Optimizing electrophysiology studies to prevent sudden cardiac death after myocardial infarction

Abstract

Aims

This study assessed associations of minimum final extrastimulus coupling interval utilized within electrophysiology study (EPS) after myocardial infarction (MI) and possible site of origin of induced ventricular tachycardia (VT) with long-term occurrence of spontaneous ventricular tachyarrhythmia and long-term survival.

Methods and results

This prospective study recruited consecutive patients with left ventricular ejection fraction (LVEF) ≤ 40% who underwent EPS days 3–5 after MI between 2004 and 2017. Positive EPS was defined as sustained monomorphic VT cycle length ≥200 ms for ≥10 s or shorter duration if haemodynamic compromise occurred. Each of the four extrastimuli was shortened by 10 ms at a time, until it failed to capture the ventricle (ventricular refractoriness) or induced ventricular tachyarrhythmia. Outcomes included spontaneous ventricular tachyarrhythmia occurrence and all-cause mortality. Shorter coupling interval length of final extrastimulus that induced VT was associated with higher risk of spontaneous ventricular tachyarrhythmia (P < 0.001). Significantly higher rates of spontaneous ventricular tachyarrhythmia (65.2% vs. 23.2%; P < 0.001) were observed for final coupling interval at EPS <200 ms vs. >200 ms. Right bundle branch block (RBBB) morphology of induced VT, with possible site of origin from the left ventricle, was associated with all-cause mortality [hazard ratio (HR) 3.2, P = 0.044] and a composite of spontaneous ventricular tachyarrhythmia recurrence or mortality (HR 1.8, P = 0.043).

Conclusion

Ventricular tachycardia induced with shorter coupling intervals was associated with higher risk of spontaneous ventricular tachyarrhythymia on follow-up, indicating that the final extrastimulus coupling interval at EPS early after MI should be determined by ventricular refractoriness. Induced VT with possible origin from left ventricle was associated with increased risk of spontaneous ventricular tachyarrhythmia recurrence or death.

Graphical abstract

VT; ventricular tachycardia.

Open in new tabDownload slide

Introduction

Ventricular tachycardia (VT) after 48 h of myocardial infarction (MI) is due to scar-related re-entry.¹ In patients with impaired left ventricular ejection fraction (LVEF) after primary percutaneous coronary intervention for ST-elevation myocardial infarction (STEMI) and inducible VT on electrophysiology study (EPS) early after MI, the prophylactic implantation of an implantable cardioverter–defibrillator (ICD) reduces all-cause mortality.² Complementing this, revascularized patients with MI and impaired LVEF, but no inducible VT on EPS, have a favourable long-term prognosis without ICD implantation.³ Accordingly, EPS presents a valuable risk stratification tool to facilitate the implantation of ICD to prevent sudden cardiac death from ventricular tachyarrhythmia after MI.⁴

For re-entrant ventricular tachyarrhythmia, programmed ventricular stimulation (PVS) at EPS can investigate the presence of electrical re-entrant substrate and risk stratify within the prediction of arrhythmic risk.^4–6 Within PVS at EPS, a ‘drive train’ of paced beats is followed by one or more extrastimuli delivered at specific intervals in relation to the previous beat. The intervals between the extrastimuli utilized for PVS and the previous paced beats are known as ‘coupling intervals’ and are usually reported in milliseconds (ms). Within the use of EPS to guide early prophylactic ICD implantation in post-MI patients with impaired LVEF, the ideal length of coupling interval to induce VT is unknown. Crucial to this consideration is the determination of relationship between coupling interval length and post-EPS ventricular tachyarrhythmia recurrence, which is also unclear.

The morphology of VT, particularly when bundle branch block (BBB) is present, can provide an indication of its site of origin.⁷ Ventricular tachycardia with left BBB (LBBB) morphology likely originates from the interventricular septum or right ventricle, whereas right BBB (RBBB) morphology likely originates from the left ventricle.⁸ The relationship between site of origin of VT induced at EPS and long-term survival has not previously been examined.

