Journal of the American College of Cardiology
© 2004 by the American College of Cardiology Foundation
Published by Elsevier Inc.
Vol. 44, No. 7, 2004
ISSN 0735-1097/04/$30.00
doi:10.1016/j.jacc.2004.06.061
The Effects of Carbon Dioxide on
Oxygenation and Systemic, Cerebral, and
Pulmonary Vascular Hemodynamics After the
Bidirectional Superior Cavopulmonary Anastomosis
Aparna Hoskote, MD,* Jia Li, PHD,†‡ Chantal Hickey, MD,§ Simon Erickson, MD,*
Glen Van Arsdell, MD,储¶ Derek Stephens, MSC,# Helen Holtby, MD,§ Desmond Bohn, MD,*
Ian Adatia, MBCHB*†‡
Toronto, Canada
We investigated the effects of different CO2 tensions on oxygenation, pulmonary blood flow
(Qp), cerebral blood flow, and systemic blood flow (Qs) after the bidirectional superior
cavopulmonary anastomosis (BCPA).
BACKGROUND Hypoxemia refractory to management of a high pulmonary vascular resistance index (PVRI)
may complicate recovery from the BCPA.
METHODS
After BCPA, CO2 was added to the inspired gas of mechanically ventilated patients. The Qp,
Qs, PVRI, and systemic vascular resistance index (SVRI) were calculated from oxygen
consumption, intravascular pressures, and oxygen saturations. Cerebral blood flow was
estimated by near infrared spectroscopy and transcranial Doppler.
RESULTS
In nine patients (median age 7.1, range 2 to 23 months), arterial oxygen tension increased
significantly (p ⬍ 0.005) from 36 ⫾ 6 mm Hg to 44 ⫾ 6 to 50 ⫾ 7 mm Hg at arterial carbon
dioxide tensions (PaCO2) of 35, 45, and 55 mm Hg, respectively and decreased to 40 ⫾ 8 mm
Hg at PaCO2 40 mm Hg. At a PaCO2 of 55 and 45 compared with 35 mm Hg, Qp, cerebral
blood flow, and Qs increased significantly, PVRI, Qp/Qs, and the ratio of Qp to inferior vena
caval blood flow were unchanged, but SVRI decreased.
CONCLUSIONS We have demonstrated that after the BCPA, systemic oxygenation, Qp, Qs, and cerebral blood
flow increased and SVRI decreased at CO2 tensions of 45 and 55 mm Hg compared with 35 mm
Hg. We suggest that hypoxemia after the BCPA is ameliorated by a higher PaCO2 and that low
PaCO2 or alkalosis may be detrimental. Hypercarbic management strategies may allow earlier
progression to the BCPA, which may contribute to reducing the interval morbidity in patients
with a functional single ventricle. (J Am Coll Cardiol 2004;44:1501–9) © 2004 by the
American College of Cardiology Foundation
OBJECTIVES
The bidirectional superior cavopulmonary anastomosis
(BCPA) increases effective pulmonary blood flow (Qp)
without increasing ventricular volume in children with
congenital heart disease and a single functional ventricle
(1,2). The success of the BCPA is considered to be
exquisitely dependant on an unobstructed, low-resistance
pulmonary vascular bed, and preoperative evaluation focuses
upon identifying adverse hemodynamic parameters (3).
Nevertheless, despite preoperative scrutiny, hypoxemia after
the BCPA may be profound and defy therapy. Hypoxemia
in the early postoperative period is considered often to be
consequent upon the transiently elevated pulmonary vascular resistance that occurs after cardiopulmonary bypass and
may limit the age at which the BCPA can be performed
safely (4,5). Paradoxically, postoperative hypoxemia is refractory to conventional treatments aimed at decreasing
pulmonary vascular resistance, especially in young infants
(5). Hyperventilation and inhaled nitric oxide have been
shown to be ineffective and do not improve oxygenation in
From the Departments of *Critical Care Medicine, †Pediatrics, §Anesthesia, and
储Surgery, Divisions of ‡Cardiology, ¶Cardiovascular Surgery, #Biostatistics, Population Health Sciences, The Hospital for Sick Children, Toronto, Canada.
Manuscript received February 8, 2004; revised manuscript received April 10, 2004,
accepted June 7, 2004.
the absence of intrapulmonary shunt (6,7). Hypercapnia
with acidosis increases cerebral blood flow and pulmonary
vascular resistance in the normal circulation, after cardiopulmonary bypass and during anesthesia (8 –10). The physiologic effects of varying CO2 and pH after the BCPA,
when cerebral venous blood returns directly to the pulmonary artery without interposition of a subpulmonary ventricle, are defined less well. Although Bradley et al. have
demonstrated that hypoventilation, without acidosis, increases systemic oxygenation, it remains incompletely understood whether hypercarbia increases cerebral blood flow
selectively or globally augments cardiac output after BCPA
(11). Therefore, we investigated the effects of a range of
CO2 tensions on systemic oxygenation and cerebral blood
flow, systemic blood flow (Qs), and Qp in the postoperative
period.
METHODS
The study protocol was approved by the research and ethics
review board at the Hospital for Sick Children. Informed
and signed consent was obtained from the parents of all
subjects. Patients scheduled for a BCPA were enrolled in
the preoperative clinic.
1502
Hoskote et al.
