Periprocedural Cardiopulmonary Bypass or Venoarterial Extracorporeal Membrane Oxygenation During Transcatheter Aortic Valve Replacement: A Systematic Review

Abstract

Background

There are limited data on the use of venoarterial extracorporeal membrane oxygenation (VA‐ECMO) or cardiopulmonary bypass (CPB) to provide hemodynamic support periprocedurally during transcatheter aortic valve replacement. This study sought to evaluate patients receiving transcatheter aortic valve replacement with concomitant use of CPB/VA‐ECMO.

Methods and Results

We systematically reviewed the published literature from 2000 to 2018 for studies evaluating adult patients requiring CPB/VA‐ECMO periprocedurally during transcatheter aortic valve replacement. Studies reporting short‐term and long‐term mortality were included. Given the significant methodological and statistical differences between published studies, meta‐analysis of the association of CPB/VA‐ECMO with mortality was not performed. Of the 537 studies identified, 9 studies representing 5191 patients met our inclusion criteria. Median ages were between 75 and 87 years with 33% to 75% male patients. Where reported, the Edwards SAPIEN™ transcatheter heart valve was the most frequently used. A total of 203 (3.9%) patients received periprocedural hemodynamic support with CPB/VA‐ECMO. Common indications for CPB/VA‐ECMO included left ventricular or aortic annular rupture, rapid hemodynamic deterioration, aortic regurgitation, cardiac arrest, and left main coronary artery obstruction. The use of CPB/VA‐ECMO was predominantly an emergent strategy and was used for durations of 1 to 2 hours. Short‐term mortality (in‐hospital and 30‐day) was 29.8%, and 1‐year mortality was 52.4%. Major complications such as bleeding, vascular injury, tamponade, stroke, and renal failure were noted in 10% to 50% of patients.

Conclusions

CPB/VA‐ECMO was used in 4% in the early experience of patients undergoing transcatheter aortic valve replacement, most commonly for periprocedural complications. There are limited data on preprocedural planned use of VA‐ECMO, and the characteristics of this population remain poorly defined.

Clinical Perspective

What Is New?

• In this systematic review of 9 studies, venoarterial extracorporeal membrane oxygenation was used in 4% in the early experience of patients undergoing transcatheter aortic valve replacement, most commonly for periprocedural complications with limited information on preprocedural indications and planning.

What Are the Clinical Implications?

• Further dedicated studies are needed to define the preprocedural role of patient, procedural, and device factors in the use of venoarterial extracorporeal membrane oxygenation before transcatheter aortic valve replacement.

Introduction

The introduction of transcatheter aortic valve replacement (TAVR) into clinical practice has revolutionized the management of aortic stenosis since its approval by the Food and Drug Administration in 2011.1 This was initially offered as an alternative to surgical aortic valve replacement in patients with a high risk for cardiac surgery but has subsequently expanded to encompass intermediate‐risk populations.1, 2, 3 With improving expertise in patient selection and procedural competence, the TAVR technology is being increasingly offered to patients with higher age, frailty, and comorbidity profiles.4 As noted in a recent study from the Transcatheter Valve Therapy registry, between 2012 and 2014, TAVR was offered to nearly 10% of patients on an urgent/emergent basis.4 These populations, among others, are at a high risk of preoperative hemodynamic instability and perioperative hemodynamic deterioration, particularly when they present at a later stage of the aortic stenosis disease progression.4

In the catheterization laboratory, mechanical circulatory support (MCS) devices are often used as hemodynamic adjuncts for coronary, structural, and electrophysiological procedures.5, 6, 7 There have been multiple studies evaluating unselected MCS devices in the perioperative management of TAVR patients.7 Despite the concomitant use of both TAVR and cardiorespiratory support using the cardiopulmonary bypass (CPB) machine or the venoarterial extracorporeal membrane oxygenation (VA‐ECMO) technologies over the past decade, there are limited data on the indications, procedural characteristics, complications, and outcomes of these patients.8 In this systematic review we sought to analyze the clinical profile, indications, complications, device parameters, and mortality outcomes in TAVR patients needing CPB/VA‐ECMO.

