer different questions.

Bone mineral density by DXA. A scan that measures mineral content per unit area at defined sites — lumbar spine, total hip, femoral neck, whole body. It is the standard clinical measure and it is slow: bone density changes over months to years, not weeks.

Z-score and T-score. A Z-score compares a person's density to others of the same age and sex. A T-score compares them to young-adult peak density. For a premenopausal woman or a man under 50, the Z-score is the relevant statistic, and it is the one the IOC's assessment tool uses. T-scores in young athletes produce impressive-looking numbers that do not mean what they appear to — a point this article returns to.

Bone turnover markers. Blood or urine measures of the two processes running constantly in bone. Formation markers include P1NP, P1CP and osteocalcin. Resorption markers include β-CTX and NTX. These move within days, which is why every controlled energy-restriction study uses them and why every such study is short.

Bone stress injury. The clinical outcome: a stress reaction or stress fracture, classified by site into high-risk (femoral neck, sacrum, pelvis) and low-risk. This is what an athlete actually experiences, and it is the outcome with the least combat-sport data of all.

The distinction between the third instrument and the first two is the single most important thing to carry through the rest of this article. Markers are fast and available; density and fractures are slow and meaningful. The literature on energy restriction and bone is overwhelmingly about markers, for the obvious practical reason that you cannot keep someone in a controlled deficit for two years.

3. Five days in a laboratory: the restriction studies

The foundational work is Ihle and Loucks, published in 2004. Twenty-nine regularly menstruating, habitually sedentary young women of normal body composition were exercised on a treadmill and held for five days at energy availabilities of 45 kcal per kilogram of lean body mass per day — the balanced control — or 30, 20 or 10.

At 30 kcal/kgLBM/d, osteocalcin fell by 0.9 ± 0.3 ng/mL and P1CP by 16 ± 8 ng/mL, with no effect on the resorption marker NTX. At 10 kcal/kgLBM/d, osteocalcin fell 2.3 ± 0.5 ng/mL, P1CP fell 48 ± 13 ng/mL, and NTX rose 17 ± 4 nM BCE/mM Cr (p<0.01). P1CP declined linearly across the restricted conditions (p<10⁻⁶), and the authors noted that most of the osteocalcin suppression "occurred abruptly between 20 and 30 kcal/kgLBM/day."

Two things follow from that study and only two. Formation is suppressed at a less severe restriction than the one that raises resorption — the building side gives way first. And the number that the whole field now repeats, 30 kcal per kilogram of fat-free mass per day, originated here: in sedentary women, over five days, on a laboratory-controlled diet and a treadmill. Not fighters. Not men. Not a season.

The later trials extend the pattern and inherit the design limits. Papageorgiou and colleagues ran 11 eumenorrheic women and 11 men through a crossover of five days at 15 versus 45 kcal/kgLBM/d with daily treadmill running at 70% of VO2peak. In the women, β-CTX was higher (P=0.03) and P1NP lower (P=0.01) in the restricted condition. In the men, neither marker moved.

A narrative review that tabulates the surrounding literature — three-day studies, dose-response studies, a 50%-restriction study in eight men — finds the same general shape: formation down, resorption up in the more severe conditions, with the female data considerably more consistent than the male.

None of these studies involves a combat athlete, a dehydration-led cut, or a period longer than five days. They are the most rigorous evidence available and they describe an exposure that only partly resembles fight week.

4. The male evidence contradicts itself, and the contradiction is the finding

The most interesting thing in this literature is a disagreement inside one research group.

The crossover described above found a clear female effect and no male effect at all — β-CTX P=0.46, P1NP P=0.12 in 11 men. Read casually, that is "men are protected."

