Heart failure (HF) is a growing medical problem in developed countries, with incidences among adult population reaching up to 2% [1]. It is well known that HF patients are simultaneously at a significantly higher risk of sudden cardiac death. The recommended and most efficient form of primary and secondary prevention against sudden cardiac death in heart failure patients is the implantation of automatic implantable cardioverter defibrillators (ICD) [2, 3]. 3.5% of patients with an ICD for primary prevention, and even 10–40% of those implanted for secondary prevention of SCD, will suffer from the most malignant form of ventricular arrhythmia, known as an electrical storm (ES) [4, 5]. An electrical storm is defined as the occurrence of three or more distinct episodes of ventricular tachycardia (VT) and/or ventricular fibrillation (VF) within 24 h [6]. In patients with ICD, its usual presentation involves multiple antitachycardia pacing therapies (ATP) and/or high-energy therapies within a short-time period. Management in this challenging group is generally focused on the elimination of potentially reversible causes of electrical instability, haemodynamic stabilisation and (once reversible causes have been excluded or dampened) modification of the arrhythmogenic substrate with catheter ablation of ventricular tachycardia [7‐9]. Most of the published data on radiofrequency catheter ablation (RFCA) for ES is based on small series or case reports, mainly describing single treatment method. No large studies have compared RFCA with conservative treatment and their effects on survival. The purpose of this study was to assess the long-term clinical outcomes after RFCA for ES and to compare this method with other forms of treatment for ES.

Seventy consecutive patients (91.4% male, age [64.71 ± 10.71] years; LVEF 26.2% ± 7.31%), the majority with ischaemic cardiomyopathy (81.43%) and ES, constituted our study population. As described in the “Methods” section, patients were divided into two groups according to execution of catheter ablation during the primary episode of ES. Among the groups, no significant difference in baseline clinical and electrocardiography characteristics were observed except BMI (group A: 28.66 ± 3.65 vs. group B: 26.76 ± 3.61; p = 0.04), myocardial infarction treated non-invasively (20 [71.4%] vs. 19 [45.2%]; p = 0.03) and duration of native QRS respectively (group A: 140.61 ± 30.42 ms vs. group B: 122.0 ± 23.88 ms; p = 0.04) (Table 1).

In group A, the majority of cases showed no clear cause of ES during primary hospitalisation (n = 24 [85.71%]), in contrast to group B, where nearly half of the patients had clear causes of ES (n = 22 [53.4%]). In group B, RFCA was abandoned in 20 patients during hospitalisation despite undetected and potentially reversible cause of electrical instability, due to the disappearance of arrhythmia as a result of the intensification of antiarrhythmic therapy (increased doses of β-blockers (20 patients) and to the inclusion of treatment with antiarrhythmic class III drugs (14 patients; 13 with amiodarone, 1 with sotalol) (Table 2). After the first ES, 28 patients underwent a catheter ablation procedure. Nine (32.1%) patients from group A and 13 (31%) from B group underwent subsequent ablation during the follow-up because of VT and ES recurrence (p = 0.92). Importantly, no statistical differences were observed in terms of VT (group A: 16 [57.1%]; group B: 31 [73.8%]; p = 0.15) and ES (group A: 9 [32.1%]; group B: 19 [45.2%]; p = 0.28) recurrence in both groups. However, there was significantly lower all-cause mortality in group A (4 [14.3%] vs. 16 [38.1%]; p = 0.03) (Table 3). Significant differences in survival between ablated and non-ablated groups are shown in Fig. 1. At the end of the ablation procedures, 23 patients underwent programmed electrical stimulation. In 2 of those patients, non-clinical VTs were inducible. There was no inducible clinical VT at the end of the procedure. Five patients (17.9%) were not tested due to haemodynamic instability after CA.

Analysis of the early post-operative period (< 30 days) revealed a non-significantly (p = 0.16) higher rate of VA recurrence rates in the group which underwent RFCA (7 [25%]) vs. the conservative group (5 [12%]), respectively. However, sub-analysis of the late (> 30 days postoperatively) period only, a significantly (p = 0.02) lower rate of recurrences was found in the ablated group (46.4%) vs. in the non-ablated group (75%; 2 patients from the non-ablated group died during the blanking period, and consequently, they were not included in the analysis).

