Outbreaks attributable to infections with CPV and CDV are common in animal shelters and are associated with high morbidity and mortality rates.1 Although vaccination against CPV and CDV is effective for prevention of disease, most dogs that are admitted to shelters have an unknown vaccination history. A study1 in which researchers investigated dogs admitted to a municipal animal shelter in Florida revealed that 142 of 431 (33%) dogs lacked PATs against CPV and 245 of 431 (57%) lacked PATs against CDV at the time of admission. The environment in animal shelters poses a high risk for transmission of infections among animals. Shelter dogs are highly susceptible to infections, which leads to a high risk of infectious disease outbreaks in animal shelters wherever CPV and CDV are endemic in dogs or wildlife in communities where animal shelters are located.
Effective management practices for the prevention and control of infectious disease outbreaks in animal shelters begin with diagnosis of infectious disease in animals that are admitted to the animal shelter and isolation of animals that have clinical signs of disease. These measures should be followed by an assessment of the risk of infection for other animals in the shelter that have been exposed to infectious diseases.2 Serologic testing can be used to assess the immunity of dogs to infection during disease outbreaks in animal shelters.2 Dogs that have a PAT against CPV and CDV are considered to be immune to infection by those viruses.3–7 Therefore, dogs that do not have clinical signs of disease, have been exposed to CPV or CDV, and have a PAT against those viruses are likely to be resistant to infection and can be assigned to a low-risk category.2 For protection of other dogs against infectious diseases during disease outbreaks, dogs without clinical signs of disease but that have been exposed to infectious diseases and that lack a PAT against those diseases should be assigned to a high-risk category. Knowledge of these infection-risk categories can aid animal shelter managers in appropriate separation of dogs during infectious disease outbreaks.
A variety of assays are used to determine antibody titers in dogs. For many assays, diagnostic samples must be shipped to a laboratory, where it may take several days for completion of the assays. This delay can compromise a timely response to a disease outbreak. Titer analysis is further complicated by interlaboratory variation in titers reported for the same samples and a lack of information provided by some laboratories regarding the antibody titer associated with protection against infection. The use of assays that can be rapidly performed to determine whether dogs have PATs against infectious agents would facilitate prompt screening of dogs that do not have clinical signs of disease but have been exposed to those agents. The use of these assays would aid animal shelter personnel in the assignment of dogs to risk groups during disease outbreaks.
We hypothesized that semiquantitative screening tests could be used to determine, in a timely manner, whether dogs housed in an animal shelter have PATs against CPV and CDV. The objective of the study reported here was to evaluate the diagnostic performance of 2 semiquantitative screening assays for identification of dogs that have a PAT against CPV and CDV, to compare results of those assays with results of quantitative criterion-referenced assays, and to determine whether the semiquantitative screening assays would be appropriate for use in assignment of dogs to infection risk categories on the basis of their PAT status.
Materials and Methods
Animals—Dogs admitted to an animal shelter in Florida were included in the present study. The study protocol was approved by the University of Florida Institutional Animal Care and Use Committee.
Sample collection and serologic testing—Blood samples (3 mL) were collected from dogs via jugular or cephalic venipuncture on the day each dog was admitted to the animal shelter and before any vaccine was administered. Blood samples were centrifuged to separate serum. Each serum sample was divided into 3 aliquots, which were stored at −20°C until analysis.
One aliquot of serum was used for determination of endpoint antibody titers with the criterion-referenced methods of HI (for determination of antibody titers against CPV) and VN (for determination of antibody titers against CDV), which were performed at a university-based diagnostic laboratory.a Antibody titers were defined as the reciprocal of the highest serum dilution that inhibited hemagglutination (CPV) or neutralized viral infection (CDV). The diagnostic laboratory reported PATs to be ≥ 80 for CPV8,9 and ≥ 32 for CDV.10
Another aliquot of serum was sent to a commercial diagnostic laboratoryb for determination of antibody titers against CPV and CDV by use of an IFA. This laboratory reported results of the IFA as negative for PATs against CPV and CDV at titers < 5 and as positive for PATs against CPV and CDV at titers ≥ 5. Results were reported within approximately 2 days after sample submission. Titer values indicating positive results for PATs against CPV and CDV (titers ≥ 5) had been established for this IFA by analysis of serum samples from dogs and comparison of these results with those of HI (CPV) and VN (CDV) assays (which had been performed at the same university-based diagnostic laboratory4,a that was used for analysis of serum samples in the present study).
