Materials and Methods

Case selection criteria—The present study was performed in conjunction with a study of radiographic patterns in dogs with pulmonary blastomycosis, rate of radiographic resolution of pulmonary infiltrates, and radiographic factors associated with outcome.⁸ Details of the case selection criteria and medical records review protocol have been published.⁸ In brief, medical records of dogs examined at the University of Minnesota Veterinary Medical Center between January 1, 1989, and January 31, 2006, were searched to identify dogs in which pulmonary blastomycosis had been identified. Dogs were eligible for inclusion in the study if the diagnosis had been confirmed cytologically, histologically, or on the basis of response to antifungal treatment in combination with results of serologic testing or identification of Blastomyces organisms in nonpulmonary tissues; orthogonal radiographic views of the thorax obtained prior to treatment were available for review; and there was no history of previous episodes of blastomycosis. Dogs with other pulmonary diseases, such as heartworm disease, primary cardiac disease, chronic airway disease, and neoplasia, were excluded from the study.

Medical records review—Information obtained from the medical records of cases included in the study consisted of results of diagnostic testing (ie, results of cytologic or histologic examination, an agar gel immunodiffusion test^a for antibodies against Blastomyces spp, and fungal and bacterial culture), type and dosage of antifungal medications administered, whether there was any history of previous treatment with antimicrobials or corticosteroids, whether supplemental oxygen or mechanical ventilator support was provided, other sites of confirmed blastomycosis-related lesions, and response to treatment as determined by examination of follow-up thoracic radiographs.

Medical records were also searched for results of routine CBCs and serum biochemical analyses, including determination of ionized serum calcium concentration, performed at the time of initial examination, prior to initiation of antifungal treatment. Results were classified as greater than, less than, or within the reference range by comparison with ranges established by the Veterinary Medical Center's Clinical Laboratory. Serum calcium concentrations were evaluated with and without correction for hypoalbuminemia or hypoproteinemia by use of published calculations,⁹ when applicable.

Statistical analysis—Demographic data (age, breed, sex, and body weight) were summarized. Severity of pulmonary disease was quantified by calculation of a weighted lobar index for each lung lobe and a lobar index for each lobe, as described.⁸ In brief, the weighted lobar index was calculated for each lobe by multiplying the weighted pattern score (alveolar pattern, 1,296; mass pattern, 216; large interstitial nodule pattern, 36; interstitial nodule pattern, 6; unstructured interstitial pattern, 1; and normal pattern, 0) for that lobe by the percentage of lobar involvement (quantified in 20% increments). The lobar index was calculated for each lobe by multiplying the pattern score (alveolar pattern, 1.5; mass pattern, 1.4; large interstitial nodule pattern, 1.3; interstitial nodule pattern, 1.2; unstructured interstitial pattern, 1.1; and normal pattern, 1) for that lobe by the percentage of lobar involvement. For each dog, the total weighted lobar index and the total lobar index were calculated by summing values for each of the 7 lung lobes examined (right cranial, right middle, right caudal, right accessory, left caudal, cranial portion of the left cranial, and caudal portion of the left cranial), and mean weighted lobar index and mean lobar index were calculated by dividing total weighted lobar index and total lobar index by the number of affected lung lobes.⁸ For dogs for which follow-up radiographs were available, the half-time for pattern resolution (ie, the time in days for a 50% reduction in weighted lobar index) was calculated for each lobe.⁸

The χ² test for homogenicity, independent t tests, and the Mann-Whitney test were used to determine whether outcome (survived vs did not survive [died or euthanized because of deteriorating clinical status despite treatment]) or response to medical treatment (evaluated on the basis of mean and total weighted lobar indices and T_1/2) was significantly associated with antifungal treatment protocol, use of specific drugs, antifungal drug dosage (below, within, or above recommended dosage range), previous antimicrobial treatment, previous corticosteroid treatment, mechanical ventilator support, and individual hematologic and biochemical parameters. The Spearman correlation method was used to determine whether use or duration of supplemental oxygen administration was correlated with outcome or response to treatment. Discriminant analyses were performed by means of forward stepwise techniques on hematologic parameters in an effort to identify factors associated with outcome.

For analyses involving outcome or survival time, only those dogs that survived and those dogs that died of blastomycosis or were euthanized because of deteriorating clinical status despite treatment (ie, dogs that did not survive) were included. Dogs that were euthanized because of financial constraints or because of a poor prognosis for recovery were excluded from these analyses. Calculations of survival times were based on the date of the initial radiographic assessment, regardless of when the diagnosis was confirmed.