The same group then published a randomised crossover pilot in seven men: mean age 23.9 ± 1.5 years, body fat 13.4 ± 2.0%, VO2peak 42.6 ± 2.4 mL/kg/min, five days at 15 kcal/kg fat-free mass per day on either 0.8 or 1.7 g/kg bodyweight of protein, against a control condition at 40 kcal/kg fat-free mass per day, with supervised daily cycling. P1NP fell 14.9 ± 6.5% on the lower protein intake and 24.8 ± 6.2% on the higher one, against −4.4 ± 5.5% in control (p=0.04). CTX-I rose 6.9 ± 6.2% and 0.2 ± 2.1% respectively, against −8.3 ± 3.9% in control (p=0.04).

That is a clear male effect. The two studies point in opposite directions in men.

Neither wins, and the reasons they do not are asymmetric in a way worth saying out loud. The first study's male null is an underpowered null at n=11 — an absence of a detected effect in a small sample is not evidence that no effect exists, and it cannot license "men are protected." The second is a self-described pilot at n=7, with a different energy-availability denominator, a different exercise mode (cycling rather than treadmill running, and cycling is low-impact), and a different control condition. It is not a replication that succeeded where the first failed. It is a different experiment.

The defensible sentence is that two small five-day studies from the same group point in different directions in men, for reasons that include design differences as much as biology. Do not average them and do not pick one.

Behind both of them sits the structural problem the IOC names directly. Roughly 80% of the REDs literature is female, and the consensus states that only a fifth of recent original studies included male athletes. The female data does not transfer to men, and the male data is too thin to reassure them. Neither gap closes with a coefficient — and specifically, the male range the consensus mentions is not a scaled-down version of the female threshold:

"While a universal cut-off of 30 kcal/kg FFM/day as a threshold of LEA leading to some REDs outcomes in females is debated, such a cut-off or range at which males experience REDs-related symptoms is even less understood, but appears to be lower (eg, ~9 to 25 kcal/kg FFM/day)."

Read the whole sentence, including "debated" and "even less understood." It is a statement about how little is known, not a number to aim at.

5. The newest trial: jumping did not rescue it

The most natural defence a fighter reaches for is that fighting is impact. Bone responds to loading; a striker's tibia, a wrestler's spine and a judoka's femur take loading no runner approximates. Surely that covers it.

The newest controlled trial in this area tested exactly that logic and did not find it.

Twelve recreational female runners, mean age 26.4 ± 3.9 years, VO2max 45.3 ± 7.2 mL/kg/min, completed a randomised crossover: five days at 15 kcal per kilogram of fat-free mass per day, running daily, with or without 50 jumps a day added. β-CTX rose over time in both conditions — main effect of time P=0.005, from 0.32 ± 0.21 to 0.36 ± 0.15 ng/mL in the running condition and 0.38 ± 0.21 to 0.42 ± 0.19 ng/mL in the running-plus-jumps condition — with no interaction between condition and time (P=0.716). P1NP did not change (time P=0.517, condition P=0.187). Sclerostin was unchanged.

In that trial, adding high-impact loading did not blunt the rise in bone resorption caused by five days of severe energy restriction. The narrative review that proposed impact loading as a countermeasure in the first place says of its own hypothesis: "It is currently unclear whether this would be of benefit to long-term bone health."

The conditions matter, as always. Twelve recreational female runners, five days, fifty jumps. That is not a Muay Thai session and it is not a season. What it establishes is narrower than the headline and more useful than it: at least once, under controlled conditions, impact did not cancel the deficit. "Impact loading protects fighters' bones from weight cutting" is not something this literature supports.

6. What a bone marker is not

Every figure in the last three sections is a blood or urine measurement taken over five days, and there is one sentence that bounds all of them. The review that tabulates this literature states that the "relationships between acute changes in markers of bone formation and resorption … and long-term bone health are unclear."

That is not a hedge. It is the state of the field. Nobody has shown that a five-day rise in β-CTX predicts a fracture, a Z-score, a broken hand, or anything a fighter will ever feel. The markers are measuring something real about the rate of a process. What that process integrates to over three years of cutting has not been established in anyone, let alone in a fighter.

