Catheter Ablation of Refractory Ventricular Fibrillation Storm After Myocardial Infarction

Introduction

Myocardial electric storm represents a life-threatening, malignant condition of clustering ventricular arrhythmias that necessitate multiple defibrillations. Ventricular fibrillation (VF) storm attributed to focally triggered VF after myocardial infarction (MI) could be recognized as a distinctive arrhythmic syndrome with specific pathophysiology and lethal characteristics that differs from scar-mediated monomorphic ventricular tachycardia (VT).¹ The initial approach to patients presenting with VF storm is usually focused on exclusion of myocardial ischemia and administration of antiarrhythmic medications. However, antiarrhythmic drugs may fail to suppress VF or may be limited by contraindications or adverse effects (eg, long-QT interval, hyperthyroidism, or labile hemodynamic compensation and severe dysfunction of the left ventricle).^1,2 The condition often requires other therapies combined with antiarrhythmic medications, including deep sedation with mechanical ventilatory support, overdrive pacing, or hemodynamic support devices to stabilize patients. Neuraxial modulation with thoracic epidural anesthesia also may be effective for electric storm management.^3,4 However, despite these extensive therapeutic efforts, suppression of lethal arrhythmias can be transient.

Over the past 2 decades, radiofrequency catheter ablation has emerged as a potentially effective treatment strategy for postinfarct VF storm. The current guideline endorses catheter ablation for the management of therapy-resistant ischemic VF storm.⁵ The main targets for ablation are identifiable focal triggers of VF. The pathogenic role of the Purkinje system is very important; discrete excitations act as main triggers of VF in patients with diseased hearts.⁶ The role of Purkinje fibers in VF maintenance is less documented than its role in VF initiation. Nevertheless, several experimental studies have revealed Purkinje in fibrillatory wave fronts during VF,⁷ and computer models showed migratory reentrant activity at the Purkinje-muscle junction as a mechanism for maintaining early VF.^8,9 Previous reports demonstrated a short-term benefit of radiofrequency catheter ablation targeting specific Purkinje potentials that precede the triggers of VF.^10–13 However, these reports were single-center series evaluating a small number of patients. The impact of VF storm ablation on patient survival has not been verified in a large population. Therefore, we sought to investigate the short- and long-term outcomes of catheter ablation for the treatment of last resort in a large series of consecutive patients with post-MI VF storm refractory to medical therapies.

Methods

The data, analytical methods, and study materials will not be made available to other researchers for purposes of reproducing the results or replicating the procedures.

Study Population

Between January 2003 and May 2017, a total of 110 consecutive patients with VF storm after MI underwent radiofrequency catheter ablation as a last resort for refractory VF storm in 15 international tertiary care arrhythmia centers. VF storm was defined as ≥3 separated episodes of VF in 24 hours. Only patients with VF without preceding monomorphic VT were included. VF storm was documented on bedside or telemetry monitors, device interrogation, or Holter recordings. The diagnosis of MI was based on medical history, clinical findings, and coronary angiography. All patients were brought to the intensive care unit and were managed by a combination of therapies, including coronary revascularization, antiarrhythmic medications, correction of electrolytes, deep sedation with mechanical ventilatory support, overdrive pacing, or circulatory support devices (extracorporeal membrane oxygenation and intra-aortic balloon pumping) at the discretion of the dedicated unit physicians. When life-threatening conditions resulting from VF storm remained despite optimal medical therapies, the decision to transfer patients to the electrophysiological laboratory for catheter ablation was made by the dedicated cardiac team. All patients or families provided written informed consent for the ablation procedures. The institutional review boards of the participating centers approved the collection of data. This study included 29 patients who had been included in previous studies.^13–15

Mapping and Ablation Strategy

The number and type of catheters varied in individual centers and individual patients. Multipolar catheters were positioned from the femoral veins into the right ventricle and at the His bundle. Mapping and ablation of the left ventricle were performed by the transaortic (retrograde) or transseptal approach. Three-dimensional electroanatomic mapping was performed with CARTO (Biosense Webster, Diamond Bar, CA) or NavX (St. Jude Medical, St. Paul, MN). A nonirrigated-tip catheter or an irrigation catheter was used for mapping and ablation. Systemic heparinization was maintained during mapping and ablation. General anesthesia with mechanical ventilatory support was used as part of the previous therapy or periprocedurally at the discretion of the operator.

