Catheterization and Cardiovascular Interventions 69:1007–1014 (2007)

Original Studies

Feasibility and Clinical Impact of Transcatheter Closure

of Interatrial Communications After a Fenestrated
Fontan Procedure: Medium-Term Outcomes
Jaana Pihkala,1 MD, Satoshi Yazaki,1 MD, Rohit Mehta,1 MD, Kyong-Jin Lee,1 MD, FRCPC,
Rajiv Chaturvedi,2 MD, Brian W. McCrindle,1 MD, MPH, FRCPC, Glen Van Arsdell,2 MD, FRCS,
and Lee N. Benson,1* MD, FRCPC, FACC
Objective: This study was to review an institutional experience with transcatheter clo-
sure of Fontan fenestrations and its impact on clinical care. Background: An interatrial
fenestration improves postoperative outcomes in high-risk children undergoing a Fon-
tan repair. While technical feasibility has been well defined, the clinical impact of
subsequent closure is not well defined. Methods: Transcatheter closure of a surgically
created or additional interatrial communication was attempted in 152 children at a
median interval of 13.8 months after surgery. The clinical records were reviewed for de-
mographic and anatomical characteristics, previous surgeries; catheterization data,
and status at latest follow-up. Results: Mean oxygen saturation and right atrial pres-
sure increased acutely from 87% 6 5% to 96% 6 3% (P < 0.001) and 12 6 2 mm Hg to
13 6 3 mm Hg (P < 0.001), respectively. Higher systemic venous atrial pressures after
occlusion correlated with higher pulmonary artery pressures (P ¼ 0.05) before the Fon-
tan procedure and with higher right (P < 0.001) and left atrial (P ¼ 0.001) and ventricu-
lar end-diastolic pressures (P < 0.001) immediately before occlusion. Complications
included device malposition in 2 children, 1 child each had an air embolism and post-
procedural bleeding, and each self-limiting and 1 child had acute ST elevation in infe-
rior ECG leads because of occlusion of the acute marginal branch which was treated
with angioplasty and placement of a stent. At follow-up (median 4.5 years), the mean
oxygen saturation was 95% 6 3%. Residual interatrial leaks were noted echocardio-
graphically in 9%. Two children developed protein-losing enteropathy after fenestration
closure. No deaths or strokes were observed in follow-up. Conclusions: Transcatheter
occlusion of Fontan fenestrations is safe with acute and persistent improvements in
oxygen saturations. ' 2007 Wiley-Liss, Inc.

Key words: fontan procedure; interventional catheterization; congenital heart disease;

pediatric cardiology

1
Department of Pediatrics, Division of Cardiology, The Variety
INTRODUCTION Club Cardiac Catheterization Laboratories, The Hospital for
Sick Children, The University of Toronto School of Medicine,
The Fontan operation is the definitive, conventional Toronto, Ontario, Canada
palliative surgical procedure for children with complex 2
Division of Cardiovascular Surgery, The Hospital for Sick
univentricular heart lesions. Reduced mortality and Children, The University of Toronto School of Medicine,
lessened morbidity rates have been achieved by creat- Toronto, Ontario, Canada
ing an interatrial fenestration at the time of surgery, *Correspondence to: Dr. Lee Benson, The Hospital for Sick Chil-
preserving cardiac output and decreasing the systemic dren, 555 University Avenue, Toronto, Ontario, Canada M5G 1X8.
venous pressure in the early postoperative period [1– E-mail: [email protected]
4]. Subsequent to surgery, the fenestration may close
spontaneously, can be closed using a snare technique Received 3 August 2006; Accepted 4 October 2006
[5] or in the catheterization laboratory utilizing a variety DOI 10.1002/ccd.20995
of devices [1,6–8]. In this regard, little data exits on Published online 9 April 2007 in Wiley InterScience (www.
the longer-term outcomes of transcatheter fenestration interscience.wiley.com).

' 2007 Wiley-Liss, Inc.

