A dual-energy X-ray absorptiometry (DEXA) scan breaks your body composition down into fat, lean soft tissue, and bone. At Longevity Benchmark, we often see readers treat a single scan as a final verdict. It's more useful as the starting point of a trend. The scan measures all three compartments in each arm and leg, as well as the trunk, and provides an estimate of visceral fat in the abdomen.
The Verdict
What a DEXA Scan Body Composition Report Measures
A DEXA report gives you three compartments broken out by region: fat mass, lean soft tissue and bone mineral. Hologic describes its own Horizon system as measuring bone density, lean mass and fat mass, and GE HealthCare Lunar machines report the same three plus a visceral estimate.
The mechanism explains most of the caveats that follow. Two X-ray energies pass through you, and each pixel of the image is solved as a two-component problem. Where the pixel contains bone, the software separates mineral from the soft tissue around it. Where it contains only soft tissue, it separates fat from lean. Bone covers roughly 40 to 45% of the scan area, and the fat in those pixels is never measured. It is interpolated from the fat-to-lean ratio in the soft-tissue pixels immediately beside them.
| What the report shows | What it is | How to read it |
|---|---|---|
| Total fat mass and percent body fat | The whole-body sum, in kilograms and as a share of your total mass | Treat it as a starting line you re-measure. The absolute figure carries every assumption the model makes |
| Lean soft tissue, by region | Everything that is neither fat nor bone: muscle, organs, connective tissue, and the water inside them | It moves with hydration and glycogen, so it is the number most sensitive to how you turned up on the day |
| Appendicular lean mass | The lean soft tissue in both arms and both legs, added together | This is the muscle figure used in sarcopenia criteria, and the one worth watching against a training block |
| Bone mineral content and density | How much mineral the skeleton holds, and its density per unit area | A body composition scan is not a bone diagnosis on its own. That needs the hip and spine scored against a reference population |
| Android and gynoid fat | Fat carried around the abdomen against fat carried around the hips and thighs | The ratio between them describes fat distribution, which tracks metabolic risk more closely than the whole-body total |
| Estimated visceral adipose tissue | A calculation that subtracts estimated subcutaneous fat from the android region total | Each manufacturer uses its own algorithm, so a GE HealthCare figure and a Hologic figure are not interchangeable |
Bone deserves one clarification, because a body composition scan and a bone density scan are not the same appointment. The whole-body scan gives bone mineral content, and osteoporosis is diagnosed from dedicated hip and spine imaging scored against a reference population. Our page on what a DEXA T-score means covers that scoring.
Precision and Accuracy Are Different Properties
DEXA is highly precise and only approximately accurate, and the two properties lead to different uses. Precision means that scanning the same person twice on the same machine on the same afternoon returns nearly the same answer. Accuracy means the answer matches the composition of the body itself, which no measurement in a living person can confirm.
Those precision figures turn into a threshold you can use. ISCD asks technologists to compute their own precision error and then multiply it by 2.77 to get the least significant change, the smallest gap between two scans that is unlikely to be noise. Take someone carrying 55 kg of lean tissue on a technologist working at the 2% minimum. That is a precision error near 1.1 kg, and a least significant change near 3 kg. A well-drilled technologist beats the minimum by a wide margin, so ask what their measured precision is rather than assuming the floor.
Accuracy is limited by the model itself. DEXA assumes fat-free tissue is about 73% water, and a 2020 review in Quantitative Imaging in Medicine and Surgery puts the resulting fat error near 1% for every 5% shift in hydration. A crossover trial published in Medicine and Science in Sports and Exercise showed the practical version. Loading muscle glycogen and creatine changed DXA lean mass estimates in trained cyclists over 48 hours, a window far too short for real muscle protein to change. Anyone starting creatine between two scans will see lean mass rise partly because creatine draws water into muscle.
The Protocol That Makes Two Scans Comparable
Standardise six things and the number becomes usable between visits. None of them cost money.
