Introduction

Nonvalvular atrial fibrillation (NVAF) is a common arrhythmia that increases the risk for ischemic stroke by 4- to 5-fold, according to current clinical reports.^1,2 Direct oral anticoagulants (DOACs) are considered as highly effective anticoagulants for thrombosis prevention in patients with NVAF recommended by the current guidelines,³ with the advantages of fixed-dose regimens without the requirement of coagulation monitoring. Previous studies indicated that DOACs were associated with better safety and effectiveness in patients across all body mass index (BMI) categories compared to warfarin.^4,5 However, these clinical trials usually lacked the effect of BMI on thromboembolism. Concerning that more and more ischemic events were reported in patients with higher BMIs,^6,7 this “one-size-fits-all” strategy in DOACs raises significant concerns about the effectiveness and safety of DOACs in these patients.

According to previous studies, a higher BMI is regarded as a well-established risk of ischemic events due to the larger body surface area or underexposed of DOACs.^8,9 Besides, positive linear relationship revealed a relation of BMI levels and the incidence of thrombosis in overweight patients with NVAF.¹⁰ However, no guidelines or recommendations for DOACs use have been published specifically among the overweight patients. To improve the estimation of thrombotic risk among overweight patients with NVAF, a new risk prediction model seems to be urgently needed.

Recently, the CHADS₂ or CHA₂DS₂-VASc risk scores are recommended as risk-predicting approaches for anticoagulation decision in NVAF.^11,12 However, studies indicated that these risk scores have only moderate ability in predicting the risk of stroke.¹³ There is even a lack of evidence on the performance of these 2 risk scores in overweight patients. Combining the significant incremental risk predictors of age, stroke, and heart failure, the major component of CHADS₂ or CHA₂DS₂-VASc score, we aimed to develop a new risk score with BMI categories (namely the AB₂S risk score) to predict the thrombotic events in overweight patients with NVAF. We also investigated the incremental value of this AB₂S risk score by comparing with the CHADS₂ or CHA₂DS₂-VASc risk scores.

Methods

Study Design and Population

The study was a retrospective single-center observational study and enrolled a total of 952 consecutive overweight patients with NVAF between January 2017 and December 2018. Nonvalvular atrial fibrillation was diagnosed according to the European Society of Cardiology criteria: absolutely irregular RR intervals and no discernible and distinct P waves presented on electrocardiogram.¹⁴

Patients with age ≥ 18, BMI ≥ 25 kg/m², high risk of stroke and systemic embolism, and DOAC therapy during hospitalization were eligible. Exclusion criteria were patients with incomplete records and severe liver or renal dysfunction or patients lost follow-up.

The study was conducted in accordance with the Basic and Clinical Pharmacology and Toxicology Policy for Experimental and Clinical Studies.¹⁵

Medication and Radiofrequency Ablation Procedure

The radiofrequency ablation procedure was performed, as described previously.¹⁶ In brief, all patients were treated with dabigatran or rivaroxaban for at least 4 consecutive weeks to achieve stable anticoagulation, discontinued 24 hours before scheduled catheter ablation, and resumed 3 to 4 hours after removing the sheath. Transesophageal echocardiography was applied during the transseptal puncture. Other medications used included beta-blockers, antihypertensives, antiarrhythmics, antiplatelet agents, and proton pump inhibitors at the physician’s discretion.

Data Collection

The following demographic and clinical baseline data were obtained from electronic medical records: (1) age, gender, and BMI; (2) comorbidity (history of hypertension, hyperlipidemia, diabetes mellitus, chronic kidney disease, liver disease, stroke, heart failure, and peripheral arterial disease; (3) biochemical blood indicators including estimated glomerular filtration rate (eGFR), platelets (PLT), hemoglobin, activated partial prothrombin time, thrombin time, and left ventricular ejection fraction (LVEF).

