Rift Valley fever potential, Arabian Peninsula.

To the Editor: Rift Valley fever (RVF) virus causes severe disease, abortion, and death in domestic animals (especially young sheep, cattle, and goats) in Africa and the Arabian Peninsula. Humans are infected by mosquitoes, which maintain epizootic transmission, or through exposure to infected animal tissue. Although human disease may be mild, sometimes severe retinitis, meningoencephalitis, or hemorrhagic fever syndromes may develop in patients. In Africa, epizootics and associated human epidemics usually follow heavy rainfall (1).

RVF was first confirmed outside Africa in September 2000. The outbreak in southwestern coastal Saudi Arabia and neighboring coastal areas of Yemen resulted in an epizootic with >120 human deaths and major losses in livestock populations from disease and slaughter (2,3). RVF virus isolated from the floodwater mosquito Aedes vexans arabiensis during the outbreak was closely related to strains from Madagascar (1991) and Kenya (1997), which suggests that the virus was imported through infected mosquitoes or livestock from East Africa (3). The Arabian outbreak followed increased rainfall in nearby highlands that flooded the coastal areas and created ideal environments for mosquito populations similar to those found in RVF-endemic regions of East Africa (4). Most RVF activity was associated with flooded wadi agricultural systems; no cases were reported in the mountains or in dry sandy regions, where surface water does not accumulate long enough to sustain mosquito breeding.

To provide early warning of conditions favorable for RVF epidemics, the National Aeronautics and Space Administration (NASA) and the Department of Defense Global Emerging Infections Surveillance and Response System (DoD-GEIS) monitor the satellite-derived normalized difference vegetation index (NDVI), which reflects recent rainfall and other ecologic and climatic factors (5-7). NDVI anomalies in the highlands east of affected areas during the 2000 outbreak (online Appendix Figure panel A, available from http:// www.cdc.gov/ncidod/EID/vol2no03/ 05-0973-G.htm) showed a spatial pattern (although of lower magnitude) similar to recent anomalies in those areas (online Appendix Figure panel B). Greater than normal NDVIs (20%-60%) were seen in the Sarawat Mountains, from just northeast of Djeddah, Saudi Arabia, and southwestward beyond Jizan and into Hodeidah governorate in Yemen during May and June 2005.

Satellite-derived rainfall estimates show that widespread rainfall occurred over most of western Saudi Arabia and Yemen from mid-April to mid-June 2005 (8) and accounts for the high magnitude and spatial pattern of observed NDVI anomalies in May and June 2005. Rainfall was concentrated in the mountainous regions east of the Red Sea coast, and was heaviest in the areas east of Djeddah and Jizan, with rainfall totals as high as 120 mm and 60-80 mm, respectively, during April 2005, compared with the same period in 2000 (10-50 mm) (online Appendix Figure panels C and D) and in southwestern Yemen, with totals as high as 120 mm during May. In the area east of Djeddah, total rainfall in April 2005 was 150 mm above the long-term average for that month. Flooding was reported in Hodeidah Governorate, Yemen during May (9) and could be expected in other Red Sea coastal areas following such heavy rainfall. This created habitats appropriate for breeding of mosquitoes capable of transmitting RVF, as occurred in 2000.

No human cases of RVF have been reported in Saudi Arabia and Yemen since the 2000 outbreak, but in September 2004 the Saudi Ministry of Agriculture reported that 5 RVF-seropositive sheep had been identified during routine surveillance in Jizan where most infected persons were exposed during the outbreak in 2000 (2). The primary infection was estimated to have occurred in April 2004 (10). The NDVIs and rainfall patterns alerted the Yemen and Saudi Arabia Ministries of Health and Ministries of Agriculture to conduct field investigations with the Food and Agriculture Organization and the World Health Organization.

Since RVF virus can be maintained in mosquito eggs for extended periods and transmitted under favorable conditions (6), the high magnitude of NDVI and rainfall patterns reported should prompt heightened veterinary and human surveillance for RVF in coastal Arabia and mass vaccination of susceptible animals. The current RVF model (7) is indicative of conditions that would promote vector breeding and could result in an outbreak of mosquitoborne diseases.

