An outbreak of respiratory disease in dogs occurred in the metropolitan Chicago area, beginning in March 2015.1 In early April 2015, the respiratory disease outbreak was found to be associated with an influenza virus infection in sheltered dogs. The virus was first recognized as canine influenza A H3N2 virus approximately 1 week later on April 13, 2015.1 Infections were initially reported for pet dogs with exposures reportedly occurring in veterinary clinics, day care and boarding facilities, and training centers. Dogs in ≥ 8 animal shelters in the area became severely affected, with almost 200 dogs in shelters testing positive for H3N2 virus in April, May, and June. Most dogs in those shelter populations had clinical signs of respiratory disease and a honking cough.
Two continuously circulating canine influenza A viruses have been identified, the H3N8 and H3N2 strains.1 The CDC, to date, has not reported human infections with either of the canine influenza viruses, but in 2004, H3N8 virus was reported to cross species from horses to dogs in North America2 and in 2013, H3N2 virus reportedly infected cats in an animal shelter in South Korea.3 Canine influenza A H3N8 virus was first identified in the United States in racing Greyhounds in 2004.2 Following the initial reports, H3N8 infections have occurred sporadically and caused some regionally endemic respiratory disease in animal shelters. Prior to the 2015 epidemic in the Chicago area, canine influenza A H3N2 virus had been only reported in China, Thailand, and South Korea since 2007.4
Reports of prolonged shedding of H3N2 virus in experimentally infected dogs prompted an evaluation of the 7-day isolation period previously recommended for H3N8 virus.5,6 Two studies5,7 from South Korea have reported on shedding of H3N2 virus. In otherwise healthy dogs, virus shedding has been shown to extend to 8 days for H3N2,5 whereas prolonged H3N2 shedding has been described in experimentally immune compromised and infected dogs7; immune suppressed dogs shed H3N2 virus for 13 days, compared with 8 days for control dogs. Duration of virus shedding may have important implications for the ecology and spread of the virus as well as have substantial implications for population management in a shelter setting. The purpose of the study reported here was to provide information on canine influenza A H3N2 virus outbreaks in dogs in shelters and companion animal practices in North America and to determine an appropriate isolation period for dogs infected with H3N2 virus on the basis of the duration of virus shedding.
Materials and Methods
Case selection
Beginning in early April of 2015, more than 300 dogs in 8 shelters (shelters A through H) were tested periodically for influenza virus by rRT-PCR assay of nasal swab specimens through the University of Wisconsin Shelter Medicine Program and the WVDL. These 300 dogs represented a sampling of dogs with respiratory disease from each shelter. The combined annual intake for dogs in these shelters is well over 20,000 dogs. Virus isolation was performed on all specimens from all dogs included in this study. Clinical data and treatment information were obtained from individual shelter-maintained medical records. Dogs that tested positive for canine influenza A H3N2 virus were repeatedly tested for H3N2 virus. Dogs were not tested at the same time points (relative to the initial positive rRT-PCR assay result) because of logistical and resource constraints at each shelter. Tracking all dogs that tested positive for H3N2 virus was not possible because shelters continued to release dogs to rescue groups and adopters as well as returning lost dogs to their owners. Dogs were tracked forward from both the time of onset of clinical signs and the time from the initial positive rRT-PCR assay result.
Dogs that tested positive for H3N2 virus within 3 days of the onset of clinical signs and had a follow-up test ≥ 15 days from the initial test were included in the study; dogs with clinical signs in excess of 3 days before the first positive assay result were excluded because the clinical signs had not been quickly followed by canine influenza virus testing, complicating interpretation of the time frame of infection.
