In North America around 2000 or earlier, the H3N8 influenza A virus jumped species from horses to dogs, and in 2004, the virus was identified from a series of outbreaks that occurred in racing Greyhounds in the Southeastern United States.1 Prior to that time, a species-adapted influenza A virus had not been detected in dogs. The CIV therefore represents a novel virus for canine species, to which the dog population is immunologically naive and highly susceptible. Outbreaks of CIV infection in this population can be expected, and indeed, infection has been spreading slowly through population clusters since 2004.2,3 However, the ongoing H3N8 CIV outbreak in North America has been associated with lower case fatality rates than initially feared, with estimates ranging from 1% to 5%.1,4–6
The H3N2 subtype of influenza A virus may have jumped to canine species in China or the Korean peninsula around the same time (early 2000s), possibly through genetic recombination events involving various avian influenza viruses known to circulate in that region.7 In 2015, an outbreak of CIV infection in and around Chicago was the first identified cluster of H3N2 CIV cases involving North American dogs.8
Dogs with CIV infection typically have pyrexia, coughing, sneezing, and signs of malaise, and these clinical signs are indistinguishable from those of more classic causes of infectious tracheobronchitis (also known as kennel cough) in dogs.9–11 In most instances, CIV infection is self-limiting or responsive to supportive care, and dogs recover within 2 to 3 weeks after initial infection. As for dogs with infectious tracheobronchitis, pneumonia is the most dangerous sequela, and the very old, very young, or previously unhealthy are at greatest risk of complications.4
Early data on H3N8 CIV infection suggest case fatality rates > 20%.1 However, those estimates included large numbers of track Greyhounds that died early in the discovery of this new disease, so subsequent estimates have continually declined as the infection has spread through the pet dog population. The 2015 Chicago outbreak of H3N2 CIV infection garnered considerable media attention and generated concern that morbidity and mortality rates associated with the H3N2 subtype may be greater than those associated with the H3N8 subtype.12–15 However, if popular press reports14,16 are accurate, then H3N2 CIV infection was diagnosed in > 1,500 dogs and 6 deaths occurred during the Chicago outbreak, equivalent to a case fatality rate of 0.4%.
In the summer of 2015, another CIV outbreak occurred in and around Atlanta that also resulted in popular press reports17,18 and, in the authors' experience, veterinary consultation calls expressing concern about severe and lethal outcomes. The submission of respiratory secretion samples from many of the affected dogs to the University of Georgia Athens Veterinary Diagnostic Laboratory for PCR testing resulted in the eventual identification of H3N2 CIV in all samples sent for typing.19 This cluster of cases of H3N2 CIV infection created an opportunity to assess the outcome of H3N2 influenza cases within a single temporal and regional outbreak.
The purpose of the study reported here was to determine the case fatality rate of dogs with positive results of PCR testing and compare clinical signs and other factors between CIV-positive dogs and CIV-negative dogs. Our hypotheses were that the case fatality rate associated with the H3N2 CIV outbreak in the Atlanta area in the summer of 2015 would be low and similar to the previously reported case fatality rate (1% to 5%) for H3N8 CIV-infected dogs and that clinical signs of H3N2 CIV infection would be similar to other forms of infectious tracheobronchitis in dogs.
Materials and Methods
Dogs
The laboratory information management system of the Athens Veterinary Diagnostic Laboratory was searched to identify dogs for which nasal swab specimens were submitted between May 1 and July 2, 2015, for PCR detection of influenza A virus (ie, CIV when in dogs) alone or influenza A and 5 other respiratory pathogens (canine adenovirus, Bordetella bronchiseptica, canine distemper virus, canine respiratory coronavirus, and Mycoplasma spp). All dogs identified by this means were eligible for inclusion in the study.
Survey
A brief survey was devised by one of the authors (PMK). Questions were designed to ascertain each dog's history prior to the veterinary visit when nasal swab specimens had been collected for PCR testing (including travel within the past 2 months, exposure to other dogs, and vaccination within the past 12 months), clinical signs, and survival status at survey completion (alive [yes or no] or unknown; Supplementary Appendix S1, available at avmajournals.avma.org/doi/suppl/10.2460/javma.252.8.959). The survey was created by use of an electronic survey program,a and a link to the survey was emailed on September 1, 2015, to all veterinarians who had submitted nasal swab specimens for CIV testing during the study inclusion period. If no email address was available, paper copies were mailed to the veterinary clinic from which the specimens had been sent. Nonresponders were then contacted via telephone, email, or both requesting completion of the survey. In some instances, veterinarians were asked to return the survey via facsimile transmission. The survey was closed to responses on December 1, 2015. Data from all returned surveys (whether via email, mail, telephone, or facsimile transmission) were entered into a spreadsheet programb and merged with the applicable PCR test results and specimen submission information (eg, dog signalment and place of residence) for statistical analysis.
