Introduction

A cholesterol granuloma is a benign nonneoplastic granulomatous structure composed of an inner cavity of cholesterol crystals surrounded by foreign body giant cells and fibrous granulation tissue, and the pathophysiology described is commonly associated with otic cholesterol granulomas.^1–3 The exact cause of formation is unknown, but it is presumed to be the result of hemorrhage, obstruction of air exchange or ventilation, and/or interference with clearance or drainage of natural secretions. For example, obstruction to cerumen drainage from the ear canal has been implicated in the creation of aural cholesterol granulomas in dogs.^1–3 As cerumen drainage is obstructed by swelling from local inflammation, the obstruction incites disruption of small capillaries. The resultant local hemorrhage increases inflammation, creating a reaction to the cerumen and the hemorrhage breakdown products, which can lead to the formation of a cholesterol granuloma. As such, cholesterol granulomas have been associated with otitis media in humans and small animals. The pathophysiology of cholesterol granuloma formation in sites other than the ear is even less well understood. Cholesterol granulomas have also been identified in the maxillary sinus and brain of dogs and in the brain, choroid plexus, and uterus of cats.^1,4–6 Several case reports describe thoracic/mediastinal cholesterol granulomas in humans, with reported concurrent hyperlipidemia, hypertension, and/or vascular diseases.^7–19 Cholesterol granulomas have also rarely been documented in the thymus, with cholesterol granuloma formation possibly associated with thymic cysts in people.¹⁷ The lack of specific imaging characteristics associated with cholesterol granulomas on computed tomography (CT) scans and the possibility of a nondiagnostic sample obtained via fine-needle aspirates can render preoperative diagnosis of these lesions challenging.

To the authors’ knowledge, this is the first report of a mediastinal cholesterol granuloma in the dog. The purpose of this report is to describe the clinical features of a mediastinal cholesterol granuloma and add to the current differential list for cranial mediastinal masses. This report also describes the treatment and outcome for a dog with a mediastinal cholesterol granuloma.

Case Report

A 10 yr old female spayed Pomeranian presented to a local emergency hospital with an acute onset of labored breathing, coughing, and lethargy. Her presenting history included boarding at a kennel facility the week prior and exhibiting repeated coughing episodes at home upon her return. She was subsequently treated for presumed kennel cough with doxycycline (3.8 mg/kg per os [PO] q 12 hr) and antitussive medication (guaifenesin 7.6 mg/kg and dextromethorphan hydrobromide 0.8 mg/kg PO every 4–6 hr). Initial examination findings at the emergency hospital showed an elevated temperature of 40.2°C, normal sinus cardiac rhythm at 150 beats per minute with no appreciable murmur, and excessive panting with loud bronchovesicular sounds. She displayed increased abdominal effort while breathing and had an oxygen saturation of 90% without supplemental oxygen. Initial diagnostics included a complete blood count (CBC) that was unremarkable and a serum biochemistry panel showing mild hyperphosphatemia 7.0 mg/dL (2.5–6.8), mild hyperglobulinemia 5.1 g/dL (2.5–4.5), and mild alanine aminotransferase enzyme elevation 206 U/L (10–125). Thoracic radiographs revealed increased soft tissue opacity within the cranial thoracic cavity suspicious for pleural effusion. The dog was administered furosemide at 4 mg/kg IV as needed, maropitant^a 1 mg/kg IV q 24 hr, enalapril 0.38 mg/kg PO q 12 hr, and pimobendan 0.38 mg/kg PO q 12 hr. She was administered 40% oxygen supplementation for 24 hr and was transitioned to furosemide 1.9 mg/kg PO q 12 hr. The dog appeared to be clinically stable and breathing comfortably and thus was discharged. Thoracic radiographs were not repeated before discharge. The dog was breathing comfortably and showing improved energy and increased stamina at home. However, she was overall still not completely normal and showed a decreased appetite, so she was brought to her primary veterinarian 3 days after being discharged from the emergency hospital. On examination she was bright and alert and panting, and a mild cough was induced with palpation of the trachea, so thoracic radiographs were taken, which showed increased soft tissue opacity within the cranial thorax. The dog was then referred by her primary veterinarian to the Virginia-Maryland College of Veterinary Medicine Veterinary Teaching Hospital for further evaluation.

