Baseline characteristics were similar between the two groups, except age which was younger in the closure group (47.7 ± 10.8 vs 51.9 ± 9.9, P = 0.013), and the presence of shunt at rest was more common in the closure group (35.8% vs 10.4%, P = 0.000). Procedural success was 94.0%. Over a mean follow-up of 27.8 months, a total of six primary end point, all of which were strokes, occurred only in the medication group (6.6% vs 0%, P = 0.039). Stroke-free survival rate was significantly higher in the closure group (P = 0.026) CONCLUSIONS: Our study showed that PFO closure may be an effective treatment strategy to prevent recurrent stroke or TIA for patients with CS if it is conducted in selective patients who have PFO shunt more than moderate grade.