Modiﬁed Blalock Taussig shunt: a not-so-simple palliative procedure
Verena Dirksa, René Prêtreb, Walter Knirschc,d, Emanuela R. Valsangiacomo Buechelc,d, Burkhardt Seiferte,
Martin Schweigera,d, Michael Hüblera,d and Hitendu Davea,d,*
Division of Congenital Cardiovascular Surgery, University Children’s Hospital Zurich, Zurich, Switzerland
Department of Cardiovascular Surgery, University Hospital Lausanne, Lausanne, Switzerland
Department of Paediatric Cardiology, University Children’s Hospital Zurich, Zurich, Switzerland
Children’s Research Centre, University of Zurich, Zurich, Switzerland
Division of Biostatistics, Institute for Social and Preventive Medicine, University of Zurich, Zurich, Switzerland
* Corresponding author. Division of Congenital Cardiovascular Surgery, University Children’s Hospital Zurich, Zurich, Switzerland. Fax: +41-44-2668021;
e-mail: email@example.com; firstname.lastname@example.org (H. Dave).
Received 7 January 2013; received in revised form 26 February 2013; accepted 27 February 2013
OBJECTIVES: Thirty-two consecutive isolated modiﬁed Blalock Taussig (BT) shunts performed in infancy since 2004 were reviewed and
analysed to identify the risk factors for shunt intervention and mortality.
METHODS: Sternotomy was the only approach used. Median age and weight were 10.5 (range 1–74) days and 2.9 (1.9–4.4) kg, respect-
ively. Shunt palliation was performed for biventricular hearts (Tetralogy of Fallot/double outlet right ventricle/transposition of great
arteries_ventricular septal defect_pulmonary stenosis/pulmonary atresia_ventricular septal defect/others) in 21, and univentricular hearts
in 11, patients. Hypoplastic left heart syndrome patients were excluded. Two procedures required cardiopulmonary bypass. Median
shunt size was 3.5 (3–4) mm and median shunt size/kg body weight was 1.2 (0.9–1.7) mm/kg. Reduction in shunt size was necessary in
5 of 32 (16%) patients.
RESULTS: Three of 32 (9%) patients died after 3 (1–15) days due to cardiorespiratory decompensation. Lower body weight (P= 0.04)
and bigger shunt size/kg of body weight (P= 0.004) were signiﬁcant risk factors for mortality. Acute shunt thrombosis was observed in
3 of 32 (9%), none leading to death. Need for cardiac decongestive therapy was associated with univentricular hearts (P< 0.001), bigger
shunt size (P= 0.054) and longer hospital stay (P= 0.005). Twenty-eight patients have undergone a successful shunt takedown at a
median age of 5.5 (0.5–11.9) months, without late mortality.
CONCLUSIONS: Palliation with a modiﬁed BT shunt continues to be indicated despite increased thrust on primary corrective surgery.
Though seemingly simple, it is associated with signiﬁcant morbidity and mortality. Effective over-shunting and acute shunt thrombosis
are the lingering problems of shunt therapy.
Keywords: Modiﬁed Blalock Taussig shunt •Palliation •Mortality •Cyanotic heart disease
While the classic Blalock Taussig shunt was a breakthrough in
treating cyanotic heart diseases , it involves sacriﬁcing ante-
grade ﬂow to the subclavian artery with its attendant risks [2,3].
In 1975, de Leval modiﬁed the technique, using a polytetra-
ﬂuoroethylene interposition graft popularly known as the modi-
ﬁed Blalock Taussig (BT) shunt (MBTS) . While MBTS has
become an established palliative procedure with progressive
improvements in the outcome , growing experience has led to
increasing thrust on primary corrective procedures. Palliative
strategy has obvious disadvantages, such as the need for two
operations, potentially two scars, possibility of distortion of the
branch pulmonary artery, volume loading of the ventricles, lower
diastolic pressures, etc. In spite of these, primary shunt palliation
continues to be indicated in neonates with physiological pul-
monary hypertension, which makes a bidirectional Glenn shunt
untenable. Many centres also consider a primary neonatal cor-
rection of Tetralogy of Fallot to be riddled with risks and hence,
still prefer to palliate neonates in blue spells.
