...  · Web viewEchocardiographic vs. balloon sizing of PFO Corresponding author: Asger Andersen, MD, PhD Dep. Cardiology, Aarhus University Hospital Palle Juul Jensens Boulevard

Title pageTitle: Comparing patent foramen ovale sizing by echocardiography with balloon sizing in patients

undergoing percutaneous closure.

Authors: Martin Christian Harmsen1, MD; Henrik Vase1, MD, PhD; Jens Erik Nielsen-Kudsk1, MD, DMSc;

Asger Andersen1, MD, PhD

Affiliations: 1. Department of Cardiology, Aarhus University Hospital, Aarhus, Denmark;

Running head: Echocardiographic vs. balloon sizing of PFO

Corresponding author:Asger Andersen, MD, PhD

Dep. Cardiology, Aarhus University Hospital

Palle Juul Jensens Boulevard 99

DK-8200 Aarhus N

E-mail: [email protected]

Conflict of interest:One or more of the authors have disclosed potential conflicts of interest regarding the content

herein.

Total word count: 2612

AbbreviationsPFO Patent foramen ovale

TEE Transesophageal echocardiography

LA Left atrium

RA Right atrium

TTE Transthoracic echocardiography

CI Confidence interval

AbstractObjective: To compare the size of a patent foramen ovale (PFO) evaluated by

echocardiography or peri-procedural balloon sizing in patients undergoing percutaneous PFO

closure.

Methods: The study was a single center retrospective analysis using a pre-specified

screening protocol for inclusion of 100 patients who underwent percutaneous PFO closure due to

left circulation thromboembolism. PFO size was evaluated by peri-procedural balloon sizing and

per-procedural echocardiography. Balloon sizing was compared using linear regression analysis to

echocardiography evaluated high-risk PFO features: PFO channel diastase, -channel length, number

of shunted bubbles and atrial septum mobility.

Results: Echocardiography underestimated the size of a PFO compared to balloon

sizing (2.19 (95% CI: 1.91–2.46) vs. 8.51 (95% CI: 8.02–9.00) mm, p < 0.001) and the correlation

between the sizing methods was poor (slope=0.80 (95%CI: 0.48-1.12, p<0.01, R=0.44). The

correlation between number of shunted bubbles and balloon sizing was poor (slope=0.04 (95%CI:

0.01-0.07), p=0.005, R=0.28). There were no correlation between balloon sizing and PFO channel

length (slope -0.018 (95%CI: -0.117 – 0.081), R = 0.036, p = 0.719) and atrial septal mobility

(slope 0.049, (95%CI: -0.043 – 0.141), R = 0.105, p = 0.297).

Conclusion: Defect and shunt size of a PFO measured by TEE showed a poor

correlation with PFO size determined by balloon sizing. PFO channel length and septal mobility did

not correlate to the size of a PFO measured by balloon sizing.

1 IntroductionPatent foramen ovale (PFO) can cause paradoxical embolism and ischemic stroke. In

selected patients with a prior ischemic stroke, closure of a PFO effectively reduces the risk of

recurrent strokes.1-4 The size of a PFO, number of microbubbles shunted to the LA and presence of

an atrial septum aneurism are considered high-risk factors for a PFO related ischemic stroke and it

has been suggested that patients with a high-risk PFO may benefit more from closure, than patients

without a high-risk PFO.5,6

Echocardiography is key in evaluating the presence and high-risk features of a PFO

and it is used in procedural planning before percutaneous PFO closure to ensure an optimal

procedural result. The anatomy of a PFO is however complex. It can vary from a slit-, tunnel- to

semilunar like shape and with varying sizes up to ~20 mm7 echocardiographic evaluation may be

difficult. Especially shunt- and anatomical size, which are some of the high-risk PFO features, may

be difficult to evaluate precisely as they are very dependent on the patients ability and timing of

Valsalva’s maneuver and whether the agitated saline is injected from an upper extremity or lower

extremity vein.8-10

Peri-procedural interrogation of the interatrial defect with a compliant balloon sizing

catheter is used for precise sizing of the defect to ensure an optimal device selection for closure.

The method is considered a valid tool for evaluating the anatomical size of a PFO as the compliant

balloon makes a precise cask of the PFO channel. It is however unknown how well balloon sizing

correlates to the size of a PFO evaluated by echocardiography.

We aimed to investigate the correlation between PFO sizing by pre-procedural

echocardiography and peri-balloon sizing in patients undergoing percutaneous PFO closure.

2 Methods2.1 Study Design and Patient Eligibility

Single center retrospective analysis using a pre-specified screening protocol for

inclusion of a total of 100 patients. Patients undergoing percutaneous PFO closure due to left

circulation thromboembolism were eligible for inclusion. Patients were excluded if

echocardiography or fluoroscopy were unavailable or if the echocardiographic imaging quality did

not allow the pre-specified analysis. In patients included in the study, balloon sizing was compared

to echocardiography evaluated high-risk PFO features: PFO channel diastase, -channel length,

number of shunted bubbles and atrial septum mobility.

