ABSTRACT
Objective:
Coronavirus disease-2019 (COVID-19) is associated with atrial fibrillation (AF) and ventricular arrhythmias. Several electrophysiological abnormalities on surface electrocardiography (ECG) are associated with AF and ventricular arrhythmias, either as markers of abnormal interatrial conduction or abnormal repolarization. The present study sought to understand whether such ECG markers are more common in patients hospitalized with COVID-19 infection during the pandemic.
Methods:
A total of 87 COVID-19 patients formed the study group, whereas 64 patients who were hospitalized for any reason other than COVID-19 infection served as controls. The frequency of partial and advanced interatrial block (IAB), QT and corrected QT (QTc) durations, QT dispersion (QTd), and T peak-to-end duration (Tpe) were measured from ECGs at admission.
Results:
Both partial and advanced IAB were more common in patients with COVID-19, although statistical significance was only observed for advanced IAB (11.5% in COVID-19 patients vs. 0.0% in controls, p=0.005). There were no differences between the groups for QTc, QTd or Tpe. On Bayesian analyses, there was strong evidence favoring an association between COVID-19 and advanced IAB (BF10:16), whereas there was no evidence for an association for partial IAB, QTc, QTd, or Tpe (BF10<1 for all).
Conclusions:
Patients hospitalized with COVID-19 were more likely to have advanced IAB, which may explain why AF is more frequent in these patients.
INTRODUCTION
Coronavirus disease-2019 (COVID-19), which is caused by the novel coronavirus severe acute respiratory syndrome coronavirus 2 (SARS-COV-2), has affected more than 750 million people and caused nearly 7 million deaths since its initial dissemination from Wuhan, China, in 20191. Although the pandemic is over, endemic infections are common worldwide, and the long-term consequences of SARS-COV-2 infection are largely unknown. Because the disease affects microvascular circulation and increases thrombogenicity during the acute phase, COVID-19 is associated with atherothrombotic and thromboembolic events, including myocardial infarction2,3,4. Moreover, there have been sporadic cases of myocarditis, and the viral genome has been isolated from myocardial cells from autopsy specimens5,6,7. Arrhythmias, particularly atrial fibrillation (AF), are frequent in hospitalized cases with a prevalence ranging between 11% and 19%8. Ventricular tachyarrhythmias, albeit less common, are still seen in a sizable part of the patients9.
A particular concern for patients hospitalized with COVID-19 is to predict their predilection for cardiovascular complications. P-wave prolongation and reversal on surface electrocardiography (ECG) indicate partial or advanced interatrial block (IAB). The presence of IAB indicates abnormal atrial depolarization due to a dysfunctional Bachmann’s bundle, and IAB is a risk factor for atrial tachyarrhythmias, particularly AF10. Corrected QT (QTc) duration and global QT dispersion (QTd) reflect prolongation of ventricular repolarization and local differences in the duration of repolarization, and both markers are linked to an increased predisposition to ventricular arrhythmias11,12. All these markers may assist in predicting the tendency for atrial and ventricular arrhythmias in COVID-19 patients.
Unfortunately, data on surface ECG markers for arrhythmias in COVID-19 patients are scarce. Therefore, we aimed to understand whether IAB, QTc, or QTd was more common in patients hospitalized with COVID-19 than in controls and to determine the strength of evidence favoring these associations.
MATERIALS and METHODS
This was a cross-sectional observational study. Patients admitted to the study center between January 2020 and June 2020 and subsequently hospitalized with either SARS-COV-2 nucleic acid test positivity with symptoms compatible with acute upper or lower respiratory track infection together with systemic inflammatory response or viral pneumonia compatible with COVID-19 disease were consecutively enrolled. Patients under 18 years old, those with a pacemaker rhythm at the time of ECG recording, and those who had an acute coronary event, pulmonary embolism, or any other type of cardiac emergency were excluded from the study. In addition, patients with a diagnosis of AF (previous or current), those with conditions associated with repolarization abnormalities (pacemaker rhythm, bundle branch block), and those with a rhythm other than sinus were excluded. After applying these inclusion/exclusion criteria, 87 out of 95 patients with COVID-19 who were initially screened were included in the analysis. 64 patients who were hospitalized for a reason other than COVID-19 served as controls. Patients’ demographic and clinical data were collected via direct interviews and/or measurements or using the institutional electronic medical database.
The study was conducted according to the principles of the 1975 Declaration of Helsinki, and all patients provided informed consent before enrollment. The study was approved by the University of Health Sciences Turkey, Haydarpasa Numune Training and Research Hospital Ethics Committee (decision no: HNEAH-KAEK 2020/152, date: 31.08.2020).
