Archives of Academic Emergency Medicine. 2023; 11(1): e8 REV I EW ART I C L E Value of N-Terminal Pro-Brain Natriuretic Peptide for Em- bolic Events Risk Prediction in Patients with Atrial Fibril- lation; a Systematic Review and Meta-Analysis Koohyar Ahmadzadeh1, Amirali Hajebi2, Hamzah Adel Ramawad3, Yaser Azizi1,4∗, Mahmoud Yousefifard1,5 † 1. Physiology Research Center, Iran University of Medical Sciences, Tehran, Iran. 2. Non-Communicable Diseases Research Center, Endocrinology and Metabolism Population Sciences Institute, Tehran University of Medical Sciences, Tehran, Iran. 3. Department of Emergency Medicine, NYC Health & Hospitals Coney Island, New York. 4. Department of Physiology, School of Medicine, Iran University of Medical Sciences, Tehran, Iran. 5. Pediatric Chronic Kidney Disease Research Center, Tehran University of Medical Sciences, Tehran, Iran. Received: October 2022; Accepted: December 2022; Published online: 1 January 2023 Abstract: Introduction: A comprehensive conclusion has yet to be made about the predictive value of serum N-terminal pro-brain natriuretic peptide (NT-proBNP) for stroke/systemic embolic events (SEE) in patients with atrial fib- rillation (AF). This study aims to review the evidence for evaluating the value of NT-proBNP in predicting the risk of stroke/SEE in patients with AF through a systematic review and meta-analysis. Methods: Two independent reviewers screened all relevant studies that were retrieved from the database of Medline, Embase, Scopus, and Web of Science until December 7th, 2021. The predictive value of NT-proBNP in the prediction of stroke/SEE was recorded as hazard ratio (HR) and 95% confidence interval (95% CI). Results: Nine articles (38,093 patients, 3.10% stroke/SEE) were included in our analysis. There was no publication bias in these studies (P=0.320). Our analysis showed that NT-proBNP can be a good predictor of stroke/SEE risk in AF patients, even at different cut-off values (HR=1.76; 95% CI: 1.51, 2.02; P < 0.001). Subgroup analysis showed that diabetes could have a possible effect on the predictive value of NT-proBNP (meta-regression coefficient = 0.042; P = 0.037). Conclu- sion: Measurement of NT-proBNP during the first admission could be used to assess the short- or long-term risk of stroke/SEE in patients with AF. Further studies are needed to evaluate the possible applicability of serum NT-proBNP measurement in the settings in which stroke is the sole outcome of the investigation. Keywords: Pro-brain natriuretic peptide; Stroke; Embolism, Paradoxical; Atrial fibrillation; Meta-analysis Cite this article as: Ahmadzadeh K, Hajebi A, Ramawad HA, Azizi Y, Yousefifard M. Value of N-Terminal Pro-Brain Natriuretic Peptide for Embolic Events Risk Prediction in Patients with Atrial Fibrillation; a Systematic Review and Meta-Analysis. Arch Acad Emerg Med. 2023; 11(1): e8. https://doi.org/10.22037/aaem.v11i1.1808. 1. Introduction Stroke and thromboembolic events are considered as one of the leading causes of death and disability worldwide (1, 2). In the past, stroke was believed to be a disease that mostly af- ∗Corresponding Author: Yaser Azizi; Physiology Research Center, Iran Uni- versity of Medical Sciences, Tehran, Iran. Tel: +982188622709; Email: az- izi.y@iums.ac.ir, ORCID: https://orcid.org/???. † Corresponding Author: Mahmoud Yousefifard; Physiology Research Cen- ter, School of Medicine, Iran University of Medical Sciences, Shahid Hemmat Highway, Tehran 14496-14535, Iran. Tel: +98 (21) 86704771; Email: yousefi- fard.m@iums.ac.ir / yousefifard20@gmail.com, ORCID: https://orcid.org/???. fected the elderly population. However, recent studies have shown that the incidence of stroke in the younger population is increasing (3-5). Regardless of the age group, there is a high prevalence of cardiovascular accidents worldwide, which re- sults in lifelong disabilities and death in severe cases (6, 7). There are many risk factors