The Extracardiac Fontan Procedure Without
Cardiopulmonary Bypass: Technique and
Intermediate-Term Results
Anji T. Yetman, MD, Jonathan Drummond-Webb, MD, William P. Fiser, MD,
Michael L. Schmitz, MD, Michiaki Imamura, MD, PhD, Sana Ullah, MD,
Ryan J. Gunselman, BSPS, Carl W. Chipman, RN, CCP, Charles E. Johnson, RN, CCP,
and Stephen H. Van Devanter, MD
Departments of Cardiovascular Surgery, Pediatric Cardiovascular Anesthesiology, and Pediatric Cardiology, The University of
Arkansas for Medical Sciences and Arkansas Children’s Hospital, Little Rock, Arkansas
Background. The extracardiac Fontan procedure (ECF)
usually requires cardiopulmonary bypass (CPB). In this
report, the results and techniques of this procedure
without CPB at a single institution are presented.
Methods. Between August 1992 and December 2001,
ECF without CPB was achieved in 24 of 44 patients
undergoing an ECF. Mean age at surgery was 5.9 ⴞ 2.9
years, and mean weight was 20.7 ⴞ 12.6 kg. Diagnoses
were tricuspid atresia in 9 patients, single-ventricle with
pulmonary outflow tract obstruction in 7, pulmonary
atresia/intact septum in 5, and other complex single-
ventricle physiology in 3. Initial palliation was by arte-
rial to pulmonary artery shunt in 21 and pulmonary
artery banding in 1. A bidirectional cavopulmonary con-
nection was created in 23 patients. A temporary inferior
vena caval–to–atrial shunt was used to complete the
procedure without CPB. Median graft size was 16 mm
(range 14 to 20 mm).
S ince the introduction of the classic Fontan procedure
in 1971 [1], a variety of surgical and technical mod-
ifications have been introduced [2– 4], with associated
improvements in morbidity and mortality. Modifications
have included a staged approach to completion of the
Fontan by means of a bidirectional cavopulmonary con-
nection (BCPC) [2], the lateral tunnel modification, and
the concept of a total cavopulmonary connection [1, 3].
Most recently, use of an extracardiac conduit to achieve a
total cavopulmonary connection (ie, the extracardiac
Fontan [ECF]) has become the strategy of many surgeons
[4]. The theoretical benefits include the following: avoid-
ance of aortic crossclamping and myocardial ischemia [5],
shortened duration of cardiopulmonary bypass (CPB) [5],
Presented at the Eighth Annual Cardiothoracic Techniques and Technol-
ogies Meeting 2002, Miami Beach, FL, Jan 23–26, 2002.
Address reprint requests to Dr Drummond-Webb, Chief, Department of
Pediatric Cardiovascular Surgery, Arkansas Children’s Hospital, 800
Marshall Street, Slot 677, Little Rock, AR 72202-3591; e-mail: drummond-
webbjonathan@uams.edu.
© 2002 by The Society of Thoracic Surgeons
Published by Elsevier Science Inc
Results. There was no early mortality, and 68% of
patients were discharged without complications. Compli-
cations included persistent cyanosis in 4 patients, persis-
tent pleural effusions in 2 (one chylous), and phrenic
nerve injury in 1. Median postoperative hospital stay was
16 days (range 10 to 50) days. At a mean follow-up of 44
ⴞ 28 months, there was no conduit obstruction. One
patient died 11 months postoperatively, and 1 patient
received a heart transplant 26 months post-ECF.
Conclusions. At intermediate term follow-up, the ECF
without CPB appears to be safe and technically reproduc-
ible in selected cases. Ongoing follow-up of these pa-
tients is necessary to document the theoretical advan-
tages of avoiding CPB.
(Ann Thorac Surg 2002;74:S1416 –21)
© 2002 by The Society of Thoracic Surgeons
improved long-term hemodynamics [6], better hydrody-
namics and flow characteristics [6], and a lower incidence
of atrial arrhythmias [7].