The same review explains why: "The lack of longitudinal and controlled trials is likely due to difficulties in accurately monitoring changes in energy availability … for periods long enough to observe structural bone change." Keeping people in a verified deficit long enough for their density to move is, in practice, close to impossible outside a clinical population.

So the marker studies are best read as evidence about direction under a severe short exposure, and as nothing at all about magnitude over a career. Anyone converting a five-day CTX change into a career-long bone loss estimate is doing arithmetic the sources refuse to do.

7. The other side: what combat athletes' bones actually look like

Now the scans, which point the other way.

Whole-body bone mineral density in male weight-classified athletes — wrestlers and judoka — has been reported as higher than in male endurance athletes and non-athletes, despite pre-competition rapid weight loss. That is a direction, reported at secondary confidence; the publisher blocked retrieval and the numbers, sites and sample sizes could not be verified, so no values appear here.

A cross-sectional DXA cohort of 135 athletes across sports reported the highest whole-body T-scores in professional MMA fighters, ahead of collegiate footballers and resistance-trained men. That figure circulates as "MMA fighters have the strongest bones in sport," and it should not.

Two reasons. First, a T-score compares a young athlete to young-adult peak density; for a premenopausal woman or a man under 50 it is the wrong statistic, and the IOC's own tool uses Z-scores. A high T-score in a 27-year-old athlete is partly a reference-range artefact. Second, a cross-sectional snapshot of who has dense bones says nothing about what weight cutting did to them — the fighters were never scanned before they started cutting, and the subgroup's size and sex breakdown could not be verified.

What survives from this section is modest and still meaningful: combat athletes, on the scans that exist, do not look like a population with damaged skeletons. That is a real observation about a real population. It is not a controlled comparison and it is not a longitudinal one.

8. The one study of a measured weight cycle

The closest the literature comes to the reader's question is a 2006 study of elite judoists.

Forty-eight elite judoists, male and female, were scanned by DXA at the lumbar spine, femur and total body and followed across the first weight cycle of the season. As a group they displayed high bone mineral density and an increased rate of bone formation. Precompetitive weight loss averaged 4 ± 0.3% of body mass and produced an acute rise in cortisol of 81% (P<0.05) and in the resorption marker CTx of 33% (P<0.0001) — a metabolic imbalance favouring breakdown. A regain of 4 ± 0.5% of body mass restored a positive balance favouring formation.

Bone mineral density was unaltered by the weight cycle.

Read that carefully, in both directions. The markers confirmed what the laboratory studies predict: a cut swings bone metabolism toward resorption, measurably, in real elite competitors rather than in sedentary volunteers. And the scans found no density change from it, with the marker swing reversing on regain.

The conditions: one cycle, about 4% of body mass, elite judoists of both sexes, published in 2006, with the sex split in the cycle result not obtainable. This article carries the study at secondary confidence — the abstract was obtained through a search-result summary because two publishers blocked direct retrieval.

One cycle at 4% is not eight cuts in three years, and 4% is a modest cut by the standards of the sport — see how much weight fighters actually cut for what the larger end of that distribution looks like. The judo study is the best evidence available and it covers a fraction of the exposure the question is about.

9. Retired martial artists, and what a cross-section cannot see

The only dataset located on combat athletes long after competition is a 2025 study of 50 men aged 40 to 50: 30 retired martial artists from taekwondo, boxing, judo, karate and kickboxing, all at least ten years since their last competition, against 20 age-matched non-athletes. DXA at the left femur (neck, Ward's area, trochanter, shaft and total) and the AP lumbar spine, L1–L4.

The Z-scores, retired athletes versus non-athletes: left femoral neck 0.32 versus −0.56 (p=0.001); left femur total 0.28 versus −0.29 (p=0.047); AP spine −0.15 versus −1.15 (p=0.0048). In the authors' words, "100% of retired athletes had normal bone density with a Z-score ≥ −1.9 and a T-score > −1.1," while among the non-athletes 65% were normal, 20% had abnormal density and 15% were osteopenic.