Utilizing two separate analyses, this study aimed to build on the authors’ previous work by (i) assessing whether the use of shorter coupling intervals to induce VT at EPS in post-MI patients with impaired LVEF carried predictive value for ventricular tachyarrhythmia recurrence on long-term follow-up and (ii) investigating the association of site of origin of induced VT with long-term survival of patients with inducible VT at EPS early after MI.

Methods

Study protocol

This study followed STROBE reporting guidelines.⁹ The protocol for data collection, including patient recruitment, EPS, ICD implantation, and programming, has previously been described in detail (see Supplementary material online). The study was approved by the Western Sydney Local Health District Human Research Ethics Committee, and all patients gave their informed consent. Consecutive patients with LVEF ≤40% on Days 3–5 after STEMI who underwent EPS were prospectively recruited. Positive EPS was defined as sustained monomorphic VT cycle length (CL) ≥ 200 ms^10,11 for ≥ 10 s or shorter duration if haemodynamic compromise occurred.¹² Inducible ventricular fibrillation or ventricular flutter with CL <200 ms CL was considered as non-inducible for VT.¹³

Patients were followed throughout their hospital admission. Post-discharge follow-up was via telephone contact at 1, 3, and 6 months, with 6-monthly intervals thereafter for at least 2 years. Long-term follow-up data were obtained from patient medical records and outpatient cardiologist reviews. Patients with an ICD were also followed in the ICD clinic, on home monitoring, and with a web application (OneView, ScottCare, Ohio, Cleveland). Within the collected data, cause of death was adjudicated by two cardiologists based on information obtained from witnesses, family members, death certificates provided by the state registry of births and deaths, hospital medical records, rhythm strips, and autopsy reports. A third cardiologist acted as arbitrator if consensus could not be reached.

Electrophysiology study protocol

Details of the EPS protocol used have been described previously.⁶ Programmed ventricular stimulation was performed at twice diastolic threshold at the right ventricular apex using a programmable stimulator. A drive train (S1S1) of eight beats at 400 ms was followed by up to four extrastimuli delivered one at time.¹⁴ Stimuli were rectangular pulses of 2 ms duration with a 3 s delay between each drive train. The initial extrastimulus was delivered at a coupling interval of 300 ms and then decreased in 10 ms steps to ventricular refractoriness. If the earliest possible extrastimulus (e.g. S1S2) failed to induce VT, that extrastimulus was delivered 10 ms outside the ventricular effective refractory period and an additional extrastimulus added (e.g. S2S3). The additional extrastimulus was decreased in 10 ms steps in the same manner. Additional extrastimuli were added in a similar manner until either VT or ventricular fibrillation or flutter was induced or refractoriness of the fourth extrastimulus was reached. There was no set lower limit for the shortest permissible extrastimulus coupling interval. Repetition of the PVS protocol to increase VT inducibility rates was used in similar fashion to that used by Cooper et al.¹⁵ If the first PVS was positive for inducible VT, the study was stopped and further PVS inductions were not performed. However, as VT can be induced on second PVS when negative on the first in a significant proportion of post-MI patients with LVEF impairment (see Supplementary material online), the same EPS protocol was repeated after a period of 5–10 min if monomorphic VT was not induced on first PVS. A third attempt at PVS was performed in a minority, when the operators had concerns with technical aspects of previous inductions. Isoproterenol infusion was not utilized for PVS. Pre-discharge ICD implantation was recommended for patients with inducible VT at EPS.^2,3

