Abstract
Background
A patent foramen ovale (PFO) has a complex anatomy, and evaluating the size before closure may be challenging. We aimed to investigate the correlation between preprocedural transesophageal echocardiography (TEE) and balloon sizing of PFO in patients undergoing percutaneous PFO closure.
Methods
A retrospective single-center study with analysis of 100 patients who, due to paradox thromboembolism in the left circulation, underwent percutaneous PFO closure. The PFO sizing was compared to measures attained by TEE and balloon sizing using linear regression analysis.
Results
PFO size measured by TEE occurred smaller than balloon sizing (2.19 mm [95% CI: 1.91 to 2.46] vs. 8.51 mm [95% CI: 8.02 to 9.00], p < 0.001). Additionally, neither the PFO channel length nor the atrial septal mobility measured by TEE correlated to the PFO size attained by balloon sizing, respectively (slope − 0.018 [95% CI: − 0.117 to 0.081], R = 0.036, p = 0.719) and (slope 0.049 [95% CI: ?0.043 to 0.141], R = 0.105, p = 0.297). Statistically significant difference in regression analysis but poor correlation was found between both TEE attained PFO and shunt size when compared to balloon sizing. Diverting patients according to the size of the PFO shunt was not statistically significant between PFO of moderate size compared, respectively, to a large and small PFO size. However, a difference was observed between a small and large PFO shunt size.
Conclusions
PFO defect and shunt size measured by TEE showed a poor correlation with balloon sizing. Neither PFO channel length nor septal mobility were correlated to the PFO size measured by balloon sizing.
Highlights
- •
Patent foramen ovale (PFO) and shunt sizing by transesophageal echocardiography correlate poorly to balloon sizing.
- •
PFO channel length and septal mobility do not correlate to the size of a PFO measured by balloon sizing.
- •
The above findings suggest caution when planning PFO closure based solely on transesophageal echocardiography measurements.
Background
Patent foramen ovale (PFO) is a known cause of paradoxical embolism and ischemic stroke. Closure of a PFO effectively reduces the risk of recurrent strokes in selected patients with a prior ischemic stroke. The size of a PFO, number of microbubbles shunted to the left atrium (LA), and presence of atrial septum aneurism are considered high-risk factors for occurrence of PFO-related ischemic stroke. Hence, it has been suggested that patients with these features benefit more from PFO closure than patients without high-risk PFO stigmata.
Transesophageal echocardiography (TEE) is key in evaluating the presence and features of a PFO, and it is used in the procedural planning prior to percutaneous closure. However, a PFO is a complex anatomical structure, where evaluating the size of the PFO using standard two-dimensional (2D) TEE is difficult. The shunt size is challenging to evaluate precisely due to dependency on the patient’s ability to cooperate with the timing of Valsalva’s maneuver and whether injection of agitated saline occurs from an upper extremity or lower extremity vein. To ensure optimal device size selection for closure, periprocedural interrogation of the interatrial defect can be performed using a compliant balloon sizing catheter. This method is considered valid for evaluation of the anatomical PFO size since the compliant balloon makes a precise cask of the PFO channel. It is, however, unknown how well different PFO sizes evaluated by balloon sizing correlate to sizing and shunting evaluated by 2D TEE.
We aimed to investigate the correlation between PFO sizing and shunting by preprocedural 2D TEE and periprocedural balloon sizing using fluoroscopy in patients who underwent percutaneous PFO closure.
Methods
Study Design and Patient Eligibility
The study was designed as a single-center retrospective analysis based on a prespecified screening protocol with the inclusion of a total of 100 patients. Patients who underwent percutaneous PFO closure due to paradoxical thromboembolism were eligible for inclusion. Patients were excluded if TEE or fluoroscopy was unavailable or if TEE imaging quality was insufficient for the prespecified analysis. A total of 192 patients underwent PFO closure during the period from January 2017 through November 2018. Of these, 92 patients were excluded ( Figure 1 ).