Although outcomes have improved with this proce-
dure [8], perioperative complications are not insignifi-
cant; some are related to the need for CPB, with the
attendant risks of activation of vasoactive cytokines sec-
ondary to exposure to a foreign surface. The technique of
performing the ECF without the use of CPB has been
reported in small series [9, 10]. Tam and colleagues [9]
reported on 21 patients undergoing ECF off CPB, but a
variety of conduit types and sizes were used. Immediate
postoperative hemodynamics appeared superior when
compared to those patients undergoing ECF with CPB
[9]. Long-term benefits and potential problems, however,
are still unknown. We report our technique of perform-
ing the ECF procedure using a polytetrafluoroethylene
conduit for the ECF without the use of cardiopulmonary
bypass. Intermediate term results of this strategy are
presented.
0003-4975/02/$22.00
PII S0003-4975(02)03922-X
Ann Thorac Surg CTT SUPPLEMENT YETMAN ET AL S1417
2002;74:S1416 –21 EXTRACARDIAC FONTAN WITHOUT BYPASS

Table 1. Indication for Cardiopulmonary Bypass at Time of

Completion of Extracardiac Fontan Procedure
Patients
Procedure (n)

Pulmonary arterioplasty/ 8
intraoperative stent placement
Right atrial reduction plasty 2
Resection of LVOTO 2
Tricuspid valvectomy 1
Damus-Stansel-Kaye procedure 1
Atrioventricular valve repair 1
Surgeon preference 5
Total 20

LVOTO ⫽ left ventricular outflow tract obstruction.

Material and Methods

Since 1982, a total of 86 Fontan procedures were per-
formed at our institution. The first ECF without the use of
CPB was performed in August 1992. Between August
1992 and December 2001, we performed 44 ECF proce-
dures, of which 24 were performed without CPB. We
perform the ECF procedure electively at a weight of 15
kg. The ECF technique without CPB is our preferred
procedure, provided that no intracardiac repair or exten-
sive pulmonary arterial reconstruction is needed. Indica-
tions for use of CPB to achieve an ECF are shown in Table
1. In 4 patients, the off-bypass procedure was aborted. In
1 patient, a faulty vascular clamp slipped off the atrium,
requiring urgent conversion to CPB. In 2 patients with a
diagnosis of left atrial isomerism with azygos continua-
tion of an interrupted inferior vena cava (IVC), exposure
of the hepatic venous-right atrial junction resulted in
significant hemodynamic compromise, requiring CPB.
One patient experienced desaturation with placement of
the pulmonary artery clamp, and CPB was used for a
10-minute period to complete the anastomosis.
Patient Population
All consecutive patients undergoing an ECF without the
use of cardiopulmonary bypass were identified from an
institutional review board–approved database. Demo-
graphic and preoperative data including anatomical di-
agnoses, age, weight, prior surgical procedures, and
interval between BCPC to ECF were analyzed. Perioper-
ative data including length of stay in the intensive care
unit, length of mechanical ventilation, need for blood
transfusion, surgical complications and the need for
reoperation/reintervention were noted. Follow-up data
including duration of follow-up, echocardiographic as-
sessment of ventricular function, atrioventricular valve
regurgitation, presence of arrhythmias, and New York
Heart Association classification of symptoms were re-
viewed from the “Cardiac Surgery Outcomes Database”
at the Arkansas Children’s Hospital and approved by the
Ethics Review Board of the University of Arkansas for
Medical Sciences on July 31, 2001 (institutional review
board record number 06874, University of Arkansas for
Medical Sciences Assurance M-1451, IORG0000345).
Fig 1. Distance X to Y determines diameter of polytetrafluoroethyl-
ene conduit that can be placed. Inferior vena caval right atrial junc-
tion is divided and oversewn.
