Postoperative Management After the Norwood Procedure

Postoperative Management After the Norwood Procedure Erle H Austin III Staged reconstruction has become the preferred approach to hypoplastic left heart syndrome at many centers In the United States. The overall results of this strategy are most adversely affected by a high mortality at the initial stage, the Norwood procedure. The hemodynamic instability of a single ventricle providing blood flow in parallel to the systemic and pulmonary circulations combined with the stresses of cardiopulmonary bypass and circulatory arrest result in a precarious postoperative condition. Diligent perioperative management at this stage is essential to survival. To help simplify the complexity of single.ventricle physiology, this article describes a mathematical model that identifies the key elements that affect systemic oxygen delivery. The importance of balancing the circulation is underscored. The value of monitoring both systemic arterial and venous oxygen saturations to assess systemic-topulmonary blood flow ratio is derived from this mathematical model and confirmed experimentally and clinically. Recent research using animal models of single·ventricle physiology is also described. Using these concepts and information, techniques for achieving adequate systemic oxygen delivery are discussed. Copyright~ 1998 by WB. Saunders Company Key words: Hypoplastic left heart syndrome, Norwood procedure, single-ventricle physiology.

O

nce a uniformly fatal condition, hypoplastic left heart syndrome can now be successfully palliated with a staged reconstructive strategy. Success with this approach has been limited by a high mortality at the first stage, the Norwood procedure. l Key elements of the reconstruction include creation of an unrestricted pathway from the right ventricle to the aortic arch, the descending aorta, and the coronary circulation; placement of a systemic to pulmonary shunt; and removal of the interatrial septum. This anatomic reconstruction requires a period of cardiopulmonary bypass and deep hypothermic circulatory arrest. An ideal surgical result at this stage is only palliative because the resultant physiology is that of a single ventricle. The effects of single-ventricle physiology, carFrom the Department r! Cardiovascular Surgery. Kosair Children S Hospital, and the Unwmi~ r!Louisville, LoUlslJIlle. KY. Supported In part by a grantftom the Alhanl Communl~ Trust Fund, M~cal Towers South , SUite 154, LoUISville, KY. Address repnnl reqlusts to Erle H. .'lustln, Ill, MD. Univemry of Louisville. Dept ofSurgery, 201 Abraham FleXT/lT Wqv. #1200, LoulSvllk, KY40202. Co/!yn.f!,ht © 1998 by WB. Saunders Comfxu!.r 1092-9126/98/0101-00 14S/J8.00/0

diopulmonary bypass, circulatory arrest, and the dynamics of newborn systemic and pulmonary vascular resistance combine to create a precarious postoperative condition. As a result, a perioperative mortality in excess of 50% has not been an uncommon experience. 2•3 Over the past 10 years, several focused and enlightened teams have reported significant improvements in outcome at the initial stage.4-6 Nevertheless, surgeons and physicians find it difficult to lower the operative mortality much below 15% to 20%.7 Over the same time period, perioperative mortality of the second (bidirectional Glenn) and third (Fontan) stages has improved to more acceptable levels. 8,!} Thus, the overall results of staged reconstruction continue to be most adversely affected by the mortality that occurs at the first stage. If the hemodynamic instability of the first stage of palliation can be minimized, the overall outcome of patients with hypoplastic left heart syndrome will be substantially improved. This article describes several investigations designed to shed light on the perioperative management of the patient undergoing the Norwood procedure.

ptdzatnc Cardzac SurJ:rr:r:lnnual f!ftht Stmmars m ThoraCIC and Cardiovascular Surgery. Vol I. 1998: PJ! 109-121

109

110

f:rte JI. ..luI/iI/Ill

Single-Ventricle Physiology Central to the hemodynamic instability after the Nonvoocl procedure is the inherent instability of a single ventricle providing blood flow in parallel to the pulmonary and systemic circulations. Minor physiological or anatomic perturbations may result in major imbalances in flow between the two circulations, especially after a period of cardiopulmonary bypass and circulatory arrest. Under these circumstances, technical imperfections in the reconstructive procedure may be greatly magnified. A Mathematical Model

To improve our understanding of the dlccts of altcrations in !low distribution between the systemic and pulmonary circulations, Barnea et al crcated a mathematical model of the circulation ill hypoplastic lelt heart syndrome III (Fig I). Our analysis was based on the obvious premise that the pllrpose of the heart and lungs is to provide oxygen to the tissut's. Recognizing that the uptake of oxygen by th e pulmon,lIY circulation should equal consumption of oxygen by th e hody, we established a set of equations Crable I) fi"om which we could derive the determinants of

