Sage Journals: Discover world-class research

Abstract

Background:

Randomized controlled trials (RCTs) that directly compare the efficacy and safety of percutaneous patent foramen ovale (PFO) closure devices have not been conducted. Thus, we performed a network meta-analysis to identify the efficacy and safety of occluder devices.

Methods:

From 1^st January, 2000 to 1^st May, 2018, we searched Embase, PubMed, and Cochrane Library for RCTs about percutaneous closure devices (such as STARFlex, GORE, and Amplatzer) and medical therapy for cryptogenic cerebral ischemic patients with PFO. The occurrence rate of recurrent stroke, atrial fibrillation (AF), major vascular complication (MVC), headache, transient ischemic attack, and bleeding were compared with the frequentist and Bayesian methods using R statistics.

Results:

We included 3747 patients from six RCTs. The GORE and Amplatzer occluders were found to be significantly associated with a decreased risk of recurrent stroke [relative risk (RR): 0.37 and 0.49; 95% confidence interval (CI): 0.17–0.81, 0.29–0.83, respectively]. Moreover, STARFlex was correlated to an increased risk of postoperative AF and MVCs (RR: 11.66 and 7.63; 95% CI: 4.87–21.91, 2.34–24.88).

Conclusions:

Among the three devices, the GORE and Amplatzer occluders are found to be the most effective in preventing secondary stroke in patients with PFO. Meanwhile, STARFlex is the least recommended device because it cannot decrease the risk of recurrent stroke and is the most likely to cause adverse events.

Keywords

patent foramen ovale (PFO)transcatheter closure stroke atrial fibrillation network meta-analysis

1 Introduction

The prevalence of cryptogenic stroke accounts for about 40% of strokes. Data have shown that the percutaneous closure of patent foramen ovale (PFO) can cause cryptogenic stroke, particularly in young individuals. Furthermore, 50% of patients with cryptogenic stroke are diagnosed with PFO based on transesophageal echocardiography reports. Previous guidelines on the prevention of stroke reoccurrence in patients with stroke and transient ischemic attack did not consider PFO closure as more beneficial than medical treatment [1, 2]. With the development of minimally invasive percutaneous approaches for PFO closure, several randomized controlled trials (RCTs) published within the last few years have pointed out a major historical turning point [3, 4]. That is, they have shown that foramen ovale closure is beneficial in patients with cryptogenic stroke and migraine, and this notion was quickly accepted by the academic community [5, 6]. Different kinds of occlusion devices for the closure of the foramen ovale are available. However, RCTs that directly compare the safety and efficacy of percutaneous PFO closure devices have not been conducted. Thus, this study aimed to perform a network meta-analysis to identify the occluder devices that have the highest and lowest efficacy by evaluating the adverse effects of various devices and drug treatments.

2 Methods

2.1 Search strategies

Two investigators independently performed an electronic search using Embase, PubMed, and Cochrane Library from 1^st January, 2000 to 1^st May, 2018. The search strategies were based on the following MeSH terms: ((“foramen ovale” OR “patent foramen ovale” OR “patent oval foramen” OR “PFO” OR “right to left shunt” OR “interatrial shunt” OR “heart septal defects” OR “atrial septal aneurysm”) OR (“percutaneous patent foramen ovale closure” OR “percutaneous PFO closure” OR “transcatheter closure”) OR (“occluder device” OR “anticoagulant” OR “antiplatelet” OR “STARFlex” OR “Amplatzer” OR “GORE” OR “Helex”)) AND (“stroke” OR “cryptogenic stroke” OR “transient ischemic attack” OR “death” OR “subarachnoid hemorrhage” OR “hemorrhage” OR “ischemic stroke” OR “atrial fibrillation”) AND “randomized controlled trial”.

The title, abstract, and main text were cautiously assessed. Language or publication status restrictions were not applied. Some supplementary materials listed in the article were also included in the assessment. Discrepancies were settled by mutual deliberations.

2.2 Description of design and selection criteria

The surgical treatment group was defined as the group who underwent foramen ovale closure operation with occluders during the treatment. Generally, anticoagulation and antiplatelet drugs were administered after closure. The drug treatment group was defined as the group who did not undergo any surgical procedures throughout the treatment period. Moreover, anticoagulant or antiplatelet drugs were administered to prevent stroke.

