Programmed electrical stimulation for risk stratification of patients with ischemic cardiomyopathy

Abstract

Background

Patients with ischemic cardiomyopathy (ICM) are at an increased risk for sudden death. Although earlier trials used programmed electrical stimulation (PES) for risk stratification, more recent data demonstrate the benefit of implantable cardiac defibrillators (ICDs) in selected patients with reduced left ventricular ejection fraction (LVEF) without performing PES. However, little is known about the outcome of non-inducible patients. The purpose of this study was to evaluate the efficacy of PES for mortality risk stratification in patients with ICM.

Methods

All consecutive patients who met the inclusion criteria (history of coronary artery disease, LVEF≤35%, and absence of documented spontaneous sustained ventricular tachycardia or aborted sudden cardiac death) were included in the study. The stimulation protocol involved up to three extrastimuli from two different sites in the right ventricle, with 180 ms as the shortest coupling interval. The primary endpoint was overall survival.

Results

A total of 198 patients were included in the study; of these, 60 exhibited negative (−)PES, and 138 had positive (+)PES and also underwent ICD implantation. The mean follow-up duration was 4.5 years. There was no difference in age or LVEF between the patient groups. We found a trend towards an increased 5-year survival rate in the (+)PES group in whom ICD implantation had been performed (p=0.058). Survival was significantly better in patients under 68 year olds in the (+)PES group in whom ICD implantation was performed (hazard ratio=0.3, p=0.01). The survival rate of patients ≥68 years old was similar in both groups (p=0.95).

Conclusions

Non-inducibility during PES does not predict the prognosis of patients with ischemic cardiomyopathy.

1. Introduction

Patients with ischemic cardiomyopathy are at an increased risk for serious ventricular arrhythmias and sudden cardiac death (SCD). Epidemiological studies indicate that more than 50% of cardiac deaths are sudden [1,2]. Antiarrhythmic drugs (AAD) do not reduce mortality in these patients [3–7]. Although earlier trials used programmed electrical stimulation (PES) for risk stratification [6–8], more recent data demonstrate the benefit of implantable cardiac defibrillators (ICDs) to treat patients with severely reduced left ventricular ejection fractions (LVEF) without performing PES [7,9]. However, as many patients will never have any ICD treatment, further risk stratification is required. The MADIT I and II studies included PES before randomization to ICD or medical therapy (PES was not mandatory but encouraged in MADIT II and was performed in 80% of the patients). In the MUSTT study, a registry of all patients recruited was maintained, and included those patients who did not experience abnormal rhythms induced during PES (non-inducible patients). Buxton et al. showed that with no antiarrhythmic therapy, the non-inducible patients had better prognosis than inducible patients [10]. However, the mortality rate of both groups was still high, and the non-inducible patients might still benefit from ICD. The purpose of this study was to evaluate the efficacy of PES for risk stratification based on mortality for inducible patients treated with ICD vs. non-inducible patients in a “real-world” registry.

2. Materials and methods

2.1. Study cohort

We performed a retrospective analysis of all consecutive patients referred for PES between 1999 and 2009, who met the following inclusion criteria: 1. Presence of coronary artery disease (CAD); 2. time from the last myocardial infarction to PES>40 days; and 3. LVEF≤35%. Patients with documented sustained ventricular arrhythmias were excluded. During this period, criteria for ICD implantation in Israel for primary prevention were similar to those of the MADIT I study; hence, this was the common practice. Since 2009, MADIT II inclusion criteria have been gradually implemented. Patients enrolled at that time had LVEFs 31–35% and therefore needed to have positive PES to be eligible for an ICD. The study protocol was approved by the institutional review board of Carmel Medical Center (IRB protocol no. 0126-09-CMC; date of approval July 7th, 2010).

2.2. Definitions

The presence of CAD was determined based on a history of myocardial infarction. We included data regarding medications at the time of hospital discharge after the procedure. Mitral regurgitation (MR) grade (0 – normal, 3 – severe) was determined using echocardiography.

