The optimal therapy for preventing recurrent stroke in people with cryptogenic stroke and patent foramen ovale (PFO) has not been defined. The choice between medical therapy (antithrombotic treatment with antiplatelet agents or anticoagulants) and transcatheter device closure has been the subject of intense debate over the past several years. Despite the lack of scientific evidence, a substantial number of people undergo transcatheter device closure (TDC) for secondary stroke prevention.
To: 1) compare the safety and efficacy of TDC with best medical therapy alone for preventing recurrent stroke (fatal or non-fatal) or transient ischemic attacks (TIAs) in people with PFO and a history of cryptogenic stroke or TIA; 2) identify specific subgroups of people most likely to benefit from closure for secondary prevention; and 3) assess the cost-effectiveness of this strategy, if possible.
We searched the Cochrane Stroke Group Trials Register (July 2014), the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library Issue 2, 2014), MEDLINE (1950 to July 2014) and EMBASE (1980 to July 2014). In an effort to identify unpublished and ongoing trials we searched seven trials registers and checked reference lists.
We included randomized controlled trials (RCTs), irrespective of blinding, publication status, and language, comparing the safety and efficacy of device closure with medical therapy for preventing recurrent stroke or TIA in people with PFO and a history of cryptogenic stroke or TIA.
Two review authors independently selected trials for inclusion, assessed quality and risk of bias, and extracted data. The primary outcome measures of this analysis were the composite endpoint of ischemic stroke or TIA events as well as recurrent fatal or non-fatal ischemic stroke. Secondary endpoints included all-cause mortality, serious adverse events (atrial fibrillation, myocardial infarction, bleeding) and procedural success and effective closure. We used the Mantel-Haenszel method to obtain pooled risk ratios (RRs) using the random-effects model regardless of the level of heterogeneity. We pooled data for the primary outcome measure with the generic inverse variance method using the random-effects model, yielding risk estimates as pooled hazard ratio (HR), which accounts for time-to-event outcomes.
We included three RCTs involving a total of 2303 participants: 1150 participants were randomized to receive TDC and 1153 participants were randomized to receive medical therapy. Overall, the risk of bias was regarded as high. The mean follow-up period of all three included trials was less than five years. Baseline characteristics (age, sex, and vascular risk factors) were similar across trials. Intention-to-treat analyses did not show a statistically significant risk reduction in the composite endpoint of recurrent stroke or TIA in the TDC group when compared with medical therapy (RR 0.73, 95% CI 0.45 to 1.17). A time-to-event analysis combining the results of two RCTs also failed to show a significant risk reduction with TDC (HR 0.69, 95% CI 0.43 to 1.13). When assessing stroke prevention alone, TDC still did not show a statistically significant benefit (RR 0.61, 95% CI 0.29 to 1.27) (HR 0.55, 95% CI 0.26 to 1.18). In a sensitivity analysis including the two studies using the Amplatzer PFO occluder, TDC showed a possible protective effect on recurrent stroke compared with medical therapy (HR 0.38, 95% CI 0.14 to 1.02); however, it did not reach statistical significance. Safety analysis found that the overall risks for all-cause mortality and adverse events were similar in both the TDC and medical therapy groups. However, TDC increased the risk of new-onset atrial fibrillation (RR 3.50, 95% CI 1.47 to 8.35) and may be associated with the type of device used.
The combined data from recent RCTs have shown no statistically significant differences between TDC and medical therapy in the prevention of recurrent ischemic stroke. TDC closure was associated with an increased risk of atrial fibrillation but not with serious adverse events.