Sero-prevalence and spatial distribution of Rift Valley fever infection among agro-pastoral and pastoral communities during Interepidemic period in the Serengeti ecosystem, northern Tanzania

The study was conducted in the Serengeti ecosystem in northern regions of Tanzania. The area included three districts of Bunda, Ngorongoro and Serengeti in the Serengeti ecosystem, (Fig. 1). However, some human subjects were enrolled from Lamadi ward in Busega district. The area is located along the border with Kenya, close to the equator, between 2o to 4o S. minimum temperature ranges between 15-21oC, while maximum ranges between 24-27oC. The rainfall is highly seasonal with peaks in March to May and November to December. Mean annual rainfall in the Serengeti varies from 1050 mm in the northwest to 550 mm in the southeast. The area is covered with rich volcanic soil that supports the growth of vegetation. Meshwork of streams of water bodies are found throughout the area. The ecosystem extends to the shores of Lake Victoria in the west, (where Bunda and Busega districts are located), Lake Eyasi in the south and the Great Rift Valley to the east, covering Ngorongoro Conservation Area (NCA) and Loliondo Game Controlled Area (LGCA) in Ngorongoro district. The study also extended to the western side in Serengeti and Bunda districts (Fig. 1). Soil type, grassy plain, temperature and rainfall of the ecosystem favor the endemicity of the disease in the area. The communities surrounding the area are mainly of pastoral and agro-pastoral communities. Pastoral communities are those deriving their income solely by keeping animals like cattle, sheep, goats, donkey and dogs. Agro-pastoral communities are those depending on crop cultivation supplemented with animal keeping mostly in small scale.

A cross-sectional, health facility based study was conducted on humans in the Serengeti ecosystem during the dry season of August–October, 2014. The health facilities where the study were conducted included Wasso District Designated Hospital (DDH), Endulen hospital, Malambo, Piyaya, Arash and Olbabal dispensaries in Ngorongoro District; Bunda DDH in Bunda district and Nyerere DDH in Serengeti district respectively. As a hospital-based study, participants from Lamadi ward (Busega district) bordering Bunda district were also enrolled as they were frequently seeking medical services at Bunda DDH. We enrolled patients aged 5 years and above attending outpatient department (OPD) regardless of their clinical presentation. Patients who could not give assent or consent and those who were critically ill to give information were excluded from the study. A total of 751 human subjects were included in the study. Probability proportion to population size was used to allocate the calculated sample size to the three districts’ health facilities. The three districts of Ngorongoro, Serengeti and Bunda districts were allocated 210, 241 and 300 study respondents respectively. Based on geographical nature and sparse population distribution of Ngorongoro district, apart from district hospital, more health facilities were included in the study to allow wider coverage of all parts of the district as much as possible. The additional health facilities were Endulen hospital, as well as Malambo, Piyaya, Arash, and Olbalbal dispensaries. These are the main health facilities in the area that provide services to all population in the study area. Simple random sampling technique was used to select participants among out-patients who visited the respective health facility for medical services or individuals who escorted relatives for medical services in the eight health facilities in the three district i.e. six health facilities in Ngorongoro district, one in Bunda and one in Serengeti districts respectively. The coordinates of the study sites were recorded using a Global Positioning System (GPS) device for subsequent use in marking the sampling sites and indicate locations with sero-positive human subjects in the study area. In Tanzania, the mean straight-line distance to a health facility is about 4.2 km (SD 3.9 km) [15].

Approximately, 4 mls of whole human blood was drawn aseptically from each patient recruited in the study. Approximately 1–2 mls of sera was obtained from each blood sample through centrifugation, and transferred into screw capped cryovials and stored at -20 °C freezer awaiting serological analysis. Sera from patients were tested for the presence of anti-RVFV IgG antibody using competitive Enzyme linked Immuno-sorbent Assay (ID.Vet Innovative Diagnostics, Grabels, France). The test has excellent sensitivity and specificity of 100%. Detection of anti-RVFV IgM was done using capture ELISA kit, to all anti-RVFV IgG positive samples as described by the manufacturer (Biological Diagnostics Supply Limited -BDSL, UK). The sensitivity test was 100% and specificity ranged from 97.4 to 99.4% [16]. Validation of the results was done by repeating all positive and 10% of the negative samples and found 100% concordance between the first test and observed validation test.

Information on social demographic characteristics including age, sex, level of education, occupation and area of residence was collected. Median, range, and proportion of seropositive samples among those tested for anti-RVFV IgG (general exposure status) and anti-RVFV IgM (recent exposure status) were calculated. A number of socio-demographic factors associated with RVFV sero-positivity were assessed using prevalence Odds Ratio (OR) as a measure of association at 95% confidence interval. Socio-demographic factors with p-value ≤0.05 were considered as statistically significant in bivariate analysis. All variables with p ≤ 0.2 at bivariate level were entered into multivariate logistic regression model. Variables with p-value ≤0.05 at multivariate level were considered independently associated with RVF seropositivity.