Use of Recombinant Tissue-Plasminogen Activator in a Dog With Chylothorax Secondary to Catheter-Associated Thrombosis of the Cranial Vena Cava

Case Report

A 4-year-old, castrated male Maltese was referred to the Cornell University Hospital for Animals for management of blunt abdominal trauma. The dog had been struck by a car the previous day. Physical examination revealed a 5.5-kg dog in good body condition. Extensive caudal abdominal contusions were present, and a 15-cm diameter raised area of subcutaneous swelling was palpable over the left thoracic wall. Thoracic radiographs revealed air-filled intestinal loops extending cranially along the left thoracic wall superficial to the rib cage, consistent with an abdominal wall hernia. Surgical exploration of the abdomen was performed the day of admission and revealed a left paracostal abdominal wall hernia, a right inguinal hernia, and a prepubic tendon avulsion. Body wall and inguinal herniorrhaphies and a routine prepubic tendon reconstruction were performed. The dog recovered uneventfully from anesthesia and was treated for 2 days with lactated Ringer’s solution (LRS; 12 mL per hour, intravenously [IV]) and oxymorphone (0.05 mg/kg body weight, IV q 2 to 4 hours).

Eight days postoperatively, progressive cellulitis developed around the caudal aspect of the abdominal incision. The dog appeared depressed and was mildly tachycardic (140 beats per minute) and tachypneic (32 breaths per minute). A complete blood count (CBC) revealed anemia (hematocrit, 28%; reference range, 39% to 57%), leukocytosis (51.9 × 10³/μL; reference range, 7.5 to 19.9 × 10³/μL), neutrophilia (41 × 10³/μL; reference range, 3.9 to 14.7 × 10³/μL) with a left shift (4.7 × 10³ bands/μL; reference range, 0 to 0 × 10³/μL), and thrombocytopenia (97 × 10³/μL; reference range, 179 to 510 × 10³/μL). Serum biochemical abnormalities consisted of hypoalbuminemia (2.7 g/dL; reference range, 3.0 to 4.5 g/dL), hypoglycemia (36 mg/dL; reference range, 60 to 120 mg/dL), hypokalemia (3.3 mEq/L; reference range, 3.8 to 5.5 mEq/L), and hyperbilirubinemia (total bilirubin, 1.7 mg/dL; reference range, 0.1 to 0.4 mg/dL). A diagnosis was made of local incisional infection with a secondary systemic inflammatory response. A 19-gauge, 20.3-cm long catheter^a was placed in the left jugular vein using aseptic technique and was advanced to the level of the cranial vena cava. Treatment was started with LRS (15 mL per hour, IV), ampicillin (30 mg/kg body weight, IV q 6 hours), and enrofloxacin (10 mg/kg body weight, IV q 24 hours).

The dog’s demeanor improved over the following 5 days. A CBC performed 13 days postoperatively showed resolution of all previous abnormalities. All serum biochemical parameters had likewise normalized, with the exception of a mild persistent hyperbilirubinemia (total bilirubin, 0.7 mg/dL). Fifteen days postoperatively, fluids were discontinued and the jugular catheter was removed. Antibiotic therapy was continued with amoxicillin-clavulanic acid (22 mg/kg body weight, per os [PO] q 12 hours). Palpable thickening of the left jugular vein was present at the time of catheter removal, consistent with thrombophlebitis or venous thrombosis.

One day following removal of the jugular catheter, the dog became tachypneic (respiratory rate, 48 breaths per minute) with an increased respiratory effort. Auscultation revealed decreased lung sounds bilaterally. Thoracic radiographs demonstrated pleural effusion. Pleurocentesis yielded 70 mL of milky, opaque fluid that did not clear upon centrifugation. The total protein and nucleated cell count of the fluid were 2.5 g/dL and 1.9 × 10³/μL, respectively. Cytopathologically, the fluid contained moderate numbers of macrophages, nondegenerate neutrophils, and lymphocytes, and a few plasma cells and reactive mesothelial cells. Fluid and serum triglyceride concentrations were 452 mg/dL and 130 mg/dL, respectively, confirming the chylous nature of the effusion. Ultrasonography of the thorax revealed an approximately 2.9-cm long thrombus within the cranial vena cava, extending into the left jugular vein [Figure 1]. Doppler color-flow imaging of the vena cava showed a nearly complete obstruction of blood flow. A thoracostomy tube was placed, LRS was reinitiated (15 mL per hour, IV through a peripheral IV line), and a continuous infusion of hetastarch (20 mL/kg body weight per day, IV) was initiated to augment plasma colloid osmotic pressure. Antibiotic therapy was continued with amoxicillin-clavulanic acid.