associated with a higher oc- currence of stroke, such as hypertension, diabetes, old age, smoking, and sedentary lifestyle (8, 9). Atrial fibrillation (AF), an irregular rapid beating of the atriums (10), increases the risk of strokes and other thromboembolic events. It is well established that patients with atrial fibrillation are at higher risk for strokes and systemic thromboembolic events com- pared to patients without AF (11, 12). The identification of This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/tem/index K. Ahmadzadeh et al. 2 high-risk patients for stroke is commonly made by using a clinical score such as the CHADS2, a scoring system designed to predict the risk of stroke in patients with atrial fibrillation based on their risk factors (heart failure, hypertension, age > 75, diabetes, prior stroke/ transient ischemic attack (TIA), vascular disease, ages 65-75, and sex category) (13, 14). In ad- dition to the availability of clinical scoring tools, there is still a need for additional simple, accurate, and reliable markers for the identification of patients with a higher risk of stroke in the clinical setting. In recent years, serum biomarkers such as natriuretic pep- tides have been mentioned as a suitable marker for predict- ing stroke and systemic embolic events (SEE) in patients with AF (15, 16). Studies have shown that NT-proBNP levels are of- ten elevated in patients with AF, which can be used as a tool for predicting the risks of adverse outcomes such as strokes (11). There are various isoforms and metabolites of natri- uretic peptides, including natriuretic brain peptide, atrial na- triuretic peptide, and N-terminal pro-brain natriuretic pep- tide (NT-proBNP). NT-proBNP is an inactive prohormone se- creted primarily by cardiac cells (17). An increase in this pro- hormone has been reported in patients with AF, and elevated levels of this protein are associated with an increased risk of stroke and mortality (16, 18, 19). Despite various existing studies, a conclusion has not yet been reached about the possible use of serum NT-proBNP levels for predicting the occurrence of stroke/SEE in patients with AF. This study aims to review the evidence for evaluating the value of N-terminal pro B-type natriuretic peptide in pre- dicting stroke/SEE in patients with atrial fibrillation through a systematic review and meta-analysis study. 2. Methods 2.1. Study design and settings The present study is a systematic review and meta-analysis designed to evaluate the predictive value of serum NT- proBNP levels for the occurrence of stroke/SEE in patient with AF. In this study, PICO was defined as: Patients (P): Pa- tients with atrial fibrillation, Index test (I): serum NT-proBNP levels, Comparison (C): AF patients who did not develop stroke/SEE Outcome (O): occurrence of stroke/SEE. 2.2. Search strategy The keywords for our search were extracted with the help of experts in the field and by reviewing the titles of relevant pa- pers. Synonyms and equivalent words were identified us- ing MeSH and Emtree databases. The search strategy was modified based on the standard Boolean operators and stan- dard tags of the Medline, Embase, Scopus, and Web of Sci- ence databases. Extensive searches were performed on each of the mentioned databases for papers published until De- cember 7th, 2021 (Online Resources). Additionally, Google and Google Scholar search engines were used to search for gray literature and non-indexed material. Finally, the cita- tions and bibliography of the eligible articles were reviewed to find any papers that might have been missed. 2.3. Selection criteria All observational or trial studies investigating the relation- ship between serum NT-proBNP levels and the occurrence of stroke/SEE in patients with AF were included. Exclusion criteria were: not having a non-stroke/non-SEE group, not having a stroke/SEE as an outcome, not reporting data for NT-proBNP, not having AF as the etiology of stroke/SEE, not reporting a hazard ratio (HR), not reporting original data, not having any valuable data, study protocol publications, edito- rials, letter to editors, duplicates, and reviews. 