Fatal Thrombosis Associated with a Hemi-Fontan Procedure, Heparin–Protamine Reversal, and Aprotinin

The patient was an 11-month-old, 8.3-kg infant with complex single-ventricle physiology consisting of double-inlet left ventricle, tricuspid atresia, L-transposition of the great arteries, hypoplastic aortic arch, and ductal-dependent systemic circulation. The child initially underwent a Damus–Kaye–Stancil procedure, repair of a hypoplastic aortic arch, and placement of a right modified Blalock–Taussig shunt at 3 weeks of age. At 11 months of age, the patient underwent a hemi-Fontan procedure and takedown of the Blalock–Taussig shunt. Postoperatively, echocardiography showed a baffle leak from the superior vena cava to the right atrium. The patient returned to surgery for baffle revision. The patient was administered aprotinin (30,000 Kallikrein inhibitor units [KIU]/kg body weight intravenously and 30,000 KIU/kg added to the cardiopulmonary bypass circuit, similar to the protocol described by Dietrich et al.  1). Continuous infusion of aprotinin was not used. The patient was anticoagulated with 300 U/kg body weight heparin, and cardiopulmonary bypass was initiated. A trabeculation was found to be the source of the leak, and the atrial patch was repaired. During cardiopulmonary bypass, additional heparin was administered as needed to maintain a kaolin activated clotting time in excess of 480 s. Activated clotting times were obtained every 30 min and were greater than 480 s, except for one, which was 408 s, just before the termination of cardiopulmonary bypass. The patient was administered an additional 500 U heparin. A follow-up activated clotting time was not obtained because bypass was discontinued shortly thereafter. Cardiopulmonary bypass was discontinued successfully with stable systemic oxygen saturation and good ventricular function as shown by transesophageal echocardiography. Protamine was administered via  the right internal jugular vein central venous catheter. Initially, 8 mg protamine (one third of the calculated required dose of 3 mg/kg body weight) was administered. Suddenly, transesophageal echocardiography revealed multiple echo-dense structures in the heart (fig. 1). Initially, these intracardiac, echo-dense structures were worm-like in appearance, but these images were not recorded. Acutely, the patient became hypotensive, and the ECG showed ST-segment elevation. Heparin was readministered, and cardiopulmonary bypass was reinitiated. Tissue plasminogen activator was administered (0.06 mg/kg load, followed by 0.7 mg · kg⁻¹· h⁻¹for 1 h, then 0.25 mg · kg⁻¹· h⁻¹for an additional 2 h) in an attempt to reverse the systemic thrombosis. The patient was again weaned from cardiopulmonary bypass. However, because of the persistent loss of the femoral arterial waveform and the onset of systemic acidosis and anuria, we suspected abdominal aortic thrombi. Thrombectomy of the abdominal aorta was performed with return of good distal perfusion and urine output. Despite maximal inotropic support, the patient remained hypotensive and hypoxic with poor ventricular function. Extracorporeal membrane oxygenation was initiated. The initial coagulation results after the patient’s catastrophic systemic thrombosis were significant for a prothrombin time greater than 90 s, an activated partial thromboplastin time greater than 150 s, a thrombin clot time greater than 90 s, a fibrinogen of 48 mg/dl, fibrinogen D-dimers greater than 2,000 ng/ml, and an antithrombin III (AT III) concentration 16% of normal. After approximately 48 h of extracorporeal membrane oxygenation support, the child had ongoing coagulopathy and evidence of an intracranial hemorrhage. At the parents’ request, support was withdrawn. A limited autopsy had findings consistent with an intracranial hemorrhage.

We hypothesize that the etiology of the thrombi formation was multifactorial: a divided venous circulation, a relatively low flow state in the pulmonary circulation, the administration of aprotinin, heparin–protamine reversal, and possibly AT III deficiency. The patient’s hemi- Fontan anatomy resulted in a divided venous circulation.For thrombi to return to the systemic ventricle, the thrombi would have to originate in the inferior vena cava system, in the pulmonary veins, or in the heart itself. We think the most likely site of origination of the thrombi was in the pulmonary veins. Before the appearance of the thrombi, cardiac function was good, making intracardiac formation of clots unlikely. The temporal correlation between clot formation and the infusion of protamine through the right internal jugular vein coupled with the relative degree of lower flow in the pulmonary circulation could provide conditions conducive to thrombi formation.