Material and Methods

Data Sources and Search Strategies

This study was performed using publicly available data from published literature. The data, analytic methods, and study materials have been made available to other researchers for purposes of reproducing the results or replicating the procedure (Data S1). A comprehensive search of several databases (Ovid MEDLINE Epub Ahead of Print, Ovid Medline In‐Process & Other Non‐Indexed Citations, Ovid MEDLINE, Ovid EMBASE, Ovid Cochrane Central Register of Controlled Trials, Ovid Cochrane Database of Systematic Reviews, and Scopus) from 2000 to February 9, 2018 was conducted. The search strategy was designed and conducted by a medical librarian with input from the study's first author. Controlled vocabulary supplemented with keywords was used to search for mortality outcomes in adult patients needing CPB/VA‐ECMO before or after TAVR. The detailed search strategy is presented in Data S1. The resultant abstracts were screened by 2 independent reviewers (H.P., H.S.). All references of included studies were evaluated for additional studies. Study inclusion was based on the consensus of the 2 reviewers. A third independent reviewer (Saarwaani V.) served as the referee in case of disagreement between the first 2 reviewers in conjunction with the first author (Saraschandra V.). The search strategy and reporting were performed using STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines.9

Inclusion and Exclusion Criteria

Studies that included adult (>18 years) patients undergoing TAVR procedures with the use of perioperative CPB/VA‐ECMO were included. Case‐control, cohort, case series, and randomized trial study designs were included. In studies reporting outcomes in unselected TAVR patients, only studies for which a 2×2 table could be constructed between CPB/VA‐ECMO and mortality were included. Abstracts presented at professional societal meetings were excluded because they are subject to a higher risk of bias due to lack of rigorous peer review. Case reports, systematic or narrative reviews, pediatric or animal studies, and studies without relevant outcomes were excluded. If multiple studies were published by the same group of authors over the same study duration, only the largest study with relevant outcomes was included. Data abstracted included study year, population, location, type of study, CPB/VA‐ECMO–related parameters, TAVR‐related parameters, echocardiography data, and clinical outcomes. Quality was assessed using the Newcastle Ottawa Scale (http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp).

Given the significant methodological and statistical differences among published studies, meta‐analysis of the association of CPB/VA‐ECMO with mortality was not performed. Therefore, available evidence was summarized using systematic review methodology.

Results

The search strategy identified 537 unique abstracts. Of those, 9 studies representing a total of 5191 patients met the inclusion criteria (Figure and Table 1).8, 10, 11, 12, 13, 14, 15, 16, 17 All studies except the study by Seco et al15 were retrospective studies of institutional or national databases. All studies were published from 2012 to 2017, with 7 out of 9 studies published from the United States and Europe. All studies demonstrated a low risk of bias (Newcastle Ottawa Scale >3) (Table S1). As noted in Table 1, the median ages across the studies varied between 75 to 87 years with 33% to 75% male patients. The aortic valve echocardiographic parameters were reflective of severe aortic stenosis (valve area <1.0 cm² and average mean gradient >40 mm Hg) (Table 2). Where reported, the Edwards SAPIEN transcatheter heart valve was the most frequently used. This was the only valve in the PARTNER (Placement of Aortic Transcatheter Valves) trial, which contributed the most patients to this systematic review.8 As noted in Table 2, a majority (≥67%) of the TAVR procedures were performed by a transfemoral approach, except in the study by Uehara et al,17 in which only 14% were performed transfemorally.

Figure 1.

Literature search strategy.

Table 1.