2.4. Data collection Two independent reviewers performed title and abstract screening of the potentially related articles, after which the retrieved full texts were studied, and the applicable studies were included. Finally, the information provided by the in- cluded studies was extracted into a checklist. Any dispute in this process was resolved using the opinion of a third re- viewer. The extracted data comprised study characteristics (name of the first author, year of publication, and country), study type, setting of included patients, sample size, age, gender distri- butions, the prevalence of underlying diseases, follow-up du- ration, and number of stroke/SEE patients. The predictive value of NT-proBNP in prediction of stroke/SEE was recorded as hazard ratio (HR) and 95% confidence interval (95% CI). 2.5. Risk of bias and certainty of evidence assess- ment The included articles were designed as cohort studies and randomized clinical trials (RCTs). All the included articles were considered prognostic studies and thus their quality was evaluated using the Quality in Prognosis Studies (QUIPS) guidelines (20). Grading of Recommendations, Assessment, Development and Evaluations (GRADE) framework was used to evaluate the level of evidence of the included articles (21). All the evaluations were done independently by two researchers and in cases of disagreement, the third reviewer was consulted. 2.6. Statistical analysis All analyses were performed using STATA 17.0 statistical pro- gram. The relationship between NT-proBNP serum values and occurrence of stroke/SEE was evaluated by reporting a pooled HR using the “meta” command. Heterogeneity be- tween the included studies was assessed by using I2 statis- This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/tem/index 3 Archives of Academic Emergency Medicine. 2023; 11(1): e8 tics, and while there was no heterogeneity, meta-regression was performed to analyze the data further. The effect of dif- ferent cut-off values, follow-up duration, prevalence of dia- betes, hypertension, heart failure and other underlying dis- eases of the population on the predictive value of NT-proBNP for stroke/SEE in patients with AF was investigated. A sensi- tivity analysis was also performed according to the quality of the included studies. Lastly, publication bias was examined using Egger’s test. 3. Results 3.1. Study characteristics The systematic search in four online databases of PubMed, Embase, Scopus and Web of Science resulted in 3,902 articles, from which 2,356 were non-duplicates. Title and abstract screening were performed on the remaining articles, 144 of which were potentially eligible. A total of 8 articles were in- cluded in this meta-analysis after the full-text screening. A manual search resulted in an additional article. Therefore, a total of 9 papers were included in our analysis (15, 22-29) (Figure 1). Four articles had a randomized clinical trial design, and five were designed as prospective cohort studies. The studies compromised data on 38,093 patients, 21,972 of which were male. During the average follow-up time of roughly three years, 3.10% of patients experienced a stroke/SEE. The gen- eral characteristics of the included papers are summarized in Table 1. 3.2. Risk of bias and publication bias The quality of the studies was assessed using the QUIPS as- sessment tool for the prognostic studies. Two studies were found to have moderate risk of bias in study participation domain due to small sample size, two studies were found to have moderate risk in study attrition domain due to no re- ports of samples lost to follow-up and in domain of outcome measurement two studies were found to have moderate risk and the risk of one study was classified as unclear, due to un- specific description of outcome assessment method. Risk of bias was low in all the remaining domains of the guideline (Table 2). As for the publication bias, Egger’s test was used, and the re- sults showed that there was no publication bias in the studies regarding the prognostic use of NT-proBNP in prediction of stroke/SEE in AF patients (p= 0.320) (Figure 2). 