Catheter Ablation in Patients with Cardiogenic Shock and Refractory Ventricular Tachycardia

Abstract

Background

There is paucity of data regarding radiofrequency ablation for ventricular tachycardia (VT) in patients with cardiogenic shock and concomitant VT refractory to antiarrhythmic drugs on mechanical support

Methods

Patients undergoing VT ablation at our center were enrolled in a prospectively maintained registry and screened for the current study (2010–2017)

Results

All 21 consecutive patients with cardiogenic shock and concomitant refractory ventricular arrhythmia undergoing “bailout” ablation due to inability to wean off mechanical support were included. Median age was 61 years, 86% were males, median LVEF was 20%, 81% had ischemic cardiomyopathy, and PAINESD score was 18 ± 5. The type of mechanical support in place prior to the procedure was intra-aortic balloon pump (IABP) in 14 patients (67%), Impella CP in 2, ECMO in 2, ECMO and IABP in 2, and ECMO and Impella CP in 1. Endocardial voltage maps showed myocardial scar in 19 patients (90%). The clinical VTs were inducible in 13 patients (62%), whereas 6 patients had PVC induced VF/VT (29%), and VT could not be induced in 2 patients (9%). Activation mapping was possible in all 13 with inducible clinical VTs. Substrate modification was performed in 15 patients with scar (79%). After ablation and scar modification, the arrhythmia was non-inducible in 19 patients (91%). Seventeen (81%) were eventually weaned off mechanical support successfully, but 6 (29%) died during the index admission from persistent cardiogenic shock. Patients who had ventricular arrhythmia and cardiogenic shock on presentation had a trend towards lower in-hospital mortality compared to those who presented with cardiogenic shock and later developed ventricular arrhythmia.

Conclusion

“Bailout” ablation for refractory ventricular arrhythmia in cardiogenic shock allowed successful weaning from mechanical support in a large proportion of patients. Mortality remains high, but the majority of patients were discharged home and survived beyond 1 year.

Graphical Abstract

Introduction

Catheter-based ablation is an effective treatment for ventricular tachycardia (VT) and is associated with reduction of VT burden, improved quality of life and improved survival in select patients. ^1–6

Cardiogenic shock and electrical storm are a major cause of morbidity and mortality in heart failure patients, despite advances in clinical management and strategies for hemodynamic support. ^7–11 In fact, cardiogenic shock secondary to refractory ventricular arrhythmia carries a very high mortality rate up to 50%. ¹²

Patients with cardiogenic shock are at risk of ventricular arrhythmias, and in some scenarios, such arrhythmias would interfere or preclude successful weaning from mechanical support, despite the use of potent antiarrhythmic drugs (AADs). The management of this high-risk group of patients is extremely challenging and would be limited to either “bailout” ablation of ventricular arrhythmia, or cardiac transplantation. The latter may be complicated by both short and long-term morbidity and limited by donor availability. Furthermore, in the setting of hemodynamic compromise, VT ablation becomes even more challenging hindering the ability to maintain the arrhythmia long enough to allow for adequate mapping and ablation, without further hemodynamic deterioration and cerebral desaturation.

In this study, we sought to evaluate the short-term, and long-term procedural, and clinical outcomes of catheter ablation for VT in patients with persistent cardiogenic shock and inability to wean off mechanical hemodynamic support due to refractory recurrent VT/VF despite AADs.

Methods

Data reported in this study will not be made available for other researchers.

Patient population

All patients undergoing catheter ablation procedures for ventricular arrhythmia at the Cleveland Clinic are enrolled in a prospectively maintained registry. All patients who were included in the registry between January 2010 and December 2017 were screened for eligibility for the current study. All procedural complications were prospectively adjudicated in a monthly cardiac electrophysiology meetings in a multidisciplinary adjudication process involving electrophysiology operators, fellows, nurses, anesthesiologists and wherever applicable intensivists, and heart failure specialists.

We retrospectively identified and included all patients who underwent “bailout” ablation for refractory VT/VF while in cardiogenic shock with inability to wean off mechanical hemodynamic support including intra-aortic balloon pump (IABP), Impella CP™ (AbioMed), and extracorporeal membrane oxygenation (ECMO). The indication of hemodynamic support in all patients was either primary cardiogenic shock, or ventricular arrhythmia with concomitant cardiogenic shock. All patients had refractory ventricular arrhythmias that were interfering with attempts of weaning mechanical support despite AADs. In addition, for the purpose of PAINESD score comparison between patients in this study and those without cardiogenic shock undergoing VT ablation, for every study patient we included two patients who underwent ablation immediately before and immediately after that case (using the chronological order of VT ablation entries into the database). For the purpose of PAINESD score comparison, we excluded any patients with idiopathic VT, PVCs or non-sustained VT. The chronological order was used as the main entry criterion given difficulty matching on other clinical criteria of this critically ill population. Patients, or family members signed informed consent prior to the mapping and ablation procedures. The study was approved by the institutional review board.

