Article Text

Original research
Atrial fibrillation following transcatheter atrial septal defect closure: a systematic review and meta-analysis
  1. Jonah Daniel Himelfarb1,
  2. Healey Shulman1,
  3. Christopher James Olesovsky1,
  4. Rawan K Rumman1,
  5. Laura Oliva2,
  6. Joshua Friedland1,
  7. Ashley Farrell3,
  8. Ella Huszti2,4,
  9. Eric Horlick5,
  10. Lusine Abrahamyan2,6
  1. 1 Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada
  2. 2 Institute of Health Policy, Management and Evaluation (IHPME), University of Toronto, Toronto, Ontario, Canada
  3. 3 Library and Information Services, University Health Network, Toronto, Ontario, Canada
  4. 4 Biostatistics Research Unit, University Health Network, Toronto, Ontario
  5. 5 Toronto Congenital Cardiac Centre for Adults, Peter Munk Cardiac Centre (PMCC), University Health Network, Toronto, Ontario, Canada
  6. 6 Toronto General Hospital Research Institute, University Health Network, Toronto, Ontario, Canada
  1. Correspondence to Dr Jonah Daniel Himelfarb, University of Toronto, Toronto, ON M5S 1A1, Canada; jonah.himelfarb{at}


Objective The ostium secundum atrial septal defect (ASD) is among the most common congenital cardiac anomalies diagnosed in adulthood. A known complication of transcatheter ASD closure is the development of new-onset atrial fibrillation and flutter (AFi/AFl). These arrhythmias confer an increased risk of postoperative stroke, thrombus formation and systemic emboli. This systematic review examines the burden of de novo AFi/AFl in adults following transcatheter closure and seeks to identify risk factors for AFi/AFl development.

Methods Studies were identified by a search of MEDLINE, EMBASE and Cochrane databases from inception until 29 April 2020. A meta-analysis of AFi/AFl incidence was performed using a random-effects model.

Results A total of 31 studies met inclusion criteria, comprising 4788 adult patients without a history of AFi/AFl. Twenty-three studies were included in quantitative synthesis and demonstrated an overall incidence rate of 1.82 patients per 100 person-years of follow-up (I2=83%). In studies that enrolled only patients ≥60 years old, the incidence was 5.21 patients per 100 person-years (I2=0%). Studies with follow-up duration ≤2 years reported an incidence of 4.05 per 100 person-years (I2=55%) compared with a rate of 1.19 per 100 person-years (I2=85%) for studies with follow-up duration >2 years.

Conclusions The incidence of new-onset AFi/AFl is relatively low following transcatheter closure of secundum ASDs. The rate of de novo AFi/AFl, however, was significantly higher in elderly patients. Shorter follow-up time was associated with a higher reported incidence of AFi/AFl.

  • atrial fibrillation
  • heart septal defects
  • atrial
  • atrial flutter
  • meta-analysis
  • systematic reviews as topic

Data availability statement

Data are available upon reasonable request.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Atrial septal defects (ASDs) have a prevalence of 1–2 per 1000 live births and are among the most common congenital cardiac anomalies diagnosed in adulthood.1 The ostium secundum defect accounts for 75% of ASDs.2 Individuals with isolated secundum ASDs have a 97% probability of survival into adulthood; however, these defects are unlikely to spontaneously close in adults and left untreated can lead to major complications later in life.1 These include right heart failure, pulmonary hypertension, paradoxical emboli and premature death.3 To mitigate these complications, current guidelines recommend closure of isolated secundum ASDs in select patients.4 Individuals with ASDs have chronic atrial and ventricular volume overload, leading to structural changes and electrical remodelling.5 Although volumes of both atria and the right ventricle decrease following closure in all patients with baseline left-to-right shunt, significant electrophysiological changes may persist.6 7 Atrial fibrillation and flutter (AFi/AFl) are particularly important after ASD closure because they confer a risk of stroke, systemic thromboembolism and reduced functional capacity in a population that is already at elevated risk for these complications.7 8 There is limited guidance to risk-stratify ASD patients for the development of postclosure arrhythmia or to inform monitoring and management decisions. This systematic review examines the incidence of new onset AFi/AFl following transcatheter closure and seeks to identify risk factors for their development.


