Table 1 Characteristics of included studies

Aissaoui et al. 2011 [15]

France;

2007–2008

Prospective cohort;

single center;

n=38

CMP (47%)

FM (6%)

Post-cardiotomy shock (22%)

Post-transplantation (10%)

Other (16%)

66% flow (15 min)

33% flow or 1–1.5 L/min (15 min)

Return to 100% if unstable any level

If stable minimal flow, VTI > 10 cm, LVEF > 20–25%, circuit clamp and decannulation

(B) ABG, Lact, Creat;

(H) MAP, SBP/DBP, PAP;

(E) LVEF, LVOT VTI, TD E, Sa, Ea, E/Ea, RA/RV size

ECMO removal and no further MCS because of recurring CS over the following 30 days

20/38 = 53%

Aissaoui et al. 2012 [52]

France;

2007

Prospective cohort;

single center;

n=22

CMP (50%)

Post-cardiotomy shock (32%)

FM (4.5%)

Post-transplantation (4.5%)

Other (9%)

See Aissaoui 2011

(B) N/A;

(H) MAP, SBP, DBP, HR;

(E) LVOT VTI, LVEDV, LVEF, TD E

VVI: TD Sa, Ea, Sv, strain, strain rate

ECMO removal and no further MCS because of recurring CS over the following 30 days

11/22 = 50%

Aissaoui et al. 2017 [55]

France;

2007–2008

Prospective study;

single center;

n=33

See Aissaoui 2012

See Aissaoui 2011

(B) Creat, pH, Lact;

(H) N/A;

(E) LV/RV size, LVEF, LVOT VTI, TD E, Ea, Sa, RVEF, MR, TR

ECMO removal and no further MCS because of recurring CS over the following 30 days

16/33 = 48%

Akin et al. [45]

Netherlands;

2014–2016

Prospective cohort;

single center;

n=13

PE (38%)

Post-cardiotomy shock (23%)

CS post-AMI (15%)

Myocarditis (15%)

Intoxication (8%)

50% flow

If stable, VTI > 10 cm, LVEF > 20–25% at minimal flow, decannulation

(B) Lact, S/L circulation (TVD, PVD, PPV);

(H) MAP;

(E) LVOT VTI, LVEF, TAPSE, TDSa

Successful VA-ECMO explantation within 48 h

10/13 = 77%

Asaumi et al. [29]

Japan;

1993–2001

Retrospective cohort;

single center;

n=14

Fulminant myocarditis (100%)

LVETc improved to > 200 ms, ECMO flow rate decreased until 1.5 L/min

If stable, decannulation

(B) CK, CK-MB, WBC, CRP, AST/ALT, Creat, BUN;

(H) CI, PCWP, RAP;

(E) LVETc, LVESD, LVEDD, LVWT, FS, MR, TR

ECMO removal

10/14 = 71%

Cavarocchi et al. [7]

USA;

2011–2012

Prospective cohort;

single center;

n=21

CMP (48%)

Myocarditis (14%)

CS post-AMI (14%)

Post-cardiotomy (10%)

PE (5%)

50% flow by 0.5 L/min decrement

Volume challenge

Minimal flow: 1–1.5 L/min

Return 100% if distension any level

Decannulation if adequate Biventricular function

(B) N/A;

(H) N/A;

(E) Qualitative (hTEE): LV/RV function and size, LV FAC

ECMO removal

14/21 = 67%

Chen et al. [5]

Taiwan;

NR

Retrospective cohort;

single center;

n=57

NR

Gradual flow reduction

Return to 100% if inotrope increase above predefined dose

If stable, decannulation

(B) CK, CK-MB, Troponins, Bun, Creat, AST, RNI, CBC;

(H) N/A;

(E) N/A

Weaning from ECMO and survival beyond 48 h

38/57 = 67%

Chommeloux et al. [44]

France;

NR

Prospective cohort;

single center;

n=14

CS post-AMI (50%)

CMP (28%)

Graft failure (14%)

FM (7%)

NR

(B) Lact, S/L circulation: SVD, PSVD, PPV, MFI, HI;

(H) MAP, HR;

(E) LVEF, VTI

ECMO removal

6/14 = 43%

Colombo et al. [22]

Italy;

2013–2017

Retrospective cohort;

single center;

n=25

CPR (71%)