Surgical Technique
Our surgical technique is as follows. A repeat median
sternotomy is performed in standard fashion. A CPB
circuit is primed, with perfusion staff in attendance for all
cases. Steroids (10 to 20 mg/kg of hydrocortisone) are
administered preoperatively to all patients. All patients
are partially heparinized, with the aim of an activated
clotting time (ACT) in excess of 300 seconds. Heparin
dosage is titrated to the preintervention ACT, with intra-
operative monitoring of the ACT. Heparinization is not
reversed at the conclusion of the procedure.
Key points of dissection include complete liberation of
the IVC junction (dissecting below the diaphragm if
necessary) and complete freeing up of the BCPC, main,
and branch pulmonary arteries. We consider it particu-
larly important to free up the lateral pericardial recess, as
this allows the conduit to lie in place without compres-
sion of the underlying pulmonary veins and atrial mass,
and also allows a longer graft to be placed with a gentle
anterior curve providing some length for somatic growth.
The polytetrafluoroethylene (PTFE) tube size is re-
stricted by the length of the pulmonary artery segment
available for construction of the superior anastomosis.
This is measured (Fig 1, distance X to Y) and the largest
size conduit chosen. In the first 7 patients, the PTFE
S1418 CTT SUPPLEMENT YETMAN ET AL
EXTRACARDIAC FONTAN WITHOUT BYPASS
Fig 2. Temporary atrial-to–inferior vena cava shunt (arrow). Bidi-
rectional cavopulmonary connection is already established. “Low”
placement of inferior vena cava cannula is shown.
conduit was anastomosed end-to-side to the IVC using a
side-biting clamp. Currently, the distal IVC to conduit
anastomosis is performed first. After heparinization, an
IVC to atrial shunt is constructed using two right-angle
venous cannulas, which are connected with the shortest
length of tubing so as to minimize resistance and to keep
it out of the field. The shunt is placed as low as possible
on the IVC (Fig 2). The shunt is very carefully deaired. A
vascular clamp is placed across the inferior portion of the
right atrium, with care taken to identify and to avoid the
coronary sinus. A running suture is placed below the
clamp before dividing the atrial–IVC junction (earlier in
our experience, a faulty clamp dislodged). The junction is
divided and the atrium is doubly oversewn. The IVC-to-
extracardiac anastomosis is completed end-to-end (Fig 3).
A side-biting vascular clamp is placed on the central
portion of the pulmonary artery confluence so as to
preserve blood flow through the BCPC. At this time,
ventilatory adjustments, a higher fraction of inspired
oxygen, and careful tracking of the systemic saturation
and arterial oxygen are essential to allow safe completion
of the ECF. The conduit-to–pulmonary artery anastomo-
Ann Thorac Surg
2002;74:S1416 –21
Fig 3. Diagram showing completion of distal anastomosis. Polytetra-
fluoroethylene conduit (*) is trimmed obliquely, and a partial occlu-
sion clamp allows the bidirectional cavopulmonary connection to
perfuse the lungs. Note offset of polytetrafluoroethylene conduit
anastomotic site relative to bidirectional cavopulmonary connection.
sis is then completed. This end of the conduit is beveled
and sewn end-to-side to the inferior aspect of the pul-
monary artery (Fig 3). Care is taken to offset this connec-
tion from the superior BCPC orifice. The conduit is
carefully deaired, and blood flow from the IVC to the
pulmonary arteries is restored. The IVC-to-atrial shunt is
removed, and a pressure monitoring line is placed
through the atrial pursestring. In patients with left atrial
isomerism, interrupted IVC and azygos continuation, the
distal end of the conduit is placed end-to-end to the
hepatic venous confluence. Fenestration is not routinely
used.
In 1 patient, placement of the clamp across the main
pulmonary artery resulted in unacceptable hypoxia. A
temporary aortic-to–pulmonary artery shunt was created
by placing a 14-mm Medtronic DLP arterial cannula
(Medtronic Inc, Minneapolis, MN) into both the ascend-
ing aorta and left pulmonary artery (Fig 4), thus allowing
completion of the ECF without CPB.
