Study reference

Study characteristics

Patient characteristics 

Intervention (I)

Comparison / control (C)

Follow-up

Outcome measures and effect size

Comments

Danial et al

Type of study:

Retrospective cohort study

Setting and country:

The study was conducted at La Pitié Salpêtrière University Hospital in Paris, France

Funding and conflicts of interest:

Pr. Combes reports receiving grant support and lecture fees from Maquet and

Baxter and consulting fees from Hemoven. The other authors declare that they

have no conflict of interest.

Inclusion criteria:

The inclusion criteria were consecutive patients aged over 16 who received peripheral femoro femoral VA ECMO between January 2015 and December 2017 at La Pitié Salpêtrière University Hospital, Paris, France

Exclusion criteria:

Patients surgically cannulated after November 2016 (n = 85, 40 under cardiopulmonary resuscitation, 15 during the transition period and 30 for failure of percutaneous approach) were excluded from the study

N total at baseline:

Intervention (surgical): 485

Control (percutaneous): 329

After propensity score matching, both groups held 266 patients.

Important prognostic factors:

Before matching

age (yrs) ± SD:

I: 53.6 ± 14.3

C: 55.8 ± 14.5

Sex, female, n (%)

I: 145 (30)

C: 88 (27)

BMI, km/m2, mean ± SD:

I: 26.4 ± 5.6

C: 26.8 ± 5.3

After matching

age (yrs) ± SD:

I: 54.2 ± 14.4

C: 55.2 ± 14.9

Sex, female, n (%)

I: 80 (30)

C: 70 (26)

BMI, km/m2, mean ± SD:

I: 26.4 ± 5.5

C: 26.5 ± 5.0

Groups comparable at baseline?

Groups comparable after propensity score matching.

Describe intervention (treatment/procedure/test):

“Femoral vessels were cannulated using an arterial cannula (Maquet®, 23 cm, 15–19 F) inserted into the common femoral artery and a venous cannula ( Maquet®, 55 cm, 23–29 F) into the common femoral vein. An additional anterograde perfusion cannula (Terumo® 5 F

or Selidcath ® 5 F) was systematically inserted into the superficial femoral artery to prevent leg ischemia. Prophylactic antibiotic therapy was not used for peripheral cannulation (unless the patient received ECMO immediately before or after sternotomy for cardiac surgery or the patient was already on antibiotics at the time of cannulation).

For the surgical approach, femoral vessels were exposed with a limited transversal incision in the groin. Then, the Seldinger’s technique was used to insert arterial, venous and leg perfusion cannulas and the skin was surgically

sutured. At the time of decannulation, a bolus of 5000 IU of heparin was injected, the operative field was reopened, vessels controlled, cannulas removed and cannulation

sutures tied. Reconstructive procedures were performed

if needed (bleeding, arterial tear) and were considered a

vascular decannulation complication.

In the days following surgical or percutaneous decannulation, Doppler ultrasound was performed in case of clinical suspicion of false aneurysm or arteriovenous fistula.”

Describe  control (treatment/procedure/test):

“Femoral vessels were cannulated using an arterial cannula (Maquet®, 23 cm, 15–19 F) inserted into the common femoral artery and a venous cannula (Maquet®, 55 cm, 23–29 F) into the common femoral vein. An additional anterograde perfusion cannula (Terumo® 5 F or Selidcath ® 5 F) was systematically inserted into the superficial femoral artery to prevent leg ischemia. Prophylactic antibiotic therapy was not used for peripheral cannulation (unless the patient received ECMO immediately before or after sternotomy for cardiac surgery or the patient was already on antibiotics at the time of cannulation).

When the percutaneous approach was used, femoral

vessels were localized by Doppler ultrasound. Then,

femoral and leg perfusion cannulas were percutaneously inserted under echo guidance with the Seldinger’s technique,

while the ipsilateral or the contralateral femoral

vein was also  percutaneously cannulated.

In the days following surgical or percutaneous decannulation, Doppler ultrasound was performed in case of clinical suspicion of false aneurysm or arteriovenous fistula.”

Length of follow-up:

30 days

No long-term

follow-up was available.

Incomplete outcome data:

N.R.