While acknowledging its role even in the modern era, mortal-
ity of the MBTS procedure is relatively high, tending to be
around 10% . This report is based on 32 consecutive ‘ﬁrst
time’MBTS palliations performed at our institution since 2004,
with a view to analysing the risk factors for mortality, shunt
thrombosis and need for decongestive therapy.
PATIENTS AND METHODS
Thirty-two MBTS procedures performed in neonates and infants
at our institution since 2004 were analysed. MBTS performed in
© The Author 2013. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.
European Journal of Cardio-Thoracic Surgery 44 (2013) 1096–1102 ORIGINAL ARTICLE
doi:10.1093/ejcts/ezt172 Advance Access publication 28 March 2013
patients with HLHS or as a part of other complex procedures,
such as unifocalization, were excluded. Twenty-six patients were
neonates and 6 were infants. Twenty-ﬁve patients were male.
The demographic data are detailed in Table 1. Diagnosis was
established using trans-thoracic echocardiography.
Primary end points were mortality and shunt thrombosis.
Secondary end points such as need for excessive inotropic
support and cardiac decongestive therapy were also analysed. A
brief analysis comparing Era 1 (2004–07) and 2 (2008–11) was
performed to see if the results had changed over time and to
identify the factors that may have accounted for that change.
All MBTS procedures were performed through a sternotomy.
Two procedures required cardiopulmonary bypass. After full
sternotomy, the right lobe of the thymus was excised; the course
of the brachiocephalic trunk up to its bifurcation was dissected.
The right pulmonary artery up to the hilum was dissected.
A bolus of 100 IU/kg crystalline heparin was administered. The
brachiocephalic trunk was clamped with a Cooley clamp, and
the distal right subclavian artery, temporarily with a ligaclip.
A longitudinal arteriotomy was performed at the ‘premarked’
undersurface of the truncus to subclavian artery continuity. An
obliquely fashioned end of the thin-wall Gore-Tex stretch vascu-
lar graft (W. L. Gore & Associates, Inc., AZ, USA) was sutured
end-to-side to the arteriotomy. Clamps were released and good
shunt ﬂow through the anastomosis ascertained. The Cooley
clamp was placed again to exclude the proximal anastomosis
from the circulation. The shunt length was trimmed so as to
avoid it being too long. The shunt lumen was ﬂushed to remove
microthrombi. The right pulmonary artery was excluded from
circulation using two vascular clamps. The transversely fashioned
distal end of the graft was anastomosed to the right pulmonary
artery. The target arterial saturation was 75–85% (Q
Patent duct: to ligate or not?
Our preference was to ligate the persistent ductus arteriosus
(PDA) in all patients having forward ﬂow through their main pul-
monary artery. In shunt-dependent circulations, the patent duct
was often circumvented and almost obliterated using a silastic
sling and ligaclips. The aim of this manoeuvre was to allow a
quick rescue by re-establishing ductal ﬂow in case of a shunt
Selection of shunt size
As a rule of thumb, a 3-mm graft was used for children around
3 kg or lower in body weight, whereas a 3.5-mm graft was used
for children around 3.5 kg. The indication, whether palliating for
a univentricular or a biventricular heart, inﬂuenced the size se-
lection in borderline weight-class children. Fine regulation of
ﬂow and pressure was inﬂuenced by displacing the proximal
inﬂow anastomosis either to the proximal subclavian artery or to
the brachiocephalic trunk, as well as by slightly titrating the
length of the shunt. The details of shunt size and positioning are
summarized in Table 2.
Postoperative left open sternum
The primary goal was to close the chest at the shunt procedure.
However, if there were any fears about the fate of the shunt
(especially in totally shunt-dependent pulmonary circulation) or
about the shunt getting squeezed behind the aorta, etc. the
sternum was left open. It was believed that an open sternum
lends itself to a quick response in case of an emergency when
compared with a closed chest.
Anticoagulation regimen was decided on a case-by-case basis.
