PFO Seen Before Noncardiac Surgery Ups Perioperative Stroke Risk

February 12, 2018

BOSTON — The risk for ischemic stroke during or within 30 days of noncardiac surgery is more than doubled in patients previously diagnosed with a patent foramen ovale (PFO), suggests an observational study.[1]

Of note, the perioperative risk in patients with known PFO went up far more steeply among those who otherwise were at low compared with high stroke risk based on a preoperative score derived from classic stroke risk factors not including PFO.

"We don't know why, but the most likely and plausible explanation is that the ones who have a high risk of stroke are on antiplatelet agents, and the ones who don't have a high stroke risk, they just fall through the cracks," said Dr Matthias Eikermann (Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, MA).

He is senior author on the study, published February 6 in the Journal of the American Medical Association with first author Dr Pauline Y Ng (Massachusetts General Hospital and Harvard Medical School).

The findings have implications for potential antithrombotic therapy or other protective measures before, during, and after noncardiac surgery in patients previously diagnosed with PFO, Eikermann told | Medscape Cardiology.

But available data aren't sufficient to recommend preoperative transcatheter PFO closure to cut their risk for perioperative stroke, he said. "You'd have to close a lot of PFOs in order to decrease the stroke risk substantially."

And the procedure is known to pose risks, as demonstrated in three trials that together appeared in the New England Journal of Medicine on September 14, 2017.

As covered here extensively, the REDUCE and CLOSE primary results and long-term outcomes from the RESPECT trial suggested that transcatheter PFO closure was associated with fewer recurrences in patients with a history of cryptogenic stroke.

Transcatheter PFO closure before such surgeries might make sense in very selected cases with increased risk for PFO-related stroke, according to Dr Matthew Price (Scripps Clinic, La Jolla, CA), who was not involved in the study.

"Screening everyone and closing PFO in everyone does not make sense. However, identifying patients and procedures where that risk is higher would be helpful to determine in which patients invasive therapies would be required," Price said when interviewed.

"We don't know from this study the mechanism of the PFO-related stroke, and there may be strokes related to microemboli which standard prophylaxis would not prevent, and from other sources, such as from the PFO tunnel itself."

This further points to "the need for judicious use of periprocedural pharmacology, to further prevent thrombus formation. But even more so, further studies to determine which types of surgeries place the patient at highest risk are needed so we can target our more aggressive therapies more wisely."

Focus Also on Long-term Risk

Rather than pursuing preventive transcatheter closure in patients with preoperative PFO but no stroke history, Eikermann said, "What I think is a lower hanging fruit is first looking at long-term stroke risk after surgery." It would be up to future trials, he said, to determine in such cases whether long-term antithrombotic therapy may be safe and effective.

An accompanying editorial agrees that the results are not necessarily a call for more transcatheter PFO closure in these patients.[2]

"At this point, clinicians should not interpret the study findings as suggesting the need for diagnostic testing and aggressive treatment of PFO prior to noncardiac surgeries," write Drs Scott E Kasner and Steven R Messe (University of Pennsylvania, Philadelphia).

"Nevertheless, the results indicating increased risk of perioperative stroke among patients with PFO remain alluring given the biological plausibility." That's despite several potential limitations both the editorialists and the authors point out, including possible referral bias in the group found to have preoperative PFO and possible non-PFO causes of perioperative stroke.

The authors acknowledge that the group with known PFO was more likely to have underlying cardiovascular or thromboembolic diseases, which themselves could have been risk factors for stroke, and also "could have biased such patients to the diagnosis of PFO through referral for a dedicated echocardiography study." Those conditions were therefore included as covariates in the adjusted analyses.

Perioperative Stroke, Other Thromboembolic Events

The study looked at 150,198 patients undergoing noncardiac surgery with general anesthesia at a major referral center and two community hospitals over 8 years; their mean age was 55 years. Their overall perioperative stroke rate was 0.6%.

Of the total cohort, 1540, or about 1%, were known before surgery to have a PFO. Their rate of perioperative stroke was 3.2%, compared with 0.5% for those without a prior PFO diagnosis.

In adjusted analyses, patients with PFO had increased risks for perioperative ischemic stroke and other embolic events, compared with patients without PFO.