3.3. Meta-Analysis Relationship between NT-proBNP serum values and occur- rence of stroke/SEE in patients with AF Our analysis showed that serum levels of NT-proBNP can be a good predictor of stroke/SEE occurrence in patients with AF, even at different cut-off values (HR= 1.76; 95% CI: 1.51, 2.02; P < 0.001). A detailed presentation of the effect size for each study is shown in Figure 3. Also, no heterogeneity was reported between the included studies using the I2 statistic (I2 = 15.2%, P = 0.50) Meta-regression Although there was no heterogeneity between the included studies, meta-regression was done to investigate the pos- sible effects of confounders on the predictive value of NT- proBNP (Table 3). The analysis showed that from all the in- vestigated factors, only diabetes could have a possible effect on the predictive value of NT-proBNP (meta-regression co- efficient=0.042; P = 0.037). Other factors such as cut-off val- ues (P = 0.970), follow-up duration (P = 0.392), hypertension (P = 0.052), heart failure (P = 0.140), prior myocardial infarc- tion (MI) (P = 0.245), history of stroke or TIA or systemic em- bolism (P = 0.329), previous peripheral artery disease (P = 0.591), previous coronary artery disease (P = 0.638) and sam- ple size (P = 0.761) did not have any noticeable effect on the predictive value of NT-proBNP for occurrence of stroke/SEE in AF patients (Figure 4). Certainty of evidence Level of evidence were evaluated using the GRADE frame- work. Since the nature of assessments in the included papers were observations of prognostic value of NT-proBNP, despite the RCT design of some of them, the initial level of evidence was considered to be low. Evidence for stroke/SEE was rated to be moderate, one point increase due to the dose-response gradient effect was observed. However, evidence for stroke alone was rated as low, one point increases due to the dose- response gradient effect, and one point decrease due to the small sample size of the studies were observed. (Table 4). Summary of findings is demonstrated in table 5. Sensitivity analysis To assess the effects of the overall quality of the included studies on the reported effect size, we divided the studies into 2 groups of “low risk” and “unclear”. The analysis showed that the reported effect size for “low risk” articles (HR = 1.774; 95% CI: 1.50, 2.05; I2 = 19.41) did not have a significant difference with the effect size reported for “unclear” articles (HR = 1.731: 95% CI: 1.02, 2.44; I2 = 20.10) (meta-regression coefficient = -0.06; 95% CI: -0.77, 0.65; P = 0.864). 4. Discussion The current study aimed to review and summarize exist- ing literature regarding the possible predictive applicability of serum NT-proBNP levels for predicting the occurrence of stroke/SEE in patients with AF. Our analysis showed that ab- normally high serum levels of NT-proBNP, at any of the cut- off points and at any follow-up duration used by the included studies, had a significant correlation with stroke/SEE occur- This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/tem/index K. Ahmadzadeh et al. 4 rence in patients with AF. Although no heterogeneity was observed between the stud- ies, we conducted a meta-regression to evaluate our results further. The value of NT-proBNP in predicting stroke/SEE in patients with AF was shown to be affected by diabetes. Therefore, diabetic patients with AF and high levels of NT- proBNP are at higher risk of stroke/SEE compared to AF pa- tients without diabetes. It has been reported that higher lev- els of NT-proBNP