A novel study on SARS‐COV‐2 virus associated bradycardia as a predictor of mortality‐retrospective multicenter analysis

Abstract Background SARS‐CoV2 has affected more than 73.8 million individuals. While SARS‐CoV2 is considered a predominantly respiratory virus, we report a trend of bradycardia among hospitalized patients, particularly in association with mortality. Methodology The multi‐center retrospective analysis consisted of 1053 COVID‐19 positive patients from March to August 2020. A trend of bradycardia was noted in the study population. Absolute bradycardia and profound bradycardia was defined as a sustained heart rate < 60 BPM and < 50 BPM, respectively, on two separate occasions, a minimum of 4 h apart during hospitalization. Each bradycardic event was confirmed by two physicians and exclusion criteria included: less than 18 years old, end of life bradycardia, left AMA, or taking AV Nodal blockers. Data was fetched using a SQL program through the EMR and data was analyzed using SPSS 27.0. A logistic regression was done to study the effect of bradycardia, age, gender, and BMI on mortality in the study group. Results 24.9% patients had absolute bradycardia while 13.0% had profound bradycardia. Patients with absolute bradycardia had an odds ratio of 6.59 (95% CI [2.83–15.36]) for mortality compared with individuals with a normal HR response. The logistic regression model explained 19.6% (Nagelkerke R2) of variance in the mortality, correctly classified 88.6% of cases, and was statistically significant X2 (5)=47.10, p < .001. For each year of age > 18, the odds of dying increased 1.048 times (95% CI [1.25–5.27]). Conclusion The incidence of absolute bradycardia was found in 24.9% of the study cohort and these individuals were found to have a significant increase in mortality.


| INTRODUCTION
In December 2019, a novel betacoronavirus was discovered from a cluster of patients identified to have pneumonia due to an unknown cause.
Discovered through unbiased sequencing from human airway epithelial cells, and named 2019-nCoV, this virus had different characteristics than the previously discovered MERS-CoV and SARS-CoV coronaviruses.
Acute hypoxic respiratory failure is the most serious manifestation of illness leading to multi organ failure and death in these cases.
While SARS-CoV2 is considered a respiratory virus, cardiac manifestations occur more frequently than previously anticipated. These include cardiomyopathy, myocardial infarction, coronary thromboembolism, and arrhythmias among others. [1][2][3] Acute cardiac injury is the most commonly reported manifestation. Mehrbod et al performed a meta-analysis of acute cardiac injury in COVID-19 and found the incidence to be 15%. 1 Less commonly reported were the complications with cardiac arrhythmias, specifically tachy-and bradyarrhythmias. In a case series of 138 patients from Wuhan, China, 17% had an arrhythmia-related complication. 2 In an analysis of 700 COVID-19 positive patients from University of Pennsylvania, they identified 8% of patients who had an arrhythmia-related event. 15 Gopinathannair et al sent a survey out to cardiac electrophysiologist specialist around the world, and of the 683 responses, the most common arrhythmia reported was atrial fibrillation (21%). Specifically with bradycardia, the most common arrhythmia found was sinus bradycardia (8%) and complete heart block (8%). 3 Bradycardia in COVID-19 appears to be an important cardiac manifestation that has a multifactorial etiology and is increasingly becoming a characteristic clinical finding in patients with COVID-19. It has been described in isolated case reports, 4,5 a limited case series 6 defined as relative bradycardia with heart rate (HR) < 90 beats per minute (BPM) 7 and through a survey sent among cardiac electrophysiologists. 3 In Southern California, there have been a multitude of patients with COVID-19 with variable presentations. Our clinicians and cardiology colleagues noticed a significant trend of bradycardia among our hospitalized COVID-19 patients during our initial case surge in March 2020. By definition, in cases of SIRS we expect a tachycardia response.
As our clinical observation differed from the above convention, our team decided to investigate the bradycardic response in COVID-19.
Below, we report the first novel study to our knowledge with a sample size of over 1000 cases that comments on the prevalence of absolute bradycardia (HR < 60 BPM) and profound bradycardia (HR < 50 BPM) within a multi-hospital COVID-19 population with a particular association with mortality.

| METHODOLOGY
We performed a multi-site retrospective analysis, which included seven Southern California hospitals, on patients that had a COVID-19 diagnosis verified by PCR from March through August 2020 ( Figure 1). A total of 1143 patients were identified. Patients who were less than 18 years of age, had end of life bradycardia, were on AV nodal blocking agents, or left against medical advice were excluded from the study. Both the inclusion and exclusion criteria of a patient was verified by two physicians. A total of 1053 patients were included in final analysis. A trend of bradycardia was noted in our cohort population. Absolute bradycardia and profound bradycardia was defined as a sustained HR < 60 BPM and < 50 BPM, respectively, on two separate occasions, a minimum of 4 hours apart during hospitalization.
Two physicians confirmed bradycardia events by reviewing telemetry strips as well as admission EKG to rule out any underlying arrhythmia.
The oxygen saturation during the bradycardic event was noted to rule out any hypoxia-driven bradycardia.
Chi-square tests were performed in intergroup comparisons of categorical variables, and categorical variables were expressed as numbers, and percentages. Event rates as descriptive statistics were calculated by dividing the total number of events by total number of cases and were reported in percentages. A logistic regression analysis was performed to study the relationship between mortality and bradycardia in the study group and the effect of age, gender, race and BMI on that relationship. The effect was expressed in terms of Odds ratio with 95% CI. The calculations were performed using IBM SPSS Statistics version 27. Table 1 shows the breakdown of baseline characteristics (Age, gender, and BMI) of the study population, all of which appeared to be F I G U R E 1 Study design. *Excluded due to end of life bradycardia, on AV nodal blockers, age < 18, or left AMA comparable. The age ranged from 19-110 years old. The mean age ± SD between the bradycardia, normal, and tachycardia heart rate groups were 61.82 ± 17.01, 61.05 ± 18.96, 59.3 ± 18.41 respectively.