Statistical analysis
Dogs with a positive or suspicious result of PCR testing for influenza A virus were categorized as CIV positive, and all other dogs were categorized as CIV negative. Continuous data acquired via survey and submission records, such as dog age at and duration of clinical signs prior to testing, were evaluated for normality of distribution. Because the data were normally distributed, values are reported as mean ± SD. Age was compared between CIV-positive and CIV-negative dogs both as a continuous variable (Student 2-tailed t test) and as a categorical variable (adult [> 1 year] vs juvenile [≤ 1 year]; Fisher exact test). For analyses regarding B bronchiseptica vaccination, SC and intranasal administration were the options provided in the survey, but because some veterinarians wrote in the use of an orally administered vaccine, orally and intranasally administered Bordetella vaccines were grouped together as mucosally administered.
Distributions of sex, reproductive status, reported historical data, reported clinical signs, and survival status were compared between groups by means of the Fisher exact test and computation of prevalence ORs and associated 95% confidence intervals. The family-wise error rate with Bonferroni correction was used to account for multiple comparisons.20 To obtain a study-wide family-wise error rate of 5%, the significance threshold for each individual putative risk factor for CIV positivity was determined by dividing 0.05 by the number of comparisons performed. Given that 15 combinations of variables were evaluated, values of P < 0.0033 were considered significant. Multivariate logistic regression was also performed to identify factors independently associated with CIV positivity.
Results
Dogs
A total of 220 dogs were included in the study. Of these, 73 (33%) had positive results of PCR testing for influenza A virus (CIV), 4 (2%) had suspicious results, and 143 (65%) had negative results. A subset of 23 CIV-positive specimens submitted during the study period had positive results of H3N2 subtype testing, as additionally performed at the Cornell University Animal Health Diagnostic Center. Results for respiratory pathogens other than CIV were not analyzed for study purposes.
Surveys
Surveys were initially sent to veterinarians via email for 158 (72%) dogs and by mail for 62 (28%) dogs, representing 88 veterinary clinics overall. Clinics from which multiple specimens had been submitted sometimes provided responses for only some surveys, particularly clinics with multiple veterinarians. A total of 120 surveys were completed, representing a response rate of 54.5% from 48 clinics. All but one of the responses were from clinics in the greater Atlanta metropolitan area; 1 response was received for a specimen that had been submitted by personnel at a humane society in Illinois. It was unknown whether the dog tested through this humane society had a travel history that included the Atlanta area.
CIV status
Of the 100 dogs (49 clinics) for which no response was received, 63 (63%) had a negative result of PCR testing for CIV, 36 (36%) had a positive result, and 1 (1%) dog had a suspicious result. Of the 120 dogs for which a completed survey was received, 80 (67%) had a negative test result (including the 1 dog from Illinois). The remaining 40 (33%) dogs had a positive (n = 37) or suspicious (3) test result, and these dogs were all treated as CIV positive for study purposes. The CIV-negative dogs had been tested a mean ± SD of 2.5 ± 2.1 days after onset of clinical signs, whereas the CIV-positive dogs had been tested a mean of 2.4 ± 1.7 days after onset of clinical signs (P = 0.76).
Signalment
The CIV-positive group included 20 (50%) neutered males, 15 (38%) spayed females, and 5 (12%) sexually intact males, and the CIV-negative group included 31 (39%) neutered males, 27 (34%) spayed females, 13 (16%) sexually intact males, and 9 (11%) sexually intact females. No difference was identified between the groups in sex or reproductive status distributions (P = 0.13). Mean ± SD age of the CIV-positive dogs was 5.3 ± 3.7 years (range, 4 months to 12 years), and that of the CIV-negative dogs was 3.4 ± 3.6 years (range, 2 weeks to 15 years; P = 0.008).
