- Email: [email protected]
A systematic review of closure versus medical therapy for preventing recurrent stroke in patients with patent foramen ovale and cryptogenic stroke or transient ischemic attack
Journal of the Neurological Sciences 337 (2014) 3–7
Contents lists available at ScienceDirect
Journal of the Neurological Sciences journal homepage: www.elsevier.com/locate/jns
Review article
A systematic review of closure versus medical therapy for preventing recurrent stroke in patients with patent foramen ovale and cryptogenic stroke or transient ischemic attack Linkun Chen 1, Shuguang Luo ⁎, Lingwan Yan, Weijia Zhao Department of Neurology, First Affiliated Hospital of Guangxi Medical University, China
a r t i c l e
i n f o
Article history: Received 20 July 2013 Received in revised form 24 October 2013 Accepted 18 November 2013 Available online 26 November 2013 Keywords: Systematic review Patent foramen ovale Cryptogenic stroke Transcatheter closure Medical therapy Recurrent stroke
a b s t r a c t The optimal treatment for secondary prevention in patients who have a patent foramen ovale (PFO) and history of cryptogenic stroke is still uncertain and controversial. In view of this, we performed a systematic review of randomized controlled trials (RCTs) to investigate whether PFO closure was superior to medical therapy for prevention of recurrent stroke or transient ischemic attack (TIA) in patients with PFO after cryptogenic stroke. We searched the Cochrane Central Register of Controlled Trials, Embase, PubMed, Web of Science, and ClinicalTrials.gov. Three randomized controlled trials with a total of 2303 patients were included and analyzed. A fixed-effect model was used by Review Manager 5.2 (RevMan 5.2) software. The pooled risk ratio (RR) of recurrent stroke or TIA was 0.70, with 95% confidence interval (CI) = 0.47 to 1.04, p = 0.08. The results were similar in the incidence of death and adverse events, and the pooled RR was 0.92 (95% CI = 0.34 to 2.45, p = 0.86) and 1.08 (95% CI = 0.93 to 1.26, p = 0.32), respectively. The data of this systematic review did not show superiority of closure over medical therapy for secondary prevention after cryptogenic stroke. Due to some limitations of the included studies, more randomized controlled trials are needed for further investigation regarding this field. © 2013 Elsevier B.V. All rights reserved.
Contents 1. 2.
Introduction . . . . . . . . . . . . . . . . . . Materials and methods . . . . . . . . . . . . . 2.1. Search strategy . . . . . . . . . . . . . 2.2. Inclusion criteria . . . . . . . . . . . . 2.3. Types of outcome measures . . . . . . . 2.4. Data extraction and statistical analysis . . 3. Results . . . . . . . . . . . . . . . . . . . . 3.1. Description of studies . . . . . . . . . . 3.2. Primary outcome of recurrent stroke or TIA 3.3. Secondary outcome of death . . . . . . . 3.4. Secondary outcome of adverse events . . . 3.5. Subgroup analysis of primary end point . . 4. Discussion . . . . . . . . . . . . . . . . . . Conflict of interest . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
3 4 4 4 4 4 4 4 5 5 6 6 6 7 7
1. Introduction ⁎ Corresponding author at: Department of Neurology, First Affiliated Hospital of Guangxi Medical University, Shuangyong Road 6, Nanning, Guangxi, China. Tel.: +86 13977182980. E-mail address: [email protected] (S. Luo). 1 First author. 0022-510X/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jns.2013.11.027
The etiology of ischemic stroke remains unknown in approximately 40% of stroke patients despite an extensive diagnostic evaluation, and such strokes are classified as cryptogenic stroke [1]. Oval foramen is an interatrial communication that serves to shunt blood from the right
4
L. Chen et al. / Journal of the Neurological Sciences 337 (2014) 3–7
atrium directly into the left atrium before birth. The foramen usually closes as the left-sided pressure rises and is typically sealed within the first year of life. However, it still exists in about 24% of healthy adults and 38% of patients with cryptogenic stroke [2]. It is widely accepted that venous thrombus can gain access to the left atrium through an oval foramen [3]. Thus, paradoxical embolism has been implied as the most likely mechanism of cryptogenic stroke related to PFO, particularly in younger patients. An embolus may enter the systemic circulation from the right to the left atrium via the PFO and cause blockage of a cerebral artery [4]. Therefore, therapeutic measures for secondary prevention are aimed at inhibiting thrombus formation, or occluding the path of the paradoxical embolism [5]. At present, the treatments for preventing recurrent stroke or TIA include medical therapy with antiplatelet agents or anticoagulants, transcatheter PFO closure, and open surgical repair. With the rapid development of interventional technologies, transcatheter device closure has become a feasible and relatively safe treatment option, with high implant success rate and low incidence of device-related complications reported in several observational studies [6,7], and primary surgical repair is rarely advocated. At present, the predominant treatments include PFO closure and medical therapy. A significant controversy surrounds the optimal strategy for treatment of cryptogenic stroke or TIA and concomitant PFO. Transcatheter device closure is superior to medical therapy in the prevention of recurrent stroke or TIA according to the conclusions of several meta-analyses [7–10]. However, all reviews relate to nonrandom trials, and it is inappropriate to answer the questions about therapeutic efficacy due to the imbalance of baseline characteristics, follow-up, and outcome assessment. Therefore, it is necessary to review all relevant randomized controlled trials on this issue to make an objective comparison of the safety and efficacy of the two treatments. 2. Materials and methods 2.1. Search strategy We searched the Cochrane Central Register of Controlled Trials, Embase, PubMed, Web of Science, and ClinicalTrials.gov, and selected potentially relevant studies through a manual search of references from all eligible studies and review articles. If necessary, we contacted authors to obtain additional unpublished data. When the same patient population was included in several publications, only the most recent or complete study would be included.
concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting and other sources of bias. We graded these items as having high, low, or unclear risk. All statistical analyses were performed using the RevMan 5.2 program. We calculated a weighted treatment effect using the Mantel– Haenszel method, and expressed results as risk ratio, 95% confidence interval, and p-value. Heterogeneity assumption was assessed by Chi2 test and I2 test. The heterogeneity was considered insignificant when p N 0.1 or I2 b 50%, and the treatment effect of each study was calculated by the fixed-effect model. Otherwise, the random effect model was utilized (http://handbook.cochrane.org). 3. Results 3.1. Description of studies The study selection process is detailed in Fig. 1. Ultimately, three randomized controlled trials [11–13] with 2303 patients in total were included, and several trials that had registered in ClinicalTrials.gov were still in progress. All the included articles that were published in The New England Journal of Medicine from 2012 to 2013 were prospective, multicenter, randomized, controlled trials. They provided a comparison of transcatheter PFO closure with medical therapy in patients with PFO and a history of cryptogenic stroke or TIA. These trials were carried out in Europe, the United States of America, Canada, Brazil, and Australia. Patients were included if they were between 18 and 60 years of age, had a cryptogenic stroke or TIA and evidence of a patent foramen ovale. PFO was diagnosed by transesophageal echocardiography with a bubble study. Exclusion criterion was any identified potential cause of stroke or TIA other than the patent foramen ovale. Patients from two groups accepted the PFO closure operation and medical therapy. The therapeutic regimen included an antiplatelet agent, anticoagulant or both of them. The follow-up period was from two to eight years. The end-point events were adjudicated by an independent, expert clinical events committee using a blind method. The distribution of patients in different study groups, along with part of their baseline characteristics, is demonstrated in Table 1. There were no significant differences between the two groups with respect to medical history,
2.2. Inclusion criteria Studies that met the following criteria were included: (1) randomized controlled trials; (2) comparing transcatheter device closure with medical therapy for prevention of recurrent stroke or TIA in patients with PFO-related cryptogenic stroke or TIA; (3) reporting the data of recurrent stroke or TIA, death, and adverse events; and (4) follow-up period was 12 months at least. 2.3. Types of outcome measures Primary outcome was defined as recurrent stroke or TIA during the follow-up period. Secondary outcomes were specified as follows: 1. Death from any cause after randomization. 2. Adverse events directly related to the device, procedure, or medical therapy during the followup period. 2.4. Data extraction and statistical analysis Two authors independently assessed each paper for relevance, eligibility and quality. We independently extracted data from each eligible trial, and assessed each study for risk of bias using the Cochrane Collaboration's tool including random sequence generation, allocation
Fig. 1. Flow diagram demonstrating selection of studies.
L. Chen et al. / Journal of the Neurological Sciences 337 (2014) 3–7
5
Table 1 Summary characteristics of studies between two groups. Study Closure group Furlan 2012 Meier 2013 Carroll 2013 Medical group Furlan 2012 Meier 2013 Carroll 2013
N
Age
Male
447 204 499
46.3 ± 9.6 44.3 ± 10.2 45.7 ± 9.7
233 92 268
462 210 481
45.7 ± 9.1 44.6 ± 10.1 46.2 ± 10.0
238 114 268
Smoke
Diabetes
Hypertension
Hypercholesterolemia
ASA
96 52 78
– 5 33
151 49 158
212 50 194
168 47 180
104 47 55
– 6 40
131 58 150
189 62 193
165 51 169
Plus–minus values are means ± SD. N = Number. ASA = Atrial septal aneurysm. There were no significant differences between the two groups in any of the characteristics listed.
prior events, or risk factors for stroke. The assessment of methodological quality for each study is displayed in Fig. 2. Three RCTs failed to demonstrate the superiority of closure over medical therapy in the intentionto-treat analysis. Although in the per-protocol analysis of Furlan et al. [13] and Meier et al. [11] the results were similar, Carroll et al. [12] concluded that closure was superior to medical therapy alone in the prespecified per-protocol and as-treated analyses. 3.2. Primary outcome of recurrent stroke or TIA The primary end point was a composite of different end-point events in three trials, including recurrent stroke, TIA, death, or peripheral embolism. The researchers utilized several methods to analyze the primary outcome, such as intention-to-treat, per-protocol, and as-treated analyses. Thus, it was inappropriate to directly combine the composite. We
selected the common index including recurrent stroke or TIA in the intention-to-treat analysis as primary end point. Data of recurrent stroke or TIA were collected from all three studies, which included 1040 patients in the closure group and 1263 patients in the medical therapy group, respectively. In the transcatheter group, there was a trend toward reduction in the incidence of recurrent stroke or TIA compared with medical therapy group, but it did not reach to significant difference (RR = 0.70, 95% CI = 0.47 to 1.04, p = 0.08, Fig. 3). Heterogeneity was low and a fixedeffect model was used (Chi2 = 1.40, p = 0.50, I2 = 0%). 3.3. Secondary outcome of death Three trials reported the data of death involving 2299 patients in total. The pooled RR of death was estimated as 0.92 (95% CI = 0.34 to 2.45, p = 0.86, Fig. 4), and it indicated that the difference between the
Fig. 2. Risk of bias graph: review authors' judgments about each risk of bias item presented as percentages across all included studies.
Fig. 3. Forest plot of comparison: recurrent stroke or TIA.
