We aimed to explore whether interventional closure of patent foramen ovale (PFO) results in reduction of composite outcome [stroke/transitory ischemic attack (TIA), death, and thrombolysis in myocardial infarction-TIMI bleeding], stroke and stroke/TIA compared to medical treatment in patients with cryptogenic stroke. Searching the PUBMED and Cochrane library database, we performed meta-analysis from all randomized controlled studies that compared effects of interventional PFO closure with medical treatment on stroke prevention. 3560 patients from six randomized trials were included. Interventional PFO closure reduced composite outcome (RR of 0.47, 0.26-0.85, p = 0.01), stroke (RR of 0.38, 0.18-0.82, p = 0.01) and stroke/TIA (RR of 0.56, 0.43-0.74, p < 0.0001). Analysis had 70.5% power to detect observed reduction of RR for the primary outcome, 70.6% for stroke and 98.7% for stroke/TIA. Bleeding rates were comparable (RR of 0.91, 0.60-1.38, p = 0.66), while there was higher burden of new AF (RR of 5.54, 3-10.2, p < 0.0001) after interventional closure. Subgroup analysis revealed that patients with large shunts had substantial less recurrent strokes over patients with small shunts (p for interaction = 0.02). Use of Amplatzer PFO device was associated with substantial less AF (RR of 2.36, p = 0.06) compared with other devices (RR of 8.93, p < 0.0001) (p for interaction = 0.04), with comparable benefit for stroke prevention (p for interaction = 0.73). Interventional closure of PFO resulted in significant reduction of stroke and stroke/TIA compared with antiplatelets/anticoagulants with comparable bleeding rates between the groups, whereas AF occurred more frequently in the intervention group. Patients with large shunts had more benefit from interventional closure.