Coming to Grips with CLOSURE I: PFO Device Plus Medical Therapy Not Superior to Medical Therapy Alone

SAN FRANCISCO, CALIF.—CLOSURE I—the first completed, prospective, randomized study of a patent foramen ovale closure device—failed to demonstrate a significant benefit of closure in patients with a PFO and cryptogenic stroke or transient ischemic attack.

Study researcher, Mark Reisman, MD, of Swedish Heart and Vascular Institute, Seattle, Wash., cited several reasons for the trial’s failure, including its inability to identify the “optimal” patient population due to insufficient specificity.


The device was not superior to medical therapy for any endpoint; the rate of stroke was comparable between the two arms (2.9% for the device vs. 3.1% for medical therapy alone; P=.79; see Table) and closure with the device (Starflex, NMT Medical) was associated with a higher incidence of atrial fibrillation (5.7% vs. 0.7%; P<.001).

The trial included 909 patients aged 60 years or younger with cryptogenic stroke or TIA and PFO documented by transesophageal echocardiography within 6 months of randomization. Patients were treated with either PFO closure with the Starflex device plus 325 mg aspirin (device arm) or with anticoagulation and/or 325 mg aspirin (medical therapy arm). The primary endpoints were 2-year incidence of stroke or TIA, all-cause mortality for the first 30 days and neurologic mortality from 31 days to 2 years.

“Alternative explanations unrelated to paradoxical embolism presented in 80% of patients with a current stroke,” Reisman said. “What is the failure mode or analysis? Too much overlap with other types of strokes; I clearly believe that.”

In addition, Reisman said the trial was not specific enough to identify the “optimal” patient population appropriate for this procedure. He also cited the timing of the strokes, noting that many of them occurred earlier than he had anticipated. Other challenging areas included the onset of AF in the device group (most were periprocedural); TIA diagnosis; and the ubiquitous nature of the second stroke.

Questions remain as to whether the 2-year timeframe was too short, and what caused the alternative explanation for stroke: “Was the first stroke not really related to the PFO? Was this a suboptimal device?” Reisman asked. “Was the patient selection not optimal? Were we too ambitious? Were these insignificant trends basically driven by transient ischemic attack? And should we not have included transient ischemic attack as an endpoint? Some of the earlier discussions were about AF not being diagnosed early on; maybe this is the cause of the original stroke leading to the second, and maybe we just need a clearer indication of high-risk patients.”

Despite the questions left unanswered and the hypotheses regarding the failure mode, Reisman said the cardiology community should accept the results of the CLOSURE I trial: currently, PFO closure is not necessary in this patient population.

  • Dr. Reisman reports serving as a principal investigator for NMT Medical, and as an advisor for and having equity in Coaptus Medical, and serving on the advisory board for Boston Scientific.