On its face that is reassuring, and it is the strongest long-horizon signal available. Four caveats have to travel with it.

It is men only, at n=30 versus 20. It is cross-sectional — one scan each, no before, no trajectory. It recorded no weight-cutting history at all, which means it cannot tell us whether the cutting cost anything; only that this group, as a group, was not osteopenic at 45. And it is subject to survivorship: the men who stayed healthy are the men who turn up for a study at 45. Anyone whose career ended badly, or who stopped identifying with the sport, is not in the sample.

The authors add their own limitations: "Small sample size may reduce the statistical power … making it challenging to identify significant differences. Due to budget constraints, participants were asked about vitamin D deficiency rather than conducting blood tests. Self-reported data are liable to recall bias and social desirability bias."

So: 30 retired men looked good on a scan. That is what the study establishes.

10. Why a whole-body number can hide a weak site

Loading is site-specific, and so is its absence.

The consensus makes the point explicitly in its discussion of para-athletes: problematic low energy availability can impair bone "secondary to factors such as altered skeletal loading experienced by para athletes (ie, the lack of loading stimulus experienced by wheelchair athletes and/or low-impact sports)," and in unilateral amputees "the affected limb may exhibit reduced bone mineral density." Bone responds where it is loaded.

Applied to combat sports, that principle cuts against the whole-body reassurance in section 7. A striker's tibia and a wrestler's spine take loading; other sites take much less. A strong total-body reading can sit on top of a weaker site, which is precisely why the IOC's assessment criteria ask for lumbar spine, total hip or femoral neck Z-scores rather than a whole-body figure.

One opinion article proposes that in combat athletes specifically, upper-limb bone density "appear[s] to be more strongly affected by prolonged low energy intake than lower limbs." That is a hypothesis published in an opinion piece with no primary data, no p-values and no comparative numbers, and it is reported here as an opinion and nothing more.

In weight-bearing endurance athletes with functional hypothalamic amenorrhoea, the pattern reported by the narrative review is lower density at all measured sites, most pronounced at hip and femoral neck, with cortical thickness reductions at the tibial epiphysis, femoral neck and trochanter and inconsistent trabecular findings. That is the closest analogue available, in a different sport and a different sex profile.

Nobody has scanned fighters' loaded and unloaded sites across a cut. That study would be straightforward to run and has not been run.

11. What has to change before a scan means anything

This is the section that stops someone panicking over a number.

Suppose a fighter gets two DXA scans a year apart and the second is lower. Is that bone loss?

Usually it is not. Every DXA facility has measurement error, and a change smaller than that error is noise. The IOC consensus prints the floor in a footnote to its assessment criteria:

"A true BMD decrease (from prior testing) is ideally assessed in comparison to the individual facilities DXA's LSC based on the facilities calculated coefficient of variation (%CV). As established by ISCD, at the very least, LSC should be 5.3%, 5.0%, and 6.9% for the spine, hip and femoral neck to detect a clinical change."

Those are minimums. A facility with worse precision needs a larger change. Below roughly 5% to 7% at those sites, a difference between two scans is measurement, not biology.

Two further conditions follow from the same footnote and from ordinary practice. The scans should be at the same facility on the same machine, because between-machine differences are not comparable in the way this calculation assumes. And the consensus asks for a scan within the previous six months when assessing current status, which means an old scan answers a question about the past.

The same section also warns that the normative datasets behind Z-scores come from able-bodied general populations, with real limitations in how population comparisons get reported. A Z-score is a comparison to a reference group, not a direct measure of how strong a bone is.

None of which means a lower second scan should be ignored. It means the person who interprets it is a clinician with the facility's own precision figures, not a fighter with a phone.

12. Bone stress injury, and the hole where the combat-sport data should be

Bone stress injury is the outcome that actually ends camps, and it is where this evidence base is emptiest.

