Longitudinal changes in hemostatic parameters and reduced pulsatility contribute to non-surgical bleeding in patients with centrifugal continuous-flow left ventricular assist devices

Bleeding and thromboembolic events are identified complications in patients supported with newer centrifugal continuous-flow left ventricular assist devices (cfLVADs). Bleeding events have been associated with acquired von Willebrand syndrome (vWS) in these patients, though longitudinal changes and the effect of pulsatility remain unquantified. We evaluated longitudinal effects of third-generation cfLVADs on hemostatic biomarkers, non-surgical bleeding, and thromboembolic events. We investigated the association between pulsatility (as defined by aortic valve opening) on von Willebrand Factor (VWF) profile and bleeding. We prospectively studied 28 patients implanted with the HeartWare (HeartWare International, Framingham, MA) cfLVAD for up to 360 days. We performed bleeding and coagulation assays 8 times from pre-implant to Day 360 (D360) post-implant, including platelet aggregometry, VWF collagen binding activity-to-antigen (CBA/Ag) ratio, thromboelastography, soluble P-selectin, platelet-specific marker soluble glycoprotein VI (sGPVI), and platelet microparticles. Aortic valve opening was assessed by echocardiography at each assessment. Bleeding and thromboembolic events were documented. Bleeding events occurred in 14 patients (50%). Maximal platelet inhibition occurred by D30. VWF profile impairment (VWF CBA/Ag < 0.8) was demonstrated in 89% of patients at D30, with subsequent recovery but further deterioration after D180. Bleeding was associated with elevated pre-implant sGPVI (p = 0.008). Pulsatility was associated with higher VWF CBA/Ag (p = 0.02) and a trend to less bleeding. Third-generation cfLVADs were associated with longitudinal changes in hemostatic markers, and bleeding was associated with elevated pre-implant plasma sGPVI. Further, pulsatility may contribute to recovery of the VWF profile and potentially lower bleeding risk.