The use of aprotinin favoring a procoagulant state must be considered. Aprotinin has many potential benefits. Aprotinin reduces blood loss after cardiac surgery in children 1–3and may ameliorate the inflammatory effects of cardiopulmonary bypass. 4In a review of patients undergoing Fontan and bidirectional cavopulmonary shunt procedures, aprotinin use was associated with a decrease in the early postoperative transpulmonary gradient. 5 

Despite the significant benefits of using aprotinin, potential complications exist. A hypercoagulable state may result from aprotinin because of its ability to inhibit plasmin, block the fibrinolytic system, inhibit the intrinsic coagulation system, and reduce thrombin formation. 6Studies have suggested an association between the use of aprotinin and an increased rate of vein graft occlusion after coronary artery surgery. 7Aprotinin has been associated with unexpected formation of thrombi on pulmonary artery catheters. 8Thrombotic episodes associated with aprotinin have occurred in children undergoing repair of congenital heart defects, as well. 2 

Reports have assessed the influence of deep hypothermia and circulatory arrest (DHCA) and aprotinin on coagulation in adults. In these studies, there is speculation that the cessation of blood flow or the reduction of blood flow combined with hypothermia led to activation of the coagulation system with adverse consequences. Sundt et al.  9compared patients undergoing aortic surgery using DHCA treated with aprotinin to those patients undergoing similar procedures not treated with aprotinin. Patients treated with aprotinin had a greater incidence of renal dysfunction, myocardial infarction, and death. Platelet-fibrin thrombi were present in multiple organs of those patients treated with aprotinin who had the morbid consequences of myocardial infarction or failure, stroke, or renal dysfunction. Sundt et al.  9proposed that, in the special situation of profound hypothermia with low flow or circulatory arrest, the use of aprotinin may predispose to disseminated intravascular coagulopathy, possibly by aprotinin’s ability to inhibit protein C. However, Dietrich and Mossinger 10reported that in more than 500 children, 60% of whom were exposed to DHCA, there were no thromboembolic complications. They concluded that DHCA does not constitute a contraindication to the use of aprotinin, at least in the pediatric population. Although the patient we describe was not exposed to DHCA, conditions in the pulmonary venous system in the current patient would be similar to those encountered systemically during DHCA.

A low plasma AT III concentration in our patient may have been a factor to favor a procoagulant state. AT III deficiency was present after thrombi formation, but a unit of fresh frozen plasma was added to the pump prime, possibly correcting a preoperative AT III deficiency, if it existed.

The temporal relation between the dosing of protamine and the formation of thrombi is remarkable because the formation of intravascular thrombi after heparin reversal is rare. A low-flow state in the pulmonary venous system, coupled with the use of aprotinin and possibly a low AT III concentration, may provide conditions that would result in the formation of thrombi after heparin–protamine reversal. Aprotinin may mask inadequate heparinization via  its ability to prolong the activated clotting time. 11,12There are reports suggesting that inadequate heparinization may result in diffuse intravascular coagulation syndromes during cardiopulmonary bypass when aprotinin is used. 13 

In summary, we present a case of a fatal thrombotic complication after cardiopulmonary bypass in an infant with complex heart disease. We hypothesize that the administration of protamine through the right internal jugular vein in a patient with hemi-Fontan anatomy may be a dangerous practice that, when coupled with a relatively low-flow state in the pulmonary venous system, may contribute to the formation of thrombosis in the pulmonary veins. The use of aprotinin and possibly a low plasma AT III concentration could further contribute to a procoagulant state. Despite the benefits observed in open-heart surgery patients after the use of aprotinin, the use of aprotinin may not be without serious complications under certain conditions.