The IOC's assessment criteria treat it seriously. As a primary indicator: a history of one or more high-risk bone stress injuries — femoral neck, sacrum, pelvis — or two or more low-risk ones in the previous two years, or an absence of six months or more from training due to one. As a secondary indicator: one low-risk injury within the previous two years together with six months or more out of training. Bone stress injury also appears in the REDs health model's bone category with evidence drawn from female athletes, male athletes and anorexia nervosa populations.

The high-risk site list — femoral neck, sacrum, pelvis — is derived from running populations. So is almost everything quantitative in this area. No bone stress injury incidence data specific to MMA, boxing, Muay Thai, BJJ, wrestling, judo or kickboxing was located. Not a null result. An absence.

That absence matters for the fighter reading this, because a bone stress injury is an injury with no incident: no bad landing, no caught kick, nothing anyone in the gym would remember, and a bone that has hurt for weeks. That category is the subject of the injuries that arrive without an incident, and the surveillance built to catch it has never been run in a combat sport either.

The consensus also lists what a clinician considers before attributing a stress injury to energy availability: "External reasons (eg, training errors, surface, shoes) or internal issues (eg, body build, medical predispositions …)." Risk factors reported generically across the wider literature include low energy availability, low BMI, low bone density and prior bone stress injury, with prior injury the strongest — but the effect sizes circulating for that were not pinned to a retrievable primary source here and no figure is printed.

13. Adolescents: the outcome is failure to accrue, not loss

Youth wrestling and youth combat sports raise a different question, and the consensus answers it differently.

For adults, impaired bone health under low energy availability shows up as loss. For children and adolescents, the impairment the consensus names is "longitudinal loss of BMD/lack of expected bone accrual (younger populations)" — and the paediatric assessment criterion is a Z-score below −1 at the lumbar spine or total body less head, or a decrease in Z-score from prior testing. Alongside it: "a negative deviation of a paediatric or adolescent athlete's previous growth trajectory (height and/or weight)" is a primary indicator, and deviation from the expected growth curve appears in the model as an impairment in its own right.

The point is structural. An adolescent skeleton is supposed to be gaining. Not gaining is the injury, and it does not announce itself as a loss on any scan. The consensus does not treat adolescents as small adults anywhere — its serious-medical-indicator thresholds are age-split, with an adolescent bradycardia threshold of under 50 bpm against under 40 for adults, for instance.

It is also restrictive about measurement in this group: "body composition assessment is recommended only for medical purposes under 18 years of age," with exceptional circumstances requiring consensus among the athlete's health and performance team and guardian consent.

And then the gap. Weight cutting in adolescent combat athletes is documented as a behaviour — one study of female adolescent combat athletes reported that 73% purposefully engaged in weight loss practices, with habitual loss averaging 8.0% of body mass, though those figures are held here at secondary confidence. Its skeletal consequences are not documented at all. No study of adolescent combat athletes' bone accrual across weight-cutting seasons was located. One paper appeared to be exactly that and turned out to be a mis-served file: the PDF hosted at the URL is a different study entirely, on an unrelated subject, so the citation could not be established and it is not used here.

14. Calcium, vitamin D, and the fighter's actual next step

Two randomised trials in military recruits tested a calcium and vitamin D combination during basic training. The larger one, in 5,201 female US Navy recruits over eight weeks of basic training, reported stress fracture incidence reduced by 25%. The other, in Army basic combat training over nine weeks, maintained parathyroid hormone, raised the osteoprotegerin-to-RANKL ratio and prevented increases in bone resorption markers; a related Marine recruit trial using a fortified snack bar reported the same marker-level effect but no change in tibial microarchitecture.

Those were recruits absorbing a sudden, uniform, very large increase in weight-bearing loading, with unknown baseline intakes — not fighters cutting weight. No trial has tested whether either nutrient protects the skeleton of someone cutting repeatedly. This article prints no amounts, because a dose inside a study protocol is still a dose once it sits next to "reduced stress fractures by 25%," and the extrapolation from a recruit to a fighter in a deficit is one the trials do not license.

