Materials and Methods

Criteria for case selection—Medical records in the Veterinary Medical and Administrative Computer System at the University of California-Davis VMTH were searched from August 15, 1988, to February 20, 2010, for cats and dogs with a diagnosis of cryptococcosis. The search term crypto* was used for initial records identification. Cats and dogs were included in the study if the diagnosis of cryptococcosis was confirmed by means of isolation and microbial culture, serology, or identification of the organism by a board-certified veterinary pathologist using cytologic or histologic techniques. Animals were classified as having possible cryptococcosis if organisms were isolated from a nonsterile site without confirmation by positive results of cytologic, serologic, or histologic analysis or if serologic titers ≤ 200 were reported in the absence of other confirmatory diagnostic test results; these cases were excluded from the statistical analysis because of the uncertainty of diagnosis.

Medical records review—Information obtained from the medical records included geographic location, whether animals were kept indoors or allowed outdoors, species, breed, sex, age at examination, clinicopathologic findings, organ involvement, and results of diagnostic imaging for each case. Information regarding the results of FeLV and FIV testing in cats was also recorded.

Sites of organ or tissue involvement were divided into nasal (within the nasal cavity or a mass protruding from the nasal cavity), cutaneous (dermal nodules, including those originating on the bridge of the nose or eyelid; other superficial dermal masses; and draining wounds of the lips), CNS (involving brain, spinal cord, or meninges), eyes (including conjunctiva), lungs (including pleura or parenchyma), lymph nodes, urinary tract (kidneys and urine), or other tissues. Sites were defined as being involved if lesions suggestive of cryptococcosis were present in conjunction with positive serologic test results, even if organisms were not identified within the lesion. For purposes of comparison with findings of a similar study in southeastern Australia,¹¹ cats and dogs were also grouped into those with nasal cavity involvement alone, nasal involvement with local invasion (ie, disease involving the nasal planum, nasal bridge, hard palate, middle ear, retrobulbar space, regional lymph nodes, and brain), nasal involvement with dissemination to distant sites, pulmonary involvement, disseminated disease without nasal involvement, and other.

Available radiographs and ultrasound, CT, and MRI images were reviewed by 2 of the authors (ERW and MSC). For thoracic radiographs, pulmonary infiltrates were categorized by pattern (interstitial, alveolar, or nodular) and extent (focal, regional, multifocal, or diffuse). Other findings recorded included the presence or absence of cranial mediastinal widening or mass effect, hilar mass, or pleural lesions. Advanced imaging analysis of the CNS in cats and dogs of this study was described elsewhere.²⁸

Diagnosis of cryptococcosis—Serologic analysis of serum and CSF samples was performed by use of a commercially available CALAS test kit^a according to the manufacturer-supplied protocol, which included a pronase step. Titers are reported as reciprocal values; titers ≥ 2 were considered a positive test result. Titers ≤ 200 were not considered diagnostic without other confirmatory results. Cryptococcal organisms were identified by means of cytologic and histologic evaluation as narrow-budding yeasts measuring 3 to 8 μm in diameter, surrounded by a variably sized capsule that did not take up H&E stain.

Smears of samples submitted for microbial culture were made and stained with Gram stain. When the presence of Cryptococcus organisms was suspected, smears were also stained with India ink. Samples for microbial culture were inoculated onto defibrinated sheep blood agar and incubated at 37°C in 5% CO₂. When direct fungal culture was requested, samples were inoculated on inhibitory mold agar^b and incubated at 30°C. Yeast colonies were Gram stained to evaluate cellular morphology and subcultured onto cornmeal Tween 80 agar. The presence of Cryptococcus spp was confirmed by use of a yeast identification kit^c according to the manufacturer's instructions. From 2006 through 2010, C neoformans was differentiated from C gattii by means of culture on L-canavanine-glycine-bromthymol blue agar.^29,d

Molecular genotyping—High—molecular weight genomic DNA was extracted and purified as previously described.³⁰ The major molecular types of the Cryptococcus spp complex were identified by URA5 restriction fragment length polymorphism analysis. The URA5 gene was first amplified via PCR with the primers URA5 (5′-ATGTCCTCCCAAGCCCTCGACTCCG-3′) and SJO1 (5′-TTAAGACCTCTGAACACCGTACTC-3′), as described previously, followed by a double digestion with restriction enzymes Sau96I and HhaI. The molecular subtype VGIIb was identified via triple digestion of the URA5 gene with restriction enzymes HhaI, DdeI, and BsrGI, as described previously.⁵ Molecular subtypes VGIIa and VGIIb were also identified and differentiated by means of DNA fingerprinting with the primer M13 for PCR amplification, as described previously.³⁰ The patterns were assigned via visual comparison with patterns obtained for standard strains of the major molecular types of the Cryptococcus spp complex, including VNI (WM 148), VNII (WM626), VNIII (WM628), VNIV (WM629), VGI (WM179), VGII (WM178), VGIII (175), and VGIV (WM779),³¹ or for representative strains of the VGIIa (CDCR265) and VGIIb (CDCR272) molecular subtypes.³²

Statistical analysis—A χ² test was used to determine whether cats and dogs with cryptococcosis differed, compared with other cats and dogs of the hospital population during the same interval, in terms of breed, sex, and (for cats) the results of FeLV and FIV testing. A χ² test was also used to compare diagnostic proportions (ie, the proportions of animals examined at the VMTH that had newly diagnosed cryptococcosis during the 12-year study period) between cats and dogs and to determine whether the distribution of involved sites differed between cats and dogs of the present study and between the present study's population and that described in an Australian study¹¹ that encompassed a 20-year period (1981 to 2001). An unpaired t test and Mann-Whitney U test were used to determine whether the continuous variables of age and cryptococcal antigen titer magnitude, respectively, differed between cats and dogs. Odds ratios and 95% confidence intervals were calculated when comparing diagnostic proportions for species (dogs compared with cats) and breeds. All analyses were performed by use of statistical software.^e Values of P < 0.05 were considered significant.

