Mesenteric Vascular Thrombosis Associated with Disseminated Abdominal Visceral Hemangiosarcoma in a Cat

Introduction

Neoplasia is a well documented predisposing factor for thromboembolic disease and hemostatic abnormalities. Neoplasia can result in thrombosis in several ways including direct vessel occlusion, presence of tumor thrombi, and coagulation cascade activation. The pathogenesis of thrombosis in malignancy is complex, and the multifactorial etiology has not yet been fully elucidated.

Feline visceral hemangiosarcoma is a highly malignant neoplasm characterized by early and aggressive metastasis, and metastasis secondary to rupture of the primary tumor has been suggested. This neoplasm has been previously reported in cutaneous, subcutaneous, and visceral (thoracic or abdominal) locations. Feline visceral hemangiosarcoma is most frequently diagnosed in the spleen, liver, and intestines.

This report describes a case of disseminated thrombosis of mesenteric vasculature associated with intra-abdominal hemangiosarcoma in a cat. To the authors' knowledge, this presentation has never been described in the literature and is a unique finding in this case.

Case Report

An 8 yr old male castrated domestic shorthair weighing 7.07 kg was examined at The Elmer & Mamdouha Bobst Hospital of the Animal Medical Center in New York City for a 4 day history of lethargy and partial anorexia. No previous pertinent medical history was reported aside from a history of feline allergic bronchitis diagnosed 2 yr prior to presentation and a grade 2/6 systolic heart murmur diagnosed 1 yr prior to presentation. There was no reported dietary indiscretion or toxin exposure. The cat was housed indoors and was receiving no medications.

On presentation, the cat was hypothermic (rectal temperature was 33.2°C) and obese (body condition score was 8/9). The cat had tacky mucous membranes and prolonged skin tent, consistent with 7–10% dehydration. Physical examination revealed a prolonged capillary refill time, moderate dental tartar, and a grade 2/6 left parasternal heart murmur with intermittent gallop rhythm. The cat was painful and tense on abdominal palpation, and percussion on one side of the abdomen transmitted a wave felt on the opposite side of the abdomen (fluid wave), which was suspected to be consistent with peritoneal effusion. The remainder of the physical examination was unremarkable.

Initial blood gas analysis of peripheral venous blood was consistent with a combined respiratory and metabolic acidosis (pH was 7.064; reference range, 7.35–7.45; partial pressure of carbon dioxide was 51.5 mm Hg; reference range, 35.0–45.0 mm Hg; partial pressure of oxygen was 25.0 mm Hg; reference range, 75.0–100.0 mm Hg; bicarbonate was 14.4 mmol/L; reference range, 18–23 mmol/L, and base excess was −16.0 mmol/L; reference range, –2 to +2 mmol/L). Other abnormalities on blood work included hyponatremia (134.8 mmol/L; reference range, 135.0–148.0 mmol/L), hypocalcemia (4.0 mg/dL; reference range, 4.5–5.3 mg/dL), hyperglycemia (125 mg/dL; reference range, 67–93 mg/dL), and hyperlactatemia (3.50 mmol/L; reference range, 0.50–2.00 mmol/L). The remainder of the baseline blood work (potassium, chloride, and anion gap) was within normal limits.

Anemia (hematocrit was 24.1%; reference range, 29–45%; hemoglobin was 7.22 g/dL; reference range, 9.5–15 g/dL; red blood cells were 4.40 million/μL; reference range, 6.0–10.0 million/μL) with evidence of regeneration (mild polychromasia and anisocytosis, 1.9% reticulocytes; reference range, 0.0–1.0%, and absolute reticulocytes were 83,600/μL; reference range, 0–50,000/μL), neutrophilia and lymphoctyopenia consistent with a stress leukogram (absolute neutrophils were 13,795/μL; reference range, 2,500–12,500/μL, and absolute lymphocytes were 1,240/μL; reference range, 1,500–7,000/μL), and thrombocytopenia (platelet count was 120thou/μL; reference range, 150–600 thou/μL), with scattered platelet clumping. The remainder of the complete blood count was within reference ranges.