2.3. Electrophysiological study

The protocol included stimulation from two right ventricular sites (the apex and septum) and two different drive trains (600 ms and 400 ms) [6]. We used up to three extrastimuli, with the shortest coupling interval being 180 ms. No drug was administered to enhance inducibility. Induction of sustained (≥30 s) or unstable VT, or ventricular flutter (VFL), was considered positive PES, while induction of ventricular arrhythmias other than monomorphic VT was considered positive only if reproducibly induced with a single or double extra-stimuli. Monomorphic VT was defined as a VT with a uniform stable QRS morphology with a cycle length>230 ms. VFL was defined as sustained monomorphic VT with a shorter cycle length (≤230 ms). Ventricular fibrillation (VF) was defined as a rapid disorganized rhythm without consistently identifiable complexes.

2.4. Device implantation

All patients with (+)PES underwent ICD implantation within the same week. The implantation and programming were not uniform but left to the discretion of the operator. However, programming was typically performed according to the PainFREE Rx II study protocol. Patients underwent dual-chamber ICD implantation if they had a history of atrial arrhythmias.

2.5. Patient follow-up

For patients who underwent ICD implantation, follow-up was conducted at our ICD clinic 3–6 months interval. Appropriate ICD therapy was defined as any therapy (anti-tachycardia pacing or DC shock) given for sustained ventricular arrhythmia. Inappropriate therapy was defined as therapy administered for supraventricular tachycardia. Survival status, record of hospitalization, and medical events were verified using the Health Maintenance Organization (HMO) database. Follow-up for all other patients utilized the HMO database.

2.6. Study endpoints

The primary study endpoint was overall survival. In addition, we evaluated complications related to the ICD implantation, such as infection, lead reposition, deep venous thrombosis, and inappropriate activation.

2.7. Statistical analysis

Analysis was performed according to an intention-to-treat model. Continuous variables were compared using Student׳s t-test or the Mann–Whitney U test, as appropriate. Continuous variables with a non-normal distribution are presented as median (interquartile range). Categorical variables are expressed as percentages and were compared using Chi-squared or Fisher׳s exact tests, as appropriate. Univariate and multivariate Cox proportional hazards models for survival with a stepwise procedure were performed. Hazards ratios with 95% confidence intervals were estimated from the models. All tests were two-sided with a significance level of 0.05. Statistical analyses were performed using SPSS software (SPSS Inc., version 16th, Chicago, Illinois).

3. Results

A total of 198 patients were included in the study. Of these, 138 exhibited (+)PES and underwent ICD implantation and 60 exhibited (−)PES and were followed clinically. The mean follow-up period was 44.2±17.6 months. The baseline characteristics upon inclusion into the study are presented in Table 1. The mean age of the (−)PES patients was similar to that of the (+)PES group (68.5±8.6 and 66.3±9 years, respectively; p=0.14). The mean LVEF of all patients was 27.8±5.3% and did not differ between the two groups. More patients in the (−)PES group had a history of atrial fibrillation (28.8% vs. 15.2%, p=0.027). β-Blockers were used less often in the (−)PES group (86.7% vs. 98.6%, p<0.001). Baseline electrocardiographic characteristics were similar in the two groups. Some characteristics of ECG and PES are presented in Table 3.

Table 1.

Baseline characteristics.

