Ventricular repolarization heterogeneity in patients with COVID‐19: Original data, systematic review, and meta‐analysis

Abstract Background Coronavirus disease‐2019 (COVID‐19) has been associated with an increased risk of acute cardiac events. However, the effect of COVID‐19 on repolarization heterogeneity is not yet established. In this study, we evaluated electrocardiogram (ECG) markers of repolarization heterogeneity in patients hospitalized with COVID‐19. In addition, we performed a systematic review and meta‐analysis of the published studies. Methods QT dispersion (QTd), the interval between T wave peak to T wave end (TpTe), TpTe/QT (with and without correction), QRS width, and the index of cardio‐electrophysiological balance (iCEB) were calculated in 101 hospitalized COVID‐19 patients and it was compared with 101 non‐COVID‐19 matched controls. A systematic review was performed in four databases and meta‐analysis was conducted using Stata software. Results Tp‐Te, TpTe/QT, QRS width, and iCEB were significantly increased in COVID‐19 patients compared with controls (TpTe = 82.89 vs. 75.33 ms (ms), p‐value = .005; TpTe/QT = 0.217 vs. 0.203 ms, p‐value = .026). After a meta‐analysis of 679 COVID‐19 cases and 526 controls from 9 studies, TpTe interval, TpTe/QT, and TpTe/QTc ratios were significantly increased in COVID‐19 patients. Meta‐regression analysis moderated by age, gender, diabetes mellitus, hypertension, and smoking reduced the heterogeneity. QTd showed no significant correlation with COVID‐19. Conclusion COVID‐19 adversely influences the ECG markers of transmural heterogeneity of repolarization. Studies evaluating the predictive value of these ECG markers are warranted to determine their clinical utility.


| INTRODUCTION
Coronavirus disease 2019  is caused by the seventh member of the coronavirus family that can infect humans. 1 The first case of COVID-19 was detected in Wuhan, China in December 2019 and the disease soon spread across the countries, causing a devastating pandemic. COVID-19 has been associated with significant cardiovascular comorbidities. 2 Numerous investigations have been conducted to find the potential risk predictors of major adverse cardiac events.
Direct myocardial damage by the virus, systemic inflammatory state, catecholaminergic response to respiratory distress, adverse effects of therapeutic agents, and deterioration of the underlying cardiovascular disorders can potentially increase the risk of malignant ventricular arrhythmias (VA). 3 In addition, QT interval prolongation has been seen in patients with COVID-19 regardless of taking medications that prolong the QT interval. [4][5][6][7] The role of cardiac muscle repolarization heterogeneity in developing VA has been well studied. 8 Electrocardiogram (ECG) repolarization markers such as QT dispersion (QTd) and the interval between T wave peak to the T wave end (TpTe) have been suggested as indicators of regional and transmural heterogeneity of myocardial repolarization. These ECG markers have been reported to be greatly influenced by systemic viral infections [9][10][11] and systemic inflammatory disorders. 12,13 However, the effect of COVID-19 on cardiac repolarization is not well established. This study was conducted to investigate the effect of COVID-19 on various ECG indicators of repolarization heterogeneity. had negative PCR tests. The patients' demographics, lab results, and admission ECGs were available for all cases and controls. Patients with uninterpretable ECG, complete bundle branch block, nonsinus rhythms (e.g., atrial fibrillation), symptoms of acute coronary syndromes, symptomatic heart failure with ejection fraction (EF) ≤ 40%, outpatient use of QT-prolonging agents (e.g., fluoroquinolones, azithromycin, etc.) and electrolyte imbalances on admission were excluded. The demographic data and medical history were extracted from the patients' charts. The medications prescribed in the outpatient setting were also collected.
The laboratory data including white blood cell count, neutrophil to lymphocyte ratio (NLR), hemoglobin concentration, platelet count, plasma level of C-reactive protein (CRP), and quantitative levels of troponin T and creatinine were recorded for COVID-19 cases. This study posed no additive charge or harm to the study population. It was approved by the ethics committee of Tehran University of medical sciences. All patients' identifiers were removed to comply with Health Insurance Portability and Accountability Act regulations. Informed consent was obtained from all patients before the study. The QT interval was considered as the distance between the first deflection of the QRS complex and the end of the T wave defined as the intersection of the tangent to the steepest downslope of the T wave and the isoelectric line. The longest QT and QRS intervals were used (usually mid-precordial V3) to prevent underestimation of isoelectric T waves. In case of a discrete U wave in a precordial lead, that value was omitted.

| ECG measurements
The TpTe interval was defined as the interval between the T wave peak to the T wave end in lead V5. 14 All 12-lead values were recorded and the difference between the longest and the shortest QT and TpTe values in a single beat was defined as QT dispersion and TpTe dispersion, respectively. QT intervals were corrected for heart rate using both Bazett's and Framingham's methods. Bazett's formula was the most common method used by previous studies and this method is the most homogenous method to perform the meta-analysis and the Framingham's formula is known to be the most accurate one. 15 To assess the contribution of repolarization heterogeneity duration to the total duration of the action potential, TpTe was divided by QT and QTc and the index of cardio-electrophysiological balance (iCEB) was defined as the QT to QRS ratio.

