DEXA Scan Explained: What Your Results Really Mean

A DEXA scan splits your body into three things it can measure directly: fat, lean soft tissue and bone mineral. Your report turns that into a body fat percentage, a visceral fat estimate, a bone density score and a left-versus-right breakdown, and the whole document is readable once you know what each row is actually counting.

Most people scroll past the technical pages to find one number and ask whether it is any good. So that is where we start.

What counts as a healthy body fat percentage

The table below gives body fat bands by age and sex. Find your sex, then your age row, then see which band your result falls into.

Source: Gallagher D, Heymsfield SB, Heo M, Jebb SA, Murgatroyd PR, Sakamoto Y. Healthy percentage body fat ranges: an approach for developing guidelines based on body mass index. American Journal of Clinical Nutrition, 2000; 72:694 to 701. Sample of 1,626 adults (1,013 women, 613 men) from the United States, United Kingdom and Japan, measured by DXA and a four-compartment model. The bands above are derived from the paper's combined white and African American data.

The limits of that table

These bands were reverse engineered from body mass index, not health outcomes. Gallagher and colleagues worked out what body fat percentage corresponds to a BMI of 18.5, 25 and 30 at each age and sex, then used those as the boundaries. The table inherits whatever BMI gets right and wrong.

The bands also rise with age, describing what happens in populations rather than setting a target. Ethnicity matters too: the paper found Asian participants carried more fat at a given BMI than white or African American participants. Treat the bands as orientation, not a verdict.

What the machine is actually doing

DEXA stands for dual energy X-ray absorptiometry, usually written DXA in the research. Two X-ray beams at different energies pass through you, and bone, fat and lean tissue absorb those energies in different ratios, so the software can work backwards to estimate how much of each you have, pixel by pixel.

That gives a three compartment model: fat mass, lean soft tissue mass and bone mineral content, which add up to your total mass. Because the measurement is spatial, the software can also report each arm, each leg, the trunk and the head separately.

Visceral adipose tissue is different. The machine estimates the fat around your organs by subtracting the subcutaneous fat it can identify from the total fat in the abdominal region. It is a model built on top of a measurement, which is worth remembering when the number moves.

A worked example: reading a report line by line

Meet Priya, an invented but plausible example. She is 41, 165 cm, lives in Melbourne, lifts twice a week and walks most days. Here is her scan.

Illustrative example only. The reference thresholds used to interpret it are cited below.

Total mass, and how it splits

Fat mass plus lean soft tissue plus bone mineral equals total mass: 22.6 plus 43.6 plus 2.2 is 68.4 kg. If those three do not add up on your report, you are looking at a region rather than the whole body.

The split is the useful part. Two people at 68 kg can differ by ten kilograms of lean tissue. Priya's 43.6 kg of lean soft tissue is muscle, organs, connective tissue and the water inside all of it, so it is not a muscle measurement and it moves with hydration.

Body fat percentage

At 33.0% and age 41, Priya sits at the very top of the healthy band in the Gallagher table, one percentage point below the overweight boundary.

This is also the number most likely to be misread. Sitting at the edge of a band does not make you unhealthy at 33.0% and healthy at 32.9%, because the boundary is a modelling artefact. What matters more is the direction of travel across repeat scans, and whether the other rows agree.

Visceral adipose tissue

Visceral fat sits around the abdominal organs and is more strongly associated with cardiometabolic risk than fat under the skin, so this row is worth reading closely.

Meredith-Jones and colleagues derived DXA specific cutoffs in 1,482 New Zealand adults aged 18 to 66, scanned on a GE Lunar Prodigy with CoreScan software, and identified the VAT mass above which cardiometabolic risk was higher: 700 g for women under 40, 800 g for women 40 and over, 1,000 g for men under 40 and 1,200 g for men 40 and over.

Priya is at 640 g, comfortably under the 800 g figure for her age and sex. So a body fat percentage at the top of the healthy band is coexisting with a visceral fat result that looks fine. That combination is common, and it is why the single percentage on the front page is not the whole story.