All analyses were performed with standard software.^b Values of P < 0.05 were considered significant.

Results

One hundred twenty-five dogs met the criteria for inclusion in the study. Demographic data have been reported previously.⁸ Seventy-nine of the 125 (63.2%) dogs survived, 38 (30.4%) died of pulmonary blastomycosis or were euthanized because of deteriorating clinical status despite medical treatment and supportive care (ie, did not survive), and 8 (6.4%) were euthanized for reasons of cost or because of a perceived poor prognosis for recovery.

A wide variety of diagnostic tests were performed in an attempt to confirm the diagnosis of blastomycosis (Table 1). In 92 of the 125 (73.6%) dogs, Blastomyces organisms were seen in transtracheal lavage fluid, a transthoracic fine-needle aspirate of the lungs, or a lung biopsy or necropsy specimen. In the remaining 33 dogs, the diagnosis had been made on the basis of response of radiographic lesions to antifungal treatment in combination with identification of Blastomyces organisms in nonpulmonary tissues or positive results for an agar gel immunodiffusion test for anti-Blastomyces antibodies. Cross-classification of the results of the most commonly used diagnostic procedures did not reveal any significant differences in the proportions of positive test results between procedures, even though results were often discordant (Table 2).

Table 1—

Results of diagnostic procedures performed in 125 dogs with pulmonary blastomycosis.

View Table

Table 2—

Cross-classification of results of diagnostic procedures performed in dogs with pulmonary blastomycosis.

View Table

A transtracheal lavage procedure was performed in 39 dogs, and cytologic examination of lavage fluid revealed budding yeast organisms in 27 of the 39 (69.2%) samples. Intracellular bacteria were seen in 1 of the 27 (3.7%) samples with yeast organisms, but extracellular bacteria were not seen in any of the samples. Eleven lavage fluid samples were submitted for aerobic bacterial culture, but none yielded any bacterial growth. However, 4 of these dogs had previously been treated with a combination of corticosteroids and antimicrobials, and 6 of these dogs had previously been treated with antimicrobials alone. Twenty-eight of the lavage fluid samples were submitted for anaerobic bacterial culture, and 2 yielded bacterial growth. Eight of these 28 dogs had previously been treated with a combination of corticosteroids and antimicrobials, and 10 had previously been treated with antimicrobials alone. By comparison, transthoracic fine-needle aspiration was performed in 57 dogs, and yeast organisms were seen in 46 of the 57 (80.7%) aspirates. No intraor extracellular bacteria were seen during cytologic examination of the 57 aspirates, and none of the 3 aspirates submitted for aerobic bacterial culture yielded any growth.

Evaluation of medical records for the 125 dogs included in the study for information regarding medications administered prior to the diagnosis of pulmonary blastomycosis revealed that 6 (4.8%) dogs had been treated with corticosteroids; 59 (47.2%) dogs had been treated with antimicrobials; 16 (12.8%) dogs had been treated with a combination of antimicrobials and corticosteroids; 2 (1.6%) dogs had been treated with nonsteroidal anti-inflammatory drugs; 5 (4%) dogs had been treated with nonsteroidal anti-inflammatory drugs and antimicrobials; and 2 (1.6%) dogs had been treated with nonsteroidal anti-inflammatory drugs, antimicrobials, and corticosteroids. It was not possible to determine whether dosages of corticosteroids used would have been considered anti-inflammatory or immunosuppressive, and duration of treatment for dogs that had been treated with corticosteroids ranged from a single dose to a year. Prior antimicrobial, corticosteroid, or nonsteroidal anti-inflammatory drug treatment alone or in combination was not significantly associated with radiographic pulmonary pattern, lobar distribution, or weighted lobar indices.