Specimen collection
Diagnostic specimens were collected by swabbing inside both the nares of each dog with sterile swabs that were individually wrapped. Veterinarians or trained veterinary technicians collected all diagnostic specimens. Sterile swabs were kept unexposed to the environment until just before sample collection, and gloves were changed between dogs. Swab specimens were immersed in individual tubes of brain-heart infusion broth, virus transport medium,a or sterile saline (0.9% NaCl) solution immediately after collection and transported chilled to the WVDL.
rRT-PCR assay
On receipt at the laboratory, specimens underwent extraction that included an internal control with a magnetic particle bead-based processorb and extraction kit.c A broadly reactive rRT-PCR assay directed to the matrix gene of influenza A virus was used to identify dogs that were positive for influenza virus. The matrix assay was developed at WVDL and validated according to AAVLD guidelines.8 The assay is sensitive to 25 genome copies and has a linearity and efficiency of 0.999 and 100%, respectively. The diagnostic sensitivity is 96.8% (95% confidence interval, 91% to 99.3%) and diagnostic specificity is 100% (95% confidence interval, 94.9% to 100%). Briefly, a 1-step reverse transcription assay followed by a probe-based amplification was performed for 40 cycles in a real-time thermocycler instrument.d A positive amplification control, extraction control, and negative control were performed with each assay. Cycle threshold levels were determined; values ≤ 36 were considered positive results, and values ≥37 but < 40 were considered weak positive results. Amplicons were sent to a commercial facilitye where they were purified, quantified, and sequenced by the Sanger DNA sequencing technique with the following primer: reverse, 5′-AGGGCATTTTGGACAAAKCGTCTA-3′.
Real-time multiplex-PCR assay
A canine respiratory real-time multiplex-PCR assay, which was developed at WVDL and validated according to AAVLD guidelines,8 was performed on the initial nasal swab specimens from the influenza virus-positive dogs. The panel included testing for canine distemper virus, canine respiratory coronavirus, canine herpesvirus-1, canine parainfluenza virus, canine pneumovirus, canine adenovirus-2, Bordetella bronchiseptica. and canine influenza A virus. The real-time multiplex-PCR assay was performed in 3 varied multiplex reactions for a total of 9 targets, 1 of which was the internal control. All assays were validated at the analytical level as described in the AAVLD guidelines.8 Briefly, the limit of detection, dynamic range, and efficiency of the reaction were determined. All but the canine pneumovirus were validated at the diagnostic level, which includes testing at least 30 known positive samples and 30 known negative samples. The diagnostic validation for canine pneumovirus has been completed with 7 known positive samples of the suggested 30.
Virus isolation
Virus isolation was performed with the use of embryonated chicken eggs9 and the MDCK cell line.f Embryonated eggs were inoculated by the allantoic inoculation route with 0.2 mL of filtered virus specimens 9 to 11 days after fertilization. Eggs were candled daily and harvested when infection was evident or at days 5 to 7. Virus growth was monitored by rRT-PCR assay for a decrease in CT. Filtered virus suspensions were inoculated on MDCK cells with Eagle minimum essential medium supplemented with 0.25% fetal bovine serum and 0.1% trypsin. Cell cultures were incubated for 5 to 11 days.10 The rRT-PCR assay was performed at days 5 and 11 to monitor CT values. Egg and MDCK cell cultures that had negative results for virus growth were subjected to 1 blind passage for 4 to 5 days.
Results
Sixteen dogs from 3 shelters (shelters A, B, and C) that developed varying degrees of clinical signs of respiratory disease and had shedding of influenza A virus confirmed by positive rRT-PCR assay results early in the course of their disease were included in the study. These 16 dogs were tested intermittently for virus shedding by rRT-PCR assay for ≥ 15 days after the first positive tests results. From each initial nasal swab specimen of dogs that tested positive for influenza A shedding, canine influenza A H3N2 virus was identified by the Sanger DNA sequencing technique.
Eleven of 16 dogs were identified as pit bull-type dogs. One dog each was identified as a Labrador Retriever, Bulldog, and Border Collie. Two dogs were identified as mixed-breed dogs. Age or estimated age ranged from 8 months to 7.5 years. Nine females and 7 males were included in the study. Nine dogs had been neutered, and 7 were sexually intact.
