Patent Foramen Ovale, Subclinical Cerebrovascular Disease and Ischemic Stroke in a Population-Based Cohort

Abstract

Objectives

To evaluate the relationship between patent foramen ovale (PFO), ischemic stroke and subclinical cerebrovascular disease in the general population.

Background

PFO is more frequently found in stroke patients than in stroke-free controls. However, the PFO-related stroke risk in the general population is not well established, and the relationship between PFO and silent brain infarcts (SBI) is not known.

Methods

PFO presence was assessed by transthoracic echocardiography with saline contrast injection in 1,100 stroke-free individuals over age 39 of a community-based sample followed for a mean of 11 years. In addition, 360 participants underwent brain magnetic resonance imaging (MRI) for SBI detection. We evaluated the risk of stroke associated with PFO after adjusting for established stroke risk factors, and examined the odds of having SBI among those with and without PFO.

Results

A PFO was present in 164 participants (14.9%). Over a mean follow up of 11.0 ± 4.5 years, 111 ischemic strokes occurred (10.1%), 15 (9.2%) in the PFO + and 96 (10.3%) in the PFO− groups. The 12.5 year cumulative risk of stroke was 10.1% (standard error 2.5%) in the PFO+ and 10.4% (standard error 1.1%) in the PFO− group (p=0.46). The adjusted hazard ratio for PFO and stroke was 1.10 (95% confidence interval 0.64–1.91). In the MRI subcohort, PFO was not associated with SBI (adjusted odds ratio 1.15, 95% CI 0.50–2.62).

Conclusions

In this community-based cohort, PFO was not associated with an increased risk of clinical stroke or subclinical cerebrovascular disease.

Introduction

Patent foramen ovale (PFO) is associated with increased risk of stroke of otherwise unknown origin (cryptogenic stroke) (1–3). This association is stronger in younger patients(1–3), but also present in the older (3, 4). The relationship between PFO and stroke has been pronounced in case-control studies, less so in community-based, prospective cohort studies. PFO is present in approximately 25% of the general population(5, 6), but only a minority develop cerebrovascular events. Prospective studies in asymptomatic individuals with incidentally detected PFO have only shown non-significant trends with lower risk estimates(6, 7). However, larger study populations and/or longer follow-up may be needed to fully evaluate the association of PFO and stroke in the general population.

Preventive measures such as antithrombotic treatment(8–10) or transcatheter PFO closure(11–13) are used to decrease stroke recurrence in patients with a PFO. However, the relationship between PFO and stroke risk has controversial aspects. In cryptogenic strokes, PFO may be an “innocent bystander” in one third of cases(14). Also, over 40% of recurrent strokes in them show concurrent etiologies(15).

Subclinical brain infarcts (SBI) are asymptomatic lesions found on brain CT or MRI in 16% to 28% of normal subjects(16–18), in the absence of stroke history. SBI share common risk factors with stroke (17, 19) and increase the risk of future stroke(16, 20), cognitive impairment(21, 22) and dementia(21, 22). The relationship between PFO and SBI has not been defined. As SBI may be multiple, occur over a long period of time, and be up to five times as prevalent as clinical strokes (17), their study may unmask a new dimension to the embolic potential of PFO that would go unnoticed if only clinical strokes were considered. The aims of the present study were to: 1) re-evaluate the relationship between PFO and ischemic stroke in a population-based cohort followed for 11 years; and 2) evaluate the relationship between PFO and SBI in a subsample of the cohort. The relationship of PFO with white matter hyperintensities (WMH), subclinical cerebrovascular lesions that are associated with increased risk of stroke and cognitive impairment(20, 23) and may share a microvascular or embolic etiology, was also evaluated.

