Sinus node dysfunction and stroke risk: a systematic review and meta-analysis

Abstract

Objectives

The role of cardiac arrhythmia in ischaemic stroke is widely studied, but the size of the stroke risk in patients with sinus node dysfunction (SND) with and without atrial fibrillation (AF) is unclear. This systematic review and meta-analysis aimed to compare the risk of stroke and its associated factors in patients with SND with and without AF.

Design

A systematic review and meta-analysis was conducted based on the Grading of Recommendations, Assessment, Development and Evaluation approach.

Data sources

PubMed, EMBASE and Cochrane Database were searched until December 2022.

Eligibility criteria for selecting studies

Studies that investigate stroke in patients with SND diagnosed with or without AF/atrial flutter.

Data extraction and synthesis

Two independent authors screened studies for inclusion and extracted data. Literature quality assessment was performed using the Newcastle-Ottawa Scale and the Cochrane Collaboration Tool. The overall risk of stroke was estimated using the random-effects model. The generic inverse variance method was used to calculate the pooled estimates of stroke-associated factors. We performed a sensitivity analysis using a fixed-effects model.

Results

Of the 929 records retrieved, 6 papers (106 163 patients) met the inclusion criteria. The average yearly stroke incidence in patients with SND was 1.542% (95% CI: 1.334% to 1.749%). The stroke incidence was similar between the isolated SND (1.587%; 95% CI: 1.510% to 1.664%) and non-isolated (SND+AF) (1.660%; 95% CI: 0.705% to 2.615%) groups. AF (HR, 95% CI: 1.53 (1.01 to 2.33)), stroke/transient ischaemia attack/other thrombotic events (HR, 95% CI: 2.54 (1.14 to 5.69)), hypertension (HR, 95% CI: 1.51 (1.11 to 2.07)) and heart failure (HR, 95% CI: 1.41 (1.01 to 1.97)) were associated with stroke in the SND population.

Conclusion

Our findings suggest that patients with SND carry a similar risk of stroke to those with combined SND and AF. Future studies are needed to investigate whether interventions targeting stroke prevention, such as anticoagulation therapy, can help to prevent stroke in patients with SND.

PROSPERO registration number

CRD42023408436.

Strengths and limitations of this study.

• The literature used in this study was thoroughly searched and reliably sourced to control potential bias.

• We used the Grading of Recommendations, Assessment, Development and Evaluation approach to evaluate the strength and quality of the evidence.

• For consistency, the average yearly stroke incidence was regarded as a standard for comparison.

• Atrial tachyarrhythmias had varying exclusion criteria among the included studies, making stroke incidence comparison impossible.

Introduction

The pathological alterations in the sinus node or atrium cause irregular heart rhythm, referred to as ‘sinus node dysfunction (SND)’. Like the atrial fibrillation (AF) population, patients with SND are considered to be at high risk of atrial cardiomyopathy.^1–4 Moreover, SND and AF share similar risk factors and pathophysiological mechanisms.

Many studies label AF population at high risk of stroke.^{5 6} Remodelling of the atrial architecture in AF promotes the risk of thromboembolism, which can eventually lead to an ischaemic stroke.^7–9 Recently, three reviews reported on the similarities between SND and AF from molecular, gene and electrophysiological perspectives confirmed that atrial myocardium in patients with SND has substrates for electrical or structural remodelling that may lead to cardioembolic stroke.^{2 3 10} However, to what extent SND populations are prone to stroke risk remains unclear, and more evidence is crucial to the scientific community.

According to recent epidemiological studies, the burden of SND is dramatically increasing due to global ageing.^{1 11} Moreover, stroke prevention in patients with SND should receive extensive attention since stroke is the second leading cause of death worldwide.^{12 13} However, per current consensus or guidelines, stroke prevention is not recommended to manage SND.^{14 15} Therefore, we conducted this meta-analysis to measure the epidemiological burden of stroke in the SND population with and without AF to provide evidence for the importance of incorporating stroke prevention protocols into SND management in the future.

Methods

We conducted a systematic review and meta-analysis using the Meta-Analyses and Systematic Reviews of Observational Studies¹⁶ and the Cochrane Handbook for Systematic Reviews of Interventions (V.6.3, updated in 2022).¹⁷ Our results are reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses,¹⁸ which can be found in (online supplemental tables S1 and S2). The principal investigator reviewed and approved the retrieved data. The study protocol has been registered in the PROSPERO database (URL: https://www.crd.york.ac.uk/prospero/).

