Arterial switch operation for transposition of the great arteries, with special reference to left ventricular function

J THORAC

CARDIOVASC SURG

1989;98:601-10

Arterial switch operation for transposition of the great arteries, with special reference to left ventricular function Between June 1984 and September 1987, 48 patients underwent Lecompte's modification of the arterial switch operation for transposition of the great arteries, including transposition with intact ventricular septum with preparatory pulmonary artery banding (n = 18~ with patent dUCM arteriosus (n = 11~ with dynamic left ventricular outflow tract obstruction (n = 4~ and transposition with ventricular septal defect (n = 15). Ages ranged from 12 days to 36 months (mean 8 months) and weights ranged from 2.7 to 12.8 kg (mean 5.7 kg). Two deaths occurred, yielding an operative mortality rate of 4.2 %. Preparatory pulmonary artery banding resulted in an increase to 65 ± 5 mm Hg in the left ventricular afterload. Linear regression of the optimum circumference of the band (Y, millimeters) against left ventricular end-diastolic volume (X, milliliters) yielded the foUowing formula: Y = 0.23X + 19.7 (r = 0.885, p < 0.001). Influence of left ventricular mass on cardiac function after anatomic correction was evaluated. The total amount of dopamine used after repair in patients in whom the left ventricular mass was less than 60 % of normal was significantly larger than that in patients with a left ventricular mass greater than or equal to 60 % of normal (p < 0.002). The left ventricular end-diastolic volume in patients with a left ventricular mass less than 60 % of normal increased significantly 2 months after operation (p < O.O~ whereas it decreased in patients with a left ventricular mass greater than 60% of normal (p < 0.01). We believe it is safe to perform tbis procedure in patients in whom the left ventricular mass is larger than 60 % of normal. Most newborn infants with simple transposition can undergo correction between 10 and 20 days of life if the ductus arteriosus is kept patent with prostaglandin £1 and the left ventricle is thereby loaded. Preparatory pulmonary artery banding, when necessary, will be satisfactory if the left ventricular pressure is greater than 65 mm Hg and/or the left ventricular/right ventricular pressure ratio is greater than 0.8.

Hisataka Yasui, MD, Hideaki Kado, MD, Kunihiro Yonenaga, MD, Manabu Hisahara, MD, Hiromi Ando, MD, Hatsuo Iwao, MD, Shousi Fukuda, MD, Yasuhiro Mizoguchi, MD, and Hiroshi Sunagawa, MD, Fukuoka, Japan

h e arterial switch operation has become a method of repairing transposition of the great arteries (TGA).1-6 The technical problems in switching the great arteries seem to have been overcome by an ingenious technical modification by Yacoub'? Lecompte.f and their associates. However, the basic questions about left ventricular

From the Departments of Cardiovascular Surgery and Pediatrics. Fukuoka Children's Hospital Medical Center. Fukuoka, Japan. Received for publication June 8. 1988. Accepted for publication Jan. 17. 1989. Address for reprints: Hisataka Yasui, MD. Fukuoka Children's Hospital Medical Center, 2-5- I. Tojin, Chuo-ku, Fukuoka 810. Japan.

12/1/11251

(L V) function to support the systemic circulation after correction have not been solved yet. Thus the indications for the arterial switch operation for TGA are still vague, especially for TGA with intact ventricular septum. This report presents our experience with the arterial switch operation for TGA with or without ventricular septal defect, with special reference to LV function. We studied the sequential changes in LV muscle mass and LV end-diastolic volume (L VEDV) before preparatory pulmonary artery banding (PAB) and before and after an arterial switch operation. Patients and methods Between June 1984 and September 1987.48 patients underwent the arterial switch operation (Table I). Their ages at op-

601

The Journal of

602

Thoracic and Cardiovascular , Surgery

Yasui et al.

Table I. Patient's initial condition and surgical results Diagnosis TGA+IVS TGA + IVS, PDA TGA + IVS, L VOTO TGA+VSD

No. of patients

Previous operation

18 II 4

PAB (18), B-T (16)

12

PAB (3), B-H (I)

48

Age (rno) 11.5 1.8 9.5 10.0 7.7

± ± ± ± ±

3.3 2.2 8.5 16.2 7.2

Weight (kg) 7.5 3.5 5.5 5.5 5.7

± ± ± ± ±

1.5 0.8 2.2 3.2 2.6

Hospital deaths 2 0 0

Q 2

IVS, Intact ventricular septum: PDA. patent ductus arteriosus: B-T, Blalock-Taussig shunt: B-H, Blalock-Hanlon operation: LVOTO. left ventricular outflow tract obstruction: PAB. pulmonary artery banding.