Three days following placement of the thoracostomy tube, pleural fluid production had increased to 100 mL/kg body weight per day. The dog’s appetite remained poor, and his body weight had decreased to 4.9 kg. The hematocrit at that time was 33%. Serum albumin had fallen to 1.2 g/dL, and the dog had developed dependent cervical and abdominal subcutaneous edema. Ultrasonography of the cranial vena cava showed no change in the size of the thrombus. Plasma antithrombin III (ATIII) activity at this time was 40% (reference range, 70% to 110% based on activity of pooled normal canine plasma). Hetastarch was discontinued, and fresh-frozen plasma (FFP) was administered (24 mL/kg body weight over 6 hours). Unfractionated (mixed molecular weight) heparin was also started (100 U/kg body weight, IV q 6 hours). The dog was anesthetized for placement of an esophagostomy tube. Following anesthetic induction, a 5-cm long catheter was placed in the right jugular vein for contrast vena caval venography. A venogram was performed by injecting 5 mL of iodinated contrast medium^b into the right jugular vein. The venogram confirmed the presence of a filling defect extending approximately 3 cm along the length of the vena cava to the border of the right atrium [Figure 2].

Due to the dog’s deteriorating condition, the decision was made to attempt caval thrombolysis using recombinant human tissue-plasminogen activator (t-PA).^c Following recovery from anesthesia, t-PA was administered at a dose of 0.4 mg/kg body weight, IV every 60 minutes for a total of four treatments. Each dose was given over 1 minute via the jugular catheter. Lactated Ringer’s solution (15 mL per hour, IV) was administered concurrently with the t-PA. Approximately 30 minutes after administration of the first dose, hemorrhage was noted at the entry sites of the jugular catheter and esophagostomy tube. The hemorrhage worsened significantly during the treatment period. By the fourth treatment, it was estimated that the dog had lost 70 to 80 mL of blood. A fifth planned treatment was withheld due to the severity of blood loss. Approximately 90 minutes after the fourth treatment, the hemorrhage resolved spontaneously. Six hours following the last treatment, the dog’s hematocrit was 23%.

Ultrasonography of the thorax performed the following day showed that the length of the thrombus had decreased to approximately 1.8 cm. A second course of t-PA was administered at a dose of 0.2 mg/kg body weight every hour for a total of five treatments. Hemorrhage recurred at the entry sites of the jugular catheter and esophagostomy tube but again resolved within 90 minutes after the last treatment. It was estimated that the dog lost an additional 40 to 50 mL of blood during this course of treatment. The dog’s hematocrit 3 hours after the last treatment was 17%. Serum albumin was 1.3 g/dL. Enteral alimentation was started at that time. The dog was fed a commercial, blenderized, low-fat diet^d supplemented with medium-chain triglycerides in hopes of minimizing intestinal production of chyle.

Ultrasonography was repeated the following day and revealed that the length of the thrombus had decreased to approximately 1.4 cm. Approximately 40% of the caval lumen was patent as assessed by color-flow imaging of the vessel in transverse orientation. Antithrombin III level at that time was 45%. Pleural fluid production was 67 mL/kg body weight per day. Total protein and nucleated cell count of the fluid were 2.0 g/dL and 2.2 × 10³/μL, respectively, at that time; cytopathologically, the composition of the fluid had not changed. Heparin was discontinued, and the dog was started on aspirin (10 mg/kg body weight, PO q 48 hours).

Pleural fluid production diminished steadily during the week following t-PA administration. One week after treatment, fluid production was 4.1 mL/kg body weight per day, and the fluid had become serosanguineous in character. At that time, the dog’s body weight was 4.4 kg and the serum albumin was 1.3 g/dL. The dog was bright and alert and eating voluntarily at that time. Ten days after treatment, the thoracostomy and esophageal tubes were removed. The dog was discharged the following day on aspirin (10 mg/kg body weight, PO q 48 hours for 30 days) and amoxicillin-clavulanic acid (22 mg/kg body weight, PO q 12 hours for 10 days). Follow-up laboratory tests were not performed; however, 2 months after discharge, the dog was physically normal and had regained its original weight.

