In Simple Terms
Weight and height alone may not tell the whole story about how obesity affects health. A large study found genetic clues linked to obesity-related health problems that BMI can miss, and these clues may help identify heart risks more accurately. The findings are promising, but the approach needs more testing before it can guide care for individual patients.
From One Number to a Broader Health Picture
Body mass index, or BMI, is calculated by dividing weight by height squared. That makes it easy to use in population studies and initial health checks. But BMI cannot directly distinguish fat from muscle, show where fat is stored, or reveal whether excess fat is affecting the function of organs or tissues. Two people with the same BMI may have different amounts and distributions of body fat—and different health effects.
That gap is at the heart of a shift toward the idea of clinical obesity. The term refers not simply to a high weight according to a formula, but to excess body fat considered alongside whether it is associated with impaired organ or tissue function. A 58-member commission published a new definition of obesity in The Lancet Diabetes & Endocrinology the year before this study, with the aim of supporting more objective diagnosis, informing clinical decisions, and helping set priorities for treatment and public health.
Under the commission’s recommendations, BMI should have a limited role: it can be a rough tool for estimating risk across populations or for screening, but it should not serve on its own as an individual verdict on someone’s health. To confirm excess body fat, clinicians can directly measure body-fat percentage when possible, or use at least one additional body measurement—such as waist circumference, waist-to-hip ratio, or waist-to-height ratio—alongside BMI. The measurement methods and cutoffs should be appropriate for a person’s age, sex, and population background.
The commission offers a practical exception for people with a BMI above 40 kg/m²: excess body fat can be presumed without an additional confirmatory test. Once obesity has been established, whether or not it is accompanied by impaired organ or tissue function, the guidance recommends assessing whether the person has clinical obesity. The distinction matters because excess fat alone does not always give a complete picture of its effects on health.
A Genetic Analysis Across Diverse Populations
The new study used this framework to investigate the genetic basis of clinical obesity. The team included nine researchers from Johns Hopkins University and one researcher from The Ohio State University. They analyzed data from 151,642 people with clinical obesity and compared them with 128,874 people who did not have recorded obesity. The participants represented five groups with different ancestral backgrounds.
The data came from UK Biobank and the US All of Us Research Program. UK Biobank contains biological, health, and behavioral information collected from about half a million participants. All of Us, run by the National Institutes of Health, was designed to advance precision medicine by collecting data from one million or more people living in the United States. These large resources make it possible to test whether findings can be seen across diverse populations, rather than relying on a small sample or a single genetic background.
In their analysis, the researchers identified 127 independent genetic loci associated with clinical obesity. Notably, 63 of these loci did not show a statistically significant association with BMI. This difference does not prove that BMI is useless. Instead, it suggests that defining obesity with a broader set of clinical measures may reveal genetic features that studies centered on weight and height do not capture.
A Genetic Signal Linked to Blood Fats
One of the clearest differences involved a genetic variant known as rs15285. It lies in a regulatory region of the LPL gene, which carries the instructions for making an enzyme called lipoprotein lipase. This enzyme helps the body process fats in the blood, including triglycerides.
The team found that the variant’s location overlapped with genetic signals associated with LPL activity, and that it was linked to higher triglyceride levels. The connection may matter because blood contains a mixture of fats and other molecules shaped by metabolic processes, and disruptions in these processes can be associated with cardiovascular risk. But observing a genetic association is not enough to prove that the variant causes a particular disease, nor does it determine what medical care is appropriate for every person who carries it.
The finding also shows why obesity may be more complicated than ranking people by BMI. Genetic differences can be associated with specific processes, such as the regulation of fats in the blood, without showing up in the same way in a measure of weight relative to height. More research is needed to understand the mechanisms linking genetic variation to changes in blood fats and, in turn, to health risks.
A Score That Combines Many Genetic Signals
The researchers did not stop at examining individual genetic loci. They also created a polygenic score for clinical obesity, combining information from many genetic variants to estimate a person’s inherited predisposition to the trait. The score was associated with BMI, body-fat percentage, and waist circumference, suggesting that it captured connections with several measures of the body—not just BMI.
