The relationship between wasting and stunting: a retrospective cohort analysis of longitudinal data in Gambian children from 1976 to 2016

Abstract

Background: The etiologic relationship between wasting and stunting is poorly understood, largely because of a lack of high-quality longitudinal data from children at risk of undernutrition.

Objectives: The aim of this study was to describe the interrelationships between wasting and stunting in children aged <2 y.

Methods: This study involved a retrospective cohort analysis, based on growth-monitoring records spanning 4 decades from clinics in rural Gambia. Anthropometric data collected at scheduled infant welfare clinics were converted to z scores, comprising 64,342 observations on 5160 subjects (median: 12 observations per individual). Children were defined as "wasted" if they had a weight-for-length z score <-2 against the WHO reference and "stunted" if they had a length-for-age z score <-2.

Results: Levels of wasting and stunting were high in this population, peaking at approximately (girls-boys) 12-18% at 10-12 months (wasted) and 37-39% at 24 mo of age (stunted). Infants born at the start of the annual wet season (July-October) showed early growth faltering in weight-for-length z score, putting them at increased risk of subsequent stunting. Using time-lagged observations, being wasted was predictive of stunting (OR: 3.2; 95% CI: 2.7, 3.9), even after accounting for current stunting. Boys were more likely to be wasted, stunted, and concurrently wasted and stunted than girls, as well as being more susceptible to seasonally driven growth deficits.

Conclusions: We provide evidence that stunting is in part a biological response to previous episodes of being wasted. This finding suggests that stunting may represent a deleterious form of adaptation to more overt undernutrition (wasting). This is important from a policy perspective as it suggests we are failing to recognize the importance of wasting simply because it tends to be more acute and treatable. These data suggest that stunted children are not just short children but are children who earlier were more seriously malnourished and who are survivors of a composite process.

Keywords: boys; concurrently wasted and stunted; girls; growth-monitoring data; seasonality; stunting; wasting.

FIGURE 1

Flow diagram of sample selection from data collected at scheduled infant welfare visits at the MRC Keneba clinic. A distinction is made between excluded participants (left) and excluded visits (right); when the criteria on the right led to exclusion of all of a participant's visits, that participant was excluded because he/she had no remaining visits to include. MRC, Medical Research Center. ¹Extreme z scores excluded as per WHO criteria (see Statistical methods). ²Extreme z score change of >3 SDs between consecutive visits, outlying observation excluded.

FIGURE 2

Stunting, wasting, and concurrence by age; cross-sectional estimates expressed as proportions, smoothed by local regression with 95% CIs, n = 5160 participants, 64,342 visits. Narrow CIs are obscured by the mean line. The top panel shows overall proportions as in Table 1. The middle and lower panels group the observations by current stunting and wasting, respectively, illustrating that wasting is more common among those who are also stunted than among those who are not, and vice versa.

FIGURE 3

Predicted WLZ trajectories, grouped by stunting status at 20–24 mo, based on mixed-effects growth models, n = 5160 participants, 64,342 visits. The mean trajectory (purple) is plotted together with predicted WLZ trajectories for children born at different times of the year, separately among those children who were or were not stunted at 20–24 mo of age. The horizontal colored bars represent ages that coincide with the wet season (July–October), for children born at different times of year. DoB, date of birth; WLZ, weight-for-length z score.

FIGURE 4

Predicted probabilities of stunting and wasting, based on the models reported in Table 3, n = 3867 participants, 28,403 visits. Being stunted increases the probability of being wasted and/or remaining stunted 3 mo later (dashed lines compared with solid lines). Being wasted also increases the probability of being stunted and/or remaining wasted 3 mo later (blue lines compared with red lines). For example, a boy who is both stunted and wasted at 12 mo has a 69% chance of still being stunted and a 9.2% chance of still being wasted at 15 mo, whereas a girl who is neither stunted nor wasted at 18 mo has a 9.7% chance of becoming stunted and a 1.7% chance of becoming wasted by age 21 mo.