Impella for Left Ventricular Unloading in Veno-Arterial Extracorporeal Membrane Oxygenation: A Meta-Analysis of Observational Studies
Presenter
Scott Mawer, The University of Nebraska Medical Center, Omaha, NE
Scott Mawer1, Arav Jhand, MBBS2, Marian Urban, MD PhD3, Cindy L. Grines, M.D., MSCAI4, Srihari S. Naidu, MD, FSCAI5, Navin K. Kapur, M.D., FSCAI6, Babar B Basir, DO, FSCAI7, Yiannis S Chatzizisis, M.D., Ph.D., FSCAI8, Michael Moulton, MD3 and Dr. Andrew Michael Goldsweig, MD, MS, FSCAI9, (1)The University of Nebraska Medical Center, Omaha, NE, (2)Mayo Clinic, Rochester, MN, (3)University of Nebraska Medical Center, Omaha, NE, (4)Northside Hospital Cardiovascular Institute, Johns Creek, GA, (5)Westchester Medical Center, New York, NY, (6)Tufts Medical Center, Boston, MA, (7)Henry Ford Health System, Detroit, MI, (8)University of Miami Miller School of Medicine, Miami, FL, (9)University of Massachusetts - Baystate Medical Center, West Hartford, CT
Keywords: Cardiogenic shock, Heart Failure and Hemodynamic support
Background:
Limited evidence supports use of Impella to unload the left ventricle (LV) in cardiogenic shock patients on veno-arterial extra-corporeal membrane oxygenation (ECMO). We conducted a meta-analysis to evaluate outcomes with Impella in these patients.
Methods:
PubMed, Scopus and Google Scholar were searched systematically for studies comparing ECMO patients treated with Impella (ECPELLA) or without (ECMO). Outcomes were all-cause mortality, stroke, bleeding, limb ischemia, hemolysis and renal replacement therapy (RRT). Risk ratios (RRs) and 95% confidence intervals (CIs) were calculated using a random effects model.
Results:
Five observational studies including 1179 patients (ECPELLA 453, ECMO 726) met inclusion criteria. Baseline demographics were similar between the 2 groups (Table 1). Outcomes were reported at 30 days (4 studies) or in-hospital (1 study). ECPELLA was associated with lower mortality compared to ECMO [RR 0.83, 95% CI 0.69-1.00, p=0.05]. However, ECPELLA was associated with a significantly higher rates of limb ischemia, hemolysis and RRT. ECPELLA was associated with a trend towards more bleeding that did not reach statistical significance. Stroke was similar in both groups.
Conclusions:
Impella for LV unloading in ECMO is associated with improved survival at the expense of increased complications.
Table 1: Patient characteristics and clinical outcomes in ECPELLA vs ECMO groups. | Patient characteristics |
| | ECPELLA (n = 453) | ECMO (n = 726) | p value |
| Age (mean ± SD), years | 56.4 ± 1.3 | 57.3 ± 3.2 | 0.62 |
| Female Sex, n (%) | 103 (22.7) | 182 (25.1) | 0.37 |
| Baseline pH (mean ± SD) | 7.26 ± 0.09 | 7.22 ± 0.12 | 0.38 |
| Baseline Lactate (mean ± SD), mmol/L | 7.8 ± 3.95 | 8.5 ± 2.19 | 0.61 |
| Timing of Impella Placement | | |
| Impella Before ECMO, n (%) | 223 (49.2) | N/A | N/A |
| Impella After ECMO, n (%) | 230 (51.1) | N/A | N/A |
| Clinical Outcomes |
| | Number of Studies Reporting Outcome | ECPELLA | ECMO | RR | 95% CI | p value |
| Mortality | 5/5 | 58.5 % | 65 % | 0.83 | 0.69 – 1.00 | 0.05 |
| Limb Ischemia | 3/5 | 20.6 % | 12.1 % | 1.63 | 1.19 – 2.25 | 0.003 |
| Hemolysis | 5/5 | 37.8 % | 19.2 % | 2.30 | 1.45 – 3.64 | 0.0004 |
| Initiation of RRT | 4/5 | 52.1 % | 31.1 % | 1.53 | 1.08 – 2.18 | 0.02 |
| Bleeding Events | 5/5 | 39.6 % | 28.8 % | 1.32 | 0.91 – 1.90 | 0.14 |
| Stroke | 3/5 | 7.9 % | 7.2 % | 1.05 | 0.65 – 1.72 | 0.84 |