Outcome measures and effect size (include 95%CI and p-value if available):

ECMO-associated infections

Propensity matched groups

I: 74 out of 266 (27.8%)

C: 44 out of 266 (16.5%)

IPTW weighted populations

I: 16.2%

C: 28.5%

Local infection

Total population

I: 133 (27.4%)

C:  49 (14.9%)

Limb ischemia

Propensity matched groups

I: 33 out of 266 (12.4%)

C: 23 out of 266 (8.6%)

IPTW weighted populations

I: 11.2%

C: 9.6%

Total population

I: 49 (10.1%)

C: 25 (8.5%)

30-day survival

Propensity matched groups

I: 150 out of 266 (56.3%)

C: 170 out of 266 (63.8%)

IPTW weighted populations

I: 55.4%

C: 62.9%

Total population

I: 267 (55%)

C: 190 (58%)

Saiydoun et al

Type of study:

monocentric prospective observational

registry study

Setting and country:

Paris, France

Funding and conflicts of interest:

The authors declare that they have no known competing financial

interests or personal relationships that could have appeared to influence

the work reported in this paper.

Inclusion criteria:

The inclusion criteria were adult patients admitted at Henri Mondor University Hospital and supported with peripheral V a ECLS for refractory cardiac arrest or refractory cardiogenic shock from January 2018 to March 2021

Exclusion criteria:

The exclusion criteria were patients supported by VA-ECLS following cardiac surgery

N total at baseline:

I (surgical): 59

C (percutaneous): 61

Important prognostic factors:

age ± SD:

I: 53 (41–64)

C: 60 (51–67)

Sex male:

I: 39 (66%)

C: 47 (77%)

BMI, kg/m2

I: 27 (23–31)

C: 25 (23–28)

Groups comparable at baseline?

“Patients in the percutaneous

group were older (60 [51–67], vs. 53 [41–64], P = 0.007) with lower LVEF (15 [11–20] vs 26 [15–45], P < 0.0001) and a higher rate

of diabetes mellitus (34% vs. 15%, P = 0.015). They also presented more frequently with STEMI, while patients in the surgical groups

had more frequently a previous history of chronic heart failure. There

was a higher prevalence of long-term anticoagulation in the surgical

group. No significant difference in the rate of cardiac arrest and

refractory cardiac arrest was observed, and the SOFA scores at

admission and before cannulation were similar.”

Describe intervention (treatment/procedure/test):

“All surgical V-A ECLS implantation were performed in the operative

room or at bedside in case of extreme emergency by trained cardiac

surgeons.

For the surgical approach, femoral vessels were exposed with a limited transversal incision in the groin. Purse string sutures were made before cannulation in the femoral artery and vein. Access to

the common femoral artery, to the femoral vein and to the superficial

femoral artery (anterograde access for distal perfusion line insertion) were obtained using the Seldinger’s technique. Then, after preparation

with adequate dilatators, femoral vessels were cannulated.

The venous cannula was advanced under echocardiographic guidance

to confirm its progression and positioning. After cannulation,

the skin was surgically sutured.

Decannulation procedures were performed in the operative room

by cardiac surgeons. At the time of decannulation, the operative field

was reopened. Then, the arterial cannula was removed and the

purse string sutures were tied and, after clamping the femoral artery,

the arteriotomy was sutured. The distal reperfusion cannula was then retrieved from the distal superficial femoral artery and the purse

string suture was tied.”

Describe  control (treatment/procedure/test):

“For the percutaneous approach, V-A ECLS was implanted in the cardiac

catheterization laboratory under fluoroscopic guidance by senior interventional cardiologists trained to echo-guided vascular puncture, preclosing and large bore vascular access as previously described.

The common femoral artery and femoral vein were punctured under

ultrasound guidance using the Seldinger’s technique. The position of

the wires was verified using angiography. Preclosing using two Perclose

ProGlides

often as possible before cannulation in both the femoral artery and

the femoral vein, as previously described. nThen, after preparation

with adequate dilatators, femoral vessels were cannulated over a stiff

wire (Amplatz Super Stiff, Boston Scientific, Marlborough, Massachusetts,

USA) under fluoroscopic guidance. V-A ECLS decannulation was performed percutaneously in the catheterization room except for exceptional cases (no preclosing

performed or removal by the cardiac surgeon at the time of LVAD

implantation or heart transplant). All percutaneous decannulation

were performed using the “the crossover balloon occlusion technique”

as previously described. Briefly, the contralateral femoral artery was punctured to obtain a crossover access. A balloon was advanced in the external iliac artery and inflated, to allow a nontraumatic occlusion of the vessel. The arterial cannula was removed

and the knots were tied. The balloon was then deflated and an

angiography of the femoral artery was performed to ensure the proper closing of the artery and the absence of complications.”

Length of follow-up:

30 days

Loss-to-follow-up:

N.R.

