Abstract
Background
Patients with history of cryptogenic stroke are more likely to have a patent foramen ovale (PFO) and should be managed with antithrombotic agents, while the alternative option is percutaneous closure of PFOs. Our aim was to perform a meta-analysis of randomized controlled trials (RCTs) comparing percutaneous closure vs. medical treatment for patients with PFO and prior cryptogenic stroke.
Methods
Medline, Scopus and Cochrane databases were reviewed. A random-effect model meta-analysis was used and I-square was utilized to assess the heterogeneity. New ischemic stroke was defined as the primary endpoint. A sensitivity analysis was performed for Amplatzer device. Subgroup analyses were performed for different patient and PFO characteristics for the composite endpoints as defined by the included RCTs.
Results
In total of 3440 patients were included in this meta-analysis. Closure devices were superior to medical therapy for prevention of recurrent ischemic strokes (HR = 0.29; CI: 0.02–0.56), but were associated with increased risk of new onset of atrial fibrillation (AF) and atrial flutter (RR = 4.67; CI: 2.22–9.81). However, in the sensitivity analysis for Amplatzer device, there was no difference between the two groups in new onset of atrial arrhythmias. Closure devices were superior across all different subgroups when compared to medical treatment with the exception of patients with a small shunt.
Conclusion
This meta-analysis shows that closure devices for patients with PFO and history of cryptogenic stroke can significantly decrease the risk of a new ischemic stroke. The use of Amplatzer device was not associated with increased risk of newly diagnosed atrial arrhythmias.
1
Introduction
Stroke is the second leading cause of death worldwide and fifth leading cause of death in the United States [ ]. The term cryptogenic stroke, which has been largely replaced by the term “embolic strokes of undetermined source (ESUS)”, is used to define strokes for which, despite the appropriate diagnostic evaluations, a cause cannot be identified [ ]. Cryptogenic strokes account for almost 40% of all the ischemic strokes and occur in about 150,000 patients each year in the United States [ ]. A mechanism that can potentially explain some of these strokes is paradoxical embolism, which is the shunting of embolic particles from the venous to the arterial circulation [ ]. An identifiable cause of this venous to arterial shunt (right to left shunt; [RLS]) is a patent foramen ovale (PFO) [ ]. The foramen ovale allows right-to-left intra-cardiac shunting in the developing fetus and has been found to remain patent after the first three months of life in almost 25% of the general population [ , ]. Patients with cryptogenic strokes were found to have a higher rate of PFO compared to the general population (40% vs. 25%) [ ]. The most recent guidelines for Prevention of Stroke in Patients with Stroke and Transient Ischemic Attack (TIA) recommend that patients with history of ischemic stroke or TIA, PFO, and potential source of embolism should be treated with oral anticoagulation (Class I; Level of Evidence A). For those without a source of embolism, antiplatelet therapy is recommended (Class I; Level of Evidence B) [ ].
The first percutaneous PFO closure was described in 1992. Since then many real-world studies in addition to three randomized controlled trials (RCTs) compared interventional vs. medical therapy with contradictory outcomes. It was only recently that the US Food and Drug Administration (FDA) approved transcatheter PFO closure with the Amplatzer PFO Occluder (Abbott, Chicago, Illinois) in October 28, 2016 to reduce the risk of recurrent cryptogenic stroke. In late 2017, two new RCTs comparing percutaneous PFO closure vs. medical therapy and the long-term results of the Randomized Evaluation of Recurrent Stroke Comparing PFO Closure to Established Current Standard of Care Treatment (RESPECT) trial were added to the previously known evidence, showing overall favorable outcomes for the percutaneous closure group. The aim of this meta-analysis was to assess for more robust conclusions for the comparative effectiveness and safety of percutaneous closure vs. medical treatment in patients with a PFO and a history of cryptogenic stroke.
2
Methods
This meta-analysis was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [ ]. Medline, Scopus and Cochrane databases were reviewed for RCTs comparing percutaneous PFO closure vs. medical therapy up to November 12, 2017.
2.1
Study selection and data extraction
The algorithm used for the Medline, Scopus, and Cochrane databases was: “patent foramen ovale” OR “percutaneous closure of patent foramen ovale” OR “foramen ovale” OR “right to left shunt” OR “atrial septal aneurysm” OR “interatrial shunt” OR “occluder device” Two reviewers (LP, DGK) independently reviewed the three databases for studies eligible for this meta-analysis. A third reviewer (IM) was involved as needed to reach consensus. In order to identify further eligible studies, manual searches of the references list of the included studies and pertinent reviews were performed. The following criteria were required for a study to be eligible for this meta-analysis: i) RCTs published in any language comparing percutaneous PFO closure vs medical therapy; ii) studies that included patients with a history of cryptogenic stroke.
