Oct 17, 2026
Most people step onto a scale, see a number, and assume it tells them something meaningful about their health. It tells them how much they weigh. It does not tell them how much of that weight is fat, how much is lean tissue, or where the fat is stored.
That is why body-composition testing has become popular. Among the available methods, dual-energy X-ray absorptiometry—usually called DXA or DEXA—is particularly useful. But there is a distinction to understand before you become too attached to a body-fat percentage printed on a report:
DEXA is a good measurement tool. It is not a perfect measurement of biological truth.
DEXA uses two X-ray energies that are attenuated differently as they pass through tissue. Software analyzes those signals to estimate three main compartments: bone mineral content, fat mass, and lean soft tissue. A whole-body scan can also divide the results into regions such as the arms, legs, and trunk. The test is relatively quick, noninvasive, and involves a low radiation dose. [1]
A modern body-composition report may include total fat mass, percentage body fat, fat mass index (FMI), appendicular lean mass, and an android-to-gynoid fat distribution ratio. Some systems also provide an estimate of visceral adipose tissue.
These are potentially useful numbers. They are not all measured with the same degree of certainty.
First, a terminology point: lean soft tissue is not synonymous with skeletal muscle. It includes muscle, but also water, blood, organs, and other nonfat soft tissue. If a report shows a change in “lean mass,” you cannot automatically conclude that the person has gained or lost that amount of actual muscle. [1]
MAN MEDICINE / BODY COMPOSITION
Two X-ray energies + software → estimates of three compartments
A change in lean soft tissue is not proof of muscle gain or loss.
Fluid and glycogen shifts can change the reading without equivalent changes in muscle tissue.
For follow-up scans: make the comparison fair
Original conceptual diagram. Sources: Chaves et al. (2022); ISCD Official Positions (2023).
Also, the whole-body scan used for body composition is not the same clinical examination as a targeted hip-and-spine bone-density study used to diagnose osteoporosis. They use related technology but answer different questions. [1]
This is where many discussions of DEXA go wrong.
Precision describes whether repeated measurements under similar conditions give similar results. Accuracy describes how closely the measurement approaches the quantity we are actually trying to determine.
A machine can produce highly repeatable results and still have a systematic bias.
For example, a study of the GE Lunar iDXA found excellent short-term repeatability: the reported coefficient of variation was 0.82% for total fat mass and 0.86% for percentage body fat. Those are impressive device- and protocol-specific precision statistics, not a guarantee that every individual's true body-fat percentage is within 0.86 percentage points of the printed number. [2]
When researchers compare DEXA with a four-compartment model, which incorporates additional measurements of body density, water, and mineral, the results become more complicated. This is a useful reference approach, but its component measurements also carry error; it is not a direct inventory of every fat cell. [3, 4]
In one small study of 32 young adults, DEXA measurements were highly repeatable but showed a significant group-level difference from the four-compartment reference method. The researchers also warned against substituting the results of one body-composition method for another when tracking changes. [3]
A broader methodological review found that DEXA can either overestimate or underestimate body fat relative to four-compartment models, depending on the population and equipment. Importantly, the disagreement for an individual can be substantially larger than the average difference between groups. [4]
So the statement “DEXA is the gold standard” needs qualification. It is an established reference technique for several clinical and research applications. It is not an error-free ground truth for an individual's body-fat percentage.
Imagine someone has a DEXA scan on Monday after a strenuous training session, relatively depleted of glycogen and fluid. They return after several days of higher carbohydrate intake and rehydration. The report may suggest that their lean mass increased.
Did they build several pounds of muscle in a few days? Almost certainly not.
DEXA is sensitive to the water contained in lean tissue. In a controlled study of 12 active men, exercise-induced dehydration was associated with an average reduction of 1.69 kg in DEXA-derived lean tissue mass. Over the next 48 hours, carbohydrate loading and recovery produced an average increase of 2.36 kg relative to the depleted measurement, not the original baseline. Fat mass did not materially change across those measurements. [5]
That is a deliberately stressful experimental protocol, not a prediction of what happens to every patient. But it demonstrates the underlying problem: changes in hydration and glycogen can look like changes in lean tissue.
