Noninvasive risk stratification of lethal ventricular arrhythmias and sudden cardiac death after myocardial infarction

Abstract

Prediction of lethal ventricular arrhythmias leading to sudden cardiac death is one of the most important and challenging problems after myocardial infarction (MI). Identification of MI patients who are prone to ventricular tachyarrhythmias allows for an indication of implantable cardioverter-defibrillator placement. To date, noninvasive techniques such as microvolt T-wave alternans (MTWA), signal-averaged electrocardiography (SAECG), heart rate variability (HRV), and heart rate turbulence (HRT) have been developed for this purpose. MTWA is an indicator of repolarization abnormality and is currently the most promising risk-stratification tool for predicting malignant ventricular arrhythmias. Similarly, late potentials detected by SAECG are indices of depolarization abnormality and are useful in risk stratification. However, the role of SAECG is limited because of its low predictive accuracy. Abnormal HRV and HRT patterns reflect autonomic disturbances, which may increase the risk of lethal ventricular arrhythmias, but the existing evidence is insufficient. Further studies of noninvasive assessment may provide a new insight into risk stratification in post-MI patients.

1 Introduction

Ventricular tachycardia (VT) and ventricular fibrillation (VF) leading to sudden cardiac death (SCD) are responsible for significant morbidity and mortality in patients with myocardial infarction (MI). Identification of MI patients who are prone to VT/VF allows for an indication of implantable cardioverter-defibrillator (ICD) placement. For this purpose, there has been an increase in research on the noninvasive risk stratification of lethal ventricular arrhythmias and SCD. To date, various noninvasive methods such as signal-averaged electrocardiography (SAECG) [1–4], microvolt T-wave alternans (MTWA) [5–7], heart rate variability (HRV) [8–10], and heart rate turbulence (HRT) [11,12] have been developed, and currently, MTWA and SAECG are widely used for risk stratification in patients with prior MI.

2 Microvolt T-wave alternans (MTWA)

T-wave alternans is a periodic beat-to-beat variation in the amplitude or morphology of the T wave on electrocardiography (ECG) (Fig. 1). Beat-to-beat T-wave alternans is considered to reflect increased dispersion of ventricular repolarization, and it is known to often precede the development of lethal ventricular arrhythmias [13,14]. A prior experimental study suggests that T-wave alternans observed at rapid rates under long-QT conditions is largely the result of alternation of the M-cell action potential duration, leading to exaggeration of transmural dispersion of repolarization during alternate beats [15]. MTWA testing is a noninvasive test, which can detect subtle beat-to-beat fluctuations in T-wave morphology and amplitude using fast Fourier transform (spectral method, Fig. 2) [16]. In general, MTWA measurement using the spectral method is commonly performed during bicycle or treadmill exercise at an optimum heart rate using dedicated noise-reducing electrodes, and the development of T-wave alternans at a heart rate <110 bpm is defined as positive. MTWA testing has been found to be able to identify patients who will benefit from ICD therapy [17]. Recent clinical trials have shown that a positive MTWA result is associated with serious ventricular arrhythmic events and SCD [18–20]. The ALPHA (T-wave alternans in patients with heart failure) study demonstrated that patients who had heart failure due to idiopathic dilated cardiomyopathy and had abnormal MTWA test results had an adjusted hazard ratio of 3.2 for the combined primary endpoint of cardiac death and lethal arrhythmias, whereas patients with negative MTWA results had a good prognosis and were unlikely to benefit from ICD therapy [18]. The Risk Estimation Following Infarction, Noninvasive Evaluation (REFINE) study showed that combined assessment of HRT, MTWA, and left ventricular ejection fraction (LVEF) <50% beyond 8 weeks after MI reliably identified patients at risk of cardiac death or resuscitated cardiac arrest, whereas that at 2–4 weeks after MI did not predict the risk of serious cardiac events [19]. In contrast, the MATER (Microvolt T-Wave Alternans Testing for Risk Stratification of Post-Myocardial Infarction Patients) trial and a prospective substudy of the Sudden Cardiac Death Heart Failure Trial (SCD-HeFT) found that MTWA testing did not predict arrhythmic events or mortality among patients with left ventricular systolic dysfunction [20,21]. A recent meta-analysis including 15 studies (5681 patients) reported that the positive predictive value (PPV) of MTWA testing during the mean 26-month follow-up period was 14%, negative predictive value (NPV) was 95%, and univariate risk ratio was 2.35 [22]. According to the literature, a positive MTWA result indicated an approximately 2.5-fold higher risk of cardiac death and life-threatening arrhythmia, and MTWA testing showed a very high NPV both in ischemic and nonischemic patients. It is notable that MTWA has been reported to have a high NPV for malignant ventricular arrhythmias after MI. Ikeda et al. reported that the sensitivity and NPV of MTWA in predicting arrhythmic events were very high (93% and 98%, respectively) [23]. The Alternans Before Cardioverter Defibrillator (ABCD) trial was conducted to test the hypothesis that MTWA testing as well as invasive electrophysiological study (EPS) should be performed to determine ICD indication for primary prevention of SCD [24]. The trial was a multicenter prospective study that enrolled patients with ischemic cardiomyopathy and nonsustained VT. All patients underwent MTWA testing and EPS, and an ICD was implanted if either test result was positive. The event rates at the mean follow-up period of 1.9 years were significantly higher in patients with either a positive MTWA (hazard ratio, 2.1; p=0.03) or a positive EPS (hazard ratio, 2.4; p=0.007) than in those with both tests negative/indeterminate. The PPV and NPV of the MTWA test for predicting appropriate ICD discharge or SCD were 9% and 95%, respectively, which were comparable to the PPV and NPV of the EPS (11% and 95%, respectively).

