Incident Atrial Fibrillation and In-Hospital Mortality in SARS-CoV-2 Patients

Round 1

Reviewer 1 Report

The manuscript addresses the relation between AF and COVID-19, the need for assistance in ICU and the risk of in-hospital death. 

The manuscript is well structured and well written.

COVID-19 is more than a respiratory disease, it is a systemic disease, with a possible unfavorable result when multiple comorbidities are associated. Because AF is the most common sustained adult arrhythmia in the general population, it is useful to assess its prevalence in this vulnerable group of patients. The results of the study may be useful in the therapeutic conduct of future cases. I encourage the publication of these results.

Detailed Comments:

Dear Editor,

I have extended my comments and I hope that now it gives my previous opinion the necessary support.

Request: Briefly summarize the content of the manuscript;

Answer: COVID-19 is a new disease, affecting millions of people worldwide, sometimes with the poor outcome despite intensive medical treatment. We know now that COVID-19 is more than respiratory disease. It is associated with systemic inflammation and thrombosis. Infection and inflammation may be triggers for atrial fibrillation. The presence of this arrhythmia may be associated with poor outcomes in patients with pre-existing diseases, cardiovascular among others, and may increase cardiovascular and overall mortality. The aim of the study was to assess the incidence of AF in patients hospitalized for COVID-19 and its role as a possible predictor of in-hospital all-cause mortality. The study was multicenter (three hospitals) and enrolled 3435 patients. Only AF with onset during hospitalization was considered. The study showed that 4.2% of patients developed AF during hospitalization and that these AF patients required more often admitted to the ICU and had higher mortality. AF was a predictor of mortality.

Request: Illustrate what are, in your opinion, the manuscript’s strengths and
weaknesses. Provide a list of recommendations for the improvement of the manuscript.

Answer:

Strengths: The study is multicenter. A significant number of patients were enrolled, allowing a conclusion to be drawn. The authors performed a comprehensive assessment of the risk factors for the onset of AF in addition to COVID-19 infection. The discussion integrates current findings with existing knowledge. The tables and figures systematize and clearly show the results of the study.

Weaknesses: The limitation of the study is fairly recognized by the authors. I consider that the most important limitation is that the study included only inpatients. This approach has two disadvantages. First of all, patients with less severe disease who have not been hospitalized are excluded. They represent a much more numerous category than those hospitalized. Hence, it is possible that the number of cases of AF incident is underestimated. This aspect is commented on by the authors. Second, patients who had AF at the time of the hospital admission were excluded from the analysis. In this way, patients who developed arrhythmia in the context of COVID-19 shortly before hospitalization were excluded.

I recommend to the authors add some comments on this issue, as it could influence the need for admission to the ICU and in-hospital mortality. Moreover, at the beginning of the pandemic, there was a shortage of hospital beds and especially ICU beds. I recommend the authors comment on whether this aspect could influence the results of the presented study.

Thank you!

Author Response

RE: biomedicines-1772817- Incident atrial fibrillation and in-hospital mortality in SARS-CoV-2 patients

Dear Editor,

we thanks the reviewers for the time and energy that invested in in revising our manuscript. We believe that the comments and suggestions provided gave us the opportunity to improve the manuscript considerably and better present the results of our study. Please find below the answers to all the comments point by point. In the revised manuscript, all changes have been highlighted.

Reviewer #1:

The manuscript addresses the relation between AF and COVID-19, the need for assistance in ICU and the risk of in-hospital death. The manuscript is well structured and well written. COVID-19 is more than a respiratory disease, it is a systemic disease, with a possible unfavorable result when multiple comorbidities are associated. Because AF is the most common sustained adult arrhythmia in the general population, it is useful to assess its prevalence in this vulnerable group of patients. The results of the study may be useful in the therapeutic conduct of future cases. I encourage the publication of these results.

COVID-19 is a new disease, affecting millions of people worldwide, sometimes with the poor outcome despite intensive medical treatment. We know now that COVID-19 is more than respiratory disease. It is associated with systemic inflammation and thrombosis. Infection and inflammation may be triggers for atrial fibrillation. The presence of this arrhythmia may be associated with poor outcomes in patients with pre-existing diseases, cardiovascular among others, and may increase cardiovascular and overall mortality. The aim of the study was to assess the incidence of AF in patients hospitalized for COVID-19 and its role as a possible predictor of in-hospital all-cause mortality. The study was multicenter (three hospitals) and enrolled 3435 patients. Only AF with onset during hospitalization was considered. The study showed that 4.2% of patients developed AF during hospitalization and that these AF patients required more often admitted to the ICU and had higher mortality. AF was a predictor of mortality.

Strengths: The study is multicenter. A significant number of patients were enrolled, allowing a conclusion to be drawn. The authors performed a comprehensive assessment of the risk factors for the onset of AF in addition to COVID-19 infection. The discussion integrates current findings with existing knowledge. The tables and figures systematize and clearly show the results of the study.

