BMI is the most widely used measure of body weight in clinical settings, and also one of the most widely criticized. The criticisms have a real basis: BMI is a single number derived from weight and height that does not directly measure body fat, body composition, or health risk in any meaningful detail. It works reasonably well at the population level. It works less well for individuals, especially individuals at the extremes of muscle mass or body shape.
This post covers what BMI does and does not measure, what body fat percentage adds, the differences among DEXA, BIA, and skinfold testing, and what “skinny fat” (normal-weight obesity) means metabolically.
What BMI is and is not
BMI = weight (kg) / height (m)². The categories used worldwide:
- Under 18.5: underweight.
- 18.5 to 24.9: normal weight.
- 25.0 to 29.9: overweight.
- 30.0 and above: obese.
These categories were derived from population studies showing that mortality risk follows a J or U shape with BMI — lowest in the normal range, rising at both extremes.
The limits of BMI for individuals:
- It does not distinguish muscle from fat. An NFL linebacker and a sedentary office worker can have the same BMI for very different reasons.
- It does not account for fat distribution. Visceral fat (around organs) carries higher metabolic risk than subcutaneous fat (under skin). BMI cannot tell them apart.
- It interacts with sex, age, and ethnicity. The same BMI value reflects different body composition in different groups. Asian populations, for example, tend to have higher body fat at the same BMI than European populations, leading some guidelines (WHO Western Pacific) to use a lower obesity threshold of 25 rather than 30.
For the population-level use BMI was designed for — comparing groups, tracking trends — it is adequate. For the individual-level use it is often pressed into — telling you specifically about your health risk — it is rough.
For more on BMI specifically in women’s health context, see BMI for women.
Body fat percentage and what it adds
Body fat percentage measures the proportion of total body mass that is fat. It separates fat mass from lean mass (muscle, bone, water, organs). Two people with the same BMI can have very different body fat percentages.
Healthy ranges for women (ACSM 2017 and ACE):
- Essential fat: 10 to 13 percent. The minimum needed for normal physiology.
- Athlete: 14 to 20 percent. Common in elite endurance and strength athletes.
- Fitness: 21 to 24 percent. Active, fit individuals.
- Average: 25 to 31 percent. Healthy for most adults.
- Obese: 32 percent and higher.
These ranges differ from men, who run roughly 8 to 10 percentage points lower at every category because women carry more essential fat for reproductive function. The ranges also drift slightly upward with age — a 50-year-old woman with 28 percent body fat is in a different relative position than a 20-year-old with the same number.
Gallagher et al. (Am J Clin Nutr 2000) developed widely used age- and sex-adjusted body fat percentage equivalents to BMI categories. The point of their work was to show that the same absolute body fat percentage means different things at different ages, and that BMI is a poor proxy for body fat at individual level.
How body fat is measured
Several methods exist, with different tradeoffs:
DEXA (dual-energy X-ray absorptiometry)
A scan that uses low-dose X-rays to differentiate fat, lean tissue, and bone.
- Accuracy: within 1 to 2 percent of true body fat in most studies. Considered the practical gold standard for clinical use.
- Cost and access: typically $50 to $250 at a sports medicine center or specialized clinic. Increasingly available.
- Pros: also gives bone density and regional fat distribution (visceral vs subcutaneous).
- Cons: small radiation exposure (very low, comparable to a few hours of background radiation). Requires a lab visit.
Underwater (hydrostatic) weighing
Once the gold standard, now largely supplanted by DEXA.
- Accuracy: within 1 to 3 percent.
- Pros: very accurate.
- Cons: cumbersome, requires specialized equipment, hard to find clinically.
Air displacement plethysmography (Bod Pod)
Similar accuracy to underwater weighing, but uses air rather than water.
- Accuracy: within 2 to 3 percent.
- Pros: comfortable, quick.
- Cons: specialized equipment, not widely available outside research and university settings.
Bioelectrical impedance analysis (BIA)
Smart scales and handheld devices that send a small electrical current through the body. Fat resists the current more than lean tissue.
- Accuracy: within 3 to 8 percent of true body fat, but highly variable depending on the device and the user’s hydration state.
- Pros: cheap, widely available, can be done at home repeatedly.
- Cons: a single measurement is not very accurate. Hydration, time of day, recent meals, and exercise all affect the reading. Useful for tracking trends in a single device under consistent conditions, less useful for absolute values.
Skinfold calipers
Pinches of skin and underlying fat are measured at standardized sites (often 3 or 7 sites), plugged into equations.
- Accuracy: within 3 to 5 percent in trained hands. Varies more widely with novice operators.
- Pros: very cheap (a caliper costs $10 to $50). No equipment needed beyond the caliper and a tape measure.
- Cons: highly operator-dependent. The same person measured by two different practitioners can get readings 4 to 6 percent apart.