Some machines report visceral fat as an area in square centimetres. The long standing threshold there is 100 cm², the Japanese diagnostic criterion for visceral fat obesity. As reported by Ryo and colleagues in World Journal of Radiology, the VACATION-J study of 12,443 Japanese adults found the average number of cardiovascular risk factors passed 1.0 at around 100 cm² in both men and women, regardless of age. Use whichever unit your report gives.

Bone mineral density, T-scores and Z-scores

This is the genuinely clinical part of the report, and the part most often misinterpreted. A T-score compares your bone density to a young adult reference population at peak bone mass. A Z-score compares you to people of your own age and sex. They answer different questions: the T-score asks how much bone you have lost relative to your own peak, the Z-score asks whether you are unusual for someone your age.

The World Health Organization diagnostic categories apply to T-scores: above -1.0 is normal, between -1.0 and -2.5 is low bone mass, often called osteopenia, and -2.5 or lower is osteoporosis. The International Society for Clinical Densitometry restricts these categories, in its 2023 Official Positions, to postmenopausal women and men aged 50 and over.

For premenopausal women, men under 50 and children, the ISCD states that Z-scores rather than T-scores are preferred, and that a Z-score of -2.0 or lower is described as "below the expected range for age", while anything above -2.0 is within the expected range.

So Priya's T-score of -1.3 would land in the osteopenia band on a chart, but she is 41 and premenopausal, and that classification is not intended for her. Her Z-score of -1.0 is within the expected range for her age, so her bone density is unremarkable and the T-score is not the row to act on.

None of this is a diagnosis. Bone density is interpreted alongside fracture history, medications, family history and other risk factors, which only a doctor can put together. If your report shows a low T-score and you are postmenopausal or a man over 50, or a Z-score of -2.0 or lower at any age, take it to your GP.

Left versus right

Priya's legs differ by 0.26 kg of lean mass and her arms by 0.13 kg. Before reading anything into that, it helps to know how precisely the machine can measure a limb. Buehring and colleagues scanned 60 Division 1 athletes twice on a GE Lunar iDXA and calculated the least significant change for each region. For women, leg lean mass had a least significant change of 310 g on the left and 389 g on the right, and arm lean mass 151 g and 164 g. In other words, differences of a few hundred grams in a limb are close to the noise floor of the instrument.

Priya's asymmetries sit around or below those figures, and a dominant side being slightly larger is normal. What is worth attention is a large, persistent difference alongside symptoms or a recent injury. That is a conversation with a physiotherapist or doctor, not a conclusion you draw from one scan.

What in this report would warrant a doctor's visit

Nothing in Priya's scan does. If any of the following show up in yours, book an appointment and bring the printout: a T-score of -2.5 or lower if you are a postmenopausal woman or a man over 50, a Z-score of -2.0 or lower at any age, a visceral fat result well above the cutoff for your age and sex, unexplained loss of lean mass, or a large change you cannot account for. A DEXA scan describes your body composition. It does not diagnose.

How much your result can move without you changing

Every measurement has error, and the honest way to read a repeat scan is to ask whether the change exceeds it. That threshold has a name: the least significant change, the smallest difference between two scans that is unlikely to be noise.

Farley, Slater and Hind measured this in 32 resistance-trained men, comparing four methods on the same day and on consecutive days. For DXA the same-day coefficient of variation was 2.4% for fat mass and 0.5% for fat-free mass, and the consecutive-day least significant change was 1,615 g for fat mass and 1,967 g for fat-free mass.

That figure reframes a lot of scan anxiety. If your fat mass drops by one kilogram between scans, it is inside the noise, and you cannot tell real change from a slightly different day.

What shifts your result on the day

Hydration is the big one, because DXA assumes lean tissue has a fixed water content. Toomey and colleagues had 12 active men cycle in 30°C heat until they lost 2.5% of body mass. Measured lean tissue mass fell by 1.69 kg, with no meaningful change in fat mass. After 48 hours of carbohydrate loading at 8 to 12 g per kilogram per day, it rose by 2.36 kg. Neither number reflects muscle appearing or disappearing.