Pre-ablation patient care and management

Prior to the ablation, patients were treated in the cardiac, heart failure, or cardiothoracic surgery intensive care units. Cardiogenic shock was defined as systolic blood pressure persistently less than 90 mmHg, with evidence of end organ damage such as altered mental status, elevated lactate, and decreased urine output. Moreover, this was further supported by hemodynamic data from right heart catheterization, such as decreased cardiac index, and decreased mixed venous oxygen saturation. Management of cardiogenic shock was performed by the respective ICU teams, with consultation from the advanced heart failure service for evaluation of the patient’s candidacy for advanced heart failure therapy, including transplant, and for input into management of the patient’s shock. Furthermore, an electrophysiology consult service was involved in management of ventricular arrhythmias. In addition to AADs, temporary atrial pacing was employed as clinically indicated. Revascularization options were also investigated and performed when indicated before adjudicating any of the cases as true refractory ventricular arrhythmias.

VT ablation procedures

Vascular access was obtained with ultrasound guidance in all patients. Recording and pacing catheters were positioned in the right ventricle (RV) in all patients with additional recording from the His bundle as needed. The use of intracardiac echocardiography was at the operator’s discretion. Irrigated tip ablation catheters, more recently with contact force sensing, were introduced to the left ventricle (LV) through a retrograde aortic approach or following standard trans-septal access. Intravenous heparin was used during the procedure to maintain an activated clotting time of 300 seconds and initiated prior to introducing catheters into the left cardiac chambers.

Endocardial electro-anatomical maps were then created. Voltage mapping was performed using a filling threshold of 12 mm. ¹³ Electrical signals were acquired during sinus rhythm, or ventricular pacing in patients with pacemaker dependency or resynchronization therapy. Unipolar electrograms were filtered at 2–240 Hz, and bipolar electrograms were filtered at 30–500 Hz. Commonly used clinical voltage criteria were employed to define areas of scar (bipolar voltage <0.5 mV), border zones (0.5–1.5 mV), and normal (>1.5 mV) myocardium. ¹³ Catheter stability during mapping and ablation was confirmed with the combination of fluoroscopy, electrogram characteristics, real-time intracardiac echocardiography whenever available, and mapping systems algorithms as applicable.

Programmed electrical stimulation was performed primarily form the RV apex, while additional sites including RV outflow tract, and LV, with additional isoproterenol infusion, were used in cases of non-inducibility with RV pacing alone. The programmed electrical stimulation protocol included using up to three drive train cycle lengths (350, 400, and 600 milliseconds) and up to three extra stimuli with minimal coupling interval of 200 milliseconds. All procedures targeted ablation of the clinical VTs as documented by 12-lead electrocardiograms, or presumed clinical VTs as assessed by VT cycle length, local RV timing to far-field electrogram, as well as far-field morphology from implantable cardiac defibrillator recordings.

Activation and entrainment mapping were attempted whenever possible. Otherwise, limited activation with additional pace-mapping were performed for hemodynamically unstable VTs despite mechanical support. Radiofrequency ablation targeted primarily areas thought to represent the critical central isthmus of clinical VTs, based primarily on activation mapping and entrainment, whenever possible and up to operator’s discretion. In addition, extensive VT substrate modification was performed targeting primarily the scar border zones including late fractionated potentials, and scar homogenization targeting abnormal potentials. In select scenarios, connection of the scar or ablation sets to an anatomic boundary such as mitral annulus was performed at the discretion of the operator. Power delivery was set to 40–50 watts, with duration of 60–90 seconds while monitoring for impedance rises to avoid steam pops. At the end of ablation, programmed electrical stimulation was repeated and successful ablation was defined as the inability to induce the clinical or presumed clinical VT. ¹⁴

Post ablation care and clinical follow-up

Patients were maintained on hemodynamic support for cardiogenic shock after the ablation procedures and transferred back to the cardiac intensive care unit. After optimization of volume status, attempts were made to wean off and eventually remove hemodynamic support, whenever possible. Clinical and follow-up data were collected from the index hospitalization and subsequent follow-up visits with both the heart failure and cardiac electrophysiology teams, as well as device clinic. Patients were encouraged upon discharge to report any new clinical events including defibrillator shocks or worsening heart failure. For patients form distant areas from our institution, all clinical assessments that took place with the local physicians were scanned into the electronic records and were available for review for the purpose of the medical records. All communication with the patients and their treating physicians was documented in electronic records as well.

