Development and Validation of a Postoperative Prognostic Nomogram to Predict Recurrence in Patients with Persistent Atrial Fibrillation: A Retrospective Cohort Study

Abstract

Background: Patients with persistent atrial fibrillation (PsAF) have a high risk of recurrence after catheter radiofrequency ablation. Nevertheless, no effective prognostic tools have been developed to identify these high-risk patients to date. This study sought to develop and validate a simple linear predictive model for predicting postoperative recurrence in patients with PsAF.

Methods: From June 2013 to June 2021, patients with PsAF admitted to our hospital were enrolled in this single-center, retrospective, observational study. The characteristics substantially associated with recurrence in patients with PsAF were screened through univariate and multivariate logistic regression analysis. The receiver operating characteristic curve was used to assess the predictive significance of the nomogram model after nomogram development. Furthermore, to assess the clinical value of the nomogram, we performed calibration curve and decision curve analyses.

Results: A total of 209 patients were included in the study, 42 (20.10%) of whom were monitored up to 1 year for recurrent AF. The duration of AF episodes, left atrial diameter, BMI, CKMB, and alcohol consumption were found to be independent risk factors (P<0.05) and were integrated into the nomogram model development. The area under the curve was 0.895, the sensitivity was 93.3%, and the specificity was 71.4%, thus indicating the model’s excellent predictive ability. The C-index of the predictive nomogram model was 0.906. Calibration curve and decision curve analyses further revealed that the model had robust prediction and strong discrimination ability.

Conclusion: This simple, practical, and innovative nomogram can help clinicians in evaluation of the risk of PsAF recurrence after catheter ablation, thus facilitating preoperative evaluation, postoperative monitoring and ultimately the construction of more personalized therapeutic protocols.

Main article text

Introduction

Atrial fibrillation (AF), one of the most common supraventricular arrhythmias characterized by the rapid and irregular rhythm of the atria, currently affects more than 33 million people globally, and its prevalence is expected to more than double in the next 40 years [1]. Untimely and non-standard AF treatment significantly affects patients’ quality of life, including progressive heart failure, stroke, sudden cardiac death, and non-cardiovascular death [2]. Therefore, prompt intervention is crucial to avoid disease progression in patients with AF.

Despite the increasing types of surgical procedures available for AF ablation, surgery remains the most effective therapeutic strategy for AF, and the cornerstone of ablation is pulmonary venous isolation [3, 4]. The postoperative recurrence rate of PsAF is significantly higher than that of paroxysmal AF, and patients are more likely to experience cardiac dysfunction and stroke [5]. Debate among experts persists regarding whether persistent AF might cause severe atrial muscle fibrosis and atrial dysfunction, which would activate atrial reversion [6–8].

Only a few predictive models for postoperative recurrence have been developed on the basis of these factors, although numerous studies have analyzed the causes of postoperative recurrence, including the linear ablation method and site [9], atrial scar [10], sex [11], duration [6], sleep apnea syndrome [12], and left atrial diameter (LAD) [13]. However, nomograms are a commonly used diagnostic tool that can predict the probability of clinical events by incorporating potential risk factors, and can also be used to determine the clinical prognosis regarding disease outcomes. The objective of the study was to retrospectively analyze the risk factors for recurrence in patients with persistent atrial fibrillation. An additional aim was to analyze and establish an appropriate predictive model for postoperative recurrence, to improve the success rates of surgery and monitoring of patients at high risk, and to enhance patient quality of life and prognosis.

Methods and Patients

Patients

Patients with PsAF who were admitted to the cardiovascular medicine department at the Affiliated Changzhou No.2 People’s Hospital of Nanjing Medical University between June 2013 and June 2021 were examined in this retrospective study. The clinical characteristics of patients with PsAF were thoroughly reviewed and analyzed. The inclusion criteria were patients with 1) persistent atrial fibrillation lasting more than 7 days (consistent with the 2014 AHA guidelines) [14]; 2) no prior AF surgical history; and 3) electrocardiogram abnormalities corresponding to atrial fibrillation.

The exclusion criteria were 1) patients with severe pulmonary, cardiovascular, hepatic, or renal diseases; 2) patients with coagulation dysfunction, left atrium and left atrial appendage thrombosis; 3) patients with hypertensive heart disease, congenital heart disease, or valvular heart disease; 4) patients with rheumatic valvular diseases, immune-associated diseases, or malignant tumors; and 5) patients with missing information or who were unreachable.

