In September 1998, several patients with fever and hemorrhagic syndrome were admitted to the Hospital of Aioun El Atrouss, Hodh El Gharbi Region, Mauritania. Sera from two of these four patients were positive for Rift Valley fever virus (RVFV) by immunoglobulin (Ig) M detection by enzyme-linked immunosorbent assay (ELISA), virus isolation on cell cultures, and reverse transcription polymerase chain reaction, focusing on the S segment of the viral genome.

Figure

Figure. Map of Hodh el Gharbi region, Mauritania.

From October to the end of December, three epidemiologic investigations were undertaken in five localities in the Hodh El Gharbi region (Figure). The Hodh El Gharbi is an extensive livestock farming region in an arid area. In September 1998, rains were exceptionally heavy, with a threefold increase over the 10-year average rainfall (86 mm vs. 26 mm).

During the investigations, serum samples were obtained from suspected cases in humans and animals (camels, goats, sheep, cattle, and donkeys) (Table). A suspected human RVFV case was defined as illness in any patient with fever (whether or not it was associated with hemorrhagic signs, icterus, or neurologic signs) occurring after September 1, 1998. Animal specimens were obtained from flocks in which abortions or stillbirths were reported in 1998. Based on a questionnaire among livestock breeders, the perinatal mortality rate (PMR) in flocks was estimated by using the ratio number of abortions, stillbirths, or deaths within 48 hours/number of females. Mosquitoes, sandflies, and biting midges were caught in CDC light traps. Mosquito larvae were also collected. Arthropods were sorted, pooled either by species and sex (for mosquitoes) or in polyspecific batches (sandflies, biting midges) in the field, and stored in liquid nitrogen. In light of the report on the possible involvement of rodents in the RVFV transmission cycle in South Africa (5), wild rodents were trapped concomitantly with the arthropods. Human and animal samples were tested for RVFV-specific IgG and IgM antibodies by ELISA as described (6) and for virus isolation in cell cultures and suckling mice. Viruses isolated were identified by indirect immunofluorescence with RVFV-specific monovalent hyperimmune mouse ascitic fluids, as well as complement fixation and seroneutralization tests (7). Arthropods were tested for the presence of RVFV by inoculating vector suspensions into cell culture, followed by identification as described for serum samples. Pooled rodent viscera were homogenized and inoculated intracerebrally into suckling mice for virus isolation.

Among the 90 human sera tested, 16.7% had evidence of recent (presence of IgM antibody or isolated virus) and 24.4% of past (presence of IgG antibody only) infection. Two virus strains were isolated. Among the 15 recently infected patients, median age was 26 years (range 10 to 45 years), male:female ratio was 2:0, and one death was reported. Proportions of hemorrhagic signs did not differ significantly among laboratory-positive and-negative cases (40.0% vs. 28.8%; p=0.5), suggesting that a cause other than RVFV should be considered to explain hemorrhages. Conversely, the prevalence of icterus and neurologic signs was significantly higher among positive cases (46.7% vs. 19.2%, p=0.04, and 53.3% vs. 20.5%, p=0.02, respectively), suggesting that these two signs were more specific indicators of RVFV in this human outbreak.

Among animals, 343 sera were tested from five species (sheep, goats, camels, cattle, and a donkey). Except for the donkey, all the species screened were positive for IgM antibodies, with prevalences ranging from 1.5% to 34.8%. These findings indicate not only widespread circulation of RVFV in the area but are also consistent with the high perinatal mortality observed in flocks, particularly among goats. The most affected species were sheep and goats, with an IgM prevalence of 34.8% and 16.3%, respectively; moreover, 11 RVFV strains were isolated from these two species (Table). However, the discrepancy between the perinatal mortality rates and the IgM-RVFV prevalence might suggest that other diseases causing abortion may have cocirculated in the area.

Among adult mosquitoes collected, Culex and Anopheles species were the most abundant. No RVFV strain was isolated, probably because the mosquito captures were undertaken at the end of the rainy season, when most breeding sites had dried up. This hypothesis is further strengthened by the small number of mosquitoes caught (546) and the absence of Aedes mosquitoes, a vector species for RVFV transmission in West Africa (8).

Seventy-three rodents belonging to five genera (Gerbillus, Desmodilliscus, Acomys, Arvicanthis, and Jaculus) and three families (Gerbillidae, Muridae, and Dipodidae) were captured in the same areas where mosquitoes were trapped, but no RVFV strains could be isolated and no serum was positive for IgM or IgG antibodies.

An outbreak of Rift Valley fever occurred in the Hodh El Gharbi region of Mauritania in 1998. Because of the proportion of hemorrhagic signs in humans and the high rate of perinatal mortality among some animals, it cannot be ruled out that some other pathogen may have been involved in the outbreak; this hypothesis merits further investigation.

Since the 1987 epidemics, RVFV circulation among livestock has been documented in this region (3,4). However, outbreaks among humans and animals have probably been underreported, emphasizing the need to strengthen surveillance in the southern area of the country to prevent the potential spread of any epidemic focus to neighboring countries through nomadic animal husbandry.

Furthermore, the unique ecologic and environmental context makes the Hodh El Gharbi region of interest for research to further understanding of factors influencing the emergence of this disease in West Africa.