	Inducible	Non-inducible	p Value
Number	138	60
Age (years)	66.3±9	68.5±8.6	0.14
Females, n (%)	51 (45.1)	29 (51.8)	0.41
Mean follow-up (months), mean±SD	42.0±19.0	43.3±18.7	0.64
LVEF, %	27.7±5.6	28±4.7	0.65
LVEDD, mm	59.3±6.9	58.2±6.8	0.3
Mitral regurgitation (0–4 scale), mean±SD	1.03±0.8	1.1±0.9	0.45
Previous PCI, n (%)	63 (46)	28 (47)	0.9
Previous CABG, n (%)	68 (49)	45 (27)	0.58
Mean time from previous MI or revascularization (years)±SD	8±6	8.1±5.9	0.88
Atrial fibrillation, n (%)	21 (15.2)	17 (28.8)	0.027
Diabetes, n (%)	30 (50)	50 (36.2)	0.07
Hypertension, n (%)	81 (58.7)	40 (66.7)	0.29
Creatinine (mg/dL), mean±SD	1.2±0.5	1.2±0.46	0.98
Medications
β-Blockers, n (%)	136 (98.6)	52 (86.7)	<0.001
ACE inhibitors, n (%)	125 (90.6)	56 (93.3)	0.53
Amiodarone, n (%)	12 (8.7)	7 (11.7)	0.51
Other antiarrhythmic drugs, n (%)	13 (9.4)	6 (10)	0.9
Aldospirone, n (%)	19 (14)	12 (20)	0.27
Digoxin, n (%)	17 (12.3)	12 (20)	0.16

LVEF, left-ventricular ejection fraction; LVEDD, left-ventricular end diastolic diameter; PCI, percutaneous coronary intervention; CABG, coronary artery bypass grafting; MI, myocardial infarction.

Table 3.

Electrophysiological characteristics.

	Inducible	Non-inducible	p Value
ECG analysis
QRS duration, ms±SD	128±31	130±27	0.65
PR duration, ms±SD	201±47	205±42	0.66
QTc duration, ms±SD	0.46±0.05	0.45±0.08	0.76
fQRS duration (%)	59 (66.3)	30 (63.8)	0.77
Number of extra-stimuli (n)±SD	2.6±0.5	2.7±0.5	0.12
Shortest coupling interval (ms)±SD	216±27	199±17	<0.001
Induction of ventricular arrhythmia
None	0 (0)	52 (87)
Ventricular tachycardia	119 (86)	0 (0)
Ventricular flutter	19 (14)	2 (3)
Polymorphic VT	0 (0)	2 (3)
Ventricular fibrillation	0 (0)	4 (7)

QTc, QT corrected according to Bazett formula; fQRS, fragmented QRS; VT, ventricular tachycardia.

The primary outcome, five-year survival rate, showed a trend towards increased survival in the (+)PES group in which ICD implantation was performed compared to the (−)PES group (20% vs.. 35%, p=0.058) (Fig. 1A, Table 2). Survival rates significantly reduced in younger patients (≤68 years old) who did not undergo ICD implantation (HR, hazard ratio=0.3; Fig. 1B), especially in young patients with severely reduced LVEF (≤25%) (HR=0.1; Table 2). Of note, the cut-off of 68 years old was chosen for subgroup analysis because it was the median age in the study.

Fig. 1.

Kaplan–Meier estimates of survival. (A) A trend towards a higher survival rate in the positive programmed electrical stimulation [(+)PES] group is apparent. (B) Subgroup analysis according to age, showing patients younger than 68 years. Patients with (+)PES who underwent implantable cardiac defibrillator (ICD) implantation experienced improved survival compared to patients in the negative (−)PES group. C. Among patients ≥68 years of age, there was no difference in mortality rates between the groups (p=0.97).

Table 2.

Clinical outcomes.

	Inducible	Non-inducible	HR	95% CI	p Value
Five-year mortality, n (%)	27 (20)	21 (35)	0.58	0.33, 1.02	0.058
Five-year mortality in patients≤68 years
All, no. of events (%)	8 (11)	10 (35)	0.3	0.12, 0.76	0.007
LVEF≤25%	3 (9)	6 (60)	0.1	0.03, 0.42	0.001
LVEF>25%	5 (12)	4 (21)	0.69	0.18, 2.57	0.58
Five-year mortality in patients>68 years
All, no. of events (%)	19 (30)	11 (36)	0.97	0.46, 2.05	0.95
LVEF≤25%	10 (40)	5 (38)	0.89	0.3, 2.6	0.83
LVEF>25%	9 (24)	6 (33)	1.08	0.38, 3.05	0.89

HR, hazard ratio; LVEF, left ventricular ejection fraction.