Eight of the 125 dogs were euthanized for reasons of cost or because of a perceived poor prognosis for recovery, and 1 dog died of respiratory failure before antifungal medication could be administered. The remaining 116 dogs received antifungal treatment. Eighty-nine of the 116 (76.7%) dogs were treated with itraconazole alone, with 31 of the 89 receiving itraconazole at a loading dosage⁵ for the first 5 to 7 days of treatment. Four (3.4%) dogs were treated with ketoconazole alone, 1 (0.9%) dog was treated with amphotericin B alone, and 1 (0.9%) dog was treated with fluconazole alone. Twenty (17.2%) dogs were treated with a combination of antifungal medications (14 were treated with itraconazole and amphotericin B; 4 were treated with amphotericin B and ketoconazole; and 2 were treated with amphotericin B, itraconazole, and ketoconazole). One (0.9%) dog was treated by means of surgical resection of a pulmonary mass; this dog did not receive antifungal medication. The dosage of itraconazole was less than the recommended range⁶ (5 to 10 mg/kg [2.3 to 4.5 mg/lb], PO, q 12 h) for 14 dogs, within the recommended range for 74 dogs, and greater than the recommended range for 17 dogs. One dog received ketoconazole at less than the recommended dosage. The dosage of amphotericin B was variable, but standard protocols^5,7 were followed. The half-time for pattern resolution did not vary significantly with antifungal treatment. However, half-times for pattern resolution for dogs treated with amphotericin B alone or in combination with any other drug (mean, 35.4 days; 55 lung lobes; P = 0.005) or ketoconazole alone or in combination with any other drug (mean, 33.8 days; 32 lung lobes; P = 0.006) were significantly shorter than half-time for dogs treated with itraconazole alone or in combination with any other drug (mean, 45 days; 222 lung lobes). Outcome (survived vs did not survive) was not significantly associated with whether dogs received itraconazole at a loading dosage (yes vs no), and time to disease resolution (ie, time to resolution of radiographic evidence of disease) was not significantly different between dogs that did or did not receive itraconazole at a loading dosage.

Forty-four of the 125 (35.2%) dogs received supplemental oxygen either by means of intranasal administration or by means of housing in an oxygen cage. Mean duration of oxygen supplementation was 67 hours (range, 10 to 240 hours). Total weighted lobar index for dogs that received supplemental oxygen was significantly (P = 0.008) higher than total weighted lobar index for dogs that did not receive supplemental oxygen, but duration of oxygen supplementation was not significantly correlated with total weighted lobar index.

Five of the 125 (4%) dogs received mechanical ventilator support. Duration of mechanical ventilator support was 1, 2, 3, 5, and 21 days. There was no difference in total weighted lobar index between dogs that were and were not given ventilator support, but there was a significant (P = 0.008) difference in total lobar index between these 2 groups, with total lobar index being higher in dogs that received mechanical ventilator support. None of the dogs that received ventilator support survived.

Results of a CBC performed at the time of initial examination were available for 75 of the 125 (60%) dogs, and results of a serum biochemical panel performed at the time of initial examination were available for 47 (37.6%) dogs. The most common hematologic abnormalities included hypoalbuminemia (n = 36 [77%]), hypoproteinemia (22 [47%]), hypocalcemia (uncorrected total calcium concentration; 23 [49%]), leukocytosis (59 [79%]; median, 21.8 × 10³ WBCs/μL; range, 13.4 to 63.9 × 10³ WBCs/μL; reference range, 4.1 to 13.3 × 10³ WBCs/μL), neutrophilia (54 [72%]; median, 18.7 × 10³ neutrophils/μL; range, 11.2 to 49.1 × 10³ neutrophils/μL; reference range, 2.1 to 11.2 × 10³ neutrophils/μL), band neutrophilia (49 [65%]; median, 0.72 × 10³ band neutrophils/μL; range, 0.19 to 3.98 × 10³ band neutrophils/μL; reference range, 0 to 0.13 × 10³ band neutrophils/μL), monocytosis (49 [65%]; median, 1.95 × 10³ monocytes/μL; range, 1.23 to 9.66 × 10³ monocytes/μL; reference range, 0 to 1.2 × 10³ monocytes/μL), and lymphocytosis (59 [79%]; median, 1.61 × 10³ lymphocytes/μL; range, 0.85 to 9.59 × 10³ lymphocytes/μL; reference range, 0 to 0.8 × 10³ lymphocytes/μL). Results of the initial CBC were within reference limits in 16 of the 75 (21%) dogs. Significant correlations were found between total weighted lobar index and band neutrophil count (P = 0.001; r = 0.392), between mean weighted lobar index and band neutrophil count (P = 0.003; r = 0.305), and between outcome (survived vs did not survive) and band neutrophil count (P < 0.001; r = 0.503). Among dogs with leukocytosis, band neutrophil count for those that did not survive (median, 1.2 × 10³ band neutrophils/μL; range, 0 to 3.98 × 10³ band neutrophils/μL; n = 16) was significantly (P < 0.001) higher than count for those that did survive (median, 0 × 10³ band neutrophils/μL; range, 0 to 2.38 × 10³ band neutrophils/μL; 38). Discriminant analysis indicated band neutrophil count as the only variable associated with outcome; accuracy of using band neutrophil count to predict outcome was 84.5%.