The inclusion criteria included the following: (1) RCTs of PFO closure or medical therapyinduced adverse outcomes and prognosis of PFO in patients with cryptogenic cerebral ischemic stroke, (2) experimental studies that included at least one group who underwent PFO closure, (3) research with a follow-up period longer than 2 years, (4) original studies of humans, and (5) studies written in English. Meanwhile, the exclusion criteria were as follows: (1) studies that did not provide detailed data about the outcomes of different devices during follow-up, (2) animal trials, (3) studies that did not have original data, and (4) duplicate reports. Discrepancies were settled by mutual deliberations.

2.3 Quality assessment

The Cochrane Collaboration Risk Assessment Tool was used to assess the methodological quality of the studies included in the analysis [7]. The methodological design has six aspects, namely, selection bias, performance bias, detection bias, attrition bias, reporting bias, and other biases, that required comprehensive evaluation. Two investigators assessed the quality of the studies, and discrepancies were settled by mutual deliberations.

2.4 Data extraction

The documents were extracted according to a pre-established data extraction table, which comprised data about the age and sex of the patients who were enrolled, treatment methods, drug used and dosage in the different groups, and follow-up time. We extracted the required data from the body of the text and supplementary materials. Two investigators independently performed data collection, and disagreements were resolved by a team discussion.

2.5 Endpoints

The primary endpoint was the occurrence of ischemic stroke, atrial fibrillation (AF), and major vascular complication (MVC). Ischemic stroke was defined as stroke that developed after patients underwent PFO closure or received medical treatments. The major MVCs included myocardial infarction, deep vein thrombosis, peripheral embolism, cardiac perforation, hematoma, and reoperation in general. The secondary outcomes were transient ischemic attacks (TIAs), major bleeding, and headache or migraine, which were common and considered serious complications in the cardiology department.

2.6 Statistical analysis

A network meta-analysis is an extension of the traditional pair-wise meta-analysis, as it combines both direct and indirect information via a common comparator, and obtains estimates of the relative interventional effects on multiple intervention comparisons [8].

Medical therapy was considered as the central reference in the whole network comparison system. Five studies conducted a direct comparison between medical therapy and one PFO closure device. The rest compared the three types of PFO closure device.

Surface under the cumulative ranking curve (SUCRA) was considered one of the most important indicators for evaluating the efficacy of different treatments. SUCRA is advantageous as it simplifies information about the effect of each treatment into a single number, and the whole complex multiple-treatment meta-analysis results were expressed with a few numbers. Moreover, it is most significant when the difference in preference between successive ranks remains the same across the entire ranking scale [9]. The study with a higher SUCRA score will have a higher recommendation level.

The Mantel–Haenszel and DerSimonian–Laird methods for fixed effects were used to estimate the pooled ratio risk (RR), and 95% confidence interval (CI) was accepted as the summary statistics, as implemented based on the frequentist model. I ², Q test, and the corresponding p value were used to estimate the heterogeneity of the studies. A p value of < 0.05 was considered statistically significant. R software 3.3.3 was used for data analysis. The whole analysis was conducted in accordance with the Cochrane Handbook and Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Checklist. Codeliberation was a prerequisite for all statistical conclusions.

3 Results

3.1 Demographic characteristics

After a detailed filtration, six RCTs that included four different interventions [medical therapy, Amplatzer^® PFO Occluder (Amplatzer), GORE^® Helex^® Septal Occluder or GORE^® CARDIOFORM^® Septal Occluder (collectively referred to as GORE), and Cardio-SEAL^® STARFlex™ device (STARFlex)] were included in this meta-analysis [4,10 –14]. Moreover, the analysis comprised 3747 patients, and the characteristics of the participants are listed in Table 1. The Amplatzer was the most commonly used closure device (n = 983). All patients received anticoagulation or antiplatelet therapy after surgical closure. The selection flow chart is depicted in Fig. S1. The characteristics of the enrolled population were similar, which indicated assumption of transitivity, and network calculations could then be conducted.

Table 1

Characteristics of the trials that were included.