Statistical analysis

All statistical analyses were performed by using the statistical software JMP pro version 10.0 (SAS, NC). A 2-sided P value <0.05 was considered statistically significant. The Student T test, and Chi-square tests or Fisher’s exact test, were used for comparison of means and proportions as appropriate. For non-normally distributed continuous variables, the nonparametric Wilcoxon rank-sum test was used.

Results

Patient characteristics

The baseline characteristics of the study population are summarized in Table 1. Twenty-one patients underwent “bailout” ablation for refractory VT/VF, in the setting of cardiogenic shock and inability to wean off mechanical hemodynamic support despite AADs. The median age was 67 [IQR 48–71] years, with 18(86%) males. The underlying cardiomyopathy was ischemic in 17 patients (81%), non-ischemic dilated cardiomyopathy in 3 (14%), and mixed in 1 patient (5%). The median LVEF was 20% [IQR 15%−32%]. Fourty-two patients were selected for the purpose of PAINESD comparison. The median PAINESD score was 18 [IQR 18–22] in the study group vs 9 [IQR 3–12] in the comparison group (p<0.0001).

Table 1.

Baseline characteristics of patients with cardiogenic shock

	Cardiogenic shock (n=21)
Age (years) (median [IQR])	67 [48–71]
Gender (male) (n (%))	18(86)
Left ventricular ejection fraction (%) (median [IQR])	20 [15–32]
Ischemic cardiomyopathy (n (%))	17(81)
Non-ischemic cardiomyopathy (n (%))	3(14)
Mixed cardiomyopathy (n (%))	1(5)
ICD present prior to ablation (n (%))	13(62)
CRT present prior to ablation (n (%))	5(24)
PAINSED score (Median [IQR])	18 [14–22]

Medical management and Mechanical support

Table 2 lists the pre-procedural details of all 21 patients, while table 3 summarizes the hemodynamic support, and arrhythmia characteristics. All patients were in cardiogenic shock with ventricular arrhythmias preventing weaning off mechanical hemodynamic support despite AADs. In 11 patients (52%), mechanical support was placed for the primary indication of cardiogenic shock, while 10 patients (48%) had mechanical support placed in an attempt to suppress ventricular arrhythmias that had resulted in decompensation and cardiogenic shock. Patients were maintained on a median of 2 antiarrhythmic drugs (IQR 1–2, mean 1.7 ± 0.5). Most patients (n=14, 67%) were on a combination of amiodarone and lidocaine. Two patients (9.5%) were on amiodarone alone, 3 (14%) were on lidocaine alone, 1 (5%) was on procainamide alone, and 1 (5%) was on amiodarone and procainamide. All anti-arrhythmic drugs were administered intravenously.

Table 2.

Detailed characteristics of all patients in the cohort

Patient	ICM/NICM*	LVEF (%)	NYHA III/IV	PAINESD score	MI prior to/on presentation	Type of VT	Total VT morphologies induced	AADs prior to ablation†
1	ICM	10	No	20	Yes	PVC induced VF	N/A	A, L
2	ICM	35	No	11	Yes	PVC induced VF	N/A	A, L
3	ICM	25	No	22	No	MMVT	2	A
4	ICM	20	Yes	26	No	MMVT	1	L
5	ICM	20	No	18	No	MMVT	1	A
6	ICM	30	No	19	No	MMVT	2	A, L
7	ICM	35	Yes	20	No	MMVT	1	A, L
8	ICM	19	No	17	No	MMVT	0	A, L
9	NICM	18	Yes	9	No	MMVT	7	A, L
10	ICM	15	No	22	No	MMVT	2	A, L
11	ICM	32	No	17	No	MMVT	1	A, L
12	ICM	9	Yes	23	No	MMVT	1	P
13	NICM	30	Yes	14	No	PVC induced VF	0	A, L
14	NICM	13	Yes	14	No	MMVT	4	L
15	ICM	15	No	14	Yes	PVC induced VF	N/A	A, P
16	Mixed	37	No	8	Yes	MMVT	2	L
17	ICM	20	Yes	23	No	PVC induced VF	N/A	A, L
18	ICM	27	Yes	25	No	MMVT	1	A, L
19	ICM	54	No	17	No	MMVT	8	A, L
20	ICM	15	No	18	Yes	PVC induced VF	N/A	A, L
21	ICM	35	No	14	No	MMVT	1	A, L

^*ICM: ischemic cardiomyopathy, NICM: non-ischemic cardiomyopathy

^†A: amiodarone, L: lidocaine, P: procainamide

Table 3.

