Atrial Fibrillation After Cardiac Surgery – A Systematic Review and Meta-Analysis

67 68 Objective: New-onset post-operative atrial fibrillation (POAF) after cardiac surgery is 69 common, with rates up to 60%. POAF has been associated with early and late stroke, 70 but its association with other cardiovascular outcomes is less known. The objective was 71 to perform a meta-analysis of the studies reporting the association of POAF with peri72 operative and long-term outcomes in cardiac surgery. 73 Methods: We performed a systematic review and a meta-analysis of studies presenting 74 outcomes for cardiac surgery based on the presence or absence of POAF. MEDLINE, 75 EMBASE and Cochrane Library were assessed, 57 studies (246,340 patients) were 76 selected. Peri-operative mortality was the primary outcome. Inverse variance method 77 and random model were performed. Leave-one-out analysis, sub-group analyses and 78 meta-regression were conducted. 79 Results: POAF was associated with peri-operative mortality (odds ratio [OR]= 1.92, 80 95% confidence interval [CI] 1.58; 2.33), peri-operative stroke (OR= 2.17, 95% CI 1.90; 81 2.49), peri-operative myocardial infarction (OR= 1.28, 95% CI 1.06; 1.54), peri82 operative acute renal failure (OR= 2.74, 95% CI 2.42; 3.11), hospital (standardized 83 mean difference [SMD]= 0.80, 95% CI 0.53; 1.07) and ICU stay (SMD= 0.55, 95% CI 84 0.24; 0.86), long-term mortality (incidence rate ratio [IRR]= 1.54, 95% CI 1.40; 1.69), 85 long-term stroke (IRR= 1.33, 95% CI 1,21;1,46) and long-standing persistent atrial 86 fibrillation (IRR= 4.73, 95% CI 3.36; 6.66). 87 Conclusion: The results suggest that POAF in cardiac surgery is associated with an 88 increased occurrence of most short and long-term cardiovascular adverse events. 89 However, the causality of this association remains to be established. 90 91


English
Background New-onset post-operative atrial fibrillation (POAF) after cardiac surgery is common, with rates up to 60%. POAF has been associated with early and late stroke, but its association with other cardiovascular outcomes is less known. The objective of this study was to perform a systematic review and a meta-analysis of the studies reporting the association of POAF with peri-operative and long-term outcomes in cardiac surgery.

Methods
We performed a systematic review and a meta-analysis of studies presenting outcomes for cardiac surgery based on the presence or absence of POAF. Three libraries were assessed (MEDLINE, EMBASE and Cochrane Library) and 57 studies (total of 246,340 patients) were selected. Peri-operative mortality was the primary outcome. Inverse variance method and random model were performed. Leave-one-out analysis, sub-group analyses and metaregression were conducted.

Conclusion
The results suggest that POAF in cardiac surgery is associated with an increased occurrence of most short and long-term cardiovascular adverse events. However, the direct causality of this association remains to be established.

Definition
New-onset postoperative atrial fibrillation (POAF) is defined as the new development of atrial fibrillation after surgery in patients with previous sinus rhythm (NSR) and no history of atrial fibrillation (AF) (Lubitz et al. 2015). It is the most important type of secondary AF (AF resulting from identifiable, primary, acute conditions) (Lubitz et al. 2015).
The POAF episodes are often brief, paroxysmal and asymptomatic (Funk et al. 2003), with a peak incidence between days 2 and 4 after surgery (Funk et al. 2003, Mathew et al. 2004).
Recurrences are frequent, especially during the first postoperative week (Lee et al. 2000).
The incidence of POAF has increased continuously over the past decades, and the reason is probably related with the aging process of the population undergoing heart surgery.
Nowadays, POAF represents approximately one-third of cases of secondary atrial fibrillat ion (Greenberg et al. 2017), and AF recurrence rate in patients who develop POAF after cardiothoracic surgery (46%) is lower than that in patients with POAF after non-thoracic surgery (64%) (Lubitz et al. 2015), supporting a greater role of transient factors than of a preexisting substrate.

