Meta-analysis Gives Edge to PFO Closure for Stroke Protection

Download this article's Factoid in PDF (& PPT for Gold Subscribers)

Contrary to the results of the randomized CLOSURE I trial, a meta-analysis of observational studies suggests that in patients with stroke of unknown origin and patent foramen ovale (PFO), percutaneous closure of the defect is more effective in preventing future neurological events than medical management alone. The findings were published in the July 2012 issue of JACC: Cardiovascular Interventions.

Researchers led by Samir R. Kapadia, MD, of the Cleveland Clinic (Cleveland, OH), evaluated data from 48 observational studies, including 39 reporting recurrent neurological events after transcatheter PFO closure (n = 8,185) and 19 after medical therapy (n = 2,142). Among them, 10 studies (n = 1,886) directly compared the 2 approaches.

Comparative, Single-Arm Studies Agree

After adjustment, the pooled incidence of neurological events in PFO closure patients was 0.8 per 100 patient-years compared with 5.0 in medical therapy patients—an 84% reduction. Similarly, the comparative studies also showed a clear advantage for closure (RR 0.25; 95% CI 0.11-0.58; P = 0.002).

Furthermore, in studies that focused on medical therapy, treatment with anticoagulants appeared superior to antiplatelet therapy, yielding an estimated rate of 2.2 neurological events per 100 patient-years vs. 4.3 for patients given antiplatelet medications (RR 0.58; 95% CI 0.41-0.82).

Transcatheter closure showed a high rate of procedural success (> 99%), and reintervention over follow-up was estimated at 0.9%. However, the overall complication rate was higher with percutaneous closure than with medical management at 4.1 vs. 0.4 events per 100 patient-years. The most common closure complication was atrial arrhythmia (3.9%). Bleeding complications were seen in 1.7% of transcatheter patients and 1.1% of medical therapy patients.

Factors That Might Affect Outcomes

The investigators also assessed how various clinical and morphological factors influenced outcomes with the different treatment strategies.

For example, the presence of atrial septal aneurysm made little difference for patients undergoing percutaneous closure but markedly increased risk for those receiving medical therapy by 0.03 (95% CI 0.01-0.05). Similarly, while an increasing proportion of patients with a large pretreatment shunt did not affect outcomes after PFO closure, it correlated with higher rates of stroke or TIA in the medical therapy arm.

Again, there was no difference in event recurrence between younger (≤ 60 years) and older subgroups for closure patients, while older age increased risk for those treated with drugs alone (RR 3.27; 95% CI1.48-7.22). Residual shunting—which declined from 25.4% post procedure to 6.3% long-term—did not predispose patients to increased events. In fact, patients who experienced a stroke or TIA were less likely to have such shunting, which the authors say suggests many recurrent events are unrelated to the PFO.

The authors conclude that “meticulous attention to patient selection with detailed history and workup to eliminate other etiologies of stroke and maximizing the likelihood of [paradoxical thromboembolism] as an etiology for stroke by detailed evaluation of PFO morphology and thrombophilia can help identify patients at risk of [recurrent neurological events] secondary to [paradoxical thromboembolism] through PFO.”

Observational Data Untrustworthy

In a telephone interview with TCTMD, Robert J. Sommer, MD, of Columbia University Medical Center (New York, NY), said that “the results are so disparate between closure and medical therapy here that there is certainly some suggestion that closure may be a valid therapy to pursue.” But as the study’s conclusion stands in “stark contrast” with the CLOSURE I trial, he added, “Right now, we’re in a sort of quagmire.”

“Like all meta-analyses, this one is fraught with the danger of making generalizations that may not be real because the inclusion and exclusion criteria or the protocols [of the studies analyzed] are not the same,” he explained.

David E. Thaler, MD, of Tufts Medical Center (Boston, MA), was more blunt. He told TCTMD in a telephone interview that while Dr. Kapadia and colleagues “pay lip service” to the limitations of observational studies, the researchers imply that these are overcome by the sheer abundance of data.

But Dr. Thaler countered that a recent meta-analysis containing most of the same observational studies and showing a comparable advantage for PFO closure, unearthed several sources of bias in the literature (Kitsios GD. Stroke. 2012;43:422-431).

For example, stroke was the index event in 76% of medical therapy studies but 63% of closure studies, putting the treatment groups at different risk for recurrent events from the outset, Dr. Thaler reported. Also, a validated screening method for recurrent stroke was used in 88% of medical studies compared with only 47% of closure studies, suggesting that events may have been missed in the latter.

“The bottom line is that there are strong biases in case series,” said Dr. Thaler, who coauthored the earlier paper. “The difference in outcome [between the PFO strategies] could have been predicted just by the way the studies were set up.”

Zeroing in on ‘Guilty’ PFOs

Dr. Thaler commented that the first step toward useful comparison is determining whether a patient’s PFO was instrumental in causing the event or simply an innocent bystander. According to a model developed by the RoPE (Risk of Paradoxical Embolism) study, the PFO is likely pathogenic in patients who are younger, do not have vascular disease, and had their index stroke in a particular location, he explained, whereas the PFO is probably incidental in older patients with multiple stroke risk factors such as diabetes and hypertension.

The next step is to identify predictors of recurrence among patients likely to have a pathogenic PFO, Dr. Thaler said. Although attractive candidates include atrial septal aneurysm, hypermobile septum, shunting at rest, and thrombophilia, he noted, none has yet been pinned down.

“I think paradoxical embolism is real and PFO closure makes sense,” he said. “But you have to identify the right patients and use the right devices, and not cause problems. And even then [paradoxical embolization] is a low-risk stroke mechanism.”

Will Closure Get RESPECT?

For the time being, CLOSURE I is a touchstone, Dr. Sommer said. The trial “definitively confirmed the opinion of people who were against PFO closure in the first place,” but at the same time it had enough shortcomings to not change the minds of believers, he commented, adding that now, however, most interventionalists who perform the procedure tend to be “much more circumspect in the patients they elect to close.”

Meanwhile, Dr. Sommer said, “everybody is waiting with bated breath” for the results of the randomized RESPECT trial, which will be unveiled at the annual Transcatheter Cardiovascular Therapeutics symposium this October. “If that trial is negative, I think PFO closure as we know it currently will go away,” he concluded.

 

Source:

Agarwal S, Bajaj NS, Kumbhani DJ, et al. Meta-analysis of transcatheter closure versus medical therapy for patent foramen ovale in prevention of recurrent neurological events after presumed paradoxical embolism. J Am Coll Cardiol. 2012:5:777-789.

Related Stories: 

Meta-analysis Gives Edge to PFO Closure for Stroke Protection

Contrary to the results of the randomized CLOSURE I trial, a meta analysis of observational studies suggests that in patients with stroke of unknown origin and patent foramen ovale (PFO), percutaneous closure of the defect is more effective in preventing
Disclosures
  • Dr. Kapadia reports no relevant conflicts of interest.
  • Dr. Thaler reports serving on the steering committee for the RESPECT trial, being a principal investigator for the REDUCE trial, and serving as a consultant for WL Gore.
  • Dr. Sommer reports serving as a physician trainer for AGA Medical and WL Gore as well as on the medical advisory board for Coherex.

Comments