Mechanical support and procedural details

	n (%)
Mechanical Support
IABP	14 (67%)
Impella CP	2 (9.5%)
ECMO	2 (9.5%)
ECMO and IABP	2 (9.5%)
ECMO and Impella CP	1 (5%)
Arrhythmia
Ventricular tachycardia	15 (71%)
PVC induced VT/VF	6 (29%)
Number of any VT induced per patient (clinical and non-clinical)	1.5 [IQR 1–2]
VT cycle length (clinical and nonclinical VT) (ms)	460 ± 86
Patients with inducible clinical VT	13 (62%)
Cycle length of clinical VT (ms)	475 ± 77
Type of VT mapping (n=13)
Full activation	8 (62%)
Entrainment	2 (15%)
Partial activation	3 (23%)
Scar modification (n=19 patients with scar)	15 (79%)

Attempts at holding antiarrhythmic immediately prior to the ablation procedure were considered whenever possible. In 12 patients (57%), it was feasible to hold at least one antiarrhythmic prior to the procedure, whereas this was not possible due to the burden of arrhythmia in the remaining 9 patients. Among the 14 patients treated with amiodarone and lidocaine, amiodarone alone was held in 3 patients (21%) prior to the procedure, lidocaine alone was held in 3 patients (21%), both medications were held in 3 patients (21%), and none of the medications were held in the remaining 5 (36%). As for the patient on amiodarone and procainamide, only procainamide was held prior to the procedure. Among patients who were on 1 antiarrhythmic medication (n=7), it was possible to hold AADs in 3 of them (1 off amiodarone, 1 off lidocaine, 1 off procainamide).

The type of mechanical support in the intensive care unit, and prior to presenting for the procedure, consisted of the following: 14 with IABP, 2 with Impella CP (AbioMed, 4.3 L/min), 2 with ECMO, 2 with ECMO and IABP, and 1 with ECMO and Impella CP. In addition, 9 patients (43%) were on a nitroprusside infusion prior to the procedure for afterload reduction, and 1 patient (5%) was on milrinone for ionotropic support. All but two patients had the procedure done under general anesthesia. In 15 patients (71%), general anesthesia was induced prior to the procedure in the respective intensive care units, in an attempt to quite the ventricular arrhythmia. No other modalities of mechanical support were added during the procedure in any of the patients.

Procedural details

All but 2 patients had endocardial scars by electroanatomic mapping (90%). (Figure. 1) CARTO mapping system was used in all patients except 1 patient where ESI/NAVX was used. All cases were done with irrigated catheters. In the 17 patients with ICM, 15 had endocardial scars, and these were localized to the anterior wall in 5 patients (33%), with one patient having an extensive area of scarring which extended from the anteroseptum to the anterolateral wall; 5 patients (33%) had a septal scar, 1 patient (7%) had an apical scar, 1 patient (7%) had a scar localized to the inferior wall, 1 patient had scars in the lateral and anterior wall, 1 patient in the inferior and lateral wall, and 1 patient in the apex and septum. In the 3 patients with NICM, 1 patient had a scar involving the RV apex, 1 patient had an LV apical scar, and 1 patient had an anterior scar. Finally, the patient with mixed ICM and NICM, scar was localized to the inferior/inferolateral region.

Figure 1.

Distribution of myocardial scar in 19 patients

Figure 2 shows a sample of voltage and VT mapping. The ablation procedure targeted the inducible clinical VTs in 13 patients (62%), and PVCs which had induced VT/VF in 6 patients (29%); giving the operators the opportunity to target the clinical arrhythmias in 91% of patients (n=19), while the clinical VT was non-inducible in the remaining 2 patients (9%, in one of them no VTs could be induced while non-clinical VT was induced in the other). In patients with inducible clinical VTs, only 1 clinical VT was present per patient with an average cycle length of 475 ± 77 ms. The total number of VTs induced in this population was 34, with a median of 1.5 VTs [IQR 1–2] per patient with inducible VT.

