Abstract
After the isolation of West Nile virus (WNV) from humans, mosquitoes, and chickens in 2007, an analysis of animal surveillance involving multiple species (horses, monkeys, sheep, dogs, and birds) used to track WNV transmission from 2006 to 2008 was performed. During this period 13.4% of all the animal samples collected were seropositive by blocking ELISA for WNV. The most complete island-wide sampling was obtained from horses of which 22% were serologically positive and 96% were confirmed as WNV infections by plaque-reduction neutralization test. Our conclusion from this 3-year study is that animal surveillance is an early indicator of WNV activity before the identification of human cases. Additionally, the results indicated that horses have a greater geographical range and should be continued to be used as sentinels for passive surveillance in the tropics.
Introduction
The first human cases of WNV in the Caribbean were reported from the Cayman Islands and the Florida keys in 2001 (CDC 2002). However, most of the evidence for WNV transmission in the Caribbean and Latin America is serological evidence in animals (Elizondo-Quiroga et al. 2005, Deardorff et al. 2006, Komar and Clark 2006). Viral isolates have been obtained from a bird in Mexico in 2003 (Estrada-Franco et al. 2003), a horse in Argentina in 2006 (Morales et al. 2006), and from mosquitoes and humans in Puerto Rico in 2007 (Hunsperger et al. 2009).
In response to the ongoing veterinary and public health threat of WNV in the Caribbean, the Centers for Disease Control and Prevention (CDC) Dengue Branch in San Juan, Puerto Rico, in coordination with the Puerto Rico Department of Health, initiated a WNV animal surveillance program. In this study, we review data from this animal surveillance system collected between November 2006 and June 2008 to determine the best early indicators for WNV transmission in a dengue-endemic tropical country following a documented WNV transmission in 2007 (Barrera et al. 2008, Hunsperger et al. 2009).
Materials and Methods
Blocking ELISA
All samples were tested at the CDC Dengue Branch for WNV using the blocking ELISA as previously described using WNV-specific 3.1112G monoclonal antibody (Chemicon) (Blitvich et al. 2003). Inhibition values of ≥30% were positive in this assay.
Plaque-reduction neutralization test
Plaque-reduction neutralization test (PRNT90) were performed on a subset of the WNV-seropositive horse samples. At least one sample from each municipality with a positive blocking ELISA result was confirmed by PRNT90 against both WNV and St. Louis encephalitis virus (SLEV). The PRNT90 assay was performed as previously described (Russell et al. 1967). ChimeraVax viral strains of WNV and SLEV were donated by Acambis (Cambridge, MA) (Pugachev et al. 2004). Serial two-fold dilutions of serum were made from 1:20 to 1:640. PRNT90 titers were calculated using a back titration of each virus and reported as the reciprocal of the serum dilution yielding ≥90% reduction in the number of plaques. At least a four-fold difference in titer between viruses was used as the infecting virus of the specimen when comparing samples with reactivity to both viruses in the assay.
Results
Sampling and distribution of WNV seropositive animals
To determine the most effective animal surveillance system for Puerto Rico, samples were collected from the widest range of species possible using a variety of collection methods. Sera were collected from primates, wild birds, horses, dogs, swine, chickens, and dead birds throughout Puerto Rico. Horse samples were collected routinely by private veterinarians, the U.S. Department of Agriculture, and Puerto Rico Department of Health from healthy, unvaccinated horses and from sick horses with neurological symptoms. All other animal samples were collected as they became available through partnerships with the Puerto Rico Department of Natural Resources, the Caribbean Primate Investigation Center, dog shelters, and private farmers throughout Puerto Rico. We collected a total of 1,058 samples from 45 of Puerto Rico's 78 municipalities during the time frame of this study (November 2006–June 2008). These results are summarized in Table 1 but do not include the results from the dead birds since this testing is not serological. However, only five dead birds were received during the study period among which one bird was positive (Hunsperger et al. 2009).
Of the samples tested, a total of 13.4% (142 animals) were seropositive by the blocking ELISA. The highest proportion of positives (58%) were observed in the sheep samples; however, this sampling was only performed in two municipalities following the known introduction of WNV in 2007. The second highest percent positivity was observed in the horse samples (22%), which had the most comprehensive geographical sampling as well as the greatest number of samples submitted. Geographical location of all the positive samples during the course of the 3-year surveillance is mapped in Figure 1.
West Nile virus (WNV)–positive samples for surveillance in (
Confirmation of WNV seropositive results with PRNT
In total, 27 presumptive positive horse samples were tested from 20 municipalities for WNV and SLEV using the PRNT90. PRNT90 titers for WNV ranged from 1:20 to 1:640; 96% of the samples had only reactivity to WNV infections and only one sample had a titer of <1:20 suggestive of a negative result. Additionally, one sample had neutralizing activity against SLEV, with a PRNT90 end-point titer of 1:20, although the predominant virus was WNV for this sample. No other samples had neutralizing activity against SLEV (Supplemental Table S1, available online at
Discussion
In this study, passive horse surveillance was the best early indicator of WNV transmission before human disease in a dengue-endemic tropical country. Despite the discovery of WNV-seropositive animals during this study, confirmed human cases of symptomatic WNV had not been reported to the passive island-wide human surveillance system. The first evidence of human infection was from asymptomatic blood donations screened by the American Red Cross in Puerto Rico in 2007 (CDC 2008). Many of the serological tests used to diagnose dengue virus (DENV) are cross-reactive with other Flaviviruses following a secondary Flavivirus infection, complicating differential diagnosis (Martin et al. 2000). Therefore, some human cases of WNV presenting with fever may have been misdiagnosed as dengue infections.
During this study period it appeared that WNV reemerged in 2007 based on seropositivity in the sampling from 2006 (0.9%) to 2007 (16.1%) and 2008 (10.6%). These results suggest that WNV is reintroduced and possibly not sustainable in Puerto Rico. Routine horse surveillance was very effective in determining WNV transmission despite several limitations. Travel history was not always available for the animals surveyed although most of the horses were locally owned. Because individual animals were not followed over the course of this study, it is difficult to assess the timeline for infection in WNV-positive animals. IgM and IgG persistence in the different animal species used in this study varies widely (Johnson et al. 2003, Wagner et al. 2008), complicating the process of establishing a timeline for infection. Only specimens from horses with no vaccine history were used in the study. Many of the locally owned horses were not vaccinated for WNV contributing to the large sampling for the study. If in the future WNV horse vaccination coverage increases, the surveillance strategy will require modification. Other animal surveillance was not as effective due to limited sampling locations or sporadic sample collection, as with sheep, dogs, wild birds, and monkeys.
Surveillance systems in Puerto Rico must be better tailored to respond to ecological differences between the mainland United States and the tropics. For instance, corvid surveillance of WNV was effective in the United States; however, these birds are far less populous in the tropics and therefore less suitable for ongoing surveillance programs (Komar and Clark 2006). The dead bird surveillance was not useful unless specific susceptible birds such as birds of prey were monitored. This study indicated that the most effective forms of animal surveillance in the tropics are active surveillance using sentinel chickens and horses with appropriate selection of ecological site, as well as passive surveillance using dead birds of prey. Surveillance programs in Puerto Rico should continue to monitor the rates of WNV prevalence along with other endemic Flaviviruses to assess outbreak risk.
Footnotes
Acknowledgment
The authors thank Elizabeth Cartozian for her helpful editing comments.
Disclosure Statement
No competing financial interests exist.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
