Materials and Methods

Dogs and samples—Blood and urine samples (3 mL each) were obtained from 70 dogs that had been categorized into 1 of 4 groups, including those with blastomycosis (n = 21), those with histoplasmosis (8), those with nonfungal pulmonary disease (21), and healthy dogs from a blastomycosis endemic area (control group; 20).

Dogs with blastomycosis were prospectively enrolled in the study if the patient had a definitive diagnosis of blastomycosis made by a board-certified veterinary clinical or anatomic pathologist on the basis of cytologic or histopathologic findings during initial diagnostic testing. Dogs in the nonfungal pulmonary disease group were prospectively selected on the basis of clinical findings closely resembling pulmonary blastomycosis. All of these dogs had blastomycosis as one of the differential diagnoses to be excluded after analysis of thoracic radiographs by a board-certified veterinary radiologist. The diagnosis of nonfungal pulmonary disease was made by evaluation of results from bronchoalveolar lavage or transtracheal washing, fine-needle aspirate with cytologic evaluation, surgical biopsy, necropsy, or resolution of clinical signs without antifungal treatment. Blood and urine samples were collected prior to definitive diagnosis of either blastomycosis or nonfungal pulmonary disease. Samples were then grouped according to results from the full diagnostic evaluation. Control samples were collected from healthy dogs recruited from the University of Minnesota Veterinary Medical Center general practice. Dogs were classified as being healthy by obtaining a thorough medical history and completing a physical examination. This part of the study was approved by the University of Minnesota Institutional Animal Care and Use Committee. Informed owner consent was obtained prior to collection and use of samples for the study. Samples collected from the University of Minnesota (for all dogs other than those with histoplasmosis) were obtained and submitted between October 2008 and December 2010.

Samples from dogs with histoplasmosis were obtained from investigators at 2 other veterinary teaching hospitals. These dogs had a definitive diagnosis of histoplasmosis made on the basis of cytologic or histologic evaluation by a board-certified veterinary clinical or anatomic pathologist and were enrolled in another study by veterinarians at the Oklahoma State University Center for Veterinary Health Sciences (Dr. Andrew Hanzlicek) and Kansas State University (Dr. Katherine KuKanich). This part of the study was approved by the Institutional Animal Care and Use Committees of Oklahoma State University Center for Veterinary Health Sciences and Kansas State University. Samples from dogs with histoplasmosis were collected from March 2012 to January 2014.

Sample handling—Blood samples were allowed to rest at room temperature (approx 21°C) for 15 minutes and centrifuged at 1,000 × g for 10 minutes, and then separated serum was collected. Serum and urine samples obtained at the University of Minnesota were stored at −20°C for variable periods of time until shipment to the diagnostic laboratory. Serum and urine samples were submitted to the primary diagnostic laboratory^a for antigen^b and antibody^c testing by EIA. Serum was also submitted to a second diagnostic laboratory^d for antibody detection by A-antigen AGID; however, a few dogs had AGID testing performed at the primary laboratory. Errors in sample handling or insufficient volume of serum caused incomplete testing for 2 dogs in the blastomycosis group (1 dog had no serum available for antigen or antibody EIA testing and 1 dog had no AGID performed) and 5 dogs in the nonfungal pulmonary group (3 dogs had insufficient serum for antigen testing, 3 dogs had no AGID performed, and 2 dogs had no antibody EIA performed). For the histoplasmosis group, blood and urine samples were obtained at Oklahoma State University and Kansas State University and processed as described for other samples. Samples were refrigerated and placed in −80°C freezers within 4 hours after collection. Specimens were stored at −20°C on arrival at the diagnostic laboratory^a until testing was performed. Serum from dogs with histoplasmosis was insufficient for testing in the quantitative antigen EIA.

Quantitative B dermatitidis antigen EIA—Use of the quantitative EIA for detection of B dermatitidis antigen^b has been described elsewhere.^5,6 Briefly, anti-Blastomyces rabbit IgG was attached to wells of a 96-well microtiter plate for antigen capture from samples. Biotinylated anti-Blastomyces rabbit IgG was applied to bind to the antigen, and streptavidin-labeled horseradish peroxidase was used as the detector enzyme. In the final step, a substance containing the enzyme's substrate was added. The subsequent reaction produced a detectable color signal, which was proportional to the amount of enzyme in the well. The OD was determined at a wavelength of 450 nm with a reference wavelength of 620 nm.

