Author + information
- Received January 27, 2009
- Revision received April 24, 2009
- Accepted May 3, 2009
- Published online July 1, 2009.
- Gianluca Rigatelli, MD⁎,⁎ (, )
- Silvio Aggio, MD†,
- Paolo Cardaioli, MD⁎,
- Gabriele Braggion, MD†,
- Massimo Giordan, MD⁎,
- Fabio Dell'avvocata, MD⁎,
- Mauro Chinaglia, MD‡,
- Giorgio Rigatelli, MD§,
- Loris Roncon, MD† and
- Jack P. Chen, MD∥
- ↵⁎Reprint requests and correspondence:
Dr. Gianluca Rigatelli, Cardiovascular Diagnosis and Endoluminal Intervention, Rovigo General Hospital, Via WA Mozart, 9, 37040 Legnago, Verona, Italy
Objectives We postulate that, in patients with large patent foramen ovales (PFO) and atrial septal aneurysms (ASA), left atrial (LA) dysfunction simulating “atrial fibrillation (AF)-like” pathophysiology might represent an alternate mechanism in the promotion of arterial embolism.
Background Despite prior reports concerning paradoxical embolism through a PFO, the magnitude of this phenomenon as a risk factor for stroke remains undefined, because deep venous thrombosis is infrequently detected in such patients.
Methods To test our hypothesis, we prospectively enrolled 98 consecutive patients with previous stroke (mean age 37 ± 12.5 years, 58 women) referred to our center for catheter-based PFO closure. Baseline values of LA passive and active emptying, LA conduit function, LA ejection fraction, and spontaneous echocontrast (SEC) in the LA and LA appendage were compared with those of 50 AF patients as well as a sex/age/cardiac risk-matched population of 70 healthy control subjects.
Results Pre-closure PFO subjects demonstrated significantly greater reservoir function as well as passive and active emptying, with significantly reduced conduit function and LA ejection fraction, when compared with AF and control patients. Furthermore, in PFO patients, 66.3% (65 of 98) had moderate-to-severe ASA and basal shunt; SEC was observed in 52% of PFO plus ASA patients before closure. Multivariate stepwise logistic regression revealed moderate-to-severe ASA (odds ratio: 9.4, 95% confidence interval: 7.0 to 23.2, p < 0.001) as the most powerful predictor of LA dysfunction. After closure, all LA parameters normalized to the levels of control subjects: no SEC, device-related thrombosis, or aortic erosion were observed on follow-up echocardiography.
Conclusions This study suggests that moderate-to-severe ASA might be associated with LA dysfunction in patients with PFO. The resultant similarities to the pathophysiology of AF might represent an additional contributing mechanism for arterial embolism in such patients.
Despite numerous reports regarding the association of arterial embolism with patent foramen ovale (PFO) (1–3), the causality of paradoxical embolism remains speculative in many cases. The true impact of this phenomenon as a risk factor for stroke in such patients with PFO is unknown, because deep vein thrombosis is frequently not identified in these individuals (4). The diagnostic accuracy for detection of small thrombi might, however, be limited by standard available techniques. Because patients with presumed paradoxical embolism frequently share functional features with those with atrial fibrillation (AF), such as impairment of the left atrial (LA) function, we postulate that an “AF-like” physiology might contribute to LA thrombosis and subsequent arterial embolism. Thus, we assessed the presence of classical LA anatomic and functional predictors of AF and stroke such as LA dysfunction, LA spontaneous echocontrast (SEC) or thrombosis, and reduced left atrial appendage (LAA) velocity amongst control subjects and AF and PFO patients before and after percutaneous closure.
We prospectively enrolled 98 consecutive patients (mean age 37 ± 12.5 years, 58 women) with previous stroke who had been referred to our center for catheter-based closure of interatrial shunts according to standard indications (5) over a 36-month period (Table 1). Written informed consent was obtained from all patients enrolled in the study.