Results

Diagnosis of cryptococcosis—Sixty-two cats and 31 dogs with a confirmed diagnosis of cryptococcosis were identified during the study period. This included 3 dogs for which clinical findings were previously reported¹⁰ because of a 2-year overlap with the study reported here. The percentage of cats with newly diagnosed cryptococcosis during this interval was higher than that of dogs in the study population (0.12% vs 0.01%, respectively; OR, 8.2 [95% confidence interval, 5.3 to 12.6]; P < 0.001). Several animals had involvement of multiple tissues or organs, and some had evaluations performed via > 1 diagnostic method.

Four cats and 2 dogs were classified as having possible cryptococcosis, and these were excluded from the study because the diagnosis of cryptococcosis could not be confirmed. Five of these 6 animals had serum cryptococcal antigen titers ≤ 200 measured by use of the CALAS test without confirmation of infection via other diagnostic methods. One cat had paraparesis associated with disk herniation determined via MRI. This cat had a serum cryptococcal antigen titer of 200 and a negative CSF CALAS test result. Two cats had persistent serum cryptococcal antigen titers of ≤ 64: one had vitreal inflammation and the other had signs of chronic rhinitis that did not respond to treatment with azole antifungal medication. The fourth cat had uveitis, a serum cryptococcal antigen titer of 2, and a partial response to treatment with fluconazole. Histologic analysis of the enucleated globes 3 and 5 months later revealed only a lymphoplasmacytic inflammatory infiltrate. Cryptococcus albidus was isolated from samples obtained with a nasal swab from 1 dog that had a mucopurulent nasal discharge. The nasal discharge resolved without treatment. The second dog had neurologic signs and chorioretinitis, with a serum cryptococcal antigen titer of 8. This dog was euthanized 2 weeks after the evaluation, and necropsy examination revealed an oligodendroglioma involving the meninges, choroid plexus, and optic nerves. Serum and CSF CALAS test results for samples obtained immediately prior to euthanasia were negative.

Signalment and risk factors—Thirty-two of 62 (52%) cats with cryptococcosis were female, and 30 (48%) were male; both sexes included neutered cats. Sex distribution in affected cats was not significantly different from that of the general hospital population. Age was recorded for 61 cats and ranged from 1 to 17 years (median, 6 years). Breeds included 54 mixed, 5 Siamese, 1 Himalayan, 1 Maine Coon, and 1 Manx cat. Compared with the general hospital population (1,985/51,446 cats), Siamese cats (5/62 cats) were not significantly (P = 0.08) overrepresented. Of the 41 cats for which the status was known, 10 (24%) were kept indoors only and 31 (76%) spent at least some time outdoors.

Results of FeLV antigen tests were negative for all 49 cats tested. Of the 47 cats for which FIV antibody test results were known, 2 (4%) tested positive. Thirteen of 62 (21%) cats had received glucocorticoids prior to diagnosis of cryptococcosis. One of these was receiving prednisone (0.5 mg/kg [0.23 mg/lb], PO, q 12 h) with cyclosporine (6 mg/kg [2.7 mg/lb], PO, IV) following renal transplantation, and another had been treated with prednisone (2.5 mg/kg [1.14 mg/lb], PO, q 12 h) for immune-mediated neutropenia and thrombocytopenia for 2 years. Two cats had been treated with anti-inflammatory doses (range, 1 to 1.6 mg/kg [0.45 to 0.72 mg/lb], PO, q 24 h) of prednisone for > 6 weeks for presumptive inflammatory bowel disease. The remaining 9 of 13 cats had been treated with glucocorticoids for clinical signs that may have been related to undiagnosed cryptococcosis for 2 to 120 days prior to evaluation at the VMTH. Other concurrent diseases identified in cats with cryptococcosis included hyperthyroidism (2 cats) and multicentric pulmonary adenocarcinomas, diabetes mellitus, dilated cardiomyopathy, urinary tract Enterococcus spp infection, and toxoplasmosis (1 cat each).

Of 31 dogs in the study, 12 (39%) were male and 19 (61%) were female; both sexes included neutered animals. No significant sex predisposition was evident. Age at diagnosis was recorded for all 31 dogs and ranged from 1 to 7 years (median, 4 years). The median age of dogs with cryptococcosis was less than that of cats with cryptococcosis (P < 0.001). Breeds included American Cocker Spaniel (n = 9), mixed (6), Labrador Retriever (5), and Australian Shepherd, Bassett Hound, Corgi, English Springer Spaniel, Golden Retriever, German Shepherd Dog, Shetland Sheepdog, Staffordshire Terrier, Standard Poodle, Doberman Pinscher, and Viszla (1 each). American Cocker Spaniels were over-represented in the study group (9/31), compared with the general hospital population of dogs (5,257/209,582 dogs; P < 0.001; OR, 16 [95% confidence interval, 7% to 35%]). Labrador Retrievers were not overrepresented in the study group, compared with the general hospital population. Of the 21 dogs for which the information was available, 1 was kept indoors only and 20 spent at least some time outdoors.