The targets of ablation were the ventricular premature beats (VPBs) triggering VF. Purkinje potentials that preceded the onset of the triggering VPBs were frequently found around the scar border zone and were carefully identified.^11–13 In addition, pace mapping was performed to localize ablation targets. In these respects, it was essential to document the triggering VPBs on a 12-lead ECG before the ablation procedure. When no spontaneous VPBs were detected during the ablation procedure, they were induced by pacing maneuvers (ventricular burst or extrastimuli) with the addition of an intravenous isoproterenol (3.0–5.0 μg/min) or epinephrine (5–10 μg) infusion. Radiofrequency energy was delivered with a maximum power of 40 W and a target temperature of 42°C for the irrigation catheter and 55°C for the nonirrigated catheters. The inducibility of VF and VT was tested at the end of procedure at the operator’s discretion. When attempted, pacing was performed from the right ventricular apex with up to triple extrastimuli decreasing down to 200 ms or to refractoriness, whichever occurred first. The procedural end point was the elimination of all clinical triggering VPBs.

In-Hospital and Long-Term Follow-Up

After the ablation procedure, patients were brought back to the intensive care unit and monitored. Intensive management to optimize the overall conditions of patients was continued after the ablation procedure. Postprocedural medical treatment included β-blockers and antiarrhythmic drugs, anticoagulation, revascularization, and management of heart failure with ventilatory support, intravenous catecholamine, hemodiafiltration, and circulatory support devices as necessary. Patients who had no previous implantable cardioverter defibrillator (ICD) were implanted with the device before hospital discharge, except for 5 patients who refused implantation. ICD programming was commonly set to include 3 zones: a VT zone (150–188 bpm; antitachycardia pacing attempts followed by shock), a fast VT zone (188–210 bpm; 1 attempt of antitachycardia pacing and subsequently shock), and a VF zone (>210 bpm; shock only, with antitachycardia pacing during charging, if available). Drug management and device programming during follow-up were at the discretion of the investigators. After hospital discharge, all recruiting centers were contacted for each patient. In addition to records of clinical events, ECG, and Holter monitoring, VF and associated intracardiac electrograms were documented by ICD interrogation logs.

Statistical Analysis

Categorical variables were expressed as numbers and percentages and were compared by the χ² or Fisher exact test as appropriate. Continuous data were expressed as mean±SD or as median (interquartile range [IQR]). Recurrent VF/VT, all-cause mortality, and cardiovascular and noncardiovascular death were assessed in the hospital and at the end of the follow-up period. Univariate logistic regression analysis was used to evaluate clinical and procedural variables associated with in-hospital mortality. Odds ratios with corresponding 95% CIs and 2-sided P values were presented. Univariate Cox proportional hazards analysis was used to assess the association of clinical and procedural variables on all-cause mortality after hospital discharge, from which hazard ratios (HRs) and 95% CIs were derived. Cumulative incidence curves were plotted for recurrent VF/VT, all-cause mortality, cardiovascular deaths, noncardiovascular deaths, and death resulting from unknown causes. All tests were 2-tailed. A value of P<0.05 was considered statistically significant. Statistical calculations were performed with IBM SPSS Statistics 21.0.