Study Characteristics

Author/Year	Study Characteristics				Patients Needing CPB/VA‐ECMO
	Country	Design	Number	Mortality End Point	Number	Indications	Age (y) Median±SD	Male, N (%)
Arlt 201210	Germany	Retrospective	14	In hospital	4	Cardiorespiratory arrest	79.7±5.6	3 (75)
Banjac 201611	United States	Retrospective	230	In hospital	10	LV perforation, VF/VT, aortic rupture, valve embolization, left main impingement, prolonged pacing, severe AR	83±8	5 (50)
Dolmatova 201712	United States	Retrospective	247	In hospital	6	VF, respiratory failure, LV perforation, aortic annulus rupture	82±7.4	2 (33.3)
Husser 201313	Germany	Retrospective	214	30‐d	18	LV perforation, cardiogenic shock, VF/VT, severe LV dysfunction, prolonged pacing, high‐dose vasopressors, high‐risk left main coronary intervention	80±5	9 (50)
Pontailler 201714	France	Retrospective	5	30‐d	5	Tamponade, hemorrhagic shock, MI, LV dysfunction	NA	NA
Seco 201415	Australia	Prospective	100	30‐d	11	Prophylactic: Heart failure hospitalization pre‐TAVR, moderate/severe LV/RV failure, hemodynamic instability during balloon aortic valvuloplasty before TAVR, CVP or PCWP >20 mm Hg, mPAP >40 mm Hg, CI <2.0 L/min per m2. Rescue: VF causing cardiogenic shock, aortic annulus rupture/tamponade with hemodynamic deterioration	75.7±9.11	8 (73)
Shreenivas 20158	United States	Prospective registry	2525	1‐y	109	Cardiogenic shock	NA	NA
Trenkwalder 201716	Germany	Retrospective	1810	In hospital, 30‐d, 1‐y	33	LV perforation, low cardiac output, hemorrhage, coronary artery impingement, ventricular arrhythmias, severe AR, aortic annulus rupture, aortic dissection	80.5±5.4	14 (42.4)
Uehara 201717	Japan	Retrospective	46	In hospital, 30‐d, 1‐y	7	Sudden systolic blood pressure decrease, mPAP elevation, VF, severe LV dysfunction	87.3±3.6	NA

AR indicates aortic regurgitation; CI, cardiac index; CPB, cardiopulmonary bypass; CVP, central venous pressure; LV, left ventricular; MI, myocardial infarction; mPAP, mean pulmonary artery pressure; NA, not available; PCWP, pulmonary capillary wedge pressure; RV, right ventricle; TAVR, transcatheter aortic valve replacement; VA‐ECMO, venoarterial extracorporeal membrane oxygenation; VF, ventricular fibrillation; VT, ventricular tachycardia.

Table 2.

Characteristics of TAVR Patients Receiving Cardiopulmonary Support

Author/Year	Risk Scores		Echocardiographic Characteristics				TAVR Characteristics
	STSa	Logistic EuroSCOREa	LVEF (%)a	AVA (cm2)a	Mean Gradient (mm Hg)	Peak Gradient (mm Hg)	Valve Type	Valve Size (mm)b	Transfemoral, N (%)	ViV, N (%)
Arlt 201210	NA	35 (25‐48)	NA	NA	NA	NA	NA	NA	4 (100)	NA
Banjac 201611	15.2 (6.96‐62.82)	NA	35 (20‐60)	NA	45±16	NA	SAPIEN: 10	29 (1), 26 (6), 23 (3)	8 (80)	NA
Dolmatova 201712	9.4±6.6	NA	39±24.5	0.55±0.15	38.4±4.5	NA	SAPIEN: 5 CoreValve: 1	NA	5 (83.3)	1 (16.7)
Husser 201313	NA	26 (18‐41)	>50%: 6 (33) 35% to 50%: 5 (29) ≤35%: 7 (39)	0.8±1.0	43±10	74±15	NA	NA	12 (66.7)	NA
Pontailler 201714	NA	NA	NA	NA	NA	NA	NA	NA	NA	NA
Seco 201415	NA	51.73±24.95	38.75±17.52	0.64±0.11	39.7±7.66	59.8±14.14	NA	NA	11 (100)	1 (9)
Shreenivas 20158	NA	NA	NA	NA	NA	NA	SAPIEN: 109	NA	NA	NA
Trenkwalder 201716	NA	21.5±14.3	52.5±13.5	NA	50±16.8	NA	NA	NA	22 (66.7)	6 (18.2)
Uehara 201717	12.2±6.2	NA	53.1±21.3	0.44±0.19	62.6±24.3	NA	SAPIEN: 7	23 (5), 26 (2)	1 (14.3)	NA

AVA indicates aortic valve area; EuroSCORE, European System for Cardiac Operative Risk Evaluation; LVEF, left ventricular ejection fraction; NA, not available; STS, Society of Thoracic Surgery; TAVR, transcatheter aortic valve replacement; ViV, valve‐in‐valve.