Discussion

In the dog of this report, catheter-associated thrombosis of the cranial vena cava was thought to be the most likely cause of the chylothorax. Obstructive lesions of the cranial vena cava, including intraluminal or extraluminal masses and thrombi, have been implicated as a cause of chylothorax in dogs; however, chylothorax appears to be an uncommon sequela of caval thrombosis per se.^1–4 The mechanism for the development of chylothorax following caval obstruction is poorly understood. Increased venous pressure or obstruction of lymphatic flow at the junction of the thoracic duct and the venous system is thought to cause a secondary increase in lymphatic back-pressure within the thoracic duct. Subsequent distention and insufficiency of intrathoracic lymphatics may result in leakage of lymph into the pleural or pericardial space or both.⁵ Lymphangiectasia is a common finding in dogs with both experimentally induced and naturally occurring idiopathic chylothorax.⁶⁷ Lymphangiectasia is presumed to occur also in dogs with chylothorax due to naturally occurring obstructive lesions of the vena cava; however, lymphangiographic or histopathological studies in such cases have not been reported.

A systemic hypercoagulable state is presumed to have contributed to thrombus formation in this dog. The incisional cellulitis that occurred on day 8 of hospitalization was followed by the development of clinical signs (i.e., tachycardia, tachypnea), hematological abnormalities (i.e., neutrophilia with a left shift, thrombocytopenia), and biochemical changes (i.e., hypoalbuminemia, hypoglycemia, hyperbilirubinemia) consistent with a systemic inflammatory response. Activation of the coagulation process is a natural component of the inflammatory response to local or generalized infection.⁸ Mechanisms by which an acute systemic inflammatory response can lead to a procoagulative state include arteriolar vasodilation and blood stasis, upregulation of tissue factor by mononuclear cells, neutrophil-mediated endothelial damage, cytokine-induced alterations in endothelial expression of thrombomodulin and vonWillebrand factor, platelet activation, and suppression of fibrinolysis.⁸ Prior to the development of the caval thrombus, clinical signs of thrombosis were not detected in this dog; however, hypercoagulability may have facilitated localized thrombus development following catheter-induced endothelial injury.

In humans, central venous thrombosis is commonly treated with systemic anticoagulation and catheter removal.⁹ Thrombolytic agents are occasionally indicated in patients with symptomatic thrombi.¹⁰ There are few reports in the veterinary literature describing use of thrombolytic agents for treatment of either venous or arterial thrombi.^11–13 To the authors’ knowledge, this is the first report of the use of t-PA for catheter-associated central venous thrombolysis in a veterinary patient.

For a given dose of t-PA, multiple bolus dosing results in more effective thrombolysis as compared with single-bolus or infusion dosing.¹⁴¹⁵ It is thought that the high affinity of t-PA for fibrin leads to rapid saturation of binding sites in the outer regions of a clot following administration. Additional binding sites are exposed as lysis of the clot progresses. Doses of t-PA that result in plasma concentrations greater than what is required for clot saturation may “waste” a fraction of the administered dose as nonfibrin-bound t-PA is rapidly cleared from the plasma.¹⁴¹⁶ In this dog, a smaller decrease in the size of the thrombus following dosing with 0.2 mg/kg boluses as compared with 0.4 mg/kg boluses was observed. Whether this represents a dose response is not clear. Many factors affect the efficacy of thrombolytic therapy, including the age and composition of the clot, plasma levels of coagulation factors, factor inhibitors and platelets, and rates of rethrombosis.¹⁰

The dose of t-PA used in the dog of this report is lower than previously described for veterinary patients and was based on experimental studies of venous thrombolysis in dogs.¹⁴ Lower doses of t-PA may be sufficient for lysis of focal venous thrombi since the drug can often be easily delivered directly into an affected vessel in the vicinity of a thrombus.⁹ By avoiding the dilutional effect of administration into the systemic circulation, rapid clot saturation and maximal thrombolytic activity can be achieved with lower doses than would be required for arterial thrombolysis.⁹¹⁷

Hemorrhage continued for approximately 90 minutes following administration of t-PA and was a serious and dose-limiting complication of therapy in this dog. A similar delay in resolution of bleeding has been observed in humans.¹⁸ This is attributed to persistent fibrinolytic activity at sites of recent vascular injury despite rapid plasma clearance of unbound drug.¹⁵ In humans, the incidence of bleeding complications requiring transfusion following t-PA therapy for myocardial infarction is reported to be ≤1%.¹⁵ Analogous figures have not been reported for treatment of central venous thrombi. The hematocrit in this dog decreased by approximately 50% during treatment as a result of hemorrhage from the entry sites of the esophageal tube and jugular catheter. Transfusion support would undoubtedly have been necessary had additional t-PA been administered.

