Pulmonary Thromboembolism Associated With Blastomyces dermatitidis in a Dog

Case Report

An 8-year-old, castrated male golden retriever was referred to the University of Illinois for dyspnea, cough, and weight loss. One month prior to presentation, the dog began coughing when excited. Two weeks later, the dog developed difficulty breathing, serous nasal discharge, and a fever ranging from 39.4° to 40.0°C. The referring veterinarian obtained thoracic radiographs and performed a transtracheal wash. A mass was noted in the perihilar region on radiographs. Cytopathology from the tracheal wash revealed suppurative inflammation, and pneumonia was diagnosed. Fungal titers (i.e., coccidioidomycosis, histoplasmosis, blastomycosis, and aspergillus) were negative. The condition slowly worsened despite concurrent therapy with enrofloxacin^a for 9 days and gentamicin^b for 6 days. During this 9-day period, the dog became hyporexic and lost 3 kg of body weight.

On presentation to the University of Illinois, the dog was depressed and febrile with a temperature of 39.6°C. Respiratory rate was 45 breaths per minute with harsh lung sounds ausculted bilaterally. No cough or nasal discharge was noted. Thoracic radiographs revealed an increased soft-tissue opacity in the right caudal lung lobe characterized by an alveolar pattern. A soft-tissue opacity noted at the tracheal bifurcation was suspected to be enlarged tracheobronchial lymph nodes. An abdominal ultrasound, done to rule out neoplasia and areas of infection, revealed no abnormalities. A complete blood count (CBC) [Table 1] revealed an elevated white blood cell count (WBC) of 29.8 × 10³/μL (reference range, 6.0 to 17.0 ×10³/μL), comprised of a mature neutrophilia and monocytosis and infection or inflammation was suspected based on the CBC. A serum biochemical profile showed elevated globulins (5.3 g/dL; reference range, 2.7 to 4.4 g/dL), elevated alkaline phosphatase (171 U/L; reference range, 12 to 110 U/L), and a decreased bicarbonate (12.6 mmol/L; reference range, 15 to 27 mmol/L). An arterial blood gas had a pH of 7.44, a partial pressure of oxygen (PaO₂) of 85, partial pressure of carbon dioxide (PaCO₂) of 18, and an alveolar-arterial (A-a) gradient of 36.9 [Table 2]. Interpretation of the blood gas results was respiratory alkalosis with compensatory metabolic acidosis. An A-a gradient >30 indicates severe ventilation perfusion mismatch.¹ The dog’s condition appeared static at this time, and the owners elected to take him home over the weekend with no treatment and return on Monday for further diagnostic testing.

Upon reexamination on day 4, the dog appeared unchanged with a temperature, pulse, and respiration of 39.8°C, 140 beats per minute, and panting, respectively. Ultrasonography was used to assess the right caudal lung lobe, which had decreased aeration and a similar echogenicity to the liver. An ultrasound-guided fine-needle aspirate of the lung was performed. Cytopathological examination of the aspirate revealed peripheral blood with only a small number of nondegenerative neutrophils, macrophages, and a few epithelial cells. No bacteria, fungal organisms, or malignant cells were seen. Ophthalmological examination, looking for fungal granulomas, revealed no abnormalities.

On day 5, thoracic radiographs were repeated [Figure 1]. The right caudal lung lobe showed a more prominent alveolar pattern with air bronchograms. Additionally, an alveolar pattern was observed in the right middle lung lobe. The perihilar soft-tissue opacity was unchanged. Differential diagnoses for the pulmonary changes included fungal infection, thromboembolism, hemorrhage, edema, pneumonia, and neoplasia. The dog was anesthetized with hydromorphone^c (0.2 mg/kg body weight, intravenously [IV]) and atropine^d (0.02 mg/kg body weight, IV), followed by propofol^e (2.2 mg/kg body weight, IV). Bronchoscopy with bronchoalveolar lavage (BAL) was performed. The tracheal bifurcation appeared severely widened. Although the bronchus of the right cranial lung lobe was easily accessed, the right principal bronchus was difficult to enter due to severe narrowing. Hemorrhagic fluid was present in the trachea and in all the airways leading to the right lung lobes. All the bronchi associated with the left lung lobes appeared normal. Bronchoalveolar lavage of right and left caudal bronchi was performed. Both samples had normal nucleated cell counts of 400 cells/mL. Seventy-eight percent of the cells were macrophages, 3% were lymphocytes, and 19% were neutrophils. Although the neutrophil count was elevated, no organisms or neoplastic cells were seen. Aerobic bacterial culture of BAL fluid yielded no growth after 2 days. Following bronchoscopy, the dog was admitted to the intensive care unit for monitoring. Recovery from anesthesia was unremarkable.

