Rift Valley Fever, Mayotte, 2007–2008

After the 2006–2007 epidemic wave of Rift Valley fever (RVF) in East Africa and its circulation in the Comoros, laboratory case-finding of RVF was conducted in Mayotte from September 2007 through May 2008. Ten recent human RVF cases were detected, which confirms the indigenous transmission of RFV virus in Mayotte.


R ift Valley fever (RVF) is a zoonotic and mosquito-
borne disease that affects animals and humans (1). In humans, RVF infection is usually asymptomatic or characterized by an acute self-limited febrile illness. However, in ≈1%-3% of case-patients, it can cause serious manifestations that may lead to death or disabling sequelae in survivors (2,3). Identifi ed in the 1930s in Kenya, RVF virus (RVFV) circulates now in many other African countries (4)(5)(6)(7), as well as on the Arabian peninsula (8), where epizootics and associated human cases have been reported.
In early 2007, RVF outbreaks were reported in several eastern African countries (9). Additionally, in August 2007, autochthonous RVF transmission was found on the Indian Ocean island of Grande Comore, Republic of Comoros. A young Comorian with a 2-month history of severe encephalitis was transferred from the Republic of Comoros to Mayotte, where diagnostic and medical care facilities are more readily available. Mayotte, a French territory, is located northwest of Madagascar, in the Comoros archipelago ( Figure).
Anti-RVFV immunoglobulin (Ig) M was detected by IgM antigen-capture ELISA (10) in the serum sample of this patient. This sentinel event had substantial public health implications for Mayotte because the Republic of Comoros and Mayotte have frequent exchanges of populations and goods, both legally and illegally.
An active laboratory-based surveillance system for RVF was implemented for livestock in Mayotte in March 2008, which has led to several recent RVF cases being detected in cattle (Bureau of Veterinary Services, Mayotte, unpub. data). Subsequently, to assess the situation of RFV among humans, we retrospectively and prospectively analyzed serum specimens from febrile patients sampled by a routine surveillance program for chikungunya and dengue viruses. We describe the clinical, biological, and epidemiologic characteristics of human RVF cases in Mayotte.

The Study
Since the 2005-2006 chikungunya outbreak (11), laboratory-based surveillance of chikungunya virus (CHIKV) and dengue virus (DENV) is routinely maintained in Mayotte. Specimens are collected from patients attending local primary healthcare facilities or hospitals who seek treatment for dengue-like illness, i.e., acute onset of high fever accompanied by at least 2 of the following signs: arthralgia, body pain, headache, malaise, fatigue, or rash. Stored patient samples collected from September 2007-March 2008 and prospective specimens from this system, which initially tested negative for CHIKV and DENV as well as for Plasmodium spp. by using the rapid diagnostic test OptiMAL (Flow Inc., Portland, OR, USA), were screened for RVFV. We defi ned a confi rmed recent human case-patient as a person with dengue-like illness and the serum positive for anti-RVFV IgM by antigen-capture ELISA or RVFV RNA by reverse transcriptase-PCR (12). Any specimen positive when tested by the local reference laboratory was then forwarded for confi rmation to the National Reference Centre for Arboviruses and the World Health Organization Collaborating Centre for Arboviruses, Pasteur Institute, Paris, France. Identifi ed case-patients were investigated by using a standardized form to gather information, particularly regarding exposure data, e.g., peridomestic environment, handling of animal products, occupational exposures, and travel history. The relevant period of exposure was 3 weeks before disease onset. Medical records of confi rmed case-patients were also reviewed if needed. Investigations were conducted in accordance with the French National Institute of Public Health (Institut de Veille Sanitaire) guidelines for studies conducted in rapid response to public health threats. From September 1, 2007, through May 31, 2008, 220 serum specimens from persons with dengue-like illness who tested negative for Plasmodium spp., CHIKV, and DENV were screened; recent RVFV infection (i.e., presence of viral RNA or IgM) was found in 10 samples (4.5%). The earliest recorded onset of dengue-like illness was September 27, 2007, and the latest was May 14, 2008. Seven cases (70%) occurred from January through April, during the hot, rainy season (Table). Confi rmed case-patients were not clustered spatially and resided in the following counties: Mamoudzou (3), Bandraboua (2), Dembeni (1), Sada (1), Chirongui (2), and Boueni (1).
Although 2 patients were admitted to the intensive care unit for medical management, no severe neurologic mani-festations of RVF were observed, and no deaths could be attributed to RVF. Case-patient 6 ( Table), with a coexisting condition of cirrhosis after hepatitis B infection, had substantial thrombopenia (< 40,000 cells/mm 3 ) and died as a result of gastrointestinal bleeding and hepatic encephalopathy. Case-patient 10 was admitted to the hospital in May 2008 with a 5-day history of high body temperature, body pain, joint pain (particularly in the shoulders), and fatigue, combined with symptoms consistent with right-sided heart failure (shortness of breath, hepatomegaly, and peripheral edema). The patient was discharged with a diagnosis of pericarditis. He had a follow-up visit 1 month later, and a sample obtained on this occasion tested positive for RVFV IgM and IgG.
Predictably, contact with ruminants (sheep, cattle, or goats) was the predominant means of exposure among reported case-patients; this mode was documented in 5 of 9 cases that were fully investigated. Furthermore, only the presence of numerous breeding sites for mosquitoes in the housing environment remained as a possible means of exposure for 3 other case-patients. The mode of transmission remains undetermined for the remaining case-patient, who had no admitted direct contact with ruminants and whose housing environment did not have numerous breeding sites (Table).