Therapeutic heparinization was performed in high-risk shunt
scenarios, such as shunt-dependent pulmonary perfusion, in
Table 1: Preoperative clinical characteristics
Variable Median (range)
Age (days) 10.5 (1–74)
Weight at operation (kg) 2.9 (1.9–4.4)
Size (cm) 48 (39–52)
Body surface area (m²) 0.2 (0.14–0.24)
(%) 85 (50–95)
atresia_VSD, Ebstein’s anomaly
Single ventricle inclusive of pulmonary
TOF: Tetralogy of Fallot; DORV: double outlet right ventricle; TGA:
transposition of great arteries; VSD: ventricular septal defect; PS:
Table 2: Shunt procedure details
Variable N(%) except otherwise
No. of shunts (mm)
3 8 (25)
3.5 19 (59)
4 5 (16)
Subclavian artery 12 (38)
Truncus brachiocephalicus 20 (63)
Right pulmonary artery 28 (88)
Left pulmonary artery 4 (13)
Heart lung machine 2 (6)
Absolute shunt size (mm) 3.5 (3–4)
Median shunt size/body weight ratio
V. Dirks et al. / European Journal of Cardio-Thoracic Surgery 1097
cases with shunt clipping (shunt size reduction) or technical pro-
blems encountered during shunt construction. Heparin infusion
starting with 5–10 IU/kg/h, followed by therapeutic dose as early
as 2 h postoperatively, was planned if surgical bleeding was not
an issue. In effect, however, the therapeutic anticoagulation was
often achieved later than 2 h. Shunts without complications and
considered normal risk received aspirin in the long term.
Three patients did not have long-term anticoagulation/platelet
inhibitor medication: 2 who died early and 1 who survived with
early shunt thrombosis in a neonatal Ebstein’s anomaly with
antegrade pulmonary ﬂow.
All patients were evaluated postoperatively with trans-thoracic
Statistical analysis was performed using IBM SPSS Statistics 19
(SPSS, Inc., Chicago, IL, USA). Categorical variables are presented
as numbers with percents and are compared using Fisher’s exact
test. Continuous variables are presented as median with range
and are compared between groups using the Mann–Whitney
test. P-values <0.05 were considered signiﬁcant.
Three of the 32 (9%) patients died at a median of 3(1–15) days.
The details of patients who died are as presented in Table 3.
Lower body weight (P= 0.04) and bigger shunt size/kg of body
weight (P= 0.004) were signiﬁcant risk factors for mortality
Acute shunt thrombosis was observed in 3 of the 32 (9%)
patients, none of whom died. These 3 patients had Ebstein’s
anomaly with antegrade pulmonary ﬂow (2.2 kg 3 mm right
MBTS), double outlet right ventricle-hypoplastic left ventricle
(LV)_pulmonary stenosis (2.9 kg 3 mm right MBTS) and unba-
lanced atrio-ventricular septal defect hypoplastic LV
d-transposition of great arteries_pulmonary stenosis (3 kg 3.5
mm left MBTS). All shunt thrombosis occurred within the ﬁrst
24 h and were diagnosed by echocardiography. Two patients
had shunt revision with evacuation of the thrombus. One patient
with stable saturation due to normal antegrade pulmonary ﬂow
was not subjected to shunt revision. Shunt thrombosis could not
be statistically related to shunt size (P= 0.1) or shunt size/kg
body weight (P= 0.92). Early anticoagulation regimen (P= 0.33),
competitive blood ﬂow (P= 1.0), diagnosis (P= 0.27), need for
cardiopulmonary bypass (P= 0.18), increased inotropic support
(P= 0.71) and shunt size reduction (P= 1.0) were not associated
with acute shunt thrombosis.
Need for adrenalin ≤0.05 and/or noradrenaline ≤0.05 and/or
milrinone ≤0.75 μg/kg/min was deﬁned as normal inotropic
support. Accordingly, 6 of the 32 (19%) patients needed normal
inotropic support and 23 of the 32 (72%) needed higher ino-
tropic support. Three patients were without any ionotropic
Need for cardiac decongestive therapy (over and above that of
diuretics) was necessary in 10 of 29 survivors to discharge.