Food does something similar. A narrative review by Kasper and colleagues reports that eating a meal can produce as much as a 2.6% increase in measured fat mass, and glycogen supercompensation a 2.5% increase in measured lean mass. The same review puts the technical error of a well-run DXA scan at about 1.9% for total fat mass and 0.4% for total lean mass.

The practical rules follow. Scan first thing in the morning, fasted overnight, before training, after emptying your bladder, in similar clothing, on the same machine. Leave at least three months between scans. Anything less and you are mostly measuring your breakfast.

Precision for bone density

Bone is measured more precisely than soft tissue, and the ISCD requires facilities to test their own precision rather than quote the manufacturer's figures. Its minimum thresholds are a precision error of 1.9% at the lumbar spine, giving a least significant change of 5.3%, 1.8% at the total hip for 5.0%, and 2.5% at the femoral neck for 6.9%. A 2% shift in your spine density between scans is not a finding.

DEXA compared with the alternatives

DEXA is not the reference standard. That title belongs to four compartment models, which account for body water and bone separately. DEXA is the most accessible method that gets close, and the only common one that also gives regional data and bone density.

In the Farley comparison, same-day precision for fat mass was 2.4% for DXA, 2.6% for the Bod Pod, 5.2% for bioelectrical impedance spectroscopy and 1.0% for skinfolds. The consecutive-day least significant change for fat mass was 1,615 g for DXA, 1,943 g for the Bod Pod, 3,607 g for impedance and 442 g for skinfolds.

Skinfolds look surprisingly good there, and that is a real result: Kasper and colleagues concluded skinfolds were the least affected by day to day variability of the methods they reviewed. The catch is that calipers measure subcutaneous fat at a handful of sites and say nothing about visceral fat, bone or regional lean mass. They track change well and give a poor absolute percentage.

Impedance scales are the weakest of the group for individual accuracy. A systematic review by Oliver and colleagues covering twelve studies of healthy adults found bioelectrical impedance body fat percentage had a mean bias against the four compartment model of -3.5 to +4.4 percentage points, with limits of agreement typically spanning 15 to 20 percentage points. For fat-free mass the bias ran from -3.9 kg to +1.8 kg, with limits often exceeding 6 kg. The averages look acceptable and the individual results do not.

That does not make your bathroom scale useless. If it is consistently wrong in the same direction it can still track a trend. It just cannot tell you your body fat percentage.

Getting a scan in Australia

Body composition scans are generally a private, out of pocket service here, and prices vary by clinic. Bone density scans are separate. Medicare item 12320 covers bone densitometry at two or more sites for people aged 70 and over who have either never had the test or have a T-score of -1.5 or better, once in any five year period, and other items cover people with specific risk factors. Your GP can tell you whether you qualify.

Where your own data fits

A scan is one frame. It cannot tell you whether your training, sleep and eating over the preceding months explain the numbers in front of you, which is the part that determines what you do next. hlth.coach reads your wearable and body composition data together and adjusts training, nutrition and recovery from the trend rather than a single reading. A DEXA result works better as an input than a scorecard.

The short version

• DEXA measures fat, lean soft tissue and bone mineral separately, then estimates visceral fat by modelling. The three should add to your total mass.

• Healthy body fat bands depend on age and sex. For women 40 to 59 the healthy range is 23% to 33%, for men the same age it is 11% to 21%, per Gallagher et al., 2000.

• Visceral fat cutoffs from Meredith-Jones et al., 2021: 700 g for women under 40, 800 g for women 40 and over, 1,000 g for men under 40, 1,200 g for men 40 and over.

• T-scores compare you to peak young adult bone, Z-scores to your own age group. The WHO categories (normal above -1.0, low bone mass -1.0 to -2.5, osteoporosis -2.5 or lower) apply to postmenopausal women and men 50 and over only.

• Under 50 or premenopausal, use the Z-score. Below -2.0 is "below the expected range for age" and is a reason to see your doctor.