- Same machine, same clinic. Differences between manufacturers are larger than most changes you are trying to detect, so a second scan elsewhere restarts your baseline.
- Morning, after an overnight fast. A meal and its fluid are counted as lean soft tissue in the trunk.
- Before training, never after. A hard session shifts fluid between compartments and out through sweat.
- Similar hydration. Drink normally the evening before, and do not scan after a sauna, a long run or a heavy salt night.
- Same clothing, no metal. Metal reads as extremely dense tissue and distorts the region around it.
- Same positioning. Ask for the positioning aids and, if the clinic allows it, the same technologist.
In the guides we publish here, the most common way a tracking plan comes apart is not the protocol at all. It is the second scan, booked six months later at whichever clinic had a slot that week, on a machine from a different manufacturer. That single decision throws away the comparability that made the first scan worth paying for. Decide where you will scan for the next two years before you book the first one, and check the running cost on our page covering what a DEXA scan costs.
DEXA Scan Body Composition Against the Alternatives
Every alternative to DEXA trades accuracy for access, and each one trades it differently. InBody and similar bioelectrical impedance analysis (BIA) devices send a small current through you and infer the compartments from the resistance. The Bod Pod from COSMED measures body volume by air displacement inside a sealed egg-shaped chamber. Hume Health, Withings and the other smart-scale makers put the same impedance method into a bathroom device with an app attached.
| Method | What it measures | How it compares with DEXA | What it is good for |
|---|---|---|---|
| DEXA | X-ray attenuation at two energies, resolved into three compartments and reported by body region | The reference standard the other methods on this list are validated against | Tracking lean mass, regional fat and bone density in one sitting |
| InBody and other BIA devices | Resistance to a small alternating current, converted into compartments by a proprietary equation | A study in 795 active-duty US Marines found the InBody 770 read body fat 3.4 percentage points below DXA, and tracked DXA closely enough to explain 96% of the variation between people | Frequent cheap re-measurement when you only need the direction of travel |
| Bod Pod | Body volume by air displacement, converted to whole-body density and then to fat | A comparison published in PLOS One found it read higher than DXA in thinner people and lower in heavier people | One body fat figure without any X-ray exposure at all |
| Smart scales | Foot-to-foot or hand-to-foot impedance in a home device, with an app on top | Hume Health says on its own site that impedance readings are sensitive to hydration, timing and calibration and can differ systematically from DXA | A daily weight and a slow trend line, taken under identical conditions each morning |
| Skinfold calipers | The thickness of pinched subcutaneous fat at set sites, run through a regression equation | Almost entirely dependent on who holds the calipers, and blind to fat inside the abdomen | A cheap repeatable measure when the same trained person takes it every time |
DEXA Scan vs InBody
InBody is wrong by a fairly consistent amount, which makes it good for a weekly trend and unreliable for a one-off verdict. A DEXA reading in January against an InBody reading in April does not describe a change in your body, because the roughly 3-point offset between the devices is larger than most genuine quarterly progress. Use DEXA for the baseline, InBody for the frequent check, and never subtract one from the other.
DEXA Scan vs Bod Pod and Smart Scales
The Bod Pod gives a whole-body fat percentage with no X-ray exposure and no regional breakdown, so it says nothing about where your fat sits or how much muscle is in each limb. Smart scales give you a reading every morning for a one-off hardware cost, which beats four data points a year for catching a trend early. They carry the same hydration sensitivity without a technologist to contain it. Weigh at the same time each morning, after the bathroom and before food, and read the fortnightly average.
Where the Visceral Fat Figure Comes From
DEXA estimates visceral adipose tissue rather than imaging it, and it does so inside the android region of the abdomen. That region is defined automatically, with its lower limit at the top of the iliac crest and its height set to 20% of the distance from there to the base of the skull. The software estimates the subcutaneous fat in that band and subtracts it from the total, leaving what sits deeper.