CO2 and BCPA
JACC Vol. 44, No. 7, 2004
October 6, 2004:1501–9
Abbreviations and Acronyms
BCPA ⫽ bidirectional superior cavopulmonary
anastomosis
FV
⫽ femoral venous/vein
IVC
⫽ inferior vena cava
JVB
⫽ jugular venous bulb
PaCO2 ⫽ arterial carbon dioxide tension
PaO2 ⫽ arterial oxygen tension
PVRI ⫽ pulmonary vascular resistance index
Qivc
⫽ inferior vena caval blood flow
Qp
⫽ pulmonary blood flow
Qs
⫽ systemic blood flow
SVC
⫽ superior vena cava
SVRI ⫽ systemic vascular resistance index
VO2
⫽ oxygen consumption
We measured sequential changes in systemic oxygenation and cerebral blood flow, Qp, and Qs at arterial
carbon dioxide tensions (PaCO2) of 35, 45, and 55 mm
Hg and then, 40 mm Hg. In the operating room, a cuffed
endotracheal tube was inserted. The intravascular monitoring lines were placed after induction of anesthesia
(superior vena cava [SVC], jugular venous bulb [JVB],
inferior vena cava [IVC], or femoral venous [FV]) and at
completion of the surgical operation (common atrial or
pulmonary venous). To cannulate the JVB, we used a
single lumen 3-F catheter inserted retrogradely into the
right internal jugular vein. Correct placement was verified
by X-ray. After arrival on the critical care unit, initial
ventilator settings were adjusted to a PaCO2 of 35 mm
Hg, guided by end-tidal CO2 obtained by mass spectrometry. Throughout the study, the transesophageal
temperature was maintained between 36°C to 37.5°C and
the fraction of inspired oxygen at 0.3. The patients
received infusions of propofol at 4 mg/kg/h, vecuronium
at 2 ␮g/kg/min, and morphine at 40 ␮g/kg/h. The
hemoglobin was maintained at or above 120 g/l. The
SVC pressure in the operating room was noted and was
maintained within 2 mm Hg by infusion of 5% albumin
(5 ml/kg) before beginning the study observations. Inotropes and vasodilators were not adjusted during the
study period. Carbon dioxide tension was adjusted by
addition of CO2 gas (initially 100 ml/min) to attain
PaCO2 of 45 mm Hg, then 55 mm Hg. Finally, additional CO2 was removed to permit the PaCO2 to return
towards baseline. We did not change ventilator parameters to avoid altering intrathoracic pressure that may
affect Qp (12).
With each increase in PaCO2 and on return towards
baseline, arterial blood gas was drawn to confirm the
PaCO2. After 15 min of stability at each level of CO2, we
recorded heart rate, mean systemic arterial, SVC, common
atrial or pulmonary venous, and low IVC or FV pressures.
Blood was drawn from the systemic artery, SVC, JVB,
pulmonary vein, and IVC or FV to measure blood gas and
O2 saturation by co-oximetry. Oxygen consumption (VO2)
was measured continuously using an Amis 2000 quadrupole
mass spectrometer (Innovision A/S, Odense, Denmark) and
adapted for use with the Servo ventilator 900C. The VO2
was measured using the mixed expiratory inert gas dilution
method (13). This requires analysis of inspired and expired
gases, together with the collection of all expired gas. Before
the study, the cuff of the endotracheal tube was inflated. A
two-point calibration of the mass spectrometer was performed. Calibration was repeated at 30-min intervals
throughout the study period.
We used the Fick equation to calculate pulmonary and
systemic flow. All patients received a postoperative chest
X-ray to exclude pleural effusions or parenchymal abnormalities. In five of nine patients, pulmonary venous oxygen
saturations were assumed according to the method of
Table 1. Clinical and Operative Details
Case
Age
(months)
Weight
(kg)
1
2
3
4
5
6
6
6.5
23
2
7
8
7.9
7.1
11.2
4.5
5.7
7.2
7
10.5
7.4
8
5
6.5
9
8
7.2
Diagnosis
HLHS
D-TGA, TA
UAVSD, HLV
HLHS
PA, IVS, HRV
Dextrocardia, DORV, D-TGA,
MA, HRV
DILV, D-TGA, CoA, LAVV
stenosis, HRV
DILV, L-TGA, AS, CoA,
HRV
DORV, PS, HRV, Bilateral
SVC
Previous
Surgery
Current Surgery
CPB
Time
(min)
Circulatory
Arrest Time
(min)
Aortic Cross
Clamp Time
(min)
Norwood I
None
PA band
Norwood I
B-T shunt
PA band
BCPA
BCPA
BCPA, DKS
BCPA, TV repair
BCPA
BCPA, DKS
102
101
90
120
57
82
0
0
24
0
0
20
0
0
58
45
17
56
Norwood I
BCPA
116
0
29
Norwood I
BCPA
56
0
14
None
Bilateral BCPA,
atrial septectomy
98
0
54
AS ⫽ aortic stenosis; BCPA ⫽ bidirectional cavopulmonary anastomosis; B-T ⫽ Blalock-Taussig shunt; CoA ⫽ coarctation of aorta; CPB ⫽ cardiopulmonary bypass; DILV
⫽ double-inlet left ventricle; DKS ⫽ Damus-Kaye-Stansel; DORV ⫽ double-outlet right ventricle; D-TGA ⫽ dextro transposition of the great arteries; HLHS ⫽ hypoplastic
left heart syndrome; HLV ⫽ hypoplastic left ventricle; HRV ⫽ hypoplastic right ventricle; IVS ⫽ intact ventricular septum; LAVV ⫽ left atrioventricular valve; L-TGA ⫽ levo
transposition of the great arteries; MA ⫽ mitral atresia; PA ⫽ pulmonary atresia; PA band ⫽ pulmonary artery band; PS ⫽ pulmonary stenosis; SVC ⫽ superior vena cava; TA
⫽ tricuspid atresia; TGA ⫽ transposition of great arteries; TV ⫽ tricuspid valve; UAVSD ⫽ unbalanced atrioventricular septal defect.
Hoskote et al.
CO2 and BCPA
JACC Vol. 44, No. 7, 2004
October 6, 2004:1501–9
1503
Table 2. Hemodynamic, Blood Gas Parameters, Oxygen Consumption, and Co-oximetry
HR (beats/min)
MAP (mm Hg)
SVCp (mm Hg)
CAp (mm Hg)
pH
PaCO2 (mm Hg)
PaO2 (mm Hg)
VO2 (ml/min/m2)
SaO2 (%)
SVC O2 Sat. (%)
JVB O2 Sat.储 (%)
IVC/FV O2 Sat. (%)
PV O2 Sat.¶ (%)
PaCO2 35
PaCO2 45
PaCO2 55
PaCO2 40
120 ⫾ 9
65 ⫾ 9
15 ⫾ 2
8⫾2
7.43 ⫾ 0.05
35 ⫾ 1.0
36 ⫾ 6
143 ⫾ 28
72 ⫾ 7
37 ⫾ 7
34 ⫾ 7
55 ⫾ 9
93 ⫾ 10
116 ⫾ 15 (p ⫽ 0.62)*
62 ⫾ 7 (p ⫽ 0.57)*
16 ⫾ 2 (p ⫽ 0.34)*
8⫾3
7.34 ⫾ 0.06 (p ⫽ 0.0003)*§
45 ⫾ 1.6 (p ⫽ 0.0003)*§
44 ⫾ 6 (p ⫽ 0.0003)*§
135 ⫾ 26 (p ⫽ 0.012)*§
77 ⫾ 5 (p ⫽ 0.029)*§
54 ⫾ 8 (p ⫽ 0.0003)*§
56 ⫾ 11 (p ⫽ 0.001)*§
66 ⫾ 5 (p ⫽ 0.002)*§
93 ⫾ 11 (p ⫽ 0.19)*
118 ⫾ 19 (p ⫽ 0.85)†
59 ⫾ 7 (p ⫽ 0.27)†
17 ⫾ 2 (p ⫽ 0.99)†
8⫾3
7.28 ⫾ 0.06 (p ⫽ 0.0003)†§
55 ⫾ 0.9 (p ⫽ 0.0003)†§
50 ⫾ 7 (p ⫽ 0.0015)†§
129 ⫾ 27 (p ⫽ 0.11)†
80 ⫾ 5 (p ⫽ 0.30)†
59 ⫾ 9 (p ⫽ 0.34)†
59 ⫾ 11 (p ⫽ 0.71)†
67 ⫾ 6 (p ⫽ 0.97)†
97 ⫾ 5 (p ⫽ 0.98)†
121 ⫾ 13 (p ⫽ 0.77)‡
59 ⫾ 8 (p ⫽ 0.98)‡
15 ⫾ 2 (p ⫽ 0.029)‡§
7⫾3
7.38 ⫾ 0.07 (p ⫽ 0.0003)‡§
40 ⫾ 4.1 (p ⫽ 0.0003)‡§
40 ⫾ 8 (p ⫽ 0.0003)‡§
141 ⫾ 27 (p ⫽ 0.02)‡§
74 ⫾ 8 (p ⫽ 0.014)‡§
42 ⫾ 9 (p ⫽ 0.0003)‡§
44 ⫾ 18 (p ⫽ 0.029)‡§
56 ⫾ 12 (p ⫽ 0.008)‡§
93 ⫾ 12 (p ⫽ 0.08)‡
Values expressed as mean ⫾ SD. *Adjusted p value for difference between PaCO2 35 and 45 mm Hg. †Adjusted p value for difference between PaCO2 45 and 55 mm Hg.
‡Adjusted p value for difference between PaCO2 55 mm Hg and back to PaCO2 40 mm Hg. §Significant p value ⱕ 0.05. 储n ⫽ 5. ¶Assumed in five patients.
CAp ⫽ common atrial pressure; CVP ⫽ central venous pressure; HR ⫽ heart rate; IVC/FV O2 Sat. ⫽ inferior vena caval/femoral venous oxygen saturation; JVB O2 Sat.
⫽ jugular venous bulb oxygen saturation; MAP ⫽ mean arterial pressure; PaO2 ⫽ arterial oxygen tension; PaCO2 ⫽ arterial carbon dioxide tension; PV O2 Sat. ⫽ pulmonary
venous oxygen saturation; SaO2 ⫽ systemic arterial oxygen saturation; SVC O2 Sat. ⫽ superior vena cava oxygen saturation; SVCp ⫽ superior vena cava pressure; VO2 ⫽ oxygen
consumption.
Tabbutt et al. (14) for patients with hypercapnia. We
calculated systemic and pulmonary vascular resistance indexed to body surface area by dividing mean pressure
change, across the respective vascular bed, by flow. Individual equations are provided in the appendix (14,15). Indirect
measurements of cerebral blood flow were made using near
infrared spectroscopy and transcranial Doppler (16 –18).
Oxygenated hemoglobin, deoxygenated hemoglobin, and
tissue oxygenation index were measured by the NIRO-300
(Hamamatsu Photonics Co., Hamamatsu City, Japan). The
flow velocity through the middle cerebral artery was measured with a 2-MHz pulse-wave ultrasound transducer,
which was fixed above the zygomatic arch with a soft rubber
holder (Medasonics Inc., Fremont, California). The transducer interrogated the portion of the middle cerebral artery
near its junction with the ipsilateral anterior cerebral artery.
We excluded from the study those patients with an
alternative source of Qp (identified at preoperative cardiac
catheterization) or decompressing venous collaterals. The
latter were excluded by postoperative saline contrast echocardiogram in all patients. Additional exclusion criteria were
postoperative bleeding (⬎5 ml/kg/h) persisting after the
first hour, fluctuations in blood pressure ⬎20%, and arrhythmias.