Decreased activity of alanine aminotransferase (15 IU/L; reference range, 28–76 IU/L), panhypoproteinemia (albumin was 2.1 g/dL; reference range, 2.5–3.9 g/dL, total protein was 3.8 g/dL; reference range, 5.9–8.5 g/dL, and globulin was 1.7 g/dL; reference range, 2.4–5.3 g/dL), hyperbilirubinemia (0.7 mg/dL; reference range, 0.0–0.4 mg/dL), hyperuremia (blood urea nitrogen was 77 mg/dL; reference range, 15–34 mg/dL), hypocalcemia (6.6 mg/dL; reference range, 8.2–11.8 mg/dL), and hyponatremia (136 mEq/L; reference range, 145–158 mEq/L) were noted on the serum biochemical profile. Total thyroxine was low at <0.5 μg/dL (reference range, 0.7–4.9 μg/dL). The remainder of the chemistry panel was within normal limits.

Thoracic radiographs (Figure 1) demonstrated a small heart size and contracted pulmonary vessels, consistent with hypovolemia. The visible portion of the abdomen on these radiographs was severely distended, with loss of serosal detail. The cranial ventral abdomen had a “ground glass” appearance, suggestive of carcinomatosis, steatitis, or pneumoperitoneum with effusion. Treatment was instituted with IV boluses of an isotonic crystalloid solution^a and a colloid solution^b.

View Figure

• Download as Powerpoint
• Download Figure

Abdominocentesis was performed, and 0.3 mL of a straw-colored flocculent fluid was withdrawn. Cytology of the effusion revealed degenerate neutrophils and scattered erythrocytes, consistent with a suppurative effusion. No intracellular bacteria were seen. Biochemical analysis of the peritoneal fluid was performed on a standard blood gas analyzer^c. Analysis of the effusion revealed a pH of 6.897, 8% hematocrit, 6.0 g/dL total protein, and 8.71 mmol/L lactate. The lactate level of the effusion was 2.4 times higher than the blood lactate level. Effusion glucose measured 10 mg/dL, which differed from the blood glucose level by 115 mg/dL.

Septic peritonitis was suspected based on the biochemical and physical examination findings. Surgical exploration of the abdomen was performed. On entrance into the peritoneal cavity, a large amount of serosanguinous fluid was noted. The entire peritoneal surface was noted to be erythematous and had petechiation. The omentum was discolored (purple) and had multiple intestinal fibrinous adhesions to the serosal surface of the gastrointestinal tract. The liver was green to tan in color, but otherwise appeared grossly normal. The gallbladder appeared to be of normal size and was easily expressible on palpation. The distal tip of the right limb of the pancreas was firm and friable. No pulses were appreciated on digital palpation of the mesenteric vasculature. Gross thromboses were noted in multiple mesenteric vessels (Figure 2). The intestines were white in color with complete ileus (Figure 3).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

During surgical exploration, the patient became acutely hypotensive (systolic blood pressure was <60 mm Hg). Fluid therapy with colloid solution boluses was instituted; however, asystole was noted shortly after. Cardiopulmonary resuscitation was attempted via injections of IV atropine(0.059 mg/kg) and IV epinephrine (0.11 mg/kg). These attempts were unsuccessful.

On gross examination of organs at necropsy, there was widespread red to red/purple discoloration of adipose tissue throughout the peritoneal cavity. Omental tissue was soft, greasy, friable, and discolored (a red/gray color). The peritoneal surface was dark red. Diffuse linear (“paintbrush”) hemorrhages covered 30% of the surface of the thoracic diaphragm. The urinary bladder serosa was a diffusely red/purple color, and the jejunal serosa was a dull, gray/tan color. Firm white nodules (suspected lymph nodes) measuring 0.1–0.2 cm in diameter were present in the mesentery adjacent to the right limb of the pancreas. The pancreas itself was noted to be small, thin, and pink.