^aRepresented as mean±standard deviation or median (interquartile range).

^bRepresented as size in mm (number of patients).

Of the 5191 TAVR patients in these 9 studies, 203 (3.9%) received periprocedural hemodynamic support with CPB/VA‐ECMO. The substudy of the PARTNER trial registry by Shreenivas et al was the largest study, contributing 109/203 (53.7%) patients, all of whom needed CPB.8 Commonly noted indications for CPB/VA‐ECMO included left ventricular (LV) or aortic annular rupture, rapid hemodynamic deterioration, severe aortic regurgitation, cardiac arrest from ventricular tachycardia or fibrillation, and obstruction of the left main coronary artery.8, 10, 11, 12, 13, 14, 15, 16, 17 The definition of hemodynamic deterioration varied across studies and included low cardiac output, need for high‐dose vasopressors and inotropes, prolonged pacing sequence, and severe LV dysfunction on echocardiography. Only the study by Seco et al specified criteria for prophylactic use of VA‐ECMO before TAVR as heart failure hospitalization pre‐TAVR, moderate to severe biventricular failure, and hemodynamic instability during balloon aortic valvuloplasty in addition to objective hemodynamic data.15

In the studies that reported prophylactic versus emergent use, a varying percentage of 17% to 73% of patients needed prophylactic CPB/VA‐ECMO support. The use of CPB/VA‐ECMO was predominantly an emergent strategy and used for durations of 1 to 2 hours except as noted in Table 3. All but 3 patients had peripheral implantation of VA‐ECMO in the studies that reported these data.11, 17 Major complications such as bleeding, vascular injury, tamponade, stroke, and renal failure were noted with varying frequency of 10% to 50% across the studies. Short‐term mortality (in‐hospital and 30‐day) was 29.8% (28/94 patients), and 1‐year mortality was 52.4% (78/149). Short‐term mortality (0‐46%) and 1‐year mortality was (14‐58%) varied widely among studies.

Table 3.

Procedural Characteristics, Complications and Outcomes of CPB/VA‐ECMO

Author/Year	Procedural Characteristics				Complications					Mortality, N (%)
	CPB/VA‐ECMO	Prophylactic, N (%)	Peripheral, N (%)	Duration (min)a	Major Bleed, N (%)	Vascular, N (%)	Tamponade, N (%)	CVA, N (%)	Hemodialysis, N (%)
Arlt 201210	VA‐ECMO	2 (50)	4 (100)	1 hour (3), 6 days (1)	0	0	NA	NA	NA	1 (25)
Banjac 201611	VA‐ECMO	0	8 (80)	87 (16‐8514)	1 (10)	NA	1 (10)	1 (10)	NA	3 (30)
Dolmatova 201712	VA‐ECMO	1 (16.7)	6 (100)	94 (20 to 4 days)	3 (50)	3 (50)	3 (50)		NA	2 (33.3)
Husser 201313	VA‐ECMO	9 (50)	18 (100)	102 (87‐148)b 116 (64‐125)c	6 (33)	5 (28)	3 (16)	2 (11)	3 (17)	4 (22)
Pontailler 201714	VA‐ECMO	NA	NA	NA	NA	NA	NA	NA	NA	2 (40)
Seco 201415	VA‐ECMO	8 (72.7)	11 (100)	NA	1 (9)	2 (18)	1 (9)	0		1 (9)
Shreenivas 20158	CPB	NA	NA	NA	NA	NA	NA	NA	NA	58 (53.5)
Trenkwalder 201716	VA‐ECMO	0	33 (100)	114 (61‐445)	27 (81.8)	8 (24.2)	NA	3 (9.1)	10 (30.3)	15 (45.5)d 19 (57.6)e
Uehara 201717	VA‐ECMO	3 (42.8)	6 (85.7)	51.8±29.9	1 (3)	0	NA	0	1 (14)	0 (0)d 1 (14.2)e

CPB indicates cardiopulmonary bypass; CVA, cerebrovascular accident; NA, not available; VA‐ECMO, venoarterial extracorporeal membrane oxygenation.

