乌鸦传媒

CHICAGO -- The traditional body mass index (BMI) measure may have misclassified millions of Americans as not having obesity, according to National Health and Nutrition Examination Survey (NHANES) data.

Among over 9,700 adults identified as having obesity according to total body fat percentage estimates using dual-energy x-ray absorptiometry (DEXA) scans, the BMI measure only accurately classified 47%, reported Aayush Visaria, MD, MPH, of Rutgers Robert Wood Johnson Medical School in New Brunswick, New Jersey, at ENDO 2023, the annual meeting of the Endocrine Society.

When ruling out obesity, BMI and DEXA measures were congruent 95% of the time, he explained.

Only 36% of participants had a BMI of 30 or more -- the traditional cutoff for obesity -- while 74% had obesity according to total body fat percentage, defined as a body fat percentage of 25% or higher in males and 32% or higher for females.

However, adding waist circumference on top of the traditional BMI measure reduced misclassifications by 23%, said Visaria. When both measures were used, only 31% of people with DEXA-confirmed obesity were misclassified as not having obesity according to a combined BMI and waist-based obesity cutoff. This combo measure and DEXA were congruent 69% of the time for classifying obesity.

"I think the biggest thing here is that BMI is used throughout population health, throughout clinical care, but what we realize here is that BMI vastly underestimates true obesity," Visaria said, "suggesting its use alone without an additional measure like waist circumference may not be sufficient."

Certain races and ethnicities had a greater chance of obesity misclassification with BMI than others. Specifically, Asian and Hispanic adults were more likely to have a normal BMI but have DEXA-confirmed obesity according to their body fat percentage compared with white adults. Asian and Hispanic adults were also more likely to have a greater proportion of abdominal fat. On the other hand, Black adults with a normal BMI were far less likely to have high body fat percentages and tended to have less abdominal fat.

Visaria pointed out that the traditional cutoffs in BMI used to separate underweight (<18.5), normal weight (18.5 to <25), overweight (25 to <30), and obesity (30+; ≥27.5 for Asians) were established decades prior in predominantly white, European men.

"Showing the racial/ethnic differences here, the gender differences, adds to the case that BMI is really maybe even exacerbating the disparities that we see," he suggested.

But it's not yet time to completely ditch BMI, Visaria noted. "I think BMI still has a lot of use -- it's one of the most accessible measures. It's simply calculated, it's trackable, it's scalable. From a population health standpoint, I do think there are some uses for it."

"But from a clinical, individual patient level, I think it's the start of the end. I think BMI should be supplemented with other measures," he added, pointing to the American Medical Association's (AMA) , which was released earlier this week.

"There are numerous concerns with the way BMI has been used to measure body fat and diagnose obesity, yet some physicians find it to be a helpful measure in certain scenarios," said Jack Resneck Jr., MD, AMA's immediate past president, in a statement. "It is important for physicians to understand the benefits and limitations of using BMI in clinical settings to determine the best care for their patients."

Looking beyond the confines of the flawed BMI measure, Visaria said clinicians should also measure waist circumference, waist-to-height ratio, and waist-to-hip ratio. "I know [BMI] is entrenched in our clinical decision making," he noted, "but as we learn more about different measures and their validity ... we will slowly move away from BMI."

Data for the analysis came from the 2011-2018 NHANES on 9,784 U.S. adults ages 20 to 59 with whole-body DEXA scan data. Mean age was 39, 49% were women, and 61% were white.

Visaria said the next steps involve looking more closely at the long-term cardiometabolic health outcomes in individuals who have obesity per DEXA scans, but not BMI cutoffs.

Comment