Echocardiographic protocols and definitions
Transthoracic (TTE) and transesophageal echocardiography (TEE) was conducted with a GE Vivid 7 (General Electric Corp., Norfolk, Virginia) 1 month before the procedure and repeated 1 month after closure: LA volumes and function as well as shunt degree as assessed by contrast injection and Valsalva maneuver under local anesthesia were recorded (6).
The LA passive and active emptying, LA conduit function, and LA ejection fraction were evaluated by TTE before and 1 month after PFO closure. Assessed parameters included: LA volumes as determined at the mitral valve opening (maximal, Vmax), at the onset of atrial systole (P-wave of the electrocardiogram, Vp), and at mitral valve closure (minimal, Vmin) from the apical 2- and 4-chamber views by means of the biplane area-length method via software within the system. The following formula was used to calculate LA volume (7,8): volume = 8 × 4 × Ach (A2ch/3Π) × common length (where A4ch and A4ch = LA area in 4- and 2-chamber views, respectively). The LA functional parameters were calculated as described in Table 2. The LA ejection fraction served as a measure of LA systolic performance; and acceleration and deceleration times of systolic phase of pulmonary venous flow corresponded to LA relaxation and compliance, respectively.
The LAA peak flow velocity as well as SEC or thrombosis in LA and LAA were also evaluated among the 3 groups, because low LAA peak flow velocity and SEC have been associated with risk of stroke (9,10). The pre- and post-operative echocardiographic findings were blindly evaluated by 2 blinded physicians.
Intracardiac echocardiography protocol
Patients who fulfilled the criteria for PFO closure underwent intraprocedural intracardiac echocardiographic (ICE) assessment with the mechanical 9-F 9-MHz UltraICE catheter (EP Technologies, Boston Scientific Corporation, San Jose, California). The ICE study was conducted as previously described (11,12), by performing a manual pull-back from the superior vena cava to the inferior vena cava through 5 sectional planes. The ICE monitoring of the implantation procedure was conducted in the 4-chamber plane. Special attention was posed in visualizing potential smoke-like phenomenon in the LA before closure with standard gray-scale set-up, although the eventual thrombogenic process might be of a minor entity than in patients with AF or rheumatic disease and thus probably difficult to visualize by TEE.
Combined antibiotic therapy (gentamicin 80 mg plus ampicillin 1 g or Vancomicin 1 g if allergy had been recorded on anamnesis) was administered intravenously 1 h before the procedure. The right femoral vein was catheterized through an 8-F sheath and used for pre-closure right heart catheterization; the sheath was subsequently replaced with a 10- or 12-F long sheath for device implantation. The left femoral vein was catheterized with an 8-F sheath and replaced with a precurved 9-F long sheath for ICE study.
Intraoperative closure criteria and device selection
On the basis of ICE study and the presence/absence of moderate-to-large ASA and long tunnel-like PFO (tunnel length ≥12 mm), the operators selected either the Amplatzer Occluder family (PFO Occluder, Cribriform Occluder, AGA Medical Corporation, Golden Valley, Minnesota) or the Premere Closure System (St. Jude Medical Inc. GLMT, St. Paul, Minnesota), as previously described (13). The Amplatzer family, a well-known device composed of 2 parallel nitinol wire-mesh disks was selected in cases of ASA, because its more rigid design offered superior interatrial septal stabilization, as previously suggested (14). The PFO Occluder was used in cases of mild ASA (1 or 2 right/left [RL]-left/right [LR], according to Olivares classification) (15) and short tunnel, whereas the Amplatzer ASD Cribriform Occluder was preferred in cases of moderate-to-severe ASA (3RL-3LR up to 5RL-5LR, according to Olivares classification). Finally, the Premere Occlusion system, a device composed of 2 portions of a single nitinol wire (right is covered by tissue), connected by an adjustable length tender, was selected in cases of long tunnel PFO without ASA, because this device can be asymmetrically opened through the long channel.