Clinical Definition

The age was categorized as <65, 65 to 74, or ≥75 years old. The BMI was categorized as 25 to 30, 30 to 34, or ≥35 kg/m² according to the World Health Organization. The eGFR was dichotomized into ≥60 versus <60 mL/min per 1.73 m². Platelet counts <125 × 10⁹/L and LVEF at <40% were derived from the previous report.¹⁰ The CHA₂DS₂-VAS_C and CHADS₂ scores were calculated to assess the risk of thrombosis.¹⁷

Follow-up and Study Outcomes

All patients were followed up once a month for 12 months. The primary endpoint was the recurrence of systemic thromboembolism, which is defined as pulmonary embolism (PE), venous thromboembolism (VTE), stroke, or cardiac embolism. Events were assessed according to the Antithrombotic Therapy and Prevention of Thrombosis Guidelines.¹⁸ The time to thrombosis (TTT)—defined as the time to the first stroke or systemic embolism from enrollment—was documented.

Model Development and Evaluation

Before constructing the predictive model, the logistic regression (LR) analysis was adopted to screen the candidate predictor variables using the “Regression Modeling Strategies” package in R software. The results in the LR analysis were represented by odds ratio with 95% CI and P-value.¹⁹ These candidate predictor variables with statistical significance were finally used to develop a clinical prediction risk thromboembolic score (namely AB₂S score) and presented as a nomogram model using R software.²⁰

The measures of model fit, including the area under curve (AUC) of receiver operating characteristic curve and the C -statistic, were calculated via the R software.^21,22 Besides, to minimize false-positive variable selection, a corrected calibration curve that includes 2000 bootstrap samples was used to verify the correlation between the calibration curve and the standard curve.²³ To verify the clinical efficacy of the AB₂S score model, the decision analysis curve (DAC) was employed by analyzing the net benefit under different risk thresholds compared with CHA₂DS₂-VASc and CHADS₂ scores.²⁴

Statistical Analysis

Quantitative data were reported as means with SDs (mean ± SD) and compared by Student’s t test. Qualitative data were presented as absolute numbers and percentages compared by chi-square test or Fisher’s exact probability test. The backward stepwise model selection procedure was applied first to evaluate the variables with best model fit by using R software (version 4.11). The variables with P < .10 in the univariate LR analysis and P < .05 in the multivariate LR analysis were determined as candidate predictor variables. The AB₂S risk score for an individual patient was determined by assigning points for each candidate predictor variables present and summing. In addition, we collapsed the AB₂S score into 3 categories on the basis of the risk score, including low-risk groups, mediated-risk groups, and high-risk groups. Cumulative systemic thromboembolic incidence was estimated by the Kaplan–Meier method, and differences were assessed with the log-rank test for the 3 categorized AB₂S risk groups using Statistical Package for the Social Sciences Statistics software (version 25.0). The performance comparison of the 3-category AB₂S score to the CHADS₂ and CHA₂DS₂-VASc scores was performed in terms of the AUC and DAC improvement by using R software (version 4.11).^22,24

Results

Patient Characteristics

A total of 952 overweight patients with NVAF were retrospectively enrolled in our study. A total of 22 (2.3%) systemic thromboembolism patients were observed during the 12-month follow-up, including 7 (31.8%) in VTE, 5 (22.8%) in stroke, 6 (27.3%) in PE, and 4 (18.2%) in cardiac embolism. Statistically significant differences were found between the systemic thromboembolism and nonsystemic thromboembolism groups in terms of age (75.82 ± 9.89 vs. 65.62 ± 12.14, P < .001), BMI (32.17 ± 4.14 vs. 29.76 ± 3.71, P = .003), and eGFR (63.50 ± 17.49 vs. 79.45 ± 27.40, P = .007). The 2 groups were comparable regarding other baseline characteristics, comorbidity, or concomitant medications. Details are listed in Table 1.

Construction and Validation of the ABS Risk Score

After LR analysis, a total of 7 candidate predictor variables were chosen into final nomogram model, included age, BMI, PLT, eGFR, LVEF, history of stroke, and heart failure. The prediction rule for AB₂S risk score assigned 2 points for BMI ≥ 35 kg/m² and history of stroke, and 1 point for age ≥ 75 years, BMI from 30 to 34 kg/m², eGFR < 60 mL/min per 1.73 m², PLT count < 125 × 10⁹/L, LVEF <40%, and history of heart failure (Table 2).