Incomplete outcome data:

N.R.

Outcome measures and effect size (include 95%CI and p-value if available):

30-day survival

I: 10 out of 59 (17%)

C: 16 out of 61 (26%)

Major bleeding

I: 18 out of 59 (31%)

C: 4 out of 61 (7%)

Lower limb ischemia

I: 5 out of 59 (8%)

C: 2 out of 61 (3%)

Groin infection

I: 6 out of 59 (10%)

C: 1 out of 61 (2%)

Wang et al

Type of study:

retrospective cohort study

Setting and country:

ELSO registry was used, which consists of data from more than 400 member centers worldwide. 

Funding and conflicts of interest:

This work was supported by the National Key Research and

Development Program of China (2016YFC1301001, to X Hou). There are no conflicts of interest.

Inclusion criteria:

We included adult patients (≥18 years)

who received femoro-femoral VA-ECMO with percutaneous (percutaneous

group) or surgical (surgical group) cannulation primarily

for refractory cardiogenic shock from January 1, 2008, through

December 31, 2019.

Exclusion criteria:

We excluded patients who received VAECMO

during cardiopulmonary resuscitation. Additional exclusion

criteria included patients with multiple VA-ECMO runs,

hybrid ECMO modalities (VVA or VAV), or missing data with

regard to key variables: cannulation site and hospital outcome.

N total at baseline:

I (surgical): 3,343 cases (27%)

C (percutaneous): 9,249 patients (73%)

Important prognostic factors:

Age, median (IQR):

I: 57 (44-65)

C: 57 (46-65)

Sex female, n (%)

I: 1077 (32)

C: 2883 (31)

Groups comparable at baseline?

There was a significant difference

between the percutaneous group and the surgical

group in age, weight, race, valvular heart disease, obesity,

prior cardiac arrest, pre-ECMO RRT, baseline pH level,

bilateral groin cannulation, and distal limb cannula placement.

Describe intervention (treatment/procedure/test):

Surgical cannulation

Describe  control (treatment/procedure/test):

Percutaneous cannulation

Length of follow-up:

Not reported.

Loss-to-follow-up:

Not reported.

Incomplete outcome data:

Data for secondary outcomes were missing in 3,951 cases

(31%). Missing values for some variables resulted in different denominators for the following variables: severe limb ischemia, cannulation site bleeding, and

systemic infection (n = 8641).

Outcome measures and effect size (include 95%CI and p-value if available):

In-hospital mortality

I: 1925 out of 3343 (58%)

C: 4881 out of 9249 (53%)

After controlling for previously described risk

factors, percutaneous cannulation was independently associated

with a lower risk of in-hospital mortality (OR 0.76;

95% CI 0.70-0.84.

Severe limb ischemia

I: 78 out of 3343 (3%)

C: 286 out of 9249 (5%)

In multivariable logistic regression analyses,

with adjustment for potential confounders, percutaneous

cannulation was not independently associated with severe

limb ischemia (OR 1.28; 95% CI 0.93-1.62.

Cannulation site bleeding

I: 582 out of 3343 (22%)

C: 1177 out of 9249  (19%)

In multivariable logistic regression analyses, percutaneous

cannulation was independently associated with a lower risk

of cannulation site bleeding (OR 0.70; 95% CI 0.60-0.80).

Systemic infection

I: 390 out of 3343 (15%)

C: 513 out of 9249 (8%)

After adjustment for potential confounders,

percutaneous cannulation was independently associated

with a lower risk of systemic infection (OR, 0.63; 95% CI

0.54-0.74).

Author, year

Selection of participants

Was selection of exposed and non-exposed cohorts drawn from the same population?

Exposure

Can we be confident in the assessment of exposure?

Outcome of interest

Can we be confident that the outcome of interest was not present at start of study?

Confounding-assessment

Can we be confident in the assessment of confounding factors? 

Confounding-analysis

Did the study match exposed and unexposed for all variables that are associated with the outcome of interest or did the statistical analysis adjust for these confounding variables?

Assessment of outcome

Can we be confident in the assessment of outcome?

Follow up

Was the follow up of cohorts adequate? In particular, was outcome data complete or imputed?

Co-interventions

Were co-interventions similar between groups?

Overall Risk of bias

Definitely yes, probably yes, probably no, definitely no

Definitely yes, probably yes, probably no, definitely no

Definitely yes, probably yes, probably no, definitely no

Definitely yes, probably yes, probably no, definitely no

Definitely yes, probably yes, probably no, definitely no

Definitely yes, probably yes, probably no, definitely no

Definitely yes, probably yes, probably no, definitely no

Definitely yes, probably yes, probably no, definitely no

Low, Some concerns, High

Danial et al.