Results

Patient Characteristics

Patient characteristics are summarized in Table 1. The mean age was 65±11 years, and 92 patients (84%) were men. At the time of VF storm occurrence, the type of MI was acute in 43 patients (39%), subacute (>1 week) in 48 (44%), and remote (>6 months) in 19 (17%). In the patients with acute MI, VF storm started 4.5±2.5 days (median, 5 days [IQR, 3–5 days]) after the onset of an infarction during the index hospitalization for MI. In the patients with subacute MI, VF storm started median 12 days (IQR, 10–20 days) after the onset of MI (during the index MI hospitalization in 43 patients [90%] and after hospital discharge in 5 patients [10%]). In the patients with a remote MI, the median interval from the onset of an infarction to VF storm occurrence was 4.2 years (IQR, 1.6–12.5 years). The mean left ventricular ejection fraction was 31±10%, and 37 patients (34%) had severe heart failure symptoms before admission (New York Heart Association class III or greater). Forty-one patients (37%) presented with cardiogenic shock caused by heart failure on admission. Amiodarone, class I antiarrhythmic drugs (lidocaine or mexiletine), and β-blockers were administered to 103 (94%), 71 (65%), and 91 (83%) patients, respectively. However, they all failed to prevent VF storm. Periprocedural hemodynamic support and deep sedation with mechanical ventilator were used in 45 (41%) and 81 (74%) patients, respectively. Catheter ablation of refractory VF storm was performed 6±6 days (median, 4 days [IQR, 2–8 days]) after the first occurrence of VF storm.

Eight patients underwent catheter ablation before coronary revascularization for the following reasons: 6 patients had remote-phase MI with chronic total occlusion and presented with VF storm without the evidence of acute infarction, and the remaining 2 patients with subacute MI had diffuse multivessel stenosis after coronary artery bypass grafting (Table I in the online-only Data Supplement).

Short-Term Results of Ablation

Ninety-four patients (86%) presented to the laboratory with frequent, spontaneous VPBs. In the remaining 16 patients (14%), pacing maneuvers with the administration of intravenous isoproterenol were performed to induce VPBs. The site of origin of the triggering VPBs was associated with the territory of infarction in all patients. The triggering VPBs were found to originate from the surviving Purkinje tissue in the dense scar area (a bipolar voltage <0.5 mV) in 15 patients (14%) and from the scar border zone (a bipolar voltage of 0.5–1.5 mV) in 88 patients (80%; Figure 1). Although VPBs were found to originate from the normal voltage area (a bipolar voltage >1.5mV) in the remaining 7 patients (6%), these sites also correlated with the territory of infarction. The site of origin of the triggering VPBs was the left ventricular septum in 78 patients (71%), papillary muscles in 10 (9%), both the left ventricular septum and papillary muscles in 17 (15%), and other scar border areas in 5 (5%). Purkinje potentials preceding the QRS complex during both sinus rhythm and VPBs were recorded at the ablated regions in 99 patients (90%). Ablation of Purkinje potentials preceding VPBs in addition to ablation guided by pace mapping eliminated triggering VPBs in 100 patients (91%). The total radio frequency energy delivery time was 22±12 minutes. The mean procedure duration was 178±66 minutes. Programmed stimulation at the end of the procedure was not performed to prevent deterioration of the hemodynamic status in 53 patients (48%). Among 57 patients who underwent programmed stimulation at the end of the procedure, noninducibility of VF and VT was achieved in 46 patients (81%). VF remained inducible in 11 patients (19%).

• Download figure
• Download PowerPoint

During the procedure, 1 patient died of electromechanical dissociation. None of the patients developed pericardial tamponade. Aggravation of heart decompensation occurred in 4 patients (4%). Cerebrovascular ischemic attacks occurred in 1 patient (1%). Complete atrioventricular block occurred in 1 patient (1%), and transient atrioventricular block that was not observed after the ablation procedure occurred in 3 patients (3%). Left bundle-branch block occurred in 8 patients (7%). Groin hematoma occurred in 1 patient (1%) who required blood transfusion and surgery.

In-Hospital Outcomes

A flow diagram of in-hospital outcomes after ablation is shown in Figure 2. After the first ablation procedure, VF storm subsided in 92 patients (84%). However, despite ablation, VF storm was not suppressed in the remaining 18 patients (16%). Of those patients with uncontrollable VF storm, 8 underwent a second ablation procedure. Among 92 patients in whom VF storm subsided after ablation, 24 patients developed single isolated episodes of recurrent VF or VT that were managed by a second ablation in 12 patients and by antiarrhythmic medications in 12 patients. Sixty-eight patients (62%) were free of any recurrent ventricular arrhythmias. The short-term impact of ablation on VF/VT frequency is shown in Figure 3. Ablation of VF storm resulted in a substantial reduction of the lethal ventricular arrhythmia burden.