Decongestive therapy was required in 2 of 19 (20%) biventricular
hearts compared with 8 of 10 (80%) univentricular hearts
(P< 0.001), obviously resulting in a signiﬁcantly longer hospital
stay (P= 0.005). Bigger shunts were associated with the need for
decongestive therapy (P= 0.054).
The sternum was left open postoperatively in 11 (34%) patients.
Postoperative complications included chylothorax, phrenic nerve
palsy, necrotizing enterocolitis and abdominal bleeding with
unclear focus in 1 patient each.
Of the 29 survivors, 28 (97%) have undergone corrective surgery
with takedown of the BT shunt at the time of this study. Ten
patients were subjected to a bidirectional Glenn anastomosis,
while 18 underwent a biventricular repair. None of the shunt
survivors died during or after the corrective surgery.
Fate of branch pulmonary artery
Seven of 28 patients (25%) had reconstruction of the branch pul-
monary artery at the distal shunt insertion site. Residual stenosis
after BT shunt takedown occurred in 6 patients. In 3 of the
patients, the stenosis occurred despite the pulmonary artery
being reconstructed, while in 3, the stenosis occurred without
the artery being reconstructed.
A brief comparative analysis between 12 shunts in Era I and 20
shunts in Era II is depicted in Table 5.
An ideal shunt helps promote uniform growth of the pulmonary
arteries, without causing distortion. An excessive shunt results in
signiﬁcant diastolic run-off in the short term and elevated pul-
monary vascular resistance or impaired ventricular and atrioven-
tricular valve performance in the long term. Although various
types of shunts have been described [1,7–11], it is the MBTS that
V. Dirks et al. / European Journal of Cardio-Thoracic Surgery1098
has become established as the procedure of choice [4,12]. MBTS
continues to be a subject of academic interest, because of per-
sistent risks associated with this simple-looking procedure.
Sternotomy or thoracotomy
MBTS was classically performed through a thoracotomy.
However, recent trends have shown increasing preference for a
sternotomy approach . A sternotomy saves the child from a
second scar, avoids morbid damage to the thorax with pro-
spects of late scoliosis, but more importantly, the target pul-
monary artery being intrapericardial, it is more accessible for
eventual reconstruction after takedown. Avoiding a thoracotomy
in the prospective Fontan patients has an added advantage of
reducing build-up of lung adhesions to the thoracic wall and
the consequent development of systemic-to-pulmonary artery
collaterals. Other disadvantages of a thoracotomy approach
enumerated in the literature include Horner’s syndrome, distor-
tion of lobar branch pulmonary arteries and preferential ﬂow to
one lung with unbalanced growth . Depending on the side of
the thoracotomy, it may not be always possible to perform
PDA ligation, but with a sternotomy, it is always possible. In the
end, whether or not to close the duct remains a strategic
The sternotomy approach does confront the surgeon with the
challenges of a central run-off from the systemic artery leading
to greater steal, low diastolic pressures, coronary malperfusion
and pulmonary hyperperfusion. In addition, the often-used
truncus brachiocephalicus to the right pulmonary artery shunt
may be at danger of being squashed between the dominant
aorta and the superior vena cava, for which the parietal pericar-
dial reﬂection over the superior vena cava to the trachea should
be divided to create space for the shunt.
Table 3: Mortality details
Cause of death Died
DORV_TGA_PS 19 2.5 3.5 Cardiorespiratory
Pulmonary atresia VSD, hepatopulmonary syndrome, catheter
perforation and emergency shunt
4 2.2 3.5 Cardiorespiratory
Pulmonary atresia intact ventricular septum 6 2.4 4 Myocardial ischaemia 3 (ECMO)
DORV: double outlet right ventricle; TGA: transposition of great arteries; PS:pulmonary stenosis; ECMO: extra corporeal membrane oxygenation.