Trials	Author, year	Closure devices	Sample size		Mean age (years)		Male sex (%)		Interventions		Follow-up time
Trials	Author, year	Closure devices	Pgroup	Mgroup	Pgroup	Mgroup	Pgroup	Mgroup	P group	M group	Follow-up time
CLOSURE I	Furlan et al., 2012	STARflex	447	462	46.3 ± 9.6	45.7 ± 9.1	52.1	51.5	Clopidogrel (75 mg daily) for 6 months and aspirin (81 or 325 mg daily) for 2 years	Warfarin (with a target international normalized ratio of 2.0-3.0), aspirin (325 mg daily) or both	2 years
RESPECT	Saver et al., 2017	Amplatzer	499	481	45.7 ± 9.7	46.2 ± 10.0	53.7	55.7	Aspirin (81-325 mg) plus clopidogrel daily for 1 month, followed by aspirin monotherapy for 5 months	Aspirin, warfarin, clopidogrel, and aspirin combined with extended-release dipyridamole. Aspirin with dopidogrel was also administered initially, but this regimen was eliminated in 2006	5.9 years
PC	Meier et al., 2017	Amplatzer	204	210	44.3 ± 10.2	44.6 ± 10.1	45.1	54.3	Acetylsalicylic acid (100-325 mg/d) for at least 5-6 months and ticlopidine (250-500 mg/d) or clopidogrel (75-150 mg/d) for 1-6 months	Antithrombotic treatment was based on the discretion of the attending physician and could have included antiplatelet or oral anticoagulation therapy provided that patients received at least one antithrombotic drug	4.1 years in P group, 4.0 years in M group
GORE REDUCE	Søndergaard et al., 2017	GORE	441	223	45.4 ± 9.3	44.8 ± 9.6	59.2	61.9	Clopidogrel (300 mg) before or immediately after the procedure, followed by 75 mg daily far 3 days; then, the antiplatelet therapy was resumed or started consistent with the medical therapy group	Aspirin alone (75-325 mg/d), a combination of aspirin (50-100 mg/d) and dipyridamole (225-400 mg/d), or clopidogrel (75 mg/d)	5.4 ±1.9 years in P group, 5.2 ± 2.1 years in M group
HORNUNG	Hornung et al., 2013	Amplatzer, STARflex, GORE	220/220 /220	N/A	49.3 ± 12.9	N/A	54.7	N/A	Aspirin (100 mg/d) and clopidogrel (75 mg/d) for the first 6 months after the procedure. Antiplatelet medication was administered for 30 days, 3 and 6 months after the intervention	N/A	5 years
DEFENSE-PFO	Lee et al., 2018	Amplatzer	60	60	49 ±15	54 ±12	55	56.7	Aspirin (100 mg/d) and clopidogrel (75 mg/d) for at least 6 months after the procedure	Aspirin and clopidogrel (75 mg/d), or aspirin and cilostazol (200 mg/d). Warfarin maintained the target international normalized ratio of 2.0-3.0	years in P group, years in M group

Amplatzer, Amplatzer® PFO Occluder; GORE, GORE® Helex® Septal Occluder or GORE® CARDIOFORM® Septal Occluder; STARFlex, Cardio-SEAL® STARFlex™ device; P group, patent foramen ovale (PFO) closure group; M group, medical therapy group.

3.2 Summary of methodological quality

Most qualified studies were of high quality in terms of methodology. Results for the risk of bias of the included articles using the Cochrane Collaboration Risk Assessment Tool are shown in Figs. S2 and S3. All studies provided complete data of the primary outcome. Some studies included the secondary endpoints. All articles described the principles of random assignment. Five of the six articles that were included have reported about random sequence generation. The blinding methods were not applied to the patients. The number of participants between the groups from the GORE REDUCE study [4] differed, which might have resulted in other potential biases.

3.3 Primary endpoint: stroke

Attention was paid to the outcome of stroke in six randomized trials. To validate the efficacy of the PFO closure devices that were included, we conducted a pair-wise estimation, and the network diagram of these studies is presented in Fig. S4(A). The GORE and Amplatzer have lower reoccurrence rates of stroke than drug therapy based on the pair-wise comparison (RR: 0.37 and 0.49; 95% CI: 0.17–0.81 and 0.29–0.83, respectively). However, a difference was not observed between STARFlex and drug-only therapy [RR: 0.89; 95% CI: 0.46–1.73; Fig. 1(A)]. The result was consistent in the network analysis, and patients who underwent PFO closure with the GORE had the best prognosis (Table 2). According to the network comparison result, the group who underwent PFO closure with the GORE (89.76%) had the highest SUCRA score, followed by the group in which the Amplatzer (73.31%) was used. Meanwhile, the medical treatment group (12.53%) had the lowest SUCRA score, which indicated that the GORE might be most effective in preventing recurrent stroke (Table 3). The heterogeneity and inconsistency tests had a low level of heterogeneity (I ² = 12.3%, τ² = 0.0632, Q = 4.56, and p = 0.3355). Eliminating any study did not significantly affect the overall outcome using the fixed effect model.