Statistical Analysis
Frequencies of preoperative, postoperative, and fol-
low-up variables were assessed. Data are expressed as
Ann Thorac Surg
2002;74:S1416 –21
Fig 4. Temporary systemic arterial to pulmonary artery shunt in
place (arrow). This allows improved pulmonary flow while the
proximal polytetrafluoroethylene conduit (*) to pulmonary artery
anastomosis is completed.
median with range and as mean ⫾ standard deviation as
appropriate.
Results
Preoperative Data
There were 8 female and 16 male patients. Anatomical
diagnoses included tricuspid atresia in 9 and single
ventricle with obstructed pulmonary blood flow in 7; of
these, 3 patients had dextrocardia, single ventricle with
left atrial isomerism, unobstructed pulmonary blood
flow, and azygos continuation of an interrupted IVC in 2,
pulmonary atresia/intact ventricular septum in 5, and
hypoplastic left heart in 1. Prior surgical procedures on
the pulmonary arteries included pulmonary artery band-
ing in 2 patients and creation of a systemic arterial to
pulmonary artery shunt in 22.
CTT SUPPLEMENT YETMAN ET AL S1419
EXTRACARDIAC FONTAN WITHOUT BYPASS
Fig 5. Relationship between body weight and polytetrafluoroethylene
conduit size.
All patients except 1 had a BCPC before completion of
the Fontan procedure at a median age of 13 months
(range 6 to 60 months). The mean interval from BCPC to
Fontan was 43.2 ⫾ 12.6 months. The ECF off CPB was
performed at a mean age of 5.7 ⫾ 2.9 years and a mean
weight of 20.3 ⫾ 10.3 kg.
Perioperative Data
The PTFE conduit sizes ranged from 14 to 20 mm (median
16 mm). The relationship of conduit size to body weight
is shown in Figure 5. The average hospital stay was 16
days (range 10 to 50 days), which was reduced to 6.5 days
for the year 2001. The median time of mechanical venti-
lation time was 20 hours (range 2 to 165 hours), which
was also reduced to 9.3 hours for the year 2001. A single
transfusion of packed red blood cells (10 mL/kg) was
required postoperatively in 10 of the 24 patients to
maintain a hematocrit of 33 or greater. One additional
patient required more than one transfusion. It is our
standard protocol to provide inotropic support with do-
pamine 5 ␮g 䡠 kg⫺1 䡠 min⫺1 and afterload reduction with
nitroglycerin 2.5 ␮g 䡠 kg⫺1 䡠 min⫺1 for 24 hours postop-
eratively after which time inotropic support is stopped
and a nitropatch placed. No patient required additional
inotropic support.
The majority of patients had an uneventful recovery
with timely discharge. Perioperative complications in-
cluded persistent cyanosis in 4 patients, with 3 patients
requiring reintervention during the postoperative period
for ligation of collaterals and 1 for unrecognized left
pulmonary artery stenoses. Persistent chylous pleural
and pericardial effusions were noted in 3 patients, with 2
requiring chemical pleurodesis and 1 requiring thoracic
duct ligation after failed chemical pleurodesis. These few
patients had lengthy stays, thus contributing to an overall
longer median hospital stay. One patient sustained dam-
age to the phrenic nerve, which was self-limiting. There
were no neurologic events.
S1420 CTT SUPPLEMENT YETMAN ET AL
EXTRACARDIAC FONTAN WITHOUT BYPASS
One patient was given a bolus dosage of digoxin for an
episode of supraventricular tachycardia while the chest
was being closed after an uneventful off-pump ECF.
Severe hemodynamic compromise with ventricular fi-
brillation ensued. The etiology was immediately recog-
nized and the patient was placed onto CPB and an
extracorporeal membrane oxygenation circuit prepared.
The patient was converted to extracorporeal membrane
oxygenation and supported for 48 hours. He was subse-
quently weaned, separated, and suffered no apparent
immediate ill effect, but died 11 months postoperatively
of plastic bronchitis.
Follow-Up
Patients were followed for a period of up to 97.2 months
with a mean follow-up duration of 44.4 ⫾ 28.8 months.
