Peripheral venoarterial extracorporeal membrane oxygenation for periprocedural Cardiogenic shock during interventional cardiology Oxigenación con membrana extracorpórea veno-arterial periférica para shock cardiogénico peri-procedimiento en cardiología intervencional

Background : Cardiogenic shock (CS) is uncommon in the cardiac catheterization laboratory (CCL) among patients undergoing coronary angiography. Periprocedural CS is more frequent in high-risk patients and in technically demanding procedures. Aim : To describe the clinical outcomes of patients who underwent peripheral venoarterial extracorporeal membrane oxygenation (pVA-ECMO) for CS associated with interventional cardiology procedures. Material and Methods : Review of clinical records of seven patients treated between January 2014 and October 2018. Results: pVA-ECMO was implanted within 6 hours of the interventional cardiology procedure. All patients had coronary artery disease and one of them also had symptomatic severe aortic stenosis. One patient entered the CCL in cardiac arrest. Percutaneous coronary intervention (PCI) was performed in all patients; four patients underwent an emergency procedure and five patients experienced PCI complications. One patient undergoing transcatheter aortic valve replacement suffered acute severe aortic regurgitation. An intra-aortic balloon pump was inserted at the CCL in five patients. Six patients experienced cardiac arrest. Mean SAVE score was -4.3 and baseline lactate 55 mg/dl. pVA-ECMO mean duration was 5 ± 4 days. Survival after both hospital discharge and 12 months of follow-up was 85.7% Regarding vascular access complications, we observed one access site hematoma and one episode of cannulation site bleeding requiring surgical repair. Conclusions: pVA-ECMO should be considered in patients with periprocedural CS as a bridge to recovery. with outcomes in this Med Chile 2020; 148: 1295-1301)


Introduction
C ardiogenic shock (CS) is uncommon in the cardiac catheterization laboratory (CCL), with an incidence of 0.24%-0.47% among patients undergoing coronary angiography 1,2 . Periprocedural CS is more frequent in high-risk patients due to baseline cardiovascular disease and the technically-demanding procedures performed. Observational studies reveal an incidence of 3.5%-4.3% in patients with acute coronary syndromes 3,4 , and a periprocedural CS incidence of 1.6% has been reported in intermediate-risk patients undergoing transcatheter aortic valve replacement (TAVR) 5 . Despite advances in percutaneous interventions and hemodynamic support, the prognosis for CS remains poor, with short-term mortality rates of 40%-60% 2,6,7 .
Short-term mechanical circulatory support (MCS) aims to improve hemodynamic status and tissue perfusion to achieve recovery or provide time to establish a prognosis and definitive treatment. Consensus statements suggest that short-term MCS may be considered in patients experiencing periprocedural CS 8,9 . Among MCS devices, peripheral venoarterial extracorporeal membrane oxygenation (VA-ECMO) has several advantages, including easy installation, highly effective hemodynamic and respiratory support, compatibility with therapeutic hypothermia and low cost 8 .
This report describes the clinical outcomes of patients undergoing peripheral VA-ECMO for CS linked to interventional cardiology procedures at our center.

Patients and Methods
We conducted a retrospective observational study at a tertiary care center. The study included adult patients (≥ 18 years) undergoing interventional cardiology procedures who developed periprocedural CS and underwent peripheral VA-ECMO from January 2014 to October 2018. The study was approved by the Chilean Eastern Metropolitan Health Service ethics committee.
CS was defined according to the Society for Cardiovascular Angiography and Interventions (SCAI) consensus statement 6,7 . Patients were retrospectively categorized into one of five CS stages at CCL admission, prior to VA-ECMO: At risk (A), Beginning (B), Classic (C), Deteriorating (D), Extremis or Refractory (E). The subscript "A" at any stage signifies cardiac arrest. Indications for peripheral VA-ECMO were at the discretion of the HEART team. Peripheral VA-ECMO was implanted by cardiac surgeons in the CCL or operating room (OR) within 6 hours of the interventional cardiology procedure in the case of deteriorating or refractory CS. Cannulation of femoral vessels was performed using a modified Seldinger technique, with 15-22 Fr arterial and 21-29 Fr venous cannulae, according to surgeon preference. Patients were postoperatively admitted to the ICU and treated per intensive care standards. VA-ECMO was removed in the OR by cardiac surgeons.
Clinical records were reviewed to record demographics, vital signs, laboratory data, clinical outcomes, and procedures performed during the hospital stay. Estimated in-hospital survival was assessed with the modified survival after VA-EC-MO (SAVE) score 10 . The Chilean Civil Registry database was reviewed for survival statistics.

Statistics
Continuous data are expressed as mean ± standard deviation or median and categorical data as absolute number and percentage.