Search strategy and data source

An electronic search of PubMed, EMBASE and Cochrane databases was conducted until 31 December 2022. The full search terms used in this study are presented in the (online supplemental table S3). Briefly, we used search terms and related items, including keywords ‘sick sinus syndrome’ and ‘stroke’, ‘sinus node dysfunction’ and ‘stroke’ and ‘sinus node disease’ and ‘stroke’ as Medical Subject Headings thesaurus without language limitations. Two investigators (HD and HC) independently used separate search items to avoid search errors. First, we screened the abstracts of the retrieved publications to identify eligible studies. Next, the full texts of these studies were assessed for compliance with the inclusion criteria. Finally, reference lists of full-text studies were also screened to identify additional relevant articles. A third experienced professor (XY) will re-evaluate the controversial studies if the first two investigators disagree on the search results.

Inclusion and exclusion criteria

The inclusion criteria were: (1) prospective and retrospective cohort studies or randomised controlled trials (RCTs) with stroke outcomes; (2) cohort studies or RCTs that involve adult patients (>18 years of age) who were diagnosed with SND only or with AF or atrial flutter (AFL); (3) studies that report risk factors for stroke (indicated clearly with the estimates of HR and 95% CI that were generated from multivariable analysis); and (4) studies that were published in English language. The exclusion criteria were: (1) animal studies or non-follow-up studies (cross-sectional studies, letters to the editor, editorials or reviews); (2) lack of information on the occurrences of SND or stroke; (3) small sample size for the SND population (n<100); and (4) duplicate data records.

Data extraction and quality assessment

Two independent authors (HC and THH) identified, reviewed and screened all potentially relevant studies based on their titles and abstracts. First, the authors performed a full-text review of the selected articles. They retrieved data on clinical characteristics, stroke incidence rate and pooled HR values and summarised it into two prespecified Microsoft Excel sheets.

Two investigators, HD, and HC, independently assessed observational study quality using the Newcastle-Ottawa Quality Assessment Scale based on the following approaches: (1) selection of the patients, (2) comparability of groups and (3) outcome evaluation for cohort studies. The Cochrane Risk of Bias tool was used by two investigators (HD and HC) to assess bias risks in RCTs independently. The assessment was carried out across seven domains of bias (random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting and other sources of bias). The risk of bias was assessed as either low (proper methods taken to reduce bias), high (improper methods creating bias) or unclear (insufficient information provided to determine the bias level). If there was any disagreement during the quality assessment of the reviewed studies, professor XY was involved.

SND diagnosis

In the past, different criteria have been used to diagnose SND. In this study, SND diagnosis was defined if the studies reported that they included patients with evidence of symptomatic bradycardia, sinus block, sinus arrest, persistent sinus bradycardia (heart rate <40 bpm), prolonged PR interval (the interval between the beginning of the P wave and the beginning of the QRS complex of an electrocardiogram) and normal QRS (the series of deflections in an electrocardiogram that represent electrical activity generated by ventricular depolarization prior to contraction of the ventricles) duration. A summary of the diagnostic criteria for SND across studies is listed in table 1. We also referred to the term non-isolated SND as SND combined with AF or AFL.

Table 1.

Baseline characteristics of the six studies included in this meta-analysis

Study (year)	Location	Study design	Mean age	Diagnostic criteria for sinus node dysfunction	Diagnostic criteria for stroke	Quality assessment*
Svendsen et al 20137	Denmark	Randomised controlled trial	72.9	See online supplemental table S6	See online supplemental table S6	Low risk
Mo et al 201725	China	Retrospective study design	74.1	See online supplemental table S6	See online supplemental table S6	7
Greenspon et al 200426	USA	Randomised controlled trial	74	See online supplemental table S6	See online supplemental table S6	Unclear
Bodin et al 2020	France	Retrospective study design	73.8	See online supplemental table S6	See online supplemental table S6	8
Brandt et al 201728	Denmark	Randomised controlled trial	72.5	See online supplemental table S6	See online supplemental table S6	Low risk
Sgarbossa et al 199329	USA	Retrospective study design	66	See online supplemental table S6	See online supplemental table S6	7

*The quality of the included cohort studies was scored 1–10 according to the Newcastle-Ottawa Quality Assessment Scale. The comprehensive risk of bias in the included trials was evaluated by Cochrane Collaboration’s tool for assessing the risk of bias (low risk, unclear, high risk).

