Welcome to Straight Talk on Health. I’m your host Dr. Chet Zelasko. Straight Talk on Health is a joint production with WGVU in Grand Rapids MI. I examine the practical application of health information. Nutrition. Exercise. Diet. Supplementation. I look at the science behind them and let you know whether it’s something to consider or not. You can check out other things that I do on my website Drchet.com and sign up for my free emails.
Body mass index has been used for decades as a way to assess whether someone is at a healthy body weight for their height. For just about as long, it’s been under attack for the people who feel it was not representative of their body composition; high muscle mass and bigger bones were two of the complaints. I’m not going to disagree, within reason, but as a way of assessing a population in large studies, it can give us some indication of how body mass is related to health.
A recently published study has called into question the use of BMI as a predictor of mortality. Researchers used the data from the 1999–2000 National Health and Nutrition Examination Survey (NHANES) to compare BMI with a method of measuring body fat called bioelectrical impendence analysis (BIA). The subjects had all metrics measured when that wave began and they were 20–49 years old. Then they obtained death certificates for everyone in the study who died in the 25 years since.
The results, presented in hazard ratios, demonstrated that BIA and waist circumference were significantly related to all-cause and CVD mortality while BMI was not significantly related to either. In the discussion, the researchers suggested that BIA for determining body fat percent be used in clinical settings to assess patients at risk for increased risk of dying. Is that a good idea? Should we throw out BMI?
This study was attempting to establish that body fat as assessed by BIA was a better predictor of the 25-year death rate than BMI. Researchers made their case by using data collected in the 1999–2000 NHANES wave to establish that a measure of body fat that can be collected with a device such as a scale or wrist monitor was better than hard data like height and weight as used in calculating BMI.
Body composition analysis was my area of expertise when I was a grad student at MSU in the 1980s. I tried to estimate the number of people I underwater weighed to assess body fat and I think it was in the thousands. I believed then, and I believe now, that underwater weighing is the best way to indirectly assess body composition. The problem is that it’s complicated and requires a significant amount of equipment to do it correctly. Getting a simple and practical way is certainly admirable if you want to test a large number of people but it’s still an indirect measurement. Even dual x-ray absorptiometry AKA DEXA is an indirect measurement because you aren’t measuring fat, bone, or muscle directly. The measurement use algorithms to come up with the percent body fat.
The only way to really assess body composition is to incinerate a body and assess the ash for the components of bone, fat, and muscle. You don’t get many volunteers for that kind of study and the human subjects review board would probably object. But decades ago, researchers took apart three cadavers into the component parts and used the data to formulate a method for indirectly assessing body fat. Since then, every method, whether it’s the Bod Pod which assesses air displacement, or BIA which uses the resistance of a low level of electricity throughout the body, has been validated by underwater weighing. If something is one level removed from an indirect method, it can’t be a direct measurement and it cannot be better.
I have a unique perspective on BIA because I was part of the laboratory that collected validation data on the original devices. The researchers continuously stated that BIA was a direct measurement of body fat in the paper. That’s not correct. BIA measures the body’s resistance to a low electrical current through the body as I said; then that number is put into an equation combined with other measures—primarily height and weight—to calculate a percentage of body fat. In that wave of the NHANES study, they also collected waist circumference and skin-fold measures to obtain the best predictive equation for use in future studies.
I’m certain that the equations have been updated over the years, but at that time, the model for calculation of body fat was 95% dependent on height and weight. Recent standard errors of the measurement of BIA for body fat range from 3.6% to 6%. There are many reasons for errors of that size, but impedance is particularly sensitive to changes in hydration and alcohol consumption. And to be clear, BIA is not a direct measurement of body fat—it’s just simple to use.
BMI is measure of surface area. It would be nice if more surface area indicated increased muscularity, but for 99.5% of us it’s not. It’s an indication, not a direct measurement, of our body fatness. It does tell us whether we are at a normal weight for our height.
BMI should not be used with hard edges. I would be hard pressed to say that there is a significant risk of anything for someone who has a BMI of 25.9 versus a BMI of 24.9; the former indicates overweight while the latter indicates normal weight. That’s a difference of just six pounds, and a good bowel movement could account for half of that. It’s not relevant in the real world. But a BMI of over 30 indicates that someone is obese, and that affects an increasing number of people in the U.S. and other industrialized nations.
BMI is still the best metric we have for assessing whether someone is at a healthy weight for their height. We can try to account for bone size and muscularity, but that doesn’t apply to most people. Getting to a healthy weight and staying there is still the best way to live the longest and the healthiest life. There’s no splitting hairs over that.
Here’s a factoid for you: my email platform tells me how many clicks we get on whatever we send readers, and by far the most-clicked page is the BMI chart on the website. And we include adjustments for frame size, so you’ll know for sure if you’re really big boned. I believe having that info easily available helps anyone make good decisions about your health, because a BMI chart is still the easiest way to determine whether your weight needs to come down. Getting to a normal body weight for height is still the optimal way of living longer while keeping good health.
I understand the rationale by physicians to be accurate in telling patients where they stand on the risk scale. While the measuring device is certainly much easier to use, it doesn’t mean that its results are any better than what BMI can provide. If we’re being honest about it, while it may be a number people want to know, we don’t need better assessments of body composition. What we need is to help them get off their good intentions and eat better, eat less, and move more.
That’s all the time I have for this episode. If you like this podcast, please hit the share button and tell your friends and colleagues about it. Until next time, this is Dr. Chet Zelasko saying health is a choice. Choose wisely today and every day.
Reference: Ann Fam Med 2025;23:Online. https://doi.org/10.1370/afm.240330
