Beyond BMI
We need new terms for a new era. Let’s start with the fact that the concept of Body Mass Index, or BMI, was created by a Belgian astronomer and statistician in 1835, as a way to describe the “average man”. Body Mass Index (BMI) is a number derived from a simple calculation: a person's weight (in kilograms) divided by height (in meters) squared: kg/m².
BMI provided a ready and convenient way to place individuals into weight status categories such as underweight, normal weight, overweight, and obese. With time its usage spread, and in 1972 the now-ubiquitous term “BMI” was coined and advocated for use in population studies. The World Health Organization (WHO), the Centers for Disease Control and Prevention (CDC), and countless other health organizations now use BMI as a basic metric to categorize weight health, and to estimate general health risk.
A metric in almost all electronic health records, BMI is frequently if not universally found in patient health assessments, and is a commonly used gauge or threshold to calculate whether a patient has a healthy body weight for their height. In everyday medical practice, it’s often one of the first numbers used to guide conversations about weight and overall health.
It is therefore important for clinicians and patients alike to be aware of what BMI actually measures, and what its value is in this new era of medical weight loss and weight management.
BMI is unquestionably useful as an epidemiologic and demographic tool, with demonstrated associations between BMI >30 and increased morbidity and mortality. But BMI is of doubtful accuracy and limited usefulness in treating specific individuals for obesity, and in measuring the progress and the success of their treatment. Magnifying the problem is the importance that BMI has been accorded by telehealth GLP-1 retailers like Ro, WW, and Medvi, and many others, who rely heavily or exclusively on BMI as both a means of determining eligibility for treatment and evaluating the success of treatment.
If BMI ain’t all that, what’s better? Decades of experience in treating obesity have led me to develop and rely on more accurate, informative, and actionable ways of assessing a patient’s weight and metabolic health. But first, let’s dig a little deeper into why BMI falls short.
BMI does not distinguish between fat and muscle. Because BMI only uses height and weight, it assumes all weight is the same. However, back on earth, muscle, fat, bone, and water all have varying densities and weigh differently and have different health effects. Why does this matter in obesity medicine? Because, for example, a highly muscular person like a professional athlete may have a relatively high BMI (>30) and thus be classified as “obese,” even though they have low body fat and overall be a low health risk. Conversely, someone with low muscle mass but a rather high percentage of body fat (sometimes called “skinny fat”) may have a normal BMI, but still be at higher metabolic risk. These examples illustrate that BMI can overestimate or misdiagnose obesity in certain fit and healthy individuals, and misdiagnose them as being at risk, while underestimating risk in people with low muscle and a higher fat percentage.
BMI does not account for fat distribution.. BMI tells you how much total weight someone has, but nothing about location - and where on the body fat is stored is of crucial importance in evaluating someone’s overall health risk. Visceral fat, which is the fat around organs in the abdomen, is strongly linked to cardiovascular disease, type 2 diabetes, and insulin resistance. On the other hand, subcutaneous fat - the fat under the skin - has much less associated health risk. You can have two individuals with the same BMI, but one could carry excess fat in the abdomen, while the other carries subcutaneous fat in the hips or thighs. The patient with fat in the abdomen is at higher metabolic health risks than the other. Consequently, BMI has the ability to miss high-risk patients who have normal weight but excess abdominal fat, or overlook the severity of obesity-related risk.
BMI ignores individual differences in age, sex, and ethnicity. This limitation is a tricky one, as it is hard to quantify these varying traits. Because BMI uses the same cutoff for everyone, it discounts the fact that body composition and health risk vary significantly across different age, sex, and ethnic groups and cohorts. This matters for a number of reasons, but I will go over the three main characteristics: gender, age, and ethnicity.
Gender: Women naturally have higher body fat percentages than men at the same BMI, and men tend to accumulate more visceral (riskier) fat.
Age: Older adults lose muscle mass (a condition called sarcopenia) and gain fat, and a “normal” BMI in older adults can still mask high body fat and frailty risk.
Ethnicity: Some populations, notably from Asia, develop metabolic diseases at lower BMI’s, while others may have higher muscle mass or different fat distribution patterns.
The clinical consequence of adhering to standard BMI thresholds in the face of these age, gender and ethnic variabilities is that it may underestimate risk in some groups, may misclassify obesity status across populations, and may lead to less personalized, less accurate, and less effective care.
BMI is a poor indicator of metabolic health. Because BMI does not measure what is happening inside the body metabolically, it is unable to provide clinically relevant or actionable information about a person’s health condition. It’s like using twenty-dollar reading glasses when what you actually need is a good microscope. Some individuals with a high BMI can in fact be metabolically healthy (they have normal blood sugar, cholesterol, blood pressure, etc.), while some individuals with a normal BMI may have poor metabolic health (insulin resistance, high triglycerides, fatty liver, etc.). So, major takeaway alert: BMI does not equate to health status.