Recording and Interpretation of Electrocardiograms
A 12-lead ECG was obtained at admission from all patients and recorded on a thermal ECG paper with a sweep speed of 25 mm/s with the standard gain setting using standard methods (10 mv/mm)12. All ECGs were interpreted by an experienced cardiologist (R.C.G.) and measurements were taken as described before13,14. IABs were classified into partial and advanced IAB according to the P-wave duration and reversal of the terminal portion of the P-wave in the inferior leads15. QT and T wave peak-to-end (Tpe) intervals were measured as described previously. The Bazett formula was used to calculate the QTc interval. QTd was calculated by extracting the maximum QT interval from the minimum QT interval.
Statistical Analysis
Continuous variables are given as mean ± standard deviation or as median and interquartile range, whereas categorical variables are given as percentages. Patterns of distribution and equality of variance for continuous variables were analyzed using Shapiro-Wilk and Levene’s tests, respectively. For continuous variables, comparisons between groups were performed using the t-test or Mann-Whitney U test as appropriate, while chi-squared or Fisher’s Exact tests were used for categorical variables. Bayesian tests were performed to understand the evidence supporting the association of COVID-19 with a given variable. To understand the independent associations between COVID-19 with any type of IAB, QTc, QTd and Tpe, logistic and linear regression models were constructed to adjust for age, gender, presence of coronary artery disease or heart failure, creatinine and electrolytes sodium and potassium. Prior to these analyses, the association between COVID-19 and the aforementioned variables was interrogated using univariate analysis, and regression models were constructed only if the initial association had a p-value <0.1 on univariate analysis. For all frequentist analyses, a p-value <0.05 was set as the limit for significance. For Bayesian analyses, a BF10 between 1 and 3 showed anecdotal evidence supporting the alternative hypothesis, whereas a BF10 approximately 3-10 and more than 10 indicated moderate and strong evidence, respectively. All statistical analyzes were performed using Jamovi 2.3.22 (The Jamovi project. Jamovi 2.3 for MacOS. Retrieved from htpss://www.jamovi.org) and R [R Core Team (2021). R: A Language and Environment for Statistical Computing. Version 4.1 for MacOS. Retrieved from https://cran.r-project.org.]
RESULTS
The mean age of the study sample was 56.5±14.2, and 36.4% (n=55) cases were females. 25.8% (n=39) and 6.6% (n=10) of the study sample had coronary artery disease and heart failure, respectively. Partial and advanced IAB were present in 9.3% (n=14) and 6.6% (n=10) of the cohort, while none of the patients had any degree of atrioventricular block.
Table 1 summarizes the baseline characteristics of the study groups. Patients in the COVID-19 group were older and more likely to be female, but significant cardiac conditions and cardiovascular risk factors such as hypertension, diabetes, and coronary artery disease were more frequent in the control group. Notably, inflammatory parameters such as C-reactive protein concentration and white blood cell count were not different between the groups.
Electrocardiographic variables are given in Table 2. The P-wave duration was significantly higher in patients with COVID-19, and advanced IAB was significantly more frequent in this group. Although partial IAB was also more common in COVID-19 patients, this finding did not reach statistical significance (Figure 1).
On univariate regression analysis, the presence of COVID-19 infection had a statistically significant association with any degree of IAB (p<0.004), whereas the p-value was >0.05 for all other associations with QTc, QTd and TPe. Multivariate logistic regression analysis of the association between COVID-19 and any degree of IAB is summarized in Table 3. After adjustment for covariates, COVID-19 remained an independent predictor of IAB [p=0.005, odds ratio (OR): 9.45, 95% confidence interval (CI): 1.95-45.76].
On Bayesian analyses, there was strong evidence to suggest an association between COVID-19 and advanced IAB (BF10: 16.05, logOR: 2.57, 95% CI: 0.77-5.87), as well as an association between COVID-19 and any type of IAB (BF10: 74.94, logOR: 1.70, 95% CI: 0.65-3.00); however, there was no evidence to suggest an association between COVID-19 and partial IAB (BF10: 0.87, logOR: 0.95, 95% CI: -0.19-2.31) (Figure 2). Likewise, there was no evidence to suggest a higher QTc, QTd, or Tpe in patients with COVID-19 than in controls (BF10<1 for all).
