Clinical, procedural and long-term outcome of ischemic VT ablation in patients with previous anterior versus inferior myocardial infarction

We analyzed the data of consecutive patients with ischemic cardiomyopathy and remote MI who underwent VT ablation at our institution between January 2010 and October 2015. The study was approved by the local research ethics committee (approval number: 2017-309-f-S). On behalf of all authors, the corresponding author states that there is no conflict of interest.

Patients were stratified by whether they had a previous inferior or anterior MI. Infarct localization was determined by patients’ history, results of coronary angiography, inferior or anterior aneurisms on echocardiography and/or magnetic resonance imaging (MRI) and/or Q waves in precordial or inferior leads. Infarct localization was confirmed by scar localization during three-dimensional (3D) mapping. Patients with both, previous anterior and inferior or lateral infarction, or uncertain infarct localization were excluded from analysis.

LVEF was determined by echocardiography in all patients, and by MRI, and/or LV angiography during coronary angiography in some. All patients had stable coronary artery disease, and acute ischemia was excluded in all cases. Clinical parameters including age, sex, time from first and last MI, number of involved coronary vessels, history of coronary artery bypass surgery (CABG), New York Heart Association (NYHA) class, comorbidities, presence and type of defibrillator, time from ICD implantation and ICD-indication as well as medication at the time of presentation were assessed.

We systematically analyzed the indications for ablation, inducibility, number and cycle length (CL) of clinical and induced VT as well as access to the left ventricle (transseptal, retrograde, and epicardial). VT was considered clinical if it was ongoing at the time of the procedure, had the same 12-lead electrocardiogram (ECG) morphology as the documented spontaneous VT (if a 12-lead VT ECG was available), or had the same CL or ≤ 30 ms difference and EGM morphology (if available) matched to that of an ICD-recorded VT. VT was presumed non-clinical, if the previous criteria were not fulfilled. We also analyzed whether activation mapping was used in addition to substrate mapping or substrate mapping only. Indications for VT ablation were defined as either electrical storm, defined as ≥ 3 VT episodes/24 h and incessant VT [14], or frequent ICD interventions [15], defined as ≥ 5 VT episodes in the previous 6 months despite antiarrhythmic drug therapy [16].

Procedures were done in the fasting state and written informed consent had been obtained prior to the study. Antiarrhythmic medication, with amiodarone and beta-blocker, was not discontinued, while intravenous therapy with ajmaline was discontinued at least six hours prior to the procedure. Conscious sedation, with midazolam, piritramide and disoprivan, was applied based on the investigator’s and the patient’s preference, with continuous invasive arterial blood pressure and oxygen saturation monitoring. In addition, repetitive blood gas analyses including lactate measurements were used for patient monitoring. Implantable defibrillators were deactivated prior to the procedure. Venous access was gained in the left groin with the placement of two sheaths [6 and 7 French (F)]. Two catheters were placed in the right ventricular apex (RVA) and the coronary sinus, respectively, except for patients with cardiac resynchronization devices in whom only a RVA catheter was placed. Venous access to the mapping and ablation catheter (Navistar Thermocool, Biosense Webster Inc., Diamond Bar, CA, USA) was gained in the right femoral groin (8 F). An arterial sheath (5 F) for invasive blood pressure monitoring was placed in the right or left groin. Access to the left ventricle was primarily achieved by transseptal puncture in the majority of cases. If required, additional access was gained retrogradely via the right femoral artery (8 F). Transseptal puncture was performed with a transseptal needle (St. Jude Medical, St. Paul MN, USA) and SL 1 sheath (St. Jude Medical, St. Paul, MN, USA) or steerable sheath (Agilis, St. Jude Medical, St. Paul, MN, USA). Once access to the left heart had been gained, heparin was administered and activated clotting time (ACT) was monitored every 30 min to maintain it above 250 s. 3D mapping (CARTO, Biosense Webster Inc., Diamond Bar, USA) was used in all the cases. In patients with ongoing VT, an activation mapping was performed. In addition to 3D mapping, conventional mapping criteria (entrainment, concealed entrainment, diastolic signals, and fractionated signals) were applied to determine critical sites of the reentrant circuit. In patients who presented with sinus rhythm, VT was induced by programmed ventricular stimulation at different CLs (500, 430, 370, and 330 ms) and for up to three extra-stimuli. If patients tolerated VT (mean invasive blood pressure ≥ 60 mmHg and no lactate during blood gas analysis), VT was targeted by a combination of activation and entrainment mapping. If VT was hemodynamically unstable or not well tolerated (symptoms during light sedation, decreasing blood pressure, increase in lactate level), VT was terminated by pacing or external defibrillator shock delivery on brief deep sedation with disoprivan bolus. A substrate mapping only was performed in patients without inducible VT or patients with hemodynamically unstable VT. Scar was defined as bipolar signal amplitude ≤ 0.5 mV and healthy tissue ≥ 1.5 mV. Values between 0.5 and 1.5 mV were considered as border zones [14].