The concentration of antigen in serum and urine samples was determined by comparison with standards containing known concentration of B dermatitidis antigen. Antigen standards were prepared from a pool of urine specimens from patients with strong-positive antigen results. Dilutions of the urine pool were prepared, and the concentration of Blastomyces galactomannan antigen in those dilutions was determined by comparing them to a standard curve of the galactomannan purified from Blastomyces yeast. The standards ranged from 0 to 14.7 ng/mL

rBAD-1 antibody EIA—The rBAD-1 EIA was modified from a recently described EIA⁷ that used a native surface adhesion protein (BAD-1) extracted from B dermatitidis for antibody capture. In the assay used for the present study, a recombinantly produced portion of the BAD-1 protein containing an immunodominant 25-amino acid repeat sequence was used for antibody capture.¹⁰ The recombinant protein was expressed in Escherichia coli as described previously.¹⁰ Microplates^e were coated with a solution containing 35 ng of rBAD-1/mL in 0.01M Tris-HCl (pH 8.0). One hundred microliters of the serum to be tested was diluted 1:1,000 and was added to each well, and plates were incubated at 37°C for 1 hour. Plates were then washed, and 100 μL of biotinylated goat anti-dog IgG antibody^f (diluted to 1:3,330) was added to each well. Following incubation at 37°C for 1 hour, plates were again washed and 100 μL of streptavidin-horseradish peroxidase was added to each well. Plates were incubated at 37°C for 1 hour and washed for a third time, and 100 μL of 3,3',5,5'-tetra-methylbenzidine substrate was added to each well. The prepared samples were incubated for 8 minutes at room temperature (approx 21°C) in the dark. Sulfuric acid (100 μL) was then added to each well to stop the reaction. Color development was assessed by measurement of OD in a plate reader at a wavelength of 450 nm with a reference wavelength of 620 nm. Results were expressed as EIA units by dividing the OD for the test specimen by the cutoff OD, which was determined by ROC curve analysis³ of serum samples from 20 dogs with blastomycosis, 21 dogs with nonfungal pulmonary disease, and 20 healthy control dogs. Results > 1.0 unit were considered positive.

Intra-assay (repeatability) and interassay (reproducibility) precision (CVs) were assessed by testing, in triplicate, calibrator samples with high, moderate, and low concentrations of antibody as well as a negative control sample on each of 10 days by 2 operators, providing 60 data points for each observation. Calibrators were prepared from pooled strong positive serum samples and diluted in normal serum to produce dilutions containing high, moderate, and low concentrations of antibody as assessed by OD. Reproducibility of individual patient results was determined by testing 27 serum samples (19 from dogs with blastomycosis and 8 from dogs with histoplasmosis) on 2 occasions.

A-antigen antibody AGID—Serum samples were tested for antibodies against the A-antigen of B dermatitidis with a conventional AGID assay. Tests were performed by 1 of 2 reference laboratories^a,d with commercial reagents according to manufacturer's recommendations.^g–i The serum specimens were tested undiluted, and outcomes were visually assessed. Samples having a line of identity with the positive control sample provided for the test kit were regarded as positive.

Statistical analysis—Statistical analysis was performed with commercially available software.^j For the rBAD-1 antibody EIA, an ROC curve analysis was used to determine the cutoff for detection of antibody with the assay. Intra-assay and interassay CVs were determined with standard calculations. Reproducibility of the individual results for samples from dogs with blastomycosis was assessed by linear regression analysis. Sensitivities of each diagnostic test were determined by evaluation of results for samples from dogs with blastomycosis, and specificities were determined by evaluation of results for samples from dogs with histoplasmosis, dogs with nonfungal pulmonary disease, and healthy control dogs. A 2-tailed Fisher exact test was used to compare the results between antibody test methods. The 95% confidence intervals were calculated for sensitivity and specificity of the various tests. Values of P < 0.05 were considered significant.

