Risk Factors for Death among Children Less than 5 Years Old Hospitalized with Diarrhea in Rural Western Kenya, 2005–2007: A Cohort Study

BACKGROUND Diarrhea is a leading cause of childhood morbidity and mortality in sub-Saharan Africa. Data on risk factors for mortality are limited. We conducted hospital-based surveillance to characterize the etiology of diarrhea and identify risk factors for death among children hospitalized with diarrhea in rural western Kenya. METHODS AND FINDINGS We enrolled all children <5 years old, hospitalized with diarrhea (≥3 loose stools in 24 hours) at two district hospitals in Nyanza Province, western Kenya. Clinical and demographic information was collected. Stool specimens were tested for bacterial and viral pathogens. Bivariate and multivariable logistic regression analyses were carried out to identify risk factors for death. From May 23, 2005 to May 22, 2007, 1,146 children <5 years old were enrolled; 107 (9%) children died during hospitalization. Nontyphoidal Salmonella were identified in 10% (118), Campylobacter in 5% (57), and Shigella in 4% (42) of 1,137 stool samples; rotavirus was detected in 19% (196) of 1,021 stool samples. Among stools from children who died, nontyphoidal Salmonella were detected in 22%, Shigella in 11%, rotavirus in 9%, Campylobacter in 5%, and S. Typhi in <1%. In multivariable analysis, infants who died were more likely to have nontyphoidal Salmonella (adjusted odds ratio [aOR] = 6·8; 95% CI 3·1-14·9), and children <5 years to have Shigella (aOR = 5·5; 95% CI 2·2-14·0) identified than children who survived. Children who died were less likely to be infected with rotavirus (OR = 0·4; 95% CI 0·2-0·8). Further risk factors for death included being malnourished (aOR = 4·2; 95% CI 2·1-8·7); having oral thrush on physical exam (aOR = 2·3; 95% CI 1·4-3·8); having previously sought care at a hospital for the illness (aOR = 2·2; 95% CI 1·2-3·8); and being dehydrated as diagnosed at discharge/death (aOR = 2·5; 95% CI 1·5-4·1). A clinical diagnosis of malaria, and malaria parasites seen on blood smear, were not associated with increased risk of death. This study only captured in-hospital childhood deaths, and likely missed a substantial number of additional deaths that occurred at home. CONCLUSION Nontyphoidal Salmonella and Shigella are associated with mortality among rural Kenyan children with diarrhea who access a hospital. Improved prevention and treatment of diarrheal disease is necessary. Enhanced surveillance and simplified laboratory diagnostics in Africa may assist clinicians in appropriately treating potentially fatal diarrheal illness.