Two people can weigh 78 kg and have wildly different health profiles. One carries 15% masse grasse and 38 kg of muscle. The other carries 32% masse grasse and 24 kg of muscle. Same scale number. Not even the same cardiovascular risk category.
Composition corporelle answers what weight cannot: what is your body actually made of? The methods range from a $30 bathroom gadget to a $150,000 clinical imaging machine. This guide explains every major method, how each works, and which one makes sense for your goals.
3-5%
BIA error margin
Balance connectée masse grasse vs. DEXA
1-2%
DEXA precision
Gold standard for composition corporelle
$30-150
Balance connectée range
Consumer BIA devices in 2026
$150-350
DEXA scan cost
Single clinical scan, out of pocket
Why Body Fat % Matters More Than BMI
BMI was designed in the 1830s for population-level statistics. A muscular 90 kg man at 180 cm registers as "overweight" (BMI 27.8) despite 12% masse grasse. A sedentary 68 kg woman at the same height registers as "normal" (BMI 21.0) at 35% masse grasse. Masse grasse percentage captures what BMI misses entirely.
The Five Methods Compared
| Method | Accuracy (BF%) | Cost | Time | Key Advantage |
|---|---|---|---|---|
| DEXA Scan | 1-2% error | $150-350/scan | 10-15 min | Full body map + bone density |
| Clinical BIA (InBody) | 2-3% error | $50-100/session | 2-5 min | Segmental analysis, 5 body regions |
| Consumer BIA (Scale) | 3-5% error | $30-150 device | 10 seconds | Daily home use, trend tracking |
| Hydrostatic Weighing | 2-3% error | $50-150/session | 20-30 min | Minimal hydration sensitivity |
| Bod Pod (ADP) | 2-3% error | $50-75/session | 5-10 min | Comfortable, no water needed |
DEXA: The Reference Standard
DEXA emits two X-ray beams at different energies. Each tissue type attenuates them differently -- bone absorbs aggressively, fat and lean tissue differ in electron density. The result is a full regional composition map: fat on your left arm vs. right, lean mass in legs vs. trunk, and visceral fat levels.
DEXA for Visceral Fat
Visceral adipose tissue (VAT) around your organs is the strongest composition corporelle predictor of metabolic disease. DEXA's CoreScan estimates VAT area with CT-scan-comparable accuracy. If your concern is metabolic risk, the VAT reading may be more valuable than total masse grasse percentage.
Limitations: Low-dose radiation (1-5 microsieverts -- less than a cross-country flight). Sensitive to hydration. Different DEXA machines produce slightly different results, so always use the same one for trend tracking.
Consumer Balance connectées (BIA)
Every "balance connectée" sends a small current through foot electrodes. Lean tissue conducts well; fat resists. The scale measures impedance and feeds it into a regression model with your height, weight, age, and sex to estimate masse grasse.
Critical point: The scale does not directly measure masse grasse. It measures impedance and uses a statistical model to predict it. Accuracy depends on how well your body matches the training population.
Never Trust a Single BIA Reading
Your scale's absolute masse grasse number is almost certainly wrong by 2-5 percentage points. What matters is the trend. If it reads 22% every morning for a month, then starts reading 20.5%, your masse grasse has almost certainly decreased -- even though neither number is your true value. Use the scale as a trend tracker, not an absolute measurement.
Clinical BIA vs. Consumer: The Gap
An InBody device uses 8 electrodes (hands + feet) and measures impedance through 5 body segments at multiple frequencies. Low frequencies travel around cells (extracellular water). High frequencies penetrate cells (total water). The ratio reveals intracellular water -- a direct proxy for muscle cell volume.
| Feature | Metric | DEXA | Clinical BIA | Consumer Scale |
|---|---|---|---|---|
| BF% Error | 1-2% | 2-3% | 3-5% | |
| Regional Analysis | Full body map | 5-segment | None | |
| Bone Density | Yes (primary use) | Estimated | Estimated (poor) | |
| Visceral Fat | Yes (CoreScan) | Estimated | Very rough | |
| Hydration Sensitivity | Moderate | Moderate | Significant | |
| Cost per Measurement | $150-350 | $50-100 | ~$0 |
body-composition-measurement
Unify all your wearable data and get personalized AI health insights in one place.