BMI oversimplifies obesity as a single number. Obesity is complicated. It is a complex, multifactorial disease, not just a matter of excess weight that can be defined and categorized according to simplistic numerical thresholds. Obesity involves many interacting factors:
Hormones (e.g. leptin, insulin, ghrelin)
Genetics (predisposition to fat storage/metabolism)
Environment and lifestyle (diet, physical activity, sleep, stress)
Psychological factors (behavior, mental health)
BMI captures none of these underlying causes or mechanisms. So, if two patients have the same BMI, they could also have completely different causes of obesity, different health risks, and different responses to treatment. Relying on BMI runs the risk of applying a one-size-fits-all approach to every patient instead of evidence-based and individualized treatment.
BMI’s accuracy drops in specific populations with different body composition or physiological states.
For example:
Older Adults
Lose muscle mass (sarcopenia) and may gain fat
Their BMI can appear “normal” while body fat is actually high
BMI does not capture frailty risk (the chance that someone is physically weak and more likely to get hurt or sick easily)
Children & Adolescents
Their bodies are still developing, so BMI has to be interpreted using age and sex-specific percentiles, not the cutoffs
Adult BMI categories are not applicable
Pregnant individuals
Weight includes fetus, placenta, prenatal fluid
BMI is therefore not a valid measure of adiposity during pregnancy
Relying on BMI without regard to differences of ethnicity, age, and gender increases the risk of over- or underestimating obesity, which can in turn lead to poorly informed clinical decisions and misdirected or ineffectual treatment.
BMI cannot measure body composition directly. Understanding the patient’s whole body composition is important in obesity medicine. Some crucial factors include:
Percent body fat
Lean muscle mass
Bone density
Water weight
Two individuals with the same BMI can still have completely different body compositions, illustrating one of BMI’s many limitations for accurately diagnosing obesity.
Relying on BMI may contribute to misdiagnosis or undertreatment of serious risk factors. Relying too heavily on BMI can mean neglecting the use of more valuable indicators, such as fat loss vs. muscle loss or gain, and may increase the risk of incorrect or incomplete clinical decision-making. As noted before, patients with a normal BMI but at high metabolic health risk can easily be missed or not treated when BMI is employed as the definitive or the stand-alone metric. This can lead to delayed diagnosis of conditions like type 2 diabetes and cardiovascular disease.
BMI, From Official Threshold to Standard Measure of Weight Loss Success
The FDA-approved indications for all anti-obesity medications, including semaglutide (Wegovy) and tirzepatide (Zepbound), are defined entirely by BMI thresholds. (Those thresholds are BMI 30+, or BMI 27+ with at least one comorbidity - hypertension, type 2 diabetes, and others). Telehealth obesity programs like Ro must prescribe within FDA-labeled indications to maintain regulatory compliance as well as to conform to insurance and pharmacy procedures. So BMI serves as the determinative criterion for screening for eligibility, and for tracking the progress of treatment. Compounding the reliability and accuracy problem is that telehealth platforms rely heavily not just on BMI, but on self-reported BMI. Yeah…what could go wrong?
There are Better Ways
Fortunately, we have better diagnostic options and systems than relying on BMI. Various techniques and measurements can be used to evaluate fat distribution. Waist circumference measures abdominal (visceral) fat, which allows it to be a strong predictor of heart disease and type 2 diabetes. Another widely used measurement is the classic Waist-to-Hip Ratio (WHR). This compares waist size to hip size, and identifies fat distribution patterns (“apple or pear shape”), making it a better indicator of metabolic risk than BMI alone. But measuring yourself can lead to inaccurate or unreliable results, as patients (and even institutions) are sometimes unsure as to where exactly to wrap the tape. (The WHO recommends measuring at the midpoint between the lowest rib and the iliac crest, while the NIH/NHLBI recommends just above the iliac crest. Got that?)
Let’s discuss body composition analysis. When it comes to full-body composition scanning technology, DEXA is the gold standard. DEXA stands for Dual-Energy X-ray Absorptiometry, which is essentially a type of test that uses low doses of X-rays to measure body fat, muscle mass, and bone density accurately. But DEXA is not often found in private clinical settings (largely due to cost and availability) and is usually only used in extreme cases or for rare conditions.
Bioelectrical Impedance Analysis (BIA) is another method of evaluating a patient's body composition, and it is the technology that we use at Forbes Health. InBody, Seca, and ImpediMed are manufacturers of widely used BIA systems. Bioelectrical Impedance Analysis (BIA) sends small electrical currents through the body. This allows the scanner to generate measurements of percentage body fat, muscle mass, and water, specific to various parts of the body. It is quick, non-invasive, and commonly used in clinics (and gyms). And while BIA scans do not directly measure body composition as DEXA scans do, they are used to validate against DEXA data and employ predictive equations and models, allowing them to provide reasonably accurate estimates.
