Waist-confirmed clinical obesity, not BMI alone, tracked higher death risk in 844,108 adults

A large multicohort analysis published in Nature, led by King's College London with collaborators in China, the United States, Spain and Italy, tested the January 2025 Lancet Diabetes & Endocrinology Commission's two-step obesity definition against BMI alone. Investigators examined 844,108 adults. The framework first confirms excess adiposity with BMI plus waist circumference or waist-to-height ratio, then separates clinical obesity, excess fat with impaired organ or daily function, from preclinical obesity, in which fat is elevated but function is preserved.

Across cohorts, BMI alone missed obesity-range adiposity in up to 8.6 percent of participants and labeled up to 3.1 percent as having obesity without that confirmation. Clinical obesity comprised 23 to 31 percent of obesity cases and 5.6 to 10.9 percent of the populations studied. It was associated with a substantially higher risk of death; preclinical obesity showed only a modest increase. Between 77 and 91 percent of people with preclinical obesity had not progressed to clinical obesity at 10 years. Press coverage omits absolute death rates and hazard ratios, so the magnitude cannot be calibrated.

Calling this a better diagnosis overstates what the design can show. This is a reclassification and prognosis study, not a comparison against dual-energy X-ray absorptiometry or imaging in all participants. Higher mortality after adding waist size is expected if waist marks visceral fat and if BMI-only obesity includes fitter, more muscular adults. Defining clinical obesity by organ dysfunction and then linking it to death risks circularity, because sicker people die more often. Ethnic cut-points, waist measurement error, and occupational cohorts such as Kailuan also limit transport to primary care. Funding is unstated, and the senior author chairs metabolic surgery.

For policy, the split cuts both ways. Capturing high-waist adults whose BMI is not in the obesity range could widen eligibility, while restricting drugs and surgery to clinical obesity could shrink it. GLP-1 indications and many payer rules still use BMI thresholds. Evidence quality: this is a multicohort observational analysis using a prespecified framework, not an RCT, and not proof that the definition improves outcomes. It can show association and reclassification. It cannot show that relabeling reduces deaths. Next studies need absolute risks, decision-curve analysis, and a pragmatic trial of care stratified by clinical versus preclinical status.