Two independent reviewers (LP, IM) extracted the data for all pertinent variables in a predefined data collection form. Discrepancies were resolved with the involvement of a third reviewer (DGK). Data for the following baseline variables were extracted: study name, inclusion criteria, exclusion criteria, device used, antiplatelet/anticoagulation regimen prescribed, antiplatelet/anticoagulation prescribed in the device group, follow-up period, age, body mass index, gender, smoking status, family history of stroke, presence of coronary artery disease, diabetes, dyslipidemia, hypertension or peripheral artery disease, migraines, history of stroke, TIA, migraines, thromboembolic events, deep venous thrombosis (DVT), pulmonary embolism (PE), shunt size, and presence or absence of septal aneurysm.
2.2
Outcomes
The primary efficacy endpoint was new ischemic stroke. The primary safety endpoint was new onset of atrial fibrillation (AF) or atrial flutter. Secondary endpoints included TIA, mortality, bleeding, new onset of DVT or PE, and any serious adverse events as defined by the original studies. A sensitivity analysis was conducted for studies using Amplatzer Occluder device only. Subgroup analyses were performed for the following characteristics: shunt size as defined by the individual study (small vs. large), presence or absence of a septal aneurysm, gender, age (<45 vs. ≥ 45 years old) for the primary composite endpoint as defined by the included studies.
2.3
Risk of bias Assessment
Two independent reviewers (LP, IM) assessed the risk of bias of the included studies according to the Cochrane risk of bias tool for RCTs [ ]. Studies were assessed as having low, high or unclear risk of bias for the following domains: random sequence generation; allocation concealment; selective reporting; blinding of participants and personnel; blinding of outcome assessment; incomplete outcome data; other bias not covered elsewhere among the domains.
2.4
Data synthesis and statistical analysis
Definitions of the included outcomes were used as defined in the original studies. Subgroup analyses were based on available data from RCTs for their primary outcomes (stroke or composite of stroke, TIA and early death). For the subgroup analysis in patients with large shunt, we used large shunt definitions as provided by the individual studies and accepted different definitions for the presence of a large shunt, including: substantial, severe, and moderate to large. A meta-analysis of hazard ratios (HR) was conducted for stroke, TIA mortality and the subgroup analyses. A meta-analysis using event rates was conducted when HRs were not available. A random effects model was selected a priori because the included studies had heterogeneous study design and baseline patient’s characteristics [ ]. Forest plots were used to illustrate the individual study findings and the random effects meta-analysis results. The I-square statistic (I 2 ) was used to assess for heterogeneity among the studies [ ]. Values <25% indicated low, 25% to 70% indicated moderate, and >70% indicated severe heterogeneity [ ]. The estimated incidence rates were expressed as percentage and 95% confidence interval (CI). The results were regarded as statistically significant at a two-sided p ≤ 0.05. We used STATA 14.1 (StataCorp, College Station, Texas) as the statistical software for all analyses.
2
Methods
This meta-analysis was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [ ]. Medline, Scopus and Cochrane databases were reviewed for RCTs comparing percutaneous PFO closure vs. medical therapy up to November 12, 2017.
2.1
Study selection and data extraction
The algorithm used for the Medline, Scopus, and Cochrane databases was: “patent foramen ovale” OR “percutaneous closure of patent foramen ovale” OR “foramen ovale” OR “right to left shunt” OR “atrial septal aneurysm” OR “interatrial shunt” OR “occluder device” Two reviewers (LP, DGK) independently reviewed the three databases for studies eligible for this meta-analysis. A third reviewer (IM) was involved as needed to reach consensus. In order to identify further eligible studies, manual searches of the references list of the included studies and pertinent reviews were performed. The following criteria were required for a study to be eligible for this meta-analysis: i) RCTs published in any language comparing percutaneous PFO closure vs medical therapy; ii) studies that included patients with a history of cryptogenic stroke.