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

Overall, there were 30 in-hospital deaths (27%) with a median survival time of 7 days (IQR, 2–17 days). Table 2 shows the clinical and procedural variables associated with in-hospital death. Of note, the duration between the occurrence of VF storm and the ablation procedure was significantly associated with in-hospital death (for each 1-day increase: odds ratio, 1.11 [95% CI, 1.03–1.20]; P=0.008). The number of in-hospital death increased with the time from VF storm occurrence to the ablation procedure (Figure 4): 1 of 25 patients (4%) in those who underwent ablation on the day of VF storm occurrence and those with ablation the day after the VF storm, 8 of 31 patients (26%) with ablation 2 to 4 days after VF storm, 7 of 22 patients (32%) with ablation 5 to 7 days after VF storm, and 14 of 32 patients (44%) with ablation ≥8 days after VF storm.

• Download figure
• Download PowerPoint

Acute recurrence of VF storm was strongly associated with subsequent in-hospital death (odds ratio, 11.47 [95% CI, 3.60–36.52]; P<0.001). However, the incidence of single isolated episodes of recurrent VF or VT was not associated with in-hospital death (odds ratio, 1.13 [95% CI, 0.41–3.07]; P=0.81).

Long-Term Outcomes

Long-term outcomes were available in 80 patients (73%) who were discharged from the hospital alive. All except 5 patients without a previously implanted ICD underwent device implantation before hospital discharge. Five patients refused implantation. Medications after hospital discharge included amiodarone in 46 patients (58%), class I antiarrhythmic agents in 10 (13%), β-blocker in 76 (95%), and oral anticoagulants in 19 (24%). The cumulative incidences of short- and long-term recurrent electric storm, isolated VF, and isolated VT are shown in Figure 5A. Of the patients who survived a VF storm after ablation and were discharged from the hospital alive, only 1 patient developed recurrent VF storm after 22 months from the index ablation procedure. Over a median follow-up of 3.7 years (IQR, 1.4–6.5 years), an isolated episode of recurrent VF occurred in 4 patients (5%), whereas an isolated episode of VT was documented in 9 patients (11%), and both VF and VT were seen in 1 patient. A median recurrence survival time was 2.2 years (IQR, 0.2–3.0 years).

• Download figure
• Download PowerPoint

In total, 29 patients (36%) died during the follow-up after discharge from the hospital with a median survival time of 2.2 years (IQR, 1.2–5.5 years). The causes of death are summarized in Table II in the online-only Data Supplement. Although short-term mortality was mainly from cardiovascular causes (mostly refractory heart failure), long-term mortality was from both cardiovascular (13 patients) and noncardiovascular (11 patients) causes. Noncardiovascular causes included sepsis, cancer, and stroke. The cumulative incidences of overall mortality, cardiovascular-related deaths, and noncardiovascular-related deaths after the index ablation procedure are shown in Figure 5B. The cumulative incidence curve of all-cause mortality had a steep slope at the beginning, mainly because of cardiovascular death, but there was a steady increase of both noncardiovascular death and cardiovascular death over the long-term follow-up.

In Cox regression analysis (Table 3), mortality during follow-up after discharge from the hospital was associated with left ventricular ejection fraction <30% (HR, 2.54 [95% CI, 1.21–5.32]; P=0.014), New York Heart Association class ≥III (HR, 2.68 [95% CI, 1.16–6.19]; P=0.021), a history of atrial fibrillation (HR, 3.89 [95% CI, 1.42–10.67]; P=0.008), and chronic kidney disease (HR, 2.74 [95% CI, 1.15–6.49]; P=0.023). None of the periprocedural therapies and intraprocedural outcomes were associated with mortality over the follow-up period. The use of a β-blocker was associated with improved survival (HR, 0.29 [95% CI, 0.085–0.97]; P=0.044).

At least 1 hospitalization for acute heart failure occurred in 17 patients (21%), which led to death in 9 patients. Over the follow-up period, 4 patients (5%) developed a stroke. Six patients (8%) patients underwent percutaneous coronary intervention for recurrent MI without VF storm.