Two of the 47 (4%) dogs for which results of serum biochemical analyses were available had hypercalcemia (ie, total serum calcium concentration > 11.5 mg/dL). With or without correction for serum albumin or protein concentration, whether dogs had an abnormal serum calcium concentration (ie, hypercalcemia or hypocalcemia) was not significantly associated with outcome (survived vs did not survive), and total weighted lobar index was not significantly different between dogs with and without a normal serum calcium concentration. Ionized calcium concentration was available for 26 of the 125 (20.8%) dogs, and only 1 dog had a high serum ionized calcium concentration (7.1 mg/dL). Serum ionized calcium concentration (high vs normal) was not significantly associated with outcome (survived vs did not survive).

Discussion

The present study was unique in regard to the attention paid to the time to radiographic resolution of pulmonary infiltrates as a function of antifungal treatment protocol. In particular, we found that the half-time for pattern resolution was significantly shorter when the treatment protocol included amphotericin B or ketoconazole than when the treatment protocol included itraconazole. However, the slightly longer half-time for pattern resolution associated with itraconazole treatment was unlikely to be of clinical relevance because of the prolonged nature of current antifungal treatment protocols. Therefore, the choice of a specific antifungal treatment protocol in any individual patient should be based on results of a cost-benefit analysis performed for that individual. Importantly, use of a loading dosage of itraconazole was not significantly associated with outcome or time to disease resolution. Thus, although some authors have suggested using a loading dosage for 5 days when beginning antifungal treatment, particularly with itraconazole,⁵ we found no added benefit of this. The addition of amphotericin B as a component of multidrug treatment protocols has also been suggested to provide a benefit in dogs with severe, life-threatening pulmonary disease owing to its rapid onset of action.¹⁰ Although we found no evidence of a higher survival rate for dogs that received amphotericin B, we did find some evidence to support a more rapid time to radiographic resolution of pulmonary infiltrates, and the effects of amphotericin B and ketoconazole on time to radiographic resolution of pulmonary infiltrates merits further investigation.

Identifying factors that could be used to predict which dogs were more likely to survive following treatment for pulmonary blastomycosis was a goal for the present study. Dogs treated for blastomycosis that were alive 4 to 5 days after diagnosis and initiation of antifungal treatment were likely to survive.⁸ Other studies^6,7 have also found that dogs that died of respiratory failure from blastomycosis did so in the first 10 days of therapy. The only hematologic factor that was found to be associated with outcome was band neutrophil count. A similar finding has been reported previously.⁷ Importantly, however, there was substantial overlap in band neutrophil counts between dogs that did and did not survive. Nevertheless, discriminant analysis indicated that the likelihood that a dog would not survive increased as the band neutrophil count increased. In addition, none of the dogs in the present study with CNS signs or that received mechanical ventilator support survived. A poor prognosis for dogs with CNS involvement has been reported previously.^6,7

The gold standard for the diagnosis of blastomycosis in dogs is identification of yeast organisms in infected tissues, but in many dogs in the present study, multiple diagnostic tests were used to establish the diagnosis. The 3 most commonly used diagnostic tests were histologic examination of biopsy or necropsy specimens of the lungs, cytologic examination of transthoracic fineneedle aspirates, and cytologic examination of transtracheal lavage fluid. Results were positive for all 28 dogs in which histologic examination of a lung specimen was performed, for 46 of 57 (81%) dogs in which a transthoracic fine-needle aspirate was examined, and for 27 of 39 (69%) dogs in which transtracheal lavage fluid was examined. Cross-classification of results of these diagnostic procedures did not reveal any significant differences in the proportions of positive test results. However, test results were often discordant. Thus, if results of 1 procedure are negative for a dog with a high index of suspicion for pulmonary blastomycosis, it may be useful to consider a second diagnostic procedure. Anecdotally, it appeared from our experience and review of medical records that the decision to perform transthoracic fine-needle aspiration versus transtracheal lavage was heavily influenced by the distribution of the infiltrate (alveolar vs interstitial) as well as by clinician experience and confidence. Importantly, the number of attempts necessary, whether with transthoracic fine-needle aspiration or transtracheal lavage, to obtain a positive test result was not controlled for in the present study, and some medical records indicated that initial samples did not contain organisms, but subsequent samples from the same patient did. Therefore, > 1 attempt may be necessary to demonstrate organisms. We recommend advising clients that numerous attempts at obtaining a diagnostic sample and, occasionally, several different diagnostic tests may be indicated to confirm the diagnosis of pulmonary blastomycosis.

