Association of QT dispersion with mortality and arrhythmic events—A meta‐analysis of observational studies

Abstract Background The risk stratification of coronary heart disease (CHD) and/or heart failure (HF) patients with easily measured electrocardiographic markers is of clinical importance. The aim of this meta‐analysis is to indicate whether increased QT dispersion (QTd) is associated with fatal and nonfatal outcomes in patients with CHD and/or HF. Methods We systematically searched MEDLINE and Cochrane databases without restrictions until August 15, 2018 using the keyword “QT dispersion”. Studies including data on the association between QTd and all‐cause mortality, sudden cardiac death (SCD) or arrhythmic events in patients with HF and/or CHD were classified as eligible. Results In the analysis including patients with CHD and/or HF, we found that QTd did not differ significantly in patients with SCD compared to no SCD patients while QTd was significantly greater in the group of all‐cause mortality patients and in patients who experienced a sustained ventricular arrhythmia. Subgroup analysis showed that in myocardial infarction studies, QTd was significantly higher in patients with an arrhythmic event compared to arrhythmic event‐free patients while a nonsignificant difference was found in QTd in patients who died from any cause compared to survivors. Similarly, in HF patients, the QTd was significantly greater in patients with an arrhythmic event while a nonsignificant difference was found regarding all‐cause mortality and SCD outcomes. Conclusions QTd has a prognostic role for stratifying myocardial infarction or HF patients who are at higher risk of arrhythmic events. However, no prognostic role was found regarding all‐cause mortality or SCD in this patient population.


| INTRODUC TI ON
The identification of patients at increased risk for sudden cardiac death (SCD) and major arrhythmic events [ventricular fibrillation (VF) and ventricular tachycardia (VT)] is of outstanding clinical importance. Noninvasive tests (signal-averaged electrocardiography, arrhythmic burden in Holter monitoring, echocardiography markers) have been used to identify patients at high risk for fatal cardiovascular events. [1][2][3] QT dispersion (QTd) (difference between maximum and minimum QT interval) is an easy measured electrocardiographic marker which correlates significantly with the dispersion of action potential duration at 90% repolarization and recovery time. 4 Several studies have studied the association of QTd with arrhythmic events and mortality in different clinical settings, 5-10 including healthy men, 11 elderly, 12 and general population. 13 However, the prognostic role of QTd is not yet fully established 0. 14,15 The risk stratification of coronary heart disease (CHD) or heart failure (HF) patients using easily measured electrocardiographic markers, such as QTd, would undoubtedly be a useful addition to the modern diagnostic arsenal.
In order to aggregate diverging evidence in the field, we performed a quantitative synthesis of the existing data about the impact of QTd on three major outcomes (all-cause mortality, SCD, and arrhythmic events) in patients CHD and/or HF.

| Search strategy
We systematically searched (WHHR, GB) MEDLINE (by using PubMed Web-based search engine) and Cochrane databases without year, language, starting date, or any other restriction until August 15, 2018. We used the keyword "QT dispersion". Furthermore, the reference lists of all included studies and relevant review studies were also searched to trace more eligible articles.

| Study selection
The studies included in our meta-analysis presented data concerning the association between QTd and all-cause mortality, SCD or arrhythmic events in patients with HF or with CHD. For the classification of HF patients two criteria were used: mean EF <50% and type of cardiomyopathy (the presence of ischemic cardiomyopathy, dilated cardiomyopathy, or valvular heart disease in more than 85% of the total population included).
Studies were excluded from our analysis, using the following rejection criteria: (a) patients <18 years old, (b) no HF or CHD patients (according to the criteria we mentioned above), (c) hypertrophic cardiomyopathy or channelopathies in >15% of the included patients, (d) congenital heart disease, and (e) studies not providing full text in English language.

| Data extraction
The data extraction performed by two independent investigators (GB and WHHR). The information extracted for each study included:

| Definitions
QTd was defined as the difference between the longest (QTmax) and the shortest (QTmin) QT intervals within a 12-lead ECG. 16 As arrhythmic events were considered VT or VF episodes. SCD was defined as an unexpected death because of cardiac causes (probably VT/VF or cardiac asystole leading to electromechanical dissociation) occurring in a short time period (generally within 1 h of symptom onset).

| Statistical analysis
Data were analyzed using Review Manager software (RevMan, version 5.3; Oxford, UK). Continuous variables were pooled as mean differences. The statistical heterogeneity of the study was assessed using the I 2 index. We considered low, medium, and high an arrhythmic event compared to arrhythmic event-free patients while a nonsignificant difference was found in QTd in patients who died from any cause compared to survivors. Similarly, in HF patients, the QTd was significantly greater in patients with an arrhythmic event while a nonsignificant difference was found regarding all-cause mortality and SCD outcomes.
Conclusions: QTd has a prognostic role for stratifying myocardial infarction or HF patients who are at higher risk of arrhythmic events. However, no prognostic role was found regarding all-cause mortality or SCD in this patient population.