A total of 192 patients who underwent PFO closure in the period from January 2017

through November 2018 were screened until 100 patients were included in the analysis (Fig. 1)

The study conformed to the declaration of Helsinki and the Danish Data Protection Act. All

patients consented prior to the intervention.

2.2 Data collectionMedical history was obtained from the electronic medical patient file. Analysis of

digitally stored echocardiography and fluoroscopy images was done in Impax Client 6.5.5.1608

(AGFA Healthcare N.V, Belgium) and/or EchoPACTM (version 203 GE Healthcare, Horten,

Norway). Study variables obtained by echocardiography and fluoroscopy were evaluated by a

single observer who were blinded to patient identification.

2.3 Echocardiographic analysisTransesophageal echocardiography (TEE) was used to analyze PFO anatomy: PFO

defect size was measured during Valsalva and defined as the separation of the primum septum from

the secundum septum at the left atrium (LA) end of the PFO tunnel during maximal septal bulge

into the LA. PFO channel length was measured when the septum was displaced into the LA and the

channel was visible in its full extent. Mobility of the interatrial septum was measured by the

maximal excursion of the primum septum from RA to LA during the cardiac cycle. An atrial septal

aneurysm was defined by a combined total excursion of the primum septum from RA to LA >15

mm or >10 mm to either LA or RA. Contrast echocardiography with Valsalva using intravenous

administration of agitated saline from a cubital vein was performed initially by transthoracic

echocardiography and repeated during TEE. The maximal number of air microbubbles on TEE or

transthoracic echocardiography (TTE) appearing in the LA was counted in a single frame within

three cardiac cycles after appearance of the first air bubbles in the RA. The shunt size was classified

into three groups: small (1–5 bubbles), medium (6– 25 bubbles), and large (>25 bubbles).

Anatomical PFO size measured by fluoroscopy in the anterior-posterior view and

defined as the minimum diameter of the sizing balloon (PTSR Numed, 25 mm). All PFO closure

procedures were guided by intracardiac echocardiography with the sizing balloon inflated using the

stop-flow method. Representative examples are shown in supplementary figure 1.

Interobserver variability was tested by measuring each variable in 10 randomly

selected patients by two observers blinded to patient identification.

2.4 StatisticsCorrelation between anatomical PFO size measured by balloon sizing and parameters

measured by TEE (PFO size, length, atrial septum mobility, shunt size) were evaluated with linear

regression and one-way ANOVA using IBM SPSS ver. 26. Regression results are expressed as

slope and continued variables as mean with 95% confidence intervals (CI) and the correlation

coefficient as Pearson’s R. The assumption of homogeneity of variances in the pre-specified shunt

size groups was tested and satisfied based on Levene’s F test. Boxplot of fluoroscopy and saline

contrast groups were created in Graph Pad Prism (v. 7.0c) and analyzed using one-way ANOVA

with multi comparison t-test between groups. Interobserver variability within each variable was

tested using Bland-Altman analysis plotted as difference versus average.

3 Results3.1 Patient characteristics

Patient characteristics are presented in table 1.

3.2 Balloon sizing and echocardiographic evaluation of PFO sizePFO size measured by TEE was smaller than measured by balloon sizing 2.19 (95%

CI: 1.91–2.46) vs. 8.51 (95% CI: 8.02–9.00) mm, p < 0.001). PFO channel length did not correlate

with PFO size by balloon sizing (slope -0.018 (95%CI: -0.117 – 0.081), R = 0.036, p = 0.719) and

neither did atrial septal mobility (slope 0.049, (95%CI: -0.043 – 0.141), R = 0.105, p = 0.297).

Regression analysis of PFO size measured by TEE showed a statistically significant, but poor

correlation to PFO balloon sizing as did the shunt size measured by TEE. Grouping of the patients

according to size of the PFO shunt size did not show a statistically significant difference between

the large and small group compared to the moderate group, but a difference was observed between

the small and large group (Figure 2).

3.3 Interobserver variability analysisBland-Altman plots of the interobserver variability analysis is presented in Figure 4.

4 DISCUSSIONOur results indicate that determining the shunt- and anatomical size of a PFO with

TEE correlates poorly to balloon sizing.

The patients included in the analysis where in their late forties, had some risk factors

with hypercholesterolemia and hypertension being the most predominant, and the primary reason

for PFO closure was a prior ischemic stroke. The mean size of their PFO was substantially larger

measured by balloon sizing compared to the echocardiographic sizing. Most shunts where moderate

or large. These baseline features are very similar to patients included in the REDUCE and

RESPECT trial.2,3 Not only the age and comorbidities, but also the anatomical PFO features were

very similar. Hypercholesterolemia was higher in our population, but as we counted patients on

statins as having hypercholesterolemia this merely reflects a prior stroke, and that statins was

initiated by the neurologist as standard care after a stroke.