The estimate is good. A 2012 validation study in Obesity compared DEXA visceral fat against computed tomography (CT) and reported an R² of 0.96 with a mean difference of 56 cm³. Visceral fat earns the attention because it drains into the portal vein and behaves differently from subcutaneous fat, which is why waist size predicts metabolic risk better than weight does.
What to Wear and How Long a DEXA Scan Takes
A body composition scan asks nothing of you beyond an overnight fast and clothing with no metal in it. Wear cotton clothing with no metal, remove jewellery and empty your pockets, or take the gown if the clinic offers one. Expect to lie flat and still for a few minutes while the scanning arm passes above you, with nothing touching you and no enclosed space involved. Fast overnight and drink normally, which is a standardisation choice rather than a medical requirement.
Radiation is the question readers ask most and worry about least once they see the number. The 2020 Quantitative Imaging review puts a whole-body scan on a current machine at roughly 4 to 5 microsieverts, against natural background exposure of about 6.7 microsieverts a day. A scan costs you less than a day of simply existing. Pregnancy is the exception, and a suspected pregnancy is a reason to postpone rather than to weigh the dose. Mention any recent barium study or nuclear medicine scan when you book, because both interfere with the measurement for days afterwards.
Appendicular Lean Mass in Sarcopenia Diagnosis
In consensus case definitions, low muscle strength comes first. The mass measurement confirms it. A lean-mass reading on its own cannot diagnose sarcopenia, however low the number appears. Diagnostic frameworks evaluate appendicular lean mass instead of total body lean mass because limb measurements isolate locomotor skeletal muscle. Total lean mass includes visceral organs and trunk water shifts that conceal true muscle loss. Dividing appendicular lean mass by height squared creates the appendicular skeletal muscle mass index, which adjusts muscle volume for stature.
Published cutpoints vary across international consensus panels. The European Working Group on Sarcopenia in Older People defines low appendicular muscle mass as under 20 kg for men and under 15 kg for women. Adjusted for height, their thresholds sit at under 7.0 kg/m² for men and under 5.5 kg/m² for women. Asian Working Group for Sarcopenia criteria, as reproduced in regional research, set height-adjusted scan cutpoints at under 7.0 kg/m² for men and under 5.4 kg/m² for women. The Foundation for the National Institutes of Health Sarcopenia Project established cutpoints of under 19.75 kg for men and under 15.02 kg for women. By contrast, the Sarcopenia Definition and Outcomes Consortium excluded scan-derived lean mass from its criteria because low tissue mass did not predict adverse health outcomes in community-dwelling older adults.
A scan result below these thresholds is a prompt for functional evaluation instead of an immediate diagnosis. Formal diagnosis belongs to a physician and begins with muscle strength, which clinicians evaluate using grip strength or a timed chair rise. When physical strength remains normal, low muscle quantity does not meet the standard for confirmed disease. Readers with low readings should consult a clinician to review the full sarcopenia diagnostic sequence and determine appropriate medical care.
Who Should Skip a DEXA Scan Body Composition Test
Anyone who wants a single body fat percentage and has no intention of scanning again should not book one. A one-off number is the weakest thing DEXA produces, because the model limits absolute accuracy far more than it limits repeatability. Take a waist measurement at a fixed landmark, a morning weight on the same scale, and two photographs in the same light and position, then repeat all three monthly. That costs nothing and answers the question a single scan was going to answer badly.
Two other groups get more elsewhere. Anyone whose real question is functional capacity should measure grip strength and a timed sit-to-stand, because a person can hold good lean mass and still be losing function. Anyone chasing bone rather than muscle wants a hip and spine scan with a score, which our page on when to get a DEXA scan sets out by age and risk.
Cross-machine calibration would change our answer. If manufacturers published a conversion that made a GE HealthCare result and a Hologic result comparable within the least significant change, the same-machine rule would relax and a one-off scan would become genuinely portable. Nothing like that exists today, and the visceral algorithms sit further apart than the fat and lean ones do.
