Can computer tomography help predict feasibilityof transseptal puncture after percutaneous closureof an interatrial septal communication?

Philipp Wagdi & Hatem Alkadhi

Received: 17 June 2011 /Accepted: 14 September 2011 /Published online: 9 February 2012# Springer Science+Business Media, LLC 2012

AbstractBackground Transseptal puncture (TSP) is the first step inpulmonary vein isolation and catheter ablation, as well as inleft atrial appendage closure in atrial fibrillation. AlthoughTSP has been reported to be successful in patients withdevice closure of interatrial septal communications, ques-tions pertinent to its feasibility in patients with large devicesstill remain. We sought to determine whether a “safe zone”for TSP could be visualised by computer tomography (CT),especially if larger device sizes for interatrial septalcommunication closure (IASC-C) had been used.Methods Retrospective observational study of 20 patientswho underwent CT for de novo chest pain occurring afterIASC-C or as a diagnostic test for suspected or provencoronary artery disease (CAD). Clinical follow-up was for20.5±17.6 (6–84) months. CT was done18±10 (2–28)weeks after IASC-C. Device size and dimensions of bothatria in the long and short axes were measured, as was theminimal distance of the device edge to the inferior andinferoposterior atrial floor.Results The calculatedminimal distance from the device edgeto the inferior aspect (at 6 o’clock) of the (right or left) atrialfloor was 7.2±6.5 (0–27) mmwhile that to the inferoposterioraspect (at 07:30 o’clock) was 5.3±4.2 (0–15) mm. In bothlocations, a distance of >6mmwas documented in ten patients

(50%) while in nine patients (45%) a space of <6 mm wasshown in both locations. There was no correlation betweenatrial dimensions or device size and minimal device distanceto either wall.Conclusion With the exception of cases with the smallestdevices (18 and 20 mm), neither device size nor atrialdimensions allow us to predict the feasibility of TSP inpatients with a clamshell-type interatrial septal device inplace, so that CT may be of help in determining whether asafe puncture space does exist in these patients.

Keywords Transseptal puncture . Interatrial septal deviceclosure . Left atrial appendage closure . Pulmonary veinisolation . Catheter ablation

While specific conditions are selectively treated surgicallyand/or interventionally, some of these therapies mayinterfere with or preclude future treatments of the same oranother organ. Theoretical examples are stripping ofvaricose veins, leaving arterial revascularisation as the onlyalternative for coronary bypass surgery, or stenting thewhole length of a coronary artery precluding its surgicalrevascularization in the event of repeated diffuse restenosis.The same theoretical concern may pertain to differenttreatment approaches used in atrial fibrillation (AF). About2% of the population is affected by AF; this prevalence willrise with increasing life expectancy. The risk of developingAF at least once in a lifetime is 25% in persons over40 years of age [1]. Percutaneous catheter ablation hasevolved as a valid treatment alternative to medication [1, 2].This approach depends on transseptal puncture (TSP) of theinteratrial septum. There has been some concern about thefeasibility of such a procedure in patients with an interatrial

P. Wagdi (*)Interventional Cardiology, HerzZentrum Hirslanden,Witellikerstrasse 36,8008 Zürich, Switzerlande-mail: wagdi@herzzentrum.ch

H. AlkadhiDepartment of Radiology, University Hospital,Rämistr 100,8091 Zürich, Switzerland

J Interv Card Electrophysiol (2012) 34:167–172DOI 10.1007/s10840-011-9625-6

septal device in place. Additionally, some patients experi-ence paroxysmal AF after interatrial septal communicationclosure (IASC-C) [3]. In the latter case, the dysrhythmiacan mostly be converted into sinus rhythm by a combina-tion of antiarrhythmics and cardioversion; AF rarelypersists [3]. Meier et al. [4] were the first to reportsuccessful TSP and catheter ablation of AF as well as leftatrial appendage closure (LAA-C) after IASC-C; however,the size of the devices in place (18 and 25 mm) was rathersmall. LAA-C is a new and potentially attractive alternativein patients with established AF and bleeding episodes underoral anticoagulants [5].