Fig. 1. T-wave alternans (ABAB oscillation) leads to ventricular fibrillation (VF).

Fig. 2. Algorithm for evaluating T-wave alternans based on the spectral method. The amplitudes of the T wave are measured for 128 beats, and the power spectrum of this time series is described using fast Fourier transform methods. Microvolt TWA represents a peak at one-half of the beat frequency (0.5 cycles per beat).

Recently, an alternative method of measuring T-wave alternans, the time-domain modified moving average (MMA) method, has been developed [25]. The MMA method applies the noise-reduction principle of signal averaging. The T-wave amplitude and morphology of odd and even beats were averaged for each 15-s period and were superimposed, and the maximum difference of each median complex was reported as the T-wave alternans values. The MMA method allows MTWA analysis during routine 24-h Holter monitoring, and a number of studies have suggested its utility for predicting malignant ventricular arrhythmias, SCD, and cardiovascular and total mortality [25–33]. The Finnish Cardiovascular Study demonstrated that the NPV of the MMA method for SCD was 98.6%, which was comparable to the results of a meta-analysis with the spectral method [26,34]. The MMA method may be superior to the spectral method in that the MMA method does not require a special protocol for maintaining the stationary heart rate. Taken together, an increasing amount of data suggests that the MTWA test is a promising risk stratifier of malignant ventricular arrhythmias. It is currently recommended as a Class IIa, Level of Evidence A risk-stratification tool among post-MI patients [35].

3 Signal-averaged electrocardiography (SAECG)

SAECG is a high-resolution electrocardiographic technique developed in the 1970s for detecting patients who are prone to SCD after MI. It is based on the theory that lethal ventricular arrhythmias leading to SCD often have a reentrant mechanism. Reentry requires a unidirectional block and conduction delay sufficient to recover excitability, and areas of delayed conduction within the infarcted ventricular myocardium can often be observed as low-amplitude, high-frequency potentials located in the late segment of the QRS complex by invasive EPSs (late potentials: LPs). However, LPs are obscured by noise in a standard body surface ECG because of their low amplitude. SAECG allows the detection of LPs by signal averaging, which improves the signal-to-noise ratio. In general, SAECG recordings are obtained from the Frank X, Y, and Z leads during sinus rhythm, and a total of 200–250 cycles are averaged to obtain a noise level of <0.2–0.4 μV. The signals are amplified, digitized, averaged, and bidirectionally filtered with a band-pass filter at frequencies between 40 and 250 Hz. The filtered QRS duration (f-QRS), root mean square voltage of the terminal 40 ms of the filtered QRS complex (RMS₄₀), and duration of low-amplitude signals of <40 μV in the terminal filtered QRS complex (LAS₄₀) are measured (Fig. 3). LPs are considered to be positive if two of the following criteria are met: (1) f-QRS>120 ms, (2) RMS₄₀<20 μV, and (3) LAS₄₀>38 ms [36].