Pathophysiology
Underlying mechanisms are not completely defined but they include intraoperative and postoperative phenomena combined with the presence of pre-existing factors. This combination results in a totally vulnerable scenario for atrial fibrillation induction and maintenance (Nattel 2002, Andrade et al. 2014, Heijman et al. 2018, Dobrev et al. 2019).
Dobrev and colleagues (Dobrev et al. 2019) described in a structured model that POAF is mainly promoted by factors inducing atrial arrhythmogenic remodeling before surgery, as well as factors increasing the substrate vulnerability or the type/number of triggers postoperatively.
The literature showed that drugs that increase sympathetic tone increase the incidence of POAF (Feneck et al. 2001, Argalious et al. 2005 and perioperative β-blocker use reduces the incidence of POAF (White et al. 1984, Lamb et al. 1988). The incidence of POAF is lower after cardiac transplantation than after other forms of cardiac surgery (Argalious et al. 2005) and pharmacological denervation with botulinum toxin prevents POAF (Pokushalov et al. 2015). Factors that reinforce the idea from a combined sympathovagal triggering of POAF (Amar et al. 2003).
In addition, elevated preoperative plasma levels of IL-2 and IL-6 (important cytokines that mediate inflammatory responses) are also reported in some studies (Gaudino et al. 2003, Ucar et al. 2007, Pretorius et al. 2007, Hak et al. 2009). Postoperative activation of C-reactive protein has been also associated with POAF occurrence (Kaireviciute et al. 2010), and corticosteroids can reduce the incidence of POAF by inhibiting cytokine release (Ho und Tan 2009).
In summary, POAF appears to be promoted by addition of transient postoperative factors on a pre-existing and surgery-induced substrate, resulting in a re-entry structural and electrical remodeling.

Risk Factors
Atrial remodeling is an important predisposing factor for POAF, therefore elements that interfere in this situation play a significant role in this pathology. Registries have been showed the independent association between some specific factors and POAF in cardiac surgery, such as: elevated age, male sex, congestive heart failure, arterial hypertension, obesity, white ethnicity, chronic obstructive pulmonary disease (COPD), mitral valve surgery, use of intraaortic balloon pump, long cross-clamp time and bicaval cannulation (Aranki et al. 1996, Almassi et al. 1997, Zaman et al. 2000, Funk et al. 2003, Mathew et al. 2004, Auer et al. 2004, Zacharias et al. 2005, Banach et al. 2006, Shen et al. 2011, Dobrev et al. 2019).

Possible Deleterious Effect
POAF as a post-operative complication has major adverse consequences for patients and the health care system, including increased risks of post-operative stroke, peri-operative acute kidney injury (El-Chami et al. 2010, Thorén et al. 2020, increased length of hospital stay (Echahidi et al. 2008, Almassi et al. 2019) and mortality (El-Chami et al. 2010, Thorén et al. 2020. Notably is also the fact that the high AF recurrence rates in patients with POAF make this condition a clear marker of subsequent risk of long-standing persistent AF.

Underestimated Risk
Despite the high incidence, POAF has generally not been considered harmful, because of its perceived reversibility. Evidence from prospective randomized trial suggests that the vast majority of patients after CABG surgery return to normal sinus rhythm within 60 days, irrespective of the therapeutic strategy (rhythm control or rate control) (Gillinov et al. 2016).
At the same time, several studies demonstrate the hazardous potential of POAF in several clinical endpoints such as mortality, stroke, heart failure and chronic AF (Almassi et al. 2019, Thorén et al. 2020, Filardo et al. 2020).
Thus, it appears unclear whether POAF is harmless or harmful and the current perception of POAF may be influenced by individual publications.