One patient was lost to follow-up after a period of 2
years. There was no early death, but one late death.
Three patients received a pacemaker for symptomatic
sinus bradycardia with junctional escape. Routine elec-
trocardiography and 24-hour ambulatory Holter moni-
toring were performed in all patients. All patients were
noted to be in sinus rhythm on most recent evaluation,
with the exception of 2 of the 3 patients who had
pacemaker implantation after ECF who received atrial
pacing. One patient who had complete atrioventricular
block before ECF remained DDD (dual chamber sensing,
pacing, and inhibited and triggered) paced. Intermittent
episodes of asymptomatic sinus bradycardia with junc-
tional escape were noted in 2 patients. No patient had
supraventricular or ventricular tachycardia. All patients
were followed-up routinely with annual echocardiogra-
phy for assessment of ventricular function, atrioventric-
ular valve regurgitation, and assessment of obstruction of
the Fontan circuit. One patient who was chronically DDD
paced developed progressive systemic ventricular dys-
function and severe atrioventricular valve regurgitation;
he underwent cardiac transplantation 26 months after
ECF. The only late mortality in the series occurred 11
months post-ECF in a patient dying of septic complica-
tions in the setting of spina bifida and plastic bronchitis.
Two patients were noted to have mild ventricular dys-
function and 8 patients mild atrioventricular valve regur-
gitation. Echocardiography revealed no evidence of pul-
monary artery distortion or obstruction of the Fontan
circuit on late follow-up. All surviving patients were in
New York Heart Association class I or II.
Long-term complications were rare. Patients were not
routinely maintained on Coumadin (Du Pont Pharma-
ceuticals, Wilmington, DE) until the last year of the study
because of a general change in management strategy.
Currently, all patients undergoing the Fontan procedure
receive anticoagulation with Coumadin. Despite lack of
anticoagulation in most patients, there were no incidents
of graft thrombosis or obstruction. No patient has re-
quired conduit replacement; no patient has developed
bacterial endocarditis; and no patient undergoing ECF
without CPB has developed protein-losing enteropathy.
Ann Thorac Surg
2002;74:S1416 –21
Comment
Surgical management of patients with single ventricle
physiology has evolved considerably over the last few
decades. Improved patient selection and understanding
of risk factors, staged preparation toward a Fontan pro-
cedure, as well as improvements in perioperative man-
agement have resulted in substantial improvements in
morbidity and mortality. The ECF procedure represents
just one such technical evolution. Avoidance of intraatrial
suture lines, intracardiac prosthetic material, and expo-
sure of the atria to increased venous pressure are just a
few of the reasons why this procedure has become the
preferred technique [8]. Azakie and colleagues [3] dem-
onstrated prolonged CPB time to be the only indepen-
dent predictor of mortality and risk of complications in
patients undergoing Fontan procedures. We have dem-
onstrated that further optimization of the ECF procedure
is possible and that completion of the procedure off CPB
can be achieved in selected patients. Importantly, we do
not compromise the Fontan principles or risk residual
lesions to achieve off-CPB status. Despite initial concerns
of an increased risk of thromboembolic complications
with this procedure, we and others [11] have documented
no thrombosis in these patients. We currently recom-
mend life-long anticoagulation, as it has been suggested
that there are abnormal coagulation factors in patients
who have had a Fontan procedure [12]. In addition,
Jahangiri and colleagues [13] have documented a 16%
incidence of thromboembolism after Fontan procedure.
Concerns over the optimal size of the conduit and the
lack of growth thereof have been raised. These issues are
valid, as conduit replacement in the setting of a Fontan
patient is a significant undertaking. It has previously
been shown that at 2 to 4 years of age, and 12 to 15 kg of
body weight, the IVC-to–pulmonary artery distance is up
to 80% that of adult size [14]. We electively perform ECF
when the patient has reached a weight of 15 kg. During a
follow-up period in excess of 8 years, we have not had a
patient require conduit replacement, nor have we docu-
mented any obstruction of the conduit within this time.