CS post-AMI (17%)

Myocarditis (7%)

PE (4%)

Tako-tsubo (2%)

Intoxication (2%)

First weaning trial at 48 h no additional details

(B) N/A;

(H) SV, CO;

(E) LV t-IVT, LVEF, LVEDD, MAPSE

Device removal without requirement for re-cannulation over the following 30 days

18/25 = 72%

Elena Puerto et al. [59]

Spain;

NR

Retrospective cohort;

NR;

n=87

NR

NR

(B) N/A;

(H) N/A;

(E) RV dysfunction, RV basal diameter

NR

NR

Frederiksen et al. [56]

Denmark;

NR

Cohort;

single center;

n=15

NR

Stable and VTI > 7 cm no additional details

(B) N/A;

(H) N/A;

(E) LVOT VTI, LVEF, TD S', TAPSE

ECMO weaning and being alive 24 h later without hemodynamic MCS

15/29 = 52%

Fried et al. [37]

USA;

2008–2018

Retrospective cohort;

Single center;

n=126

CS post-AMI (100%)

Daily flow reduction to 1L/min once on low-dose inotrope. Hemodynamic and echocardiographic follow up to decide decannulation

(B) Lactates, Create, peak CK;

(H) MAP;

(E) LVEF

Ventricular recovery defined as survival to discharge without durable LVAD or HT

39/126 = 31%

Gambaro et al. [57]

Italy;

NR

Prospective cohort;

single center;

n=14

NR

NR

(B) N/A;

(H) HR, MAP, CO, SV;

(E) CSt/LS (LV, by STE) LVEF, LVOT VTI

ECMO weaning without adverse outcome within 1 year (MCS, transplant, CV death)

N/A

Gonzalez Martin et al. [27]

Spain;

2013–2020

Cohort;

Single center;

n=85

CS (47%)

ECPR (9%)

Electrical storm (9%)

Post-cardiotomy CS (33%)

Other (1%)

NR

(B) N/A;

(H) N/A;

(E) LVEDD, LVEF, LVOT VTO, RV basal diameter, RV qualitative function, 1:1 aortic valve aperture

Survival > 24 h after explant and no mortality from cardiogenic shock/heart failure or cardiac arrest during admission

52/85 = 61%

Hsu et al. [39]

Taiwan;

NR

Cohort;

single center;

n=133

NR

Flow reduction trial (< 1.5L/min)

If tolerated, ECMO removal

(B) ABG, Lact, Bic;

(H) SBP/DBP, CVP, SVO₂;

(E) LVEF

ECMO removal and survival to discharge

73/133 = 55%

Huang et al. [61]

Taiwan;

2014–2015

Retrospective cohort;

single center;

n=46

NR

Weaning trial when stable

Flow reduction to 0.5 L/min (5 min)

If tolerated, circuit clamp and decannulation

(B) N/A;

(H) HR, CVP, SV (LV/RV);

(E) LVEF, CSt/LS (LV), LV size, MR, RVEF, RV FAC, GLS (RV), RV size, TAPSE, TR

ECMO removal and no mortality and/or MCS because of recurring CD over the following 48 h

28/46 = 61%

Joseph et al. [49]

USA;

NR

Retrospective cohort;

single center;

n=30

NR

NR

(B) N/A;

(H) RA/PCWP, TPG, PAPi;

(E) LVEF, LVEDD, FS

NR

NR

Kim et al. 2021 (JASE) [58]

South Korea;

2016–2018

Prospective cohort;

multicenter;

n=92

CS post-AMI (48%)

Ischemic cardiomyopathy

30–50% flow (15 min)

If unstable back to previous flow

(B) N/A;

(H) N/A;

(E) LVEF, Mitral E/A, Mitral TDI (S' e' a'), LVOT VTI, RVFAC, TAPSE, Tricuspid TDI (S')

ECMO removal and not requiring further MCS over the following 30 days

64/92 = 70%

Kim et al. 2021 (JACC-imaging) [60]

South Korea;

2016–2019

Prospective cohort;

single center;

n=79

Post-MI CMP (52%)

Idiopathic dilated CMP (18%)

Fulminant myocarditis (4%)

Stress-induced CMP (4%)