Table II. LV training in simple TGA No. of patients Age (mo) Weight (kg) LVSP (mm Hg) Preop. Pastop. Postop. maximum pulse rate (beats/min) Pastop. lowest arterial P0 2 (mmHg) Blalock-Taussig shunt flow

20 5.9 ± 3.8 6.1 ± 1.6 36 ± 9 65 ± 5' 183 ± 19 35 ± 6 669 ± 485

(ml/min/m-) Percent decrease in L VSV

48 ± 13

LVSP. Left ventricular systolic pressure: Po 2, oxygen tension: LVSV. left ventricular stroke volume. of'

< 0.001 versus preoperative value.

eration ranged from 12 days to 36 months with a mean of 8 months. The patients weighed 2.7 to 12.8 kg (mean 5.7 kg). Twenty-nine patients had TGA with intact ventricular septum (simple TGA). Eleven of these had primary repair between the ages of 12 days and 6 months, and 18 had repair at an older age after preliminary PAB (Table I). All but two of the patients undergoing PAB had a Blalock-Taussig shunt at the same operation. Eleven patients had a patent ductus arteriosus in addition to simple TGA. In seven of the II patients the ductus had been kept patent by the infusion of prostaglandin EJ, 0.05 !J.g/ kg/min, which had been started within 5 days of life, and they underwent correction between the ages of 12 and 25 days. Four patients had dynamic LV outflow tract obstruction." We decided that their LV pressures were high enough to allow correction without preparatory PAB. Fifteen patients had TGA with ventricular septal defect. Three of them had a previous PAB and Blalock-Hanlon procedure and the others had no previous procedures. L V training. Twenty patients with simple TGA underwent PAB with a 4 mm wide Teflon tape. A classic left BlalockTaussig shunt was performed in all but two of these patients. The shunt flow was measured intraoperatively by an electromagnetic flowmeter (Narcomatic, model 700-1500, Narco Bio-Systems, Houston, Texas). Age at PAB ranged from 19 days to 18 months (mean 6 months). The PAB was tightened until the LV pressure exceeded 60 mm Hg and/or the left ventricularfright ventricular (LV /RV) pressure ratio exceeded

0.75 (Table II). All patients survived. Eighteen of them were subjected to an arterial switch operation 3 to 15 months later (mean 5 months), and two patients did not undergo the correction because of the small LV mass. Arterial switch operation. Forty-eight patients underwent an arterial switch operation. A standard median sternotomy was made. The aorta was widely dissected distally to the left carotid artery along with a division of the ligamentum arteriosum or patent ductus arteriosus. The bilateral pulmonary arteries were dissected as far as possible behind the pericardial reflection. Pulsatile high-flow (2.5 L/m 2/min) cardiopulmonary bypass was initiated (with a Kontron bypass pump, model 20, Kontron Instruments, Inc., Everett, Mass.; pulse rate 100 beats/min, driving pressure 350 mm Hg), and patients were cooled to a rectal temperature of 28° C. The aorta was crossclamped and cold potassium cardioplegic solution (K+ 20 mEq) was infused in retrograde fashion from the coronary sinus every 20 minutes. Ventricular septal defects, when present, were closed with a Teflon patch through the tricuspid valve. The ascending aorta was divided 5 mm distal to the upper border of the sinuses of Valsalva. The pulmonary artery was divided at its bifurcation. The coronary ostia were mobilized with as much rim of the aortic wall as possible, starting at the edge of the transected aorta. The pulmonary artery was prepared for coronary anastomosis by a vertical incision made at a transected edge. The defects in the aorta were repaired with equine pericardium treated with glutaraldehyde. The cephalic segment of the ascending aorta was brought behind the pulmonary bifurcation and was anastomosed end to end to the posterior vessel as described by Lecompte and associates.t The distal segment of the pulmonary artery was anastomosed to the anterior vessel. The atrial septal defect was closed and the left atrial vent was inserted through the left atrial appendage to avoid distention of the LV until LV contraction became forceful. The aortic crossclamp was removed. The mean durations of cardiopulmonary bypass and anoxic arrest in 48 patients were 185 ± 26 and 142 ± 25 minutes, respectively. Cardiac catheterization and angiography. Cardiac catheterization was performed before PAB and before and 2 months after the arterial switch operation with the patient lightly sedated with pethidine, 5 mg/kg, Roentgenograms of the left atrium or pulmonary artery were obtained. LVEDV and LV mass were estimated according to Simpson's rule!" and Rackley's method, 11 respectively and were expressed in percentage against the predicted normal value (% N).12, 13 Echocardiogram. Sequential changes in LV internal dimension in end-diastole after correction were monitored by two-di-

