Across 127 low- and middle-income countries, the dominant dietary staple — not merely caloric availability — independently predicts child wasting and stunting prevalence. Sorghum-dominant food supplies showed the strongest positive association with wasting (β=0.361, p<0.001), followed by rice (β=0.086, p=0.004), while maize was negatively associated (β=−0.139, p=0.004). Cassava showed a borderline positive association with stunting (β=0.975, p=0.092). Secondary analyses across 46 nationally representative surveys revealed that wasting phenotype also differed by staple: MUAC-predominant wasting clustered in cassava and maize regions, while WHZ-predominant wasting dominated sorghum and millet regions.
This ecological analysis is notable for operationalizing protein quality via DIAAS scores and zinc bioavailability across cereal classes — variables largely absent from prior cross-national undernutrition modeling since Frongillo et al. (1997). The finding that staple type predicts distinct wasting phenotypes challenges the one-size-fits-all anthropometric surveillance approach currently deployed in DHS and MICS surveys, which capture height and weight but not MUAC or individual dietary data. If confirmed, this would argue for staple-stratified nutrition interventions rather than uniform supplementation protocols. Critical limitations apply: this is an ecological design incapable of establishing individual-level causation, confounding by poverty, sanitation, and healthcare access is substantial, and regression coefficients reflect population averages. As a preprint not yet peer-reviewed, these findings require independent validation before informing policy. Still, the mechanistic framing around protein quality represents a genuinely novel analytical lens for global undernutrition research.