is associated with higher rates of vascular complications among individuals who developed type 2 dia- betes (30). This is a possible explanation for the confound- ing effect of diabetes on the predictive value of serum NT- proBNP. Various cut-off points used by the studies did not affect the capability of serum NT-proBNP levels for prediction of stroke/SEE in patients with AF. NT-proBNP levels could be measured even with the lowest cut-off value and still be help- ful in stroke/SEE prediction. The lowest cut-off used in the included studies was 169 pg/ml. Appropriately, this cut-off can be used to evaluate the risk of stroke/SEE in patients with AF. The follow-up duration in the included studies varied be- tween 1 and 5 years. Our results show that serum levels of NT- proBNP is a good predictor of stroke/SEE in patients with AF at any follow-up duration. Therefore, a one-time measure- ment of serum NT-proBNP levels during the first admission to the hospital can be a good predictor of stroke/SEE even at longer follow-ups. Considering that neither cut-off values nor follow-up duration differences influenced the predicting capabilities of serum NT-proBNP, it can be implied that mea- surement of NT-proBNP can be a helpful risk assessment tool in clinical settings. To the best of our knowledge one meta-analysis has studied the association between natriuretic peptides (BNP and NT- proBNP) and atrial fibrillation (31). In their study, Hong et. al, have reported a hazard ratio of 2.53 (95% CI 2.00, 3.19) for the pooled value of BNP and NT-proBNP in prediction of stroke/SEE in AF patients; however, the predictive value of NT-proBNP alone has only been evaluated with four in- cluded studies with a reported risk ratio of 2.43 (95% CI 1.90, 3.11). Our study further solidifies the predictive value of NT- proBNP in evaluation of stroke/SEE in AF patients with a larger number of included articles and a larger sample size. Additionally, we have also considered the medical character- istics of the included populations and have provided a possi- ble cut-off for the use of NT-proBNP in the evaluation of the risk of stroke/SEE in patients with AF. 5. Limitation This systematic review and meta-analysis has its limitations. The primary goal of the study was to investigate the possi- ble predictive value of serum NT-proBNP for stroke as the sole outcome; but only three studies met this criterion, while other studies had pooled the data of stroke with SEE. As a result, this study should be used cautiously in the setting of stroke alone. Further studies could shed more light on the predictive applicability of NT-proBNP for stroke as the sole outcome. Another limitation of this study was the incom- plete report of some of the demographic information and the fact that the full text of one of the included articles could not be found, even after multiple email inquiries to their corre- sponding authors. As for the strength of this study, it is worth noting that no het- erogeneity and no publication bias were observed in the in- cluded studies. Different sample sizes of the included stud- ies did not affect the predictive value reported by the articles; this shows that none of the included articles had insufficient sample size. 6. Conclusion The results of our study show that there is a significant asso- ciation between serum levels of NT-proBNP and the risk of stroke/SEE in patients with AF. Measurement of serum NT- proBNP can be used as a clinical tool to assess the short term or long-term risk of stroke/SEE in patients with AF during first the admission to the hospital. Despite our findings, fur- ther studies are needed to evaluate the possible applicability of serum NT-proBNP measurement in settings with stroke as the sole outcome of the investigation. 7. Declarations 7.1. Acknowledgments Not applicable. 7.2. Conflict of interest The authors declare that there is no conflict of interest. 