Your BMI may be wrong about you. A study of 844,000 adults says the tape measure, and your organs, tell the real story

For decades, one number has decided who counts as obese: weight divided by height squared. A study of 844,108 adults across four countries now shows how often that number gets it wrong, and who it is wrong about.

The research, published in the journal Nature on Wednesday 11 November, is the largest test yet of a new way to diagnose obesity proposed in January 2025 by a Lancet Diabetes & Endocrinology commission. It matters because obesity medicines and surgery are becoming more powerful, more popular and more expensive, and health systems need to know who should get them first.

What did the researchers do?

The team, led by An Pan of Huazhong University of Science and Technology in Wuhan with senior author Francesco Rubino of King’s College London, worked through health records from four large groups of adults aged 18 to 74 in China, Spain, the UK and the US. None had had a stroke, a heart attack or cancer when they joined.

They applied the commission’s two-step test. First, confirm that a person really carries excess fat, by adding a waist measurement or waist-to-height ratio to BMI. Second, split those people into two groups. “Clinical obesity” means the extra fat is already impairing how organs work or how a person gets through daily life. “Preclinical obesity” means the fat poses a risk but the body is still functioning normally.

How often did BMI get it wrong?

In every one of the four populations, BMI both overcounted and undercounted. It missed obesity-level body fat in up to 8.6% of participants, the kind of person whose weight looks unremarkable but whose fat sits around the middle. It also classed up to 3.1% as obese when their body fat did not bear that out, a familiar complaint from muscular people.

Adding the waist measurements identified obesity in an extra 3.3% to 7.4% of participants, depending on the country.

“BMI can be a useful starting point, but it doesn’t tell us enough,” said Robert Eckel, an endocrinologist at the University of Colorado Anschutz and one of the authors.

Who is actually at risk?

This is the finding that could change how doctors talk to patients. People flagged as overweight or obese by BMI alone, without the waist measurements to back it up, did not show a higher risk of death.

Once the full test was applied, clinical obesity accounted for between 23% and 31% of all obesity cases, or 5.6% to 10.9% of each population. That group carried a substantially higher risk of dying over 10 years, about 69% higher than people without obesity or organ problems, while people with preclinical obesity carried about a 19% higher risk. The researchers did not run the death analysis on the US group, whose follow-up was too short.

Most people with preclinical obesity did not get worse: between 77% and 91% were still free of clinical obesity a decade later. “Most people with preclinical obesity, but not all, remained free of clinical obesity for a decade,” Professor Rubino said.

The study also tested the popular rule of thumb that your waist should be less than half your height. Using that single cut-off for everyone pulled in far more people, but those caught only by the lower line did not show extra risk of death. Waist limits matched to each ethnic group sorted risk more usefully, because people store fat differently.

What does it not show?

Independent experts called the work large and carefully done, and named its limits. It looked back over existing health records, which were not collected the same way in every country. Whether obesity caused a given person’s organ problems was estimated statistically, not checked one by one. Body fat was not measured directly, and the effect of the new GLP-1 weight-loss medicines was not studied. Marie Spreckley, a weight management researcher at the University of Cambridge, noted the study does not show that using the new definition leads to better health for patients.

It also offers no new advice for individuals. Anyone worried about their weight or waist should talk to their doctor.

What does it mean in Australia?

On Australian Bureau of Statistics surveys, about two in three adults are overweight or obese by BMI and close to one in three are obese, figures that drive everything from public health campaigns to who is offered weight-loss medicines. No Australian adults were in this study. But if doctors here adopt the new test, the counting would change: some people now labelled obese would be reassured, some with normal-looking BMIs would be flagged by their waistline, and treatment could be steered towards those whose bodies are already paying the price.

“These findings reinforce that obesity is not a one-size-fits-all diagnosis,” Dr Eckel said.