GIVE

Predicting the Risk for West Nile using Sentinel Chickens or Monitoring Mosquito Infection Rates

As of September 25, 2012, there were more than 3,545 West Nile human cases in the U. S. with 145 deaths. Texas had more human cases than any other state with about 1,350 cases by this date and 43 deaths. 2012 will be one of the worst years for West Nile in U. S. history.

Florida fortunately dodged such an epidemic of West Nile in 2012 having a total of 43 human cases as of Sept. 22, 2012 with 23 of these in Duval County. The reasons for this are varied but almost certainly this is the result of the absence of environmental conditions in Florida that would support epidemic transmission of West Nile virus (WNV) to humans. Readers of BuzzWords and those attuned to FMEL's Encephalitis Information System (http://eis.ifas.ufl.edu/) should be familiar with the environmental triggers for WNV transmission. These include a bird breeding cycle that provides susceptible hosts timed with an increase in mosquito vector populations. Both are influenced by a cycle that includes a wet season, followed by drought that forces birds and mosquitoes around scarce water sources, followed by wetting that spreads the infected mosquitoes and infected birds. Thankfully there have been few Florida areas with such a cycle.

Over the years Florida's public health and mosquito control professionals have improved Florida's capabilities against WNV, as well as St. Louis and Eastern equine encephalitis viruses. Since these three viruses are dependent on infected birds to maintain them in nature, sentinel chicken surveillance has played a prominent role in forecasting their transmission. The importance of Florida's sentinel chicken surveillance program has been explained elsewhere in BuzzWords. Table 1 shows 20 BuzzWords articles about sentinel chicken surveillance for WNV that were published between 2001 and 2007.

Questions are still asked about the suitability of sentinel chicken WNV surveillance. Should Florida make a greater effort to get information about WNV using other methods such as by sampling field collected mosquitoes and analyzing these mosquitoes for WNV? Readers should be familiar with detecting WNV in mosquito pools as a measure of the mosquito infection rate. Rather than repeating the details I direct readers to the previously published articles in Table 1 following this article. They can be read online at: http://fmel.ifas.ufl.edu/buzz/archive.shtml or http://mosquito.ifas.ufl.edu/. For example see Tabachnick, WJ. 2004. Florida's Sentinel Chicken Surveillance Program: Smart Chickens. BuzzWords 4(1): 7-8 for why Florida's sentinel chickens are "smart" (Fig. 1).

smartchicken
Figure 1. A smart Florida sentinel chicken.

It is critical to understand that the primary goal of WNV surveillance is to obtain information in the most efficient manner to predict the risk for West Nile human cases. This is a different goal than simply detecting the presence of WNV in an area that was important only a few years ago when WNV had yet to spread throughout the U. S. Now one may likely "detect" WNV in many areas in the U. S. if one is able to test enough bird or mosquito samples. We now need information to assess when and where there is enough transmission in an area to cause substantial risk to humans. Therefore Florida and most areas in the U. S. need an actual estimate of WNV transmission to assess human risk in an area. One can estimate mosquito transmission from mosquito infection rates that can be obtained by sampling thousands of mosquitoes in an area, by the extent of sentinel chickens that seroconvert to WNV, or the appearance of West Nile disease in humans and other animals. However, it is well established that once human cases begin to occur transmission is well underway and control measures will likely be after when they would be most effective.

Consider the options of using mosquitoes or sentinel chickens for WNV surveillance. One must test thousands of mosquitoes for virus to obtain information about mosquito infection rates that is only indirect information about mosquito transmission. Mosquitoes must be collected in the area where there are infected mosquitoes. A similarly well-placed sentinel chicken can be bitten by thousands of mosquitoes and provide direct information about transmission through a single test for antibody in the chicken's blood. More mosquitoes can be sampled by using more chickens and actual transmission can be assessed for less cost and effort. Details about the efficiency of sentinel chickens vs. mosquito analyses for WNV can be found elsewhere (Tabachnick, WJ. 2006. Mosquito surveillance and West Nile prediction: Lessons learned. BuzzWords 6(4): 6-8).

