What Is a DEXA Scan, and How Accurate Is It?

Key takeaways

• A whole-body DEXA scan gives a radiation dose of about 4 to 8 microsieverts, roughly half a day to one day of natural background radiation.¹²³
• DEXA lean mass is not the same as muscle. It is a chemical grouping that also counts organs and body water, so it can move when muscle does not.⁵
• Against a four-compartment model, individual DEXA body fat readings ranged from 2.6 points too high to 7.3 points too low.⁷
• A repeat scan only shows a real change when the change is bigger than the clinic's least significant change, which works out to about 2.8 times its precision error.⁴¹⁵
• In a hospital study of three consumer scales, the median fat mass error was 2.2 to 4.4 kilograms too low compared with DEXA.¹³

What is a DEXA scan?

A DEXA, or DXA, scan passes two low-energy x-ray beams through your body and uses the difference to split each spot into fat, lean soft tissue and bone. A whole-body scan takes a few minutes and gives a radiation dose of about 4 to 8 microsieverts, which is roughly half a day to one day of natural background radiation. It is good at showing whether your body changed between two careful scans. It is less exact about your true body fat level than its reputation suggests.²¹³

What does a DEXA scan measure?

The report gives the totals in the table, broken down by arms, legs and trunk.²

Lean mass is not the same as muscle. It is a chemical grouping that also counts organs and body water, so it can move when muscle does not.⁵

In one study, two days of very high carbohydrate intake raised DEXA lean mass by about 2 percent in trained cyclists, and by 3 percent when combined with creatine loading, at a time when little real muscle change was likely.⁶

Many scanners also print an estimate of visceral fat, the fat around your organs. DEXA does not see that fat directly. It estimates it, and the professional body for DEXA, the International Society for Clinical Densitometry, or ISCD, says in its own words that the clinical value of that number is currently uncertain.⁴

How much radiation is in a DEXA scan?

The whole-body row comes from two studies. One measured 4.2 and 8.4 microsieverts in adults on two Hologic models, and the other reported about 4 to 5 microsieverts on newer machines. All the other rows come from RadiologyInfo.¹²³

The dose per scan is small, but it does add up. Four scans a year come to about 16 to 32 microsieverts, close to one cross-country round trip by air. Ten years of scans every three months comes to about 160 to 320 microsieverts, which is more than one chest x-ray but less than two months of natural background.¹²³

The ISCD lists pregnancy as a reason not to have a body composition scan. Recent contrast material, from some CT or other imaging tests, can also throw off the results, so tell the clinic about any recent scans.⁴

How accurate is DEXA?

DEXA is often called the gold standard, but researchers use a four-compartment model, which splits the body into water, bone mineral, fat and the remainder, as the stronger reference. In 152 healthy adults, DEXA read 18.9 percent body fat on average against 20.7 percent from the four-compartment model. For individuals, the gap ranged from 2.6 points too high to 7.3 points too low, and DEXA tended to read leaner people lower.⁷

So treat a single DEXA body fat number as accurate to within a few points, not to the decimal.⁷

Hydration matters too. When athletes lost sweat during exercise in the heat, DEXA lean mass dropped by about 1.3 kilograms while fat mass did not change. Eating breakfast and going about a normal day both raised lean mass readings compared with a fasted morning scan.⁸⁹

DEXA vs smart scale, which is more accurate?

Home smart scales use bioelectrical impedance, or BIA. They send a small current through your feet and use equations to estimate body water and fat-free mass. Fat is what is left over.¹¹

In a hospital study of three consumer scales, body weight was close to DEXA, but the median fat mass error was 2.2 to 4.4 kilograms too low, with a wide spread between people. Another study found that for one person, a foot-to-foot scale could read body fat up to about 14 points lower or 9 points higher than DEXA.¹³¹⁴

A gap of several points between a smart scale and DEXA is common and does not mean either device is broken.¹³¹⁴

A smart scale is better for direction than for level. European clinical nutrition guidelines say BIA can follow body composition over time in people without unusual hydration, but in the guideline's words, the results must be interpreted with caution. Carbohydrate intake also shifts the reading. Glycogen is stored with water, and in one trial two days of high carbohydrate intake raised measured body water by about 2 percent. Since the scale works from body water, the next morning's lean mass reading can rise without any change in muscle.¹²⁶¹¹

How often should you get a DEXA scan?