All patients were administered aspirin 100 mg/day for 6 months after the procedure. Follow-up evaluations consisted of TEE at 1 month, with repeat study at 6 months if even minimal shunt was detected. Post-procedural assessment further included TTE at 1, 6, and 12 months; transcranial Doppler (TCD) at 1 month; Holter monitoring at 1 month; and combined cardiologic and neurological visit at 1, 6, 12 months. Residual shunt was assessed by contrast TEE and TCD (16).
During transthoracic and TEE, basal values were compared with those of a sex/age/heart rate-matched population of 70 healthy volunteers and 50 sex-matched AF patients. The latter were enrolled during the same study period; these subjects all demonstrated normal left ventricular function and size and were optimally medicated with respect to heart rate control and oral anticoagulant levels (Table 1). Additionally, pre-procedural values of PFO patients were compared with those obtained at 6-month follow-up. All values were corrected for R-R interval and body surface area.
Chi-square, analysis of variance, and paired Student t tests were used to compare frequencies and continuous variables between groups. Statistical analysis was performed with a statistical software package (SAS for Windows, version 8.2, SAS Institute, Cary, North Carolina). A probability value of <0.05 was considered to be statistically significant. Stepwise logistic regression analysis was used to determine independent determinants of preoperative LA dysfunction. We considered only patients with at least 1 altered parameter as patients with “LA dysfunction”. The variables subject to analysis were: age, sex, moderate-severe ASA (3RL, 3LR up to 5RL, according to Olivares classification), shower pattern on TCD, curtain pattern on TCD, medium-to-large shunt with Valsalva on TEE, and basal shunt without Valsalva on TEE. A dispersion graphic with linear regression was generated to determine the correlation between LA functional parameter.
When compared with healthy subjects, pre-procedural PFO patients demonstrated greater reservoir function and passive and active emptying but lower conduit function and LA ejection fraction (Table 3). These abnormal values were similar to those found in AF subjects.
When compared with isolated PFO patients, those with PFO plus ASA were observed to have worse functional parameters and higher percentage of spontaneous right-to-left shunt (Fig. 1). Interestingly, patients with both PFO and ASA had more coagulative abnormalities than PFO alone patients: 67.3% (35 of 52) versus 19.5% (9 of 46), p < 0.01. This observation might account for the more frequent history of recurrent cerebral events (>2 stroke/transient ischemic events or multiple magnetic resonance imaging ischemic foci) in the former versus latter groups: 78.8% (41 of 52) versus 34.8% (16 of 46), p >0.01. In the PFO plus ASA group, a smoke-like phenomenon was observed by at least 1 imaging tool in the LA before closure of 27 of 52 patients (52%, 22 of 27 patients by intraprocedural ICE) (Fig. 2): all these patients had moderate-to-severe ASA. No SEC was observed in the PFO alone group (p < 0.01). The LAA peak flow velocity was not statistically different between the entire cohort of PFO patients and healthy control subjects (48.1 ± 21.0 cm/s vs. 49.8 ± 23.2 cm/s, p = 0.18), with both being higher than AF patients (21.8 ± 8.2 cm/s, p < 0.001).
Multivariate stepwise logistic regression revealed moderate-to-severe ASA (odds ratio [OR]: 9.4, 95% confidence interval [CI]: 7.0 to 23.2, p < 0.001) and basal shunt on TEE (OR: 5.6, 95% CI: 3.0 to 12.1, p < 0.001) as predictors of LA dysfunction, whereas moderate-to-severe ASA (OR: 11.7, 95% CI: 8.0 to 29.2, p < 0.001) was the strongest predictor of SEC in the LA.
All patients underwent successful PFO transcatheter closure with the Premere Occlusion System in 28 patients, Amplatzer Cribriform Occluder in 40 patients, and Amplatzer PFO Occluder in 30 patients. Complete PFO closure was achieved in 90 patients (91.8%, 8 patients with a persistent small shunt, all with an Amplatzer ASD Cribriform Occluder) on TEE and TCD. Aside from 4 episodes of AF lasting up to 48 h, no other perioperative or long-term complications—including ictus or transient ischemic attack recurrence—were observed during mean follow-up of 26.1 ± 6.7 months: in particular, no LA SEC, LA thrombosis, or device-related thrombosis or aortic erosions were observed on follow-up echocardiography.