7.3. Fund This study has been funded and supported by the Iran Uni- versity of Medical Sciences (IUMS); [Grant No: 1400-1-32- 20043.] 7.4. Authors’ contribution Study design and conception: Yousefifard M, Azizi Y; Data gathering: All authors; Analysis: Yousefifard M; Drafting and revising: All authors. 7.5. Data availability Data can be shared at the request of any qualified investiga- tor for purposes of replicating procedures and results. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/tem/index 5 Archives of Academic Emergency Medicine. 2023; 11(1): e8 References 1. Johnson CO, Nguyen M, Roth GA, Nichols E, Alam T, Abate D, et al. Global, regional, and national bur- den of stroke, 1990–2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet Neurol. 2019;18(5):439-58. 2. Wendelboe AM, Raskob GE. Global Burden of Thrombo- sis. Circ Res. 2016;118(9):1340-7. 3. Smajlović D. Strokes in young adults: epidemiology and prevention. Vasc Health Risk Manag. 2015;11:157-64. 4. George MG. Risk Factors for Ischemic Stroke in Younger Adults: A Focused Update. Stroke. 2020;51(3):729-35. 5. Renna R, Pilato F, Profice P, Della Marca G, Broccolini A, Morosetti R, et al. Risk Factor and Etiology Analysis of Ischemic Stroke in Young Adult Patients. J Stroke Cere- brovasc Dis. 2014;23(3):e221-e7. 6. Lui SK, Nguyen MH. Elderly Stroke Rehabilitation: Over- coming the Complications and Its Associated Chal- lenges. Curr Gerontol Geriatr Res. 2018;2018:9853837. 7. Lloyd A, Bannigan K, Sugavanam T, Freeman J. Expe- riences of stroke survivors, their families and unpaid carers in goal setting within stroke rehabilitation. JBI Database System Rev Implement Rep. 2018;16(6):1418- 53. 8. Boehme AK, Esenwa C, Elkind MSV. Stroke Risk Factors, Genetics, and Prevention. Circ Res. 2017;120(3):472-95. 9. The Stroke Risk in Atrial Fibrillation Working Group. In- dependent predictors of stroke in patients with atrial fib- rillation: A systematic review. Neurology. 2007;69(6):546- 54. 10. Staerk L, Sherer JA, Ko D, Benjamin EJ, Helm RH. Atrial Fibrillation: Epidemiology, Pathophysiology, and Clini- cal Outcomes. Circ Res. 2017;120(9):1501-17. 11. Petersen P. Thromboembolic complications in atrial fib- rillation. Stroke. 1990;21(1):4-13. 12. Bassand J-P, Virdone S, Goldhaber SZ, Camm AJ, Fitz- maurice DA, Fox KAA, et al. Early Risks of Death, Stroke/Systemic Embolism, and Major Bleeding in Pa- tients With Newly Diagnosed Atrial Fibrillation. Circula- tion. 2019;139(6):787-98. 13. Sepehri Shamloo A, Bollmann A, Dagres N, Hindricks G, Arya A. Natriuretic peptides: biomarkers for atrial fib- rillation management. Clin Res Cardiol. 2020;109(8):957- 66. 14. Best JG, Bell R, Haque M, Chandratheva A, Werring DJ. Atrial fibrillation and stroke: a practical guide. Pract Neu- rol. 2019;19(3):208-24. 15. Roldán V, Vílchez JA, Manzano-Fernández S, Jover E, Gálvez J, Puche CM, et al. Usefulness of N-Terminal Pro–B-Type Natriuretic Peptide Levels for Stroke Risk Prediction in Anticoagulated Patients With Atrial Fibril- lation. Stroke. 2014;45(3):696-701. 16. Li L, Selvin E, Lutsey PL, Hoogeveen RC, O’Neal WT, Soli- man EZ, et al. Association of N-terminal pro B-type natri- uretic peptide (NT-proBNP) change with the risk of atrial fibrillation in the ARIC cohort. Am Heart J. 2018;204:119- 27. 17. Bay M, Kirk V, Parner J, Hassager C, Nielsen H, Krogs- gaard K, et al. NT-proBNP: a new diagnostic screen- ing tool to differentiate between patients with normal and reduced left ventricular systolic function. Heart. 2003;89(2):150-4. 18. Chi G, Januzzi JL, Korjian S, Daaboul Y, Goldhaber SZ, Hernandez AF, et al. N-terminal pro-B-type natriuretic peptide and the risk of stroke among patients hospital- ized with acute heart failure: an APEX trial substudy. J Thromb Thrombolysis. 2017;44(4):457-65. 19. Ebihara K, Yamagishi K, Umesawa M, Muraki I, Cui R, Imano H, et al. Moderate Levels of N-Terminal Pro-B- Type Natriuretic Peptide is Associated with Increased Risks of Total and Ischemic Strokes among Japanese: The Circulatory Risk in Communities Study. J Atheroscler Thromb. 