Introduction

Coccidioidomycosis, commonly called Valley Fever, is a fungal infection caused by Coccidioides immitis or Coccidioides posadasii and is endemic to the southwest region of the United States.¹ Like other pathogenic fungi, Coccidioides spp. are primarily acquired via inhaling arthroconidia following soil disturbances. Infection with the organism causes clinical disease in a variety of mammalian species including humans, dogs, cats, and several livestock and exotic species.^1–5 The majority of clinical infections are reported in humans and dogs.^4,5 Infection initially affects the lungs and can thereafter disseminate to other tissues.¹ Reportedly, 20–42% of clinical cases demonstrate lesions from disseminated disease.^1,6 Clinical infection can present with varied and nonspecific signs including fever, lethargy, weight loss, and respiratory signs.^1,2 Disseminated coccidioidomycosis can manifest with dermatologic, ocular, or neurologic signs, cardiac lesions, and osteomyelitis.^1–3,5,7 This varied presentation of disease creates a diagnostic challenge to veterinarians.⁴

Clinical suspicion of coccidioidomycosis can be based on presentation, clinical signs, radiography, and/or serology. Antibody tests are considered the traditional screening test for infection. Serology tests usually evaluate both immunoglobulin (Ig) G and IgM, with serial dilutions performed of IgG antibody to quantify a titer.² IgM is thought to respond early in infection, whereas IgG levels elevate later in the disease process and remain persistently elevated.⁸ Some literature suggests that a presumptive diagnosis of coccidioidomycosis can be made with an IgG titer ≥1:16 in a patient with clinical signs of disease.¹ Other studies have found a lack of correlation between titer values and disease, with some dogs remaining seronegative despite having lesions of disseminated disease.^4,5 Additional research reports that IgG antibodies may remain elevated for years following infection.⁴ Essentially, studies thus far have not demonstrated a consistent association between disease and antibody tests. Therefore, a definitive diagnosis of coccidioidomycosis can only be confirmed by observing fungal organisms on cytology, histopathology, or fungal culture.^4,5

The objective of this study was to analyze the signalment, serologic test results, and lesion location of cases of canine coccidioidomycosis confirmed via histopathology. As part of this evaluation, we sought to establish the sensitivity of serology testing used in the included cases. We hypothesized that there would be a significant number of seronegative cases with confirmed histopathologic diagnosis of coccidioidomycosis. An additional goal of this study was to evaluate whether a particular lesion associated with disseminated disease was related to serology results. We hypothesized that lesion location would affect antibody test results in that bone lesions would be more likely to have negative antibody testing compared with other lesions reported. To the authors’ knowledge, this is the first study evaluating the correlation between histopathology and serology results including serology sensitivity for coccidioidomycosis in dogs.

Materials and Methods

The database of a commercial laboratory was searched for canine samples with a diagnosis of coccidioidomycosis on histopathology and had prior serology results available over a 2 yr period inclusive of 2012 and 2013. Cases were included if serology testing was performed and reported within 3 mo prior to histopathology of lesions during the 2 yr search period specified. Reports from the cases that met the inclusion criteria were reviewed. Data collected included patient signalment (age, breed, gender), anatomic location of lesion, type of biopsy (incisional or excisional), and results of agar gel immunodiffusion assay (AGID) testing for IgG and IgM with serial dilution to determine IgG titers for Coccidioides spp.

All samples submitted for histopathology were routinely processed and stained with hematoxylin and eosin. In some cases, periodic acid Schiff and/or Grocott methenamine silver were used to identify spherules in the tissue. All cases included in the study had characteristic pyogranulomatous or granulomatous inflammation with identification of Coccidioides spp. spherules and/or endospores. Histopathology was performed by multiple board-certified veterinary pathologists.

Serology tests were performed using a commercially available kit for AGID that detects IgM and IgG antibodies against Coccidioides. When IgG antibody tests were positive, immunodiffusion with serial dilution was then further used to quantitate the titers.

The sensitivity of serology testing in the case population was determined for both IgG and IgM testing by dividing the number of cases with positive antibody tests by the total number of cases to determine the percent of true positives identified.⁹

Fisher exact tests were used to evaluate if there was an association between lesion location and IgG serology results, or more specifically, if biopsy results of a particular location of disseminated disease were associated with a negative IgG antibody test.⁹ Lesion sites were selected for evaluation if the biopsy site was identified to have a diagnosis of coccidioidomycosis from histopathology but had negative serology testing. Four Fisher exact tests were individually performed to compare biopsy results from skin, bone, lymph node, and digit with biopsies of all other locations and IgG serology results. Two-tailed P values were calculated for the Fisher exact tests using the method of summing small P values.⁹ Values of P < .05 were considered significant.

Results

Seventy-six cases of canine coccidioidomycosis were diagnosed via histopathology during the 2 yr period reviewed. Twenty-four cases were excluded as a result of incomplete data sets, leaving 52 cases with complete data sets that met the inclusion criteria.