After closure, active and passive emptying as well as conduit function and LA ejection fraction tended to normalize (Table 3) to the levels of healthy subjects; moreover, after closure, LA functional parameters varied according to device type. The Premere Occlusion System was associated with a greater reduction of LA passive, LA conduit, and active emptying as well as larger increase in LA ejection fraction, as compared with Amplatzer PFO and ASD Cribriform Occluders. This observation might well be related to the more favorable baseline anatomic features (Table 4) (Fig. 3). The inter-observer agreement was 99.4% globally.
Our study suggests an alternative mechanism for arterial embolism in patients with PFO and large ASA. Either with or without concurrent potential for paradoxical embolism, LA dysfunction—simulating that of AF—might further predispose these patients to cardiogenic thrombosis. Similar to those found in patients with chronic AF, LA parameters of active and passive emptying as well as ejection fraction are clearly altered in PFO patients with large ASA.
Mechanistically, the isolated existence of a PFO is not per se a risk factor for stroke; paradoxical embolism requires a venous source of thrombus, typically from the lower extremity deep veins (17). However, deep venous thrombosis has been documented in only a small percentage of presumed paradoxical embolic cases (4). Thus, the certainty of this phenomenon as a cause of recurrent stroke in the PFO patient without deep venous thrombosis remains unclear. Alternative hypotheses in such cases have included platelet and fibrin aggregates on the ASA surface or undiagnosed deep venous thrombosis in atypical sites (such as uterine or prostate venous plexus) (18). Although all or some might be contributory, the relative importance of each is difficult to quantitate.
More recently, the impact of coagulation abnormalities in the genesis of paradoxical embolism in PFO patients has been evaluated (19). The importance of such coagulopathies, not infrequent concurrent findings, in the pathophysiology of arterial embolism in PFO patients is not diminished by our findings; such pro-coagulable states further enhance the thrombogenicity of the dysfunctional LA.
The role of LA function in the modulation of left ventricular diastolic filling is well-described (20). The LA functions as a reservoir for the collection and storage of blood during left ventricular systole and as a conduit for the blood passage from the pulmonary veins to the left ventricle during early left ventricular diastole. In addition to serving as a passive conduit and reservoir for pulmonary venous inflow, the LA further augments left ventricular end-diastolic filling through active contraction (21).
Grant et al. (22) estimated that 42% of the left ventricular stroke volume is stored in the LA during ventricular systole; the subsequent dissipation of this potential energy acts to accentuate left ventricular filling during diastole. The efficiency of this mechanism is governed by atrial distensibility during ventricular systole. The LA conduit function occurs primarily but not exclusively during ventricular diastole and represents the blood volume passing through the LA, which is not attributable to reservoir or booster pump function. This amount accounts for 35% of LA flow (23). The booster pump function represents LA contraction and is dependent on several factors, including timing of atrial systole, vagal stimulation, magnitude of venous return, left ventricular end-diastolic pressures, and left ventricular systolic reserve (24).
On the basis of these considerations, it is likely that some degree of LA dysfunction, such as impairment of active or passive emptying or perhaps conduit function, might be present in patients with PFO, especially those with moderate-large ASA. Our study further suggests that, after percutaneous closure, these functional aberrations return to levels observed in healthy subjects. The pathophysiologic similarities to AF patients are particularly provocative. Left atrial wall stiffness resulting from loss of active and passive emptying as well as contraction and reservoir functions have been correlated with long-term risk of LA thrombosis and subsequent arterial embolism (9,10,25). The presence of SEC alone versus that of frank LA or LAA thrombosis might be dependent upon the degree and form of LA functional alterations. Although some measurable impairments were insufficient to produce visible thrombus, they might nonetheless promote platelet-fibrin aggregation and micro-thrombi, manifested as SEC. This phenomenon might be further augmented in patients with pro-coagulative tendencies.
