Net clinical benefit of patent foramen ovale closure in patients with cryptogenic stroke: Meta-analysis and meta-regression of randomized trials

Abstract Background Controlled randomized trials (CRTs) comparing the efficacy of patent foramen ovale (PFO) closure and medical therapy in patients with cryptogenic stroke have yielded heterogeneous results. No data are available on the net clinical benefit with the two strategies. Methods We pooled data of 3440 patients enrolled in five CRTs, randomized to PFO closure ( n  = 1829) or medical therapy ( n  = 1611) and followed for a mean of 4.1 years. Results The net composite endpoint of stroke, major bleeding or atrial fibrillation (AF)/flutter was not different among PFO closure and medical therapy (OR 1.06; 95% CI 0.63–1.77; p = 0.83). PFO closure was associated with similar bleeding rates and with a significant 59% relative reduction of recurrent stroke versus medical therapy; in the intervention group this stroke prevention was counterbalanced by a significant 4.7-fold higher risk of AF/flutter. Meta-regression analysis showed that odds ratios for the net composite endpoint were related to prevalence of severe shunt at baseline (p = 0.002), percentage of procedural success (p = 0.002), stroke incidence in the medical therapy arm (p = 0.012) and to follow-up duration (p = 0.001). Conclusions This study-level meta-analysis of CRTs demonstrates that, compared to medical therapy, PFO closure prevents recurrent ischemic cerebral events, but increases the risk of AF/flutter in patients with cryptogenic stroke; as a result, the net clinical benefit with the two strategies was similar. Our results support an individualized therapeutic approach, tailored on the evaluation of the patient's risks (anatomical PFO risk, clinical risk of recurrent stroke, bleeding risk, and risk of AF).