Methods

As part of the Northern Manhattan Study (NOMAS), 1,148 stroke-free subjects over age 39 underwent transthoracic echocardiography with saline contrast injection for PFO detection between May 1993 and November 1999, 1,100 (95.9%) of whom had diagnostically adequate tests. They represented a subgroup of the total 3,298 NOMAS participants, as the performance of contrast studies was stopped after November 1999. Study design and methodological details regarding NOMAS have been described previously(24). Briefly, community subjects from Northern Manhattan were eligible if they (1) had never been diagnosed with stroke, (2) were over age 39, and (3) resided in Northern Manhattan for at least 3 months in a household with a telephone. Stroke-free subjects were identified by random digit dialing of published and unpublished telephone numbers. In-person assessment was performed in 68% of responders. Details of enrollment in the PFO study have been previously published(7).

From 2003–2008, NOMAS participants over the age of 55 years with no contraindications were asked to undergo a brain MRI for the detection of subclinical cerebrovascular disease. The present article reports on two groups: 1) Overall cohort: all 1,100 participants; 2) MRI subcohort: 360 stroke-free participants (at the time of the MRI) who had been tested for PFO and underwent brain MRI.

All participants provided informed consent. The study was approved by the Institutional Review Boards of Columbia University Medical Center and the University of Miami.

Diagnostic Evaluation

Data were collected through interview, review of the medical record, physical and neurological examination, in-person measurements, and fasting blood specimens. Arterial hypertension was defined as positive history, antihypertensive treatment, or blood pressure >140/90 mmHg at enrollment. Hypercholesterolemia was defined as total serum cholesterol >240mg/dl or presence of appropriate drug treatment. Diabetes mellitus was defined as fasting glucose >125 mg/dl or presence of oral or insulin treatment. Coronary artery disease (CAD) included history of myocardial infarction or typical angina or history of positive diagnostic test (stress test, coronary angiography) and/or drug treatment. Atrial fibrillation was classified by history (current or past ECG or Holter) or ECG detection at enrollment.

Echocardiography

Transthoracic echocardiography (TTE) was performed according to the recommendations of the American Society of Echocardiography. Contrast injection (aerated saline solution) was used for PFO detection(1, 3, 25), with Valsalva maneuver and coughing to increase sensitivity(25). PFO was considered present if any microbubble appeared in the left-sided chambers within three cardiac cycles from maximum right atrial opacification(25). Atrial septal aneurysm (ASA) was defined as 10 mm or greater protrusion of the septum from the midline into the left or right atrium(25).

MRI

Brain imaging was performed on a 1.5-T MRI system (Philips Medical Systems) according to a previously published protocol(26). SBI were defined as cavitations on the fluid-attenuated inversion recovery sequence of at least 3 mm in size, and distinct from a vessel due to the lack of signal void on T2 sequence, and of equal intensity to cerebrospinal fluid(18). Interobserver agreement for SBI detection was 93.3%(18). WMH analysis was performed as previously described using a custom-designed image analysis package (QUANTA 6.2 on a Sun Mycrosystem Ultra 5 workstation)(18). WMH volume was calculated after correcting for total cranial and log-transformed to achieve a normal distribution for analysis (log-WMHV). All measurements were performed blinded to participant clinical information, including echocardiographic findings.

Follow-up and Outcome Evaluation

Subjects were followed annually by telephone. Information on outcomes was available in 98% at 5 years and in 97.3% at 10 years. Any vascular event or neurological or cardiac symptoms triggered an in-person assessment. Active hospital surveillance of admission and discharge ICD-9 codes was performed. Stroke was defined as first symptomatic stroke occurrence as defined by Trial of Org 10172 in Acute Stroke Treatment (TOAST) criteria(27). Diagnosis of ischemic stroke was determined by two neurologists independently, and NOMAS principal investigators (RLS/MSVE) adjudicated disagreements. Ischemic stroke, fatal and non-fatal, was used as the primary outcome of the present study. Stroke diagnostic subtypes (embolic vs. non-embolic; cryptogenic vs. known cause) were defined according to modified Stroke Data Bank criteria(28).