The consensus itself frames these nutrients the other way round. "Poor micronutrient intake (eg, calcium and vitamin D)" appears in its list of differential diagnoses to exclude for low bone density, alongside genetic bone disorders, hyperparathyroidism, malabsorption such as coeliac disease, malignancies, renal disease and medications including anabolic steroids. They are things a clinician rules out, not things an athlete self-treats. The broader question of what a long deficit costs at the micronutrient level is its own subject — what a long deficit costs beyond calories covers it.

It is also worth knowing how the one long treatment trial went. In the 12-month intention-to-treat randomised trial the consensus cites, which increased energy intake in exercising females with REDs-related biomarkers, the drop-out rate was 57% and there was improvement in some outcomes — resumption of menstrual function in selected participants — "but not all symptoms (eg, inability to retard bone loss)." Eating more, over a year, in the only trial of that length, did not stop the bone loss.

So where does a worried fighter actually go? To a physician. The findings that put bone into the consensus's moderate or high categories — a Z-score below −1 at the lumbar spine, total hip or femoral neck; one high-risk or two low-risk bone stress injuries in two years; six months out of training for one — are clinical assessments made by clinicians. The tool carries its own printed warning:

"these guidelines are not to be used in isolation and are not to be solely used for diagnosis. Furthermore, these guidelines are less reliable when it is impossible to assess all indicators in table 4. These guidelines are not a substitute for professional clinical diagnosis, advice and/or treatment from a team of REDs health and performance experts led by a physician"

It is not something a fighter scores themselves, and nothing in this article is either.

That is the limit of everything a fighter can track at home, and it is worth stating plainly because the tools that feel most precise are the ones furthest from this question. A body-fat figure and a bodyweight trend describe soft tissue. Bone is not in the readout. A fighter who has had a stress injury, who has restricted hard for years, or who simply wants to know, needs a referral and possibly a DXA scan at a facility that can tell them its own precision figures. Fighter Cut logs what happened in a camp; it does not and cannot assess a skeleton.

What we could not verify

This section is longer than usual because the subject is mostly absence, and naming the absences precisely is more useful than the findings around them.

