A Case of Severe Diffuse Venous Thromboembolism Associated with Aprotinin and Hypothermic Circulatory Arrest in a Cardiac Surgical Patient with Factor V Leiden

A 61-yr-old man presented for aortic valve replacement for severe calcific aortic stenosis with mild aortic insufficiency. He had undergone myocardial revascularization 10 yr previously and had bilateral patent internal mammary artery grafts. The surgical plan was aortic valve replacement through a limited sternotomy using cardiopulmonary bypass. Immediately preoperatively, his vital signs were normal; his coagulation, hematology, and biochemical profiles were all normal; and he was taking the following medications: ramipril, aspirin, atenolol, hydrochlorothiazide, and amlodipine. He had a history of minor allergic reactions to penicillin and sulfa drugs.

Anesthesia and neuromuscular blockade were achieved using 100% oxygen, etomidate, midazolam, sufentanil, and vecuronium and maintained using inhalational isoflurane and infusions of midazolam, sufentanil, and vecuronium. Clindamycin, 600 mg, was administered for antibiotic prophylaxis. Monitoring included standard anesthesia monitors, a radial intraarterial line, a pulmonary artery catheter, and a transesophageal echocardiography probe. The transesophageal echocardiography examination confirmed the severe nature of the aortic stenosis with a peak instantaneous transvalvular gradient of 102 mmHg and a calculated aortic valve area of 0.47 cm². The examination results were also notable for diffuse atherosclerosis of the ascending aorta, transverse arch, and descending thoracic aorta. Before skin incision, a test dose of 10,000 kallikrein inhibiting units aprotinin was well tolerated; a 2,000,000-kallikrein inhibiting unit loading dose was infused, followed by an infusion of 500,000 kallikrein inhibiting units per hour. Aprotinin, 2,000,000 kallikrein inhibiting units, was also added to the cardiopulmonary bypass priming solution (full-dose Hammersmith regimen).4 

Kaolin-activated clotting time was maintained at greater than 500 s with bovine lung heparin, and heparin concentration was maintained at greater than 2.7 U/ml. The initial dose of heparin was calculated using an in vitro  heparin dose–response curve generated using the patient's own blood. Kaolin-activated clotting time testing was performed at least every 30 min thereafter.

Because of the high risk of disrupting the bilateral internal mammary arterial grafts with sternotomy, femoral arterial and venous cannulation was performed, and cardiopulmonary bypass was instituted before limited median sternotomy. The limited median sternotomy and the initial mediastinal dissection were conducted without difficulty during a 37-min period of normothermic cardiopulmonary bypass. The patient was weaned from cardiopulmonary bypass for several minutes for relocation of the arterial cannula to a region of the ascending aorta with minimal atherosclerosis under epiaortic ultrasonic guidance. Cardiopulmonary bypass was reinstituted, the aortic cross clamp was placed with epiaortic ultrasonic guidance, cardioplegia was administered, and aortic valve replacement was performed using a Hewlett-Packard St. Jude (St. Paul, MN) prosthesis. At the completion of the valve replacement, the aortic cross clamp was removed, the endothelial surface of the aorta was inspected, and the aortotomy was closed during a 13-min interval of deep hypothermic circulatory arrest (esophageal temperature 19.2°C.) After aortic valve prosthesis insertion, rewarming and resuscitation of the heart were remarkable for ventricular arrhythmias controlled with amiodarone and ventricular pacing.

The patient was weaned from cardiopulmonary bypass using moderate doses of epinephrine and norepinephrine by continuous infusion. A vasopressin infusion was added. Moderate pulmonary hypertension was treated with inhaled nitric oxide, 40 parts per million. The venous cannula was removed from the right femoral vein, and a slow protamine infusion was begun. Hypotension and worsening pulmonary hypertension required discontinuation of the protamine infusion after administration of 120 mg—approximately 50% of the calculated dose. Large reticular thrombi that nearly filled the inferior vena cava and extended into the right atrium, right ventricle, and pulmonary arteries were noted on transesophageal echocardiography. The patient was reheparinized, aprotinin was discontinued, and cardiopulmonary bypass was reinstituted.

The thrombosis extended inferiorly into the left femoral vein and was evacuated upon replacement of the femoral venous cannula. Using moderately hypothermic perfusion, the right atrium and pulmonary artery were opened, and massive quantities of friable thromboembolic material were removed. After rewarming, the patient did not wean from cardiopulmonary bypass due to severe biventricular failure. He could not be adequately oxygenated despite a fractional inspired oxygen concentration of 1.0, high levels of positive end-expiratory pressure, and inhaled nitric oxide therapy, presumably because of distal pulmonary embolization of the friable thromboembolic material. Extracorporeal membrane oxygenation was instituted, but severe hemorrhage from the aortic cannulation site was present. The ascending aorta was replaced using a Hemashield (Medi-Tech, Natick, MA) graft during a second period of hypothermic circulatory arrest.

Massive bleeding ensued intraoperatively. The patient developed a consumptive coagulopathy that was diagnosed by persistent increase of the Kaolin-activated clotting time to greater than 1,000 s, despite heparin concentrations less than 2.0 U/ml in the face of continued heparin administration. The patient received further extracorporeal membrane oxygenation support and massive transfusion of packed cells, platelets, fresh frozen plasma, and cryoprecipitate. He died in the operating room.

The family underwent counseling and testing for hypercoagulable states. The patient's two sons both tested positive as heterozygotes for the Leiden mutation of factor V (FV:R506Q), and their biologic mother (the patient's wife) tested negative for this mutation. We conclude that the patient was a carrier of the factor V Leiden mutation.