Volume 98 Number 4

Arterial switch operation for TGA

October 1989

mensional and M-mode echocardiograms (Hewlett-Packard model 77020AC, Hewlett-Packard Company, Palo Alto, Calif.) and values were expressed in percentage against the predicted normal value.!" All values were expressed as mean ± standard deviation of the mean unless otherwise indicated. Statistical evaluation was performed by an analysis of variance followed by Student's t test. Differences in total amount of dopamine was tested by the Wilcoxon rank sum test. Differences were considered significant at p < 0.05.

Results LV training. All patients survived. LV systolic pressure rose from 35 ± 6 mm Hg preoperatively to 65 ± 5 mm Hg postoperatively. Pulse rate rose to 183 ± 19 beats/min postoperatively, continued at that rate for about 2 days, and then gradually decreased. The lowest arterial oxygen tension was 35 ± 6 mm Hg, which improved spontaneously 24 to 48 hours after operation (Table II). On the basis of the Emax theory (the slope of the endsystolicpressure-volume line), 15. 16 as shown in Fig. I, the percent decrease in LV stroke volume was calculated on the assumption that preload to the LV would not change significantly in the acute postoperative state; all patients had a large atrial septal defect because of the previous balloon atrial septostomy, resulting in equal pressure in the two atria. The percent decrease in LV stroke volume ranged from 31% to 67% (mean 48%) and tended to correlate with the postoperative lowest arterial oxygen tension in patients in whom flow through the Blalock-Taussig shunt was less than 800 ml/rnin/rrr' (p < 0.01) (Fig. 2). Flow through the Blalock-Taussig shunt ranged from 143 to 1900 ml/rnin/rn? with a mean of 663 ml/min/rn? (Table II). There was a significant correlation between shunt flow and the lowest postoperative arterial oxygen tension (p < 0.005) (Fig. 3). To predict the optimum circumference of the PAB in an attempt to raise the LV systolic pressure to 60 mm Hg, we plotted the circumference (Y, millimeters) of the pulmonary artery band against the preoperative body weight (X, kilograms) (Fig. 4, A) and the preoperative LVEDV (X, milliliters) (Fig. 4, B). The better correlation was noted against the LVEDV (p < 0.00 I) and its regression equation was as follows: Y=0.23X+ 19.7

LVEDV did not change significantly, but LV mass increased significantly in the mean postoperative interval of 5 months (Table III). LV mass (% N) was plotted against LV systolic pressure (Fig. 5, A) and the LV /RV pressure ratio (Fig. 5, B).

LV pressure

E max I

Pmax ----------------------------

I

• I

a ----

I

I

I

I

I

f -----r---

I

I

I

I

I

I

I

f

I

I

I

f

I

I

I

I

"

" I

b -------

EDP

603

I

I

I

I

I

f

I

I

I

I

I

I

I

post PAS

I

I

pre PAS

I

-

----------

----------_ ....... -.. '--_ _~: LV volume

I

LVSV

% decrease in LVSV

=: 1-

(Pmax-b)/(Pmax-a) f X

100

Fig. 1. Calculation of percent decrease in LV stroke volume (LVSV) caused by PAS according to Emax theory. EDP. Enddiastolic pressure; Pmax, end-systolic pressure when the pulmonary artery is completely occluded; a. end-systolic pressure before PAS; b. end-systolic pressure after PAB.

