INTRODUCTION
Avian mortality surveillance has been established as a useful and sensitive early detection system for West Nile virus (WNV; genus Flavivirus, family Flaviviridae) activity.1–3 WNV-positive birds have often provided the earliest indication of WNV activity and frequently precede human WNV case reports.4 The American crow (Corvus brachyrhynchos) has proven to be a sensitive avian sentinel for WNV surveillance in the northeast United States5–7 and southern California,8 but reports documenting the use of crows and other species for surveillance are lacking from other regions. The primary value of testing avian carcasses for WNV infection is the documentation of local WNV activity, which can be quantified for risk determination using a variety of other surveillance systems, including mosquito collection and testing, sentinel animals, or dead bird cluster analyses.9–11
American crow populations have been reduced in some areas because of WNV-related die-offs.12–15 In regions where crows are sparse, whether naturally or because of WNV-related depletion, carcasses of other bird species may be useful WNV surveillance tools. Furthermore, the VecTest WNV Antigen Assay (VecTest; Medical Analysis Systems, Camarillo, CA) has been shown to be a sensitive diagnostic tool in testing oral and cloacal swabs from American crows and some other corvids, but few data exist on the use of VecTest for rapid WNV detection in samples from non-corvids.16–20
To make recommendations on appropriate avian mortality surveillance techniques for WNV in a High Plains state (Colorado), we tested carcasses of multiple bird species to 1) determine the spectrum of avian species affected by WNV in Colorado, 2) compare WNV detection rates in carcasses of the American crow and other birds in Colorado, 3) compare WNV detection assays (virus isolation versus TaqMan reverse transcriptase-polymerase chain reaction [RT-PCR]) and WNV detection in samples (heart versus oral swab) from avian carcasses, and 4) evaluate VecTest applied to oral swabs from corvid and non-corvid bird species.
MATERIALS AND METHODS
Sampling and origin of dead birds.
Dead birds were collected during 2002–2005, mostly from private citizens through local health departments, wildlife rehabilitators, and the Centers for Disease Control and Prevention, Arbovirus Diseases Branch in Fort Collins, CO. In 2002, dead bird collection began in mid-August, after WNV was first detected by the Colorado Department of Public Health and Environment, and collection was carried out until November. In subsequent years, dead birds were collected from April until October. American crows were not included in the study in 2002, and corvids were not included in 2004 because they were tested exclusively by the state health laboratory for their own surveillance purposes.
The vast majority of birds in this study were from Colorado. However, several raptors were submitted that originated from New Mexico, Wyoming, or Nebraska. Although birds from numerous counties were tested in all years, in 2003–2005, the study focused mostly on two northern Colorado counties (Larimer and Weld).
Birds deemed dead ≤ 48 hours were refrigerated and sampled the same day when possible or were placed at −20°C and thawed for sampling within 2–4 days. After thawing, the oropharyngeal cavity was swabbed with a cotton-tipped applicator, and several pieces of heart were removed aseptically with a sterile surgical blade. The heart was deemed a convenient tissue for sampling and has consistently proven to be one of the most reliable tissues for WNV detection in birds.21–26 In all years except 2003, oral swabs were broken off into 1 mL BA-1 diluent (Hank’s M-199 salts, 1% bovine serum albumin, 350 mg/L sodium bicarbonate, 100 units/mL penicillin, 100 mg/L streptomycin, 1 mg/L Fungizone in 0.05 mol/L Tris, pH 7.6); in 2003, oral swabs were submerged in 1 mL of VecTest grinding buffer. In all years, heart samples were placed in cryovials with 1 mL BA-1 diluent and BB pellets. Heart samples (0.5 cm3) were ground in a Qiagen mixer mill (Qiagen, Valencia, CA) at 25 cycles/s for 5 minutes and clarified by centrifugation (12,000 × g for 3 minutes). Oral swab samples were tested by VecTest immediately after sampling, after which oral swab and heart samples were held at −80°C until further testing was performed.
Plaque assay.
Virus isolation was performed on heart homogenates (2002–2005) and oral swabs (2002 and 2004–2005) by Vero cell plaque assay as previously described.23
VecTest.
The VecTest WNV Antigen Assay was performed on oral swabs from avian carcasses submitted in 2002–2005. In 2002 and 2004–2005, a 225-μL aliquot of the oral swab sample was added to 25 μL of grinding B solution (a 10× strength detergent provided by Medical Analysis Systems). In 2003, oral swabs were placed into 1 mL VecTest grinding buffer, and 250 μL was aliquoted for VecTests. The VecTest strip was added to the solution for 15 minutes at room temperature and dried on a paper towel for ~5 minutes. Strips were independently read by three people, and a strip was interpreted positive if at least two of three readers recorded the presence of a line, scored between 1 (lightest band) and 3 (darkest band).
TaqMan RT-PCR.
TaqMan RT-PCR was performed on oral swabs and heart homogenates. RT-PCR methods for detection of WNV RNA followed those of Lanciotti and others27 except for use of the Viral RNA Minikit (Qiagen) for RNA extraction and use of the Bio-Rad Icycler IQ Real-time Detection system (Bio-Rad, Hercules, CA) for cDNA amplification. A Ct value of 37 or less was considered positive for target sequence amplification. Samples were screened with one pair of primers and positives confirmed with a second pair of primers.27
Statistical analyses.
We calculated 95% confidence intervals (CIs for carcass infection proportions using the Wilson score method [S-PLUS 6.1 Professional software; Insightful, Seattle, WA]). To evaluate whether species or year data could be combined, we used log-linear analysis (Statistica for Windows 99 edition; StatsSoft, Tulsa, OK) to test for partial or mutual independence among the variables “species” and “year.” A WNV-positive sample was defined as a sample that tested positive by either plaque assay or TaqMan RT-PCR, because both of these tests have comparable sensitivity and have become standard diagnostic tests for WNV detection.27 We used Pearson χ2 test to compare the proportion positive of each species for which N > 5 and each species-group against all other species or species-groups, respectively. When one cell of the 2 × 2 table was < 5, we used a right-tailed Fisher exact test instead. To determine significant differences for species comparisons, we used Bonferroni-adjusted α = 0.0031 (for 16 comparisons). For species-group comparisons (N ≥ 80), we used Bonferroni-adjusted α = 0.017 (for three comparisons). We used Pearson χ2 test to compare test agreement values (kappa statistic [κ]) for plaque assay and TaqMan RT-PCR of both oral swab and heart and for κ values for oral swab and heart samples tested by both assays. Sensitivity, specificity, false positivity, and κ values of VecTest were derived by comparison with standard test (Vero plaque assay or RT-PCR of oral swab or heart) results. Sensitivity was defined as the proportion of true positives that was positive by VecTest; specificity was the proportion of true negatives that was negative by VecTest; false positivity was the proportion of true negatives that was positive by VecTest; and test agreement was the proportion of true positives and negatives with matching results using VecTest. We compared VecTest sensitivity for each species against that of all others combined using a right-tailed Fisher exact test, and a Bonferroni-adjusted α = 0.0038 (for 13 comparisons).