Acknowledgments

Systeme Probatoire pour l'Observation de la Terre (SPOT) vegetation data were provided by United States Department of Agriculture/Foreign Agricultural Service under an analysis agreement with the Global Inventory Monitoring and Modeling Systems group at the NASA Goddard Space Flight Center. Satellite rainfall estimates were provided by the National Oceanic and Atmospheric Administration Climate Prediction Center.

This study was supported in part by the DoD-GEIS, Silver Spring, Maryland.

Assaf Anyamba, * Jean-Paul Chretien, ([dagger]) Pierre B.H. Formenty, ([double dagger]) Jennifer Small, * Compton J. Tucker, * Joseph L. Malone, ([dagger]) Hassan El Bushra, ([section]) Vincent Martin, ([paragraph]) and Kenneth J. Linthicum (#)

* NASA/Goddard Space Flight Center, Greenbelt, Maryland, USA; ([dagger]) Department of Defense Global Emerging Infections Surveillance and Response System, Silver Spring, Maryland, USA; ([double dagger]) World Health Organization, Geneva, Switzerland; ([section]) World Health Organization, Cairo, Egypt; ([paragraph]) Food and Agricultural Organization, Rome, Italy; and (#) United States Department of Agriculture, Gainesville, Florida, USA

References

(1.) Davies FG, Linthicum KJ, James AD. Rainfall and epizootic Rift Valley fever. Bull World Health Organ. 1985;63:941-3.

(2.) Centers for Disease Control and Prevention. Outbreak of Rift Valley fever--Saudi Arabia, August-October, 2000. MMWR Morb Mortal Wkly Rep. 2000;49:905-8.

(3.) Miller BR, Godsey MS, Crabtee MB, Savage HM, Al-Mazrao Y, Al-Jeffri MH, et al. Isolation and genetic characterization of Rift Valley fever virus from Aedes vexans arabiensis, Kingdom of Saudi Arabia. Emerg Infect Dis. 2002;8:1492-4.

(4.) Food and Agriculture Organization. Update on RVF outbreaks in Saudi Arabia and Yemen. EMPRES Transboundary Animal Diseases Bulletin. 2000 Jul-Dec [cited 2005 Jul 11]. Available from http://www. fao.org/documents/show_cdr.asp?url_file=/ DOCREP/003/X9550E/X9550E03.HTM

(5.) Linthicum KJ, Anyamba A, Tucker CJ, Kelley PW, Myers MF, Peter CJ. Southern oscillation index, sea surface temperature and satellite vegetation index indicator to forecast Rift Valley fever epizootics/epidemics in Kenya. Science. 1999;285: 397-400.

(6.) Anyamba A, Linthicum KJ, Mahoney R, Tucker C J, Kelley PW. Mapping potential risk of Rift Valley fever outbreaks in African savannas using vegetation index time series data. Photogrammetric Engineering and Remote Sensing. 2002;68:13745.

(7.) Department of Defense Global Emerging Infections Surveillance and Response System. Climate and disease connections: Rift Valley fever monitor; 2000-2005. [cited 2005 Jul 11]. Available from http://www.geis.fhp.osd.mil/GEIS/Surveill anceActivities/RVFWeb/indexRVF.asp

(8.) International Research Institute for Climate Prediction. Malaria early warning system [cited 2005 Jul 11]. Available from http://iridl.ldeo.columbia.edu/maproom/.R egional/.Africa/.MEWS/

(9.) ReliefWeb. Yemen: floods--May 2005. [cited 2005 Jul 11]. Available from http://www.reliefweb.int/rw/dbc.nsf/doc10 8?OpenForm&rc=3&emid=FL-2005-000070-YEM

(10.) World Organization for Animal Health. Rift Valley fever in Saudi Arabia--serological findings. ProMed. 2004 Oct 3 [cited 2005 Dec 30]. Available from http://www. promedmail.org (archive no. 20041003. 2723)

Address for correspondence: Kenneth J. Linthicum, Center for Medical, Agricultural, and Veterinary Entomology, US Department of Agriculture, 1600/1700 SW 23rd Dr, Gainesville, FL 32608, USA; fax: [phone]; email: [e-mail]. ufl.edu