Outcome

The primary outcome of interest was stroke. The included studies confirmed stroke based on the International Classification of Diseases (ICD), 9th or 10th edition coding, and/or brain imaging.

Statistical analysis

Data analysis was performed using Stata V.16.0 software. The average annual incidence rate was calculated using the following formula: (Average yearly incidence=(stroke screened/total study population)/follow-up time (year)×100%). The incidence of stroke was categorised according to cohort size (<1000 or >1000 patients), stroke subtypes, pacing methods and SND status (isolated or non-isolated SND). A meta-analysis from comparable studies was performed to assess the potential risk factors for stroke associated with SND. The generic inverse variance method was used to calculate the pooled estimates of stroke-associated factors.

A test for heterogeneity was performed to assess whether there was variation in the actual effects underlying the studies. The I² statistic was used to select the appropriate effect model for pooling estimates. The level of heterogeneity was classified as low, moderate and high when the percentage was 25%, 26–74% and 75%, respectively. The overall HR of stroke was estimated using the random-effects model. We used fixed-effects models for studies with low/moderate heterogeneity (I²<50%) to confirm results from random-effects models. Publication bias was evaluated using funnel plots and the Egger test. Statistical significance was defined as a two-tailed p value of 0.05 for pooled estimates and a two-tailed p value of 0.1 for the heterogeneity test.

Grading the evidence

The grading of the Recommendations Assessment, Development and Evaluation (GRADE) approach was used to assess the certainty of the evidence. Evidence was graded as high, moderate, low or very low quality. The RCTs were graded as high-quality evidence by default and downgraded based on prespecified criteria. Criteria to downgrade evidence included risk of bias (weight of studies show risk of bias assessed by the Cochrane Risk of Bias tool), inconsistency (substantial unexplained heterogeneity, I²>50%, p<0.10), indirectness (presence of factors that limited the generalisability of the results), imprecision (the 95% CI for effect estimates were wide or crossed prespecified minimally important differences for harm) and publication bias (significant evidence of small-study effects).

Patient and public involvement

None.

Results

Search result

A flow diagram of the data search and study selection process is detailed in figure 1. A total of 929 records were identified from the three databases. Of these, 143 replicated studies, 773 irrelevant studies and 1 study missing the original text were excluded. Again, six studies were excluded because of the inadequacy of follow-up time,¹⁹ duplicate data of the Mode Selection (MOST) trial^{20 21} and recruitment of small sample size (<100 population).^22–24 Finally, six studies were included in this meta-analysis.^{7 25–29}

Figure 1.

Flow diagram of identifying the studies in the meta-analysis. MOST, Mode Selection Trial.

Baseline characteristics

Patient characteristics and study-related data are presented in tables 1 and 2, including the first author, year of publication, location, study design, mean age, diagnosis criterion of SND and stroke, literature quality scores, the total number of patients with SND, number of patients with SND with stroke, mean follow-up time, average yearly stroke incidence and HRs (95% CIs) of the stroke risk factors. In addition, the quality score data for the observational studies is shown in the (online supplemental table S4). (Online supplemental figure 1 and 2) show the individual Cochrane Risk of Bias tool assessments for the included RCTs. No serious risk of bias was detected.

Table 2.

The patient characteristics of the six studies included in this meta-analysis

Study (year)	Total patients with SND (n)	Patients with SND with stroke (n)	Mean follow-up time (years)	Average yearly stroke incidence (%)	HR (95% CI)*
Svendsen et al 20137	1415	92	4.30	See online supplemental table S7	See online supplemental table S7
Mo et al 201725	481	46	3.33	See online supplemental table S7	See online supplemental table S7
Greenspon et al 200426	2010	90	2.76	See online supplemental table S7	See online supplemental table S7
Bodin et al 2020	100 366	NA	0.59	See online supplemental table S7	See online supplemental table S7
Brandt et al 201728	1384	121	8.90	See online supplemental table S7	See online supplemental table S7
Sgarbossa et al 199329	507	32	8.42	See online supplemental table S7	See online supplemental table S7

*All were analyzed in multivariate analysis.

