Patent Foramen Ovale and Recurrent Stroke: Closure Is the Best Option: No
Patients with patent foramen ovale (PFO) experiencing ischemic cerebrovascular symptoms should not routinely undergo closure. A clear relationship between PFO and stroke has not yet been proven. Even if PFOs are shown to predispose to stroke, medical therapies for stroke prevention in patients with PFOs have not been adequately tested, making comparisons with invasive treatment difficult, and probably premature.
Where’s the Evidence?
Several small, uncontrolled studies have suggested a relationship between PFO and stroke. Recent data, however, indicate that these studies may overestimate the association. In one study, PFOs were found in 20.8% of 519 randomly selected asymptomatic community-based controls compared with 16.5% of 158 patients referred for evaluation of cryptogenic stroke, demonstrating no increase in the prevalence of PFO among patients with stroke compared with a random nonhospitalized reference population.1
Only 2 prospective multicenter studies of substantial size have evaluated the stroke recurrence risk in patients with PFO. These studies provide the best data for guiding the management of patients with PFO and ischemic cerebrovascular events.
The French PFO-ASA Study Group evaluated 216 young patients (aged 18 to 55, mean age 40) with PFO and cryptogenic stroke and compared them with 304 cryptogenic stroke patients without PFO.2 All patients were extensively screened for alternative stroke etiologies, including coagulation testing and transesophageal echocardiography. Treatment consisted of aspirin (300 mg) in all cases. In this study, patients with PFO alone had a nonsignificantly lower stroke risk than those without a PFO at 4-year follow-up (2.3% PFO[+] versus 4.2% PFO[-]). Only patients with both PFO and an atrial septal aneurysm (ASA) experienced an increased risk of stroke (15.2% at 4 years; odds ratio 4.17, range 1.47 to 11.84)
The second major study was the PFO in Cryptogenic Stroke Study (PICSS). PICSS was a substudy of the Warfarin Aspirin Recurrent Stroke Study (WARSS) and evaluated 630 older (age range 30 to 85, mean 59) patients with PFO and stroke who underwent transesophageal echocardiography in a blinded fashion.3 The majority of strokes in PICSS were cryptogenic (42%) or lacunar (39%). Patients were randomized to aspirin (325 mg) or warfarin (INR 1.7 to 2.2). In this study, the rate of recurrent stroke or death in patients with PFO was not significantly different, regardless of treatment (Figure). Moreover, the recurrence risk was not significantly different from that observed in patients without PFO. Among patients with cryptogenic stroke (n=265), the stroke or death rate was about 50% lower in warfarin treated patients, although this difference did not reach statistical significance (Figure). The presence of a PFO, however, did not influence this lower rate of stroke recurrence on warfarin, which was 9.5% in patients with PFO and 8.3% in patients without PFO. Of interest, larger PFOs were associated with a lower, not higher, overall rate of recurrent stroke or death (18.5% with small PFOs versus 9.5% with large PFOs), in contradiction to the common belief that larger septal defects should be associated with higher stroke risk. In the PICSS, the coexistence of PFO and ASA did not increase stroke risk as in the French PFO-ASA study. These findings question the causal relationship between PFOs and stroke in most patients.
Are We Doing the Right Studies?
Despite these data, several closure trials have been initiated. Unfortunately, design issues may make the results difficult to interpret. For example, some studies include primarily younger patients who have had only a single neurological event. Since the best information show a very low stroke recurrence rate in this population, these studies may have difficulty demonstrating a difference between treatments. In one study, the end point includes both TIAs and strokes. Given the subjective nature of TIAs, their inclusion as an end point could be particularly problematic in a trial in which adequate blinding cannot be achieved due to the invasive nature of the therapy. This trial also allows enrollment of patients with TIAs, risking the dilution of treatment effects, as well as subjecting individuals to potentially unnecessary therapy. Moreover, it is unclear whether a reduction of transient ischemic events is sufficient justification to perform an invasive procedure, especially if a reduction in the firmer end points of stroke or death is not detected. This latter possibility is of particular concern given that such trials are probably underpowered to detect significant differences in these hard end points alone.
Finally, at least one proposed closure trial is an “equivalency study” in which a positive result will be declared if the recurrent event rate for the device is no worse than that of “best medical therapy.” Because the “best medical therapy” is unknown and will not be standardized, the results of these trials will be difficult to interpret. Moreover, given the extremely low rate of stroke seen in the French PFO-ASA study and the uncertain effects of medical therapy found in PICSS, equivalency seems an inappropriate criterion by which to judge success.
The assumed association between PFO and stroke may be analogous to that between mitral valve prolapse and stroke reported over the last decade. While early studies suggested an association between mitral valve prolapse and stroke, a number of subsequent investigations failed to confirm this relationship.4 One can image the harm that might have occurred if an invasive treatment to “fix” the mitral valve had been available at that time.
In summary, it is imperative that before embarking on a trial of PFO closure appropriate studies be designed to better characterize who is at sufficient risk to warrant interventional therapy. While it is likely that such a population exists, the challenge is to identify this group before exposing them to an invasive treatment.
The opinions expressed in this editorial are not necessarily those of the editors or of the American Stroke Association.
Petty GW, Khanderia BK, Meissner I, Whisnant JP, Rocca WA, Sicks JD, O’Fallon WM, McClelland RL, Christianson JH, Wiebers DO. A population-based study of the relationship between patent foramen ovale and cerebrovascular ischemic events. Neurology. 2003; 60: A256.Abstract.
Homma S, Sacco RL, Di Tullio MR, Sciacca RR, Mohr JP. Effect of medical treatment in stroke patients with patent foramen ovale: patent foramen ovale in Cryptogenic Stroke Study. Circulation. 2002; 105: 2625–2631.