Five of 31 (16%) dogs had other infections concurrent with cryptococcosis: these included Dirofilaria immitis, Neospora caninum, Dipylidium caninum, urinary tract Escherichia coli infection, and papillomavirus infection (1 each). Seven dogs had been treated with prednisone for durations ranging from 3 to 150 days before their evaluation at the VMTH; 1 dog treated with prednisone (0.6 mg/kg [0.27 mg/lb], PO, q 24 h) for 150 days was also being treated with ketoconazole (2.8 mg/kg [1.27 mg/lb], PO, q 24 h) and cyclosporine (5.6 mg/kg [2.55 mg/lb], PO, q 12 h) for immune-mediated hemolytic anemia and thrombocytopenia.

History and physical examination—The most commonly reported signs in cats resulted from nasal cavity involvement (Table 1). In contrast, the most commonly reported signs in dogs were neurologic. Four of 20 cats and 6 of 14 dogs with ocular abnormalities lacked detectable vision in ≥ 1 eye. High rectal temperatures (range, 39.3° to 39.8°C [102.7° to 103.6°F]) were detected during physical examination in 3 of 46 (7%) cats and 4 of 23 (17%) dogs. The distribution of affected sites determined on the basis of history and physical examination findings was significantly (P < 0.001) different between cats and dogs.

CALAS test results—Fifty-one of 53 cats had positive serum CALAS test results. Initial titers ranged from 0 to 524,288 (Figure 1). One of the 2 cats with negative serum test results had localized nasal involvement with signs of rhinitis and swelling on the bridge of the nose; the diagnosis of cryptococcosis was made on the basis of histologic examination of nasal biopsies and isolation of Cryptococcus spp from nasal discharge. The other cat had localized ocular involvement, and a diagnosis of cryptococcosis was made on the basis of histologic examination of an enucleated eye.

View Full Size

Scatterplot indicating magnitude of serum cryptococcal antigen titers measured via CALAS assay for 53 cats and 18 dogs with cryptococcosis in California. The y-axis represents reciprocal titers at time of diagnosis expressed on a logarithmic scale. The horizontal line represents median titer for each species.

Citation: Journal of the American Veterinary Medical Association 239, 3; 10.2460/javma.239.3.357

• Download Figure
• Download figure as PowerPoint slide

Fifteen of 18 dogs had positive serum CALAS test results. Initial titers ranged from 0 to 65,536 (Figure 1). One of the 3 dogs with negative serum test results had a CSF cryptococcal antigen titer of 16 measured by use of the CALAS test; however, organisms were observed in cytospin preparations of the CSF. Another dog had ocular involvement, and Cryptococcus spp were isolated from a urine sample. Organisms were not observed in CSF smears from the remaining dog, which had a retrobulbar mass, but the CSF cryptococcal antigen titer was 128, and C gattii was cultured from samples obtained from CSF and from the mass. The median serum cryptococcal antigen titer was significantly (P = 0.03) higher in cats than in dogs.

Cytologic and histologic evaluations—Cryptococcal organisms were detected via cytologic examination in 41 of 62 (66%) cats. Organisms were identified in smears made of fine-needle aspirates of cutaneous lesions (n = 15 cats), kidneys (2), mediastinal or pulmonary masses (6), enlarged lymph nodes (2), and an orbital mass (1); CSF (9); impression smears of ulcerative lesions or nasal biopsy samples (2); nasal exudates or washes (3); pleural effusion (1); and urine (1). Organisms were not seen in cytospin preparations of 2 CSF samples. Organisms were detected via histologic examination of samples from 30 cats, which included biopsy samples for 19 cats and samples obtained at necropsy for 15 cats (4 cats that had biopsies performed were later necropsied). Biopsy samples included intranasal masses (n = 9 cats), cutaneous masses (6), gingival tissue (2), eye (1), and kidney (1).

Cryptococcal organisms were detected via cytologic examination in 17 of 31 (55%) dogs. Organisms were identified in cytospin preparations of CSF (n = 11 dogs) and smears of fine-needle aspirates of enlarged lymph nodes (3) and retina, pancreas, cutaneous mass, retrobulbar mass, spleen, frontal sinus, and kidney (1 dog each). Organisms were not seen in cytospin preparations of 4 CSF samples. Organisms were detected via histologic examination of samples from 20 dogs, which included biopsy samples for 5 dogs, and samples obtained at necropsy for 17 dogs. Two dogs for which the diagnosis of cryptococcosis included evaluation of a biopsy sample were subsequently necropsied. Biopsy samples included intranasal masses (n = 3 samples) and cutaneous lesion, gastrointestinal tract, liver, and perihilar mass (1 sample each). Histologic evaluation of a frozen section obtained during thoracotomy in the dog with a perihilar mass revealed changes suggestive of malignant neoplasia. Examination of samples obtained at necropsy of this dog revealed an intense pyogranulomatous inflammatory process with small numbers of Cryptococcus spp within the mass and associated pleura.

Isolation and molecular typing of Cryptococcus spp—Microbial culture was attempted in aspirates, tissue samples, or body fluid specimens from 27 cats, and organisms were isolated from 25 cats. Cryptococcus spp were isolated from the urine of 2 of 10 cats for which aerobic microbial culture and susceptibility testing was performed. Negative culture results were obtained from aspirates of 1 cutaneous mass and 1 mediastinal mass that contained cryptococcal organisms confirmed via cytologic examination. Species differentiation was performed for isolates from 9 cats. Seven of these isolates were C gattii, and 2 were C neoformans. All 7 C gattii isolates were analyzed for molecular type; 1 was type VGIIa, and 6 were type VGIII (Figure 2; Table 2). Three VGIII isolates were identified as having M13 fingerprinting pattern 3, 1 as having pattern 2, and 2 as having pattern 1 (data not shown).