Target Species Index.
We conceived the Target Species Index (TSI) to evaluate which species of bird was most efficient for detecting WNV infections. We define the TSI as a relative numerical expression that considers abundance and infection rate of carcasses for each species tested. We developed a simple mathematical equation describing the TSI as follows:
where T represents the index value, a is a proportion (ranging from 0 to 1) describing the relative abundance of carcasses of any one species compared with the total number of carcasses tested, and f is the proportion (ranging from 0 to 1) of species-specific carcasses that tested positive. To make the value for T relative to unity, we divided each species-specific T value by the lowest non-zero T value (Tmin).
For the case where the diagnostic methods are experimental, we conceived the test-specific TSI and modified the equation for T as follows:
where Texp is the test-specific index value, and s and κ are test-specific sensitivity and agreement parameters, respectively, derived from comparison with the standard methodology. In this study, Texp was calculated using experimental data derived from VecTest.
Differences among index values were considered significant if their 95% CIs failed to overlap.
RESULTS
A total of 1,549 dead birds (of 104 species) were tested for WNV; dead birds of 42 species tested positive (Table 1). The frequency of infection of the most common species tested (American crow [N = 72], American kestrel [Falco sparverius, N = 76], American robin [Turdus migratorius, N = 139], common grackle [Quiscalus quiscula, N = 143], European starling [Sturnus vulgaris, N = 57], great horned owl [Bubo virginianus, N = 82], house finch [Carpodacus mexicanus, N = 79], house sparrow [Passer domesticus, N = 152], mourning dove [Zenaida macroura, N = 110], northern flicker [Colaptes auratus, N = 41], rock pigeon [Columba livia, N = 63], and red-tailed hawk [Buteo jamaicensis, N = 46]) varied significantly by both species and year (P < 0.05).
To study differences in WNV infection by species, we focused on carcasses tested during peak transmission months (June–September) during the 2003 epizootic year (Table 2). Range of percents positive for species with N > 5 was 67–91% for corvids, 14–85% for non-corvid passerines, and 23–45% for raptors. When compared against all other species, American crow (91%) and house finch (85%) had significantly greater proportions of positives (P < 0.003). We also evaluated differences in proportions positive among species-groups including corvids, non-corvid passerines, and raptors (included both falconiforms and strigiforms), and only corvids were significantly more likely to be positive (P < 0.01). To determine which species or species-groups should be targeted for surveillance in our study area (northern Colorado), we calculated TSI values (Table 2). American crow scored significantly higher than all other species evaluated. For species-groups, corvids and raptors scored highest, followed by non-corvid passerines and then all others. Grouping all birds together yielded the highest target species index score of 180 (95% CI, 160–200).
In 2002, 227 dead birds of 52 species were tested, and 22 birds tested positive (14 species) for WNV by at least one standard technique. One of the 22 positives was a raptor (4.5%) and 15/22 (68.2%) were non-corvid passerines. The earliest positives were a mourning dove and house finch collected on August 30, and the latest was a Brewer’s blackbird (Euphagus cyanocephalus) collected on October 11, 2002.
In 2003, 417 dead birds of 58 species were tested, and 176 birds tested positive for WNV (34 species). The majority of positives were corvids (40.3%; 71/176), whereas 36.4% (64/176) were raptors and 17.0% (30/176) were non-corvid passerines. Of corvids tested, 84.5% (71/84) were positive, whereas 34.8% (64/184) of raptors and 34.1% (30/88) of non-corvid passerines tested positive. The earliest positive was an American kestrel collected on June 3, and the latest was a sharp-shinned hawk (Accipiter striatus) on October 9, 2003.
In 2004, 356 dead birds of 49 species were tested, and 11 birds tested positive for WNV (10 species). Seven of the 11 (63.6%) positive dead birds were raptors, and the remainder was non-corvid passerines. Approximately 2% (4/203) of non-corvid passerines tested positive, whereas 6.3% (5/80) of raptors tested positive. The earliest positives were a red-tailed hawk and great horned owl collected on July 28 and the latest an American kestrel on September 19, 2004.
In 2005, 549 dead birds were tested of 54 species, and 33 birds tested positive for WNV, including 15 species. Thirteen of the 33 (39.4%) positive dead birds were raptors, 9/33 (27.3%) were corvids, and 7/33 (21.2%) were non-corvid passerines. The earliest positive was an American kestrel sampled on July 31, and the latest an American crow on September 23, 2005.
We compared plaque assay and TaqMan RT-PCR using test agreement, κ, to ensure that our use of virus isolation and RNA detection as standard tests was appropriate. The κ for hearts tested by both assays (N = 1,462) was 0.97 and for oral swabs (N = 1,085) was 0.99. No significant differences were detected between species or years, so tests were combined across species and years for this analysis. We also compared heart and oral swabs as samples for WNV detection; the κ for sample type (N = 1,104) was 0.97. No significant difference in detection rate was observed between sample types (Pearson χ2 test; P = 0.15).
VecTest results are presented for species-groups (Table 3) and individual species for which the number of true positives was ≥ 5 (Table 4). Sensitivity values among species-groups varied and were highest for corvids (0.80). False positivity was generally low (range, 0.03–0.12 among species-groups), and test agreement was high (0.85–0.87) among species-groups. Species-specific sensitivities were highest for the corvid species (range, 0.69–0.83), as well as house sparrow (0.60) and house finch (0.62), but were very low for most raptors (0.00–0.33, excluding northern saw-whet owl; Aegolius acadicus), American robin, and mourning dove (0.00 each). Test agreement was good to excellent for most species (> 0.75). The test-specific TSI values derived from the VecTest results confirmed that American crows were significantly more valuable for WNV surveillance than the other species tested by VecTest (Table 4).
