Data of atrial arrhythmias in hospitalized COVID-19 and influenza patients

Atrial arrhythmias (AA) are common in hospitalized COVID-19 patients with limited data on their association with COVID-19 infection, clinical and imaging outcomes. In the related research article using retrospective research data from one quaternary care and five community hospitals, patients aged 18 years and above with positive SARS-CoV-2 polymerase chain reaction test were included. 6927 patients met the inclusion criteria. The data in this article provides demographics, home medications, in-hospital events and COVID-19 treatments, multivariable generalized linear regression regression models using a log link with a Poisson distribution (multi-parameter regression [MPR]) to determine predictors of new-onset AA and mortality in COVID-19 patients, computerized tomography chest scan findings, echocardiographic findings, and International Classification of Diseases–Tenth Revision codes. The clinical outcomes were compared to a propensity-matched cohort of influenza patients. For influenza, data is reported on baseline demographics, comorbid conditions, and in-hospital events. Generalized linear regression models were built for COVID-19 patients using demographic characteristics, comorbid conditions, and presenting labs which were significantly different between the groups, and hypoxia in the emergency room. Statistical analysis was performed using R programming language (version 4, ggplot2 package). Multivariable generalized linear regression model showed that, relative to normal sinus rhythm, history of AA (adjusted relative risk [RR]: 1.38; 95% CI: 1.11–1.71; p = 0.003) and newly-detected AA (adjusted RR: 2.02 95% CI: 1.68–2.43; p < 0.001) were independently associated with higher in-hospital mortality. Age in increments of 10 years, male sex, White race, prior history of coronary artery disease, congestive heart failure, end-stage renal disease, presenting leukocytosis, hypermagnesemia, and hypomagnesemia were found to be independent predictors of new-onset AA in the MPR model. The dataset reported is related to the research article entitled “Incidence, Mortality, and Imaging Outcomes of Atrial Arrhythmias in COVID-19” [Jehangir et al. Incidence, Mortality, and Imaging Outcomes of Atrial Arrhythmias in COVID-19, American Journal of Cardiology] [1].


Specifications
Cardiology and Cardiovascular Medicine Specific subject area Cardiac Electrophysiology Type of data Table  Figure Supplementary datasheet How the data were acquired Retrospective chart and data review of hospitalized COVID-19 patients and influenza patients meeting the inclusion criteria. Data format Raw Analyzed Description of data collection Data were collected for COVID-19 and influenza patients fulfilling the inclusion criteria. COVID-19 patient records were retrospectively examined. Data pertaining to vital signs, laboratory values, baseline demographics, comorbid conditions, in-hospital COVID-19 treatments, and in-hospital events were electronically extracted from the electronic medical record. Social history, pre-admission medications, chest computed tomography findings, and echocardiographic findings were extracted manually from the electronic medical record for a subset of cases. Likewise, influenza patents records were examined and data were electronically extracted from the EMR on baseline demographics, comorbid conditions, and in-hospital events.
( continued on next page )

Value of the Data
• These data are useful as they provide insights into large and diverse populations of COVID-19 and influenza patients with atrial arrhythmias (atrial fibrillation and atrial flutter) admitted at six hospitals in Southeast Michigan. • Our data identify risk factors of new-onset atrial arrhythmias and mortality in SARS-CoV-2 infection. In addition, we report crucial clinical and imaging findings and in-hospital treatments of COVID-19 patients. • The data is of value to clinicians as there is limited published data on the impact of atrial arrhythmias on chest computerized tomographic and echocardiographic findings in COVID-19 infection. Moreover, literature comparing atrial arrhythmias in COVID-19 to influenza patients is deficient. • Our data present a broad range of outcomes, including mortality, new-onset heart failure, and myocardial infarction in atrial arrhythmia patients with COVID-19, and compare the results to a propensity-matched cohort of influenza patients. These data suggest that COVID-19 is associated with a higher risk of new-onset atrial arrhythmias than influenza. • These data have important clinical implications as new-onset atrial arrhythmias confer an unfavorable prognosis in viral pneumonia, with mortality higher in influenza than COVID-19 infection. Cardiologists, infectious disease specialists, and internists may find this data useful as early identification and treatment of atrial arrhythmias can potentially improve outcomes in viral pneumonia. • Researchers can use data from our multicenter registry to further investigate COVID-19 and influenza patients, compare our results with other studies and perform systematic reviews and meta-analyses.