K E Y W O R D S
all-cause mortality, arrhythmic events, coronary artery disease, heart failure, QT dispersion heterogeneity to have approximate values: 25% (I 2 = 25), 50% (I 2 = 50), and 75% (I 2 = 75), respectively. 17 Funnel plots were constructed to assess publication bias. Random effect models were utilized in the analysis because they provide a more conservative estimate of the overall results. Subgroup analysis regarding the underlying cardiac disease was a priori projected (MI studies, HF studies).

| Patients with coronary heart disease and/or heart failure
Our search retrieved seven studies (n = 2582 patients, age: 64.3 years old, males: 76.3%) including data regarding the association between QTd and SCD. No statistically significant difference was observed between patients with and without SCD (mean difference [95% CI]: 4.33 [−4.08, 12.75], P = .31, I 2 : 64%) (Figure 2A). Funnel plot showed no significant publication bias.
We found 14 studies (n = 2362 patients, mean age: 58.9 years old, males: 79.6%) including data regarding the association between QTd and arrhythmic events. The analysis showed that patients who at some point suffered from an arrhythmic event, have a significantly higher QTd compared to patients without arrhythmic events (mean difference [95% CI]: 19.34 [8.94, 29.75], P = .0003, I 2 : 91%) ( Figure 2B). Funnel plot showed no significant publication bias.

| Subgroup analysis
In an attempt to review the literature in depth, we proceeded in a subgroup analysis, essentially separating the study population, previously described, in two major subgroups: Myocardial infarction (MI) patients and HF patients.

| Quality assessment
The Newcastle-Ottawa Quality Assessment Scale (NOS) was used for quality assessment of the included studies. 37 The NOS which indicated that studies were graded as poor quality if the score was <5, fair if the score was 5 to 7, and good if the score was >8. All included studies were graded with a score >5 while none of the included studies was characterized as having a poor quality (Table S1).

| Rational
Existing literature hinds toward the role of QT dispersion in predicting the trajectory of cardiovascular disease. Yet, a definitive model that would shape current guidelines and would help clinicians with stratification of risk for patients with HF and CAD remains a challenge.
Analyzing the 22 included studies we concluded to the following (a) QTd was significantly associated with arrhythmic events and all-cause mortality but not with SCD in CHD and/ or HF patients (b) In the MI group, we found that higher QTd was associated with increased incidence of arrhythmic events but not associated with increase in all-cause mortality. iii) In the HF group, we found a significant association between QTd and arrhythmic events but we failed to prove an association between

| Pathophysiologic mechanism
Identification of the exact role of QTd in arrhythmogenesis remains a challenge. QTd contributes the heterogeneities of repolarization time in the three-dimensional structure of the ventricular myocardium, which are secondary to regional differences in action potential duration and activation time. 38 The increased QTd seems to act in the absence of abnormalities that have been proposed by Packer et al as mechanisms of sudden death in patients with congestive HF (abnormality in neurohormones, electrolytes, or wall stress). 39,40 An explanation about the association between QTd and arrhythmogenesis is the increased disparity of regional ventricular repolarization times which predisposes to sustained ventricular arrhythmias. 41 The causative role of this inhomogeneity on ventricular arrhythmia has been shown in experimental studies and during programmed electrophysiological studies in humans. 42,43 QTc dispersion has been found to be prolonged in acute MI while the QTc dispersion influenced by the site and size of infarction. 44,45 Myocardial fibrosis could be one local factor leading to greater QTd in these patients. 40 It has been found that in the acute setting of MI, the QTd gradually increases, peaks at day 3 and then falls after few days in most cases. 46,47 As a result, the timing of QT measurements has great importance after MI. Another interesting finding is that the beneficial role of classic HF medications may be attributed to the effect of these drugs in QTd shortening. 48 For example, the beneficial role of carvedilol in HF patients has been attributed to the dose-dependent reduction of QTd. 48 The beneficial effect of drug combinations in QTd in HF patients has also been studied. 49

| Associations
QTd has been associated with arrhythmic events in long QT syndrome, HF patients, CHD, post-MI or hypertrophic cardiomyopathy. 14 A positive association between QTd and left ventricular ejection fraction has been demonstrated in MI patients. 50 Additionally, the computerized measurements of QTd have been proposed as a tool for noninvasive risk stratification of patients at higher risk of cardiovascular mortality as indicated by the results of the community-based Strong Heart Study, 51 while the prolongation of the corrected QTd after hemodialysis has been found to predict cardiovascular mortality in hemodialysis patients. 52 Other repolarization markers such as T peak -T end interval have been found to be significant higher in individuals who are at elevated risk for adverse events in congenital LQTS. 53 In the same context, late potentials have been associated with ventricular arrhythmia in patients with MI but there was no correlation between late potentials and QTd, possibly because of the fact that they reflect different electrophysiological disorders (late potentials identify mainly depolarization disorders while QT dispersion is related to repolarization disorders). 45 Several studies reported a significant shortening in QTd following successful coronary revascularization or thrombolysis. 47,54 The QTd shortening and the subsequent reduction of ventricular arrhythmia inducibility may be one of the mechanisms of arrhythmic events reduction in patients undergo thrombolytic or percutaneous coronary intervention (PCI) treatment. In particular, the absolute corrected QTd change after PCI has been found to be significantly correlated with major adverse cardiac events in patients with ST-elevation MI. 55 Another study concluded that early QTd reduction after primary PCI has been found to be closely related to the restoration of reperfusion at the microvascular level and provides additional prognostic information in ST-elevation MI patients. 56 Furthermore, a significant decrease in QTd has been proposed to provide an additional electrocardio-

| CON CLUS IONS
QTd is associated with a higher incidence of major arrhythmic events in patients with HF or MI but is not associated with an increased incidence of SCD or all-cause mortality in this population.

| LI M ITATI O N S
Our study included observational studies some of them with retrospective analysis of their data which may introduce an element of bias. However, by excluding the retrospective studies, we retrieved similar results. Potential unmeasured external confounding factors could pose another limitation to our analysis.

ACK N OWLED G M ENTS
We thank Antonis Armoundas for the proofreading and his significant comments which improved our study.