Hypothesis and Aim of the Study
Based on the perceived controversy from the recent randomized trial evidence (Gillinov et al. 2016) and the associations presented from propensity matched studies (Schwann et al. 2018, Filardo et al. 2020), we set out to systematically review the entire literature and thoroughly address the impact of POAF on clinical outcomes, focusing on all classic cardiovascular adverse events during the peri-operative period and long-term follow-up. Objective: New-onset post-operative atrial fibrillation (POAF) after cardiac surgery is 69 common, with rates up to 60%. POAF has been associated with early and late stroke, 70 but its association with other cardiovascular outcomes is less known. The objective was 71

Original
to perform a meta-analysis of the studies reporting the association of POAF with peri-72 operative and long-term outcomes in cardiac surgery. New-onset postoperative atrial fibrillation (POAF) is defined as the new development of atrial 95 fibrillation after surgery in patients with previous sinus rhythm (NSR) and no history of atrial 96 fibrillation (AF) 1 . It is the most important type of secondary AF (AF resulting from identifiable, 97 primary, acute conditions) 1 . 98 POAF is a common complication of surgery, with an incidence varying from 10-63% for 99 cardiac surgeries (38-63% for valve and 10-33% for coronary artery bypass graft surgery) 2 . 100 Despite the high incidence, POAF has generally not been considered harmful, because of its 101 perceived reversibility. Evidence from prospective randomized trials suggests that the vast 102 majority of patients after CABG surgery return to normal sinus rhythm within 60 days 3 . 103 In contrast, other non-randomized evidence suggests that POAF may be associated with 104 increased risks of post-operative stroke, peri-operative acute kidney injury 4, 5 , increased length 105 of hospital stay 6, 7 and mortality 4, 5 . Even meta-analyses have addressed this topic, but limited 106 their assessment to stroke and peri-operative mortality 8-11 , mixed cardiac and non-cardiac 107 surgery patients 10 or are already outdated because many important studies appeared only 108 recently [8][9][10][11][12] . 109 Based on the perceived controversy from the recent randomized trial evidence 3 and the 110 associations presented from propensity matched studies 13, 14 and meta-analyses 8-11 , we set out 111 to systematically review the entire literature and thoroughly address the impact of POAF on 112 clinical outcome, focusing on all classic cardiovascular adverse events during the peri-113 operative period and long-term follow-up.  Table 1. 127

Study selection and eligi bility criteria 128
The study selection was guided by Preferred Reporting Items for Systematic Reviews and 129 Meta-Analyses (PRISMA) strategy. After de-duplication, records were screened by two 130 independent reviewers (TC and HK). Any discrepancies and disagreements were resolved by a 131 third author (TD). All titles and abstracts were reviewed against pre-defined inclusion and 132 exclusion criteria. Studies were considered for inclusion if they were written in English and 133 reported direct comparison between POAF patients and NSR patients following cardiac surgery 134 and had at least 1 outcome of interest reported. Studies evaluating non-cardiac surgeries, 135 conference abstracts and proceedings and case reports were excluded. Included studies 136 indicated clearly that the patients were evaluated on admission and that they had also no 137 previous history of alleged atrial fibrillation. 138 Following the first round of screening, full text was pulled for selected studies for a second 139 round of eligibility screening. Reference lists for articles in these selected studies were also 140 searched for any relevant articles not captured by the original search strategy. 141 Data abstraction and quality assessment 142 8 The data extraction and the quality assessment were performed independently by two different 143 investigators (TC and HK) and verified by a third investigator (TD) for accuracy. The following 144 variables were extracted: age, sex, left ventricular ejection fraction, hypertension, diabetes, 145 chronic obstructive pulmonary disease, prior cerebrovascular accident, prior myocardial 146 infarction, prior use of beta-blockers, previous heart surgery, chronic renal failure, serum 147 creatinine level. 148 For short term binary outcomes, number of events were extracted from the included studies and 149 expressed as odds ratio, for long term outcomes, incidence rate ratio was estimated. 150