1008 Pihkala et al.

closure [9,10]. The purpose of this study was to review number of devices implanted, additional interventions
our institutional experience with transcatheter closure performed, complications, and duration of fluoroscopy.
of Fontan fenestrations and to determine individual Data documented at the latest follow-up assessment
child and procedural related factors associated with included child growth, oxygen saturation by pulse oxi-
early and late outcomes over a 14 year experience. metry, presence of residual interatrial shunts (docu-
mented by echocardiography), evidence of stenosis
within the Fontan circuit, presence of effusions or vas-
METHODS
cular thromboses, subsequent interventions, or deaths.
Study Population
The study population consisted of 152 children (84 Fenestration Occlusion
males) who had previously undergone a fenestrated Written informed consent for cardiac catheterization
Fontan procedure, and subsequently a postoperative and intervention was obtained from each parent (or
cardiac catheterization and attempted transcatheter child as appropriate). The catheterization was per-
device closure of the interatrial communication bet- formed under general anesthesia in room air. Femoral
ween December, 1992 and February, 2006. Children or jugular venous access was obtained and a complete
were screened in the outpatient setting for functional hemodynamic assessment of the Fontan circuit was
(decreased ventricular ejection fraction) or other ana- performed in each child. Systemic venous saturations
tomical concerns (i.e., pulmonary artery hypoplasia and pressures were measured throughout the Fontan
or stenosis) by echocardiography, which would exclude pathway. Angiography was performed in superior and
them from fenestration closure. At catheterization, inferior caval veins and pulmonary arteries to visualize
those with unfavorable hemodynamics during a test the Fontan pathway, surgical fenestration, additional
balloon occlusion of the fenestration (>4 mm Hg rise interatrial leaks, and possible venous collateral vessels.
in systemic venous atrial pressure, >10% decrease in Subselective injections were obtained in the innomi-
mixed venous saturation or systemic arterial pressure) nate vein (when present) and right hepatic vein to fur-
were also ineligible for device closure and excluded. ther define the presence of venovenous shunting. A
balloon-tipped side-hole catheter (Berman, Critikon,
Data Set USA) was subsequently passed across the fenestration
The medical records were reviewed for demographic and test occlusion of the fenestration performed while
and anatomical characteristics, previous surgeries and monitoring systemic venous pressure and oxygen satu-
catheterizations. Findings noted at the cardiac catheter- ration, systemic arterial oxygen saturation, and blood
ization before the Fontan procedure included hemody- pressure [2]. Fenestration occlusion was undertaken in
namic measurements, oxygen saturations, and pulmo- the absence of a significant (>4 mm Hg) increase in
nary artery dimensions, the Mair et al. [11], Nakata mean systemic venous pressure or reduction in oxygen
et al. [12], and McGoon and coworkers [13] indices, saturation (>10%). All children received heparin sul-
and the presence of subaortic obstruction and pulmo- phate 150 IU/kg (5,000 IU maximum) and antibiotic
nary artery stenosis or distortion. Ventricular function prophylaxis (cephazolin 40 mg/kg) at the beginning of
(ejection fraction) as assessed either echocardiographi- the study. After fenestration occlusion, an angiogram
cally or angiographically by volumetric analysis were was performed in the neo-right atrium (Fontan baffle)
detailed. to assess adequacy of closure.
Surgical data included the type of material used in
the Fontan connection as well as the type and size of Data Analysis
the fenestration. Duration of hospitalization and use of Data are described as frequencies, medians with
anticoagulation therapy was also noted. ranges, and means with standard deviations. Where
At the time of the interventional catheterization, the there was missing data the number of non-missing val-
type of interatrial communication (surgical fenestration, ues is given. Changes in oxygen saturations and hemo-
suture line leak, or direct venous connection between dynamic measurements before and after fenestration
right and left atrium) was determined, as was the pres- closure were tested with paired t tests. Factors associ-
ence of right-to-left shunting venous collateral vessels. ated with higher pressures in the Fontan baffle after
Oxygen saturations and hemodynamic measurements fenestration occlusion were sought with Fisher’s exact
were detailed before, during temporary balloon occlu- tests, v2 tests, and correlation coefficients as appropri-
sion, and after fenestration closure (see below). Oxy- ate. Additionally, factors associated with the presence
gen saturation data was used, as oxygen tension data of residual leak after device implantation were sought
was available only on a limited number of children. with a Fisher’s exact test, v2 test, Kruskal–Wallis anal-
Noted technical features included the type, size, and ysis of variance and t-tests as appropriate. All analyses
Catheterization and Cardiovascular Interventions DOI 10.1002/ccd.
Published on behalf of The Society for Cardiovascular Angiography and Interventions (SCAI).
 Outcomes of Fontan Fenestration Closure 1009