Implantable cardioverter–defibrillator settings

For the implanted ICDs, whenever possible, three detection zones were set: ventricular fibrillation zone of CL < 250 ms programmed to deliver therapy via shock, very fast VT zone within the ventricular fibrillation zone of CL 200–250 ms programmed to deliver anti-tachycardia pacing followed by tiered shock, and VT zone of CL 251–360 ms programmed to deliver anti-tachycardia pacing followed by tiered intensity shock. If three detection zones could not be set, ventricular fibrillation zone of CL <250 ms and VT zone of CL 251–350 ms were set. Parameters for ventricular tachyarrhythmia detection were standardized according to the ICD manufacturer. In general, for Medtronic ICDs, detection of VT CL <250 ms required 18/24 beats and subsequently 12/16 beats for redetection; for VT CL > 250 ms, detection required 16 beats and subsequently 12 beats for redetection. For Boston Scientific/Guidant devices, ventricular tachyarrhythmia detection was in accordance to the ‘sliding window principle’, with 8/10 intervals for commencement of window detection and 6/10 for continuation. For VT <240 ms, the duration timer was 1 s, with redetection for 1 s. For VT >240 ms, the duration timer was 2.5 s, with redetection for 1 s. For St. Jude devices, detection of ventricular tachyarrhythmia was based on the binning system of current interval and running interval average. Ventricular tachycardia detection required 12 binned intervals at a rate >250 ms, and ventricular fibrillation required 12 binned intervals at a rate <250 ms. Ventricular tachyarrhythmia that did not reach the set number of detection intervals was classified as non-sustained and did not meet the primary endpoint. Discriminators for supraventricular tachycardia were standardized based on arrhythmia onset, stability, QRS morphology, and ventriculo-atrial dissociation.

Outcomes

For the separate analyses for the two respective aims of this study, separate cohorts of patients, with different sample sizes, within this population were analysed, based on relevant data availability. The outcomes used differed across the two separate analyses within the present study. For the first analysis about the coupling interval, the primary outcome was the occurrence of a spontaneous ventricular tachyarrhythmia after discharge, indicated by appropriate ICD activation. Secondary outcomes included all-cause mortality and total number of ICD activations. The association of these outcomes with the coupling interval length responsible for the induction of VT during EPS was estimated. For the second analysis about the possible site of origin of induced VT, the primary outcome was time to mortality from any cause. The secondary outcome was time to a composite endpoint of first spontaneous ventricular tachyarrhythmia recurrence or mortality.

Statistical analysis

Two separate statistical analysis approaches were undertaken to investigate the two respective aims within this study.

For the first aim (final coupling interval that induced VT), the cohort of patients was described by reporting means and standard deviations (SD), medians and interquartile ranges (IQR), and proportions as appropriate. Coupling interval lengths were dichotomized into above and below the thresholds of 200 and 250 ms, respectively. These coupling interval thresholds for analysis were selected by the study investigators based on prior experience conducting PVS and EPS within this patient population. The proportion of patients with a follow-up ventricular tachyarrhythmia was plotted by coupling interval length showing the associated 95% confidence interval (CI). Kaplan–Meier survival curves graphed time free of ventricular tachyarrhythmia censored at death and time free of the composite event of death and ventricular tachyarrhythmia. A sensitivity analysis treating death and ventricular tachyarrhythmia as competing risks was undertaken to assess the effect of censoring by death. This was graphically presented showing the cumulative incidence, the probability of surviving and experiencing ventricular tachyarrhythmia recurrence prior to that time point, and the probability of experiencing ventricular tachyarrhythmia–free mortality prior to that time point. A Cox proportional hazards model was fitted to explore the association of time to ventricular tachyarrhythmia recurrence and all baseline covariates (including coupling interval length) using a backwards selection method. Baseline covariates with P < 0.2 in the univariate analysis were included in the model. Data regarding death were also summarized across time for each group stratified according to coupling interval length. Analyses were performed using R (version 4.0.2) packages Gmisc for plot and table output and knitr for reproducible research. P values of <0.05 were considered statistically significant unless stated otherwise.

