Meta-analysis of T-wave indices for risk stratification in myocardial infarction.

Ventricular tachy-arrhythmias are the commonest cause of sudden cardiac death in patients with ischemic heart disease.[1] It is estimated that up to 20% patients with acute myocardial infarction suffer from these arrhythmias.[2] There is a growing need for developing tools for risk assessment for sudden cardiac death (SCD) in this population. Several ECG indices have been proposed for risk stratification and for prognostication after myocardial infarction (MI).[3]–[5] These can divided broadly into those based on repolarization, conduction or both, as reviewed recently.[3] Repolarization indices include QT interval, QT dispersion, Tpeak-Tend interval, Tpeak-Tend/QT ratio, JTpeak/JT, (Tpeak-Tend)/ JTpeak and Tpeak/JT ratios. Conduction indices include QRS duration and QRS dispersion. Conduction-repolarization indices are exemplified by index of Cardiac Electrophysiological Balance (iCEB: QT/QRS),[6] and QRS-T angle, amongst others.[7]–[9]

Of the repolarization markers, Tpeak-Tend interval, the difference between the peak and the end of the T wave, and Tpeak-Tend/QT ratio, represent the dispersion of repolarization.[10] A meta-analysis recently published by our group demonstrated its predictive value in several cardiac conditions that included Brugada syndrome, heart failure and myocardial infarction.[11] Other investigators have proposed dividing Tpeak-Tend by the QT interval. This yields the Tpeak-Tend/QT ratio, which has a relatively constant normal range between 0.17 and 0.23.[12] Theoretically, increases in this ratio represent a higher dispersion of repolarization, which should promote ventricular arrhythmogenesis, however its value for predictive arrhythmic or mortality risk after MI remains controversial. Thus, three studies also did not demonstrate differences in this ratio between high and low risk groups.[13]–[15] Therefore, we performed a systematic review and meta-analysis to examine whether Tpeak-Tend interval and Tpeak-Tend/QT ratio can distinguish patients with MI who are at high risk of arrhythmic or mortality events from those free from these events.

A total of 42 and 95 entries were retrieved from Pubmed and Embase, respectively, of which 10 studies met the inclusion criteria and were therefore included. A total of 1967 MI patients were included in this meta-analysis. The patients had a mean age of 68 and 78% were male with a mean follow-up period of 15 ± 12 months. We found that Tpeak-Tend/QT ratios were significantly higher in high risk patients compared to the low risk group (mean difference: 0.06, standard error 0.02, P < 0.01; I2 = 88%; Figure 1). The Cochran's Q value was greater than the degrees of freedom (58 vs. 7), indicating that the true effect size was different between studies. I2 was 88%, which indicates the presence of substantial heterogeneity. Begg and Mazumdar rank correlation analysis demonstrated that Kendall's Tau took a value of -0.1 with P > 0.05, which suggests no significant publication bias. Egger's test demonstrated no significant asymmetry (intercept 2.2, t-value 2.1; P > 0.05). To identify the source of the heterogeneity, sensitivity analysis was performed by removing one study at a time, but this did not significantly influence the mean difference. This suggests that none of the studies was singly responsible for the heterogeneity observed. ORs/HRs were available for five studies and are shown in Figure 2. In terms of risk quantification, all five studies reported a positive association between increased Tpeak-Tend/QT and increased risk of VT/VF or mortality (two studies used multivariate analysis and three studies used univariate analysis). The pooled meta-analysis demonstrated that a higher Tpeak-Tend/QT ratio is associated with approximately 3.16 times higher risk of these endpoints (95% CI: 1.13 to 8.82; P < 0.05). The Cochran's Q value was greater than the degrees of freedom (21 vs. 4), indicating the true effect size were different among different studies. I2 of 81% suggested substantial heterogeneity. Begg and Mazumdar rank correlation analysis demonstrated that Kendall's Tau took a value of 0 with P > 0.05, which suggests no significant publication bias. Egger's test demonstrated significant asymmetry (intercept 2.2, t-value 5.0; P < 0.05). To identify the source of the heterogeneity, sensitivity analysis was performed by removing one study at a time, but this did not significantly influence the mean difference. This suggests that none of the studies was singly responsible for the heterogeneity observed.

Figure 1.

Flow diagram of the study selection process for studies investigating Tpeak-Tend interval and Tpeak-Tend/QT ratio in myocardial infarction patients.

