Background: Patent foramen ovale closure represents a potential secondary prevention strategy for cryptogenic stroke, but available trials have varied by size, device studied, and follow-up. Methods: We conducted a systematic search of published randomized clinical trials evaluating patent foramen ovale closure versus medical therapy in patients with recent stroke or transient ischemic attack using PubMED, EMBASE, and Cochrane through September 2017. Weighting was by random effects models. Results: Of 480 studies screened, we included 5 randomized clinical trials in the meta-analysis in which 3440 patients were randomized to patent foramen ovale closure (n = 1829) or medical therapy (n = 1611) and followed for an average of 2.0 to 5.9 years. Index stroke/transient ischemic attack occurred within 6 to 9 months of randomization. The primary end point was composite stroke/transient ischemic attack and death (in 3 trials) or stroke alone (in 2 trials). Patent foramen ovale closure reduced the primary end point (0.70 vs 1.48 events per 100 patient-years; risk ratio [RR], 0.52 [0.29-0.91]; I 2 = 55.0%) and stroke/transient ischemic attack (1.04 vs 2.00 events per 100 patient-years; RR, 0.55 [0.37-0.82]; I 2 = 42.2%) with modest heterogeneity compared with medical therapy. Procedural bleeding was not different between study arms (1.8% vs 1.8%; RR, 0.94 [0.49-1.83]; I 2 = 29.2%), but new-onset atrial fibrillation/flutter was increased with patent foramen ovale closure (6.6% vs 0.7%; RR, 4.69 [2.17-10.12]; I 2 = 29.3%). Conclusions: In patients with recent cryptogenic stroke, patent foramen ovale closure reduces recurrent stroke/transient ischemic attack compared with medical therapy, but is associated with a higher risk of new-onset atrial fibrillation/flutter.