Another aliquot of serum was used for detection of antibodies against CPV and CDV with a point-of-care microwell plate ELISA.c The color intensity in the sample well that was equal to or greater than the color of the positive control well was considered a positive result for the presence of a PAT. The color in the positive-control well of the ELISA corresponded to an anti-CPV antibody titer > 80 (as determined by comparison with HI assay results) or an anti-CDV antibody titer > 16 (as determined by comparison with VN assay results).11 This ELISA was designed for simultaneous analysis of up to 10 serum samples for the identification of dogs with a PAT against CPV and CDV. The ELISA was performed in accordance with the manufacturer's instructions11 by 1 investigator (PCC) who had no knowledge of the results of any of the other assays. Results of the ELISA were available within approximately 20 minutes.
Statistical analysis—Results of the criterion-referenced (HI and VN) assays were compared with results of the ELISA and IFA to determine measures of the diagnostic performance of the assays. The sensitivity (true-positive results/[true-positive results + false-negative results]), specificity (true-negative results/[true-negative results + false-positive results]), PPV (true-positive results/[true-positive results + false-positive results]), NPV (true-negative results/[true-negative results + false-negative results]), overall accuracy ([true-positive results + true-negative results]/all results), and 95% confidence intervals for each of these measures were calculated for the IFA and ELISA. Because predictive values change when the prevalence changes, the PPV and NPV were also calculated for hypothetical populations of dogs with a prevalence of PATs against CPV or CDV of 25%, 50%, and 75%, which approximated conditions that may be found in various populations of dogs. Differences between results of the assays were compared by use of the χ2 test.d Values of P < 0.05 were considered significant.
Results
Serum samples were obtained from 431 dogs admitted to the animal shelter. Of the 431 dogs, 289 (67%) were identified on the basis of results of the HI assay to have a PAT against CPV and 186 (43%) were identified on the basis of the VN assay to have a PAT against CDV.
Specificity of the ELISA was significantly (P < 0.001) higher than specificity of the IFA for correct identification of dogs that had a PAT against CPV and CDV (Table 1). Sensitivity of the IFA was not significantly (P = 0.62) different from sensitivity of the ELISA for correct identification of dogs that had a PAT against CPV. Sensitivity of the IFA was significantly (P < 0.001) higher than that of the ELISA for correct identification of dogs that had a PAT against CDV. Overall accuracy of the ELISA was significantly (P < 0.001) higher for correct identification of dogs that had a PAT against CPV and CDV than was overall accuracy of the IFA. Diagnostic performance of the IFA was improved by designating titers of > 5 to be indicative of a PAT instead of the laboratory-defined value of ≥ 5. By use of a titer cutoff value of > 5, sensitivity of the IFA for correct identification of dogs that had a PAT decreased slightly from 97% to 96% for CPV and decreased from 97% to 95% for CDV; however, specificity for correct identification of dogs that had a PAT increased from 82% to 92% for CPV and increased from 70% to 83% for CDV.