DISCUSSION
Avian mortality–based surveillance for WNV is valuable for passive early detection of WNV activity. WNV avian mortality surveillance has advantages (e.g. potential for early detection, passive data collection) and disadvantages (e.g. reliance on outside sources for reporting and submitting dead birds with associated biases and difficulty of coordinating dead bird reports and collection).9,14,28,29 Although much WNV surveillance literature has focused on the American crow,1,2,5,7,30 percentages of positive crows versus other bird species in New York indicated that testing carcasses of all bird species was important.31 Common urban or suburban species (e.g. house sparrow, house finch, and northern cardinal [Cardinalis cardinalis]) were useful for WNV avian mortality surveillance in areas of New York State where the density of corvid populations was low.19 Our results confirm that the American crow is more useful than other species for WNV detection in a region of intense WNV transmission risk (Colorado),32 but also show the use of testing many species. For example, we ranked the most useful species for WNV avian mortality surveillance using a novel TSI and found that after American crow, several other species ranked highly using this index: American kestrel, black-billed magpie (Pica hudsonia), house finch, red-tailed hawk, great horned owl, house sparrow, and blue jay (Cyanocitta cristata; Table 2). Using the TSI highlights the use of testing multiple species. TSI values for species-groups were much higher than individual species with the possible exception of American crow. The TSI score for all birds combined was 3- to 4-fold greater than that for American crow, which indicates that testing all birds would be the most sensitive WNV avian mortality surveillance system in Colorado.
As groups, raptors and corvids scored similar TSI values in our study, in part because of coordinated submissions of carcasses by raptor rehabilitators. In New York, 57% of American kestrels, 43% of red-tailed hawks, and 19% of great horned owls were WNV positive,31 compared with 45%, 45%, and 28%, respectively, in our study. Applying the TSI to the New York data, however, we found that these species scored lower relative to American crow than in Colorado (data not shown). Many factors contribute to this difference, including selection biases (e.g. some species such as crows are more likely to be submitted for testing than others) and differences in species distributions between the two states. An important limitation to the TSI is season, because birds that die when WNV is inactive would bias the index values. We avoided this bias by limiting the carcasses used for these calculations to those submitted during peak WNV transmission season. TSI values could change temporally and spatially.
Collaboration with wildlife rehabilitators and agencies can help maximize the usefulness of avian mortality surveillance in a given region. Such efforts are mutually beneficial because along with fulfilling surveillance purposes, testing avian carcasses provides diagnostic information to rehabilitators while increasing the understanding of how WNV affects different bird species. Live raptors admitted to rehabilitation facilities have also shown promise for early detection of WNV.33
Although data collected from our testing of dead birds did not specifically evaluate the effects of WNV on avian populations of Colorado, they did point to some concerns. We report WNV detections in carcasses of 42 avian species of 19 families, which included rare reports of WNV infection in species such as Clark’s grebe (Aechmophorus clarkii) and flammulated owl (Otus flammeolus). We did not detect WNV-infected carcasses of locally threatened species such as bald eagle (Haliaeetus leucocephalus) or burrowing owl (Athene cunicularia). However, we feel that the probability of these being submitted for testing would be very low given their reduced populations. Our data do show potential WNV-associated population-reducing effects on corvids and raptors in Colorado, although we cannot quantify these reductions. Some species for which we tested few carcasses, but may have been substantially affected by WNV, would include blackcapped chickadee (Poecile atricapilla; 3 positive/4 tested), northern goshawk (Accipiter gentilis; 2/2), and ferruginous hawk (Buteo regalis; 4/7). Although these species will not be useful targets for surveillance purposes because of low numbers, they may be very efficient indicators of WNV activity. Other vulnerable species in Colorado would include greater sage-grouse (Centrocercus urophasianus),34 as well as endangered species such as Gunnison sage-grouse (Centrocercus minimus), piping plover (Charadrius melodus), and least tern (Sterna antillarum).35 Our data do permit us to weigh WNV as a cause of mortality against other causes within species. For example, during the 2003 epizootic, WNV was frequently associated with crow deaths (91% of carcasses tested positive) and house finch deaths (85%). This changed in 2005 (a year with less transmission), in which WNV was detected in 60% of American crow carcasses and only 6% of house finch carcasses. Causes of death among the WNV-negative carcasses in this study remain unknown, but numerous possibilities exist; infectious diseases that may cause bird mortality in Colorado include avian pox, salmonellosis, ornithosis, trichomoniasis, and aspergillosis. Additionally, we detected several as yet unidentified viruses in some carcasses.
Decreased detection of WNV in avian carcasses after 2003 corresponded with a decrease in human WNV cases and deaths reported in Colorado within the same time frame.36 The months of highest WNV activity in northern Colorado from 2002 to 2005 included August and September in most years, whereas in the epizootic year of 2003, July and August were the months with highest proportions of WNV positive dead birds (results not shown). We observed a low prevalence of positive birds in early and late months of the WNV transmission season in Colorado, so we recommend increasing the avian carcass sample size if an emphasis on early and late detection is desired. In addition, the earliest and latest species detected varied among years, and often included raptors. Therefore, we also recommend testing multiple bird species for earliest detection of WNV in a given season. Our TSI, calculated from locally derived data, can be used to determine which species should be included for greatest efficiency of detection. Local resources will determine how many additional species to include. For example, from our data and assuming limited resources, we might recommend expanding to seven other species in addition to American crow, using a TSI value of 6.0 as a cut-off (see Table 2). This expansion would add black-billed magpie, blue jay, house finch, house sparrow, great horned owl, American kestrel, and red-tailed hawk and more than doubles probability of detection (group relative TSI for 2003 June–September = 4.5; 95% CI, 4.0–5.0) relative to the rejected species (group relative TSI for 2003 June–September = 1.0; 95% CI, 0.7–1.4), including commonly reported carcasses such as American Robin, common grackle, European starling, rock pigeon, and mourning dove. We recognize that expanding avian mortality surveillance to include multiple species will increase the cost and overall burden of the surveillance program on the public health system. However, we believe that the passive nature of this surveillance makes it more cost effective than other systems, such as virus testing of mosquito pools during periods of very low transmission (e.g., early and late in the transmission season).