Potential publication bias was assessed using funnel plots and Egger’s test (online supplemental figure S3). All the included studies had a low risk of bias, except the study by Brandt et al.²⁸ We observed heterogeneity in the study by Brandt et al,²⁸ which may be attributed to differences in diagnostic criteria. The stroke diagnosis in the study by Brandt et al was based on ICD-10, which is more rigorous than diagnosis by hospital record, clinician assessment and medical imaging evidence.

Average yearly stroke incidence in patients with SND

After a meta-analysis of the single-group rate combined with these six studies, we found that the average yearly stroke incidence was 1.542% (95% CI: 1.334% to 1.749%) with insignificant heterogeneity (I²=20.7%; p=0.278; figure 2) in patients with SND.

Figure 2.

Forest plot of average yearly stroke incidence in patients with sinus node dysfunction.

Larger sample sizes are the cornerstone of epidemiological research, and smaller sample sizes can lead to selection bias. In our meta-analysis, the sample sizes of the studies conducted by Mo et al²⁵ and Sgarbossa et al²⁹ were significantly smaller than those of other studies. Thus, we compared stroke rates in these two small cohorts with other studies. We observed no significant difference in the incidence of stroke based on the cohort size (figure 3A).

Figure 3.

Subgroup analysis for different (A) sample cohorts, (B) stroke subtypes, (C) pacing methods and (D) SND populations AF, atrial fibrillation; SND, sinus node dysfunction; AAIR, atrial pacing mode with rate response (atrial stimulation, atrial sensing, and pacing inhibition by atrial events outside of refractory period); VVIR, single chamber ventricular pacing mode with rate response enabled (ventricular stimulation, ventricular sensing, and stimulation inhibition by ventricular sensed event (outside of refractory periods)); DDDR, rate-modulated pacing mode with dual chamber stimulation, dual chamber sensing, and dual response to detection (inhibition and triggering).

To explore the epidemiological characteristics of different stroke subtypes in patients with SND, we extracted different types of strokes. We categorised them into ischaemic, haemorrhagic and other stroke subtypes. Our analysis demonstrated that ischaemic stroke accounts for a higher proportion than haemorrhagic and other stroke types. The average yearly incidence of ischaemic stroke is 7 times that of haemorrhagic stroke and 13 times that of other types of strokes (figure 3B). Among the included studies, Mo et al²⁵ reported a relatively high incidence of ischaemic stroke, leading to a high heterogeneity. The study by Mo et al was retrospective in design and enrolled only 481 patients, which may explain the difference in heterogeneity.

We analysed stroke incidence in patients with SND with different pacing modes: AAIR (atrial pacing mode with rate response (atrial stimulation, atrial sensing, and pacing inhibition by atrial events outside of refractory period), two studies), VVIR (single chamber ventricular pacing mode with rate response enabled (ventricular stimulation, ventricular sensing, and stimulation inhibition by ventricular sensed event (outside of refractory periods), one study) and DDDR (rate-modulated pacing mode with dual chamber stimulation, dual chamber sensing, and dual response to detection (inhibition and triggering), three studies). The combined stroke incidence rates and 95% CIs for AAIR, VVIR and DDDR were 1.239% (0.655% to 1.823%), 1.784% (0.962% to 2.606%) and 1.264% (0.818% to 1.710%), respectively (figure 3C). Statistically, there was no significant difference among AAIR, VVIR and DDDR in the incidence of stroke (p=0.511), suggesting different pacing modes in patients with SND have similar effect on the occurrence of stroke.

