The seroprevalence of African horse sickness virus, and risk factors to exposure, in domestic dogs in Tshwane, South Africa

Dogs are the only non-equid species to develop the fatal form of African horse sickness (AHS). Research conducted in 2013 questioned the long-held belief that naturally occurring cases of AHS in dogs were contracted exclusively through the ingestion of contaminated horse meat. Culicoides midges, the vector of AHS virus (AHSV) for horses, have an aversion to dog blood meals and dogs were believed to be dead-end or incidental hosts. More recently, dog mortalities have occurred in the absence of horse meat consumption and vector transmission has been suspected. The current study is a retrospective serological survey of AHSV exposure in dogs from an endemic area. Dog sera collected from dogs (n = 366) living in the city of Tshwane, Gauteng Province, South Africa, were randomly selected from a biobank at a veterinary teaching hospital, corresponding to the years 2014-2019. The study used a laboratory in-house indirect recombinant VP7 antigen-based enzyme-linked immunosorbent assay (iELISA) with a test cut-off calculated from AHSV exposure-free dog sera (n = 32). Study AHSV seroprevalence was 6 % (22/366) with an estimated true prevalence of 4.1 % (95 % confidence interval (CI) = 1.3-8.1 %). Incidence was estimated for dogs with multiple serological results with seroconversion occurring at a rate of 2.3 seroconversions per 10 dog years at risk (95 % CI = 0.6-6.2). A subsection of the study sera was tested with AHSV viral neutralisation test (VN) (n = 42) for serotype determination. Antibodies to AHSV serotype 6 were most prevalent (90 %) in VN seropositive dogs (n = 20) with most dogs seemingly subclinically infected (>95 %). Seroprevalence descriptively varied by year and identified risk factors were annual rainfall > 754 mm (odds ratio (OR) = 5.76; 95 % CI = 2.22 - 14.95; p < 0.001), medium human population densities, 783-1663 people/km2 (OR = 7.14; 95 % CI = 1.39 - 36.73; p = 0.019) and 1664-2029 people/km2 (OR = 6.74; 95 % CI = 1.40 - 32.56; p = 0.018), and the month of March (OR = 5.12; 95 % CI = 1.41 - 18.61; p = 0.013). All identified risk factors were consistent with midge-borne transmission to dogs. The relatively high seroprevalence and seroconversion rates suggest frequent exposure of dogs to AHSV and indicates the need to investigate the role dogs might play in the overall epidemiology and transmission of AHSV.