Outcomes and effect summary 154
The primary outcome was peri-operative mortality. 155 Secondary outcomes were long-term mortality, peri-operative and long-term stroke, peri-156 operative myocardial infarction, acute renal failure, long-standing persistent AF, hospital 157 length of stay and intensive care unit length of stay. The peri-operative outcomes were defined 158 as in-hospital or 30-day events. 159 Subgroup analysis of recent studies (published year: up to 2010 and after 2010), a sub-group 160 analysis based on the method used for rhythm monitoring (comparison between continuous 161 monitoring in intensive care unit and irregular electrocardiography until discharge vs. 162 continuous monitoring during entire hospital stay) and a subgroup analysis addressing the type 163 of surgery (CABG, CABG and valve surgery or valve surgery) were performed to test the 164 solidity of the main analysis. 165

Data analysis 166
Peri-operative binary outcomes were reported as odds ratios (OR) while long-term outcomes, 167 9 were reported as incidence rate ratio (IRR); for both estimates the generic inverse variance 168 method was used and 95% confidence intervals (CIs) were also presented. Sub-group analyses 169 of the primary outcome were reported as OR associated with subgroup difference P-interaction 170 (SGD-P) with 95% CI. Continuous outcomes were expressed as standardized mean difference 171 (SMD) with 95% CI. 172 Random effect meta-analysis was performed using "metafor" and "meta" package 16, 17 . NSR 173 was the reference for all pairwise comparisons. Heterogeneity was reported as low (I2= 0%-174 25%), moderate (I2= 26%-50%), or high (I2 > 50%) 18 . Leave-one-out analysis for the primary 175 outcome was performed to assess the robustness of the obtained estimate. Meta-regression was 176 used to explore the effects of age, sex, left ventricle ejection fraction, comorbidities, use of 177 beta-blockers and previous heart surgery on the OR of the primary outcome. States, 3 from Canada, 3 from Netherlands, 3 from Sweden, 3 from Brazil, 2 from Finland, 2 192 from Turkey, 2 from Iran, 2 from Australia, 2 from Korea, 2 from the United Kingdom, 1 each 193 from Denmark, Serbia, Austria, Japan, Taiwan, Israel, Saudi Arabia, Malaysia,and Colombia. 194 The number of patients in each study ranged from 44 to 49,264. The mean age ranged from 195 54.6 to 77.4 years. The percentage of female sex in each study ranged from 0.9 to 68.0%. In 196 terms of patient comorbidities, the prevalence of hypertension ranged from 30.4 to 97.0%, the 197 prevalence of diabetes ranged from 3.4 to 66.7%, the prevalence of chronic obstructive 198 pulmonary disease ranged from 0.5 to 41.8%, the prevalence of prior cerebrovascular accident 199 ranged from 0.6 to 29.2%, the prevalence of prior myocardial infarction ranged from 2,2 to 200 73,0% and the prevalence of chronic renal failure ranged from 1.0 to 15.0% (Supplementary 201  (Supplementary Figu re 2). At meta-regression, the proportion of female patients was inversely associated with the OR for 231 the primary outcome (beta=-0.0293, p=0.0043). The proportion of diabetes and presence of 232 prior myocardial infarction were associated with "higher" OR for the primary outcome (beta= 233 0.0144, p=0.0056 and 0.0122, p=0.0303; respectively -Supplementary Table 5). 234