Radiofrequency energy (RF) was used for ablation with an irrigated tip ablation catheter (Navistar Thermocool, Biosense Webster Inc., Diamond Bar, CA, USA) with a flow rate of 30 ml/min at a maximum of 50 W and maximum temperature set at 45 °C. Ablation targeted all inducible VT. In patients with ongoing VT, RF was delivered over several minutes at the site where VT slowing occurred and extended in a linear fashion to transsect the scar or connect scar to an anatomical barrier, typically the mitral annulus. In patients with hemodynamically unstable VT or several different VT morphologies, ablation was guided by pace mapping (matching QRS morphology, long stimulus QRS interval, local fractionated signals) if VT morphology was available. Scar was defined by a combination of 3D scar mapping and pacing maneuvers to define channels and exits. Substrate ablation was performed by linear ablation extended to anatomic barriers like the mitral annulus, parallel to the scar at sites with fractionated signals and transsecting the scar. Programmed ventricular stimulation was repeated after ablation. It was deferred in patients with hemodynamic instability or in whom VT could not be terminated by the first external defibrillator shock, and after long procedures (> 6 h). Acute ablation outcome was analyzed separately as (1) complete success with no inducible VT, (2) partial success if clinical VT is no longer inducible but non-clinical VT is still inducible, or (3) unsuccessful if the clinical VT was still inducible. ICD were interrogated and reprogrammed after the procedure and pericardial effusion was excluded by echocardiography. Amiodarone was started, continued or discontinued after VT ablation is at the discretion of the treating physician.

Patients were seen either in the outpatient ICD clinic or by their primary cardiologist every 3 months. Clinical status and ICD interrogation were performed at each visit. In addition, patients and referring physicians were contacted by telephone at the time of manuscript preparation. Patients’ status was obtained including current medication, symptoms, ICD shocks, and re-ablation procedures. Documents of ICD interrogation were collected to assess the VT episodes and ICD therapies. For deceased patients, date of death was documented and cause of death was classified as cardiac (arrhythmic/non-arrhythmic), non-cardiac or unknown.

The cohort was divided into two subgroups according to the location of MI. Categorical variables are presented as absolute numbers (n) and percentages (%) of the total numbers for each subgroup; statistical comparisons for these were made by the Chi-squared test. Continuous variables were tested by the ANOVA-F test. The impact of the infarct localization on cardiovascular and overall mortality was tested by multivariable Cox regression models adjusted for the number of affected coronary vessels, previous bypass surgery, NYHA class, diabetes, and amiodarone use. Results were displayed as cumulative event curves. The impact of infarct localization on VT recurrence was displayed by Kaplan–Meier model as cumulative event curve. All tests performed were two-sided, and p values of < 0.05 were considered statistically significant.

Patient characteristics are summarized in Table 1. There were more patients with inferior than anterior MI (88 (58%) vs. 64 (42%) patients). Ejection fraction was significantly lower in patients with anterior MI (28 ± 10 vs. 34 ± 10, p = 0.001), and more patients with inferior compared to anterior MI had undergone (CABG) (47% vs. 30%, p = 0.05). Other parameters including sex, age, time from MI, NYHA class, and presence of comorbidities were not significantly different between groups.

Procedural characteristics are summarized in Table 2. The indication for VT ablation was electrical storm in 43% of patients while 57% of patients presented had frequent ICD therapies. This was not different between the two groups. A mean of 2.8 ± 2 VT morphologies were inducible per patient, with a trend towards more inducible VT in previous anterior versus inferior MI. Acute ablation outcome was not significantly different between the groups.