^aMean±SD or median (interquartile range).

^bProphylactic.

^cEmergency.

^dIn‐hospital/30‐d mortality.

^eOne‐year mortality.

Discussion

In this systematic study, preprocedural or postprocedural CPB/VA‐ECMO was used in 4% in the early experience of patients undergoing TAVR, primarily as an emergent strategy for procedural complications. Short‐term mortality (in‐hospital or 30‐day mortality) was 29.8% and 1‐year mortality was 52.4% in TAVR patients requiring VA‐ECMO or CPB. Due to the improvements in device technology, technique, and success rates, TAVR has evolved into a relatively low‐risk procedure despite the high comorbidity and relative frailty of patients needing this procedure.18 As noted in a recent article, less than 5% of all TAVR procedures require conversion to open surgery, making this procedure both safe and feasible in high‐risk populations.19 However, in patients presenting with concomitant cardiogenic shock, TAVR is associated with nearly 33% 30‐day and 60% 1‐year mortality.20 As noted in a recent study from the Transcatheter Valve Therapy registry, nearly 10% of contemporary TAVRs are urgent or emergent and nearly 8% are performed in patients with LV ejection fraction <30%.4 It is therefore of the utmost importance to optimize the hemodynamics in these patients to prevent periprocedural complications. Such patients may receive lower‐support MCS devices such as the intra‐aortic balloon pump and Impella; however some cases ultimately require escalation to the VA‐ECMO due to persistent hemodynamic compromise.7, 21 In a study from the National Inpatient Sample, Singh et al noted 7.4% of TAVR procedures performed in 2011‐2012 required VA‐ECMO support that is likely reflective of early experience and learning curve associated with TAVR.21

Intra‐ or postprocedurally, CPB/VA‐ECMO was used as an emergent maneuver with a variable incidence of 33% to 83% as noted in a majority of the studies in this systematic review. Mechanical complications such as LV or aortic annular rupture, obstruction of the left main coronary artery, and physiological complications such as rapid hemodynamic deterioration, severe aortic regurgitation, and cardiac arrest from ventricular tachycardia or fibrillation were the chief indications for CPB/VA‐ECMO. Need for high‐dose vasopressors, hemorrhagic shock, periprocedural myocardial infarction, and aortic dissection were quoted as other indications for CPB/VA‐ECMO in this population. The relative advantages of VA‐ECMO include rapid bedside access, high‐flow circuit, relatively low expense, and concomitant pulmonary support that make it an attractive option in postoperative emergencies.11 Importantly, unlike the percutaneous LV assist devices, the peripheral VA‐ECMO does not require transseptal placement or the crossing of the aortic valve.13 Cardiac tamponade, particularly from the rupture of a highly calcified and fragile aortic annulus, is an independent predictor of mortality in patients with TAVR.22 The CPB/VA‐ECMO has been noted to be extremely efficient in establishing rapid circulatory support in the treatment of cardiac tamponade in patients with complications of catheter‐based interventions.12 Access for VA‐ECMO can either be obtained before TAVR procedure or emergently intraprocedurally. Cannulation is performed carefully in the femoral artery and femoral vein using ultrasound‐guided technique, stiff 0.035 inch wires, and serial dilation to the desired VA‐ECMO cannula size (typically 16F‐18F arterial and 21F venous). If ECMO is required emergently during TAVR, large bore arterial access is usually already available, and the TAVR access sheath can be replaced with the ECMO arterial cannula. Venous access is often already available as it is used for temporary pacing.

Despite its stated benefits, CPB/VA‐ECMO does have an inherent risk of complications. It is associated with significant need for blood transfusions either due to hemorrhage or hemolysis.23 Transfusion, independently, in patients with VA‐ECMO and TAVR has been associated with worse outcomes and needs to be carefully titrated against the need to maintain a robust cardiac output.23, 24 The need for large‐bore femoral arterial cannulation is associated with a higher risk of vascular complications, such as retroperitoneal hemorrhage, distal limb ischemia, and arterial laceration.25 Importantly, the use of either CPB or VA‐ECMO has not shown convincing evidence of mortality advantage in patients undergoing TAVR. In a retrospective review of the PARTNER registry, Shreenivas et al demonstrated that, regardless of an emergent versus planned strategy, CPB was associated with nearly 2‐fold higher mortality.8 The Society of Thoracic Surgeons Predicted Risk of Mortality and Logistic European System for Cardiac Operative Risk Evaluation Scores have shown poor calibration for predicting MCS use, and have advocated for further research into optimal patient, procedural, and device factors to develop individualized risk scores.8 Neurological complications in TAVR have been receiving increasing attention in recent years due to their long‐term consequences. The use of emergent VA‐ECMO in patients with TAVR has been associated with a higher rate of major strokes and necessitates further study into mechanistic aspects.13