Table 4: Risk factors for mortality
Weight (kg) 2.94 (1.9–4.4) 2.37 (2.2–2.49) 0.04
Biventricular hearts 19/29 (66) 2/3 (67) 1.00
Univentricular hearts 10/29 (34) 1/3 (33)
Competitive pulmonary flow 21/29 (72) 1/3 (33) 0.22
Use of heart lung machine 2/29 (7) 0/3 (0) 1.00
Size of shunt (mm):
3 8/29 (28) 0/3 (0) 0.47
3.5 17/29 (59) 2/3 (67)
4 4/29 (14) 1/3 (33)
Shunt size/kg body weight 1.19 (0.88–1.58) 1.59 (1.41–1.69) 0.004
1 LDH 7/28
2 ETH 4/28
3 LTH 17/28
Long-term anticoagulation 7/28
Postoperative high ionotropes 20/29 (69) 3/3 (100) 0.52
Shunt size reduction 3/29 (10) 2/3 (67) 0.056
Shunt thrombosis 3/29 (10) 0/3 (0) 1.00
postoperative (day of operation) 83 (77–94) 85 (80–90) 0.97
Hospital stay 23 (5–95) 3 (1–15) 0.02
LDH: low dose (10 IU/kg/h) heparin; ETH: early therapeutic heparin; LTH: late therapeutic heparin.
One patient data missing.
One patient on immediate ECMO was not analysed for acute anticoagulation regimen.
Two patients who died early were not analysed for long-term anticoagulation.
V. Dirks et al. / European Journal of Cardio-Thoracic Surgery 1099
While the proposed alternative shunts such as Potts or
Waterston/Cooley shunts were difﬁcult to regulate, the MBTS
ﬂow is restricted by the size of the graft as also by the size of the
While the Boston group  reported four times higher risk of
shunt failures through a thoracotomy when compared with a
sternotomy, Shauq et al. have reported signiﬁcantly longer
ventilation time, inotropic support, intensive care unit (ICU) stay
and hospital stay in the sternotomy group. These ﬁndings reﬂect
the learning curve involved with shunts created through a
Competitive ﬂow and PDA strategy
A completely left open duct may be difﬁcult to regulate in the
presence of a MBTS. With our technique of duct obliteration
using a silastic sling, one retains the possibility of quickly restor-
ing duct patency in the case of a shunt thrombosis. While a
patent duct imparts signiﬁcant safety in the case of a shunt
failure, some reports have associated patent duct with shunt
thrombosis . Petrucci et al. (Society of Thoracic Surgeons
[STS] database) have shown no association between closed duct
and the risk of composite morbidity . Closing or keeping the
duct open during the MBTS procedure has advantages and dis-
advantages and, hence remains in the end, an individual
In spite of overall improvement in results , mortality reported
ranges from 2.3 to 16% . Our overall postoperative mortality
was 9.4%. Low body weight (P= 0.041) and bigger shunt size/kg
body weight (P= 0.011) were factors associated with post-
operative mortality. There was a trend towards signiﬁcance
between the need for postoperative shunt size reduction and
mortality (P= 0.056). These ﬁndings point towards over-shunting
as a possible indicator of mortality in our series. The Boston
group has reported a mortality of 9 of 102 (8.7%) patients, with
indications that excessive pulmonary blood ﬂow could have con-
tributed to mortality in the sternotomy group. Multivariate risk
factors for mortality in their analysis included small graft size, left
MBTS and male sex . The same group also suggested the use
Table 5: Comparison between eras
Variable Era I (2004–07)
Era II (2008–11)
N12 (38) 20 (62)
Biventricular hearts 6/12 (50) 15/20 (75) 0.25
Univentricular hearts 6/12 (50) 5/20 (25)
Weight (kg) 2.89 (2.2–4.4) 2.93 (1.9–3.8) 0.99
Size of shunt (mm) 3.5 (3.5–4) 3.5 (3–3.5) 0.002
Shunt size/weight 1.24 (0.91–1.69) 1.16 (0.88–1.58) 0.13
Presence of competitive blood flow 7/12 (58) 15/20 (75) 0.44
Use of HLM 0/12 (0) 3/20 (10) 0.52
Shunt distribution, mm (in %)
3 0/12 (0) 8/20 (40) 0.001
3.5 7/12 (58) 12/20 (60)