Our data provided an opportunity to re-evaluate sample types and assay methods for WNV detection among avian carcasses. We corroborated previous findings by other authors that oral swabs versus hearts are equally valuable for WNV detection, as are TaqMan versus plaque assay.20,27,37 However, we recognize that testing multiple specimen types per bird carcass, as well as using a variety of assays, increases detection rates.
Rapid antigen detection assays may improve the usefulness of avian mortality surveillance for WNV.20,38 For this reason, we evaluated the sensitivity, false positivity, and test agreement of VecTest compared with standard virus detection methods and found that VecTest sensitivities varied among species. VecTest sensitivity of American crow oral swabs (83%) in our study is comparable with previous field studies; as in other studies, VecTest sensitivity in the American crow is higher than that of other species, especially non-corvids.16,19 Stone and others19 reported that sensitivity of VecTest applied to oral swabs was favorable in some bird species, such as American crow (0.67), blue jay (0.78), and house sparrow (0.82), whereas northern cardinal, common grackle, and house finch had sensitivities ≥ 0.50 but were limited by small sample sizes. Lower sensitivities (≤ 0.20) occurred for mourning dove, fish crow (Corvus ossifragus), American robin, and raptors (seven species combined), which concurred with our observations (excluding fish crow, which we did not test). We observed that VecTest was insensitive for raptors and non-corvid passerines (0.30 and 0.46, respectively). However, house finch and house sparrow sensitivities (0.63 and 0.60, respectively) may be considered useful, especially given the relative abundance of these peridomestic species. When we combined sensitivity and test agreement results, we observed that corvids, house finch, and house sparrow all scored higher than other species (Table 4). We also modified our TSI for application with experimental diagnostic tests such as VecTest. The modified index was termed test-specific TSI. Applying this index, we found that VecTest was useful for WNV detection in some corvids (American crow and black-billed magpie) and some other passerines (house finch and house sparrow) but should not be used with raptors, except for American kestrels (Table 4). Both the TSI and the test-specific TSI should have use for other disease surveillance programs in which multiple species may be targets for surveillance, such as avian influenza and plague.
We report VecTest false positivity in our study (~10% overall). All of the false positives we observed were considered weak positives (scored as 1 or 0 by all three observers). A more stringent definition of a VecTest positive result, such as a score of ≥ 2 by at least one observer, would have eliminated all false positivity in our hands but at the expense of sensitivity. Stone and others19 observed false-positive VecTest results, mostly in gray catbirds (Dumetella carolinensis) and green herons (Butorides virescens), but otherwise found high specificity for the VecTest assay applied to oral swabs of bird carcasses (0.98) compared with a standard test (WNV RNA detection in kidney or brain). We observed similar specificity for VecTest in our study (~0.90). Yaremych and others17 evaluated VecTest usefulness in American crows and detected some false positives; they suggested that more extensive testing of VecTest in American crows is needed and a confirmatory test be used in conjunction. We also recommend confirmatory testing by a second technique, especially for early detection purposes.
In conclusion, avian carcass testing complements quantitative WNV surveillance methods as an effective means of detecting early WNV activity. We believe that avian mortality surveillance in the High Plains states such as Colorado should continue focusing on corvids and raptors, although inclusion of multiple species would enhance early and late detection and detection in regions of low corvid density. Eidson and others6 recommended “unrestricted testing” of bird species to provide earliest warning of WNV activity in an area and encouraged continued surveillance of viral activity. Wildlife rehabilitation centers can greatly enhance and simplify surveillance efforts in some areas by concentrating many samples in few locales. We compared a novel rapid test with standard virus detection methods for avian carcasses and found that VecTest is especially useful when applied to oral swabs from corvid carcasses but is less sensitive when used with other species.
Proportion of WNV–positive carcasses (N > 30) for common bird species in Colorado by year*†
| 2002 | 2003 | 2004 | 2005 | |||||
|---|---|---|---|---|---|---|---|---|
| Species | N | Proportion positive (CI) | N | Proportion positive (CI) | N | Proportion positive (CI) | N | Proportion positive (CI) |
| * All other species (with N ≤ 30) that tested positive for WNV: American goldfinch (Carduelis tristis), American white pelican (Pelecanus erythrorhynchos), black-capped chickadee (Poecile atricapilla), blue jay (Cyanocitta cristata), Brewer’s blackbird (Euphagus cyanocephalus), Canada goose (Branta canadensis), cedar waxwing (Bombycilla cedrorum), Clark’s grebe (Aechmophorus clarkii), cliff swallow (Petrochelidon pyrrhonota), Cooper’s hawk (Accipiter cooperii), eastern screech-owl (Megascops asio), ferruginous hawk (Buteo regalis), flammulated owl (Otus flammeolus), golden eagle (Aquila chrysaetos), MacGillivray’s warbler (Oporornis tolmiei), mallard (Anas platyrhynchos), merlin (Falco columbarius), northern goshawk (Accipiter gentilis), northern harrier (Circus cyaneus), northern rough-winged swallow (Stelgidopteryx serripennis), northern saw-whet owl (Aegolius acadicus), peregrine falcon (Falco peregrinus), prairie falcon (Falco mexicanus), red-naped sapsucker (Sphyrapicus nuchalis), rough-legged hawk (Buteo lagopus), sharp-shinned hawk (Accipiter striatus), Swainson’s hawk (Buteo swainsoni), western tanager (Piranga ludoviciana), western scrub jay (Aphelocoma californica), Wilson’s warbler (Wilsonia pusilla). | ||||||||