For the second aim (possible site of origin of induced VT), IBM SPSS Statistics version 28 (IBM Corp. Armonk, NY) was used to analyse the data. Continuous variables were summarized using mean ± SD or median and lower to upper quartile as appropriate. Frequencies and percentages were used for categorical variables. Chi-squared, or exact permutation tests when appropriate, was used to assess univariable association between morphology of induced VT (RBBB v LBBB) and each of the categorical variables. Mann–Whitney tests were used to test for univariable associations between morphology of induced VT and continuous variables. Cox proportional hazards models were used to assess the association between potential risk factors and the two right-censored outcome variables, namely time from EPS to death (overall survival) and time from EPS to the composite endpoint of first activation or death (activation-free survival). Hazard ratios with 95% CIs from these models were used to quantify the association between each potential risk factor and outcome variable. Hazard ratios for age and LVEF were expressed per 10 unit change in value to facilitate clinical interpretation. All potential risk factors associated with an outcome at the P < 0.2 level were considered as candidate variables for inclusion in multivariable Cox proportional hazard models. Backward stepwise variable selection using likelihood ratio tests (P > 0.1 for removal) was used to identify the independent predictors of the outcome. Kaplan–Meier curves were used to illustrate overall survival and activation-free survival distributions by group. Two-tailed tests with a significance level of 5% were used throughout. This was a longitudinal observational study, and no corrections have been made for multiple comparisons.

Results

Results of the two respective analyses are presented according to aim.

Final coupling interval that induced VT

Based on available data regarding coupling intervals from a total sample of 410 patients that underwent EPS after MI between 2004 and 2017, 122 patients (29.8%) with VT induced were included in this study (Figure 1). Of these 122 patients, 38 (31.1%) experienced follow-up ventricular tachyarrhythmia. Median duration of follow-up was 5.7 years (IQR: 2.5–8.8). In the study population, the mean age was 58.9 (±11.2) years, the majority was male (87.7%), and mean LVEF was 30.0% (±7.1). The majority of patients (70.5%) had the left anterior descending as the infarct artery responsible for MI. These details and other characteristics for this analysis are shown in Table 1. The median coupling interval of the final extrastimulus responsible for induction of VT was 230 ms (IQR: 210–260 ms) with details in Table 2.

Within the cohort, 63.9% had VT induced with the final extrastimulus at a coupling interval below 250 ms, while 36.1% had VT induced with the final extrastimulus at a coupling interval above 250 ms. A total of 45.1% of the cohort had VT induced with the final extrastimulus at a coupling interval between 200 and 250 ms, and 18.9% of the cohort had VT induced with the final extrastimulus at a coupling interval below 200 ms. Median CL of induced VT across the cohort was 220 ms (IQR: 210–240). Sixty five (53.3%) patients had VT induced on first induction, 51 (41.8%) on second induction, and 6 (4.9%) on third induction. On average, 3.4 (±0.7) extrastimuli were required to induce VT. Regarding mode of termination of VT at EPS, 50.8% of patients were via shock/countershock, 41.8% were via anti-tachycardia/overdrive pacing, and 7.4% terminated spontaneously.

Patients who had VT induced at EPS with the final extrastimulus at shorter coupling intervals were more likely to have spontaneous ventricular tachyarrhythmia on follow-up (P = 0.0004). When PVS concluded with the final extrastimulus at coupling interval lengths under 200 ms, ventricular tachyarrhythmia recurrence occurred in (15/23) 65.2% of patients; however, when it concluded at coupling intervals above 200 ms, the rate of ventricular tachyarrhythmia recurrence was lower at (23/99) 23.2% (P < 0.001). When PVS was ceased with the final extrastimulus at a coupling interval length of under 250 ms, ventricular tachyarrhythmia recurrence occurred in (31/78) 39.7% of patients; however, when it concluded at coupling interval lengths above 250 ms, the rate of ventricular tachyarrhythmia recurrence was lower at (7/44) 15.9% (P = 0.006). These results are shown in Figure 2.

There was no significant difference in all-cause mortality in patients above vs. below both 200 ms (17.2% vs. 13.0%; P = 0.76) and 250 ms (18.2% vs. 15.4%; P = 0.80) coupling interval. All-cause mortality occurred in 16.4% of the entire cohort. Within the secondary outcomes, only results relating to number of ICD activations were significant for final extrastimulus coupling interval cut-off above and below 200 ms (P = 0.003) and 250 ms (P = 0.005), respectively. A greater mean number of ICD activations occurred below compared with above final extrastimulus coupling interval 200 ms (1.2 vs. 0.7; P = 0.003) and 250 ms (1.1 vs. 0.4; P = 0.005).

