Pulmonary Atresia, Ventricular Septal Defect, and
Major Aortopulmonary Collaterals: Neonatal
Pulmonary Artery Rehabilitation Without
Unifocalization
Matthew Liava’a, MBChB, Christian P. Brizard, MD, Igor E. Konstantinov, MD, PhD,
Terry Robertson, MBBS, Michael M. Cheung, MD, Robert Weintraub, MBBS, and
Yves d’Udekem, MD, PhD
Departments of Cardiac Surgery and Cardiology, Royal Children’s Hospital, and Department of Pediatrics, University of
Melbourne and the Murdoch Children’s Research Institute, Melbourne, Victoria; and Department of Cardiology, Adelaide
Women’s and Children’s Hospital, Adelaide, South Australia, Australia
Background. This study analyzed a protocol of neona-
tal rehabilitation of hypoplastic pulmonary arteries in
the management of pulmonary atresia, ventricular septal
defect (VSD), and major aortopulmonary collateral arter-
ies (MAPCAs). Ideal management of patients with pul-
monary atresia, VSD, and MAPCAs is the subject of
controversy.
Methods. From June 2003 to December 2008, 25 consec-
utive patients were diagnosed with pulmonary atresia,
VSD, and MAPCAs, and 20 were entered into a neonatal
shunting regimen. The median age at the first operation
was 3.6 weeks (range, 0.7 to 17 weeks). All patients
underwent an initial central or modified Blalock-Taussig
shunt, or both. Further preparatory procedures included
26 pulmonary artery patch reconstructions, 19 right ven-
tricle-to-pulmonary artery conduits, 4 MAPCA ligations,
and 4 further shunts. No patient underwent translocation
of the collateral arteries.
P atients born with pulmonary atresia, ventricular sep-
tal defect (VSD), and major aortopulmonary collat-
erals (MAPCAs) have atresia of the pulmonary valve,
with hypoplastic or absent native pulmonary arteries
(PAs), and pulmonary blood supply depends on collat-
eral arteries arising from the aorta or its branches [1, 2].
Various surgical approaches have been applied to restore
the best possible pulmonary circulation to these patients.
The most popular technique involves anastomosing the
collateral vessels to the native pulmonary vessels in a
procedure termed “unifocalization” [3–5]. Although ini-
tially performed as staged procedures over years, unifo-
calization can now be performed at a single operation in
the neonatal period with low mortality [6].
Presently, surgical units around the world, including
ours, have vacillated between the staged approach and
Accepted for publication Aug 30, 2011.
Address correspondence to Dr d’Udekem, Department of Cardiac Surgery,
Royal Children’s Hospital, Flemington Rd, Parkville, Melbourne 3052,
Victoria, Australia; e-mail: yves.dudekem@rch.org.au.
© 2012 by The Society of Thoracic Surgeons
Published by Elsevier Inc
PEDIATRIC CARDIAC
Results. At the latest follow-up, 12 of 20 patients have
had a complete repair at a median age of 18 months
(range, 11 to 48 months), 6 are awaiting repair, and 2 are
considered unlikely to be repaired. No patient was miss-
ing follow-up. Median pulmonary artery indices had
grown from 14.51 to 118.7 in 17 patients. Postoperative
angiograms were performed at a median of 8 months
(range, 1.9 to 32.7 months) in 9 of 12 completely repaired
patients. The median right/left ventricular pressure ratio
was 0.64 (range, 0.54 to 0.91).
Conclusions. Rehabilitation of hypoplastic native pul-
monary arteries by a neonatal shunting regimen, without
MAPCA translocation, for pulmonary atresia, VSD, and
MAPCAs, provides encouraging results with excellent
early survival.
(Ann Thorac Surg 2012;93:185–92)
© 2012 by The Society of Thoracic Surgeons
the neonatal complete repair. Both techniques, however,
rely on collateral artery translocation to achieve an ade-
quate pulmonary vascular bed. We recently reviewed our
center’s experience with the staged approach and were
disappointed with the patency rate, lack of growth, and
the final contribution to the pulmonary circulation of the
translocated collateral vessels [7]. As a consequence, we
have largely abandoned the translocation of collateral
arteries and initiated a policy of rehabilitation of the
pulmonary circulation by the early shunting of the hyp-
oplastic native pulmonary vessels [8]. This article reports
our review of the safety and initial benefits of this change
in management in death, completion of biventricular
repair, hemodynamic results, and restoration of the pul-
monary vascular bed.
