Ambulatory ECG-based T-wave alternans and heart rate turbulence can predict cardiac mortality in patients with myocardial infarction with or without diabetes mellitus

Numerous studies over the past decades have provided a sound scientific basis for employing T-wave alternans (TWA), monitored during exercise testing, pacing, daily activities or from defibrillator electrogram as an index of vulnerability to ventricular arrhythmias and sudden cardiac death, as it represents an increased heterogeneity of ventricle repolarization on a beat-to-beat basis, and may provide a substrate for reentry [1‐6]. Currently, there are both frequency domain- and time domain-based analysis methods of determining TWA. Our study used a time domain-based modified moving average (MMA) to analyze TWA measured on ambulatory electrocardiograms (AECGs) [7, 8].

Following a premature ventricular contraction, there is a compensatory sinus pause, followed by sinus acceleration that typically overshoots the baseline heart rate before gradually decelerating back to baseline. This sequence of cycle length change has been labeled ‘heart rate turbulence’ (HRT). Abnormal findings for HRT are defined as TO ≥0% and TS ≤2.5 ms/RR interval [9]. Specifically, blunted post-premature ventricular contraction, sinus acceleration (less negative TO), and reduced subsequent deceleration (lower TS) have been introduced as an autonomic risk stratification marker for cardiac events after MI or ischemic cardiomyopathy [9‐12].

As is well known, diabetes mellitus (DM) not only causes coronary artery disease and cardiac neuropathy, but also increases the risk of cardiac death in post-MI patients. Numerous studies have since established definitively that hyperglycemia is highly prevalent and that it is associated with an increased risk of death and in-hospital complications in patients with MI [13, 14].

In our study, we prospectively assessed the utility of TWA and HRT, measured from 24-hour AECGs, in identifying the risk of cardiac mortality in post-MI patients with or without DM.

Patients with a positive microvolt-level TWA and HRT are characterized by an increased risk of ventricular tachyarrhythmias and cardiac mortality. We observed that TWA was elevated in patients following MI as our prior study [15]. Our data showed that patients in post MI with DM have higher TWA values than those without DM. MI patients with or without DM had higher association with positive results for TWA and HRT. A HRT-positive outcome, as well as the combination of abnormal TWA and positive HRT had significant association with cardiac mortality during the follow-up period.

Our study found that TWA was elevated in patients following MI, HRT parameters-TO and TS were significantly impaired in MI patients. A HRT-positive outcome, as well as the combination of abnormal TWA and positive HRT had significant association with the endpoint. Recently, the Consensus Guideline about TWA by International Society for Holter and Noninvasive Electrocardiology concluded the utility of TWA [18]. To date, over 100 studies enrolling a total of more than 12,000 patients support the predictive of TWA testing (including the frequency-domain Spectral Method and the time-domain Modified Moving Average method) for cardiovascular mortality and SCD. The HRT evaluation has thus been found appropriate in risk stratification after acute myocardial infarction (AMI), risk prediction, and monitoring of disease progression in heart failure, as well as in several other pathologies [12]. It is worth mentioning that in the REFINE (Risk Estimation Following Infarction Noninvasive Evaluation) trial they investigated the capacity of combined assessment of autonomic tone and cardiac electrical substrate to predict the development of serious outcomes after MI, and found that patients with impaired HRT plus abnormal Holter TWA at 10 to 14 weeks after MI were at higher risk for cardiac death or cardiac arrest, death from any cause, and fatal or nonfatal cardiac arrest [19].