Measures of diagnostic performance of 2 assays for identification of dogs (n = 431) that had a PAT against CPV and CDV at the time they were admitted to a municipal animal shelter.
| Test method | Sensitivity (95% CI) | Specificity (95% CI) | Overall accuracy (95% CI) | PPV (95% CI) | NPV(95%CI) |
|---|---|---|---|---|---|
| CPV | |||||
| Point-of-care ELISA | 98 (96–99) | 98 (96–100)* | 98 (96–99)* | 99 (98–100)* | 95 (92–99) |
| Reference laboratory IFA | 97 (94–99) | 82 (76–89) | 92 (89–94) | 92 (89–95) | 92 (87–97) |
| CDV | |||||
| Point-of-care ELISA | 88 (83–92) | 95 (92–98)* | 92 (89–94)* | 93 (89–97)* | 91 (88–95) |
| Reference laboratory IFA | 97 (95–100)* | 70 (64–76) | 82 (78–86) | 71 (66–77) | 97 (95–100)* |
Values reported are percentages.
Within a virus, value is significantly (P < 0.001) greater than the value for the alternate testing method.
CI = Confidence interval.
The predictive value of a positive or negative result for a PAT was calculated by means of the actual prevalence of PATs against CPV and CDV in serum samples from the present study (67% for CPV and 43% for CDV; Table 1). The PPV for correct identification of dogs that had a PAT against CPV or CDV was significantly (P < 0.001) higher for the ELISA than it was for the IFA, but the NPV for correct identification of dogs that did not have a PAT against CDV was significantly (P = 0.018) higher for the IFA than it was for the ELISA. The PPV and NPV of the ELISA and IFA were also calculated for hypothetical populations of dogs that had a prevalence of PATs against CPV or CDV of 25%, 50%, and 75% (Table 2).
Calculated predictive values of 2 assays for identification of dogs that had a PAT against CPV and CDV in 3 hypothetical populations of dogs with a prevalence of PATs against CPV or CDV of 25%, 50%, and 75%.
| 25% prevalence | 50% prevalence | 75% prevalence | ||||
|---|---|---|---|---|---|---|
| Test method | PPV | NPV | PPV | NPV | PPV | NPV |
| CPV | ||||||
| Point-of-care ELISA | 94 | 99 | 98 | 98 | 99 | 94 |
| Reference laboratory IFA | 64 | 99 | 84 | 96 | 94 | 90 |
| CDV | ||||||
| Point-of-care ELISA | 85 | 96 | 95 | 89 | 98 | 73 |
| Reference laboratory IFA | 52 | 99 | 76 | 96 | 91 | 89 |
Values reported are percentages.
Differences between results of the ELISA and IFA and results of the criterion-referenced assays were evaluated. For results of the ELISA, most differences in the results for the criterion-referenced assays were detected for samples that had an antibody titer that was within a single serum dilution of the cutoff value for a PAT against CPV (titers ≥ 80 [as determined with the HI assay]) and CDV (titers ≥ 32 [as determined with the VN assay]). For anti-CPV antibody assays, 7 of 10 (70%) results of the ELISA that differed from results of the criterion-referenced (HI) assay had a titer of 40 through 160 in the HI assay; false-positive results were detected for samples that had an anti-CPV antibody HI titer of 10 through 40, and false-negative results were detected for samples that had an anti-CPV antibody HI titer of 80 through 320. For anti-CDV antibody assays, 20 of 35 (57%) results of the ELISA that differed from results of the criterion-referenced (VN) assay had a titer of 24 through 64 in the VN assay; false-positive results were detected for samples that had an anti-CDV antibody VN titer of 4 through 24, and false-negative results were detected for samples that had an anti-CDV antibody VN titer of 32 through 1,024.