The total score of each variables was calculated in the final nomogram model (Figure 1A), which ranged from 0 to 300, and the corresponding risk rate ranged from 0.1 to 0.9. The higher the total score, the higher the risk of systemic thromboembolism in overweight patients with NVAF. The score was 180 points for “BMI ≥ 35 kg/m²” in this nomogram model. The c-index for the final model was 0.885 (95% CI, 0.814-0.954), and AUC was 0.885 (95% CI, 0.815-0.955) (Figure 1B). The corrected calibration curve showed a goodness of fit of the model, with a mean absolute error of 0.01 and a mean squared error of 0.00042 (Figure 1C).

Distributions of ABS score

The AB₂S score ranged from 0 to 8, with the majority of patients in 2 points (n = 273, 28.7%). The majority of patients were classified as 1 and 2 points in the CHADS₂ and CHA₂DS₂-VASc risk scores, respectively (Figure 2A). By collapsing the AB₂S score into low (0 points), moderate (1-2 points), and high (3-8 points) risk categories, the incidence of systemic thromboembolism was 2 (9.1%), 9 (40.9%), and 11 (50%) in the low-, mediated-, and high-risk groups, respectively (Figure 2B). While the incidence was 4 (18.2%), 13 (59.1%), and 5 (22.7%) in the 3-category CHADS₂ scores and 0, 1 (4.5%), and 21 (95.5%) in the 3-category CHA₂DS₂-VASc scores.

Predicting Clinical Outcome of ABS Score

Predicting the 1-year systemic thromboembolism, the AUC score was 0.76 (95% CI, 0.74-0.79) for the 3-category AB₂S score. While predicting the PE, stroke, and PTE events, the AUC AB₂S score was 0.76 (95% CI, 0.74-0.79), 0.76 (95% CI, 0.74-0.79), and 0.76 (95% CI, 0.74-0.79), respectively (Figure 3).

Predicting the TTT, the Kaplan-Meier results illustrated high-risk category of AB₂S score experienced a shorter TTT than the low-risk category (11.1 ± 0.32 months vs. 11.96 ± 0.03 months, P < .001). Details are listed in Figure 4.

Comparison of the ABS, CHADS and CHADS-VASc Scores

The 3-category AB₂S score had a greater AUC in predicting 1-year systemic thromboembolism, PE, stroke, and PTE events than both CHADS₂ and CHA₂DS₂-VASc risk scores (Figure 3). At the same time, there were statistical differences in TTT between the 3-category AB₂S and CHADS₂ scores (P < .0001, P < .002, respectively). However, TTT were similar in the 3-category CHA₂DS₂-VASc scores (P = .082) (Figure 4).

In addition, the DCA curve showed that the 3-category AB₂S score revealed quite good clinical efficacy than both CHADS₂ and CHA₂DS₂-VASc scores in predicting 1-year systemic thromboembolism, when the risk threshold was between 35% and 80% (Figure 5).

Discussion

In the present study, we developed a AB₂S score that accurately predicted 1-year systemic thromboembolism in overweight patients with NVAF and showed good discrimination in risk stratification than CHADS₂ and CHA₂DS₂-VASc scores.

In our study, a total of 952 overweight patients with NVAF were included. We found that the 1-year systemic thromboembolic event rate was 2.3%, although every individual was anticoagulated. Such higher thromboembolic event rate in our study may be understood through several explanations. First, both the factors of NVAF and overweight highly increased the risk for ischemic events according to current clinical reports.^1,2,10 Besides, the thromboembolic event was defined as systemic thromboembolic event, which included PE, VTE, stroke, or cardiac embolism. Anyway, we call for high-quality researches to continue exploring the relationship of overweight and thromboembolic events.