Postoperative recurrence was defined by signs and symptoms of atrial tachycardia, atrial fibrillation, and atrial flutter 3 months after AF ablation (as confirmed by a definite electrocardiogram or dynamic electrocardiogram). The Ethics Review Committee of the Affiliated Changzhou No.2 People’s Hospital of Nanjing Medical University approved this study. Because this was a retrospective study, informed consent from eligible patients was not required.

Clinical characteristics

The electronic medical record system at the hospital was used to extract the fundamental clinical characteristics, laboratory test results, treatments, and outcomes of individuals. Sex, smoking status, alcohol consumption, age, body weight, and body mass index (BMI) were among the basic characteristics analyzed. Medical history included complications (such as hypertension, diabetes, hyperthyroidism, pulmonary hypertension, sleep apnea syndrome, acute myocardial infarction, coronary heart disease, stroke, syncope, and heart failure); calculated CHA2DS2-VASc Scores (two points each were assigned for age ≥75 years and for history of stroke, TIA, or thromboembolism, and 1 point was assigned for each of the following items: congestive heart failure, hypertension, diabetes mellitus, age 65–75 years, vascular disease (defined as previous myocardial infarction, complex aortic plaque, carotid stenosis, and peripheral artery disease), and female sex category); types of anticoagulants taken (warfarin, dabigatran, rivaroxaban, and others); whether antiarrhythmic medication (amiodarone) was taken; and whether mechanical electrocardiography was performed during the operation (if sinus rhythm was not restored at the completion of radiofrequency ablation, electrical cardioversion was performed on the operating table). Left atrial anterior and posterior diameter, ejection fraction (EF), left ventricular end-diastolic diameter (LVEDD), ventricular septal thickness (IVS), left ventricular diameter shortening fraction, and other cardiac ultrasound parameters were among the clinical and examination information included. Preoperative venous testing was conducted for laboratory indicators such as white blood cells, neutrophil ratio, red blood cell distribution width, creatine kinase, troponin, c-reactive protein, blood creatinine, brain natriuretic peptide (BNP), cholesterol, triglycerides, low-density lipoprotein cholesterol, thyroid-stimulating hormone, and three free iodine thyroid-related.

Catheter ablation procedure

After insertion of three catheters through the right femoral vein, a ten-pole catheter into the coronary sinus, two long shears into the atrial septum twice, a beacon electrode (PentaRay), and an ablation catheter (Johnson’s ST head), the procedure was completed. An electrode (PentaRay) for mapping was used to create the left atrium model. The site of radiofrequency ablation was isolated at a distance of 4 mm from the pulmonary vein entrance. The complete isolation of both bilateral pulmonary veins was considered the hallmark of successful surgery, and auxiliary ablation lines such as the left atrium, posterior wall, atrial low-voltage scar area, and even mitral isthmus were added when required. Electrical cardioversion was used to ensure sinus rhythm after surgery if the sinus rhythm of patients with PsAF was not restored after isolation. Anticoagulants were administered preoperatively for at least 3 weeks, with esophageal ultrasound, CT, or intracardiac ultrasound, to ensure the absence of thrombi in the atria. Anticoagulants were taken postoperatively for at least 3 months until reevaluation.

Patient follow-up and study endpoint

In addition to outpatient visits, all patients were followed up by telephone. Ambulatory follow-up was conducted every 3 months during the first year, and follow-up and examinations were conducted every 3–6 months thereafter. All patients were followed up for 1 year. The definition of recurrence was as stated above, but was supported by clear ECG or ambulatory ECG evidence.

Statistical analysis

Data analysis was conducted in SPSS version 22.0 (IBM, Chicago, IL, USA) and R (version 4.0.3), and a two-sided P-value <0.05 was considered statistically significant. Categorical variables were examined with the Pearson chi-square or Fisher exact test, and are expressed as frequency and percentage. Univariate analysis was used to compare the baseline characteristics of the recurrence group and non-recurrence group to assess independent predictors of recurrence. Multivariate logistic regression analysis was then used to further analyze significantly different variables. The final included variables were obtained with the stepwise forward method, and the variance expansion factor was calculated to determine whether each factor displayed multicollinearity. Finally, with these variables, a predictive model was developed.