View Full Size

Agarose gel electrophoretograms of PCR products used to discriminate molecular types of Cryptococcus gattii isolated from cats and dogs in California. In panel A, restriction fragment length polymorphism patterns of C gattii DNA obtained after digestion of the URA5 gene with restriction enzymes Sau96I and HhaI are shown for isolates from 5 cats and 1 dog. Lanes 1, 2, 3, 5, and 6 = C gattii feline isolates WM09.43, WM09.44, WM09.45, WM09.47, and WM09.48, respectively, typed as VGIII. Lane 4 = C gattii canine isolate WM09.46, typed as VGII. Lanes 7 through 16 = Molecular type standards VNI (WM148), VNII (WM626), VNIII (WM628), VNIV (WM629), VGI (WM179), VGII (WM178), VGIII (WM175), VGIV (WM779), VGIIa (CDCR265), and VGIIb (CDCR272), respectively. In panel B, PCR fingerprinting patterns were determined for C gattii isolates obtained from 2 dogs after amplification of C gattii DNA by use of the M13 primer. Lanes 1, 2, and 4 = Molecular type standards VGII (WM178), VGIIa (CDCR265), and VGIIb (CDCR272), respectively. Lane 3 = C gattii isolate WM10.16, typed as VGIIa. Lane 5 = C gattii isolate WM09.46, typed as VGIIb. Values to the left of the gels represent molecular weight (kDa) of bands. M = Molecular weight marker.

Citation: Journal of the American Veterinary Medical Association 239, 3; 10.2460/javma.239.3.357

• Download Figure
• Download figure as PowerPoint slide

Microbial culture was attempted in samples from 24 dogs, and organisms were isolated from all except 2 of these: one was a CSF sample with a neutrophilic pleocytosis that had no organisms detected via cytologic examination, and the other was a nasal biopsy sample in which Cryptococcus spp were identified via histologic examination. Cryptococcus spp were isolated from the urine of 4 of 11 dogs for which aerobic microbial culture and susceptibility testing was performed. Species differentiation was performed for isolates from 8 dogs; 6 of these were C neoformans, and 2 were C gattii. The molecular type of the 2 C gattii isolates was VGII, with one subtyped as VGIIa and the other as VGIIb (Figure 2; Table 2).

Sites of tissue or organ involvement—Sites of tissue or organ involvement for cats and dogs in the study were assessed and categorized (Table 3; Figure 3). The most common site of confirmed tissue or organ involvement for cats was nasal tissue (n = 35 cats). Sites classified as other tissues or organs included mediastinal masses (n = 4 cats), gingival tissue (3), spleen (2), and right atrium, liver, and thyroid gland (1 each). All cats with mediastinal masses had concurrent pulmonary involvement.

View Full Size

Distribution of tissue or organ involvement in 62 cats (white bars) and 31 dogs (black bars) with cryptococcosis in California. Some animals had involvement of multiple sites. Percentages indicate proportion of the total for each species. Significant (*P < 0.05; †P < 0.01) differences between cats and dogs are indicated. LN = Lymph node.

Citation: Journal of the American Veterinary Medical Association 239, 3; 10.2460/javma.239.3.357

• Download Figure
• Download figure as PowerPoint slide

Histologic evaluation of the eye in 7 of 19 cats with ocular involvement revealed chorioretinitis (n = 4 cats), optic neuritis (4), and uveitis, endophthalmitis, panophthalmitis, retinal folding, retinal degeneration, and retinal separation (1 each). Fifteen of these 19 cats had concurrent neurologic involvement, and 11 of 35 cats with nasal involvement also had neurologic involvement. Imaging features and the results of clinicopathologic and necropsy examination of cats and dogs that had neurologic involvement were described elsewhere.²⁹

Of the 28 cats with cutaneous tissue involvement, 15 had solitary or multiple nodular soft tissue masses involving the bridge of the nose, 8 had multiple subcutaneous dermal nodules over the face and body, and 5 had solitary subcutaneous masses or draining wounds on the head, trunk, digits, or neck. Two cats with cutaneous masses also had gingival masses in which Cryptococcus spp were identified. For 3 of the 28 cats, ulceration of cutaneous lesions was reported.

Eleven cats had lymph node involvement. Affected lymph nodes were hilar (n = 3 cats), mandibular (2), mesenteric (2), and mediastinal, cervical, superficial cervical, axillary, tracheobronchial, and medial iliac nodes (1 each).

Images from CT scans of the head were available for review in the records of 3 cats. Two cats had increased soft tissue and fluid opacification in the rostral nasopharynx and nasal cavities without a mass effect. One of these 2 cats also had partial fluid opacification in the frontal sinus. The third cat had a small, peripherally contrast-enhancing mass at the nasal planum with underlying focal lysis of the nasal bone. Results of analysis of radiographs of the skull in another cat revealed osteomyelitis of the nasal bone. Invasion of the cribriform plate was detected in 2 cats by use of MRI as described elsewhere.²⁸

Radiographs were available for review for 26 of 44 cats for which thoracic radiography was performed. No clinically relevant lesions were reported for 26 of the 44 cats, and 13 of these had radiographs available for review. Eighteen of the 44 cats were reported to have thoracic lesions, and 13 had radiographs available for review. Radiographic abnormalities consisted of interstitial infiltrates (n = 8 cats; regional in 3 and diffuse or multi-focal in 5), cranial mediastinal widening resulting from mass lesions or lymph-adenopathy (8), pulmonary nodules or masses (3), hilar masses (3), regional alveolar infiltrates (2), and pleural effusion (2). One cat had a diffuse miliary pattern throughout the lungs. Rib fractures were evident in association with a large mediastinal mass in 1 cat (Figure 4). Thoracic ultrasonography with static image review was performed in 5 cats. Findings included cranial mediastinal masses (n = 4 cats) and diffuse comet-tail artifact at the pleural surface consistent with pulmonary infiltrates (2). A thoracic MRI examination performed in one of these cats revealed a lobular mass dorsally located in the cranial mediastinum that partially enveloped the distal trachea. The mass was isointense to muscle on T1-weighted sequences and hyperin tense on T2-weighted sequences, with moderate heterogenous contrast enhancement.