Statistics. The data, collected at four levels of PaCO2 (35,
45, 55 mm Hg, and then 40 mm Hg) were analyzed by
repeated measures analysis of variance for an effect over
time. A quadratic effect was indicated by a statistically
significant parameter estimate for the quadratic time sequence effect. The actual estimate for the inverted parabolas
was negative, indicating a plateau or a return to baseline for
the outcome. Pair-wise comparison was performed between
the data at different levels of CO2; the overall p value and
the adjusted p value for multiple comparisons were calculated using the Tukey-Kramer adjustment. A p value of
⬍0.05 was considered significant. We used the statistical
software SAS version 8.2 (Cary, North Carolina). The
results are expressed as mean ⫾ standard deviation or
medians with a range.
RESULTS
Between November 1, 2002, and June 1, 2003, we enrolled
11 patients and studied 9 patients. Two patients were
excluded because of postoperative bleeding that persisted
after the first hour. The median age was 7.1 months (range
2 to 23 months), and the median weight was 7.2 kg (range
4.5 to 11.2). Diagnostic and operative details are displayed
in Table 1. The median duration of cardiopulmonary bypass
Table 3. Calculated Flows, Flow Ratios, and Resistances
2
Qp (l/min/m )
Qivc (l/min/m2)
Qs (l/min/m2)
Qp/Qivc (l/min/m2)
Qp/Qs (l/min/m2)
PVRI (WU·m2)
SVRI (WU·m2)
PaCO2 35
PaCO2 45
PaCO2 55
PaCO2 40
1.4 ⫾ 0.4
1.7 ⫾ 0.6
3.1 ⫾ 0.6
0.9 ⫾ 0.4
0.5 ⫾ 0.1
6.0 ⫾ 1.9
19 ⫾ 3.1
1.8 ⫾ 0.5 (p ⫽ 0.01)*§
4.3 ⫾ 4 (p ⫽ 0.02)*§
6.0 ⫾ 4.2 (p ⫽ 0.009)*§
0.6 ⫾ 0.4 (p ⫽ 0.14)*
0.3 ⫾ 0.2 (p ⫽ 0.19)*
5.3 ⫾ 2.4 (p ⫽ 0.67)*
11 ⫾ 5.9 (p ⫽ 0.0003)*§
1.8 ⫾ 0.4 (p ⫽ 0.85)†
3.1 ⫾ 2.9 (p ⫽ 0.44)†
4.9 ⫾ 3.0 (p ⫽ 0.48)†
0.8 ⫾ 0.4 (p ⫽ 0.42)†
0.4 ⫾ 0.1 (p ⫽ 0.44)†
5 ⫾ 1.9 (p ⫽ 0.96)†
12 ⫾ 3.6 (p ⫽ 0.95)†
1.5 ⫾ 0.4 (p ⫽ 0.02)‡§
1.4 ⫾ 0.5 (p ⫽ 0.76)‡
2.7 ⫾ 0.5 (p ⫽ 0.34)‡
1.2 ⫾ 0.5 (p ⫽ 0.62)‡
0.6 ⫾ 0.2 (p ⫽ 0.11)‡
5.5 ⫾ 2.7 (p ⫽ 0.95)‡
18 ⫾ 2.2 (p ⫽ 0.006)‡§
Values expressed as mean ⫾ SD. *Adjusted p value for difference between PaCO2 35 and 45 mm Hg. †Adjusted p value for difference between PaCO2 45 and 55 mm Hg.
‡Adjusted p value for difference between PaCO2 55 mm Hg and back to PaCO2 40 mm Hg. §Significant p value ⱕ 0.05.
PaCO2 ⫽ arterial carbon dioxide tension; PVRI ⫽ pulmonary vascular resistance index; Qivc ⫽ inferior vena caval blood flow; Qp ⫽ pulmonary blood flow; Qs ⫽ systemic
blood flow; SVRI ⫽ systemic vascular resistance index; WU ⫽ Wood Units.
1504
Hoskote et al.
CO2 and BCPA
JACC Vol. 44, No. 7, 2004
October 6, 2004:1501–9
Table 4. Cerebral Blood Flow Markers
PaCO2 35
Transcranial A-V O2 difference
NIRS-TOI
TCD peak velocity (cm/s)
TCD mean velocity (cm/s)
35 ⫾ 4
38 ⫾ 10
90 ⫾ 19
40 ⫾ 7
PaCO2 45
23 ⫾ 6 (p ⫽ 0.0003)*§
48 ⫾ 8 (p ⫽ 0.0009)*§
107 ⫾ 27 (p ⫽ 0.01)*§
52 ⫾ 13 (p ⫽ 0.009)*§
PaCO2 55
PaCO2 40
21 ⫾ 6 (p ⫽ 0.68)†
51 ⫾ 9 (p ⫽ 0.54)†
114 ⫾ 23 (p ⫽ 0.42)†
55 ⫾ 11 (p ⫽ 0.83)†
32 ⫾ 4 (p ⫽ 0.0003)‡§
36 ⫾ 12 (p ⫽ 0.0003)‡§
88 ⫾ 19 (p ⫽ 0.02)‡§
45 ⫾ 22 (p ⫽ 0.34)‡
Values expressed as mean ⫾ SD. *Adjusted p value for difference between PaCO2 35 and 45 mm Hg, †Adjusted p value for difference between PaCO2 45 and 55 mm Hg,
‡Adjusted p value for difference between PaCO2 55 mm Hg and back to PaCO2 40 mm Hg, §Significant p value ⱕ 0.05.