Upon presentation to the Virginia-Maryland College of Veterinary Medicine Veterinary Teaching Hospital, the dog was anxious, febrile (temperature of 39.8°C), and panting, with a heart rate of 130 beats per minute, strong synchronous pulses, and a normal sinus rhythm. She had a body condition score of 9/9. The dog was panting heavily but did not exhibit an increase in respiratory effort, and thoracic auscultation was unremarkable with no appreciable cardiac murmurs, respiratory crackles, or wheezes. She was mildly tense on abdominal palpation and had a left medial luxating patella (grade 2/4). No other abnormalities were detected on physical examination. Systemic blood pressure measurements were not obtained. A CBC and serum biochemistry panel showed a persistent hyperphosphatemia 6.1 mg/dL (1.9–4.4), hypercalcemia 10.8 mg/dL (9.4–10.7), hyperglobulinemia 4.1 g/dL (2.1–3.8), elevated alanine aminotransferase 182 U/L (16–75), elevated alkaline phosphatase 195 U/L (8–70), hypernatremia 157 mg/dL (143–152), and elevated creatine kinase 304 U/L (32–193). Total T4 was within normal limits, activated partial thromboplastin time was normal, and prothrombin time was slightly decreased 5.5 sec (6.3–8.4). The clinical suspicion for a coagulopathy was low, and the decreased prothrombin time was presumed to be a sampling error. An echocardiogram was performed, which showed normal cardiac structure and function, as well as a cavitated mass cranial to the cardiac silhouette. Subsequently, thoracic and abdominal CT scans were performed to further delineate the mass in the thoracic cavity and for staging purposes. The dog’s blood pressure measurements while under anesthesia for her CT scans demonstrated normotension. The CT scans revealed a soft tissue attenuating mass (4 × 4 cm) cranial to the base of the heart with pinpoint foci of mineralization along the periphery with thin rim contrast enhancement (Figure 1). The mass was noted to have caused right dorsolateral displacement and compression of the cranial vena cava. Additional findings included mild bilateral pleural effusion, right adrenomegaly, multiple small poorly defined splenic nodules, bilateral renal cortical cysts, multifocal thoracolumbar intervertebral disc disease, and multiple subcutaneous nodules. Ultrasound-guided tru-cut biopsies using a 16 gauge, 9 cm needle, and fine needle aspirates of the mass were obtained for histopathological and cytological evaluation and culture/sensitivity testing. Cytological evaluation showed a predominance of degenerate neutrophils with fewer small lymphocytes, a few scattered sheets of poorly preserved, polygonal epithelial cells with dark basophilic cytoplasm, and pyknotic nuclei with minimal anisocytosis along with a few foamy macrophages and frequent cholesterol crystals (Figure 2). The cytopathologist’s opinion was mixed inflammation, necrosis, and probable thymoma based on the presence of epithelial cells and small lymphocytes in a cranial thoracic mass. Four biopsy samples ranging in size from 1 × 1 × 1 mm to 1 × 2 × 3 mm were submitted for evaluation. The histology report revealed small slightly polygonal densely staining cells arranged in nests and trabeculae and supported by a connective tissue stroma that was often necrotic. Cells had distinct borders, scant eosinophilic cytoplasm, and round nuclei with finely stippled chromatin and indistinct nucleoli. Anisocytosis and anisokaryosis were minimal and no mitoses were observed in ten 40× fields. A small number of acicular clefts suggestive of the presence of cholesterol were noted at one border of a biopsy sample (Figure 3). All histological findings apart from the small number of acicular clefts were suspicious for thymoma, with carcinoma less likely owing to little atypia. Culture of the aspirate samples was negative for microbial organisms. Given the lack of cardiac abnormalities, the previously prescribed pimobendan, furosemide, and enalapril were discontinued. Surgical resection of the intrathoracic mass was recommended, and the dog returned to the Virginia-Maryland College of Veterinary Medicine Veterinary Teaching Hospital 6 wk later for surgery. Preoperative thoracic radiographs, CBC, and serum biochemistry did not reveal any clinically significant changes from the previous visit except for a mildly elevated cholesterol level of 334 mg/dL (192–332).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