Fig. 1

Direct meta-analysis of the calculation results. (A) All closure strategies compared with medical therapy for stroke. (B) All closure strategies compared with medical therapy for atrial fibrillation (AF). (C) All closure strategies compared with medical therapy for major vascular complications (MVC). (D) All closure strategies compared with medical therapy for headache. (E) All closure strategies compared with medical therapy for transient ischemic attack (TIA). (F) All closure strategies compared with medical therapy for bleeding. Amplatzer, Amplatzer^® PFO Occluder; GORE, GORE^® Helex^® Septal Occluder or GORE^® CARDIOFORM^® Septal Occluder; STARFlex, Cardio-SEAL^® STARFlex™ device.

Table 2

Network meta-analysis results in different patients with cryptogenic cerebral ischemic stroke using weighted mean difference, RR, and 95% CI.

Items		Comparisons			Quantifying heterogeneity
Items		Comparisons			I²	τ²	Q	p
Recurrent stroke	Medical Therapy	1.1228 (0.5791–2.1769)	2.0374 (1.2118–3.4256)	2.6797 (1.2298–5.8388)
	0.8906 (0.4594–1.7268)	STARFlex	1.8146 (0.8183–4.0239)	2.3866 (0.9898–5.7546)	12.30%	0.0632	34.56	0.3355
	0.4908 (0.2919–0.8252)	0.5511 (0.2485–1.2221)	Amplatzer	1.3152 (0.5373–3.2193)
	0.3732 (0.1713–0.8131)	0.419 (0.1738–1.0103)	0.7603 (0.7603–1.8611)	GORE
AF	Medical Therapy	0.0858 (0.1186–0.2053)	0.2973 (0.1186–0.7456)	0.3315 (0.1123–0.9783)
	11.6585 (4.8698–27.9106)	STARFlex	3.4665 (1.7171–6.9984)	3.865 (1.6252–9.1915)	16.20%	0.1146	3.58	0.3106
	3.3632 (1.3412–8.4334)	0.2885 (0.1429–0.5824)	Amplatzer	1.115 (0.4104–3.0290)
	3.0164 (1.0222–8.9013)	0.2587 (0.1088–0.6153)	0.8969 (0.3301–2.4366)	GORE
MVC	Medical Therapy	0.1311 (0.0402–0.4277)	0.5137 (0.2086–1.2652)	0.1817 (0.0572–0.5773)
	7.6277 (2.3382–24.8835)	STARFlex	3.9187 (1.2449–12.3347)	1.3862 (0.7807–2.4614)	64%	1.4653	8.34	0.0395
	1.9465 (0.7904–4.7938)	0.2552 (0.0811–0.8033)	Amplatzer	0.3537 (0.1137–1.1005)
	5.5026 (1.7321–17.4812)	0.7214 (0.4063–1.281)	2.8269 (0.9087–8.7944)	GORE
Headache	Medical Therapy	4.8378 (0.2329–100.489)	0.6031 (0.3273–1.1116)	2.6368 (0.5953–11.6793)
	0.2067 (0.01–4.2935)	STARFlex	0.1247 (0.0056–2.7525)	0.545 (0.0186–15.9917)	0	0	0.19	0.6602
	1.658 (0.8996–3.0555)	8.021 (0.3633–177.0859)	Amplatzer	4.3717 (0.8748–21.8475)
	0.3793 (0.0856–1.6798)	1.8348 (0.0625–53.8335)	0.2287 (0.0458–1.1431)	GORE
TIA	Medical Therapy	1.1229 (0.5718–2.2052)	1.5394 (0.9119–2.5989)	1.4636 (0.4394–4.8755)
	0.8905 (0.4535–1.7488)	STARFlex	1.3709 (0.5934–3.1672)	1.3034 (0.4393–3.8673)	0	0 2.	82	0.5878
	0.6496 (0.3848–1.0966)	0.7294 (0.3157–1.6853)	Amplatzer	0.9508 (0.2614–3.4582)
	0.6832 (0.2051–2.2759)	0.7672 (0.2586–2.2764)	1.0518 (0.2892–3.8257)	GORE
Bleeding	Medical Therapy	0.387 (0.1223–1.225)	2.0114 (0.5337–7.5812)	1.4832 (0.521–4.2223)
	2.5839 (0.8163–8.1785)	STARFlex	5.1973 (0.8966–30.1275)	3.8324 (0.8083–18.1704)	0	0	0.89	0.6394
	0.4972 (0.1319–1.8738)	0.1924 (0.0332–1.1153)	Amplatzer	0.7374 (0.1361–3.9948)
	0.6742 (0.2368–1.9194)	0.2609 (0.055–1.2372)	1.3562 (0.2503–7.3472)	GORE

