You got the baseline scan, you did the work, and now the second report is in your hands: two columns of numbers that are supposed to tell you whether the last few months went the way you hoped. The trick is knowing which of those numbers actually moved, and which just wobbled.
If you haven't had your first scan yet, start with our companion guide on getting a DXA scan and what it measures. This one picks up afterward: how to read a report against your last one and tell real change from measurement noise.
The numbers that matter, and what "better" looks like
A DXA report hands you more figures than you need. The ones worth watching over time are:
- Total mass: your weight, the same number the scale gives you, and the least informative one here.
- Fat mass: total fat in kilograms or pounds, plus a breakdown by region (arms, legs, trunk). This is the number you most want to see fall.
- Lean mass (sometimes labeled lean soft tissue or fat-free mass): muscle and everything else that isn't fat or bone. The number you most want to protect.
- Body fat percentage: fat mass divided by total mass. Useful, but it can move just because one of its two inputs moved, so read it alongside the raw kilograms.
- Visceral fat (VAT): the metabolically active fat around your organs. It's also the least precise number on the report (more on that below).
- Bone mineral density: carried over from DXA's original job as an osteoporosis screen. It changes slowly, so don't expect meaningful movement between scans a few months apart. For what weight loss does to bone and how to protect it, see protecting your bones on a GLP-1.
On a GLP-1, a good trajectory is easy to describe and harder to achieve: fat mass down, lean mass holding as steady as you can keep it. A drop in total weight that is almost entirely fat, with lean mass barely changed, is the outcome you're training and eating for. Some lean loss alongside the fat is expected, but treat it as a signal to push harder on protein and resistance work, and don't celebrate the smaller number on the scale just yet. (One caveat if you've lost a lot of weight: loose skin is fat-free tissue that DXA files under "lean," which can flatter that number. See does loose skin throw off your DXA, BIA, or Seca results?)
The precision trap: small changes may be noise
Before you react to a single report, know how much the machine itself wobbles. DXA is precise, but not perfect. A 2026 study that scanned people twice in a row put DXA's precision error (coefficient of variation) at roughly 1% for whole-body fat mass and under 1% for fat-free mass (DOI (external link)). In practical terms, an earlier study found the repeat-scan error was around 0.5 kg for whole-body fat mass and around 0.6–0.7 kg for lean mass. It used the International Society for Clinical Densitometry method to calculate a "least significant change," the smallest difference between two scans that reflects real change and not machine variability (DOI (external link)).
So if your lean mass reads half a kilogram lower than last time, that may be noise. A change only becomes trustworthy once it clears the least-significant-change threshold, which is why the trend across three or four scans tells you far more than any single scan-to-scan difference. Regional numbers are noisier still, and visceral fat is the noisiest of all: one precision study measured a VAT error near 16%, versus about 1.4% for whole-body fat (DOI (external link)). Treat a big VAT swing as a rough direction, not a decimal-point fact.
Scan the same way every time
Because you're chasing changes that can be smaller than the machine's own error, consistency matters a lot. Two rules do most of the work:
- Same machine, ideally the same clinic. Different DXA models don't read identically. One study found systematic differences between two densitometers large enough to require conversion equations before their numbers could be compared (DOI (external link)). A "loss" that's really just a change of scanner wastes a checkpoint.
- Same hydration and fasting state. Lean-tissue readings partly reflect body water, so a scan after a big meal, a hard workout, or a lot of fluid will differ from a fasted, rested one. Researchers who standardized food, fluid, and activity in the 24 hours before scanning concluded it has the potential to shrink the measurement error further, though that still needs confirming (DOI (external link)). Book your scans at a similar time of day, fasted, before training, and keep that routine each time.
How much lean loss is normal, and how to shift the ratio
Some lean-mass loss is expected on a GLP-1. A 2024 review in Diabetes Care reports that incretin medications cause a rapid loss of lean mass in trials, roughly 10% or about 6 kg, which the authors compare to a decade or more of aging (DOI (external link)). So if your report shows part of the loss coming from lean tissue, you're in the expected range and you haven't failed.
What you can change is how big that share is. The same review notes that supervised resistance training lasting more than 10 weeks can add about 3 kg of lean mass and about 25% more strength, and its authors recommend tailored resistance training alongside incretin therapy to preserve lean mass while you lose fat. Pair it with adequate protein. Our guides on why strength training matters and protein, fat, and carbohydrates cover the how, and your serial DXA reports are how you confirm it's working. A lean-loss trend that flattens out once you add training and protein is exactly what these scans exist to show.
This is general education and peer information, not medical advice. Interpreting your specific results belongs with your prescriber or a clinician who can see the full report. Research findings here are drawn from PubMed; see the linked studies for details.