What to add to BMI, which reads crowds better than it reads you

Key takeaways

• BMI is your weight in kilograms divided by your height in meters squared, and CDC calls it a screening tool, not a diagnostic one, that can’t tell fat from muscle and bone or show where fat sits⁴⁵.
• In a Lancet analysis of 239 studies covering 10.6 million people, death rates were lowest at a BMI of 20 to 25, and each 5 points above 25 went with a 31% higher death rate³.
• In June 2023, the American Medical Association called BMI an imperfect measure of body fat, noted it rests mainly on data from earlier generations of non-Hispanic white people, and urged doctors to pair it with other measures⁶.
• Cutoffs by ancestry differ by country: CDC applies the same adult categories to everyone, while the U.K.’s NICE starts overweight at 23 and obesity at 27.5 for people of Asian, Middle Eastern, Black African or African-Caribbean background⁴⁷.
• A tape measure adds a lot: NICE advises adults with a BMI under 35 to keep their waist under half their height⁷.

On June 3, 1998, about 29 million Americans became overweight without gaining an ounce¹.

That day, new federal guidelines from the National Institutes of Health set the line for “overweight” at a body mass index, or BMI, of 25¹. Until then, the National Center for Health Statistics had used about 27 for women and 28 for men¹. Under the new rule, some 97 million adults, about 55%, counted as overweight or obese¹.

Nothing about those people’s bodies had changed. Only the cutoff had.

That story captures BMI’s double life. It’s a cheap, useful way to track risk across large groups³⁴. Applied to one person, it can mislead⁶.

In June 2023, the American Medical Association said as much, calling BMI an imperfect way to measure body fat and urging doctors to use it alongside other measures⁶. Below, we cover what BMI gets right, where it goes wrong, and which two measures to add.

A formula from 1832 got its modern name in 1972

BMI’s math goes back to Adolphe Quetelet, who first proposed it in 1832². He was interested in how human traits are spread across whole populations². He noticed that as people grow, weight rises roughly with the square of height, not the cube you’d expect if bodies simply grew larger in every direction².

In 1972, Ancel Keys and four colleagues tested four weight-for-height formulas in the Journal of Chronic Diseases². They wanted one that tracked weight closely, had little to do with height, and matched measures of body fat². Weight divided by height squared won. They named it body mass index and called it ideal for all human populations².

The cutoffs came later. A 1985 paper proposed 25 and 30 as adult lines for overweight and obesity². As a 2014 commentary in the International Journal of Epidemiology noted, those lines rest on statistics rather than on a direct measure of how the body works².

To compute yours, divide your weight in kilograms by your height in meters squared⁵. CDC’s example: an adult who is 5 feet 8 inches tall and weighs 200 pounds has a BMI of 30.4, in the obesity range⁴.

BMI’s strength is tracking risk across millions

Some of the strongest evidence for BMI comes from the Global BMI Mortality Collaboration, which pooled 239 studies of 10,625,411 people in Asia, Australia and New Zealand, Europe and North America³. Its 2016 report in The Lancet followed people for a median of 13.7 years³.

The main analysis focused on 3,951,455 people who had never smoked, had no chronic disease at the start, and survived the first five years³. That design is meant to keep smoking and existing illness from muddying the link between weight and death. Of that group, 385,879 died, about 1 in 10 by our math³.

Death rates were lowest at a BMI of 20 to 25 and rose on both sides³. Compared with that range, the hazard ratios, which compare death rates, looked like this³:

BMI range	Hazard ratio for death
15 to under 18.5	1.51
18.5 to under 20	1.13
25 to under 27.5	1.07
27.5 to under 30	1.20
30 to under 35	1.45
35 to under 40	1.94
40 to under 60	2.76

That’s a J-shaped curve: risk rises at the low end and climbs further at the high end. Above 25, each 5 points went with a 31% higher death rate³. As an illustration, if 10 in 100 people at a healthy BMI died over a study, a 31% higher rate would mean about 13 in 100, by our math.

The National Institutes of Health and three British funders, including the British Heart Foundation, paid for the work³.