The dog was premedicated with methadone 0.39 mg/kg and acepromazine 0.02 mg/kg intramuscularly and maropitant^a 1 mg/kg subcutaneously. Anesthesia was induced with alfaxalone (1.2 mg/kg) IV and maintained with an inhalant anesthetic (sevoflurane) throughout the procedure. The patient was also placed on fentanyl (20 μg/kg/hr), lidocaine (2 mg/kg/hr), and ketamine (1 mg/kg/hr) constant rate infusions (CRIs) throughout anesthesia. Cefazolin was administered at 22 mg/kg IV starting 60 min before the first incision and every 90 min thereafter until the end of the procedure. The patient was placed in dorsal recumbency and a median sternotomy was performed as described, sparing the xiphoid process to preserve a point of stability.²⁰ A 4 cm round and well-encapsulated mass was present within the cranial mediastinum, moderately adherent to the mediastinal pleura, cranial vena cava, and the base of the heart. The outer rim of the mass was characterized by a pale tan color punctuated with multifocal areas of yellow caseous material. Before mass dissection, a sternal lymph node was identified, dissected, and excised using monopolar electrocautery. The mass was carefully dissected away from the surrounding tissues with a combination of digital dissection, blunt dissection using Lahey forceps, and a vessel-sealing device^b. Feeder vascular branches to the mass were double ligated with 3-0 polydioxanone suture (PDS) and sharply divided using Metzenbaum scissors. The mass was resected without complications. An image of the resected mass is represented in Figure 4. No iatrogenic trauma to the surrounding structures or phrenic nerve during mass resection were observed. The pulmonary parenchyma and heart appeared grossly normal. A 12 g chest tube^c was inserted in approximately the 8th intercostal space and secured to the chest wall with 3-0 prolene using simple interrupted sutures. The intercostal muscles and subcutaneous tissues were injected with liposomal bupivacaine^d at 5.3mg/kg, diluted with saline to a total volume of 20 ml. The sternebrae were apposed with 0 PDS in a figure of eight pattern. The pectoral muscles were apposed with 3-0 PDS in a simple interrupted pattern. The subcutaneous layers were apposed in multiple simple continuous layers of 3-0 monocryl. The subcutis was apposed with 3-0 monocryl in a continuous vertical mattress pattern followed by a skin closure with 3-0 prolene in a ford interlocking pattern.

View Figure

• Download as Powerpoint
• Download Figure

Upon anesthetic recovery the dog was stable but oxygen dependent and placed on 40% oxygen support for 24 hr. Postoperative care included analgesia via carprofen at 2.2 mg/kg subcutaneously twice daily and a fentanyl CRI (3 μg/kg/hr), lidocaine CRI (25 μg/kg/min), and ketamine CRI (0.6 mg/kg/hr). Cefazolin was continued at 22 mg/kg IV q 8 hr until 24 hr after surgery. Pain scoring was performed every 6 hr, and the patient maintained pain scores of less than 2 out of 4 on the Colorado State Canine Acute Pain Scale. The patient remained normotensive throughout the period of her post-surgical recovery in the hospital. The chest tube did not yield any air production or any significant fluid production after surgery. The dog was weaned off her IV medications over 2 days and was transitioned to the following oral medications: codeine 1.14 mg/kg PO q 8 hr, carprofen 1.9 mg/kg mg PO q 12 hr, and trazodone 3.8 mg/kg PO q 12 hr as needed for anxiety. The dog was discharged 3 days postoperatively with no signs of respiratory difficulty.

Histopathological evaluation of the mass revealed a large central region of coagulation necrosis with scattered neutrophils. The periphery of the necrosis had large clusters of acicular clefts surrounded by macrophages and multinucleate giant cells. The macrophages were large with abundant foamy cytoplasm and the granuloma was surrounded by a thick rim of mature fibrous tissue. No thymic tissue was identified on histopathological evaluation of the mass. All of the histopathological findings of the mass were consistent with a cholesterol granuloma (Figure 5). The histopathological evaluation of the lymph node revealed subcapsular and medullary sinuses filled with large macrophages with abundant foamy cytoplasm, indicative of sinus histiocytosis.