A-V O2 ⫽ arterio-venous oxygen difference; NIRS ⫽ near infrared spectroscopy; PaCO2 ⫽ arterial carbon dioxide tension; TCD ⫽ transcranial Doppler; TOI ⫽ tissue
oxygenation index.
was 98 min (range 56 to 120 min), and aortic cross clamp
time was 29 min (range 14 to 58 min). Circulatory arrest
was used in two patients. Nine patients received an infusion
of milrinone (0.33 to 0.66 ␮g/kg/min), and two received
dopamine 5 ␮g/kg/min postoperatively. Phenoxybenzamine
(0.25 mg/kg) was administered to two patients at the start of
cardiopulmonary bypass and in one patient continued postoperatively (2 mg/kg/day).
Jugular venous bulb catheters and pulmonary venous
catheters either could not be inserted or were incorrectly
positioned in four and five of the nine patients, respectively.
The end of the FV line was situated in the IVC in five of
Figure 1. Arterial pH, arterial carbon dioxide tension (PaCO2), oxygen saturation (SaO2), and PaO2 at PaCO2 35, 45, 55, and 40 mm Hg. The lines
represent individual patient values with the bold line indicating the mean value. *Adjusted p value for difference between PaCO2 35 and 45 mm Hg.
†Adjusted p value for difference between PaCO2 45 and 55 mm Hg. ‡Adjusted p value for difference between PaCO2 55 and 40 mm Hg.
Hoskote et al.
CO2 and BCPA
JACC Vol. 44, No. 7, 2004
October 6, 2004:1501–9
1505
Figure 2. Superior vena cava (SVC) oxygen saturation and inferior vena cava (IVC)/femoral venous (FV) oxygen saturation at CO2 35, 45, 55, and 40 mm Hg.
The lines represent individual patient values, with the bold line indicating the mean value. *Adjusted p value for difference between arterial carbon dioxide tension
(PaCO2) 35 and 45 mm Hg. †Adjusted p value for difference between PaCO2 45 and 55 mm Hg. ‡Adjusted p value for difference between PaCO2 55 and 40
mm Hg.
nine patients and in the common femoral vein in the
remainder. In four patients, for technical reasons we were
unable to collect IVC saturations and transcranial Doppler
traces at a PaCO2 of 40 mm Hg. No patient experienced an
adverse event as a result of the study procedure.
The results are summarized in Tables 2 to 4 and Figures
1 to 4.
We found that arterial oxygen tension (PaO2) increased
significantly from 36 ⫾ 6 to 44 ⫾ 6 to 50 ⫾ 7 mm Hg at
PaCO2 of 35, 45, 55 mm Hg, respectively and decreased to
40 ⫾ 8 mm Hg when PaCO2 returned towards baseline at
40 mm Hg. The systemic O2 saturation increased significantly from 72 ⫾ 7% to 77 ⫾ 5% at PaCO2 of 35 compared
with 45; the increase was sustained at 80 ⫾ 5% at a PaCO2
of 55 mm Hg and decreased to 74 ⫾ 8% when PaCO2
returned towards baseline at 40 mm Hg. At a PaCO2 of 45
compared with 35 mm Hg, Qp and cardiac output (Qs)
increased significantly (1.4 ⫾ 0.4 to 1.8 ⫾ 0.5 and 3.1 ⫾ 0.6
to 6.0 ⫾ 4.2 l/min/m2); increases were sustained at a PaCO2
of 55 mm Hg [(Qp) 1.8 ⫾ 0.4 and (Qs) 4.9 ⫾ 3.0
l/min/m2] and decreased to [(Qp) 1.5 ⫾ 0.4 and (Qs) 2.7 ⫾
0.5 l/min/m2] at a PaCO2 of 40 mm Hg. Pulmonary
vascular resistance was 6.0 ⫾ 1.9 WU/m2 and remained
unchanged. At a PaCO2 of 45 compared with 35 mm Hg,
systemic vascular resistance decreased from 19 ⫾ 3.1 to 11
⫾ 5.9 Um2, remained 12 ⫾ 3.6 Um2 at PaCO2 55 mm Hg,
and increased 18 ⫾ 2.2 Um2 at a PaCO2 of 40 mm Hg. The
ratio Qp/Qs and the ratio of Qp to inferior vena caval blood
flow (Qp/Qivc) were unchanged. Cerebral blood flow increased at a PaCO2 of 45 and 55 mm Hg and decreased
when PaCO2 returned to 40 mm Hg. Between a PaCO2 of
35 and 45 mm Hg, the SVC, JVB, and IVC/FV saturations
increased significantly but did not increase further between
a PaCO2 of 45 and 55 mm Hg and decreased significantly
at a PaCO2 of 40 mm Hg. The lowest pH was 7.28 ⫾ 0.06.
The median duration of study was 2.5 h (range of 1.5 to
3.25 h).
DISCUSSION
In this study, we demonstrated that increasing PaCO2 from
35 to 55 mm Hg with respiratory acidosis improved
systemic oxygenation, Qs, cerebral blood flow, and Qp and
decreased systemic vascular resistance without increasing
pulmonary vascular resistance after a BCPA. These changes
were marked when PaCO2 increased from 35 to 45 mm Hg.
Increasing the PaCO2 from 45 to 55 mm Hg augmented
PaO2 and maintained favorable pulmonary, systemic and
cerebral blood flows without increasing pulmonary vascular
resistance. Furthermore, decreasing PaCO2 from 55 to 40
mm Hg, by withdrawing exogenous carbon dioxide, caused
the return of all parameters towards baseline. These results
may have important implications in the management of the
hypoxemic child after the BCPA, and a permissive hypercapnic strategy may improve the postoperative course, particularly in young infants. It is equally evident that a PaCO2
lower than 45 mm Hg has a deleterious effect on oxygenation and hemodynamics and may be counterproductive
after the BCPA.