2020;27(8):751-60. 20. Hayden JA, van der Windt DA, Cartwright JL, Côté P, Bom- bardier C. Assessing Bias in Studies of Prognostic Factors. Ann Intern Med. 2013;158(4):280-6. 21. Schünemann HJ, Higgins JPT, Vist GE, Glasziou P, Akl EA, Skoetz N, et al. Completing ‘Summary of findings’ ta- bles and grading the certainty of the evidence. Cochrane Handbook for Systematic Reviews of Interventions: Wi- ley; 2019. p. 375-402. 22. Berg DD, Ruff CT, Jarolim P, Giugliano RP, Nordio F, Lanz HJ, et al. Performance of the ABC Scores for Assessing the Risk of Stroke or Systemic Embolism and Bleeding in Patients With Atrial Fibrillation in ENGAGE AF-TIMI 48. Circulation. 2019;139(6):760-71. 23. Ha S-I, Choi D-H, Ki Y-J, Yang J-S, Park G, Chung J-W, et al. Stroke prediction using mean platelet volume in pa- tients with atrial fibrillation. Platelets. 2011;22(6):408-14. 24. Hamatani Y, Iguchi M, Ueno K, Aono Y, Esato M, Tsuji H, et al. Prognostic significance of natriuretic peptide levels in atrial fibrillation without heart failure. Heart. 2020;107(9):705-12. 25. Hijazi Z, Oldgren J, Andersson U, Connolly SJ, Ezekowitz MD, Hohnloser SH, et al. Cardiac Biomarkers Are Associated With an Increased Risk of Stroke and Death in Patients With Atrial Fibrillation. Circulation. 2012;125(13):1605-16. 26. Hijazi Z, Wallentin L, Siegbahn A, Andersson U, Chris- tersson C, Ezekowitz J, et al. N-Terminal Pro–B-Type Na- triuretic Peptide for Risk Assessment in Patients With Atrial Fibrillation. J Am Coll Cardiol. 2013;61(22):2274-84. 27. Kuronuma K, Okumura Y, Morikawa T, Yokoyama K, This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/tem/index K. Ahmadzadeh et al. 6 Matsumoto N, Tachibana E, et al. Prognostic Value of Serum N-Terminal Pro-Brain Natriuretic Peptide Level over Heart Failure for Stroke Events and Deaths in Pa- tients with Atrial Fibrillation. Int Heart J. 2020;61(3):492- 502. 28. Singleton MJ, Yuan Y, Dawood FZ, Howard G, Judd SE, Zakai NA, et al. Multiple Blood Biomarkers and Stroke Risk in Atrial Fibrillation: The REGARDS Study. J Am Heart Assoc. 2021;10(15):e020157. 29. Tomasdottir M, Hijazi Z, Lindback J, Benz A, Connolly SJ, Eikelboom JW, et al. Abstract 10162: NT-proBNP is Associated with Stroke, Heart Failure Hospitalizations, and Death in Patients with Atrial Fibrillation Without Oral Anticoagulation Regardless of Heart Rhythm. Circu- lation. 2021;144(Suppl_1):A10162. 30. Birukov A, Eichelmann F, Kuxhaus O, Polemiti E, Fritsche A, Wirth J, et al. Opposing Associations of NT-proBNP With Risks of Diabetes and Diabetes-Related Complica- tions. Diabetes Care. 2020;43(12):2930-7. 31. Du H, Yang L, Zhang H, Zhang X, Shao H. Association of natriuretic peptide and adverse outcomes in patients with atrial fibrillation: A meta-analysis. Clin Exp Pharma- col Physiol. 2020;48(2):161-9. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/tem/index 7 Archives of Academic Emergency Medicine. 2023; 11(1): e8 Table 1: Characteristics of included studies Study, Year, Reg- istry, Country Design Participants No. cases Age* Male Medical History Follow- up (year) Outcome N event HF HTN DM Stroke/ TIA/SE MI CAD PAD Berg, 2018, EN- GAGE AF-TIMI 48, Multinational RCT AF and CHADS2 score ≥2. 8705 72 (64-78) 5330 59.4 95.5 37 28 NR 34.8 4.4 2.8 Stroke/ SEE 139 Ha, 2011, NA, Korea PCS AF patients 200 68.9 ± 11.7 38 NR 49 18.5 15 NR NR NR 1.26 Stroke 14 Hamatani, 2021, FUSHIMI, Japan PCS AF patients w/o HF 1159 72.1 ± 10.2 718 NR 69.7 20.2 14.3 NR 13.2 8.2 5 Stroke/ SEE 113 Hijazi, 2012, RELY, Multina- tional RCT AF patients + at least one of the following risk factors: previous stroke or TIA, CHF or reduced LVEF <40%, age > 75, age > 65 with DM or HTN or CAD 6189 72 (67-77) 3944 30 78.4 21.4 19.6 17.4 24.9 NR 2.2 Stroke/ SEE 183 Hijazi, 2013, ARISTOTLE, Multinational RCT AF and at least 1 CHADS2 risk factor 14892 