Fig. 3. A representative record of abnormal signal-averaged electrocardiography (SAECG). Signals are obtained from the Frank X, Y, and Z leads during sinus rhythm, and a total of 200–250 cycles are averaged to obtain a noise level of [less than]0.2–0.4μV. The signals are amplified, digitized, averaged, and bidirectionally filtered with a band-pass filter at frequencies between 40 and 250Hz. The arrow indicates late potential (LP).

To date, previous studies have reported that abnormal SAECG is useful for identifying patients with VT after MI [37–39]. Gomes et al. evaluated 1925 patients with asymptomatic nonsustained VT, coronary artery disease, and left ventricular dysfunction in a multicenter trial. An f-QRS>114 ms (abnormal SAECG) independently predicted arrhythmic death, cardiac arrest, and cardiac death, independent of clinical variables, cardioverter-defibrillator implantation, and antiarrhythmic drug therapy. The 5-year rates of arrhythmic death or cardiac arrest (28% versus 17%, P=0.0001), cardiac death (37% versus 25%, P=0.0001), and total mortality (43% versus 35%, P=0.0001) were significantly higher in patients with abnormal SAECG. They also assessed the prognostic utility of SAECG at different levels of LVEF. The results showed that the combination of an LVEF <30% and an abnormal SAECG identified a particularly high-risk subset. It was concluded that SAECG is a powerful predictor of poor outcomes, and the noninvasive combination of an abnormal SAECG and a reduced LVEF may help identify high-risk patients [39]. Many studies on LPs in MI patients were conducted before the reperfusion era [1–3]. Some investigators suggested that LPs are less common among patients after thrombolysis and percutaneous coronary intervention, and they limit the utility of SAECG for arrhythmia risk stratification [40–43]. In addition, Ikeda et al. suggested that LPs were independent predictors of sustained VT after MI but had no significant prognostic role in predicting SCD or resuscitated cardiac arrest [44]. Furthermore, we previously investigated the prevalence of LPs in patients with VT and VF after MI. The positive rate of LPs was 92% in VT patients and 50% in VF patients [45]. LPs may reflect an arrhythmogenic substrate, which is responsible for macroreentrant VT and is not useful for prediction of VF. Besides, LPs are more common in patients with inferior infarction than in those with anterior infarction. This may be because peri-infarct tissue in an anterior infarction is activated during the early phase of ventricular activation, and delayed potential may be hidden within the QRS complex.

In summary, SAECG may play an important role as a screening test in post-MI patients because of its high negative predictive accuracy, but its role in detecting patients at risk of lethal arrhythmic events is limited by its poor positive predictive accuracy.

At present, routine use of SAECG for identifying patients at high risk of SCD is not adequately recommended [35].

4 Heart rate variability (HRV)

HRV is a physiological phenomenon of beat-to-beat variation in cardiac cycle length, which is influenced by autonomic tone. Depressed HRV reflects autonomic disturbances, which may increase the risk of lethal ventricular arrhythmias. In 1987, Kleiger et al. analyzed HRV in 808 patients with MI. They found that HRV was the strongest univariate predictor of mortality. The relative risk of mortality was 5.3 times higher in the group with HRV <50 ms than in the group with HRV >100 ms. HRV remained a significant predictor of mortality even after adjusting for clinical, demographic, and other Holter features and ejection fraction [8]. Since then, HRV has been widely investigated and has subsequently been well established as a risk factor for poor clinical outcome in MI patients. Autonomic Tone and Reflexes After Myocardial Infarction investigators evaluated the prognostic value of HRV in post-MI patients. They found that patients with low standard deviation of normal-to-normal relative risk (RR) intervals (SDNN) and LVEF <35% carried an RR of 6.7 (95% confidence interval [CI], 3.1–14.6) compared with patients with LVEF >35% and less compromised SDNN (≥70 ms) [10]. Zuanetti et al. assessed the prognostic value of HRV for total and cardiovascular mortality in the fibrinolytic era. All indexes of low HRV could identify patients with a higher total mortality, and the independent predictive value of low HRV was confirmed by the adjusted analysis [9]. Camm et al. evaluated HRV in 3717 post-MI patients with left ventricular dysfunction. In their cohort, patients with low HRV had a significantly higher 1-year mortality than did those with high HRV (15% versus 9.5%, P<0.0005), despite nearly identical LVEF values [46]. They concluded that low HRV independently identified a subpopulation at high risk of mortality. Depressed HRV is currently considered a strong predictor of mortality and lethal ventricular arrhythmias in post-MI patients, but the evidence is lacking. Currently, HRV is recommended as a Class IIb, Level of Evidence B risk-stratification tool among post-MI patients [35]. Further studies are needed to establish its role in risk stratification.