As diagnostics had failed to reveal a cause of the pulmonary disease, the owners elected surgical lung biopsy. The following day (day 6), preoperative evaluation included an arterial blood gas, prothrombin time (PT), activated partial thromboplastin time (PTT), and CBC [Table 1]. The WBC count was 40.0 × 10³/μL, again comprised of a mature neutrophilia and monocytosis. Changes in the leukogram were attributed to infection or inflammation in the pulmonary tissue, with possible blood loss from the lungs accounting for the decreased hematocrit. Differential diagnoses for the pulmonary disease were unchanged from those prior to bronchoscopy.

On day 7, the dog was anesthetized for a thoracotomy with hydromorphone^c (0.2 mg/kg body weight, IV) and atropine^d (0.02 mg/kg body weight, IV), followed by propofol^e (2.2 mg/kg body weight, IV). Anesthesia was maintained with isoflurane^f in oxygen (O₂). A standard right lateral thoracotomy was performed at the sixth intercostal space. The pleural surface of the right caudal lung lobe had numerous 10- to 20-mm diameter, circular black lesions with bright red perimeters. The right cranial lobe contained two smaller (5- to 7-mm), red circular lesions on the pleural surface. The remainder of the lobes appeared grossly normal. The right caudal lung lobe, an enlarged sternal lymph node, and hilar lymph node were resected and submitted for histopathology and bacterial culture. A polypropylene chest tube was placed in the right pleural space, and a routine closure was performed. Postoperative analgesia was provided with a single dose of a combination of IV and intramuscular (IM) hydromorphone and intrathoracic bupivacaine^g (60 mg). Analgesia was maintained with subcutaneous morphine^h (0.5 mg/kg body weight, q 4 hours). Postoperatively, arterial blood gases were performed [Table 2]. Severe respiratory compromise was noted on room air, so the dog was recovered in an O₂ cage with 40% O₂. The chest tube was manually aspirated every 2 hours for 24 hours and then removed. The following day, the dog remained stable in the O₂ cage.

On day 9, the patient was hypoxic on room air [Table 2]. Bacterial cultures of the lung lobe and lymph nodes were negative. The pathology report noted that during sectioning of the right caudal lung lobe, pulmonary vessels contained large thrombi that were firmly adhered to the intimal surfaces. Histopathologically, thrombi consisted of fibrin intermixed with neutrophils and macrophages. Thrombi were multifocally covered with endothelium and perpendicularly oriented fibroblasts, indicating at least a subacute duration. Arterial walls were mildly infiltrated by neutrophils, were cuffed by haphazardly arranged collagen fibers, and contained rare, focal regions of endothelial hyperplasia. Occasional 5- to 10-mm diameter areas of coagulative necrosis, consistent with infarction, were present within the sub-pleural parenchyma. Marked type-II pneumocyte hyperplasia was noted, and the alveolar spaces contained moderate amounts of fibrin admixed with foamy macrophages. Small numbers of neutrophils were present within bronchial and bronchiolar spaces as well as within the alveolar septa. These findings were interpreted as severe pulmonary thrombosis with chronic-active, fibrinous, interstitial pneumonia.

Histopathologically, the hilar and sternal lymph nodes contained numerous secondary follicles within their cortices, consistent with reactive lymph nodes. Additionally, a severe, pyogranulomatous lymphadenitis was noted within the mediastinal lymph nodes. Gram’s and Steiner’s stains for bacterial organisms and periodic acid-Schiff and Grocott’s Gomori’s methenamine silver stains for fungal organisms were negative.

Thoracic radiographs were repeated due to the continued hypoxia. An alveolar pattern persisted in the middle lung lobe and now also involved the right cranial lung lobe. As expected postoperatively, the middle lung lobe was shifted caudally and mild pleural effusion and pneumothorax were noted. The perihilar soft-tissue opacity was still present and unchanged. Pulmonary thromboembolism was suspected, and a nuclear perfusion scan was performed. However, due to technical difficulties, the scan was not successful.