6
L. Chen et al. / Journal of the Neurological Sciences 337 (2014) 3–7
Fig. 4. Forest plot of comparison: death.
Fig. 5. Forest plot of comparison: adverse events.
two treatments was not significant. The heterogeneity was low (Chi2 = 1.65, p = 0.44, I2 = 0%). However, only Meier et al. reported two deaths in the closure group as the secondary outcome. Two patients from the closure arm and four patients from the medical therapy arm of Furlan et al. died due to reasons other than end point. Three deaths in the closure group and six in the medical therapy group occurred after the early postrandomization period and were adjudicated as not being study related in Carroll et al. 3.4. Secondary outcome of adverse events Adverse event data were available from all three trials with 2299 patients. But they did not report adverse events in a consistent manner, as it was not considered appropriate to quantitatively synthesize each adverse event. Thus we only analyzed the total amount of adverse events. The pooled RR was estimated as 1.08 (95% CI = 0.93 to 1.26, p = 0.32, Fig. 5), and no significant difference was observed between the two treatments. This analysis had low heterogeneity (Chi2 = 0.53, p = 0.77, I2 = 0%). While Carroll et al. [12] emphasized that devicerelated adverse events occurred in 21 out of 499 patients in the closure group (4.2%), which seemed significant, this rarely appeared in the other two trials. 3.5. Subgroup analysis of primary end point The result of Carroll et al. indicated that patients who had an atrial septal aneurysm or substantial residual shunt might benefit from a closure operation. Another two meta-analyses [8,9] suggested that anticoagulants are superior to antiplatelet agents for preventing the recurrence of stroke or TIA. But the conclusions of the other two trials were negative. Thus we conducted several subgroup analyses to explore the impact of patient characteristics that might govern the recurrence of stroke or TIA in the follow-up period. These factors included the presence of an atrial septal aneurysm, age, gender, and residual shunt size. All three trials provided a primary end point subgroup analysis about atrial septal aneurysm. Only two of the three trials analyzed sex, age, shunt size, cardiovascular event, and planned medical regimen. Due to limited space, we did not demonstrate the figures of each subgroup analysis, only listed the total results in Table 2. This suggested that the
data did not support a treatment effect in patients with an atrial septal aneurysm who received PFO closure. The conclusions were similar when analyzed by age, gender, and residual shunt size.
4. Discussion This systematic review compared transcatheter device closure with medical therapy for the prevention of recurrent stroke or TIA in patients with PFO and a history of cryptogenic stroke or TIA. No significant difference was observed between the two treatments with respect to the prevention of recurrent stroke or TIA. Wahl et al. [14] conducted a comparative study on long-term comparison of PFO closure with medical treatment after paradoxical embolism. This prospective cohort study had the longest follow-up duration available to date with a maximum follow-up time of 15 years, and its procedure was similar to the three included trials. They concluded that transcatheter PFO closure was more effective than medical therapy for the prevention of recurrent cerebrovascular events among patients with PFO and a history of cryptogenic stroke or TIA. They analyzed the clinical outcomes of 103 propensity score
Table 2 Results of subgroup analyses. Subgroup
ASA (total) Present Absent Age (total) 18–45 years 45–60 years Gender (total) Male Female Shunt size (total) None–moderate Substantial
Test for overall effect
Heterogeneity
RR
95% CI
p
p
I2
0.72 0.67 0.76 0.59 0.44 0.69 0.75 0.51 1.05 0.70 0.90 0.33
0.49–1.08 0.34–1.31 0.46–1.25 0.32–1.09 0.15–1.27 0.33–1.48 0.48–1.17 0.26–1.01 0.57–1.92 0.44–1.13 0.52–1.56 0.12–0.92
0.11 0.24 0.27 0.09 0.13 0.34 0.20 0.05 0.88 0.14 0.71 0.04
0.27 0.11 0.38 0.39 0.25 0.23 0.37 0.90 0.41 0.09 0.60 0.21
21% 55% 0% 1% 23% 31% 4% 0% 0% 26% 0% 37%
ASA = Atrial septal aneurysm. RR = Risk ratio. CI = Confidence interval.
