Serial Assessment of Magnetically Levitated Centrifugal-Flow LVAD Pump Position and Inflow Angle on Chest Radiograph and Effects on Adverse Events

Abstract

We analyzed the position of a fully magnetically levitated centrifugal-flow left ventricular assist device (LVAD) on serial radiographs to assess positional change and possible correlation with adverse events. We retrospectively analyzed 59 LVAD recipients who had serial chest radiographs at 1 month, 6 months, and 12 months after implantation between November 2014 and June 2018. We measured pump angle, pump-spine distance, and pump-diaphragm depth (figure) and investigated their relationship to a composite outcome of heart failure readmission, low flow alarms, or inflow/outflow occlusion requiring surgical repositioning through recurrent event survival modeling. There were 26 patients who experienced the composite outcome; in these patients, the median 1 month pump angle was 66.2 degrees (IQR 54.5-78.0) as compared to 59.0 degrees (IQR 47.0-65.0) in the 33 patients who did not have adverse events (p = 0.04). Pump depth and pump-spine distance at 1 month were not associated with the composite outcome. Between 1 month and 6 months, the absolute pump-spine distance changed by 10.00 mm (p<0.01) and the absolute pump-diaphragm depth changed by 18.80 mm (p<0.01). These parameters did not change significantly between 6-12 months post-implantation. Pump angle did not change significantly over any time period. Change in pump depth between 1-6 months (HR = 1.019; 95%CI = 1.000-1.039; p = 0.05) and between 6-12 months (HR = 1.020; 95%CI = 1.000-1.040; p = 0.05) were weakly associated with the composite outcome. Larger pump angles are associated with the composite outcome of heart failure readmission, low flow alarm, or inflow/outflow obstruction. At 6 months, the absolute values of pump-spine distance change and pump-diaphragm depth change are significant. Movement in pump depth is weakly associated with the composite outcome.