DISCUSSION
COVID-19 remains a disease with many unknowns, particularly with respect to cardiovascular involvement and outcomes. The present findings suggest a strong association between COVID-19 infection and IAB (particularly advanced IAB), which may explain why AF is so frequent in patients with COVID-19. In contrast, we have not observed an association between COVID-19 and various ECG parameters considered as risk markers for ventricular tachyarrhythmias, despite the use of various statistical methods to analyze the dataset.
Although the existence of an interconnecting muscle bridge that allows the fast conduction of impulses from right to left atrium - the so-called Bachmann bundle - was long known, the understanding that IAB may be a cause of atrial arrhythmias, particularly AF, is a relatively recent development16. Although left atrial enlargement (LAE), hence prolonged P-wave duration on ECG, is a major risk factor for AF, the association between IAB and AF is largely independent of LAE17,18. The present findings suggest an independent association between COVID-19 disease and IAB, thus providing a potential mechanistic explanation for the higher than expected prevalence of AF in this population. A similar finding was previously noted by Yenerçağ et al.19, although in their study, the investigators only assessed P-wave duration without a detailed analysis of the patterns of LAE of IAB. Thus, the present findings are incremental over these past results and suggest that IAB is at least partly responsible for this increase in P-wave duration. However, because we have not collected echocardiographic data for the purposes of this study, it remains unclear to what degree this increase in P-wave duration can be attributable to IAB or LAE.
Contrary to the findings on IAB, the present study did not suggest a difference in indices of ventricular repolarization in patients with or without COVID-19. Indeed, no statistical differences were found between the groups in adjusted or unadjusted analyzes nor was there any evidence on Bayesian analyzes that would support an increased duration or heterogeneity of repolarization in the COVID-19 group. This is in contrast with previous reports, where it has been suggested that measures of repolarization (including QTc, QTd and Tpe) were significantly prolonged in COVID-19 patients and this finding correlated with overall mortality20,21. A possible explanation for this finding is that the control group was used to assess the statistical significance of changes in the COVID-19 patients. In contrast to other studies, which used healthy volunteers or patients who were presumably less sick than COVID-19 patients, the control group in the present study consisted of patients who either had already established cardiovascular diseases or were at a high risk for cardiovascular disease. Thus, the present findings suggest that most of the variability in the ECG markers of repolarization can be attributable to coexisting conditions in COVID-19 patients, as suggested by others22.
The present findings have practical implications. Given that IAB provides a substrate for the development of AF, it would be reasonable to closely follow-up COVID-19 patients with IAB and sinus rhythm to prevent the devastating consequences of AF. Diuretics and ventricular resynchronization may reverse IAB in patients with heart failure; however, it is unclear whether IAB is reversible in other circumstances23,24. Thus, patients with a history of COVID-19 with IAB may be at an increased life-long risk of AF. Given that other physiological abnormalities, including autonomic dysfunction, are also prevalent in patients with past COVID-19 infection, this risk may be even more pronounced than anticipated25. Conversely, we do not recommend the routine use of indices of ventricular repolarization for COVID-19 patients who are not otherwise at risk of QT prolongation. This recommendation was not solely based on the current findings, as there are significant methodological concerns about the usefulness of these parameters for predicting ventricular arrhythmias26.
The present study has several limitations that should be stated. Patients were enrolled from a single center, and the sample size was rather limited. Because this was an all-comers study and prior data on the prevalence of ECG abnormalities in COVID-19 patients were limited, an a priori power analysis was not feasible. However, a post-hoc power analysis found that the study had a power (1-b) of 0.94 and 0.99 to detect a significant difference between the advanced and any IAB groups, respectively. The diagnosis of COVID-19 was based on nucleic acid tests only in approximately one-third of the sample, whereas the diagnosis was based on the presence of clinical and imaging findings for viral pneumonia in the remaining cases. Although the study was conducted during the peak of the COVID-19 pandemic, some cases might have been infected with an agent other than SARS-COV-2. No data on the severity of COVID-19 were collected; therefore, it was not possible to analyze whether the degree of ECG abnormalities was related to the severity of COVID-19. Although predictive usefulness and prognostic implications of IAB have been demonstrated, no follow-up data were collected for the purposes of the present study; therefore, the outcomes related to IAB in COVID-19 patients remain unknown18.
CONCLUSION
Both partial and advanced IAB are more common in hospitalized COVID-19 patients, which may provide a mechanistic explanation for the higher incidence of AF in this patient population. Because of the proven connection of IAB with AF in other patient populations, a close follow-up is reasonable for patients with COVID-19 and IAB (particularly advanced IAB). The prognostic implications of IAB specific to patients with COVID-19 or the long-term implications of IAB in this patient population remain unknown, and further research is warranted on these topics.