Table 3

Results of the SUCRA rankings for different treatments.

Endpoint		SUCRA Rankings	SUCRA Score
Recurrent stroke	1	GORE	0.8976
	2	Amplatzer	0.7331
	3	STARFlex	0.2439
	4	Medical Therapy	0.1253
AF	1	Medical Therapy	0.9908
	2	GORE	0.5354
	3	Amplatzer	0.4734
	4	STARFlex	0.0005
MVC	1	Medical Therapy	0.9747
	2	Amplatzer	0.6759
	3	GORE	0.3019
	4	STARFlex	0.0476

SUCRA, surface under the cumulative ranking curve; AF, atrial fibrillation; MVC, major vascular complications; Amplatzer, Amplatzer^® PFO Occluder; GORE, GORE^® Helex^® Septal Occluder or GORE^® CARDIOFORM^® Septal Occluder; STARFlex, Cardio-SEAL^® STARFlex™ device.

Amplatzer, Amplatzer^® PFO Occluder; GORE, GORE^® Helex^® Septal Occluder or GORE^® CARDIOFORM^® Septal Occluder; STARFlex, Cardio-SEAL^® STARFlex™ device.

3.4 Primary endpoint: AF

Direct and network comparisons: five RCTs were included in the analysis of AF after PFO closure or medical therapy. The network diagram of these studies is depicted in Fig. S4(B). A direct comparison showed that patients who underwent PFO closure generally had an increased incidence of AF compared with patients who received treatment with drugs [Fig. 1(B)]. The STARFlex had an increased risk of AF by more than 10 folds compared with treatment with drugs alone (RR: 11.66; 95% CI: 4.87–21.91; Table 2). As for network calculation, while medical therapy (99.08%) is advantageous in the entire hierarchy, the STARFlex (0.05%) had the lowest rank in the recommended order, which indicated that the use of the device for foramen ovale closure may have an increased risk of AF (Table 3). Quantifying heterogeneity was acceptable (I ² = 16.2%, τ ² = 0.1146, Q = 3.58, and p = 0.3106).

3.5 Primary endpoint: MVC

A pair-wise comparison of six studies showed that the STARFlex had an increased risk of vascular accidents in patients with PFO compared with conservative treatment with drugs [RR: 7.63; 95% CI: 2.34–24.88; Fig. 1(C)]. The difference in vascular complications between the Amplatzer device and drug therapy was not statistically significant (RR: 1.95; 95% CI: 0.79–4.79; Table 2). The network diagram of the selected studies is depicted in Fig. S4(C). By contrast, the use of medications (97.47%) had the highest recommendation ranking in the SUCRA score in the network analysis, indicating that it might be safer than surgery for PFO patients in terms of vascular adverse effects. The STARFlex (4.76%) might cause the highest number of vascular complications. Meanwhile, the Amplatzer (67.59%) was considered the safest device (Table 3). Notably, the heterogeneity was relatively high (I ² = 16.2%, τ² = 0.1146, Q = 3.58, and p = 0.3106). The heterogeneity of MVC was attributed to the GORE REDUCE trial, and after excluding the studies by Søndergaard et al., the network comparison heterogeneity decreased (I ² = 0%, τ² = 0, Q = 0.13, and p = 0.9348). We speculated that the source of heterogeneity might be correlated to the varying number of individuals included in each study group. Correspondingly, the results of the pairwise and network comparisons and the SUCRA rank findings remained relatively stable (Tables S1 and S2, Fig. S5).

