Study Challenges Accuracy of BMI Measures

 A new study published in the Journal of the American Medical Association Pediatrics shows that tri-ponderal mass index (TMI) estimates body fat more accurately than the traditional body mass index (BMI) in adolescents. In essence, the BMI calculations that have been relied on for decades may not be an accurate scale for assessing body fat in adolescents between the ages of eight and 17.

These new findings are timely as diagnosing, treating, and tracking the prevalence of children and adolescents with obesity is a high public health priority. Moreover, many school districts are sending home report cards labeling adolescents as overweight—a practice that has been controversial because children and adolescents tend to be more vulnerable to weight bias and fat shaming than adults.

“Treating obesity in adolescents requires an accurate diagnosis first,” said lead scientist Dr. Courtney Peterson, an assistant professor at the University of Alabama at Birmingham. “We found that TMI is both more accurate and easier-to-use than BMI percentiles. These new findings have the power to potentially change the way we diagnose obesity in children and adolescents aged eight through 17.”

BMI is the standard currently used worldwide to screen for obesity in both children and adults, despite prior evidence that it does not work as well in adolescents. To test BMI’s accuracy in adolescents, the researchers compared body composition data from 2,285 Caucasian individuals aged eight to 29 who participated in the 1999-2004 U.S. National Health and Nutrition Examination Survey (NHANES). Their findings challenge the accuracy of BMI (weight in kilograms divided by height in meters squared) in adolescents aged eight through 17 and show that the tri-ponderal mass index, or TMI (weight in kilograms divided by height in meters cubed), estimates body fat more accurately than BMI in adolescents.

Using three different calculations (stability with age, accuracy in estimating percent body fat, and accuracy in classifying adolescents as overweight vs normal weight), the researchers compared BMI to several different obesity indices. They found that TMI is the best overall body fat index to use in Caucasian adolescents between the ages of eight and 17 years old.

The researchers found that TMI better estimates body fat percentage, especially in male adolescents, for which the investigators found BMI to be particularly inaccurate. Their analysis also showed TMI to be a better index for diagnosing overweight adolescents than the current BMI percentiles. Using current BMI percentiles, the researchers noted that adolescents are incorrectly diagnosed as overweight 19.4% of the time, versus only an 8.4% incorrect overweight diagnosis rate for TMI. The data showed this is especially true for lean adolescents, a significant fraction of whom are incorrectly being diagnosed as overweight.

The researchers also used mathematical strategies to show how the relationship between body weight and height is much more complex in children and adolescents than it is in adults, particularly when adolescents are rapidly growing. The authors explained that for decades this complexity made it challenging to figure out the optimal body fat index for adolescents.

“Historically, BMI was used in children in part because it worked so well in adults,” said Dr. Steven Heymsfield, one of the team’s physician scientists at LSU’s Pennington Biomedical Research Center. Heymsfield added that to make BMI work in children, complicated BMI percentiles (called “Z scores”) were developed to diagnose overweight status and obesity based on BMI levels specific to a child’s age and gender. But the researchers found that using percentiles does not solve BMI’s accuracy problems. Peterson mentions that percentiles are problematic because they change over time and can become outdated. However, she emphasizes that even if BMI percentiles were updated to be as accurate as TMI for diagnosing adolescents as lean versus overweight, TMI still inherently estimates levels of body fat in adolescents more accurately than BMI does, while also eliminating the need for complicated percentiles.

“The percentiles make BMI difficult to use because they involve separate diagnostic numbers for every age for both boys and girls. TMI is an improvement because TMI involves only single numbers for boys and girls, respectively,” explained mathematician Dr. Diana Thomas, a professor at the U.S. Military Academy, who also worked on the study.

Obesity increases the risk of several chronic diseases in adults, including diabetes, cardiovascular disease, and cancer. Children who are overweight are at an even higher risk than adults for chronic disease. With nearly one in six children in the U.S. classified as overweight, several national and global initiatives are now underway to screen and diagnose children as a first step in curbing the obesity epidemic and the chronic diseases that follow.

The investigators agree that further research is needed to assess the effectiveness of TMI in broader audiences, including wider age ranges and ethnicities.

“We look forward to collaborating with other existing national and global health organizations to analyze additional data and current policies for diagnosing weight status among children and adolescents,” said Peterson. “Ultimately, we hope this research lays the foundation for improving health outcomes in the younger population.”

Researchers from LSU’s Pennington Biomedical Research Center, The University of Alabama at Birmingham, Montclair State University, The United States Military Academy, The Albert Einstein College of Medicine and Verona University Medical School collaborated on this research and publication.

Investigators and affiliations:

Courtney M. Peterson, PhD, University of Alabama at Birmingham, Birmingham, Alabama

Haiyan Su, PhD, Montclair State University, Montclair, New Jersey

Diana M. Thomas, PhD, United States Military Academy, West Point, New York

Moonseong Heo, PhD, Albert Einstein College of Medicine, Bronx, New York

Amir H. Golnabi, PhD, Montclair State University, Montclair, New Jersey

Angelo Pietrobelli, MD, Verona University Medical School, Verona, Italy

Steven B. Heymsfield, MD, LSU’s Pennington Biomedical Research Center, Baton Rouge, LA