A retrospective analysis of the epidemiology of Rift Valley fever in Namibia

Abstract

Rift Valley fever (RVF) is a peracute or acute disease of domestic ruminants and humans in sub-Saharan Africa, caused by a mosquito-borne virus. It is a high priority pathogen because of its potential to cause severe economic harm to the livestock industry and to cause life threatening haemorrhagic disease in humans. The disease was first recorded in southern Africa when a large epidemic occurred in the South Africa in 1950, and the first recorded outbreak in Namibia was in 1957. Since then, occasional large epidemics have occurred in southern Africa, with long interepidemic periods. The epidemiology of RVF is complex and many questions regarding the movements of the virus and its survival during the interepidemic period remain unanswered. The aim of this study was to compile a comprehensive description of the history of RVF in Namibia and to describe its epidemiological characteristics. This was accomplished using information available in the scientific literature, annual reports, disease reports and reports to the OIE. The geographical location and temporal occurrence of each outbreak was recorded as accurately as allowed by available records. Also recorded were suspected RVF outbreaks, defined as those outbreaks in which samples were not collected for laboratory analysis or RVF was not confirmed on submitted samples but where the clinical picture was suggestive of the disease. Serological surveys done in humans and animals were also included in the study. The collected data were analysed descriptively, by risk mapping and by cluster analysis. The relatively low number of recorded outbreaks and the poor spatial resolution of much of the data prevented more detailed multivariable analysis. Maps were produced to show the districts affected for the outbreaks with no coordinates and the exact location of the outbreaks which had coordinates. This was then followed by a detailed description of each outbreak showing the species affected and the mortalities caused. Risk mapping was done to identify areas of the country which are at high risk of having outbreaks. A quarter degree square grid was used to show the cumulative number of confirmed outbreaks occurring from 1957 to 2011. The accuracy of this was, however, limited due to the poor spatial resolution of data prior to 1986, which recorded only the district(s) affected. The risk map was visually compared with maps of sheep and cattle density and rainfall. A space-time permutation model, using case-only data, was used to detect space-time clusters with high rates, using SaTScan software on all the confirmed outbreaks with GPS coordinates. The objective was to detect areas of significantly high rates of RVF in Namibia, testing whether the outbreaks were randomly distributed over space and time. Space time permutation requires the use of precise geographic coordinates; therefore the only confirmed outbreaks that could be used for this analysis were those occurring during 2010 and the 2011. A total of six years had outbreaks of RVF in Namibia, the major outbreaks occurring in 1957, 1974, 1984, 2010 and 2011. Rift Valley fever was confirmed in the Karas, Hardap, Khomas, Erongo, Otjozondjupa, Omaheke and Oshikoto regions, with suspected outbreaks occurring in the Kavango and Caprivi regions. SaTScan analysis showed that there were two statistically significant outbreak clusters observed, one in the Hardap region in 2010 and the other in the Oshikoto region in 2011. The south-eastern part of the country was shown to be predisposed to RVF outbreaks; this correlated with sheep population density. The southern part of Namibia receives less rainfall and is hotter than the north, with colder winters, factors which may reduce vector and virus survival and therefore limit continuous viral circulation. This likely renders livestock highly susceptible to infection and if there is an introduction of the virus a severe epidemic may occur. In the Northern Communal Areas and adjacent Etosha National Park the positive serological results in humans and wildlife show that there is continuous or intermittent low level circulation of the virus. This could be leading to high levels of herd immunity and hence no confirmed outbreaks recorded in these areas to date. Nevertheless, all suspected cases should be tested for RVF to avoid misdiagnosis and under-reporting of cases.