Definitely yes

Reason:

All consecutive patients aged over 16 who received

peripheral femoro-femoral VA-ECMO between January

2015 and December 2017 at our institution (La

Pitié-Salpêtrière University Hospital, Paris, France) were

Included.

Definitely yes

Reason:

During the study period, the same surgical team (senior

or assistant cardiac surgeons) performed all VA-ECMO

implantations (both surgical and percutaneous) and all

surgeons performed the same proportion of procedures.

No information

Definitely yes

Reason:

Baseline variables considered for propensity score estimation were chosen based on empirical knowledge for their expected influence on survival and other outcomes, and included: age, sex, weight, height, BMI, SOFA score, peripheral artery disease, diabetes mellitus, dyslipidemia, ECMO indication, reperfusion cannula, long-term anticoagulation before VA-ECMO, and chronic cardiac failure. Variable entered into the propensity score model were age, weigh, seize, body mass index, sex, SOFA score, peripheral artery disease, diabetes mellitus, dyslipidemia, VA-ECMO indications, chronic heart failure, reperfusion cannula and long-term anticoagulation before VA-ECMO

Definitely yes

Reason:

Clinical endpoints were compared between surgical

and percutaneous groups using propensity score (PS)

framework. The PS approach aims at creating a new

dataset in which the probability to receive surgical or

percutaneous ECMO is equal (as in a pure randomized

trial) to balance patients’ baseline characteristics.

Probably yes

Reason:

Reported variables were also reported in the results. Method of analyses reported.

Probably yes

Reason:

Follow of 30 days probably appropriate for the measured outcomes.

Missing data not reported, but data of matched participants was present.

Definitely yes

Reason:

Comparable after PS matching.

Low (all outcome measures)

Saiydoun et al.

Definitely yes

Reason:

Patients were selected from a registry

Probably yes

Reason:

All surgical V-A ECLS implantation were performed in the operative

room or at bedside in case of extreme emergency by trained cardiac

surgeons.

No information

Only information reported:

“Access related complications occurring at the time of cannulation,

during V-A ECLS support and after decannulation were recorded.”

Definitely yes

Reason:

Variables

included in the model were based on their clinical relevance (regardless

of their p-values), with others variables when their p-values was

inferior or equal to 0.15 on univariate analysis. The variables

included were age, body mass index, chronic heart failure, Sequential

Organ Failure Assessment (SOFA) score at cannulation, LVEF,

preclosing, cardiac arrest and refractory cardiac arrest.

Probably yes

Reason:

Log-Rank test was used for univariate survival analysis. A Cox proportional hazard regression model was performed to identify independent

multivariable factors predictive of survival. Variables

included in the model were based on their clinical relevance.

No information

Probably yes

Reason:

Follow of 30 days probably appropriate for the measured outcomes.

Missing data not reported.

Probably yes

Reason:

Correction for confounding was done, so co-interventions should be comparable.

Low (all outcome measures)

Wang et al.

Definitely yes

Reason:

Two authors (L.W and F.Y) independently reviewed the deidentified

dataset from the ELSO Registry database and identified all

patients meeting the inclusion criteria

Definitely yes

Reason:

Patients were included from the ELSO registry, which contains information on procedures, techniques, etc.

Probably yes

Reason:

The following

information was extracted for all included patients: demographic

variables; diagnoses associated with cardiogenic shock; comorbidities

associated with atherosclerosis; pre-ECMO support; pre-

ECMO cardiac arrest; pre-ECMO hemodynamic variables;

pre-ECMO arterial blood gases (ABGs); year of ECMO run; cannulation

site; cannulation techniques; cannula type; VA-ECMO

duration; and ECMO-associated mortality and morbidities, including

limb complications, cannulation site bleeding and infections.

Definitely yes

Reason:

variables found to be associated with in-hospital

mortality in previous studies:17,19 age, sex, weight, race, pre-

ECMO cardiac arrest, congenial heart disease, acute myocarditis,

post-heart or lung transplantation, and pre-ECMO RRT.

Probably yes

Reason:

adjusted analyses to control for confounders in the

evaluation of patients’ outcomes were performed.

Probably yes

Reason:

Data was subtracted from the ELSO registru

No information

Probably yes

Reason:

No statistical significant difference at baseline, and correction for confounding was done, so co-interventions should be comparable.

Low (all outcome measures)