Clinical utility of fungal culture and antifungal susceptibility in cats and dogs with histoplasmosis

Abstract Background Culture can be used for diagnosis and antifungal susceptibility testing in animals with fungal infections. Limited information is available regarding the diagnostic performance of culture and the susceptibility patterns of Histoplasma spp. isolates. Hypothesis/Objectives Describe the clinical utility of culture and the susceptibility patterns of Histoplasma spp. isolates causing histoplasmosis in cats and dogs. Animals Seventy‐one client‐owned animals, including 33 cats and 19 dogs with proven or probable histoplasmosis. Methods Culture was attempted from tissue or fluid samples. Diagnostic performance of culture, cytopathology, and antigen detection were compared with final diagnosis. Susceptibility to antifungal agents was determined for a subset (11 from dogs, 9 from cats) of culture isolates. Results Culture had a diagnostic sensitivity of 17/33 (52%; 95% confidence interval [CI], 34%‐69%) and 15/19 (79%; 95% CI, 61%‐97%) and specificity of 6/6 (100%; 95% CI, 54%‐100%) and 10/10 (100%; 95% CI, 69%‐100%) in cats and dogs, respectively. Culture was not positive in any animal in which cytopathology and antigen testing were negative. Target drug exposure (area under the concentration curve [AUC]/minimum inhibitory concentration [MIC] >25) should be easily achieved for all isolates for itraconazole, voriconazole, or posaconazole. Five of 20 (25%) isolates had fluconazole MIC ≥32 μg/mL and achieving target drug exposure is unlikely. Conclusions and Clinical Importance Fungal culture did not improve diagnostic sensitivity when used with cytopathology and antigen detection. Susceptibility testing might help identify isolates for which fluconazole is less likely to be effective.


| INTRODUCTION
Histoplasmosis, caused by the dimorphic fungus Histoplasma spp., is an enzootic invasive fungal infection affecting mammals worldwide. In North America, most infections are found east of the Rocky Mountains and in central California. Infection is by inhalation of microconidia found in the environment, and development of disease reflects the interactions of inoculum size, host immunity, and fungal virulence. 1 Disease can be localized to the respiratory tract or disseminate to any organ in the body. [2][3][4] After dissemination, disease often is multisystemic but can be localized. [2][3][4] Domestic cats are more commonly affected as compared to domestic dogs, but histoplasmosis is an important clinical consideration in both species.
Clinical signs are variable and often nonspecific, overlapping with other infectious or inflammatory diseases. 2,5 Diagnosis often is delayed, with an average duration of clinical signs of 6 and 4 weeks in cats and dogs, respectively. 2,5 Identification of Histoplasma spp. organisms within tissue specimens or by fungal culture or PCR is considered the diagnostic standard in human medicine. 6 Although veterinary references note the limitations of culture, it is used in certain cases. [7][8][9] Limitations include unknown diagnostic sensitivity, long turnaround time, and risk to laboratory personnel from exposure to infectious mycelia. In addition, 1 study reported that 81/449 (18.0%) cats and 145/397 (36.5%) dogs from an enzootic area had positive culture results, primarily from tracheobronchial lymph nodes, without finding organisms on histopathology or evidence of clinical disease. 10 These data bring into question the diagnostic specificity of culture for active disease. Cytopathology has limitations including issues related to the risk associated with sample collection and the possibility of a missed diagnosis if organisms are present in low numbers. Identification of Histoplasma yeast organisms in cytology or histopathology samples requires experience because it can appear similar to Sporothrix, Candida, Leishmania, capsule-deficient Cryptococcus, endospores of Coccidioides, and small variant Blastomyces. 11,12 In addition, treatment with antifungal drugs can alter yeast morphology, making identification more difficult. 13 Because of these limitations, serum and urine biomarkers, such as Histoplasma antigen, commonly are used to establish a diagnosis in clinically affected cats and dogs. [14][15][16][17] To determine the clinical utility of fungal culture, a better understanding of its diagnostic performance in combination with cytopathology and antigen detection is needed.
One potential benefit of culture is the ability to evaluate the isolate's susceptibility to antifungal agents. This is important when fungi have unpredictable susceptibility patterns. In human medicine, the susceptibility of Histoplasma spp. to commonly used drugs is predictable and therefore testing is not routinely recommended. 18 Several studies have reported susceptibility patterns, but these are either from clinical isolates obtained from humans or the source was not reported. [19][20][21] For example, for 23 clinical mycelial isolates from humans the mean minimum inhibitory concentrations (MICs) were 7.0, 0.04, 0.05, and 0.1 μg/mL for fluconazole, itraconazole, posaconazole, and voriconazole, respectively. 22 Similar patterns have been found from isolates around the world. 22 Little data is available for clinical isolates from veterinary species. Based on limited genomic data, isolates causing histoplasmosis in cats have substantial genetic divergence from those infecting humans. 23 Thus, cryptic (only differentiated genetically) species infecting cats or dogs are possible and might have unique susceptibility patterns. A better understanding of the antifungal susceptibility of veterinary isolates is needed to guide diagnostic and treatment decisions in these species.
Our primary aim was to determine the utility of fungal culture when combined with cytopathology and antigen detection for the diagnosis of histoplasmosis in cats and dogs. A secondary aim was to report the antifungal susceptibility patterns of isolates causing histoplasmosis in cats and dogs.

