Impella versus Non-Impella for Nonemergent High-Risk Percutaneous Coronary Intervention.

The benefit of mechanical circulatory support (MCS) with Impella (Abiomed, Inc, Danvers, MA) for patients undergoing nonemergent, high-risk percutaneous coronary intervention (HR-PCI) is unclear and currently the subject of a large randomized clinical trial (RCT), PROTECT IV. While contemporary registry data from PROTECT III demonstrated improvement of outcomes with Impella when compared with historical data (PROTECT II), there is lack of direct comparison to the HR-PCI cohort that did not receive Impella support. We retrospectively identified patients from our institution meeting PROTECT III inclusion criteria (left ventricular ejection fraction [LVEF] <35% with unprotected left main or last remaining vessel or LVEF <30% undergoing multivessel PCI), and compared this group (NonIMP) to the published outcomes data from the PROTECT III registry (IMP). Baseline differences were balanced using inverse propensity weighting (IPW). The co-primary outcome was major adverse cardiac events (MACE) in-hospital and at 90 days. Secondary outcomes included in-hospital post-PCI complications. We identified 284 high-risk patients who did not receive Impella support; 200 patients had 90-day event ascertainment and were included in IPW analysis, with 504 patients in the IMP group. After calibration, few residual differences remained; patients in the NonIMP group were older (73.4 vs. 69.3, p <0.001) with higher prevalence of coronary artery bypass grafting (65.0% vs. 13.7%, p <0.001). Unprotected left main intervention was performed in 43% of patients in both groups. The primary outcome was not different in-hospital (3.0% vs. 4.8%, p = 0.403), but lower in the NonIMP group at 90 days (7.5% vs. 13.8%, p = 0.033). Peri-procedural vascular complications, bleeding, and transfusion rates were not different between groups. However, acute kidney injury occurred more frequently in the NonIMP group (10.5% vs. 5.4%, p = 0.023). Under identical HR-PCI inclusion criteria for Impella use in PROTECT III, an institutional non-Impella supported HR-PCI cohort demonstrated similar MACE in-hospital but lower MACE at 90 days. There was no signal for peri-procedural harm with Impella use. These results do not support routine usage of Impella for HR-PCI patients. Careful patient selection is critical until a large RCT demonstrates benefits in a broad HR-PCI population.