Statistical Analysis

Differences between proportions were assessed by chi-square or Fisher exact test, differences between mean values by unpaired Student’s t-test.

Univariate logistic regression analysis was used to calculate the unadjusted odds ratios (OR) for the association between PFO and SBI/log-WMHV.

Multivariate logistic regression analysis was used to test the independent association between PFO and SBI or log-WMHV adjusting for age, sex and stroke risk factors. Variables significantly associated with SBI/log-WMHV at univariate analysis were entered in the model. Log-WMHV was categorized in quartiles of the observed distribution, and the upper quartile (log-WMHV Q4) was used in the analyses.

Kaplan-Meier event-free analysis by PFO status was performed for ischemic stroke/combined vascular events. Differences were evaluated by log-rank test. Cox proportional hazards models were used to assess the PFO-related risk of stroke/combined vascular events. Hazard ratios (HR) and 95% confidence intervals (CI) were calculated after adjusting for other stroke risk factors. To test the effect of age (≥ vs. <60 years), sex and race-ethnicity, separate variables were fit in the model. Differences between subgroups were tested using interaction terms.

In all analyses, 0.05 was used as significance level. Statistical analyses were performed by an investigator (ZJ) using SAS software (version 9.2). Plots were generated with S-PLUS software; power assessment utilized PASS software.

Results

Demographics and clinical characteristics of the study population are summarized in Table 1. PFO was detected in 164 of 1,100 participants (14.9%), and in 60 of 360 participants (16.7%) of the MRI subcohort. ASA was observed in 27 (2.5%) and 8 (2.2%), respectively. In the overall cohort, hypertension was significantly less frequent in participants with a PFO (60.4% vs. 68.5%, p=0.04). In the MRI subcohort, no significant differences were observed in demographics and stroke risk factors between subjects with and without PFO. ASA was significantly more frequent among subjects with PFO in both cohorts.

Table 1.

Participant characteristics by PFO status in the overall study cohort and in the MRI subcohort

	Overall Cohort			MRI Subcohort
	PFO + N = 164	PFO − N = 936	P-value*	PFO + N = 60	PFO − N = 300	P-value*
Age (years)	68.4 ± 9.4	68.7 ± 10.1	0.67	65.0 ± 8.7	63.9 ± 8.1	0.36
Male, n (%)	77 (47.0)	383 (40.9)	0.15	31 (51.7)	131 (43.7)	0.26
Race-ethnicity			0.92			0.99
Black	44 (26.8)	237 (25.3)		13 (21.7)	65 (21.7)
Hispanic	78 (47.6)	454 (48.5)		35 (58.3)	174 (58.0)
White	38 (23.2)	228 (24.4)		10 (16.7)	52 (17.3)
Other	4 (2.4)	17 (1.8)		2 (3.3)	9 (3.0)
Hypertension	99 (60.4)	641 (68.5)	0.04	39 (65.0)	229 (76.3)	0.07
Diabetes	30 (18.3)	214 (22.9)	0.19	13 (21.7)	95 (31.7)	0.12
Hypercholesterolemia	68 (41.5)	429 (45.8)	0.30	30 (50%)	181 (60.3)	0.14
Ever smoker	91 (55.5)	497 (53.1)	0.57	29 (48.3)	161 (53.7)	0.45
CAD	15 (9.2)	99 (10.6)	0.58	1 (1.67)	24 (8.0)	0.08
Aspirin use	41 (25.0)	243 (26.1)	0.77	17 (28.3)	68 (22.7)	0.35
ASA	19 (11.6)	8 (0.9)	<0.001	7 (11.7)	1 (0.3)	<0.001
SBI	-	-	-	10 (16.7)	42 (14.0)	0.59
Possibly embolic SBI †	-	-	-	5 (8.5)	16 (5.6)	0.39
Log-WMHV Q4 ‡	-	-	-	16 (26.7)	73 (24.4)	0.71