| † All other species tested that did not test positive for WNV (in addition to northern flicker, listed in the table): American coot (Fulica americana), bald eagle (Haliaeetus leucocephalus), Baltimore oriole (Icterus galbula), barn swallow (Hirundo rustica), belted kingfisher (Ceryle alcyon), black-crowned night heron (Nycticorax nycticorax), black-headed grosbeak (Pheucticus melanocephalus), Brewer’s sparrow (Spizella breweri), broad-tailed hummingbird (Selasphorus platycercus), brown-headed cowbird (Molothrus ater), budgerigar (Melopsittacus undulatus), Bullock’s oriole (Icterus bullockii), burrowing owl (Athene cunicularia), calliope hummingbird (Stellula calliope), common canary (Serinus canaria), chipping sparrow (Spizella passerina), cockatiel (Nymphicus hollandicus), cockatoo (Cacatua sp.), barn owl (Tyto alba), common nighthawk (Chordeiles minor), common poorwill (Phalaenoptilus nuttallii), Wilson’s snipe (Gallinago delicata), domestic chicken (Gallus gallus), downy woodpecker (Picoides pubescens), European collared-dove (Streptopelia decaocto), Gambel’s quail (Callipepla gambelii), great blue heron (Ardea herodias), green-tailed towhee (Pipilo chlorurus), grey catbird (Dumetella carolinensis), gyrfalcon (Falco rusticolus), hermit thrush (Catharus guttatus), house wren (Troglodytes aedon), killdeer (Charadrius vociferus), lark bunting (Calamospiza melanocorys), lark sparrow (Chondestes grammacus), lesser goldfinch (Carduelis psaltria), loggerhead shrike (Lanius ludovicianus), long-eared owl (Asio otus), mourning warbler (Oporornis philadelphia), northern pygmy owl (Glaucidium gnoma), osprey (Pandion haliaetus), pine siskin (Carduelis pinus), red-breasted nuthatch (Sitta canadensis), red-winged blackbird (Agelaius phoeniceus), ring-billed gull (Larus delawarensis), ruby-crowned kinglet (Regulus calendula), rufous hummingbird (Selasphorus rufus), sage thrasher (Oreoscoptes montanus), short-eared owl (Asio flammeus), spotted towhee (Pipilo maculatus), Swainson’s thrush (Catharus ustulatus), tree swallow (Tachycineta bicolor), wild turkey (Meleagris gallopavo), turkey vulture (Cathartes aura), violet green swallow (Tachycineta thalassina), western bluebird (Sialia mexicana), western kingbird (Tyrannus verticalis), western meadowlark (Sturnella neglecta), western screech-owl (Megascops kennicottii), western wood-pewee (Contopus sordidulus), wood duck (Aix sponsa). | ||||||||
| American crow/Corvus brachyrhynchos | 0 | – | 56 | 0.91 (0.8–1.0) | 2 | 0.0 (0.0–0.7) | 14 | 0.6 (0.3–0.8) |
| American kestrel/Falco sparverius | 2 | 0.0 (0.0–0.7) | 42 | 0.43 (0.29–0.58) | 20 | 0.05 (0.1–0.2) | 12 | 0.25 (0.09–0.5) |
| American robin/Turdus migratorius | 7 | 0.0 (0.0–0.4) | 12 | 0.33 (0.1–0.6) | 45 | 0.0 (0.0–0.08) | 75 | 0.01 (0.002–0.07) |
| Black-billed magpie/Pica hudsonia | 10 | 0.0 (0.0–0.3) | 17 | 0.76 (0.5–0.9) | 0 | – | 5 | 0.0 (0.0–0.43) |
| Common grackle/Quiscalus quiscula | 23 | 0.13 (0.05–0.3) | 10 | 0.5 (0.2–0.8) | 45 | 0.02 (0.004–0.1) | 65 | 0.03 (0.01–0.1) |
| European starling/Sturnus vulgaris | 24 | 0.04 (0.01–0.2) | 7 | 0.14 (0.03–0.5) | 5 | 0.0 (0.0–0.4) | 21 | 0.05 (0.01–0.2) |
| Great horned owl/Bubo virginianus | 5 | 0.0 (0.0–0.4) | 45 | 0.22 (0.13–0.36) | 14 | 0.07 (0.01–0.3) | 18 | 0.06 (0.01–0.3) |
| House finch/Carpodacus mexicanus | 13 | 0.23 (0.1–0.5) | 13 | 0.85 (0.6–1.0) | 22 | 0.0 (0–0.15) | 31 | 0.06 (0.02–0.2) |
| House sparrow/Passer domesticus | 16 | 0.06 (0.01–0.3) | 30 | 0.23 (0.12–0.41) | 47 | 0.0 (0.0–0.8) | 59 | 0.02 (0.003–0.09) |
| Mourning dove/Zenaida macroura | 20 | 0.1 (0.03–0.3) | 15 | 0.13 (.04–0.4) | 30 | 0.0 (0.0–0.1) | 45 | 0.07 (0.02–0.2) |
| Northern flicker/Colaptes auratus | 9 | 0.0 (0–0.3) | 2 | 0.0 (0–0.7) | 15 | 0.0 (0.0–0.2) | 15 | 0.0 (0.0–0.2) |
| Rock pigeon/Columba livia | 4 | 0.5 (0.2–0.9) | 14 | 0.07 (0.01–0.3) | 14 | 0.0 (0.0–0.2) | 31 | 0.03 (0.006–0.16) |
| Red-tailed hawk/Buteo jamaicensis | 2 | 0.5 (0.1–0.9) | 23 | 0.43 (0.26–0.63) | 9 | 0.22 (0.06–0.6) | 12 | 0.25 (0.09–0.5) |
Proportion of carcasses, by bird species (N > 5) and species-group, testing positive for WNV during June–September 2003, and calculation of a TSI for WNV avian mortality surveillance
| Species | N | Proportion of all carcasses (a) | Proportion positive (f) (CI) | Relative TSI [T = (a × f)/Tminimum] (CI) |
|---|---|---|---|---|
| * Values considered significantly greater compared to that of all other species (P < 0.0031) or species-groups (P < 0.017). | ||||
| † Includes western scrub-jay (one pos/one tested). | ||||
| ‡ Includes: American goldfinch (1 pos/1 tested), barn swallow (0/1), brewer’s blackbird (1/1), house wren (0/3), mourning warbler (0/1), red-winged blackbird (0/5), sage thrasher (0/1), western meadowlark (0/1), Wilson’s warbler (0/1). | ||||
| § Includes bald eagle (0 pos/2 tested), barn owl (0/3), eastern screech-owl (1/1), ferruginous hawk (1/3), flammulated owl (1/1), golden eagle (2/5), gyrfalcon (0/1), long-eared owl (0/3), merlin (1/2), northern goshawk (2/2), northern harrier (1/1), northern pygmy-owl (0/2), northern saw-whet owl (5/5), peregrine falcon (1/2), prairie falcon (1/3), rough-legged hawk (1/1), sharp-shinned hawk (1/3), short-eared owl (0/1), and western screech-owl (0/2). | ||||