The time to ventricular tachyarrhythmia Kaplan–Meier graph shows recurrence for patients with final extrastimulus below 200 ms compared with those above to be statistical significant (P = 0.0012; Figure 3) and similarly using the 250 ms cut-off (P = 0.018; Figure 4). We also did a competing risk analysis, where the difference in cumulative incidence for the competing risks of death and ventricular tachyarrhythmia recurrence for patients with VT induced with final extrastimulus coupling interval below 200 ms compared with those above (P = 0.001) and for patients below 250 ms compared with those above (P = 0.019). Similar results were also found using a composite outcome of all-cause mortality and ventricular tachyarrhythmia recurrence for patients with VT induced with final extrastimulus coupling interval below 200 ms compared with those above (P = 0.036).

Within the results of the multivariable Cox proportional hazards model, final PVS coupling interval length below 250 ms was an independent predictor of shorter time to ventricular tachyarrhythmia recurrence (P = 0.002) with a hazard ratio (HR) of 6.0 (95% CI: 1.9,18.7; Figure 5), indicating an increase in the hazard of earlier subsequent spontaneous ventricular tachyarrhythmia by 500% relative to above 250 ms. Left ventricular ejection fraction was also significantly associated with follow-up ventricular tachyarrhythmia (P = 0.002) and remained in the model with a HR of 0.91 (95% CI: 0.86, 0.97), indicating that every 1% increase in LVEF reduced the hazard of subsequent spontaneous ventricular tachyarrhythmia by 9%.

Site of origin of induced VT

Based on available data, a total of 126 patients with inducible VT on EPS early after MI were included. Of these, 54% had LBBB, and the remaining 46% had RBBB morphology during induced VT. Mean age was 58 years, with 10% female and a median follow-up of 6.7 years. Mortality was 8.9% in the LBBB group, compared with 25.0% in the RBBB group (P = 0.034). The characteristics of the study population for this analysis are presented according to VT morphology in Table 3. Electrophysiology study details are summarized by VT morphology in Table 4. Details on follow-up according to VT morphology are summarized in Table 5. A total of five patients declined to have an ICD inserted. Distributions of age, sex, LVEF, cardiovascular risk factors, infarct artery, and EPS characteristics did not differ significantly by induced VT morphology.

Kaplan–Meier curves demonstrating overall survival and survival free from ventricular tachyarrhythmia recurrence by induced VT morphology are presented in Figures 6 and 7, respectively. There was a significant difference in overall survival (P = 0.043) and activation-free survival (P = 0.040) according to the morphology of induced VT, those with RBBB doing worse than those with LBBB.

Univariable Cox proportional hazard models of overall survival are presented in Table 6. Only a history of previous ischaemic heart disease had a HR that was statistically significant (HR 3.0 P = 0.027) although increasing age (HR 1.5 per 10-year increase, P = 0.060) and RBBB morphology of VT induced at EPS (HR 2.6, P = 0.051) showed borderline significance. Univariable Cox PH models of survival free from spontaneous ventricular tachyarrhythmia recurrence are presented in Table 7. Only RBBB morphology of VT induced at EPS had a HR that was statistically significant (1.8, P = 0.043).

Age, LVEF, previous ischaemic heart disease, infarct artery, and morphology of VT induced at EPS demonstrated univariable association with overall survival significant at the P < 0.2 level. They were selected as candidates for inclusion in a multivariable Cox model. After backwards selection, the estimated instantaneous risk of death increased by a multiplicative factor of 1.6 for every 10 years of increase in age (adjusted HR 1.6, 95% CI 1.0–2.6, P = 0.05) and in those with RBBB at EPS was 3.2 times that of those with LBBB (adjusted HR 3.2, 95% CI 1.0–10.0, P = 0.044).