69 (62-75) 9590 35.9 87.5 30.9 18.7 12.8 NR 4.9 1.9 Stroke/ SEE 393 Kuronuma, 2020, SAKURA, Japan PCS Non-valvular AF, >20 year, receiving oral anticoagulant 2417 72 (66-79) 1777 25.9 72.3 23.5 11.4 NR NR NR 3.25 Stroke/ SEE 107 Roldan, 2014, NA, Spain PCS Permanent or paroxysmal AF patients, taking OAC with stable INR for > 6 months 1172 76 (71-81) 575 30 82 26 19 NR 19 NR 2.75 Stroke/ SEE 51 Singleton, 2019, REGARDS, USA PCS AF patients 175 66 ± 9 NR NR NR NR NR NR NR NR 5.2 Stroke 81 Tomasdottir, 2021, AVEROES & ACTIVE A, Multinational RCT AF patients, receiving aspirin with no oral anticoagulation 3184 NR NR NR NR NR NR NR NR NR 2.5 Stroke NR * Age was reported as mean ± SD or median (interquartile range). Data of medical history were reported as a percentage of the whole population. RCT: Randomized Clinical Trial, PCS: Prospective Cohort Study, HF: Heart Failure, HTN: Hypertension, DM: Diabetes Mellitus, TIA: Transient Ischemic Attack, SEE: Systemic Embolic Event, MI: Myocardial Infarction, CAD: Coronary Artery Disease, PAD: Peripheral Artery Disease; NR: not reported; AF: atrial fibrillation; CHADS: Congestive heart failure, Hypertension, Age, Diabetes, prior Stroke; CHF: congestive heart failure; LVEF: left ventricular ejection fraction; OAC: oral anticoagulant; INR: international normalized ratio. Table 2: Risk of bias assessment Study Study Participation Study Attrition Prognostic Factor Measurement Outcome Measurement Study Confounding Statistical Analysis/ Reporting Berg, 2018 Low Low Low Low Low Low Ha, 2011 Moderate Low Low Moderate Low Low Hamatani, 2021 Low Moderate Low Moderate Low Low Hijazi, 2012 Low Low Low Low Low Low Hijazi, 2013 Low Low Low Low Low Low Kuronama, 2020 Low Moderate Low Unclear Low Low Roldan, 2014 Low Low Low Low Low Low Singleton, 2019 Moderate Low Low Low Low Low Tomasdottir, 2021 Low Low Low Low Low Low This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/tem/index K. Ahmadzadeh et al. 8 Figure 1: PRISMA flow diagram of the present study. SEE: Systemic Embolic Event; AF: atrial fibrillation. Figure 2: Assessment of publication bias. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/tem/index 9 Archives of Academic Emergency Medicine. 2023; 11(1): e8 Figure 3: Effect size of the predictive effect of NT-proBNP in different cut-off values for the occurrence of stroke/SEE in AF patients. Table 3: Meta-regression analysis for the effect of confounding factors on the predictive value of NT-proBNP for the occurrence of stroke/Systemic embolic events (SEE) in patients with atrial fibrillation Variable Meta-regression coef. 95% CI P value Cut-off point 0.00001 -0.00055, 0.00057 0.970 Follow-up duration 0.11156 -0.14374, 0.36687 0.392 Diabetes 0.04203 0.00253, 0.08153 0.037 Hypertension 0.02357 -0.00023, 0.04738 0.052 Heart Failure 0.01932 -0.00635, 0.04500 0.140 Prior MI -0.08606 -0.23119, 0.05907 0.245 Stroke/TIA/SEE 0.02979 -0.03008, 0.08967 0.329 Peripheral Artery Disease -0.07108 -0.33050, 0.18833 0.591 Coronary Artery Disease 0.01336 -0.04233, 0.06906 0.638 Sample Size 0.00001 -0.00004, 0.00005 0.761 CI: confidence interval; coef.: coefficient; TIA: Transient Ischemic Attack; MI: Myocardial Infarction. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/tem/index K. Ahmadzadeh et al. 10 Figure 4: Meta-regression for assessment of the effect of cut-off values (A) and percentage of diabetic patients (B) in included studies on the occurrence of stroke/SEE following atrial fibrillation. Table 4: Certainty of Evidence Outcome Number of studies (analysis) N N event Quality of Evidence Stroke/SEE 6 (14) 34534 986 Moderate ⊕⊕⊕⊕⊕ + dose-response gradient Stroke 3 (5) 375 (+3184) 95 (+NR) Low ⊕⊕⊕⊕⊕ª +dose-response gradient - small sample size SEE: systemic embolic events; NR: Not Reported; N: sample size. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/tem/index Introduction Methods Results Discussion Limitation Conclusion Declarations References