DEXA

DEXA (Dual-energy X-ray Absorptiometry, sometimes spelled DXA) is a low-dose medical X-ray scan that measures bone mineral density and soft-tissue composition. The reference clinical method for measuring body fat percentage at ±1-2% accuracy.

The scan takes 10-20 minutes; the radiation dose is roughly 1% of a chest X-ray (about a day of natural background radiation). Cost: $50-200 in most US markets, often not covered by insurance unless ordered for diagnostic purposes (osteoporosis screening, post-menopausal bone health).

DEXA distinguishes three tissue compartments: fat mass, lean mass (muscle + organs), and bone mineral content. The body-fat-percentage number reported is fat-mass divided by total mass. The regional breakdown (arms vs legs vs torso) is more clinically useful than the global average for tracking changes from training or weight loss.

For at-home body composition estimation, the US Navy tape method (used by our body fat calculator) reaches ±3% accuracy — close to DEXA at zero cost.

Worked example

A 35-year-old male, 80 kg, gets a DEXA scan. Results: 12.4 kg fat mass, 64.3 kg lean mass, 3.3 kg bone mineral content. Body fat % = 12.4 / 80 = 15.5%. Regional breakdown: arms 18% lean / 8% fat, legs 27% lean / 12% fat, trunk 32% lean / 20% fat. Bone density T-score: +0.4 (normal, slightly above the young-adult reference). Six months later after a structured training programme, the rescan shows 10.1 kg fat, 66.0 kg lean, 3.3 kg bone — same total weight 79.4 kg, but body fat dropped to 12.7% and lean mass gained 1.7 kg, all of it in legs and trunk (the trained regions). The bathroom scale would have shown “basically no change” (−0.6 kg); DEXA shows the recomposition clearly. That regional precision is why bodybuilders, recovering injury patients, and anyone tracking specific muscle-group response uses DEXA over BIA scales.

When and why it matters

DEXA matters in three populations: (1) osteoporosis screening in post-menopausal women and men over 70, where T-scores guide whether to start bisphosphonate therapy — the test is covered by Medicare and most insurance for this indication; (2) athletes and bodybuilders tracking lean-mass and fat-mass changes precisely enough that the ±5% noise of BIA scales would mask their progress; (3) clinical research on sarcopenia, cancer cachexia, and pediatric growth disorders where total-body composition is a primary endpoint. It matters less for casual fitness tracking where ±3% accuracy from a tape measurement plus the Navy formula is usually sufficient and free. The mistake to avoid is rescanning on a different brand of DEXA machine and comparing the numbers — Hologic, GE Lunar, and Norland machines cross-calibrate within ±2-3%, which can be larger than the change you’re trying to measure. Reference: ISCD — Official Positions on DXA.

How DEXA works, briefly: the scanner emits X-rays at two different energy levels (hence “dual-energy”). Bone, fat, and lean tissue each absorb the two energies in distinct ratios, so the absorption pattern at each pixel of the scan uniquely identifies the tissue composition along that line. Bone density is measured in grams per cm² (areal density, not true volumetric density), and the headline output for osteoporosis screening is the T-score — your bone density relative to a healthy 30-year-old of the same sex, in standard deviations. T-scores below −2.5 define osteoporosis under WHO criteria; between −1.0 and −2.5 is osteopenia.

DEXA pitfalls and competitor methods: hydration status, recent meals, and very recent intense exercise can shift the lean-mass reading by several hundred grams, so serial scans should be done under the same conditions (morning, fasted, before training). Manufacturers (Hologic, GE Lunar, Norland) calibrate slightly differently and cross-machine comparisons are unreliable; always rescan on the same machine for trend tracking. Hydrostatic weighing and air-displacement plethysmography (BodPod) are competing reference methods with comparable accuracy but neither gives regional breakdowns. Bioelectrical impedance (BIA) scales and InBody machines are convenient but only ±5-8% accurate, with poor reproducibility against DEXA in any population not built like the calibration cohort. Related: BMI, BMR, and BMI vs body-fat percentage.