L. Chen et al. / Journal of the Neurological Sciences 337 (2014) 3–7
(PS)-matched pairs, all 308 patients, and the 43 medically treated patients who crossed over to PFO closure. In all the outcomes, only TIA was significantly different. The improved clinical outcome was mainly attributed to TIA, and it might mean that PFO closure was superior to medical treatment in the prevention of recurrent TIA. Though the investigators analyzed the outcomes with a PS-matched comparison for the sake of balancing the baseline characteristics between the two groups, selection bias might still exist due to lack of randomization. The analysis, which was based on the time after PFO closure compared with the time before PFO closure, might yield bias since PFO closure during the observation period was entirely decided by physicians and patients, and some of the index events (particularly TIA) might not be related to PFO. The above factors might influence the authenticity of the results. The other two meta-analyses [8,9] regarding the same issue suggested that transcatheter PFO closure was superior to medical therapy for the prevention of recurrent neurological events after cryptogenic stroke. They also implied that anticoagulants are superior to antiplatelet agents for preventing stroke recurrence. Nonetheless, the three RCTs did not provide data comparing anticoagulants with antiplatelet agents. We could not conduct such an analysis in this review. All the studies included in the two meta-analyses were single-arm and comparative trials, with different baseline characteristics and risk factors. Systematic pooling of different nonrandom controlled studies with different baseline characteristics might induce serious imprecision in the results due to heterogeneity. Several studies included in the reviews had a small number of patients, short follow-up period, and various biases. These factors might strongly influence the validity of the conclusion. The included trials of this review had some biases as well. None of the three trials used the blind method in patients and investigators, and the allocation concealment of Furlan et al. was unclear. These factors could induce selection and performance biases that might exaggerate curative effect. Though all the included trials had a relatively high dropout rate after follow-up, intention-to-treat analysis was used to avoid attrition bias. Other biases displayed in Fig. 2 were low. There were some differences between the three trials with respect to study design, the population included, the device tested, and the follow-up period. First, the primary end point of Furlan et al. and Meier et al. was a composite of recurrent stroke and TIA, but the primary end point of Carroll et al. was only recurrent stroke. TIA was a less clear-cut end point than stroke. Including TIA as a component could not only increase the event rate but might also dilute the effects [11]. Second, there were two types of closure device used in the three trials. Furlan et al. performed the transcatheter PFO closure with a StarFlex Occluder, while the device used in the other two trials was an Amplatzer Occluder. Transcatheter closure of PFO with the Amplatzer was quicker and provided a higher rate of complete occlusion and a lower rate of complication than with the StarFlex device [15]. The procedural success rate of Furlan et al. was 86.1%, obviously lower than the other two trials (96.1%, 95.9%). Furthermore, the therapeutic regimen of the medical therapy group was different with diverse medicines, courses or both. Anticoagulants might be superior to antiplatelet agents for preventing stroke recurrence. These factors might influence the treatment effects. Third, the follow-up period of Furlan et al. was two years, which was obviously shorter than the other two trials. Recurrent events were more likely to be observed in the longer follow-up period. The abovementioned differences might increase the heterogeneity. But we found
7
that there was no significant heterogeneity between the three studies after analyzing by RevMan 5.2. In conclusion, although the included trials had several limitations, they could still be considered to be excellent studies with high methodological quality. Therefore, there were reasons to believe that our conclusion might be more credible. Since all three randomized trials failed to prove which therapy was more effective, PFO closure was at least competitive to medical treatment and it should be offered to the patients as a choice between a simple once-in-a-lifetime operation and lifelong medical therapy that might increase bleeding risk. In addition, only three trials were included, most of the participants were white and there was a lack of other races. Therefore, more randomized controlled trials about this issue are needed in the future to investigate which treatment is superior and which specific subgroups of patients are most likely to benefit from closure for the prevention of recurrent stroke or TIA.
Conflict of interest There is no financial support. The authors declare no conflicts of interest.
References [1] Windecker S, Wahl A, Nedeltchev K, Arnold M, Schwerzmann M, Seiler C, et al. Comparison of medical treatment with percutaneous closure of patent foramen ovale in patients with cryptogenic stroke. JACC 2004;44:750–8. [2] Mattle HP, Meier B, Nedeltchev K. Prevention of stroke in patients with patent foramen ovale. Int J Stroke 2010;5:92–102. [3] Thaler DE, Wahl A. Critique of closure or medical therapy for cryptogenic stroke with patent foramen ovale: the hole truth? Stroke 2012;43:3147–9. [4] Almekhlafi MA, Wilton SB, Rabi DM, Ghali WA, Lorenzetti DL, Hill MD. Recurrent cerebral ischemia in medically treated patent foramen ovale: a meta-analysis. Neurology 2009;73:89–97. [5] Homma S, Sacco RL. Patent foramen ovale and stroke. Circulation 2005;112:1063–72. [6] Wohrle J. Closure of patent foramen ovale after cryptogenic stroke. Lancet 2006;368:350–2. [7] Khairy P, O'Donnell CP, Landzberg MJ. Transcatheter closure versus medical therapy of patent foramen ovale and presumed paradoxical thromboemboli: a systematic review. Ann Intern Med 2003;139:753–60. [8] Kitsios GD, Dahabreh IJ, Abu Dabrh AM, Thaler DE, Kent DM. Patent foramen ovale closure and medical treatments for secondary stroke prevention: a systematic review of observational and randomized evidence. Stroke 2012;43:422–31. [9] Agarwal S, Bajaj NS, Kumbhani DJ, Tuzcu EM, Kapadia SR. Meta-analysis of transcatheter closure versus medical therapy for patent foramen ovale in prevention of recurrent neurological events after presumed paradoxical embolism. JACC Cardiovasc Interv 2012;5:777–89. [10] Fathi AR, Eshtehardi P, Meier B. Patent foramen ovale and neurosurgery in sitting position: a systematic review. Br J Anaesth 2009;102:588–96. [11] Meier B, Kalesan B, Mattle HP, Khattab AA, Hildick-Smith D, Dudek D, et al. Percutaneous closure of patent foramen ovale in cryptogenic embolism. N Engl J Med 2013;368:1083–91. [12] Carroll JD, Saver JL, Thaler DE, Smalling RW, Berry S, MacDonald LA, et al. Closure of patent foramen ovale versus medical therapy after cryptogenic stroke. N Engl J Med 2013;368:1092–100. [13] Furlan AJ, Reisman M, Massaro J, Mauri L, Adams H, Albers GW, et al. Closure or medical therapy for cryptogenic stroke with patent foramen ovale. N Engl J Med 2012;366:991–9. [14] Wahl A, Juni P, Mono ML, Kalesan B, Praz F, Geister L, et al. Long-term propensity score-matched comparison of percutaneous closure of patent foramen ovale with medical treatment after paradoxical embolism. Circulation 2012;125:803–12. [15] Butera G, Carminati M, Chessa M, Delogu A, Drago M, Piazza L, et al. CardioSEAL/ STARflex versus Amplatzer devices for percutaneous closure of small to moderate (up to 18 mm) atrial septal defects. Am Heart J 2004;148:507–10.