Usually, SND combines with atrial tachyarrhythmia, called tachycardia-bradycardia syndrome or bradycardia-tachycardia syndrome.³ Hence, isolated SND was defined as SND without evidence of atrial tachyarrhythmia (including AF/AFL). In the subgroup analysis, the average yearly stroke incidence in the isolated SND and the non-isolated (SND+AF) groups were 1.587% (1.510% to 1.664%) and 1.660% (0.705% to 2.615%), respectively. In the subgroup of isolated SND, heterogeneity decreased to 0 (I²=0%; p=0.674), while another group remained significant (I²=85.2%; p<0.001), indicating that heterogeneity was mainly derived from the subgroup of SND with AF studies (figure 3D). Only 167 patients with non-isolated SND were included in the survey by Mo et al,²⁵ leading to a considerable bias and high heterogeneity in the SND with the AF subgroup. Stroke incidence was still high in patients with isolated SND and was not statistically different from stroke incidence in patients with SND with AF. Note that despite the absence of AF (at least not detected by conventional methods), attention still needs to be paid to the risk of stroke in patients with SND. Anticoagulation should be appropriately initiated in patients with isolated SND at high risk of ischaemic stroke.

The risk factors of stroke

We extracted data from five studies significantly associated with stroke events, including previous history of stroke/transient ischaemia attack (TIA)/other thrombotic events, hypertension, heart failure and CHA₂DS₂-VASc score (age (1 point for 65–74 years and 2 points for ≥75 years), sex (1 point for female), congestive heart failure (1 point), hypertension (1 point), diabetes mellitus (1 point), vascular disease (1 point), and thrombotic event (2 points)).

AF history was significantly associated with stroke (the HR and 95% CI of AF history: 1.53 (1.01 to 2.33), (online supplemental figure S4A).

Our meta-analysis indicated that some components of the CHA₂DS₂-VASc score were still significantly associated with stroke risk in patients with SND (the HRs and 95% CIs of stroke/TIA/other embolisms, hypertension and heart failure are 2.54 (1.14 to 5.69), 1.51 (1.11 to 2.07) and 1.41 (1.01 to 1.97), respectively; (online supplemental figure S4B-D). In addition, higher values of the CHA₂DS₂-VASc score were significantly associated with the occurrence of stroke (HR and 95% CI per 1 score change, 1.32 (1.15 to 1.52); p<0.001; (online supplemental figure S4E).

Sensitivity analysis

We used random-effects models in all steps regardless of heterogeneity. However, when the I²<50%, the results of the fixed-effects models are more accurate and the random-effects models are more conservative. Thus, we applied fixed-effects models in analysis with I²<50% (online supplemental figure S5). The incidence of stroke in SND remained statistically significant (annual rate: 1.582%). The stroke incidence was similar among the three different pacing modes. Factors including the previous history of AF, hypertension, heart failure and increased CHA₂DS₂-VASc score were significantly associated with stroke in patients with SND. The HRs and 95% CIs for AF, hypertension, heart failure and CHA₂DS₂-VASc score were 1.457 (1.069 to 1.986), 1.514 (1.106 to 2.071), 1.407 (1.007 to 1.966) and 1.299 (1.180 to 1.430), respectively.

GRADE assessment

(Online supplemental table S5) summarises the GRADE assessments for the average yearly stroke incidence in patients with SND and the stroke risk factors. The levels of certainty for the annual stroke incidence in cohorts and RCTs are low and moderate, respectively. In cohort studies, the certainty of the evidence for previous AF history was graded as low. In RCT studies, it was graded as moderate. The quality of the evidence for the CHA₂DS₂-VASc score was graded moderate due to indirectness. Likewise, the certainty of the evidence for the previous history of hypertension and heart failure in RCTs was moderate.

Discussion

This systematic review and meta-analysis included six studies (106 163 patients with SND). Our results show a high stroke risk in patients with SND (annual rate: 1.54%), with ischaemic stroke being the most common subtype, occurring at a yearly rate of 1.40%. Furthermore, the incidence rates of stroke in patients with isolated and non-isolated SND were similar (1.59% and 1.66%, p=0.511). In addition, AF, previous stroke/TIA/other thrombotic events, hypertension, heart failure and elevated CHA₂DS₂-VASc score were factors that promote stroke development. These results emphasise the need for prompt stroke risk assessment in patients with SND.