If HD stable with low/no vasopressor support, MAP ≥ 65 mmHg, lactate < 2 mmol/L, CVP ≤ 15 mmHg, then gradual weaning

(B) N/A;

(H) N/A;

(E) Tricuspid annular S′/RVSP

RVFAC/RVSP

TAPSE/RVSP

[RV FWLS]/RVSP

Successful removal of VA-ECMO and no further mechanical circulatory support in the following 30 days

50/79 = 63%

Li et al. [16]

China;

2011–2012

Retrospective cohort;

single center;

n=123

Post-cardiotomy shock

Gradual flow reduction to 1 L/min

If stable, decannulation

(B) Lact, Lact clearance;

(H) N/A;

(E) N/A

ECMO removal and no HD deterioration within 48 h after

69/123 = 56%

Lim et al. [48]

South Korea;

2010–2018

Cohort;

NR;

n=122

NR

NR

(B) N/A;

(H) HR, MAP, PP;

(E) LVEF, LVOT VTI, TDSa

NR

72/122 = 59%

Ling et al. [63]

China;

2010

Observational study;

single center;

n=30

Post-cardiotomy shock (57%)

Myocarditis (14%)

CMP (29%)

Reduce speed to target retrograde flow of 0.5–1 L/min and Sweep gas off (1 h)

If tolerated, decannulation

(B) N/A;

(H) PCRTO;

(E) N/A

N/A

7/7 decannulated = 100%

Luyt et al. [23]

France;

2009–2010

Prospective cohort;

single center;

n=41

CS post-AMI (27%)

Myocarditis (17%)

Post-cardiotomy (15%)

Graft failure (17%)

Septic shock (10%)

CPR (7%)

Rhythm disturbance (7%)

66% flow (15 min)

33% flow or minimum of

1–1.5 L/min (15 min)

If unstable, return to 100% flow

If stable minimal flow, LVEF > 20–25%, VTI > 12 cm, Mitral systolic velocity > 6 cm

Decannulation

(B) NT-proBNP, MR-proANP, proADM, Copeptin, TNIc;

(H) N/A;

(E) N/A

ECMO removal and survival without MCS for > 30 days

18/41 = 44%

Matsumoto et al. [30]

Japan;

1995–2014

Retrospective cohort;

single center;

n=37

Myocarditis (100%)

Weaning trial when LVETc > 200 ms

Gradual flow reduction to 1.5 L/min

If stable, decannulation

(B) CK, CK-MB, ABG, Lact, Bun, Creat, Bili;

(H) HR, MAP;

(E) LVEF, LV size, LVPWT

ECMO removal

22/37 = 59%

Mazet et al. [24]

France;

2014–2016

Cohort;

single center;

n=31

CS (71%)

CPR (29%)

Gradual decrease to < 2 L/min

(60 min) If stable, decannulation

(B) Lact;

(H) N/A;

(E) LVEF

NR

NR

Mongkolpun et al. [36]

Belgium;

NR

Cohort;

NR;

n=22

CS post-AMI (64%)

Post-cardiotomy (14%)

Myocarditis (14%)

PE (8%)

Gradual flow 1 L/min

If VTI > 10 cm, decannulation

(B) Lact, SBF;

(H) MAP, CI, SVO₂;

(E) N/A

ECMO removal and HD

Stabilization without the need to increase the vasopressor dose within 24 h

12/22 = 55%

Morisawa et al. [38]

Japan;

2006–2008

Retrospective cohort;

single center;

n=29

CS post-AMI (100%)

NR

(B) BE;

(H) HR, Peak BP, SVO₂;

(E) LVEF

ECMO removal and survival for more than one month

15/29 = 52%

Mork et al. [25]

Denmark;

2017–2019

Prospective cohort;

single center;

n=38

CPR (61%)

Heart failure (5%)

PE (5%)

Post-cardiotomy shock (11%)

CS (16%)

66% flow (5 min)

33% flow (1–2 h)

If stable, decannulation

If unstable any level, back to full flow

(B) N/A;

(H) N/A;

(E) LVEF, LVOT VTI, TAPSE, Mitral S'

ECMO removal and survival without MCS for > 24 h

25/38 = 66%

Moury et al. [68]

France;

2018–2019

Prospective cohort;

single center;

n=15

Post-cardiotomy shock (60%)

AMI (40%)

