Determinants of linear growth faltering among children with moderate-to-severe diarrhea in the Global Enteric Multicenter Study

Abstract

BackgroundModerate-to-severe diarrhea (MSD) in the first 2\u2009years of life can impair linear growth. We sought to determine risk factors for linear growth faltering and to build a clinical prediction tool to identify children most likely to experience growth faltering following an episode of MSD.MethodsUsing data from the Global Enteric Multicenter Study of children 0–23\u2009months old presenting with MSD in Africa and Asia, we performed log-binomial regression to determine clinical and sociodemographic factors associated with severe linear growth faltering (loss of ≥\u20090.5 length-for-age z-score [LAZ]). Linear regression was used to estimate associations with ΔLAZ. A clinical prediction tool was developed using backward elimination of potential variables, and Akaike Information Criterion to select the best fit model.ResultsOf the 5902 included children, mean age was 10\u2009months and 43.2% were female. Over the 50–90-day follow-up period, 24.2% of children had severe linear growth faltering and the mean ΔLAZ over follow-up was −\u20090.17 (standard deviation [SD] 0.54). After adjustment for age, baseline LAZ, and site, several factors were associated with decline in LAZ: young age, acute malnutrition, hospitalization at presentation, non-dysenteric diarrhea, unimproved sanitation, lower wealth, fever, co-morbidity, or an IMCI danger sign. Compared to children 12–23\u2009months old, those 0–6\u2009months were more likely to experience severe linear growth faltering (adjusted prevalence ratio [aPR] 1.97 [95% CI 1.70, 2.28]), as were children 6–12\u2009months of age (aPR 1.72 [95% CI 1.51, 1.95]). A prediction model that included age, wasting, stunting, presentation with fever, and presentation with an IMCI danger sign had an area under the ROC (AUC) of 0.67 (95% CI 0.64, 0.69). Risk scores ranged from 0 to 37, and a cut-off of 21 maximized sensitivity (60.7%) and specificity (63.5%).ConclusionYounger age, acute malnutrition, MSD severity, and sociodemographic factors were associated with short-term linear growth deterioration following MSD. Data routinely obtained at MSD may be useful to predict children at risk for growth deterioration who would benefit from interventions.