View Full Size

Right lateral and dorsoventral radio-graphic images of a cat in the present study infected with C gattii. In panel A, ventral deviation of the trachea (arrow) is evident. A moderate volume of pleural effusion and multifocal pulmonary infiltrates are also present. In panel B, marked widening of the mediastinum is present (arrows), and fractures with minimal displacement of ribs 10 through 13 are evident on the left side (arrowheads); only ribs 12 and 13 had evidence of fracture in thoracic radiographs obtained 2 days prior to this evaluation. Necropsy examination revealed a large dorsal mediastinal cryptococcal granuloma and granulomatous pneumonia. R = Right.

Citation: Journal of the American Veterinary Medical Association 239, 3; 10.2460/javma.239.3.357

• Download Figure
• Download figure as PowerPoint slide

Abdominal ultrasonography was performed in 28 of the 62 cats in the study; images were available for review in 21 cases. No clinically relevant findings were reported in 18 cats. Two had renal masses that were iso- to hypoechoic and involved the renal pelvis, with evidence of pelvic dilatation. These masses were identified as cryptococcal granulomas via cytologic evaluation of fine-needle aspirates. One of these 2 cats had a diagnosis of renal failure. The third cat with abdominal ultrasonographic abnormalities had ascites, mesenteric lymphadenopathy, and mesenteric hyperechogenicity, but renal abnormalities detectable via ultrasonography were not reported. Cryptococcus spp were isolated from the urine of this cat.

In dogs, the CNS was the most commonly involved site (Table 3; Figure 3). Of 15 dogs for which the CNS was examined at necropsy, all had CNS involvement. Sites classified as other tissues or organs included the pancreas (n = 6 dogs); liver (6); myocardium (6); gastrointestinal tract (5); spleen (3); peritoneum, mesentery, thyroid gland, urinary bladder wall, and oral cavity including the tongue (2 each); and retrobulbar space, prostate, pleura, elbow joint, mediastinum, and adrenal gland (1 each). American Cocker Spaniels were the only breed with cryptococcal involvement of the liver, myocardium, spleen, thyroid gland, and adrenal gland. The 2 dogs infected with C gattii VGII had nasal cavity involvement with destruction of the cribriform plate, and lesions were not found in other organs. Necropsy in 5 dogs infected with C neoformans had dissemination of the infection to multiple parenchymal organs, and breaches of the cribriform plate were not identified despite the presence of nasal cavity or frontal sinus involvement for 3 of the dogs. The remaining dog infected with C neoformans was determined to have CNS involvement on the basis of results of CSF analysis, and necropsy was not performed.

There was no significant (P = 1.0) difference in the proportion of American Cocker Spaniels with neurologic involvement, compared with other dog breeds. Histopathologic changes reported for dogs with ocular involvement included pyogranulomatous chorioretinitis (n = 6 dogs), optic neuritis (3), retinal separation (3), and conjunctivitis (1); all had intralesional Cryptococcus spp identified. Ten of 13 dogs with ocular involvement and 9 of 12 dogs with nasal tissue involvement had concurrent CNS involvement.

Eleven dogs had lymph node involvement. Involved lymph nodes were hilar (n = 5 dogs), mesenteric (3), pancreatic (2), and tracheobronchial, retropharyngeal, mandibular, gastric, ileocolic, medial iliac, and popliteal nodes (1 dog each).

Two of 5 dogs with cutaneous involvement had solitary cutaneous lesions (1 on the head and 1 on a pelvic limb), 1 dog had granulomatous dermatitis involving the nostrils, and 1 dog had multiple lesions distributed on the head, neck, and limbs. The remaining dog had nail bed and tarsal pad involvement, with invasion of the underlying bone.

A CT scan of the nasal passages was performed in 3 dogs, and linear tomography was performed on the frontal sinus of a fourth dog. Mass lesions were detected in 3 dogs; 1 had a frontal sinus mass with associated osteolysis, and 2 had peripherally contrast-enhancing masses. In 1 of these 2 dogs, masses were located in the caudal nasal cavity and nasopharynx. In the other dog, masses were detected in the left retrobulbar space, the ventral aspect of the left frontal sinus, and the lateral aspect of the left nasal cavity; this dog also had osteomyelitis of the frontal bones, lacrimal bones, and cribriform plate, with extension into the olfactory lobe region of the calvarium. The dog evaluated via linear tomography had periosteal thickening and irregularity associated with the right frontal bone, consistent with osteomyelitis. Examination of the nasal cavity at necropsy in 7 dogs revealed cryptococcal osteomyelitis of the cribriform plate (n = 3 dogs) or intranasal plaques or masses that contained Cryptococcus spp within the caudal nasal cavity or nasopharynx (5), turbinates (3), or frontal sinus (3).

Thoracic radiographs were available for review for 15 dogs. Only 3 dogs had clinically relevant lesions including multifocal interstitial infiltrates (n = 3 dogs), cranial mediastinal widening (2), pleural effusion (2), hilar mass (1), and multifocal alveolar infiltrates (1). Lesions found at necropsy in dogs with pulmonary involvement were located throughout the pulmonary parenchyma (6 dogs) or pleura (5).