3.6 Secondary endpoint

Headache, TIAs, and major bleeding were the outcomes of the comparison after a long-term follow-up. Some studies included the incidence of the three endpoint events. The network diagrams of the studies about these endpoints are depicted in Figs. S4(D)–(F). A pair-wise comparison of the enrolled studies did not show any difference in the three evaluation index [Figs. 1(D)–(F)]. The indirect comparison results were consistent with those of the direct comparison, which did not indicate any significant difference between all treatments (Table 2). The quantifying heterogeneity was low. The stability and reliability of the combined results were acceptable.

4 Discussion

Study results about whether closure is required for patients with PFO and cryptogenic stroke switched from negative to affirmative. Studies with longer follow-up time and larger sample size make conclusions more reliable. Patients with PFO can benefit from closure. Thus, choosing an appropriate closure device became a problem worthy of attention. At present, only one study has conducted a direct comparison of RCTs of occlusion devices [13], which is insufficient for clinical guidance. Thus, the present study aimed to further compare the advantages and disadvantages of using closure devices, not to reassess the notion whether foramen ovale closure is more effective than drug treatment in preventing the recurrence of cryptogenic stroke. Compared with previous meta-analyses [15, 16], this systemic review and network meta-analysis did not only include a higher number of studies and evaluation indicators to obtain comprehensive conclusions from indirect comparisons but also chose the most suitable and clinically effective closure device, and this is considered the strength of the study. The ideal device for transcatheter PFO closure should facilitate complete and rapid defect closure, and it must be safely, easily implanted and adaptable [17]. Although the device might be associated with coagulation disorders, preventing the formation and shedding of thromboembolism in patients who underwent PFO closure cannot be ignored. PFO closure with the GORE is superior to other devices in terms of preventing recurrent stroke; the STARflex has less protective effects than treatment with medications. This result is consistent with that of previous studies. According to previous reports, both the Amplatzer and the GORE^® HELEX^® devices can be considered the least thrombogenic [14, 18 –20]. However, the STARflex device resulted in frequent thrombus formation [18].

Compared to medical treatment, percutaneous closure, particularly with the STARflex device, was more likely to lead to atrial arrhythmias and MVCs, and this notion has been proven in a previous research [18]. Based on the characteristics of the surgery, we assessed the cause of multiple postoperative AF. Most atrial arrhythmias and MVCs occurred during the surgical and perioperative periods. Residual shunt after closure, left atrial enlargement, and old age might be the risk factors for the increased risk of AF [19]. In addition, compared with the medication group, the perioperative electrocardiogram monitoring of the PFO closure groups had greater accuracy, which might lead to an increase in the detection rate of AF [21]. Therefore, in patients with PFO and cryptogenic stroke, proper monitoring and prevention of AF are important after closure during the perioperative period [22]. PFO occlusion-related AF is usually self-limiting. Perioperative AF events are not likely associated with new strokes because an increase in stroke events was not observed in the interventional occlusion group. The occurrence of MVCs, including bleeding and vascular hematoma formation, might differ between surgical procedures and operators’ manipulations. Moreover, considering that the MVCs vary in different studies and the incidence of adverse events is relatively low, heterogeneity and volatility of statistical results might occur.

No difference was observed in the occurrence rate of headache, TIA, and major bleeding between different therapies. TIA is a transient neurological deficit attributed to focal brain, retinal, or spinal cord ischemia. The clinical symptoms of TIA do not exceed 1 h, and there is generally no evidence of cerebral infarction. Some neurological disorders, such as epileptic seizure, Meniere’s syndrome, and migraine, are also known to recur [23 –25]. TIA lacks a specific diagnosis; thus, accurate assessments are required to prevent misdiagnosis and missed diagnosis. The statistics of the studies about TIA may not be completely reliable. Major bleeding events include spontaneous intracerebral hemorrhage and hemorrhagic transformation of acute ischemic stroke during the anticoagulant or antiplatelet period. Most major bleeding events occur during the medication period [26], which indicate that the process of a simple PFO closure might not increase the risk of bleeding.