Patient status was known in 97% of patients after a mean follow-up of 3 ± 2 years (range 0–7 years). More patients with anterior MI compared to inferior MI had died (39% vs. 24%, p = 0.062, Chi-square test) with no difference regarding cardiac death between the groups. In addition, cumulative survival between patients with anterior and inferior MI was compared by multivariable Cox regression analysis adjusted for the number of affected coronary vessels, NYHA class, diabetes, previous CABG, and amiodarone use showing a trend for worse outcome in patients with anterior MI regarding total mortality (HR 1.78, 95% CI 0.97–3.28, p = 0.065) and cardiovascular mortality (HR 1.90, 95% CI 0.80–4.49, p = 0.15). When also adjusted for ejection fraction, there was no difference between infarct localization regarding total mortality (Fig. 1a, HR 1.104, 95% CI 0.55–2.23, p = 0.78) and cardiovascular mortality (panel b, HR 0.76, CI 0.26–2.23, p = 0.61). More patients with anterior MI compared to inferior MI received ICD shocks during follow-up (46% vs. 34%, p = 0.22) and more anterior MI patients had undergone a second VT ablation procedure (49% vs. 33%, p = 0.09). Significantly more patients with anterior MI compared to inferior MI patients were using amiodarone (83% vs. 66%, p = 0.04). In contrast, there was no significant difference regarding beta-blocker (90% and 93%, p = 0.52), ACE-inhibitor and angiotensin II receptor blocker (77% vs. 70%, p = 0.42), and aldosterone antagonist use (47% vs. 58%, p = 0.2) between anterior and inferior MI patients. VT-free survival is depicted in Fig. 2. There was no significant difference between inferior and anterior MI (p = 0.38). We tested whether there was a difference between the patients who presented with electrical storm versus patients with frequent ICD therapies but did not find any differences regarding survival, time of first VT recurrence and ICD shocks. Furthermore, an association between VT inducibility at the end of the procedure and cumulative mortality was assessed by multivariable Cox regression models and found no differences between patients (p = 0.64).

Previous data suggest that inferior MI may favor the occurrence of VT [8]. In line with this observation, more patients with inferior than anterior MI presented for VT ablation at our center. There may be referral bias and the possibility that anterior MI is associated with worse acute outcome preventing survival to develop late complications like sustained VT. Many studies on ischemic VT ablation do not report localization of the previous infarction. In those that do, though, more patients with inferior than anterior MI were included as observed in our study [7, 18‐23].

A lower ejection fraction in patients with anterior infarction has been reported previously [8, 12, 24, 25]. In line with these reports, LVEF in our cohort was significantly lower in patients with anterior MI compared to those with inferior MI. In addition, patients differed with regard to cardiac resynchronization therapy with more biventricular ICD in anterior than inferior MI patients. This supports the notion that patients with previous anterior MI may be sicker than patients with previous inferior MI.

Different ablation strategies have evolved over the last two decades [26]. Recently, substrate ablation was reported to be superior over ablation of the clinical VT alone [27]. Our mapping and ablation strategy involves a combination of substrate and activation mapping to define the scar and critical parts of the reentrant circuit if VT is ongoing and hemodynamically stable. Our mean RF duration was 24 min, which is somewhat shorter than that in the multicenter study by Stevenson et al. ([23]; 36 min), a study by Sacher et al. ([7]; 33 min), and one by Yamashita et al. ([28]; 32 min). As in other published studies, there is a wide range of RF duration, and studies aiming at complete substrate ablation [27, 28] certainly require more ablation than shorter linear lesions or even focal ablation in a narrow isthmus (mean RF time 32 min, range 20–50 min [28], and 68 ± 21 min [27]).

We use individualized ablation strategies, depending on inducibility, hemodynamic stability, and VT stability. This limits comparison of the ablation approach with other studies beyond acute and long-term success. This is not only a limitation of our analysis but that of VT ablation studies in general. Even apparently standardized strategies like linear ablation or “complete” LAVA ablation are limited by sampling errors and technical limitations (i.e. type and size of mapping catheters, reliable proof of complete linear lesions) [26]. Outcome, on the other hand, is most important for the patient.