Results

The 21 dogs with blastomycosis included 11 Labrador Retrievers, 3 Golden Retrievers, 2 Cocker Spaniels, and 1 each of 5 other breeds (Basenji, Shar-Pei, Iceland Sheepdog, English Springer Spaniel, and Beagle). Median age of dogs with blastomycosis was 5.3 years (range, 0.9 to 10.6 years), and median weight was 28.6 kg (62.9 lb; range, 10.5 to 40.5 kg [23.1 to 89.1 lb]); 6 dogs were sexually intact males, 9 were neutered males, 2 were sexually intact females, and 4 were spayed females. Only 1 dog with blastomycosis had received immunosuppressive medications prior to diagnosis; this dog had been treated with prednisone (0.55 mg/kg [0.25 mg/lb], PO, q 24 h) at the time of sample collection, but the medication had been discontinued after diagnosis.

The 21 dogs with nonfungal pulmonary disease included 6 Golden Retrievers, 2 Labrador Retrievers, 3 mixed-breed dogs, 2 Bernese Mountain Dogs, and 1 each of 8 other breeds (Doberman Pinscher, Siberian Husky, German Shorthair Pointer, Standard Poodle, Mastiff, Yorkshire Terrier, English Pointer, and Vizsla). Median age of these dogs was 7.75 years (range, 0.7 to 14.2 years), and median weight was 28.2 kg (62.0 lb; range, 3.2 to 75.0 kg [7.04 to 165.0 lb]); 1 dog was a sexually intact male, 9 were neutered males, and 11 were spayed females. Diagnoses for these dogs included carcinoma (n = 8), sarcoma (2), bacterial pneumonia (2), hemangiosarcoma (2), and eosinophilic bronchopneumopathy (2) as well as lymphoma, squamous cell carcinoma, histiocytic sarcoma, intrapulmonary hemorrhage secondary to anticoagulant rodenticide toxicosis, and melanoma (1 each).

The 8 dogs with histoplasmosis included 2 Schnauzers, 2 mixed-breed dogs, 1 Boston Terrier, 1 Shih Tzu, and 1 dog of unknown (hound-type) breed. Median age of these dogs was 6 years (range, 1 to 13 years), and median weight was 7.6 kg (16.7 lb; range, 5 to 10 kg [11 to 22 lb]). One was a sexually intact male, 3 were neutered males, and 3 were spayed females; age, sex, breed, and weight were not recorded for one dog, and weight was not recorded for another.

The 20 healthy dogs from which control samples were obtained included 8 Labrador Retrievers, 2 Golden Retrievers, 1 Cocker Spaniel, 2 mixed-breed dogs, and 1 each of 7 other breeds (German Shorthair Pointer, Miniature Schnauzer, Maltese, English Setter, West Highland White Terrier, Coonhound, and Standard Poodle). Median age of the healthy dogs was 5.8 years (range, 1.1 year to 9.3 years), and median weight was 26.6 kg (58.5 lb; range, 5.0 to 42.3 kg [11 to 93.06 lb]). Four were sexually intact males, 8 were neutered males, and 8 were spayed females.

Results of the 3 assays for assisting in the diagnosis of blastomycosis were summarized (Table 1). Sensitivity of the quantitative B dermatitidis antigen EIA was 100% in serum and urine samples from dogs with blastomycosis. Specificity of this assay was 100% in serum samples from the nonfungal pulmonary disease and healthy control groups, and 95%, 25%, and 100% in urine samples from the nonfungal pulmonary disease, histoplasmosis, and healthy control groups, respectively. Cross-reactive antigen was detected in 6 of 8 urine samples from dogs with histoplasmosis. A urine sample from 1 dog in the nonfungal pulmonary disease group initially tested positive (indicating an antigen concentration of 2.1 ng/mL) with the quantitative antigen EIA. This dog had a diagnosis of eosinophilic bronchopneumopathy, and a urine specimen obtained 21 days later tested negative for the antigen. A serum sample from the same dog had negative results for the quantitative antigen EIA, rBAD-1 antibody EIA, and A-antigen antibody AGID assay, and the patient had clinical improvement following corticosteroid treatment.

