The first survey, intended to clarify the Rift Valley fever epidemiologic situation on the island, was undertaken in the M’Tsangamouji area (northern part of Mayotte). It examined samples from 29 illegally introduced goats and 79 cattle born on the island and living near the goats. Among the 29 goats, competitive IgG ELISA found IgG against RVFV in 4 goats that had been introduced illegally during November 2007–April 2008 (13), and IgM-capture ELISA found IgM against RVFV in 2 goats (14), suggesting recent infection. Among the 79 cattle, IgG against RVFV was found in 29 (37%) and IgM against RVFV was found in 3 (4%).

These data led us to conduct the second survey, a retrospective study on the whole island to define the geographic distribution of the infection and to trace back the period of introduction. This survey analyzed 301 cattle serum samples collected during June 2007–May 2008 on 104 farms in 17 districts. Exposure to RVFV was indicated by competitive IgG ELISA detection of RVFV-specific antibodies. Positive results were found for 32 samples from cattle in 9 districts (Table). The overall apparent RVFV seroprevalence of 10.6% (95% CI 7%–14%) was supported by the high specificity of the ELISA (14). The 32 positive samples came from cattle distributed all over the island (Figure 2, panel A).

Because RVFV circulation had been confirmed as early as 2007–2008 in Mayotte, a third cross-sectional and retrospective study was conducted to trace previous virus circulation. The 120–130 samples that had been collected from cattle since 2004 were randomly selected every year over a 4-year period and analyzed by IgG ELISA; results were confirmed by neutralization tests (15). These results helped evaluate RVFV circulation on Mayotte island before the 2007–2008 outbreak on the eastern Africa mainland. In 2004, a total of 29 of 130 cattle had IgG against RVFV; thus, seroprevalence was high (22.66%). In 2005, seroprevalence rates fell; IgG against RVFV was found in only 4 of the 130, suggesting low levels of RVFV circulation. In 2006 and 2007, seroprevalence increased; IgG against RVFV was found in 16 of 130 and 39 of 126 cattle, reaching seroprevalence rates of 12.31% and 30.95% for 2006 and 2007, respectively. Specific IgM against RVFV was not detectable during this cross-sectional and retrospective study.

The fourth survey, intended to evaluate the recent virus dynamics in Mayotte, was a longitudinal serologic survey of goat farms. In June 2008, a total of 13 goat farms were selected and all 272 animals were screened for antibodies against RVFV. Of the 13 farms, 8 had seropositive animals (herd prevalence 62%, 95% CI 35%–88%) (Figure 2, panel B). Intraherd prevalence ranged from 6% to 42%. The 5 farms with seronegative goats (total 70 goats) were included in the longitudinal study. During August 2008–August 2009, the seronegative goats were sampled every 6–8 weeks. Only 1 goat, located in Bouyouni and sampled in February 2009, had seroconverted and was confirmed IgM positive for RVFV. Virus isolation attempts were unsuccessful.

The 4 serologic surveys conducted in Mayotte revealed medium to high rates of RVFV prevalence all over the island. The high rates obtained with the first survey in the M’Tsangamouji area suggest that illegal animal movements from the Comoros are a likely source for RVFV introduction onto Mayotte. Results of the 4-year survey show that the virus was already present in 2004. After a low seroprevalence rate in 2005, the increased seroprevalence rates for 2006 and 2007 suggest that the virus had recirculated or had been newly introduced.

It is unclear why relatively high circulation of RVFV in Mayotte and an increased rate of seroprevalence to 22% did not result in detectable clinical cases in animals while Rift Valley fever was diagnosed for humans with brain disorders (11). This finding might be because the density of susceptible animals on the island was high enough to support virus circulation but too low to support waves of epidemic abortion and death. These study findings, coupled with epidemics in eastern Africa, illustrate the risk for introduction of infectious agents from the African mainland to Mayotte or other Comoros islands. Entomologic studies need to be conducted to identify all potential vector species on the island and to better understand the ecologic and climatic factors that favor RVFV dissemination. The ecologic factors in favor of Rift Valley fever outbreaks might be comparable between Mayotte, the other Comoros islands, Madagascar, and the eastern African mainland (Kenya, Tanzania, and Mozambique) but need to be looked at more closely. In Mayotte, an entomologic surveillance program is being developed to help define the distribution of potential vectors in association with virus circulation and provide better understanding of disease spread mechanisms. The role of wildlife should also be investigated. These data highlight the need for extensive studies to determine RVFV distribution and to evaluate the effect of Rift Valley fever on the susceptible livestock populations.