-(Qs

eVa

-(

CaO:z . Q, = 1 + Qp/Qs CO . C PVO:.? 1 . - Qp/Qs CVO:.? This equation indicates that systemic oxygen delivelY (CaO:.?· Qs) is a complex function of cardiac output (CO), pulmonary venous blood oxygen content (C pvo:.?), whole-body oxygen consumption (CV0 2), and Qp/Qs ratio. This complex relationship can be studied by alternating each variable individually, The effect of alterations in Qp/Qs on systemic oxygen availability and percent oxygen saturation for several differe nt cardiac outputs are depicted in Figs 2 and 3 respectively, Figure 2 indicates that systemic oxygen availability (or oxyge n delivery) peaks at a Qp/Qs less than or equal to 1. For high cardiac outputs, the optimal Qp/Qs is less than 1 and the critical range of Qp/Qs in which oxygen delivery exceeds minimal oxygen consumption is broad. A,> cardiac output becomes limited, the optimal Qp/Qs approaches 1 and the critical range of Qp/Qs narrows significantly. The clinical implication of this graph is the importance of

Ductus Anenosus

Ca02

Atrl!lIc Valve

systemic oxygen delivery m a univentricular heart:

~

~ teo

2

Foramen Ovall:

CpV02

Figure l. SdH'lll;lIi .. llsl'd to d .... in· lIlatlu' lIlati ..al llUKkl or till' circulalioll ill hypoplasti(' left heart svndroml' . CaO!, ox ygC '1I .... 111 .. 111 C,r I hI' nlivd hlcMKI .. jcTI .. d fr"l11 lilt" right n'lllrid .. (RV); ( :(), righ} \"("lIlricular outPllt ; ' CI'\"(~! (CSHIJ, ClX\".t: C·1I .... 'It c'lIl (Ill!. oxygc 'n/ lIl!' blo(KI ) of tllC" plIllIlClllary (,,'slc'mid \'l"nOliS blood; C:V()~, whok-blKly oxygen consumption; LV, 1.. 11 \("111 ririe; I' (S), plIllluJilary (syst .. mic) rirrulatiClII; Q. (Us), pulmollary (systcmic) flow ; SV()~ , ratc ofoxygcll supplY or ''ill ;lkc' ill tile" illlIgs, (Rl"prilll c·d "il h p.... lIlissioll from I hc ' :\mlTi .. an Collc'gc' of C;lrdiolol.,'Y IJ :\111 (;,,11 Cardiul 11 : 117Ii-i:IHI,II)
111

Mana,l!,emenl After Nontwd Procfdure

Table 1. Mathematical Model ofUniventricular Circulation: Equations (I) Cao2' Qs - CVo 2 = CSV02' Qs (2) CaOl' Qp

+ SVo 2 =

(3) CO = Qs

+ Qp

CPV02 . Qp

(4) SV02 = CV0 2 (5) Cach' Qs = I

I

+ Qp/Qs

I , CO· CPV(h - - - CVoz Qp/Qs

Abbreviations: Cao), oxygen content of the mixed blo
trying to balance systemic and pulmonary blood flow (Qp/QS = I), especially when cardiac output is marginal. A secondary but valuable finding shown in this mathematical model is the effect of varying Qp/Qs on systemic arterial and systemic venous oxygen saturations (Fig 3). As expected, systemic arterial oxygen saturation increases with increasing Qp/Qs, but as percent oxygen satura-

tion increases, relatively small increases in systemic arterial oxygen saturation represent increasingly large increases in Qp/Qs. As such, when systemic arterial oxygen saturation exceeds 80%, a significant increase in Qp/Qs may go unrecognized. Mixed vellous systemic oxygen saturation shows a very different relationship with Qp/Qs. Interestingly, for any given cardiac output (or pulmonary venous oxygen saturation), systemic mixed venous oxygen saturation peaks when Qp/Qs = I. This finding indicates the potential value of monitoring both systemic arterial and venous oxygen saturation as a way to determine relative pulmonary and systemic flows. A high systemic arterial oxygen saturation combined with a low systemic venous oxygen saturation would indicate that pulmonary blood flow is excessive, whereas high values for both would indicate a satisfactory balance. Furthermore, because systemic mixed venous oxygen saturation peaks at a Qp/Qs or I, any increase in systemic mixed venous oxygen saturation can be interpreted as an improvement in cardiac output or blood flow balance. Any decrease must be interpreted as an adverse change. Such f'lTd-

S,.o,-96"It Oz ~OII- .1 mlOzlmin

Figure 2. Systemic oxygen availability (deli\'ery) as a fimction of pulmonary/systemic flow ratio (QI'/Q~) for different values of cardiac output (CO). SI'VOb pl'rcent oxygen saturation of pulmonary H'nous blood. (Reprinted with permission hum the American College of Cardiology [.J Am Coli Cardiol U: 13 76-13H I. 1994],)

2

s

10

112

Erte H. AlLI'tin III

80

-e

SpvOz-96% 0 1 COIIlllmpIIOII- 18 mlO{min

.§ ::s

60

~

en

5

~ 40

-~

cS

cf

20

0.2

0.5

2

back may be useful in recognizing a problem early as well as determining the effectiveness of any intervention applied to correct the problem. To test the validity of our mathematical model, we created a parallel circulation in a piglet and varied Qp/Qs from 0.1 to 6.5. 11 As predicted, the systemic venous oxygen saturation was optimal at a Qp/Qs of 1 (Fig 4).