Two independent reviewers (LP, IM) extracted the data for all pertinent variables in a predefined data collection form. Discrepancies were resolved with the involvement of a third reviewer (DGK). Data for the following baseline variables were extracted: study name, inclusion criteria, exclusion criteria, device used, antiplatelet/anticoagulation regimen prescribed, antiplatelet/anticoagulation prescribed in the device group, follow-up period, age, body mass index, gender, smoking status, family history of stroke, presence of coronary artery disease, diabetes, dyslipidemia, hypertension or peripheral artery disease, migraines, history of stroke, TIA, migraines, thromboembolic events, deep venous thrombosis (DVT), pulmonary embolism (PE), shunt size, and presence or absence of septal aneurysm.
2.2
Outcomes
The primary efficacy endpoint was new ischemic stroke. The primary safety endpoint was new onset of atrial fibrillation (AF) or atrial flutter. Secondary endpoints included TIA, mortality, bleeding, new onset of DVT or PE, and any serious adverse events as defined by the original studies. A sensitivity analysis was conducted for studies using Amplatzer Occluder device only. Subgroup analyses were performed for the following characteristics: shunt size as defined by the individual study (small vs. large), presence or absence of a septal aneurysm, gender, age (<45 vs. ≥ 45 years old) for the primary composite endpoint as defined by the included studies.
2.3
Risk of bias Assessment
Two independent reviewers (LP, IM) assessed the risk of bias of the included studies according to the Cochrane risk of bias tool for RCTs [ ]. Studies were assessed as having low, high or unclear risk of bias for the following domains: random sequence generation; allocation concealment; selective reporting; blinding of participants and personnel; blinding of outcome assessment; incomplete outcome data; other bias not covered elsewhere among the domains.
2.4
Data synthesis and statistical analysis
Definitions of the included outcomes were used as defined in the original studies. Subgroup analyses were based on available data from RCTs for their primary outcomes (stroke or composite of stroke, TIA and early death). For the subgroup analysis in patients with large shunt, we used large shunt definitions as provided by the individual studies and accepted different definitions for the presence of a large shunt, including: substantial, severe, and moderate to large. A meta-analysis of hazard ratios (HR) was conducted for stroke, TIA mortality and the subgroup analyses. A meta-analysis using event rates was conducted when HRs were not available. A random effects model was selected a priori because the included studies had heterogeneous study design and baseline patient’s characteristics [ ]. Forest plots were used to illustrate the individual study findings and the random effects meta-analysis results. The I-square statistic (I 2 ) was used to assess for heterogeneity among the studies [ ]. Values <25% indicated low, 25% to 70% indicated moderate, and >70% indicated severe heterogeneity [ ]. The estimated incidence rates were expressed as percentage and 95% confidence interval (CI). The results were regarded as statistically significant at a two-sided p ≤ 0.05. We used STATA 14.1 (StataCorp, College Station, Texas) as the statistical software for all analyses.
3
Results
3.1
Study Selection Process and Studies Characteristics
In total 9570 studies were screened and 7 full-text articles were reviewed for eligibility. A PRISMA flow diagram with the selection process is shown in the Data Supplement ( Supplementary Fig. A ). Six studies met the inclusion criteria [ ]. One of them (RESPECT 2013) reported the earlier results of another included study (RESPECT 2017) and thus was not included in our analysis. The characteristics of the studies are summarized in Supplementary Table 1 . In total, 3440 patients were included. From the CLOSE (Patent Foramen Ovale Closure or Anticoagulants versus Antiplatelet Therapy to Prevent Stroke Recurrence) trial, only the arms that directly compared PFO closure vs. medical therapy were included. The mean age ranged from 42.9 ± 10 to 46.3 ± 9.6 while the percentage of female patients ranged from 47.3% to 60.1%. Details on baseline patient characteristics are presented in Supplementary Table 2 . The risk of bias assessment of included studies is presented in detail in the data supplement ( Supplementary Fig. B ), while the overall risk of bias for each included study is summarized briefly in Supplementary Table 1 .
3.2
Primary Outcomes
Patients treated with a closure device had a lower risk of stroke compared to patients treated with medical therapy alone (HR: 0.29; 95% CI: 0.02–0.56; I 2 = 67.1%) ( Fig. 1 ). The benefit associated with the closure devices was maintained regardless of device used. Patients treated with a closure device had a higher risk of atrial fibrillation (RR: 4.66; 95% CI: 2.22–9.81; I 2 = 24.4%) ( Fig. 2 ). Studies where Amplatzer devices were exclusively used did not show a significant increase in the risk of AF/Atrial flutter (RR: 2.11; 95% CI: 0.80–5.56; I 2 = 0) ( Fig. 3 ).
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