Discussion

This multicenter study reviewed the largest consecutive series of patients undergoing catheter ablation for therapy-resistant VF storm after MI. The data indicated that catheter ablation targeting the focal Purkinje-related triggers frequently arising from the scar border zone at the left ventricular septum was effective in suppressing VF storm in most cases; however, despite ablation, uncontrollable VF storm occurred in 16% of patients. In-hospital death occurred in 27% of patients. Uncontrollable VF storm was associated with a high risk of subsequent death, and besides the severity of heart failure, time from the beginning of the storm to catheter ablation was associated with in-hospital mortality. When ablation resulted in short-term survival from a VF storm, patients developed only sporadic events of recurrent VF over the long follow-up period. However, a steady increase in mortality was noted during that period, which was associated with poorer clinical status, including left ventricular ejection fraction <30%, New York Heart Association class ≥III, and a history of atrial fibrillation and chronic kidney disease.

Timing of the Catheter Ablation Procedure

In clinical practice, ablation of VF storm is usually implemented as a last-resort strategy; however, the optimal timing of catheter ablation remains unknown. Although VF storm uncontrollable by ablation placed the patients at an extremely high risk of subsequent mortality, single isolated episodes of recurrent VF were not associated with either in-hospital death or long-term mortality. On the other hand, the delay between the onset of the storm and ablation was associated with poor short-term outcome. The reason in part is the prolonged exposure to electric storm, which can further compromise the cardiac function and result in worsened overall conditions of patients. In this regard, using catheter ablation for the management of VF storm (not for single isolated episodes of VF) in patients after revascularization and medical therapies should be considered earlier. Experience suggests that patients with VF storm should be transferred to a dedicated critical care unit that involves not only electrophysiologists who are experienced in VF/VT ablation but also specialists for managing heart failure, coronary interventionists, thoracic surgeons, and cardiac anesthesiologists to optimize the overall condition of these patients and to perform ablation as early as possible.

Strategy of Catheter Ablation

The primary goal of the current ablation strategy for focally triggered VF is to eliminate the Purkinje-related ectopic focus. Therefore, determining the earliest Purkinje potentials preceding the target ectopic beats is the key to successful ablation. Because the proarrhythmic and profibrillatory effects of β-adrenergic stimulation have been demonstrated in the setting of ischemic heart disease,^16,17 administration of isoproterenol or epinephrine to induce target VPBs may facilitate ablation, especially in the case of a paucity of VPBs during the procedure. However, this study demonstrated that the culprit Purkinje sources were most commonly distributed over the border zone of the ischemic scar at the left ventricular septum, which is in line with findings from previous clinical and experimental studies.^{10–13,18–21} This observation about the dominant domain of VF triggers may serve as a road map for ablation to target the culprit Purkinje network at the specific regions in sinus rhythm.

None of the ablation procedures performed except for 1 was interrupted because of complications related to hemodynamic intolerance and respiratory failure. However, atrioventricular block and left bundle-branch block occurred in ≈10% of the patients. It should be kept in mind that the triggering VPBs often originate from the Purkinje network at the left ventricular septum, where ablation carries a potential risk of damage to the left His-Purkinje system. On the other hand, catheter ablation of VF storm is usually a bail-out procedure, and risk of damage to the conduction system may not be a factor that influences the ablation strategy.

Inducibility of VF and VT was tested because of the following reasons. First, patients undergoing VF ablation may develop newly emergent monomorphic VT that often originates from the Purkinje network in the ischemic scar area close to the VF ablation sites.¹⁴ Second, ablation targeting Purkinje-related triggers might modify the VF substrate. It remains to be determined whether the favorable effect of ablation was the result of the elimination of VPBs triggering VF or modification of the VF substrate at the Purkinje network. Ablation targeting the Purkinje potentials in the vicinity of the ectopic focus may change the milieu for maintenance of VF. However, this could not be clarified in this retrospective observational study. Further investigations are needed to assess the prognostic significance of noninducibility after ischemic VF ablation.