This systematic review’s protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO) database on 21 April 2020 (registration number: CRD42020181257). The study reporting followed the Preferred Reporting Items in Systematic Reviews and Meta- Analyses (PRISMA) guideline (see online supplemental table I).9 As all analyses were based on previous studies, neither patient consent nor ethical committee approval was required.

Supplemental material

Study selection criteria

This review included studies providing information on new-onset AFi/AFl in adult patients after transcatheter ASD closure. Due to inconsistent study definitions of ‘new-onset’ AFi/AFl after closure, studies were only included if they reported both a baseline rate of AFi/AFl and the number of cases of new-onset AFi/AFl after transcatheter closure, so that an incidence rate for de novo AFi/AFl could be determined. Several studies do not specify if cases of AFi/AFl developed during the procedure or postprocedurally. To maintain consistency, all reported cases of new-onset AFi/AFl were included regardless of timing during or after ASD closure. Studies were excluded if they were designed for a special ASD population (eg, exclusively patients with deficient rims, large defects or a history of pulmonary hypertension), as these cohorts have additional risks for AFi/AFl compared with the general ASD population. Studies with mixed populations, such as both PFO closure and ASD closure, were included if the AFi/AFl incidence for the ASD closure group was reported separately. To avoid analysing the same data twice, studies from the same institution with enrolment from overlapping years were assessed and only the larger, more inclusive study was included. Case reports, cross-sectional studies, conference abstracts, reviews, meta-analyses, animal studies, paediatric studies, non-English language studies and those with less than 20 adults undergoing transcatheter closure for ASD were excluded.

Information sources

All eligible studies were identified through a systematic comprehensive search performed by an information specialist of the databases MEDLINE, Ovid MEDLINE Epub Ahead of Print and In-Process & Other Non-Indexed Citations, Cochrane Database of Systematic Reviews, and EMBASE all on the OvidSP platform. Databases were searched from their inception dates to 29 April 2020.

Search strategy

The search strategy was designed with an information specialist. The search included Medical Subject Headings terms and keywords pertaining to ASD and transcatheter closure (see online supplemental table II). Bibliographies of selected studies were reviewed to identify additional relevant articles. When relevant, authors of included studies were contacted to request clarifications or additional data not reported in their original article.

Selection of studies and data extraction

All stages of the review including abstract screening, full-text review, data extraction and quality assessment were completed in duplicate by independent reviewers (JDH, HS, CJO and JF). Abstracts identified as relevant by either reviewer were considered for full-text review. Disputes during full-text review, data extraction and quality assessment were resolved by consensus or, when needed, through consulting a third reviewer (LA). Information extracted included study characteristics (eg, study design, country), baseline participant characteristics (eg, age, sex, comorbidities), details about ASD closure (eg, defect size, device size, device type) and echocardiographic data. Data extraction prioritised information about baseline AFi/AFl rate, new-onset AFi/AFl, time to AFi/AFl onset postclosure and methods used to diagnose AFi/AFl. All studies underwent quality assessment using the Joanna Briggs Institute appraisal tool for prevalence and incidence studies.10

Data synthesis and statistical analysis

Study and patient characteristics were summarised using counts and proportions or minimum and maximum values. Only studies reporting mean follow-up duration were included in the meta-analysis. Meta-analysis using a random-effects model with a Freeman-Tukey arcsine transformation was performed. The Freeman-Tukey arcsine transformation is useful to normalise proportions so that sample variance can be better approximated.11–13 Inferences were made regarding the measure of heterogeneity (I 2) as well as the variances of effect sizes (τ2) across all studies. The pooled estimates were expressed as event count per 100 person-years of follow-up. We conducted a sensitivity analysis with studies published in 2015 and later to evaluate if there was an impact from changes in AFi/AFl monitoring and reporting following a 2014 AHA/ACC/HRS atrial fibrillation guideline.14 All statistical analyses were performed using R V.