Endpoints for VT ablation are not clearly defined [29]. Non-inducibility of clinical VT and any VT have both been used to define acute ablation success and are probably the best acute outcome predictors we have. In a meta-analysis, non-inducibility of any VT has been shown to be associated with improved arrhythmia free survival and all-cause mortality [30]. Other data question the predictive value of non-inducibility in post MI patients [29]. In our study, inducibility at the end of the procedure was not clearly associated with an increase in total mortality during follow-up. 40% of patients in our study were non-inducible at the end of VT ablation, while in 57% of patients clinical VT was no longer inducible but non-clinical VT was still inducible. Non-inducibility of any VT ranges between 29 and 93% in the literature [28] with a variable study size (range 15–231 patients). Our results are in the mid-range of these reports. There is certainly a large diversity of patient and substrate characteristics as well as ablation strategies limiting direct comparisons even in prospective randomized studies.

Follow-up was 3 ± 2 years in our study and thereby much longer than in other reports on VT ablation. Follow-up periods typically range between six and 36 months [29]. Recently, Sapp et al. [2] reported outcome data of patients randomized to VT ablation or escalation of medical therapy. During a mean follow-up of 27 ± 17 months, 42% of patients in the catheter ablation group had appropriate ICD shocks and 27% of patients in this group died, which is comparable to our data despite a longer follow-up period in our study. Studies evaluating ablation of late potentials report a VT recurrence rate between 19 and 42% over follow-up periods of 21 and 39 months [29, 32]. In the VISTA trial, 48% of patients with stable VT ablation had VT recurrence during follow-up of 12 months [27]. Of them, 11.9% died during that time. Overall, our outcome data appear to be comparable to those reported in the literature. Recognizing the limitations of a retrospective analysis, our very long-term follow-up data reveal a rather low cardiovascular mortality supporting generous use of VT ablation to improve the quality of life as recommended by current guidelines. In addition, VT that was not induced or targeted during a first procedure may be causing recurrent ICD interventions and trigger a redo procedure rather than recurrence of the same VT. This was described by Tokuda et al. [33] who compared details of repeat procedures to an initial VT ablation. Only 30% of VT were induced at both the initial and the redo procedure. Not only initially unsuccessful or incomplete VT ablation but also changes in the substrate may, therefore, trigger redo procedures irrespective of the ablation strategy used.

To the best of our knowledge, no study on ischemic VT ablation has yet evaluated the role of infarct localization. As discussed earlier, inferior MI may be more prone to VT based on the patients presenting with VT ablation which is also reflected in a larger proportion of patients with inferior MI in our study. In addition, outcome of patients with anterior MI may be worse compared to inferior MI. This is supported by a recent study that compared the outcome of ischemic mitral regurgitation in anterior versus inferior ST elevation myocardial infarctions (STEMI) [24]. In this study, 30-day, 1-year, and 5-year mortality rates were higher in anterior versus inferior MIs irrespective of the grade of ischemic mitral regurgitation. Differences in quality of scar, e.g. infarct size and density due to early, late or no reperfusion during the initial infarction may also impact the later occurrence of VT and cause differences in outcome. Detailed information regarding reperfusion therapy or enzyme levels during previous infarction was not available for our patient cohort, precluding this analysis. Antz et al. [10] correlated substrate size and ablation success in post MI patients. They reported that 73% of patients with small substrates had a history of inferior infarction that appeared to be better amenable to catheter ablation compared to larger substrates. Recurrence rate, though, did not differ between patients with small, medium and large substrates.

Despite comparable acute ablation outcome in our study, total and cardiovascular mortality were higher in anterior MI patients, and more patients with previous anterior MI presented for a re-ablation procedure. Anterior MI patients also had a significantly lower ejection fraction which may explain worse long-term outcome, irrespective of infarct localization itself. When adjusting for ejection fraction, there was no differences in total mortality and cardiovascular mortality between infarct localizations. The size and quality of scar associated with worse LVEF in anterior versus inferior infarction likely determines patient outcome. Further studies are needed to elucidate the interaction between infarct localization, related quality of scar, ejection fraction, VT incidence and outcome as well as possible underlying mechanisms.