To generate forest plots, we performed univariate and multivariate analyses with the “Forest Plot” software package in R Software (version 4.0.3). The original data were resampled with the bootstrap method, and the nomogram model was used to forecast recurrence after resampling. With the “rms” package, nomogram analysis for predicting recurrence in patients with PsAF integrated duration of AF episodes, LAD, BMI, CKMB, and alcohol consumption. Receiver operating characteristic curve analysis was performed with the “pROC” package software in R to evaluate the ability of the combined model to predict postoperative recurrence in patients with PsAF. Subsequently, the calibration curve and decision curve were plotted with the “ggDCA” package to verify our nomogram’s predictive accuracy and reliability.

Results

Patient Characteristics

Retrospective analysis was performed on all clinical baseline characteristics of the 238 patients with PsAF. Among them, 29 patients were excluded on the basis of the exclusion criteria, and 209 patients were included in this study (Figure 1). Table 1 summarizes the clinical data for the two groups of patients. After radiofrequency catheter ablation surgery, 42 (20.10%) patients were unable to maintain sinus rhythm continuously. The recurrence group consisted of 27 men and 15 women, with a median age of 66.42 years. The average body weight of patients in the recurrence group was 75.4±11.1 kg, and echocardiography revealed an average LAD of 49.4±4.6 mm; the group included 15 patients (35.7%) who experienced sustained atrial fibrillation for 3–6 years, 8 patients (19%) for more than 6 years; 29 patients (69%) who had hypertension, 21 patients (50%) who had a long history of drinking, and 31 patients (73.8%) who underwent electrical conversion during surgery.

Figure 1

Flowchart of Patient Enrollment.

Table 1

Comparison of Clinical Data between Patients with Recurrent and Non-Recurrent Atrial Fibrillation after Surgery.

	Recurrence (n=42)	Non-recurrence (n=167)	P value
Age, years	66 (57–69)	67 (57–72)	0.77
Male, n(%)	27 (64.3)	107 (64.1)	0.98
BMI (kg/m2)	27.3 (24.7–29.1)	23.6 (21.9–26.1)	<0.001
COPD, n(%)	1 (2.4)	4 (2.4)	0.1
PAH, n(%)	0 (0)	2 (1.2)	0.48
SAS, n(%)	2 (4.8)	5 (3)	0.57
Smoke, n(%)	17 (40.5)	46 (27.7)	0.1
Syncope, n(%)	0 (0)	1 (0.6)	0.62
Drink, n(%)	21 (50.0)	44 (26.5)	0.003
HBP, n(%)	29 (69.0)	90 (54.2)	0.08
CHD, n(%)	4 (9.5)	20 (12)	0.66
OMI, n(%)	0 (0)	3 (1.8)	0.38
DM, n(%)	8 (19.1)	30 (17.9)	0.87
Stroke, n(%)	0 (0)	8 (4.8)	0.15
Hyperlipidemia, n(%)	10 (23.8)	33 (19.8)	0.56
HF, n(%)	9 (21.4)	35 (21.0)	0.82
Hyperthyroidism, n(%)	2 (4.8)	2 (1.2)	0.13
Hypothyroidism, n(%)	0 (0)	1 (0.6)	0.62
Prior stent implantation, n(%)	1 (2.4)	1 (0.6)	0.3
CKMB (U/L)	14.6 (12.4–16.5)	13.1 (11.0–15.6)	0.014
TNI (μg/L)	0.01 (0.01–0.01)	0.01 (0.01–0.01)	0.92
BNP (pg/mL)	546 (190–985)	782 (374–1390)	0.38
CR (μmol/L)	85.9 (66.0–106.5)	71.0 (63.0–86.0)	0.01
TCH (mmol/L)	3.7 (3.4–4.7)	3.9 (3.5–4.6)	0.32
TG (mmol/L)	1.3 (1.0–2.1)	1.3 (1.1–2.0)	0.84
LDL (mmol/L)	2.1±0.6	2.2±0.7	0.26
LVEDD (mm)	51.0 (49.5–54.5)	50.0 (47.0–54.0)	0.05
LVPWT (mm)	10.0 (9.3–10.0)	9.0 (9.0–10.0)	0.05
LVESD (mm)	36.0 (34.5–37.5)	35.0 (32.0–37.0)	0.04
LAD (mm)	49.4±4.6	42.1±5.3	<0.001
LDFS (%)	29.0 (27.5–31.5)	31.0 (28.0–33.0)	0.24
EF (%)	55.0 (52.5–58.0)	58.0 (54.0–61.3)	0.1
IVS (mm)	10.0 (10.0–11.0)	10.0 (9.0–10.0)	0.019
Electrical cardioversion, n(%)	29 (70.7)	88 (52.7)	0.08
CHA2DS2-VASc Scores	2.0(2.0–3.0)	2.0(2.0–3.0)	0.69
Antiarrhythmic drugs (Amiodarone), n(%)	17(40.5)	55(33.1)	0.37
Duration grouping, n	42	167	<0.001
<1 year, n(%)	13 (31.0)	69 (41.6)
1–3 year, n(%)	6 (14.3)	72 (43.4)
3–6 year, n(%)	15 (35.7)	20 (12.0)
>6 year, n(%)	8 (19.0)	5 (3.0)
Anticoagulants, n	42	167	0.04
Warfarin, n(%)	14 (33.3)	41 (24.6)
Dabigatran, n(%)	14 (33.3)	31 (18.6)
Rivaroxaban, n(%)	11 (26.2)	81 (48.5)
Others, n(%)	3 (7.1)	14 (8.4)

