Maternal Height and Risk of Child Mortality and Undernutrition

Maternal undernutrition, embodied by short stature and a low body mass index (BMI) and caused by chronic energy and micronutrient deficiencies, is highly prevalent in many developing countries. Short stature (<145 cm) affects more than 10% of women of reproductive age across south-central Asia, but only 1% to 2% in sub-Saharan Africa, whereas a low BMI (<18.5) is found among 20% or more women in both regions.¹ Both indicators can predict adverse pregnancy outcomes. Maternal height is a strong predictor of birth size, independent of pre-pregnancy BMI and weight gain during pregnancy. Short maternal stature is highly associated with uterine volume and blood flow and is associated with risks of fetal growth restriction, cesarean delivery,² and cephalo-pelvic disproportion, the risk of which is likely modified by newborn size.³

Of these 2 indices of undernutrition, attained adult height in women reflects a cumulative outcome measure of environmental exposures from fetal to adult life encompassing nutritional, infectious, sociocultural, and economic influences. The critical windows for linear growth include the fetal period, early childhood (0-24 months), and pubertal and adolescent periods, each of which influences attained adult height. The long-described theory of intergenerational cycle of growth failure posits that a small adult woman will deliver a small baby with fetal growth restriction, who will subsequently experience growth faltering as a child and low stature and body weight as an adolescent, ultimately leading to a small adult.⁴ However, few empirical data have directly linked maternal stature to undernutrition over a generational cycle.

Drawing on cross-sectional data from 109 Demographic Health Surveys (DHS) conducted between 1991 and 2008 in 54 developing countries, covering 2.6 million children aged 0 to 59 months, Özaltin and colleagues⁵ in this issue of JAMA show that maternal height is inversely associated with risks of child mortality, underweight, stunting, and wasting. With every 1-cm increase in height, the relative and absolute risks of each of these adverse outcomes was significantly decreased. Compared with the highest maternal height category of more than 160 cm, women with short stature (<145 cm) had an approximately 40% higher risk of any of their offspring dying, after adjusting for confounders. A similar analysis revealed risks of stunting and underweight in offspring to be 2-fold greater among short mothers, whereas that of wasting was only 17% higher.

These findings extend the authors' previous analysis of the same relationship from DHS data in India.⁶ Although causality cannot be inferred from cross-sectional data, the consistent, dose-response findings across countries and regions, and the stability of estimates when assessed in numerous sensitivity analyses, suggest that there exists an undoubtedly complex but true set of biological pathways for this relationship. The association of maternal stature with survival and nutritional outcomes of offspring may be mediated by the influence of maternal stature on birth weight, including intrauterine growth,⁷ and perhaps gestational age at birth.⁸ Low birth weight is associated with increased neonatal and infant mortality, as well as childhood wasting and stunting.⁹ That Özaltin et al⁵ found maternal stature to be more strongly associated with neonatal and infant mortality than with mortality beyond the first year of life also implicates fetal growth restriction or preterm birth as possible antecedent risk factors for mortality.

A few cohort studies have examined the relationship between maternal stature and childhood outcomes. In a study from Nepal, in which offspring of pregnant women enrolled in a micronutrient intervention trial were assessed at birth and prospectively followed up at 39 to 50 months, maternal height was associated with stunting (adjusted odds ratio [AOR], 0.90; 95% confidence interval [CI], 0.87-0.93) and underweight (AOR, 0.92; 95% CI, 0.89-0.93), but not wasting (AOR, 0.98; 95% CI, 0.94-1.03).¹⁰ In the study by Özaltin et al,⁵ the association between maternal height and childhood wasting was also weak, indicating that other concurrent and direct factors are more likely to influence the risk of acute undernutrition.

One of the indices assessed in this analysis, underweight (weight <−2 Z score below the international reference median for age) is determined both by weight and height, and therefore does not distinguish between wasting and stunting. A low weight-for-age after 3 years is largely due to stunting, except in acute food insecure settings in which wasting might largely explain low values of this index.¹¹ However, underweight is one of the indices being used to track the first of the United Nations Millennium Development Goals (MDGs),¹² which calls for halving the number of individuals experiencing hunger between 1990 and 2015.

Özaltin et al⁵ show high mortality rates among children younger than 5 years in the surveyed countries: 11.9% overall, with the highest rate of 24.7% in Niger. The authors also report high rates of stunting at 38.3% overall (55% in Ethiopia) and underweight at 21.5% overall (44% in Bangladesh). This enormous burden of child mortality and undernutrition suggests limited progress toward achieving the MDGs by 2015, although these estimates are averaged over the entire period between 1991 and 2008 and may mask progress, if any, in specific countries.

Few interventions are known to influence linear growth faltering. Although maternal calorie/protein or micronutrient supplementation during pregnancy has shown to be associated with an increase in mean birth weight and a reduction in low birth weight and small for gestational age birth, such supplementation is not known to affect birth length.^{13 - 14} However, birth length has not been assessed systematically in studies. Food supplementation has been associated with improved linear growth between 6 months and 24 to 36 months in randomized studies, but with little or no improvement beyond this age.¹¹ Interventions during the prepubescent or adolescent period or adoption of children from deprived into affluent environments appear to work only by modulating the onset of menarche to an earlier age.¹⁵ Consequently, there is a reduction in the duration of growth, which normally extends into the early 20s in many resource-poor settings, resulting in little to no overall gain in attained adult height. Also, early pregnancy can have an adverse influence on adolescent girls' nutritional well-being. For example, in Bangladesh where a large percentage of women are short,¹⁵ pregnancy during adolescence appeared to halt linear growth and was associated with lower BMI of young mothers compared with nonpregnant, age-matched controls.¹⁶ Delaying age of marriage and first pregnancy could have a benefit of increasing attained adult height of women.

Although the causes of childhood undernutrition are well described and exclusive breastfeeding, appropriate complementary feeding, hygiene and sanitation, and access to health care are known interventions likely to improve growth in young children, little attention has been given to the prenatal origins of childhood undernutrition. Specifically, maternal stature, overall maternal nutritional status, and the factors that give rise to them needs more emphasis, especially because the early growth failure is known to track into childhood and beyond. The speculation by Özaltin et al⁵ that the intergenerational transmission from the mother's environmental milieu during her own childhood (and also fetal life) to her offspring's growth and survival may be plausible in explaining the relationship between maternal stature and child mortality and undernutrition, although further investigation of this phenomenon in cohort studies is needed. Also, because short maternal stature is less common in sub-Saharan Africa, the risk of child mortality and undernutrition attributable to maternal height may be expected to be correspondingly lower in this region. This needs examination as does the independent and adjusted association of maternal BMI to childhood outcomes.

In conclusion, the findings of Özaltin et al⁵ call important attention to the vast challenge and potential public health value of addressing maternal undernutrition, and specifically maternal short stature, as a still-neglected approach to reducing child mortality and undernutrition in many low- and middle-income countries.