| Animals
Ours was a prospective cohort study. Client-owned cats and dogs sus-

| Clinical diagnosis
The diagnosis of proven histoplasmosis was defined by the identification of Histoplasma yeast associated with inflammation on cytopathology. A diagnosis of probable histoplasmosis was defined as findings consistent with ocular or pulmonary involvement (tissues less commonly sampled) and detection of Histoplasma antigen in urine by EIA. Evidence of intrathoracic involvement included radiographic changes consistent with parenchymal lung disease or lymphadenopathy. Evidence of ocular involvement included inflammatory disease of the posterior segment, anterior segment, or both. If cytopathology and antigen testing were negative but fungal culture was positive, this situation was defined as proven histoplasmosis as long as additional supportive clinical evidence was present, including pyogranulomatous inflammation on pathology and a positive clinical response to antifungal drug treatment. Relapse was defined as the diagnosis of proven or probable histoplasmosis after discontinuation of all antifungal agents at the recommendation of the attending veterinarian.
Control animals were those that tested negative for histoplasmosis and were classified as either having a definitive alternative diagnosis or unknown diagnosis. Animals with a definitive alternative diagnosis required negative cytopathology for Histoplasma organisms along with clinical signs consistent with the alternative diagnosis and specific additional supportive evidence. More specifically, an animal with respiratory disease was required to have cytopathology (lung lavage or lung aspirate) or histopathology (nasal biopsy) supportive of the alternative diagnosis. Cardiac disease required echocardiography.
Animals with an alternative diagnosis causing gastrointestinal (GI) disease were required to have supportive findings on GI histopathology. For patients with nonregenerative anemia, a bone marrow aspirate or biopsy was required. For the diagnosis of primary liver disease, an aspirate and cytopathology or liver biopsy and histopathology documenting the alternative diagnosis was required. If joint effusion or lameness was present, synovial fluid cytopathology was required.
The diagnosis of cancer required appropriate findings on cytopathology or histopathology. Diagnosis of an infectious disease required cytopathology, histopathology, microbial culture, or other supportive infectious disease testing. An animal with an unknown diagnosis was defined as 1 that did not meet the definition of proven or probable histoplasmosis or an alternative diagnosis.

| Fungal culture
Samples suspected to be infected by Histoplasma were collected by fine needle aspirate (FNA) for tissues, needle centesis for body fluids, swab for cutaneous or rectal mucosa samples, and peripheral venipuncture for blood. Samples were transferred immediately to separate culturettes for each sample containing either liquid Amies or Stuart media on foam or Amies agar gel (Transystem, Copan Diagnostics, Murrieta, California). Whole blood was transferred immediately to standard aerobic blood culture broth (Bactec Standard Aerobic medium, BD, Franklin Lakes, New Jersey). Samples were submitted to 1 of 2 service laboratories (MiraVista Diagnostics, Indianapolis, Indiana; Oklahoma Animal Disease Diagnostic Laboratory, Stillwater, Oklohoma) with experience in fungal culture. Samples were plated on Sabouraud's dextrose, Mycobiotic, and brain heart infusion with 5% blood agar or potato dextrose and brain heart infusion with 5% blood agar and incubated at 25 C and 37 C in room air atmosphere. If fungal growth was present, identification of Histoplasma was based on morphology, 28S rDNA sequencing (Eurofins, Genomics, Louisville, Kentucky), or both. Culture was considered negative if Histoplasma was not grown within 6 weeks of plating.

| Antifungal susceptibility testing
For a subset of Histoplasma culture isolates, antifungal susceptibility testing was performed at a commercial laboratory (Fungal Testing Laboratory, University of Texas Health, San Antonio, Texas). Clinical and Laboratory Standards Institute (CLSI M38-A2) broth dilution methods were used to determine MICs for fluconazole, itraconazole, posaconazole, voriconazole, terbinafine, and amphotericin-B. These were reported as μg/mL. For MICs below or above the lowest or highest tested concentration, respectively, that concentration was used for statistical analysis.

| Statistical analysis
Statistical analysis was performed using commercial software Fisher's exact test was used to compare frequency of positive fungal culture between cats and dogs. The Mann-Whitney U test was used to test differences between MICs of fluconazole and voriconazole for isolates from animals with clinical relapse that were previously exposed to fluconazole vs those not previously exposed. The ratio of area under the concentration curve (AUC) from previously published pharmacokinetic studies over the median MIC determined in our study was reported. Statistical significance was set as P ≤ .05.