were performed using StatView Version 9.1 statistical The mean size of the fenestration was 4.2 6 0.5 mm
software (SAS Institute, Cary, N.C.), and a P < 0.05 (n ¼ 142). Warfarin was administered to 117 (85%)
was set as the level of statistical significance. children, acetylsalicylic acid to 19 (14%), and both to
1 (1%) child (n ¼ 137) after hospital discharge.
RESULTS Postoperative Catheterization Procedure
Patient Characteristics Attempted transcatheter closure of the interatrial
The underlying diagnoses of the children are listed communication was performed at a median 13.8
in Table I. months (range, 1.1–113.1 months, n ¼ 152) after the
Fontan procedure with a median child weight of 16.0
Fontan Procedure Risk Factors kg (range, 9.5–62 kg, n ¼ 147). The mean room air
Cardiac catheterization before the Fontan operation oxygen saturation, by finger tip oximetry, before the
occurred at a median age of 2.4 years (range, 1.0–15.4 catheterization was 87% 6 5% (n ¼ 148) (Fig. 1),
years, n ¼ 133), with a median weight of 12.4 kg mean hemoglobin concentration 136 6 16 g/l (n ¼
(range, 8.4–84.5 kg, n ¼ 136). Calculated pulmonary 124). Transthoracic echocardiography performed before
artery resistance was >2 Wood units in 20 (24%) chil- catheterization detected the persistence of an interatrial
dren, with a mean of 1.5 6 0.9 Woods units (n ¼ 85). communication in 104 children (91%), while it was
The calculated mean Mair index was 2.4 6 1.0 (n ¼ unable to define a communication in 10 children (9%,
60), the mean Nakata index 280 6 146 (n ¼ 98), and n ¼ 114). Two children (1.3%) had a history of a
the mean McGoon index 2.0 6 0.9 (n ¼ 92). Mean stroke in the perioperative period after the Fontan pro-
ventricular ejection fraction was 62 6 8% (n ¼ 62) cedure, and 3 (2%) children had evidence of a clot in
and the ejection fraction was <50% in 4 (6%) chil- the Fontan circuit at the time of the catheterization.
dren. The calculated mean cardiac index was 4.8 6 All these children had been on anticoagulation therapy
1.9 ml/min/m2 (n ¼ 101). (4 on warfarin, 1 on acetylsalicylic acid) since surgery.
The hemodynamic findings at catheterization before
Fontan Procedure and after fenestration occlusion are presented in
Table II.
The median age at the Fontan procedure was 3 years
Angiographically, a single surgical fenestration was
(range, 1.3–15.4 years, n ¼ 144) with the median
seen in 129 children (86%, n ¼ 150). In addition to
weight of 13.1 kg (range, 8.6–62.8 kg, n ¼ 135). The
the identified surgical fenestrations, 11 (7%) children
type of the Fontan connection was interatrial lateral
had suture line leaks and 4 (3%) children had venous
tunnel or interatrial grafting in 50 (34%) children, and
collateral vessels (thebesian veins) shunting from right-
an extracardiac tunnel in 98 children (66%, n ¼ 148).
to-left atrium. Three (2%) additional children had only
Foreign material (polytetrafluoroethylene or bovine
suture line defects and no surgical fenestration identi-
pericardium) was used for the connection in 109
fied. Venovenous collateral vessels with right-to-left
(77%) children, aortic or pulmonary homograft in 30
shunting were seen in 67 children (45%, n ¼ 150) and
(21%), and pericardium in 3 children (2%, n ¼ 142).
subaortic obstruction in 3 children. In all children, atri-
oventricular valve regurgitation was mild or nil.
TABLE I. The Underlying Anatomic Diagnosis and the
Fenestration occlusion was performed with a variety
Predominant Ventricular Morphology of the Children of implants listed in Table III. An additional interven-
Undergoing Occlusion of Fontan Fenestration tion was performed at the same catheterization in 59
Diagnosis n (39%) children and included embolization of venous
collateral vessels using coils in 38 (25%), and dilation
Double inlet left ventricle 34
Tricuspid atresia 29 of the Fontan circuit in 15 (10%) (placement of a stent
Hypoplastic left heart syndrome 21 in 8, redilation of existing stents in 4, balloon dilation
Pulmonary atresia with intact ventricular septum 15 in 3). Additionally, in 1 child each had a balloon
Complex form of double outlet right ventricle 16 angioplasty of a coarctation of the aorta, stenosis of
Common inlet right ventricle 9
Unbalanced atrioventricular septal defect 7
the left pulmonary artery and main pulmonary artery,
Complex transposition of the great arteries 7 stenting of the right pulmonary artery, coil occlusion
Other complex lesions 14 of an aortopulmonary collateral artery, a pulmonary to
The predominant ventricular morphology ventricular connection, a communication between the
Left 89 (59%) left superior caval vein and coronary sinus and com-
Right 42 (27%) plete occlusion of a left superior vena cava with an
Indeterminate 21 (13%)
Amplatzer device. Test fenestration occlusion was
Catheterization and Cardiovascular Interventions DOI 10.1002/ccd.
Published on behalf of The Society for Cardiovascular Angiography and Interventions (SCAI).
1010 Pihkala et al.

Complications
Device malposition occurred in 2 children: in 1, a
12 mm RashkindÔ device embolized to the innomi-
nate artery and was successfully retrieved with a snare
catheter. Another 12 mm RashkindÔ device embolized
to the femoral artery and was retrieved surgically. In
this child, additional surgical repair of the femoral ar-
tery was required 2 months after the catheterization. In
both children, successful occlusion of the fenestration
using a 17 mm RashkindÔ device was carried out sub-
sequent to the retrieval. Other complications included
1 child each with an air embolism of a coronary ar-
tery, post-procedural bleeding and 1 child with acute
ST elevation in inferior ECG leads at the end of the
procedure due to an occlusion of the distal aspect of
the acute marginal branch, which was treated with
angioplasty and placement of a stent. The overall com-
Fig. 1. Systemic arterial saturation measured in room air im-
plication rate was 3.3% (5 of 152). The median fluo-
mediately before and after occlusion of the fenestration and at roscopy time was 31 min (range, 13–96 min, n ¼ 118)
the latest follow-up (F–U). At follow-up, the children are and median hospital stay after the procedure was 1
grouped according to the presence or absence of residual day (range, 1–9 days, n ¼ 148).
leak across the interatrial septum.

Factors Associated With Higher Systemic Venous

Pressure After Occlusion
performed after all such procedures to qualify for Systemic venous pressure immediately after occlu-
closure. All children were able to have subsequent sion was 15 mm Hg or above in 27 children (21%, n
closure. ¼ 138), and correlated significantly with a higher pul-
Mean oxygen saturations in room air increased from monary artery pressure (r ¼ 0.20; P ¼ 0.05, n ¼ 94)
87% 6 5% to 96% 6 3% (P < 0.001, n ¼ 148) (Fig. at the catheterization performed prior to Fontan sur-
1, Table II), and mean systemic venous atrial pressure gery. It was not significantly related to other known
increased from 11.9 6 2.3 mm Hg to 12.8 6 2.6 mm Fontan procedure risk factors, ventricular morphology,
Hg (P < 0.001, n ¼ 148) (Fig. 2, Table II). The the age at the Fontan procedure, or the interval from
increase in oxygen saturation in children with addi- the Fontan procedure to the catheterization for occlu-
tional collateral occlusions did not differ from those sion. Higher systemic venous pressures after the occlu-
children that did not have abnormal venous channels. sion correlated significantly with higher pressures
Immediately after the procedure (within 5 min), a re- immediately before occlusion, including higher baffle
sidual leak across the atrial septum was seen in 78 pressure (r ¼ 0.86, P < 0.001, n ¼ 103) (Fig. 2), left
(58%) children. After the procedure, acetylsalicylic atrial pressure (r ¼ 0.68, P < 0.001, n ¼ 83), and ven-
acid was administered to 39 (26%) children, warfarin tricular end-diastolic pressure (r ¼ 0.53, P < 0.001, n
to 18 (12%), and both in 10 (7%) children. The major- ¼ 54), but not related to age at catheterization, arterial
ity of children (55%, n ¼ 83) received no anticoagula- oxygen saturations before occlusion or the presence of
tion or antiplatelet therapy after the catheterization. venous collateral vessels or ventricular morphology.

TABLE II. The Hemodynamic Findings Before and After Occlusion of the Fontan
Fenestration
Before After
Arterial oxygen saturation 87 6 5% (n ¼ 152) 96 6 3% (n ¼ 148)*
Left atrial pressure 7.6 6 3.0 mm Hg (n ¼ 107)
Systemic venous atrial pressure 12.0 6 2.4 mm Hg (n ¼ 148) 12.7 6 2.6 mm Hg (n ¼ 148)*
Aortic pressure 88 6 12 mm Hg (n ¼ 116)
Pulmonary artery pressure 11.6 6 2.4 mm Hg (n ¼ 104)
Ventricular end-diastolic pressure 8.0 6 3.1 mm Hg (n ¼ 73)
The values are given as mean 6 SD,*P < 0.001.
Catheterization and Cardiovascular Interventions DOI 10.1002/ccd.
Published on behalf of The Society for Cardiovascular Angiography and Interventions (SCAI).
 Outcomes of Fontan Fenestration Closure 1011

TABLE III. Type and Size of Occluder Device

Device Size Number of children
Single implants
Rashkind 17 mm 45
12 mm 30
Cardioseal 23 mm 2
17 mm 15
Amplatzer Septal Occluder 8 mm 5
6 mm 4
5 mm 29
4 mm 5
PDA coil 12 mm 4
5 mm 1
Gianturco coil 5 cm–5 mm 1
Device Size and number of devices Number of children
Multiple similar implants
Rashkind 17 mm 3 2 2
12 mm 3 2 1
Gianturco coil 5 cm–5 mm 3 1 1 5 cm–3 mm 3 1 1
Multiple different implants (1 child each)
Rashkind 12 mm 3 2 1 Rashkind 17 mm 1 Amplatzer Septal Occluder 10 3 mm
Rashkind 17 mm 3 2 1 Amplatzer Duct Occluder 10 mm
Rashkind 12 mm 3 2 1 Detachable PDA coil 8 mm-4 loops
Amplatzer Septal Occluder 4 mm 1 Detachable PDA coil 5 mm–5 loops 3 3

residual leak and 95% 6 3% in children (n ¼ 113)

without an identified residual leak (P ¼ 0.55) (Fig. 1).
Reintervention was performed for a residual leak in 1
child. Multiple interatrial communications had been
seen at the time of the primary fenestration occlusion,
and 2 RashkindÔ umbrellas had been implanted. At
the second procedure, a third Rashkind umbrella and a
coil were implanted across a deficient suture line. No
residual leaks were noted at the latest follow-up. From
transthoracic echocardiography, there was no evidence
of stenosis within the Fontan circuit in any child (n ¼
123) or a history of suspected cerebrovascular acci-
dent, distant from fenestration closure (n ¼ 123). Two
(1.6%) children developed protein-losing enteropathy
(PLE) after fenestration closure, 1 developing it 4
months after the occlusion, stabilizing on diuretics.
The other child developed PLE 39 months after the
Fig. 2. Relation between Fontan baffle pressure measured occlusion, requiring corticosteroids, diuretics, and hep-
before and after occlusion of the Fontan fenestration. (There arin therapy for control. The systemic venous pressures
are multiple overlapping data points. of these 2 children after fenestration closure were 16
and 14 mm Hg respectively. There was 1 child who
developed PLE soon after the Fontan procedure with
Follow-up After Device Closure recovery, but closure of fenestration 14 months after
At the latest follow-up, a median interval of 4.5 the Fontan operation did not result in a relapse. There
years (range, 7 days to 11.4 years, n ¼ 125) after the were no deaths in the 123 children who were seen in
catheterization, the mean weights of the children were follow-up.
34 6 16 kg (n ¼ 117) and a residual interatrial leak
was noted in 11 children (9%, n ¼ 125) by echocardi-
ography. Overall, mean arterial oxygen saturations Factors Associated With Residual Leaks
measured by fingertip oximetry were 95% 6 3% (n ¼ At follow-up (n ¼ 123), 11 children were noted to
124), 92% 6 3% in those children (n ¼ 11) with a have persistent residual interatrial leaks. The mean
Catheterization and Cardiovascular Interventions DOI 10.1002/ccd.
Published on behalf of The Society for Cardiovascular Angiography and Interventions (SCAI).
1012 Pihkala et al.

