Atrial appendage closure is associated with increased risk for postoperative atrial fibrillation Authors: Jeffrey H Shuhaiber, Mostafa Abbas, Thomas Morland, H Lester Kirchner, Yasser El-Manzalawy Published: 2024 ABSTRACT AbstractObjectivesThis study aims to examine the relationship between left atrial appendage closure (LAAC) and post-operative atrial fibrillation (POAF) in cardiac surgery patients with no pre-operative atrial fibrillation (AF).MethodsWe analyzed a cohort of 2059 adult patients in our Society of Thoracic Surgery (STS) database who underwent at least one of the following procedures between 2018 and 2021: coronary artery bypass grafting (CABG), aortic valve replacement, or mitral valve replacement. All patients had no pre-operative AF, and 169 (8.2%) of them received a left atrial appendage closure (LAAC). Primary outcome was new-onset POAF and secondary outcomes included 1-year mortality, 30-day readmission, 1-year incident stroke, and post-operative hospital length of stay (LOS). Patients without an LAAC were matched to patients with LAAC using a 1:1 nearest neighbor propensity score method to reduce the bias due to potential confounding. Associations between LAAC and postoperative adverse outcomes were assessed using appropriate statistical tests for matched analyses.ResultsThe matched cohort included 162 pairs. LAAC was associated with elevated risk of new-onset POAF (41.4% vs. 25.3%, p = 0.003), and postoperative hospital LOS (142.1 vs. 120.5 h, p = 0.001). The LAAC did not significantly impact all-cause 1-year mortality, 30-day readmission, and 1-year incident stroke.ConclusionsIn cardiac surgery patients with no pre-operative history AF, LAAC was associated with substantially higher rates of new-onset POAF, without a corresponding impact on risk of 1-year incident stroke, 30-day readmission, or 1-year mortality.Keywords: Cardiac surgery, Left atrial appendage closure, Atrial fibrillation, Outcomes Introduction Left atrial appendage closure (LAAC) is a minimally invasive surgical procedure to permanently close the left atrial appendage in order to prevent blood clots from entering the bloodstream and causing a stroke [1–5]. Recently, there has been a growing interest in LAAC as a non-pharmacological adjunct to lifelong anticoagulation for stroke prophylaxis in cardiac surgery patients with history of atrial fibrillation (AF) [6–12]. Several randomized controlled trials (RCTs) and observational studies have demonstrated that the surgical LAAC concomitant to planned cardiac surgery is associated with reduced risk of stroke [6, 9, 10, 12], systemic embolism [9, 12],mortality [6, 7, 10, 12], and long-term bleeding [10, 12]. For cardiac surgery patients without a history of pre-operative AF, the evidence supporting LAAC is limited [13–15]. In an RCT of 187 open heart surgery patients by Park-Hansen et al. [16], LAAC was associated with reduced incidence of ischemic stroke, regardless of history of AF. The AtriClip Left Atrial Appendage Exclusion Concomitant to Structural Heart Procedures (ATLAS) trial, enrolled 562 patients without a history of pre-operative AF, but at risk of stroke (i.e., CHA2DS2VASC score > = 2) and bleeding (i.e., HAS-BLED score > = 2), and found that patients in the LAAC group had significantly higher rates of POAF and non-significant reduction in the incidence of peri-operative stroke [13]; these findings have not been replicated. Moreover, it is unknown whether cardiac surgery patients without a history of AF and with low stroke risk would benefit from LAAC. In this retrospective real-world study, our objective was to retrospectively determine the POAF incidence and related complications among patients who underwent open heart surgery and received LAAC with no history of AF prior to surgery. Methods Study population The Geisinger Institutional Review Board approved this retrospective study with a waiver of consent, in conjunction with our institutional patient