Data Description
In this study, we collected data of patients who were hospitalized with SARS-CoV-2 and influenza virus infections. The prevalance of atrial arrhythmias (AA) in COVID-19 is reported from 15.8 to 19.6% across academic centers in the United States [2][3][4][5] . In COVID-19 hospitalized patients, AA are independently associated with higher in-hospital mortality [3 , 5] . Moreover, respiratory viruses such as influenza, severe acute respiratory syndrome coronavirus, and SARS-CoV-2 can be associated with decompensated congestive heart failure (CHF) [6][7][8][9][10] . The incidence of AA, mortality, and clinical outcomes associated with AA, including new-onset CHF, were compared between COVID-19 and influenza populations after propensity matching. Moreover, the imaging outcomes, including computerized tomography chest scan findings and transthoracic echocardiographic findings, were studied in COVID-19 patients. Supplementary File: Analyzed data of COVID-19 and influenza patients admitted to the hospitals in Southeast Michigan. Patients were stratified into normal sinus rhythm (NSR), new-onset AA, and history of AA. Detailed data on baseline demographics, comorbid conditions, and inhospital events are reported for COVID-19 and influenza cohorts. Moreover, data on vital signs, laboratory values, social history, pre-admission medications and in-hospital medical treatments, computerized chest tomography (CT), and echocardiographic data are reported for COVID-19 patients. The file can be found on Mendeley. Fig. 1 -Figure showing COVID-19 treatments received during hospitalization including lopinavir, ritonavir, remdesivir, tocilizumab, hydroxychloroquine, azithromycin, and cumulative dosage of steroids including hydrocortisone, dexamethasone, and methylprednisolone. COVID-19 patients were stratified into three groups: NSR, new-onset AA, and AA. Remdesivir and hydroxychloroquine usage were more common in NSR compared to patients with history of AA and new-onset AA. Cumulative dosage of steroids was defined as methylprednisolone 40 mg twice daily or ≥80 mg daily for ≥3 or more days; dexamethasone ≥6 mg daily for ≥3 days, and hydrocortisone ≥50 mg daily for ≥3 days. If a patient on the aforementioned dosage of steroids died before reaching 3 days length of stay, they were included in the steroids regimen group as they likely had severe COVID-19 disease. Patients on the same dosage who were discharged before reaching 3 days length of stay likely did not have severe disease and were excluded from the steroid group. The use of all steroids including hydrocortisone, dexamethasone, and methylprednisolone was more frequent in patients with new-onset AA. The difference in the usage of lopinavir, ritonavir, and tocilizumab did not reach statistical significance.  stratified into three groups in each cohort based on the status of AA. Matches were made within population, first between new-onset and history of AA groups, then between AA (new + history) and NSR groups. Finally matches were made between COVID-19 and influenza populations. After completing the propensity matching, the cohorts had 1632 patients each. Table 1 -Table showing International Classification of Diseases-Tenth Revision (ICD-10) codes and other identification methods used in the study. The ICD-10 codes were used to identify the patients with influenza, AA, chronic heart failure (HF), and the outcomes including new-onset HF, ischemic stroke, transient ischemic stroke, myocardial infarction (MI), deep vein thrombosis (DVT), other arterial thromboembolism, pulmonary embolism, ventricular tachycardia (VT) and ventricular fibrillation (VF), acute renal failure (ARF), requirement for new renal replacement therapy (RRT), and minor and major bleeding (using International Society on Thrombosis and Haemostasis definition) [11] . Standardized text variables were also used to identify patients with AA, chronic and new-onset HF, requirement for RRT, along with transfusion and drop in hemoglobin ≥2 mg/dL. Table 2 -Table showing the baseline demographics and comorbidities of 14,174 hospitalized influenza patients. Patients were stratified into three groups based on the status of AA. Among the influenza cohort, 12,325 remained in NSR, 1,499 patients had history of AA, whereas 350 patients experienced new-onset AA. The Kruskal-Wallis test was used for age, and Chi-square tests were used otherwise. All the baseline characteristics were significantly different between the groups ( p < 0.001). Table 3 -Table showing the medical treatments received by COVID-19 patients with newonset AA and history of AA during hospitalization. Chi-square test or Fisher-exact test was used based on expected cell counts. A total of 67% of patients with new-onset AA and 65% patients with history of AA received rate controlling agents during hospitalization with no statistical difference between the groups. The usage of rhythm controlling agents was more frequent in patients with new-onset AA (25%) compared to history of AA (21%).   Anticoagulation was not given in 13.1% of patients with new-onset AA, 9.2% with history of AA, and 2.8% with NSR. Table 5 -Table showing three groups within the COVID-19 population after propensity matching between the three groups. The Kruskal-Wallis test was used for age, and Chi-square tests were used otherwise. A 1:1 match (history of AA vs new-onset AA) was used, followed by a 1:1 match between AA (new-onset AA + history of AA) and NSR groups. All p -values are insignificant indicating successful matching. Table 6 -Table showing three groups within the influenza population after propensity matching between the three groups. The Kruskal-Wallis test was used for age, and Chi-square tests were used otherwise. Since there were few new-onset AA cases in influenza population, a 2:1 match (history of AA vs new-onset AA) was used, followed by a 1:1 match between AA (newonset AA + history of AA) and NSR. All p-values are insignificant indicating successful matching.   ( Tables 5 and 6 ). All p-values are insignificant indicating successful matching. Table 8 -Table showing the pre-hospital medications in the study population. Usage of statins, warfarin, direct oral anticoagulants, digoxin, beta-blockers, and diuretics were more common in patients with history of AA whereas antiplatelets usage was more common in patients with new-onset AA. There was no difference in the use of angiotensin-converting-enzyme inhibitors, angiotensin receptor blockers, calcium channel blockers, azithromycin, and hydroxychloroquine between the study groups. Table 9 - Table showing multivariable generalized linear regression model using a log link with a Poisson distribution (multi-parameter regression [MPR]) built to identify predictors of new-onset AA. The model was built using baseline demographic characteristics (age, sex, race, body mass index), comorbid conditions (hypertension, diabetes mellitus, CHF, cerebrovascular accident, kidney disease, pulmonary disease, pulmonary hypertension, liver disease, cancer, thyroid disease, and history of DVT) and on-arrival labs (white cell count, aspartate aminotransferase (AST), D-dimer, potassium, and magnesium) which were significantly different between the groups, and hypoxia in the emergency room. Adjusted relative risk (RR) with 95% confidence intervals (CI) were calculated. Significant variables included: age in increments of 10 years (RR:  Table 10 -Table showing chest CT findings in study patients. Data on pleural effusion, ground-glass infiltrates, multifocal pneumonia, pulmonary edema, and pulmonary vascular congestion was collected. Pleural effusions were most common in patients with history of AA (54.6%) compared to patients with NSR (13.8%) and new-onset AA (13.3%); p -value = 0.02. The difference in findings of pulmonary edema, pulmonary vascular congestion, ground-glass opacities and multifocal pneumonia did not reach statistical significance.   length of stay and higher incidence of intensive care unit (ICU) admission, need for mechanical ventilation, usage of vasopressors and inotropes, new-onset CHF, ARF, and VT compared to patients with history of AA and NSR. Similarly, in the influenza population, the need for mechanical ventilation, usage of vasopressors and inotropes, new-onset CHF, ST-segment elevation MI, non-ST segment elevation MI, ARF, VF, and VT were more common in patients with new-onset AA. Table 13 -Table showing odds ratios (OR) of the in-hospitals events in 3 groups after propensity matching between COVID-19 and influenza populations. Odds ratios were calculated for each 2-group comparison using univariate logistic regression within the COVID-19 and influenza populations. Group 1 includes patients with NSR, group 2 includes patients with new-onset AA, whereas group 3 includes patients with history of AA. Within the COVID-19 population, newonset AA had higher ICU admission rate, 90 day-readmission, need for mechanical ventilation, vasopressors and inotropes usage, new-onset CHF, non-ST-segment elevation MI, ARF, and VT as evident by OR with 95% CI not crossing 1 and p -value < 0.05. Similar in the influenza population, new-onset AA had higher ICU admission, 90 day-readmission, need for mechanical ventilation, vasopressor and inotropes usage, new-onset CHF, ST-segment elevation MI, and requirement for new RRT, VF, and VT as shown by statistically significant OR with 95% CI.

