PARENT SESSION
Contributed Oral Session 94: Toxicology and Disease : Viruses and Epidemics
Wednesday, August 10, 1:30 PM - 5:00 PM, Meeting Room 514 A, Level 5, Palais des congrès de Montréal

Inter-annual associations between precipitation and human incidence of West Nile Virus in the United States.

Landesman, William^*,1, Allan, Brian², Knight, Tiffany², Chase, Jonathan², ¹ Rutgers University, New Brunswick, NJ² Washington University, St. Louis, MO

ABSTRACT- A prevailing dogma in epidemiology is that higher-than-average levels of precipitation lead to mosquito outbreaks, which in turn cause high levels of mosquito-borne illness in humans. However, recent investigations suggest that droughts in one year can lead to mosquito outbreaks in the following year, possibly as a result of ecological changes in aquatic communities that reduce predators and competitors of mosquito larvae. Here, we evaluate the role of precipitation in contributing to spatial and temporal variation in the incidence of West Nile Virus in humans during the recent U.S epidemic. We compared county-level human per capita WNV incidence with precipitation from both the year concurrent with and prior to human incidence of WNV in the U.S. County-level Standardized Precipitation Index (SPI) data from 2001 was positively correlated with WNV incidence in 2002 and SPI in 2002 and 2003 were negatively correlated with WNV incidence in 2003 and 2004, respectively. For all three analyses, the previous year's SPI produced a stronger association with human incidence than did the SPI of the concurrent year. WNV was more prevalent in the eastern U.S. in 2002 and in the western U.S. in 2003 and 2004. We hypothesize that the change between the positive relationship found in 2002 and the negative relationships found in 2003 and 2004 may be due to differences between mosquito communities in the eastern and western U.S. Container-breeding mosquitoes such as Culex pipiens and Cx. restuans are believed to be the primary vectors of WNV in the eastern U.S. Precipitation may increase the availability of larval mosquito habitat, which would result in a positive correlation between the previous year's SPI and WNV. Alternatively, Cx. tarsalis, which breeds in ponds, may be the primary vector of WNV in the western U.S., and this species may be more likely to show high abundances following a drought year. In tests of this hypothesis, we found that the strength of these relationships improved when only eastern (2001 SPI x 2002 WNV) or western (2002 SPI x 2003 WNV) counties were considered. We suggest that increased precipitation could potentially create more habitat for container-breeding mosquitoes like Cx. Pipiens, while droughts may create ecological conditions that lead to outbreaks of pond-breeding mosquitoes like Cx. tarsalis.

Key words: West Nile Virus, Drought