Table. Adjusteda Odds Ratio for Perioperativeb Outcomes in  Patients With vs Without Preoperatively Diagnosed PFO

Endpoint Odds Ratio (95% CI) P Value
Ischemic stroke 2.66 (1.96–3.63) <.001
Large-vessel-territory stroke 3.14 (2.21–4.48) <.001
Systemic embolic event 2.31 (1.74–3.08) <.001
Acute limb ischemia 2.31 (1.35–3.94) .002
Renal artery embolism 5.15 (1.88–14.13) .001
Acute MI 1.60 (1.13–2.27) .008
aAge, sex, body mass index, smoking history, Charlson Comorbidity Index, diabetes, cardiopulmonary comorbidities, migraine, chronic kidney disease, hypercoagulable state, deep-vein thrombosis, pulmonary and systemic embolic events, recent use of cardiovascular drugs, emergency surgery status, inpatient vs outpatient surgery, high-risk surgical service, surgery duration, intraoperative hypotensive minutes, intraoperative vasopressor dose and fluid volumes, need for packed red blood cells, and work relative value units.

bDuring and within 30 days after surgery.


Supporting evidence came from an analysis that divided the entire cohort into two groups according to perioperative stroke risk as determined by standard preoperative criteria but excluding PFO.

Among the 50% of the study population who were at lower than the median perioperative stroke risk, the adjusted odds ratio (OR) for stroke was 15.92 (95% CI, 4.92–51.53; P<0.001) for those with preoperatively diagnosed PFO compared with those without a known PFO.

That was sharply higher than the adjusted OR of 3.80 (95% CI, 2.81–5.15; P<0.001) among the other 50% who scored at higher than the median perioperative stroke risk.

There were significantly more strokes in large-vessel cerebral territories in the preoperative PFO group (2.4% vs 0.4%), and that group also showed more severe neurologic deficits related to their strokes (P=0.02) according to National Institutes of Health Stroke Scale criteria.

Further Sensitivity Analyses

In an analysis that propensity-matched each patient with a preoperative PFO diagnosis to five patients without a PFO, those with PFO showed an OR for ischemic stroke of 3.16 (95% CI, 2.19–4.51; P<0.001).

Another analysis addressed potential confounding from referral bias: 19.7% of the total combined cohort had been evaluated with echocardiography before the surgery, yet all patients with a preoperative diagnosis of PFO by definition had undergone echocardiography.

The perioperative stroke rates only among the study's patients with preoperative echocardiography confirmed the primary findings of the study, according to the authors. The rates were 2.8% for those with preoperative PFO and 1.2% for those without, for an OR of 1.86 (95% CI, 1.28–2.72; P =0.001).

The association between PFO and perioperative stroke was also confirmed in analyses further limited solely to patients who had undergone the forms of echocardiography considered most sensitive for PFO.

In those groups, the PFO rates probably more closely resembled the rate in the general population, Eikermann pointed out. This part of the study, he said, makes "the most striking argument that we have" that the overall results are not the result of referral bias.

Among patients with preoperative transesophageal echocardiography (TEE), 22.1% were identified as having PFO. The perioperative stroke rates were 4.1% and 1.4% for those with vs without PFO (adjusted OR, 2.06; 95% CI, 1.15–3.68; P=0.02).

The overall rate of PFO was 28.2% for those who had undergone TEE with microbubble contrast. Perioperative strokes occurred in 4.0% and 2.3% of those with vs without PFO (adjusted OR, 1.86; 95% CI, 1.26–2.74; P=0.002).

PFO a Risk Equivalent

To some extent, Eikermann said, previously diagnosed PFO is a perioperative-stroke risk equivalent on par with a history of stroke. The current study's patients with prior PFO but no stroke history showed a lower perioperative stroke rate than "even the control groups" in the REDUCE, CLOSE, and RESPECT trials that entered patients with a history of cryptogenic stroke, he said.

"So we are really looking at a very high risk group, and that group needs more attention." It may be that preventive antithrombotic therapy, or adjustments to the anticoagulation protocol during surgery, he said, will be protective in those with known PFO. All that should be explored in future clinical trials.

Currently, he noted, "If patients have recent stroke and undergo surgery, we now are already liberal with use of antithrombotic drugs, even though there are no guidelines." So for now, clinicians should consider putting patients with previously diagnosed PFO who are slated for surgery, especially those at increased stroke risk by other criteria, "more liberally on antithrombotic agents."

Eikermann discloses receiving grants and consulting fees from Merck; receiving funding from Jeffrey and Judith Buzen; holding equity in Calabash Bioscience. Ng had no disclosures. Kasner discloses receiving grants from WL Gore and Associates and Bayer and consulting fees from Bristol-Myers Squibb, Boehringer Ingelheim, and Daiichi Sankyo. Messe discloses receiving grants from WL Gore and Associates and GlaxoSmithKline and consulting fees from Claret Medical.

Follow Steve Stiles on Twitter: @SteveStiles2. For more from | Medscape Cardiology, follow us on Twitter and Facebook.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: