Researchers in Italy and Lebanon scanned 1,351 adults with dual-energy X-ray absorptiometry (DXA), the medical gold standard for measuring actual body fat, and compared the results with each person's Body Mass Index. The mismatch was striking. Among adults labeled obese by BMI, 34 percent belonged in the overweight category once fat was measured directly. Among those labeled overweight, 53 percent were misclassified.
The findings were published in the journal Nutrients in June 2025 and were selected for presentation at the European Congress on Obesity in Istanbul in May 2026. The European Association for the Study of Obesity announced the study on April 3, 2026.
What BMI actually measures
BMI is a simple ratio of weight to height squared. It says nothing about how much of that weight is fat, how much is muscle or bone, and where the fat sits on the body. A muscular 40-year-old can score in the "overweight" range with very little body fat. An older adult with low muscle mass can score as "normal" while carrying enough fat to raise metabolic risk.
DXA scans, by contrast, use low-dose X-rays to separate fat mass, lean mass, and bone. That is why obesity specialists treat DXA as the reference method when they want to know how much fat someone actually carries.
Also read: Just 22 hours of one specific brain training may cut your dementia risk by a quarter
Where BMI got it wrong
The sample included adults aged 18 to 98, 60 percent of them women, referred to the University of Verona's Department of Neurosciences, Biomedicine and Movement Sciences. When the team lined up BMI categories against DXA results, the disagreements clustered around the middle of the scale.
In the overweight group, about three-quarters of the misclassified adults actually had a normal fat percentage, while the remaining quarter carried enough fat to qualify as obese. In the obese group, a third dropped down to overweight once fat was measured directly. Even the normal-weight group was not spared: 22 percent landed in a different category on DXA, 9.7 percent were actually underweight, 11.4 percent were overweight, and 0.8 percent were obese.
Overall, DXA put combined overweight and obesity at about 37 percent of the sample, compared with 41 percent by BMI. The headline totals were close, but as co-author Chiara Milanese of the University of Verona put it, "even though both systems identify a similar overall prevalence of overweight and obesity, we are talking in some cases about different people."
What that means for your number
The study was led by Professor Marwan El Ghoch of the University of Modena and Reggio Emilia, with collaborators at the University of Verona and Beirut University. His summary was blunt: a large proportion of adults, more than a third, sit in the wrong weight-status category when only BMI is used.
Also read: Cooling your pasta or rice overnight may lower its blood sugar impact
The researchers do not argue for scrapping BMI, which is cheap, fast, and useful at the population level. They recommend pairing it with a direct body-composition measure or a simple shape marker such as waist-to-height ratio, which captures where fat sits and correlates more closely with cardiovascular risk. DXA remains the most accurate option, but it typically requires a referral and is not part of a routine checkup.
For anyone who has been told their BMI is fine, borderline, or elevated and wants a clearer read on metabolic risk, the practical step is to ask a primary care doctor about adding a waist measurement, a body-composition assessment, or a DXA scan to the conversation. The sample studied was White Caucasian adults in northern Italy, so the exact misclassification rates may differ in other populations, but the underlying limitation of BMI is the same everywhere it is used.
Also read: More than 8 in 10 people with prediabetes don't know it. Do you have these signs?
This article is made and published by Jesper Bengtson, who may have used AI in the preparation.