Figure 2.

Voltage and VT mapping showing inferior wall myocardial scar (panel A), and critical isthmus and VT circuit (panels B and C), respectively

Activation mapping was performed in all patients with inducible clinical VTs (n=13), and in patients with PVC induced VT/VF (n=6). Out of the 13 patients with inducible clinical VTs, full activation mapping was possible in 10 (77%) patients, while only partial activation mapping was performed in the remaining 3 patients (23%) due to worsening hemodynamics despite mechanical support (one patient on ECMO and IABP, and two patients on IABP). In addition, entrainment mapping was performed in 2 patients with full activation maps. As for patients without inducible clinical VT (n=2), a strategy of pace-mapping and extensive scar modification was followed.

Out of the 13 patients in whom ablation of the clinical VT was pursued, the VT was localized to the septal area in 6 (46%), to the lateral wall in 4 (31%), mitral annulus in 1 (8%), inferior wall in 1 (8%), and anterior wall in 1 (8%). These clinical VTs originated from a scar border zone in 7 patients (54%), and from within a scar in 4 patients (31%). In the remaining 2 (15%), the VT did not originate from a region of endocardial scar (1 patient with NICM, and 1 patient with ICM).

In addition, scar modification aiming at scar homogenization was performed in 15 of 19 patients with scar (79%). (Table 3) The remaining 4 patients had only partial scar modification near the clinical arrhythmia sites of origin.

Of note, 2 patients in this series underwent epicardial access and ablation. One of those patients had IABP for mechanical support which was exchanged for an Impella prior to the procedure (patient #9 in table 2). Epicardial access was obtained before placement of the Impella, and subsequently anticoagulation was initiated. The second patient had IABP for mechanical support (patient #14).

Acute procedural outcomes

Following targeted ablation and scar modification, the clinical arrhythmias were acutely confirmed to be successfully ablated in 15 patients (71%, 9 patients with non-inducibility of clinical VT after ablation, and 6 patients with PVC induced VT/VF). In 4 patients (19%) VT re-induction was not attempted after ablation due to multiple hemodynamically unstable VTs and extensive scar homogenization.

In the remaining 2 patients (10%), the clinical VT continued to be inducible following ablation. One of these patients was on ECMO, and despite mapping and extensive ablation of the clinical VT endocardially, it continued to be inducible and required several shocks to terminate. Several “steam pops” occurred during the procedure and it was decided to not pursue further ablation and re-induction. Given his recent history of open heart surgery (coronary artery bypass grafting and mitral valve replacement), and the fact that an epicardial source of the VT was suspected, the patient had an open epicardial ablation by cardiothoracic surgery 2 days after the index ablation. He was found to have an extensive scar over the lateral wall, and cryoablation was applied through the scar from the apex to the mitral annulus. Following this ablation, VT was no longer inducible. The second patient was supported with an IABP, and a total of 8 VTs, including the clinical VT, were induced. Some were hemodynamically stable, while others were not. Following extensive scar modification, unstable VTs continued to be induced and therefore the procedure was terminated.

In addition to the clinical VTs, non-clinical VTs were targeted for ablation when possible. This involved successful mapping and ablation of 5 non-clinical VTs in 3 patients.

Post-ablation outcomes

Out of the 21 patients, 17 (81%) were successfully weaned off mechanical support after 2 ± 1.4 days following VT ablation, and 15 patients (71%) were discharged alive from the hospital after a median of 20 days [IQR 10.5–26] following the ablation procedure. Six patients (29%) died during the index admission following ablation. Four of these patients died while still on mechanical support. Baseline and procedural characteristics of these patients are summarized in table 4 (note that patients’ numbers are the same from Table 2). The first patient (patient #5) was on ECMO and IABP. The ECMO was weaned off 3 days after the ablation, but the patient continued to require support from the IABP secondary to persistent cardiogenic shock and multi-organ failure, and finally succumbed 8 days after the ablation. The second patient (patient #9) was supported with an Impella CP. The Impella CP was removed at the end of the procedure, but the patient decompensated a few hours later requiring insertion of an IABP. He died on the same evening of the ablation from cardiogenic shock and cardiac arrest secondary to pulseless electrical activity. The third patient (patient #12) was supported with ECMO and Impella CP. This patient had suffered a motor vehicle accident with resultant traumatic left anterior descending artery dissection and ischemic cardiomyopathy. He had a PVC ablation for PVC induced VF, and after the ablation he continued to require mechanical hemodynamic support for persistent cardiogenic shock and multi-organ failure, and died after 11 days after the ablation. The fourth patient (patient #15) had a PVC ablation done with ECMO support. Prior to the ablation she was on ECMO and IABP. ECMO was decanulated a few days later, and the IABP was kept. She remained in cardiogenic shock and had recurrent PVC induced VF. Her course was complicated by multi-organ failure and lower extremity ischemia, and she eventually passed away 10 days after the ablation.

