Who benefits most from device PFO closure after a stroke?

It has been well established that device closure has, on average, prevented stroke recurrence in people who’ve had patent foramen ovale–associated stroke, but a meta-analysis has drilled down into clinical trials to advance a potentially practice-changing principle: that, while device closure shows an overall benefit, not all patients derive a benefit and some may actually be harmed by the procedure.

What’s more, the researchers developed a scoring system that helps determine which patients are likely to benefit from device closure.

“What was unknown was how to treat individual patients because the decision to close the patent foramen ovale (PFO) is still preference sensitive because the risk of a recurrent stroke is low, and most of the strokes that recur are not terribly severe,” lead study author David M. Kent, MD, MS, said in an interview.

“On top of this,” he said, “it was still suspected that some of the PFOs, even in trials of well-selected patients, may not be causally related to stroke; the stroke may still have another occult cause, such as paroxysmal atrial fibrillation or aortic arch atheroma.” Dr. Kent is a professor of medicine at Tufts University in Boston and director of the Predictive Analytics and Comparative Effectiveness Center there.

The meta-analysis, conducted by the Systematic, Collaborative, PFO Closure Evaluation (SCOPE) consortium, analyzed data from six randomized clinical trials that compared device closure and medical therapy to medical therapy alone in 3,740 patients who had PFO-associated stroke from 2000 to 2017. It was published in JAMA.

Overall, the rate of recurrent ischemic stroke was less than half that in patients who had device closure, compared with those who were on medical therapy: 0.47% (n = 39 of 1,889) vs. 1.09% (n = 82 of 1,851).

The researchers also applied two tools designed to calculate the probability of recurrent stroke in individual patients: Risk of Paradoxical Embolism (RoPE), an index that assigns a score of 0-10 to stratify cryptogenic stroke patients with PFO by the likelihood that the stroke was associated with their PFO; and the PFO-Associated Stroke Causal Likelihood (PASCAL) classification system, which integrates the RoPE score with physiological and anatomical features – namely, the size of the PFO shunt and the presence of an atrial septal aneurysm.

“We came up with a way to more accurately identify those patients who are likely to get the most benefit from PFO closure based on mathematic modeling that estimates an individual’s probability that the PFO is causally related to the stroke,” Dr. Kent said.
 

Multivariate analysis determines risk

The study used a multivariate classification system that Dr. Kent had been developing to perform subgroup analyses of the clinical trials. It assigned patients to three different risk groups based on the likelihood that the PFO was causally related to their stroke: PASCAL categories of unlikely, possible, and probable.

The PASCAL unlikely group had a risk of stroke recurrence in the first 2 years of 3.4% (95% confidence interval, 1.1%-5.7%) if they were on medical therapy, and 4.1% (95% CI, 1.7%-6.4%) if they had device closure. In the PASCAL possible group, those risks were 3.6% (95% CI, 2.4%-4.9%) and 1.5% (95% CI, 0.7-2.3%), respectively. For the probable group, device closure represents “a near perfect therapy” with a 90% risk reduction, Dr. Kent said. “Moreover,” he said, “adverse events of device closure, such as atrial fibrillation, appear to be concentrated in those patients who fall into the unlikely classification, who appear to get no benefit.”

The ideal patient for device closure is age 60 years or younger and without vascular risk factors such as hypertension, diabetes, a history of smoking, or a prior stroke, but has high-risk PFO features such as a large shunt or atrial septal aneurysm, Dr. Kent said.

“We think these findings should be practice changing now,” Dr. Kent said.

Faisal M. Merchant, MD, director of cardiac electrophysiology at Emory Healthcare in Atlanta, concurred with that statement. “This is in my mind probably as good as any data we’re going to get on this,” he said in an interview. “The results support what’s been a general gestalt in the clinical world, but [also] really provide an evidence base on how to make decisions.”

He noted that guidelines, including those of the American Academy of Neurology, recommend medical therapy or device closure to prevent recurrent stroke in people who’ve had PFO-associated ischemic stroke. “But they hedge a bit,” he said of the guidelines. “We haven’t had data that’s as robust as this. I think this really solidifies those recommendations.”

He also credited the “unique” study design to extract findings from clinical trials and apply them to personalized medicine. “Clinical trial results give you an average treatment effect of the patients included, but who are ones who really benefit? Who are the ones that don’t benefit? Who are the ones who are harmed?” Dr. Merchant said. “It’s rare that you can parse out this nicely between the people who both benefit and are less likely to be harmed and the people who don’t benefit and are more likely to be harmed.”

The study received funding from the Patient-Centered Outcomes Research Institute. Dr. Kent disclosed relationships with PCORI, W.L. Gore and the Canadian Stroke Consortium. Dr. Merchant has no relevant disclosures.

It has been well established that device closure has, on average, prevented stroke recurrence in people who’ve had patent foramen ovale–associated stroke, but a meta-analysis has drilled down into clinical trials to advance a potentially practice-changing principle: that, while device closure shows an overall benefit, not all patients derive a benefit and some may actually be harmed by the procedure.

What’s more, the researchers developed a scoring system that helps determine which patients are likely to benefit from device closure.

“What was unknown was how to treat individual patients because the decision to close the patent foramen ovale (PFO) is still preference sensitive because the risk of a recurrent stroke is low, and most of the strokes that recur are not terribly severe,” lead study author David M. Kent, MD, MS, said in an interview.

“On top of this,” he said, “it was still suspected that some of the PFOs, even in trials of well-selected patients, may not be causally related to stroke; the stroke may still have another occult cause, such as paroxysmal atrial fibrillation or aortic arch atheroma.” Dr. Kent is a professor of medicine at Tufts University in Boston and director of the Predictive Analytics and Comparative Effectiveness Center there.

The meta-analysis, conducted by the Systematic, Collaborative, PFO Closure Evaluation (SCOPE) consortium, analyzed data from six randomized clinical trials that compared device closure and medical therapy to medical therapy alone in 3,740 patients who had PFO-associated stroke from 2000 to 2017. It was published in JAMA.

Overall, the rate of recurrent ischemic stroke was less than half that in patients who had device closure, compared with those who were on medical therapy: 0.47% (n = 39 of 1,889) vs. 1.09% (n = 82 of 1,851).

The researchers also applied two tools designed to calculate the probability of recurrent stroke in individual patients: Risk of Paradoxical Embolism (RoPE), an index that assigns a score of 0-10 to stratify cryptogenic stroke patients with PFO by the likelihood that the stroke was associated with their PFO; and the PFO-Associated Stroke Causal Likelihood (PASCAL) classification system, which integrates the RoPE score with physiological and anatomical features – namely, the size of the PFO shunt and the presence of an atrial septal aneurysm.

“We came up with a way to more accurately identify those patients who are likely to get the most benefit from PFO closure based on mathematic modeling that estimates an individual’s probability that the PFO is causally related to the stroke,” Dr. Kent said.
 

Multivariate analysis determines risk

The study used a multivariate classification system that Dr. Kent had been developing to perform subgroup analyses of the clinical trials. It assigned patients to three different risk groups based on the likelihood that the PFO was causally related to their stroke: PASCAL categories of unlikely, possible, and probable.

The PASCAL unlikely group had a risk of stroke recurrence in the first 2 years of 3.4% (95% confidence interval, 1.1%-5.7%) if they were on medical therapy, and 4.1% (95% CI, 1.7%-6.4%) if they had device closure. In the PASCAL possible group, those risks were 3.6% (95% CI, 2.4%-4.9%) and 1.5% (95% CI, 0.7-2.3%), respectively. For the probable group, device closure represents “a near perfect therapy” with a 90% risk reduction, Dr. Kent said. “Moreover,” he said, “adverse events of device closure, such as atrial fibrillation, appear to be concentrated in those patients who fall into the unlikely classification, who appear to get no benefit.”

The ideal patient for device closure is age 60 years or younger and without vascular risk factors such as hypertension, diabetes, a history of smoking, or a prior stroke, but has high-risk PFO features such as a large shunt or atrial septal aneurysm, Dr. Kent said.

“We think these findings should be practice changing now,” Dr. Kent said.

Faisal M. Merchant, MD, director of cardiac electrophysiology at Emory Healthcare in Atlanta, concurred with that statement. “This is in my mind probably as good as any data we’re going to get on this,” he said in an interview. “The results support what’s been a general gestalt in the clinical world, but [also] really provide an evidence base on how to make decisions.”

He noted that guidelines, including those of the American Academy of Neurology, recommend medical therapy or device closure to prevent recurrent stroke in people who’ve had PFO-associated ischemic stroke. “But they hedge a bit,” he said of the guidelines. “We haven’t had data that’s as robust as this. I think this really solidifies those recommendations.”

He also credited the “unique” study design to extract findings from clinical trials and apply them to personalized medicine. “Clinical trial results give you an average treatment effect of the patients included, but who are ones who really benefit? Who are the ones that don’t benefit? Who are the ones who are harmed?” Dr. Merchant said. “It’s rare that you can parse out this nicely between the people who both benefit and are less likely to be harmed and the people who don’t benefit and are more likely to be harmed.”

The study received funding from the Patient-Centered Outcomes Research Institute. Dr. Kent disclosed relationships with PCORI, W.L. Gore and the Canadian Stroke Consortium. Dr. Merchant has no relevant disclosures.

It has been well established that device closure has, on average, prevented stroke recurrence in people who’ve had patent foramen ovale–associated stroke, but a meta-analysis has drilled down into clinical trials to advance a potentially practice-changing principle: that, while device closure shows an overall benefit, not all patients derive a benefit and some may actually be harmed by the procedure.

What’s more, the researchers developed a scoring system that helps determine which patients are likely to benefit from device closure.

“What was unknown was how to treat individual patients because the decision to close the patent foramen ovale (PFO) is still preference sensitive because the risk of a recurrent stroke is low, and most of the strokes that recur are not terribly severe,” lead study author David M. Kent, MD, MS, said in an interview.

“On top of this,” he said, “it was still suspected that some of the PFOs, even in trials of well-selected patients, may not be causally related to stroke; the stroke may still have another occult cause, such as paroxysmal atrial fibrillation or aortic arch atheroma.” Dr. Kent is a professor of medicine at Tufts University in Boston and director of the Predictive Analytics and Comparative Effectiveness Center there.

The meta-analysis, conducted by the Systematic, Collaborative, PFO Closure Evaluation (SCOPE) consortium, analyzed data from six randomized clinical trials that compared device closure and medical therapy to medical therapy alone in 3,740 patients who had PFO-associated stroke from 2000 to 2017. It was published in JAMA.

Overall, the rate of recurrent ischemic stroke was less than half that in patients who had device closure, compared with those who were on medical therapy: 0.47% (n = 39 of 1,889) vs. 1.09% (n = 82 of 1,851).

The researchers also applied two tools designed to calculate the probability of recurrent stroke in individual patients: Risk of Paradoxical Embolism (RoPE), an index that assigns a score of 0-10 to stratify cryptogenic stroke patients with PFO by the likelihood that the stroke was associated with their PFO; and the PFO-Associated Stroke Causal Likelihood (PASCAL) classification system, which integrates the RoPE score with physiological and anatomical features – namely, the size of the PFO shunt and the presence of an atrial septal aneurysm.

“We came up with a way to more accurately identify those patients who are likely to get the most benefit from PFO closure based on mathematic modeling that estimates an individual’s probability that the PFO is causally related to the stroke,” Dr. Kent said.
 

Multivariate analysis determines risk

The study used a multivariate classification system that Dr. Kent had been developing to perform subgroup analyses of the clinical trials. It assigned patients to three different risk groups based on the likelihood that the PFO was causally related to their stroke: PASCAL categories of unlikely, possible, and probable.

The PASCAL unlikely group had a risk of stroke recurrence in the first 2 years of 3.4% (95% confidence interval, 1.1%-5.7%) if they were on medical therapy, and 4.1% (95% CI, 1.7%-6.4%) if they had device closure. In the PASCAL possible group, those risks were 3.6% (95% CI, 2.4%-4.9%) and 1.5% (95% CI, 0.7-2.3%), respectively. For the probable group, device closure represents “a near perfect therapy” with a 90% risk reduction, Dr. Kent said. “Moreover,” he said, “adverse events of device closure, such as atrial fibrillation, appear to be concentrated in those patients who fall into the unlikely classification, who appear to get no benefit.”

The ideal patient for device closure is age 60 years or younger and without vascular risk factors such as hypertension, diabetes, a history of smoking, or a prior stroke, but has high-risk PFO features such as a large shunt or atrial septal aneurysm, Dr. Kent said.

“We think these findings should be practice changing now,” Dr. Kent said.

Faisal M. Merchant, MD, director of cardiac electrophysiology at Emory Healthcare in Atlanta, concurred with that statement. “This is in my mind probably as good as any data we’re going to get on this,” he said in an interview. “The results support what’s been a general gestalt in the clinical world, but [also] really provide an evidence base on how to make decisions.”

He noted that guidelines, including those of the American Academy of Neurology, recommend medical therapy or device closure to prevent recurrent stroke in people who’ve had PFO-associated ischemic stroke. “But they hedge a bit,” he said of the guidelines. “We haven’t had data that’s as robust as this. I think this really solidifies those recommendations.”

He also credited the “unique” study design to extract findings from clinical trials and apply them to personalized medicine. “Clinical trial results give you an average treatment effect of the patients included, but who are ones who really benefit? Who are the ones that don’t benefit? Who are the ones who are harmed?” Dr. Merchant said. “It’s rare that you can parse out this nicely between the people who both benefit and are less likely to be harmed and the people who don’t benefit and are more likely to be harmed.”

The study received funding from the Patient-Centered Outcomes Research Institute. Dr. Kent disclosed relationships with PCORI, W.L. Gore and the Canadian Stroke Consortium. Dr. Merchant has no relevant disclosures.