We practised TSP approach for LAA-C in a patient inwhom a device closure of an IASC had been performed. Thepatient had a medium-sized device (30-mm right atrial discdiameter) in place. We sought to determine whether a “safezone” for TSP could be visualised by computer tomography(CT), especially if larger device sizes had been used.

1 Patients and methods

Twenty patients in whom an IASC-C had been performedpreviously underwent CT. The IASC consisted of an atrialseptal defect in five patients and of a patent foramen ovale(PFO) in 15 patients, four of whom presented with a largeaneurysm and two with a cribriform defect. Balloon sizingrevealed a defect of 15.4±6.9 (6–32) mm. Althoughballoon sizing of PFO is not routine in all cath labs, theauthor does size all defects after undersizing based onechocardiograpic assessment alone may have contributed tothe embolization of a device [6]. Two of the patients in thisseries had no balloon sizing; they underwent IASC-C priorto implementation of routine sizing. Clinical characteristicsare shown in Table 1. In ten patients, CT was indicated forprecordial chest pain (we do not routinely perform coronaryangiography before IASC-C in asymptomatic patients). Thechest pain was severe and acute in seven of the patients andoccurred 18 weeks (±10; 2–28) after IASC-C. In fiveasymptomatic patients a pathological stress test duringfollow-up prompted CT. In one patient with progressive

dyspnoea, CT was indicated to rule out pulmonaryembolism or neoplasm. Four patients with previouslydocumented coronary artery disease (CAD) underwent CTfor follow-up. We recently demonstrated the reliability ofCT in the diagnosis of CAD in our respective institutions[7]. All CT scans were performed on a second generationdual-source CT scanner. Depending on the heart rate,patients were scanned in the high-pitch mode (heart rate,<62 bpm) or in the step-and-shoot mode (heart rate,>62 bpm). Radiation dose applied in this study was 3.1±1.2 (0.5–4.4) millisievert. Parameters that were analysedretrospectively include:

Maximal device size was noted as stated by the manufac-turer. For example, the maximal diameter of the (left atrial)disc of a 32-mm Amplatzer™ ASD device measures 46 mmwhile the right atrial disc measures 42 mm; whereas, the rightatrial disc of a 25-mm Amplatzer™ PFO device measures25 mm while the left atrial one measures 18 mm.

The following measurements were defined in the four-chamber view, analogous to suggested measurements usedin echocardiography [8]. Right atrial short axis: from themiddle of the septum perpendicularly to the lateral rightatrial wall. Right atrial long axis: from the origin oftricuspid annulus perpendicularly to the right atrial poste-rior wall. Left atrial short axis: from the middle of theseptum perpendicularly to the lateral left atrial wall. Leftatrial long axis: from the origin of mitral annulusperpendicularly to the left atrial posterior wall. Table 1summarises the atrial dimensions. The devices used were12 Amplatzer™, four Atriasept™, three Figulla™, and oneSolysafe™. Device types, sizes, and correlation to defectsize are shown in Table 2.

The minimum distance between the device and the atrialfloor (at a 6-o’clock position) in millimetres, as well as thedistance at a 7:30-o’clock position as shown in Fig. 1 whileTable 3 summarises the data. The 6 o’clock position waschosen because it corresponds to the intuitive puncture siteas described by Meier et al. [4]. The 7:30 o’clock positionwas chosen because it may offer an alternative site if theproximity of the device to the atrial floor precludes a TSP;furthermore, an inferior and slightly posterior puncture may

Table 1 Clinical charateristics of patients and measurements

Age in years Gender FU BMI Max DS

56±8 (38–70) 16 females 21±18 (6–84) 27±6.1 (19.5–42.6) 30.4±7.4 (18–46)

LA short echoa LA short CTa LA long CT RA long CT RA short CT

35±5.2 (27–48) 34±5.5 (24–46) 51.6±6 (44–65) 50.8±5.1 (42–63) 50.6±8 (36–64)

BMI body mass index, DS device size in mm (= maximal diameter of the largest disc as indicated by the manufacturer), CT measurement bycomputer tomography, FU follow-up in months after IASC-C, LA left atrium, Long long-axis dimension in millimetres, RA right atrium, Shortshort-axis dimension in millimetresa Correlation coefficient, r2 =0.52, p=0.07

168 J Interv Card Electrophysiol (2012) 34:167–172

Table 2 Defect size and devicetype used (nomenclature as bymanufacturer)

C cribriform defectaA 14-mm Amplatzer ASDdevice did not fitbLarger device than would beexpected was chosen because oflarge atrial septal aneurysm

Defect size and type Device used Defect size and type Device used

20-mm ASD 21-mm ASD Figulla™ 9-mm PFO 25-mm PFO Atriasept™

18-mm ASD 20-mm ASD Amplatzer™ 8-mm PFO 20-mm Solysafe™

14-mm ASD 14-mm ASD Amplatzer™ 12-mm PFOb 35-mm PFO Amplatzer ™

32-mm ASD 32-mm Atriasept™ 12-mm PFO 30-mm PFO Amplatzer™

34-mm ASD 32-mm ASD Amplatzer™ 10-mm PFO C 30-mm Atriasept™

6-mm PFO 18-mm PFO Amplatzer™ PFO; no sizing 25-mm PFO Amplatzer™

6-mm PFO 20-mm PFO Atriasept™ 14-mm ASD+C 18-mm ASD Amplatzer™

12-mm PFOb 27/30 PFO Figulla™ 12-mm PFOb 35-mm PFO Amplatzer™

11-mm PFO 25-mm PFO Amplatzer™ PFO, no sizing 27/30-mm PFO Figulla™

14-mm ASD 30-mm PFO Amplatzer™a 7-mm PFOb 25-mm PFO Amplatzer™

Sagittal plane of a CT scan showing an Amplatzer ™ 35 mm device in place, with no space for safe TSP on the right atrial side floor.

(b)

(a)

A 21 mm Amplatzer ™ASD device in place with no space for safe TSP inferiorly (small arrow) or inferoposteriorly (large arrow).

Fig. 1 (a) Sagittal plane of aCT scan showing an Amplat-zer™ 35-mm device in place,with no space for safe TSP onthe right atrial side floor. (b) A21-mm Amplatzer ™ASDdevice in place with no space forsafe TSP inferiorly (smallarrow) or inferoposteriorly(large arrow)

J Interv Card Electrophysiol (2012) 34:167–172 169

be useful—at least for LAA-C—in large atria. An arbitrarydistance of 6 mm was chosen as a cut-off value for TSPsuitability. This was based on the fact that some sheathsneeded for LAA-C gauge 12 French (=4 mm), allowing amargin of safety for puncture, although usually a needlepuncture at the lower end of the device stretches the septumcaudally to a greater degree allowing additional safemanoeuvring space.

All data were analysed using commercially availablestatistical software (StatView 5.0; ASA Institute Inc., Cary,NC, USA).

2 Results

Apart from information concerning the IASC closure, patencyof a previously implanted stent in the left anterior descendingcoronary artery (LAD) of one patient was confirmed, and intwo patients a de novo 60% LAD stenosis was diagnosed. Inall other patients, CAD was excluded, as was a pulmonarypathology in one of the patients.

Transthoracic echocardiographic and CT measurements ofleft atrial short axis correlated fairly (r2=0.52; Table 1). Lackof perfect correlation may be explained by the fact that CTand echocardiography are different imaging modalities,interobserver variability, and the fact that both measurementswere not taken simultaneously, some before, others afterIASC-C. The other dimensions (RA and LA long axis) werenot compared because they were not routinely available onechocardiographic studies. Furthermore, CT studies usuallycompare atrial volumina, not dimensions.

The calculated minimal distance from the device edge tothe inferior aspect (at 6 o’clock) of the (right or left) atrialfloor was 7.2±6.5 (0–27) mm while that to the inferopos-terior aspect (at 07.30 o’clock) was 5.3±4.2 (0–15) mm. Inboth locations, a distance of >6 mm was documented in tenpatients (50%) while in nine patients (45%) a distance of<6 mm was shown in both locations. There was nocorrelation between atrial dimensions and minimal devicedistance either to the posterior or to the inferoposterior wall.The best correlation was given by the right atrial short axisdimensions and this was already relatively weak (Fig. 2(a)). Also, there was no correlation between device size andits proximity to the atrial wall (Fig. 2 (b)). Patients with a

Table 3 Minimal measureddistance between device edgeand atrial floor

MD to AF minimal distancefrom device to atrial floor, Nnumber of patients, D>6 mmdistance over 6 mm

MD to inferior AF MD to inferoposterior AF

Distance (mm) 7.2±6.5 (0–27) 5.3±4.2 (0–15)

N (D>6 mm) 10 10

N (D<6 mm) 10 10

N (D>6 mm in both planes) 9

N (D<6 mm in both planes) 9

(a)

(b)

0

5

10

15

20

25

30

15 20 25 30 35 40 45 50Device max diam

0

5

10

15

20

25

30

35 40 45 50 55 60 65

RA short

Fig. 2 (a) Weak (r2=0.4) or no (r2=0.1) correlation between devicesize and its proximity to the inferior (blue circles) or inferoposterior(red circles) atrial wall. (b) Bivariate regression chart showing nocorrelation between right atrial short axis dimensions and the minimaldevice distance to posterior (blue circles) or inferoposterior (redcircles) atrial floor (r2=0.3 and 0.2, respectively)

170 J Interv Card Electrophysiol (2012) 34:167–172

device size of 18 (n=1) or 20 mm (n=1) both presenteddevice edge to atrial wall distances of ≥10 mm. In the fivepatients with a 25-mm device in place, both distancesmeasured <6 mm in both locations in one patient while twoshowed only one location with ≥6 mm. In the other twopatients, both locations offered >6 mm potential space forTSP. The question has been raised as to whether thedistance between device edge and atrial floor might bedifferent between left and right side in some patients. Thismight be the case, first of all because of anatomicalvariation, but also because left and right atrial discs of theAmplatzer™ and the Figulla™ devices are of differentdiameter: the right atrial discs are larger than the left ones inPFO devices, and vice versa for the ASD devices. Wetherefore decided to consider only the smaller of bothmeasurements. The minimal distance is the limiting factorfor TSP that has to be considered and overcome, indepen-dently of the side (Fig. 3).

3 Discussion

TSP is only the first step preceding LAA-C and AFablation, yet it is one of the potentially critical stages ofthe procedure. Although safe in experienced hands, thepresence of a device in the IAS may present an additionalchallenge in terms of limiting the space available forpuncture. Furthermore, once the puncture is successful,the freedom of movement of the LAA-C sheath as well asthat of the ablation catheters may be limited due to theextremely low position and may lead to an unsatisfactoryintervention. It is very probable that unsuccessful attemptsat TSP due to device presence are underreported or not

reported at all. One would expect a correlation betweendevice size and its proximity to the atrial wall on the onehand, and/or a correlation between atrial dimensions andthe device proximity to the wall on the other. Somewhatunexpectedly, CT failed to show such a correlation. Thismeans that neither device size nor atrial dimensions enableus to extrapolate the feasibility of TSP in patients with anatrial septal device in place. At times, transesophagealechocardiography yields reliable measurements (Fig. 3),yet in other circumstances an unequivocal affirmationcannot be obtained. We thus suggest CT prior to LAA-Cor pulmonary vein isolation in patients with an IAS devicein place. If the minimal distance between the device edgeand one of the two locations measured exceeds 6 mm, thena safe puncture may be expected. In cases where locationsshow a distance below 6 mm, then a surgical interventionwith LAA-exclusion and catheter ablation may be consid-ered [9]. Of course it is always possible to take the patientto the catheter laboratory without previous CT, yet bothLAA-C as well as pulmonary vein isolation and catheterablation are invasive, and at times challenging interventionsmay carry an even higher risk of adverse events if they haveto be performed in suboptimal circumstances even undertransoesophageal or intracardiac echocardiography guid-ance. Performing a CT prior to attempting TSP for AFablation and/or LAA-C seems thus to be a valid alternativeto taking the patient directly to the cath lab for an attempt.If CT shows no adequately safe space for TSP, theelectrophysiologist or the interventional cardiologist maywell choose not to attempt the procedure. CT is alreadybeing widely used before attempting LAA-C or pulmo-nary vein isolation and catheter ablation in patientswithout an atrial septal device in place, in order to better

RA

LA

Fig. 3 Intrainterventional TEEduring implantation of a smallsized device (white arrow),showing unequivocally a largeinferior space for potential TSP(between the two markingpoints). A case in which obvi-ously no CT scan is needed toanswer the question whether asafe space is available for TSP.LA left atrium, RA right atrium

J Interv Card Electrophysiol (2012) 34:167–172 171

define the morphology of the structures, so the extensionof the examination to clarify this crucial aspect should notraise any additional questions in terms of appropriateness,cost or radiation exposure. Some special consideration isof course warranted during TSP with an IASC device inplace. For example, if the device has been implantedrecently (i.e. before complete endothelialisation and firmembedding), it may get partially distorted or dislocated if theprocedure is not conducted with special care. It has beensuggested that intracardiac echocardiography may have arole to play in locating a “safe” spot for TSP, one of its majorlimitations in our practice is its prohibitive price; furthermoreour institution does not allow re-sterilisation for second use.

One acknowledged limitation of the study is of coursethe arbitrary choice of 6 mm as a “cut-off” value forperforming “safe” TSP. We do not dispose of clinical data(even as a personal communication) concerning patientswith an IASC device in place, in whom an attempt for TSPwas unsuccessful. Thus validity of both, our approach andour statement concerning the safe zone, needs to beconfirmed by the publication of data of centres confrontedwith the problem. The fact that none of the patientsexamined by CT for feasibility of TSP, experienced AFduring follow-up, may represent a selection bias. It may beargued that they may represent a different population thanthe patients with paroxysmal AF selected for PV isolation.

4 Conclusions

In patients with large-sized IAS devices for whom catheterablation of AF and/or implantation of a LAA-C device areindicated, prior CT yields valuable information as to thefeasibility of TSP. If adequate safety margins for TSP are

not documented on CT, minimal invasive surgery may beconsidered for pulmonary vein isolation and LAA-C.

Acknowledgements This work is dedicated to C. Rupp.

References

1. Camm, J. A., Kirchhof, P., Lip, G. Y. H., Schotten, U., Savelieva.I., Ernst S, et al. (2011). Guidelines for the management of atrialfibrillation. The task force for the management of atrial fibrillationof the European Society of Cardiology (ESC) of the EuropeanSociety of Cardiology. European Heart Journal (in press).

2. Weerasooriya, R., Khairy, P., Litalien, J., Macle, L., Hocini, M.,Sacher, F., et al. (2011). Catheter ablation for atrial fibrillation.Journal of the American College of Cardiology, 57, 160–166.

3. Wagdi, Ph. (2010). Incidence and predictors of atrial fibrillationfollowing transcatheter closure of interatrial septal communicationsusing contemporary devices. Clinical Research in Cardiology, 99,507–511.

4. Zaker-Shahrak, R., Fuhrer, J., & Meier, B. (2008). Transseptalpuncture for catheter ablation of atrial fibrillation after deviceclosure of patent foramen ovale. Catheterization and CardiovascularInterventions, 71, 551–552.

5. Holmes, D. R., Reddy, V. Y., Turi, Z. G., Doshi, S. K., Sievert, H.,Buchbinder, M., et al. (2009). Percutaneous closure of the left atrialappendage versus warfarin therapy for prevention of stroke inpatients with atrial fibrillation: a randomised non-inferiority trial.Lancet, 374, 534–542.

6. Wagdi, Ph. (2011). Interatrial septal communication closure:adverse events and lessons learned. Cardiology Research, 2, 7–15.

7. Wagdi, Ph., & Alkadhi, H. (2010). The impact of cardiac CT on theappropriate utilization of catheter coronary angiography. Interna-tional Journal of Cardiovascular Imaging 26(3):333–344.

8. Feigenbaum, H., Armstrong, W. F., & Ryan T. (2005). Echocar-diography, 6th edition. Baltimore: Lippincot Williams and Wilkins,pp 184–195.

9. Wagdi, Ph., & Siebenmann, R. (2011). Unkonventionelle Hybrid-Intervention bei etabliertem Vorhofflimmern: optimierte Behand-lung durch interdisziplinäre Zusammenarbeit. KardiovaskuläreMedizin 14, 159–162.

172 J Interv Card Electrophysiol (2012) 34:167–172