So set the cadence before the first appointment. Twelve weeks is long enough for a training block to clear the least significant change and short enough to correct course. Readers pairing scans with a protein target usually want our page on protein timing alongside it. Book your DEXA scan body composition test for a morning slot at one clinic. Note the machine model and the technologist's name, and put the repeat in your calendar before you leave the building.
Frequently Asked Questions
Does a DEXA scan measure body fat?
Yes, as a mass in kilograms and as a percentage, for each arm, each leg, the trunk and the whole body. It gets there indirectly. Two X-ray energies pass through you, and each pixel is solved for two components at a time. That separates fat from lean where there is only soft tissue, and mineral from soft tissue where there is bone. Fat sitting behind bone is never measured. It is interpolated from the pixels beside it, which is one reason two machines can disagree about the same person on the same afternoon.
Is a DEXA scan accurate for body fat?
It is the most accurate method you can book, and it is still an estimate rather than a true value. The model assumes fat-free tissue holds a roughly fixed share of water, near 73%, and the 2020 Quantitative Imaging review puts the resulting error near 1% for every 5% that hydration shifts. Direct chemical analysis is the only true reference and nobody offers it to living customers. For deciding whether fat mass fell over twelve weeks, DEXA is the right tool. For quoting one true number, nothing available delivers that.
Can DEXA scans be inaccurate?
Yes, in three ways worth knowing about. Manufacturers apply different algorithms, so a GE HealthCare result and a Hologic result are not directly comparable. A meal, a litre of water or a carbohydrate load adds mass the software books as lean soft tissue. And anyone too tall or too broad for the scanning field gets a partial scan with the missing side estimated by mirroring. ISCD asks each technologist to measure their own precision for exactly this reason, so ask yours what theirs is.
Does a DEXA scan show muscle mass?
It shows lean soft tissue, which is close to muscle without being the same thing. Lean soft tissue also includes organs, connective tissue, blood and the water held inside all of it. The figure closest to skeletal muscle is appendicular lean mass, the lean tissue in the four limbs added together, because limbs carry almost no organ tissue. Sarcopenia criteria use that number, and it responds to a training block, which makes the limb-by-limb breakdown more useful than the whole-body lean figure.
Should a DEXA scan be done on an empty stomach?
Fasting is not required for safety, and it is worth doing for comparability. Food and fluid sit in the trunk as mass the software must classify, and most of it lands in lean soft tissue. A large breakfast and half a litre of water can look like lean gain that took ten minutes to acquire. An overnight fast with a little water, repeated the same way before every scan, removes that variable. A clinic telling you fasting is unnecessary is right about the medicine and beside the point on tracking.
What should you wear for a DEXA scan?
Wear light clothing with no metal in it: cotton shorts and a t-shirt, no zips, buttons, belt buckles, underwire or hooks. Take off jewellery, watches and hair clips, and empty your pockets. Metal absorbs X-rays strongly and the software reads it as very dense tissue, which distorts the whole region it sits in. Some clinics hand out a gown to sidestep the problem. Wearing the same outfit every time removes one more difference between visits and costs nothing.
How long does a DEXA scan take?
The scan itself is a few minutes of lying flat and still on a padded table while an arm passes above you. There is no tunnel, no noise and nothing touches you. Most of the appointment is paperwork, changing, positioning and the walk through your results afterwards. Ask what slot the clinic books when you make the appointment, because a place scheduling ten minutes end to end is not going to explain the report to you.
Is there a DEXA scan home scale?
No. DEXA needs an X-ray source, a detector array and a licensed operator, and no home device reproduces that measurement. Scales marketed with DEXA comparisons use bioelectrical impedance, which measures resistance to a current and infers the rest. That is a genuinely useful home tool for tracking a trend under fixed conditions. It is not a substitute, and the visceral fat rating a smart scale prints is a proprietary index rather than a measurement in cm² or litres.