Guidelines for Arbovirus Surveillance Programs in the United States
Click here for CONTENTS Page
For contact details regarding publication see Title Page

Surveillance systems quantify disease activity at a given time, predict the probable future course of the disease cycle, and indicate when control should be started to prevent epizootic or epidemic transmission. This requires that surveillance programs be long-term, proactive projects, gathering and analyzing data in epidemic and nonepidemic years to provide a basis for setting thresholds and decision making. No single technique can collect all of the data needed for a rational assessment of the risk of vector-borne disease.

Because arbovirus cycles are complex, and components of the cycle vary regionally, threshold levels and indicator parameters must be determined individually for each surveillance region. Current-year data should be compared with historical data for the same region or locality, rather than looking for absolute index values. The appearance of human or equine cases is unlikely to be associated with a specific value of a single index (e.g., vector females per light trap night) over large geographic areas. However, such indices may prove locally useful.

The following is a brief summary, by disease, indicating the methods we feel are most appropriate for an ideal surveillance program. The realities of local, state, and regional resources will often restrict the extent to which these recommendations can be fully implemented. For an overview of the types of surveillance systems currently employed in various states, see Appendix I.

The distribution of EEE is intimately associated with the distribution of the enzootic vector, Cs. melanura. Thus, the presence of this mosquito, or of habitat capable of supporting this species marks areas with the potential for EEE transmission. The density of Cs. melanura has often been related to the intensity of EEE activity. However, monitoring Cs. melanura population density alone is not a reliable surveillance tool; other mosquito species are responsible for transmission to horses and humans. In addition, a susceptible bird Successful EEE surveillance programs will monitor components of both the enzootic cycle (vector population, bird population, virus prevalence) and of the epizootic cycle (bridge vector populations). population is required for amplification of the virus.

Meteorologic data: 

Both local and regional weather patterns are important. The ideal program will monitor rainfall and temperature patterns that promote the development and survival of large mosquito populations, especially Cs. melanura, in each area. It should examine annual rainfall patterns for the previous 2-3 years. It should compare monthly rainfall quantities to local and regional averages, especially during fall and spring. It also should look for early temperatures that permit mosquito development. At least in the northeast, programs will monitor ground water levels in freshwater swamps as a method of predicting subsequent Cs. melanura populations.

Vector data:

Surveillance programs should monitor current and historical patterns in density and age structure of Cs. melanura populations in swamp foci. Collections of Cs. melanura are made by using CO2-baited CDC light traps and black resting boxes are effective for collecting Cs. melanura. Parity rates can be determined with sufficient accuracy to establish crude age structure by using the tracheation method of Detinova.⁸⁰ The program also should monitor field infection rates in Cs. melanura populations by submitting pools to the state or regional laboratory for virus isolation.

The ideal surveillance program also will monitor the density and age structure of epizootic vector species. These include Cq. perturbans and Ae. canadensis in swamp habitats, Ae. vexans in upland floodwater sites near swamps, and Ae. sollicitans in areas where enzootic foci are adjacent to coastal salt marshes.

Vertebrate host data:

The ideal surveillance program will measure the prevalence of EEE viral antibody in wild passerine birds located near swamp foci during the current season (monthly) and compare to EEE antibody levels during the previous 2-3 years.

Other data:

In areas where they are known to be effective predictors, seroconversion in sentinel chickens should be monitored. Programs should conduct active or passive surveillance for EEE in unvaccinated horses.

The LaCrosse virus cycle differs somewhat from that of other viruses discussed here. The primary vector is the tree hole mosquito, Ae. triseriatus. The virus is maintained in a focus by vertical (transovarial) transmission in the mosquito. The primary amplification hosts are chipmunks and squirrels. The virus is limited to wooded areas by the ecological requirements of the mosquito and vertebrate hosts. Ae. triseriatus does not disperse great distances from wooded areas. Human cases of LAC have been associated with the presence of artificial containers (i.e., discarded tires) in adjacent wooded areas. These containers can produce very large Ae. triseriatus populations.

