Gender Differences in Prognosis and Risk Stratification of Brugada Syndrome: A Pooled Analysis of 4,140 Patients From 24 Clinical Trials

Background: Male gender has been consistently shown to be a risk factor for a greater number of arrhythmic events in patients with Brugada Syndrome (BrS). However, there have been no large-scale comprehensive pooled analyses to statistically and systematically verify this association. Therefore, we conducted a pooled analysis on gender differences in prognosis and risk stratification of BrS with a largest sample capacity at present. Methods: We searched PubMed, Embase, Medline, Cochrane Library databases, Chinese National Knowledge Infrastructure, and Wanfang Data for relevant studies published from 2002 to 2017. The prognosis and risk stratification of BrS and risk factors were then investigated and evaluated according to gender. Results: Twenty-four eligible studies involving 4,140 patients were included in the analysis. Male patients (78.1%) had a higher risk of arrhythmic events than female patients (95% confidence interval: 1.46–2.91, P < 0.0001). Among the male population, there were statistical differences between symptomatic patients and asymptomatic patients (95% CI: 2.63–7.86, P < 0.00001), but in the female population, no statistical differences were found. In the female subgroup, electrophysiological study (EPS) positive patients had a tendency toward a higher risk of arrhythmic events than EPS-negative patients (95% CI: 0.93–29.77, P = 0.06). Conclusions: Male patients are at a higher risk of arrhythmic events than female patients. Within the male population, symptomatic patients have a significantly higher risk profile compared to asymptomatic patients, but no such differences are evident within the female population. Consequently, in the female population, the risk of asymptomatic patterns cannot be underestimated.


INTRODUCTION
Brugada syndrome (BrS) is an inherited arrhythmic disorder generally characterized by a distinct electrocardiogram (ECG) pattern: the presence of ST-segment elevation in the right precordial leads (V1-V3), which may carry an increased risk of sudden cardiac death (SCD) due to malignant ventricular arrhythmias (Bayés et al., 2012). That typical "syndrome" was firstly presented by Nava et al. in 1988 at the National Congress of Italian Cardiologists, which subsequently named by Brugada brothers in 1992Nava et al., 1988;Brugada and Brugada, 1992). In current common consensus, BrS was described as a functional abnormality of repolarization, but theory proposed by Nava et al. believed that the true syndrome is not only a primary electrical disease performed particular ECG but a conduction disturbance at the right ventricular outflow tract (RVOT) related to clinical events (Martini and Nava, 2004;Marras et al., 2009). Recent focal therapeutic radiofrequency ablation (RFA) strategy indirectly proved the theory (Brugada et al., 2015). According to the expert consensus in 2013, patients with Brugada type 1 ECG induced by Class I antiarrhythmic drugs are included (Priori et al., 2013). Type 1 ECG induced by drug may occur false positive Brugada (Konigstein et al., 2016;Mizusawa et al., 2016).
Male sex has consistently been shown to be associated with a higher risk of arrhythmic events (Benito et al., 2008;Priori et al., 2013). However, the lack of large-scale samples and systematic comprehensive analysis have contributed to weak conformance and statistical power. In addition, there have been no comprehensive pooled analyses examining prognosis and risk stratification for BrS. Several clinical variables are considered to be potentially associated with worse outcome in patients with BrS. Electrophysiological study (EPS) might be the most controversial factor, and there remains no consensus on whether its inducibility is valuable in predicting outcome (Priori et al., 2002(Priori et al., , 2012Eckardt et al., 2005). Large registries have consistently shown that patients with spontaneous type 1 ECG have a high risk of cardiac arrhythmic events at follow-up (Brugada et al., 2002(Brugada et al., , 2004(Brugada et al., , 2005. The presence of symptoms is a significant predictor of arrhythmias (Priori et al., 2002). SCN5A mutation and recent positive family history of SCD have debatable feasibility as risk markers (Kanda et al., 2002). Lack of examination for documented auricular fibrillation (AF) status might lead to new agitation. However, the gender differences between these variables and whether men and women experience disparate outcomes remain indeterminate. Variables differing between the sexes, and how these manifests in certain sex groups, remain to be elucidated.
Abbreviations: AF, auricular fibrillation; BrS, Brugada Syndrome; ECG, electrocardiogram; EPS, electrophysiological study; ER, early repolarization; f-QRS, fragmented QRS; ICD, implantable cardioverter-defibrillator; RFA, radiofrequency ablation; RVOT, right ventricular outflow tract; SCD, sudden cardiac death; VF, ventricular fibrillation; VT, ventricular tachycardia. Therefore, based on a largest sample capacity of 4,140 patients from 24 clinical trials at present, we conducted a comprehensive pooled analysis of gender differences, including the following aspects: risk of arrhythmic events, EPS status, family history of SCD, spontaneous type 1 ECG pattern, SCN5A mutation, diagnosis status, and documented AF status.