The AB₂S score comprised 7 independent predictors: age, BMI, PLT, eGFR, LVEF, history of stroke, and heart failure, which were simple to use in clinical practice. Several thrombotic predictors have been reported previously, including age, heart failure, or stroke.^11,25 However, we found BMI was significantly correlated with the incidence of 1-year systemic thromboembolism in overweight patients with NVAF (P < .001), together with eGFR, PLT, and LVEF. Previous study demonstrated that higher BMIs usually indicated poor clinical outcomes for NVAF patients treated with DOAC by altering peak plasma concentrations.^7,26 Based on the evidence that the higher BMIs experienced a shorter TTT (HR = 3.716, P  = .001)¹⁰ and increased thrombosis occur in these patients,²⁷ one meta-analysis even concluded that a weight-based dosage adjustment may be necessary for thromboembolic prevention in overweight patients with NVAF.²⁸ Most importantly, we used a broader range of BMI categories, a decision consistent with multiple prior reports.^10,16 We found strong amplification of 1-year systemic thromboembolism across the entire BMI range, with individuals’ BMI ≥ 35 kg/m² at nearly double the risk of those BMI < 30 kg/m² (P < .001). However, individuals who had a history of stroke were at elevated risk regardless of BMI. Thus, BMI and history of stroke were the dominant risk factors in our AB₂S risk model.

Other factors, including eGFR, PLT, and LVEF, were also observed significantly associated with the incidence of 1-year systemic thromboembolism and were selected as potent predictors of systemic embolism in overweight patients with NAVF in our study. This results were partially consistent with previous researches.^29,30 Unfortunately, as with multiple prior studies, we did not observe significant incremental risk prediction from gender, hypertension, and diabetes mellitus, the major component of CHA₂DS₂-VASc score.^30,31 Therefore, a larger sample of randomized controlled studies is needed for further analysis.

Compared to CHA₂DS₂-VASc scores, we found AB₂S score performed better in predicting 1-year systemic thromboembolism in overweight patients with NVAF. The AUC was greater, and there was a positive net reclassification improvement. Although CHA₂DS₂-VASc score may be helpful in several clinical settings, for example, a CHA₂DS₂-VASc score of 0 point indicated a small net clinical benefit from warfarin anticoagulation.^32,33 In addition, this score could aid in decision-making for patients with atrial fibrillation or acute coronary syndrome undergoing percutaneous coronary intervention to balance the risk of hemorrhage from the underlying thrombotic risk.³⁴ However, CHA₂DS₂-VASc score showed a poor predictive performance (AUC = 0.62) in the 1-year systemic thromboembolism and TTT (P = .082) than the AB₂S score in our study. It is highly likely that the traditional CHA₂DS₂-VASc risk score overestimates the thromboembolic risk in overweight patients with NVAF. This results may be comprehensible with several explanations. First, our study did not include the patients who received warfarin anticoagulation. It is known that CHA₂DS₂-VASc scores have superior prediction for the underlying thrombotic risk of using warfarin.³⁵ Second, our study population included only overweight patients (BMI ≥ 25 kg/m²) with NVAF. To date, no BMI parameter is included in any risk stratification scheme for thromboembolism in atrial fibrillation patients.³⁶ Thus, the predictive performance of CHA₂DS₂-VASc scores may be significantly eliminated, as with multiple prior studies.^36,37 Third, 1-year systemic thromboembolism was defined as our endpoint event, while CHA₂DS₂-VASc risk scores are well known in predicting the risk of ischemic stroke.^38,39

Study Limitations

Our study has the following limitations: (1) the study had a small sample size with a short follow-up, (2) the incidence of in the 1-year systemic thromboembolism could be underestimated as data were obtained from the medical records, and (3) our data were obtained from a Chinese NVAF overweight population treated with DOACs. The results may not apply to other races or settings.

Conclusion

Our results highlighted the importance of including multiple categories of BMI in determining systemic thromboembolism risk in overweight patients with NVAF. The AB₂S score’s performance showed a greater AUC and clinical efficacy value than the widely used CHADS₂ and CHA₂DS₂-VASc score in terms of systemic thromboembolism prediction, which may aid in decision-making for these patients to balance the effectiveness of anticoagulation from the underlying thrombotic risk.