Images obtained during abdominal ultrasonography were reviewed for 15 dogs. Eight dogs had clinically relevant lesions, which included intra-abdominal lymphadenopathy (n = 5 dogs), renal lesions (4), multifocal gastrointestinal thickening (3), ascites (3), pancreatic abnormalities (3), and mild splenomegaly (2). Four dogs also had mesenteric lymphadenopathy, which was moderate to severe in 3 dogs and mild in 1 dog. Severe medial iliac lymphadenopathy in 1 dog was associated with obstructive hydronephrosis of the left kidney and ureter. Two of the 4 dogs with renal lesions had a diagnosis of renal failure; one had bilateral renomegaly with decreased corticomedullary distinction, and the other had mass-like lesions in the renal medulla. Cytologic evaluation of fine-needle aspirates of these lesions revealed Cryptococcus spp.

Pancreatic abnormalities were identified during abdominal ultrasonographic evaluation of 1 of 6 dogs that had pancreatic involvement at necropsy. The pancreas was enlarged, hypoechoic, and surrounded by hyperechoic mesentery, and the duodenum had a corrugated appearance. Another dog that had pancreatic involvement identified at necropsy did not have detectable evidence of pancreatic lesions during abdominal ultrasonography. Abdominal ultrasonography was used to identify multifocal gastrointestinal thickening in 3 dogs with gastrointestinal involvement, which was verified at necropsy. Loss of normal gastric wall layering was observed during the ultrasonographic evaluation in one of these dogs. Sites of gastrointestinal involvement identified at necropsy included the stomach (n = 2 dogs), cecum (3), jejunum (3), and duodenum (3).

The distribution of tissue or organ involvement was significantly (P < 0.001) different between cats and dogs of the present study. Involvement of multiple tissue or organ sites and involvement of sites that were categorized as other tissues were significantly more common in dogs than in cats (P = 0.02 and P = 0.003, respectively). In cats, this distribution was significantly (P < 0.001) different from that reported in the study¹¹ in southeastern Australia, with disseminated disease being more commonly detected in cats of the present study in California and nasal disease predominating in cats of the Australian study (Figure 5). The distribution of sites involved in dogs was also significantly (P < 0.001) different than that reported in dogs in southeastern Australia; a greater proportion of dogs in the present study in California had disseminated disease with or without nasal involvement, and nasal disease with local extension predominated in the dogs of the Australian study.

View Full Size

Distribution of tissue or organ involvement in cats and dogs with cryptococcosis in the present study in California (black bars) and in an earlier study¹¹ in southeastern Australia (white bars). In panel A, n = 62 cats in the present study and 153 cats in the Australian study; in panel B, n = 31 dogs in the present study and 40 dogs in the Australian study. Percentages indicate proportion of the total in each study for each species. Local invasion was defined as disease involving the nasal planum, nasal bridge, hard palate, middle ear, retrobulbar space, regional lymph nodes, and brain. Disseminated disease was infection that was present in tissue or organ sites distant from these locations. Significant (*P < 0.001; †P < 0.01) differences between study groups at various sites of involvement are indicated. Diss = Disseminated. Local = Local invasion. Nasal = Nasal cavity. Resp = Respiratory tract.

Citation: Journal of the American Veterinary Medical Association 239, 3; 10.2460/javma.239.3.357

• Download Figure
• Download figure as PowerPoint slide

Discussion

Although several retrospective studies of canine and feline cryptococcosis have been published, to our knowledge, no previous study has compared canine and feline cryptococcosis in California over a single period. In the study reported here, cats were 8 times more likely to have a diagnosis of cryptococcosis than were dogs. The proportion of dogs with newly diagnosed cryptococcosis in the present study was 100 times that reported in the Veterinary Medical Database at Purdue University from 1964 to 1990.^10,25 The lack of predisposition for either sex to have cryptococcosis in the present study generally agrees with findings in other studies^{9–11,16,20,24} performed in the United States, British Columbia, and Australia, with the exception of 1 study²⁶ in the United States that reported a sex predisposition for male cats. A previously reported breed predilection in Siamese cats^11,12 was not evident in the study reported here. Similar to the results of another study,^10,25 a strong breed predisposition for cryptococcosis in American Cocker Spaniels was identified in our study population. In contrast to other studies,^9–11,25,33 we did not find any other breeds to be at increased risk in this study, including German Shepherd Dogs, which were previously reported³⁴ to be at increased risk for systemic aspergillosis among dogs of our hospital population.

Species determination was performed for 17 cryptococcal isolates in the present study. Of those isolates, C gattii predominated (7/9) as a pathogen of cats and C neoformans was isolated from most (6/8) dogs. In a study¹¹ in southeastern Australia, 29% of cats and 18% of dogs had C gattii infections, with an increased prevalence of C gattii infections in animals from rural areas. In contrast, approximately equal numbers of cats and dogs from western Australia were infected with C gattii and C neoformans.¹⁸ Since 1999, C gattii serotype B has been isolated from at least 26 cats and 19 dogs in southwestern British Columbia.¹⁶

Case numbers and follow-up were insufficient in this study to allow a comparison of responses to antifungal drugs between C gattii- and C neoformans-infected animals. In humans, infections with C gattii have been associated with immunocompetence and resistance to antifungal drug therapy, whereas C neoformans mainly causes disease in the immunocompromised.³⁵ The apparent difference in host preference for the Cryptococcus spp in the present study may reflect the degree of immune competence of the hosts or differences in exposure to pathogens. No other obvious underlying immunosuppressive conditions were identified in any of the dogs for which cryptococcal species determination was performed, and there was no apparent breed association with the cryptococcal species isolated. In contrast to cats, dogs in the study reported here frequently had dissemination of cryptococcal organisms to multiple sites. Some (6/31 [19%]) dogs had concurrent opportunistic infections or immunosuppressive disease.