follow-up interval in those children (2.0 6 1.3 years, n original descriptions, a number of different devices
¼ 11) was shorter than that in children without a leak have been used for fenestration occlusion, including
(5.3 6 3.5 years, n ¼ 112, P ¼ 0.42). The presence of the RashkindÔ umbrella [18], GianturcoÔ coils [6],
a leak was not significantly related to characteristics at and the Amplatzer Septal OccluderÔ [7]. Despite
the catheterization prior to the Fontan or age at the observations made by others, this study found that
Fontan procedure. Leaks were less prevalent with a these devices can be used in both intracardiac and ex-
single surgical fenestration (5%, n ¼ 139) versus other tracardiac Fontan fenestrations [19].
types of communications (13%, n ¼ 18; P ¼ 0.04). Higher systemic atrial pressures were observed after
Persistent leaks were not related to the type of connec- occlusion of the interatrial communication [16,18] that
tion or material used for baffle construction or the size were related to higher pulmonary arterial pressures at
of the fenestration. Leaks were seen in 9 (11%) of 78 the time of the catheterization performed before the
children with a Rashkind device, 2 (5%) of 43 children Fontan procedure. The other known risk factors for
with an Amplatzer Septal OccluderÔ device, and not mortality and morbidity after a Fontan procedure was
seen in children with a CardiosealÔ device or coils not predictive of higher central venous pressures or
however, there were no significant differences amongst outcomes after occlusion of the interatrial communica-
these rate’s. Likewise, leaks were not significantly tion.
related to size or number of devices placed. Systemic Our study confirms the feasibility and safety of
venous pressure and anticoagulation therapy after de- transcatheter closure of interatrial communications after
vice implantation was not associated with the presence Fontan type surgery. The fenestration was device
of residual leaks. occluded successfully in all children, consistent with
that reported previously [9,10,14,16,18]. The rate of
complete closure was also comparable to that previ-
DISCUSSION
ously published [1,6,9,10,18]. In 58% of children, re-
The most common factors associated with undesir- sidual interatrial shunting was seen immediately after
able outcomes after a Fontan procedure are poor ven- the device implantation, but only 9% of children had
tricular function, hypoplasia or distortion of the pulmo- residual interatrial shunts at the time of the last fol-
nary arteries, and increased pulmonary vascular resist- low-up. Despite residual shunting, the arterial oxygen
ance which, in an atrial dependent circulation, can lead saturation increased significantly immediately after the
to a low cardiac output and high systemic venous pres- occlusion, although differences were noted in systemic
sures [1,3,14–16]. Earlier, nonrandomized studies dem- arterial oxygen saturation between those children with
onstrated decreased mortality and morbidity rates after and without residual leaks at the last follow-up visit
a Fontan procedure if an interatrial communication (Fig. 1). Higher systemic atrial pressures after occlu-
was created at the time of surgery [1–4]. Creation of sion of the interatrial communication were related to
an atrial fenestration allowed a variable degree of the pressures immediately before occlusion (Fig. 2)
right-to-left shunting in the immediate postoperative and no late deaths or device associated strokes were
period, presumably preserving ventricular preload and observed, while 2 children developed PLE. In previous
cardiac output while limiting the rise in central venous studies, a higher baffle pressure has been shown to be
pressure. Although a moderate degree of arterial desa- a predictor of late morbidity and lower functional class
turation results, theoretically the increased cardiac out- [1,16]. Unfortunately, in this study, most of the chil-
put compensated has a favorable effect on oxygen dren were too young to undergo exercise testing,
transport. which would have added important information regard-
The factors contributing to early difficulties in the ing functional status. This review confirms earlier
postoperative course may improve with time after the observations [9], that mortality is not increased, there
Fontan repair. As such, late occlusion of the surgical is no increased need for anticoagulant therapy after
fenestration achieves separation of the systemic and defect closure, and the overall health of the population
pulmonary venous circulations, \completing" the Fon- remains good without an increased incidence of signs
tan pathway, at the expense of a reduction in cardiac or symptoms of clinically significant systemic venous
index and tissue oxygen delivery [16,17]. The concept congestion or PLE.
of transcutaneous fenestration occlusion was applied to Postoperative cardiac catheterization after Fontan
children with univentricular hearts by Laks and cow- surgery is encouraged, not only in the children with a
orkers, using a snare device encircling the interatrial known fenestration but to examine potential areas for
communication [5], and the method of transcatheter residual shunts or pulmonary artery stenosis, and when
occlusion using a ClamshellÔ double umbrella device noted, these can be addressed with transcatheter techni-
was first described by Bridges et al. [1]. Since these ques [4,20]. The etiologies of such right-to-left shunting
Catheterization and Cardiovascular Interventions DOI 10.1002/ccd.
Published on behalf of The Society for Cardiovascular Angiography and Interventions (SCAI).
 Outcomes of Fontan Fenestration Closure 1013

