Is BMI a useless measure?

No, but it is useless for the thing people most often use it for. As a way of tracking populations — comparing countries, monitoring trends over decades, identifying where the burden of obesity is rising — it is cheap, reproducible and correlates well enough with body fat at group level to do that job. As a way of assessing an individual person in a consulting room, it is weak: it cannot tell muscle from fat, says nothing about where the fat is, and misclassifies both muscular people as overweight and thin people with excess visceral fat as healthy. Most of the argument about BMI comes from applying criticism that is entirely valid for the second use in order to dismiss the first, which does not follow.

What should be measured instead?

Ideally BMI plus a waist measurement, which takes an extra thirty seconds and adds genuinely independent information about where fat is stored — and visceral fat around the organs is far more strongly linked to diabetes and heart disease than fat under the skin. The simplest version is waist-to-height ratio: your waist should be less than half your height, a rule that works consistently across sexes and ethnic groups without needing separate thresholds. More sophisticated methods exist — DXA scanning, bioimpedance, CT for visceral fat — but they are variously expensive, unavailable, or sensitive to hydration and recent meals. In practice, a tape measure used routinely would improve clinical assessment far more than any advanced technology that is not actually accessible.

Why do BMI thresholds differ between ethnic groups?

Because the relationship between body weight and health risk genuinely differs. People of South Asian, Chinese, Japanese and several other Asian backgrounds tend to carry a higher percentage of body fat and more visceral fat around the organs at any given BMI, and their risk of type 2 diabetes and cardiovascular disease rises at lower body weights. A WHO expert consultation reviewed this evidence and supported lower action points — commonly applied at 23 rather than 25 for increased risk and 27.5 rather than 30 for high risk. Applying the standard European thresholds to these populations systematically misses people who are already at substantial risk. This is a specific instance of a general problem: reference ranges derived in one population should be validated before being transferred to another.

What is the Lancet Commission proposing?

A significant change in what obesity means as a diagnosis. Rather than defining it by a BMI number, the 2025 Commission proposed two categories. Preclinical obesity is excess body fat with organs still working normally — a risk state that warrants monitoring and attention to risk factors, but not a disease. Clinical obesity is excess body fat together with objective evidence that organs are not working properly, or substantial limitation of everyday activities — that is the illness. Confirming excess adiposity requires more than BMI alone: either a waist measurement alongside it, or direct measurement of body fat, with BMI accepted by itself only at very high values. The effect is to define obesity by demonstrated harm rather than by a number, which changes who is diagnosed and, potentially, who is eligible for treatment.

Does criticising BMI mean weight does not matter?

No, and this is where the argument most often goes wrong in both directions. Weight stigma in healthcare is real and causes measurable harm: people delay seeking care, skip screening, and have unrelated symptoms wrongly attributed to their weight, with worse outcomes that are partly independent of the adiposity itself. That is a strong argument for changing how weight is discussed, for asking permission before weighing someone, and for not turning a consultation about an ankle injury into a lecture. It is not an argument that excess adiposity has no metabolic consequences, which the physiological and trial evidence does not support. Both things are true at once: the measure is crude, the stigma is harmful, and adiposity still affects health. Holding all three is the only position the evidence supports.

How does GMJ News cover BMI and obesity measurement?

GMJ News covers the distinction between population and individual use, what BMI misses including body composition and fat distribution, Quetelet's original derivation and intent, WHO-supported ethnic-specific thresholds, waist circumference and waist-to-height ratio as under-used additions, body composition measurement methods and their practical limitations, visceral versus subcutaneous adipose tissue physiology, the obesity paradox and the biases that explain it, the 2025 Lancet Commission clinical obesity framework, and documented weight stigma in healthcare alongside the physiological evidence. Part of PHIG's contested science, obesity and health equity programmes.

Home › Topics › Body Mass Index

Body Mass Index

GMJ News knowledge hub · last reviewed September 2026 · Georgian Medical Journal

Summary Statistics Glossary GMJ news FAQ Documents Organizations Research

Body mass index is simultaneously a poor measure of any individual and a useful instrument for populations, and most of the argument about it fails to distinguish these two uses. It cannot separate fat from muscle, says nothing about fat distribution, misclassifies muscular individuals as overweight and normal-weight individuals with excess visceral fat as healthy, and was never designed as a clinical diagnostic tool (WHO). Its persistence reflects genuine practical advantages — it requires only height and weight, is reproducible across settings, and correlates well enough with adiposity at population scale to support surveillance and comparison. The substantive developments are threefold: WHO recommends lower BMI action thresholds for South Asian, Chinese and other populations in whom cardiometabolic risk rises at lower body weights; waist circumference and waist-to-height ratio add independent predictive value and are trivially easy to measure yet routinely omitted; and the 2025 Lancet Commission proposed replacing BMI-defined obesity with clinical obesity diagnosed by confirmed excess adiposity plus organ dysfunction, a genuinely significant reframing of what the diagnosis means.