A separate controlled dietary study also found that feeding and acute carbohydrate manipulations could alter DEXA estimates. The authors recommended standardized pretest conditions to improve the interpretability of serial scans. [6]
This is especially relevant for people using DEXA to monitor resistance training, weight loss, or changes in medication. A single small increase in lean mass does not prove hypertrophy, just as a small decrease does not establish muscle loss.
Here DEXA offers a potentially useful advantage over a bathroom scale or BMI calculation. Where fat is stored changes its clinical interpretation. Abdominal visceral adipose tissue is metabolically distinct from subcutaneous fat and is associated with cardiometabolic risk.
Some DEXA systems estimate visceral fat by applying an algorithm to regional abdominal measurements. That is different from directly visualizing visceral fat with CT or MRI.
In one validation study involving 124 adults, DEXA-derived visceral-fat volume correlated strongly with CT-derived measurements. The combined coefficient of determination was 0.957. But the individual limits of agreement remained fairly wide. In other words, a strong correlation across a group should not be mistaken for interchangeable measurements in every person. This manufacturer-funded study examined adults with BMI between 18.5 and 40, so its results should not automatically be extended beyond the population or equipment tested. [7]
DEXA-derived visceral-fat estimates can provide context for longitudinal monitoring, alongside blood pressure, glycemic measures, lipid assessment, waist circumference, and the rest of a cardiovascular risk evaluation.
It depends on the question.
BMI is crude as a measure of body composition. It cannot distinguish a muscular athlete from a person carrying substantial adipose tissue at the same height and weight. Waist circumference adds useful information about central adiposity. Neither measure becomes worthless simply because more elaborate imaging is available.
Bioelectrical impedance analysis (BIA) is inexpensive and accessible, but readings depend on the instrument, its prediction equations, and fluid status. DEXA offers more detailed regional information and often better-controlled laboratory measurement. That does not mean every BIA device underestimates body fat or that the difference between DEXA and BIA has one universal percentage. [3, 4]
An FMI—fat mass divided by height squared—can be helpful when BMI obscures the distinction between fat and lean tissue. Still, no single cutoff or scan-derived number should be treated as a complete assessment of obesity-related risk.
Percentage body fat is a ratio: fat mass divided by total mass. It can be high because fat mass is high, because lean tissue is low, or both. FMI separates the fat component from the amount of lean tissue by dividing fat mass by height squared.
A 2026 analysis of 14,850 adults in the 1999–2006 NHANES dataset illustrates the distinction. Among people with low appendicular lean mass, the proportion classified as obese varied markedly depending on whether researchers used body-fat percentage, FMI, or BMI. The study supports looking at fat and lean compartments separately. It does not establish a universal FMI threshold or demonstrate that changing diagnostic labels improves outcomes. Its operational definition of sarcopenia used low lean mass, rather than a complete clinical assessment of muscle strength and performance. [9]
If you are repeating scans to evaluate an intervention, the most important question is not, “What is my body-fat percentage today?” It is, “Can I distinguish a real change from measurement variability?”
Use the following principles:
The International Society for Clinical Densitometry specifically emphasizes standardized analysis, appropriate cross-calibration, and facility-level precision assessment for body-composition DXA. It also notes that the clinical utility of several optional indices, including visceral adipose tissue and FMI, remains uncertain and that specific DXA adiposity thresholds defining obesity are not established. [8]
Low radiation exposure does not mean zero exposure. Body-composition DXA is contraindicated during pregnancy. Tell the facility about recent contrast examinations or radiopharmaceutical studies, which can interfere with interpretation, and check its scanner size and weight limits. [8]
Choose a repeat-testing interval that allows a plausible treatment-related change to emerge. More frequent scans are not automatically more informative. A clinician can help decide whether the result is likely to change your care.
DEXA is a valuable tool for assessing body composition, particularly when someone is losing weight, attempting to preserve lean tissue, or trying to understand changes that the scale alone cannot explain.
A major clinical decision should not rest on a tiny change in a DEXA-derived body-fat percentage or lean-mass estimate. The scan is an estimate built from physical measurements and mathematical assumptions. Its reliability depends on the device, the analysis, and the conditions under which it was performed.
Use DEXA to improve the conversation about body composition, not to end it. The goal is not to chase a precise-looking number on a printed report. The goal is to understand whether a person is losing excess fat, maintaining functional muscle, and improving their long-term metabolic health.