A majority (90/113 [80%]) of the differences between results of the IFA and results of the criterion-referenced assays were detected for samples that had an anti-CPV or anti-CDV antibody IFA titer result that differed from the criterion-referenced assay titer result by more than a single dilution. This indicated that results of the IFA deviated more from the criterion-referenced assay results than did results of the ELISA. For anti-CPV antibody assays, 10 of 35 (29%) results of the IFA that differed from results of the criterion-referenced (HI) assay had a titer of 40 through 160 in the HI assay; false-positive results were detected for samples that had an anti-CPV antibody HI titer of 10 through 40, and false-negative results were detected for samples that had an anti-CPV antibody HI titer of 80 through 640. For anti-CDV antibody assays, 13 of 78 (17%) results of the IFA that differed from results of the criterion-referenced (VN) assay had a titer of 24 through 64 in the VN assay; false-positive results were detected for samples that had an anti-CDV antibody VN titer of 4 through 24, and false-negative results were detected for samples that had an anti-CDV antibody VN titer of 32 through 96. Overall, results that incorrectly classified the anti-CPV and anti-CDV PAT status of dogs were significantly (P < 0.001) more likely to be close to the criterion-referenced assay antibody titer cutoff value for the ELISA (27/45 [60%]) than for the IFA (23/113 [20%]).
Discussion
Results of the point-of-care ELISA were more similar to results of the criterion-referenced assays than were results of the diagnostic laboratory IFA for determination of CDV and CPV PAT status of dogs. The ELISA had higher specificity for identification of dogs that had CDV and CPV PATs than did the IFA, which resulted in fewer false-positive results for the ELISA. The predictive value of a positive result for a CDV or CPV PAT was greater for the ELISA than for the IFA, regardless of the prevalence of PATs in the hypothetical populations of dogs that were analyzed.
Each assay had limitations. At the time the present study was conducted, the point-of-care ELISA cost approximately $10/sample, not including the cost for staff time, when samples were assayed as a batch (ie, all available ELISA assay wells were used). This ELISA could be performed at a shelter on an as-needed basis and requires a serum or plasma sample volume of only 1 μL. Completion of the assay required approximately 20 minutes, but it was a multistep procedure requiring performance by a proficient technician, and it could be a time-consuming process for staff at an animal shelter if there are many dogs to evaluate. The IFA cost $41/sample, not including the cost of shipping, regardless of the number of samples submitted. The laboratory requested a serum or plasma sample volume of 1 mL, but less staff time was needed to perform the IFA, compared with performance of the ELISA, because the samples were sent to a diagnostic laboratory not associated with the animal shelter. Results of the IFA were not available as quickly as results of the ELISA because of the time required for shipping and laboratory analysis and the number of days the laboratory was closed. This delay in availability of results, which can range from 2 to 4 days, could be problematic during disease outbreaks in which efficient and prompt triage of dogs on the basis of their PAT status is important.
The point-of-care ELISA, which had high diagnostic accuracy in the present study and could be performed by animal shelter staff in < 1 hour, can be a useful tool for efficient triage of dogs that are exposed to infectious agents during outbreaks in shelters as a result of infection with CPV or CDV.12 Use of the ELISA may enable shelters to improve the chances of survival of many dogs and may aid in preventing the spread of infection to other dogs and inadvertent adoption of infected dogs. It may also help prevent shelter personnel from having to use drastic measures such as depopulation of animals. Healthy dogs that are at least 5 months old, that have been exposed to infectious agents, and that have PATs against those agents are considered to have a low risk for infection and can be adopted or transferred to other agencies.12 Puppies that are < 5 months old, that have been exposed to infectious agents, and that have PATs against those agents pose a challenge for correct classification into low-risk or high-risk categories of susceptibility to infection because passively acquired maternal antibodies make it difficult to interpret antibody titers. Maternally derived antibodies provide some protection against infection, but this protection is not as complete as it is in dogs that have both humoral and cell-mediated immunity (as a result of previous exposure to infectious agents or prior vaccination), which allow a rapid anamnestic response to occur.12 In addition, the concentrations of circulating maternal antibodies rapidly decrease in puppies, which results in a loss of PATs and renders the puppies susceptible to infection.9 Therefore, puppies that are < 5 months old should be considered to have a high risk of infection attributable to CPV or CDV, regardless of their antibody titers. Healthy dogs that are exposed to infectious agents and lack PATs against those agents should be considered to have a high risk for infection, regardless of their age. High-risk dogs should be revaccinated and quarantined (in a segregated area of a shelter or in an off-site location such as a foster home) for the incubation period of the virus to which they have been exposed.