BMI, body mass index; BNP, brain natriuretic peptide; CHD, coronary heart disease; CKMB, creatine kinase-mb; COPD, chronic obstructive pulmonary disease; CR, creatinine; DM, diabetes mellitus; EF, ejection fraction; HBP, high blood pressure; HF, heart failure; IVS, thickness of the interval; LAD, left atrial diameter; LDFS, the left indoor diameter shortens the fraction; LDL, low density lipoprotein; LVEDD, left ventricular end diastolic diameter; LVESD, left ventricular end contractile diameter; LVPWT, posterior wall thickness of the left ventricle; OMI, old myocardial infarction; PAH, pulmonary arterial hypertension; SAS, sleep apnea syndrome; CH, total cholesterol; TG, triglyceride.

Evaluation of Independent Prognostic Factors

Basic demographic factors, vital signs, and laboratory test results in the primary cohort were further analyzed with the univariate logistic model to identify potential indicators (Table 2). Univariate regression analysis (P<0.05) demonstrated that variables such as weight, BMI, TT, CKMB, LVESD, LAD, IVS, drinking history, CR, type of anticoagulant, and duration of AF episodes were potential predictors. Five prognostic factors – BMI, CKMB, LAD, drinking history, and duration of AF episodes – were included in the final predictive model (each P<0.05) after further analysis of all candidate factors with a multivariable logistic regression model (Table 2 and Figure 2). The forward stepwise variable selection method was used, wherein the variables were sequentially entered into the model, as shown in Figure 3.

Table 2

Univariate and Multivariate Regression Analysis of Recurrence Risk Factors for Patients with Atrial Fibrillation after Radio Frequency Ablation.

	Univariate model			Multivariable model
	OR	95% CI	P value	OR	95% CI	P value
Weight	1.075	1.040–1.110	<0.001	0.977	0.905–1.055	0.554
TT	1.008	0.998–1.017	0.112
BMI	1.327	1.174–1.500	<0.001	1.181	1.006–1.386	0.043
CR	1.021	1.003–1.040	0.023	1.023	0.995–1.052	0.113
CKMB	1.093	1.018–1.174	0.014	1.098	1.006–1.386	0.037
LVESD	1.045	0.970–1.126	0.242
LAD	1.251	1.156–1.355	<0.001	1.261	1.141–1.394	<0.001
IVS	1.416	1.017–1.970	0.039	1.012	0.908–1.128	0.832
Drink			0.004			0.032
No	1			1
Yes	0.358	0.178–0.717	0.004	0.354	0.136–0.934	0.032
Duration grouping			<0.001			<0.001
<1 year	1			1
1–3 year	0.118	0.033–0.417	0.001	0.073	0.015–0.362	0.001
3–6 year	0.052	0.013–0.210	<0.001	0.04	0.007–0.222	<0.001
>6 year	0.127	0.127–1.274	0.254	0.354	0.354–0.069	0.214
Anticoagulants			0.05
Warfarin	1
Dabigatran	1.593	0.398–6.377	0.51
Rivaroxaban	2.108	0.521–8.526	0.296
Others	0.634	0.157–2.562	0.522

BMI, body mass index; CKMB, creatine kinase-mb; CR, creatinine; IVS, thickness of the interval; LAD, left atrial diameter; LVESD, left ventricular end contractile diameter; TT, thrombin time.