5 Heart rate turbulence (HRT)

HRT was introduced by Schidmt in 1999 [11]. It refers to the physiological phenomenon in response to a premature beat. After a ventricular premature contraction (VPC) and a compensatory pause, there is an increase in blood pressure because of the prolonged filling. Reflex parasympathetic activation accelerates and then slows down the heart rate, which eventually returns to normal. The relative change in RR intervals before and after the VPC is defined as HRT onset (TO), and the slope of that regression is defined as the turbulence slope (TS). To date, HRT has been examined mainly in post-MI patients, and it is suggested that abnormal HRT is associated with increased mortality after MI [11,47,48]. A large prospective study validated HRT in post-MI patients. In that study, 1455 survivors of an acute MI in sinus rhythm were enrolled, and TO and TS were calculated from Holter monitoring. Multivariate analysis showed that the presence of both abnormal TO and TS was the strongest predictor of death (hazard ratio, 5.9; 95% CI, 2.9–12.2). It was concluded that HRT is a strong predictor of subsequent death in post-MI patients of the reperfusion era [49]. Although HRT is currently regarded as a strong predictor of mortality and malignant ventricular arrhythmia after MI, the available evidence is insufficient to support the routine use of HRT testing in post-MI patients. Further studies are necessary to establish its clinical utility in risk stratification.

6 Newly developed noninvasive techniques

Deceleration capacity (DC) is a new risk stratifier, which is based on the assessment of deceleration-related modulations of heart rate and is considered to reflect the vagal activity of the autonomic nervous system. It was introduced by Bauer et al. in 2006 by using a signal-processing algorithm to separately characterize the deceleration and acceleration of heart rate [50]. They demonstrated that an impaired heart rate DC is a powerful predictor of mortality after MI and is more accurate than LVEF and the conventional measures of HRV.

Wavelet transform analysis is a novel time–frequency technique that detects transient changes in an ECG even if they are superimposed on the high-gain QRS complex. High-frequency components in the QRS complex detected by this technique are considered to reflect conduction abnormality hidden within the QRS complex, which may be responsible for malignant ventricular arrhythmias. Previously, Morlet et al. developed a wavelet-transformed ECG for the detection of patients at risk of VT. They found that detection of local conduction abnormality in the QRS complex by wavelet transform was useful for VT risk stratification in post-infarction patients and reported 85% specificity and 90% sensitivity for VT detection [51]. Recently, we evaluated the utility of a wavelet-transformed ECG in post-MI patients. The sensitivity of the abnormal high-frequency components within the QRS complex for identifying VT/VF patients was higher than that of SAECG (96% versus 72%), although the specificity was similar (68.5% versus 64.3%) [44]. Furthermore, Tsutsumi et al. suggested that the combined use of LPs and intra-QRS high-frequency potentials hidden within the QRS complex improved the prediction of lethal ventricular arrhythmias in post-MI patients [52].

7 Conclusions

Noninvasive risk stratification techniques may be useful for identifying patients at risk of lethal ventricular arrhythmias, but further studies are needed. Further refinement of noninvasive studies may provide a new insight into risk stratification in post-MI patients.

Disclosures

None.

Conflict of interest

None.

Financial support

None.