For pooled serum samples with high, moderate, and low anti–rBAD-1 concentrations (mean ODs of 0.830, 0.501, and 0.260, respectively), the intra-assay CVs were 16.8%, 10.3%, and 9.0%, with interassay CVs of 21.4%, 18.7%, and 14.5%, respectively. For negative controls (mean OD, 0.020), the intra-assay CV was 11.9% and interassay CV was 20.2%. The optimal OD cutoff for detection of anti–rBAD-1 antibody with the EIA was determined to be 0.143 (Figure 1).

View Full Size

Receiver operating characteristic curve showing sensitivity and specificity of the rBAD-1 antibody EIA in dogs with and without blastomycosis. Serum samples obtained from 20 dogs with blastomycosis, 21 dogs with nonfungal pulmonary disease, and 20 healthy control dogs were used in the analysis. The optimal OD cutoff for the test was determined to be 0.143. The area under the curve was 0.995 (95% confidence interval, 0.907 to 1.000; P < 0.001).

Citation: Journal of the American Veterinary Medical Association 247, 10; 10.2460/javma.247.10.1133

• Download Figure
• Download figure as PowerPoint slide

Sensitivity of the rBAD-1 antibody EIA for detection of blastomycosis was 95.0%, compared with 65% for the A-antigen antibody AGID (P = 0.011; Table 1). Anti–rBAD-1 antibody concentrations as determined by EIA ranged from 1.16 to 21.30 EIA units (Figure 2). Serum from the 1 dog that received glucocorticoid treatment had an antibody concentration of 1.75 units. Specificity of the rBAD-1 antibody EIA was 100%, 88%, and 95% in serum samples from the nonfungal pulmonary disease, histoplasmosis, and healthy control groups, respectively. The 2 false-positive results with this assay (1 each in the histoplasmosis and healthy control groups) were weakly positive (1.7 and 1.34 EIA units, respectively). Specificity of the A-antigen antibody AGID was 100% in serum samples from all groups of dogs. Specificity did not differ significantly (P = 1.000) between the 2 antibody tests in histoplasmosis group samples.

View Full Size

Concentrations (EIA units) of anti–rBAD-1 antibody in serum samples from dogs with blastomycosis (n = 20), histoplasmosis (8), or nonfungal pulmonary disease (19) and from healthy control dogs from a blastomycosis endemic area (20). Dashed lines indicate the cutoff (determined by ROC curve analysis) for a positive test result. Numbers in parentheses below the line indicate the number of dogs in each group with a negative test result.

Citation: Journal of the American Veterinary Medical Association 247, 10; 10.2460/javma.247.10.1133

• Download Figure
• Download figure as PowerPoint slide

The rBAD-1 antibody EIA results were reproducibly positive for samples from 19 of 20 dogs with blastomycosis and reproducibly negative for 1. Individual results for the 2 test runs were closely correlated (R² = 0.99; P < 0.001) in this group. Results were reproducibly positive for samples from 1 of 8 dogs with histoplasmosis and reproducibly negative for the remaining 7 with this assay.

Discussion

The rBAD-1 antibody EIA evaluated in the present study had significantly greater sensitivity than did the A-antigen antibody AGID assay (95% vs 65%, respectively) in dogs with blastomycosis. Historically, commercially available tests for detection of anti– B dermatitidis antibodies had been limited to A-antigen antibody AGID.^1,3,11 The AGID method has been widely viewed to have a low degree of utility in the diagnosis of blastomycosis owing to concerns about poor sensitivity and specificity.^1,12,13 Our results were in agreement with previous observations of poor sensitivity of the A-antigen antibody AGID assay (17%, 41%, and 50%) when used to test canine serum.^1,3,14 Also, an EIA for detection of antibodies against the A antigen (although not available for commercial testing) was previously reported to be less sensitive, compared with a BAD-1 antibody EIA, in serum from dogs with blastomycosis.³