Search Strategy
A comprehensive literature search of relevant studies published in PubMed, Embase, Medline, Cochrane Library databases, Chinese National Knowledge Infrastructure, and Wanfang Data was performed by two reviewers independently and systematically. We searched relevant published studies from 2002 to 2017 using the keywords: "Brugada" and "syndrome" or "Brugada syndrome" and "risk stratification." The titles, abstracts, and reference lists of all articles were carefully reviewed for potential and additional publications regarding this topic. Full text assessment of potential relevant studies was conducted for compliance with the inclusion criteria and to prevent duplication of data by the same group of authors (Figure 1).

Inclusion Criteria
All studies had to meet following criteria for inclusion: (a) full-text English language studies published in peerreviewed journals; (b) prospective or retrospective observational study design; (c) follow-up duration sufficiently long to detect arrhythmic events; (d) information included regarding clearly defined endpoint events (appropriate shocks, ventricular fibrillation/ventricular tachycardia, and SCD); (e) risk ratio (RR), hazard ratio (HR), odds ratio (OR), corresponding 95% confidence intervals (CIs), or necessary raw data were reported.
Upon sending e-mails to the principal authors of identified studies to request data sharing with a standardized form and definitions, we received original data for two of the studies (Sacher et al., 2013;Tokioka et al., 2014). Some of the data could not be found in the articles because the original data might be different from that published, owing to additional patients and longer follow-up times.

Quality Assessment
The Methodological Index for Non-Randomized Studies (MINORS) (Slim et al., 2003) was used to assess the quality of all included studies. The maximum value with this index is 24 points, with each item scored from 0 to 2 on the following aspects: (a) a clearly stated aim; (b) inclusion of consecutive patients; (c) prospective collection of data; (d) endpoints appropriate to the aim of the study; (e) unbiased assessment of the study endpoint; (f) follow-up period appropriate to the aim of the study; (g) loss to follow up <5%; and (h) prospective calculation of the study size. Both reviewers independently scored the included publications, then used the average MINORS score for final assessment. Based on MINORS scores of <16 and ≥16 points, studies were defined to be low-quality and high-quality studies, respectively.

Study Selection
The systematic review of the literature yielded a total of 5,648 potentially relevant studies with our search criteria.
After screening of the titles and abstracts, 2,534 studies were excluded, leaving 50 for full-text assessment. Twentysix duplicate studies were excluded, while 21 did not provide clear data pertaining to sex-related differences. Two studies did not clearly define the endpoint, while three had the same author with data included. Eventually, 24 of the original qualifying studies from the databases were included. Seven of the 24 studies were separated for independent depth analysis (Figure 1).

Male and Female
Overall, among 4,140 patients with BrS, 3,222 male patients (event rate 12.4%) and 918 female patients (event rate 4.4%) were included, because BrS is a male predominance syndrome (Priori et al., 2013). All 24 studies were included in this pooled gender analysis. An increased risk of arrhythmic events was observed in the male population compared to the female population (OR 2.06, 95% CI: 1.46-2.91, P < 0.0001; heterogeneity: P = 0.70, I 2 = 0%, Figure 2). The calculations showed a statistically significant difference between the two groups. Males had a higher risk of arrhythmia compared to females. At the same time, we conducted sensitivity analysis, excluding any set of data that would have no effect on the results.
In the male subgroup, also, there were no significant differences based on documented AF status (OR 1.67, 95% CI: 0.92-3.04, P = 0.09; heterogeneity: P = 0.12, I 2 = 53%, Figure 9C). In the female subgroup, the result was the same (OR 1.50, 95% CI: 0.15-14.99, P = 0.37, Figure 9D). Heterogeneity was not applicable for some outcomes because only one study provided suitable data for documented AF status.