Key messages

THE CORE DISTINCTION: poor for individuals, useful for populations

Most of the argument about BMI fails to separate two entirely different uses. As a population surveillance instrument it is defensible — cheap, reproducible, comparable across time and countries, and correlated well enough with adiposity at group level to track trends and compare regions. As an individual diagnostic it is weak: it cannot distinguish fat from muscle, says nothing about fat distribution, and misclassifies both muscular individuals and normal-weight people with excess visceral fat. Criticism that is valid for the second use is routinely applied to invalidate the first.

WHAT BMI ACTUALLY MISSES

Body composition: a muscular athlete and a sedentary person of the same height and weight receive identical scores. Fat distribution: visceral adipose tissue is metabolically active and far more strongly associated with cardiometabolic risk than subcutaneous fat, and BMI is blind to the difference. Normal-weight obesity — normal BMI with high body fat percentage and metabolic dysfunction — is common and entirely invisible. Age-related change: older adults lose lean mass while gaining fat at unchanged BMI. Origin: it was derived by Quetelet in the 1830s from data on European men and was never intended as a measure of individual health.

WHO RECOMMENDS LOWER THRESHOLDS FOR SOME POPULATIONS

Cardiometabolic risk rises at lower BMI in South Asian, Chinese, Japanese and several other Asian populations, reflecting higher body fat percentage and greater visceral adiposity at equivalent BMI. WHO has issued guidance supporting lower action points, with public health thresholds commonly applied at 23 for increased risk and 27.5 for high risk rather than 25 and 30. Applying European thresholds unmodified to these populations systematically under-identifies risk, and this is a concrete example of a measurement standard failing to transfer between populations.

WAIST CIRCUMFERENCE ADDS INDEPENDENT VALUE AND IS ROUTINELY OMITTED

Waist circumference and waist-to-height ratio capture central adiposity and predict cardiometabolic risk independently of BMI, adding information precisely where BMI is weakest. A waist-to-height ratio above 0.5 — the simple rule that waist should be less than half of height — performs well across populations, sexes and ethnicities without requiring separate thresholds. Both measures require only a tape measure and thirty seconds. Their near-universal omission from routine clinical practice, while BMI is calculated automatically, is difficult to justify on any grounds other than habit.

THE 2025 LANCET COMMISSION: clinical obesity as a diagnosis

A substantial reframing. The Commission proposed abandoning BMI-defined obesity as a diagnosis and distinguishing preclinical obesity — excess adiposity with preserved organ function, a risk state — from clinical obesity, defined as confirmed excess adiposity plus objective evidence of organ dysfunction or substantial limitation of daily activities. Confirmation requires either a second anthropometric measure alongside BMI, direct body fat measurement, or a very high BMI where excess adiposity can be assumed. This converts obesity from a category based on a number into a diagnosis based on demonstrated harm, with significant implications for treatment eligibility.

THE STIGMA DIMENSION IS REAL AND CUTS BOTH WAYS

Weight stigma in healthcare is well documented and causes measurable harm: delayed presentation, avoidance of care, misattribution of unrelated symptoms to weight, and worse outcomes independent of adiposity. That is a genuine argument for changing how weight is discussed. It is not an argument that adiposity is metabolically neutral, and the two propositions are frequently conflated in both directions — with critics using stigma to dismiss physiology, and clinicians using physiology to dismiss stigma. Both are true simultaneously, and holding both is the only defensible position.

Key statistics

1830s

Quetelet derived the index from European male data; never intended as an individual health measure

Historical

23 / 27.5

WHO-supported lower action points for many Asian populations rather than 25 and 30

WHO expert consultation

WHtR >0.5

waist-to-height ratio threshold performing consistently across sexes and ethnicities

NICE/meta-analyses

Normal-weight obesity

normal BMI with high body fat and metabolic dysfunction is common and invisible to BMI

Eur Heart J

Lancet 2025

Commission proposed clinical obesity defined by adiposity plus organ dysfunction, not BMI alone

Lancet Diabetes Endocrinol 2025

Documented harm

weight stigma in healthcare delays presentation and worsens outcomes independent of adiposity

Obesity Reviews

Body mass index — where the disagreement actually lies

Source: Bars show strength of supporting evidence. BMI is defensible for populations and weak for individuals; conflating the two drives most of the argument.

