A suspected human RVFV case-patient was defined as a person with fever associated or not with hemorrhagic, jaundice, or neurologic symptoms or any person who died who had had overt hemorrhagic fever symptoms from September through December 2003. A confirmed human RVFV case-patient was defined as a person for whom laboratory tests confirmed an acute or recent RVFV infection, e.g., by >1 of these results: immunoglobulin M (IgM), reverse transcription–PCR (RT-PCR) or virus isolation positive results. A human contact was defined as any person or relative who had been directly in contact with a confirmed human or animal case-patient, or with a person who died who had had overt hemorrhagic fever symptoms from September through December 2003.

The different cases recorded and the linkages between them are represented in Figure 2. The 9 confirmed case-patients, identified before the investigation, were from the Assaba, Brakna, Trarza, and Tagant provinces. Subsequent discussions with the head of the sanitary district enabled the localization of the residences and the relatives for 2 of them (index case-patients 1 and 2). Several confirmed cases-patients (index case-patients 3 to 9) and their relatives were not found due to the great distances between localities or nomadic behaviors of populations. However, for all suspected case-patients (S), further investigation was conducted in the provinces where the confirmed case-patients lived.

In total, 98 persons (66 contacts, 23 with suspected cases, and 9 with confirmed cases) were included in this study. Of these persons, 25.5% (25/98) had evidence of recent RVFV infection (i.e., presence of IgM, viral RNA or virus, or >1 of these results), and 10% had evidence of past infection (i.e., presence of IgG alone). Seven viral strains were isolated. For the 25 patients who were recently infected (9 with confirmed cases before the investigation, 6 contacts, and 10 with suspected cases at the time of investigation), the median age was 21 years (range 7–50 years), 4 died, and 16 had hemorrhagic signs (hematemesis, vaginal bleeding, severe hemoptysis, bleeding from the gums and venipuncture sites, petechial rashes, and ecchymoses of the skin). Among the 23 suspected case-patients from whom blood samples were collected, 10 were infected by RVFV; only 2 patients were positive for malaria parasites.

In Assaba Province, 2 confirmed case-patients were recorded before the investigation (index case-patients 1 and 2), and 1 suspected case-patient (S1) was found during the investigation. The index case-patient 1 was dead at the time of the investigation; however, RVFV (strain SHM169867) was isolated from a blood sample taken during his illness. No RVFV infection was noticed in contact persons associated with this patient. However, IgM against RVFV was detected in animals living in and near the residence of index case-patient 1. For index case-patient 2 (infected with the strain SHM169872), a second blood sample was taken, and presence of IgM antibodies against RVFV was confirmed. No evidence of RVFV infection was detected in contact persons or in animals living in or near the residence of this patient. The suspected case-patient 1 was identified in Kiffa Hospital, where he was admitted on November 3, 2003. This 50-year-old patient had onset of fever on October 24, 2003, with asthenia, jaundice, nausea, hematemesis, epistaxis, and gingival hemorrhage. Serologic tests showed IgM against RVFV. No evidence of virus infection was detected in those who had accompanied this patient to the hospital.

In Brakna Province, where 3 cases were confirmed before the investigation (index case-patients 3–5), 6 suspected case-patients (S2–S7) were found during the investigation. Laboratory testing of samples from the 3 index case-patients showed IgM against RVFV by ELISA and RVF viral RNA by RT-PCR. A virus strain (SHM169898) was isolated from index case-patient 4. Among the suspected case-patients from this province, 3 deaths (S2–S4) were recorded, but no samples were available from those patients. Nevertheless, animals in the vicinity of S1, S2, and S3, were found to be infected (5/7, 5/8, and 3/10 animals, respectively). In addition, infection was detected in 4 of 6 persons who had been in contact with suspected case-patient 3. In contrast, no evidence of infection by RVFV was detected in contacts of S2 and S4. The suspected case-patient 5 was a student living in Makhtar Lahjar, who had fever onset on October 20, 2003, was admitted to the National Hospital Center (NHC) of Nouakchott on November 6, and from whose blood RVFV (strain SHM172776) was subsequently isolated. During the investigation, 2 suspected case-patients (S6 and S7) were discovered in the Healthcare Center of Makhtar Lahjar. Suspected case-patient 6 was a 17-year-old female patient who had onset of fever on October 10 and who was admitted to the center on November 1 with headache, abdominal pain, vomiting, hemorrhages, and epistaxis; virus was isolated from this patient (strain SHM172805). The contacts associated with suspected case-patient 6 were also found to be infected with RVFV. Suspected case-patient 7 was a 35-year-old man. He had a fever on October 15 and was admitted to the Health Center of Makhtar Lahjar on November 1 with headache, nausea, vomiting, and epistaxis. The RVF IgM test result for this patient was positive for RVFV.