A significant correlation existed between them (p < 0.001 in both). The increase in LV systolic pressure in the mean 5-month interval just after PAB and before the arterial switch operation was analyzed in relation to LVEDV at the time of PAB (Table III, Fig. 6). The LV systolic pressure increased with time in patients with a large LVEDV (> 120% N) (p<0.001), whereas in patients with a small LVEDV ( < 120% N) it did not. The increase in LV mass of the patients with a small LVEDV was significantly less than that of patients with a large LVEDV (Table III). Arterial switch operation. Two of 48 patients died because of kinking of the translocated coronary artery (Table I). No postoperative myocardial infarctions were observed and normal sinus rhythm was present immediately after the operation in the surviving patients. LV mass, LVEDV, total amount of dopamine hydrochloride used postoperatively, and duration of postoperative mechanical ventilatory support in patients classified

The Journal of

604

Thoracic and Cardiovascular Surgery

Yasui et al.

(mmHg)

45

y = - 0.20x + 44.6

(r= -0.7615, p
40 N

0

ro o, 35

+> (/) ())

~ 0

30 25 20 40

30

% decrease

70 (%)

60

50 In

LVSV

Fig. 2. Correlation between percent decrease in LV stroke volume (L VSV) and the postoperative lowest arterial oxygen tension (PaD2) in patients with Blalock-Taussig shunt flow less than 800 ml/m 2/min.

(mmHg)

•

45 40 N

0 ro 35 Q.

+> (/) ())

~ 0

30 y = 6.97x + 30.6

25 20

• o

(r=0.6204. p<0.005)

• 0.2

0.4

0.6

0.8

1.0

1.2

B-T shunt flow

1.4

1.6

1.8

( L/min/BSA)

Fig. 3. Correlation between Blalock-Taussig shunt flowand the postoperative lowest arterial oxygen tension (PaD!).

by associated lesion are summarized in Table IV. The patients whose LV mass could not be measured (n = 5) and who had intraoperative transient kinking of the translocated coronary arteries (n = 6) were excluded from this study. Generally, the postoperative recovery was smooth. The duration of mechanical ventilatory support ranged from 1 to 12 days with a mean of 3 days (Table

IV). The total amount of dopamine used postoperatively ranged widely from none to 102 mg /kg (mean 16 mg/kg). There were no significant differences in these values among the four groups. To evaluate the influence of LV mass on LV function, we plotted LV mass (% N) against the total amount of dopamine (milligrams per kilogram) used postoperatively

Volume 98 Number 4

Arterial switch operation for TGA

October 1989

(mm) (])

o c

(]) L (])

30

(%N)

A

...

28

E 26 :::J

o

c 22 tU

20 2.0

4.0

6.0

80 (/)

8.0

10.0

(])

o c

(]) L (])

30

B

•

28

•

20

• 5

10

~5

0

LVEDV

100

B

0

0.2

0.4

(mi)

Fig. 4. Optimum circumference of PAB related to body weight (A) and LVEDV (B) to raise the LV systolic pressure to 65 ± 5 mm Hg.

(Fig. 7). There was a tendency that the smaller the LV mass, the larger the amount of dopamine used. The total amount of dopamine used for the patients with an LV mass less than 60% N was 30 ± 32 mg /kg, significantly larger than the 6 ± 9 mg/kg dose used in the patients with an LV mass more than 60% N (p < 0.002, Wilcoxon rank sum test). Cardiac catheterization data obtained 2 months after correction in all patients revealed that the mean right atrial pressure and the mean pulmonary wedge pressure were all within the normal range. Changes in LV mass and LVEDV from the prerepair stage to 2 months after repair are illustrated in Fig. 8. The LVEDV of the patients with a preoperative LV mass less than 60% N was significantly increased from 142% ± 32% to 165% ± 34% N (p < 0.05). Their LV mass was increased as well from 51% ± 7% to 81% ± 15% N (p < 0.001), which was the normal range for LV mass. On the other hand, the LVEDV of the patients with a preoperative LV mass of more than 80% N was remarkably decreased from 247% ± 64% to

120 (mmHg)

'"

40 20

30

80

LV systolic pressure

60

(r = 0.8853, p
25

60

0

y = 0.23x + 19.7

20

40

80 (/)