Results
Peripheral VA-ECMO was implanted in seven patients. Baseline characteristics are shown in Table 1. Mean age was 56.9 ± 16.9 years; 71.4% were male; 57.1% had hypertension; 42.9% had dyslipidemia; 42.9% had prior tobacco consumption. All patients had coronary artery disease. One patient had symptomatic severe aortic valve stenosis and severe stenosis in the middle segment of the left descending coronary artery and proximal segment of the circumflex artery, considered a candidate for elective TAVR and PCI after assessment by the HEART team. One patient entered the CCL in cardiac arrest (SCAI E A ); six patients (85.7%) were at risk for CS (SCAI A). Percutaneous coronary intervention (PCI) was performed in all patients. Four patients (57.1%) had emergency PCI due to ST-elevation myocardial infarction. Two patients (28.6%) with chronic coronary syndrome were scheduled for elective PCI for left main coronary artery bifurcation disease. Five patients (71.4%) experienced PCI complications: one dissection at the distal stent edge; one abrupt left main coronary artery closure; one coronary perforation managed with a covered stent with later intraprocedural stent thrombosis; and two intraprocedural stent thromboses. One patient undergoing TAVR suffered acute severe aortic regurgitation. All complications were associated with hemodynamic instability and managed percutaneously during the procedure, achieving revascularization of target vessels or TAVR implant accordingly.
Upon hemodynamic deterioration, patients received vasopressor/inotropic medications, with a median of two drugs. An intra-aortic balloon pump (IABP) was inserted in the CCL in five patients (71.4%). Six patients (85.7%) experienced cardiac arrest, with a cardiopulmonary resuscitation time of 23 ± 12 minutes. Peripheral VA-EC-MO was implanted in the CCL in five patients (71.4%); in two cases as extracorporeal life support and in the others due to CS. Two patients (28.6%) completed PCI under VA-ECMO. Two patients (28.6%) had peripheral VA-ECMO implanted in the OR. Mean SAVE score was -4.3 ± 4.9 (median -5); baseline lactate 55 ± 36 mg/dl (median 36 mg/ dl); and modified SAVE score 6.4 ± 10.5 (median 7). VA-ECMO pump flow was 3.7 ± 0.6 L/min. Table 2 summarizes clinical outcomes and complications. Peripheral VA-ECMO duration was 5 ± 4 days. Concomitant IABP was used in three patients (42.8%). At initial echocardiographic assessment, mean left ventricular ejection fraction was 38% ± 11%. Lactate clearance returned to normal values within 48 hours in six patients (85.7%) (Figure 1). Median invasive mechanical ventilation time, ICU and hospital stays were 14, 18 and 26 days, respectively. Survival to both hospital discharge and 12 months of follow-up was 85.7%. The deceased patient developed multi-organ failure and severe hypoxic-ischemic encephalopathy and was declared brain dead at 48 hours from VA-ECMO implant. One patient developed pulmonary edema refractory to adjustments in therapy, requiring bridging to central MCS (CentriMag TM ) and transplant (performed after 86 days). The other patient with pulmonary edema was successfully weaned. Five patients (71.4%) had ventilator-associated pneumonia. Regarding vascular access complications, we observed one access site hematoma and one episode of cannulation site bleeding requiring surgical repair. Two patients required hemodialysis (28.6%).

Discussion
We describe an initial experience with peripheral VA-ECMO for periprocedural CS in a heterogeneous population undergoing interventional cardiology procedures. The use of peripheral VA-ECMO was effective in restoring systemic perfusion, allowing for completion of procedures; bridge to recovery, transplantation or decision and overall improved clinical outcomes.
Patients with periprocedural CS subjected to peripheral VA-ECMO had in-hospital and 12-month survival rates of 85.7% in our series.
Prior studies in patients undergoing interventional cardiology procedures have described lower in-hospital survival rates despite implementation of VA-ECMO programs in the CCL. In patients with acute myocardial infarction experiencing refractory CS, several cases series reported survival rates of 22%-65% after peripheral VA-ECMO [11][12][13][14][15][16][17] . In patients suffering hemodynamic complications during TAVR, rescue VA-ECMO support produced in-hospital survival rates of 44%-75% [17][18][19][20] . Differences in survival might be partially explained by variability in definitions of refractory CS and associated multi-organic failure. This problem  All patients experienced at least one complication associated with peripheral VA-ECMO. In terms of peripheral VA-ECMO complications in the context of periprocedural CS, previous studies report vascular complications requiring surgery in 11%-37%, bleeding related to ECMO in 8%-34%, neurologic complications in 9%-32% and dialysis requirement in 30%-46% of patients. Our results were similar for the aforementioned complications. Interestingly, installation by experienced operators in the CCL does not seem to produce a higher rate of vascular complications as compared to surgical implantation in the OR 11 .

Limitations
Study limitations include the small number of patients and retrospective design, which preclude further analysis. There was inconsistent availability of some variables, including time to pump installation and transfusions. Comparisons with other types of MCS were not feasible, given the concomitant IABP and lack of availability of other devices.

Conclusion
Peripheral VA-ECMO was associated with improved clinical outcomes in this series. This procedure should be considered in patients with periprocedural CS as a bridge to recovery or decision. Prospective multicenter registries evaluating the safety, efficacy or futility of MCS in different settings are of significant interest.