Case Report

A 72-year-old man with a tachycardiomyopathy, atrium flutter, and atrial fibrillation underwent an elective radiofrequency ablation procedure. In addition to 5 mg of apixaban twice daily, his regular medication consisted of 2.5 mg of bisoprolol, 100 mg of amiodarone, 1 mg of bumetanide, 25 mg of spironolactone, and 40 mg of omeprazole. He had a normal renal function and no other relevant comorbidities. During the radiofrequency ablation procedure, the patient developed cardiac tamponade. Initially, a percutaneous pericardial drain was placed, with 500 mL of bloody pericardial fluid removed. However, transthoracic echocardiography still showed an accumulation of bloody fluid in the pericardium. Since the patient had taken his last dose of apixaban early on the day of the intervention and an apixaban plasma level of 108 ng/mL (30-353 ng/mL) was estimated, a reversal therapy with low-dose andexanet alfa (400 mg bolus followed by 4 mg/minute infusion for two hours) was initiated. Due to further hemodynamic deterioration, the decision was made to place a surgical drain. After median sternotomy and pericardiotomy, an inspection of the heart revealed a left ventricular free wall rupture. The cardiac surgeon decided to close the defect under cardiopulmonary bypass. According to local procedures, an intravenous bolus of 25,000 IE UFH was administered to achieve an activated clotting time (ACT) between 400 and 480 seconds before initiation of CPB. In addition, a standard dose of 2,000 mg of tranexamic acid was given. The first measured ACT (I-STAT Alinity, Abbott, Princeton, NJ) was 214 seconds (reference range, 80-160 seconds). An additional 25,000 IE of UFH were given, with minimal impact on further ACT controls (263 seconds). Over the next 20 minutes, a total of an additional 30,000 IE of UFH were administered without a significant effect on the ACT. CPB was initiated despite an ACT of 258 seconds due to hemodynamic instability. As surgery went on, thrombi were found in the operating field and in the CPB circuit as a sign of inadequate anticoagulation. Additionally, next to ACT, ROTEM was used to verify the heparin effect. A clotting time of 773 seconds (INTEM) versus 237 seconds (HEPTEM) indicated a sufficient dose of UFH. It was supposed that the observed heparin resistance had to be linked to the inhibitory effect of andexanet alfa on the antithrombin–heparin complexes. Consequently, the continuous infusion of andexanet alfa was terminated immediately, and a dose of 1,000 IU of antithrombin (Antithrombin III, Takeda Pharmaceutical Company) was administered. Subsequently, the following ACT measurements were out of range (>999 seconds), and no additional heparin was applied. A timeline of the operative course of the ACT is outlined in Figure 1. Separation from CPB was attempted after 143 minutes on bypass and 68 minutes of aortic cross-clamp time. Weaning was unsuccessful due to a global left ventricular hypokinesia and akinesia of the left lateral wall after the placement of a bovine patch. For postoperative hemodynamic support, the CPB was converted into a veno-arterial extracorporeal membrane oxygenation (VA ECMO) and the patient was admitted to the intensive care unit. The patients received one pool of platelets, 5 g of fibrinogen, and 1,250 mL of recovered blood from the surgical field via Cellsaver before intensive care unit transfer. Due to the complexity of anticoagulation in this specific patient and the use of a VA ECMO, the authors carefully administered 250 mg of protamine sulphate for the reversal of UFH at the end of the procedure, considering the dose and time-dependent elimination half-life of heparin; afterwards the ACT was 153 seconds.