Maternal Height-standardized Prevalence of Stunting in 67 Low- and Middle-income Countries

Abstract

Background: Prevalence of stunting is frequently used as a marker of population-level child undernutrition. Parental height varies widely in low- and middle-income countries (LMIC) and is also a major determinant of stunting. While stunting is a useful measure of child health, with multiple causal components, removing the component attributable to parental height may in some cases be helpful to identify shortcoming in current environments. Methods: We estimated maternal height-standardized prevalence of stunting (SPS) in 67 LMICs and parental height-SPS in 20 LMICs and compared with crude prevalence of stunting (CPS) using data on 575,767 children under-five from 67 Demographic and Health Surveys (DHS). We supplemented the DHS with population-level measures of other child health outcomes from the World Health Organization's (WHO) Global Health Observatory and the United Nations' Inter-Agency Group for Child Mortality Estimation. Prevalence of stunting was defined as percentage of children with height-for-age falling below -2 z-scores from the median of the 2006 WHO growth standard. Results: The average CPS across countries was 27.8% (95% confidence interval [CI], 27.5-28.1%) and the average SPS was 23.3% (95% CI, 23.0-23.6%). The rank of countries according to SPS differed substantially from the rank according to CPS. Guatemala, Bangladesh, and Nepal had the biggest improvement in ranking according to SPS compared to CPS, while Gambia, Mali, and Senegal had the biggest decline in ranking. Guatemala had the largest difference between CPS and SPS with a CPS of 45.2 (95% CI, 43.7-46.9%) and SPS of 14.1 (95% CI, 12.6-15.8%). Senegal had the largest increase in the prevalence after standardizing maternal height, with a CPS of 28.0% (95% CI, 25.8-30.2%) and SPS of 31.6% (95% CI, 29.5-33.8%). SPS correlated better than CPS with other population-level measures of child health. Conclusion: Our study suggests that CPS is sensitive to adjustment for maternal height. Maternal height, while a strong predictor of child stunting, is not amenable to policy interventions. We showed the plausibility of SPS in capturing current exposures to undernutrition and infections in children.