The 2014 commentary offers a twist. BMI is a poor gauge of fat, its author argued, but studies in children and teens tie height and lean mass to heart and metabolic risk on their own². So part of the risk BMI captures has nothing to do with fat, and its weakness as a fat gauge may be part of its strength as a risk index².

What BMI can’t see in your body

CDC lists the blind spots plainly. BMI isn’t a direct measure of body fat, can’t separate fat from muscle and bone, and can’t show where fat is stored⁴. A muscular person or a trained athlete can land in the overweight or obesity range⁴⁵.

The American Medical Association’s 2023 policy makes the same point from the other side: across a whole population, BMI rises and falls with body fat, but for any one patient it’s a much shakier guide⁶.

Picture two hypothetical adults, both 5 feet 8 inches and 200 pounds, with a BMI of 30.4⁴. One trains for strength competitions, and the other carries much of the weight around the middle. BMI rates them the same.

Where fat sits matters, too. NICE, the U.K.’s National Institute for Health and Care Excellence, recommends a waist-to-height ratio because it gives a truer estimate of fat around the middle⁸. Our metabolic syndrome article explains U.S. waist cutoffs.

In fairness, CDC says that for most people BMI is a good indicator of too much or too little body fat, and that it lines up moderately to strongly with direct fat measures⁴. The trouble is the people it misreads.

Age and sex change what a number means

At the same BMI, older adults tend to carry more body fat than younger adults, and women tend to carry more than men, according to CDC guidance for clinicians⁵.

Yet BMI’s link to death is weaker in old age. In the Lancet analysis, each 5 points above 25 raised death rates 52% at ages 35 to 49 but 21% at ages 70 to 89³. The link was also stronger in men than in women: 51% versus 30% higher per 5 points³.

NICE also tells clinicians to interpret BMI with caution in people 65 and older, partly because a slightly higher BMI may protect older people⁷. Muscle loss and spine conditions can also reduce how well older adults function⁸.

NICE gives the same warning for adults with high muscle mass⁷.

Ancestry: the same BMI can carry different risk

In 2004, a World Health Organization expert consultation published in The Lancet found that many Asian people face high risk of type 2 diabetes and heart disease at BMIs below 25⁹. Across Asian populations, the BMI where risk began ranged from 22 to 25⁹.

WHO kept its standard categories worldwide but added “public health action points” at 23, 27.5, 32.5 and 37.5, and left each country to decide how to use them⁹.

NICE went further. Its obesity guideline, updated in January 2025 with ancestry thresholds first set in 2022, uses lower lines for people of South Asian, Chinese, other Asian, Middle Eastern, Black African or African-Caribbean background⁷:

• Overweight: BMI 23 to 27.4
• Obesity: BMI 27.5 or higher

The evidence NICE reviewed showed rising risk at a BMI of 21 to 26 in people of South Asian and Chinese background, compared with 25 to 29 in white people⁸. In Middle Eastern, Black African, Black Caribbean and other Asian groups, risk rose anywhere from 21 to 30, though mostly below 25⁸.

U.S. data point the same way. In a 2009 California survey of 45,946 adults, Vietnamese, Korean, Filipino and South Asian adults with a “normal” BMI of 23 to 24.9 had more diabetes than white adults in that range¹⁰.

CDC, by contrast, says to use the standard adult categories regardless of age, sex or race⁴. And the American Medical Association has acknowledged BMI’s use in racist exclusion and its roots in data from earlier generations of non-Hispanic white people⁶. If your family background is on NICE’s list, ask your clinician which threshold applies to you.

Children are measured against a growth curve

For ages 2 through 19, CDC compares a child’s BMI with other children of the same sex and age, using percentiles¹¹. Race isn’t part of the calculation¹¹.

Category	BMI-for-age
Underweight	Below the 5th percentile
Healthy weight	5th to below the 85th percentile
Overweight	85th to below the 95th percentile
Obesity	95th percentile or higher
Severe obesity	120% of the 95th percentile or higher, or a BMI of 35 or more

The American Academy of Pediatrics’ 2023 guideline split severe obesity further into class 2 and class 3, and CDC lists those classes¹¹.

