Risk factors for arboviral seropositivity in children in Teso South Sub County, Western Kenya

Abstract Arboviruses are responsible for epidemics and are emerging and re-emerging in sub-Saharan Africa. However, the risk factors for arboviral diseases are poorly described in Kenyan children. Knowledge of risk factors can facilitate earlier diagnosis and better treatment and implementation of effective prevention in children. This study determined risk factors for seropositivity to Yellow fever (YFV), Dengue (DENV), Chikungunya (CHIKV) and West Nile (WNV) viruses among children at two facilities in Teso Sub-County in Western Kenya. In a hospital-based cross-sectional survey, the risk factors for seropositivity to the arboviruses were assessed. Eligible children aged 1 to 12 (n = 656) who visited Alupe Sub County Hospital and KEMRI Alupe Clinic in Teso Sub County were recruited. Socio-demographic, environmental, behavioural and medical information was collected using a questionnaire. Blood drawn from these children was screened for antibodies to YFV, DENV, CHIKV and WNV using Indirect Enzyme-Linked Immunosorbent Assays. Descriptive statistics were used to summarise seroprevalence, socio-demographic, clinical and environmental variables. Binomial logistic regression described the relationship between the risk factors and arbovirus seropositivity. Seropositivity to at least one arbovirus was found in 27.7%, with 15.7% being positive for DENV, 9.6% for WNV, 5.6% for CHIKV and 4.4% for YFV. The factors that significantly increased the risk to at least one of the arboviruses were: age 6-9 years (by 18%, p=0.006) compared to those 1-3 years, school attendance (by 66%, p=0.000) compared to none, the primary caregiver being "Other" (by 17%, p=0.026) and not the parent, the use of Olyset (by 7%, p=0.039), or an unknown mosquito net (by 26%, p=0.020) compared to Permanet. The risk of yellow fever seropositivity was increased where vegetation was close to the house (by 5%, p=0.042) compared to where vegetation was far. The risk was decreased by the use of an unknown bed net (by 4%, p=0.046) compared to Permanet and having a past history of rash (by 6%, p=0.018). For Dengue Fever, females were at an increased risk (by 8%, p=0.002) compared to males and having water bodies near the house (7%, p=0.030). The risk of chikungunya was increased by school attendance (by 25%, p=0.021) compared to not, the use of mosquito repellents (by 10%, p=0.006) compared to no interventions and having had a rash in the past (by 6%, p=0.043). The risk was decreased by roofing with iron sheets (by 3%, p=0.048) compared to grass-thatching. WNV seropositivity risk was higher in those aged 3-6 years (by 8%, p=0.004) and 6-9 years (by 15%, p=0.004) than in those aged 1-3 years. It was increased in those attending school (by 37%, p=0.006) compared to those not, and those using Olyset (by 11%, p=0.000) or an unknown bed net (by 30%, p=0.001) compared to Permanet. The risk was lower by between 25% and 33% (p<0.003) in those in pre-school, in lower and upper primary compared to those not in school. These factors are amenable to interventions that can be implemented to prevent and reduce arbovirus infections in children in endemic areas in Kenya.