As of 2019, stroke remained the second-leading cause of death worldwide, with ischaemic stroke accounting for the majority (62.4%–67%).^{12 13} Extensive epidemiological studies showed a substantial increase in the annual incidence rates of stroke around the world.12 13 30–34 Likewise, the rate of death and disability caused by stroke is sharply increasing.¹² One of the leading causes of ischaemic stroke is cardioembolic stroke. Preventing stroke, especially ischaemic stroke, may significantly reduce stroke-related death and disability. However, less focus has been given to SND compared with AF in the past.¹³ Previous studies observed a similar incidence of ischaemic stroke between AF (yearly rate: 1.25%–2.22%) and SND-related AF (yearly rate: 1.66%–2.06%) patients.^{8 9 27} Compared with general populations, patients with isolated SND have 3.12-fold to 12.23-fold higher risk for stroke (yearly rate: 1.59% vs 0.13%–0.51%, p<0.001 by χ² test).12 13 30–34 Therefore, ischaemic stroke risk should not be underestimated in patients with isolated SND. In current clinical practice, except for permanent pacemaker therapy,^{14 15} additional anticoagulation therapy may be needed when SND is combined with AF.⁶ We recommend conducting nested case–control studies or clinical trials to investigate further the benefits of prophylactic anticoagulation therapy in preventing ischaemic stroke in patients with SND.

This meta-analysis determined significant stroke risk factors in patients with SND, including previous stroke/TIA/other thrombotic events, hypertension and heart failure, similar to AF. However, no relationship was observed between implanted cardiac pacemakers (AAIR, VVIR or DDDR) and stroke occurrences in patients with SND.

Studies indicate that AF can increase the risk of stroke by five times.^{8 9} In addition, SND has been found to be associated with AF.^{2 3} The link between SND, AF and atrial cardiomyopathy is reflected in anatomic and electrophysiological remodelling of the atrium,² mainly promoted by ageing,^{1 2} atrial stretching,2–4 35–37 inflammation,2 3 38 39 oxidative stress injury1 2 40–42 and abnormal electrical activity of atrial myocytes.²

The stroke incidence rate of isolated and non-isolated SND was similar in this meta-analysis, suggesting that SND is also equally important to AF. The history of AF was associated with a stroke risk in this meta-analysis. Recent studies and guidelines have proposed that AF screening in the general and post-stroke populations can help reduce new stroke events through appropriate anticoagulant intervention in patients with AF.^{6 43–46} The STROKESTOP study is the first randomised population-based AF screening trial with long-term follow-up, providing the latest evidence on the benefits of AF screening in the general population.⁴⁶

Experts suggest that SND can be divided into two categories based on potential pathophysiological mechanisms. The first is primary or permanent SND, which is caused by intrinsic substrates that result in atrial myopathy. This type of SND can lead to isolated sinus bradycardia/arrest (isolated SND) or atrial arrhythmia due to the fibrous tissue separating the sinoatrial node from the atrial or pulmonary vein. The second category is secondary SND, which is reversible and results from the sinoatrial node inhibition by atrial tachyarrhythmias, such as paroxysmal AF with a long RR interval (The interval between two adjacent R waves on the electrocardiogram) after termination (tachycardia-bradycardia syndrome). However, conclusive evidence as to which condition developed first remains vague in scenarios where SND and atrial tachyarrhythmia coexist, such as in AF. A feasible method is to perform pulmonary vein isolation (PVI) and observe if SND persists. Further research is needed to determine if there are differences in stroke occurrence between patients with primary and secondary SND.

It is crucial to examine whether successful PVI leads to an increased risk of stroke in individuals with SND. Some experts propose that patients with primary SND may have a higher risk of stroke than those with secondary SND. Hence, even after successful PVI, patients with primary SND may still require anticoagulation therapy. This hypothesis may explain why the meta-analysis found similar stroke rates for isolated and non-isolated SNDs. It is possible that many cases of AF in the studies were caused by primary SND.

It is important to note that individuals diagnosed with isolated SND are at risk of stroke (1.542%) and may fall into the high-risk category for AF. Therefore, it is highly recommended to prioritise AF screening for these patients. Additionally, ischaemic strokes may occur before the onset of AF, as indicated by atrial arrhythmia monitoring.^{47 48} Hence, cardiologists may be required to assess the risk of ischaemic stroke in patients with SND before the onset of AF. Recent reviews have also highlighted the potential of various atrial cardiomyopathy-related features to predict stroke, which should be taken into consideration. Yet, no studies have explored the stroke-predictive power of these markers in patients with SND. Future research should focus on the predictive role of novel atrial cardiomyopathy features in patients with SND.