DI SCUSSION 236
Our analysis suggests that POAF in cardiac surgery is associated with an increased occurrence 237 of most short and long-term cardiovascular adverse events. Specifically, POAF appears to be 238 12 associated with peri-operative mortality, peri-operative stroke, peri-operative myocardial 239 infarction, peri-operative acute renal failure and long-term mortality, long-term stroke, long-240 standing persistent AF, as well as hospital length of stay and intensive care unit length of stay 241 (Video 1). 242 However, our results cannot prove causation and it is unclear if POAF was involved in the 243 pathogenesis of the associated outcomes, or if it was only a marker of increased cardiovascular 244 risk. Recent data suggest that a pre-existing arrhythmogenic substrate exists before surgery and 245 discriminates who is going to develop POAF. It may potentially explain the long-term 246 recurrence rate of AF and the occurrence of other cardiovascular events 21 . 247 It is known that factors such as inflammation, myocardial ischemia and autonomic nervous 248 system activation are thought to be superimposed on susceptible atrial substrates, making the 249 atrium vulnerable to AF induction and maintenance 22 . In addition, pre-existing atrial fibrosis 250 may predispose patients to developing atrial fibrillation, which may have implications for the 251 timing of cardiac interventions 23 . 252 Different clinical factors have also been described that may contribute to the development of 253 POAF. They range from pre-operative (hypertension, myocardial ischemia, valvular 254 abnormalities), through peri-operative (surgical trauma, local inflammation, large fluid shifts, 255 electrolyte disturbances) to post-operative conditions and events (inotropic drugs, atrial pacing, 256 pneumonia) 6, 22 . Thus, the combination of pre-disposing substrates for the natural development 257 of AF with peri-operative events would then increase the risk of POAF. This conceptual model 258 could explain the differences between classic surgery and TAVI for aortic valve replacement 259 24, 25 and the significant rate of POAF in non-cardiac surgery 26 . 260 Although previous meta-analyses addressed this topic, they had significant limitations. For 261 instance, even the latest publication in the field 11 analyzed only 2 outcomes (stroke and 262 mortality) and did not include some of the important contemporary publications on POAF post-263 13 cardiac surgery with more than 20,000 patients 4,7,12,13,[27][28][29][30] . To the best of our knowledge, our 264 comprehensive meta-analysis is the first to assess all important cardiovascular adverse events. American Heart Association (AHA) statistics assuming a POAF incidence of 30%, the extra 287 cost due to POAF can be calculated to exceed $2 billion/year 6 . 288 14 Considering the possible negative effects of POAF, patients and physicians may be reluctant 289 to recommend/undergo surgery. However, it is important to note that current indications for 290 cardiac surgery are often free of alternatives (e.g., endocarditis), and the generated outcomes 291 are often still superior to their existing interventional (i.e., TAVI) or conservative alternatives 292 with regards to long-term perspective 25,33,34 . It therefore appears that the positive effects of 293 surgery outweigh the negative influence of POAF. 294 Given the association of POAF with worse peri-operative and long-term outcomes, the interest 295 in this topic has recently grown. Based on the assumption that successful prevention or 296 treatment of AF may be able to further improve outcomes of cardiac surgery, randomized trials 297 have already been performed or recently initiated. Some of them have concentrated on 298 preventing or reducing the incidence of POAF 35 and others on the treatment and prevention of 299 adverse events. The Anticoagulation for New-Onset Post-Operative Atrial Fibrillation after 300

CABG (PACES) trial of the cardiothoracic surgical trials network (CTSN) [NCT04045665] is 301
a good example of the latter. 302 In this context, our comprehensive meta-analysis provides a broad overview on POAF and its 303 association with the most important clinical outcomes. Thus, the information we present might 304 be useful when building future hypotheses or designing future randomized control trials on this 305 topic. 306

STUDY STRENGTH AND LIM I TATI ONS 308
This analysis was conducted at study level rather than patient level. All studies were 309 observational in nature. However, there are no randomized trials addressing this issue, which 310 is not unexpected since equipoise for an adequate conservative control group does not exist. 311 Another limitation of this study is the fact that patients with preexisting episodes of silent atrial 312 fibrillation preoperatively might have also been included in the individual studies. However 313 15 this possibility exists for both investigated groups (NSR and POAF). 314 Since one of the inclusion criteria for the review was general studies concerning POAF, the 315 aggregate study population was potentially heterogeneous. We pooled related outcomes and 316 included the definitions of others that may be different among different studies, as acute renal 317 failure and long-persistent AF (Supplementary Table 6 Figure 6. Graphical Abstract. POAF after cardiac surgery appears to be associated with increased occurrence of peri-operative mortality, peri-operative stroke, peri-operative myocardial infarction, peri-operative acute renal failure, hospital length of stay, intensive care unit length of stay, long-term mortality, long-term stroke and long-standing persistent atrial fibrillation.