The mean age of patients was 6 yr (range 6 mo to 11 yr). Twenty-eight (54%) of the cases were male, and 24 (46%) cases were female. Twenty-six different dog breeds were represented, with mixed-breeds (11; 21.1%), Labrador retrievers (5; 9.6%), Weimaraners (3; 5.8%), and Rhodesian ridgebacks (3; 5.8%) being the most commonly represented breeds. Two cases (3.9%) did not have the breed identified in the record. There were 2 dogs each (3.9%) of the following breeds: golden retriever, Pit bull terrier, rottweiler, cocker spaniel, and Yorkshire terrier. One dog (1.9%) from each of the following breeds was included in the study: shih tzu, Italian Spinone, Scottish terrier, Lhasa apso, schnauzer, Cavalier King Charles spaniel, German shepherd dog, Doberman pinscher, boxer, Australian shepherd, Chinese shar pei, Carin terrier, border collie, whippet, terrier (not specified), deerhound, dachshund, and vizsla.

Thirty-eight (73%) of the diagnostic biopsy samples were collected by incisional biopsy, and the remaining 14 (27%) were collected via lesion excision. The most common anatomic site for coccidioidomycosis lesions in this population were skin biopsies (22; 42%) followed by bone biopsies (11, 21.1%) and then lymph node biopsies (5; 9.6%). Other lesions biopsied include digit samples (3; 5.8%), testes (3; 5.8%), lungs (3; 5.8%), pericardium samples (2; 3.9%), oral cavity biopsy (1; 1.9%), mammary tissue (1; 1.9%), and spleen (1; 1.9%). A summary of location of confirmed coccidioidomycosis lesions and respective IgG-negative serology results is included in Table 1. All biopsied lesions were described as granulomatous to pyogranulomatous inflammation with identification of fungal organisms consistent with Coccidioides spp. Additional stains were applied to biopsy sample to aid in diagnosis and identification of Coccidioides spp. in 15 cases.

TABLE 1 Summary of Coccidioidomycosis Lesion Biopsy Location and Negative IgG Serology Results

View Fullscreen

Serology results were evaluated for both IgG and IgM antibody testing, and sensitivity was calculated for both IgG and IgM. A total of 45 out of the 52 dogs had positive IgG results reported. The sensitivity of IgG antibodies for detecting infection in the study population was calculated to be 87%. Only 24 of the 52 cases had positive IgM antibody tests which corresponded to a sensitivity of 46%. Twenty-one cases had negative IgM antibody testing but had positive IgG antibody tests. Of the 45 cases with positive IgG serology, 42 had quantitative immunodiffusion performed with results provided. Three cases were reported to have a recent history of positive antibody tests, but the quantitative titer results were not reported. Quantitative IgG titers ranged from ≤1:2 to 1:256 with median and mode titer results of 1:32 and 1:16, respectively. Two cases had IgG quantitative titers of ≤1:2 and were reported as “weak positives.” Seven cases (13%) were negative for both IgG and IgM antibody testing, and coccidioidomycosis was diagnosed with histopathology alone. Of the 7 cases with negative antibody tests, 3 were biopsy samples of skin lesions, 2 were biopsies from bone lesions, and 1 of each were from a digit and lymph node biopsy.

The 2 × 2 tables used to perform the four Fisher exact tests and their respective P values are summarized in Table 2. Compared with a significance value of .05, none of the P values calculated were significant.

TABLE 2 Summary of 2 × 2 Tables and P Values for Fisher Exact Tests for Coccidioidomycosis Lesions and IgG Serology

View Fullscreen

Discussion

Our study sought to report the sensitivity of a commercially available AGID serologic test for canine coccidioidomycosis. The sensitivity measured the proportion of the study population that was correctly identified or diagnosed with a disease process.⁹ The sensitivity of IgG serology for coccidioidomycosis in dogs in the study population was 87%, whereas the sensitivity for IgM was 46%. The results of this study support the hypothesis that a significant number of cases (7; 13%) were seronegative and diagnosed with coccidioidomycosis solely on the basis of histopathology results. Our results and those of previous studies support that positive antibody testing can favor the suspicion of coccidioidomycosis, but a negative result cannot rule out disease.⁴ The sensitivity of IgM (46%) and IgG (87%) was similar to what is reported in the human literature with sensitivities ranging from 56 to 83% depending on the antibody molecule evaluated (IgM or IgG, respectively) and various testing methods.¹⁰ Our findings are consistent with previous research indicating that elevated serum IgM is more common in acute disease, with IgG being the predominant antibody in chronic and/or disseminated disease.^2,8 The difference in duration and timing of IgM and IgG expression likely explains why more cases had positive IgG antibody tests in our patients with disseminated disease.