4 5/12 (42) 0/20 (0)
Right vs left 11/20 (92) right 17/20 (85) right 1.00
1/12 (8) left 3/20 (15) left
Site of take-off (truncus brachiocephalicus vs subclavian artery) 10/12 (83) vs 2/12 (17) 10/20 (50) vs 10/20 (50) 0.08
Early anticoagulation strategy
Low dose 3/10 (30) 4/20 (20) 0.71
Early therapeutic 1/10 (10) 4/20 (20)
Late therapeutic (as defined in Table 4) 6/10 (60) 12/20 (60)
Aspirin 8/10 (80) 14/19 (74) 1.00
Therapeutic 2/10 (20) 5/19 (26)
Postoperative ionotropic support
None 1/12 (8) 2/20 (10) 0.48
Normal 1/12 (8) 5/20 (25)
10/12 (83) 13/20 (65)
Need for shunt reduction 3/12 (25) 2/20 (10) 0.34
Decongestive therapy (more than diuretics) 6/9 (67) 4/20 (20) 0.03
Duration of ventilation 2 (1–15) 1.5 (0–9) 0.53
ICU stay 4 (1–15) 5 (1–13) 0.39
Duration of hospital stay 18.5 (1–95) 22.5 (5–84) 0.69
Mortality 3/12 (25) 0/20 (0) 0.04
Shunt thrombosis 0/12 (0) 3/20 (15) 0.27
before takedown 82 (72–93) 81.5 (73–94) 0.66
Residual branch PA stenosis at shunt insertion site 4/9 (44) 1/15 (7) 0.05
Normal ionotropic support is defined as adrenalin ≤0.05 and/or noradrenaline ≤0.05 and/or milrinone ≤0.75 μg/kg/min.
V. Dirks et al. / European Journal of Cardio-Thoracic Surgery1100
of smaller (3.5 mm) shunts through a sternotomy approach
instead of the 4-mm shunts for the thoracotomy approach. An
STS database harvest study  has identiﬁed preoperative venti-
lation, pulmonary atresia_intact ventricular septum, univentricu-
lar hearts and weight <3 kg as risk factors for mortality.
Pulmonary atresia with intact ventricular septum, when speciﬁc-
ally analysed, did not come out as a signiﬁcant risk factor for
mortality in our cohort, probably because of small numbers.
While Alkhulaiﬁet al. identiﬁed weight <2 kg and preopera-
tive ventilation, Rao et al. identiﬁed restrictive atrial septal
defect, univentricular physiology and postoperative intervention
as risk factors for mortality.
Shunt thrombosis is a grave complication of the MBTS proced-
ure. Our overall acute shunt thrombosis of (3 of 32) 9.4% corre-
sponds with those of (9 of 76) 11.8% reported from Bristol and
(14 of 102) 13.7% reported from Boston . We could not show
an association between smaller shunt size and occurrence of
thrombosis, probably because of the small numbers. Tsai et al.
 and Tamisier et al. have suggested that young age and
smaller size are signiﬁcantly related to shunt thrombosis. Other
reports have also linked weight <2 kg  and weight <3.6 kg 
to shunt thrombosis. Gedicke et al. have found weight <3 kg,
high preoperative haemoglobin (>18 g/dl) and a postoperative
patent duct as signiﬁcant factors for shunt thrombosis.
Although an association between an anticoagulation regimen
and shunt thrombosis could not be established in our study, it
does not belittle the role of postoperative anticoagulation, par-
ticularly in high-risk patients. Al Jubair et al. have shown,
less-shunt failure occurs if heparin is given before clamping. An
early postoperative phase with a fresh anastomosis, coupled with
phases of low systemic pressures, pulmonary hypertension, ex-
ternal compression and resulting stasis, can initiate thrombus
formation. It is these uncertainties that can be positively inﬂu-
enced by early anticoagulation. Li et al. have demonstrated a
beneﬁcial effect of acetylsalicylic acid in infants palliated with a
shunt, with reduced incidence of shunt thrombosis and death.
Another prospective study has shown the beneﬁcial effect of
haemodilution with a signiﬁcantly higher shunt patency rate .
Rare coagulopathies, such as protein C deﬁciency , and
primary antiphospholipid syndrome  have also been
reported to cause shunt thrombosis.
