What does a DEXA scan show, and is it worth the money?

A DEXA scan shows four numbers: how dense your bones are, how much fat you carry, how much lean tissue (muscle, organs, water) you carry, and how much visceral fat you have, meaning the fat packed deep around the organs rather than sitting under the skin. The machine is a low-dose X-ray that fires two beam energies at once, and the gap between how the two beams are absorbed is what tells bone from fat from everything else. The full name, dual-energy X-ray absorptiometry, describes precisely that [11].

Bone density is the number with the deepest evidence behind it. A review in JAMA of research in postmenopausal women found that bone density predicts who goes on to fracture, and that density measured at the hip predicts hip fracture more strongly than density measured anywhere else in the body [12]. Newer methods read bone better but cost more. A 2018 review of computer models built from CT scans concluded they beat DEXA at estimating how much force a thigh bone can take, yet are not yet worth the money for predicting which patients fracture [9].

Results come back as comparison scores rather than raw grams: your density set against a healthy young adult, and against people of your own age and sex. That is a reporting convention, not a finding, and it is why the same bone produces two different-looking scores on the same page.

The body composition half is where the money question gets harder. In an analysis of US national survey data, visceral fat above roughly 327 cubic centimetres marked the point where cardiovascular disease became more common - it was found in 2.41% of adults above that line and 1.12% below, so about 24 in every 1,000 people rather than 11 [8]. The line sat higher in men, at about 387 cubic centimetres, than in women, at about 312 [8].

That same analysis is the strongest argument against paying for a scan just to get a body fat percentage. A model using nothing but waist circumference and two blood tests matched the scan almost exactly at identifying people with high organ fat [8]. The two tests were for triglycerides and apolipoprotein B, a protein that rides on the particles carrying fat through the blood.

For bone, the case turns on age. A German analysis for the country's statutory health insurance found that adding a scan to a questionnaire-based risk check cost about 20,235 euros for every extra year of healthy life it bought in 60-year-old women, and that in women over 70 the scan plus questionnaire was both cheaper and more effective than the questionnaire alone [10]. Where scans are scarce, other routes exist. In Singapore, running an algorithm over ordinary chest X-rays of 10,000 women aged 50 and over, alongside a short risk questionnaire, was projected to prevent 17 fractures and add 16 years of healthy life across that group [13].

Two X-ray beams of different strengths pass through you. Mineral absorbs them in a different pattern from soft tissue, and fat differently from muscle, so the scanner works backwards from that pattern to how many grams of each sit at every point of the image [11]. It is also why the output is areal density, meaning mineral per square centimetre of the picture rather than a true measure of volume [9].

The bone half and the body composition half of the report are not independent of each other. In 876 Boston children, fat-free mass went with higher bone density, while abdominal fat went with lower density only in the heaviest 15% [1]. In a separate study of 49 children aged 6 to 11, lean mass was the only measure that explained the differences in bone density between them: the children with obesity had denser bones, but because they carried more muscle, not more fat [3]. Among 118 adolescents, grip strength and fat-free mass together accounted for roughly half the variation in bone density in girls and closer to two-thirds in boys [2].

The pattern holds into later life. In 356 postmenopausal women, lean mass went with higher bone density at the spine, hip and femoral neck in those of normal weight and those who were overweight [4]. The authors of the children's study read this as a case for muscle: they concluded that building the muscular system through physical activity is what matters most for bone [3].

Visceral fat gets its own line on the report because it is the compartment where the survey data found risk rising. The researchers described their threshold as the point that divides fat the body appears to tolerate from fat that tracks with disease [8].

Bone density is not bone strength. When researchers scanned 50 reindeer thigh bones and then loaded them until they broke, density predicted only part of the force each bone could take, a moderate link rather than a tight one [7]. Those were animal bones tested in a laboratory, so the figure does not transfer directly to people, but it shows that a density number is an estimate of fragility, not a measurement of it.

The evidence behind the body composition half is thinner than the bone half, and much of it sits in the wrong people:

• Three of the studies linking lean mass to bone density were done in children and adolescents, not adults [1][2][3].
• In postmenopausal women with obesity, no significant link between body composition and bone density showed up at all [4].
• The JAMA review found there was not enough data to turn a bone density result into a fracture risk for men, or for women who are not white [12].
• The visceral fat thresholds came from US survey data and did not hold in non-Hispanic Black or Hispanic participants [8].

The cost figures are health-system figures, not price tags. They describe what a national insurer in Germany or Singapore gets back for screening a whole population [10][13], which is a different question from what a single scan is worth to one person paying privately.

None of these studies tested whether scanning healthy adults under 50 with no risk factors changes anything about what happens to them. Bone densitometry is a clinical test, and whether it belongs in your care is a conversation with a doctor rather than a purchase decision.

Two people book the same scan for opposite reasons. A 34-year-old wants a body fat percentage to track a training block, and she will get a detailed one: the scan was the reference standard the cheaper waist-and-blood-test model was measured against [8]. It is just that the cheaper model told researchers nearly as much about cardiovascular risk. A 68-year-old woman whose mother fractured a hip is in the group where the health economics work out, because her result can change what her doctor does next [10][12]. One machine, two very different reasons to pay for it.