are variable, and include residual shunting at atrial related to the catheterization were transient and with-
level (surgical fenestrations, suture line leaks, or collat- out consequence.
eral or thebesian vessels), the development of systemic In summary, we found that a catheter based strategy
venous collaterals with connection to the pulmonary for closure of Fontan baffle right-to-left shunt results
veins or left atrium, the development of pulmonary ar- persistent shunt reduction and improved oxygen satura-
teriovenous fistulae (rare), or unrecognized connection tions in the longer-term (mean 4.5 years, up to 14
of a left superior caval or phrenic vein to left atrium, years follow-up). A baffle fenestration or other sources
or to the coronary sinus incorporated within the pul- of right-to-left shunting can be safely eliminated if the
monary venous compartment [21]. At our institution, child is desaturated at rest (<92%) or demonstrates
systemic venous collaterals have previously been significant desaturation with exercise, in the absence of
shown to develop in 31% of children after a bidirec- a significant (>4 mm Hg) increase in mean systemic
tional cavopulmonary anastomosis, and found more venous pressure or reduction in mixed venous oxygen
commonly in those children with higher postoperative saturation (>10%) during a test occlusion. Late out-
gradients between the superior caval vein and right come complications were rare. Slightly higher systemic
atrium [22]. In this regard, 23% of children had multi- venous pressures can be anticipated after defect closure
ple interatrial shunts and 40% had additional systemic in those children with higher pulmonary artery pres-
venous collateral vessels shunting right-to-left. Such sures before surgery and higher right and left atrial
collateral vessels were embolized at the time of the and ventricular end-diastolic pressures immediately
fenestration occlusion in 26% of children. Interestingly, before occlusion. However, they appear not to have an
no differences in the increase in oxygen saturation increased incidence of clinically significant symptoms
were noted in children who had an additional collateral of systemic venous congestion, which in part may be
occlusion as compared to the children with fenestration due to the timing of closure [9].
occlusion only. Study limitations inherent in a retrospective study
Children after a fenestrated Fontan procedure, they and the manner in which children were selected and
have both persistent right-to-left shunts and a tendency treated were subject to physician bias. Selection for
to form venous thrombi, may be at increased risk for fenestration closure was also biased, as only children
paradoxical embolic events [23–28] and transcatheter with non-obstructive systemic venous pathways and ac-
occlusion of the fenestration may decrease the risk of ceptable ventricular function had attempted closure. As
such systemic thromboembolization. It has been sug- such, this cohort represents the \best" outcome group.
gested that prophylactic anticoagulation therapy with Furthermore, the entire cohort of fenestrated Fontan
warfarin should continue for 4–6 months in children at procedures was not analyzed, so the incidence or risk
increased risk for postoperative thrombosis after the of spontaneous fenestration closure could not be
Fontan procedure. For infants and young children at addressed. Use of an atrial baffle fenestration with sub-
low risk for postoperative thrombosis, acetylsalicylic sequent transcatheter closure, while thought to be ad-
acid has been recommended for 6 months postopera- vantageous for children at increased risk has not been
tively [23]. In this study, 2 children with a history of a decisively proven. The long-term benefits including
postoperative stroke and 3 with thrombus found at functional status before and after transcatheter occlu-
catheterization had been on anticoagulation therapy. So sion requires further definition. Although children tol-
far, no specific and reliable risk factors for thrombus erate an open fenestration well, a patent interatrial
formation have been identified and no guidelines communication is not without risks, including acute
regarding anticoagulation based on controlled trials and chronic hypoxemia at rest, intensified with exer-
have been published [24]. Routine evaluation with cise, and the potential for paradoxical thromboembolic
transthoracic echocardiography for detection and subse- stroke.
quent treatment of asymptomatic thrombi has previ-
ously been reported to improve outcome and reduce
stroke rate. However, transthoracic echocardiography
may fail to detect some thrombi and therefore, trans- REFERENCES
esophageal echocardiography may prove more useful 1. Bridges ND, Lock JE, Castaneda AR. Baffle fenestration with
technique [25–28]. subsequent transcatheter closure. Modification of the Fontan
A variety of minor complications occurred at the operation for patients at increased risk. Circulation 1990;82:
interventional catheterization. In 2 children, we were 1681–1689.
2. Bridges ND, Lock JE, Mayer JE Jr, Burnett J, Castaneda AR.
unable to achieve a stable position of the device across Cardiac catheterization and test occlusion of the interatrial com-
the fenestration at first, but successful retrieval and munication after the fenestrated Fontan operation. J Am Coll
reimplantation was performed. Other complications Cardiol 1995;25:1712–1717.
Catheterization and Cardiovascular Interventions DOI 10.1002/ccd.
Published on behalf of The Society for Cardiovascular Angiography and Interventions (SCAI).
1014 Pihkala et al.