Footnotes

Ethics Committee Approval: The study was approved by the Ethics Committee of Zhongshan Hospital, Fudan University (approval number B2021-030).

Informed Consent: Written informed consent was obtained from the patients who agreed to take part in the study.

Peer-review: Externally peer-reviewed.

Author Contributions: Concept – Q.Z.L.; Design – Q.Z.L.; Supervision – M.X.Y., X.Y.L.; Resources – L.L.Z., Z.W., Q.Z.L.; Materials – L.L.Z., Z.W.; Data Collection and/or Processing – M.X.Y., X.Y.L.; Analysis and/or Interpretation – M.X.Y., X.Y.L.; Literature Search – L.L.Z., Z.W.; Writing – M.X.Y., X.Y.L.; Critical Review – M.X.Y., X.Y.L., L.L.Z., Z.W., Q.Z.L.

Declaration of Interests: The authors have no conflict of interest to declare.

References

1. Benjamin EJ, Muntner P, Alonso A. Heart disease and stroke Statistics-2019 update: a report from the American Heart Association. Circulation. 2019;139(10):e56-e528. https://doi.org/10.1161/CIR.0000000000000659
2. Group M, January CT, Wann LS. 2019 AHA/ACC/HRS focused update of the 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 clinical practice guidelines and the heart rhythm society. Heart Rhythm. 2019;16(8):e66-e93.
3. Burnett AE, Mahan CE, Vazquez SR, Oertel LB, Garcia DA, Ansell J. Guidance for the practical management of the direct oral anticoagulants (DOACs) in VTE treatment. J Thromb Thrombolysis. 2016;41(1):206-232. https://doi.org/10.1007/s11239-015-1310-7
4. Kaplan RM, Tanaka Y, Passman RS. Efficacy and safety of direct oral anticoagulants for atrial fibrillation across body mass index categories. J Am Heart Assoc. 2020;9(24):e017383-. https://doi.org/10.1161/JAHA.120.017383
5. Barakat AF, Jain S, Masri A. Outcomes of direct oral anticoagulants in atrial fibrillation patients across different body mass index categories. JACC Clin Electrophysiol. 2021;7(5):649-658. https://doi.org/10.1016/j.jacep.2021.02.002
6. Breuer L, Ringwald J, Schwab S, Köhrmann M. Ischemic stroke in an obese patient receiving dabigatran. N Engl J Med. 2013;368(25):2440-2442. https://doi.org/10.1056/NEJMc1215900
7. Lucijanic M, Jurin I, Jurin H. Patients with higher body mass index treated with direct/novel oral anticoagulants (DOAC/NOAC) for atrial fibrillation experience worse clinical outcomes. Int J Cardiol. 2020;301():90-95. https://doi.org/10.1016/j.ijcard.2019.10.035
8. Hanley MJ, Abernethy DR, Greenblatt DJ. Effect of obesity on the pharmacokinetics of drugs in humans. Clin Pharmacokinet. 2010;49(2):71-87. https://doi.org/10.2165/11318100-000000000-00000
9. Wang HJ, Si QJ, Shan ZL. Effects of body mass index on risks for ischemic stroke, thromboembolism, and mortality in Chinese atrial fibrillation patients: a single-center experience. PLoS One. 2015;10(4):e0123516-. https://doi.org/10.1371/journal.pone.0123516
10. Li X, Zuo C, Ji Q, Xue Y, Wang Z, Lv Q. Body mass index influence on the clinical outcomes for nonvalvular atrial fibrillation patients admitted to a hospital treated with direct oral anticoagulants: a retrospective cohort study. Drug Des Devel Ther. 2021;15():1931-1943. https://doi.org/10.2147/DDDT.S303219
11. Gage BF, Waterman AD, Shannon W, Boechler M, Rich MW, Radford MJ. Validation of clinical classification schemes for predicting stroke: results from the National Registry of atrial fibrillation. JAMA. 2001;285(22):2864-2870. https://doi.org/10.1001/jama.285.22.2864