Material and Methods
The study was approved by the Royal Children’s Hospi-
tal Ethics Committee. Hospital records of all patients
0003-4975/$36.00
doi:10.1016/j.athoracsur.2011.08.082
 186 LIAVA’A ET AL
NATIVE PA REHAB FOR PA, VSD, MAPCAS
diagnosed with PA, VSD, and MAPCAs admitted from
June 2003 to December 2008 were reviewed. We identi-
fied 25 patients. Twenty neonates were entered into an
early shunting protocol aiming to rehabilitate native PAs
PEDIATRIC CARDIAC
without translocation of collateral arteries and constitute
the core of this study. Patient characteristics are reported
in Table 1.
Five patients were managed using alternate
techniques:
● One patient, referred at the age of 4 years, had a
normal left PA perfused by a ductus, previously
stented at birth, and an extremely hypoplastic
right PA with the right lung perfused by 3 collat-
erals. He underwent a right ventricle (RV)-to-PA
conduit, and is waiting for repair aiming at direct-
ing the RV blood flow to 1 lung.
● The second patient, born with normally sized PAs
supplied by 1 collateral vessel arising from the
descending aorta, underwent a complete repair at
the age of 14 months.
● The pulmonary circulation in the third patient
depended on 2 major collaterals feeding normally
sized confluent hilar PA branches. She underwent
a one-stage repair and unifocalization with a
Contegra (Medtronic Inc, Minneapolis, MN)
valved conduit used in the RV outflow tract.
● A fourth patient, with Alagille syndrome, was
diagnosed at the age of 3 months. He underwent
an RV-to-PA conduit but died 4 months later of
liver failure.
● The last patient underwent staged rehabilitation
of native PAs starting at the age of 10 years,
because of late referral. Two separate shunting
procedures were followed by complete repair
when he was 15.
PA Nakata indices were calculated for the 17 shunted
patients whose serial measurements of PA size were
available. Catheter data after complete repair was ob-
tained to determine right and left ventricular pressures.
The following formulas were used:
Body surface area 共BSA兲
⫽ 兹共height in cm ⫻ weight in kg兲 ⁄ 3, 600
Pulmonary artery Nakata Index ⫽ right PA ⫹ left PA
cross-sectional areas in mm2/BSA in m2.
Surgical Strategy
All neonates were planned for elective central shunting
within the first 4 weeks of life, regardless of their oxygen
Table 1. Patient Demographics
No. (%)
Variable (N ⫽ 20)
Male 14 (70)
22q11 deletion 9 (45)
Right aortic arch 9 (45)
Antenatal diagnosis 13 (65)
Ann Thorac Surg
2012;93:185–92
saturations. The technique of central shunting initially
described by Gates and colleagues [9] was our favored
approach. The distal opening was made at the level of the
bifurcation of the PAs and extended down to the main
PA. A Gore-Tex shunt (W. L. Gore and Associates,
Flagstaff, AZ) was anastomosed to this opening end-to-
side with 8-0 Prolene sutures (Ethicon, Somerville, NJ).
The proximal anastomosis consisted of a side-to-side
anastomosis on the ascending aorta.
The patients subsequently underwent further shunting
procedures or RV-to-PA artery conduit implantations
until the team had the subjective impression that the PAs
had attained sizes that would allow complete repair.
Statistical Analysis
Data are reported as median (range) or mean (standard
deviation), as appropriate.
Results
Twenty patients with a diagnosis of pulmonary atresia,
VSD, and MAPCAs underwent an initial neonatal shunt-
ing procedure. At the time of last follow-up, the median
age of patients was 34.8 months (range, 5 to 79 months),
with no deaths.