Weaning from ECMO was performed if no onset of a new respiratory, neurologic, or cardiovascular failure was clinically assessed

(B) N/A;

(H) N/A;

(E) Diaphragm thickening fraction (TF)

LVEF

ECMO weaning failure was defined by the death of the patient while being treated with assistance, the need for heart transplantation, and the need for an LVAD

9/15 = 60%

Naruke et al. 2010 [50]

Japan;

1996–2008

Retrospective cohort;

single center;

n=25

Myocarditis (52%)

CS post-AMI (36%)

ACHF (12%)

Gradual flow reduction to 1.0 L/min

If stable, decannulation

(B) CK, BNP, Creat, CRP;

(H) HR, MAP, PAP, PCWP, CVP, CI, ETCO₂;

(E) LVET, LVEF

ECMO weaning

18/25 = 72%

Naruke et al. 2012 [42]

Japan;

NR

Cohort;

NR;

n=30

NR

NR

(B) N/A;

(H) SVO₂, ETCO₂;

(E) N/A

VA-ECMO weaned off without severely deteriorated cardiac output indicated by ETCO2 < 10 mmHg or LVET < 100 ms

19/30 = 63%

North et al. 2018 [46]

USA;

2012–2017

Retrospective cohort;

single center;

n=60

NR

Gradual flow reduction to 0.5–1.5 L/min (10 min). If stable according to precise criteria’s, decannulation

(B) N/A;

(H) MAP, CI, CVP, PA Systolic pressure, PA saturation;

(E) LVEF

Successful wean was defined by the following parameters: MAP > 60 mmHg; cardiac index > 2.2 L/min; CVP ≤ 16 mmHg; and EF ≥ 20% on low doses of inotropes or/and pressors followed by decannulation

42/60 = 70%

North et al. 2022 [31]

USA;

2012–2019

Retrospective cohort;

single center;

n=62

CS post-AMI (100%)

Gradual flow reduction by 0.5–1 L decrement (1–2 min), with echocardiographic evaluation, until 0–0.5 L/min. Decannulation if MAP > 60 mmHg, cardiac index > 2.0, CVP⩽16 mmHg, and LVEF ⩾20% on low-dose inotrope

(B) Troponin I, Creat, CK, AST, ALT, lact;

(H) N/A;

(E) LVEF

ECMO removal without further mechanical circulatory support defined a successful weaning from ECMO

45/62 = 73%

Omar et al. [33]

USA;

2014–2018

Retrospective cohort;

Single center;

n=238

Arrhythmia (37%)

MI (24%)

HF (45%)

PE (33%)

Post-cardiotomy (25%)

NR

(B) Lact (baseline, 1,3,5, 10 days);

(H) N/A;

(E) N/A

NR

98/238 = 41%

Oshima et al. [35]

Japan;

1997–2004

Retrospective cohort;

single center;

n=32

Post-cardiotomy (47%)

PE (13%)

CS post-AMI (9%)

Myocarditis (9%)

CMP (3%)

Gradual flow reduction to 1.5–2 L/min

If stable, decannulation

(B) Lact;

(H) N/A;

(E) N/A

ECMO removal and discharged from the ICU

12/32 = 38%

Ouazani et al. [54]

USA;

NR

Prospective cohort;

single center;

n=12

NR

Removal considered when LVEF > 25% and VTI > 10 cm

If unstable, weaning trial stopped

(B) N/A;

(H) N/A;

(E) LVEF, LVOT VTI, TD SaL, SaS, EaL, EaS, LS (LV)

ECMO removal without requiring any further MCS

9/12 = 75%

Pappalardo et al. [9]

Italy;

2008–2013

Observational study;

single center;

n=42

ECPR (29%)

Post-cardiotomy (24%)

CS post-AMI (14%)

Arrhythmia (13%)

PE (4%)

Trauma (2%)

Weaning by 0.5 L/min decrement every 6–24 h to 2 L/min

If stable, decannulation

(B) BNP, Bili, Creat, CRP;

(H) MAP,SBP/DBP, HR, CI, PSP, PDP, PCWP, CVP, SVO₂;

(E) LVEF, LVEDD, LVOT VTI, TR, TD S tricuspid annulus, TAPSE, RVEDD

ECMO removal

49/129 = 38%

Park et al. [17]

South Korea;