Investigators in Australia described FIV coinfection in 28% of cats with cryptococcosis in one study¹² and 21% in another,¹¹ which was reported to be similar to the prevalence of FIV infection within populations of cats with a variety of other illnesses in Australia. Similar to findings of another study²⁴ in the United States, the proportion of cats with retroviral infection in the study reported here was low and similar to that of our general hospital population³⁶ and in cats with clinical signs of disease from other parts of the United States.³⁷ Other immunosuppressive conditions, such as renal transplantation, neoplasia, or immune-mediated disease, were recognized in a small percentage of cats in this study.

In the present study in California, molecular types of Cryptococcus spp in cats differed from those detected in dogs. One of the 2 C gattii isolates obtained from dogs was of the molecular subtype VGIIb, which was also identified for cryptococcal strains of low virulence that were uncommonly isolated from human and veterinary patients in British Columbia.³⁸ Most patients in the British Columbia study were infected with strains of the more virulent VGIIa molecular subtype, which has recently been isolated from cats and dogs in Washington and Oregon^39,40 and an alpaca in California⁷; the VGIIa subtype was also isolated from 1 dog and 1 cat in the study reported here. In contrast, the remaining C gattii isolates from cats in our study for which molecular typing was performed were type VGIII. To the authors' knowledge, this molecular type was reported in only 1 Australian cat,⁴¹ whereas most reported C gattii isolates from Australia have been type VGII on the west coast and type VGI on the east coast.^19,41 Thus, the various molecular types of Cryptococcus spp isolated from cats and dogs may also reflect differences in environmental exposure between these species. Approximately one-fourth (10/62) of cats in the present study were housed strictly indoors, reinforcing observations of other investigators in the United States that indoor cats are not protected from cryptococcosis.^24,26 One C gattii isolate in the present study was from a cat housed exclusively indoors in the absence of house plants, although the cat lived with a dog and an African Grey parrot. Whether the high proportion of cats with C gattii VGIII infections in the present study reflects a similar epidemiology in humans infected with C gattii in the geographic region of California requires investigation.

The sensitivity of the serum CALAS assay in cats has been reported to range from 95% to 98%, with a specificity of 100%.^21,22 Sensitivity and specificity of the assay has not been evaluated in a large number of dogs. In humans, sensitivity and specificity of the CALAS assay was reported as 97% and 95%, respectively, for evaluation of serum samples and as 100% and 96%, respectively, for evaluation of CSF samples.²³ False-negative CALAS test results were reported for serum samples in 2 cats and 3 dogs of the present study that had nasal, ocular, or neurologic tissue or organ involvement. Other studies^21,22 have reported small numbers of false-negative serum CALAS assay results in cats with localized cryptococcal infections or with ocular and CNS lesions experimentally induced via inoculation with small numbers of organisms.¹³ When cryptococcosis is a differential diagnosis for animals with neurologic signs that have negative serum CALAS test results or lack detectable cryptococcal organisms in CSF, testing of the CSF for cryptococcal antigen should be considered. During records review of cases in this study, a few animals were identified that likely had false-positive results of the CALAS assay, although only 1 dog was confirmed to have a false-positive result on the basis of results of a full necropsy examination. We chose to set a serum cutoff value of > 200 for reciprocal titers determined via CALAS assay for diagnosis of cryptococcosis in the absence of other confirmatory findings in the present study because we identified 1 cat with a serum cryptococcal antigen titer of 200 for which the diagnosis of cryptococcosis could not be confirmed despite an extensive workup, and an alternate diagnosis was determined. This did not result in exclusion of a large number of cases. False-positive cryptococcal antigen test results have been recognized in human patients that were treated with hetastarch⁴² or had Klebsiella spp infections.⁴³ False-positive test results of low cryptococcal antigen titers have also been identified in humans in the absence of these variables.²³

In the study reported here, median values of serum cryptococcal antigen titers measured via CALAS assay were higher in cats than in dogs. The magnitude of serum cryptococcal antigen titers determined via this method has been positively correlated with disease severity, and in 1 study,²² cats and dogs with disseminated skin disease, lymph node involvement, or both had significantly higher titers, compared with those that did not have this tissue distribution. The increased proportion of dogs with disseminated disease in the present study would be expected to predict higher, rather than lower, cryptococcal antigen titers in dogs than in cats if this applied across species and geographic locations. One study⁴⁴ in cats showed that initial titers determined via CALAS assay did not differ significantly on the basis of infection site. The lower titer values in dogs of the present study may reflect differences between cryptococcal isolates that infected cats versus dogs or differences in the immune response to infection between cats and dogs. Cryptococcus gattii infections have been associated with high antigen titers in humans.⁴⁵ Despite their relatively high magnitude, the cryptococcal antigen titers for cats in the present study were lower than that reported previously in the United States for cats.²¹ This may reflect variation among laboratories or differences in organism burden in cats located in different geographic regions.

Cytologic and histologic examination of fluids or tissue from affected sites and microbial culture of samples appeared to be sensitive methods for diagnosis of cryptococcosis, although a few animals had negative or misleading results. Histologic examination of a frozen section of affected tissue from 1 dog was suggestive of malignant neoplasia rather than cryptococcosis. Microbial culture of 4 of 50 (8%) samples (3 aspirates or biopsy specimens and 1 CSF sample) had negative results. The negative results in these cases may have related to size of samples submitted. Cryptococcus albidus was isolated from nasal discharge of 1 dog that was excluded from the study. Although C albidus has been reported to cause disease in dogs,⁴⁶ Cryptococcus spp can be isolated from the nasal cavity from apparently healthy cats and dogs⁴⁷; therefore, isolation of the organism from nasal swabs in the absence of other supportive findings was not used as an inclusion criterion in this study.