In the univariate subgroup analysis, the factors that predicted reduced survival were increased age, significant mitral regurgitation (≥ grade 2), and creatinine level (Table 4). After multivariate analysis, (+)PES followed by ICD implantation was a strong predictor of reduced mortality, along with creatinine level and MR grade (Table 5).

Table 4.

Univariate analysis of five-year mortality rates.

Variable	HR	95% CI	p Value
(+) PES	0.58	0.33–1.03	0.06
Age>68 years	2.15	1.2–3.8	0.01
LVEF≥25%	0.78	0.4–1.4	0.38
Mitral regurgitation≥grade 2	2.8	1.5–5.1	0.01
Creatinine level	2.1	1.5–2.8	<0.0001

PES, programmed electrical stimulation; LVEF, left ventricular ejection fraction; HR, hazard ratio.

Table 5.

Multivariate analysis of five-year mortality rate.

Variable	HR	95% CI	p value
(+) PES	0.29	0.1–0.83	0.02
AGE>68 years	1.24	0.45–3.46	0.68
LVEF≥25%	0.67	0.36–1.25	0.21
Mitral regurgitation≥grade 2	4.44	2.26–8.71	<0.0001
Creatinine	2.43	1.65–3.57	<0.0001

PES, programmed electrical stimulation; HR, hazard ratio.

Ten (16.7%) patients in the (−)PES group underwent “cross-over” ICD implantation during the mean follow-up period of 38±19 months. Of these, three underwent repeat (+)PES at an average of 44±15 months after the first PES and did not experience any arrhythmic symptoms; two received a pacemaker or underwent internal loop recorder documentation of non-sustained VT; three developed an indication for CRT implantation; and in two patients, a change in the HMO policy enabled ICD implantation despite the PES result. None of the patients experienced cross-over owing to symptomatic arrhythmia.

4. Discussion

The main findings of our study are that patients with inducible VT during PES and underwent ICD implantation exhibited a trend towards improved survival compared to patients with non-inducible PES. Therefore, PES cannot reliably predict mortality in the general population of patients with ischemic cardiomyopathy. Interestingly, subgroup analysis according to age showed poor survival among non-inducible patients<68 years old (the median age in the study) who did not undergo ICD implantation compared to patients who underwent ICD implantation, especially in those with severely reduced LVEF (≤25%) (Table 2 and Fig. 1B). On the other hand, survival of elderly patients (>68 years old) was similar in both groups.

Several large studies evaluated the effectiveness of PES for the risk stratification of patients with CAD who have moderately to severely reduced LVEF. The MUSTT study compared EP-guided therapy vs. conventional therapy in inducible patients with CAD and reduced LVEF, and found that ICD implantation, but not AAD, reduced the risk of sudden death in inducible patients [6]. In addition to the main MUSTT trial, the non-inducible patients were included in a registry. The mortality rate of the non-inducible patients was significantly lower than that of the inducible patients receiving non-EP-guided therapy [10], but this rate was still high (21% in two years). Treatment with β-blockers was prescribed to less than half of the patients. The authors of the MUSTT trial concluded that non-inducibility during PES is a good predictor for improved survival [11]. The MADIT I study examined the effectiveness of ICD on mortality rate in CAD patients with LVEF≤35%, with inducible but not suppressible VT during PES [8]. However, the outcome of non-inducible patients in this study is not known. β-Blockers were administered to less than 10% of the patients. In the MADIT II trial, patients with CAD and severely reduced LVEF (≤30%) were randomized to ICD implantation or conventional therapy without requiring PES [9]. Nonetheless, PES was encouraged by the study committee, and 82% of the patients in the ICD arm underwent PES [12]. In the subgroup analysis, appropriate ICD therapy was more common in the inducible patients, but was still high in the non-inducible group (25.5% at two years). However, appropriate ICD therapy for ventricular arrhythmia occurs more frequently than death among those patients and therefore cannot be used as a surrogate for mortality. Furthermore, β-blockers and angiotensin-converting-enzyme inhibitors were administered to approximately 70% of the patients. Thus, we lack updated data regarding the outcome of non-inducible patients who were treated based on current guidelines and recommendations, including the use of β-blockers. Recently, Zaman et al. showed that PES is a strong predictor of mortality early after acute MI [13].