Clinical Outcomes
PROCEDURE 1. The median age at the first procedure was 3.6
weeks (range, 0.7 to 17 weeks), and mean preoperative
oxygen saturations were 85% (range, 72% to 95%). One
patient’s first operation was delayed until age 17 weeks
because of a bleeding peptic ulcer. All other patients
underwent an initial shunting procedure before age 8
weeks, regardless of oxygen saturation level. Initial
shunting consisted of 15 central shunts only (3 mm, n ⫽
2; 3.5 mm, n ⫽ 11; 4 mm, n ⫽ 2), 4 received a right
modified Blalock-Taussig shunt (3 mm, n ⫽ 2; 3.5 mm,
n ⫽ 1; 4 mm, n ⫽ 1), and 1 patient had a central and left
modified Blalock-Taussig shunt (both 3 mm). Concomi-
tant PA reconstruction with 0.4-mm-thick Gore-Tex
patches (W. L. Gore and Associates) was performed in 2
of 20 patients (10%).
PROCEDURE 2. Nineteen patients have undergone a second
procedure at a median age of 7.9 months (range, 2.6 to
20.7 months). Complete repair was achieved in 2 patients
(11%), and a RV-to-PA conduit was inserted in 12 (63%)
using 6- or 8-mm Gore-Tex grafts, or 12-mm Contegra
valved conduits.
Three of 19 patients (16%) had a second systemic-to-
pulmonary shunt (3.5 mm, n ⫽ 1; 4 mm, n ⫽ 2), without
creation of an RV-to-PA connection. These were all per-
formed for origin stenosis of the left PA. One patient
underwent tricuspid valve repair for severe tricuspid regur-
gitation, and 1 patient underwent pulmonary valve com-
missurotomy. Concomitant patch reconstruction of the PAs
was performed in 9 of 19 (47%) and 3 (16%) underwent
MAPCA ligation for excessive pulmonary blood flow.
PROCEDURE 3. Fourteen patients underwent a third proce-
dure at median age of 19.5 months (range, 4.8 to 40.4
Ann Thorac Surg
2012;93:185–92
Table 2. Postoperative Length of Stay and Ventilatory Time
Length of Stay
(days)a
a
Ventilator Time
Operation (hours) ICU Hospital
Initial neonatal shunt 61 (8–166) 4 (1–9) 16 (7–53)
Intermediate procedures 28 (8–692) 2 (1–30) 10 (4–86)
Complete repair 19 (8–78) 1 (1–4) 8 (4–11)
a
Data are presented as median (range).
ICU ⫽ intensive care unit.
months). Complete repair was performed in 7 patients,
RV-to-PA conduit change in 5, and ligation of MAPCA
for excessive pulmonary blood flow in 1. One patient had
tricuspid valve repair with RV-to-PA conduit and a
systemic-to-pulmonary shunt to the left PA. Nine pa-
tients had concomitant PA patch reconstruction.
PROCEDURE 4. Seven patients had a fourth operation at a
median of 25 months (range, 19.4 to 48.4 months). Com-
plete repair was performed in 3 patients, and all had
concomitant patch reconstruction of the PAs. Three pa-
tients underwent further PA reconstructions associated
with RV-to-PA conduit change in 1 and MAPCA ligation
in 1. An exploratory thoracotomy was done in 1 patient to
assess the possibility of MAPCA translocation; however,
no suitable vessels were found for translocation.
Hospital Morbidity
No major airway complications occurred after any of
these procedures. The median ventilatory times, inten-
sive care unit stay, and hospital stay after the initial
shunting, intermediate procedures, and final repair are
given in Table 2.
Follow-Up
After a mean of 39 ⫾ 19 months, 12 patients reached a
complete repair and 6 are awaiting repair. The VSDs
were closed in all patients, although 1 patient had a
fenestrated flap patch that has never showed any shunt-
ing, and 1 patient was left with a small muscular VSD that
is expected to close spontaneously. In 2 patients, an atrial
septal defect was left open because the pulmonary ves-
sels seemed small. Two patients are currently not con-
sidered repairable (Fig 1). All patients who attained
complete repair had native PA arborization to at least 14
lung segments, as identified on preoperative angiograms.
LIAVA’A ET AL 187
NATIVE PA REHAB FOR PA, VSD, MAPCAS
In 3 patients, treatment with sildenafil was instituted
after their control angiograms.
