. 2022 Jul 7;11(14):3955. doi: 10.3390/jcm11143955

Table 1.

Qualitative analysis and characteristics of included studies.

Author (Year)	Study Design	Country (Time)	Population; Age in Years; % Male	Type of Impella	Ascertainment of Exposure and Outcome	Outcomes	Follow Up Period	Adjustment for Potential Confounders
Mourad (2018)	Retrospective cohort	France (January 2009–April 2015)	n = 42 (VA-ECMO 23 vs. Impella 19). VA-ECMO 54 (50–60) vs. Impella 57 (51–65). VA-ECMO 78% vs. Impella 89%.	Impella CP Impella 2.5 Impella 5.0	Medical records	Mortality, 30 days: VA-ECMO 7 (30%) vs. Impella 7 (37%). Mortality, 6 months: VA-ECMO 11 (48%) vs. Impella 8 (42%). Bleeding requiring surgery, in-hospital: VA-ECMO 2 (9%) vs. Impella 1 (5%). Thromboembolic events, in-hospital: VA-ECMO 2 (9%) vs. Impella 1 (5%). Peripheral limb ischaemia, in-hospital: VA-ECMO 3 (13%) vs. Impella 1 (5%).	6 months	Unadjusted. Stratified by device first used. Exclusion criteria: delayed MCS after AMI (>72 h) or after CS onset (>48 h), refractory cardiac arrest, mechanical AMI complication, aortic valve pathology, early surgical revascularization, other causes of shock.
Garan (2019)	Prospective cohort	United States (April 2015–March 2017)	n = 51 (VA-ECMO 20 vs. Impella 31). VA-ECMO 64 ± 9 vs. Impella 61 ± 11. VA-ECMO 80% vs. Impella 74%.	Impella CP	Prospective	Mortality, in-hospital: VA-ECMO 9 (45%) vs. Impella 14 (45%). Mortality, 12 months: VA-ECMO 12 (60%) vs. Impella 14 (45%). Peripheral limb ischaemia, in-hospital: VA-ECMO 2 (20%) vs. Impella 4 (13%).	24 months	Unadjusted. Stratified by device first used. Exclusion criteria: not stated.
Karami (2020)	Retrospective cohort	The Netherlands (2006–January 2018)	n = 128 (VA-ECMO 38 vs. Impella 90). VA-ECMO 55 ± 9 vs. Impella 60 ± 10. VA-ECMO 79% vs. Impella 73%.	Impella 5.0	Medical records	Mortality, in-hospital: VA-ECMO 19 (50%) vs. Impella 48 (53%). Mortality, 12 months: VA-ECMO 22 (61%) vs. Impella 50 (56%). Bleeding complications, in-hospital: VA-ECMO 5 (13%) vs. Impella 9 (10%). Peripheral limb ischaemia, in-hospital: VA-ECMO 2 (5%) vs. Impella 2 (2%).	12 months	Unadjusted. Exclusion criteria: MCS after revascularisation with CABG, IABP, Impella 2.5, cardiac arrest without ROSC.
Lemor (2020)	Retrospective cohort	United States (October 2015–December 2017)	n = 6290 (VA-ECMO 560 vs. Impella 5730). VA-ECMO 61 ± 12 vs. Impella 66 ± 12. VA-ECMO 83% vs. Impella 73%.	Impella CP Impella 2.5 Impella 5.0	National database, ICD-10 codes	Mortality, in-hospital: VA-ECMO (46%) vs. Impella (41%) (adjusted OR 2.34, 95% CI 1.31–4.14, p = 0.004). Propensity matched mortality, in-hospital (450 in each group): VA-ECMO 43% vs. Impella 27%, OR 2.10, 95% CI 1.12–3.95, p = 0.021).	In-hospital	Adjusted for: age, gender, race, type of AMI, comorbidities, Charlson Comorbidity Index, resuscitation status, median income, teaching status, hospital region, hospital bed size, insurance, APR-DRG Severity Classification. Exclusion criteria: Impella and ECMO, no PCI, <18 years of age, missing mortality data.
Karatolios (2021)	Retrospective cohort	Germany (September 2014–September 2019)	n = 423 (VA-ECMO 123 vs. Impella 300). VA-ECMO 61 ± 10 vs. Impella 69 ± 12. VA-ECMO 78% vs. Impella 76%. NB—83% VA-ECMO AMI, 87% Impella AMI. NB—cohort overlap with Syntila (2021).	Impella CP Impella 2.5	Medical records	Mortality, in-hospital: VA-ECMO 77 (63%) vs. Impella 157 (52%). Mortality, 6 months: VA-ECMO 79 (64%) vs. Impella 163 (54%). Propensity matched mortality, in hospital (83 in each group): VA-ECMO 51 (61%) vs. Impella 41 (49%). Propensity matched mortality, 6 months (83 in each group): VA-ECMO 51 (61%) vs. Impella 45 (54%). Bleeding requiring transfusion, in-hospital: VA-ECMO 21 (17%) vs. Impella 22 (7%). Peripheral limb ischaemia, in-hospital: VA-ECMO 21 (17%) vs. Impella 23 (8%).	6 months	Adjusted for: age, Charlson Comorbidity Index, vasoactive score, creatinine, pH, aetiology of shock, PaO2/FiO2, prior cardiopulmonary resuscitation. Exclusion criteria: refractory cardiac arrest in whom insertion of MCS took place under ongoing cardiopulmonary resuscitation.
Syntila (2021)	Retrospective cohort	Germany (May 2015–May 2020)	n = 159 (VA-ECMO 54 vs. Impella 105). VA-ECMO 62 ± 10 vs. Impella 68 ± 14. VA-ECMO NB—cohort overlap with Karatolios (2021).	Impella CP Impella 2.5	Medical records	Mortality, in-hospital: VA-ECMO 38 (71%) vs. Impella 61 (58%). Mortality, 12 months: VA-ECMO 38 (71%) vs. Impella 64 (61%). Bleeding requiring transfusion, in-hospital: VA-ECMO 19 (35%) vs. Impella 16 (15%). Peripheral limb ischaemia, in-hospital: VA-ECMO 11 (20%) vs. Impella 8 (8%). Propensity matched mortality, in hospital (40 in each group): VA-ECMO 27 (68%) vs. Impella 22 (55%). Propensity matched mortality, 12 months (40 in each group): VA-ECMO 27 (68%) vs. Impella 24 (60%).	12 months	Adjusted for: Charlson comorbidity index, vasoactive score, pH, PaO2/FiO2, lactate, first rhythm, time from collapse to ROSC. Exclusion criteria: non-AMI causes of OHCA, biventricular support.