The CIV-positive group included 9 (22%) mixed-breed dogs, 6 (15%) Labrador Retrievers, 3 (8%) Golden Retrievers, 3 (8%) Boxers or Boxer mixes, 2 (5%) Golden Retriever-Poodle mixes, 2 (5%) Great Danes, 2 (5%) Poodles, and 1 (2%) each of 13 other breeds or breed types. The CIV-negative group included 15 (22%) mixed-breed dogs, 11 (14%) Labrador Retrievers, 8 (10%) German Shepherd Dogs, 4 (5%) Blue Heelers or Blue Heeler mixes, 4 (5%) Boxers, 4 (5%) Golden Retriever-Poodle mixes, 3 (4%) Golden Retrievers, 3 (4%) Bulldogs, 2 (2%) Border Collies, 2 (2%) Dachshunds, 2 (2%) French Bulldogs, 2 (2%) Maltese, 2 (2%) terrier types, and 1 (1%) each of 20 other breeds or breed types.
Clinical signs, vaccination status, and survival status
Univariate analysis revealed that CIV-positive dogs were significantly more likely than CIV-negative dogs to be adult (> 1 year of age) and to have coughing or sneezing, signs of lethargy or inappetence, pyrexia, and a history of having been vaccinated SC against B bronchiseptica infection ≤ 12 months prior to testing (Table 1). Multivariate logistic regression revealed that adult (vs juvenile) status (P < 0.001), inappetence (P = 0.047), and SC vaccination against B bronchiseptica infection (P = 0.004) were independently significant. However, with correction for multiple comparisons, only adult status and inappetence were significantly (P < 0.0033) associated with a positive CIV test result.
Comparison of the prevalence of various factors between CIV-positive and CIV-negative dogs for which a nasal swab specimen was submitted to an Atlanta diagnostic laboratory for PCR assay detection of CIV alone or CIV and 5 other respiratory pathogens between May 1 and July 2, 2015.
| Factor | Proportion (%) of CIV-positive dogs | Proportion (%) of CIV-negative dogs | Prevalence OR (95% CI) | P value |
|---|---|---|---|---|
| Adult* | 38/40 (95) | 50/80 (62) | 11.4 (2.6–50.7) | < 0.001 |
| Survived ≥ 2 mo after initial evaluation | 38/39 (97) | 66/72 (92) | 3.5 (0.4–29.8) | 0.42 |
| Clinical signs | ||||
| Coughing or sneezing | 39/40 (98) | 67/80 (84) | 7.6 (1.0–60.1) | 0.030 |
| Lethargy | 31/40 (78) | 37/80 (46) | 4.0 (1.7–9.5) | 0.002 |
| Inappetence | 31/40 (78) | 30/80 (38) | 5.7 (2.4–13.7) | < 0.001 |
| Pyrexia | 29/38 (76) | 29/80 (36) | 5.7 (2.4–13.6) | < 0.001 |
| Canine housemates | 14/36 (39) | 36/76 (47) | 0.7 (0.3–1.6) | 0.42 |
| CIV-positive canine housemates | 6/14 (43) | 2/36 (6) | 12.7 (2.1–75.3) | 0.004 |
| Vaccines received ≤ 12 mo prior to testing | ||||
| Bordetella bronchiseptica, SC | 17/39 (44) | 14/78 (18) | 3.5 (1.5–8.3) | 0.004 |
| B bronchiseptica, mucosal† | 13/40 (32) | 37/80 (46) | 0.6 (0.3–1.2) | 0.17 |
| B bronchiseptica, any route | 30/40 (75) | 49/80 (61) | 1.9 (0.8–4.4) | 0.16 |
| CIV | 6/39 (15) | 19/78 (24) | 0.6 (0.2–1.6) | 0.34 |
Adult was defined as > 1 year of age.
Mucosal administration includes both oral and intranasal routes.
CI = Confidence interval.
The OR represents the odds of a dog with the indicated factor being CIV positive relative to the odds for a dog without that factor.
In the free-text field of the survey regarding clinical signs other than those specifically listed (lethargy, inappetence, coughing or sneezing, and pyrexia), ocular or nasal discharge was reported for 22 (10%) dogs, diarrhea for 9 (4%) dogs, vomiting for 4 (2%) dogs, and respiratory distress for 4 (2%) dogs. No significant difference was identified between CIV-positive and CIV-negative dogs with regard to the prevalence of these additional clinical signs. Overall, 78 of the 80 (98%) CIV-negative dogs and all 40 (100%) CIV-positive dogs had ≥ 1 clinical sign (P = 0.55).
