BACKGROUND

Few studies have had sufficient longitudinal data to track how different malnourished states relate to mortality at different ages and interrelate over time.

OBJECTIVES

This study aims to describe the RRs and proportions of mortality associated with wasting and stunting and the pathways into and out of these nutritional states.

METHODS

Longitudinal growth data sets collected for children ages 0-24 months from Malawi, South Africa, and Pakistan were combined (n = 5088). Children were classified as deceased, wasted (weight for height < -2 SD; 1-4%), stunted (length < -2SD; 20-47%), or wasted and stunted (WaSt; 2-5%) at ages 3, 6, 9, 12, 18, and 24 months. Mixed-effects Cox models were used to study the association between nutritional status and mortality.

RESULTS

By age 3 months, 20% of children were already stunted, rising to 49% by 24 months, while wasting (4.2% and 2.2% at 3 months, respectively) and WaSt (0.9% and 3.7% at 24 months, respectively) were less common. The HR for mortality in WaSt was 9.5 (95% CI, 5.9-15), but 60% of WaSt-associated mortality occurred at 3-6 months. Wasting or WaSt was associated with 10-23% of deaths beyond 6 months, but in the second year over half of deaths occurred in stunted, nonwasted children. Stunting persisted in 82% of children and wasting persisted in 44%. Wasted children were more likely than nonwasted, nonstunted children to become stunted (RR, 1.93; 95% CI, 1.7-2.2), but 94% of children who progressed to stunting had not been wasted in the prior period.

CONCLUSIONS

WaSt greatly increased the risk of death, particularly in very young infants, but more deaths overall were associated with stunting. Most stunting appeared to be either intrauterine in origin or arose in children without prior wasting. Either stunting and wasting represent alternative responses to restricted nutrition, or stunting also has other, nonnutritional causes.