2009–2011

Retrospective cohort;

single center;

n=69

AMI (31.9%)

Respiratory failure (18.8%)

Sepsis (15.9%)

PE (4.3%)

Trauma (2.9%)

Gradual flow reduction to 1 L/min/m²

If stable, decannulation

(B) ABG, Creat, Hb;

(H) SBP, MAP, mean PP;

(E) N/A

Survival for 48 h after weaning with mean systolic blood pressure > 90 mmHg

27/69 = 39%

Sawada et al. [40]

Japan;

2013–2017

Retrospective cohort;

single center;

n=50

CS post-AMI (54%)

FM (24%)

CMP (10%)

other heart disease (12%)

Weaning trial when stable

Flow reduction to 1.5–2 L/min

Then 0.5–1 L/min

If unstable, return to full flow

If stable, decannulation

(B) pH, Bic, Lact, Bili, AST,ALT, Creat;

(H) PAP, PADP, PCWP, RAP, SVO₂;

(E) LVETc, LVOT VTI, FS, LVEDD/SD

ECMO removal and survival beyond 30 days without needs for further MCS

24/50 = 48%

Sawamura et al. [43]

Japan;

2000–2016

Retrospective cohort;

multicenter;

n=99

Myocarditis (100%)

NR

(B) CK, BUN, Creat, AST/ALT, LDH, Bili;

(H) N/A;

(E) LVEF, LVEDD

VA-ECMO decannulation and subsequent discharge

46/99 = 46%

Sugiura et al. [32]

Japan;

2012–2016

Retrospective cohort;

multicenter;

n=55

CS post-AMI (100%)

Weaning trial when stable

Flow reduction to 0.5–1.5 L/min

If stable, decannulation

(B) Lact, Creat, Bili;

(H) SBP, MAP;

(E) LVEF

CE-CT LV wall enhancement

ECMO removal

28/55 = 51%

Suhr et al. [34]

Germany;

2006–2017

Retrospective cohort;

single center;

n=258

NR

NR

(B) Lactate 1, 6, 12, 24 and 36 h;

(H) N/A;

(E) N/A

NR

136/258 = 53%

Vuthoori et al./Heaney et al. [51]

USA;

NR

Prospective cohort;

single center;

n=34

NR

Weaning trial with close monitoring

Flow reduction by 1 L/min decrements

If stable, circuit clamped, decannulation

(B) macrophage migration inhibitory factor;

(H) CI;

(E) LVEF, LV size

ECMO removal and free from pharmacologic and MCS at 30 days post-explant

8/34 = 24%

Wu et al. [47]

Taiwan;

2003–2008

Retrospective cohort;

single center;

n=72

Post-cardiotomy shock (100%)

Weaning trial when stable

Gradual flow reduction to 1 L/min. If stable, decannulation

(B) Creat;

(H) MAP, SVO2, MPAP, MAP:MPAP ratio;

(E) N/A

ECMO removal

41/72 = 57%

Xu et al. [28]

China;

2019–2021

Retrospective cohort;

single center;

n=20

Myocarditis (27%)

CMP (23%)

Ischemic heart disease (21%)

ECPR (19%)

Other (10%)

Weaning trial when HD stable, signs of cardiac and pulmonary recovery. Flow reduction to 1.5 L/min, If HD stable with low-dose inotrope, PCRTO begins, target—0.5–1 L/min for 30 min with arterial gas, HD and respiratory monitoring. Decannulation if successful

(B) N/A;

(H) HR, CVP, PAWP, MAP;

(E) LVEDV, LVEF, LVOT VTI, Mitral TD e′ E and lat s′, GLS

Patients who survived for 48 h after withdrawal and did not require ECMO assistance

13/20 = 65%

Yi et al. [41]

China;

2018–2020

Retrospective cohort;

single center;

n=24

NR

NR

(B) Create, Lact;

(H) MAP;

(E) LVEF, LVOT VTI, Mitral TD lat s′, GLS

NR

16/24 = 67%

Yoshida et al. [26]

Japan;

2002–2003

Cohort;

single center;

n=15

CS post-AMI (80%)

CPR (20%)

NR

(B) ABG, Lact;

(H) HR, MAP, CVP, MPAP, PCWP, SVO₂, ETCO₂;

(E) N/A

ECMO removal

6/15 = 40%