The answer depends on how precise your clinic's scanner is and how fast your body is changing. A repeat scan only tells you something when the change is bigger than the scanner's own noise.⁴¹⁵

That threshold is called the least significant change, or LSC. The ISCD says every DEXA facility should measure its own precision error and calculate its LSC at 95 percent confidence. It sets the worst acceptable precision for one technologist at 3 percent for total fat mass and 2 percent for lean mass and body fat percent. The LSC works out to about 2.8 times the precision error.⁴¹⁵

The table shows what that means for someone carrying 30 kilograms of fat.⁴¹⁵

Same-day precision of about 1 percent for fat mass has been reported on newer scanners. Between days, food, fluid and activity add more noise.¹⁶⁹

A practical rule is to ask your clinic for its LSC, estimate how much your fat mass is likely to change per month, and space scans so the expected change clears the LSC. If fat mass is changing by about 1 kilogram a month, a clinic with a 2.5 kilogram LSC needs about two and a half months or more between scans to show a real difference.⁴¹⁵

How to get a repeat scan you can compare

A review from the Australian Institute of Sport lays out a standard protocol.¹⁰

Do not compare scans from different brands of machine. The ISCD says no calibration object removes the systematic differences between manufacturers, and a study in people is needed before results can be compared.⁴

• Scan in the morning after an overnight fast, rested, with no exercise beforehand.¹⁰
• Wear minimal clothing.¹⁰
• Lie in the same position, ideally with positioning aids.¹⁰
• Go back to the same scanner each time.¹⁰

How to track it

DEXA tells you what changed. A food log and regular weigh-ins tell you why. Weigh at the same time each morning and look at the weekly average rather than any single day, since water and food move the scale from one day to the next. Keep a steady food log in the days before each scan too, since food and fluid change the lean reading. An app like TrueCal can keep your food log and weight history in one place, so you can line up two scans with what you ate between them.⁹

Common myths

• Myth, DEXA is the gold standard, so the number is exact. Against a four-compartment model, individual DEXA readings were off by up to about 7 points of body fat.⁷
• Myth, DEXA lean mass is muscle. It includes organs, water and glycogen, and it moves with hydration and carbohydrate intake.⁶
• Myth, years of regular scans add up to less than one chest x-ray. Ten years of scans every three months is about 160 to 320 microsieverts, more than one 100 microsievert chest x-ray. The total is still low, but the comparison does not hold.¹²³
• Myth, my smart scale and DEXA should match. Consumer scales were often several kilograms off for fat mass compared with DEXA.¹³

A note on medical advice

This article is for general information and isn't medical advice. Talk with your clinician about your own situation.

Worked example

Take someone carrying 30 kilograms of fat who scans at a clinic with 3 percent precision for fat mass, the worst the ISCD accepts. The LSC for that clinic is about 8.3 percent, which is about 2.8 times the 3 percent precision error. Then 8.3 percent of 30 kilograms is about 2.5 kilograms, so a repeat scan can only confirm a fat change of about 2.5 kilograms or more. If fat mass is changing by about 1 kilogram a month, two months gives about 2 kilograms, which is still short of 2.5. Two and a half months gives about 2.5 kilograms, which reaches it. So scans at that clinic need to be about two and a half months or more apart to show a real difference.⁴¹⁵

A closer look

DEXA lean mass is not pure muscle, and food can move it. In one study, two days of very high carbohydrate intake raised DEXA lean mass by about 2 percent in trained cyclists, and by 3 percent when combined with creatine loading, at a time when little real muscle change was likely.⁶⁵

Questions and answers

Related guides

• FFMI calculator
• TDEE calculator
• What visceral fat is
• Water weight vs fat loss

Sources

1. Bone. Blake. 2006.
2. Quantitative Imaging in Medicine and Surgery. Messina. 2020.
3. Radiation Dose in X-Ray and CT Exams. RSNA, ACR. 2026.
4. Official Positions. ISCD. 2023.
5. Journal of the Endocrine Society. Tinsley, Heymsfield. 2024.
6. Medicine and Science in Sports and Exercise. Bone. 2017.
7. Journal of Applied Physiology. Van Der Ploeg. 2003.
8. International Journal of Sport Nutrition and Exercise Metabolism. Rodriguez-Sanchez. 2015.
9. Medicine and Science in Sports and Exercise. Nana. 2012.
10. International Journal of Sport Nutrition and Exercise Metabolism. Nana. 2015.
11. Clinical Nutrition. Kyle. 2004.
12. Clinical Nutrition. Kyle. 2004.
13. JMIR mHealth and uHealth. Frija-Masson. 2021.
14. Obesity Facts. Bosy-Westphal. 2008.
15. Obesity. Meredith-Jones. 2018.
16. Journal of Clinical Densitometry. Rothney. 2012.