Meteorologic data:

The relationship, if any, between rainfall and Ae. triseriatus density is not known, but frequent rainfall will repeatedly flood treeholes and containers and produce frequent hatches. Therefore, surveillance programs should monitor seasonal rainfall.

Vector data:

The density and field infection rate of Ae. triseriatus should be monitored. Adults can be collected at bait or resting in the understory of the woodlot. Ovitraps can be used to determine the number of eggs produced by the population. Eggs from the ovitraps can then be used to determine the proportion of offspring transovarially infected with LAC. Because ovitraps compete with naturally occurring oviposition sites for egg deposition, results should be interpreted with caution. Ovitrap results are useful for comparing density within a site over time, but comparisons of population density between woodlots are not reliable. Discarded tires and other artificial containers often serve as LAC virus foci near human habitations, and these should be inspected. Where Ae. albopictus is abundant, collect and process specimens for virus isolation.

Vertebrate host data:

The ideal surveillance program will monitor current and historical patterns in presence, density and seroconversion rate of chipmunks and tree squirrels in LAC virus foci.

Other data:

Surveillance data can be supplemented by serosurveys of humans living near LAC foci. Areas at greatest risk can be identified and mapped by identifying hardwood forest habitats where Ae. triseriatus and chipmunks or squirrels are abundant.

At least three, and probably four, geographically distinct patterns of SLE transmission can be distinguished, based on the primary vector species (see Chapter 5). Techniques used to monitor SLE activity will vary depending on whether the vector is Cx. tarsalis, Cx. p. pipiens, Cx. p. quinquefasciatus, or Cx. nigripalpus.

Meteorologic data:

The amount of rainfall, interval between rainfall events (Florida), and January - July cumulative precipitation (California) have been useful predictors of SLE activity. Complex seasonal temperature and rainfall patterns have been found for SLE transmitted by Cx. pipiens complex mosquitoes.²⁴⁷

Vector data:

Surveillance programs should sample populations of the important local vector or vectors (Appendix II lists sampling methods for particular species). Mosquito pools should be submitted for arbovirus isolation to a state or regional laboratory. Programs should monitor vector abundance in peridomestic container habitats when Cx. pipiens complex is involved in transmission.

Vertebrate host data:

Passeriform and columbiform birds that are locally important in the enzootic SLE cycle (see p. ?) should be bled to obtain serum samples. Programs may or may not choose to use sentinel chicken flocks, depending on whether seroconversions precede or are concurrent with human infections. This appears to vary with region and vector species.

Other data:

Using census maps, the program should identify areas characterized by large elderly populations or by low socioeconomic status, as clinical disease tends to be more frequent in these locations.

Cx. tarsalis is the primary vector of WEE throughout the range of the virus. Thus, the ecology of WEE is more uniform than with arboviruses that have regionally differing vectors. Differences in disease dynamics are more likely to be linked to north-south seasonal differences in temperature and rainfall. Differing enzootic avian hosts also may alter the dynamics of WEE transmission.

Meteorologic data:

The ideal surveillance program will monitor meteorologic data to estimate the likelihood of increased WEE activity. In California, climatologic data provide an early-season gauge of the likelihood of WEE activity.²⁹⁵ Accumulated degree-days (defined as the sum of daily mean temperature minus the developmental threshold temperature) served as a predictor in the Rocky Mountain region.¹³⁰ Such data are readily obtained from the local weather service.

Vector data:

Surveillance programs will measure relative vector densities based on CO2 -baited light trap or lard can trap collections, and will correlate light trap data with levels of WEE virus activity.²²⁴ Pools of vector species should be submitted for processing for virus isolation at a state or regional laboratory.

Vertebrate host data:

Programs should sample wild and peridomestic passerine birds that are known or suspected to be locally important for enzootic or epizootic transmission.

Other data:

There is some question regarding whether sentinel chickens provide sufficient lead time to react to the appearance of WEE virus. In some areas (e.g., Imperial County, California), high seroconversion rates are observed annually without the appearance of human or equine cases. Passive or active surveillance for equine cases may be useful, but reaction by health agencies must be rapid to have an impact on transmission once equine cases have been diagnosed.