As someone who performs a good number of PFO closures in those believed to be at high-risk of recurrent cryptogenic stroke despite medical therapy, I was certainly interested in hearing preliminary results from the CLOSURE trial released a few weeks ago. Initiated in 2003, during the fledgling days of closure technology and expertise, the study randomized patients with cryptogenic stroke and PFO to closure using the StarFlex device (NMT Medical) versus best medical therapy, the latter comprising full dose aspirin or warfarin at physician discretion. The study failed to meet its primary endpoint, a reduction in stroke or TIA at 2 years, although the device “provided a small, but not statistically significant, benefit over current best medical therapy.” Perusing what limited information has been released, I’m left with several thoughts pertaining to case selection, safety, procedural success, and efficacy. Case Selection: Since the trial began, it has become increasingly clear that not all PFO are high risk for recurrent events. Those with concomitant inter-atrial septal aneurysm, long tunnel, Eustachian valve, and/or resting large shunting appear to be cohorts at elevated risk. Since CLOSURE included all PFO, regardless of co-existent high risk features, it is possible that event rates were too low to show significant differences between arms. It will therefore be interesting to see what percent of patients had high-risk anatomic features. Safety: The press release stated the device exhibited “an excellent safety profile,” and as such I am assuming there was no difference between arms from a safety standpoint. This, of course, is reassuring given the fact that the trial included very early experience, when the learning curve for PFO closure was still rather steep. Procedural Success: As it has been shown that residual shunting after PFO closure increases risk of recurrent events, I was struck by the roughly 85% complete closure rate seen in this trial. Other devices, namely those from AGA Medical, have evidenced superior rates of complete closure > 95%, which my own experience mirrors. Efficacy: My reading of the observational data to date is that primarily high-risk anatomic cohorts benefit from closure, due to the high rates of recurrence despite aspirin. And, in these patients, warfarin is almost certainly superior to anti-platelet therapy, while closure is possibly superior (or at least non-inferior) to warfarin. Thus, trials that randomize to a medical arm that includes either warfarin or aspirin, such as CLOSURE, might find it hard to show a benefit to closure, especially if a significant portion received warfarin. In addition, one would expect the higher complete closure rates of the Amplatzer (AGA Medical) device to be more efficacious, and thereby more likely to show a significant difference compared to either medical therapy. Taken together and placing everything in practical terms, I continue to perform PFO closure in patients with cryptogenic stroke and one or more high-risk anatomic features, after an exhaustive neurologic and cardiac evaluation, unless they are willing to take long-term warfarin. For many patients, however, the desire to avoid warfarin and its accompanying lifestyle effects is significant. For those with low risk anatomy, I recommend aspirin (or other anti-platelet therapy), and feel comfortable doing so. Finally, for those with recurrent events despite medical therapy (anti-platelet or warfarin), PFO closure appears not only reasonable but indicated. Dr. Srihari S. Naidu is Director of the Cardiac Catheterization Laboratory, Interventional Cardiology Fellowship Program and Hypertrophic Cardiomyopathy Center at Winthrop University Hospital on Long Island.