Diabetes mellitus (DM) is a well-established risk factor for ischemic heart disease. Insulin resistance is a precursor and a characteristic feature of T2DM and it is also associated with higher risk of cardiovascular diseases [20, 21]. Rodríguez-Colón et al. reported that among persons with T2DM have more severe form of insulin resistance, and the circadian mechanisms of cardiac autonomic modulation are impaired [22]. Prognosis after MI is worse in patients with DM as compared to patients without DM [13]. Autonomic nervous dysfunction is the main complication of DM and associated with an increased risk of mortality in patients with diabetes and survivors of myocardial infarction [23]. In our prior study, TWA was moderate correlated with heart rate variability (HRV), and seemed to be associated with the sympathetic component of the autonomic nervous system [15]. Elevated R wave amplitudes, widening of QTc intervals and decreased HRV in an electrocardiogram (ECG) are early markers of diabetic autonomic neuropathy. The severity of diabetic autonomic neuropathy has a direct relationship with mortality risk. VanHoose et al. found that aerobic exercise training may attenuate the ECG changes, included R wave amplitudes, QTc intervals and HRV in the model of T2DM-Zucker Diabetic Fatty rat [24]. Nieminen et al. found a significant increase in TWA among diabetics in the FINCAVAS (The Finnish Cardiovascular Study) population of 1,037 consecutive patients [16]. The prevalence of DM was the only parameter that differed among the groups with statistical significance. Molon et al. have shown that in patients with DM and no known cardiovascular disease, abnormal TWA prevalence was a very common condition (approximately 25%) among people with T2DM without manifest cardiovascular disease and was closely correlated to glycemic control [25, 26]. In multivariate regression logistic analysis, HbA1c (OR = 13.6, 95% CI =2.0-89.1) predicted abnormal TWA independent of other potential confounders. Jouven et al. conducted a population-based case–control study at the Group Health Cooperative [27]. Cases (n = 2040) experienced out-of-hospital cardiac arrest due to heart disease. These investigators observed that a progressively higher risk of SCD was associated with borderline diabetes (OR = 1.24, 95% CI = 0.98-1.57), diabetes without microvascular disease (OR = 1.73, 95% CI = 1.28-2.34), and diabetes with microvascular disease (OR = 2.66, 95% CI = 1.84-3.85). Bonapace et al. studied 50 consecutive, well controlled T2DM outpatients without a history of ischemic heart disease and with normal systolic function [28]. All patients underwent a complete echocardiographic Doppler evaluation with spectral tissue Doppler analysis. TWA analysis was performed noninvasively during submaximal exercise. Multivariable logistic regression analysis revealed that higher E/e′ ratio was the only independent correlate of abnormal TWA (OR = 3.52, 95% CI = 1.19-10.6) after controlling for glycemic control and other potential confounders. They found that early diastolic dysfunction is independently associated with TWA abnormality in T2DM individuals with normal systolic function. Hennersdorf et al. found that abnormal TWA was found in 33% of patients with both hypertension and left ventricular hypertrophy (LVH) [29]. Hakalahti et al. investigated the possible association between the beta-1 adrenergic receptor (b1AR) Arg389Gly polymorphism (which plays a fundamental role in the regulation of cardiovascular functions) and LVH among non-diabetic and diabetic AMI survivors [30]. They found that the b1AR Arg389 variant seem to confer higher risk of developing LVH among non-diabetic patients who have suffered AMI. But it does not exist among diabetic AMI survivors. They hypothesized that this negative finding is caused by the strong association between DM and LVH, which may mask the presumably weaker effect of the b1AR Arg389 variant on the left ventricular structure.

Miwa et al. assessed the utility of HRT as a risk marker in post-MI patients with and without DM [31]. They prospectively enrolled 231 consecutive DM patients and 300 non-DM patients after acute MI. HRT was considered positive when both TO ≥0% and TS ≤2.5 ms/RRI criteria were met. The endpoint was defined as cardiac mortality. Forty-two of 222 patients (19%) were HRT positive. During follow-up of 876 ± 424 days, 26 patients (22%) reached the endpoint. HRT- positive outcome had significant value with a hazard ratio of 3.5 (95% CI = 1.4-8.8). Stein et al. investigated 481 hospitalized patients (24% of whom had DM) after AMI with HF and left ventricular dysfunction [32]. Over a 1-year follow-up, 55 died, 49 of cardiovascular causes. The study showed that abnormal HRT predicted cardiovascular mortality in high-risk patients with AMI and left ventricular dysfunction. Bauer et al. found a higher percentage of abnormal values for both HRT parameters in patients with DM than without DM in a subgroup analysis [33]. Barthel et al. enrolled 1455 survivors of an a AMI who were in sinus rhythm [17]. The primary endpoint was all-cause mortality. During a follow-up of 22 months, 70 patients died. These investigators showed that the combined assessment of HRT, LVEF ≤ 30%, age ≥ 65 years, and DM can predict the population at high risk. Thus, abnormal HRT is associated with DM in post-MI patients, and HRT can detect cardiac neuropathy due to DM. Balcioğlu et al. investigated the detection of diabetic neuropathy in patients with T2DM and no obvious heart disease by autonomic markers that included HRT and HRV, and found that HRT was the most powerful indicator of cardiac diabetic neuropathy [34]. Thus, HRT might be more sensitive indicators of cardiac autonomic abnormality in post-MI patients with DM because these patients have disturbance of both parasympathetic and sympathetic tone.

However, negative studies have also appeared. Martin et al. performed a case–control cross-sectional study in 140 patients [35]. Patients performed a symptom-limited Bruce protocol exercise test with assessment of TWA by the spectral method. Logistic regression analysis in all patients showed that abnormal TWA was not related to prevalent DM (OR = 0.9, 95% CI =0.4–1.8). Jackson et al. investigated the utility of TWA treadmill testing in an unselected population of patients with HF and found that half of patients with HF are eligible for TWA testing and the most common result is an indeterminate test [36]. It concluded that TWA treadmill testing is not widely applicable in typical HF patients and is unlikely to refine risk stratification for sudden death. Wijers et al. summarized the main reason for indeterminate test of TWA treadmill testing [37]. They reported that inability to achieve the target heart rate is the main reason for an indeterminate result of TWA testing. Therefore, it is questionable whether TWA alone is able to be of predictive value and for what patient population. But they concluded that combine a number of invasive and non-invasive electrophysiological risk markers (including TWA and short-term variability) could be the solution.

The results of this study indicated that patients with MI, especially those complicated with diabetes, have higher TWA values and abnormal HRT, and have a high risk of life threatening ventricular arrhythmias and cardiac death, possibly because of increased repolarization heterogeneity and abnormal autonomic function. Combined analysis using MMA-based TWA and HRT may be a convenient and useful method of predicting patients at high risk for arrhythmic events and cardiovascular death.

RL, design and data collection and drafted the manuscript. FX and WY, data analysis. GJ and RL, statistics. QG, design, critical revision of article and approval of article. All authors read and approved the final manuscript.