CI: confidence inervals; MD: Mean difference.

Figure 2.

Forest plot demonstrating the mean differences for between Tpeak-Tend/QT ratios obtained in event-positive and event-negative groups.

CI: confidence inervals; OR: odds ratio.

Our study is the only meta-analysis that evaluates the role of Tpeak-Tend/QT ratio for prognostication in 1967 patients with prior MI. The main findings are: (1) Tpeak- Tend/QT ratio is higher by 0.06 in patients who suffered from malignant ventricular arrhythmias and/or death compared to those who were free from these events; (2) a higher Tpeak-Tend/QT ratio with a mean cut-off of 0.26 are associated with 3.16-fold increase in the risk of ventricular arrhythmias and/or death, respectively. Tpeak-Tend is the interval between the peak and the end of the T-wave on the ECG, which is a reflection of the dispersion of repolarization.[10] As this interval varies with heart rate,[12] dividing it by the QT interval yields the Tpeak-Tend/QT ratio, which has a relatively constant normal range between 0.17 and 0.23.[12] Theoretically, higher values of both indices reflect increased dispersion of repolarization, which is expected to increase arrhythmogenicity.[16] Pre-clinical experiments conducted in coronary-perfused, canine wedge preparations showed that the end of action potential (AP) repolarization at the epicardium coincided with the Tpeak and end of AP repolarization at the M-cell coincided with Tend.[17] In other words, Tpeak- Tend can be used as a surrogate marker of transmural dispersion of repolarization (TDR). Increases in TDR are pro-arrhythmic owing to the increased likelihood of developing unidirectional conduction block and therefore to re-entry. Increased TDR is also thought to underlie phase 2 reentry in Brugada syndrome.[16] Tpeak-Tend is lead-dependent as the dispersion of repolarization varies with cardiac regions.[18] This interval can be determined generally using two methods, the tangent method and the tail method. This is important because one study reported that Tpeak-Tend interval, when determined using the tangent method, was not significantly between high risk and low risk groups.[13] However, using the tail method, Tpeak-Tend interval was significantly prolonged in high risk MI patients compared to the low risk cohort.[13]

Recently, our group conducted a comprehensive systematic review with meta-analysis into the value of Tpeak-Tend interval in predicting arrhythmic and mortality outcomes in different clinical conditions.[11] However, it did not examine the relationship between Tpeak-Tend/QT ratio and adverse outcomes. Therefore, the present study goes on to investigate the mean differences in Tpeak-Tend/QT ratio in a cohort of myocardial infarction patients. The central hypothesis is that higher Tpeak-Tend/QT ratios increase arrhythmic risk due to an exacerbation of the transmural or global dispersion of repolarization, and that measures that reduce this exacerbation would reduce arrhythmic risk.[3] The results of this meta-analysis clearly show that prolongations do indeed distinguish high risk patients from patients who did not suffer from adverse cardiac events. It should be noted that three studies reported no value of this ratio in risk stratification.[13]–[15] There are several reasons why this might have been the case. Firstly, in normal hearts, a degree of transmural repolarization heterogeneity ensures the normal unidirectional conduction of the cardiac impulse through the heart. Normally the endocardium depolarizes first and repolarizes last whereas the epicardium depolarizes last and repolarizes first. Disturbances in this activation and recovery sequence could potentially lead to arrhythmogenesis.[19] Decreases in transmural dispersion of repolarization, reflected by shorter Tpeak-Tend intervals, may also be pro-arrhythmic. Indeed, this notion is supported by a study published by Porthan and colleagues demonstrating that a reduction in this interval was associated with a higher likelihood of ventricular arrhythmic events in the general population.[20] Other studies found no significant difference in Tpeak-Tend intervals[15] or in Tpeak-Tend/QT ratios[13]–[15] between high and low risk groups. These findings suggest that increased transmural dispersion of repolarization is only one mechanism through which arrhythmogenesis takes place. Other electrophysiological abnormalities, such as reduced conduction velocity and increased dispersion of conduction are also important, as suggested previously.[11]

In conclusion, this meta-analysis shows that Tpeak-Tend/QT ratio is a significant predictor of ventricular tachy-arrhythmias and/or death in myocardial infarction. Our study adds to growing body of evidence supporting utility of repolarization abnormalities seen on a standard electrocardiogram for prognostication and risk stratification in patients with prior myocardial infarction. These indices are should be utilized in clinical practice to aid risk stratification.