Weight-for-height is associated with an overestimation of thinness burden in comparison to BMI-for-age in under-5 populations with high stunting prevalence

Abstract Background Thinness at <5 years of age, also known as wasting, is used to assess the nutritional status of populations for programmatic purposes. Thinness may be defined when either weight-for-height or body-mass-index-for-age (BMI-for-age) are below –2 SD of the respective World Health Organization standards. These definitions were compared for quantifying the burden of thinness. Methods Theoretical consequences of ignoring age were evaluated by comparing, at varying height-for-age z-scores, the age- and sex-specific cut-offs of BMI that would define thinness with these two metrics. Thinness prevalence was then compared in simulated populations (short, intermediate and tall) and real-life data sets from research and the National Family Health Survey-4 (NFHS-4) in India. Results In short (–2 SD) children, the BMI cut-offs with weight-for-height criteria were higher in comparison to BMI-for-age after 1 year of age but lower at earlier ages. In Indian research and NFHS-4 data sets (short populations), thinness prevalence with weight-for-height was lower from 0.5 to 1 years but higher at subsequent ages. The absolute difference (weight-for-height – BMI-for-age) for 0.5–5 years was 4.6% (15.9–11.3%) and 2.2% (19.2–17.0%), respectively; this attenuated in the 0–5 years age group. The discrepancy was higher in boys and maximal for stunted children, reducing with increasing stature. In simulated data sets from intermediate and tall populations, there were no meaningful differences. Conclusions The two definitions produce cut-offs, and hence estimates of thinness, that differ with the age, sex and height of children. The relative invariance, with age and stature, of the BMI-for-age thinness definition favours its use as the preferred index for programmatic purposes.