LVEF, sex, elevated blood pressure, smoking history, and morphology of VT induced at EPS demonstrated univariable association with survival free of spontaneous ventricular tachyarrhythmia recurrence at the P < 0.2 level. They were selected as candidates for inclusion in a multivariable Cox model. Backward stepwise variable selection identified only morphology of VT induced at EPS as an independent predictor of survival free of spontaneous ventricular tachyarrhythmia recurrence. The estimated instantaneous risk of the composite endpoint of spontaneous ventricular tachyarrhythmia recurrence or death in those with RBBB at EPS was 1.8 times that of those with LBBB (HR 1.8, 95% CI 1.0–3.2, P = 0.043).

Discussion

It is a common belief that shorter coupling intervals of the final extrastimulus at PVS result in non-specific inducible ventricular tachyarrhythmias and compromise the predictive value of EPS.¹⁶ This was the first study to investigate the predictive value of shorter coupling intervals of the final extrastimulus to induce VT at EPS after MI in patients with LV dysfunction, on long-term follow-up. Amongst those with VT induced, the majority was detected at the coupling interval of the final extrastimulus below 250 ms. In addition, patients with VT induced by coupling intervals below 250 and 200 ms vs. above 250 and 200 ms had higher rate of spontaneous ventricular tachyarrhythmia during long-term follow-up. This suggests that the characteristics of the EPS and specifically the length of the coupling interval at which VT is induced is a determinant of future VT recurrence.

We also found that patients with inducible VT displaying RBBB morphology, with VT probably originating from the left ventricle, were more likely to experience poorer long-term survival and higher rates of spontaneous ventricular tachyarrhythmia, compared with those with LBBB morphology of induced VT. This study is the first to characterize the association between site of origin of VT induced at EPS after MI, indicated by RBBB or LBBB morphology, and overall survival and activation-free survival.

Many centres still conduct EPS with the final coupling interval selected through a standardized and arbitrary cut-off and not by ventricular refractoriness. Extrapolating from our results, if PVS had been ceased at a coupling interval of 200 ms, then VT would not have been induced during EPS in 23 patients, who would have not gone on to receive an ICD. Of these patients, 15 (65.2%) experienced spontaneous ventricular tachyarrhythmia. Similarly, if PVS had been ceased at a coupling interval of 250 ms, then VT would not have been induced during EPS in 78 (64%) patients, who would have not gone on to receive an ICD. Of these patients, 31 (60%) experienced ventricular tachyarrhythmia recurrence.

VT after MI is due to a scar-related re-entrant circuit.^1,4 The use of multiple wavefront pacing has recently been shown to be beneficial for the electroanatomic characterization of VT substrate when less remodelling has occurred.¹⁷ Patients with impaired LVEF have a larger amount of scar, and thus, re-entrant VT is more likely to recur (see Supplementary material online). For re-entry to occur during PVS, sufficient conduction delay is necessary to avoid stimulation while the myocardium is still in a refractory state (unable to be paced). Shorter coupling intervals during PVS are associated with a reduction in conduction velocity and an increase in conduction delay within the re-entrant circuit.¹⁸ There is an inverse relationship between coupling interval length and extent of conduction delay, as long as the coupling interval length is above the effective refractory period. This explains our finding of inverse relationship between coupling interval length to induce VT at EPS and chance of subsequent ventricular tachyarrhythmia.

QRS morphology is determined by the site of origin of focal VT or where a re-entrant circuit exits the central isthmus to activate normal myocardium.¹⁹ Left bundle branch block morphology VT induced at EPS study after MI, with possible origin from interventricular septum or right ventricle, was associated with better overall survival and reduced occurrence of spontaneous ventricular tachyarrhythmia or death. This cannot be accounted for by differences in LVEF or culprit infarct artery between LBBB and RBBB groups, as none were apparent in the study cohort. Neither did the distribution of CL nor duration of the induced VT differ significantly between the two groups. Previous studies have shown that it is the rate of VT that appears to be the main factor that determines haemodynamic instability rather than baseline left ventricular function.²⁰ Haemodynamic instability during VT has been proposed to be caused by either uncoordinated ventricular contraction or reduced diastolic filling. Decreased coordination of contraction and relaxation is critical in patients with depressed left ventricular function. Decreased systolic left ventricular function during VT is dependent on left ventricular filling, which is further compromised by impaired diastolic left ventricular relaxation resulting from abnormal contraction patterns after MI.²¹ It is also known that the right and left ventricles are not haemodynamically affected equally during VT. The fall in systolic pressure and rise in end-diastolic pressure during VT is less pronounced in the RV.²⁰ It is possible that VT with RBBB morphology, originating from the left ventricle, might be associated with greater haemodynamic instability, which might explain the increased mortality in these patients. However, within this analysis, there was a surprising lack of mortality effect of ICD shocks found within the multivariable analysis, and this should be explored in future research of larger sample size than the present study.