Table 4.

Baseline and procedural characteristics of patients with in-hospital mortality

	Gender	Age (years)	EF (%)	PAINESD	Substrate	Number of VTs induced	Clinical VT re-inducibility	Cause of death*
5	Male	71	20	18	Ischemic	1	Non-inducible	CS
9	Male	39	18	9	Non-ischemic	7	Not attempted	CS, PEA
12	Male	47	15	23	Ischemic	PVC induced VF	PVC induced VF	CS, MOF
15	Female	67	20	14	Ischemic	PVC induced VF	PVC induced VF	CS, PMVT, MOF
17	Male	77	9	23	Ischemic	1	Non-inducible	CS
18	Female	73	27	25	Ischemic	4	Not attempted	Septic shock

^*CS: cardiogenic shock; PEA: Pulseless electrical activity; MOF: Multi-organ failure; PMVT: Polymorphic VT

The remaining 2 patients (patient #17 and patient #18) were weaned off mechanical support but died later during the course of the index admission. Patient #17 was on IABP support. He was weaned off the IABP 1 day after the ablation but later during the day decompensated again and went back into cardiogenic shock. The patient declined further invasive treatment measures, and died on the same day from persistent cardiogenic shock. As for patient #18, she was on ECMO and IABP. She was successfully weaned off mechanical support 6 days after the ablation, but subsequently developed sepsis and septic shock and died 5 days later.

The median duration in cardiogenic shock was 5 days [IQR 2–9 days] prior to the ablation, and 1 day [IQR 1–1 days] after the ablation in patients who survived till discharge. In addition, the median ICU length of stay was 8 days [IQR 5–12.5 days] following the ablation in patients who were discharged alive. Antiarrhythmic therapy for the 15 patients who were discharged was as follows: 9 were discharged on amiodarone (60%), 1 (7%) on amiodarone and mexiletine, and 5 (33%) were not discharged on any antiarrhythmic medications. In addition, 9 (60%) were discharged on either an angiotensin converting enzyme inhibitor or an angiotensin receptor blocker (7 on Lisinopril, 1 on enalapril, and 1 on losartan), 13 (87%) were discharged on a beta-blocker (2 on metoprolol succinate, 10 on carvedilol, and 1 on metoprolol tartrate), 3 (20%) were discharged on hydralazine (2 of those in combination with isosorbide dinitrate), and 4 (27%) on spironolactone. To note that, one of these 15 patients (a patient with NICM) received a heart transplant during the index admission, and therefore was not discharged on any heart failure or antiarrhythmic medications.

Out of the 15 patients who were discharged, 1 patient (7%) had received a heart transplant during the index admission, and 5 patients (33%) received left-ventricular assist device implantation within 3 months of discharge. Upon one year of follow-up after the ablation procedures, only one of the 15 patients had recurrent ventricular arrhythmias requiring defibrillator shocks, and this occurred 96 days after discharge. The recurrent VT was different from the clinical VT targeted during the ablation procedure. Given that his VT occurred in the setting of heart failure decompensation, occurrence of multiple VT morphologies, and the fact that the patient had a history of amiodarone-related pulmonary fibrosis, it was decided to manage him conservatively with optimization of his heart failure regimen without repeating a VT ablation procedure. This patient passed away 362 days after the ablation procedure from sudden cardiac death. In addition, another patient passed away more than 6 months after the date of the ablation. The cause of death is unknown. All remaining 13 patients survived through 1 year.

Differences between patients discharged alive versus those who died during the index admission

Table 5 summarizes the baseline, procedural, and clinical differences between patients discharged alive from the hospital, versus those who died during the index admission. Patients alive at discharge were more likely to be male (93% vs 67%), and had higher LVEF (30% vs 19%). There was no difference in the proportion of ischemic and non-ischemic cardiomyopathy between the groups. PAINESD score in patients alive at discharge was 17 [IQR 14–20], versus 20.5 [IQR 15–23] in patients who died during the index admission. As for the duration in cardiogenic shock, the deceased group had a longer duration of cardiogenic shock prior to the ablation (9 days vs 4 days), and after the ablation (9 days vs 3 days). In addition, the total duration in cardiogenic shock was longer in the group who died during the index admission (13 vs 8 days).