| Statistical analysis for original study
Continuous variables are presented as mean ± SD and categorical variables are reported as frequency and percentage. Student t test and χ 2 test were used to compare continuous variables and categorical variables, respectively. Pearson correlation analysis was applied to assess the correlation of ECG markers with laboratory and clinical data. A 2-sided p value of < .05 was considered statistically significant. Statistical analyses were performed using the SPSS software version 25 (IBM Corp.).

| Systematic review search strategy
The current review followed the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guideline. The The studies that presented the mean and standard deviation for any of the four ECG variables including "QTd," "TpTe," "TpTe to QT ratio," and "TpTe to QTc ratio" in COVID-19 patients were included. All types of observational studies were included. The search results were screened using their titles, abstracts, and fulltexts in more relevant studies. Duplications were removed and the full texts of all the included studies were obtained. The quality of the studies was assessed using Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) checklists at the study level (available as Supporting Information Material). The number of cases and controls, the mean and standard deviation of QTd, TpTe, TpTe/QT, and TpTe/QTc in the study groups, and the characteristics of the study population including age, gender, hypertension, diabetes mellitus, smoking, study location, and COVID-19 hospitalization status were extracted from each study by two independent investigators. In case of a discrepancy, a consensus was achieved by discussion.

| Meta-analysis of overall data
The values were summarized as mean differences (MDs) with a 95% confidence interval. MDs were standardized using Cohen's method and random-effects meta-analysis was performed on each target ECG marker by the Stata software (Stata/MP 16.0; StataCorp LLC).
Statistical heterogeneity was evaluated using the Higgins' I 2 test based on Cochrane's Q. Higher Q values with p values less than .1 were considered as significant heterogeneity. An I 2 value of <25%, 25%-50%, 50%-75%, and >50% indicated absent, low, moderate, and high heterogeneity, respectively. 16 Subgroup analysis and metaregression were performed in case of severe heterogeneity: 1. Age equal to or above 18 and lower than 18; 2. Male sex more than 55% of the study population, less than 45%, and equal to 50 ± 5%; 3.
Country where the study was conducted; 4. Only hospitalized cases were included, only outpatient cases were included, both types of patients were included. Publication bias was evaluated through visual assessment of funnel plots as well as the Egger, Begg, and "trim-andfill" tests with p values < .1 indicating publication bias. Sensitivity analysis was performed by excluding individual studies from the meta-analysis.

| RESULTS
The study population included 101 patients hospitalized due to COVID-19 and 101 gender-and age-matched controls. Seventy percent of the cases and controls were male with a mean age of 60.11 ± 16.16 years for cases and 61.1 ± 17.22 years for controls.
The patients' characteristics are summarized in Table 1.
T A B L E 1 Demographic characteristics of our study population ECG parameters were also corrected using the Bazett's equation for use in the meta-analysis. These values are reported separately in Table S1. Possible correlations between ECG markers and laboratory and clinical data were investigated. The oxygen saturation level was positively correlated with the TpTe interval and TpTe/QTc (r 2 = .28, p-value = .049 for TpTe and r 2 = .30, p-value = .035 for TpTe/QTc, Figure S1A). The platelet count had a negative correlation with TpTe/ QT and TpTe/QTc (r 2 = −.28, p-value = .021 for TpTe/QT and r 2 = −.27, p-value = .025 for TpTe/QTc, Figure S1B) and the CRP level had positive correlation with TpTe dispersion (r 2 = .28, p-value = .025 for TpTe, Figure S1C).

