Body mass index (BMI) is calculated by dividing weight in kilograms by height in meters squared. It was developed in the 1830s by the Belgian statistician Adolphe Quetelet as a population-level measurement of body size, he explicitly noted it was not intended to measure individual body fatness. Yet it became the standard clinical tool for classifying weight in individual patients, which is a use it was not designed for.

The limitations of BMI are well documented in the research literature. These limitations do not mean BMI is useless, it remains a reasonable initial screen for population-level health risk. They do mean that treating BMI as a definitive individual health assessment misses important information.

What BMI Gets Right

At the population level, BMI categories correlate with the risk of cardiovascular disease, type 2 diabetes, certain cancers, and all-cause mortality. Studies covering millions of people consistently find this relationship. For identifying trends in obesity prevalence across populations or decades, BMI is practical and consistent.

For clinical use as an initial screen, BMI is fast, free, and non-invasive. A BMI above 30 in an adult should prompt further metabolic evaluation. A BMI below 18.5 signals underweight that warrants nutritional assessment. At the extremes, BMI is reasonably reliable.

Where BMI Fails at the Individual Level

Muscle mass: BMI cannot distinguish muscle from fat. A muscular person, a competitive athlete, for example, may have a BMI of 28 (classified as overweight) with minimal body fat and no metabolic risk. Conversely, a sedentary person with a BMI of 24 (classified as normal) may carry enough visceral fat to have significant metabolic risk.

Ethnic variation: The relationship between BMI and body fatness, and between BMI and metabolic risk, differs by ethnic group. Research consistently shows that Asian populations have higher body fat percentages and greater metabolic risk at lower BMI thresholds. The WHO recommends lower BMI cutoffs for Asian populations: 23 for overweight instead of 25, and 27.5 for obesity instead of 30. Many US clinicians do not apply these adjusted thresholds.

Fat distribution: BMI provides no information about where fat is distributed. Two people with identical BMIs may have very different visceral-to-subcutaneous fat ratios, and visceral fat is the metabolically active component that drives most health risk. The same total weight with predominantly visceral fat carries higher risk than the same weight with predominantly subcutaneous fat.

Sex differences: Women have higher body fat percentage than men at any given BMI. At a BMI of 25, the average woman has approximately 30-35% body fat while the average man has approximately 20-25%. Using identical BMI thresholds for both sexes treats these different metabolic situations identically.

Better Metrics for Individual Assessment

Waist circumference: Captures abdominal fat accumulation directly. The NIH defines elevated risk at above 40 inches (102 cm) in men and 35 inches (88 cm) in women. It adds information about fat distribution that BMI cannot provide.

Waist-to-height ratio: Dividing waist circumference by height normalizes for body size better than waist circumference alone. A ratio above 0.5 is associated with elevated cardiometabolic risk in most populations. A 2012 meta-analysis in Nutrition Research Reviews found waist-to-height ratio outperformed both BMI and waist circumference for predicting cardiometabolic risk.

DEXA body composition: Provides precise measurement of fat mass, lean mass, and bone density. It directly measures what BMI attempts to infer. DEXA is increasingly available at fitness centers and specialized clinics at relatively low cost ($50-150 per scan). For anyone wanting to track body composition changes during weight loss or training, DEXA is the most informative practical option.

Metabolic markers: HbA1c, fasting insulin, triglycerides, HDL cholesterol, and blood pressure collectively provide a direct window into metabolic health that BMI cannot. Metabolic syndrome is defined by a cluster of these markers, not by BMI.

BMI in Clinical Practice: What It Should and Should Not Determine

BMI should be used as an initial screening tool, not as a definitive assessment. A BMI over 25 should prompt further evaluation including waist circumference, metabolic markers, and family history. A BMI under 25 does not exclude metabolic risk, particularly in those with high waist circumference or Southeast Asian, South Asian, or East Asian ancestry.

For GLP-1 medication prescribing, FDA approval is based on BMI thresholds because BMI was the metric available in the clinical trials. Using BMI for this eligibility decision is appropriate as a regulatory matter, even if BMI is an imperfect metabolic health measure.

For context on what drives metabolic risk at the tissue level, see Visceral Fat vs. Subcutaneous Fat: Why Where You Store Fat Matters.