The implications of inaccurate results must be considered when choosing assays for detection of antibodies against infectious agents. Failure to identify dogs that have a PAT against infectious diseases (false-negative results) may result in a low-risk dog being kept unnecessarily in quarantine at a shelter or being euthanized if quarantine space is not available. Failure to detect dogs that do not have PATs against infectious diseases (false-positive results) may result in a high-risk dog that has been exposed to an infectious agent being kept in the shelter population or being adopted, which would potentially facilitate the spread of infection to other dogs and undermine interventions by shelter personnel to prevent disease outbreaks. Use of a screening test that has a high specificity for detection of PATs, such as the point-of-care ELISA used in the present study, provides the best opportunity for successful management of a disease outbreak. The higher rate of false-positive results of the diagnostic laboratory IFA (vs the ELISA) used in the present study suggests that the IFA should not be used to identify dogs that have a PAT against CPV or CDV during an infectious disease outbreak in an animal shelter.
ABBREVIATIONS
| CDV | Canine distemper virus |
| CPV | Canine parvovirus |
| HI | Hemagglutination inhibition |
| IFA | Immunofluorescence assay |
| NPV | Negative predictive value |
| PAT | Protective antibody titer |
| PPV | Positive predictive value |
| VN | Virus neutralization |
Animal Health Diagnostic Center, College of Veterinary Medicine, Cornell University, Ithaca, NY.
Antech Diagnostics, Irvine, Calif.
TiterCHEK, Synbiotics Corp, San Diego, Calif.
EpiInfo, version 3.5.1, CDC, Atlanta, Ga.
References
- 1.↑
Lechner ES, Crawford PC, Levy JK, et al. Prevalence of protective antibody titers for canine distemper virus and canine parvovirus in dogs entering a Florida animal shelter. J Am Vet Med Assoc 2010; 236:1317–1321.
- 2.↑
Hurley KF. Outbreak management. In: Miller L, Hurley K, eds. Infectious disease management in animal shelters. Ames, Iowa: Wiley-Blackwell, 2009;39–48.
- 3.
McCaw DL, Thompson M, Tate D, et al. Serum distemper virus and parvovirus antibody titers among dogs brought to a veterinary hospital for revaccination. J Am Vet Med Assoc 1998; 213:72–75.
- 4.↑
Twark L, Dodds WJ. Clinical use of serum parvovirus and distemper virus antibody titers for determining revaccination strategies in healthy dogs. J Am Vet Med Assoc 2000; 217:1021–1024.
- 5.
Moore GE, Glickman LT. A perspective on vaccine guidelines and titer tests for dogs. J Am Vet Med Assoc 2004; 224:200–203.
- 6.
Horzinek MC. Vaccine use and disease prevalence in dogs and cats. Vet Microbiol 2006; 117:2–8.
- 7.
Burr P. Serological testing—an alternative to boosters? Vet Microbiol 2006; 117:39–42.
- 8.
Carmichael LE, Joubert JC, Pollack RV. Hemagglutination by canine parvovirus: serologic studies and diagnostic applications. Am J Vet Res 1980; 41:784–791.
- 9.↑
Pollack RV, Carmichael LE. Maternally derived immunity to canine parvovirus infection: transfer, decline, and interference with vaccination. J Am Vet Med Assoc 1982; 180:37–42.
- 10.↑
Appel M, Robson DS. A microneutralization test for canine distemper virus. Am J Vet Res 1973; 34:1459–1463.
- 11.↑
Canine distemper-parvovirus antibody test kit [package insert]. Kansas City, Mo: Synbiotics Corp, 2010.
- 12.↑
Larson LJ, Newbury S, Schultz RD. Canine and feline vaccinations and immunology. In: Miller L, Hurley K, eds. Infectious disease management in animal shelters. Ames, Iowa: Wiley-Blackwell, 2009;61–82.