Contents lists available at ScienceDirect
Journal of the Neurological Sciences journal homepage: www.elsevier.com/locate/jns
Review article
A systematic review of closure versus medical therapy for preventing recurrent stroke in patients with patent foramen ovale and cryptogenic stroke or transient ischemic attack Linkun Chen 1, Shuguang Luo ⁎, Lingwan Yan, Weijia Zhao Department of Neurology, First Affiliated Hospital of Guangxi Medical University, China
a r t i c l e
i n f o
Article history: Received 20 July 2013 Received in revised form 24 October 2013 Accepted 18 November 2013 Available online 26 November 2013 Keywords: Systematic review Patent foramen ovale Cryptogenic stroke Transcatheter closure Medical therapy Recurrent stroke
a b s t r a c t The optimal treatment for secondary prevention in patients who have a patent foramen ovale (PFO) and history of cryptogenic stroke is still uncertain and controversial. In view of this, we performed a systematic review of randomized controlled trials (RCTs) to investigate whether PFO closure was superior to medical therapy for prevention of recurrent stroke or transient ischemic attack (TIA) in patients with PFO after cryptogenic stroke. We searched the Cochrane Central Register of Controlled Trials, Embase, PubMed, Web of Science, and ClinicalTrials.gov. Three randomized controlled trials with a total of 2303 patients were included and analyzed. A fixed-effect model was used by Review Manager 5.2 (RevMan 5.2) software. The pooled risk ratio (RR) of recurrent stroke or TIA was 0.70, with 95% confidence interval (CI) = 0.47 to 1.04, p = 0.08. The results were similar in the incidence of death and adverse events, and the pooled RR was 0.92 (95% CI = 0.34 to 2.45, p = 0.86) and 1.08 (95% CI = 0.93 to 1.26, p = 0.32), respectively. The data of this systematic review did not show superiority of closure over medical therapy for secondary prevention after cryptogenic stroke. Due to some limitations of the included studies, more randomized controlled trials are needed for further investigation regarding this field. © 2013 Elsevier B.V. All rights reserved.
Contents 1. 2.
Introduction . . . . . . . . . . . . . . . . . . Materials and methods . . . . . . . . . . . . . 2.1. Search strategy . . . . . . . . . . . . . 2.2. Inclusion criteria . . . . . . . . . . . . 2.3. Types of outcome measures . . . . . . . 2.4. Data extraction and statistical analysis . . 3. Results . . . . . . . . . . . . . . . . . . . . 3.1. Description of studies . . . . . . . . . . 3.2. Primary outcome of recurrent stroke or TIA 3.3. Secondary outcome of death . . . . . . . 3.4. Secondary outcome of adverse events . . . 3.5. Subgroup analysis of primary end point . . 4. Discussion . . . . . . . . . . . . . . . . . . Conflict of interest . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
3 4 4 4 4 4 4 4 5 5 6 6 6 7 7
1. Introduction ⁎ Corresponding author at: Department of Neurology, First Affiliated Hospital of Guangxi Medical University, Shuangyong Road 6, Nanning, Guangxi, China. Tel.: +86 13977182980. E-mail address: [email protected] (S. Luo). 1 First author. 0022-510X/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jns.2013.11.027
The etiology of ischemic stroke remains unknown in approximately 40% of stroke patients despite an extensive diagnostic evaluation, and such strokes are classified as cryptogenic stroke [1]. Oval foramen is an interatrial communication that serves to shunt blood from the right
4
L. Chen et al. / Journal of the Neurological Sciences 337 (2014) 3–7
atrium directly into the left atrium before birth. The foramen usually closes as the left-sided pressure rises and is typically sealed within the first year of life. However, it still exists in about 24% of healthy adults and 38% of patients with cryptogenic stroke [2]. It is widely accepted that venous thrombus can gain access to the left atrium through an oval foramen [3]. Thus, paradoxical embolism has been implied as the most likely mechanism of cryptogenic stroke related to PFO, particularly in younger patients. An embolus may enter the systemic circulation from the right to the left atrium via the PFO and cause blockage of a cerebral artery [4]. Therefore, therapeutic measures for secondary prevention are aimed at inhibiting thrombus formation, or occluding the path of the paradoxical embolism [5]. At present, the treatments for preventing recurrent stroke or TIA include medical therapy with antiplatelet agents or anticoagulants, transcatheter PFO closure, and open surgical repair. With the rapid development of interventional technologies, transcatheter device closure has become a feasible and relatively safe treatment option, with high implant success rate and low incidence of device-related complications reported in several observational studies [6,7], and primary surgical repair is rarely advocated. At present, the predominant treatments include PFO closure and medical therapy. A significant controversy surrounds the optimal strategy for treatment of cryptogenic stroke or TIA and concomitant PFO. Transcatheter device closure is superior to medical therapy in the prevention of recurrent stroke or TIA according to the conclusions of several meta-analyses [7–10]. However, all reviews relate to nonrandom trials, and it is inappropriate to answer the questions about therapeutic efficacy due to the imbalance of baseline characteristics, follow-up, and outcome assessment. Therefore, it is necessary to review all relevant randomized controlled trials on this issue to make an objective comparison of the safety and efficacy of the two treatments. 2. Materials and methods 2.1. Search strategy We searched the Cochrane Central Register of Controlled Trials, Embase, PubMed, Web of Science, and ClinicalTrials.gov, and selected potentially relevant studies through a manual search of references from all eligible studies and review articles. If necessary, we contacted authors to obtain additional unpublished data. When the same patient population was included in several publications, only the most recent or complete study would be included.
concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting and other sources of bias. We graded these items as having high, low, or unclear risk. All statistical analyses were performed using the RevMan 5.2 program. We calculated a weighted treatment effect using the Mantel– Haenszel method, and expressed results as risk ratio, 95% confidence interval, and p-value. Heterogeneity assumption was assessed by Chi2 test and I2 test. The heterogeneity was considered insignificant when p N 0.1 or I2 b 50%, and the treatment effect of each study was calculated by the fixed-effect model. Otherwise, the random effect model was utilized (http://handbook.cochrane.org). 3. Results 3.1. Description of studies The study selection process is detailed in Fig. 1. Ultimately, three randomized controlled trials [11–13] with 2303 patients in total were included, and several trials that had registered in ClinicalTrials.gov were still in progress. All the included articles that were published in The New England Journal of Medicine from 2012 to 2013 were prospective, multicenter, randomized, controlled trials. They provided a comparison of transcatheter PFO closure with medical therapy in patients with PFO and a history of cryptogenic stroke or TIA. These trials were carried out in Europe, the United States of America, Canada, Brazil, and Australia. Patients were included if they were between 18 and 60 years of age, had a cryptogenic stroke or TIA and evidence of a patent foramen ovale. PFO was diagnosed by transesophageal echocardiography with a bubble study. Exclusion criterion was any identified potential cause of stroke or TIA other than the patent foramen ovale. Patients from two groups accepted the PFO closure operation and medical therapy. The therapeutic regimen included an antiplatelet agent, anticoagulant or both of them. The follow-up period was from two to eight years. The end-point events were adjudicated by an independent, expert clinical events committee using a blind method. The distribution of patients in different study groups, along with part of their baseline characteristics, is demonstrated in Table 1. There were no significant differences between the two groups with respect to medical history,
2.2. Inclusion criteria Studies that met the following criteria were included: (1) randomized controlled trials; (2) comparing transcatheter device closure with medical therapy for prevention of recurrent stroke or TIA in patients with PFO-related cryptogenic stroke or TIA; (3) reporting the data of recurrent stroke or TIA, death, and adverse events; and (4) follow-up period was 12 months at least. 2.3. Types of outcome measures Primary outcome was defined as recurrent stroke or TIA during the follow-up period. Secondary outcomes were specified as follows: 1. Death from any cause after randomization. 2. Adverse events directly related to the device, procedure, or medical therapy during the followup period. 2.4. Data extraction and statistical analysis Two authors independently assessed each paper for relevance, eligibility and quality. We independently extracted data from each eligible trial, and assessed each study for risk of bias using the Cochrane Collaboration's tool including random sequence generation, allocation
Fig. 1. Flow diagram demonstrating selection of studies.
L. Chen et al. / Journal of the Neurological Sciences 337 (2014) 3–7
5
Table 1 Summary characteristics of studies between two groups. Study Closure group Furlan 2012 Meier 2013 Carroll 2013 Medical group Furlan 2012 Meier 2013 Carroll 2013
N
Age
Male
447 204 499
46.3 ± 9.6 44.3 ± 10.2 45.7 ± 9.7
233 92 268
462 210 481
45.7 ± 9.1 44.6 ± 10.1 46.2 ± 10.0
238 114 268
Smoke
Diabetes
Hypertension
Hypercholesterolemia
ASA
96 52 78
– 5 33
151 49 158
212 50 194
168 47 180
104 47 55
– 6 40
131 58 150
189 62 193
165 51 169
Plus–minus values are means ± SD. N = Number. ASA = Atrial septal aneurysm. There were no significant differences between the two groups in any of the characteristics listed.
prior events, or risk factors for stroke. The assessment of methodological quality for each study is displayed in Fig. 2. Three RCTs failed to demonstrate the superiority of closure over medical therapy in the intentionto-treat analysis. Although in the per-protocol analysis of Furlan et al. [13] and Meier et al. [11] the results were similar, Carroll et al. [12] concluded that closure was superior to medical therapy alone in the prespecified per-protocol and as-treated analyses. 3.2. Primary outcome of recurrent stroke or TIA The primary end point was a composite of different end-point events in three trials, including recurrent stroke, TIA, death, or peripheral embolism. The researchers utilized several methods to analyze the primary outcome, such as intention-to-treat, per-protocol, and as-treated analyses. Thus, it was inappropriate to directly combine the composite. We
selected the common index including recurrent stroke or TIA in the intention-to-treat analysis as primary end point. Data of recurrent stroke or TIA were collected from all three studies, which included 1040 patients in the closure group and 1263 patients in the medical therapy group, respectively. In the transcatheter group, there was a trend toward reduction in the incidence of recurrent stroke or TIA compared with medical therapy group, but it did not reach to significant difference (RR = 0.70, 95% CI = 0.47 to 1.04, p = 0.08, Fig. 3). Heterogeneity was low and a fixedeffect model was used (Chi2 = 1.40, p = 0.50, I2 = 0%). 3.3. Secondary outcome of death Three trials reported the data of death involving 2299 patients in total. The pooled RR of death was estimated as 0.92 (95% CI = 0.34 to 2.45, p = 0.86, Fig. 4), and it indicated that the difference between the
Fig. 2. Risk of bias graph: review authors' judgments about each risk of bias item presented as percentages across all included studies.
Fig. 3. Forest plot of comparison: recurrent stroke or TIA.
6
L. Chen et al. / Journal of the Neurological Sciences 337 (2014) 3–7
Fig. 4. Forest plot of comparison: death.