View Figure

• Download as Powerpoint
• Download Figure

Four days after discharge, the dog was brought to her primary veterinarian because of an acute increase in respiratory effort. For the previous 2 days at home, she had developed intermittent wheezing episodes that were self-limiting after long periods of rest. On the morning of presentation to her primary veterinarian, she exhibited persistent wheezing and coughing with possible vomiting. At her primary veterinarian’s clinic, the dog was mildly dyspneic and wheezing with pink mucous membranes and moderate serosanguinous discharge from the incision site. Thoracic radiographs showed scant pleural effusion but no obvious signs of pulmonary contusions, pneumonia, or pneumothorax. The dog was referred to the Virginia-Maryland College of Veterinary Medicine Veterinary Teaching Hospital for supportive care. Repeat thoracic radiographs were performed by the emergency service upon the dog’s arrival at the Virginia-Maryland College of Veterinary Medicine Veterinary Teaching Hospital, which showed mildly progressive pleural effusion but no evidence of pneumonia, tracheal collapse, or pulmonary consolidation. Cytological evaluation of the pleural fluid was consistent with a modified transudate, suggestive of routine postsurgical fluid production. The dog was placed on 40% oxygen support and IV fluid therapy at 40 mL/kg/day and started on ampicillin-sulbactam at 30 mg/kg IV every 8 hr for presumptive aspiration pneumonia. The dog’s clinical signs improved greatly after 24 hr of oxygen supplementation and the dog was discharged 2 days later with oral Clavamox^e (19mg/kg PO q 12 hr). The differentials postulated for the acute respiratory signs included a mild pulmonary thromboembolism, aspiration pneumonia, or tracheal collapse. The dog continued to do well at home after discharge. Five months later, the dog presented to her primary veterinarian for vomiting and lethargy for 2 days. Thoracic radiographs showed no obvious abnormalities and no signs of mass recurrence, and she improved with medical management for possible pneumonia. Nearly 7 mo postoperatively, the owners reported no recurrence of respiratory signs, a normal activity level, and an improved (reduced) body condition.

Discussion

Cholesterol granulomas are benign lesions that have been reported in dogs, cats, horses, and people, but non-otic cholesterol granulomas are rare.^{1,4–6,17,19,21–23} Earlier reports often interchanged terminology between cholesterol granuloma and cholesteatoma, with the latter distinctly different histologically.^1,2,5 A cholesteatoma lacks any presence of cholesterol granules and is an epidermoid cyst composed of keratinizing stratified squamous epithelium lining a cyst-like structure containing a lamellae of keratin.² Cholesterol granulomas on the other hand are lesions composed of large numbers of cholesterol crystals surrounded and engulfed by foreign body giant cells further encased by fibrous granulation tissue.² Histologically, instead of cholesterol crystals, cholesterol granulomas contain acicular clefts, which are indicative of the presence of cholesterol. The processing of histology sections for evaluation removes lipid, thus leaving the cleft caused by the cholesterol but not the crystal. The processing of samples for cytological evaluation does not remove lipid, thus leaving the cholesterol crystals intact for identification. Cholesterol granulomas and cholesteatomas have been documented to occur concurrently with certain comorbidities such as otitis media in humans and dogs.^1–3 Overlapping histopathological changes between both of these conditions and their associations with similar disease processes such as otic disease have made differentiation between cholesterol granulomas and cholesteatomas difficult in early reports with limited descriptions.^{1,2,4,6,21,24,25} These factors and the scarcity of information, especially on cholesterol granulomas, have led to confusion in terminology between the two disease entities, as well as potential inaccuracy in the determination of the historical prevalence of cholesterol granulomas in domestic animals.^{1,2,4,6,21,24,25}

The pathogenesis of cholesterol granulomas remains unclear; however, the main contributory components to their development include hemorrhage, occlusion of air flow or ventilation, and/or interference with clearance or drainage of natural secretions or inflammatory debris. With trauma or severe inflammation, damaged erythrocytes can release large amounts of cholesterol from their membranes that coalesce to form a foreign body–like inflammatory reaction walled off by phagocytes and multinucleated giant cells. These features are consistent with previous histopathology reports describing hemorrhage noted within and adjacent to a cholesterol granuloma, leading to mass discoloration.^1,25 The contributory effects of air flow obstructions, impaired drainage of secretions, and inflammation have been supported by experimental studies in which cholesterol granuloma formation was experimentally induced in the cat, chinchilla, and squirrel monkeys by obstructing the Eustachian tube (with or without introduction of foreign material into the middle ear).^26–28 Despite these studies, the exact mechanisms by which air flow obstruction and impaired secretory drainage contribute to cholesterol granuloma formation remain poorly understood, especially in locations apart from the ear. Most clinical reports of cholesterol granulomas in domestic animals have been associated with the tympanic cavity; however, other reported locations include the choroid plexus, maxillary sinus, brain, and uterus.^1,4,5 Furthermore, there are several reports of thoracic/mediastinal cholesterol granulomas in people.^7–19 Within some of these reports and those of domestic animals, additional factors that may be associated with cholesterol granuloma formation include hyperlipidemia and hypercholesterolemia.^{9,10,16,17,19,25,29} Moreover, multiple human case reports demonstrate concurrent hyperlipidemia, hypertension, and/or vascular disease such as aortic aneurism, and coronary artery disease.^{9,10,14–17,19} One theory postulates that high-fat diets and high accumulation of peripheral blood lipid content could lead to buildup of components of lipid deposits within the tissues. However, the exact association of the pathogenesis of cholesterol granulomas with hyperlipidemia and hypercholesterolemia has not been thoroughly investigated. In human reports, a relation between hyperlipidemia and vascular disease has been proposed.¹⁶ Hyperlipidemia may lead to atherosclerosis, thereby weakening vessel walls.¹⁶ The inflammatory infiltrates and hemorrhage that surround cholesterol clefts may suggest the presence of vascular disease.¹⁶ Hyperlipidemia may also lead to hypertension within people.¹⁶ However, any association between hypertension and cholesterol granuloma formation in dogs is undetermined at this time. With the dog in this case report, her obesity as well as her elevated cholesterol levels at the time of surgical consultation may potentially be contributory factors to the formation of her mediastinal cholesterol granuloma. The dog’s blood pressure measurements while she was under anesthesia for her presurgical CT scan and during the immediate postsurgical recovery period in the hospital demonstrated normotension.