The determinants of systemic oxygenation after the
BCPA are multifactorial and depend upon pulmonary and
cerebral blood flow, cardiac output, as well as intrapulmonary shunting (4,11,15,19). A unique aspect of the physiology of Qp after BCPA is the interaction of two highly
regulated vascular beds—the cerebral and the pulmonary,
which have opposite responses to changes in carbon dioxide
and acid base status (20 –22). We found that by three
1506
Hoskote et al.
CO2 and BCPA
JACC Vol. 44, No. 7, 2004
October 6, 2004:1501–9
Figure 3. Transcranial arterio-venous oxygen difference, near infrared spectroscopy (NIRS)-tissue oxygenation index, peak transcranial Doppler velocity,
and mean transcranial Doppler velocity at CO2 35, 45, 55, and 40 mm Hg. The lines represent individual patient values, with the bold line indicating the
mean value. Peak and mean transcranial Doppler measurements were not available at an arterial carbon dioxide tension (PaCO2) of 40 mm Hg in four
patients. *Adjusted p value for difference between PaCO2 35 and 45 mm Hg. †Adjusted p value for difference between PaCO2 45 and 55 mm Hg.
‡Adjusted p value for difference between PaCO2 55 and 40 mm Hg.
indirect methods of measurement (the JVB saturation, near
infrared spectroscopy, and transcranial Doppler) the cerebral blood flow significantly increased between PaCO2 of 35
and 45 mm Hg but did not increase further at a PaCO2 of
55 mm Hg. An increase in cerebral blood flow and decrease
in cerebral vascular resistance is a well-known response to
increasing arterial CO2 tension (20). However, although
cerebral blood flow and Qp increased, so too did cardiac
output. Therefore, there was no change in the ratio of
Qp/Qs or Qp/Qivc, suggesting that the effect of CO2 on
oxygenation after the BCPA was not the result of selective
cerebral vasodilation with increased flow distribution to the
brain but rather, through an increase in total Qs and
decrease in systemic vascular resistance index (SVRI). This
caused both an increase in Qp as well as increased IVC
saturations and improved systemic oxygenation. Others
have documented that modestly elevating CO2 increases
cardiac output after cardiac surgery and during anesthesia
(23,24), and at least one report suggests that despite an
increase of 64% in cardiac output selective cerebral vasodilation does not occur (25). Our study concurs with the
preliminary observations of Fogel et al. (26) that additional
CO2 after BCPA improves oxygenation. In contrast to
Fogel et al. (26), we did not demonstrate redistribution of
Qs to the brain, although it is unclear whether Fogel et al.
(26) measured descending aorta flow velocities directly.
However, Fogel et al. studied patients with BCPA remote
from the perioperative period, which may account for the
different findings (26).
There are inherent difficulties in the practical application
of the Fick principle that may have influenced our results.
We may have underestimated Qp, especially at PaCO2
tensions of 35 and 40 mm Hg, as directly measured
pulmonary venous oxygen saturations were lower than the
JACC Vol. 44, No. 7, 2004
October 6, 2004:1501–9
Hoskote et al.
CO2 and BCPA
1507
Figure 4. The pulmonary vascular resistance index (PVRI) remained unchanged at arterial carbon dioxide tension (PaCO2) 35, 45, 55, and 40 mm Hg.
There was a significant decrease in superior vena cava pressure between PaCO2 55 and 40 mm Hg but no change in pulmonary vascular resistance as
pulmonary blood flow decreased.
assumed value. In addition, the difficulties in obtaining a
true mixed oxygen saturation from the IVC are well known
(27). We attempted to overcome the vagaries of streaming
in the IVC by obtaining oxygen saturations from below the
renal veins. However, it remains possible that we have
overestimated Qivc and that the improvement in systemic
oxygenation at a higher PaCO2 may have reflected a more
favorable distribution of flow to the brain with an increased
ratio of superior vena caval blood flow to inferior vena caval
blood flow (Qsvc:Qivc) as described in the theoretical model
of Santamore et al. (19). There was a significant decrease in
the IVC, SVC, JVB, and arterial oxygen saturation on
return to 40 mm Hg, which again suggests global decrease
in cardiac output on withdrawal of CO2, though the
calculated Qs did not reach statistical significance, probably
because of a small sample size owing to missing data at a
PaCO2 of 40 mm Hg. Nevertheless, Qs demonstrated a
significant trend with a quadratic effect correlating with the
four different tensions of PaCO2 time measurements,
whereas the ratio of Qp to Qs remained unchanged. The
VO2 decreased with additional CO2 and complemented the
improvement in oxygenation and Qs. To our knowledge,
the measurement of VO2 by mass spectroscopy should not
be confounded by the addition of CO2 to the inspired gas.
Elevating arterial CO2 and particularly decreasing pH
have been shown to cause pulmonary vasoconstriction in
animals, in humans, and after surgical correction of congenital heart disease using cardiopulmonary bypass (8,21,22).
Calculated baseline pulmonary vascular resistance index
(PVRI) was increased in our study but did not increase with
the addition of CO2, suggesting that pulmonary vasoconstriction does not predicate Qp after the BCPA. This is
supported by the observation that treatment aimed at
pulmonary vasodilation such as hyperventilation with alkalosis or inhaled nitric oxide, despite an increase in cyclic
guanosine monophosphate, does not improve systemic oxygenation after the BCPA (6,7). This has potentially important implications in the management of hypoxemia after
the BCPA and suggests that measures aimed at increasing
cardiac output will influence systemic oxygenation and
improve oxygen delivery to a greater degree than pulmonary
vasodilators. Hypercapnia has been shown to increase cerebral, mesenteric, and skin blood flow and tissue oxygenation
(28). Furthermore, permissive hypercapnia in the management of adult respiratory distress syndrome has been shown
to be beneficial not only by reducing lung injury but also by
improving systemic cardiac output and oxygen delivery
(29 –32).
