BMI: When It Stops Being Useful
Published 2026-10-31
BMI is weight in kilograms divided by height in metres squared. A 1.75 m person weighing 70 kg has a BMI of 22.9. That is the whole formula.
It is useful because, across populations, higher BMI correlates with higher rates of metabolic disease. It is limited because, for one person, it does not know what the weight consists of.
Two sets of thresholds, and they differ
| Category | WHO | Chinese standard |
|---|---|---|
| Underweight | < 18.5 | < 18.5 |
| Normal | 18.5 – 24.9 | 18.5 – 23.9 |
| Overweight | 25 – 29.9 | 24 – 27.9 |
| Obese | ≥ 30 | ≥ 28 |
The Chinese thresholds are lower because at the same BMI, East Asian populations tend to carry more visceral fat and show higher diabetes risk. Which set applies depends on the guideline you are being measured against, not on the calculator.
Three things BMI cannot see
1. Muscle and fat weigh the same
Two people, 1.80 m and 85 kg, both BMIs of 26.2:
- A lifter at 15% body fat → classified overweight
- A sedentary person at 30% body fat → also classified overweight
Completely different risk profiles. Athletes routinely land in the "obese" range; "skinny fat" — normal BMI, high body fat, low muscle — lands in the normal range and gets missed.
2. Where the fat is matters more than how much
Visceral fat (around the organs) is far more dangerous than subcutaneous fat. BMI is blind to distribution.
A cheap and better proxy: waist-to-height ratio.
| Ratio | Reading |
|---|---|
| < 0.4 | Underweight |
| 0.4 – 0.5 | Healthy |
| 0.5 – 0.6 | Increased risk |
| > 0.6 | High risk |
Measure at the navel, at the end of a normal exhale, without pulling the tape tight. Waist-to-height outperforms BMI for cardiovascular risk prediction and costs one tape measure.
3. Age changes the interpretation
- Over 65: a slightly higher BMI (22–27) is associated with lower mortality, because nutritional reserve matters more. A BMI under 20 in an older adult warrants a check for malnutrition and sarcopenia.
- Children and adolescents: use age- and sex-specific percentile curves, not adult thresholds. A normal BMI at 7 and at 17 are different numbers.
- Pregnancy: BMI is only for pre-pregnancy assessment; gestational weight gain has its own chart.
A decision path that works
Compute BMI
├─ Normal range + regularly active → fine, also check waist ratio
├─ Normal range + sedentary → check body fat and waist ratio; watch for "skinny fat"
├─ Overweight + strength training + normal waist → probably muscle; confirm with body fat %
└─ Overweight/obese + large waist → this is the combination that warrants action
Body fat scales (bioimpedance) swing 3–5 percentage points with hydration, so read the trend, not the absolute number.
Setting a target
Going from BMI 28 to 24 at 1.70 m means 80.9 kg → 69.4 kg, a loss of 11.5 kg.
A sustainable rate is 0.5–1 kg per week (roughly 0.5–1% of body weight), so three to five months. Chasing a "standard weight" from an old formula (height − 105, or similar) ignores frame size, which varies by several kilograms at the same height.
Unit friction
- Height must be in metres for the formula; convert from cm or from feet and inches first
- 1 lb ≈ 0.4536 kg; 1 ft = 12 in
- A "stone" is 14 lb — still used in UK body-weight conversation
Unit Converter handles the conversions; BMI Calculator shows both the WHO and Chinese classifications side by side, so you can see where they disagree.
Summary
BMI is a screening tool, not a diagnosis. Use it once: if it lands in the normal range and you are active, move on. If it lands outside, look at waist ratio, body fat, blood pressure and fasting glucose before drawing conclusions from that one number.
BMI finds people who might have a problem. It does not establish that a particular person does.