The shorter mean and median CLs of the induced VT in this study cohort echoed established findings in the literature for patients receiving early reperfusion after STEMI.²² Faster induced VT (and shorter CL between 200 and 250 ms) likely indicates less scar within the myocardial substrate that has to be navigated by the re-entry circuit, representing successful reperfusion (see Supplementary material online). The mean and median CLs of the induced VT within this study cohort fell between 200 and 250 ms, a range that has been demonstrated to be a clinically significant arrhythmia with a high risk of spontaneous ventricular tachyarrhythmia equivalent to CLs between 251 and 300 ms.¹³ A considerable proportion of the VT induced in this study cohort was able to be terminated by anti-tachycardia/overdrive pacing, confirming re-entry as the underlying pathophysiological mechanism and differentiating from ventricular flutter as the causative ventricular tachyarrhythmia (see Supplementary material online).

Our findings are consistent with the findings of Yokokawa et al.²³ regarding the negative predictive value of PVS utilizing up to four extrastimuli with coupling intervals down to 200 ms, which showed that non-inducibility of VT was independently associated with lower mortality in patients that underwent catheter ablation for post-infarction VT. The negative predictive value of our protocol was consistent with results from Belhassen et al.²⁴ who utilized coupling intervals down to refractoriness but limited to up to three extrastimuli and demonstrated that those without ventricular tachyarrhythmia were safe without implantation of an ICD. Our finding of requiring a lower coupling interval for the final extrastimulus is supported by the findings of Morady et al.²⁵, who reported the shortest coupling interval that induced clinically relevant VT was 180 ms, but their findings were limited to use of up to three extrastimuli. Similarly, our findings are also consistent with those from Hummel et al.²⁶ who used a minimum coupling interval of 200 ms and four extrastimuli and reported the value of PVS in patients with previous MI.

The post-MI electrophysiology protocol utilized in this study represents a progression from that used in past landmark trials. The MADIT trial used stimulation at the right ventricular apex and outflow tract, two drive CLs of 600 and 400 ms, up to three extrastimuli, and no interpulse CL <200 ms (see Supplementary material online). These data were not elaborated for MADIT II.²⁷ The MUSTT trial used stimulation at the right ventricular apex and outflow tract, an eight-beat drive train with CLs of 600 and 400 ms, and up to three extrastimuli and did not specify minimum coupling interval (see Supplementary material online). In the more recent BEST + ICD trial, pacing occurred at the right ventricular apex and outflow tract; up to three extrastimuli were utilized with different drive CLs of 600, 460 and 375 ms; and the minimum coupling interval was 200 ms.²⁸ These differences in post-MI EPS protocol compared with the prior literature may explain why PVS in our cohort was of greater predictive value for recurrence of ventricular tachyarrhythmia.

Our study’s findings have considerable utility for the optimal implantation of ICDs in patients at risk of arrhythmic sudden cardiac death after MI and thus the reduction of global rates of sudden cardiac death. In our study cohort, ceasing PVS during EPS at final extrastimulus coupling interval of 200 ms would have missed over one in ten patients who would have experienced future ventricular tachyarrhythmia. Further, ceasing PVS during EPS at final extrastimulus coupling interval of 250 ms would have missed over four in every five patients who would go on to experience future ventricular tachyarrhythmia. In these patients who will go on to experience ventricular tachyarrhythmia and have not been risk stratified to have an ICD implanted, chances of sudden cardiac death after out-of-hospital cardiac arrest are high. Conversely, utilization of our study’s findings within a PVS protocol that includes shorter coupling intervals will lead to patients at risk of future ventricular tachyarrhythmia being detected at a greater rate, ICDs being implanted accordingly, and their overall chances of sudden cardiac death and all-cause mortality being considerably reduced.² As sudden cardiac death represents a lead cause of death worldwide, this study’s findings may be beneficial for post-MI care that improves survival in the global population.