Patient and public involvement

Patients and the public were not involved in our meta-analysis.


Study selection

Our search returned a total of 8293 records with an additional 11 identified from references of reviewed articles. After removal of duplicates, 6098 records remained for title and abstract screening (figure 1). During review of titles and abstracts, 5632 were excluded and 466 full-text articles underwent full review. From these, 31 studies met our inclusion criteria, comprising of 4788 adult patients without a history of AFi/AFl.2 3 5–8 16–40 The most common reason for study exclusion was lack of clarity in defining new-onset AFi/AFl as an outcome (eg, not reporting baseline AFi/AFl, not reporting new-onset AFi/AFl, not separating AFi/AFl from other atrial tachycardias). Eight studies were excluded from quantitative analysis because they did not report mean follow-up duration.

Figure 1

Flow chart of study selection.

Summary of study characteristics

Summary statistics and patient characteristics for the 31 studies included in the qualitative synthesis are shown in table 1. Full study characteristics including disclosure statements are detailed in online supplemental table III. The study characteristics were quite heterogeneous: sample sizes ranged from 23 to 1062 patients, average follow-up from 6 to 110 months, age from 15 to 90 years, and percentage of males ranged from 6.2% to 43%. Studies were conducted across a variety of countries, including Germany (n=4), South Korea (n=3) and the USA (n=3). Most studies had a retrospective cohort design (n=26) or prospective cohort design (n=4). AFi/AFl was most commonly assessed through clinical assessments (n=24) and ECG (n=20). Six of the studies had an inclusion criterion of age ≥60 or ≥65 years.6 7 16 20 22 24 Two of the studies had an inclusion criterion of age ≥40 years.29 40

Table 1

Descriptive summary statistics of study and patient characteristics

Critical appraisal of quality of studies

All 31 studies were deemed to have a low risk of bias. Nearly all studies (30/31) were found to be lacking in one or more of the Joanna Briggs Institute critical appraisal criteria; however, this was largely due to insufficient explanation regarding choice of sample size (figure 2). Other criteria absent in studies included valid methods of identifying AFi/AFl (which we defined as using scheduled ECG or Holter monitoring for diagnosis) and clarity of statistical analysis. Several studies did not report sufficient information about loss to follow-up or methods to address this. One study had a high rate of loss to follow-up (>20%). The impact of these omissions on bias were uncertain due to incomplete reporting by authors.

Figure 2

Risk of bias assessment for included studies, as assessed by the Joanna Briggs Institute critical appraisal checklist for prevalence and incidence studies.


A meta-analysis was performed for the 23 studies that reported mean follow-up duration, comprising 3796 patients. The overall incidence of new-onset AFi/AFl following ASD closure was 1.82 (95% CI 1.23 to 2.69) patients per 100 person-years based on the random-effects model (figure 3). A high level of heterogeneity was observed between studies for the overall meta-analysis (I2=83% and τ2=0.6400). The meta-analysis restricted to 12 publications from 2015 and later resulted in a similar overall incidence rate of 1.68 (95% CI 0.95 to 2.97) per 100 person-years (I2=87% and τ2=0.8035) for the random-effects model (online supplemental figure I).

Figure 3

Atrial fibrillation/flutter incidence in all studies included in quantitative synthesis.

Meta-analysis by age at closure

Four of the studies included in the meta-analysis were designed for an elderly population, with three studies designed for age ≥60 and one study designed for age ≥65 years. Meta-analysis of these four studies showed the incidence of new-onset AFi/AFl was 5.21 (95% CI 3.34 to 8.13) per 100 person-years (I2=0% or no heterogeneity, figure 4A). Meta-analysis of the remaining 19 studies demonstrated that the incidence of new-onset AFi/AFl was 1.62 (95% CI 1.07 to 2.45) per 100 person-years with an I2 of 84% (figure 4B).

Figure 4

Atrial fibrillation/flutter incidence by age. (A) In studies with only patients ≥60 years old. (B) In the remaining studies, not limited by age ≥60 years old.

Meta-analysis by follow-up duration

After removal of the four studies designed for age ≥60 years, meta-analysis of the remaining 19 studies had a high degree of heterogeneity (I2=84%). Six of the studies had a mean follow-up duration ≤2 years. Meta-analysis of these studies showed an AFi/AFl incidence of 4.05 (95% CI 2.03 to 8.09) per 100 person-years with I2 of 55% (figure 5A). Meta-analysis of the remaining 13 studies, with follow-up duration >2 years, showed an incidence of 1.19 (95% CI 0.75 to 1.89) per 100 person-years with an I2 of 85% (figure 5B).

Figure 5

Atrial fibrillation/flutter incidence by follow-up duration for studies not limited by age ≥60 years old. (A) In studies with follow-up ≤2 years. (B) In studies with follow-up >2 years.

Meta-analysis by method for rhythm assessment

Twelve studies in the quantitative meta-analysis incorporated scheduled ECG or Holter monitoring as part of their follow-up after ASD closure. Meta-analysis of these studies had an AFi/AFl incidence of 1.45 (95% CI 0.64 to 3.28) per 100 person-years with I2 87% (figure 6A). The remaining nine studies did not incorporate scheduled ECG or Holter monitoring for all patients. Meta-analysis of these studies had an AFi/AFl incidence of 2.22 (95% CI 1.49 to 3.31) with I2 77% (figure 6B).

Figure 6

Atrial fibrillation/flutter incidence by method for rhythm assessment. (A) In studies with scheduled ECG or Holter monitoring. (B) In studies without scheduled ECG or Holter monitoring.

Risk factors reported for the development of postprocedural atrial fibrillation

Only four studies evaluated independent risk factors for the development of postprocedural AFi/AFl (online supplemental table IV). Park et al identified age ≥48 years, deficient posteroinferior rim, and AFi/AFl during closure as significant risk factors (p<0.05) for the development of AFi/AFl >30 days after closure.18 Praz et al identified smoking and presence of a residual shunt on the first postoperative day as significant risk factors for new-onset AFi/AFl.17 Chiu et al demonstrated that mitral valve disease, multiple occluder implantation and a short run of atrial tachycardia during Holter examination were significantly associated factors.21


This review focused on the development of new-onset AFi/AFl in adults following transcatheter ASD closure. Closure may increase the risk of developing new arrhythmias through several mechanisms including local inflammation and stretching of the atrial septum.8 Over time mechanical irritation from contact with the atrial walls may also contribute to inflammation and arrhythmia development. It is important to note that even prior to closure, patients with ASDs are at increased risk for AFi/AFl development due to long-standing elevated right heart pressures, atrial and ventricular dilatation, and electrical remodelling.41 Following closure, sequelae of these changes can contribute to AFi/AFl independent of occluder placement. AFi/AFl should be distinguished from other postprocedure arrhythmias because a new diagnosis of AFi/AFl confers unique risks when compared with other supraventricular arrhythmias. A recent study identified AFi/AFl as the most common reason for 30-day readmission following transcatheter secundum ASD closure.42

AFi/AFl incidence in the general population and in medical/surgical secundum ASD management