Fig. 5. Forest plot of comparison: adverse events.
two treatments was not significant. The heterogeneity was low (Chi2 = 1.65, p = 0.44, I2 = 0%). However, only Meier et al. reported two deaths in the closure group as the secondary outcome. Two patients from the closure arm and four patients from the medical therapy arm of Furlan et al. died due to reasons other than end point. Three deaths in the closure group and six in the medical therapy group occurred after the early postrandomization period and were adjudicated as not being study related in Carroll et al. 3.4. Secondary outcome of adverse events Adverse event data were available from all three trials with 2299 patients. But they did not report adverse events in a consistent manner, as it was not considered appropriate to quantitatively synthesize each adverse event. Thus we only analyzed the total amount of adverse events. The pooled RR was estimated as 1.08 (95% CI = 0.93 to 1.26, p = 0.32, Fig. 5), and no significant difference was observed between the two treatments. This analysis had low heterogeneity (Chi2 = 0.53, p = 0.77, I2 = 0%). While Carroll et al. [12] emphasized that devicerelated adverse events occurred in 21 out of 499 patients in the closure group (4.2%), which seemed significant, this rarely appeared in the other two trials. 3.5. Subgroup analysis of primary end point The result of Carroll et al. indicated that patients who had an atrial septal aneurysm or substantial residual shunt might benefit from a closure operation. Another two meta-analyses [8,9] suggested that anticoagulants are superior to antiplatelet agents for preventing the recurrence of stroke or TIA. But the conclusions of the other two trials were negative. Thus we conducted several subgroup analyses to explore the impact of patient characteristics that might govern the recurrence of stroke or TIA in the follow-up period. These factors included the presence of an atrial septal aneurysm, age, gender, and residual shunt size. All three trials provided a primary end point subgroup analysis about atrial septal aneurysm. Only two of the three trials analyzed sex, age, shunt size, cardiovascular event, and planned medical regimen. Due to limited space, we did not demonstrate the figures of each subgroup analysis, only listed the total results in Table 2. This suggested that the
data did not support a treatment effect in patients with an atrial septal aneurysm who received PFO closure. The conclusions were similar when analyzed by age, gender, and residual shunt size.
4. Discussion This systematic review compared transcatheter device closure with medical therapy for the prevention of recurrent stroke or TIA in patients with PFO and a history of cryptogenic stroke or TIA. No significant difference was observed between the two treatments with respect to the prevention of recurrent stroke or TIA. Wahl et al. [14] conducted a comparative study on long-term comparison of PFO closure with medical treatment after paradoxical embolism. This prospective cohort study had the longest follow-up duration available to date with a maximum follow-up time of 15 years, and its procedure was similar to the three included trials. They concluded that transcatheter PFO closure was more effective than medical therapy for the prevention of recurrent cerebrovascular events among patients with PFO and a history of cryptogenic stroke or TIA. They analyzed the clinical outcomes of 103 propensity score
Table 2 Results of subgroup analyses. Subgroup
ASA (total) Present Absent Age (total) 18–45 years 45–60 years Gender (total) Male Female Shunt size (total) None–moderate Substantial
Test for overall effect
Heterogeneity
RR
95% CI
p
p
I2
0.72 0.67 0.76 0.59 0.44 0.69 0.75 0.51 1.05 0.70 0.90 0.33
0.49–1.08 0.34–1.31 0.46–1.25 0.32–1.09 0.15–1.27 0.33–1.48 0.48–1.17 0.26–1.01 0.57–1.92 0.44–1.13 0.52–1.56 0.12–0.92
0.11 0.24 0.27 0.09 0.13 0.34 0.20 0.05 0.88 0.14 0.71 0.04
0.27 0.11 0.38 0.39 0.25 0.23 0.37 0.90 0.41 0.09 0.60 0.21
21% 55% 0% 1% 23% 31% 4% 0% 0% 26% 0% 37%
ASA = Atrial septal aneurysm. RR = Risk ratio. CI = Confidence interval.