As mentioned previously, the STARFlex is commonly associated with the development of thrombus and atrial arrhythmias. Based on previous studies, the main technical difficulty of the device was its nonretrievable nature; thus, complete removal is required to correct its position [18]. During this process, the myocardium and blood vessels could be damaged. Currently, no direct evidence has shown which treatment option is better. Studies with larger sample size and longer follow-up time must be conducted to validate the reliability of this conclusion. The extended follow-up trial of RESPECT in 2017 has shown a diametrically opposite result compared with the study in 2013 [11, 27]. Thus, we only included studies with follow-up periods of more than 2 years. Long-term follow-up and observations are more robust than the evaluation of short-term outcomes.

Oral anticoagulants, single antiplatelet therapy, and dual antiplatelet therapy were included and compared in different medical therapies [16]. These regimens have differential effects on the recurrence of cryptogenic stroke. Since the closure group used anticoagulant and antiplatelet therapy after operation and the therapies included in the study were different, we could not further subdivide the medication group to explore which drug treatment is more effective in this meta-analysis, and this is one of the limitations of this study. However, this current research aimed to identify the most appropriate closure devices via comparison, and a detailed drug treatment can be used as a reference. From a statistical point of view, different kinds of drug treatments do not affect the safety and efficacy of closure devices.

Moreover, in terms of limitation, although more studies were included in the current study than in previous studies, our sample size was still relatively small. Thus, larger multi-arm studies that directly compare different treatments must be conducted to obtain evidence-based medical conclusions. The outcomes included in each study were not completely consistent, and the difference might be attributed to small-sample effects.

5 Conclusion

PFO closure has advantages as it prevents secondary stroke. Among the devices, the GORE and Amplatzer had the best performance with cumulative ranking probability. However, the use of occluders increases the risk of adverse events, such as AF and MVCs, due to occlusion of the foramen ovale. Meanwhile, the STARFlex is the least recommended device as it is most likely to cause adverse events. However, no difference was observed in terms of the occurrence of headache and TIAs between the use of closure devices and treatment with medications.

Supplemental Material

Supplemental Material, BSA20190014_Supplementary_material - Efficacy and safety of percutaneous patent foramen ovale closure devices for recurrent stroke: A systemic review and network metaanalysis

Supplemental Material, BSA20190014_Supplementary_material for Efficacy and safety of percutaneous patent foramen ovale closure devices for recurrent stroke: A systemic review and network metaanalysis by Hao Nie, Yang Hu and Zhouping Tang in Brain Science Advances

Footnotes

Conflict of interests

The authors declare that they have no conflict of interests.

Financial support

This study was supported by grants from the National Natural Science Foundation of China (No. 81471201, No. 81873750).

References

Kernan

Ovbiagele

Black

, et al. Guidelines for the prevention of stroke in patients with stroke and transient ischemic attack: a guideline for healthcare professionals from the American Heart Association/American Stroke Association. Stroke. 2014, 45(7): 2160–2236.

Messé

Gronseth

Kent

, et al. Practice advisory: Recurrent stroke with patent foramen ovale (update of practice parameter): Report of the Guideline Development, Dissemination, and Implementation Subcommittee of the American Academy of Neurology. Neurology. 2016, 87(8): 815–821.

Mas

Derumeaux

Guillon

, et al. Patent foramen ovale closure or anticoagulation vs. antiplatelets after stroke. N Engl J Med. 2017, 377(11):1011–1021

Søndergaard

Kasner

Rhodes

, et al. Patent foramen ovale closure or antiplatelet therapy for cryptogenic stroke. N Engl J Med. 2017, 377(11): 1033–1042.

Stortecky

da Costa

Mattle

, et al. Percutaneous closure of patent foramen ovale in patients with cryptogenic embolism: a network meta-analysis. Eur Heart J. 2015, 36(2): 120–128.

Kent

Dahabreh

Ruthazer

, et al. Device closure of patent foramen ovale after stroke: pooled analysis of completed randomized trials. J Am Coll Cardiol. 2016, 67(8): 907–917.

Higgins

Altman

Sterne

. Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0. http://handbook.cochrane.org (Available March, 2011).