Current guidelines recommend ICD for patients with EF≤35% and New York Heart Association (NYHA) class II or III status, without the need to induce VT during PES. This recommendation is based on the large trials mentioned above. However, MADIT II included patients with EF≤30%. Although the SCD-Heft study [7] recruited patients with EF≤35%, the subgroup with EF>30% was small (285 patients) and subgroup analysis failed to demonstrate a survival benefit with ICD in this group. In the MUSTT study, patients underwent PES for risk stratification. However, the MUSTT study excluded non-inducible patients while the MUSTT registry compared only between non-inducible patients to inducible patients whom did not underwent ICD implantation. Thus, our study provides information that supports the current guidelines and recommendations for ICD implantation for the primary prevention of ventricular arrhythmias without using PES for risk stratification. Because device implantation carries a substantial cost and potential for complications, further risk stratification is needed [14]. Thus far, many predictors of SCD risk have been identified, including age [15], renal function, degree of LV dysfunction, and electrocardiographic parameters such as QRS duration and notching [14]. However, none of these parameters could clearly influence the guidelines and recommendations owing to the lack of specificity. Subgroup analysis of our results shows that in elderly patients, the survival rate was similar in both groups, despite the lack of ICD implantation in the (−)PES group (Table 2 and Fig. 1C). Mortality causes in this group may be different, and thus, the ICD benefit may reduce [16]. In contrast, younger patients (≤68 years old) who did not undergo ICD implantation had a significantly lower survival rate despite showing (−)PES, with a hazard ratio of 0.3. However, this finding should be confirmed in larger patient populations, as the current study was underpowered for this type of analysis.

Our two groups of patients were similar in all, but two, parameters: the non-inducible group had a significantly higher prevalence of atrial fibrillation (Table 1) and a significantly lower use of β-blockers at the time of enrollment into the study (discharge time after PES only or ICD implantation). The latter difference may be related to the lack of back-up pacing, because some these patients had bradycardia.

5. Study limitations

We performed an intention-to-treat statistical analysis including patients in the (−)PES group analysis who underwent ICD implantation (“crossover”). Although this type of analysis may result in overestimation of the survival rate in the (−)PES group, the odds are low owing to the relatively long interval between enrollment and crossover. Furthermore, crossover is common in many ICD trials, and intention-to-treat is still the preferred method of analysis. The follow-up of our registry did not include documentation of repeated ischemic events and repeated LV function assessments, which could deteriorate or improve over time and have an effect on patient outcome. However, neither of the large studies discussed above included such data. We used total mortality rates as an end-point and did not have information about the cause of death. Therefore, non-cardiac causes of death may have influenced our results. However, most trials involving ICD implantation used all-cause mortality as the primary end-point. In addition, non-arrhythmic causes of death may also influence the benefit of ICD in clinical practice. Finally, we performed a subanalysis according to age, which suggested that elderly patients may have similar survival for both (+)PES with ICD implantation and (−)PES. The study was underpowered for such an analysis, but we chose to describe it because the results were significant and are supported by the results of previous studies (15).

In conclusion, (−)PES cannot be used for risk stratification in the general population with ischemic cardiomyopathy. However, future studies should evaluate the role of PES in the elderly and in patients with less severe LV dysfunction (LVEF>30%), to improve the cost-effectiveness of ICD.

Conflict of interest

The authors report no conflicts of interest.