^*PFO + vs. PFO −

^†SBI in location compatible with an embolic mechanism

^‡Upper quartile of the distribution of white matter hyperintensities volume (log transformed) in the cohort

CAD = Coronary artery disease

ASA = Atrial septal aneurysm

SBI = Silent brain infarcts

CAD = Coronary artery disease

ASA = Atrial septal aneurysm

SBI = Silent brain infarcts

SBI were observed in 52 of 360 subjects (14.4%). Subjects with SBI were significantly older, more frequently male and less frequently Hispanic, and had higher frequencies of hypertension and aspirin use (Table 2). Subjects in the log-WMH Q4 subgroup were older and more often hypertensive (also Table 2). Blacks were more represented in log-WMHV Q4 subgroup, whereas Hispanics were less represented. Age and hypertension were associated with both SBI and log-WMHV Q4, male sex with SBI only (Table 3). Hispanic race-ethnicity was associated with lower risk of SBI compared to Whites, and lower risk of log-WMHV Q4 compared to Blacks.

Table 2.

Characteristics of participants with and without silent brain infarcts (SBI) or white matter hyperintensities in the upper quartile (Log-WMHVQ4)

	SBI + N = 52	SBI− N = 308	P-value	Log-WMH V Q4 N = 89	Log-WMH V Q1–3 N = 270	P-value
Age (years)	68.0 ± 8.0	63.4 ± 8.1	0.0002	69.2 ± 7.0	62.4 ± 7.9	<0.0001
Male, n (%)	32 (61.5)	130 (42.2)	0.01	40 (44.9)	122 (45.2)	0.97
Race-ethnicity*			0.05			0.002
Black	13 (25.0)	65 (21.1)		31 (34.8)	46 (17.0)
Hispanic	22 (42.3)	187		41 (46.7)	168
White	14 (26.9)	(60.7)		16 (18.0)	(62.2)
Other	3 (5.8)	48 (15.6)		1 (1.1)	46 (17.0)
		8 (2.6)			10 (3.7)
Hypertension	46 (88.5)	222 (72.1)	0.01	76 (85.4)	192 (71.1)	0.007
Diabetes	18 (34.6)	90 (29.2)	0.43	25 (28.1)	82 (30.4)	0.68
Hypercholesterolemia	27 (51.9)	184 (59.7)	0.29	45 (50.6)	165 (61.1)	0.08
Ever smoker	24 (46.2)	166 (53.9)	0.30	53 (59.6)	136 (50.4)	0.13
Atrial Fibrillation	2 (3.9)	13 (4.2)	0.90	2 (2.3)	13 (4.8)	0.29
Aspirin use	19 (36.5)	66 (21.4)	0.02	21 (23.6)	63 (23.3)	0.96
PFO	10 (19.2)	50 (16.2)	0.59	16 (18.0)	44 (16.3)	0.71
ASA	2 (3.9)	6 (2.0)	0.39	2 (2.3)	6 (2.2)	0.99
PFO+ASA	1 (1.9)	6 (2.0)	0.99	2 (2.3)	5 (1.9)	0.82

^*p= 0.01 Hispanic vs. White; p= 0.93 Black vs. White; p= 0.99 Black vs. Hispanic

Table 3.

Association of demographics, vascular risk factors and aspirin treatment with SBI or log-WMHV