| Systematic review and meta-analysis
Eight studies were considered eligible for inclusion in the current meta-analysis and the original data of the present study were added to these eight studies. 4,[17][18][19][20][21][22][23] The flow diagram of the screening process is presented in Figure 1.
After excluding a study by Rubin et al. 4 (whose characteristics were not reported), the overall mean age of the subjects was 42.43 ± 11.75 years for patients and 42.97 ± 11.70 years for controls.
About 58% of the patients and 54% of the controls were male. The prevalence of hypertension, diabetes mellitus, and cigarette smoking was 36%, 27%, and 28% in patients and 29%, 14%, and 21% in controls, respectively. Six studies were performed in six different Turkish cities, one study was conducted in Iraq, one was carried out in the United States, and our study was performed in Iran. The characteristics of the studies are summarized in Table S2.
The overall results of the meta-analysis are summarized in Figure 2.
We observed that the QTd tends to be higher in COVID-19 cases compared with controls with a p value of .1 (QTd SMD = 0.84, p-value = .10, Figure S2); however, there was substantial heterogeneity among studies and subgroup and meta-regression analysis did not improve it.
The TpTe, TpTe/QT and TpTe/QTc were significantly increased in COVID-19 patients compared with controls (TpTe SMD = 0.75, p-value = .01, Figure S3; TpTe/QT MD = 0.03, p-value < .01, Figure S4; TpTe/QTc MD = 0.02, p-value = .01, Figure S5).  26 TpTe interval was introduced as the interval between termination of repolarization in the first and the last cells across a transmural section of the myocardium (epicardial and Mcells, respectively). 27 TpTe and its proportion to QT interval (TpTe/ QT) provide a relative (but not absolute 28  Myocardial injury in the setting of ischemia and viral myocarditis have been previously shown to increase repolarization heterogeneity and prolong the TpTe and TpTe/QT intervals. 8,11,31 In terms of COVID-19, cardiac injury is a common complication during hospitalization 32 and it is not only a cause of death due to myocardial dysfunction and malignant VAs but it is also a prognostic indicator of poor outcomes. 2  Sympathetic activation due to respiratory distress, hypoxia, and fever can increase the markers of repolarization heterogeneity. 38 In addition, the particular influence of inflammation on repolarization heterogeneity has been previously ascertained in patients suffering from systemic inflammatory disorders 12,13,39,40 and the TpTe interval has been correlated with plasma levels of inflammatory biomarkers in HIV patients. 9 In the setting of COVID-19, the CRP level was correlated with TpTe and TpTe/QT ratio in studies conducted by Yenercag et al. 23 and Koc et al. 21 ; however, the results of a study by Colkesen et al. 19 and our findings did not indicate such a correlation. Our current study also found a significant positive correlation between TpTe dispersion and plasma CRP level and a nonsignificant tendency of the TpTe dispersion to be higher in patients with increased NLR (r 2 = .185, p-value = .128). In addition, TpTe/QT and TpTe/QTc ratios were negatively correlated with the platelet count, which can be secondary to bone marrow suppression and the resultant thrombocytopenia in the setting of septicemia and systemic inflammation.
The multifaceted effects of COVID-19 on myocardial repolarization heterogeneity markers are illustrated in Figure 3. Together with the focus on the pathophysiology of COVID-19, the baseline characteristics of patients should be also cautiously explored to prevent misconceptions due to their confounding effects on the ECG markers. 41 The age is a positive moderator of the TpTe and TpTe/QT intervals in low-risk populations. 42 This association was not found in our study population, which may be caused by the fact that even our controls were not selected from a low-risk population.
In terms of gender-specific differences in myocardial electrophysiologic properties, androgen deprivation and anti-androgen treatment have been associated with a prolonged TpTe interval and MAHMOUDI ET AL.
| 115 administration of testosterone can shorten this interval. 43 In addition, women have shown greater drug-induced prolongation in the TpTe interval. 44 We did not observe a direct correlation between the gender status of subjects and their ECG values; although subgroup analysis of studies with male predominance considerably reduced between-study heterogeneity.
Among the common comorbidities, diabetes mellitus has been associated with a tendency to increase the TpTe dispersion and TpTe/ QT ratio. 45 LV hypertrophy, diastolic dysfunction, LV mass index, and elevated systemic blood pressure are also associated with an increase in the TpTe interval and TpTe/QT ratio. 46,47 Moreover, chronic tobacco users develop a transient increase in TpTe interval following each episode of cigarette smoking. 48 As the studies included in the present meta-analysis had heterogeneous populations in terms of age, sex, and the mentioned comorbidities (Table S2), metaregression analysis moderated by these factors considerably improved the heterogeneity.
Eventually, the drug-induced changes in repolarization heterogeneity markers and their role in the development of VA are well studied 49 and so is the repolarization heterogeneity associated with malignant VA in the setting of acute coronary events. 31 In this study, these confounding factors were relatively corrected by excluding the affected cases. The same exclusion strategy was followed by all the studies included in our meta-analysis except for two studies by Ece et al. 20 and Cevik et al., 18 which were conducted in pediatric age groups. to the surgery ward whose COVID-19 was rolled out to prevent complications and they all had ECG by routine. In addition, investigating the prognostic value of ECG markers in predicting the in-hospital or long-term adverse cardiovascular events including arrhythmias and mortality could have been of great value but a larger sample of patients was needed to perform a statistically reliable analysis of these relatively rare outcomes.

| CLINICAL PERSPECTIVE
• According to previous studies, cardiovascular complications have imposed considerable burden to patients with COVID-19.
• Repolarization heterogeneity is increased in patients with COVID-19, before initiation of QT-prolonging agents.
• Myocarditis, inflammation, Thrombophilia, Coronary ischemia, Hypoxia, adrenergic stimulation and therapeutic agents are possible mediators influencing repolarization duration and heterogeneity in patients with COVID-19.
• TpTe, TpTe/QT, QRS width and iCEB should be considered as risk markers of arrhythmia in patients with COVID-19.

ACKNOWLEDGMENTS
The authors want to appreciate all healthcare workers and researchers who stood still besides the humanity during this pandemic.

CONFLICT OF INTERESTS
The authors declare that there are no conflict of interests.

AUTHOR CONTRIBUTIONS
Pejman Mansouri, Shayan Mirshafiee, Behnam Hedayat, and Mojtaba Salarifar provided the original data summarized in Tables 1 and 2

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.