Figure 2

Forest Plots based on Univariate and Multivariate Analysis of Risk Factors for Patients with PsAF under Radiofrequency Ablation.

Figure 3

Changes in AUC Values of the Model after Gradual Inclusion of Variables with the Forward Method.

The inclusion of clinical variables improved the AUC from 0.814 to 0.906. The final model, consisting of LAD, duration of AF episodes, BMI, drinking history, and CKMB level, was a good predictor of postoperative recurrence in patients with PsAF.

Construction and Performance Evaluation of the Prognostic Nomogram

The duration of AF episodes, LAD, BMI, CKMB, and drinking history were used to construct a nomogram for predicting recurrence in patients with PsAF, as shown in Figure 4. Each independent prognostic parameter was given a weighted score to create the nomogram, and each predictor was then connected by a straight line to the points axis to represent a particular point. A higher recurrence probability was represented by a higher total score, which was calculated by summation of the points allotted to the nomogram’s prognostic factors.

Figure 4

Nomogram Model of Recurrence Risk after Radiofrequency Ablation for Patients with Atrial Fibrillation.

To evaluate the nomogram’s discrimination ability, calibration, and clinical practicability, we performed receiver operating characteristic curve, C-index, calibration curve, and decision curve analyses with 1000 bootstrap resamplings. The C-index of the prognostic nomogram was 0.906, and the calibration curves for probabilities in recurrence prediction showed a good fit of the nomogram for recruitment prediction. The AUC was 0.895, and the sensitivity and specificity were 93.3% and 71.4%, respectively, thus implying that the nomogram model had good discrimination ability. Moreover, the nomogram predictive model was valuable in making effective judgments, and the decision curve analysis curve demonstrated its net benefit across a wide range of threshold probabilities (Figure 5).

Figure 5

Evaluation of the Nomogram Model.

Discussion

Catheter ablation is the mainstay therapy for PsAF. However, this procedure is limited by the high rate of long-term recurrence. The clinical data for 209 patients with PsAF were retrieved after database screening for this single-center retrospective investigation. To create a prognostic nomogram, we examined factors that significantly differed between the groups with and without recurrence. This novel nomographic chart showed satisfactory performance according to C-index, AUC, calibration curve, and decision curve analyses. Thus, the nomographic chart can be effectively implemented clinically.

We followed patients for 1 year after surgery: 42 of the 209 patients had AF recurrence, based on ECG or Holter data. In the analysis of recurrent PsAF, we identified five risk factors on the basis of statistics: BMI, CKMB, LAD, duration of AF episodes, and alcohol consumption history, in agreement with previous observations. Higher BMI has previously been shown to increase the risk of atrial fibrillation by at least 50%, and BMI is used as a proxy for obesity [15], possibly because of its effects on cardiac muscle structure by increasing oxidative stress or fat cell apoptosis [16, 17]. Furthermore, our statistical results revealed that high BMI was an independent risk factor and was more prevalent in the recurrence group. Moreover, most patients with AF were clinically obese. Atrial structural remodeling [18] is therefore a key etiology for persistent AF [1], thus leading to a series of physiological and pathological changes, because the complexity of atrial muscle conduction determines the electrophysiological basis of AF [19]. Furthermore, atrial electrical remodeling [1] and atrial fibrosis [18] affect redox signaling [20]. Studies on the effects of atrial diameter and AF duration on the recurrence rate in patients with AF have been conducted in centers worldwide. Both factors have been shown to affect cardiac function recovery and mortality in patients with AF and heart failure [21–23]. We observed that recurrence was more prevalent in patients with larger atrial size and longer duration of AF episodes, thus corroborating previous results. We divided the patients into four groups in our study, on the basis of AF episode duration. The recurrence rate was significantly higher in patients at 1–3 years and 3–6 years. Thus, our findings provide a solid basis and a useful tool for follow-up before and after surgery.

