Changing Etiologies of Febrile Illness in Areas of Differing Malaria Transmission in Rural Kenya

Thesis (Ph.D.)--University of Washington, 2014"<bold>Background</bold>": As malaria continues to decline in many parts of sub-Saharan African (SSA) and malaria rapid diagnostic tests become increasingly used; a growing number of febrile children are now being diagnosed as not having malaria. In order for these children to receive appropriate treatments, information on common non-malaria causes of fever, and treatment practices for febrile children without malaria is critically required to inform judicious use of antimalarials and antibiotics. At sites of high and low malaria endemicity in western Kenya; Kisii and Homa Bay, we conducted a 2-year cross-sectional surveillance study of febrile illness among children aged 6 months to 15 years to determine the: 1) frequency and correlates of malaria overtreatment and, 2) prevalence, causative organisms, and predictors of bacteremia due to any pathogen and specific bacterial pathogens among children seeking for fever. "<bold>Methods</bold>": Sociodemographic, environmental and clinical data were collected, and children tested for malaria, HIV, and bacteremia. Correlates of malaria overtreatment, and bacteremia were evaluated using multivariate logistic regressions. "<bold>Results</bold>": Nearly 7% of the 685 children enrolled in Kisii, and 45.8% of 677 Homa Bay had laboratory-confirmed malaria; p<0.001. Malaria overtreatment was more common in Homa Bay (57.2%), a high malaria endemic area with entomological infection rates (EIR) ≥300 than in Kisii (7.0%), a low malaria endemic area with EIR <1.5. Predictors of overtreatment in Homa Bay included presence of ≥1 Integrated Management of Childhood Illness (IMCI) danger signs (aOR=8.5; 95% CI: 4.8-14.9), fever lasting ≥7 days (aOR=4.9; 95%CI: 1.9-12.9), and fever ≥390C (aOR=3.1; 95%CI: 1.6-6.0). In Kisii, only fever ≥390C predicted overtreatment (aOR=2.1; 95%CI: 1.0-4.5). Malaria endemicity influences the clinical management of febrile children and may result in missed opportunities to treat alternative causes of fever. There is need to strengthen adherence to treatment guidelines to improve management of febrile children, to reduce risk of missed treatment opportunities for non-malaria fevers, particularly in malaria endemic areas. The prevalence of bacteremia was 3.3% (48/1478); including 3.1% (24/734) in Kisii and 3.4% (24/742) in Homa Bay. Salmonella spp. (19 typhi and 21 nontyphoidal salmonella [NTS]) accounted for 83% of isolates, and did not differ between study sites. Bacteremia prevalence in children with and without malaria was 1.9% and 3.8% respectively (p=0.05). Co-infection with bacteremia and malaria was uncommon,