A patent foramen ovale (PFO) is a remnant interatrial communication, best diagnosed with transoesophageal echocardiography (TOE) and bubble study. Although quite common and often asymptomatic, PFO is associated with cryptogenic stroke and migraine. Approximately one-half of patients with a cryptogenic stroke have a PFO, and the dilemma regarding whether or not to proceed with percutaneous device closure, to reduce the risk of future recurrent events due to paradoxical embolism, has been subject to debate for nearly two decades. Despite promising observational data, initial randomised clinical trials failed to demonstrate superiority of closure over medical therapy. However, long-term follow-up data from one of these early trials, combined with two new randomised trials, have provided more evidence for the benefits of closure in selected patients. This new evidence suggests that younger patients with high-risk features such as an atrial septal aneurysm (ASA) or large interatrial shunt are more likely to benefit from PFO closure, after fastidious exclusion of an alternative cause for the index stroke. However, issues which require further clarification include whether anticoagulant therapy is preferable to antiplatelet therapy for medical management, and which particular type of closure device is optimal. Finally, despite promising retrospective observational data suggesting improvement in migraine attacks after PFO closure, high quality evidence is lacking in this regard.