L. Chen et al. / Journal of the Neurological Sciences 337 (2014) 3–7
(PS)-matched pairs, all 308 patients, and the 43 medically treated patients who crossed over to PFO closure. In all the outcomes, only TIA was significantly different. The improved clinical outcome was mainly attributed to TIA, and it might mean that PFO closure was superior to medical treatment in the prevention of recurrent TIA. Though the investigators analyzed the outcomes with a PS-matched comparison for the sake of balancing the baseline characteristics between the two groups, selection bias might still exist due to lack of randomization. The analysis, which was based on the time after PFO closure compared with the time before PFO closure, might yield bias since PFO closure during the observation period was entirely decided by physicians and patients, and some of the index events (particularly TIA) might not be related to PFO. The above factors might influence the authenticity of the results. The other two meta-analyses [8,9] regarding the same issue suggested that transcatheter PFO closure was superior to medical therapy for the prevention of recurrent neurological events after cryptogenic stroke. They also implied that anticoagulants are superior to antiplatelet agents for preventing stroke recurrence. Nonetheless, the three RCTs did not provide data comparing anticoagulants with antiplatelet agents. We could not conduct such an analysis in this review. All the studies included in the two meta-analyses were single-arm and comparative trials, with different baseline characteristics and risk factors. Systematic pooling of different nonrandom controlled studies with different baseline characteristics might induce serious imprecision in the results due to heterogeneity. Several studies included in the reviews had a small number of patients, short follow-up period, and various biases. These factors might strongly influence the validity of the conclusion. The included trials of this review had some biases as well. None of the three trials used the blind method in patients and investigators, and the allocation concealment of Furlan et al. was unclear. These factors could induce selection and performance biases that might exaggerate curative effect. Though all the included trials had a relatively high dropout rate after follow-up, intention-to-treat analysis was used to avoid attrition bias. Other biases displayed in Fig. 2 were low. There were some differences between the three trials with respect to study design, the population included, the device tested, and the follow-up period. First, the primary end point of Furlan et al. and Meier et al. was a composite of recurrent stroke and TIA, but the primary end point of Carroll et al. was only recurrent stroke. TIA was a less clear-cut end point than stroke. Including TIA as a component could not only increase the event rate but might also dilute the effects [11]. Second, there were two types of closure device used in the three trials. Furlan et al. performed the transcatheter PFO closure with a StarFlex Occluder, while the device used in the other two trials was an Amplatzer Occluder. Transcatheter closure of PFO with the Amplatzer was quicker and provided a higher rate of complete occlusion and a lower rate of complication than with the StarFlex device [15]. The procedural success rate of Furlan et al. was 86.1%, obviously lower than the other two trials (96.1%, 95.9%). Furthermore, the therapeutic regimen of the medical therapy group was different with diverse medicines, courses or both. Anticoagulants might be superior to antiplatelet agents for preventing stroke recurrence. These factors might influence the treatment effects. Third, the follow-up period of Furlan et al. was two years, which was obviously shorter than the other two trials. Recurrent events were more likely to be observed in the longer follow-up period. The abovementioned differences might increase the heterogeneity. But we found
7
that there was no significant heterogeneity between the three studies after analyzing by RevMan 5.2. In conclusion, although the included trials had several limitations, they could still be considered to be excellent studies with high methodological quality. Therefore, there were reasons to believe that our conclusion might be more credible. Since all three randomized trials failed to prove which therapy was more effective, PFO closure was at least competitive to medical treatment and it should be offered to the patients as a choice between a simple once-in-a-lifetime operation and lifelong medical therapy that might increase bleeding risk. In addition, only three trials were included, most of the participants were white and there was a lack of other races. Therefore, more randomized controlled trials about this issue are needed in the future to investigate which treatment is superior and which specific subgroups of patients are most likely to benefit from closure for the prevention of recurrent stroke or TIA.
Conflict of interest There is no financial support. The authors declare no conflicts of interest.
References [1] Windecker S, Wahl A, Nedeltchev K, Arnold M, Schwerzmann M, Seiler C, et al. Comparison of medical treatment with percutaneous closure of patent foramen ovale in patients with cryptogenic stroke. JACC 2004;44:750–8. [2] Mattle HP, Meier B, Nedeltchev K. Prevention of stroke in patients with patent foramen ovale. Int J Stroke 2010;5:92–102. [3] Thaler DE, Wahl A. Critique of closure or medical therapy for cryptogenic stroke with patent foramen ovale: the hole truth? Stroke 2012;43:3147–9. [4] Almekhlafi MA, Wilton SB, Rabi DM, Ghali WA, Lorenzetti DL, Hill MD. Recurrent cerebral ischemia in medically treated patent foramen ovale: a meta-analysis. Neurology 2009;73:89–97. [5] Homma S, Sacco RL. Patent foramen ovale and stroke. Circulation 2005;112:1063–72. [6] Wohrle J. Closure of patent foramen ovale after cryptogenic stroke. Lancet 2006;368:350–2. [7] Khairy P, O'Donnell CP, Landzberg MJ. Transcatheter closure versus medical therapy of patent foramen ovale and presumed paradoxical thromboemboli: a systematic review. Ann Intern Med 2003;139:753–60. [8] Kitsios GD, Dahabreh IJ, Abu Dabrh AM, Thaler DE, Kent DM. Patent foramen ovale closure and medical treatments for secondary stroke prevention: a systematic review of observational and randomized evidence. Stroke 2012;43:422–31. [9] Agarwal S, Bajaj NS, Kumbhani DJ, Tuzcu EM, Kapadia SR. Meta-analysis of transcatheter closure versus medical therapy for patent foramen ovale in prevention of recurrent neurological events after presumed paradoxical embolism. JACC Cardiovasc Interv 2012;5:777–89. [10] Fathi AR, Eshtehardi P, Meier B. Patent foramen ovale and neurosurgery in sitting position: a systematic review. Br J Anaesth 2009;102:588–96. [11] Meier B, Kalesan B, Mattle HP, Khattab AA, Hildick-Smith D, Dudek D, et al. Percutaneous closure of patent foramen ovale in cryptogenic embolism. N Engl J Med 2013;368:1083–91. [12] Carroll JD, Saver JL, Thaler DE, Smalling RW, Berry S, MacDonald LA, et al. Closure of patent foramen ovale versus medical therapy after cryptogenic stroke. N Engl J Med 2013;368:1092–100. [13] Furlan AJ, Reisman M, Massaro J, Mauri L, Adams H, Albers GW, et al. Closure or medical therapy for cryptogenic stroke with patent foramen ovale. N Engl J Med 2012;366:991–9. [14] Wahl A, Juni P, Mono ML, Kalesan B, Praz F, Geister L, et al. Long-term propensity score-matched comparison of percutaneous closure of patent foramen ovale with medical treatment after paradoxical embolism. Circulation 2012;125:803–12. [15] Butera G, Carminati M, Chessa M, Delogu A, Drago M, Piazza L, et al. CardioSEAL/ STARflex versus Amplatzer devices for percutaneous closure of small to moderate (up to 18 mm) atrial septal defects. Am Heart J 2004;148:507–10.