5

10

Figure 3. Percent oxygen saturation of systemic arterial (Sa~) and venous (Ssv~) blood as a function of the pulmonary/systemic flow ratio (QplQs) for different values of cardiac output (CO). Systemic arterial oxygen saturation continually increases as QplQs increases, whereas venous oxygen saturation peaks at a QplQs of 1. The combination of a high So:! and a low Ssv~, therefore, is an indication of a high QplQs. (Reprinted with permission from the American College of Cardiology [J Am COU Cardio124: 1376-1381, 1994].)

Clinical Monitoring of Systemic Venous Oxygen Saturation Shortly before our mathematical model was developed and published, the value of measuring mixed venous oxygen saturation after the Norwood procedure was demonstrated by Rossi et al l2 at Mt. Sinai Medical Center in New York.

Qp/Qs vs Systemic Venous Oxygen Saturation 100~----------------------------1

•

• Figure 4. Systemic venous oxygen saturations and pulmonary/systemic now ratios (QplQ<;) in a piglet model of parallel circulations. As predicted by the mathematical model, venous oxygen saturation pt'aks at a Q.,fQ~ of I. (Reprinted with permission." )

O~--~---r---r--~--~---,--~~~

o

1

2

6

7

8

Management ~ Norwood Procedure

In their study, blood was sampled intermittently from the superior vena cava in 13 consecutive infants undergoing stage 1 palliation for hypoplastic left heart syndrome. Abnormalities in Qp/Qs were noted in 12 patients. A high Qp/Qs was noted even in patients with acceptable arterial oxygen saturations and was associated with early mortality. Knowledge of mixed venous oxygen saturation in addition to systemic arterial oxygen saturation permits bedside calculation of Qp/Qs based on the following formula: systemic arterial O 2 saturation - mixed venous O 2 saturation Qp/Qs = pulmonary venous O 2 saturation - systemic arterial O 2 saturation For the purposes of this estimation, a pulmonary venous saturation of 96% is assumed. 12 With this information, Qp/Qs can be monitored to allow early and appropriate adjustments in ventilator or pharmacological management. Although total body mixed venous oxygen saturation cannot be measured in infants with hypoplastic left heart syndrome, sampling from the midinferior vena cava or high superior vena cava provides a reasonable correlation. Intermittent sampling may provide adequate feedback, 12 but continuous monitoring of mixed venous oxygen saturation can be obtained with an appropriately placed 4 Fr oximetric catheter (Abbott Biomedical, North Chicago, IL). The catheter may be inserted preoperatively via the umbilical vein or intraoperatively through the right atrial wall. Preoperative stabilization of hemodynamically unstable infants may, in fact, be facilitated by inserting the catheter soon after initial presentation at the time when an umbilical arterial catheter is placed. 13 The vagaries of catheter positioning, of course, must be recognized.

Animal Models Direct measurements of cardiac output and blood flow distribution in newborn infants with single-ventricle physiology have been limited by the lack of safe, reliable, and easily applied measuring devices. As such, most management techniques have been derived from trial and

113

error and have been subject to personal and institutional biases. An appropriate animal model of single-ventricle physiology permits more vigorous analysis of the factors that afTect the balance of flows and delivery of oxygen in this abnormal circulation. Within the past 3 years, three independent groups have described techniques of creating single-ventricle physiology in neonatal animals. Mora et al H created a univentricular circulation in piglets using cardiopulmonary bypass and deep hypothermic circulatory arrest. In this model, the interatrial septum is removed. The orifice of the tricuspid valve is closed with a patch, and a systemic-to-pulmonary arterial shunt is inserted (Fig 5). As a result, the left ventricle becomes the "single ventricle." Appropriately placed flow probes and pressure transducers allow direct measurements. In an effort to avoid the costs and morbidity of cardiopulmonary bypass, our group15 described a closed technique in neonatal piglets for creating a similar parallel circulation based on the left ventricle. A Rashkind septostomy catheter is introduced from the femoral vein to perform an atrial septostomy as well as render the tricuspid valve incompetent. A 6-mm expanded polytetraflouroethylene (PTFE) graft is then anastomosed between the innominate artery and the main pulmonary artery. Once the right ventricular outflow tract is occluded with a tourniquet, a univentricular circuit results (Fig 6). Reddy et al 16 described an elegant technique for creating single-ventricle physiology in a fetal lamb. At approximately 140 days of gestation, a 100mm PTFE graft is anastomosed between the aorta and pulmonary artery. The main pulmonary artery is then occluded distal to the graft with a large vascular clip. A 5-mm PTFE graft is then placed between the ascending aorta and the left pulmonary artery (Fig 7). In this model, both the right and left ventricle contribute to the cardiac output with mixing occurring at the aortic level. The fetus is allowed to complete gestation and spontaneous delivery. At 2 to 3 days of life, the lamb is anesthetized, instrumented, and subjected to cardiopulmonary by-

