PFO closure for stroke prevention: how does the evidence add up?

Three RCTs had failed to show benefit, so a meta-analysis was done and recently published:

Results

Three RCTs met inclusion criteria. The pooled data provided 2,303 patients, of which 1,150 were in the PFO closure group and 1,153 in the medical therapy group. In the ITT analysis, there were 43 events (3.7%) of the composite end point in the closure group compared with 61 events (5.3%) in the medical therapy group, with a trend in favor of the PFO closure (OR = 0.70; 95% CI, 0.47–1.05, P = 0.08). The incidences of TIA, ischemic CVA, and bleeding were not statistically different between the groups. There was a trend for the more frequent occurrence of atrial fibrillation in the PFO closure group (OR = 3.29; 95% CI, 0.86–12.60, P = 0.08). In the AT analysis, the composite end point was significantly less frequent in the PFO closure group (OR = 0.62; 95% CI, 0.41–0.94, P = 0.02).

Here is a discussion of the article at Medpage Today.

This will be spun various ways. My summation is that while the evidence is inconclusive and no one can make a high level claim for closure over medical therapy the as treated analysis suggests that patients who successfully completed device therapy did better. The true practice of evidence based medicine in this situation will require a nuanced conversation with the patient. Statements like “we need to fix this hole in your heart to keep you from having another stroke” are unfounded. Dismissive statements like “PFO closure doesn't work” only reflect one's misunderstanding of the proper application of EBM.

The prevention of stroke in patients with PFO is, to say the least, controversial. It would seem plausible for device therapy to be superior to medical treatment and for systemic anticoagulation over antiplatelet therapy though neither has been proven convincingly. The UptoDate authors “suggest” medical over device therapy and “suggest” antiplatelet therapy over anticoagulation unless there is a specific reason to favor anticoagulation in the individual patient.