In patients with severe LV dysfunction, acute heart failure, and cardiogenic shock, TAVR remains an attractive option to treat severe aortic stenosis, however only after clinical stabilization.26 Currently, there is limited evidence on the prophylactic use of VA‐ECMO in critically ill patients needing a TAVR procedure. As noted in this article, only 1 out of 9 studies defined a high‐risk cohort for preprocedural VA‐ECMO use.15 Patients with pulmonary hypertension and biventricular failure would conceivably benefit from prophylactic VA‐ECMO; however this high‐risk cohort needs to be better defined to optimize the clinical outcomes.17 We recommend a multidisciplinary team approach for the care of these patients to decide among available therapies including medical management, use of balloon aortic valvuloplasty, durable LV assist devices, and palliative care.26 Ideally, these teams should comprise physicians from cardiology, interventional cardiology, cardiac surgery, critical care medicine, heart failure, palliative medicine, and anesthesiology.26, 27 It is important to incorporate input from palliative medicine physicians to balance the need for using sophisticated MCS devices against the use of futile, resource‐intensive therapies that may be unlikely to improve clinical outcomes.

Limitations

This systematic review included only 9 studies, most of which were published in 2017. Heterogeneous study inclusion and exclusion criteria, high confounding from periprocedural complications, and inconsistent criteria for the use of CPB/VA‐ECMO prevented a meta‐analysis on these data. Most of the included studies restricted their mortality outcomes to <30 days; therefore, there are limited data on the prediction of long‐term outcomes in this population. It is conceivable that the timing from patient collapse to cannulation for cardiopulmonary support and the location of placement (operating room versus catheterization laboratory) have prognostic implications in patients needing CPB/VA‐ECMO intraoperatively; however these data were inconsistently reported across studies preventing meaningful conclusions in this review. Higher hospital volumes for TAVR and VA‐ECMO have independently been shown to be associated with better outcomes; however these data were not analyzed in this study, thereby limiting the assessment of clinical outcomes.28, 29

Conclusions

In this systematic review, CPB/VA‐ECMO was used in 4% in the early experience of patients undergoing TAVR, most commonly for periprocedural complications. There are limited data on preprocedural planned use of VA‐ECMO, and the characteristics of this population remain poorly defined. In patients with cardiogenic shock, biventricular failure, and cardiopulmonary compromise, the use of VA‐ECMO provides an attractive option for hemodynamic support. Further dedicated studies are needed to balance the need for using sophisticated MCS devices against the use of futile, resource‐intensive therapies that may be unlikely to improve clinical outcomes in this high‐risk population.

Author Contributions

Study design, literature review, data analysis, and statistical analysis were done by Saraschandra V., H.S., H.P., and Saarwaani V.; data management, data analysis, and drafting of the manuscript were by Saraschandra V., H.S., H.P., and Saarwaani V.; access to data were by Saraschandra V., H.S., H.P., Saarwaani V., G.W.B., S.M.D., K.L.G., D.R.H., and M.F.E.; manuscript revision, intellectual revisions, and mentorship were done by G.W.B., S.M.D., K.L.G., D.R.H., and M.F.E.; and final approval was given by Saraschandra V., H.S., H.P., Saarwaani V., G.W.B., S.M.D., K.L.G., D.R.H., and M.F.E.

Disclosures

None.

Supporting information

Data S1. Detailed Search Strategy

Table S1. Newcastle Ottawa Scale for Assessment of Bias

(J Am Heart Assoc. 2018;7:e009608 DOI: 10.1161/JAHA.118.009608.)