The 2020 European Society of Cardiology Guidelines for diagnosing and managing AF recommends using the CHA₂DS₂-VASc score to predict ischaemic stroke risk in patients with AF.⁶ Also, recent meta-analyses suggested that the CHA₂DS₂-VASc score can predict stroke events in patients with sinus rhythm.^{5 49} These meta-analyses further demonstrated the independent association of the CHA₂DS₂-VASc score with SND-related stroke, indicating its potential for stroke prediction.

Although atrial cardiomyopathy increases the risk of ischaemic stroke, aggressive use of anticoagulation can be dangerous for patients with SND due to adverse effects like fatal bleeding events. Besides, indiscriminate overuse of anticoagulants compromises the physical and economic benefits for patients with SND with low ischaemic stroke risk. Hence, it is essential to evaluate the ischaemic stroke risk in isolated SND and determine which patients might benefit from anticoagulation therapy.

In this meta-analysis, the CHA₂DS₂-VASc score and some of its components, including previous stroke/TIA/other thrombotic events, hypertension and heart failure, were associated with stroke. Likewise, AF was associated with stroke. Unfortunately, the CHA₂DS₂-VASc score exhibits discriminatory limitations in assessing atrial-related strokes because it does not consider atrial electrical or structural remodelling.⁵⁰ Combining atrial cardiomyopathy-related markers with the CHA₂DS₂-VASc score may improve the ability to predict stroke in patients with SND.

Future perspectives

While the data presented in this meta-analysis suggests a potential association between SND and stroke, further research is needed to confirm and clarify this relationship. In light of the current guidelines, which do not address stroke prevention strategies specifically for patients with SND, it is essential to gather solid evidence regarding the safety and effectiveness of anticoagulant drugs in this population. Therefore, real-world, long-term and large-cohort RCTs are necessary to assess the benefit of anticoagulation in patients with SND who are at high risk of ischaemic stroke. Additionally, investigating the role of atrial cardiomyopathy-related markers and gene mutations may provide valuable insights and improve stroke prediction in the SND population.

Further real-world evidence is needed to determine which factors most influence stroke in patients with SND, the extent of the impact on stroke, and whether modifiable risk factors still contribute to stroke events after correction. In addition, future research should investigate clinical and epidemiological evidence on cardiomyopathy-related markers because adding atrial cardiomyopathy markers to the already-known traditional risk factors for cardiovascular disease or algorithms may accurately stratify the risk of stroke. Moreover, further research should focus on whether left atrial appendage occlusion decreases the stroke risk in patients with SND.

What is more, gene mutations are associated with some SND, AF and stroke, but the extent to which gene mutations play a role in the progression of SND to AF or stroke is unclear. For example, do SNDs due to genetic mutations have different risks of AF or stroke? Will mutations of different types or sites lead to different outcomes? Future research is needed to answer the above questions.

Limitations

While we conducted a thorough systematic review and meta-analysis, it is important to note some limitations. One of the studies we analysed was retrospective, which could have led to recall bias. Our findings suggest that factors such as history of AF, stroke/TIA/other embolisms, hypertension, heart failure and CHA₂DS₂-VASc score are linked to stroke in the SND population. Additional research should include other risk factors not covered in the current systematic review and meta-analysis. We searched multiple engines to find all relevant studies; however, there may still present publication bias, like other meta-analyses. Thus, our results should be interpreted cautiously. Additionally, we could only obtain data on the average yearly incidence rate of stroke from the study by Bodin et al.²⁷

Conclusion

This meta-analysis and systematic review of observational studies found that patients with SND carry a similar risk of ischaemic stroke to those patients with combined SND and AF. This suggests that, like the patients with AF, patients with SND should be considered at high risk of stroke. In addition, our findings suggest that paying attention to previous history of AF, stroke/TIA/other embolisms, hypertension, heart failure and increased CHA₂DS₂-VASc score may be a valuable addition to stroke prevention strategies. Health practitioners have an important role to play in identifying high-risk patients. Future studies are needed to investigate whether interventions targeting stroke prevention, such as anticoagulation therapy, can help to prevent stroke in patients with SND.

Data availability statement

Data are available upon reasonable request. The data underlying this article will be shared on reasonable request to the corresponding author.

Ethics statements

Patient consent for publication

Not applicable.

Ethics approval

Not applicable.