Glossary of key terms

Body composition measurement methods
Clinical measurementDual-energy X-ray absorptiometry is the practical reference standard, giving fat mass, lean mass and regional distribution with low radiation, though results vary between manufacturers and it is not portable. Bioelectrical impedance analysis is cheap and widely available but sensitive to hydration, recent food and exercise, and consumer devices vary considerably in accuracy. Air displacement plethysmography and hydrostatic weighing are accurate but confined to research settings. CT and MRI quantify visceral adipose tissue directly and are the true reference for fat distribution, but cost and radiation limit routine use. Skinfold calipers are operator-dependent. For most clinical purposes, BMI combined with waist circumference provides more information than any single sophisticated measure that is not actually available.

Visceral versus subcutaneous adipose tissue
PhysiologyVisceral adipose tissue surrounding the abdominal organs is metabolically distinct from subcutaneous fat: it is more lipolytically active, drains directly into the portal circulation delivering free fatty acids to the liver, secretes a different profile of adipokines and inflammatory cytokines, and is far more strongly associated with insulin resistance, hepatic steatosis, dyslipidaemia and cardiovascular risk. This explains why two people with identical BMI can have entirely different metabolic profiles, why waist circumference adds predictive value, and why South Asian populations carry higher risk at lower BMI — they have greater visceral adiposity at equivalent weight. It also explains the metabolically healthy obesity phenotype, in which subcutaneous fat predominates, though longitudinal data suggest this state frequently converts over time.

The obesity paradox
EpidemiologyThe repeated observation that among people with established heart failure, coronary disease, chronic kidney disease, COPD or cancer, those with higher BMI sometimes have better survival than those in the normal range. Several explanations account for most of it without any protective effect of adiposity: reverse causation, since serious illness causes weight loss; collider stratification bias, arising from conditioning on disease status; the inability of BMI to distinguish protective muscle mass from fat, with sarcopenia driving mortality in the low-BMI group; smoking, which lowers weight and raises mortality; and unmeasured confounding. The paradox is a useful teaching case in epidemiological bias, and it is frequently cited in popular argument as evidence that weight does not matter, which the analysis does not support.

Ethnic-specific thresholds
Global healthA WHO expert consultation examined evidence that Asian populations exhibit higher cardiovascular risk and type 2 diabetes prevalence at BMI levels below conventional thresholds, driven by higher body fat percentage and greater visceral deposition at equivalent BMI. The resulting guidance supports lower action points, widely operationalised at 23 for increased risk and 27.5 for high risk in South Asian, Chinese, Japanese and several other populations, and adopted in national guidelines including NICE. Comparable considerations apply in reverse for some populations with greater lean mass at given BMI. The broader principle matters beyond obesity: measurement standards derived in one population require validation before transfer, and the historical default of European-derived reference ranges is a recurring source of systematic misclassification.

The Lancet Commission on clinical obesity
NosologyThe 2025 Commission proposed distinguishing preclinical obesity — excess adiposity with preserved organ function, understood as a risk state warranting monitoring and risk reduction — from clinical obesity, a chronic illness defined by excess adiposity together with objective signs of reduced organ or tissue function, or substantial limitation of age-appropriate daily activities. Confirming excess adiposity requires BMI plus at least one additional anthropometric measure such as waist circumference or waist-to-height ratio, or direct body fat measurement, with BMI alone accepted only at very high values. The framework addresses two criticisms simultaneously: that BMI over-diagnoses in people without dysfunction, and that a purely anthropometric definition fails to identify who actually needs treatment. Implications extend to treatment eligibility, insurance coverage and trial design.

Weight stigma in healthcare
Ethics/Clinical practiceDocumented consequences include delayed or avoided presentation, reduced uptake of screening including cervical and breast cancer screening, diagnostic overshadowing in which unrelated symptoms are attributed to weight, shorter consultations, less patient-centred communication, and reduced likelihood of appropriate investigation. Effects on outcomes appear partly independent of adiposity itself, mediated through avoidance and stress pathways. Practical mitigation includes appropriately sized equipment and seating, asking permission before weighing and before discussing weight, using person-first language, focusing on health behaviours and measurable outcomes rather than weight alone, and recognising that a patient presenting with an unrelated complaint has not asked for a weight consultation. None of this requires denying that adiposity affects health.