Histologic evaluation of the abnormal tissues listed above revealed extensive infiltration of short, plump, irregular neoplastic spindle cells in the peritoneal adipose tissue, as well as the pancreas, urinary bladder, intestine, and diaphragm. The neoplastic population had round to oval nuclei with coarsely stippled chromatin and prominent nucleoli and had moderate quantities of eosinophilic cytoplasm. Cells displayed a high nucleus-to-cytoplasm ratio with moderate anisocytosis and anisokaryosis. Mitoses averaged 1–3/high-power field. Cells were present in disorganized sheets and indistinct bundles and were sometimes associated with branching collagenous septae. In many areas, rounding of the neoplastic cells, dissociation/loss of cohesion, and occasional small foci of necrosis were noted, consistent with degeneration of the population. Hemorrhage was frequently present. Rare areas of rudimentary vascular differentiation were seen. A lymph node adjacent to the pancreas was almost completely effaced by neoplastic cells. Several small to medium diameter mesenteric arterioles in areas of neoplasia contained organizing fibrin thrombi. No areas of intestinal infarctive necrosis were identified, and no neoplastic cells were identified within vessels.

Cell morphology was felt to be consistent with a poorly differentiated angiosarcoma. Immunohistochemistry for factor VIII antigen was performed on sections of affected urinary bladder, diaphragm, and omentum, which showed moderate to strong, but diffuse, positive cytoplasmic staining of neoplastic cells, confirming endothelial cell origin. The frequent presence of hemorrhage in association with neoplasia together with the gross appearance of the tumor, were consistent with hemangiosarcoma. To summarize, the most noteworthy findings were disseminated abdominal visceral hemangiosarcoma with associated mesenteric vascular thrombosis.

Discussion

Feline visceral hemangiosarcoma has been previously reported, and neoplasia is a well-documented predisposing factor for thrombosis.^1,2 To the authors’ knowledge, disseminated mesenteric thrombosis associated with intra-abdominal neoplasia in a cat has never been described in the literature and is a unique finding in this case.

Thromboembolic disease and hemostatic abnormalities are common in patients with cancer.² Numerous reports exist of neoplasia in association with systemic arterial thrombosis in veterinary patients.^3,4 The pathogenesis of thrombosis in malignancy is complex, and the multifactorial etiology has not yet been fully elucidated.

Tumor cells can indirectly and directly enhance coagulation cascade activation. Neoplasia can predispose to thrombosis in numerous ways, including increased coagulation (due to platelet activation, thromboplastin release from tumor cells, or production of an activator of factor X), reduced clotting factor clearance, reduced clotting factor neutralization, and decreased fibrinolysis.² In addition, malignant cells produce cytokines, such as interleukin-1, interleukin-6, and tumor necrosis factor-α, which can suppress anticoagulant activities and promote coagulation.^5–7 Furthermore, neoplasms may arise from or erode into blood vessels, leading to the direct disturbance of endothelial integrity and disruption of blood flow.³ Masses of tumor cells can also break off of the primary tumor and circulate as tumor thrombi.⁸

Thrombosis of the external iliac artery and terminal aorta with secondary ischemic necrosis of the hind limb muscular and dermal tissue has been reported in a dog with angiosarcoma of the skeletal muscle of the limb.⁹ In that case, thrombosis was suspected to be the result of systemic hypercoagulability following endothelial damage to the vascular tumor bed, as well as altered blood flow characteristics.⁹ To the authors’ knowledge, there are no reports of thrombosis of mesenteric vasculature associated with hemangiosarcoma in the cat. In this present case, thrombosis was accompanied by evidence of intestinal ischemia at surgery, but not by histologic evidence of intestinal infarction. This suggests either acute development of the vascular lesion or the presence of collateral circulation to affected segments. True infarctive necrosis is a coagulative necrosis of the full thickness of an entire segment of intestine. The microscopic appearance of coagulation necrosis is such that the normal tissue architecture with its cellular components is still recognizable, but the nuclei display characteristic changes, such as karyolysis, pyknosis, and/or karyorrhexis of nuclei.¹⁰

Although rare in humans, mesenteric thrombosis of the superior mesenteric artery with subsequent bowel ischemia has been reported in association with primary and metastatic abdominal neoplasia of the upper gastrointestinal tract (i.e., pancreas, liver).¹¹ Diagnosis of this condition can be difficult, but can be achieved using abdominal sonography, angiography, or computerized tomography.¹² The mortality rate is high (59–100%), with rapid diagnosis after onset of symptoms being associated with significantly improved survival.¹²