12. Lip GY, Nieuwlaat R, Pisters R, Lane DA, Crijns HJ. Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the euro heart survey on atrial fibrillation. Chest. 2010;137(2):263-272. https://doi.org/10.1378/chest.09-1584
13. Van Staa TP, Setakis E, Di Tanna GL, Lane DA, Lip GY. A comparison of risk stratification schemes for stroke in 79,884 atrial fibrillation patients in general practice. J Thromb Haemost. 2011;9(1):39-48. https://doi.org/10.1111/j.1538-7836.2010.04085.x
14. Kirchhof P, Benussi S, Kotecha D. 2016 ESC Guidelines for the management of atrial fibrillation developed in collaboration with EACTS. Eur Heart J. 2016;37(38):2893-2962. https://doi.org/10.1093/eurheartj/ehw210
15. Tveden-Nyborg P, Bergmann TK, Jessen N, Simonsen U, Lykkesfeldt J. BCPT policy for experimental and clinical studies. Basic Clin Pharmacol Toxicol. 2021;128(1):4-8. https://doi.org/10.1111/bcpt.13492
16. Yu M, Wang Z, Zong L, Xu Q, Li X, Lv Q. A retrospective cohort study of the effectiveness and safety of dabigatran versus Rivaroxaban in overweight patients with nonvalvular atrial fibrillation. Int J Clin Pharm. 2022;44(5):1149-1157. https://doi.org/10.1007/s11096-022-01443-2
17. Güler E, Babur Güler G, Demir GG, Hatipoğlu S. A review of the fixed dose use of new oral anticoagulants in obese patients: is it really enough?. Anatol J Cardiol. 2015;15(12):1020-1029. https://doi.org/10.5152/AnatolJCardiol.2015.6532
18. Falck-Ytter Y, Francis CW, Johanson NA. Prevention of VTE in orthopedic surgery patients: antithrombotic therapy and prevention of thrombosis, 9th ed: American college of chest physicians evidence-based clinical practice guidelines. Chest. 2012;141(2):e278S-e325S. https://doi.org/10.1378/chest.11-2404
19. Wang Y, Zhang Y, Wang K. Nomogram model for screening the risk of Type II diabetes in western Xinjiang, China. Diabetes Metab Syndr Obes. 2021;14():3541-3553. https://doi.org/10.2147/DMSO.S313838
20. Peng B, Min R, Liao Y, Yu A. Development of predictive nomograms for clinical use to quantify the risk of amputation in patients with diabetic foot ulcer. J Diabetes Res. 2021;2021():6621035-. https://doi.org/10.1155/2021/6621035
21. Pencina MJ, D'Agostino RB. Overall C as a measure of discrimination in survival analysis: model specific population value and confidence interval estimation. Stat Med. 2004;23(13):2109-2123. https://doi.org/10.1002/sim.1802
22. Lätti S, Niinivehmas S, Pentikäinen OT. Rocker: open source, easy-to-use tool for AUC and enrichment calculations and ROC visualization. J Cheminform. 2016;8(1):45-. https://doi.org/10.1186/s13321-016-0158-y
23. Kramer AA, Zimmerman JE. Assessing the calibration of mortality benchmarks in critical care: the Hosmer-Lemeshow test revisited. Crit Care Med. 2007;35(9):2052-2056. https://doi.org/10.1097/01.CCM.0000275267.64078.B0
24. Fitzgerald M, Saville BR, Lewis RJ. Decision curve analysis. JAMA. 2015;313(4):409-410. https://doi.org/10.1001/jama.2015.37
25. Reilly PA, Lehr T, Haertter S. The effect of dabigatran plasma concentrations and patient characteristics on the frequency of ischemic stroke and major bleeding in atrial fibrillation patients: the RE-LY Trial (Randomized Evaluation of Long-Term Anticoagulation Therapy). J Am Coll Cardiol. 2014;63(4):321-328. https://doi.org/10.1016/j.jacc.2013.07.104
26. Piran S, Traquair H, Chan N, Bhagirath V, Schulman S. Peak plasma concentration of direct oral anticoagulants in obese patients weighing over 120 kilograms: a retrospective study. Res Pract Thromb Haemost. 2018;2(4):684-688. https://doi.org/10.1002/rth2.12146
