ABSTRACT
Background: Overnourished under-five children are anthropometrically classified as either being at possible risk of overweight, overweight or obese and defined so, when either weight for height or body mass index for age (BMI-for-age) are >1SD to 2SD, >2SD to 3SD and >3SD, respectively of the analogous World Health Organization standards. Aim: To compare weight for height and BMI for age definitions for quantifying overnutrition burden. Methods: Theoretical consequences of ignoring age were evaluated by comparing, at varying height for age z-scores, the age- and sex-specific cutoffs of BMI that would define overnutrition with these two metrics. Overnutrition prevalence was then compared in simulated populations (short, intermediate and tall) and real-life datasets from India. Results: In short (-2SD) children, the BMI cutoffs with weight for height criteria were lower in comparison to BMI for age till 7-8 months, but higher at later ages. In National Family Health Survey-4, India dataset (short population), overnutrition (>1SD) prevalence with weight for height was higher from 0-0.5 years (exclusive breastfeeding age), but lower at subsequent ages. The prevalence difference (weight for height - BMI for age) in 0.5-5 years was -2.26% (6.57% vs 8.83%); this attenuated in 0-5 years (-1.55%; 7.23% vs 8.78%). The discrepancy was maximal for stunted children and was lower in girls. A similar pattern, of lower magnitude, was observed for overweight (>2SD) comparison. In intermediate and tall populations, there were no meaningful differences. Conclusion: The two definitions produce cutoffs, and hence estimates of overnutrition, that differ with the age, sex, and height of under-five children. The relative invariance, with age and height, of BMI for age, favors its use.
ABSTRACT
Background: Wasting and stunting commonly coexist, supposedly due to biological and social mechanisms. In under-five children, low-ponderosity is defined as <–2SD of WHO standards for either weight for height (wasted) or body mass index for age (thin) metrics. Unlike body mass index for age, weight for height ignores physiological changes in ponderosity with age, resulting in overestimation of wasting in comparison to thinness in under-5 populations with high stunting prevalence. This suggests a plausible statistical explanation for the wasting-stunting association. Aim: To test the null hypothesis that wasting-stunting (WaSt) and thinness-stunting (ThSt) associations are similar. Methods: Demographic Health Survey datasets (2010-2020) from South and South-East Asia (7 countries) and Sub-Saharan Africa (13 countries) were evaluated. WaSt and ThSt associations were estimated as odds ratio (OR) for individual datasets, which was pooled (random-effects meta-analysis). Stratified analyses were done for sex, age and region. Results: Young infants (0-6 months) comprised 8-14% of underfive children, with equal representation of boys and girls. Participants, especially Asians, were mostly shorter with lower ponderosity than WHO standards. WaSt prevalence was higher than ThSt in the 6-59 months age group, but lower in young infants. Pooled WaSt estimates were not significant: Asia (OR 0.95; 95% CI 0.75-1.14), Africa (1.17; 0.95-1.40), and combined (1.09; 0.93-1.24). In contrast, pooled ThSt associations were significantly negative: Asia (0.63; 0.50-0.76), Africa (0.82; 0.68-0.96), and combined (0.75; 0.65-0.85). In girls, these associations were attenuated for WaSt (0.96; 0.8-1.1), but enhanced for ThSt (0.6; 0.5-0.7). Conclusion: WaSt and ThSt associations are dissimilar. This suggests a primary statistical explanation for the reported wasting-stunting association, originating from ignoring physiological changes with age.