114

Erie H Austin III

Figure 5. The piglet model of single ventricle physiology described by Mora et aI. The interatrial septum was removed, the tricuspid annulus was closed with a patch, and a 4-mm P1FE tube was interposed between the innominate artery and the main pulmonary artery. (Reprinted with permission. 14 )

pass and deep hypothermic circulatory arrest. The effects of a variety of manipulations before and after bypass can then be measured. Each model has its own advantages and disadvantages. In none of the models is a single right ventricle created. In two of the models, the left ventricle becomes the single ventricle with mixing occurring at the atrial level before ejection, whereas in the fetal model, both ventricles contribute to the cardiac output with mixing occurring after ejection at the level of the great vessels. The open-heart model of Mora et al 14 requires the neonatal animal to survive a period of the cardiopulmonary bypass and circulatory arrest, whereas the closed heart model avoids these additional insults. The use of cardiopulmonary bypass in the open model, however, more closely reproduces the clinical situation after the Norwood procedure. It also assures a nonrestrictive atrial communication that may be less likely with the closed technique. It is currently not known whether closure of the tricuspid valve or

rendering it incompetent importantly affects the physiology of the model. Both of the above models have been created in animals that, although very young, have lived for several days or weeks with a normal postnatal circulation. A model created in utero, such as that described by Reddy et al,16 results in an animal with parallel pulmonary and systemic circulations that experiences the transitional phase of circulation that occurs with birth. Thus, at least in terms of maturational and transitional changes, this fetal model may more accurately resemble clinical conditions. The fetal model does, of course, represent a technical tour de force and may be difficult to reproduce without first acquiring significant experience in fetal surgery. Despite their differences and respective shortcomings, all three animal models have provided insight into single-ventricle physiology. Using their open model, Mora et aP4 measured the effect of adding CO2 to the inspired gas on pulmonary vascular resistance. The addition of

Management 4fter Norwood Procedure

115

Figure 6. The piglet model of Randsbaek et al. Atrial sept ostomy and tricuspid valvotomy were performed with a Rashkind balloon catheter. A PTFE graft was placed from the innominate artery to the pulmonary artery, and the proximal main pulmonary artery was occluded with a tourniquet. (Reprinted with permission. 1.1)

-DAo LPA

Figure 7. The fetal lamb model of Reddy et al. A DamusKaye-Stansel anastomosis was created with a IO-mm FfFE graft between the aorta and pulmonary artery with placement of a large vascular clip on the main pUlmonary artery distal to the Damus anastomosis. A 5-mm PTFE graft was then placed between the ascending aorta and left pulmonary artery. A, aortopulmonary shunt; Ao, ascending aorta; BT, brachiocephalic trunk; C, clip closure of main pulmonary artery; D, Damus-Kaye-Stansel anastomosis; DA, ductus arteriosus; DAo, descending aorta; LPA,Ieft pulmonary artery; PA, main pulmonary. (Reprinted with permission. lb )

increasing levels of CO2 to the inspired gases resulted in corresponding increases in pulmonary vascular resistance (Fig 8). Infusion of a buffer (tromethamine) to offset the decrease in pH caused by the added C02 diminished but did not eliminate this effect of C02 on the pulmonary vascular bed (Fig 8). Using our single-ventricle model, our group examined the effects of F102, positive endexpiratory pressure, and C02 on QplQs, vascular resistance, and oxygen deliveryl7 (Fig 9). Confirming what had been learned empirically from clinical experience, we showed that decreasing levels of F102 resulted in increasing levels of pulmonary vascular resistance and decreasing ratios of Qp/Qs. Supplemental CO2 was also noted to increase pulmonary vascular resistance and decrease the Qp/Qs ratio. In addition, positive end-expiratory pressure was noted to increase pulmonary vascular resistance independent of inspired oxygen tension. We also confirmed in this study the value of combining arterial and venous oxygen saturations to determine the Qp/Qs ratio and oxygen delivery.

116

in

•

E

.• U

Erie H. Austin III

8000

7000

III

8000 •

C

•

-"• E u u

U

•

....on

PVR/PIC02 No THAW

-.•, c >-

-"

11000 4000 •

".....

> ~

3000~

> Do

>-

IE:

•

CII

III

0:

•. •>

2000

CII

1'1

1000

~

II

>

c(

o·

0

c(

14

21

21

35

PIC02 (mm HII)

PVR I PIC02 THAM group 8000 7000 6000

I

6000 4000

I

3000

2000 1000 0

0

.