Table 5.

Baseline, procedural, and clinical differences between patients deceased versus alive at the time of discharge

	Died during index hospitalization (n=6)	Alive at discharge (n=15)
Male gender	4 (67%)	14 (93%)
Age	69 [52–72.5]	67 [49.5–68]
LVEF	19 [15.75–20]	30 [17–35]
PAINESD score	20.5 [15–23]	17 [14–20]
Clinical VT non-inducible*	2/4 (50%)	6 (60%)
ICM	5 (83%)	13 (87%)
NICM	1 (17%)	3 (20%)
Prior VT ablation	2 (33%)	5 (33%)
Number of home AADs	1 [0.25–1]	0 [0–1]
Duration in cardiogenic shock prior to procedure (days)	9 [5.25–12.75]	4 [1.5–6.5]
Duration in cardiogenic shock after to procedure (days)	9 [3.5–10.75]	3 [1–4.5]
Total duration in cardiogenic shock (days)	13 [11.25–21.5]	8 [4.5–9.5]

^*In patients with baseline inducible ventricular tachycardia

Differences between patients who presented with cardiogenic versus those who presented with VT and cardiogenic shock

Results are summarized in Table 6. Ages were comparable between groups. Patients who primarily presented with cardiogenic shock only as the primary initial indication for mechanical support had lower EF (18% [IQR 13%−27%] vs 31% [20%−35%], p=0.03), and were more likely to require more advanced mechanical support than IABP alone (6(55%) vs 1(10%)), when compared to patients who presented with both VT and cardiogenic shock. However, both groups had comparable PAINESD scores. In-hospital mortality was 45% in the former group, compared to 10% in the latter group, p=0.06.

Table 6.

Differences between patients who presented with cardiogenic versus those who presented with VT and cardiogenic shock

	Cardiogenic shock (n=11)	VT and cardiogenic shock (n=10)
Age in years (median [IQR])	67 [49–72]	67 [52–68]
EF (%) (median [IQR])	18 [13–27]	31 [20–35]	p=0.03
Level of mechanical support
IABP alone (n (%))	5 (45)	9 (90)
Higher level of support* (n (%))	6 (55)	1 (10)
PAINESD score (mean ± std)	18 ± 6	17 ± 4	p=0.6
In-hospital mortality (n (%))	5 (45)	1 (10)	p=0.06
Post-discharge 1-year mortality (n (%))	0	2 (22)

^*Higher level of support that includes Impella or ECMO

Discussion

The current study reports on the procedural and the clinical outcomes of “bailout” ventricular arrhythmia ablation in a critically ill population with cardiogenic shock and inability to wean off mechanical support due to refractory ventricular arrhythmias despite AADs. To our knowledge this is the first study investigating this topic in this particular high-risk population, reflected by the clinical status of the patients, and the elevated PAINESD scores. ¹⁵ The study demonstrates that catheter mapping and ablation are feasible in this group of patients, with good acute procedural outcomes including successful ablation of the clinical arrhythmias in the majority of patients. This allowed liberation from mechanical hemodynamic support shortly after the ablation in patients who were otherwise dependent on it. However, despite the favorable acute outcomes of the procedure, an elevated morbidity and mortality rate continued to be observed in this critically ill population, and was related mostly to worsening cardiogenic shock. The observed in-hospital mortality rate was 30%, and the 1-year mortality was about 40%, while a large portion of survivors required advanced heart failure therapies. The findings confirm that “bailout” VT ablation is a viable option in this population allowing weaning off mechanical support, and survival to discharge in the majority of patients who would otherwise die or require cardiac transplantation.