• The reader's actual question: what repeated cuts do to bone. No study in any combat sport has measured cumulative cut count against any bone outcome. There is no dose-response between number of cuts and density, fracture, marker or Z-score. The nearest study covers one cycle in elite judoists. No number for eight cuts exists, and any number offered for it is invented.
• Bone stress injury rates in combat sports. None exist for MMA, boxing, Muay Thai, BJJ, wrestling, judo or kickboxing. Every quantified figure in that area comes from runners, endurance athletes, military recruits and clinic-referred populations. The high-risk site list in the IOC criteria is derived from running.
• Adolescent combat athletes' bone accrual. No study located. The paper that appeared to be exactly that — a seasonal bone density study in high school wrestlers — could not be cited, because the PDF served at the hosting URL is a different paper entirely on an unrelated subject. The citation could not be established and it is refused rather than reported. The weight-cutting prevalence figures for adolescent combat athletes (73% engaging in weight loss practices; habitual loss averaging 8.0% of body mass) were obtained at search-summary level only and are held at secondary confidence.
• The reverse sex gap, which runs in both directions at once. Roughly 80% of the REDs literature is female, and the IOC says only a fifth of recent original studies included male athletes — so the general evidence base is female and its transfer to men is uncertain. Meanwhile the combat-sport bone data is the opposite: the retired-athlete study is men only, the weight-classified BMD comparison is male, and the judo weight-cycling study includes both sexes but its cycle result could not be obtained with a sex split. The field has the usual gap in one place and its mirror image in the other, and the two do not cancel. Neither closes with a coefficient.
• Whether men are affected by low energy availability at the bone level. Two studies from the same research group reach opposite conclusions. The 2017 crossover found a female effect and a male null at n=11 per sex (β-CTX P=0.46, P1NP P=0.12 in men). The 2021 pilot found a clear male suppression at n=7, with a different energy-availability denominator, a different exercise mode and a different control condition. This article does not average them, does not choose between them, and does not read the null as protection, because an underpowered null is not evidence of absence.
• Whether five-day marker changes mean anything over a career. The review that tabulates this literature says the relationships between acute marker changes and long-term bone health "are unclear," and explains the reason: nobody can keep athletes in a verified deficit long enough to observe structural bone change. No marker figure in this article has been shown to predict a fracture, a Z-score or anything a fighter can feel.
• Any energy-availability figure as a safe floor. The 30 kcal/kg fat-free mass per day threshold came from short-term laboratory work in a small sample of sedentary women and was intended as a guide rather than a diagnostic endpoint. The consensus's own position is that "there are risks in setting a definitive clinical threshold of EA due to many moderating factors," and that real-world observation has "identified large differences in the EA level associated with health and performance concerns between individuals, the sexes, and among different body systems." The male range it mentions is described as "even less understood." No number here is a target and no body-fat percentage appears as one either.
• Numbers from two papers that could not be retrieved. The Clinical Nutrition ESPEN comparison of weight-classified, endurance and non-athlete men returned publisher errors on repeated attempts, so only its direction of effect is reported and no values, sites or sample sizes. A 2026 Sports Medicine paper directly on sport-specific impact loading and bone density in relation to REDs assessment is paywalled; its DOI and title verify in Crossref, so it is listed as existing, but nothing from its contents appears here.
• The T-score figure for professional MMA fighters. Reported at secondary confidence from a cross-sectional cohort whose MMA subgroup size and sex breakdown could not be verified — and a T-score is the wrong statistic in a young athlete regardless. No headline about fighters having the strongest bones in sport is supportable from it.
• Site-specific effects in combat athletes. What exists is one opinion article's hypothesis about upper limbs, with no primary data, plus the general loading principle from the consensus's para-athlete discussion. Nobody has scanned fighters' loaded and unloaded sites across a cut.
• Whether the method of cutting matters. A dehydration-led cut and an energy-restriction-led cut are physiologically different exposures, and no bone study located separates them. The judo study reports about 4% of body mass lost by unspecified means.
• Amateurs, and anyone with a day job. Every combat-sport bone dataset located is elite, collegiate, professional or retired-elite. The amateur training four nights a week around work, cutting eight times in three years, appears in no cohort here.
• Any prior-injury odds ratio. The frequently quoted threefold figure for a prior bone stress injury raising the odds of a future one could not be pinned to a retrievable primary source, so no figure is printed.
• Anything about you. This article describes a body of evidence and the holes in it. It does not diagnose, it does not assess a scan, and it does not override a physician, a coach or gym medical staff.