| ¶ Includes American coot (0 pos/1 tested), belted kingfisher (0/3), black-crowned night heron (0/1), calliope hummingbird (0/1), Canada goose (2/5), Clark’s grebe (1/1), common nighthawk (0/2), common poorwill (0/2), domestic chicken (0/1), downy woodpecker (0/1), mallard (3/4), northern flicker (0/2), turkey vulture (0/2). | ||||
| Corvids | 83 | 0.206 | 0.855 (0.764–0.915)* | 72 (65–78) |
| American crow/Corvus brachyrhynchos | 56 | 0.139 | 0.911 (0.807–0.961)* | 52 (46–55) |
| Black-billed magpie/Pica hudsonia | 17 | 0.042 | 0.765 (0.527–0.904) | 13 (9.1–16) |
| Blue jay/Cyanocitta cristata | 9 | 0.022 | 0.667 (0.354–0.879) | 6.0 (3.2–8.0) |
| Other corvids† | 1 | – | – | – |
| Non-corvid passerines | 87 | 0.216 | 0.345 (0.253–0.449) | 31 (22–40) |
| Americn robin/Turdus migratorius | 12 | 0.03 | 0.333 (0.138–0.609) | 4.1 (1.7–7.5) |
| Common grackle/Quiscalus quiscula | 10 | 0.025 | 0.500 (0.237–0.763) | 5.1 (2.4–7.8) |
| European starling/Sturnus vulgaris | 7 | 0.017 | 0.143 (0.026–0.513) | 1.0 (0.18–3.6) |
| House finch/Carpodacus mexicanus | 13 | 0.032 | 0.846 (0.578–0.957)* | 11 (7.6–13) |
| House sparrow/Passer domesticus | 30 | 0.075 | 0.241 (0.118–0.409) | 7.4 (3.6–13) |
| Other non-corvid passerines‡ | 15 | – | – | – |
| Raptors | 171 | 0.425 | 0.368 (0.300–0.443) | 64 (52–77) |
| American kestrel/Falco sparverius | 40 | 0.100 | 0.450 (0.307–0.602) | 19 (13–25) |
| Cooper’s hawk/Accipiter cooperii | 6 | 0.015 | 0.333 (0.097–0.700) | 2.1 (0.60–4.3) |
| Great horned owl/Bubo virginianus | 43 | 0.107 | 0.233 (0.132–0.377) | 10 (5.8–17) |
| Red-tailed hawk/Buteo jamaicensis | 22 | 0.055 | 0.455 (0.269–0.653) | 10 (6.1–15) |
| Swainson’s hawk/Buteo swainsoni | 17 | 0.042 | 0.294 (0.133–0.531) | 5.1 (2.3–9.2) |
| Other raptors§ | 43 | – | – | – |
| Other | 61 | 0.152 | 0.180 (0.104–0.295) | 11.3 (6.5–18) |
| Am. white pelican/Pelecanus erythrorhynchos | 6 | 0.015 | 0.333 (0.097–0.700) | 2.1 (0.60–4.3) |
| Mourning dove/Zenaida macroura | 15 | 0.037 | 0.133 (0.037–0.379) | 2.0 (0.56–5.8) |
| Rock pigeon/Columba livia | 14 | 0.035 | 0.142 (0.013–0.315) | 2.0 (0.19–4.5) |
| All other species¶ | 26 | – | – | – |
VecTest sensitivities, false positive rates, and test agreement for corvid, non-corvid passerines (non-corvids), and raptor species-groups
| No. VecTest true positives*/standard test positives (sensitivity, CI) | No. VecTest false positives/standard test negatives (false positive rate, CI) | κ | |
|---|---|---|---|
| * VecTest true positives were among those that also tested positive by a standard technique (see methods). | |||
| † Eighteen raptor oral swabs were not tested by VecTest. | |||
| Corvids | 65/81 (0.80, 0.70–0.87) | 6/69 (0.09, 0.04–0.18) | 0.85 |
| Non-corvids | 26/57 (0.46, 0.33–0.58) | 75/648 (0.12, 0.09–0.14) | 0.85 |
| Raptors | 20/66† (0.30, 0.21–0.42) | 12/387 (0.03, 0.02–0.05) | 0.87 |
VecTest sensitivity and test agreement by species for which at least five true WNV positives were detected (2002–2005)
| Species | No. true positives* | Subset VecTest positive | VecTest sensitivity (s) (CI) | κ | Relative test–specific TSI† [Texp = (a × f × s × κ)/Texp minimum] (CI) |
|---|---|---|---|---|---|
| * Positives determined by RT-PCR or plaque assay of heart or oral swab. | |||||
| † Data for a, relative abundance of carcasses, and f, proportion of carcasses positive by the standard test, are not presented in this table. | |||||
| ‡ Includes only carcasses for which oral swabs were tested by VecTest. | |||||
| § All were juveniles from the same nest. | |||||
| American crow/Corvus brachyrhynchos | 59 | 49 | 0.83 (0.72–0.91) | 0.86 | 43 (32–51) |
| American kestrel/Falco sparverius | 12‡ | 4 | 0.33 (0.14–0.61) | 0.79 | 5.9 (1.1–13) |
| American robin/Turdus migratorius | 5 | 0 | 0.00 (0.00–0.43) | 0.86 | 0 (0–4.4) |
| Black-billed magpie/Pica hudsonia | 13 | 9 | 0.69 (0.42–0.87) | 0.68 | 6.4 (2.5–11) |
| Blue jay/Cyanocitta cristata | 7 | 5 | 0.71 (0.36–0.92) | 0.81 | 4.2 (1.0–10) |
| Common grackle/Quiscalus quiscula | 11 | 5 | 0.46 (0.21–0.72) | 0.84 | 4.3 (1.1–12) |
| Great horned owl/Bubo virginianus | 12 | 3 | 0.25 (0.09–0.53) | 0.87 | 2.7 (0.57–9.6) |
| House finch/Carpodacus mexicanus | 16 | 10 | 0.62 (0.39–0.82) | 0.80 | 8.2 (3.2–16) |
| House sparrow/Passer domesticus | 9 | 6 | 0.67 (0.35–0.88) | 0.90 | 5.5 (1.6–14) |
| Mourning dove/Zenaida macroura | 7 | 0 | 0.00 (0.00–0.35) | 0.85 | 0 (0–4.3) |
| Northern saw-whet owl/Aegolius acadicus | 5§ | 5 | 1.00 (0.57–1.00) | 1.00 | 5.3 (1.6–6.2) |
| Red-tailed hawk/Buteo jamaicensis | 12‡ | 0 | 0.00 (0.00–0.24) | 0.71 | 0 (0–3.3) |
| Swainson’s hawk/Buteo swainsoni | 8‡ | 1 | 0.12 (0.02–0.47) | 0.80 | 1.0 (0.07–5.4) |
| Other species | 48 | 19 | – | – | – |
Address correspondence to Nicole M. Nemeth, Department of Microbiology, Immunology and Pathology, Colorado State University, Fort Collins, CO 80523. E-mail: nnemeth@colostate.edu
Authors’ addresses: Nicole M. Nemeth, Department of Microbiology, Immunology and Pathology, Colorado State University, Fort Collins, CO 80523, Telephone: 970-491-8165, Fax: 970-491-3557, E-mail: nnemeth@colostate.edu. Susan Beckett, Eric Edwards, Kaci Klenk, and Nicholas Komar, Arbovirus Diseases Branch, Centers for Disease Control and Prevention, Fort Collins, CO 80521, Telephone: 970-221-6400, Fax: 970-221-6476, E-mails: susanmb79@yahoo.com, ede2@cdc.gov, Kaci.Klenk@aphis.usda.gov, and nkomar@cdc.gov.