Variables evaluated in the univariate analysis were as follows: age, gender, BMI, NYHA class, LVEF, severe MVR, implantation of ICD/CRTD in secondary prevention of SCA, presence of an implanted CRT-D (during the storm), ischemic cardiomyopathy, previous myocardial infarction treated conservatively, complete revascularisation after angiography, diabetes type 2, chronic kidney disease, stroke and/or transient ischaemic attack, chronic obstructive pulmonary disease, atrial fibrillation, haematocrit level, C-reactive protein level, creatinine level, GFR, NT-pro-BNP level, potentially reversible causes of ES, primary RFCA procedure, VT recurrence and ES recurrence.

The univariate analysis revealed that severe mitral valve regurgitation (MVR), presence of CRT-D during ES, diabetes type 2, stroke and/or a transient ischaemic attack, a lower baseline haematocrit level, a higher NT-pro-BNP level, a higher C-reactive protein (CRP) level and infective endocarditis as a potentially reversible cause of ES were significant prognostic predictors for all-cause mortality. Further details are provided in Table S1 in the Supplementary Appendix.

The multivariate Cox regression model, incorporating parameters which had significance in the univariate model < 0.05 as covariates (model 1), demonstrated that both the presence of CRT-D during ES and stroke and/or a transient ischaemic attack were independently associated with the increased risk of total mortality (HR 7.99, 95%CI 2.35–27.16; p = 0.001; and HR 4.78, 95%CI 1.43–15.93; p = 0.01, respectively). In the second model (including all variables with p < 0.1), the presence of CRT-D during ES (HR 6.83, 95%CI 1.93–24.24; p = 0.003), stroke and/or a transient ischaemic attack (HR 9.86, 95%CI 1.71–56.65; p = 0.01) and a lower baseline haematocrit level (HR 0.84, 95%CI 0.72–0.99; p = 0.04) were all independently associated with the higher risk of all-cause mortality. A thorough analysis of these multivariates analyses (both models 1 and 2) is also provided in Table S1 in the Supplementary Appendix.

Recurrent ventricular arrhythmias, which as they are the cause of electrical storms, are the most common problem among ICD/CRT-D recipients. In our study, catheter ablation in such patients was more effective than any other form of therapy in reducing death at any time. Despite the interdisciplinary approach to the treatment, long-term mortality remains high. In our study, more than a quarter of patients died during the course of the study, with the majority of deaths taking place during a few months after an episode of electrical storm. This is concordant with several studies, which have shown that ES occurrence in patients with ICD/CRT-D is an independent risk factor for death, with an increase in risk of at least 2–3-fold in long-term follow-up. The risk is highest in the first 3 months post an electrical storm episode [5, 10‐14]. For many years, consensus statements and guidelines have recommended the use of catheter ablation (CA) when antiarrhythmic therapy does not prevent recurrences of ventricular arrhythmias; however, these recommendations have been largely based on expert opinion and non-randomised case series [15‐17]. These recommendations were in force up until the publication of the 2015 ESC guidelines for the management of patients with ventricular arrhythmias and the prevention of sudden cardiac death where urgent catheter ablation is recommended (classification of recommendation II B) in patients with scar-related heart disease with incessant VT or electrical storm [18]. However, these guidelines were mainly created based on the Carbucicchio et al. study, where catheter ablation for the management of ES was evaluated in detail [9]. The authors included 95 patients with structural heart disease and electrical storm refractory to antiarrhythmic drug therapy. At a median follow-up of 22 months, ES recurred in 8% and cardiac mortality was significantly higher in patients in whom at least one clinical VT could not be abolished, when compared with patients after successful CA defined as non-inducible VT in programmed electrical stimulation. Similar findings were reported by Nayyar et al. [14] in their meta-analysis of 39 publications including a total number of 471 patients with structural and non-structural heart disease, who underwent invasive ablation management of ventricular arrhythmia (VA) storm with total follow-up 61 ± 37 weeks. The odds of death were found to be four times higher after a failed procedure, compared to those with a successful procedure (95%CI 2.04–8.01; p < 0.001). The value of information from our study is that it focuses on patients with structural heart disease and with a first episode of ES, excluding those who had already undergone previous RFCA. Our trial provides evidence that catheter ablation is the most effective first-line life-saving therapy for ES and should not be used only as a bailout therapy. Interestingly, no significant benefit of RFCA with respect to composite of all-cause death, electrical storm recurrence and ventricular tachycardia/ventricular fibrillation recurrence was observed in the group treated with RFCA, compared with those in the remaining population. We hypothesise that the observed reduction of all-cause mortality in the RFCA group (despite lack of statistical significance in recurrence rates) was due to trigger and substrate modification caused by ablation. This modification in turn might have led to different arrhythmias (with more benign characteristics: slower and with lower burden). We can speculate that tissue oedema, which is created during ablation, could render a new reentrant circuit in the surrounding myocardium responsible for creation of new non-clinical VT (e.g. with lower cycle and burden). Indeed, our study has shown a high rate of early VA recurrence after RFCA procedure only within the 30 days after procedure compared with the conservative group (7 [25%] vs. 5 [12%], p = 0.16, respectively). On the other hand, if blanking periods were used (30 days), the recurrence rate was 46.4% in RFCA vs. 75% in non-ablated group; p = 0.02.