Klein and Jones¹⁵ described use of a radioimmunoassay to test human serum for antibodies against BAD-1 extracted from B dermatitidis organisms; in that study, sensitivity of the assay was 85% in samples from 68 people with blastomycosis, and specificity was 94% in samples from 34 individuals with histoplasmosis.¹⁵ Klein et al¹⁴ later reported that the same radioimmunoassay had 92% sensitivity and 100% specificity in samples from 112 dogs with and 99 dogs without blastomycosis, respectively, but dogs with histoplasmosis were not evaluated. In that same study,¹⁴ sensitivity of A-antigen antibody AGID, used in samples from 70 dogs with blastomycosis, was 41%. Boyd et al¹⁶ reported detection of antibodies in canine samples by EIA where B dermatitidis yeast lysates were used as the antigen but did not report sensitivity or specificity of the tests. The results of the present study support the utility of the rBAD-1 antibody EIA as an aid in diagnosis of blastomycosis in dogs. Precision and reproducibility of the rBAD-1 antibody EIA results for serum samples from dogs in the present study compared favorably to those of other fungal antibody EIAs in human medicine.⁷ Our findings that the rBAD-1 antibody EIA had high degrees of sensitivity and specificity for the detection of anti-B dermatitidis antibody in canine serum will require further verification in a prospective study that includes a larger number of dogs.

The quantitative B dermatitidis antigen EIA evaluated in the present study is currently used in the diagnosis of blastomycosis in dogs; however, the assay fails to differentiate B dermatitidis antigen from Histoplasma capsulatum antigen.⁵ Therefore, the test is essentially nonspecific in dogs with either fungal disease. In the present study, the rBAD-1 antibody EIA had good specificity (88%) in samples from 8 dogs with histoplasmosis, similar to findings reported for BAD-1 antibody EIA testing of larger number of humans (n = 50) with histoplasmosis (94%).⁷ Specificity of the rBAD-1 antibody EIA was 100% in samples from dogs with nonfungal pulmonary disease (n = 19) and 95% in healthy control dogs (20) from a blastomycosis endemic area. These findings support the idea that testing for antibodies against rBAD-1 may aid in determining the correct diagnosis (histoplasmosis or blastomycosis) in dogs with positive results for the quantitative antigen EIA.

Antibody detection may also be useful for diagnosis of blastomycosis in dogs with false-negative antigen test results. This hypothesis could not be addressed in the present investigation because all samples from dogs with blastomycosis tested positive for the B dermatitidis antigen. Although sensitivity of this antigen test was 100% in this study, a somewhat lower value (93%) has been previously reported.³ Furthermore, veterinarians in clinical practice have communicated with the authors regarding specific cases of blastomycosis in dogs that tested negative for B dermatitidis antigen with this EIA. In clinical practice, false-negative antigen results (if test sensitivity is incorrectly assumed to be 100%) may lead to treatment delays or owner-elected euthanasia. Improved antibody assays used in conjunction with antigen tests may help to avoid such poor outcomes.

Several limitations of the study reported here should be noted. The sample size was small, reducing the power of the study to determine sensitivity and specificity of the assays used. Samples from only 8 dogs with histoplasmosis were included, which compromised the assessment of test specificity for patients with that disease. A large proportion (16/21) of dogs with nonfungal pulmonary disease that contributed samples for this study had various forms of malignancy. Although malignancy is an important differential diagnosis when evaluating dogs for blastomycosis, the study did not completely address the at-risk population, which would be more diverse. Only dogs with definitive diagnoses were included in the study, thus limiting the application of the findings to clinical practice. A preferential enrollment of dogs with systemic blastomycosis (vs cutaneous or ocular blastomycosis only) likely biased toward enrollment of patients that were likely to have circulating concentrations of B dermatitidis antigen. Where all dogs with nonfungal pulmonary disease were not necropsied, overestimation of sensitivity may have resulted. It should also be considered that administration of high doses of glucocorticoids or other immunosuppressive medications may hinder the ability to detect sufficient antibody concentrations in dogs with blastomycosis by use of the rBAD-1 antibody EIA; this effect could not be evaluated in the present study and requires further investigation.