Sources

Sourced to

1. 2023 International Olympic Committee's (IOC) consensus statement on Relative Energy Deficiency in Sport (REDs) — Mountjoy M, Ackerman KE, Bailey DM, Burke LM, Constantini N, Hackney AC, Heikura IA, Melin A, Pensgaard AM, Stellingwerff T, Sundgot-Borgen JK, Torstveit MK, Jacobsen AU, Verhagen E, Budgett R, Engebretsen L, Erdener U, British Journal of Sports Medicine, 2023;57(17):1073–1097. DOI 10.1136/bjsports-2023-106994, PMID 37752011
2. Dose-Response Relationships Between Energy Availability and Bone Turnover in Young Exercising Women — Ihle R, Loucks AB, Journal of Bone and Mineral Research, 2004;19(8):1231–1240. DOI 10.1359/JBMR.040410
3. Effects of reduced energy availability on bone metabolism in women and men — Papageorgiou M, Elliott-Sale KJ, Parsons A, Tang JCY, Greeves JP, Fraser WD, Sale C, Bone, 2017;105:191–199. DOI 10.1016/j.bone.2017.08.019, PMID 28847532
4. Low Energy Availability with and without a High-Protein Diet Suppresses Bone Formation and Increases Bone Resorption in Men: A Randomized Controlled Pilot Study — Papageorgiou M, Martin D, Colgan H, Cooper S, Greeves JP, Tang JCY, Fraser WD, Elliott-Sale KJ, Sale C, Nutrients, 2021;13(3):802. DOI 10.3390/nu13030802, PMID 33671093
5. Effects of High-Impact Exercise on Bone Marker Concentrations During Controlled Low Energy Availability in Recreational Female Runners — Calcified Tissue International, 2026;117. DOI 10.1007/s00223-026-01564-0, PMCID PMC13303319
6. Effects of Low Energy Availability on Bone Health in Endurance Athletes and High-Impact Exercise as a Potential Countermeasure: A Narrative Review — Hutson MJ, O'Donnell E, Brooke-Wavell K, Sale C, Blagrove RC, Sports Medicine, 2021;51(3):391–403. DOI 10.1007/s40279-020-01396-4, PMCID PMC7900047
7. Comprehensive health assessment of retired martial arts athletes: bone density, dietary intake, physical activity, and wellbeing — Alshaer T, Battikhi N, Amawi AT, Trabelsi K, Jahrami H, Bouedo P, Ghazzawi HA, Frontiers in Aging, 2025;6:1513936. DOI 10.3389/fragi.2025.1513936
8. Bone Density in Elite Judoists and Effects of Weight Cycling on Bone Metabolic Balance — Prouteau S, Pelle A, Collomp K, Benhamou L, Courteix D, Medicine & Science in Sports & Exercise, 2006;38(4):694–700. DOI 10.1249/01.mss.0000210207.55941.fb, PMID 16679985 (obtained at secondary confidence; direct retrieval blocked by two publishers)
9. Bone mineral density in male weight-classified athletes is higher than that in male endurance-athletes and non-athletes — Clinical Nutrition ESPEN, 2020. DOI 10.1016/j.clnesp.2020.01.006 (publisher returned errors on repeated attempts; direction of effect only, no values taken)
10. Bone Mineral Density in Competitive Athletes — Antonio J, Leaf A, et al., Journal of Exercise and Nutrition, 2018. DOI 10.53520/jen2018.10314 (search-summary level; T-score figures not used as a health claim, see article text)
11. Differential Risks of the Duration and Degree of Weight Control on Bone Health and Menstruation in Female Athletes — Uchizawa A, Kondo E, Lakicevic N, Sagayama H, Frontiers in Nutrition, 2022;9:875802. DOI 10.3389/fnut.2022.875802 (Opinion article; no primary data)
12. Calcium and Vitamin D Supplementation Decreases Incidence of Stress Fractures in Female Navy Recruits — Lappe J, Cullen D, Haynatzki G, Recker R, Ahlf R, Thompson K, Journal of Bone and Mineral Research, 2008;23(5):741–749. DOI 10.1359/jbmr.080102 (search-summary level)
13. Calcium and vitamin D supplementation maintains parathyroid hormone and improves bone density during initial military training: a randomized, double-blind, placebo-controlled trial — Gaffney-Stomberg E, Lutz LJ, Rood JC, Cable SJ, Pasiakos SM, Young AJ, McClung JP, Bone, 2014;68:46–56. DOI 10.1016/j.bone.2014.08.002, PMID 25118085 (search-summary level)
14. Sport-Specific Impact Loading is Associated with Bone Mineral Density in Athletes: Implications for Relative Energy Deficiency in Sport (REDs) Assessment — Sports Medicine, 2026. DOI 10.1007/s40279-026-02478-5 (DOI and title verified in Crossref; full text paywalled and not retrieved, no contents used)
15. Rapid weight loss practices in Olympic combat sports: a systematic review and meta-analysis — Frontiers in Physiology, 2022;13:830229. DOI 10.3389/fphys.2022.830229 (search-summary level; prevalence figures held at secondary confidence)

Read next