No significant differences were identified between CIV-positive and CIV-negative dogs in proportions that had received a mucosal vaccine (intranasal [n = 12] or oral [3]) or any vaccine against B bronchiseptica infection ≤ 12 months prior to testing, had received a vaccine against CIV infection, had travelled ≤ 2 months prior to testing, or had other dogs in household. One (3%) CIV-positive dog and 6 (8%) CIV-negative dogs died during the 2- to 7-month period between specimen submission and survey completion. The difference in these survival rates was not significant (P = 0.42). The CIV-positive dog that died was reported to have had a 3-year history of chronic bronchitis for which inhalant steroid treatment had been prescribed. This dog had developed bacterial pneumonia that responded clinically to antimicrobial treatment 4 months prior to the veterinary visit that prompted respiratory pathogen testing. It had reportedly once again developed clinical signs of respiratory disease at home 2 days prior to evaluation by the veterinarian who had collected the specimen for respiratory pathogen testing, and it went into respiratory arrest shortly thereafter.
Of the 6 CIV-positive dogs reportedly vaccinated against H3N8 CIV infection ≤ 12 months prior to testing, results of serotyping were available for 5, indicating infection with the H3N2 strain. No serotyping had been performed for the sixth dog.
Forty-seven percent (36/76) of CIV-negative dogs for which the number of canine housemates was reported came from a multidog household, whereas 39% (14/16) of the CIV-positive dogs came from a multidog household (P = 0.42). Among dogs with canine housemates, 2 of 36 (6%) CIV-negative dogs had CIV-positive housemates, whereas 6 of 14 (43%) CIV-positive dogs had CIV-positive housemates (P = 0.004).
Discussion
In the few studies1,4,6 in which case fatality rates have been determined for H3N8 influenza A infection in dogs, results have ranged from 5% to 36%. Results of the present study indicated that the case fatality rate of dogs with presumed H3N2 CIV infection in and around Atlanta during the 2015 summer outbreak was low as well. Only a single (3%) CIV-positive dog was reported to have died, whereas all other CIV-positive dogs (97%) recovered from their illness. We suggest that this finding indicated that the novelty of this virus in the American dog population is not necessarily associated with high lethality.
Different and more severe outcomes have been reported for dogs experimentally infected with H3N2 CIV.c The reason for the discrepancy in findings between that studyc and the present study are unknown, but could involve differences in field versus laboratory strains, size of inoculum between field and experimental conditions, or study end points in clinical trials versus clinical practice.
Of the 120 dogs for which surveys were completed in the study reported here, 40 (33%) had positive or suspicious results of PCR testing for CIV, and 80 (67%) dogs had negative results. All 23 CIV-positive dogs in this study that had their nasal swab specimens submitted for viral subtype determination were confirmed to be infected with H3N2 and not H3N8 influenza A virus, further substantiating the existence of an H3N2 outbreak in the Atlanta area in the summer of 2015.
Comparisons of the distributions of sex and reproductive status between CIV-positive and CIV-negative dogs yielded no significant differences. No breed predilection was identified. Mean age of CIV-positive dogs was 5.3 years, which was significantly greater than the mean age of CIV-negative dogs (3.4 years). Severe manifestations of influenza A infection in humans are considered most likely among the very old, the very young, and those with prior respiratory ailments.21 In parallel, infectious tracheobronchitis in dogs appears most likely to become complicated in similar risk groups.22,23 Severity of clinical signs as a function of age was not assessed in the present study.
A few factors may explain the age distribution of CIV-positive dogs in the present study. A similar distribution was identified in the Chicago outbreak of CIV infection earlier the same year.24 First, dogs that are very young or old may not be routinely exposed to settings with high dog densities, such as daycare or boarding facilities, where transmission of respiratory pathogens is likely and age restrictions may apply. An overall lack of exposure to other dogs for those at the ends of the age spectrum could result in a false suggestion of an increased risk of CIV exposure and subsequent infection for middle-aged dogs.