The incidence rate of AFi/AFl in our meta-analysis was 1.82 (95% CI 1.23 to 2.69) per 100 person-years. In comparison, a recent population-based retrospective cohort study reported an AFi incidence of 0.47–0.68 per 100 person-years in the general adult population.43 A recent retrospective cohort study of patients with transcatheter secundum ASD repair compared outcomes against a matched population without ASDs and also reported an increased risk for development of new-onset AFi in patients with ASD (HR=3.73, 95% CI 2.79 to 4.98).44 In a study assessing new-onset AFi for medical management or surgical closure of ASDs (91% secundum type), the incidence rate was similar over a follow-up of 8.9 years at 17% and 15%, respectively.45 Another study assessing medically and surgically managed patients over a mean follow-up of 25 years found a similar incidence of chronic AFi at 35% and 33%, respectively.46 A large study reported similar AFi incidence between transcatheter and surgical closure (HR 1.5, 95% CI 0.6 to 3.5).47 The comparable incidence of AFi/AFl regardless of medical management, surgical closure or transcatheter closure may indicate that the greatest risk for arrhythmia development is the defect itself and its long-term structural and electrical sequelae.


Increasing age is a known risk factor for AFi/AFl in the general population and has been associated with development of AFi/AFl after transcatheter closure of ASDs.8 18 In our subgroup analysis, we found the studies designed for elderly patients had significantly higher incidence of AFi/AFl than the remaining studies (5.21 vs 1.62 per 100 person-years). The study by Ortega et al was the only study designed for a younger patient population (age 18–40) and reported a very low incidence of AFi/AFl.37

Follow-up duration

We know from clinical experience the periprocedural period is a time of increased risk of arrhythmia development. While performing the review, it was noted that these arrhythmias often developed within the first 2 years of follow-up. Through subgroup analysis, we found studies with follow-up duration ≤2 years reported much higher AFi/AFl incidence than studies with longer follow-up duration (4.05 vs 1.19 per 100 person-years). This association may indicate that the risk of developing AFi/AFl is not constant throughout the follow-up period, but rather is highest in the months and years immediately following intervention. This is expected due to the arrhythmogenic nature of device placement. Following this initial period, there may be favourable remodelling of the atria related to reduction in volume overload resulting in lower AFi/AFl risk over subsequent years. It is likely that for studies with very long follow-up, incidence rate would begin to rise as the population develops arrhythmias related to ageing. The association between shorter follow-up time and higher reported AFi/AFl incidence may have been confounded by increased monitoring in the early postprocedural period. During this period, patients often have more frequent clinic visits, ECGs and Holter monitoring.

Method of AFi/AFl monitoring

AFi/AFl incidence was not significantly different in studies that incorporated scheduled ECG or Holter monitoring during follow-up compared with those that did not. This may be because many studies without scheduled ECG/Holter monitoring still had frequent clinical assessments, which may have prompted investigations for rhythm assessment. Appraisal of AFi/AFl monitoring is also incomplete without incorporating frequency of assessments, which varied widely and were difficult to compare across modalities.

Implications for future research

More research is needed to characterise the incidence, epidemiology and risk factors for AFi/AFl before and after transcatheter ASD closure. It is crucial to identify these arrhythmias prior to closure since certain therapeutic options, including ablation, may be offered prior to closure but are more difficult to provide after occluder placement.38 Following closure, rigorous monitoring is needed to diagnose AFi/AFl and institute management to mitigate further complications. Implantable cardiac monitors are becoming increasingly recognised as a sensitive modality to monitor for AFi/AFl in at-risk populations where shorter monitoring modalities may miss the diagnosis; however, they are expensive and impractical for most applications. Wearable medical technology is also becoming accessible with the development of smartwatch devices, which may allow for wide-scale monitoring of arrhythmias. An effort must then be made to not only diagnose these patients, but to follow outcomes and complications from new-onset AFi/AFl. Once more information is available, patients can be risk-stratified and provided with tailored recommendations for postoperative monitoring and management.