Reoperative cardiac surgery is associated with a more complex and prolonged surgical intervention that results in a higher incidence of pathologic bleeding and perioperative requirement for allogeneic transfusions. The inflammatory, coagulation, and fibrinolytic cascades are activated to a greater degree than in primary cardiac surgery due to incision through the thromboplastin-rich mediastinal scar, extensive tissue dissection, and prolonged cardiopulmonary bypass duration. To minimize hemorrhagic complications associated with complex surgical procedures, data support the use of prophylactic antifibrinolytic therapy.5–8The risks of hypercoagulability or overt thrombosis in association with antifibrinolytic therapy are theoretical, at best. The mechanisms of activity of these agents are not “procoagulant,” but are “antifibrinolytic,” in a population of patients who are hypocoagulable and have active fibrinolysis. However, a small subset of patients experience thrombosis after cardiopulmonary bypass despite standard care. We previously reported two cases of complete aortic thrombosis after aortic repair and deep hypothermic circulatory arrest in which ϵ-aminocaproic acid was used.3Perimortem blood samples were drawn, and one patient was found to have a factor V Leiden gene mutation (FV:R506Q). It seems reasonable to propose that patients with known predispositions to thrombotic disease should not receive prophylactic antifibrinolytic therapy, even in association with cardiopulmonary bypass. However, in patients with a heterozygous mutation or a subclinical presentation of such an illness, preoperative recognition would be impossible to detect by clinical history and examination.

In operations requiring deep hypothermic circulatory arrest, antifibrinolytic therapy is frequently used, although its efficacy in this population has not yet been demonstrated. There are early reports of renal failure and thrombosis when using antifibrinolytic therapy in deep hypothermic circulatory arrest, but many of these reports are retrospective, anecdotal, or both.9–14Aortic surgery is itself associated with a risk of renal dysfunction, and this is not increased by the concomitant use of aprotinin.15The retrospective nature of these studies cannot identify those patients with thrombosis who had presumed inadequate heparin levels during cardiopulmonary bypass because the threshold diatomaceous earth (Celite) kaolin-activated clotting time was maintained at only 480 s. Nor can retrospective reporting identify which patients may have received two different antifibrinolytic agents in combination.9The risks of thrombosis in deep hypothermic circulatory arrest have prompted some investigators to avoid the use of aprotinin. Others recommend using the drug after  deep hypothermia and circulatory arrest have been completed.16Most other clinicians continue to use the drug without a change in practice.17 

High concentrations of aprotinin inhibit many serine proteases, including kallikrein and protein C.18Factor V Leiden mutations, which occur in 3–5% of the population, yield a resistance to activated protein C, which results in impaired signaling for anticoagulation and fibrinolysis. Using a clot-based assay, in vitro  analyses evaluating the response to activated protein C in cardiac surgical patients indicate that aprotinin does induce a factor V Leiden–like defect in normal plasma. In vitro  analyses from factor V Leiden patients suggest that aprotinin further exacerbates this defect in the plasma.19Corroborating clinical data demonstrate that factor V Leiden patients have lesser amounts of mediastinal tube drainage and allogeneic transfusions.20,21A case report describes a patient with factor V Leiden who experienced thrombosis of coronary artery revascularization grafts within a month of surgery.22 

In the current case, the patient experienced a thrombotic event before the overt consumptive coagulopathy to which he succumbed was diagnosed. The inherited factor V Leiden mutation and abnormality of the protein C anticoagulant pathway predisposed this patient to hypercoagulability, which became evident when the protective effects of heparin were reversed with protamine. After the consumptive coagulopathy ensued, the patient would certainly have been at high risk for further thrombotic and bleeding complications. The thrombus in the inferior vena cava at the same anatomical site occupied by the venous cannula (from the femoral vein to the right atrium) is rare and suggests that endothelial integrity disruption provided a nidus for thrombus formation. Therefore, the possible etiologies for this thrombotic event include endothelial injury, disseminated intravascular coagulation after hypothermic cardiopulmonary bypass and deep hypothermic circulatory arrest, factor V Leiden mutation, and the use of an antifibrinolytic agent. Elimination of just one of these risk factors might have avoided this outcome.

Factor V Leiden mutation has been associated with an increased risk for venous  thrombosis, although large-scale observational studies have not demonstrated the presence of factor V Leiden as an independent risk factor for the development of arterial  atherosclerotic cardiovascular disease.23–25In population-based studies, the elderly seem to have no increased risk of arterial thrombosis due to factor V Leiden, whereas women and, specifically, obstetric patients do have significant thrombotic risks.26Although no direct link between factor V Leiden and atherosclerosis has been shown, resistance to activated protein C has been identified as a laboratory marker that is linked to an increased risk of advanced atherosclerosis.27,28 

Screening for subclinical procoagulant states such as factor V Leiden abnormality may be considered in patients undergoing deep hypothermic circulatory arrest in which the use of antifibrinolytic therapy is planned. The benefits of reduced blood loss due to pharmacologic therapy need to be carefully weighed against the cost of testing in any population.29Because no causative agents or clinical conditions have been definitively identified, the potential risk of intravascular thrombosis is not currently known.

At our institution, we have had four fatal thrombotic events in cardiothoracic surgical patients in a 3-yr span. The common factors include antifibrinolytic therapy (aprotinin or ϵ-aminocaproic acid), a period of hypothermic circulatory arrest, and documented predisposition to hypercoagulable state (factor V Leiden mutation in two cases and antiphospholipid antibody in one). Based on this experience, we have begun screening elective surgical procedures requiring hypothermic circulatory arrest for factor V Leiden mutation at a cost of $175 per patient. We plan to withhold antifibrinolytic therapy in any patient testing positive for factor V Leiden mutation despite the absence of scientific evidence of a causal relation between antifibrinolytic therapy and adverse outcomes, in patients with this genetic trait.