Stroke recurrence in patients with patent foramen ovale

During the last decade, there has been a progressive emphasis on the role of interatrial septa1 abnormalities, including patent foramen ovale PFO) and interatrial septum aneurysm (ISA), in the genesis of ischemic stroke in young adults. ^1–6 Indeed, there is a significant association between PFO, ISA, and stroke of otherwise undetermined origin in case-control studies. ^7–9 Paradoxical embolism from a venous source through a right-to-left shunt is usually incriminated, ^1,5,8 but direct evidence for paradoxical embolism is commonly lacking, ^3,9,10 and systematic screening for deep venous thrombosis in the lower limbs or pelvis has led to extremely variable estimates. ^9,11 Despite these ongoing controversies, the possibility of paradoxical embolism is commonly retained in young patients with ischemic stroke and PFO without another cause of stroke. ^5,6,9 Moreover, ISA may contribute to increased risk of paradoxical embolism in patients with PFO, and may even act as a direct source of embolism. ^9,12 For these reasons, patients with stroke and PFO or ISA are commonly given antithrombotic therapy (platelet antiaggregating drugs, anticoagulants) for secondary prevention. ^13–15 There has been advocacy for more invasive treatments, such as surgical or transcatheter closure of PFO, in selected cases. ^16–19 However, the natural history of patients with stroke and interatrial septum abnormality is not known, so that preventive therapy is based only on a presumed risk of stroke recurrence.

We studied and followed all patients with stroke and interatrial septum abnormality (PFO, with or without ISA), admitted to our center since 1988, to assess the rate of event recurrence and patterns of recurring stroke.

Patients and methods

We studied all patients with PFO with and without ISA among 340 patients aged ≤ 60 years with acute cerebral infarct or transient ischemic attack (TIA) admitted consecutively to our population-based primary-care stroke center from 1988 to 1994. All 280 patients were systematically assessed following a standard protocol, ^4,20 including neurologic examinations, brain CT (usually two or more), extracranial Doppler ultrasonography with frequency analysis and B-mode real-time imaging, transcranial Doppler, echocardiography, 12-lead ECG, 24-to 48-hour 3-lead ECG monitoring following admission, and standard blood tests. Brain MRI and angiography as well as conventional catheter angiography were performed in selected cases. M-mode and B-mode echocardiography included transthoracic examination before and during microbubble test (antebrachial venous contrast injection of 10 mL of 5% glucose with no more than 0.3 cc of air, both with and without Valsalva maneuver and also during coughing) and transesophageal examination (mono/biplane) in order to assess the morphology of the interatrial septum. A right-to-left shunt during contrast injection was diagnosed when at least four microbubbles were seen in the left atrium or ventricle, not later than five cycles after the appearance of the microbubbles in the right atrium (to avoid false positives caused by re-circulation). ¹⁴ ISA was diagnosed with a > 1.5-cm bulging of the septum beyond the septa1 plane or a > 1.5-cm phasic course of the septum during the respiratory cycle, when the width of the septum basis was at least 1.5 cm. ^21–24 Other cardiac abnormalities, including interatrial communication (ostium secundum already visible on transthoracic examination, with dilatation of right atrium), were also assessed by echocardiography. ^4,20 Deep venous thrombosis was not looked for systematically without local symptoms, given the very low yield of venous Doppler and phlebography in this setting, from our own experience and that of others. ^9,15 Lung scintigraphy for pulmonary embolism was performed only in patients with clinically suspect pulmonary embolism.

We also assessed risk factors such as hypertension (blood pressure > 160/90 mm Hg at least twice before the stroke), diabetes mellitus (known fasting blood glucose > 120 mg/dL before the stroke), oral contraceptive use, cigarette smoking, and hypercholesterolemia (cholesterol > 180 mg/dL). Alternative causes of stroke, including coexisting extra-or intracranial arterial disease, general clinical findings, topography of stroke, and disability were evaluated following the protocol of the Lausanne Stroke Registry. ^4,20

We used three treatment regimens: (1) Surgical closure of the PFO was considered in patients with no alternative cause of stroke and at least two of the following: ISA, major right-to-left shunt (>50 microbubbles), multiple cerebral infarcts, multiple clinical events, and Valsalva maneuver preceding the neurologic event; (2) Long-term (indefinite) oral anticoagulant therapy (acenocoumarol, target INR 3 to 4) in surgical candidates who declined the operation and in patients with no alternative cause of stroke and only one of the factors listed under (1); (3) antiaggregant therapy (acetylsalicylic acid 250 mg/day) in the remaining patients, except if an alternative cause of stroke justified another treatment. In the latter group, a 3-month anticoagulation course was initiated in the patients with deep venous thrombosis or pulmonary embolism before switching to antiaggregant therapy.