Charts had to change, too. CDC’s 2000 charts drew on children measured between 1963 and 1980, and their scale stopped at a BMI of 37¹². In December 2022, CDC released extended charts that add higher percentile curves and plot BMI up to 60¹². Severe obesity among children and teens rose from 1% in 1971–1974 to 6.1% in 2017–2018¹².

A percentile is still a screen. CDC’s guidance for clinicians says a child’s assessment should also weigh diet, activity, family history and other checks⁵.

Why BMI still runs the rules

The American Medical Association’s 2023 policy says BMI shouldn’t be the only reason an insurer denies appropriate payment⁶. Yet coverage rules still lean on it:

• Medicare obesity counseling. Since 2011, Medicare has covered intensive behavioral therapy in primary care for adults with a BMI of 30 or more¹³. Visits run weekly the first month, then every other week through month six¹³. Monthly visits for months 7 through 12 continue only if you’ve lost at least 3 kilograms, about 6.6 pounds by our math¹³.
• Medicare’s 2026 drug bridge. Starting July 1, 2026, a temporary Medicare GLP-1 Bridge program covers a tablet called Foundayo, Wegovy injections or tablets, and Zepbound in its KwikPen form only, for a $50 copay per one-month supply¹⁴. Adults qualify with a BMI of 35 or more; 30 or more plus uncontrolled high blood pressure, a type of heart failure or chronic kidney disease stage 3a or higher; or 27 or more plus prediabetes, a past heart attack or stroke, or leg or arm artery disease with symptoms¹⁴. A clinician must certify that the drug is part of a lifestyle program, and approvals run through December 31, 2027¹⁴. People with type 2 diabetes, moderate-to-severe sleep apnea or fatty liver disease aren’t eligible, though their Part D plan might cover the drugs¹⁴.

Those thresholds echo 1998, when NIH guidelines suggested diet drugs at a BMI of 30, or 27 with risk factors such as diabetes or high blood pressure, after six months of diet and exercise¹.

A 2025 global commission called for BMI to serve as a stand-in for populations rather than a measure of individual risk¹⁵. Our companion article on weight and metabolic health covers its proposals. A later article in this series covers what these medicines do, including side effects and what happens when people stop.

BMI plus two: a tool for your next checkup

BMI is easy to compute at home, and so is a useful partner for it: NICE encourages adults to measure their own waist-to-height ratio⁷. That suits anyone who likes to understand their own body.

This tool is our own, built from the sources cited. Find your situation, add two measures and bring the question.

If this describes you	Why BMI may mislead	Two measures to add	What to ask
Muscular or strength-trained	BMI can’t tell muscle from fat4	Waist-to-height ratio7; blood pressure	“Do my waist and blood pressure match what my BMI suggests?”
Age 65 or older	More fat at the same BMI, and a slightly higher BMI may protect57	Waist-to-height ratio; a check of strength and daily function8	“Is my weight helping or hurting how well I move?”
South Asian, Chinese, other Asian, Middle Eastern, Black African or African-Caribbean background	Risk can rise at a lower BMI79	Waist-to-height ratio; a blood sugar test	“Should a lower threshold apply to me?”
Normal BMI with a family history of diabetes or a growing waist	BMI can’t show where fat sits4	Waist-to-height ratio; a blood sugar test	“Could I have metabolic risk at a normal weight?”
BMI of 35 or more	NICE’s waist ratio advice covers BMIs under 357	Blood pressure; blood sugar and cholesterol tests	“Is my weight affecting my organs or daily life?”
A child or teen	Adult categories don’t apply; children are compared by age and sex11	BMI-for-age percentile over time; a review of diet, activity and family history5	“How has my child’s percentile changed over time?”

If tracking body numbers stirs up worry about food or weight, for you or your child, ask your clinician to handle the measurements. The National Alliance for Eating Disorders helpline, 1-866-662-1235, offers weekday support and referrals from licensed therapists. In a crisis, call or text 988.

How strong is the evidence?