27. López-López JA, Sterne JAC, Thom HHZ. Oral anticoagulants for prevention of stroke in atrial fibrillation: systematic review, network meta-analysis, and cost effectiveness analysis. BMJ. 2017;359():j5058-. https://doi.org/10.1136/bmj.j5058
28. Malik AH, Yandrapalli S, Shetty S. Impact of weight on the efficacy and safety of direct-acting oral anticoagulants in patients with non-valvular atrial fibrillation: a meta-analysis. Europace. 2020;22(3):361-367. https://doi.org/10.1093/europace/euz361
29. Go AS, Fang MC, Udaltsova N. Impact of proteinuria and glomerular filtration rate on risk of thromboembolism in atrial fibrillation: the anticoagulation and risk factors in atrial fibrillation (ATRIA) study. Circulation. 2009;119(10):1363-1369. https://doi.org/10.1161/CIRCULATIONAHA.108.816082
30. Natsuaki M, Morimoto T, Yamaji K. Prediction of thrombotic and bleeding events after percutaneous coronary intervention: CREDO-Kyoto thrombotic and bleeding risk scores. J Am Heart Assoc. 2018;7(11):-. https://doi.org/10.1161/JAHA.118.008708
31. Singer DE, Chang Y, Borowsky LH. A new risk scheme to predict ischemic stroke and other thromboembolism in atrial fibrillation: the ATRIA study stroke risk score. J Am Heart Assoc. 2013;2(3):e000250-. https://doi.org/10.1161/JAHA.113.000250
32. Eckman MH, Singer DE, Rosand J, Greenberg SM. Moving the tipping point: the decision to anticoagulate patients with atrial fibrillation. Circ Cardiovasc Qual Outcomes. 2011;4(1):14-21. https://doi.org/10.1161/CIRCOUTCOMES.110.958108
33. Singer DE, Chang Y, Fang MC. The net clinical benefit of warfarin anticoagulation in atrial fibrillation. Ann Intern Med. 2009;151(5):297-305. https://doi.org/10.7326/0003-4819-151-5-200909010-00003
34. Wu Y, Xie Z, Liang W. Usefulness of CHADS2, R2CHADS2, and CHA2DS2-VASc scores for predicting incident atrial fibrillation in heart failure with preserved ejection fraction patients. ESC Heart Fail. 2021;8(2):1369-1377. https://doi.org/10.1002/ehf2.13217
35. Bernaitis N, Ching CK, Badrick T, Anoopkumar-Dukie S. Identifying warfarin control with stroke and bleed risk scores. Heart Lung Circ. 2018;27(6):756-759. https://doi.org/10.1016/j.hlc.2017.11.009
36. Overvad TF, Rasmussen LH, Skjøth F, Overvad K, Lip GY, Larsen TB. Body mass index and adverse events in patients with incident atrial fibrillation. Am J Med. 2013;126(7):640.e9-640.17. https://doi.org/10.1016/j.amjmed.2012.11.024
37. Cohoon KP, McBane RD, Ammash N, Slusser JP, Grill DE, Wysokinski WE. Relationship between body mass index and left atrial appendage thrombus in nonvalvular atrial fibrillation. J Thromb Thrombolysis. 2016;41(4):613-618. https://doi.org/10.1007/s11239-015-1266-7
38. Chan YH, Yiu KH, Lau KK. The CHADS2 and CHA2DS2-VASc scores predict adverse vascular function, ischemic stroke and cardiovascular death in high-risk patients without atrial fibrillation: role of incorporating PR prolongation. Atherosclerosis. 2014;237(2):504-513. https://doi.org/10.1016/j.atherosclerosis.2014.08.026
39. Komatsu T, Sato Y, Ozawa M. Comparison between CHADS2 and CHA2DS2-VASc score for risk stratification of ischemic stroke in Japanese patients with non-valvular paroxysmal atrial fibrillation not receiving anticoagulant therapy. Int Heart J. 2014;55(2):119-125. https://doi.org/10.1536/ihj.13-242