Follow-up was performed in our stroke clinic, with regular clinical assessment every 6 or 12 months, or earlier in case of event recurrence. Death, stroke, TIA, and other vascular events (cardiac, peripheral arterial, pulmonary embolism, venous thrombosis) were assessed. In particular, brain CT or MRI was performed in case of stroke recurrence.

Results

Of the 340 consecutive patients aged ≤ 60 years with cerebral infarct or TIA, 140 (41%) had PFO [83 males (59%), 57 females (41%); mean age 44 ± 14 years], 118 (84%) had a brain infarct, and 22 (16%) only had TIA. Seventy-seven patients (55%) had involvement of the anterior (carotid) circulation or the border zones, whereas 63 patients (45%) had involvement of the posterior (vertebrobasilar) circulation (table 1). In 35 patients (25%), an ISA was associated with the PFO.

Discussion

Our study confirms the high prevalence (41%) of PFO in stroke patients ≤ 60 years, ^7,8 usually without any coexisting mechanism for stroke. Indeed, only one in six patients had an alternative cause of stroke, which was usually cardiac. Besides, our findings emphasize that ISA is commonly (1 of 4) associated with PFO in patients with stroke. Although the significance of this association is not clear, in some instances ISA may act as a potential source of paradoxical embolism in patients with PFO but without peripheral venous source. ^9,12 Clinical evidence for deep venous thrombosis was scarce in our patients, and a Valsalva maneuver at onset of stroke was present in only one-sixth of the patients. Some authors have advocated an investigative search for deep venous thrornbosis, ^ll but this remains controversial, with tremendous variations in diagnostic yield, ^11,12,25 including our previous experience. ^15,26 Also, coagulation disorders (including antiphospholipid antibodies) ²⁵ were extremely rare in our patients. This absence of a clear source of embolism and of an underlying thrombotic process is in contrast with the demonstration of intracranial embolic occlusions in nearly 90% of our patients with PFO and stroke who underwent cerebral angiography in the acute (<12 hours) phase of stroke. Thus, embolism may explain most strokes in patients with PFO, although an embolic source either is rarely found (deep venous thrombosis) or remains speculative (ISA). With embolism, the high frequency of posterior circulation stroke (45%) must also be emphasized, as previously noted by some, ²⁷ but not all studies. ²⁸

Our study is the first to report late prognosis in a large series of non-selected patients with PFO and stroke admitted to a primary-care center. Hanna et al. ²⁹ found no recurrence over 28 months of follow-up, but included only 15 patients. The French Study ³⁰ followed 132 patients from several primary, secondary, and tertiary care centers for 22.6 months, with a brain infarct rate of 1% per year; most patients had received acetylsalicylic acid or no treatment. In our patients, the early brain infarct or death rate was 2.4%, whereas the brain event (infarct or TIA) rate was 3.8% per year. Of note, no clinical embolic or thrombotic event in sites other than brain occurred. Logistic regression showed that independent correlates of stroke recurrence included interatrial communication, posterior cerebral artery territory involvement, positive history of recent active migraine, and an alternative cause of stroke. The last two items (migraine and alternative cause of stroke) emphasize non-PFO factors, whereas the first item (interatrial communication) may highlight the importance of right-to-left shunt. Indeed, some have suggested that the number of microbubbles in the left circulation during the microbubble test and the size of PFO on echocardiography may reflect a particular stroke risk. ^31–33 The association of ISA with PFO did not emerge as a factor for stroke recurrence in the multifactorial analysis, in contrast with preliminary data, ^30,34 in which the association of ISA with PFO multiplied the yearly stroke recurrence rate by two to three. However, these findings must be interpreted with caution, given the very low number of events during follow-up.

We did not randomize treatment in our patients, but antiaggregant therapy, anticoagulation, or surgical closure of PFO were not statistically significant markers of a decreased recurrence rate. However, this lack of difference between treatments must be interpreted with caution since there were only 16 brain events during follow-up and one-half were only TIAs. Major complications of treatment were not present, as no major bleeding developed over 998 months of oral anticoagulant therapy, and none of the 11 operated patients had perioperative complications.

In summary, our study shows that the first brain event in patients with PFO may often be devastating, as one-half the patients suffered a severe initial stroke. On the other hand, the overall risk of stroke recurrence seems rather low. The demonstration of factors associated with an increased risk of recurrence in our study (interatrial communication, posterior cerebral artery stroke, coexisting cause of stroke, history of migraine) and other studies ^30,33 (ISA) suggests that high-risk patients with PFO exist. Further attempts to better delineate the factors associated with a higher risk of stroke seem necessary before launching a clinical trial because of the high number of patients with a low recurrence risk who would be included in a trial randomizing non-selected patients with PFO and stroke.