Systemic oxygenation after the BCPA also depends on
the degree of intrapulmonary shunt and subsequent decrease
in pulmonary venous saturation. In our study, there was a
tendency for pulmonary venous saturations to increase with
increasing CO2. A reciprocal relationship between alveolar
to arterial oxygen difference and CO2 has been described in
the anesthetized, mechanically ventilated human (33).
Elevations in CO2 have been reported to increase heart rate,
blood pressure, and cardiac contractility as a result of endogenous catecholamine release (34). However, in the current study
we did not observe changes in heart rate or blood pressure,
perhaps because the changes in CO2 were modest and contin-
1508
Hoskote et al.
CO2 and BCPA
ued anesthesia blunted catecholamine release. Thus, our findings may not be reproducible in the awake, spontaneously
breathing patient or with CO2 above 55 mm Hg. However,
they concur with previous studies in the anesthetized, mechanically ventilated human (31,32).
Troublesome hypoxemia may occur particularly in young
infants after BCPA (5,35), and thus the strategy of permissive hypercapnia may be most applicable in infants less than
six months of age.
Conclusions. We have demonstrated that after the BCPA
systemic oxygenation, Qp, Qs, and cerebral blood flow increased and SVRI decreased at CO2 tensions of 45 and
55 mm Hg compared with 35 mm Hg. We suggest that
hypoxemia after the BCPA is ameliorated by a higher PaCO2
and that low PaCO2 or alkalosis may be detrimental. Hypercarbic management strategies may allow earlier progression to
the BCPA, which may contribute to reducing the interval
morbidity in patients with a functional single ventricle.
Acknowledgment
The authors gratefully acknowledge the technical assistance
of Cameron Slorach.
Reprint requests and correspondence: Dr. Ian Adatia, UCSF
Children’s Hospital, 505 Parnassus Avenue, Room M-655, San
Francisco, California 94143-0106. E-mail: [email protected].
REFERENCES
1. Glenn WL. Superior vena cava-pulmonary artery anastomosis. Ann
Thorac Surg 1984;37:9 –11.
2. Hopkins RA, Armstrong BE, Serwer GA, Peterson RJ, Oldham HN.
Physiological rationale for a bidirectional cavopulmonary shunt a
versatile complement to the Fontan principle. J Thorac Cardiovasc
Surg 1985;90:391– 8.
3. Bridges ND, Jonas RA, Mayer JE, Flanagan MF, Keane JF, Castañeda
AR. Bidirectional cavopulmonary anastomosis as interim palliation for
high-risk Fontan candidates. Circulation 1990;82:IV-170 – 6.
4. Aeba R, Katogi T, Kashima I, Omoto T, Kawada S, Omae K. Factors
influencing arterial oxygenation after bidirectional cavopulmonary
shunt without additional sources of pulmonary blood flow. J Thorac
Cardiovasc Surg 2000;120:589 –95.
5. Bradley SM, Mosca RS, Hennein HA, Crowley DC, Kulik TJ, Bove
EL. Bidirectional superior cavopulmonary connection in young infants. Circulation 1996;94:II5–11.
6. Bradley SM, Simsic JM, Mulvihill DM. Hyperventilation impairs
oxygenation after bidirectional superior cavopulmonary connection.
Circulation 1998;98:11372– 6.
7. Adatia I, Atz AM, Wessel DL. Inhaled nitric oxide does not improve
systemic oxygenation after the bidirectional superior cavopulmonary
anastomosis. J Thorac Cardiovasc Surg 2004. In press.
8. Chang AC, Zucker HA, Hickey PR, Wessel DL. Pulmonary vascular
resistance in infants after cardiac surgery: role of carbon dioxide and
hydrogen ion. Crit Care Med 1995;23:568 –74.
9. Karsli C, Lugenbuehl I, Farrar M, Bissonnette B. Cerebrovascular
carbon dioxide reactivity in children anesthetized with propofol.
Paediatr Anaesth 2003;13:26 –31.
10. McNeill BR, Murkin JM, Farrar JK, Gelb AW. Autoregulation and
the CO2 responsiveness of cerebral blood flow after cardiopulmonary
bypass. Can J Anaesth 1990;37:313–7.
11. Bradley SM, Simsic JM, Mulvihill DM. Hypoventilation improves
oxygenation after bidirectional superior cavopulmonary connection.
J Thorac Cardiovasc Surg 2003;126:1033–9.
12. Penny DJ, Hayek Z, Redington AN. The effects of positive and
negative extrathoracic pressure on pulmonary blood flow after the total
cavopulmonary shunt procedure. Int J Cardiol 1991;30:128 –30.
JACC Vol. 44, No. 7, 2004
October 6, 2004:1501–9
13. Davies NJ, Denison DM. The measurement of metabolic gas exchange and minute volume by mass spectrometry alone. Respir Physiol
1979;36:261–7.
14. Tabbutt S, Ramamoorthy C, Montenegro LM, et al. Impact of
inspired gas mixtures on preoperative infants with hypoplastic left
heart syndrome during controlled ventilation. Circulation 2001;104
Suppl 1:I159 – 64.
15. Salim MA, Case CL, Sade RM, Watson DC, Alpert BS, DiSessa TG.
Pulmonary/systemic flow ratio in children after cavopulmonary anastomosis. J Am Coll Cardiol 1995;25:735– 8.
16. Watzman HM, Kurth CD, Montenegro LM, Rome J, Steven JM,
Nicolson SC. Arterial and venous contributions to near-infrared
cerebral oximetry. Anesthesiology 2000;93:947–53.