CKMB is a marker of myocardial injury and can also cause atrial inflammation [24]. Moreover, CKMB has been found to be associated with early postoperative recurrence [25] and to specifically predict the incidence of atrial fibrillation in patients with acute coronary syndrome within 3–5 years [26]. However, insufficient in-depth studies have been performed on this mechanism. Numerous large population-based meta-analyses have demonstrated that drinking history – whether light to moderate or heavy – is an AF risk factor [27, 28], whereas abstinence from alcohol significantly decreases the risk of AF recurrence [29]. Alcohol consumption has complex pathological effects on AF and significantly correlates with the duration and amount of alcohol consumed [30], probably because of electrical remodeling, which affects the autonomic nervous response and alters excitation-contraction coupling [31, 32].

Here, we built a strong prognostic nomogram including five significantly distinct variables in the predictive model for postoperative recurrence. However, our nomogram did not include several previously reported predictors, such as BNP, hypertension, diabetes, white blood cells, and red blood cell distribution width. The following factors may account for errors in research: 1) Most previous reports have argued that BNP is associated with recurrence in patients with AF with heart failure complicated by insufficient EF [33, 34]. Although many of our patients had cardiac dysfunction, most showed normal ejection fractions. The variation might have been due to the different patient populations examined. 2) Hypertension, diabetes, and other comorbidities are risk factors for AF. However, because of the current disease detection rate, and because the disease risk factors did not accurately reflect the risk factors for postoperative recurrence of the disease, most of our patients had either both or only one of these comorbidities [35]. Thus, further testing in larger sample-size studies might be warranted. 3) Differences in instruments and techniques might have influenced certain test indicators, such as the neutrophil/lymphocyte ratio and RBC distribution width. Although some studies have linked these factors to relapse [36, 37], most of these variables were not included in the predictive models developed and validated in a large prospective cohort. 4) Patients with PsAF made up most of the study population, and alternative study topics might have produced different conclusions.

Nevertheless, our study has several limitations. First, this was a single-center retrospective study with a small sample size. Second, our analysis did not quantify the indicators, such as alcohol consumption, whose specific dose was not determined; there are individual differences in how much each person drinks and how much each person defines as habitual drinking. Additionally, the study period commenced in 2013, and surgical indications and technology have continually been updated, thus resulting in a non-standardized patient baseline. Finally, our nomogram has not been validated in other hospitals. Therefore, additional study is required to evaluate the reliability of our proposed nomogram model.

Conclusion

In conclusion, we developed a nomogram with five baseline predictors that are easy to use, then performed internal validation. To improve outcome management, this tool would allow physicians to assess and predict postoperative recurrence in patients with PsAF and identify high-risk individuals. According to the findings, our nomogram predictive model has good clinical translational value. We recommend further validation with external cohorts with large sample sizes.

Ethics Statement

The studies involving human participants were reviewed and approved by Ethics Committee of The Second People’s Hospital of Changzhou, Nanjing Medical University. The patients/participants provided their written informed consent to participate in this study.

Authorship Contribution Statement

Cong-Ying Deng: Conceptualization, Methodology, Software analysis, Investigation and Writing of the original draft. Ling Sun: Validation, Software analysis, Visualization and Writing. Ailin Zou: Data collection and processing, and Review and editing. Yuan Ji: Validation, Formal analysis and Visualization.

Conflicts of Interest

None.

Citation Information

Download Citation.

References

1. FE Mason, J Pronto, K Alhussini, C Maack, N Voigt. Cellular and mitochondrial mechanisms of atrial fibrillation. Basic Res Cardiol 2020;115(6):72.

2. H Calkins, G Hindricks, R Cappato, YH Kim, EB Saad, L Aguinaga, et al. 2017 HRS/EHRA/ECAS/APHRS/SOLAECE expert consensus statement on catheter and surgical ablation of atrial fibrillation: executive summary. Europace 2018;20(1):157–208.

3. JG Andrade, J Champagne, M Dubuc, MW Deyell, A Verma, L Macle, et al. Cryoballoon or radiofrequency ablation for atrial fibrillation assessed by continuous monitoring: a randomized clinical trial. Circulation 2019;140(22):1779–88.