In conclusion, the ECF, with many of its inherent
benefits, can be achieved without the use of CPB in
selected patients. The surgical procedure and protocols
are simple and easily reproducible. At intermediate term
follow-up, this strategy has resulted in acceptable out-
comes. Ongoing evaluation of this cohort is required to
validate whether this strategy offers any objective advan-
tage over those protocols that routinely use CPB.
This work was supported by funding from the Arkansas Log-A-
Load Endowed Chair of Pediatric Cardiovascular Surgery at
Arkansas Children’s Hospital. The surgical input of James E.
Harrell, MD, Jean Anne Phillips, RN, Carole J. Wilson, RN,
Kalen Rogers, RN, Patrick Young, RN, James Traylor, RN,
Shawn Hill, RN, Michael Tucker, RN, and the anesthesia con-
tribution of Gerald A. Bushman, MD, are recognized in this
series. All data used in this report were obtained from an
institutional review board–approved database.
Ann Thorac Surg
2002;74:S1416 –21
References
1. Fontan F, Baudet E. Surgical repair of tricuspid atresia.
Thorax 1971;26:240 –8.
2. DeLeval M, Kilner P, Gewellig M, Bull C. Total cavopulmo-
nary connection: a logical alternative to atrial pulmonary
connection for complex Fontan operations. J Thorac Cardio-
vasc Surg 1988;96:682–95.
3. Azakie A, McCrindle BW, Van Arsdell GS, et al. Extracardiac
conduit versus lateral tunnel cavopulmonary connections at
a single institution: impact on outcomes. J Thorac Cardio-
vasc Surg 2001;122:1219 –28.
4. Marcelletti CF, Hanley FL, Mavroudis C. Revision of previ-
ous Fontan connections to total extracardiac cavopulmonary
anastomosis: a multicenter experience. J Thorac Cardiovasc
Surg 2000;199:340 –6.
5. Petrossian E, Thompson LD, Hanley FL. Extracardiac con-
duit variation of the Fontan procedure. Advances in Cardiac
Surgery 2000;12:175–98.
6. Lardo AC, Webber SA, Friehs I, et al. Fluid dynamic com-
parison of intraatrial and extra cardiac total cavopulmonary
connections. J Thorac Cardiovasc Surg 1999;117:697–704.
7. Amodeo A, Galletti L, Marianeschi S, et al. Extracardiac
Fontan operation for complex cardiac anomalies: seven
CTT SUPPLEMENT YETMAN ET AL S1421
EXTRACARDIAC FONTAN WITHOUT BYPASS
years experience. J Thorac Cardiovasc Surg 1997;114:1020 –
31.
8. Van Arsdell GS, McCrindle BW, Einarson KD, et al. Inter-
ventions associated with minimal Fontan mortality. Ann
Thorac Surg 2000;70:568 –74.
9. Tam VK, Miller BE, Murphy K. Modified Fontan without use
of cardiopulmonary bypass. Ann Thorac Surg 1999;68:1698 –
1704.
10. McElhinney DB, Petrossian E, Reddy VM, et al. Extracardiac
conduit Fontan procedure without cardiopulmonary bypass.
Ann Thorac Surg 1998;66:1826 –8.
11. Marcelletti CF, Iorio FS, Abella RF. Late results of extracar-
diac Fontan repair. Semin Thorac Cardiovasc Surg Pediatr
Card Surg 1999;2:131–42.
12. Cromme-Dijkhuis AH, Hess J, Hahlen K, et al. Specific
sequelae after Fontan operation at mid- and long-term
follow up. J Thorac Cardiovasc Surg 1993;106:1126 –32.
13. Jahangiri M, Ross DB, Redington AN, et al. Thromboembo-
lism after the Fontan procedure and its modifications. Ann
Thorac Surg 1994;58:1409 –14.
14. Alexi-Meskishvili V, Ovroutski S, Ewert P, et al. Optimal
conduit size for extracardiac Fontan operation. Eur J Cardio-
thorac Surg 2000;18:690 –5.