Claim	Evidence level	What the research shows
BMI predicts death risk across large populations	Strong evidence	239 studies, 10.6 million people3
BMI accurately measures one person’s body fat	Evidence says no, or unsafe	It can’t separate fat from muscle and bone246
Many Asian groups face diabetes and heart risk at lower BMIs	Strong evidence	WHO review, NICE evidence review and U.S. survey data8910
Black African and African-Caribbean groups face risk at lower BMIs	Promising but early	NICE’s review found wide-ranging thresholds, mostly below 258
Waist-to-height ratio predicts weight-related conditions about as well as BMI, with one cutoff for everyone	Strong evidence	NICE evidence review8
A slightly higher BMI protects older adults	Promising but early	NICE flags the possibility; BMI’s link to death weakens with age37

Questions readers ask

The bottom line

BMI was built to compare populations, and it still does that well: across millions of people, death rates climb at both very low and high BMIs. For one adult, though, it can’t see muscle, can’t see where fat sits, and doesn’t adjust for age, sex or, in U.S. practice, ancestry.

Treat your BMI as a starting point, not a verdict. Pair it with a waist-to-height ratio and the blood pressure, blood sugar and cholesterol checks that show how your body is working, and ask which thresholds fit you. Coverage rules may still run on BMI, but your health picture doesn’t have to.

Keep exploring: metabolic syndrome criteria and insulin resistance.

Related reading

• Your weight and your metabolic health overlap without being the same thing
• Metabolic syndrome is just five numbers that together still predict trouble
• Insulin resistance is common even without obesity, and its popular test isn’t standard
• What helps a teenager’s body image, and what backfires
• What a medical test can and cannot tell you

Sources

1. CBS News. “Think You’re Thin? Think Again.” June 4, 1998. cbsnews.com
2. Wells JCK. “Commentary: The paradox of body mass index in obesity assessment: not a good index of adiposity, but not a bad index of cardio-metabolic risk.” International Journal of Epidemiology, 2014;43(3):672–674. doi.org
3. Global BMI Mortality Collaboration (Di Angelantonio E, Bhupathiraju ShN, Wormser D, et al.). “Body-mass index and all-cause mortality: individual-participant-data meta-analysis of 239 prospective studies in four continents.” The Lancet, 2016;388(10046):776–786 (University of Oxford repository). ora.ox.ac.uk
4. Centers for Disease Control and Prevention. “BMI Frequently Asked Questions.” June 28, 2024. cdc.gov
5. Centers for Disease Control and Prevention. “Body Mass Index: Considerations for Practitioners.” Undated. stacks.cdc.gov
6. American Medical Association. “AMA adopts new policy clarifying role of BMI as a measure in medicine.” Press release, June 14, 2023. ama-assn.org
7. National Institute for Health and Care Excellence (NICE). “Overweight and obesity management” (NG246). January 14, 2025. nice.org.uk
8. National Institute for Health and Care Excellence (NICE). “Overweight and obesity management (NG246): Rationale and impact.” 2025. nice.org.uk
9. WHO Expert Consultation. “Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies.” The Lancet, 2004;363(9403):157–163 (MEDLINE record). pubmed.ncbi.nlm.nih.gov
10. Jih J, Mukherjea A, Vittinghoff E, et al. “Using appropriate body mass index cut points for overweight and obesity among Asian Americans.” Preventive Medicine, 2014;65:1–6. pmc.ncbi.nlm.nih.gov
11. Centers for Disease Control and Prevention. “Child and Teen BMI Categories.” June 28, 2024. cdc.gov
12. Centers for Disease Control and Prevention. “Background: CDC Extended BMI-for-Age Growth Charts.” Growth Chart Training, September 13, 2024. cdc.gov
13. Centers for Medicare & Medicaid Services. “Decision Memo for Intensive Behavioral Therapy for Obesity (CAG-00423N).” November 29, 2011. cms.gov
14. Medicare.gov. “Weight loss drugs” (Medicare GLP-1 Bridge). Accessed October 2026. medicare.gov
15. PACE-CME. “Reframing obesity: updated definition and diagnostic criteria for clinical and preclinical obesity” (summary of Rubino F, et al. Lancet Diabetes & Endocrinology, 2025). pace-cme.org