7

U

21

28

PI C02 (mm Hg)

Figure 8. Effect of addition of carbon dioxide on pulmonary vascular resistance (PVR) in a piglet model of single ventricle.

Left panel shows an increase in pulmonary vascular resistance with increases in partial pressure of inspired carbon dioxide (PI~). Right panel shows an increase in pulmonary vascular resistance with increases of PI~ when pH is maintained between 7.3 and 7.5 with infusion oftromethamine buffer (Tham). (Reprinted with permission. I.)

Using their elegant fetal model, Reddy et aI I6 measured the effects of 100% oxygen, 10% oxygen (hypoxia), 5% CO2, and nitric oxide (80 ppm of air) before and after a period of cardiopulmonary bypass and deep hypothermic circulatory arrest (Fig 10). Confirming clinical experience and information derived from the other animal models, 100% oxygen significantly increased Qp/Qs, whereas hypoxia significantly decreased it. The addition of CO2 also significantly decreased Qp/Qs. Interestingly, in this model, nitric oxide resulted in only a modest increase in Qp/Qs, no greater than what was observed ""ith I 00% oxygen. Surprisingly, responses to these manipulations were essentially the same before and after cardiopulmonary bypass with no statistical differences observed between the preb}pass and postbypass states. Our group has also used the closed animal model to examine the effect of inotropic agents on single-ventricle physiology. IS Because different agents may vary in their effects on Qp/Qs, we hypothesized that they may not be equally effective at increasing oxygen delivery. In fact, in our comparison between dopamine, dobutamine, and epinephrine, we discovered that all three agents increased cardiac output of the single ventricle, but only epinephrine significantly increased oxygen delivery. This occurred because of epinephrine's tendency to decrease QplQs, compared with dobutamine's tendency

to increase it (Fig 11). Increasing doses of dopamine had no appreciable effect on QplQs. This animal study lends further support to the clinically applicable concept of following up systemic venous oxygen saturations to determine the effectiveness of therapeutic interventions. Hopefully, additional laboratory investigations using animal models such as these will further increase our understanding of neonatal singleventricle physiology and will improve clinical outcomes.

Current Techniques in Postoperative Management After the Norwood Procedure The object of first-stage palliation of hypoplastic left heart syndrome is to establish a parallel circulation with adequate systemic oxygen delivery to sustain tissue viability without metabolic acidosis. During the early clinical experience with the Norwood procedure, the importance of respiratory management on Qp/Qs was discovered. 19,20 High levels of inspired oxygen were noted to decrease pulmonary vascular resistance, increase systemic vascular resistance, and tip the balance of blood flow away from the body and toward the lungs. Hyperventilation and its resultant low PeD:! had a similar effect. As noted above, these effects of oxygen and carbon dioxide have been confirmed in animal models of

117

,\;fana,t;emenl Ajler NomJOOd Prort,dure

Qp/Qs vs Inspired Oxygen Tension 1.4

12

j

1.0

"

III 0.8

~

06

CJ

0.4 0.2 -I

0.0

75..

15..

21"

~

Inepiftd OxygOl!n TOI!n.ion

Resistance vs Inspired Oxygen Tension

o

100'II

21 ..

~

15'110

-. fl.jPVRl

·~~I

Inspired Oxygen Tension

Oxygen Delivery V8. Inspired Oxygen Tension 120 -

. ~c ~

Q ·S f} CIlI_

g

100 J 80

C

00'

«>1 I

20·

I

o .. 100'II

75'11.

~

21'11.

15'11.

Inspired Oxygen Tension

Figure 9. Effects of inspired oxygen tensions on QI'/Q~ ratio, vascular resistances, and oxy~en dcliVt'ry in a pi~kt modcl of single ventricle. Top panel, Q';Q~ decreased as inspired oxy~en tension dC(,I'<'ased. r..liddk panel, pulmonary vascular resistance (PVR) increased as inspired-oxygen tension de('reased, whereas no significant ('han~e in systemic vascular resistance (SVR) was noted. Lower panel, maximal oxygen delivery in this model OlTlIITnl at an inspired oxygen tension of 50%. (Reprinted with permission. II)

single-ventricle physiology. Understanding the effects of respiratory manipulations on the resistances of the two parallel circulations is fUJ1(lamental to the postoperative management of the Norwood patient. By adjusting the resistances to achieve a balanced distribution of blood flow (Qp/Qs = 1), satisfactory O:z delivery is achieved. The commonly cited indicator of this state is an "ideal" set of arterial blood gases (pH 7.+0, PO:z

+0 mm Hg, Peo:! +0 mOl Hg). An arterial oxygen saturation of' 7Y}i) to HO% and a mixed venous oxygen saturation 01'+5% to 60% are also indicative of a hal anced circulation with adequate oxygen delivery. These inCants typically appear well perfused with good urine output. Although a well-balanced circulation at the conclusion of the operative procedure is likely to stabilize further \..-ith time, it is important to monitor

118

Erie H Austin III

I P-.12 \ 0

;:: IV

0:: ~ 0

ii:

.~ GI

iii

>-

60

40

~

10

c: 0 E

"3

~ GI

C»

P·.01 P-.42

30 20

IV

21.7%

50

en 0

36.6% P-.001

7.1%

P..014

I

4.2'1. P-.05

I P"'.40 \

I P-.67 \

.32.2'1. .27.4'1. P<.0001 P-.0003

·26.9'1. ·26.20/. P-.001 P-.OOO9

Hypoxia

5'10 Carbon

0 ·10

c:

.c "' ·20

-.

