Incidence and Risk Factor Analysis for Gastrointestinal Bleeding and Pump Thrombosis in Left Ventricular Assist Device Recipients

Left ventricular assist devices (LVADs) are associated with hemostatic complications. We describe the incidence and risk factors for gastrointestinal bleeding (GIB) and pump thrombosis (PT) to optimize patient selection/management. An IRB-approved retrospective review of first LVAD implants between October 1, 2011 and September 30, 2013 at a single center was conducted. Endpoints included epidemiological and risk factor analyses for GIB and PT. Descriptive statistics, chi-squared, and t-tests were used. Sixty-four patients received continuous-flow LVADs. The 12-month incidence of GIB and PT was 23.4% and 12.5%. Time to first GIB was 72.6 days (9-160). The 1-, 3-, and 6-month rate of PT was 1.6%, 6.25%, and 12.5%, respectively. All PT required pump exchange. Females (50% vs. 16%, P = 0.026) and patients without antiplatelet therapy (12.5% vs. 50%, P = 0.046) were at increased risk of PT. No pre-implant comorbidities were associated with PT. Infection was not identified as a risk factor in our cohort (25% vs. 51.8%, P = 0.156). Mean INR preceding event was not different from nonevent patients (2.1 vs. 2.24, P = 0.24). Regarding biomarkers preceding event, elevated plasma free hemoglobin (pfHg) did not reach significance (75% vs. 58%, P = 0.383) while lactate dehydrogenase was elevated significantly (744 vs. 298, P < 0.001). Receiver operating characteristic (ROC) analysis demonstrated that an LDH of >500 was highly sensitive and specific for PT. No pre-implant factors were associated with GIB. Post-implant risk factors for GIB included infection (80% vs. 38.8%, P = 0.005) and infrequent elevations in pfHg (13.3% vs. 63.3%, P < 0.001). Increased pump speed as a GIB risk factor was confirmed (HeartMate II 9560 rpm vs. 9490 rpm, P < 0.001; HeartWare 2949 rpm vs. 2710 rpm, P < 0.001). Anticoagulation/antiplatelet therapy did not affect GIB: mean INR preceding event was not different from nonevent patients (2.21 vs. 2.27, P = 0.67) and antiplatelet use was not different (46.7% vs. 46.9%, P = 0.985). LVADs are associated with early hemostatic-related morbidity. Few pre-implantation risk factors were elucidated; however, post-implantation factors including antiplatelet therapy, infection, and pump speed were identified.