:J

•

20

y = 0.58x + 19.52 (r = 0.7800, p
(/)

24

tU

0

ro

c

•

20

E

1)

..Q

40

100

E 26

o

:J

(%N)

:::J

o

60

12.0

'+-

L

ro

E

(kg)

weight (mm)

A

(/)

•• •• • • y = 0.68x + 16.7 (r=0.6133. P
24

1)

..Q

100

• • •• •

'+-

oL

60 5

0.6

•

y = 49.70x + 17.71 (r=0.7564, p
0.8

1.0

1.2

LVp/RVp Fig. 5. Correlation between LV systolic pressure (A~ the LV / RV systolic pressure ratio (LVp/RVp) (B), and LV muscle mass in patients with LV training. 0, Before PAB. e, Before arterial switch operation.

Table III. Sequential changes in LV pressure, LV mass, and LVEDV after LV training LVEDV before PAS

No. of patients Interval: PAB-repair (mo) LVSP (mm Hg) After PABt At repair LV mass (% N) Before PAB At repair LVEDV (% N) Before PAB At repair LVSP. Left ventricular systolic pressure. 'Two patients did not have repair. f Aftcr PAB, Intraoperative measurement.

+p < 0.001 versus after (before) §p < 0.001 versus;" 120% N. IIp 1207< N. ~

< 0.005 versus;" 1200/, N.

PAB.

2:120% N

<120% N

16 4.5 ± 2.0

7.5 ± 5.9

65 ± 5 80 ± lOt

63 ± 7 53 ± 12§

42 ± 11 65 ± 12

37 ± 15 45 ± 1111

157 ± 25 150 ± 28

87 ± 9 93 ± 33~

4"

The Journal of

606

Thoracic and Cardiovascular Surgery

Yasui et al.

(mmHg)

100 Q)

t-

o

:::J

(/) (/)

Q)

t-

80

o, (J

o

70

>,

(/)

::J

!

I ll I J ---~--------------- ----01-0

+' (/)

o

or

90

60 ----------

L

50 40

~~

0*

e: Post PAS 0: at repair

---il--

50

---I._

---L

100

150

200 (%N)

preop. LVEDV

Fig. 6. Sequential changes in LV systolic pressure after PAS related to preoperative LVEDV. The LV systolic pressure in the patients with small LVEDV did not rise with time. * Patients who did not have arterial switch operation.

Table IV. Preoperative LV morphology, duration ofpostoperative respiratory support, and total amount of dopamine used postoperatively Diagnosis

No.

TGA+ IVS TGA + IVS, PDA TGA + IVS, LVOTO TGA+VSD

15 9 3

LQ

37

LVEDV

LV mass

(% N)

(% N)

144 202 148 197 170

± ± ± ± ±

29 48 38 64 48

67 70 63 75 69

± ± ± ± ±

20 24 23 26 23

Respiratory support (days) 2.6 2.5 4.3 3.9 3.3

± ± ± ± ±

3.1 0.9 3.0 3.4 3.2

Total amount of dopamine (lIIg/kg) 12.7±290 20.9 ± 18.8 16.0 ± 12.4 16.2 ± 26.4 16.4 ± 26.1

For abbreviations sec Table I.

169% ± 42% N (p < 0.02) and their LV mass was also decreased from 101% ± 23% to 89% ± 27% N. In the patients with a preoperative LV mass of 60% to 80% N, the LVEDV slightly decreased and LV mass increased. To follow the sequential changes in LVEDV after repair, we plotted the LV diastolic dimension estimated by echocardiography against time interval after correction in 33 patients whose postoperative follow-up was longer than 1 year (Fig. 9). The LV diastolic dimension of the patients with a preoperative LV mass less than 60% N increased significantly from the preoperative value of 83% ± 20% to 113% ± 15% N 2 months after correction (p < 0.01). It decreased gradually thereafter and reached the normal value of 104% ± 9% N 12 months after correction (p < 0.05). On the other hand, the LV diastolic

dimension of the patients with a preoperative LV mass more than 80% N decreased from the preoperative value of 126% ± 15% to the normal range of 109% ± 20% N at 2 months and 108% ± 12% N at 6 months after correction. Discussion Compared with intraatrial correction of TGA, anatomic correction at the arterial level has the advantage of normalizing the reversed role of the two ventricles and is expected to have better long-term results. 17 However,the operative mortality rate with this type of repair was high, ranging from 30% to 51% in the reported large series. U Ingenious technical modification 7. s. I x and physiologic stud y l 9-25 of this repair have improved the results.l" In

Volume 98 Number 4

(mg/kg)

(%N)

••

100 90

Q)

•

70

E 60

0

"0 l\l

~

0

~

Q)

E :::::l

>

I

50

o o

30

• •

20

I I

I I

I

:

•

l\l

..