Since conditions favorable for development of a thrombus persist following successful vessel recanalization, adjunctive systemic anticoagulation is indicated to prevent rethrombosis.¹⁹ Heparin is routinely administered to human patients receiving thrombolytic therapy; however, the optimum dose, route of administration, timing with respect to thrombolytic therapy, and duration of treatment remain subjects of ongoing investigation.²⁰ Controlled studies of the efficacy of heparin as an adjunct to lysis of venous thrombi have not been reported. However, large studies of human patients with coronary arterial thrombosis have generated mixed results with respect to the ability of heparin to prevent rethrombosis.¹⁹ The efficacy of adjunctive heparin may depend on the composition of the thrombus. For example, heparin has been shown to inhibit t-PA-mediated lysis of platelet-rich thrombi in vitro.²¹ Heparin may interfere with the binding of plasminogen to the surface of activated platelets, thereby limiting the efficacy of fibrinolytic agents for treatment of platelet-rich thrombi.²¹ Heparin was used in this dog to potentiate inhibition of thrombin activity. However, the ATIII deficiency present in this dog is likely to have blunted the anticoagulant effect of heparin; thus, its efficacy and potential contribution to the dog’s bleeding complications are unknown. Serial measurement of the dog’s activated partial thromboplastin time would have been desirable but was not possible due to the small size of the dog and the paucity of venipuncture sites from which adequate quantities of blood for coagulation testing could be repeatedly obtained.

Platelet activation and adhesion to areas of residual thrombus or endothelial injury are thought to be primary events in vascular reocclusion following thrombolysis.¹⁹²² Therefore, antiplatelet therapy is an important component of adjunctive anticoagulation following successful vessel recanalization.²² Platelets have been shown both to inhibit and accentuate the activity of fibrinolytic drugs.^23–25 Activated platelets may inhibit fibrinolysis through release of plasminogen-activator inhibitors (PAI-1, α-2 antiplasmin) as well as factor XIII.²³²⁴ However, activation also leads to increased platelet expression of endogenous t-PA and upregulation and secretion of the plasminogen-binding protein, thrombospondin.²⁵ Thus, plasmin is able to be concentrated at high levels within the platelet’s local environment. In a reciprocal manner, thrombolytic drugs can either enhance or suppress the activity of platelets.²⁶ Platelet aggregability may increase immediately following administration of t-PA.^26–28 Thrombin is a potent activator of platelets; thus, the increase in its activity following administration of fibrinolytic agents may promote platelet aggregation.²⁹ However, platelet aggregation may also be inhibited by fibrinolytic drugs.²⁷²⁸³⁰ Fibrin degradation products may inhibit platelet aggregation by interfering with the binding of fibrinogen to the platelet fibrinogen receptor, GPIIb/IIIa.²⁸³⁰ Plasmin may also cause proteolytic alterations in vonWillebrand factor, leading to decreased platelet adherence to sites of endothelial injury.³⁰ The balance of the interaction between platelets, the vascular system, and fibrinolytic drugs is likely to vary depending on many factors, including the nature and extent of the vascular injury and the current status of the patient’s hemostatic system. Aspirin is commonly used for its antiplatelet effects in conjunction with fibrinolytic drugs, and large studies of human patients have shown aspirin to be unequivocally effective at reducing rates of rethrombosis following fibrinolytic therapy for myocardial infarction.³¹ The efficacy of adjunctive aspirin following venous thrombolysis is less clear, and its impact on the dog of this report is not known.

Conclusion

Despite complications, t-PA appeared effective in this dog at hastening resolution of chylothorax secondary to cranial vena caval thrombosis. While indications for the use of t-PA in veterinary patients are not clearly established, the authors suggest that its use may be warranted in select individuals with clinically significant central venous thrombi.

Acknowledgment

The authors thank Dr. Thomas Schermerhorn for his review of the manuscript.