Our study has several limitations. For most included studies, incidence of new-onset AFi/AFl was not the primary outcome. As a result, the quality of rhythm assessment was a significant limitation and varied greatly between studies. Many studies did not include ECG or Holter monitoring, and instead relied on clinical or telephone assessments to diagnose AFi/AFl or prompt investigations leading to the diagnosis. One of the included studies, Kadirogullari et al, required AFi to last 24 hours despite medical treatment to diagnose new-onset AFi. It is possible other studies also implemented strict criteria, but failed to specify this in their methods. The variability in follow-up methodology and the lack of uniform diagnostic criteria could have contributed to ascertainment bias and underestimation of the true incidence of AFi/AFl. In our analysis, procedural, periprocedural and postprocedural AFi/AFl were considered new-onset AFi/AFl, which may overestimate the incidence of clinically relevant AFi/AFl. Although our search strategy was designed with an information specialist, some relevant articles might have been missed. Meta-analysis was limited to studies reporting mean follow-up duration. Further, there was a high degree of heterogeneity across studies, partially related to features including differing definitions of new-onset AFi/AFl, varying detection methods, and differing frequency of screening assessments for AFi/AFl. Meaningful subgroup analysis based on cardiac risk factors, occluder type, defect/occluder size and echocardiographic features was not possible due to inconsistent reporting. In some studies, the baseline characteristics included patients that had a history of AFi/AFl prior to closure. Given these limitations, meta-analysis remained the most appropriate modality to pool and compare these studies.


This analysis found the incidence of new-onset AFi/AFl is relatively low following transcatheter closure of secundum ASDs; however, these arrhythmias were significantly more common in elderly patients. The incidence of AFi/AFl was noted to be significantly higher in studies with shorter follow-up time. Given the low incidence of postclosure AFi/AFl, the haemodynamic benefits of closure, and the competing risk of AFi/AFl due to dilation and remodelling in untreated patients, transcatheter closure of secundum ASDs should continue to be recommended for patients with indications for closure. This review provides useful information for both patients and clinicians about the risk of development of AFi/AFl after transcatheter closure in patients without a history of these conditions. More comprehensive follow-up monitoring is needed to better assess the incidence of AFi/AFl, identify risk factors and measure the impact of these conditions.

Key messages

What is already known on this subject?

  • The ostium secundum atrial septal defect (ASD) is one of the most common congenital cardiac anomalies in adulthood.

  • Transcatheter closure of the ostium secundum ASD has been associated with atrial arrhythmias, including atrial fibrillation and flutter (AFi/AFl).

  • Studies vary widely in their reported incidence of AFi/AFl following transcatheter closure, as well as reported risk factors for AFi/AFl development.

What might this study add?

  • This meta-analysis distinguishes AFi/AFl from other atrial arrhythmias following transcatheter ASD closure, recognising that AFi/AFl confer unique risks to this patient population.

  • We report the incidence of these important arrhythmias and identify that age ≥60 is significantly associated with their development.

How might this impact on clinical practice?

  • While atrial fibrillation and flutter are common disease entities in the general population, these arrhythmias are incompletely understood in the unique ASD patient population.

  • This study quantifies the incidence of AFi/AFl and begins to identify risk factors for their development.

  • Future studies can build on these findings to develop strategies for risk stratification, monitoring and tailored management of AFi/AFl in ostium secundum ASD patients following transcatheter closure.

Data availability statement

Data are available upon reasonable request.

Ethics statements

Patient consent for publication

Ethics approval

This study complies with the Declaration of Helsinki. Due to the review nature of this work, approbation by an ethics committee was not necessary.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • Contributors All authors were involved in the conceptualisation and design of the review. RKR established the search strategies with assistance from AF. JDH, HS, CJO and JF screened the titles, abstracts and full articles as per the protocol. JFH and HS assessed the quality of the included studies. JDH, LO, E Huszti and LA analysed the data. All authors discussed the findings. JDH drafted the manuscript. All authors provided critical feedback and contributed to the final manuscript. JDH and LA are responsible for the overall content as guarantors.

  • Funding This study was supported through funding from the Peter Munk Chair in Structural Heart Disease Interventions.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.