| Baseline study characteristics
There were 468 female patients and 585 male patients, total. Within the cohorts there were 115 females, and 147 males in the bradycardia group.
In the group with a normal HR throughout admission there were 113 females, and 102 males. In the final group of patients with tachycardia there were 240 females, and 336 males. Race was evaluated and divided into white, and non-white. There was a total of 668 white patients and a total of 385 non-white patients. The mean BMI across all three cohorts was 30, which classically is characterized as obese. Mean ± SD admission heart rates, systolic blood pressures, and temperatures were 98.6± 28.3, 130.2 ± 25.2 , and 100.3 ±1.8 respectively for the data set.
In this study, there were 477 patients (45.3%) who had a nontachycardic response to COVID-19. There were 262 (24.9%) verified COVID-19 patients with absolute bradycardia, and 137 (13.0%) who had profound bradycardia. The COVID-19 database was then separated into three cohort groups based on heart rate; bradycardia, normal heart rate, and tachycardia as defined above, to compare the findings of absolute bradycardia and normal heart rate response.

| Associated mortality in patients with COVID-19 and absolute bradycardia
In this study population the overall mortality, (Supplemental Figure 2)

| Associated mortality in patients with COVID-19 and age > 60
The above regression revealed that for every year of age > 18, the odds of mortality due to COVID-19 increased by 1.048 times. Table 4  were not ordered daily on patients, so we were unable to correlate the inflammatory markers with events of bradycardia (Supplemental Table 5).
Statistical analysis of limited data demonstrated that LDH and CRP were significantly higher in the non-survivor versus survivor group. LDH and CRP were also significantly higher in the marked bradycardia group versus normal HR group. In the non-survivor group D-Dimer and CPK were also significantly higher when compared with the survivor group. This finding differs from the marked T A B L E 3 Regression coefficients obtained from the logistic binomial regression model studying the predictors of mortality in the study group In addition, the authors attempted to look retrospectively at EKGs or echocardiograms obtained at the time of a marked bradycardia event. Unfortunately, a complete data set could not be obtained as these patients were part of the initial COVID surge when the main focus was to limit staff exposure to COVID. Only 34 echocardiograms were obtained among individuals who had a profound bradycardia event, with a mean ejection fraction of 62% with a SD ± 0.08%.
There is limited data on COVID-19 patients with bradycardia and conflicting recommendations on the management. While The European Society of Cardiology suggested an initial trial of isoprenaline and atropine, a recent Italian paper proposed an early permanent pacemaker implantation to avoid the risk of infection with a temporary pacemaker. 11,12 Our study was limited by its retrospective, cohort design. All patients with documented COVID-19 infection from the seven southern California hospitals were initially included in our study. The data collection was dependent on the results manually entered into the EMR, which caused some discrepancies. One such discrepancy noted was the subjective reporting of race in the EMR; patients were categorized as white vs non-white based on information entered by admission personnel, and we were unable to independently verify the accuracy of the information.
Lastly, this study can only comment on an association between bradycardia and mortality and not a causation, due to the retrospective design. In the future, the timing of the inflammatory markers with the bradycardic event, the timing of the onset of bradycardia, and COVID-19 medications, such as Remdesivir and Dexamethasone, should be researched further to help us determine the pathophysiology of the bradycardia.

| CONCLUSION
The clinical implications of this novel descriptive study include the following: (1) This is the first large scale study to report an absolute bradycardia response in COVID-19 patients.
(2) Within the bounds of our study, no obvious cause for bradycardia was identified in these patients suggesting a primary toxic effect of COVID-19 infection on heart rate.
(3) Absolute bradycardia events during hospital admission were associated with an odds ratio of 6.59 for mortality as compared with patients who had a normal heart rate response, therefore we recommend early triage and appropriate level of care in this highrisk group.

| Clinical perspective
This retrospective cohort study reports that a quarter of hospitalized COVID-19 patients had an absolute bradycardia event. Cases with a bradycardia event are particularly associated with higher odds of mortality compared to those with normal heart rates. There was no obvious cause of bradycardia identified in these patients, suggesting a primary toxic effect of COVID-19 infection on heart rate. Therefore, it is imperative to identify and triage these patients earlier as they are a high-risk group for potentially adverse outcomes. In the future, elevation of inflammatory marker in correlation with onset of bradycardia can be elucidated in detail. The impact of COVID-19 medications, such as Remdesivir and Dexamethasone, can also be investigated further to help determine the effect on bradycardia.

CONFLICT OF INTEREST
The author(s) declare no conflict of interest for this work.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.