Neupane

Richer

Bonner

, et al. Network meta-analysis using R: a review of currently available automated packages. PLoS One. 2014, 9(12): e115065.

Salanti

Ades

Ioannidis

. Graphical methods and numerical summaries for presenting results from multiple-treatment meta-analysis: an overview and tutorial. J Clin Epidemiol. 2011, 64(2): 163–171.

10.

Furlan

Reisman

Massaro

, et al. Closure or medical therapy for cryptogenic stroke with patent foramen ovale. N Engl J Med. 2012, 366(11): 991–999.

11.

Saver

Carroll

Thaler

, et al. Long-term outcomes of patent foramen ovale closure or medical therapy after stroke. N Engl J Med. 2017, 377(11): 1022–1032.

12.

Meier

Kalesan

Mattle

, et al. Percutaneous closure of patent foramen ovale in cryptogenic embolism. N Engl J Med. 2013, 368(12): 1083–1091.

13.

Hornung

Bertog

Franke

, et al. Long-term results of a randomized trial comparing three different devices for percutaneous closure of a patent foramen ovale. Eur Heart J. 2013, 34(43): 3362–3369.

14.

Lee

Song

Kim

, et al. Cryptogenic stroke and high-risk patent foramen ovale: the DEFENSE-PFO trial. J Am Coll Cardiol. 2018, 71(20): 2335–2342.

15.

Tsivgoulis

Katsanos

Mavridis

, et al. Percutaneous patent foramen ovale closure for secondary stroke prevention: Network meta-analysis. Neurology. 2018, 91(1): e8–e18.

16.

Mir

Siemieniuk

RAC

, et al. Patent foramen ovale closure, antiplatelet therapy or anticoagulation in patients with patent foramen ovale and cryptogenic stroke: a systematic review and network meta-analysis incorporating complementary external evidence. BMJ Open. 2018, 8(7): e023761.

17.

Hardt

Eicken

Berger

, et al. Closure of patent foramen ovale defects using GORE® CARDIOFORM septal occluder: Results from a prospective European multicenter study. Catheter Cardiovasc Interv. 2017, 90(5): 824–829.

18.

Taaffe

Fischer

Baranowski

, et al. Comparison of three patent foramen ovale closure devices in a randomized trial (Amplatzer versus CardioSEAL-STARflex versus Helex occluder). Am J Cardiol. 2008, 101(9): 1353–1358.

19.

Krumsdorf

Ostermayer

Billinger

, et al. Incidence and clinical course of thrombus formation on atrial septal defect and patient foramen ovale closure devices in 1000 consecutive patients. J Am Coll Cardiol. 2004, 43(2): 302–309.

20.

von Bardeleben

Richter

Otto

, et al. Long term follow up after percutaneous closure of PFO in 357 patients with paradoxical embolism: Difference in occlusion systems and influence of atrial septum aneurysm. Int J Cardiol. 2009, 134(1): 33–41.

21.

Gaillard

Deltour

Vilotijevic

, et al. Detection of paroxysmal atrial fibrillation with transtelephonic EKG in TIA or stroke patients. Neurology. 2010, 74(21): 1666–1670.

22.

Freedman

Potpara

Lip

. Stroke prevention in atrial fibrillation. Lancet. 2016, 388(10046): 806–817.

23.

Lamberink

Otte

Geerts

, et al. Individualised prediction model of seizure recurrence and long-term outcomes after withdrawal of antiepileptic drugs in seizure-free patients: a systematic review and individual participant data meta-analysis. Lancet Neurol. 2017, 16(7): 523–531.

24.

Gürkov

Pyykö

Zou

, et al. What is Menière’s disease? A contemporary re-evaluation of endolymphatic Hydrops . J Neurol. 2016, 263(Suppl 1): S71–S81.

25.

Schankin

Viana

Goadsby

. Persistent and repetitive visual disturbances in migraine: a review. Headache. 2017, 57(1): 1–16.

26.

Butera

Saracino

Danna

, et al. Transcatheter PFO closure with GORE septal occluder: early and mid-term clinical results. Catheter Cardiovasc Interv. 2013, 82(6): 944–949.

27.

Carroll

Saver

Thaler

, et al. Closure of patent foramen ovale versus medical therapy after cryptogenic stroke. N Engl J Med. 2013, 368(12): 1092–1100.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.64 MB