(J

c: ·30GI

u

GI

~

-40

Nitric oxide

100% Oxygen

dioxide

Figure 10. Percent change in Q/Qs after administration of nitric oxide, 100% oxygen. 10% oxygen (hypoxia), and 5% carbon dioxide before (shaded bars) and after (open bars) cardiopulmonary bypass in a fetal lamb model of single ventricle. (Reprinted with permission. 16)

arterial blood gases and arterial and venous 02 saturations frequently in the initial 24 to 48 hours to allow rational manipulations in F102 and ventilation as patients are allowed to awaken from sedation and wean from ventilatory support. Because each infant's early response to the procedure may vary, frequent measurements allow the physician to adjust FI02 and minute ventilation to achieve optimal arterial blood gases and O 2 saturations. Inadequate Systemic Oxygen Delivery

The most disturbing finding after the completion of the Norwood procedure is the development of metabolic acidosis (base deficit >6). This is often associated with the clinical finding of diminished peripheral pulses, poor capillary refill, and oliguria. These findings, of course, are indicators of inadequate systemic oxygen delivery. Mter the Norwood procedure, the three most common causes of inadequate oxygen delivery are (I) excessive pulmonary blood flow, (2) inadequate pulmonary blood flow, and (3)

inadequate cardiac output. A technical problem or residual anatomic defect may be the basis for any of these conditions and must be ruled out or immediately corrected. The metabolic acidosis should be corrected by administering sodium bicarbonate, but the cause of the acidosis, whether it is technical or physiological, should be determined and rectified. Excessive Pulmonary Blood Flow

When pulmonary blood flow is excessive, the p~ generally exceeds 45 mm Hg and the systemic arterial oxygen saturation exceeds 85%. The mixed venous saturation is less than 40%, and the calculated Qp/Qs is greater than 2: 1. To tip the balance back to the systemic circulation, the inspired oxygen concentration should be decreased to room air (FIo 2 = 0.21), and the minute ventilation decreased to allow the Pc~ to increase to between 40 and 45 mm Hg. The decrease in minute ventilation is best achieved by diminishing the ventilator rate while maintaining relatively high tidal volumes (20 to 25

Management ¥r Norwood Procedure

119

Epinephrine vs Qp/Qs

2

Fi02 100 %

o 1.6

Fi02 50 %

1:1

~ 12

~0.8 0'04

o Base

0.05

0.1

Epinfphrioe (pg/mio)

Base

0.1

Dobutamine vs Qp/Qs

Fi02 100 %

3 ~

2.S

~

2

•

Fi02 50%

•

oU C5. 1 0' 0.5

o Base

5

15 Base Dobutamine (lI&Ik&/min)

Dopamine vs Qp/Qs

z

Fi02 100

15

Fi02 SO %

~ 1.6

~ 1.2

Figure 11. Effects of inotropes on Q/Qs ratio

in a piglet model of single ventricle. Epinephrine decreased Q.IQs, whereas dobutamine increased it. No significant change occurred with dopamine. (Reprinted with permission from the Society of Thoracic Surgeons [Ann Thorac Surg 62:83-90, 1996].)

~

0.•

0' 0."

o

mUkg) to prevent atelectasis. The addition of positive end-expiratory pressure (5 to 15 mm Hg) may also help diminish pulmonary blood flow. If Qp/Qs remains greater than 2: I after these relatively simple ventilator changes, the pulmonary vascular resistance can be increased further by decreasing the inspired oxygen concentration to less than 0.21. This is achieved by adding nitrogen to the inflow of the ventilator. An in-line oximeter should continuously monitor the F102 • F102 values as low as 0.15 may be used. Carbon dioxide may also be added to the ventilator circuit to increase pulmonary vascular resistance. 21 The flow of CO2 into the ventilator should be adjusted to achieve a F1C02 of 2% to

But

5

15

Ba.