Ten dogs were housed in shelter A, 5 dogs were housed in shelter B, and 1 dog was housed in shelter C. All affected dogs received supportive care and a variety of antimicrobial regimens for secondary infections. Clinical signs resolved, even though rRT-PCR assay results for virus shedding continued to be persistently positive or intermittently positive for all dogs that remained in shelter care during the entire testing period.
For the 16 dogs, the interval between the first and last positive rRT-PCR assay result for influenza A virus shedding ranged from 13 to 24 days (Tables 1 and 2). Dogs were tested for canine influenza virus shedding via rRT-PCR assay between 2 and 9 times during the course of signs of respiratory disease and after resolution of clinical signs.
Canine influenza A rRT-PCR assay results (CT values) for 10 of 16 affected dogs in a Chicago area animal shelter (shelter A [dogs 1 through 10]) during an outbreak of respiratory disease.
| CT values by dog No.* | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Time (d)† | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 |
| 1 | 20.8‡ | 34.7 | Neg | 29.1 | Neg | 16.7‡ | 19.0‡ | 34.5 | 25.2‡ | 25.1‡ |
| 13 | 39.3 | 38.1 | 31.3 | 35.1 | 33.9 | 31.4 | 31.9 | Neg | 29.6 | 37.3 |
| 14 | Neg | Neg | 32.1 | Neg | 32.0 | 35.3 | 35.4 | 38.1 | 32.3 | Neg |
| 15 | Neg | Neg | 36.2 | 37.1‡ | Neg | 39.1 | Neg | Neg | 38.5 | Neg |
| 17 | Neg | Neg | 37.2 | Neg | Neg | 31.1 | Neg | Neg | Neg | Neg |
| 19 | Neg | NT | Neg | Neg | Neg | Neg | Neg | Neg | Neg | Neg |
| 20 | Neg | NT | Neg | NT | Neg | Neg | Neg | 38.2 | 37.9 | 37.8‡ |
| 22 | Neg | NT | Neg | NT | Neg | Neg | 36.5 | 39.2 | Neg | Neg |
| 24 | Neg | NT | Neg | NT | Neg | 39.3 | Neg | Neg | Neg | 37.3 |
| 29 | Neg | NT | Neg | NT | Neg | Neg | Neg | Neg | Neg | Neg |
| 31 | Neg | NT | Neg | NT | Neg | Neg | Neg | Neg | Neg | Neg |
CT values ≤ 36 are considered positive results and values ≥ 37 but < 40 are considered weak positive results.
Day after first positive result for influenza A virus shedding determined by rRT-PCR assay of nasal swab specimens in 16 shelter dogs.
Indicates successful isolation of canine influenza A H3N2 from this specimen. isolation of H3N2 virus was successful in 12 of 16 dogs with positive rRT-PCR assay results for influenza A virus shedding.
Neg = Negative rRT-PCR assay result for influenza A virus shedding. NT = Not tested.
Canine influenza A rRT-PCR assay results (CT values) for 6 of 16 affected dogs in Chicago area animal shelters (ie, shelters B [dogs 11 through 15] and C [dog 16]) during an outbreak of respiratory disease.
| CT values by dog No.* | ||||||
|---|---|---|---|---|---|---|
| Time (d)† | 11 | 12 | 13 | 14 | 15 | 16 |
| 1 | Neg | 37.4 | 25.1 | 37.9 | 26.0 | 29.5‡ |
| 5 | NT | NT | 22.7 | 24.6‡ | NT | NT |
| 6 | 24.6† | 23.7‡ | NT | NT | NT | NT |
| I0 | 36.2 | 39.5 | NT | NT | NT | NT |
| II | NT | NT | NT | NT | 37.1 | NT |
| I5 | NT | NT | 38.6‡ | 35.9 | 38.2‡ | 36.0 |
| I9 | NT | NT | Neg | NT | NT | NT |
| 20 | 37.3 | 37.2 | NT | NT | NT | NT |
| 23 | NT | NT | Neg | NT | NT | Neg |
| 24 | Neg | Neg | NT | NT | NT | NT |
| 28 | Neg | Neg | NT | NT | NT | NT |
See Table 1 for key.