Late shunt obstruction
We have not observed any late shunt thrombosis in this series of
patients. This has been reported as a cause in up to 15% of
out-of-hospital mortalities [14,15]. Wells et al. have observed
>50% obstruction of the MBTS in 21% of their patients and have
identiﬁed a shunt size of <4 mm to be a risk factor for high-
grade stenosis (>50%).
While the cohort did not change over time in terms of most
demographic and procedural variables, shunt size was signiﬁ-
cantly lower in the latter half of the series when compared with
the former half. Shunt thrombosis was higher in the later era,
but did not reach statistical signiﬁcance. These ﬁndings may indi-
cate that small shunts are prone to shunt thrombosis. Mortality
as well as need for cardiac decongestive therapy was signiﬁcantly
higher in the previous era. With time, while mortality was
avoided, shunt thrombosis remained worrisome. While shunt
thrombosis morbidity could be partially attributed to the tech-
nique, the importance of optimal intensive postoperative man-
agement cannot be overemphasized. Interestingly, in spite of
smaller shunt size selection in the later era, transcutaneous satur-
ation before shunt takedown was 82% in both the eras. This
implies that the shunt ﬂow was adequately regulated by the
artery from which the shunt was sourced.
This is a retrospective study with a small patient cohort, which
may not be powered enough to identify all risk factors contribut-
ing to the various end points. The series being spread over a
time frame of 8 years, even generalized improvements in opera-
tive technique and perioperative care may alone account for the
improvements in outcome.
In spite of increasing conﬁdence with primary neonatal intracardiac
repairs, the MBTS continues to be indicated for malformations
of the univentricular pathway. Although seemingly innocuous,
the MBTS procedure is associated with signiﬁcant morbidity and
mortality. While small shunts may have a tendency to shunt throm-
bosis, large shunts may lead to pulmonary over-circulation and
volume loading of the heart. Various studies have identiﬁed
low body weight, small shunt size, over-shunting, univentricular
hearts—speciﬁcally pulmonary atresia with intact ventricular
septum, to be risk factors associated with postoperative morbidity
and mortality. It appears that timely and efﬁcient early anticoagula-
tion as well as long-term antiplatelet therapy may help reduce the
risk of early and late shunt dysfunction.
Conﬂict of interest: none declared.
 Blalock A, Taussig HB. Landmark article May 19, 1945: the surgical treat-
ment of malformations of the heart in which there is pulmonary stenosis
or pulmonary atresia. By Alfred Blalock and Helen B Taussig. JAMA 1984;
 Mavroudis C, Backer CL. Palliative operations. In: Mavroudis C, Backer CL
(eds). Pediatric Cardiac Surgery. 3rd edn. Philadelphia: Mosby, 2003,
 Geiss D, Williams WG, Lindsay WK, Rowe RD. Upper extremity gangrene: a
complication of subclavian artery division. Ann Thorac Surg 1980;30:487–9.
 de Leval MR, McKay R, Jones M, Stark J, Macartney FJ. Modiﬁed
Blalock-Taussig shunt. Use of subclavian artery oriﬁce as ﬂow regulator
V. Dirks et al. / European Journal of Cardio-Thoracic Surgery 1101
in prosthetic systemic-pulmonary artery shunts. J Thorac Cardiovasc Surg
 Williams JA, Bansal AK, Kim BJ, Nwakanma LU, Patel ND, Seth AK et al.
Two thousand Blalock-Taussig shunts: a six-decade experience. Ann
Thorac Surg 2007;84:2070–5; discussion 2070–5.
 Odim J, Portzky M, Zurakowski D, Wernovsky G, Burke RP, Mayer JE Jr
et al. Sternotomy approach for the modiﬁed Blalock-Taussig shunt.
 Potts WJ, Smith S, Gibson S. Anastomosis of the aorta to a pulmonary
artery: certain types in congenital heart disease. J Am Med Assoc 1946;
 Waterston DJ. Treatment of Fallot’s tetralogy in children under 1 year of
age. Rozhl Chir 1962;41:181–3.
 Cooley DA, Hallman GL. Intrapericardial aortic-right pulmonary arterial
anastomosis. Surg Gynecol Obstet 1966;122:1084–6.