Clinical signs at the time of examination were different between cats and dogs in the present study; a greater number of cats had nasal or upper respiratory signs and cutaneous masses, and a greater number of dogs had neurologic signs and ocular lesions. Involvement of abdominal organs, including the pancreas, kidneys, gastrointestinal tract, and abdominal lymph nodes, was more frequently detected during diagnostic evaluation or necropsy in dogs, which had a higher occurrence of disseminated disease than did cats, especially for sites other than the nasal tissues, skin, lungs, CNS, eye, and lymph nodes. In some dogs that had no abnormalities detected during abdominal ultrasonography, cryptococcal infiltration was detected at necropsy in organs such as the kidney and pancreas. Pulmonary lesions were more commonly detected via radiography in cats than in dogs; these lesions were similar in cats and dogs, except that pulmonary nodules were seen only in cats. Interestingly, C gattii has been incriminated as a cause of pulmonary nodules in humans.⁴⁸ Because many of the radiographic images that were unavailable for review were from animals that had pulmonary involvement, we may have underestimated the frequency of pulmonary abnormalities in the present study.

According to the classification scheme described by O'Brien et al,¹¹ 48% of cats and 70% of dogs in the present study had disseminated disease with or without nasal involvement, which was 3 and 2 times as high as the percentages reported in the retrospective study¹¹ in southeastern Australia, respectively. This result may reflect referral bias, the extent of diagnostic testing, or the relatively large proportion of patients that were necropsied in the present study. Other possible explanations include a difference in host susceptibility or cryptococcal strain virulence in the California geographic locale.

Similar factors may explain an apparently higher proportion of cats (26/62 [42%]) and dogs (21/31 [68%]) with CNS involvement in the study reported here, compared with that reported in studies in southern California and Canada.^9,17,24 Only 6% of cats from southern California were reported²⁴ to have neurologic involvement on the basis of clinical signs. In 1 Canadian study, 26% and 42% of dogs had CNS involvement. In another report²⁰ from British Columbia, 11 of 20 cats examined during an outbreak of cryptococcosis were reported to have neurologic signs. A high prevalence of neurologic involvement in dogs was also described in studies^10,25 that used the Veterinary Medical Database at Purdue University. Referral bias may have contributed to some extent to the proportion of animals with neurologic involvement in the present study because the hospital includes a large neurology referral service, although not all cases were referred to this service. The outcomes for animals with CNS involvement in this study are reported elsewhere.²⁸

By use of the classification system of O'Brien et al,¹¹ nasal involvement was detected in 58% of cats in the present study; 15% of cats had localized nasal cavity involvement without extension to other sites; 27% had nasal involvement with local invasion of sites including the nasal bridge, retrobulbar space, regional lymph nodes, and brain; and 16% had nasal involvement with dissemination to distant sites, compared with 40%, 41%, and 9%, respectively, in cats of the southeastern Australia study.¹¹ Similarly, nasal involvement assessed via the same method was less common in dogs in our study (12/31 [39%]) than in the southeastern Australia study,¹¹ in which it was 60%. We used 2 methods of classification for site involvement in the present study because we could not be certain whether site involvement was the result of extension from one site to another or of hematogenous dissemination, yet we wanted to compare our results with those previously reported. In another study,¹³ intracarotid inoculation of cats with Cryptococcus spp was followed by development of distortion and swelling of the nostril as well as frontal sinusitis on the inoculated side, which supports the hypothesis that nasal disease may not always result from localized infection.

More than one-third (8/21) of the dogs with nasal cavity involvement in the southeastern Australia study¹¹ had evidence that suggested the spread of cryptococcal infection across the cribriform plate. Osteomyelitis of the cribriform plate was confirmed at necropsy in 3 dogs of the present study, and an additional 2 dogs had invasion of the frontal sinus or frontal bone as determined via diagnostic imaging or necropsy; all of these dogs had CNS involvement as well. However, in other animals with CNS involvement, no evidence of compromise of the calvarium was detected at necropsy. Hematogenous spread or local invasion from the nasal cavity may contribute to CNS involvement. Of interest, only the 2 dogs infected with C gattii in the present study had lesions localized to the caudal nasal cavity, cribriform plate, and CNS, without evidence of lesions elsewhere. In contrast, 5 dogs infected with C neoformans had widespread dissemination of infection. Although more cases are required to verify such an association between Cryptococcus spp and lesion distribution, lesion distribution may reflect organism virulence attributes or the nature of the host immune response.

An important limitation of the study reported here was its retrospective nature. Not all tests were performed on all patients, and a full necropsy examination was not performed on every animal to confirm tissue or organ involvement. Some involved sites may have been overlooked. Some animals were classified as having tissue or organ involvement on the basis of typical clinical signs or radiographic findings, and it is possible that other causes of the lesions may have existed, which could result in overestimation of site involvement. Nevertheless, 15 of 62 (24%) cats and 17 of 31 (55%) dogs had necropsy confirmation of organ or tissue involvement. False-negative (5/93) and false-positive (1/93 as confirmed via necropsy; up to 5 other false-positive results were suspected but were not confirmed) results of the CALAS assay were recorded, which emphasizes that although this is a sensitive test, it should not be relied upon solely for the diagnosis of cryptococcosis. Finally, cryptococcal species and strains commonly detected in dogs differed from those in cats among the present study population in California. The C gattii strains in cats of the present study varied from those reported in studies^7,8,38–40 in British Columbia and the Pacific northwest. Whether the differences in clinical features between dogs and cats of the present study and the differences in distribution of involved tissues between the study populations in California and southeastern Australia¹¹ reflect these variations, differences in host response to these organisms, or referral bias requires further study.