	SBI	Log-WMHV Q4*
	Unadjusted OR (CI)	Unadjusted OR (CI)
Age (per year)	1.07 (1.03–1.12)	1.12 (1.08–1.17)
Male sex	2.19 (1.20–4.00)	0.99 (0.61–1.60)
Race-ethnicity (vs. White)
Black	0.69 (0.30–1.59)	1.94 (0.94–4.02)
Hispanic	0.40 (0.19–0.85)	0.70 (0.36–1.36)
Other	1.29 (0.30–5.51)	0.29 (0.03–2.43)
SBP (per mmHg)	1.02 (1.00–1.03)	1.00 (0.99–1.02)
DBP (per mmHg)	1.01 (0.98–1.04)	1.00 (0.97–1.02)
Hypertension	2.97 (1.22–7.20)	2.38 (1.25–4.52)
Diabetes	1.28 (0.69–2.39)	0.90 (0.53–1.52)
Cigarette smoking	0.73 (0.41–1.32)	1.45 (0.89–2.36)
HDL cholesterol (per mg)	0.99 (0.97–1.01)	1.02 (1.01–1.04)
LDL cholesterol (per mg)	0.99 (0.98–1.00)	1.00 (0.99–1.01)
Hypercholesterolemia	0.73 (0.40–1.31)	0.65 (0.40–1.05)
Aspirin use	2.11 (1.13–3.95)	1.02 (0.58–1.79)

OR = Odds Ratio

CI = 95% Confidence Interval

^*Upper quartile of the distribution in the cohort

PFO, SBI and WMH

SBI were present in 16.7% of subjects with a PFO and in 14.0% of those without it (p=0.59; Table 1). With sample size 360, the study had 80% power to detect an OR of 2.56 for PFO and SBI at 0.05 significance level. The detectable OR for PFO and log-WMHV Q4 was 2.24.

PFO was not a significant predictor of SBI or log-WMHV Q4 in either unadjusted or adjusted analysis (Table 4). The combination of PFO and ASA was not associated with a significant risk of SBI or log-WMHV Q4 in either unadjusted or adjusted models (also Table 4).

Table 4.

Risk of SBI or log-WMHV Q4 associated with the presence of PFO or PFO+ASA

	SBI	Log-WMHV Q4*
	OR (CI)	OR (CI)
PFO
Model 1	1.23 (0.58–2.61)	1.13 (0.60–2.11)
Model 2	1.15 (0.50–2.62)	0.93 (0.46–1.91)
PFO+ASA
Model 1	1.05 (0.12–8.94)	1.23 (0.23–6.49)
Model 2	0.37 (0.04–3.54)	0.60 (0.10–3.57)

Model 1 = unadjusted

Model 2 = adjusted for age, sex, race-ethnicity, hypertension, time from PFO detection to MRI, aspirin usage (for SBI) and for age, hypertension, HDL cholesterol (for Log-WMHV Q4)

OR = Odds Ratio

CI = 95% Confidence Interval

^*Upper quartile of the distribution in the cohort

PFO and Vascular Outcomes

Based on sample size 1,100, PFO prevalence 14.9% and available follow-up time, the study had 80% power to detect a minimum relative risk of 2.11 for ischemic stroke (10.1% stroke rate) and 1.58 for combined vascular events (26.9% rate of stroke, myocardial infarction and vascular death) at a 0.05 significance level.

Follow-up mean duration was 11.0±4.5 years (range 0.04 to 18.6 years). An ischemic stroke occurred in 111 subjects (10.1%), 15 in PFO+ group (9.2%) and 96 in PFO− group (10.3%). The 12.5 year cumulative risk was 10.1% (standard error 2.5%) in the PFO+ and 10.4% (standard error 1.1%) in the PFO− group (p=0.46). Stroke-free survival PFO did not differ significantly by PFO status (log-rank test 0.92; Figure 1A). The HR of PFO for stroke was 0.97 (95% CI 0.56 to 1.68) in unadjusted analysis and 1.10 (95% CI 0.64 to 1.91) after adjustment (Table 5).

Figure 1. PFO and Vascular Outcomes.

Kaplan-Meier event-free survival curves according to PFO status for ischemic stroke (A) and combined vascular events (B)

Table 5.