4. M Mansour, H Calkins, J Osorio, SJ Pollak, D Melby, FE Marchlinski, et al. Persistent atrial fibrillation ablation with contact force-sensing catheter: the prospective multicenter PRECEPT trial. JACC Clin Electrophysiol 2020;6(8):958–69.

5. M Alkhouli, PA Friedman. Ischemic stroke risk in patients with nonvalvular atrial fibrillation: JACC review topic of the week. J Am Coll Cardiol 2019;74(24):3050–65.

6. HS Lim, M Hocini, R Dubois, A Denis, N Derval, S Zellerhoff, et al. Complexity and distribution of drivers in relation to duration of persistent atrial fibrillation. J Am Coll Cardiol 2017;69(10):1257–69.

7. G Cheniti, K Vlachos, T Pambrun, D Hooks, A Frontera, M Takigawa, et al. Atrial fibrillation mechanisms and implications for catheter ablation. Front Physiol 2018;9:1458.

8. S Zahid, H Cochet, PM Boyle, EL Schwarz, KN Whyte, EJ Vigmond, et al. Patient-derived models link re-entrant driver localization in atrial fibrillation to fibrosis spatial pattern. Cardiovasc Res 2016;110(3):443–54.

9. JS Sutter, Y Lokhnygina, JP Daubert, T Bahnson, K Jackson, JI Koontz, et al. Safety and efficacy outcomes of left atrial posterior wall isolation compared to pulmonary vein isolation and pulmonary vein isolation with linear ablation for the treatment of persistent atrial fibrillation. Am Heart J 2020;220:89–96.

10. U Canpolat, S Mohanty, C Trivedi, Q Chen, H Ayhan, C Gianni, et al. Association of fragmented QRS with left atrial scarring in patients with persistent atrial fibrillation undergoing radiofrequency catheter ablation. Heart Rhythm 2020;17(2):203–10.

11. H Sugumar, S Nanayakkara, D Chieng, GR Wong, R Parameswaran, RD Anderson, et al. Arrhythmia recurrence is more common in females undergoing multiple catheter ablation procedures for persistent atrial fibrillation: time to close the gender gap. Heart Rhythm 2020;17(5 Pt A):692–8.

12. AS Fein, A Shvilkin, D Shah, CI Haffajee, S Das, K Kumar, et al. Treatment of obstructive sleep apnea reduces the risk of atrial fibrillation recurrence after catheter ablation. J Am Coll Cardiol 2013;62(4):300–5.

13. A Berruezo, D Tamborero, L Mont, B Benito, JM Tolosana, M Sitges, et al. Pre-procedural predictors of atrial fibrillation recurrence after circumferential pulmonary vein ablation. Eur Heart J 2007;28(7):836–41.

14. CT January, LS Wann, JS Alpert, H Calkins, JE Cigarroa, JC Cleveland Jr, et al. 2014 AHA/ACC/HRS guideline for the management of patients with atrial fibrillation: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and the Heart Rhythm Society. J Am Coll Cardiol 2014;64(21):e1–76.

15. TJ Wang, H Parise, D Levy, RB D’Agostino Sr, PA Wolf, RS Vasan, et al. Obesity and the risk of new-onset atrial fibrillation. J Am Med Assoc 2004;292(20):2471–7.

16. T Feng, M Vegard, LB Strand, LE Laugsand, B Mørkedal, D Aune, et al. Weight and weight change and risk of atrial fibrillation: the HUNT study. Eur Heart J 2019;40(34):2859–66.

17. M Al-Rawahi, R Proietti, G Thanassoulis. Pericardial fat and atrial fibrillation: epidemiology, mechanisms and interventions. Int J Cardiol 2015;195:98–103.

18. J Jalife, O Berenfeld, M Mansour. Mother rotors and fibrillatory conduction: a mechanism of atrial fibrillation. Cardiovasc Res 2002;54(2):204–16.

19. D Li, L Zhang, J Kneller, S Nattel. Potential ionic mechanism for repolarization differences between canine right and left atrium. Circ Res 2001;88(11):1168–75.

20. JN Simon, K Ziberna, B Casadei. Compromised redox homeostasis, altered nitroso-redox balance, and therapeutic possibilities in atrial fibrillation. Cardiovasc Res 2016;109(4):510–8.