Our strong feeling has been based on the assumption that not every early recurrence will lead to late recurrence, and as such, it does not necessarily represent treatment failure. Lack of late recurrences in the ablated group, but their presence in the conservative group, might be responsible for survival benefit (what is also reflected by progressively more divergent with time Kaplan–Meier curves). However, this is only a hypothesis, which has to be confirmed in further studies.

Our concept is supported by an important recent study by Muser et al. [19]. The authors compared the long-term outcomes after VT ablation for ES in patients with non-ischemic dilated cardiomyopathy and ischemic cardiomyopathy. Compared to our results, the authors reported a lower incidence of VT recurrence (33%) in median follow-up 45 months. This difference may be related to the fact that in contrast to our study, they also included patients who underwent endo-epicardial RFCA (22%). Overall, this outcome may favourably affect long-term outcomes of RFCA. Although VT recurred in 33% patients after the ablation, a substantial reduction in VT burden was observed in most of these cases. Moreover, the majority of patients (69%) with VT recurrence had only isolated (≤ 3 VT) episodes during the 6 months after the RFCA procedure.

The open question remains, what is the rate of clinical to non-clinical arrhythmias among patients with recurrences and what is the impact of non-clinical VT or of early VA recurrence on clinical outcome?

This issue is even more interesting because, in contrast to our study, Sapp et al. recently reported a significantly lower rate of composite of death as the primary outcome, three or more episodes of ventricular tachycardia within 24 h (ventricular tachycardia storm) or appropriate ICD shock (59.1%) in groups of patients with ventricular tachycardia, who were randomly assigned to receive CA with continuation of baseline antiarrhythmic medication vs. 68.5% of those in the escalated pharmacological therapy group, p = 0.04 [20]. There was also no significant difference in mortality between the groups [20]. Their results may be explained by several reasons. The authors only conducted their analysis with a follow-up period of 27.9 ± 17.1 months in the group of patients with ischaemic cardiomyopathy after myocardial infarction and implanted with ICD, and who had experienced an episode of VT within the 6 months before enrolment. During follow-up, they analysed only the VT episodes and appropriate ICD shock after the 30-day treatment period. There is also lack of data on the percentage of previous RFCA and ES, which may reflect on their outcomes. Notably and surprising is the fact that our study, based on multivariable Cox regression analysis, provides the evidence that the presence of CRT-D is an independent predictor of all-cause mortality in patients with electrical storm. We suppose that this may be attributed to the fact that in contrast to ICD recipients, patients who initially demand the implantation of CRT-D are characterised by increased severity of heart failure, and may be much more exposed to the worsening of heart failure, resulting from multiple high-energy therapies within a short time period.

By comparing long-time effects of ablation to other forms of electrical storm treatment among patients with heart failure and implanted ICD/CRT-D, who had a first episode of electrical storm, a statistically significant reduction in terms of total morality in the group treated with RFCA compared with those of the remaining population was observed.