Risk of ischemic stroke by PFO/ASA status, in the overall study population and in age, sex and race-ethnicity subgroups

	Ischemic	Stroke	Combined	Vascular Events
	Unadjusted HR (CI)	Adjusted HR* (CI)	Unadjusted HR (CI)	Adjusted HR* (CI)
PFO (regardless of ASA presence)	0.97 (0.56–1.68)	1.10 (0.64–1.91)	1.02 (0.73–1.41)	1.13 (0.81–1.57)
PFO alone	1.06 (0.60–1.86)	1.23 (0.70–2.16)	1.01 (0.71–1.43)	1.19 (0.84–1.69)
PFO+ASA	0.50 (0.07–3.58)	0.48 (0.07–3.50)	1.08 (0.48–2.44)	0.87 (0.38–1.98)
ASA alone	2.77 (0.68–11.25)	2.10 (0.51–8.65)	1.54 (0.49–4.79)	1.09 (0.35–3.42)
PFO and stroke in subgroups
Men	1.23 (0.60–2.55)	1.35 (0.65–2.82)	1.12 (0.71–1.76)	1.20 (0.76–1.89)
Women	0.74 (0.32–1.72)	0.87 (0.38–2.03)	0.90 (0.56–1.45)	1.06 (0.66–1.71)
Age <60 yrs	1.56 (0.47–5.15)	2.08 (0.62–6.91)	1.50 (0.60–3.74)	1.86 (0.74–4.67)
Age ≥ 60 yrs	0.93 (0.51–1.69)	0.99 (0.54–1.81)	1.04 (0.74–1.47)	1.07 (0.76–1.52)
Whites	1.04 (0.36–2.99)	1.10 (0.38–3.19)	1.03 (0.58–1.81)	1.14 (0.65–2.02)
Blacks	0.87 (0.31–2.48)	1.01 (0.35–2.92)	0.86 (0.44–1.67)	0.96 (0.49–1.87)
Hispanics	1.01 (0.45–2.23)	1.20 (0.54–2.66)	0.97 (0.56–1.67)	1.14 (0.66–1.96)

^*adjusted for age, sex, atrial fibrillation, diabetes mellitus, hypertension, hypercholesterolemia and ever smoking

Stroke subtypes did not differ between the PFO+ and PFO− groups. Specifically, cryptogenic stroke was not significantly different (21.4% vs. 17.9%; p = 0.75), nor was embolic stroke (21.4% vs. 34.5%; p= 0.33).

Table 5 also shows the stroke risk associated with ASA, alone or in combination with PFO, and the association of PFO and stroke across age, sex and race-ethnic subgroups. No significant associations were observed in univariate or multivariate analyses. Combined vascular events occurred in 296 subjects (26.9%). PFO, alone or in association with ASA, was not found to be associated with combined vascular events in the overall cohort or in any subgroup (Figure 1B and Table 5).

Discussion

This is the first study that followed a cohort of asymptomatic individuals with and without PFO for an average of 11 years, and evaluated the relationship of PFO not only with the occurrence of stroke or other vascular events, but also with the presence of subclinical brain disease potentially related to PFO. The results do not confirm the hypothesis that PFO increases the risk of cerebrovascular events in the general population.

