Body mass index (BMI) may not measure obesity accurately

by Kristina Fiore

The scale of the obesity epidemic may be much worse than currently believed, because the usual measure, body mass index (BMI), is a very insensitive measure of excess body fat, researchers said.

In a single-center study, 66% of patients classified as obese on the basis of DEXA scanning had BMI values in the nonobese range, according to Eric Braverman, MD, of New York Presbyterian Hospital/Weill Cornell Medical Center.

Among more than 1,000 patients, 56% were obese according to the DEXA results, versus 20% using the standard BMI-based definitions.

Braverman and colleagues presented the findings during a press conference here at the American Association of Clinical Endocrinologists meeting.

Scoffing at BMI as the “baloney mass index,” Braverman said it’s “very likely that obesity is a much bigger epidemic than the 300 million people acknowledged by the World Health Organization.”

Currently 23% of Americans are labeled obese by BMI.

He explained that BMI is just a mathematical equation based solely on height and weight that is too general for diagnosing anything, especially in such an exacting field of clinicians.

“If any endocrinologist would rely on math to calculate thyroid stimulating hormone, for instance, he would be laughed at,” Braverman said.

DEXA scans, on the other hand, provide a direct measurement of body fat percentage. It can spot fat exactly, in every part of the body.

It’s particularly effective, Braverman said, for that part of the population that is known as “thin-but-unfit.” Their condition is known as normal-weight obesity, in which the BMI is low but they have a high percentage of body fat, especially compared with more favorable tissue like muscle.

These patients are at higher risk of dyslipidemia, as well as hypertension among men and cardiovascular disease among women.

The researchers conducted a retrospective analysis of 1,234 patients seen at a private medical practice in the U.S. from 2003 and 2009. All had both BMI measurements and DEXA scans available.

The BMI threshold for obesity was 30; with DEXA, a score of 25% body fat or higher in males and 30% or higher in women marked obesity.

The researchers found that DEXA identified 56% of patients as obese while BMI identified 20% as such.

Some 37% of patients were misclassified with BMI, the researchers indicated.

Among those classified as obese by DEXA, only 34% were classified as obese by BMI.

Also, 5% of patients identified as obese by BMI actually weren’t obese according to DEXA scans.

“These individuals were muscular and large, so they look like they’re high weight but they really had high muscle mass,” Braverman said.

So the researchers concluded that BMI is a highly insensitive measure of obesity prone to under-diagnosis, while direct fat measurements are superior because they show distribution of body fat.

“This means that we may have more health problems, as individuals are delaying treatment because they don’t think they’re obese,” he said. “They think they’re thin and ‘just a little flabby.’”

They called for additional studies to confirm the results and to “determine the true nature of the obesity epidemic.” Further analyses are also needed to identify which patients may be at risk of mislabeling by BMI.

Kristina Fiore is a MedPage Today staff writer.

Originally published in MedPage Today. Visit for more endocrinology news.

Comments are moderated before they are published. Please read the comment policy.

  • Tex Bryant

    I agree that BMI is a poor way to measure obesity. Yet, in the proposed “meaningful use” criteria for EHR’s, BMI is one of the measures to be tracked for adolescents of age 2 to 20.

    The Wall Street Journal online on January 26, 2010 had a very good article on this topic–”The Scales Can Lie: Hidden Body Fat.” It also mentioned home scales that measure body fat by using an electrical pulse sent through the feet.

  • Doc99

    According to BMI, Shaq, even in his prime, is obese.


    It’s not just your BMI, but the fat distribution that matters. It has been shown that the patients with predominantly abdominal obesity do worse than the patients with just a high BMI when admitted to the hospital.

  • Eddy

    BMI is a complete crock. I am six feet tall, 185 pounds, and athletic. I have a 33 inch waist. The BMI calculators label me “overweight”. Not so much.

  • Payne Hertz

    Perhaps the definition of obesity itself is flawed. As bad as BMI is, these scans seem to be exaggerating the misdiagnosis of obesity rather than minimizing it, and may well be detecting a “problem” where none exists.

    Reminds me of my time in the service as a paratrooper. The Army was absolutely obsessed with weight, and used the “pinch test” as a measure of body fat and thus obesity. Most of us were pretty gaunt from our rigorous training, which involved 3 to 5 mile runs everyday and an 8 to 15 mile run at the end of the week, as well as a lot of physical labor, marching and walking. No one capable of this kind of training could be said to be “unfit.”

    Yet anyone with any degree of body fat or who appeared to be “overweight,” particularly bodybuilders, was forced to diet or do remedial PT. It was a complete disaster as they eventually decided everyone was too fat and cut the meat ration, while allowing all the bread and soda you could consume, resulting in the whole battalion becoming sick and weak.

    I can see a similar obsessiveness in patients being cajoled into excessive weight loss to get their DEXA index within accepted limits, and of course the usual moralizing about “personal responsibility” when they fail to do so.

  • Bladedeoc

    This stuff is insane. BMI is an awful measurement — but usually obviously so (see muscular male) and therefore only an idiot applies it indiscriminately. That being said if you define the vast majority of the population as being either overweight or obese, maybe you’ve screwed up the definition of “normal”.

  • Kim

    I at one time lost a lot of weight and while my BMI was 18, I was gaunt. The medical professionals I saw were only concerned about weight-and were of no help in gaining the weight back.

  • Alex

    Countless studies have found an association between BMI and risk for cardiovascular disease. Perhaps more studies should be done to calibrate risk with these DEXA measurements. With increased accuracy, we may be able to refine risk stratification based on the distribution of fat, as suggested by one of the comments above, and not just the overall score.

  • Mary Brighton, MS, RD

    I recently attended a medical conference where one lecture addressed exactly the same issue as this post.
    BMI to assess obesity isn’t perfect, but it is a quick and easy measurement as part of an assessment (in my case as a dietitian).
    I look at waist hip ratio, waist circumference, and use calibers or a body fat machine (which uses bioelectrical impedance) to get further information on how “healthy” a client actually is.
    BMI should just be one part of the puzzle when assessing overweight (even underweight!) or obesity.

    Mary Brighton

Most Popular