3. Gentles TL, Mayer JE Jr, Gauvreau K, et al. Fontan operation in 16. Senzaki H, Naito C, Kobayashi T, et al. Hemodynamic evalua-
five hundred consecutive patients: Factors influencing early and tion for closing interatrial communication after fenestrated
late outcome. J Thorac Cardiovasc Surg 1997;114:376–391. Fontan operation. J Thorac Cardiovasc Surg 2001;121:1200–
4. Kopf GS, Kleinman CS, Hijazi ZM, Fahey JT, Dewar ML, Hel- 1202.
lenbrand WE. Fenestrated Fontan operation with delayed trans- 17. Coles JG, Kielmanowicz S, Freedom RM, et al. Surgical experi-
catheter closure of atrial septal defect. Improved results in high- ence with the modified Fontan procedure. Circulation 1987;76(3):
risk patients. J Thorac Cardiovasc Surg 1992;103:1039–1048. III61–III66.
5. Laks H, Pearl JM, Haas GS, et al. Partial Fontan: Advantages of 18. Redington AN, Rigby ML. Transcatheter closure of interatrial
an adjustable interatrial communication. Ann Thorac Surg 1991;52: communications with a modified umbrella device. Br Heart J
1084–1095. 1994;72:372–377.
6. Sommer RJ, Recto M, Golinko RJ, Griepp RB. Transcatheter 19. Kim SJ, Lee JY, Song JY, Kim WH. Catheter closure of tube
coil occlusion of surgical fenestration after Fontan operation. graft fenestration in extracardiac Fontan operation. Int J Cardiol
Circulation 1996;94:249–252. 2005;104:224–226.
7. Tofeig M, Walsh KP, Chan C, Laudsans E, Gladman G, Arnold 20. Hsu H, Nykanen D, Williams WG, Freedom RM, Benson LN.
R. Occlusion of Fontan fenestrations using the Amplatzer septal Right-to-left interatrial communications after the modified Fon-
occluder. Heart 1998;79:368–370. tan procedure: Identification and management with transcatheter
8. Rueda F, Squitieri C, Ballerini L. Closure of the Fenestration in occlusion. Br Heart J 1995;74:548–552.
the extracardiac fontan with the amplatzer duct occluder device. 21. Sugiyama H, Yoo SJ, Williams W, Benson LN. Characterization
Cathet Cardiovasc Interv 2001;54:88–92. and treatment of systemic venous to pulmonary venous collater-
9. Goff DA, Blume ED, Gauvreau K, Mayer JE, Lock JE, Jenkins als seen after the Fontan operation. Cardiol Young 2003;13:
KJ. Clinical outcome of fenestrated fontan patients after closure 424–430.
the first 10 years. Circulation 2000;102:2094–2099. 22. Magee AG, McCrindle BW, Mawson J, Benson LN, Williams
10. Kuhn MA, Jarmakani JM, Laks H, et al. Effect of late postoper- WG, Freedom RM. Systemic venous collateral development af-
ative atrial septal defect closure on hemodynamic function in ter the bidirectional cavopulmonary anastomosis. Prevalence and
patients with a lateral tunnel Fontan procedure. J Am Coll Car- predictors. J Am Coll Cardiol 1998;32:502–508.
diol 1995;26:259–265. 23. Kaulitz R, Ziemer G, Bergmann F, Luhmer I, Kallfelz HC.
11. Mair DD, Hagler DJ, Puga FJ, Schaff HV, Danielson GK. Fon- Atrial thrombus after the Fontan-operation: predisposing factors,
tan operation in 176 patients with tricuspid atresia. Results and treatment and prophylaxis. Cardiol Young 1997;7:37–43.
a proposed new index for patient selection. Circulation 1990;82(4): 24. Monagle P, Cochrane A, McCrindle B, Benson L, Williams W,
IV164–IV169. Andrew M. Thromboembolic complications after Fontan proce-
12. Nakata S, Imai Y, Takanashi Y, et al. A new method for the dure—The role of prophylactic anticoagulation. J Thorac Cardi-
quantitative standardization of cross-sectional areas of the pul- ovasc Surg 1998;115:493–498.
monary arteries in congenital heart disease with decreased pul- 25. Rosenthal D, Friedman A, Kleinman S, Kopf G, Rosenfeld L,
monary blood flow. J Thorac Cardiovasc Surg 1984;88:610–619. Hellenbrand W. Thromboembolic complications after Fontan
13. Piehler JM, Danielson GK, McGoon DC, Wallace RB, Fulton operations. Circulation 1995;92(2):II287–II293.
RE, Mair DD. Management of pulmonary atresia with ventricu- 26. Jahangiri M, Ross DB, Redington AN, Lincoln C, Shinebourne
lar septal defect and hypoplastic pulmonary arteries by right EA. Thromboembolism after Fontan procedure and its modifica-
ventricular outflow construction. J Thorac Cardiovasc Surg tions. Ann Thorac Surg 1994;58:1409–1414.
1980;80:552–567. 27. Fyfe DA, Kline CH, Sade RM, Gillette PC. Transesophageal
14. Hijazi ZM, Fahey JT, Kleinman CS, Kopf GS, Hellenbrand WE. echocardiography detects thrombus formation not identified by
Hemodynamic evaluation before and after closure of fenestra- transthoracic echocardiography after the Fontan operation. J Am
tion. An acute study of changes in oxygen delivery. Circulation Coll Cardiol 1991;18:1733–1737.
1992;86:196–202. 28. Stumper O, Sutherland G, Geuskens R, Roelandt J, Bos E, Hess
15. Caspi J, Coles JG, Rabinovitch M, et al. Morphological findings J. Transesophageal echocardiography in evaluation and manage-
contributing to a failed Fontan procedure. Twelve-year experi- ment after a Fontan procedure. J Am Coll Cardiol 1991;17:
ence. Circulation 1990;82(4):IV177–IV182. 1152–1160.

Catheterization and Cardiovascular Interventions DOI 10.1002/ccd.

Published on behalf of The Society for Cardiovascular Angiography and Interventions (SCAI).