Weaknesses: The limitation of the study is fairly recognized by the authors. I consider that the most important limitation is that the study included only inpatients. This approach has two disadvantages. First of all, patients with less severe disease who have not been hospitalized are excluded. They represent a much more numerous category than those hospitalized. Hence, it is possible that the number of cases of AF incident is underestimated. This aspect is commented on by the authors. Second, patients who had AF at the time of the hospital admission were excluded from the analysis. In this way, patients who developed arrhythmia in the context of COVID-19 shortly before hospitalization were excluded.

I recommend to the authors add some comments on this issue, as it could influence the need for admission to the ICU and in-hospital mortality. Moreover, at the beginning of the pandemic, there was a shortage of hospital beds and especially ICU beds. I recommend the authors comment on whether this aspect could influence the results of the presented study.

We thank the reviewer for the kind comments. We hope that the revised manuscript meets also the expectations of the reviewer. Regarding the limitations highlighted by the reviewer:

I consider that the most important limitation is that the study included only inpatients. This approach has two disadvantages. First of all, patients with less severe disease who have not been hospitalized are excluded. They represent a much more numerous category than those hospitalized. Hence, it is possible that the number of cases of AF incident is underestimated.

Our study is based on a population of hospitalized patients, and we think it would have been impossible to characterize so precisely a population of outpatients. Outpatients COVID-19 infected patients are probably affected by a less severe disease and, likely, present fewer incident AF. We think, therefore, that if we had examined a population that also included outpatients the incidence of AF would have been lower and not higher. So the overall incidence of AF in our population might be overestimated and not underestimated compared to the total COVID patient population. The paragraph on this limitation have been expanded in order to better explain these concepts (page 9, line 330).

Second, patients who had AF at the time of the hospital admission were excluded from the analysis. In this way, patients who developed arrhythmia in the context of COVID-19 shortly before hospitalization were excluded.

Patients who had AF at the time of the hospital admission were not excluded from the mortality analysis but only from the analysis of AF incidence. These patients, since the onset of arrhythmia could not be dated and we could not know whether the AF found at the time of admission was due to COVID infection or not, were not considered as patients with incident AF and were classified as patients with history of AF.

We agree with the reviewer that some patients presenting with arrhythmia at the time of admission may have COVID-related AF, and therefore it is possible that patients with incident AF were underestimated in our analysis. This point is now better reported in the limitations (page 9, line 338).

Moreover, at the beginning of the pandemic, there was a shortage of hospital beds and especially ICU beds. I recommend the authors comment on whether this aspect could influence the results of the presented study.

The well known problem of ICU beds management in the first phases of the pandemic may have influenced our results. In fact, diagnosis have been done both at a 12-lead electrocardiogram following patients’ symptoms onset or the identification of a not regular pulse at the clinician visit, or as an episode on electrocardiogram monitoring system lasting for at least 30 seconds. The majority of non-ICU hospitalized patients doesn’t present electrocardiograhic continuous monitoring that is a more sensible methods to diagnosing also not symptomatic episodes. So this could lead to an under estimation of incident AF. These points are now stated in the methods (page 2, line 94) and in the limitation section (page 9, line 338).

Reviewer 2 Report

-) The whole manuscript including the abstract needs revising in terms of English grammar and style by a native speaker.

-) Please use the correct, overall applicable full name for COVID-19.

-) Methods: Why were only data until mid-2021 included?

-) Methods: Is it correct that patients were only included if they had de-novo AF, and patients developing AF during their hospital stay but had a history of AF were excluded? If so, why? And also, how can then "history of AF" be a covariate in your analysis?

-) How exactly was AF defined and diagnosed?

-) Is there any data what happened to the patient (AF-wise)? Were they cardioverted?

-) The discussion section needs to be more structured, probably via subheadings.

-) Discussion: What about AF simply occurring in patients being septic due to COVID-19? A lot of comparisons with big sepsis studies could be done here.

Author Response

RE: biomedicines-1772817- Incident atrial fibrillation and in-hospital mortality in SARS-CoV-2 patients

Dear Editor,

we thanks the reviewers for the time and energy that invested in in revising our manuscript. We believe that the comments and suggestions provided gave us the opportunity to improve the manuscript considerably and better present the results of our study. Please find below the answers to all the comments point by point. In the revised manuscript, all changes have been highlighted.

Reviewer #2:
We thank the reviewer for the kind comments. We hope that the revised manuscript meets also the expectations of the reviewer.

The whole manuscript including the abstract needs revising in terms of English grammar and style by a native speaker.

The whole paper has been reviewed for grammar and style by a native English speaker.

Please use the correct, overall applicable full name for COVID-19.

As suggested, COVID-19 has been used instead of COVID.

Methods: Why were only data until mid-2021 included?

Data have been collected till the second wave of COVID-19 infection. After that, SARS-CoV-2 has changed significantly and it becomes a less severe, but more infectious, disease. We maintain those data in order to avoid including patients with omega mutation of the virus. However, its inclusion could determine different results with, probably due to the reduction in mean severity, a less frequent presence of incident AF. The explanation on the chosen inclusion period is now stated in the results (page 2, line 82) while the possibility of lower incidence of AF with the newer SARS-CoV-2 variation have been added to the limitation section (page 9, line 341).