The separation between the primum and secundum septum of a PFO have been used

to measure PFO size on TEE in key clinical trials.1-4 By using the this method, the size of a PFO

was underestimated fourfold based on means compare to balloon sizing. Not only did 2 dimensional

TEE massively underestimate the anatomical size of the PFO, but the regression analysis show that

it correlates weakly to balloon sizing of a PFO suggesting that patient selection and procedural

planning based on this entity alone should be done very cautiously. In itself this finding is not

surprising. The anatomical size of a PFO is very complex7 and using a 2-dimensional measure may

be too simple a method to evaluate the PFO anatomical size. This is also supported by findings

from others, where advanced peri-procedural 3D TEE estimates a much larger anatomical PFO size

than 2D echocardiography.11

The ability of right-to-left shunting through a PFO may be the most important risk

factor when considering the pathophysiology of paradoxical embolism. It correlates poorly with the

anatomical size of a PFO measured by balloon sizing emphasizing that the number of microbubbles

shunted is not solely determined by the size of a PFO. We know that the number of shunted

bubbles is very dependent on the patients ability and timing of Valsalva’s maneuver and whether

the agitated saline is injected from an upper extremity or lower extremity vein. 8-10 The poor

correlation can to some extend be explained by this.

The mobility of the atrial septum and the length of the PFO channel evaluated by TEE

does not correlate to the anatomical size of a PFO suggesting that these measures are independent of

the anatomical PFO size. This could explain recent results from the close and DEFENSE-PFO trial

showing that ASA was the most important predictor of recurrent stroke, not shunt size,5 but we still

do not know the how the true anatomical size of a PFO would correlate to recurrent stroke. These

findings emphasize that a PFO is a complex anatomical structure and that decision-making and

procedural planning based on PFO size evaluated with TEE should be done carefully.

4.1 LimitationsData was collected retrospectively. Echocardiographic parameters were measured

from clinical scans where no strict echocardiographic protocol was used. This may also confound

shunt evaluation, since the ability of the patient to perform a proper Valsalva maneuver, the

presence of an eustachian valve and timing of injection of the agitated saline are important to

visualize the shunting potential through a PFO. All patients had injection of agitated saline through

a cubital vein, knowing that injection though a femoral vein has higher sensitivity for PFO detection

and likely shunts more microbubbles across the PFO.11

Our test of interobserver variability showed reasonable results for PFO channel length

and shunt size, but not for measuring PFO size (separation between primum and secundum septum)

further emphasizing that this parameter should be used with caution in decision-making.

5 Conclusion

Defect and shunt size of a PFO measured by TEE showed a poor correlation with PFO

size determined by balloon sizing. PFO channel length and septal mobility did not correlate to the

size of a PFO measured by balloon sizing.

Impact on daily practice

There is a poor correlation between the size of a PFO evaluated by TEE and the

anatomical size evaluated by balloon sizing. It emphasizes that a PFO is a complex anatomical

structure that is difficult to fully evaluate by TEE and that PFO size based on TEE should be used

cautiously in decisions and per-procedural planning of closure.

Funding

Asger Andersen is funded by the Alfred Benzon FoundationJens Erik Nielsen-Kudsk is funded by

the Novo Nordisk Foundation (NNF17OC0024868/NNF17OC0029510).

Conflict of interest

Asger Andersen have retrieved travel grants from ABBOTT. Jens Erik Nielsen-Kudsk is a proctor

and consultant for Abbott and Boston Scientific. Martin Christian Harmsen and Henrik Vase have

no conflicts of interest to declare.

References1. Mas JL, Derumeaux G, Guillon B, et al. Patent foramen ovale closure or

anticoagulation vs. antiplatelets after stroke. N Engl J Med 2017;377:1011–21.

2. Sondergaard L, Kasner SE, Rhodes JF, et al. Patent foramen ovale closure or

antiplatelet therapy for cryptogenic stroke. N Engl J Med 2017;377:1033–42.

3. Saver JL, Carroll JD, Thaler DE, et al. Long-term outcomes of patent foramen ovale

closure or medical therapy after stroke. N Engl J Med 2017;377:1022–32.

4. Lee PH, Song JK, Kim JS, et al. Cryptogenic stroke and high-risk patent foramen

ovale: the DEFENSE-PFO J Am Coll Cardiol 2018;71:2335–42.

5. Turc G, Lee JY, Brochet E, Kim JS, Song JK, Mas JL; CLOSE and DEFENSE-PFO

Trial Investigators. Atrial Septal Aneurysm, Shunt Size, and Recurrent Stroke Risk in

Patients With Patent Foramen Ovale. J Am Coll Cardiol. 2020 May 12;75(18):2312-

2320. doi: 10.1016/j.jacc.2020.02.068. PMID: 32381162.