Heparinⁱ therapy (200 U/kg body weight, subcutaneously q 8 hours, initial dose IV) was initiated after coagulation times and CBC were repeated [Table 1]. The CBC noted continuing leukophilia and a declining hematocrit most likely due to pulmonary blood loss and surgery. Diagnostic tests were performed over the next 24 hours to evaluate the dog for common causes of pulmonary thromboembolism. Differential diagnoses for pulmonary thromboembolism include hyperadrenocortism, hypothyroidism, protein-losing nephropathy, immune-mediated hemolytic anemia, cardiac disease, heartworm disease, disseminated intravascular coagulation (DIC), and neoplasia.^2–4 Echocardiography and electrocardiography revealed no abnormalities. Thyroxine (T₄) was decreased at 2.4 nmol/L (reference range, 15 to 48 nmol/L), with normal triiodothyronine (1.0 nmol/L; reference range, 1 to 1.6 nmol/L) and normal thyroid-stimulating hormone (0.04 ng/mL; reference range, 0.03 to 1.0 ng/mL). This combination of findings may be due to sick euthyroid syndrome or true hypothyroidism. A heartworm antigen test was negative. An adrenocorticotropin stimulation test, to evaluate for hyperadrenocorticism, was not performed due to lack of clinical signs and the difficulty in interpreting the test in a stressed animal. Coagulation times [Table 1] were repeated and were similar to preoperative values. Further coagulation testing for detection of DIC, including fibrin degradation products and d-dimer, were also not performed due to the normal platelet count and unchanged PT and PTT. A urinalysis and urine protein to creatinine ratio for detection of protein-losing nephropathy was unfortunately not performed.

On day 10, a blood gas was repeated [Table 2]. Later that day, the dog developed sudden onset of dyspnea, followed by respiratory arrest and death.

At necropsy, a firm mass was wrapped around the right main-stem bronchus, compressing the bronchi to the right lung lobes. On cut section, the mass was tan, very firm, and contained a central area of necrosis. Several tracheobronchial lymph nodes were enlarged. Numerous blood vessels within the remaining lung lobes were expanded by fibrin thrombi. Results of necropsy indicated the probable cause of death was pulmonary thromboembolism.

Histopathologically, the peribronchial mass consisted of dense, haphazardly arranged collagen fibers interspersed with fibroblasts. Randomly dispersed throughout the mass were variably sized infiltrates of macrophages, lymphocytes, plasma cells, neutrophils, and multinucleated giant cells admixed with abundant necrotic cellular debris, indicating a pyogranuloma. Within the leukocellular infiltrates were rare, 25- to 40-μm diameter yeast bodies with thick, refractile cell walls, exhibiting occasional broad-based budding. These organisms were morphologically consistent with Blastomyces dermatitidis (B. dermatitidis) [Figure 2]. A single yeast body surrounded by a similar inflammatory cell infiltrate was also noted within the left ventricular myocardium. Adjacent cardiac myofibrils were degenerate and necrotic. No organisms were seen in the lymph nodes or any other tissue.

Other findings included a moderate membranous glomerulonephritis with associated protein cast formation within both kidneys. Lung sections showed numerous pulmonary arteries and alveolar capillaries with fibrin thrombi [Figure 3] and a mild, suppurative, histiocytic pneumonitis. Within the spleen, a fibrin thrombus was present in a single large artery, and the adjacent parenchyma was acutely necrotic, consistent with infarction. Histopathological changes within the tracheobronchial lymph nodes were similar to those described at biopsy.

Discussion

Blastomyces dermatitidis is a dimorphic fungus, with a mycelial form and a yeast form, that causes a systemic infection. Blastomycosis is endemic in the Mississippi, Missouri, and Ohio River valleys. Spores enter through the lungs and then disseminate via the blood and lymphatic systems. Rarely, the organism will cause a focal lesion when directly inoculated in a wound.

Although the majority of affected dogs (80%) have pulmonary involvement, pulmonary thromboembolism has not been noted in the literature as a sequel to blastomycosis infection. Radiographic findings of blastomycosis include diffuse nodular interstitial and bronchointerstitial patterns, solitary to multiple nodules, and tracheobronchial lymphadenomegaly. Many other tissues are commonly affected by blastomycosis, including the eye, lymph nodes, skin, and bone.⁵ Pathology of lesions include purulent to pyogranulomatous inflammation with neutrophils, macrophages, and multinucleated giant cells. Yeast bodies are often identified in these lesions.⁵

Identifying the organism on cytopathology or histopathology is the primary method of diagnosing blastomycosis. Cytopathology of fine-needle aspirates of lymph nodes, draining tracts, lung tissue, or other affected tissue often reveals pyogranulomatous inflammation and Blastomyces dermatitidis. When the organism is not identified, fungal titers may be performed. The agar-gel immunodiffusion test is 90% sensitive and specific.⁶ Early in the course of disease, the titer may be negative. Serological testing, along with clinical signs and suggestive radiographic findings, may be substituted for identification of the organism, although this method is not as accurate as demonstration of the organism.⁵ In this case, the organism was not identified on cytopathology or histopathology of the lung and lymph nodes. The perihilar mass where the organism was identified on necropsy was not able to be visualized at surgery. Additionally, fungal titers performed by the referring veterinarian were negative.