Because the reason for respiratory pathogen testing (CIV alone or respiratory pathogen panel) was not determined in the present study, it was unknown whether CIV infection was considered likely by the attending clinician for any dog. Indeed, such testing could have been requested simply for dogs sharing a home with a known CIV-positive dog, regardless of their age. No clinical signs have been reported to help differentiate CIV infection specifically from other causes of infectious tracheobronchitis in dogs, and clinical signs can also vary greatly among dogs. Infection with H3N2 CIV has been anecdotally believed to have a more rapid onset and more severe manifestation than infection with H3N8 CIV and other components of the canine infectious tracheobronchitis complex.14,16
In the present study, no significant difference between CIV-positive and CIV-negative dogs was identified in duration of clinical signs prior to the veterinary visit when the nasal swab specimen was collected. Although the retrospective nature of the study did not allow for assessment of the interval between potential CIV exposure and onset of clinical signs, the lack of a difference in the interval from onset of clinical signs to testing suggested that cases of CIV infection were indistinguishable in timeline or severity of onset from other disease processes with similar clinical signs.
Clinical signs associated with infectious tracheobronchitis, and with CIV infection in particular, commonly include coughing, sneezing, pyrexia, lethargy, ocular or nasal discharge, and inappetence.11,25 As expected, such clinical signs were observed in both the CIV-positive and CIV-negative groups of dogs in the present study. Coughing and sneezing were the most common signs in both CIV-positive and CIV-negative dogs. However, these and other signs were significantly less common in CIV-negative dogs than in CIV-positive dogs, without controlling for other factors. Whereas a large but smaller proportion of CIV-negative dogs (84%) had respiratory signs such as coughing and sneezing than did CIV-positive dogs (98%), the lower prevalence of respiratory signs in CIV-negative dogs could have been simply attributable to inclusion in that group of dogs that had no clinical signs but were tested for respiratory pathogens because of exposure to known CIV-infected dogs. Indeed, this variable did not maintain a significant association with infection status in the multivariate analysis. However, that possibility could not be explored with the available data.
Although typically considered a common finding in dogs with CIV infection, ocular and nasal discharge was no more prevalent in CIV-positive dogs than in CIV-negative dogs. The omission of these signs from the list of clinical signs provided in the survey could have led to underestimation of the prevalence of this factor. Similarly, no significant differences were identified between groups in the prevalence of other signs not included on the list (ie, vomiting, diarrhea, and respiratory distress).
Given the ability of CIV to be transmitted via aerosol secretions, contaminated objects, or people moving between infected and uninfected dogs, and given the possibility of viral shedding prior to onset of clinical signs, it would be reasonable to conclude that exposure to other dogs would be a risk factor for CIV positivity. However, no significant association was identified between CIV positivity and residence in a multidog household in the present study. No questions were included in the survey regarding exposure to boarding facilities, shelters, and daycare facilities, so no inferences can be drawn regarding the transmissibility of CIV within the study population. No information was requested regarding the nature of any dog-to-dog contact, and the number of dogs for which such information was volunteered was too low to allow statistical comparisons.
Vaccination against CIV infection has increased in popularity since reports23,26 were published that use of an H3N8 vaccine can significantly reduce the severity and duration of clinical signs of CIV infection and may decrease viral shedding. Whether this H3N8 CIV vaccine provides any cross protection for dogs infected with the H3N2 subtype remains unknown, and no vaccine against H3N2 CIV infection was available during the study period. The percentage of dogs in the present study receiving a CIV vaccine in the 12 months prior to the veterinary visit that prompted specimen collection (11% of all dogs) was statistically similar between CIV-positive and CIV-negative dogs. The low number of CIV-positive dogs (n = 6) that had received the H3N8 CIV vaccine likely had minimal cross protection from the vaccine when later exposed to H3N2 CIV.
Prevalence of CIV positivity also did not differ significantly between the small number of dogs that had received an intranasal (n = 12) or oral (3) B bronchiseptica vaccine and dogs that had received neither of these. The lack of a specific option in the survey for oral vaccination may have led to underreporting of the use of this vaccine by respondents that did not voluntarily disclose it as free text. Interestingly, dogs vaccinated SC against B bronchiseptica infection were more likely to be CIV positive (n = 17) than dogs that received no such vaccine in univariate analysis, although this variable did not remain significant on multivariate analysis when significance criteria for multiple comparisons were applied. Dogs that routinely received the Bordetella vaccine, which is considered noncore, may have had owners who believed their dog was at greater risk of exposure to respiratory pathogens than dogs that did not receive it. If this were true, then one would expect that dogs that were mucosally vaccinated would have been at increased odds of CIV positivity because they could be considered to have similarly minded owners; however, the data did not support this possibility. Perhaps the nonspecific stimulation of innate mucosal immunity induced by intranasal or oral Bordetella vaccination27 afforded some degree of protection against other pathogens, whereas SC vaccination did not, resulting in an increased risk of CIV positivity in dogs vaccinated SC, which were already members of a group at risk for CIV exposure. Additional research involving a larger number of dogs would be needed to investigate relationships among vaccination type, other concurrent vaccinations, date of vaccine administration prior to illness, and CIV exposure.