Acknowledgments: The authors thank the volunteers and staff at the Rocky Mountain Raptor Program (especially G. Kratz, J. Scherpelz, C. Avila, L. Winta, R. Bates, M. Grove, and C. Phillips), Larimer County Humane Society (especially B. Nightwalker, J. Plunkett, S. Breuilly, and J. Crick), Birds of Prey Foundation (S. Ueblacker), Raptor Education Foundation (A. Price and P. Reshetniak), Laramie Raptor Refuge (C. Symchych and N. Prior), Colorado Division of Wildlife (especially L. Baeten), Greenwood Wildlife Rehabilitation Sanctuary, The Wildlife Center, Inc. in New Mexico, and the many private citizens who contributed dead birds to the study. We also thank J. Velez, K. Burkhalter, and N. Panella for laboratory assistance, K. Huyvaert for assistance with statistical analyses, and R. Bowen for his support.
Financial support: This research was funded by the Centers for Disease Control and Prevention.
REFERENCES
- 1↑
Julian KG, Eidson M, Kipp AM, Weiss E, Petersen LR, Miller JR, Hinten SR, Marfin AA, 2002. Early season crow mortality as a sentinel for West Nile virus disease in humans, northeastern United States. Vector Borne Zoonotic Dis 2 :145–155.
- 2↑
Guptill SC, Julian KG, Campbell GL, Price SD, Marfin AA, 2003. Early-season avian deaths from West Nile virus as warnings of human infection. Emerg Infect Dis 9 :483–484.
- 3↑
Mostashari F, Kulldorff M, Hartman JJ, Miller JR, Kulasekera V, 2003. Dead bird clusters as an early warning system for West Nile virus activity. Emerg Infect Dis 9 :641–646.
- 4↑
O’Leary DR, Marfin AA, Montgomery SP, Kipp AM, Lehman JA, Biggerstaff BJ, Elko VL, Collins PD, Jones JE, Campbell GL, 2004. The epidemic of West Nile virus in the United States, 2002. Vector Borne Zoonotic Dis 4 :61–70.
- 5↑
Eidson M, Komar N, Sorhage F, Nelson R, Talbot T, Mostashari F, McLean R, West Nile virus Avian Mortality Surveillance Group, 2001. Crow deaths as a sentinel surveillance system for West Nile virus in the northeastern United States, 1999. Emerg Infect Dis 7 :615–620.
- 6↑
Eidson M, Kramer L, Stone W, Hagiwara Y, 2001. Avian mortality surveillance as an early warning system for West Nile virus. Emerg Infect Dis 7 :631–635.
- 7↑
Johnson GD, Eidson M, Schmit K, Ellis A, Kulldorff M, 2006. Geographic prediction of human onset of West Nile virus using dead crow clusters: an evaluation of year 2002 data in New York State. Am J Epidemiol 163 :171–180.
- 8↑
Reisen WK, Barker CM, Carney R, Lothrop HD, Wheeler SS, Wilson JL, Madon MB, Takahashi R, Carroll B, Garcia S, Fang Y, Shafii M, Kahl N, Ashtari S, Kramer V, Glaser C, Jean C, 2006. Role of corvids in epidemiology of West Nile virus in southern California. J Med Entomol 43 :356–367.
- 9↑
Eidson M, 2001. “Neon needles” in a haystack: the advantages of passive surveillance for West Nile virus. Ann N Y Acad Sci 951 :38–53.
- 10
Gubler DJ, Campbell GL, Nasci R, Komar N, Petersen L, Roehrig JT, 2000. West Nile virus in the United States: guidelines for detection, prevention, and control. Viral Immunol 13 :469–475.
- 11↑
Nasci RS, Komar N, Marfin AA, Ludwig GV, Kramer LD, Daniels TJ, Falco RC, Campbell SR, Brookes K, Gottfried KL, Burkhalter KL, Aspen SE, Kerst AJ, Lanciotti RS, Moore CG, 2002. Detection of West Nile virus-infected mosquitoes and seropositive juvenile birds in the vicinity of virus-positive dead birds. Am J Trop Med Hyg 67 :492–496.
- 12↑
Caffrey C, Weston TJ, Smith SC, 2003. High mortality among marked crows subsequent to the arrival of West Nile virus. Wildl Soc Bull 31 :870–872.
- 13
Yaremych SA, Warner RE, Mankin PC, Brawn JD, Raim A, Novak R, 2004. West Nile virus and high death rate in American crows. Emerg Infect Dis 10 :709–711.
- 14↑
Hochachka WM, Dhondt AA, McGowan KJ, Kramer LD, 2004. Impact of West Nile virus on American crows in the northeastern United States, and its relevance to existing monitoring programs. EcoHealth 1 :60–68.
- 15↑
Ward MP, Raim A, Yaremych-Hamer S, Lampman R, Novak RJ, 2006. Does the roosting behavior of birds affect transmission dynamics of West Nile virus? Am J Trop Med Hyg 75 :350–355.
