Prevalence and correlates of stunting at hospital discharge among children 1-59 months in Western Kenya

Abstract

Background: The risk of mortality among recently hospitalized children in sub-Saharan Africa is 6 to 8-fold higher than similarly aged children in the community. Stunting (length/height-for-age z-score [LAZ/HAZ] <-2 SD) is associated with significant morbidity and mortality among children under the age of 5 in sub-Saharan Africa. Children who are stunted and recently hospitalized for acute illness may be at elevated risk for post-discharge morbidity and mortality. Few studies have examined predictors of stunting at hospital discharge among recently hospitalized children, who represent a population at high risk of poor outcomes and may benefit from targeted intervention. Methods: We determined correlates of stunting among children 1-59 months at discharge from 4 hospitals in in the Nyanza Province of Western Kenya enrolled in the Toto Bora Trial (NCT02414399), using log-Poisson regression to estimate prevalence ratios (PRs), adjusted for child age and study site. Results: Of the 1400 children enrolled in the parent trial (median age 18 months (IQR: 9-32)), 23% were stunted. Age over one year was associated with a higher prevalence of stunting at discharge (aPR12-23m: 1.5 [95%CI: 1.1-2.1]; and 24-59 months aPR24-59m: 1.5 [95% CI: 1.1-2.1]). Lack of exclusive breastfeeding in the first 6 months of life was associated with higher stunting prevalence as compared to exclusive breastfeeding (aPRnever: 2.6 [95% CI: 1.4-4.7]; aPRpartial: 1.3 [95% CI: 1.0-1.7]). Children who were HIV-infected or HIV-exposed, uninfected were more likely to be stunted than HIV-unexposed children (aPRHIVinfect: 2.7 [95% CI: 1.4-5.0]); aPRHEU: 1.9 [95% CI: 1.3-2.6]). Children whose caregivers reported having the equivalent of a primary school education or less were 2 times more likely to be stunted relative to those whose caregivers reported completing secondary school (aPR: 2.0 [95% CI: 1.6-2.5]). Having unimproved sanitation (pit latrine without slab floor or open defecation) was associated with a higher prevalence of stunting relative to having a flush toilet (aPRpit latrine without slab floor: 2.0 [95% CI: 1.2-3.4]; aPRopen defecation: 3.5 [95% CI: 1.8-7.1]). Child sex, shared toilet facility, enteric pathogen exposure (Salmonella, Shigella, and Campylobacter, Giardia and Cryptosporidium), and number of previous hospitalizations in the last year were not associated with stunting at discharge in this population. Conclusions: HIV exposure, poor sanitation, and lack of exclusive breastfeeding were important risk factors of stunting. Children hospitalized with an acute infectious illness who are stunted at discharge are at high risk for adverse health outcomes and death. Hospital discharge may represent an important opportunity for administering targeted interventions to prevent further growth faltering and post-discharge mortality in this highly vulnerable population.