Data Collection
The patient’s medical history was obtained from the electronic medical patient file. Analysis of TEE and fluoroscopy images was done in Impax Client 6.5.5.1608 (AGFA Healthcare N.V, Belgium) and/or EchoPAC (version 203 GE Healthcare, Horten, Norway). Furthermore, study variables obtained by TEE and fluoroscopy were evaluated by a single observer blinded to patient identification.
Echocardiographic Analysis
Analysis of the PFO anatomy was performed by 2D TEE during the Valsalva maneuver. The section plane was determined to best visualize the PFO channel. Subsequently, measurements of the PFO height and channel length were performed. The height or size was defined as the separation of the primum septum from the secundum septum at the LA end of the PFO channel during maximal septal bulge into the LA, while the channel length was measured when the interatrial septum was displaced into the LA, hence visible to its full extent. Another feature of a PFO is interatrial septal aneurysm, which is defined by a total excursion of the primum septum from the right atrium (RA) into LA of >15 mm or an excursion beyond the septal plane into either LA or RA of >10 mm. Thus, to evaluate the mobility of the interatrial septum, measurements were conducted of the maximal excursion of the primum septum during a cardiac cycle. The shunt size of the PFO was examined through contrast echocardiography using intravenous administration of agitated saline injected into a cubital vein, followed by performance of the Valsalva maneuver. Initially, the contrast echocardiography was performed by transthoracic echocardiography with subsequent TEE repetition. Depending on the number of air bubbles passing through the PFO, the size of the shunt was classified into three sizes according to the classification used in the REDUCE trial, respectively small (1-5 bubbles), medium (6 to 25 bubbles), and large (>25 bubbles). The maximal number of air bubbles appearing in the LA was counted during a single frame within three cardiac cycles, starting from the first appearances of air bubbles in RA.
All PFO closure procedures were guided by intracardiac echocardiography with sizing balloon inflated using the stop-flow method ( Figure 2 ). Periprocedural measurements of the anatomical PFO size were conducted by fluoroscopy in anterior-posterior view. The size was defined as the minimum diameter of the sizing balloon (PTS Numed, 25 mm).
An interobserver variability test was performed by two observers, both blinded to patient identification, measuring all variables in 10 randomly selected patients.
Statistics
The correlation between the measurements obtained by balloon sizing and the different parameters measured from TEE (length, height, interatrial septum mobility, and shunt size) was evaluated through linear regression and one-way analysis of variance using IBM SPSS version 26. Thus, with a 95% CI and correlation coefficient defined as Pearson’s R, the results were expressed as slope and continued variables as mean with the correlation.
The assumption of homogeneity of variances between the prespecified shunt size was analyzed using Levene’s F test. Subsequently, a boxplot comparing fluoroscopy and the three different PFO sizes defined by the numbers of bubbles occurring using saline contrast was created in Graph Pad Prism (version 7.0c) and analyzed using one-way analysis of variance as well as multi comparison t-test between each group.
Interobserver variability was tested for the TEE measurements regarding PFO channel length, height, and shunt size using Bland-Altman analysis plotted as difference vs. average.
Results
Patient Characteristics
As indicated in the flowchart ( Figure 1 ), 100 patients fulfilled the inclusion criteria and had both TEE and fluoroscopy imaging available for analysis. The patient characteristics are presented in Table 1 . There was a predominance of men and a mean age of 48.7 ± 11.3 years. The most frequently observed comorbidity was hypercholesterolemia (56%), whereas diabetes mellitus (5%), hypertension (27%), migraine (11%), previous myocardial infarction (4%), and current smoking status (18%) occurred to a lesser extent. The primary thromboembolic event was due to ischemic stroke (74%), while only 19% occurred as transient ischemic attack and 7% as a systemic embolism. Echocardiographic morphology of the interatrial septum observed included atrial septal aneurism (11%), fenestrations (2%), Chiari network (7%), and Eustachian valve (26%). PFO size attained by balloon sizing had a mean of 8.5 ± 2.5 mm, while size measured by TEE were smaller, 2.2 ± 1.4 mm. In 23 patients a small shunt was present, 45 patients had a medium shunt and 32 patients had a large shunt.