Short- and Long-Term Outcomes

Previous studies on smaller patient cohorts demonstrated encouraging outcomes of VF storm ablation, wherein nearly all patients achieved elimination of the VF storm and were discharged from the hospital alive.^10–13 This study provided a realistic appreciation of ablation outcomes in VF storm. Patients presenting with VF storm often have multiple comorbidities, including heart failure decompensation, multiple coronary artery diseases, diabetes mellitus, atrial fibrillation, and renal dysfunction. Periprocedural management of heart failure decompensation and hemodynamic optimization with mechanical support when clinically indicated is of great importance, even in cases in which urgent catheter ablation successfully suppressed the electric storm.

After patients survived the acute phase of VF storm, most of them remained free from recurrent VF, as observed over the long follow-up period. This is in contrast to a prior report investigating outcomes of ablation for scar-mediated VT in which the VT recurrence remained steady over 3 years after ablation.²² This is probably related to different characteristics of arrhythmia mechanisms. Although patients with VF storm present with focal Purkinje arrhythmogenicity at the specific endocardial areas around the scar border, patients with VT often have multiple coexisting VT circuits surrounded by large scars at the endocardium, epicardium or deep within the myocardium, which may make durable suppression of VT recurrence difficult to achieve in the long term.

Previous studies have reported that the recurrence of VT after ablation of scar-mediated VT predicts a poor outcome in the short and long terms.^23–27 This most probably reflects progressive heart failure associated with a worsened hemodynamic state caused by irrepressible VT and negative inotropic consequences of ICD shocks for VT. Consistent with these previous studies on VT population, the early recurrence of VF storm in this study was strongly associated with subsequent in-hospital death. These findings support the notion that short-term suppression of VF storm by catheter ablation reduces cardiovascular mortality in the short term. However, a gradual increase in long-term mortality occurred despite a low rate of recurrent life-threatening arrhythmias. Although heart failure was the dominant cause of in-hospital death, long-term mortality resulted from both cardiovascular and noncardiovascular reasons and was not associated with the short-term outcomes of ablation. These results underscore the importance of careful management over a long follow-up period globally beyond the use of catheter ablation.

Study Limitations

This study was not randomized and hence was subject to the limitations inherent to observational studies. Although all patients in this study required multiple defibrillations for VF storm that was refractory to coronary revascularization and antiarrhythmic agents, the preprocedural therapies were not uniform. The referral for catheter ablation might be influenced by the severity of arrhythmia burden and the overall conditions of patients. However, it would be difficult to conduct large randomized trials in this specific patient population because of the lethal characteristics and the relative scarcity of cases with refractory VF storm after MI. Thus, nonrandomized observational studies with a large population could be a source of evidence.

In this study, no patients used left ventricular assist devices that could be of potential benefit to improve the hemodynamic state, which might affect the timing of the ablation procedure. Furthermore, none of the patients underwent heart transplantation. Because most of the patients in this study had end-stage intractable heart failure, heart transplantation may have improved their survival rates and their functional status. Urgent catheter ablation may be performed as a bridge to heart transplantation in these sickest patients.

We also acknowledge the limited clinical data available from this observational study, resulting from the potential substantial heterogeneity of procedures and techniques used during the ablation and management of patients after ablation. Long-term outcomes were available in only 80 patients (73%) who were discharged from the hospital alive. Furthermore, this study was not able to capture changes in ICD programming during follow-up. ICD reprogramming according to type of recurrent arrhythmias may influence the mortality outcome.²⁸

Conclusions

In patients with MI presenting with focally triggered VF storm refractory to medical therapies, catheter ablation targeting the culprit triggers is lifesaving and appears to be associated with short- and long-term freedom from recurrent VF storm. The triggering ectopic activities most commonly originate from the surviving Purkinje tissue over the border zone of ischemic scar in the left ventricular septum. Early intervention after the occurrence of VF storm may lead to a reduced risk of in-hospital cardiovascular mortality. However, the outcome over long-term follow-up is limited by steady increases in both cardiovascular and noncardiovascular mortalities. The poor long-term prognosis reflects primarily the severity of underlying cardiovascular disease and comorbidities in this specific patient population.

Acknowledgments

The authors gratefully acknowledge Iwanari Kawamura, MD, Takashi Kurita, MD, Meiso Hayashi, MD, Kenji Kurosaki, MD, and Satoshi Aita, MD, for their help and support in preparing the manuscript.

Footnotes