Early thrombus in a HeartMate II™ Left Ventricular Assist Device: A Potential Cause of Hemolysis and Diagnostic Dilemma

Thromboembolic events occur in 2–3% of patients receiving the HeartMate II™ (HMII) left ventricular assist device (LVAD). Even though several surfaces of the HMII are textured and thromboresistant, long-term anticoagulation is indicated.¹ We report the case of an early postoperative distal inflow-conduit thrombus resulting in hemolysis, in a patient undergoing HMII and concomitant bioprosthetic aortic valve replacement (AVR).

Clinical Summary

A 61-year-old female with idiopathic cardiomyopathy, severe aortic stenosis, New York Heart Association III heart failure and left ventricular ejection fraction of 15–20% (disproportionate to her aortic stenosis), was admitted for syncope and worsening dyspnea. Her comorbidities included obesity, diabetes, paroxysmal atrial fibrillation, previous stroke, and stage III chronic renal insufficiency. An AVR and HMII LVAD were performed, and the expectation was that the LVAD would serve as a bridge to recovery or a bridge to transplant. The LVAD and AVR proceeded uneventfully with satisfactory unloading of the left ventricle. Although fastidious surgical hemostasis was achieved, phytonadione, aminocaproic acid, and multiple blood product transfusions were necessary through the first 12 hours post-operatively; recombinant factor VIIa was not required. The patient was extubated within 24 hours and progressed well. Systemic anticoagulation with warfarin was initiated within 72 hours (Figure 1).

Figure 1. Anticoagulation and Hemolysis timeline from HMII and AVR implantation through HMII pump replacement.

Vitamin K administration (dark arrow) on POD #1 due to suspected hemorrhage/excessive post-op bleeding, and POD #26 due to evolving hematomas around pump pocket site. Vitamin K on POD #71 administered to reverse therapeutic anticoagulation in preparation for HMII replacement. Warfarin and Aspirin (ASA) administration are identified by grayscale arrows with the INR trend (dark curve) consistent with warfarin administration. LDH activity (data points) remained significantly elevated (dashed regression trend) with the concomitant Hp level <8mg/dl throughout, taken together were suggestive of hemolytic anemia. Regression trend suggests LDH activity increased with lowered systemic anticoagulation.

The day before planned discharge (POD #18) she developed acute onset hematuria and dyspnea. Transthoracic echocardiography showed occasional opening of the aortic valve with no paravalvular leak and normal flow into the LVAD. She subsequently developed a large left pleural effusion and required reintubation. Repeated red blood cell transfusions in response to decreasing hematocrit led to the temporary cessation of warfarin. Transesophageal echocardiography with HMII rotor speed adjustments showed expected variations in left ventricular volume. Cardiac catheterization with left ventriculography confirmed what appeared to be normal, but perhaps slightly delayed flow through the LVAD. There appeared to be no inflow-conduit or outflow-graft patency limitations. She subsequently developed jaundice, hemoglobinuria, and elevated lactate dehydrogenase (LDH). Hematologic studies, including von Willebrand factor, Glucose-6-phosphate dehydrogenase, fibrinogen, haptoglobin (Hp), and D-Dimer levels suggested ongoing hemolysis. The LVAD rotor speed was reduced to attenuate any potential mechanical source. The HMII controller showed no change in pump-power or differential pressure with a pulsatility index minimally above expectations.

Due to ongoing hemolysis a decision was made to replace the HMII. Intra-operatively, a thrombus on the inflow side of the bearing cup was identified (Figure 2). The HMII pump was replaced and the previously placed inflow-conduit and outflow-graft were reattached. Systemic anticoagulation was initiated post-operatively and an INR >2.0 maintained. She had an uneventful postoperative course, was discharged home within two weeks of her second operation, and made satisfactory progress at her 3 month follow-up visit.

Figure 2. Thrombus at the distal end of the inflow-conduit as seen through the pump body of the HMII.

Note the red arrow in the lateral profile of the HMII LVAD (inset). The inlet bearing cup is identified as the location of the distal inflow-conduit thrombus. Multiple LVAD studies showed normal function despite the presence of thrombus suggesting current modalities may be inadequate to evaluate HMII LVAD pump body lesions. (Image use approved by Thoratec, Inc.)

Discussion

While thrombus formation is a known late complication of the HMII,² early postoperative distal inflow-conduit thrombus formation has not been reported, and in our patient was a unique cause of hemolysis. The most widely used anticoagulation protocol in patients receiving the HMII is that reported by the Texas Heart Institute.³ The decision to delay systemic anticoagulation until POD #3 in our patient was based on post-operative bleeding, coagulopathy, and the low thromboembolic profile of the HMII. The use of phytonadione to correct coagulopathy in the face of chronic renal insufficiency challenges anticoagulation, with the early use of intravenous phytonadione having a putative delay in therapeutic systemic anticoagulation. The unclear source of hemolysis in the face of numerous false-negative LVAD interrogations compounded the difficult diagnosis.

Reportedly von Willebrand factor-dependent platelet aggregation and related thrombocytopenia is another potential source of thrombus.⁴ Upon further review of the HMII pump, the thrombus around the inlet bearing cup appeared as a formation of tissue and denatured blood that had undergone thermal degradation due to prolonged exposure to bearing heat. The thrombus demonstrated evidence of laminated layering, indicative of growth over an undetermined amount of time, presumably in the face of adequate anticoagulation. The size of the thrombus may have caused non-laminar flow to the rotor and subsequent hemolysis.

Absolute anticoagulation requirements following the HMII are unclear. One single-center analysis found the HMII to have an extremely low thromboembolic risk, allowing for less stringent anticoagulation requirements in certain patients.⁵ Those at high risk for bleeding can sometimes be followed safely with either no or extremely low anticoagulation requirements.⁵ We have seen gastrointestinal (GI) bleeding after HMII implantation, requiring cessation of anticoagulation for two weeks while the bleeding was managed with no resultant thrombotic complications. Despite the low thromboembolic profile of the HMII, early thrombotic complications such as hemolysis can occur, posing a diagnostic challenge.