The dog in this report had a cholesterol granuloma within the cranial mediastinum, where occlusion of air flow does not seem likely as a contributory factor. However, impaired lymphatic drainage and hemorrhage could have occurred as a result of a primary mass growing in a confined space, leading to inflammation and rupture of small capillaries from local invasion. The dog in this report had no history or physical exam findings consistent with trauma causing hemorrhage, thus making local infiltration/obstruction of blood vessels and lymphatic drainage more likely factors contributing to the formation of a cholesterol granuloma in this dog.

Cholesterol granulomas have been found to be associated with the thymus in humans, albeit the reported incidence is very rare.^17,19 In these reports, rupture of the walls of thymic cysts and resultant inflammation were theorized to lead to cholesterol granuloma formation. The initial histopathological diagnosis obtained from biopsies of the mediastinal mass in this report was a suspected thymoma with necrosis, which differs from the final histopathological diagnosis of a cholesterol granuloma obtained after surgical resection of the mass. A possible explanation for the discrepancy in histopathological diagnoses of the mass in this report is that representative samples were not obtained during the initial biopsy. The biopsy samples may not have contained scattered granuloma clefts, which can be seen with more common lesions in the mediastinum such as multilocular thymic cyst or thymomas, and the combination of lesion location and a nonrepresentative biopsy sample may have led to the discrepancy in histological diagnosis of the biopsy sample versus the mass after surgical resection.¹⁹ Another hypothesis for this difference in histopathological diagnoses is the original thymic mass experienced a compromise to its blood supply, leading to capillary rupture and subsequent granuloma formation. However, unlike the reports of cholesterol granulomas in humans associated with the thymus,^17,19 where residual cells of thymic origin were identified in conjunction with the cholesterol granuloma, the histopathological evaluation of the resected mass in this case report did not reveal any residual thymic cells.

The treatment of choice for cholesterol granulomas is surgical excision. The purpose of excision is to remove the initial inciting inflammation and to establish drainage and air flow to the site, as the obstruction of drainage and air flow is frequently the issue with those cholesterol granulomas associated with otitis media. In the few veterinary and human reports of cholesterol granuloma excision of lesions in locations appropriate for surgical removal, the overall outcome was excellent with no recurrence of clinical signs, even in humans with thymic cholesterol granulomas.^1,17,25

For cranial mediastinal masses removed via a median sternotomy, potential complications of surgery include persistent pneumothorax, development of pleural effusion, osteomyelitis of the sternotomy site, and incisional dehiscence. The dog in this report had mild residual pleural effusion a few days postoperatively, was moderately dyspneic, and responded to oxygen therapy. A mild pulmonary thromboembolus is a likely differential to explain these signs and response to therapy, especially because no evidence of aspiration pneumonia or tracheal collapse was noted with thoracic radiographs. The dog improved with supportive care, and the exact cause of her acute respiratory decline 1 wk after surgery remains undetermined. Repeated radiographs 5 mo after surgery showed no recurrence of the cranial mediastinal mass, and the dog exhibited no further respiratory signs.

Conclusion

To the authors’ knowledge, this is the first report of a cholesterol granuloma in the cranial mediastinum of the dog. The pathogenesis of the mediastinal cholesterol granuloma in this dog remains undetermined. Surgical excision successfully removed the mass and resulted in resolution of clinical signs with no signs of recurrence at 7 mo after surgery. The current differential list for a cranial mediastinal mass should be expanded to include cholesterol granulomas. Further investigations such as a multicenter retrospective study into the biologic behavior of canine mediastinal cholesterol granulomas can provide valuable client information regarding the outcomes of surgical excision and prognosis following complete resection.