Methods: Is it correct that patients were only included if they had de-novo AF, and patients developing AF during their hospital stay but had a history of AF were excluded? If so, why? And also, how can then "history of AF" be a covariate in your analysis?

We apologize to the reviewer for not being clear on this point. All patients except those who already had permanent AF were included in the analysis reported in Table 2 (Incident AF analysis). Patients who had a previous history of AF (non-permanent, i.e., paroxysmal or persistent) but were in sinus rhythm at the time of admission were included in the analysis. In the analysis in Table 2, the variable "history of AF" thus refers to these patients. For clarity, we changed the name of the variable in “history of non-permanent AF”.

All patients, including those with a history of AF, both non-permanent and permanent, were included in the analysis shown in Table 3 (Mortality analysis). This variable (“history of AF”) is not present in Table 3 since it is not significant at multivariable analysis. This is now more clearly stated in the methods section (page 2, line 94).

How exactly was AF defined and diagnosed?

AF has been diagnosed following international guidelines as an arrhythmia without P waves with an arrhythmic R-R interval lasting for at least 30 second. It has been diagnosed both at a 12-lead electrocardiogram following patients’ symptoms onset or the identification of a not regular pulse at the clinician visit, or as an episode on electrocardiogram monitoring system. This definition has been added to the methods section (page 2, line 94).

Since not all the patients had an electrocardiogram monitoring system during their hospitalization, not symptomatic episodes could be not diagnosed with a total under estimation of incident AF. This has been added to the limitation section (page 9, line 337).

Is there any data what happened to the patient (AF-wise)? Were they cardioverted?

Details on post-incident AF treatment (pharmacological or electrical cardioversion vs spontaneous rhythm restoration or rate control management) were not collected in our dataset. This is now stated as a limitation in the relative section (page 10, line 357).

The discussion section needs to be more structured, probably via subheadings.

Discussion section have been structured with subheadings as suggested.

Discussion: What about AF simply occurring in patients being septic due to COVID-19? A lot of comparisons with big sepsis studies could be done here.

The reviewer is more than right when stated that AF probably simple occurring due to septic state more than from a direct viral effect. In fact, we agreed with this and also in this paper we try to transfer this message. A discussion on comparison with studies on incident AF in patients without COVID-19 infection was already present in our paper have been expanded (page 8, line 251).

Reviewer 3 Report

The present study shows that incident atrial fibrillation (AF) is a common complication in COVID-19 patients being hospitalized. Moreover, AF is associated with a significantly worse clinical outcome. Allthough not really surprizing these data are of actual clinical relevance, as these patients need appropriate and intensive diagnostic procedures as well as an intensive clinical supervision.

Unfortunatelly relevant echocardiographic features like left ventricular function, atrial and ventricular dimensions, valve function and signs  of inflammation like pericardial effusion were not presented. If possible, these data should be included and commented.

Moreover, I do not see relevant myocardial markers like BNP and troponin. If available, these data should be included and discussed.

Author Response

RE: biomedicines-1772817- Incident atrial fibrillation and in-hospital mortality in SARS-CoV-2 patients

Dear Editor,

we thanks the reviewers for the time and energy that invested in in revising our manuscript. We believe that the comments and suggestions provided gave us the opportunity to improve the manuscript considerably and better present the results of our study. Please find below the answers to all the comments point by point. In the revised manuscript, all changes have been highlighted.

Reviewer #3:
We thank the reviewer for the kind comments. We hope that the revised manuscript meets also the expectations of the reviewer.

The present study shows that incident atrial fibrillation (AF) is a common complication in COVID-19 patients being hospitalized. Moreover, AF is associated with a significantly worse clinical outcome. Allthough not really surprizing these data are of actual clinical relevance, as these patients need appropriate and intensive diagnostic procedures as well as an intensive clinical supervision.

Unfortunatelly relevant echocardiographic features like left ventricular function, atrial and ventricular dimensions, valve function and signs of inflammation like pericardial effusion were not presented. If possible, these data should be included and commented.

Moreover, I do not see relevant myocardial markers like BNP and troponin. If available, these data should be included and discussed.

Echocardiographic and biochemical variables suggested by the reviewer are very important but, unfortunately, they were not collected systematically and, due to the fact that these evaluations were performed only on a minotiry of patients, they could not be used for our analysis.

Furthermore, particularly regarding proBNP and TnT, also a risk of selection bias is present. In fact, proBNP and TnT were assessed mainly in patients with a more severe disease and this would lead to a significant selection bias on the subsample with these information.

We now stated this as a limitation in the appropriate section (page 10, line 347) and quote some of the studies that emphasize the importance of these clinical variables.

Round 2

Reviewer 2 Report

Dear authors,

Thank you for having addressed all my comments so thoroughly. In fact, through your adaptations, the whole manuscript is presented in a different light and I think adding content to your limitations section was also very important.