Body mass index gets smack down: AMA calls out harms and “racist exclusion”

Body mass index has for decades been used as a shorthand for assessing body fat and weight-related health risks. But for about just as long, critics have noted the simple calculation is laden with problems; BMI doesn’t actually measure fat mass, account for its distribution, or how those differ by age, gender, ethnicity, race, and how those differences affect health risks. Calculations and cutoffs are largely based on past generations of non-Hispanic white people. And BMI classifications mislead people on their individual risks of disease and death and can lead to substandard care for eating disorders. Now, it seems the hefty criticism has finally reached a critical mass. During the annual meeting of the American Medical Association (AMA) this week in Chicago, physicians and medical students voted to adopt a strongly worded policy acknowledging the calculation's "significant limitations" and "historical harms," including "racist exclusion." While the massive medical group acknowledged that BMI remains useful for population-level trends and associations, the policy called for doctors to become familiar with the calculation's problems and explore alternative measurements for diagnosing obesity and assessing health risks. "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," AMA Immediate Past President Jack Resneck, Jr., said 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."

History

The shift will likely inspire cheers from critics, whose numbers have been increasing for decades. Even before the rise of BMI—aka the Quetelet Index as it was called in the 1970s—medical researchers had realized that simple features like leg length, bone mass, and body frame influenced weight-to-height ratios. BMI is a simple calculation of weight in kilograms divided by height in meters squared. In 1993, an expert advisory committee for the World Health Organization came up with four BMI-based weight categories: underweight (below 20), normal weight (20 to 24.9), overweight (25 to 29.9), and obese (30 and up). In the US then, the National Institutes of Health considered the threshold for overweight to be a BMI of 27.8  for men or 27.3 for women. Anything under those cutoffs was "normal." This was based on an 85 percent cutoff point from data in a large national NIH survey. But in 1998, the NIH lowered the cutoff to 25, aligning with WHO categories. And overnight, millions of Americans went from having normal weight to overweight. That cutoff is still used. BMI is a simple, easy-to-obtain number that has strong, established associations with body fatness, morbidity, and mortality data. But on an individual level, it's loaded with limitations. For one, it doesn't actually measure body fat mass—it can't distinguish lean mass from fat.

Disadvantages

In a report prepared by an AMA council before the vote, council members noted that updated data from a national NIH survey compared BMI to body fat percentages based on bioelectric impedance. In subjects all with a BMI of 25, the percent of body fat in men varied between 14 percent and 35 percent, and in women, it varied between 26 percent and 43 percent, the report noted. Based on the NIH's criteria for defining obesity based on percent of body fat, subjects with the same BMI—which was at the cutoff for overweight—were associated with a body fat mass that varied between "low normal" to "obese." BMI also doesn't account for fat distribution and a person's life stage. For decades, researchers have noted that the accumulation of fat in the upper part of the body ("apple-shaped" body) is linked to higher risks of diseases, like coronary heart disease and diabetes, compared to when fat accumulates in the lower body ("pear-shaped" body). Hormonal changes in puberty may be linked to some fat distribution differences. For instance, females generally develop fat stores around their pelvis and thigh during puberty, which is not seen in males, who develop a relatively larger amount of lean mass during puberty. And as women age, they tend to lose bone and muscle mass. Thus, as the AMA report notes, a 65-year-old woman who weighs the same as she did when she was 25 will actually have a higher percentage of body fat than she had in the past. Beyond gender, racial and ethnic minorities also have different weight-related health risks. For example, a 2006 study found that Asian women had more than double the risk of developing diabetes than white women who had the same BMI. Other studies have also noted that Asians have higher risks of cardiovascular disease at the same BMIs as white Europeans. One possible explanation is differences in fat mass. A 2002 study found that, when compared to white Europeans of the same BMI, Asian people have 3 percent to 5 percent higher body fat. By contrast, a 2007 study suggested that Black women have higher lean mass than white women. The AMA report also noted that BMI doesn't account for other factors that can significantly affect health, such as family history of disease, genetic factors, smoking, excessive alcohol use, mental health, stability of body weight, medication-induced obesity, and a person's occupation. The simple BMI-based cutoffs for underweight can miss eating disorders, and the AMA notes that an "overemphasis on bodily thinness is as deleterious to one's physical and mental health as obesity."

Alternatives

To move away from the harms and limitations of BMI, the AMA calls for researchers to become familiar with alternative ways of assessing fat and research these as potential new standards. These include: abdominal circumference measurements; waist-to-hip ratio; skinfold measurements; an empirical mathematical model called the Visceral Adiposity Index; a linear equation for the relative fat mass, which is based on a height-to-waist ratio; underwater weighting to look at body density (hydrostatic weighting); and air displacement plethysmography (aka Bod Pod). There's also bioelectrical impedance, which models the body as five cylinders of mass (the trunk and four limbs) and estimates body fat mass. Last, there's a variety of imaging methods including dual-energy X-ray absorptiometry (DXA), computed tomography (CT) scans, and magnetic resonance imaging (MRI). The combination of CT and MRI scans is currently considered the gold standard for body composition analysis, the AMA notes. In all, the AMA's new policy acknowledges "the significant limitations associated with the widespread use of BMI in clinical settings and suggests its use be in conjunction with other valid measures of risk."