On the last postoperative echocardiography performed
a median of 12 months (range, 4 to 48 months) after
PEDIATRIC CARDIAC
repair, the systolic RV function appeared to be normal in
all patients. The RV appeared to be of normal size in 3 of
12 patients, was mildly dilated in 7, and was moderately
dilated in 2.
One patient left with palliation has 22q11 deletion and
underwent a 3-mm central shunt at age 1.7 weeks,
RV-to-PA conduit and left PA reconstruction at 5 months,
with left PA balloon angioplasty at 9 months. The Nakata
index increased from 4.92 at birth to 67.5 at 9 months.
Although his native PAs have increased in size and are
distributed to both lungs, their arborization was poor,
supplying only 10 lung segments. They also carried
multiple stenoses, and the vascularization of both lungs
seemed to mainly depend on multiple small MAPCAs.
The second patient left with palliation underwent ini-
tial central shunting at age 2.7 weeks, RV-to-PA conduit
at 4.5 months, and a valved RV-to-PA conduit at 12
months. The native PA distribution was to only 9 lung
segments, with multiple bilateral MAPCAs providing the
bulk of the pulmonary circulation. In both of these
patients, the MAPCAs were left untouched because they
were providing most of the pulmonary blood flow.
PA Growth
Measurements of right and left branch PAs were re-
corded from angiograms, direct measurements at opera-
tion, computed tomography, magnetic resonance imag-
ing, and echocardiography. These were plotted against
BSA, as described by Nakata, to provide a measure of
pulmonary vascular arborization [10]. Data were avail-
able for 17 of 20 patients.
The median Nakata index before the first procedure
was 14.51 (range, 1.77 to 55.15). This increased to 118.7
(range, 16.4 to 377.7) at a median of 17 months (range, 5 to
53 months). The last recordings before complete repair
gave a median Nakata index of 101.4 (range, 29.44 to
226.9; Fig 2).
Angiography After Complete Repair
Complete repair was achieved in 12 patients. Nine were
studied with angiography after a median of 13 months
(range, 4 to 33 months; Table 3; Fig 3 and Fig 4). Their
median RV systolic pressure was 53 mm Hg (range, 43 to
65 mm Hg), median left ventricular pressure was 82 mm
Fig 1. Flow chart shows outcome of all pa-
tients with pulmonary atresia (PA), ventricu-
lar septal defect (VSD), and major aortopul-
monary collaterals (MAPCAs).
 188 LIAVA’A ET AL Ann Thorac Surg
NATIVE PA REHAB FOR PA, VSD, MAPCAS 2012;93:185–92

Fig 2. Growth of pulmonary arteries as as-

sessed by the Nakata index is shown, with
each line representing one patient.
PEDIATRIC CARDIAC

Hg (range, 68 to 100 mm Hg), and the median right/left
ventricular pressure ratio was 0.64 (range, 0.54 to 0.71).
The median mean pulmonary artery pressure was 25 mm
Hg (range, 22 to 33 mm Hg).
PA artery stenosis resulted in 6 patients undergoing 15
balloon angioplasties. One patient also underwent stent-
ing of the left PA. Results in these 6 patients were
variable, ranging from no improvement (stenosis re-
mained) in 2 patients, small to moderate improvement in
3, and complete resolution of gradient across branch
pulmonary stenosis in 1.
Complications
There were 17 major surgical complications in 57 opera-
tions (Table 4). Rapid oxygen desaturation in 2 patients in
the immediate postoperative period required chest re-
opening to relieve shunt compression in the intensive
care unit. One patient required 4 days of extracorporeal
support for poor ventricular function.
Comment
Most of the surgical literature published during the last
15 years concerning the treatment of patients born with
pulmonary atresia, VSD, and MAPCAs has promoted the
unifocalization of collateral arteries to create a new
pulmonary vascular bed [2– 4]. This strategy has resulted
in improved survival for these patients [4, 6].