Of course a surveillance program is useful only if the information is then used to make timely decisions to mitigate a potential outbreak. In 2005 Florida's Pinellas County Mosquito Control District initiated increased mosquito control based on their sentinel chicken surveillance information before the appearance of human cases of West Nile (Tabachnick, WJ and JF Day. 2005. West Nile surveillance and Florida Mosquito Control: Acting on surveillance information. BuzzWords 5(5): 6-8). In 2012 mosquito surveillance in Dallas County in Texas found high WNV mosquito infection rates beginning in June and continuing throughout the summer.

There were 181 West Nile cases and 9 deaths by August 10 and by August 15 several Dallas County cities agreed to aerial spraying for mosquito control. Aerial spraying began on August 16 in north Dallas, and August 22 in south Dallas cities. It will be interesting to assess how the mosquito surveillance information in Dallas County was used to make decisions about the extent of the risk in the county, how it was used as part of the public health decision-making process and the effectiveness of the mosquito control efforts that were implemented. Dallas County had the highest number of West Nile cases in Texas but this outbreak had a lower incidence than the 2010 dengue epidemic in Key West. The highest incidence in Dallas County was in Highland Park at 45.8/100,000 as of Sept. 7, 2012 ( www.dallas-cms.org/community_health/DCHHS/). This is only about 1/10 the incidence of dengue (325/100,000) that occurred in Key West in 2010. Disease incidence is essential to gauging human risk.

Florida's scant resources for WNV surveillance must be used to provide useful information in the most effective manner. Unfortunately there are still far too many areas of Florida where mosquito-borne disease surveillance is woefully inadequate. Florida may pay a heavy price for the absence of useful surveillance information should WNV strike one of these areas. There will be circumstances when one or the other of the available surveillance tools may be more useful, efficient, and/or effective. Which one to employ must be based on an understanding of the specific surveillance goals for the area under surveillance and the effectiveness and efficiency of the tool in meeting these goals.

Walter J. Tabachnick
Florida Medical Entomology Laboratory
Professor, Department of Entomology and Nematology
University of Florida, IFAS
Vero Beach, FL

Table 1. Articles on Sentinel Chickens and WNV Surveillance published in Buzz Words


  • Tough Choices for West Nile Virus Surveillance Programs Mar/Apr 2007
  • Sentinel Chicken Surveillance for West Nile Virus Nov/Dec 2006
  • Sentinel Chicken Surveillance: Some Pitfalls in Analyzing the Data Sep/Oct 2006
  • Mosquito Surveillance and West Nile Prediction: Lessons Learned Jul/Aug 2006
  • West Nile Surveillance and Florida Mosquito Control: Acting on Surveillance Information Sep/Oct 2005
  • Florida West Nile Surveillance: Estimating Mosquito Transmission Frequencies Mar/Apr 2005
  • A Florida Mosquito Control Arbovirus Response Plan Jan/Feb 2005
  • West Nile Epidemic? These are the questions we will ask Jul/Aug 2004
  • Sentinel Surveillance and Human Risk for West Nile Virus Mar/Apr 2004
  • Florida's Sentinel Chicken Surveillance Program: Smart Chickens Jan/Feb 2004
  • Great Move by the Indian River County Health Department and the Florida Department of Health, September 2003 Nov/Dec 2004
  • Mosquito pooling: Getting a return on your investment Mar/Apr 2003
  • The relationship of WN dead birds and human cases Mar/Apr 2003
  • West Nile virus in Florida: The Calm Before the Storm? Jan/Feb 2003
  • Predicting the "Big Event" May/Jun 2002
  • West Nile virus Detection: The details are important. Apr/May 2002
  • Sentinel Chicken Surveillance and West Nile Virus in Florida Nov/Dec 2001
  • West Nile in Florida: Getting Ready for 2002 Sep/Oct 2001
  • Finding WN Positive Birds in Florida: Getting Serious Mar/Apr 2001
  • A West Nile Virus Positive Bird: WN Response by the Numbers Jan/Feb 2001