Balance connectée Metrics: Trust Levels
| Metric | Trust Level | Reality |
|---|---|---|
| Body Fat % | Moderate (trend) | Off by 3-5% absolute, but directionally reliable |
| Muscle Mass | Low-Moderate | Estimates lean mass, not actual muscle |
| Body Water % | Moderate | Reasonable impedance-based estimate |
| Bone Mass | Low | Regression estimate, not direct measurement |
| Visceral Fat | Low | Foot-to-foot BIA cannot assess trunk fat well |
| BMR / Metabolic Age | Low | Fun but not precise |
| Protein % | Very Low | Compounding estimation errors |
Balance connectées + DEXA = Best Combo
These methods are complementary, not competing. DEXA every 3-6 months gives you an accurate absolute baseline. Daily balance connectée readings track whether your interventions are moving that baseline in the right direction. The combination is more informative than either alone.
The Right Method for Your Goal
| Goal | Best Method | Frequency |
|---|---|---|
| General suivi de santé | Consumer balance connectée | Daily, same conditions |
| Athletic optimization | Clinical BIA + consumer scale | InBody monthly, scale daily |
| Weight loss monitoring | Consumer scale + quarterly DEXA | Scale daily, DEXA every 3 months |
| Osteoporosis screening | DEXA (mandatory) | Annually or as prescribed |
| Body recomposition | DEXA + consumer scale | DEXA every 8-12 weeks, scale daily |
| Casual curiosity | Consumer balance connectée | As desired |
Gagnants par catégorie
Most Accurate
DEXA Scan Winner
1-2% masse grasse error with full regional mapping, bone density, and visceral fat quantification.
Runner-up: Hydrostatic Weighing
Best for Daily Tracking
Consumer Balance connectée Winner
The only method you can use every morning at home. Individual readings are noisy, but the 30-day trend is remarkably informative.
Runner-up: N/A
Best Accuracy-to-Convenience Ratio
Clinical BIA (InBody) Winner
2-3% error with segmental analysis in under 5 minutes. Available at most gyms for $50-100 per session.
Runner-up: Bod Pod
Best Value
Consumer Balance connectée Winner
A $50 Eufy P3 provides unlimited daily estimates for years. A single DEXA costs more than the scale itself.
Runner-up: Clinical BIA
Body Fat Reference Ranges
| Category | Men (20-39) | Men (40-59) | Women (20-39) | Women (40-59) |
|---|---|---|---|---|
| Essential Fat | 2-5% | 2-5% | 10-13% | 10-13% |
| Athletic | 6-13% | 11-17% | 14-20% | 18-23% |
| Fit | 14-17% | 18-22% | 21-24% | 24-28% |
| Average | 18-24% | 23-28% | 25-31% | 29-34% |
| Above Average | 25%+ | 29%+ | 32%+ | 35%+ |
The Morning Protocol
Wake up. Bathroom. Stand on the scale barefoot, minimal clothing, before food or drink. Record. Walk away. Check your 7-day and 30-day trend lines weekly. That trend is your real data -- individual readings are just samples.
Frequently Asked Questions
Every 6-12 months for general monitoring. Every 3-4 months if actively changing composition corporelle. More frequently than every 8 weeks is rarely justified -- changes are too slow for even DEXA to reliably detect.
Absolutely. The value is in frequency and consistency, not single-reading accuracy. 365 morning readings par an produce a trend line that is remarkably informative. Buy a balance connectée for trends. Get a DEXA for absolute truth.
No. Different scales use different algorithms, electrodes, and regression models. Pick one device and stick with it. Your trend data is only valid within a single device's framework.
Yes. After age 30, adults lose 3-8% of masse musculaire per decade (sarcopenia), typically replaced by fat at a similar rate. A 70-year-old at the same weight as age 30 may carry 15-20 fewer kg of muscle. This is why resistance training is the most impactful long-term metabolic health intervention.
Hydration is the single largest confounder for BIA. When hydrated, impedance drops and the algorithm estimates lower masse grasse. When dehydrated, the reverse. The swing can be 1-3 percentage points. This is why consistent morning measurement matters more than device precision.