During unifocalization procedures, collateral vessels
are joined to native pulmonary vessels or to newly
created central vessels. It was demonstrated many years
ago that hypoplastic central PAs could be developed or
“rehabilitated” by shunting procedures [11, 12]. A num-
ber of patients in the published series of patients man-
aged with unifocalization strategies will also have had
previous shunting operations on their native pulmonary
vessels. Because of the complexity and the heterogeneity
of the procedures included under the term unifocaliza-
tion, it is very difficult to evaluate the relative benefits of
integrating collaterals vessels into the pulmonary circu-
lation and comparing it with the benefits acquired by the
recruitment of native vessels.
In 2005 we reviewed our experience of multistage
unifocalization procedures in 82 of our initial patients
and reported our disappointing results with the translo-
cation of collateral arteries [7]. Half of the identified
translocated collateral arteries thrombosed, significant
stenoses developed in a third of those that remained
patent, and the collaterals that remained patent did not
seem to show any growth. For these reasons, we decided
to reserve translocation of collateral arteries for excep-

Table 3. Catheter Hemodynamics After Complete Repair

Time from Repair Pressures (mm Hg)
to Angiography Residual Residual RV/LV Oxygen
Pt (months) ASD VSD Mean MPA RV Systemic Ratio Saturations

1 13 0 0 28 59 92 0.64 96
2 25 0 0 37 64 82 0.79 94
3 13 0 0 29 53 80 0.66 97
4 8 1 0 25 47 72 0.65 99
5 8 0 0 22 43 68 0.63 100
6 25 0 1 22 43 80 0.54 97
7 3 0 0 25 47 85 0.55 100
8 4 1 1 24 58 85 0.68 99
9 33 0 0 26 58 100 0.58 98

ASD ⫽ atrial septal defect; LV ⫽ left ventricle; MPA ⫽ main pulmonary artery; RV ⫽ right ventricle; VSD ⫽ ventricular septal defect.
Ann Thorac Surg
2012;93:185–92
Fig 3. Patient 2: (A) Initial angiography at birth, (B) after shunting, then right ventricular-to-pulmonary artery conduit, and (C) after complete
repair with pulmonary artery patching.
tional cases where large collaterals connect to pulmonary
vessels of normal caliber.
We have therefore initiated a strategy focusing on
rehabilitation of the native pulmonary vessels. We hy-
pothesized that PAs would have the best chance to grow
if the initial shunting procedure was done early in life;
therefore, we planned the initial operation for the first
few weeks of life, regardless of patient oxygen satura-
tions. We thereafter proceeded with repeated shunting
procedures or implantation of RV-to-PA conduits until
we were satisfied we had obtained PAs of sizes that were
large enough to allow definitive repair.
Rehabilitation of the native PAs in patients with pul-
monary atresia, VSD, and MAPCAs is an old concept but
does not seem to have been reported as a single strategy
in recent times, especially in asymptomatic neonates. We
initially considered this an investigational approach;
however, by the end of this initial period, we feel justified
in continuing with the protocol. The numbers of patients
enrolled with this rare condition are still small, and a
larger cohort of patients with longer follow-up will be
necessary to ascertain any superiority of this approach.
The fact that we encountered no deaths related to the
condition or to the operation is certainly encouraging.
LIAVA’A ET AL 189
NATIVE PA REHAB FOR PA, VSD, MAPCAS
PEDIATRIC CARDIAC
We hypothesized that the survival of some patients
was dependent on their aortopulmonary collaterals.
Therefore, translocation of collaterals may put the pa-
tient at risk of postoperative cyanosis and death because
pulmonary blood flow may become compromised. We
believe that if the network of aortopulmonary collaterals is
left intact, the initial procedures to rehabilitate the native
PAs will be largely innocuous, because even if the central
shunt occludes, the patients will be left with the same
amount of pulmonary blood flow they had before the
procedure. Exceptionally, a patient’s pulmonary circula-
tion may have very large collaterals, and increasing
pulmonary blood flow can put them into heart failure. In
these patients, ligation or embolization of the collateral
has to be envisioned preoperatively or postoperatively.
The main interest in this strategy of native PA rehabil-
itation is that it demonstrates the true potential of the
native pulmonary circulation to develop and to connect
to enough segmental pulmonary branches to allow sub-
sequent repair.
