Extended BMI-for-Age Growth Charts: Putting Them to Use

Imagine this scenario: A primary care pediatrician has been following a 10-year-old female with an initial body mass index (BMI)-for-age of 36 with monthly visits for the past 6 months. The patient and family have worked hard to make and sustain healthy lifestyle changes, and want to see how her growth has changed. The pediatrician wants to view her weight change in the context of her height change and BMI norms for her age. The pediatrician pulls up the 2000 Centers for Disease Control and Prevention (CDC) growth charts in the electronic health record and has no easily explainable data to share with the family.

The brief report by Ogden et al¹  in this issue of Pediatrics provides welcomed insight for pediatric health care providers caring for the nearly 6% of US youth with severe obesity.²  Before the recent release of the CDC extended BMI-for-age growth charts, Gulati and colleagues designed growth charts depicting mathematically derived cut points of 120% and 140% above the 95th percentile to aid in clinical management of these youth.³  Ogden and colleagues report the differences between these 2 clinical tools when applied to data from the 2017–March 2020 National Health and Nutrition and Examination Survey.

The prevalence of comorbidities rises with excess adiposity and obesity severity,⁴  highlighting the need for accurate assessment and ongoing monitoring and communication of weight status with families and colleagues. Before the release of the Gulati et al growth charts, pediatric health care providers lacked a visual way of tracking and sharing treatment progress for youth with severe obesity (defined as ≥120% of the 95th BMI percentile). The 2000 CDC growth charts were based on data from an era when child obesity prevalence was much lower, and percentiles above the 97th were not depicted. In fact, the CDC had cautioned against using estimates of percentiles >97th because the underlying data set contained too few children with high BMIs to determine 98th or 99th percentile.⁵  The 120% and 140% above the 95th percentiles shown on the Gulati et al growth charts align in late adolescence with class II and III severe obesity in adults, and changes in growth in youth with severe obesity are more easily visualized.

Below, we describe contributors to the potential clinical utility of the new CDC extended BMI-for-age growth charts and possible future research and clinical directions. Benefits (and limitations) of the new extended BMI growth charts include the following:

1. Unity of units: When providers care for children with obesity, the new extended BMI charts keep BMI percentile the same unit of measure across the range of BMI values for each child, a benefit from the convolutions required to switch from BMI percentile to percentage of the 95th percentile when BMI moves above the 97th percentile.

2. A new severity cut point: Although the authors do not explicitly make this conclusion, this study suggests that the new 98th percentile could replace the current definition of severe obesity, at least for children aged 6 years and above. Although the percentage of children with BMI ≥98th percentile is somewhat higher than percentage with BMI ≥120% of the 95th percentile overall and in all subcategories, the difference reaches statistical significance only among children aged 2 to 5 years.

Conceptually, the label severe should reflect substantial health risk in the population, and a useful definition should be applicable across ages, sex, and other categories. Therefore, a definition tied to age and sex, such as percentile, is valuable. However, we understand that a cut point is imperfect; health risks vary within a cohort with the same BMI status, and health risks increase continuously, if not linearly, as BMI increases, but not in quantum levels. Cut point assignment is artificial, whether on the basis of percentage above the 95th percentile curve or on the basis of the somewhat round number of 98th percentile (or, in adults, on the basis of similarly convenient numbers of 35 or 40). Further studies might help us understand the ways risks change as each assigned threshold is crossed and across different populations, but ultimately, 98th percentile may be close enough to 120% of the 95th, which we know delineates different obesity-related comorbidity prevalence. In practice, use of 98th percentile as the severity cut point is interpretable and explainable.

• 3. Improved interpretation of severity: The darker shades of the higher categories on the hard copy/PDF versions can help providers communicate weight status over time. Parents understand green, yellow, and red shading for healthy weight, overweight, and obesity,⁶  and parents will likely find the increasingly dark shades on the new charts helpful in understanding severity change, although studies to confirm would be welcome. Assuming shading improves interpretation, at least qualitatively, we hope that, as electronic health records incorporate the new charts, the display showing an individual’s BMI trajectory will add shading as an aid to interpretation.

However, the challenge remains of explaining to patients any measures of weight relative to height, age, and sex, and BMI percentile is only slightly easier to explain than percentage of the 95th percentile. Interpretation is further complicated by the fact that, despite more current data underlying the extended curves, far more than 2% of children have BMIs above the 98th percentile cut point.

• 4. Poor utility in quantifying change, whether improvement or worsening over time: These new charts do not solve the problem of the compression of percentiles at higher BMI percentile levels, a limitation which drove the creation of the percentage of the 95th percentile metric. A change in BMI percentile is not uniform across different obesity severity levels. As an example, change in BMI percentile in the highest severity level, from 99.99 to 99.9 (−0.1), reflects substantial weight and BMI improvement. This change in a 10-year-old girl describes a decrease in BMI from 36 to 33, about a 15 lb weight loss. However, a change in BMI percentile from 98.0 to 97.9 (also −0.1) describes a BMI decrease from 26.5 to 26.3, about a 0.5 lb weight loss. Although the visual displays help communicate the change, families do ask for quantification. This limitation affects outcomes reports of studies and clinical programs, which typically include children with a range of ages and of obesity severity. The CDC has provided z score charts (and formulas) to help researchers, but the lay public and many clinicians are not knowledgeable about the concept of SD scores, and the challenge of disseminating outcomes to nonscientists remains.

Now that the new extended BMI-for-age growth charts have been released, there is more work to do. Clinicians and researchers can evaluate the health risks associated with different BMI severity levels and the impact of change (improvement or worsening) on health markers. Also important will be evaluation of the meaning of these growth charts for patients, parents, and clinicians, and their utility in communication and decision-making.

CDC

Centers for Disease Control and Prevention