Experimental Design, Materials and Methods
We collected data for COVID-19 and influenza patients from one quaternary care and five community hospitals at Henry Ford Health System and Trinity Health System. The first hospital admission per case was retained for both COVID-19 and influenza patients. For COVID-19 patients, clinical data were abstracted from the Epic, Inc. electronic medical record (EMR) at contributing hospitals, deidentified and stored in the Southeast Michigan COVID-19 Consortium Reg-  istry Database (SMCRD) using REDCap (software hosted at Vanderbilt University Medical Center in Nashville, Tennessee). The two systems submitted Michigan Health Information Network ID numbers (MiHIN) so that data from patients receiving care at both institutions could be linked. COVID-19 data were collected retrospectively and concurrently from 1st March 2020 to 31st March 2021. Hospitalized patients aged 18 years and above with polymerase chain reactionproven SARS-CoV-2 infection were included. Out of 6943 patients in the SMCRD registry, 16 patients were excluded because of lack of data on inpatient diagnoses, 6927 patients met the inclusion criteria. Data were collected for patients hospitalized with a diagnosis of influenza (identified using International Classification of Diseases-Tenth Revision codes) at Henry Ford Health System. Data were then deidentified and stored. The study period for influenza patients was from 1st January 2014 to 31st December 2019. A total of 14,174 influenza patients met the inclusion criteria. The EMR queries used for characteristics of the hospital stay and clinical history for the COVID-19 consortium data were the basis of the influenza data queries, so that the two data sources were compatible in definition.
The study patients in both COVID-19 and influenza populations were divided into three groups based on history of atrial arrhythmias (AA): group 1 was the normal sinus rhythm (NSR) group-these patients did not have history of AA and remained in NSR throughout hospitalization; group 2 was the new-onset AA group which did not have a prior history of AA but developed atrial fibrillation or atrial flutter during hospitalization; group 3 patients had a prior history of AA and may have stayed in NSR or experienced AA during hospitalization. The incidence of AA in COVID-19 population was 20.3% with 9% patients having new-onset AA. Within influenza population, 13.1% patients had AA with incidence of new-onset AA at 2.5%.
Hospital records of patients included in the study were reviewed to identify:   Table 13 Odds ratios of the in-hospitals events in 3 groups after propensity matching on AA groups within and between COVID-19 and influenza populations. ST-segment elevation myocardial infarction for COVID-19 population had too few events for valid estimation of odds ratio and thus not reported.