In most of our patients, the pulmonary vasculature
had developed enough to allow a successful repair.
That the central PAs in most of our patients, once
developed, were connected to most of the lung seg-
Fig 4. Angiography after repair of (A) patient
6 and (B) patient 7.
 190 LIAVA’A ET AL
NATIVE PA REHAB FOR PA, VSD, MAPCAS
Table 4. Complications
Complication
Shunt compression requiring urgent reoperation
PEDIATRIC CARDIAC
Sepsis
Thoracocentesis
Pericardial effusion requiring draining
Wound dehiscence
Vocal cord palsy
Chylothorax
Bowel perforation
Bleeding requiring reoperation
Extracorporeal membrane oxygenation
Prolonged ventilation requiring tracheostomy
Total
ments demonstrates that unless these arteries are
rehabilitated, it is impossible to determine their poten-
tial to contribute to the final pulmonary vasculature.
Some have described that up to 16% of patients with
the condition are born with no central PAs, which
would preclude a strategy of early shunting of the
central PAs [2]. This finding, however, has not been our
experience, with all patients having central PAs. Two
of our 20 patients did have PAs that were so small that
they were considered disconnected and required two
separate shunts. Only one patient had a left branch PA
that did not achieve growth with shunting. It is possi-
ble that we were able to identify central PAs in a larger
proportion of patients because they were operated on
earlier in life, and that subsequently, these arteries
continue to atrophy and disappear.
Our present results were obtained in a small number
of patients compared with the larger presented series,
and one cannot exclude that we were favored by a cluster
of patients with a more favorable anatomy and that some
patients may still be born with no identifiable central
PAs. The very large published series are arising from
centers that have specific referral for this condition, and
the spectrum of the disease they encounter may be
different from our unselected population.
Because in the last years we have adopted a unified
initial strategy of central shunting in neonates, we have
abandoned initial investigation by diagnostic catheteriza-
tion. Our patients are now initially evaluated by com-
puted tomography or magnetic resonance imaging, and
diagnostic catheterization is reserved for the preopera-
tive evaluation of the subsequent staged operations. In
our experience, none of these examinations will accu-
rately identify the real distribution of the native vessels to
the pulmonary circulation before these native vessels are
rehabilitated.
There is a general agreement that the anatomic sub-
strate of this condition varies, even among patients
whose lung circulation depends exclusively on aortopul-
monary collaterals. Some may only have minimal hyp-
oplasia of otherwise normal PAs, and at the other ex-
treme, some patients’ lung circulation, even after
Ann Thorac Surg
2012;93:185–92
rehabilitation of native vessels, will exclusively depend
on vessels with more similarities to bronchial arteries
No.
than pulmonary vessels [13]. It could be argued that the
2 patients at the worst end of the spectrum will not benefit
3 from a strategy of rehabilitation of the native pulmonary
1 vessels and that they are likely to be left with palliation.
1 Only 2 of our 20 patients so far have been left to
2 palliation, so that it seems that at this time, this approach
3 would potentially allow an achievable rate of repair up to
1 90%. This matches even the best-reported outcomes of
1 unifocalization strategies [4 – 6].
1 It has been suggested that after unifocalization,
1 death is independent from the severity of the disease
1 or the morphology of the pulmonary vasculature. It
17 seems unlikely to us that the 2 patients whom we
consider to have the worst form of the disease and that
we left palliated would not be at risk of dying after a
unifocalization procedure. Because their symptoms are
still limited, we have not yet offered them further
intervention. It has already been suggested that appro-
priate palliation may allow better survival in this
subset of patients [2].
Determining what is an acceptable RV pressure after
complete repair remains a contentious issue. Typically,
the systolic RV pressure is compared with the systolic left
ventricular or systemic pressure, with ratios of less than
0.6 considered acceptable [6]; however, some groups will
accept ratios as high as 0.8 to 1 [14]. We were slightly
disappointed by the mean residual RV pressures in these
patients after repair. It seems comparable to most series
of the staged approach [4, 15] but inferior to the neonatal
one-stage approach [6]. Unfortunately, whether such
pressures will affect the long-term survival of these
patients is unknown.