Compared with a genetic score based on BMI, the new score showed stronger associations with cardiovascular risk factors and protein markers linked to inflammation. Those markers included interleukin-6, fibroblast growth factor 21, and hepatocyte growth factor. The score was also associated with indicators related to insulin resistance, including adiponectin, resistin, leptin, and the leptin receptor.
This comparison does not mean the new score can diagnose inflammation or insulin resistance, or replace established medical measurements. It is a statistical method for combining genetic signals and then testing how well they align with other health traits. The findings could eventually help improve risk estimates for some groups, but the score’s practical value will depend on showing that it adds useful information to what clinicians already learn from tests and a patient’s health history.
A Possible Link to Heart Attack and Stroke Risk
The researchers reported that the clinical-obesity score was better than the BMI-based score at distinguishing risk levels for three important events: myocardial infarction, or heart attack; heart failure; and stroke. Better discrimination means that a statistical model may sort people into relative risk levels more accurately. It does not mean the model can predict with certainty that a particular person will develop a disease.
The team tested the improvement in four long-running US health studies. The first was the Atherosclerosis Risk in Communities study, known as ARIC. It expanded its focus from risk factors for atherosclerosis that is not yet clinically apparent to the risks of heart disease, including heart attack, coronary artery disease, stroke, and heart failure. About 16,000 adults in the United States took part.
The second was the Multi-Ethnic Study of Atherosclerosis, or MESA, sponsored by the National Institutes of Health’s National Heart, Lung, and Blood Institute. MESA examines features of cardiovascular disease before symptoms appear, as well as factors that may predict progression to overt disease or worsening early changes. The study included 6,814 adults at baseline between 2000 and 2002.
The comparison also included the Framingham Heart Study, which began in 1948 to identify common traits and factors that contribute to cardiovascular disease. Since 1971, Boston University School of Medicine has served as the study’s contractor and academic partner to the National Institutes of Health. More than 15,400 participants have been enrolled across its different study cohorts.
The fourth study was the Women’s Health Initiative, which received its initial funding in 1992 and focuses on strategies to prevent heart disease, breast cancer, colorectal cancer, and osteoporosis in postmenopausal women. A total of 161,808 women were enrolled. Testing the score across these different projects offered a way to assess whether the association held up in independent datasets and across populations with varied characteristics and research goals.
Promising Results, but Not a Ready-Made Clinical Test
In a preliminary version of the study, the researchers reported that classifying participants into low, medium, and high genetic-predisposition groups led to 35% of people being reclassified when the clinical-obesity score was used instead of the BMI-based score. Reclassification means that some people moved into a different risk category under the new model. It does not mean that 35% developed a disease or that changing their classification would automatically improve their health outcomes.
The team emphasized that more research is needed to establish whether these scores offer added value for diagnosis, risk classification, or management beyond assessments based on BMI. Genetic testing for some rare forms of obesity can already help guide specific treatments, but that does not show that a general polygenic score will lead to an appropriate or better treatment decision for every person. The score’s usefulness may also depend on how well the populations used to build it represent different groups.
The findings were presented in a research poster at the American Society of Human Genetics 2026 annual meeting, held in Montreal from October 20 to 24. The work was titled Genome-Wide Analysis and Polygenic Prediction of Clinical Obesity. The study’s co-corresponding authors were Soheil Zahid, MD, PhD, a researcher at the Ciccarone Center for the Prevention of Cardiovascular Disease at Johns Hopkins University, and Alexis Patel, PhD, director of the Malone Center for Engineering in Healthcare and director of strategy and partnerships at the university’s Data Science and AI Institute. The team also posted a preliminary version of the study on May 22.
The Bottom Line
The findings support the idea that clinical obesity may carry genetic and health information that BMI alone misses. The researchers identified 127 associated genetic loci, 63 of which were not significantly associated with BMI, and developed a genetic score linked to markers of cardiovascular health, inflammation, and insulin resistance. In the analyses described, the score also distinguished risk for heart attack, heart failure, and stroke better, and it reclassified 35% of participants in the preliminary analysis. The results were seen across four large US studies, but they are a research advance—not a ready-to-use clinical tool. The score’s added value still needs to be demonstrated. More broadly, obesity may be assessed more accurately by combining measures of body fat and body shape with organ function and a person’s health context, rather than reducing the picture to a single number.