The role of PFO as a risk factor for ischemic stroke has mainly been demonstrated in case-control studies, with an approximately 4-fold increase in PFO prevalence in stroke patients younger than 55 years and an approximately 2-fold increase in older patients compared with controls of similar age(25). The role of PFO as a risk factor for stroke and vascular events in the general population has never been demonstrated, but inadequate sample sizes and/or follow-up durations might have masked associations. In the Stroke Prevention: Assessment of Risk in a Community (SPARC) study, over a median follow-up of 61 months a non-significant HR of 1.46 (95% CI 0.74–2.88) for PFO and cerebrovascular events was observed in 585 volunteers(6). In the cohort of 1,100 stroke-free participants of the present study, we previously reported a similar, non-significant HR of 1.64 (CI 0.87–3.09) over a mean follow up of over 7 years(7). However, follow-up was relatively short. Also, the possibility existed that PFO might cause mild or asymptomatic cerebrovascular events that escaped detection. The present report addresses both concerns. The longer follow-up duration allowed to exclude that too short time of observation may have hidden a significant association. In fact, the difference in stroke risk between individuals with and without a PFO was further attenuated with longer follow-up (adjusted HR 1.10, CI 0.64–1.91). Moreover, PFO was not associated with potential sequelae of cardiac embolism such as SBI or WMH. Together, these findings suggest that PFO should not be considered a significant risk factor for cerebrovascular events in the general population. This observation contrasts with the results of the case-control studies mentioned earlier, but should not affect the management of secondary stroke prevention in patients with PFO. In the case-control studies, patients “selected” themselves as a subgroup at high risk among all PFO carriers by virtue of having a stroke as the presenting event, and cannot be compared with asymptomatic subjects with PFO as incidental finding. The combined information of the negative population studies and positive case-control ones suggests that a small subgroup of individuals with a PFO at high stroke risk may exist, but their effect on the stroke incidence in a general population sample is diluted when all individuals with PFO are considered. While considerable efforts are paid to assessing which of established preventive options (antithrombotic treatment, PFO closure) may be superior in preventing recurrent PFO-related strokes(13), far less attention is devoted to identifying asymptomatic individuals with PFO who may be at increased risk of first stroke. Several conditions (PFO size and degree of shunt(29); coexisting ASA(3); right atrial morphologic features(30); hypercoagulability(31) and presence of deep venous thrombosis(32); increased potential for atrial arrhythmias(33)) have been proposed as potential cofactors in PFO-related stroke risk, but some have been questioned, and none has been unequivocally identified as a reliable indicator of stroke risk. At this time, the prospective identification of individuals with PFO at increased risk of ischemic stroke remains elusive.

Secondary stroke prevention in patients with PFO include antithrombotic treatment with warfarin or antiplatelet agents as well as PFO closure. However, the level of evidence to support a specific treatment has been low, and the American Heart Association/American Stroke Association (AHA/ASA) and American College of Chest Physicians (ACCP) have published guidelines recommending antiplatelet therapy in first-ever ischemic stroke in this setting(34, 35). In addition, a randomized trial showed no significant benefit for percutaneous PFO closure compared with medical therapy among cryptogenic stroke patients(13), while two other randomized trials showed a trend favoring PFO closure, which did not reach statistical significance in an intention-to-treat analysis(36, 37). Stroke recurrence while on treatment is infrequent, and often secondary to causes other than PFO(15). The results of our study reaffirm that no preventive treatment is needed in asymptomatic subjects with an incidentally detected PFO.

Unlike previous reports(8), we did not observe an increased risk of cerebrovascular events in individuals with coexisting PFO and ASA, confirming previous reports in the general population(6, 7). The number of individuals with isolated ASA was too small to assess ASA’s independent stroke risk.

Our study has limitations. The mean age of our cohort was rather high, and the effect of PFO on stroke risk is stronger in younger subjects, who also have fewer competing stroke mechanisms. However, we did not observe an increased stroke risk even in the subset younger than 60. We used TTE for PFO detection, which is less sensitive than transesophageal echocardiography (TEE), but mainly misses small, possibly less significant PFOs(25). Moreover, our results are similar to those of SPARC(6), which were obtained by TEE, albeit on a smaller, predominantly White sample and with shorter follow-up.

In conclusion, we found that PFO is not associated with an increased risk of stroke or subclinical cerebrovascular disease in an urban multi-ethnic cohort living in the same community. The identification of possible associated factors that may increase the PFO-related stroke risk deserves further investigation.

Abbreviations

TTE

Transthoracic echocardiography

TEE

Transesophageal echocardiography

PFO

Patent foramen ovale

ASA

Atrial septal aneurysm

MRI

Magnetic resonance imaging

SBI

Silent brain infarcts

WMH

White matter hyperintensities

NOMAS

Northern Manhattan Study

CAD

Coronary artery disease

TOAST

Trial of Org 10172 in Acute Stroke Treatment