Most groups would be concerned about eventual RV
dysfunction if the RV pressure is left greater than two-
thirds systemic for the long-term. We were encouraged,
however, to notice that our patients’ median mean PA
pressures were low. We can therefore expect that their
RV pressures will decrease after the subsequent proce-
dures of replacement of the RV-to-PA conduits. This will
only be shown with ongoing follow-up, but we believe
that native PAs should have better growth potential than
unifocalized MAPCAs [7, 13, 16].
Our limited experience with angioplasty of pulmonary
stenoses has been mixed, with initial improvements not
long-lasting. Others have shown PA stenting using tra-
ditional or hybrid approaches can have better success
rates [17, 18], but our unresolved concerns regarding
possible limitations of PA growth have precluded exten-
sive use of this modality.
In conclusion, a strategy focusing on rehabilitation of
PAs of patients born with pulmonary atresia, VSD, and
MAPCAs by early shunting procedure in neonates re-
stores a pulmonary vasculature, allowing repair in most
patients. Low early mortality and a high rate of repair
have encouraged us to pursue further application and
investigation of this strategy.
Ann Thorac Surg
2012;93:185–92
Yves d’Udekem is a Career Development Fellow of The National
Heart Foundation of Australia Research Program (CR 10M 5339).
This research project was supported by the Victorian Govern-
ment’s Operational Infrastructure Support Program.
References
1. Amark K, Karamlou T, O’Carroll A, et al. Independent
factors associated with mortality, reintervention, and
achievement of complete repair in children with pulmonary
atresia with ventricular septal defect. J Am Coll Cardiol
2006;47:1448 –56.
2. Bull K, Somerville J, Ty E, Speigelhalter D. Presentation and
attrition in complex pulmonary atresia. J Am Coll Cardiol
1995; 25:491–9.
3. Reddy V, Liddicoat J, Hanley F. Midline one-stage complete
unifocalization and repair of pulmonary atresia with ventric-
ular septal defect and major aortopulmonary collaterals.
J Thorac Cardiovasc Surg 1995;109:832– 45.
4. Duncan B, Mee R, Prieto L, et al. Staged repair of tetralogy of
Fallot with pulmonary atresia and major aortopulmonary col-
lateral arteries. J Thorac Cardiovasc Surg 2003;126:694 –702.
5. Brawn W, Jones T, Davies B, Barron D. How we manage
patients with major aorta pulmonary collaterals. Semin
Thorac Cardiovasc Surg Pediatr Card Surg Ann 2009;12:
152–7.
6. Malhotra S, Hanley F. Surgical management of pulmonary
atresia with ventricular septal defect and major aortopulmo-
nary collaterals: a protocol-based approach. Semin Thorac
Cardiovasc Surg Pediatr Card Surg Ann 2009;12:1445–151.
7. d’Udekem Y, Alphonso N, Norgaard M, et al. Pulmonary
atresia with ventricular septal defects and major aortopul-
monary collateral arteries: unifocalization brings no long-
term benefits. J Thorac Cardiovasc Surg 2005;130:1496 –502.
8. Brizard C, Liava’a M, d’Udekem. Pulmonary atresia, VSD
and MAPCAs: repair without unifocalization. Semin Thorac
Cardiovasc Surg Pediatr Card Surg Ann 2009;12:139 – 44.
INVITED COMMENTARY
Patients with ventricular septal defect, pulmonary
atresia, and major aortopulmonary collateral arteries
(MAPCAs) are not uncommon. There is still controversy
regarding surgical treatment for these patients. The au-
thors’ experience in this article [1] was disappointing
regarding patency rate, lack of growth, and the contribu-
tion to the pulmonary circulation of translocated collat-
eral vessels. They planned to rehabilitate hypoplastic
native pulmonary arteries by neonatal shunting without
MAPCA translocation in 25 consecutive patients. These
patients underwent an initial central or modified Blalock-
Taussig shunt, or both, and excellent results were
achieved. The authors are to be congratulated.