All but 1 of 10 dogs housed in shelter A had detectable virus shedding 13 days after the initial positive rRT-PCR assay result for canine influenza A virus shedding. Two dogs in shelter A were transferred out and lost to follow-up 17 and 19 days after the first positive assay result. Four of the remaining 8 dogs in shelter A had detectable canine influenza virus shedding at 20 to 24 days after the initial positive assay result. Two of 5 dogs housed in shelter B had detectable virus shedding at 20 days after the first positive rRT-PCR assay result and 24 days after the onset of clinical signs. Three of 5 dogs in shelter B had positive assay results for virus shedding at 15 days after their initial positive assay result. The 1 dog from shelter C had a positive rRT-PCR assay result for virus shedding at 15 days after the first positive assay result. In the 8 dogs in shelter A that were tested with the greatest frequency, the use of rRT-PCR assay resulted in detection of intermittent shedding of canine influenza virus, with as many as 4 negative assay results (with 2- to 3-day intervals between sample collection) followed by single or multiple positive assay results (Tables 1 and 2).
Cycle threshold values increased over time, indicating decreased virus nucleic acid content in the nasal swab specimens for 14 of the 16 study dogs following the initial positive specimen. Two of the dogs in shelter B had weak positive results (37.4 and 37.9 CT values) on initial testing followed by strong positive results (23.7 and 24.6 CT values) and then increasing CT values. One of these dogs was the first to test positive in shelter B. In shelter A, 2 dogs with the strongest initial positive results (16.7 and 19.0 CT values) also had the longest confirmed shedding periods (22 and 24 days). In shelter C, the 1 dog with virus shedding confirmed over 15 days had received a series of 2 vaccinesg for canine influenza with the first dose given 30 days prior to the positive assay rRT-PCR result and a second dose given 4 days prior to the development of clinical signs. This dog also was suspected of having spindle cell or mixed cell mammary neoplasia and had been treated with prednisone.
Isolation of canine influenza A H3N2 virus was attempted from nasal swab specimens with positive rRT-PCR assay results for all 16 dogs. Isolation of H3N2 virus was successful from ≥ 1 nasal swab specimen from 12 of 16 dogs (Tables 1 and 2). Isolation of H3N2 was successful in 26.8% (15/56) of all nasal swab specimens tested, including those obtained from 4 dogs during the late shedding period of 15 to 24 days. Two of these dogs were in shelter A (dog Nos. 4 and 10), and 2 dogs were in shelter B (dog Nos. 13 and 15). The H3N2 virus was isolated from 1 nasal swab specimen taken 20 days after the first positive rRT-PCR assay result. For this dog, day 1 of sample collection was the same as the first day clinical signs were reported. The H3N2 virus was also isolated from nasal swab specimens from 3 dogs obtained 15 days after the first positive rRT-PCR assay result, which was 18 days after the first clinical signs of illness. Cycle threshold values for these specimens were 37.1, 37.8, 38.6, and 38.2, respectively.
Sequence analysis of the virus was performed on the penultimate or last nasal swab specimens with positive rRT-PCR assay results for each dog. The specimens from 9 dogs (Nos. 3, 6, 7, 9, 10, 11, 12, 14, and 16) yielded an amplicon that gave quality sequence data, and in each case the data were homologous to A/canine/Illinois/12191/2015 (H3N2) GenBank No. KT002539. Therefore, results from 12 dogs (Nos. 3, 4, 6, 7, 9, 10, 11, 12, 13, 14, 15 and 16) were independently confirmed as canine influenza A H3N2 virus by either sequence analysis alone or virus isolation followed by sequence analysis of the penultimate or last nasal swab specimen with positive rRT-PCR assay result.
Two of 8 shelters (shelters D and E) maintained adoption facilities, separated from their isolation housing, where infected dogs were transferred when recovered and presumed to be no longer shedding. Newly admitted, apparently healthy dogs were also housed in the adoption facilities. When dogs were transferred to the holding facility after 7 days or 14 days in isolation, shelter directors, veterinarians, and managers anecdotally reported many new cases of respiratory disease in dogs within the adoption facilities, suggesting that infections continued to spread to naïve dogs. After extending the isolation period to 21 days, the number of new cases of respiratory disease appeared to lessen and transmission within the shelters between dogs appeared to stop.
The canine respiratory real-time multiplex-PCR assay findings revealed that 10 of 16 dogs had negative results for all pathogens tested, other than canine influenza A virus. Two of 10 dogs in shelter A had positive canine respiratory real-time multiplex-PCR assay results. In addition to canine influenza A virus, 1 dog had positive results for canine parainfluenza virus and canine respiratory coronavirus and 1 dog had positive results for canine pneumovirus. Four of 5 dogs at shelter B had positive real-time multiplex-PCR assay results for ≥ 2 pathogens other than canine influenza virus; these dogs had positive results for a combination of canine parainfluenza virus, canine pneumovirus, and canine respiratory coronavirus. The dog in shelter C did not have positive real-time multiplex-PCR assay results for any pathogens other than canine influenza A virus.
Discussion
The data reported here suggest that the duration of virus shedding from dogs naturally infected with canine influenza A H3N2 virus is longer, with intermittent negative and positive test results, compared with dogs infected with H3N8 virus6 or dogs with experimentally induced H3N2 virus7 infections. The duration of virus shedding reported here is 2 to 3 times as long as the 1 to 7 days that has been reported for H3N8.6 In addition, duration of shedding in the naturally infected dogs of this report is also greater than that described for experimental challenges with influenza H3N2 that included immune suppression with prednisone treatment.7 Although it is possible that the stress of shelter housing mimics the immune suppression caused by prednisone administration, duration of shedding observed in the present report exceeded that reported in the other study.7 We have consistently found in the shelters from our study that there were new infections within their populations when they followed either a 7- or 14-day isolation protocol for dogs with suspected or confirmed canine influenza virus infections during the outbreak. New infections were reported to decrease substantially when shelters switched to a 21-day isolation protocol before commingling with dogs thought to be previously unexposed.
We report shedding time as the day from first positive rRT-PCR assay result test because use of onset of clinical signs could be associated with coinfections or other pathogens, both of which are common in the shelter setting. We also report range of time between first signs and last positive rRT-PCR assay result, only when the first clinical signs began within 3 days of the first confirmed positive rRT-PCR assay result. The longer duration of virus shedding of canine influenza A virus found in the dogs of the present report would correlate to a similar time frame for potential contagion. In the present report, the rRT-PCR assay results and the limited virus isolation data suggested viable virus from prolonged shedding. In terms of the virus isolation results, it is noteworthy that although successful virus isolation was achieved from nasal swab specimens with decreased virus load (ie, high CT values), testing of some specimens with a moderate virus load (eg, CT value of 31) did not result in successful virus isolation, perhaps because nasal swab specimens had been submitted on swabs made of wood and cotton, which may have residual chlorine known to degrade virus nucleic acid.11,12
Prolonged virus shedding is a critical driver of transmission and has been associated with increased severity and weakened host defenses in humans infected with influenza A virus.13 Lee et al13 defined prolonged virus shedding as positive detection of influenza A H1N1 virus by rRT-PCR assay 7 days after diagnosis. An extended period of shedding of influenza A virus has also been reported in an immune compromised child who shed virus for > 1 year.14 Prolonged virus shedding has also been reported in 25% of hospitalized patients with pandemic influenza A H1N1 virus infection.15
Documentation of the extended shedding period of canine influenza A H3N2 virus in the affected dogs of the present report helps to explain, at least in part, why such a large number of dogs were affected by H3N2 infections in this outbreak, as prolonged shedding increases the likelihood of exposure. These data and anecdotal information reported by shelters affected by H3N2 virus suggest that, to reduce the risk of transmission of H3N2 virus, infected dogs should be isolated from uninfected, unexposed dogs for a period of ≥ 21 days following onset of illness.
Acknowledgments
Supported by Maddie's Fund and the WVDL.
ABBREVIATIONS
| AAVLD | American Association of Veterinary Laboratory Diagnosticians |
| CT | Cycle threshold |
| MDCK | Madin-Darby canine kidney |
| rRT-PCR | Real-time reverse transcriptase PCR |
| WVDL | Wisconsin Veterinary Diagnostic Laboratory |
Footnotes
Remel M-6, Thermo Fisher Scientific, Lenexa, Kan.
Kingfisher, Thermo Fisher Scientific, Vanta, Finland.
MagMAX-96 Viral RNA Isolation Kit (AM1836), Thermo Fisher Scientific, Vilnius, Lithuania.
Applied Biosystems 7500 Fast, Thermo Fisher Scientific, Singapore.
Functional Biosciences, Madison, Wis.
CCL-34 ATCC Manassas, Va.
H3N8 killed virus, Zoetis Vanguard, Florham Park, NJ.
References
1. CDC. Update on Canine Influenza (Dog Flu) Outbreak Reported in Chicago Area. Available at: www.cdc.gov/flu/news/canine-influenza-update.htm. Accessed Apr 13, 2015.
2. Crawford PC, Dubovi EJ, Castleman WL, et al. Transmission of equine influenza virus to dogs. Science 2005; 310:482–485.
3. Jeoung HY, Lim SI, Shin BH, et al. A novel canine influenza H3N2 virus isolated from cats in an animal shelter. Vet Microbiol 2013; 165:281–286.
4. Zhu H, Hughes J, Murcia PR. Origins and evolutionary dynamics of H3N2 canine influenza virus. J Virol 2015; 89:5406–5418.
5. Song D, Kang B, Lee C, et al. Transmission of avian influenza virus (H3N2) to dogs. Emerg Infect Dis 2008; 14:741–746.
6. Pecoraro HL, Bennett S, Garretson K, et al. Comparison of the infectivity and transmission of contemporary canine and equine H3N8 influenza viruses in dogs. Vet Med Intern 2013; 2013:874521.
7. Hong M, Kang B, Na W, et al. Prolonged shedding of the canine influenza H3N2 virus in nasal swabs of experimentally immunocompromised dogs. Clin Exp Vaccine Res 2013; 2:66–68.
8. Laboratory Technology Committee. Crossley B, Toohey-Kurth K, co-chairs. Guidelines for assay development and performance of PCR in veterinary diagnostic laboratories. Kansas City Mo: American Association Veterinary Laboratory Diagnosticians, 2014.
9. Dufour-Zavala L. A laboratory manual for the isolation, identification, and characterization of avian pathogens. Athens, Ga: American Association of Avian Pathologists, 2008.
10. Meguro H, Bryant JD, Torrence AE, et al. Canine kidney cell line for isolation of respiratory viruses. J Clin Microbiol 1979; 9:175–179.
11. Johnson FB. Transport of viral specimens. Clin Microbiol Rev 1990; 3:120–131.
12. Spackman E. Avian influenza virus. Totowa, NJ: Humana Press, 2008.
13. Lee N, Chan PK, Hui DS, et al. Viral loads and duration of viral shedding in adult patients hospitalized with influenza. J Infect Dis 2009; 200:492–500.
14. Pinsky BA, Mix S, Rowe J, et al. Long-term shedding of influenza A virus in stool of immunocompromised child. Emerg Infect Dis 2010; 16:1165–1167.
15. Giannella M, Alonso M, Garcia de Viedma, et al. Prolonged viral shedding in pandemic influenza A(H1N1): clinical significance and viral load analysis in hospitalized patients. Clin Microbiol Infect 2011; 17:1160–1165.