6. Pristipino C, Sievert H, D’ascenzo F, Mas J-L, Meier B, Scacciatella P, et al.

European position paper on the management of patients with patent foramen ovale.

General approach and left circulation thromboembolism. EuroIntervention. 2019

Jan;14(13):1389–402.

7. Rana BS, Shapiro LM, McCarthy KP, Ho SY. Three-dimensional imaging of the atrial

septum and patent foramen ovale anatomy: defining the morphological phenotypes of

patent foramen ovale. Eur J Echocardiogr. 2010 Dec;11(10):i19-25. doi:

10.1093/ejechocard/jeq122. PMID: 21078835.

8. Mojadidi MK, Bogush N, Caceres JD, Msaouel P, Tobis JM. Diagnostic accuracy of

transesophageal echocardiogram for the detection of patent foramen ovale: a meta-

analysis. Echocardiography. 2014 Jul;31(6):752-8. doi: 10.1111/echo.12462. Epub

2013 Dec 23. PMID: 24372693.

9. Marriott K, Manins V, Forshaw A, Wright J, Pascoe R. Detection of right-to-left atrial

communication using agitated saline contrast imaging: experience with 1162 patients

and recommendations for echocardiography. J Am Soc Echocardiogr. 2013

Jan;26(1):96-102. doi: 10.1016/j.echo.2012.09.007. Epub 2012 Oct 13. PMID:

23072711.

10. Hamann GF, Schätzer-Klotz D, Fröhlig G, Strittmatter M, Jost V, Berg G, Stopp M,

Schimrigk K, Schieffer H. Femoral injection of echo contrast medium may increase

the sensitivity of testing for a patent foramen ovale. Neurology. 1998

May;50(5):1423-8. doi: 10.1212/wnl.50.5.1423. PMID: 9595999.

11. Demulier L, Paelinck BP, Coomans I, Hemelsoet D, De Backer J, Campens L, De

Wolf D. A new dimension in patent foramen ovale size estimation.

Echocardiography. 2020 Jul;37(7):1049-1055. doi: 10.1111/echo.14696. Epub 2020

Jul 7. PMID: 32634292.

Figure legends

Figure 1: Flow chart of study inclusion.PFO (Patent foramen ovale). TTE (Transthoracic echocardiography). TEE (Transesophageal

echocardiography).

Figure 2 Echo analysis A: Patent foramen ovale (PFO) Diastasis. B: PFO Length. C: Bubbles shunted through a PFO D1

& D2: Atrial septum

aneurysm with a combined deviation of 22 mm. E1 and E2: Large eustachian valve

inhibiting the contrast from reaching the septum. F: Balloon sizing. G PFO occluder in situ

Figure 3: Correlation between PFO size measured by TEE, shunt size and PFO balloon sizing.Linear regression analysis of PFO size (A) and shunt size (B) measured by TEE and balloon sizing.

Boxplot of shunt size and balloon sizing (C). Each column illustrates the mean and the quartile

distribution of balloon sizing.

Figure 4 Interobserver variabilityBland-Altman plots of difference versus average assessing echocardiographic interobserver

variability of 10 randomly selected patients analyzed by 2 observers blinded to outcome. A)

Persistent foramen ovale (PFO) channel length B) PFO size by trans esophageal echocardiography

(TEE) C) Bubbles shunted

Figures Figure 1 Flow chart of study inclusion

Figure 2 Echo and fluoroscopy analysis

Figure 3: Correlation between PFO size measured by TEE, shunt size and PFO balloon sizing.

Figure 4 Interobserver variability

Table 1Patient Characteristics n = 100

Age, years 48.7 ± 11.3

Male gender, n (%) 70 (70%)

Medical history

Diabetes, n (%) 5 (5%)

Hypertension, n (%) 27 (27%)

Current smoker, n (%) 18 (18%)

Hypercholesterolemia, n (%) 56 (56%)

Migraine, n (%) 11 (11%)

Myocardial infarction, n (%) 4 (4%)

Tromboembolic Event

Ischemic stroke, n (%) 74 (74%)

Transient ischemic attack, n (%) 19 (19%)

Systemic embolism, n (%) 7 (7%)

Morphology of the Atrial Septum

Atrial septum aneurism, n (%) 11 (11%)

Fenestrations, n (%) 2 (2%)

Chiari network, n (%) 7 (7%)

Eustachian valve, n (%) 26 (26%)

PFO balloon sizing, mm 8.5 ± 2.5

PFO echocardiography sizing, mm 2.2 ± 1.4

Shunt size small/medium/large, n (%) 23/45/32 (23/45/32%)

Table 1 Patient Characteristics