Practical advice
For most people:
- A one-time DEXA scan gives you a reliable baseline.
- A consistent BIA scale used at the same time of day, in the same hydration state, can track trends usefully.
- Skinfolds are reasonable if you have access to a trained practitioner.
You do not need to obsess over the exact number. The trend over months matters more than any single reading.
What “skinny fat” really means
The lay term “skinny fat” describes someone who looks slim, has a normal BMI, but has higher than expected body fat percentage and lower than expected muscle mass. The medical literature calls this normal-weight obesity (NWO).
The metabolic risk pattern in normal-weight obesity is closer to overweight than to lean. Studies have linked NWO to:
- Higher rates of insulin resistance.
- Elevated rates of metabolic syndrome.
- Increased visceral fat accumulation.
- Higher all-cause mortality compared to BMI-matched peers with normal body composition.
The pattern is more common than people realize: estimates put it at 10 to 30 percent of normal-BMI adults depending on the definition and population. Women, sedentary individuals, and older adults are more often affected.
What drives the pattern? Mostly low muscle mass and high relative fat mass. Causes include:
- Chronic under-eating combined with low physical activity (especially low resistance training).
- Yo-yo dieting that loses muscle along with fat each cycle.
- Sedentary lifestyle without compensating dietary intake.
- Aging without resistance training.
The fix is not weight loss — that often makes the picture worse by losing more lean mass. The fix is body recomposition: resistance training, adequate protein intake (typically 1.2 to 1.6 g per kg body weight per day for active individuals), and sufficient overall energy.
Waist circumference and waist-to-height ratio
Beyond body fat percentage, fat distribution matters. Visceral fat (around organs) is much more metabolically active and harmful than subcutaneous fat (under skin).
Two simple measures track distribution:
- Waist circumference. Risk thresholds for women: above 80 cm (about 31.5 inches) increases risk; above 88 cm (about 34.5 inches) is considered high risk.
- Waist-to-height ratio. A simpler, more universal measure. Aim for under 0.5 (i.e., your waist should be less than half your height). See our dedicated post on waist-to-height ratio.
These are useful adjuncts to BMI. A normal BMI with elevated waist circumference is the typical signature of normal-weight obesity. A “high” BMI with normal waist circumference and high muscle mass is often not the cardiometabolic concern BMI alone would suggest.
The Body Shape Calculator walks through several of these measures together. The BMI Calculator gives the BMI number with appropriate context for women.
When the BMI number is misleading
A few archetypes where BMI does not capture the real picture:
Muscular athletes. Strength athletes, sprinters, and rugby players often have BMI in the overweight range (25 to 29) with body fat percentages in the 12 to 18 range. They are not overweight in any meaningful sense.
Older adults with sarcopenia. People who have lost significant muscle mass may have BMI in the normal range but body fat percentages of 35 to 40 percent and high cardiometabolic risk.
Asian and South Asian populations. Higher body fat at the same BMI compared to European populations. Some guidelines lower the overweight cutoff to 23 and obesity to 27.5 for these groups.
Pregnancy and postpartum. BMI changes here are normal physiology. See our post on postpartum period return for context on body changes after delivery.
Eating disorder recovery. Weight restoration often outpaces muscle restoration, leading to elevated body fat percentage at low absolute weights. Resistance training and continued recovery typically resolve this.
When body fat percentage testing is worth doing
Practical use cases:
- Goal-setting in fitness contexts. A baseline DEXA helps set realistic body composition goals beyond just weight.
- Confirming or refuting a BMI label. If your BMI is “overweight” but you train hard and look lean, body fat percentage will clarify.
- Tracking body recomposition. When you are trying to lose fat while maintaining or building muscle, scale weight is misleading. Body composition tracks the actual changes.
- Endocrine or fertility workup contexts. Severe under- or overweight states affect cycles and ovulation. See weight and fertility evidence.
- Suspected normal-weight obesity. Family history of metabolic disease with a normal BMI but symptoms of insulin resistance.
If you are not in any of those situations, the marginal value of body fat percentage testing over a sensible read of BMI plus waist measurement is small. Trends over years matter more than the precise number.
The bottom line
BMI is a population screening tool that performs adequately at the group level and roughly at the individual level. Body fat percentage adds meaningful information for individuals, particularly those at the extremes of muscle mass or those with normal BMI but suspected normal-weight obesity. DEXA is the practical gold standard for measurement, with BIA acceptable for tracking trends and skinfolds reasonable in trained hands.
The biggest practical addition to BMI is not always body fat percentage — it is waist circumference or waist-to-height ratio, which captures fat distribution and is highly correlated with cardiometabolic risk. The combination of BMI, waist measurement, and body composition gives a much fuller picture than any single number alone. The BMI Calculator and Body Shape Calculator put these together for a starting assessment.