15

Dopamine (lia/k1fmin)

3% (14 to 21 mm Hg) as measured by an infrared capnometer on the patient limb of the circuit. This commonly results in an increase in arterial PC02 to a level of 50 to 60 mm Hg, which is usually well tolerated if the Qp/Qs has been successfully lowered. Whether N2 or CO2 is used in this situation is a matter of physician and institutional preference. It is not known whether their efTects are additive. Whichever technique is used, the efTectiveness of the intervention is assessed by monitoring arterial and venous oxygen saturations and calculating the Qp/Qs. Inadequate Pulmonary Blood Flow When pulmonary blood flow is inadequate, arterial P02 is less than 25 mm Hg and arterial

120

Erie H. Austin III

oxygen saturation is less than 60%. Mixed venous oxygen saturation is less than 40%, and estimated Qp/Qs is less than 0.5. In this situation, the inspired oxygen concentration is increased to an FI02 of 1.0, and the minute ventilation is increased to lower the PC02 to below 30 mm Hg. If systemic arterial blood pressure is relatively high (>65 mm Hg systolic), and the Po.z remains less than 25 mm Hg, the shunt should be carefully evaluated for any anatomic constriction. If none is found, a larger shunt should be placed. When the P02is 25 to 30 mm Hg, pulmonary blood flow and systemic 02 delivery may be improved by an infusion of phenylephrine (1 to 5 mg/kglmin). In these patients, the P02 generally improves within 24 hours and the phenylephrine can be weaned off. Inadequate Cardiac Output

When the estimated QplQs is close to 1 but systemic venous oxygenation saturation is less than 40%, it is likely that cardiac output is inadequate. Because the systemic venous oxygen saturation is low and the volume of pulmonary venous blood return is limited by the total output of the ventricle, the systemic arterial saturation tends to be low in this condition. The systemic arterial oxygen saturation is usually less than 70%, with a Po.z less than 35 mm Hg. If only systemic arterial oxygen saturation is being monitored, an inadequate cardiac output may be misinterpreted as inadequate pulmonary blood flow. When cardiac output is the problem, however, respiratory maneuvers to improve pulmonary blood flow (increase FI02 , decrease PC02) only make the condition worse. The level of systemic arterial blood pressure generally indicates which condition exists. Blood pressure is typically high (systolic blood pressure >65 mm Hg) in infants with inadequate pulmonary blood flow and low or marginal (systolic blood pressure <60 mm Hg) in infants with inadequate cardiac output. In addition, the degree of metabolic acidosis is generally greater (base deficit >8) in infants with inadequate cardiac output. Low cardiac output is the most difficult of the three conditions to manage and is more likely related to a technical problem with the proce-

dure. Poor myocardial protection andlor compromised coronary blood flow are likely explanations for this situation. It is imperative that a thorough evaluation of the operative repair be performed with echocardiography. Proper function of the tricuspid and pulmonary valves as well as an absence of any obstruction throughout the aortic reconstruction must be assured. If there is any evidence of aortic obstruction or limitation to coronary blood flow, cardiopulmonary bypass must be reinstituted and the anatomic problem corrected immediately before further myocardial injury results. Assuming no anatomic defects are identified, the Qp/Qs ratio should be closely monitored and kept as close to 1 as possible using the respiratory techniques described above. Standard maneuvers to increase the cardiac output of a failing heart are also applied. These include assurance of an adequate filling pressure (central venous pressure = 10 to 15 mm Hg), the addition of inotropic agents, and the use of vasodilators. Based on information derived from our animal model, 18 we prefer epinephrine (0.05 to 0.2 mg/kglmin) as the inotropic agent for these patients, although dopamine (5 to 10 mg,lkg/min) may be equally effective. Empirically, nitroprusside (0.1 to 5.0 mg/kg/min) appears to reduce afterload and improve tissue perfusion in this situation. Again, the effectiveness of each intervention can be assessed by monitoring both arterial and venous oxygen saturations.

Summary Although the mortality of first-stage reconstruction for hypoplastic left heart syndrome remains higher than that for most neonatal cardiac procedures, improvements in surgical technique and perioperative management have justified its application to the majority of infants born in the United States with this otheIWise lethal malformation. Building on experience and empirical management strategies, several centers have been able to show notable improvement in early outcome with these infants.6,8 Smaller centers are also reporting improved success with this challenging condition. 12.22 This improve-

Management Afler Norwood Procedure

ment in results is in large part secondary to increased understanding of the uniqueness of single-ventricle physiology. A mathematical model of single-ventricle physiology has helped identify the role of Qp/Qs in the delivery of oxygen to the tissues. Furthermore, this model has shown the value of combined monitoring of systemic arterial and venous oxygen saturations to provide meaningful feedback regarding the state of the univentricular circulation. Recent animal models of neonatal single-ventricle physiology have allowed more rigorous measurements, confirming clinically determined effects of respiratory management on systemic-topulmonary blood flow ratio. These models hopefully will permit further insight into the complexity of this abnormal circulation. Armed with an improved understanding of the unique physiology of the single ventricle, a knowledge of the factors that affect pulmonary and systemic vascular resistance, and a systematic form of feedback, clinicians now have an improved appreciation of perioperative events and are better prepared to initiate rational therapeutic responses. As more information and experience is derived from laboratory and clinical experience, outcomes will continue to improve.