The presence of moderate-to-large ASA seemed to be the most important determinant of impaired LA function, as further suggested in a recent small study by Goch et al. (26). This hypothesis is additionally substantiated by analysis of results based upon device types, which were selected according to the presence or absence of moderate-to-large ASA. The apparent superior results achieved by the Premere Occlusion system might be attributable to the more favorable pre-selection anatomic features at baseline (because they were implanted in patients with no ASA), whereas the septal rigidity caused by use of the Amplatzer Occluder family might account for the trend to near normalization of LA functional parameters. The apparent association with basal right-to-left shunt is more difficult to explain but might be attributable to slight alterations in LA volumes and function resulting from increased volume.
Moreover, the subsequent post-closure improvements in LA functional parameters suggest that PFO closure or ASA stabilization might be beneficial not only in protection from recurrent paradoxical embolism but also in preservation of LA function. Given the similarities to findings in AF, prevention of such sequelae by prophylactic or primary percutaneous closure might reduce the risk of LA thrombosis in the long-term. Although chronic anticoagulation might be proposed as a more conservative treatment in these situations, such therapy is clearly not without associated morbidity. Furthermore, actual reversal of the atrial pathophysiology as reported in our evaluation might be, at least intuitively, a more attractive option.
Although our study was a prospective evaluation, the relatively small sample size, lack of randomization, and short-term follow-up all tend to limit the widespread applicability of our findings. A placebo PFO arm was not ethically possible in our study, because all PFO subjects had history of minor or major recurrent stroke and were referred for percutaneous closure. Moreover, in concordance with present indications, all closures were performed for secondary prevention in patients who have already suffered a stroke. Finally, the visualization of smoke-like phenomenon by TEE or ICE might be somewhat questionable, depending on different equipment set-up, operators' skill, and probably has limited significance compared with LA dysfunction parameters.
To our knowledge, the present study is the first to suggest an alternative mechanism, involving AF-like LA pathophysiology, for arterial embolism in patients with PFO. Furthermore, large prospective randomized trials might help elucidate any potential benefits of PFO closure for primary stroke prevention in asymptomatic patients with PFO and moderate-to-large ASA.
- Abbreviations and Acronyms
- atrial fibrillation
- confidence interval
- intracardiac echocardiographic
- left atrium/atrial
- left atrial appendage
- odds ratio
- patent foramen ovale
- spontaneous echocontrast
- transcranial Doppler
- transesophageal echocardiography
- transthoracic echocardiography
- Received January 27, 2009.
- Revision received April 24, 2009.
- Accepted May 3, 2009.
- American College of Cardiology Foundation
- Lamy C.,
- Giannesini C.,
- Zuber M.,
- et al.
- ↵(2005) SPREAD—Stroke Prevention and Educational Awareness Diffusion: Stroke Italian Guidelines IV Edition, http://www.spread.it. Accessed June 2009.
- Lang R.M.,
- Bierig M.,
- Devereux R.B.,
- et al.,
- American Society of Echocardiography's Nomenclature and Standards Committee,
- Task Force on Chamber Quantification,
- American College of Cardiology Echocardiography Committee,
- American Heart Association,
- European Association of Echocardiography,
- European Society of Cardiology
- Leung D.Y.,
- Black I.W.,
- Cranney G.B.,
- Hopkins A.P.,
- Walsh W.F.
- Goldman M.E.,
- Pearce L.A.,
- Hart R.G.,
- et al.
- Zanchetta M.,
- Onorato E.,
- Rigatelli G.,
- et al.
- Rigatelli G.,
- Cardaioli P.,
- Braggion G.,
- Giordan M.,
- Aggio S.,
- Roncon L.
- Sloan M.A.,
- Alexandrov A.V.,
- Tegeler C.H.,
- Therapeutics and Technology Assessment Subcommittee of the American Academy of Neurology
- Sastry S.,
- Riding G.,
- Morris J.,
- et al.
- Hoit B.D.
- Dernellis J.,
- Stefanidis C.,
- Toutouzas P.