Despite the hemodynamic support used in this series, many VTs were poorly tolerated hemodynamically which could reflect a drop in right ventricular output, and/or sub-optimal hemodynamic left ventricular support relative to the VT cycle lengths and severity of LV dysfunction. Nevertheless, the presence of hemodynamic support at the time of ablation may have improved the acute procedural outcomes as suggested by prior studies performing VT ablation with mechanical support. ^16–19

The clinical management of cardiogenic shock remains challenging, with high morbidity and mortality. This holds true, irrespective of whether cardiogenic shock is secondary to acute myocardial infarction, decompensation of chronic heart failure, or refractory ventricular arrhythmia. ^{12, 20–24}

The observed elevated morbidity and mortality in the study population reflect the underlying critical status of these patients. This is further supported by the observation of worsening cardiogenic shock as the predominant cause of death. In fact, patients who died during the index hospitalization were more likely to have a lower EF, and were in cardiogenic shock for a longer period of time before and after the ablation. However, it remains unclear if deaths were related to longer times in cardiogenic shock prior to ablations, and whether or not sooner interventions would have allowed better rates of survival to discharge.

Furthermore, results showed that patients who initially presented with cardiogenic shock requiring hemodynamic support and subsequently developed VT, had a higher in-hospital mortality (although not statistically significant, possibly due to lack of statistical power), than patients who required mechanical support for concomitant VT and cardiogenic shock on presentation. Although the retrospective nature of this study does not permit us to confirm whether or not VT occurred before cardiogenic shock in the latter subgroup, these results can be explained by a worse cardiac substrate in the former group.

The arguments in some cases against very early ablation would be to optimize hemodynamics with hope to reduce the burden of arrhythmia. Further hemodynamic decompensation could be observed at the time of VT ablation procedures even in patients who were otherwise hemodynamically stable. In a recent study, ²⁵ the clinical outcomes were investigated in patients with electrical storm who develop acute periprocedural hemodynamic decompensation around the time of VT ablation procedures. The study showed very good acute success rates, but with 62% in-hospital mortality. Although our patient population is different from that described in the aforementioned study, particularly because our patients had VT ablation as a last resort for liberation from mechanical hemodynamic support, the current findings confirm the prior observations of poor outcomes when ventricular arrhythmias and cardiogenic shock coexist. However, when offered as a “bailout” intervention, an ablation was still associated with survival to discharge in most patients.

Although the study was derived from a prospective registry with prospectively adjudicated procedural and complications data, the retrospective nature of data collection is one of the major limitations. Some patients who were discharged alive, may have had recurrence of VT below the detection limit of their ICD, however, data on ICD settings were not available for all patients. In addition, the retrospective nature of the study did not allow us to identify the temporal relationship between VT and cardiogenic shock in patients who had both ventricular tachycardia and cardiogenic shock on presentation.

Given the retrospective nature of the study we cannot completely rule out the possibility that there were patients with VT and cardiogenic shock requiring mechanical support, who did not undergo ablation for their arrhythmias secondary to the severity of their illness. However, all the patients in the study group were severely ill, with advanced heart failure, and had VT ablation only as a last resort. Clinical decision making is difficult in these settings and need to account for co-morbid conditions and baseline neurological and functional status on a case by case basis.

The population reflects experience from a tertiary care center with highly experienced practitioners in intensive care, VT ablation and cardiac surgery. As such, the results may not be generalizable to smaller or non-tertiary care settings.

Finally, the small population size may have underpowered the study, and may have contributed to the statistically nonsignificant results when comparing in-hospital mortality between patients who presented in cardiogenic shock alone, versus those who presented with both VT and cardiogenic shock.

Conclusion

In patients with refractory ventricular arrhythmias precluding weaning from hemodynamic support in the setting of cardiogenic shock, “bailout” ablation procedures represent a viable option with very good acute procedural success rate; allowing liberation from mechanical support in most patients.

What is Known

• In patients with cardiogenic shock, and concomitant refractory ventricular arrhythmia preventing discontinuation of mechanical hemodynamic support, catheter ablation has favorable acute procedural outcomes, and facilitates weaning off mechanical support.

• Long-term outcomes for this patient population continues to be poor.

What the Study Adds

• Catheter ablation of ventricular arrhythmia in patient with cardiogenic shock on mechanical hemodynamic support, and refractory ventricular arrhythmia can be used to allow for liberation off mechanical support, and bridge to advanced heart failure therapy

Nonstandard Abbreviations and Acronyms

VT

Ventricular Tachycardia

VF

Ventricular Fibrillation

IABP

Intra-aortic balloon pump

ECMO

Extra-corporeal membrane oxygenation

RV

Right ventricle

LV

Left ventricle

LVEF

Left ventricular ejection fraction

PVC

Premature ventricular contraction

NICM

Non-ischemic cardiomyopathy

ICM

Ischemic cardiomyopathy

AADs

Antiarrhythmic drugs