Three major risk factors exist for the development of thrombosis: endothelial injury, hypercoagulability, and alteration or stasis of blood flow. These factors are referred to as Virchow’s triad.⁷ Endothelial injury has been associated with infectious, immunological, and toxic insults and can result in initiation of either the extrinsic or intrinsic coagulation pathway. Predisposing diseases in the dog have been reviewed.^2–4 Endothelial damage can result from macrophage and neutrophil chemotaxis and phagocytosis when cellular products are released into the extracellular space. The products, including lysosomal enzymes, oxygen-derived active metabolites, and arachidonic acid metabolites, are potent mediators of endothelial injury. Inflammatory mediators, such as interleukin-1 and tumor necrosis factor released by lymphocytes and macrophages, also cause endothelial damage.⁸ Endothelial damage exposes the subendothelial collagen and leads to activation of platelets, complement, and the intrinsic coagulation cascade. These processes can lead to DIC.⁷ In this case report, inflammatory events secondary to the blastomycosis and granulomatous inflammation may have resulted in endothelial damage that could lead to DIC.

A diagnosis of DIC may be difficult to make, as no one test is diagnostic. The platelet count and clotting times are the minimal tests to diagnose DIC, with a decreased platelet count and prolonged coagulation times suggesting DIC. Further tests for products of fibrin degradation, including fibrin degradation products and d-dimer, may support the diagnosis of DIC.⁹¹⁰ In this dog, the platelet counts remained within reference range, and the PT and PTT were not prolonged, so no further tests were performed. However, the PT and PTT were shortened, which could be indicative of early DIC.⁹

Several factors may have predisposed this dog to pulmonary thrombosis. The inflammatory peribronchial mass may have resulted in endothelial damage. In addition, this mass significantly altered the normal architecture of the pulmonary tissues and may have resulted in local blood stasis. Also, cellular damage associated with the surgical procedure most likely resulted in tissue thromboplastin release, which could initiate the coagulation pathway.

While this dog did have histopathological evidence of glomerular disease, neither clinical nor hematological signs of nephrotic syndrome (such as edema, severe hypoalbuminemia, or hypercholesterolemia) were evident. Additionally, a urinalysis performed 2 days prior to the dog’s presentation at the Veterinary Teaching Hospital did not contain protein. As a result, glomerular function was considered to be intact, and loss of antithrombin III was unlikely to be the primary factor in the development of thrombosis in this animal. Unfortunately, a urinalysis, urine protein to creatinine ratio, or antithrombin III level were not repeated while the dog was hospitalized to definitively rule out proteinuria and glomerular disease.

Only one other case of embolic disease in a dog with blastomycosis is noted in the literature. In this case, the blastomycosis infection located in the hilar lymph node had eroded through the left atrial wall, resulting in thrombi formation. Sterile thrombi were found in the terminal aorta and iliac arteries. The kidney also had signs of thrombosis.¹¹

In humans, cases of embolism secondary to pulmonary fungal infections have been noted, most commonly in immunocompromised patients. In these cases, fungal emboli, usually aspergillus or Zygomycetes, result when the mycotic infection erodes into a blood vessel.¹²¹³ However, these fungi are angioinvasive, whereas blastomycosis is not. No fungal organisms were seen in the sections of the emboli from the dog in this case. No cases of pulmonary thromboembolism secondary to blastomycosis in humans or dogs were revealed in a literature search.

Conclusion

In the dog of this case report, the infection with B. dermatitidis most likely contributed to pulmonary thromboembolism through multiple pathological pathways. Inflammatory products causing endothelial damage, potential low-grade DIC, and alterations in blood flow caused by the perihilar mass all may have contributed to thrombosis. Although many differential diagnoses for pulmonary thromboembolism have been noted, any other causes of inflammation, such as blastomycosis, may cause embolic disease.