One limitation of the study reported here was the potential for recall bias. Distribution of the survey at the time of specimen submission may have led to a larger sample size and limited the amount of recall bias. The survey was designed to be concise, requiring a minimal time commitment from responding veterinarians. Therefore, it was designed as a combination of both multiple choice and fill-in-the-blank questions, with the option to provide additional details if desired. The prevalence of unspecified clinical signs of or unlisted putative risk factors for CIV infection could therefore not be determined. In addition, no survey was returned for a large (100/220 [45%]) proportion of tested dogs, and the reason for this was unknown.
The study findings, particularly regarding the prevalence of CIV infection in certain groups and the case fatality rates, must also be considered in the context in which they were obtained. To identify CIV-positive dogs from the Atlanta area in the summer of 2015, we chose to use records of PCR testing for respiratory pathogens from a single regional reference laboratory that was extensively used by clinics in this region. Therefore, dogs identified as CIV infected through other means would have been excluded, as would dogs with subclinical infection or others that were not taken for veterinary evaluation. Even if dogs were taken for veterinary care, their veterinarians may not have been aware of the option to submit nasal swab specimens for PCR testing for CIV by this regional laboratory. Additionally, tested nasal swab specimens included 1 submitted from a humane society in Illinois, and it was unknown whether the associated CIV-negative dog had a travel history that included the Atlanta area. The study findings should therefore be interpreted with these limitations in mind.
Reported cases of CIV infection can often be traced back to high-volume, high-turnover facilities such as shelters, kennels, and daycare.11,23 Some of the participating veterinarians in the present study provided information regarding dog exposure to such facilities; however, because this information was not specifically requested in the survey, no conclusions could be drawn regarding these data.
The present study provided valuable information concerning the 2015 H3N2 CIV outbreak in the Atlanta area. Given that only 1 CIV-positive dog was known to have died, the virus appeared to be of low lethality and was certainly no more lethal than in dogs with reported H3N8 CIV infection. The CIV-positive dogs were more likely to be adults and have signs of inappetence than CIV-negative dogs. Additional studies involving a broader and larger dog population are warranted to determine whether the interval between exposure to CIV and onset of clinical signs, vaccination history, and immune status of vaccinates (vs nonvaccinates) are associated with a diagnosis of CIV infection.
Acknowledgments
This research was not supported by grant funding. The authors declare that there were no conflicts of interest.
ABBREVIATIONS
| CIV | Canine influenza virus |
Footnotes
Qualtrics 2015, Qualtrics LLC, Provo, Utah.
Microsoft Excel 2013, Microsoft Corp, Santa Rosa, Calif.
LeFleur R. Demonstration of protection against canine influenza virus H3N2 infection following vaccination with inactivated CIV H3N2 (abstr), in Proceedings. Am Coll Vet Intern Med Forum 2016;159.
References
1. Crawford PC, Dubovi EJ, Castleman WL, et al. Transmission of equine influenza virus to dogs. Science 2005;310:482–485.
2. Anderson TC, Crawford PC, Dubovi EJ, et al. Prevalence of and exposure factors for seropositivity to H3N8 canine influenza virus in dogs with influenza-like illness in the United States. J Am Vet Med Assoc 2013;242:209–216.
3. Pecoraro HL, Bennett S, Huyvaert KP, et al. Epidemiology and ecology of H3N8 canine influenza viruses in US shelter dogs. J Vet Intern Med 2014;28:311–318.
4. Payungporn S, Crawford PC, Kouo TS, et al. Influenza a virus (H3N8) in dogs with respiratory disease, Florida. Emerg Infect Dis 2008;14:902–908.
5. Castleman WL, Powe JR, Crawford PC, et al. Canine H3N8 influenza virus infection in dogs and mice. Vet Pathol 2010;47:507–517.