privacy policies. We created a retrospective study cohort encompassing 2,059 adult patients from our institutional Society of Thoracic Surgery (STS) database who underwent at least one of the following procedures between 2018 and 2021: coronary artery bypass grafting (CABG), aortic valve replacement, or mitral valve replacement at Geisinger HealthCare System. None of the patients had a pre-operative AF, and 169 (8.2%) of them received Left atrial appendage closure (LAAC) using (Atricure Inc.) atriclip. Clinical variables and outcomes Patients’ STS data were linked with our electronic health records (HER) in order to enrich our analysis with additional clinical variables and outcomes. Patients comorbidities were determined using the Charlson Comorbidity Index based on the International Classification of Diseases (ICD) diagnosis codes. Pre-operative AF was defined as a documented AF in patient records including ECG findings and diagnosis codes. The primary outcome of the current study was new-onset POAF, defined as an AF event captured within 30 days after the cardiac surgery in a patient with no prior history of AF. A post-operative AF event had to be documented in at least one of the following sources: 12-lead ECG, STS post-operative records, and post-operative clinical notes. We also included the following secondary outcomes: mortality, 30-day re-admission, hospital/ICU length of stay (LOS), and incident stroke. Our practice for ligation of atrial appendage Our surgical practice akin to many others held a strong conviction regarding the ligation of the atrial appendage in patients undergoing cardiac surgery, despite the lack of evidence-based support for this practice. Many claimed experts without robust evidence claim that if cardiac surgery is performed, the left atrial appendage should be ligated at the time. They posited that post-cardiac surgery, the majority of patients develop atrial fibrillation, and that ligating the atrial appendage could potentially mitigate both short-term and long-term complications associated with atrial fibrillation post-surgery. While this argument holds logic it was firmly followed in our practice however, it lacks supportive data-driven evidence. Propensity score matching Twenty-one baseline characteristic covariates (see Table 1 for the list of variables) were incorporated into a logistic regression model with LAAC as the dependent variable. The model was then used to estimate a propensity score for each patient. To select controls, we restricted the distance between a matched pair to be at most 0.2 of the standard deviation of the logit of the propensity score distribution. A nearest neighbor, without replacement, greedy 1-to-1 caliper matching process was used to match LAAC patients with patients from the no-LAAC group. To assess the balance of the matched groups on the baseline variables, we used an absolute standardized mean difference (ASMD) < 0.10 to indicate acceptable difference between groups for each variable. Matching was accomplished using the R package MatchIt (version 4.5.5) [17]. Table 1Characterization of the patients in the original unmatched cohortCharacteristicLAAC (n = 169)no-LAAC(n = 1890) p Age (Years)65 (56, 71)65 (57, 71)0.936Female56 (33.1%)451 (23.9%)0.007BMI (KG/m2)28.4 (25.1, 32.3)29.7 (26.1, 34)0.002Charlson Comorbidity Index5 (3, 7)5 (3, 7)0.228CHA2DS2VASc3 (2, 4)3 (2, 4)0.180HAS-BLED Score3 (2, 4)3 (2, 4)0.507Ejection fraction (EF)57 (47, 62)57 (47, 60)0.430Left ventricular end-systolic diameter (LVESD)34 (29.5, 38)32 (28, 38)0.257Left ventricular end-diastolic diameter (LVEDD)50 (45, 57)48 (43, 53)0.005Coronary artery bypass graft (CABG) only86 (50.9%)1424 (75.3%)< 0.001Aortic valve (AV) only21 (12.4%)279 (14.8%)0.410Mitral valve (MV) only34 (20.1%)79 (4.2%)< 0.001CABG and (AV or MV)21 (12.4%)81 (4.3%)< 0.001Acute myocardial infarction55 (32.5%)669 (35.4%)0.457Congestive heart failure69 (40.8%)570 (30.2%)0.004Peripheral vascular disease60 (35.5%)568 (30.1%)0.140Cerebrovascular disease51 (30.2%)413 (21.9%)0.013Chronic pulmonary disease (COPD)51 (30.2%)544 (28.8%)0.702Hypertension122 (72.2%)1519 (80.4%)0.011Diabetes63 (37.3%)787 (41.6%)0.270Renal disease39 (23.1%)376 (19.9%)0.323 Characterization of the patients in the original unmatched cohort Statistical analysis For the unmatched cohort, descriptive statistics for categorical variables were reported as percentages, and the chi-square test was used to assess whether two proportions are different from each other. Continuous variables were summarized as medians and interquartile ranges (IQR), and the non-parametric Wilcoxon Mann–Whitney rank sum test was used to assess differences in distribution between the two sex groups. For the matched cohort, the following statistical tests were utilized that accounted for the paired samples. Differences in categorical and continuous variables in the two groups were assessed using McNemar and Wilcoxon signed rank tests, respectively. Independent associations between LAAC and targeted binary and continuous outcomes were established using conditional logistic and median regressions, respectively. All statistical analyses were performed using R version 4.0.3. Results The unmatched study cohort included 169 LAAC and 1890 no-LAAC patients. Table 1 summarizes the characterization of the patients in the unmatched cohort. The median age in both groups was 65 years. The percentage of women was significantly (p = 0.007) higher in LAAC group (33.1% vs. 23.9%). Patients in LAAC group had significantly higher rates of congestive heart failure (40.8% vs. 30.2%, p = 0.004) and cerebrovascular disease (30.2% vs. 21.9%, p = 0.013). Combined CABG and valve procedures were significantly higher in the LAAC group (12.4% vs. 4.3%, p < 0.001). Charlson comorbidity index, CHA2DS2VASC, and HAS-BLED score were not different in the two groups. The matched cohort included 162 matched LAAC and no-LAAC pairs. Table 2 reports the frequencies of LAAC and no-LAAC characteristics in the unmatched and matched cohorts. The ASMD for all covariates in the matched cohort was less than < 0.10 which indicates a well-matched cohort. Table 2Frequencies of LAAC and no-LAAC characteristics in the unmatched and propensity-matched cohortsCharacteristicUnmatched cohortPSM cohortLAAC (n = 169)no-LAAC (n = 1890)ASMDLAAC (n = 162)no-LAAC (n = 162)ASMDAge (Years), mean (SD)63.2 (11.9)63.4 (10.9)0.01963.4 (11.7)63.1 (10.8)0.021Female, n (%)56 (33.1)451 (23.9)0.19749 (30.2)46 (28.4)0.039BMI (KG/m2), mean (SD)29 (5.7)30.5 (6.2)0.27129.2 (5.6)29.1 (5.4)0.008Charlson Comorbidity Index, mean (SD)5.3 (3)5.1 (3)0.0685.3 (3)5.3 (3.3)0.013CHA2DS2VASC, mean (SD)3.1 (1.6)2.8 (1.5)0.1313 (1.6)2.9 (1.7)0.053HAS-BLED Score, mean (SD)2.9 (1.4)2.9 (1.3)0.0642.9 (1.4)2.8 (1.5)0.048Ejection fraction (EF), mean (SD)53.1 (11.1)52.2 (11.4)0.07952.8 (11.1)53.5 (11.2)0.061LVESD, mean (SD)33.1 (5.7)32.9 (6.6)0.02933.1 (5.8)33.3 (6.6)0.042LVEDD, mean (SD)49.7 (7.2)48.3 (7.1)0.20249.6 (7.2)49.8 (6.7)0.027Gender (Female), n (%)56 (33.1)451 (23.9)0.19749 (30.2)46 (28.4)0.039Coronary artery bypass graft (CABG) only, n (%)86 (50.9)1424 (75.3)0.48986 (53.1)84 (51.9)0.025Aortic valve (AV) only, n (%)21 (12.4)279 (14.8)0.07121 (13)23 (14.2)0.037Mitral valve (MV) only, n (%)34 (20.1)79 (4.2)0.39827 (16.7)29 (17.9)0.031CABG and (AV or MV), n (%)21 (12.4)81 (4.3)0.24721 (13)22 (13.6)0.019Acute myocardial infarction, n (%)55 (32.5)669 (35.4)0.06153 (32.7)58 (35.8)0.066Congestive heart failure, n (%)69 (40.8)570 (30.2)0.21764 (39.5)63 (38.9)0.013Peripheral vascular disease, n (%)60 (35.5)568 (30.1)0.11457 (35.2)52 (32.1)0.064Cerebrovascular disease, n (%)51 (30.2)413 (21.9)0.18146 (28.4)51 (31.5)0.067Chronic pulmonary disease (COPD), n (%)51 (30.2)544 (28.8)0.0349 (30.2)54 (33.3)0.067Hypertension, n (%)122 (72.2)1519 (80.4)0.183119 (73.5)122 (75.3)0.041Diabetes, n (%)63 (37.3)787 (41.6)0.0960 (37)58 (35.8)0.026Renal disease, n (%)39 (23.1)376 (19.9)0.07638 (23.5)35 (21.6)0.044 Frequencies of LAAC and no-LAAC characteristics in the unmatched and propensity-matched cohorts Comparison of outcomes in the matched cohort are shown in Table 3. LAAC was independently associated with new-onset POAF (41.4% vs. 25.3%, p = 0.003). Also, LAAC was independently associated with longer post-operative hospital stay (142.1 vs. 120.5 h, p = 0.001). Interestingly, the LAAC was not associated with elevated risks of 1-year mortality, 30-day re-admission, or 1-year incident stroke. Finally, Table 4 shows estimates of the treatment effects in terms of odds ratio (OR) and standardized coefficients for binary and continuous outcomes (respectively). Notably, out of the eight outcomes, LAAC was an independent predictor of new-onset POAF (OR = 2.18,CI = 1.32–3.61, p = 0.002) and postoperative hospital stay (SC = 0.019, CI=-0.012-0.034, p = 0.006). Table 3Comparison of postoperative outcomes in the propensity-matched cohortOutcomeLAAC(n = 162)no-LAAC(n = 162) p new-onset POAF67(41.4%)41(25.3%)0.00330-day readmission4(2.5%)3(1.9%)0.9991-year mortality7(4.3%)7(4.3%)0.9991-year incident stroke6(3.7%)5(3.1%)0.999postoperative hospital LOS (hours)142.1(116.8-191.8)120.5(97.9-145.6)0.001 Comparison of postoperative outcomes in the propensity-matched cohort Table 4Conditional logistic and median regression models. For numeric outcomes, standardized coefficient (SC), 95% CI for the SC, and p-value are reportedOutcomeOR/SC95% CI p new-onset POAF2.181.32–3.610.00230-day readmission1.330.3–5.960.7061-year mortality10.35–2.850.9991-year incident stroke1.20.37–3.930.763postoperative hospital LOS (hours)0.019-0.012-0.0340.006 Conditional logistic and median regression models. For numeric outcomes, standardized coefficient (SC), 95% CI for the SC, and p-value are reported Discussion POAF is a common complication occurring in 20-50% of cardiac surgery patients [18–23]. Episodes of POAF are often transient and follow a typical time course, peaking at 2–4 days after surgery [19] and resolving within 4–6 weeks [24]. Nonetheless, those patients with transient POAF might frequently have AF recurrences [25–27] as well as elevated risk of stroke and mortality [25–27]. The popularity of the left appendage clip has triggered interest of applying it in patients who are at risk of stroke and have no history of AF [13]. However, whether patients with no pre-operative AF may benefit from LAAC remains unknown [14]. The ongoing left atrial appendage closure by surgery-2 (LAACS-2) trial [14] is the first study to enroll patients with and without history of AF. Results from this ongoing trial are expected to determine if the prevention of stroke events using LAAC is achievable in first-time open-heart surgery patients despite their AF status and CHA2DS2VASC score. Another ongoing trial is the Left Atrial Appendage Exclusion for Prophylactic Stroke Reduction Trial (LeAAPS), a randomized multicenter trial that aims to evaluate the effectiveness of LAAC for the prevention of ischemic stroke or systemic arterial embolism in cardiac surgery patients who have no history of AF, but have risk factors for atrial fibrillation and ischemic stroke (ClinicalTrials.gov identifier: NCT05478304). Because LAAC could be a promising alternative to oral anticoagulation (OAC) therapy for preventing or decreasing the risk of stroke, most prospective and retrospective clinical studies evaluating the efficacy and safety of LAAC focused on patient populations with history of AF. Recently, the ATLAS trial was a feasibility study monitoring a subset of cardiac surgery patients for whom the risks of thromboembolic events and bleeding intersected and thus had equipoise regarding LAA management and anticoagulation limitations [13]. The ATLAS inclusion criteria included a CHA2DS2VASc score > = 2, encompassing 376 patients undergoing LAAC and 186 without LAAC. POAF among those who received LAAC compared to no-LAAC was (47.3 vs. 38.2%; P = 0.047). Of note, all the patient demographics and patient characteristics among LAAC and no-LAAC were not statistically significant. Despite differences in the study design, findings from our study were in agreement with the findings from the ATLAS trial. Specifically, both studies demonstrated significant increase in the risk of new-onset POAF after cardiac surgery in patients underwent LAAC. While the ATLAS study suggested lower risk for thromboembolic events in LAAC patients with POAF, our study demonstrated non-significant lower risk of mortality and 30-day readmission in LAAC patients. Further, our results showed significant increase in hospital LOS. Possible explanations for the increased risks of POAF following atrial clip exclusion include (but are not limited to) local atrial infarction and inflammation from the atrial, serial pericarditis, and a reduction in atrial compliance, resulting in more effective volume in the left atrium [6, 13]. The outcome of more POAF in those without a prior history of AF needs to be emphasized to the patient because POAF may lead to higher long-term risk of atrial fibrillation [28]. Given increasing understanding of implications of AF across domains (e.g., risk of heart failure), subjecting a patient without a prior history of AF to LAAC may place that patient at risk of long-term health conditions. Similarly, while the incidence of short-term complications related to LAAC is low, heart torsion, left atrial appendage bleeding, and circumflex artery pinching have been reported 29,30 . The potential justification for performing LAAC in patients without a history of atrial fibrillation is that they will be protected from prothrombotic events if they do develop POAF and may be able to forego anticoagulation and its associated risks. It is our belief, based on the accumulated data, including the findings of our study herein, that preemptive LAAC in cardiac surgery with sinus rhythm is not justified unless future high-quality RCTs should definitively demonstrate benefits that outweigh the risk [16]. The current study has limitations inherent in a small, single center, retrospective study. The small size of the participants in the LAAC group (162 patients) may have implications for the statistical power and generalizability. Further, the study might lack sufficient power to detect some endpoints with low prevalence, such as the 1-year risk of incidence stroke. Finally, there might be confounding variables that were not included in our propensity matching analysis because these variables were unknown to us or not included in our cohort. Conclusion This single center retrospective matched cohort study of 162 LAAC and no-LAAC open heart-surgery patients demonstrated a significant association between LAAC and new-onset POAF. However, LAAC did not impact other outcomes including 30-day re-admission, 1-year mortality, and 1-year incident stroke. Acknowledgements All the surgeons at Geisinger who managed cardiac surgery patients. Sarah Knorr-STS Data Coordinator for extracting STS records for patients in the study cohort. Author contributions JS: conceptualization, methodology, writing- original draft -review & editing. MA: data curation, formal analysis, writing -review & editing. TM: methodology, writing -review & editing. HLK: methodology, writing -review & editing. YE: methodology, supervision, writing- original draft -review & editing. Funding The authors declare no funding relevant to the study. Data availability No datasets were generated or analysed during the current study. Declarations Ethical approval This study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board at Geisinger Medical Center, with the approval number: 2020–0101. The need for informed consent was waived by the IRB due the retrospective nature of the study. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement No datasets were generated or analysed during the current study.