In Trarza Province, 2 confirmed case-patients were observed during the period of surveillance (index case-patients 6 and 7) and 4 suspected case-patients (S8–S11) were identified during the investigation. Viral isolation was positive for the index case-patient 7 (strain SHM169885). Among the suspected case-patients, 2 (S8 and S9) died before blood samples could be obtained. However, blood testing of samples from 2 animals living in the vicinity of S8 and from 2 human contacts of S9 found recent RVFV infection. Suspected case-patient 10 was 28-year-old man, with onset of fever on October 22, who came for consultation to Keur Macene Healthcare Center. He had a prolonged cough without hemorrhagic symptoms, and an IgM ELISA result for RVFV was positive. The suspected case-patient 11 is a 26 year-old woman, with onset of fever on October 22, who was admitted to the healthcare center of Rkiz with headaches, asthenia and anorexia without hemorrhagic signs. The RVF IgM test result of this patient was positive. In Tagant Province, index case-patient 8, whose condition was diagnosed before the investigation, was a man who came for consultation at the provincial hospital on September 24, exhibiting fever and hematemesis. RVFV (strain SHM169868) was isolated from a blood sample taken on September 29.

In Gorgol Province, 2 suspected case-patients (S12 and S13) were evacuated to the NHC of Nouakchott. The onset of their symptoms dated to October 23 and October 25, respectively. Samples from each were positive for IgM against RVFV by October 30, and RVFV (strain SHM172768) was isolated from S13, who died on October 30.

In Dakar, Senegal, an “imported” case (index case-patient 9) in a person from Rosso, Mauritania, was diagnosed by positive results by ELISA IgM and RT-PCR. This patient was first admitted to the NHC of Nouakchott, Mauritania, before being transferred to the Hôpital Principal de Dakar.

Serum samples were obtained and tested from 54 domestic animals (48 goats and 6 sheep) living in the visited localities (Table 1). The median age was 4 years (range 1–15), and the abortion rate was 70% during the last gestation period. IgM against RVFV was detected by ELISA in 25 of 54 animals; no IgG against RVFV was found in any of the 54. Among the animals with a positive test result, the abortion rate was 92%.

A total of 22,201 mosquitoes, belonging to 4 genera and 17 species, were collected. Culex poicilipes was the most frequent species (43.8%), followed by Cx. antennatus (23%) and Mansonia uniformis (9%). A total of 544 monospecific pools were constituted and submitted for viral isolation. Only Cx. poicilipes was found to be associated with RVFV. Three strains (ArD 174367, ArD 174303, and ArD 174347) were isolated from the 146 pools constituted in Guimi Province, giving rise to a minimum infection rate of 0.04% for this locality and 0.01% for the whole study site (Table 2).

RNA was extracted from the 8 viral strains isolated from humans, and fragments of the S, M, and L segments were amplified and sequenced. No amino acid (aa) differences were found between the fragments of the S or L segments analyzed (198 and 51 aa, respectively). A single amino acid difference was found between the M fragment (255 aa) of the 5 viral strains analyzed. Results of phylogenetic analyses of the nucleotide sequences of amplified fragments from 3 segments belonging to 2 representative strains (H1MAU03 [SHM169867] and H2MAU03 [SHM169868]) isolated during this epidemic and previously described nucleotide sequences of RVFV are shown in Figure 3. The strains identified in Mauritania 2003 are consistently located within the East/Central lineage for all trees. This lineage contains viral strains that circulated in Madagascar (1991), Kenya (1997), Chad (2001), and Saudi Arabia (2001).