Patients with ventricular septal defect, pulmonary atresia,
and MAPCAs have a widely variable pathology and hemody-
namics, especially pulmonary vascularity, which include the
native pulmonary artery and the MAPCAs. Even in each lobe
or segment, lung pulmonary vascularity may develop differ-
ently, and the number, position, size, and morphology of the
MAPCAs can vary. The origin of the aorta also may emerge
differently from the left or right ventricle and with different
degrees of overriding. The pathology of pulmonary vascularity
also changes with the growth of the patient.
© 2012 by The Society of Thoracic Surgeons
Published by Elsevier Inc
LIAVA’A ET AL 191
NATIVE PA REHAB FOR PA, VSD, MAPCAS
9. Gates RN, Laks H, Johnson K. Side-to-side aorto-Gore-Tex
central shunt. Ann Thorac Surg 1998;65:515– 6.
10. Nakata S, Imai Y, Takanashi Y, et al. A new method for the
quantitative standardization of cross-sectional areas of the
pulmonary arteries in congenital heart diseases with de-
PEDIATRIC CARDIAC
creased pulmonary blood flow. J Thorac Cardiovasc Surg
1984;88:610 –9.
11. Haworth SG, Rees PG, Taylor JFN, Macartney FJ, de Leval
M, Stark J. Pulmonary atresia with ventricular septal defect
and major aortopulmonary collaterals. Effect of systemic
pulmonary anastomosis. Br Heart J 1981;45:133– 41.
12. Rome J, Mayer J, Castaneda A, Lock J. Tetralogy of Fallot
with pulmonary atresia, rehabilitation of diminutive pulmo-
nary arteries. Circulation 1993;88:1691– 8.
13. Norgaard M, Alphonso N, Cochrane A, Menahem S, Brizard
C, d’Udekem Y. Major aorto-pulmonary collateral arteries of
patients with pulmonary atresia and ventricular septal de-
fect are dilated bronchial arteries. Eur J Cardiothorac Surg
2006;29:653– 8.
14. Metras D, Chetaille P, Kreitmann B, Fraisse A, Ghez O,
Riberi A. Pulmonary atresia with ventricular septal defect,
extremely hypoplastic arteries, major aorto-pulmonary col-
laterals. Eur J Thorac Cardiovasc Surg 2001;20:590 –7.
15. Ishibashi N, Shin’oka T, Ishiyama M, Sakamoto T, Kurosawa
H. Clinical results of staged repair with complete unifocal-
ization for pulmonary atresia with ventricular septal defect
and major aortopulmonary collateral arteries. Eur J Cardio-
thorac Surg 2007;32:202– 8.
16. Song S, Park H, Park Y, Cho B. Pulmonary atresia with
ventricular septal defects and major aortopulmonary collat-
eral arteries, 18-year clinical experience and angiographic
follow-up of major aortopulmonary collateral arteries. Circ J
2009;73:516 –22.
17. Nakanishi T. Balloon dilation and stent implantation for
vascular stenosis. Pediatr Int 2001;43:548 –52.
18. Holzer R, Chrisolm J, Hill S, et al. “Hybrid” stent delivery in
the pulmonary circulation. J Invasive Cardiol 2008;20:592– 8.
The following factors have to be considered for better
results of surgical repair:
● an adequate pulmonary arterial bed that results in
reduced pulmonary vascular resistance after sur-
gical repair;
● interruption of MAPCA or patent ductus arterio-
sus blood flow from the aorta;
● use of an ideal conduit from the right ventricle to
the pulmonary artery.
Because each patient has his or her own characteristics
of pulmonary vascularity, there is no universal method
suitable for all patients; thus, each patient must be
managed individually.
A hypoplastic native pulmonary artery can be rehabil-
itated by a central shunt or Blalock-Taussig shunt. This
procedure is required for many patients. Similarly, the
MAPCAs or patent ductus arteriosus helps to develop
the pulmonary artery and may even cause pulmonary
arterial hypertension in a lobe or lung segment if the
MAPCAs do not have proximal stenoses. Proximal ste-
nosis of the MAPCAs usually protects against excessive
distal pulmonary vascularity. Therefore, the shunt pro-
0003-4975/$36.00
doi:10.1016/j.athoracsur.2011.10.009