Date:
Mar 12, 2021 To: "Torsten Doenst" doenst@med.uni-jena.de;benjamin.may@med.uni-jena.de cc: rakesh arora (rakeshcarora@gmail.com), stephen.fremes@sunnybrook.ca From: "Journal of Thoracic and Cardiovascular Surgery" jtcvs@aats.org Subject: Acceptance of your Submission JTCVS-20-3301R2 The The editorial staff of The Journal of Thoracic and Cardiovascular Surgery is pleased to inform you that, after careful review, your Origina l Manuscript "Atrial Fibrillation after Cardiac Surgery -A Systematic Review and Meta-Analysis" has been accepted for publication.
You will receive an electronic copyright form and elec tronic page proofs directly from Elsevier Science Publishing. Once received, please return these proofs with all necessary corrections to Elsevier within 48 hours.

Discussion
The meta-analysis suggests that POAF in cardiac surgery is associated with an increased occurrence of most short and long-term cardiovascular adverse events. Specifically, POAF appears to be associated with peri-operative mortality, peri-operative stroke, peri-operative myocardial infarction, peri-operative acute renal failure and long-term mortality, long-term stroke, long-standing persistent AF, as well as hospital length of stay and intensive care unit length of stay. We addressed the specific aspects of the meta-analysis, its strengths and limitations in the discussion of the original manuscript in the last section. I here discuss the general meaning of the associations found in our meta-analysis and meta-regression.
These results suggest that POAF should no longer be understood as transitory, self-sufficient and potentially clinical insignificant complication in cardiac surgery. The main concern arising from these findings shows that the cardiovascular community can benefit from better surgical outcomes if they comprehend the underlying mechanism of this condition. A key question addresses the role of POAF as an indicator of a particular combination of risk factors or as the main causative element itself.

Causality Aspects
A key question is whether POAF is a cause or merely an indicator of the described complications. It is conceivable that POAF occurs most frequently in patients with comorbidities and is therefore associated with increased mortality. It will currently be impossible to provide a definitive answer to this question because of the lack of an ideal control group of patients not having undergone surgery with otherwise the same demographic characteristics.
Anyway, cardiac surgery patients have higher incidence of POAF compared with those undergoing non-cardiac surgery (Christians et al. 2001, Villareal et al. 2004, Echahidi et al. 2008, Turagam et al. 2016. In this context, a certain degree of causality may be expected, mainly based on an exposure-response relationship: the incidence of POAF increases as the cardiac surgical approach becomes more invasive (Mihos et al. 2013, Tanawuttiwat et al. 2014). This notion is supported also in randomized trials comparing CABG and PCI where similar patients in both groups show an increased incidence of POAF between procedures based on its level of invasiveness (i.e., CABG compared with PCI procedure -18.0% versus 0.1%) (Stone et al. 2016, Kosmidou et al. 2018. These findings reinforce the idea that reducing the degree of operative trauma may result in the reduction of POAF incidence. Thus, minimally invasive approaches may have to be investigated for their ability to reduce this adverse event.
Observational evidence suggests that minimally-invasive procedures may indeed be associated with lower rates of POAF in comparison with the traditional approach (Tabata et al. 2008, Murzi et al. 2012, Glauber et al. 2013, Doenst und Lamelas 2017, Faerber et al. 2020.

Race and Sex Difference
POAF also appears to be influenced by differences in the genetic background. For instance, POAF was observed to be a stronger predictor of operative mortality in Black individua ls compared with White patients undergoing elective CABG (Efird et al. 2013 (Lahiri et al. 2011, Sun et al. 2011, Rader et al. 2011, Efird et al. 2013).
Male gender is also one of the possible risk factors described in the literature. The apparent protective effect of female sex shows not to be a mere coincidence, because male sex is frequently identified as an independent risk factor for POAF in CABG patients (Zacharias et al. 2005, Filardo et al. 2009, Alam et al. 2013, Almassi et al. 2019).
The race influence and how female sex protects against POAF are statements that require further research. Sex-specific studies have historically been missing from the evidence base but have been called for in recent years to help address continuing sex-related disparities in health care issues and key outcomes (Girardi et al. 2019, Gaudino et al. 2020a, Gaudino et al. 2020b).