 Gazzaniga AB, Elliott MP, Sperling DR, Dietrick WR, Eiseman JT,
McRae DM et al. Microporous expanded polytetraﬂuoroethylene
arterial prosthesis for construction of aortopulmonary shunts: experi-
mental and clinical results. Ann Thorac Surg 1976;21:322–7.
 Amato JJ, Marbey ML, Bush C, Galdieri RJ, Cotroneo JV, Bushong J.
Systemic-pulmonary polytetraﬂuoroethylene shunts in palliative opera-
tions for congenital heart disease. Revival of the central shunt. J Thorac
Cardiovasc Surg 1988;95:62–9.
 Tamisier D, Vouhe PR, Vernant F, Leca F, Massot C, Neveux JY. Modiﬁed
Blalock-Taussig shunts: results in infants less than 3 months of age.
Ann Thorac Surg 1990;49:797–801.
 Shauq A, Agarwal V, Karunaratne A, Gladman G, Pozzi M, Kaarne M
et al. Surgical approaches to the Blalock shunt: does the approach
matter? Heart Lung Circ 2010;19:460–4.
 Gedicke M, Morgan G, Parry A, Martin R, Tulloh R. Risk factors for acute
shunt blockage in children after modiﬁed Blalock-Taussig shunt opera-
tions. Heart Vessels 2010;25:405–9.
 Petrucci O, O’Brien SM, Jacobs ML, Jacobs JP, Manning PB, Eghtesady P.
Risk factors for mortality and morbidity after the neonatal
Blalock-Taussig shunt procedure. Ann Thorac Surg 2011;92:642–51;
 AlkhulaiﬁAM, Lacour-Gayet F, Serraf A, Belli E, Planche C. Systemic
pulmonary shunts in neonates: early clinical outcome and choice of
surgical approach. Ann Thorac Surg 2000;69:1499–504.
 Rao MS, Bhan A, Talwar S, Sharma R, Choudhary SK, Airan B et al.
Modiﬁed Blalock-Taussing shunt in neonates: determinants of immediate
outcome. Asian Cardiovasc Thorac Ann 2000;8:339–43.
 Tsai KT, Chang CH, Lin PJ. Modiﬁed Blalock-Taussig shunt: statistical
analysis of potential factors inﬂuencing shunt outcome. J Cardiovasc
Surg (Torino) 1996;37:149–52.
 Bove EL, Kohman L, Sereika S, Byrum CJ, Kavey RE, Blackman MS et al.
The modiﬁed Blalock-Taussig shunt: analysis of adequacy and duration
of palliation. Circulation 1987;76:III19–23.
 Al Jubair KA, Al Fagih MR, Al Jarallah AS, Al Yousef S, Ali Khan MA,
Ashmeg A et al. Results of 546 Blalock-Taussig shunts performed in 478
patients. Cardiol Young 1998;8:486–90.
 Li JS, Yow E, Berezny KY, Rhodes JF, Bokesch PM, Charpie JR et al.
Clinical outcomes of palliative surgery including a systemic-to-
pulmonary artery shunt in infants with cyanotic congenital heart disease:
does aspirin make a difference? Circulation 2007;116:293–7.
 Sahoo TK, Chauhan S, Sahu M, Bisoi A, Kiran U. Effects of hemodilution
on outcome after modiﬁed Blalock-Taussig shunt operation in children
with cyanotic congenital heart disease. J Cardiothorac Vasc Anesth 2007;
 Watanabe M, Aoki M, Fujiwara T. Thrombotic occlusion of Blalock-
Taussig shunt in a patient with unnoticed protein C deﬁciency. Gen
Thorac Cardiovasc Surg 2008;56:544–6.
 Deally C, Hancock BJ, Giddins N, Hawkins L, Odim J. Primary antipho-
spholipid syndrome: a cause of catastrophic shunt thrombosis in the
newborn. J Cardiovasc Surg (Torino) 1999;40:261–4.
 Wells WJ, Yu RJ, Batra AS, Monforte H, Sintek C, Starnes VA. Obstruction
in modiﬁed Blalock shunts: a quantitative analysis with clinical correl-
ation. Ann Thorac Surg 2005;79:2072–6.
V. Dirks et al. / European Journal of Cardio-Thoracic Surgery1102