The sensitivity results reported here may not be representative of other antibody testing methods or agents. We evaluated a commercially available AGID serology testing method. We selected this test to study as it is currently the diagnostic performed at the commercial laboratory used by the authors’ institution. However, a new enhanced enzyme immunoassay (EIA) for detection of Coccidioides antibodies has been reported to have a higher sensitivity and specificity compared with AGID serology in both humans and dogs.^11,12 Evaluation and validation of this EIA test for canine coccidioidomycosis is limited but has the potential for improved serology detection.¹²

We also evaluated whether lesion location was associated with negative IgG antibody test results. Considering that IgM antibodies are reported to be poorly sensitive, according to both our results and those of previous studies, evaluating the association between lesion location and IgM antibody tests was not performed.⁵ Based on the analysis using Fisher exact tests, there was no association between anatomic location of lesion and negative serology results. Therefore, the second hypothesis was rejected. Our results are likely another example of the lack of association between clinical disease and antibody testing.^3,6

Immunodiffusion was used to quantitate IgG titers in 42 of the 45 cases that had positive serology. The IgG titer results in our population ranged from ≤1:2 to 1:256. Although it is tempting to associate more significant disease with higher titers, previous studies have not found an association between severity of clinical disease and the magnitude of titer values.^3,6 Shubitz et al. demonstrated that significant overlap exists between titer values for clinically ill and subclinical cases of coccidioidomycosis.⁴ The data from our study and previous literature support that coccidioidomycosis can have a wide range of serology results even in the face of disseminated disease.^4,13

Histopathologic diagnosis of coccidioidomycosis was specifically evaluated in this study and is based on the combination of inflammation and observation of the organism in biopsy samples. Diagnostic lesions of coccidioidomycosis have been well reported in previous literature.^1–3,9,6,14 Coccidioides spp. typically incite pyogranulomatous, granulomatous, or less commonly, suppurative lesions in infected tissues.^9,14 Histologically, Coccidioides spp. are observed as spherules that range in size from 10 to 100 μm and contain endospores of 2–5 μm in size. The spherules are double walled and can often be identified with routine hematoxylin and eosin stain.^2,9,15 The sensitivity of spherule identification can be improved with the use of special stains including periodic acid Schiff or with Grocott methenamine silver; the latter is considered to be the most sensitive stain for detection of Coccidioides in tissue.^9,15 A combination of these stains was used to identify Coccidioides organisms in the biopsy samples studied.

Lesion location varied greatly in the study population (Table 1). The most common location of biopsy was skin followed by bone. Given that bone is reported to be the most common location for disseminated disease in dogs, it is interesting that, in our study population, skin biopsy samples were twice as common as bone biopsy samples.^1,2,6 We speculated that this may be because of the relative simplicity in obtaining either incisional or excisional skin biopsy samples compared with the more invasive nature of bone biopsy. Furthermore, because this study was not intended to evaluate the incidence of disseminated disease, no significance can be attributed to the high numbers of skin biopsy samples in the study population. Three samples evaluated were described as digit biopsy samples. These samples were of digits that had a combination of pyogranulomatous dermatitis and osteomyelitis with organisms identified and thus were evaluated as digit biopsy samples independent of bone or skin lesions.

Additional diagnostics for evaluation of suspected canine coccidioidomycosis include fungal culture, cytology, antigen detection, and polymerase chain reaction (PCR). In general, cytology is thought to be relatively low yield for the diagnosis of coccidioidomycosis.¹ Fungal culture of lesions can be performed but does carry infectious biohazard risks.^1,3 A Coccidioides antigen EIA exists that uses antibodies to Coccidioides to detect organisms in serum and urine.¹⁶ Antigen detection is commonly used for diagnosis of Coccidioides infection in humans as well as other fungal infections in dogs including blastomycosis and histoplasmosis.^16–18 Unfortunately, antigen testing in both urine and serum was found to be an insensitive diagnostic for coccidioidomycosis in dogs.¹⁹ PCR tests have been developed and appear to have diagnostic potential.^20,21 Initial animal studies and case reports support PCR and could be a rewarding veterinary diagnostic for coccidioidomycosis.^22,23

The main limitation of this study is the number of cases and its retrospective nature. Because this study evaluated cases with confirmed diagnosis of coccidioidomycosis, evaluating the specificity of serology testing could not be performed. Also, serial serology was not reported or evaluated in this study. Serial serologic testing is advocated for infectious diseases, including coccidioidomycosis, for both diagnosis and monitoring response to therapy.^2,4 A subject for future study could be to compare the sensitivity of repeat serology with confirmed diagnoses of coccidioidomycosis and evaluate if serial serology could improve overall sensitivity. Further studies are also needed to identify the specificity and sensitivity of PCR and enhanced antibody EIA in veterinary patients and evaluate the use of these diagnostic tests in practice.

Conclusion

The results of this study emphasize the challenges veterinarians face in obtaining a definitive diagnosis of canine coccidioidomycosis. The sensitivity of IgG serology in identifying coccidioidomycosis in our study population was 87%, and the sensitivity of IgM was 46%. A total of 7 of 52 cases (13%) were diagnosed on results of histopathology alone. In our study population, negative serology results were not associated with anatomic location of coccidioidomycosis lesions.

Although the Coccidioides fungal organism is geographically limited within the United States to the southwest region, the increasingly mobile nature of our society will likely cause the diagnosis of Valley Fever to spread to both humans and dogs who visit or have temporary residence in this area. Cases of fungal disease outside of their respective anticipated geographical areas have been reported.^24–26 Therefore, it is imperative that all practicing veterinarians, even those practicing outside of the endemic region, be familiar with the complex diagnosis of coccidioidomycosis and that negative serology results cannot rule out Coccidioides infections in dogs.