Long-standing Atrial Fibrillation as a Consequence of POAF
An important point of our meta-analysis is the fact that it did not only show an association with short-term outcomes, but especially with long-term clinical adverse events. One of these implications that deserves specific attention is the fact that the meta-analysis demonstrates increased occurrence of long-standing AF in individuals that developed POAF (Ahlsson et al. 2010, Pillarisetti et al. 2014, Melduni et al. 2015, Tulla et al. 2015, Konstantino et al. 2016, Lee et al. 2017, Carter-Storch et al. 2019, Thorén et al. 2020. Patients with POAF can present increased incidence of AF not only compared with patients without POAF but also compared with matched presumably healthy controls (Thorén et al. 2020). This increase in AF compared with controls persisted over time and was valid after more than 10 years of follow-up. On the other hand, the non-POAF cohort showed no increase in AF beyond the first postoperative year (Thorén et al. 2020). This finding supports the notion that a certain substrate for the development of AF is present at an increased prevalence in those patients developing POAF. Thus, surgical trauma triggers AF earlier than it would appear naturally anyway (Dobrev et al. 2019). The fact that POAF may lead to chronic AF then of course introduces all adverse events associated with chronic AF into the POAF arena. Patients with chronic AF show a dramatic increase in the incidence of pathologies such as stroke and heart failure , Vintila et al. 2019). Based on the economic, health and social impact from atrial fibrillation, these results are certainly alarming and accentuate the harmful potential of POAF , Vintila et al. 2019. Therefore, it is clear that POAF is a clinical entity that deserves attention and probably the investment of efforts to combat it through new medical therapies, new surgical approaches and new therapeutic alternatives.

Medication
Nowadays, perioperative beta-blocker treatment is the main pharmacologic therapy with the objective to reduce rates of POAF (Echahidi et al. 2008, Dobrev et al. 2019. Other therapies such as amiodarone, verapamil, diltiazem, and digoxin are used less frequently and are generally less effective (Buckley et al. 2007). Due to the low efficacy of traditional drug therapy, new solutions have been proposed in order to act not only in the causative mechanism, but also in the prophylaxis of possible deleterious effects.
In this context, the use of anticoagulants proved to be a potential treatment in POAF due to its classic use in chronic atrial fibrillation. Recently, El-Chami and colleagues analyzed the connection between anticoagulation and survival in cardiac surgery patients with POAF, showing mortality reduction in patients treated with warfarin, after adjusting for age, sex, and medical comorbidities (El-Chami et al. 2010).
Generally, anticoagulation in atrial fibrillation has been aimed at reducing stroke risk and minimizing other side effects. However, anticoagulation in patients with POAF is an unexplored topic. No guidelines provide specific recommendations for initiation of anticoagulation for POAF in the post-cardiac surgery population (Macle et al. 2016, Kirchhof et al. 2016). The majority of evidence for anticoagulation in AF emerges mostly from the nonsurgical community, which have a substantially different risk profile compared to surgical patients in terms of bleeding risk and, above all, as possible trigger for AF. Accordingly, as the profile of patients is different and especially the etiology of AF, it is difficult to draw conclusions about the therapeutic efficacy of these drugs. Thus, undoubtedly the role of anticoagulation in POAF is a topic that needs to be explored with an accurate study design through a randomized clinical approach.
In the long term, for the patient with long-standing persistent atrial fibrillation after cardiac surgery, the use of Novel Oral Anticoagulants (NOAC) may be plausible, as a number of studies have shown NOACs reduce bleeding risk and other severe complications in comparison with warfarin (Hicks et al. 2016, Aimo et al. 2018). However, no data has shown the effect of NOACs in the specific population of patients developing POAF.

