909 patients were randomized in a 1:1 ratio to either percutaneous PFO closure with the STARFlex Septal Occluder plus antiplatelet therapy with aspirin and clopidogrel, or medical therapy alone with aspirin, warfarin or a combination of both. The primary endpoint was stroke and/or TIA during a 2-year follow up, all-cause mortality within the first 30 days, and death from a neurological cause occurring between 31 days and 2 years. Secondary endpoints included all-cause mortality, stroke, TIA, significant bleeding, and other transient neurologic events of undetermined etiology.

Of 909 patients, 447 were randomly assigned to undergo percutaneous closure with the STARFlex device and 462 were randomized to the medical therapy group (Fig. 22.2). Of the patients who were assigned to treatment, 402 underwent attempted implantation of the STARFlex device and 458 received medical therapy (Fig. 22.3). There were no significant differences between patients randomly assigned to the STARFlex device compared to the medical therapy group in terms of their baseline demographics and co-morbidities.

The results were initially presented at the American Heart Association scientific sessions 2010. Contrary to the many observational studies that suggested a benefit of PFO closure for the prevention of recurrent cryptogenic stroke, the results of CLOSURE I did not show any significant differences for the endpoint of stroke within the 2-year follow up (2.9 % for PFO closure group and 3.1 % for medical therapy group in the intention-to-treat population, p = 0.79; 3.2 % for PFO closure group and 3.5 % for medical therapy group in the per-protocol population; p = 0.80) or TIA (3.1 % for PFO closure group and 4.1 % for medical therapy group in the intention-to-treat population, p = 0.44; 3.2 % for PFO closure group and 4.6 % for medical therapy group in the per-protocol population, p = 0.31). In the intention-to-treat population (defined as all randomized patients who received a study treatment and who had no major inclusion/exclusion criteria violations, regardless of length of follow-up), the collective primary endpoint of stroke, TIA, or death within 2-year follow-up occurred in 5.5 % of patients in the Closure group and 6.8 % of patients in the Medical Therapy group (p = 0.37). In the per-protocol population (defined as all randomized patients who received the treatment to which they were randomized, who had no major inclusion/exclusion criteria violations, and who had a follow-up of at least 22 months), there was similarly no statistical difference in the collective endpoint of stroke, TIA, or death between the PFO closure group (5.8 %) and the medical therapy group (7.7 %) within 2-year follow up (p = 0.28). The results were not significant even for the per-protocol groups that actually received the implant or correct medical therapy (Tables 22.1 and 22.2). Major vascular procedural complications occurred only in the Closure group: in 13 of 402 (3.2 %) patients compared to none of 458 (0 %) in the Medical Therapy group (p < 0.001). Of the 13 patients who had a major vascular event in the Closure group, 4 had a hematoma larger than 5 cm in diameter at the access site, 3 needed a procedure-related transfusion, 3 had a retroperitoneal hemorrhage, 1 had a perforation of the left atrium, 1 needed a vascular surgical repair, and 1 suffered from a peripheral-nerve injury. In addition, during follow up, atrial fibrillation occurred at a much higher rate in the Closure group (23/402 [5.7 %]) compared to the Medical Therapy group (3/458 [0.7 %]) (p < 0.001). Of the 23 incidents of atrial fibrillation in the Closure group, 14 were periprocedural.

When assessed by shunt size at baseline and the presence of an atrial septal aneurysm, there was also no advantage with using the STARFlex device in preventing the primary endpoint (Table 22.3).