Assessing BMI change in Research

Although extended bmiz should work well in most analyses, the best metric for assessing BMI change is uncertain. Several studies (1–6) have shown that expressing BMI relative to the CDC 50th or 95th percentile is better than the LMS-based version of BMIz in assessing changes in adiposity. This is because of the curvilinear relation of BMI to BMIz, but would likely apply to both the original and extended (7) versions of BMIz. Further, a recent commentary (8) suggested that until consensus is reached, “investigators can consider consistently reporting more metrics (e.g., absolute and percentage of change in BMI, change in percentage of the 95th BMI percentile, change in percentage of the median).”

As of December 15, 2022, the SAS program includes 2 additional variables for assessing BMI change:
(1) bmip95: BMI expressed as a percentage of the 95th percentile of the CDC growth charts.
(2) bmip50: BMI expressed as a percentage of the 50th percentile (median) of the CDC growth charts. If desired, 100 can be subtracted from bmip50 so that it is the %distance from the median (9).

Note that some care should be taken in using these metrics as it is likely that their interpretations differ by age. Because the dispersion of BMI values is much greater at age 18 y than at age 2 y, it is possible that a bmip50 value of 150, for example, should be interpreted differently at these two ages. These metrics can be adjusted for differences by sex and age by specifying the ‘all=TRUE’ option in the R program.