Treatment options for tumor-induced thrombosis are limited. Successful treatment of mesenteric vascular thrombosis is dependent on rapid, early diagnosis and aggressive treatment.¹² Resection of infarcted bowel is necessary, and anticoagulant therapy is warranted in most cases.¹¹ In humans, revascularization techniques can be performed via thrombectomy, angioplasty, aortomesenteric bypass, or vascular stenting.^11–14 Utilization of these surgical techniques may not be applicable to feline patients at this time due to the requirement for highly specialized equipment and/or highly technical microprocedures. In the cat reported here, bowel ischemia was too diffuse to allow successful surgical resection, had the cat survived.

Angiosarcomas (i.e., hemangiosarcomas or lymphangiosarcomas) are rare malignant sarcomas of vascular endothelial cell origin.¹⁵ A survey of 3,145 necropsies at one institution found only 18 cases of hemangiosarcoma and 1 case of lymphangiosarcoma highlighting the extremely low incidence of these tumors in the feline population, but this statistic has not been re-examined since 1977.¹⁵ Immunohistochemistry for vascular endothelial markers (e.g., factor VIII, CD31, PROX-1) may be required for definitive diagnosis.^16–19 In cases of poor cellular differentiation, morphologic features alone are insufficient to distinguish hemangiosarcoma from lymphangiosarcoma.^16,17 Hemangiosarcoma can often be differentiated from lymphangiosarcoma by the presence of hemorrhage and numerous erythrocytes within vascular structures.¹⁵

Cutaneous, subcutaneous, and visceral (thoracic or abdominal) hemangiosarcoma have all been reported in cats.¹⁸ Although these neoplasms arise from vascular endothelium and can occur in any tissue, feline visceral hemangiosarcomas are most frequently diagnosed in the spleen, liver, and intestines.^1,19 As in the current case, hemangiosarcoma associated with the pancreas, diaphragm, and omentum has also been reported.¹

Feline hemangiosarcoma is a highly malignant neoplasm characterized by early and aggressive metastasis (as high as 77% at the time of diagnosis) most commonly to the lungs, liver or abdominal lymph nodes.^1,15,19 Intra-abdominal spread of the tumor is also possible, and metastasis secondary to rupture of the primary tumor has been suggested.¹ This may have occurred in this cat as no primary site of the tumor could be identified.

Clinical signs in this case included abdominal pain and peritoneal effusion. Septic peritonitis was suspected based on analysis of peritoneal effusion. No intracellular bacteria were observed on cytologic examination of the peritoneal effusion, but this method of diagnosis is only 86% accurate in diagnosing septic effusion.²⁰ Peritoneal effusion lactate concentration and blood to fluid lactate difference has been used as a marker for septic peritonitis in dogs; however, peritoneal fluid lactate concentration and a blood-to-fluid lactate comparison are not accurate at detecting septic peritoneal effusion in cats.²¹ Bonczynski and colleagues (2003) reported that a blood-to-peritoneal fluid glucose difference >20 mg/dL was 86% sensitive and 100% specific for the diagnosis of septic peritonitis in cats.²⁰ In this case, the blood-to-fluid difference in glucose was 115 mg/dL and was the primary indication for abdominal exploratory surgery.

No primary infectious focus or rupture of the gastrointestinal or urogenital tract was discovered either at the time of surgery or on gross exam of organs at necropsy. It is suspected that peritoneal effusion was secondary to extensive abdominal neoplasia and vascular thromboses, which caused inflammation and occlusion of vasculature.

Conclusion

Tumor-associated vascular thrombosis is rare. Thromboses can occur due to direct vessel occlusion, presence of tumor thrombi, and indirect effects on coagulation by the tumor. Management of these cases must be aggressive, including anticoagulant therapies, removal of the primary tumor, and, when possible, specific treatment of thrombosis. Complete thrombosis of the mesenteric vasculature secondary to feline visceral hemangiosaroma is possible and can be fatal. Early intervention is vital for positive outcomes.