Bi-atrial Pacing
Bi-atrial pacing has emerged as a promising approach to reduce the incidence of POAF. The rationale is that the electric maintenance of atrio-ventricular synchrony has shown to reduce the incidence of atrial fibrillation by suppressing premature atrial complexes and runs of supraventricular re-entry rhythm (Saksena et al. 1996, Gillis et al. 1999. A recent work on the topic, showed in a pairwise and network meta-analysis involving 14 trials that bi-atrial pacing, compared to other pacing modalities, is associated with lower rates of POAF following CABG (Ruan et al. 2020). These findings demonstrate that an effective therapy for POAF is plausible and accessible. However, bi-atrial pacing has not yet been accepted in routine practice, possibly for its technical need for the tedious need to place two epicardial pacemaker wires. We are currently starting a trial in our Department to assess the impact of bi-atrial pacing and the possible therapeutic impact of atrial cardioversion on the incidence and clinical course of POAF (Defi-Pace trial). Other trials are currently ongoing in this field.

CABG-AF Trial
CABG-AF is a multicenter trial where study patients undergoing CABG, without previous history of atrial fibrillation or other complex rhythm disorders are receiving an event-recorder implantation at the end of surgery. Their heart rhythms are being continuously monitored for up to 3 years. Data concerning the development of atrial fibrillation, atrial fibrillation burden, atrial fibrillation density, number and length of atrial fibrillation, episodes, silent vs.
symptomatic episodes, stroke and mortality will be collected and evaluated. Our center is one of the four surgical centers participating in the study within the GermaN HeaRTS Network.

PACES Trial
The Anticoagulation for New-Onset Post-Operative Atrial Fibrillation after CABG (PACES) trial [NCT04045665] is a multicenter randomized controlled trial with the aim to evaluate the effectiveness (prevention of thromboembolic events) and safety (major bleeding) of adding oral anticoagulation (OAC) to background antiplatelet therapy in patients who develop newonset post-operative atrial fibrillation (POAF) after isolated coronary artery bypass graft (CABG) surgery. In the trial, 3200 patients are been randomized in 2 arms:  OAC-based strategy (experimental arm): OAC with vitamin K antagonist (VKA) with international normalized ratio (INR) target 2-3 or any approved direct oral anticoagulant (apixaban, rivaroxaban, edoxaban or dabigatran) in addition to background antiplatelet therapy with aspirin 75-325mg once-daily or a P2Y12-inhibitor (clopidogrel or ticagrelor);  Antiplatelet-only strategy (control arm): with aspirin 75-325mg once-daily or a P2Y12inhibitor (clopidogrel or ticagrelor).
The primary outcome of the study is a composite of death, stroke, transient ischemic attack, myocardial infarction, systemic arterial thromboembolism or venous thromboembolism.

Conclusions
POAF after cardiac surgery appears to be associated with increased occurrence of a plethora of cardiovascular adverse events. While a directly causal relationship between POAF and these adverse cardiovascular events cannot be totally established. The information represents an important milestone for future hypotheses about etiology from POAF and for designing randomized trials addressing its prevention/treatment.

(ehrenwörtliche Erklärung)
I hereby declare that I am familiar with the doctoral regulations of the Medical Faculty of the Friedrich Schiller University.
I wrote the dissertation myself and all aids, personal communications and sources I used are given in my work.
The following people supported me in the selection and evaluation of the material as well as in the preparation of the manuscript: Univ. Prof. Dr. med. Torsten Doenst and Dr. med. Hristo Kirov as well as the other co-authors of the manuscript.
The help of a doctoral advisor was not used and that third parties did not receive any direct or indirect monetary benefits from me for work related to the content of the submitted dissertation.
I have not yet submitted the dissertation as an examination paper for a state or other scientific examination.
I have not submitted the same, essentially similar or a different dissertation to another university as a dissertation.

Acknowledgments
I would like to thank Univ. Prof. Dr. med. Torsten Doenst for guidance, trust and patience. His competent support had fundamental importance for this work, and his commitment with research and with teaching moulds admirable values that make him a great example for me not only in the clinical activities, but especially in the scientific field.
I would like also to thank all the co-authors of the publication, in special Dr. med Hristo Kirov.
They supported me with technical advice and instructions, which were fundamental to perform this work.
Finally, I am totally grateful for the constant support from my parents, my brother and my friends.