Statistical Methods
Categorical data were summarized as percentages and fraction of occurrence. Continuous data were summarized as median with interquartile range or means with standard deviations. Variable distributions were compared using Chi-square tests or Fisher's exact tests for categorical data and ANOVA or Kruskal-Wallis tests for continuous data, as appropriate. Generalized linear models were used to estimate odds ratios and risk ratios. A p -value of < 0.05 was considered significant.
We matched the hospitalized COVID-19 population to a pre-COVID hospitalized viral influenza cohort. Propensity scoring was used serially to generate balanced groups, within the COVID-19 study set, within the influenza study set, and between the COVID-19 and influenza study sets. Logistic regression was used to generate the propensity scores for each stage, using demographics and past medical history variables as predictors. Matching was done using a 0.1sd caliper, without replacement, and with ties broken randomly [12] . Within each study set, logistic regression models were used to first estimate the probability of new-onset AA given that AA were observed (history of AA vs new-onset AA). Since there were few new-onset AA cases in the influenza study set, a 2:1 match was used; 1:1 matching was used for the COVID-19 set. Logistic regression was then used in each study set to model the probability of AA (history/new AA vs none) and cases were matched 1:1. For comparison across the influenza and COIVD-19 cases, we used propensity matching to further align the two study sets. Logistic regression was used to model the probability of a COVID-19 diagnosis as the cause of hospitalization (COVID-19 vs influenza). The studies were matched 1:1. The final study set was achieved with balanced data across AA groups and study sets. Data summaries and analysis were performed with the R programming language (version 4, ggplot2 package) [13] .

Ethics Statements
The study was approved as a retrospective study by institutional review boards at Henry Ford Health System (protocol # 13785) and Trinity Health System (protocol # 2021-009). The need for informed consent was waived for the use of deidentified medical records.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.