17. Schell RM, Cole DJ. Cerebral monitoring: jugular venous oximetry.
Anesth Analg 2000;90:559 – 66.
18. Minassian A, Melon E, Leguerinel C, Lodi CA, Bonnet F, Beydon L.
Changes in cerebral blood flow during PaCO2 variations in patients
with severe closed head injury: comparison between the Fick and
transcranial Doppler methods. J Neurosurg 1998;88:996 –1001.
19. Santamore WP, Barnea O, Riordan CJ, Ross MP, Austin EH.
Theoretical optimization of pulmonary-to-systemic flow ratio after a
bidirectional cavopulmonary anastomosis. Am J Physiol 1998;274:
H694 –700.
20. Kety SS, Schmidt CR. The effects of active and passive hyperventilation on cerebral blood flow, cerebral oxygen consumption, cardiac
output, and blood pressure of normal young men. J Clin Invest
1946;25:107–19.
21. Rudolph AM, Yuan S. Response of the pulmonary vasculature to
hypoxia and H⫹ ion concentration changes. J Clin Investig 1966;45:
399 – 411.
22. Bergofsky EH, Lehr DE, Fishman AP. The effect of changes in
hydrogen ion concentration on the pulmonary circulation. J Clin
Invest 1962;41:1492–502.
23. Prys-Roberts C, Kellman GR. Haemodynamic influence of graded
hypercapnia in anaesthetised man. Br J Anaesth 1966;38:661–5.
24. Hewitt PB, Chamberlain JH, Seed RF. The effect of carbon dioxide
on cardiac output in patients undergoing mechanical ventilation
following open heart surgery. Br J Anaesth 1973;45:1035– 41.
25. Djurberg HG, Seed RF, Evans DAP, et al. Lack of effect of CO2 on
cerebral arterial diameter in man. J Clin Anesth 1998;10:646 –51.
26. Fogel MA, Durning S, Wernovsky G, Pollock AN, Gaynor JW,
Nicolson SC. Hierarchy of feedback loops in single ventricle patients
with superior cavopulmonary connections (abstr). Circulation 2003;
108:410.
27. Rudolph AM. Cardiac Catheterization and Angiography: Congenital
Diseases of the Heart. Chicago, IL: Year Book Medical Publishers
Inc., 1974:109.
28. Akça O, Doufas AG, Morioka N, Iscoe S, Fisher J, Sessler DI.
Hypercapnia improves tissue oxygenation. Anesthesiology 2002;97:
801– 6.
29. Hickling KG, Joyce C. Permissive hypercapnia in ARDS and its
effects on tissue oxygenation. Acta Anesthesiol Scand 1995;107:
201– 8.
30. Hickling KG, Walsh J, Henderson S, Jackson R. Low mortality rate in
adult respiratory distress syndrome using low-volume, pressure-limited
ventilation with permissive hypercapnia: a prospective study. Crit Care
Med 1994;22:1564 –78.
31. Mas A, Saura P, Joseph D, et al. Effect of acute moderate changes in
PaCO2 on global hemodynamics and gastric perfusion. Crit Care Med
2000;28:360 –5.
32. Thorens JB, Jolliet P, Chevrolet JC. Effects of rapid permissive
hypercapnia on hemodynamics, gas exchange, and oxygen transport
and consumption during mechanical ventilation for the acute respiratory distress syndrome. Intensive Care Med 1996;22:182–91.
33. Kelman GR, Prys-Roberts C. Variation of alveolar-to-arterial oxygen
tension difference with changes of PaCO2 in anesthetized man.
J Physiol 1967;194:13P–4P.
34. Suutarinen T. Cardiovascular response to changes in arterial carbon
dioxide tension. Acta Physiol Scand 1966;67:9 –77.
35. Edwards WS, Bargeron LM. The superiority of the Glenn operation
for tricuspid atresia in infancy and childhood. J Thorac Cardiovasc
Surg 1968;55:60 –9.
JACC Vol. 44, No. 7, 2004
October 6, 2004:1501–9
APPENDIX
Formulas used to calculate flows and resistances
(15,19):
Qs ⫽ cardiac output ⫽ (Qp ⫹ Qivc)
Qp ⫽ {O2 consumption/[(PV O2 sat ⫺ SVC O2
sat)·Hb·1.39·10]}/BSA
(SVC sat ⫽ pulmonary arterial saturation in BCPA;
pulmonary venous oxygen saturation [PV O2 sat] assumed
0.99 in five of nine patients)
Qs ⫽ {O2 consumption · (PV O2 sat ⫺ IVC O2 sat)/
[(PV O 2 sat ⫺ SVC O 2 sat)·(SaO 2 ⫺ IVC O 2
sat) · Hb · 1.39·10]}/BSA
Hoskote et al.
CO2 and BCPA
1509
(FV O2 saturation substituted for IVC O2 saturation in
four cases)
Qivc ⫽ Qs ⫺ Qp
SVRI ⫽ (MAP ⫺ CAp)/Qs
PVRI ⫽ (SVCp ⫺ CAp)/Qp
BSA ⫽ body surface area; CAp ⫽ common atrial pressure;
Hb ⫽ hemoglobin; MAP ⫽ mean arterial pressure; O2 sat ⫽
oxygen saturation; Qivc ⫽ calculated IVC flow; Qp ⫽ calculated pulmonary blood flow; Qs ⫽ calculated systemic blood
flow; PVRI ⫽ pulmonary vascular resistance index; SaO2 ⫽
systemic arterial oxygen saturation; sat ⫽ saturation; SVC ⫽
superior vena cava; SVCp ⫽ superior vena caval pressure;
SVRI ⫽ systemic vascular resistance index.