21. SI Sarvari, KH Haugaa, TM Stokke, HZ Ansari, IS Leren, F Hegbom, et al. Strain echocardiographic assessment of left atrial function predicts recurrence of atrial fibrillation. Eur Heart J Cardiovasc Imaging 2016;17(6):660–7.

22. A Njoku, M Kannabhiran, R Arora, P Reddy, R Gopinathannair, D Lakkireddy, et al. Left atrial volume predicts atrial fibrillation recurrence after radiofrequency ablation: a meta-analysis. Europace 2018;20(1):33–42.

23. JG Andrade, MW Deyell, A Verma, L Macle, J Champagne, P Leong-Sit, et al. Association of atrial fibrillation episode duration with arrhythmia recurrence following ablation: a secondary analysis of a randomized clinical trial. JAMA Netw Open 2020;3(7):e208748.

24. L Wu, RW Emmens, J van Wezenbeek, W Stooker, CP Allaart, ABA Vonk, et al. Atrial inflammation in different atrial fibrillation subtypes and its relation with clinical risk factors. Clin Res Cardiol 2020;109(10):1271–81.

25. HS Lim, C Schultz, J Dang, M Alasady, DH Lau, AG Brooks, et al. Time course of inflammation, myocardial injury, and prothrombotic response after radiofrequency catheter ablation for atrial fibrillation. Circ Arrhythm Electrophysiol 2014;7(1):83–9.

26. H Zhang, P Dong, X Yang, L Du, K Wang, P Yan, et al. Prognostic indicators of new onset atrial fibrillation in patients with acute coronary syndrome. Clin Cardiol 2020;43(6):647–51.

27. SC Larsson, N Drca, A Wolk. Alcohol consumption and risk of atrial fibrillation: a prospective study and dose-response meta-analysis. J Am Coll Cardiol 2014;64(3):281–9.

28. S Kodama, K Saito, S Tanaka, C Horikawa, A Saito, Y Heianza, et al. Alcohol consumption and risk of atrial fibrillation: a meta-analysis. J Am Coll Cardiol 2011;57(4):427–36.

29. A Voskoboinik, JM Kalman, A De Silva, T Nicholls, B Costello, S Nanayakkara, et al. Alcohol abstinence in drinkers with atrial fibrillation. N Engl J Med 2020;382(1):20–8.

30. A Voskoboinik, S Prabhu, LH Ling, JM Kalman, PM Kistler. Alcohol and atrial fibrillation: a sobering review. J Am Coll Cardiol 2016;68(23):2567–76.

31. MR Piano, C Rosenblum, RJ Solaro, D Schwertz. Calcium sensitivity and the effect of the calcium sensitizing drug pimobendan in the alcoholic isolated rat atrium. J Cardiovasc Pharmacol 1999;33(2):237–42.

32. T Maki, L Toivonen, P Koskinen, H Näveri, M Härkönen, H Leinonen. Effect of ethanol drinking, hangover, and exercise on adrenergic activity and heart rate variability in patients with a history of alcohol-induced atrial fibrillation. Am J Cardiol 1998;82(3):317–22.

33. R Rorth, PS Jhund, MB Yilmaz, SL Kristensen, P Welsh, AS Desai, et al. Comparison of BNP and NT-proBNP in patients with heart failure and reduced ejection fraction. Circ Heart Fail 2020;13(2):e006541.

34. D Grieco, Z Palama, A Borrelli, E De Ruvo, L Sciarra, A Scarà, et al. Diabetes mellitus and atrial remodelling in patients with paroxysmal atrial fibrillation: role of electroanatomical mapping and catheter ablation. Diab Vasc Dis Res 2018;15(3):185–95.

35. Q Wang, X Sun, C Zhuo, M Chen, M Wu, J Zhao, et al. Investigation of proper home blood pressure with atrial fibrillation recurrence in patients undergoing radiofrequency catheter ablation. Minerva Cardioangiol 2020;68(3):224–33.

36. AC Valenti, M Vitolo, JF Imberti, VL Malavasi, G Boriani. Red cell distribution width: a routinely available biomarker with important clinical implications in patients with atrial fibrillation. Curr Pharm Des 2021;27(37):3901–12.

37. Q Shao, K Chen, SW Rha, HE Lim, G Li, T Liu. Usefulness of neutrophil/lymphocyte ratio as a predictor of atrial fibrillation: a meta-analysis. Arch Med Res 2015;46(3):199–206.