In dogs suspected to have disseminated blastomycosis, clinicians will often choose to perform cytologic examinations of peripheral tissues, such as lymph nodes and skin, when there is evidence of their involvement to avoid potential complications associated with obtaining samples from the respiratory tract. Although no complications requiring treatment were reported for any of the dogs in the present study that underwent transthoracic fine-needle aspiration or transtracheal lavage, there is a potential for tracheal or pulmonary laceration, pneumomediastinum, pneumothorax, hemorrhage, and infection associated with these procedures. Thus, in patients with cutaneous or lymph node involvement, sampling of these tissues may be an alternative to collection of respiratory tract samples and may be associated with fewer potential risks for patients. In dogs in which results of cytologic examination of cutaneous or lymph node samples are negative and in dogs in which there is a concern that pulmonary changes could represent a combination of disease etiologies, such as neoplasia and blastomycosis, our preferred diagnostic test is transthoracic fine-needle aspiration. In particular, we have had success with ultrasound-guided fine-needle aspiration of peripheral lung lesions in dogs with both fungal and neoplastic pulmonary disease.

The agar gel immunodiffusion test is used to detect antibodies to the A-antigen of B dermatitidis; reported sensitivity ranges from 41% to 90%.^5,10-12 In the present study, results of the agar gel immunodiffusion test were positive in only 12 of the 24 (50%) dogs in which it was performed, and we currently do not recommend the use of the agar gel immunodiffusion test as the sole diagnostic test for confirming a diagnosis of pulmonary blastomycosis in dogs. A radioimmunoassay for the WI-1 antigen of B dermatitidis has been described and is reported to have a higher sensitivity (92%) than the agar gel immunodiffusion test¹² but is not currently available for clinical use. Preliminary results for an enzyme immunoassay to detect B dermatitidis antigen in urine or serum samples seem promising as well.^c

Concurrent bacterial pneumonia as a cause of the alveolar radiographic pattern was uncommon in the dogs in the present study. Only 3 of 94 (3.2%) dogs in which cytologic or histologic examinations or bacterial culture of pulmonary samples were performed had any evidence of concurrent bacterial infection. Therefore, we believe that an alveolar radiographic pattern affecting a single lung lobe or in combination with mixed radiographic patterns in dogs with pulmonary blastomycosis is most likely a result of the underlying fungal infection and associated inflammation, rather than concurrent or secondary bacterial pneumonia. Previous antimicrobial treatment in some dogs may have decreased the sensitivity of cytologic examination and bacterial culture, but the lack of resolution of the alveolar pattern with antimicrobial treatment supports our interpretation.

As evidenced in the present study, corticosteroids are sometimes administered to dogs with various pulmonary diseases including pneumonia. Despite the potential immunosuppressive effects of corticosteroids in dogs with blastomycosis, we were surprised to find that dogs that had been treated with corticosteroids did not have higher total or mean weighted lobar indices than did dogs that had not been treated with corticosteroids. This may, however, have been a function of the dosage used, duration of corticosteroid administration, or stage of disease when corticosteroids were administered. We are not suggesting that corticosteroid administration in dogs with blastomycosis is without hazard. We were, however, unable to demonstrate any specific effects associated with their use.

Previous authors have suggested that blastomycosis should be considered as a potential differential diagnosis in dogs with hypercalcemia.^13-15 However, hypercalcemia, regardless of whether total or ionized calcium concentrations was or was not corrected for albumin or protein concentration, was not a common finding among dogs in the present study. A detailed discussion of the role and relevance of hypercalcemia in dogs with blastomycosis has been published.¹⁶ However, corrected and uncorrected total serum calcium concentration, serum albumin concentration, and serum total protein concentration were not significantly associated with outcome in the present study.