This study has some limitations. The study design did not include randomization and comparison to a control group; however, to avoid bias, the investigators recruited consecutive patients for inclusion and optimized population homogeneity by only including post-STEMI patients with LV dysfunction who had received the same post-MI care and EPS protocol. The potential confounders of atrial fibrillation and inappropriate ICD shocks were limited by ICD settings designed to limit therapy for these. A comparative EPS protocol arm involving a different number of extrastimuli and fixed coupling interval cut-off was not included in this study, thus limiting the additional conclusions that can be drawn. A lack of control arm limits the ability to draw conclusions more broadly. A relatively small sample size was utilized, particularly for the group of patients who had VT induced with a final coupling interval of <200 ms on PVS; however, statistically significant results were able to be obtained. This study was conducted in a single centre, and it is unknown whether similar findings would result from a large multi-centre clinical trial. Within the analysis of the site of origin associated with induced VT, no data are available on the prevalence and severity of multi-vessel coronary artery disease, or previous MI, both of which are strong predictors of long-term mortality. Further in this analysis, there were no follow-up data regarding recurrent ischaemic events or LVEF evolution, both of which might influence survival. Similarly within this analysis, there was no statistical investigation regarding the statistical significance of the differences on ventricular arrhythmia type and/or ICD therapy type, which presents an avenue for clinically useful future research. Within the present study, the ICD programming implemented was standard, and it is currently well known that more permissive programming reduces the number of adequate, but unnecessary, ICD interventions for self-limited ventricular arrhythmias; therefore, this is likely to overestimate the predictive value and impact on clinically relevant sustained ventricular arrhythmias.

While this study and past evidence⁴ imply that the utilized post-MI clinical care protocol confers a large survival benefit in this patient population, relevant international randomized control trial data are lacking. To address this gap, this post-MI clinical care protocol is currently being evaluated in the international PROTECT-ICD trial.²⁹ PROTECT-ICD seeks to optimize risk stratification for future sudden cardiac death after MI and is investigating a combination of early EPS with an optimal induction protocol and early ICD implantation (in those with inducible VT) for this purpose. Future research into the optimization of EPS protocols should build on the findings from the present study via further analysis investigating the effects of the coupling intervals within PVS. A potential analysis plan to investigate effects of coupling interval on VT inducibility could comprise dichotomizing EPS cohorts into two groups of coupling intervals below 200 ms and coupling intervals above 200 ms and comparing associated outcomes. Similar analyses should also be conducted to further explore the impacts of site of origin of induced VT investigated in the present study.

Conclusions

In patients with left ventricular dysfunction following STEMI, the use of shorter coupling intervals for the final extrastimulus at PVS early after MI was associated with higher rates of spontaneous recurrence of ventricular tachyarrhythmia, and RBBB morphology of induced VT (VT originating from the left ventricle) was associated with poorer long-term survival and higher rates of spontaneous recurrence of ventricular tachyarrhythmia. These data suggest that during electrophysiology studies in these situations, no minimum limit should be used for final extrastimulus coupling interval and PVS should continue until ventricular refractoriness (when the paced beat fails to stimulate the ventricular myocardium) or induction of ventricular tachyarrhythmia. Further, when induced VT has RBBB morphology, more aggressive ICD programming could be considered. These findings can be utilized to optimize EPS early after MI and the appropriate implantation of ICDs within clinical care after MI to reduce rates of sudden cardiac death.

Supplementary material

Supplementary material is available at Europace online.

Acknowledgements

P.K. is currently principal investigator of the PROTECT-ICD trial, which is supported by funding from Biotronik Australia via a research grant to Western Sydney Local Health District, Sydney, Australia.

Data availability

Data are available upon reasonable request made to the corresponding author.

References

Author notes