• A fat mass change under about 1.6 kg between scans is inside the measurement noise, and dehydration alone shifted measured lean mass by 1.69 kg in one controlled study. Scan fasted, before training, first thing, on the same machine, at least three months apart.

• Impedance scales have limits of agreement spanning 15 to 20 percentage points of body fat. Use them for trends, not for a number.

Common questions

What do my DEXA results actually mean?

The scan gives you body fat percentage, fat mass, lean mass and where each sits on your body, plus an estimate of visceral fat. The worked example above walks through a full result line by line so you can read your own rather than fixating on the headline percentage.

What is a healthy body fat percentage?

Healthy ranges differ by sex and age, and the reference ranges above give them. Body fat percentage on its own is a poor verdict on health, which is why the scan reports lean mass and visceral fat as well.

How accurate is a DEXA scan?

It is the most accessible reference method for body composition, but it is not exact. Hydration, food and even which machine you use shift the result. Scan under the same conditions each time and compare against your own previous scan, not against someone else's.

How often should I get a DEXA scan?

Every three to six months is enough for most people. Body composition moves slowly, and scanning more often mostly measures noise and hydration rather than real change.

What is visceral fat and why does it matter?

It is the fat around your abdominal organs rather than under the skin. It is reported separately because it tracks differently with metabolic risk than total body fat does, and it can change while your overall weight barely moves.

Is DEXA better than smart scales?

For a single accurate measurement, yes. Bioimpedance scales are heavily affected by hydration and give a wide error band. Scales are still useful for a daily trend, provided you read them as a trend rather than a measurement.

Sources

• Gallagher D, Heymsfield SB, Heo M, Jebb SA, Murgatroyd PR, Sakamoto Y. Healthy percentage body fat ranges: an approach for developing guidelines based on body mass index. American Journal of Clinical Nutrition, 2000; 72:694 to 701.

• Meredith-Jones K, Taylor R, Brown R, et al. Age- and sex-specific visceral fat reference cutoffs and their association with cardio-metabolic risk. International Journal of Obesity, 2021; 45(4):808 to 817. doi:10.1038/s41366-021-00743-3

• Ryo M, Kishida K, Nakamura T, Yoshizumi T, Funahashi T, Shimomura I. Clinical significance of visceral adiposity assessed by computed tomography: a Japanese perspective. World Journal of Radiology, 2014; 6(7):409 to 416. doi:10.4329/wjr.v6.i7.409 (source for the 100 cm² threshold and the VACATION-J and J-VFS findings)

• International Society for Clinical Densitometry. 2023 ISCD Official Positions, Adult. iscd.org/official-positions-2023

• Farley A, Slater GJ, Hind K. Short-term precision error of body composition assessment methods in resistance-trained male athletes. International Journal of Sport Nutrition and Exercise Metabolism, 2021; 31(1), article p55. doi:10.1123/ijsnem.2020-0061

• Buehring B, Krueger D, Libber J, et al. Dual-energy X-ray absorptiometry measured regional body composition least significant change: effect of region of interest and gender in athletes. Journal of Clinical Densitometry, 2013; 16(2):193 to 200.

• Toomey CM, McCormack WG, Jakeman P. The effect of hydration status on the measurement of lean tissue mass by dual-energy X-ray absorptiometry. European Journal of Applied Physiology, 2017; 117:567 to 574.

• Kasper AM, Langan-Evans C, Hudson JF, et al. Come back skinfolds, all is forgiven: a narrative review of the efficacy of common body composition methods in applied sports practice. Nutrients, 2021; 13(4):1075. doi:10.3390/nu13041075

• Oliver CJ, Del Vecchio L, Minehan M, et al. The validity of bioelectrical impedance analysis compared to a four-compartment model in healthy adults: a systematic review. Journal of Functional Morphology and Kinesiology, 2026; 11(1):65. doi:10.3390/jfmk11010065

• Australian Government Department of Health. Medicare Benefits Schedule, item 12320.

This article is general information and is not medical advice; speak to your GP or an accredited practising dietitian about your own results.