When BMI doesn’t apply: limitations every reader should know
BMI is one of the most repeated numbers in health journalism. It is also one of the most over-interpreted. The formula — weight in kilograms divided by height in metres squared — was created by Adolphe Quetelet in 1832 as a way to characterise populations, not to assess individuals. That distinction matters, because the cutoffs (18.5, 25, 30) become misleading in several specific groups. Below are the five I see misread most often.
1. Athletes and high-muscle-mass individuals
Muscle is denser than fat. A trained bodybuilder at 1.83 m (6 ft) and 100 kg (220 lb) sits at BMI 30 — officially “obese” — with body fat often below 10%. The number lies because BMI doesn’t see body composition. For athletes, recreationally muscular adults, or anyone who lifts seriously, the Body Fat Calculator (U.S. Navy method or Jackson–Pollock skinfold) is a more honest measure.
2. Children and adolescents
A child’s body composition changes constantly with growth. Adult BMI thresholds simply don’t map onto a 10-year-old. The right reference is the CDC age-and-sex BMI-for-age growth chart (percentiles, not raw numbers): “overweight” in children is defined as the 85th–95th percentile, “obese” as above the 95th. A raw BMI of 17 in a 12-year-old might sit at the 50th percentile and be entirely typical.
3. Older adults (65+)
Bone density decreases with age, sarcopenia (muscle loss) sets in, and the body’s reserves matter more for surviving illness. Several large meta-analyses suggest that for adults over 70, the BMI range associated with the lowest all-cause mortality shifts up, often to around 23–30, instead of the standard 18.5–25. A 75-year-old at BMI 22 may actually be undernourished, not “ideal”.
4. Ethnic and geographic variation
The 25-cutoff for “overweight” was derived from mostly European-ancestry populations in the 20th century. The WHO Expert Consultation (2004) formally acknowledged that for South Asian, East Asian and many Pacific populations, cardiovascular and diabetes risk rises at lower BMIs — cutoffs around 23 (overweight) and 27.5 (obese) are now recommended in those populations. A single global cutoff is a convenience, not a truth.
5. Pregnancy
BMI tracked during pregnancy gives no useful clinical signal. Pre-pregnancy BMI is used to calibrate gestational-weight-gain recommendations, but the live calculator on this page should not be used to evaluate weight changes mid-pregnancy. Talk to your obstetric provider instead.
Takeaway: BMI is a free, fast screening signal — useful for population studies, public-health policy, and a starting point for individual conversations. It is not a diagnosis. If your number puts you outside the “normal” range and you fit one of the categories above, treat it as a prompt for a real conversation with a healthcare provider, not a verdict. For a more nuanced view of body composition, try the Body Fat Calculator with the U.S. Navy or Jackson–Pollock method.
Sources: CDC growth charts · WHO Expert Consultation on BMI cutoffs (2004) · CDC adult BMI guidance.