- 16↑
Lindsay R, Barker I, Nayar G, Drebot M, Calvin S, Scammell C, Sachvie C, Scammell-La Fleur T, Dibernardo A, Andonova A, Artsob H, 2003. Rapid antigen-capture assay to detect West Nile virus in dead corvids. Emerg Infect Dis 9 :1406–1410.
- 17↑
Yaremych SA, Warner RE, Van de Wyngaerde MT, Ringia AM, Lampman R, Novak RJ, 2003. West Nile virus detection in American crows. Emerg Infect Dis 9 :1319–1321.
- 18
Henson G, Hicock P, 2004. Rapid detection of West Nile Virus in birds using the VecTest WNV Antigen Assay. Clin Lab Sci 17 :218–220.
- 19↑
Stone WB, Okoniewski JC, Therrien JE, Kramer LD, Kauffman EB, Eidson M, 2004. VecTest as diagnostic and surveillance tool for West Nile virus in dead birds. Emerg Infect Dis 10 :2175–2181.
- 20↑
Panella NA, Burkhalter KL, Langevin SA, Brault AC, Schooley LM, Biggerstaff BJ, Nasci RS, Komar N, 2005. Rapid West Nile virus antigen detection. Emerg Infect Dis 11 :1633–1635.
- 21↑
Steele KE, Linn MJ, Schoepp RJ, Komar N, Geisbert TW, Manduca RM, Calle PP, Raphael BL, Clippinger TL, Larsen T, Smith J, Lanciotti RS, Panella NA, McNamara TS, 2000. Pathology of fatal West Nile virus infections in native and exotic birds during the 1999 outbreak in New York City, New York. Vet Pathol 37 :208–224.
- 22
Kramer LD, Bernard KA, 2001. West Nile virus infection in birds and mammals. Ann N Y Acad Sci 951 :84–93.
- 23↑
Panella NA, Kerst A, Lanciotti RS, Byrant P, Wolf B, Komar N, 2001. Comparative West Nile virus detection in organs of naturally infected American crows (Corvus brachyrhynchos). Emerg Infect Dis 7 :754–755.
- 24
Wünschmann A, Shivers J, Carroll L, Bender J, 2004. Pathologic and immunohistochemical findings in American crows (Corvus brachyrhynchos) naturally infected with West Nile virus. J Vet Diagn Invest 16 :329–333.
- 25
Gibbs SEJ, Ellis AE, Mead DG, Allison AB, Moulton JK, Howerth EW, Stallknecht DE, 2005. West Nile virus detection in the organs of naturally infected blue jays (Cyanocitta cristata). J Wildl Dis 41 :354–362.
- 26↑
Komar N, Langevin S, Hinten S, Nemeth N, Edwards E, Hettler D, Davis B, Bowen R, Bunning M, 2003. Experimental infection of North American birds with the New York 1999 strain of West Nile virus. Emerg Infect Dis 9 :311–322.
- 27↑
Lanciotti RS, Kerst AJ, Nasci RS, Godsey MS, Mitchell CL, Savage HM, Komar N, Panella NA, Allen BC, Volpe KE, Davis BS, Roehrig JT, 2000. Rapid detection of West Nile virus from human clinical specimens, field-collected mosquitoes, and avian samples by a TaqMan reverse transcriptase-PCR assay. J Clin Microbiol 38 :4066–4071.
- 28↑
Centers for Disease Control and Prevention, 2003. Epidemic/Epizootic West Nile Virus in the United States: Guidelines for Surveillance, Prevention, and Control. 3rd ed. Available at http://www.cdc.gov/ncidod/dvbid/westnile/resources/wnv-guidelines-aug-2003.pdf. Accessed February 10, 2006.
- 29↑
Ward MR, Stallknecht DE, Willis J, Conroy MJ, Davidson WR, 2006. Wild bird mortality and West Nile virus surveillance: biases associated with detection, reporting, and carcass persistence. J Wildl Dis 42 :92–106.
- 30↑
Eidson M, Schmit K, Hagiwara Y, Anand M, Backenson PB, Gotham I, Kramer L, 2005. Dead crow density and West Nile virus monitoring, New York. Emerg Infect Dis 11 :1370–1375.
- 31↑
Bernard KA, Maffei JG, Jones SA, Kauffman EB, Ebel GD, Dupuis AP II, Ngo KA, Nicholas DC, Young DM, Shi P-Y, Kulasekera VL, Eidson M, White DJ, Stone WB, New York State West Nile virus Surveillance Team, Kramer LD, 2001. West Nile virus infection in birds and mosquitoes, New York State, 2000. Emerg Inf Dis 7 :679–685.
- 32↑
Hayes EB, Komar N, Nasci RS, Montgomery SP, O’Leary DR, Campbell GL, 2005. Epidemiology and transmission dynamics of West Nile virus disease. Emerg Infect Dis 11 :1167–1173.
- 33↑
Nemeth N, Kratz G, Edwards E, Scherpelz J, Bowen R, Komar N, 2007. West Nile virus surveillance in clinic-admitted raptors, Colorado. Emerg Infect Dis 13 :305–307.
- 34↑
Naugle DE, Aldridge CL, Walker BL, Cornish TE, Moynahan BJ, Holloran MJ, Brown K, Johnson GD, Schmidtmann ET, Mayer RT, Kato CY, Matchett MR, Christiansen TJ, Cook WE, Creekmore T, Falise RD, Rinkes ET, Boyce MS, 2004. West Nile virus: pending crisis for greater sage-grouse. Ecol Lett 7 :704–713.
- 36↑
Colorado Department of Public Health and Environment. Disease Control and Environmental Epidemiology Division: West Nile virus. Available at http://www.cdphe.state.co.us/dc/zoonosis/wnv/index.html Accessed December 28, 2005.
- 37↑
Komar N, Lanciotti R, Bowen R, Langevin S, Bunning M, 2002. Detection of West Nile virus in oral and cloacal swabs collected from bird carcasses. Emerg Infect Dis 8 :741–742.
- 38↑
Hunt AR, Hall RA, Kerst AJ, Nasci RS, Savage HM, Panella NA, Gottfried KL, Burkhalter KL, Roehrig JT, 2002. Detection of West Nile virus antigen in mosquitoes and avian tissues by a monoclonal antibody-based capture enzyme immunoassay. J Clin Microbiol 40 :2023–2030.