| Cats and dogs without histoplasmosis
Eight cats without histoplasmosis were included; median age was 4 years (range, 1.5-16.5) with 4 spayed females and 4 castrated males.
Cats without histoplasmosis (alternate or unknown diagnosis) had samples collected from a single organ in 5/8 (62.5%) and multiple organs in 3/8 (37.5%). Histoplasma spp. was not isolated from any sample. Histoplasma antigen was detected in urine of 1/7 cats tested, with an antigen concentration below the limit of quantification (<0.4 ng/mL).

| DISCUSSION
Our study showed that fungal culture did not provide any additional benefit when combined with cytopathology and antigen detection for the diagnosis of histoplasmosis in cats and dogs. In addition, antifungal susceptibility patterns were similar to those previously reported from Histoplasma clinical isolates obtained from humans. [19][20][21][22] Susceptibility testing provided evidence of fluconazole resistance, which could be acquired or intrinsic.
The diagnosis of histoplasmosis can be challenging, and several tests might be needed to confirm the diagnosis. Our study investi- No animal in our study without proven or probable histoplasmosis had a positive culture. Although inclusion numbers were small, our findings do not corroborate previous reports of apparently healthy animals with positive Histoplasma tissue cultures. 10 There were differences between sample types however as in the previous study samples were collected at necropsy and positive cultures often came from intrathoracic lymph nodes and less often from other lymph nodes, spleen, liver, or adrenal glands. Histoplasma spp. are not colonizers or commensals, and a positive culture result from inflamed tissues is considered proof of active disease in humans. 6 In general, the same interpretation should hold true in veterinary species. The clinical relevance of a positive culture from nondiseased tissue is unclear.
The antifungal activity of azoles is time and concentration-dependent, whereas that of amphotericin-B is primarily concentration dependent. 24 For azoles, this effect is best described using the AUC/MIC ratio. Target AUC/MIC for commonly used azoles, except posaconazole, is >25, and target maximum blood concentration (C max )/MIC for amphotericin-B is >10. 24 The pharmacokinetic parameter that best predicts terbinafine efficacy has not been established, but the duration of time the blood drug concentration exceeds MIC has been used. 25 [34][35][36] Four of these five isolates were from animals with relapses after previous treatment with fluconazole. These data suggest acquired resistance to fluconazole, which has been described previously in humans after treatment failure. 19 This conclusion is speculative, because without susceptibility data before treatment, it is not possible to definitively differentiate acquired from intrinsic resistance.
Our findings also support the possibility of intrinsic resistance because Cross resistance between fluconazole and voriconazole has been reported in humans after fluconazole treatment failure. 19 Significantly higher voriconazole MICs in isolates previously exposed to fluconazole in our study suggest cross resistance might also occur in cats and dogs. These higher MICs however are unlikely to have clinically relevant impact on outcome because the increases in MICs were modest. Certain factors not accounted for by this target, such as plasma protein binding (which decreases active drug exposure), anatomic location of infection (body system), and microanatomic location of infection (intracellular vs extracellular) also might affect efficacy. Secondly, clinical disease from Histoplasma is caused by the yeast form, but our study used mold for antifungal susceptibility testing. This practice is common because the mold form can be grown more quickly and is thus better suited for diagnostic purposes. Conversion to the yeast form in the laboratory can take up to 8 weeks. 22 In addition, mold has good agreement with yeast for susceptibility testing. 22 Differences in MIC can occur between mold and yeast forms for selected isolates however, which could explain a positive clinical response in some animals infected with strains that have high MICs. Higher tissue permeation leading to higher drug concentrations in infected tissues also could play a role. Thirdly, because the inclusion criteria required a sample to be submitted for cytopathology and culture, our study was likely biased against including animals with pulmonary histoplasmosis.
In retrospective reviews, this form of histoplasmosis accounts for 18% and 9% of cases in cats and dogs, respectively. 2,5 Because of risks associated with lung aspiration or anesthesia and bronchoalveolar lavage, many animals with suspected pulmonary histoplasmosis are diagnosed based on clinical signs, thoracic radiographs, and antigen detection. The underrepresentation of pulmonary histoplasmosis in our study could affect external validity. Fourthly, histoplasmosis is a worldwide disease and several Histoplasma genomospecies exist, with at least 2 in the United States. 51 The relatively small number of animals included in our study were naturally infected in 4 US states (Kansas, Oklahoma, Mississippi, and Texas), representing a small portion of the entire world-wide enzootic area. Histoplasma from other geographic regions might have different antifungal susceptibility patterns.
Finally, we did not measure blood drug concentrations, and relied on published pharmacokinetic data for interpretation of MIC results.
Many pharmacokinetic studies use healthy animals, often of similar age, size, and breed, that are expected to have less variability in drug absorption and metabolism as compared to animals with naturallyoccurring fungal infections. With variable absorption or metabolism, therapeutic drug monitoring is likely required to optimize and individualize dosing.
In conclusion, we found that fungal culture provided no clinical benefit when combined with antigen detection and cytopathology for the diagnosis of histoplasmosis in cats or dogs. Our study provides valuable antifungal susceptibility data using clinical isolates from cats and dogs.

ACKNOWLEDGMENT
Funding provided by the Joan Kirkpatrick Chair in Small Animal Medicine, Oklahoma State University.