Seroprevalence and associated risk factors of West Nile virus in selected equine populations in South Africa

Abstract

The family Flaviviridae consists of 94 species, which are distributed worldwide. Viruses in this family share aetiological properties such as membrane envelope proteins, resulting in diagnostic cross-­‐reactivity. Flaviviruses are divided into two clades; arthropod-­‐borne and non-­‐arthropod borne viruses. The most predominant flavivirus seen in horses is West Nile virus (WNV), belonging to the arthropod-­‐borne clade. West Nile virus has been divided into two lineages, of which, the second lineage is primarily found in South Africa. West Nile virus is transmitted through vector species, of which mosquito species are most predominant. Two genera of mosquitoes in particular, Aedes and Culex are widely distributed and are fundamental in the transmission of the virus to both reservoir hosts and incidental hosts. Other vectors included in the transmission he family Flaviviridae consists of 94 species, which are distributed worldwide. Viruses in this family share aetiological properties such as membrane envelope proteins, resulting in diagnostic cross-­‐reactivity. Flaviviruses are divided into two clades; arthropod-­‐borne and non-­‐arthropod borne viruses. The most predominant flavivirus seen in horses is West Nile virus (WNV), belonging to the arthropod-­‐borne clade. West Nile virus has been divided into two lineages, of which, the second lineage is primarily found in South Africa. West Nile virus is transmitted through vector species, of which mosquito species are most predominant. Two genera of mosquitoes in particular, Aedes and Culex are widely distributed and are fundamental in the transmission of the virus to both reservoir hosts and incidental hosts. Other vectors included in the transmission Blood samples were initially tested using the serum neutralization test (SNT). Documented positive and negative samples were then subjected to a capture IgG sandwich enzyme-­‐linked immunosorbant assay (ELISA). The results of the two assays were compared with one another, proving to correlate efficiently, giving the resultant seroprevalence percentages for each province, and then were used in comparison with the outcome of the questionnaires to determine significant associated risk factors. Results were analyzed with both univariable and multivariable analyses, taking clustering into consideration, to determine both apparent and prevalence estimate for each province and significance of seropositivity with associated risk factors. A small population of mosquitoes was collected in both Gauteng and Mpumalanga Provinces and were identified and separated into species. A nested SYBR green real-­‐time PCR assay was conducted on the pooled species of mosquitoes. All species presented negative for the presence of WNV, which could be a result of a low number of mosquitoes or a low prevalence of WNV in each species. The species identified included: Culex spp., Aedes spp. and Anopheles spp. The SNT were used to determine the apparent seroprevalence of WNV in the collected serum samples, thereafter, the prevalence estimate was calculated with a 95% confidence interval, taking clustering into consideration, for each province. The Free State Province had a high seroprevalence of 73% (95% CI 64-­‐81%), the Western Cape Province had a seroprevalence of 65% (95% CI 51-­‐79%) and Gauteng Province had a seroprevalence of 61% (95% CI 61-­‐62%). Limpopo Province had a seroprevalence of 60% (95% CI 45-­‐74%), followed by Northern Cape Province with 57% (95% CI 48-­‐66%), KwaZulu-­‐Natal Province with 54% (95% CI 43-­‐65%) and Mpumalanga Province with 56% (95% CI 40-­‐73%). The North West and Eastern Cape Provinces had lower seroprevalences of 43% (95% CI 34-­‐52%) and 48% (95% CI 43-­‐54%) respectively. Overall, the apparent seroprevalence for South Africa was 59% (95% CI 54-­‐64) using the SNT. The ELISA assays showed similar results to the SNT, with a 61% (95% CI 44-­‐79%) seroprevalence of WNV for South Africa. Gauteng Province had a seroprevalence of 47% (95% CI 44-­‐79%), KwaZulu-­‐Natal Province had a seroprevalence of 24% (95% CI 44-­‐79%), Northern Cape Province had 78% (95% CI 44-­‐79%) seroposivitiy, Eastern Cape Province had a seroprevalence of 68% (95% CI 44-­‐79%), North West Province with 59% (95% CI 44-­‐79%) and Mpumalanga Province had a seropositivity of 79% (95% CI 44-­‐79%). The results obtained using the ELISA had a moderate agreement with the SNT results (Kappa = 0.5).The univariable analysis showed association of WNV seropositive horses with; various agricultural activities, contact with different animal species, presence of annual frost, assorted water sources, occurrence of standing water pools and presence of Culicoides midges. These variables were subjected to multivariab analysis. The variables that indicated a p-­‐value of less than 0.05 were considered significant. Among these values were agricultural activities, such as livestock in the Free State Province, forestry in Mpumalanga Province and vineyards in the Western Cape Province. Contact with small ruminants and other species were the only significant species associated with WNV seropositive horses. Both standing pools of water and river sources were associated with seroprevalence in different provinces. Lastly, annual frost was only associated with seroprevalence in the Limpopo Province. Of the medical history and symptoms, fever was the singular variable associated with seropositivity. It is evident that many positive cases of infected horses are either not being reported or are not presenting with substantial clinical signs. The horses included in this study were from various age groups, different sexes and breeds and participated in various disciplines. Racehorses were excluded from the study due to their movement throughout the country making them bad sentinels for the study. The high seroprevalence of WNV in horse populations, determined in this study, indicates a subsequent high exposure rate throughout South Africa, varying amongst provinces. The risk factors associated with seroprevalence were all area specific, indicating the importance of habitats and the role it plays in transmission due to the presence of potential vectors. This study also noted a lack of knowledge about the transmission and prevalence of WNV amongst horse owners in South Africa.