6. Yoon KJ, Cooper VL, Schwartz KJ, et al. Influenza virus infection in racing Greyhounds. Emerg Infect Dis 2005;11:1974–1976.
7. Song D, Kang B, Lee C, et al. Transmission of avian influenza virus (H3N2) to dogs. Emerg Infect Dis 2008;14:741–746.
8. CDC. Key facts about canine influenza (dog flu). Available at: www.cdc.gov/flu/canineflu/keyfacts.htm. Accessed Nov 17, 2015.
9. Deshpande M, Abdelmagid O, Tubbs A, et al. Experimental reproduction of canine influenza virus H3N8 infection in young puppies. Vet Ther 2009;10:29–39.
10. Priestnall SL, Mitchell JA, Walker CA, et al. New and emerging pathogens in canine infectious respiratory disease. Vet Pathol 2014;51:492–504.
11. Dubovi EJ. Canine influenza. Vet Clin North Am Small Anim Pract 2010;40:1063–1071.
12. DVM360. Canine influenza outbreak caused by Asian virus, scientists say. Available at: veterinarynews.dvm360.com/canine-influenza-outbreak-caused-asian-virus-scientists-say. Accessed Aug 1, 2015.
13. AVMA. Canine influenza FAQ: questions, answers, and interim guidelines. Available at: www.avma.org/KB/Resources/FAQs/Pages/Control-of-Canine-Influenza-in-Dogs.aspx?PF=1. Accessed Aug 23, 2016.
14. Cornish A. Chicago-area dog flu outbreak rises to over 1,000 cases. Available at: www.npr.org/2015/04/15/399937728/chicago-area-dog-flu-outbreak-rises-to-over-1-000-cases. Accessed Aug 1, 2015.
15. Firozi P, Rodriguez M, Cancino A. Chicago canine influenza epidemic traced to Asian strain. Available at: www.chicagotribune.com/news/local/breaking/ct-dog-flu-met-0415-20150414-story.html. Accessed Jul 1, 2015.
16. Lourgos A. Dog flu outbreaks temporarily close shelters, suspend adoptions. Available at: www.chicagotribune.com/news/ct-dog-flu-update-met-20150501-story.html. Accessed Jan 8, 2016.
17. Goheen W. Tests confirm canine flu confirmed in Georgia linked to Chicago strain. Available at: www.cbs46.com/story/29083866/first-case-of-canine-flu-confirmed-in-metro-atlanta. Accessed Aug 20, 2015.
18. Tejeda N. Canine flu rising fast in metro Atlanta. Available at: www.fox5atlanta.com/news/2349787-story. Accessed Jul 13, 2015.
19. University of Georgia Veterinary Diagnostic Laboratories. Positive canine influenza case in Georgia. Available at: vet. uga.edu/dlab/news/view/positive_canine_influenza_case_in_georgia. Accessed Nov 17, 2015.
20. Benjamini Y, Hochberg Y. Controlling the false discovery rate—a practical and powerful approach to multiple testing. J R Stat Soc Series B 1995;57:289–300.
21. Rothberg MB, Haessler SD, Brown RB. Complications of viral influenza. Am J Med 2008;121:258–264.
22. Schulz BS, Kurz S, Weber K, et al. Detection of respiratory viruses and Bordetella bronchiseptica in dogs with acute respiratory tract infections. Vet J 2014;201:365–369.
23. Larson LJ, Henningson J, Sharp P, et al. Efficacy of the canine influenza virus H3N8 vaccine to decrease severity of clinical disease after cochallenge with canine influenza virus and Streptococcus equi subsp zooepidemicus. Clin Vaccine Immunol 2011;18:559–564.
24. Niedziela K. New canine flu: what we are learning today. Available at: www.veterinarypracticenews.com/what-were-learning-about-new-canine-flu. Accessed Nov 17, 2015.
25. Buonavoglia C, Martella V. Canine respiratory viruses. Vet Res 2007;38:355–373.
26. Deshpande MS, Jirjis FF, Tubbs AL, et al. Evaluation of the efficacy of a canine influenza virus (H3N8) vaccine in dogs following experimental challenge. Vet Ther 2009;10.
27. Davis R, Jayappa H, Abdelmagid OY, et al. Comparison of the mucosal immune response in dogs vaccinated with either an intranasal avirulent live culture or a subcutaneous antigen extract vaccine of Bordetella bronchiseptica. Vet Ther 2007;8:32–40.