References I. Norwood WI, Kirklin JK, Sanders SP: Hypoplastic left heart syndrome: Experience with palliative surgery. Am J Cardiol 45:87-91,1980 2. Sade RM, Crawford FAJr, Fyfe DA: Symposium on hypoplastic left heart syndrome.J Thorac Cardiovasc Surg 91:937-939, 1986 3. Gutgesell HP, l\1assaro TA: Management of hypoplastic left heart syndrome in a consortium of university hospitals. AmJ Cardiol 76:809-811, 1995 4. lannettoni MD, Bove EL, Mosca RS, et al: Improving results with first-stage palliation for hypoplastic left heart syndrome.J Thorac Cardiovasc Surg 107:934-940, 1994 5. Forbess ]M, Cook N, Roth S], et al: Ten-year institutional experience with palliative surgery for hypoplastic left heart syndrome: Risk factors related to stage I mortality. Circulation 92:262-266, 1995 (suppI2) 6. KernjH, Hayes C], MIchler RD, et al: SulVival and risk factor analysis for the Norwood procedure for hypoplastic left heart syndrome. Am] Cardiol80: 170-174, 1997

121

7. Mosca RS, Bove EL, Crowley DC, et al: Hemodynamic characteristics of neonates following first stage palliation for hypoplastic left heart syndrome. Circulation 92:267-271. 1995 (suppI2) 8. Bove EL, Lloyd TR: Staged remnstruction for hypoplastic left heart syndrome. Ann Surg 224:387-395. 1996 9. Forbess]M, Cook N, Serraf A, d al: An institutional experience with second- and third-stage palliative procedurrs for hypoplastic left heart syndrome: The impact of the bidIrectional cavopulmonary shunt..J Am Coil Cardiol 29:665-670, 1997 10. Barnea 0, Austin Ell III, Richman B. et al: Balancing the cirrulation: Theoretic optimization of pulmonary/systemicnow ratio in hypoplastie left heart syndromr. J Am Coli CardioI24:1376-1381,1994 II. Riordan CJ, Randsbaek F, Storey.JII, et al: Balandng pulmonary and systemic arterial flows in parallel rirculations: The value of monitoring system venous oxygen saturations. Cardiol Young 7:74-79, 1997 12. Rossi AF, Sommer RJ, Lotvin A, et al: Usefulness of intermittent monitoring of mixed venous oxygen sat uration after stagr I palliation for hypoplastic left hrart syndrome. Am] Cardinl 73:1118-1123,1994 13. Riordan ('J, Locher.IP Jr, Santamore WP, et al: Monitoring systemic venous oxygen saturations in infants with the hypoplastic left heart syndrome. Ann Thorac Surg 63:835-83 7, 1997 14. Mora GA, Pizarro C,.Jacobs to.fL, et al: Experimental model uf single ventricle. Influence of carbon dioxide on pulmonary vascular dynamics. Circulation 90:43-46, 1994 (suppl 2) IS. Randsbaek F, Riordan C]. Storey JII. et al: Animal model of the univentricular heart and single ventricular physiology. J Invest Surg 9:375-384, 1996 16. Reddy VM, Liddicoat JR, Fineman ]R, et al: Fetal moclel of single ventricle physiology: Hemodynamic effects of oxygen, nitric oxide, carbon dioxide, and hypoxia in the t'arly postnatal period.JThoracCardiovascSurg 112:437-449, 1996 17. Riordan CJ, Randsbaek F, StoreyJH. et al: Effects of oxygen, positive end-expiratory pressure, and carbon dioxide on oxygen delivery in an animal model of the univentricular hrart.] Thorac Cardiovasc Surg 112:644-654, 1996 18. Riordan Cj, Randsbaek F, Storey JII, et al: Inotropes in the hypoplastic left heart syndrome: Effects in an animal model. Ann Thorac Surg 62:83-90, 1996 19. Jonas RA, Lang P, Hansen D, et al: First-stage palliation of hypoplastic left heart s}ndrome. The importance of marclation and shunt size.] Thorac Cardiovasc Surg 92:6-13, 1986 20. Murdison KA, BafTa]~f, Farrell PEJr, et al: HypoplastIc left heart syndrome. Outcome after initial remnstruction and before modified Fontan procedure. Circulation 82:199-207. 1990 (suppI4) 21. Jobes DR, Nicolson SC, Steven JM, et al: Carbon diOXIde prevents pulmonary overcirculation in hypoplastic left heart syndrome. Ann Thorac Surg 54: 150-151, 1992 22. Starnes VA, Griffin r-rr., Pitlick PT, et al: Current approach to hypoplastic left heart syndrome. Palliatiun, transplantation, or both? J Thorac Cardiova~c Surg 104: 189-195. 1992