A new study builds on evidence that body mass index (BMI) alone does not fully capture the risk of death, particularly for a US population that is increasingly obese and racially diverse.
Associations between BMI and mortality have been inconsistent in the past, with some studies showing an elevated risk and others finding a lower risk for overweight adults, the study authors wrote in PLOS ONE . And most studies to date have used data from the 1960s to 1990s that predominantly include non-Hispanic white men and women.
To offer additional insight, Aayush Visaria, MD, MPH, analyzed recent health survey data from 1999 to 2018 to extrapolate BMI and found that although obesity was associated with an increased risk of mortality, being at a healthy weight or being overweight It had almost no effect on the person’s risk of death.
“Especially in the overweight range, BMI does not differentiate the risk of mortality , so I think it is important to include other measures of adiposity to try to better stratify people’s risk,” Visaria, of the Institute of Rutgers Health in New Jersey. He co-authored the study with Soko Setoguchi, MD, DrPH, of Rutgers Robert Wood Johnson Medical School.
BMI’s role as an independent metric for obesity gained new attention in June when the American Medical Association urged doctors not to use it solely to assess healthy weight.
The new analysis only examined all-cause mortality, not the association between BMI and the risk of cardiovascular disease or diabetes. These diseases, along with hypertension, are the leading causes of premature death in the US, and it is well established that high BMI contributes to these conditions, the authors noted in their study.
Analyzing the numbers
Visaria’s interest in helping people at risk for cardiometabolic diseases began 9 years ago as an undergraduate student at Rutgers University. She founded an organization, now called the American Preventive Screening and Education Association, that has trained 1,300 undergraduate, graduate and medical students to provide free blood pressure and diabetes testing throughout New Jersey. In addition to earning degrees in public health and medicine at Rutgers, Visaria completed a postdoctoral fellowship, during which she spent time working on the current study.
He and Setoguchi retrospectively analyzed health data from approximately 554,000 U.S. residents, including nearly 200,000 Asian, non-Hispanic black, Hispanic, multiracial, and Native American adults. The majority of participants, 69%, were non-Hispanic white adults. Participants were, on average, 46 years old, with equal numbers of men and women. The data came from the National Health Interview Survey, which asks U.S. households a variety of questions, including self-reported height and weight, which the authors used to calculate BMI. BMI is calculated by dividing weight in kilograms by height in meters squared.
During a median follow-up of 9 years , 75,807 people died. The risk of death remained about the same for adults whose BMI ranged from 22.5 to 29.9 . But the risk of mortality increased significantly for adults with a BMI of 30 or more , considered obese or severely obese, and adults with a BMI less than 18.5 , considered underweight.
Examining the risk by age group, Visaria said he was surprised to find that among people 65 and older, the risk of mortality was similar for those who were a healthy weight, overweight or obese with a BMI up to 34.9 . Among younger adults, the risk of death increased significantly for those with a BMI greater than 27.5.
“It points out that BMI is not everything, since some groups can have good survival with higher BMIs. Additionally, very low BMIs , especially in underweight people, are generally associated with high mortality, and very high BMIs above 35 are as well,” Carl J. Lavie, MD, medical director of cardiac prevention and rehabilitation at John Ochsner Heart and Vascular Institute, he wrote in an email to JAMA. He was not involved in the PLOS ONE study.
Francisco Lopez-Jimenez, MD, MBA, chair of the division of preventive cardiology and director of the cardiometabolic program at the Mayo Clinic, noted in an interview with JAMA that overweight Hispanic adults had a higher risk of death than non-Hispanic whites and black adults who were also overweight. López-Jiménez was not involved in the study.
“For Hispanics, there is some risk [of death] that starts to increase when the BMI is 27 or 28, something that is not seen in the entire cohort or in whites in particular,” López-Jiménez said. One possible reason for this, she said, is that Hispanics have a tendency to develop central obesity , which is associated with a higher risk of mortality from all causes.
A limitation of the study, Jiménez-López and Lavie agreed, is its reliance on self-reported weight and height measurements to calculate BMI. Research has shown that people tend to incorrectly report their own height and weight in national surveys.
Other analyzes have also found that people who fall into the overweight BMI category may not have an increased risk of death from all causes. A 2013 meta-analysis in JAMA of nearly 100 studies that included measured and self-reported weight and height found that overweight BMI was associated with lower all-cause mortality.
However, several large studies, including analyzes published in 2010 and 2016, have found the opposite. In those studies, people with overweight BMI had a higher risk of death from all causes.
Lavie said he would like to see more studies in large populations with other real-world measures, such as waist circumference, muscle strength and cardiorespiratory fitness . Measured physical fitness is perhaps one of the strongest predictors of the risk of death, particularly from cardiovascular disease, she added.
John A. Batsis, MD, wrote in an email to JAMA that one important element not considered in the study is the changes in body composition that occur with aging. Batsis is an associate professor of medicine at the University of North Carolina at Chapel Hill and was not involved in the study.
“BMI does not capture differences in age-related loss of muscle mass, strength or function (called sarcopenia ) and does not differentiate between people with central adiposity or adipose deposition in other tissues ,” he said. “This type of deposit is probably to blame for the adverse effects of obesity.”
Think Phenotype, Not Just BMI
“For me, as a cardiologist, this study underscores the importance of using alternative measures of fat such as waist circumference or waist-to-hip ratio,” in addition to BMI, López-Jiménez said. These other measures are relatively easy and inexpensive to take during patient exams, she added.
Waist circumference is strongly associated with cardiovascular and all-cause mortality, with and without adjustment for BMI, according to a 2020 consensus statement from the International Atherosclerosis Society and the International Chair Working Group on Risk Cardiometabolic on Visceral Obesity.
Batsis added that in addition to simple in-office measurements such as waist circumference or waist-to-hip ratio, doctors can use specific scales that measure body fat or fat-free mass and muscle strength.
“We really need to think about people’s ’phenotypes’ ; For example, we know that older adults with obesity and sarcopenia are at risk for greater adverse outcomes than either alone. "So not all ’obesity’ is the same and it may be that specific individuals with specific characteristics are treated differently."
The extra kilos are not the only problem either. Lopez-Jimenez and Lavie noted that the risk of death increased among adults in the low weight category . It’s a difficult finding to communicate amid the current obesity epidemic, López-Jiménez said, but doctors and patients should consider it when talking about weight. A low BMI can indicate underlying conditions such as cancer, depression and other illnesses.
“We need to move away from BMI or use it in conjunction with other measures; it’s not the gold standard,” Batsis said.