DISCUSSION
We drew the following conclusions from the pooled analysis: (i) male patients display a higher risk of arrhythmic events than female patients; (ii) in the male population, symptomatic patients display a higher risk profile of arrhythmic events compared to asymptomatic patients, but there are no significant differences within the female population. Consequently, in the female population, the risk of asymptomatic patterns cannot be underestimated. According to our systematically comprehensive analysis of 24 trials, male patients display a higher risk profile compared with female patients. Although this conclusion has been consistently recognized in the HRS/EHRA/APHRS expert consensus statement (Priori et al., 2013), our study is the largest at present, including 4,140 patients, to analyze gender differences in prognosis and risk stratification for BrS.
Similar outcomes were found in other studies (Gehi et al., 2006;Benito et al., 2008). New studies have confirmed those acknowledged results, and outlined a complex relationship between sex distribution and patient ethnicity and age (Milman et al., 2018).  Many studies have shown that syncope was an independent predictor of risk, and provided sufficient evidence (Brugada et al., 2004;Priori et al., 2012;Calvo et al., 2016). The presence of symptoms in patients was significantly associated with arrhythmic events (23 vs. 3.8%, P < 0.00001) in our analysis. These results might explain the conclusion that in the male subgroup, symptomatic patients displayed a higher risk of arrhythmic events than asymptomatic patients. Surprisingly, in the female population, there were no significant differences between symptomatic patients and asymptomatic patients. We can infer that symptomatic status might only be a risk factor for men, and that asymptomatic women may be in a potentially dangerous situation. The risk of asymptomatic patterns cannot be underestimated. Although these results may be due to the lower incidence (11%) of women with BrS, the findings offer new insights for further research to combine with the new syncope episodes (Olde Nordkamp et al., 2015).
In our results, EPS-positive patients had a tendency toward a higher risk of arrhythmic events than EPS-negative patients only in the female subgroup(p = 0.06), which presented a potential risk factor to women. We can infer that the result may turn positive when the sample size is enlarged. Whether EPS inducibility is a predictor of arrhythmic events in BrS patients with previous syncope/sudden death or an independent character remains in dispute (Brugada et al., 2002(Brugada et al., , 2004Priori et al., 2002;Giustetto et al., 2009). In the 2017 AHA/ACC/HRS guideline for ventricular arrhythmias and SCD, an EPS with programmed ventricular stimulation using single or double extrastimuli may be considered for further risk stratification in asymptomatic and spontaneous type 1patients (Kusumoto et al., 2017). Newly studies suggested that extent of substrate is the only independent predictor of inducibility of VT or VF and may contribute to a new marker for risk stratification and therapy (Pappone et al., 2018). The differences of sexrelated cardiac electrophysiological characteristics may be the main reason contributing to the result, that women have lower expression of KChIP2 which is the main accessory subunit of transient outward current in right ventricular epicardium (Tadros et al., 2014). Besides women have greater sinoatrial node automaticity and enhanced atrioventricular node function than men (Burke et al., 1996;Shaowen Liu and Ole Kongstad, 2001).
Spontaneous type 1 ECG was regarded as a risk factor for arrhythmic events in most studies (Brugada and Brugada, 1992;Brugada et al., 1998Brugada et al., , 2002Brugada et al., , 2004Brugada et al., , 2005Priori et al., 2002Priori et al., , 2012Benito et al., 2008). Many reporters overserved that men with cardiac events had greater rates of spontaneous type 1 ECG, and among male patients with spontaneous type 1 ECG, cardiac events were more frequent (Benito et al., 2008;Sacher et al., 2013;Shi et al., 2018). Recent studies have indicated that females have less type 1 BrS ECG and lower inducibility rates than males (Milman et al., 2018). However, interestingly no statistically significant sex-related differences were found in our result. In a report by the European Society of Cardiology, family history of SCD is regarded as one of three factors for the events (Priori et al., 2001), but in the family history of the SCD group, we obtained absolutely negative results in all four subgroups. We also observed negative results in the SCN5A group, which was consistent with the HRS/EHRA/APHRS expert consensus statement (Priori et al., 2013).
The limitations of the study should be acknowledged. Although we included 4,140 patients from 24 studies incorporating the original data from two articles, there remain limitations in subgroup analysis to a certain extent. The number of women with BrS is relatively small, especially in some small samples. This situation limits the statistical power.

AUTHOR CONTRIBUTIONS
YG and YX defined the research theme. Dr. Frédéric Sacher and KK contributed the original data. MY and XL wrote the manuscript. XY, YY, and NL designed the methods. CT and XW analyzed the data. HB-M, DH, and HS interpreted the results. All authors discussed the results and commented on the manuscript.