.. ....:.

10

:. I

~

0 20

40

60

200

+J lJl

• •

40

0

250

o

l\l

Q.

300

•

80 c

60 7

Arterial switch operation for TGA

October 1989

80

LV mass

.?'.

~.

"U I "U C

150

Q)

:J

...

100

. : Preop. LV mass: >80%N (n=14) 6.: Preop. LV mass: 80 2 >60%N (n=11) 0: Preop. LV mass: S60%N (n=17)

(%N)

Fig. 7. Correlation between LV mass and total amount of

dopamine used postoperatively in patients with arterial switch operation. Values are mean ± standard error of the mean. *Significantly greater dose ofdopamine was required inpatients with LV mass less than 60% N (p < 0.002).

our series the operative mortality rate was 4.2% in the patients with simple TGA with or without preparatory PAB and in those with TGA plus ventricular septal defeet. In simple TGA the LV becomes incapable of sustaining the systemiccirculation shortly after birth because of the rapid diminution in LV mass owing to the physiologic fall in LV pressure.19. 20, 22-25 Yacoub and associates/ trained the LV by constricting the main pulmonary artery to raise LV pressure to systemic levels. LV mass was therebyincreased to above normal limits.Three questions arise. How much should the LV be loaded? What is the proper timing for surgical treatment? How much LV mass is necessary to support the systemic circulation? Tight constriction of the pulmonary artery decreases pulmonarybloodflow, the extent of which is an important influence on intercirculatory mixing.i" The decrease in pulmonary blood flow induces severe hypoxemia, as shownin Figs. 2 and 3, which will increase the risk of op-

n =42

total

100 120 140

20

40

60

80

LV mass

100

120 (%N)

Fig. 8. Changes in LVEDV and LV mass from preoperative stateto 2 months afterarterial switch operation in patients with preoperative LV mass less than 60% N, 60% to 80% N, and greater than 80% N. Values are mean ± standard errorof the mean. *p < 0.05 versus preoperative LVEDV; p < 0.001 versus preoperative LV mass. **p < 0.02 versus preoperative LVEDV. LVEDV inthepatients with LVmass less than 60% N increased significantly, whereas itdecreased intheothergroups.

eration. Indeed, Yacoub and associates- reported a 15% mortality rate with this procedure. Their technique requires a large amount of bloodflow by an aortopulmonary shunt (Fig. 3). To decrease the risk of PAB for LV training, we banded the pulmonary artery looser than in Yacoub's report. However, the decrease in arterial oxygen tension was significantand the early postoperative course was precarious in several patients (Table II). Pulmonary blood flow in patients with simple TGA gradually increases after birth and becomes approximately twice normal after 2 months of life.27 Consequently, the risk of PAB in patients older than 2 months is lower than in younger patients because oftheir greater pulmonary blood flow. Moreover, an increase in pressure

The Journal of

608

Thoracic and Cardiovascular

Yasui et at.

Surgery

(%N)

140 c 0 .iii

. : Preop. LV mass: >80%N (n=ll) Preop. LV mass: 80:;O;>60%N (n= 8) O:Preop.LVmass: ;S60%N (n=13)

130

1'>.:

c C1.l 120 E

"0 0

0

n=33

110

+-' If) CO

100

"0

90

:J

total

*

80 preop.

2

6

12

(month)

months after 0atene operation Fig. 9. Comparison of postoperative changes in LV diastolic dimension among patients with LV mass less than 60o/r N, 60% to 80% N, and greater than 80% N. LV diastolic dimension in patients with LV mass less than 60% N increased significantly, whereas it decreased in the other groups. **p < 0.001 versus patients with LV mass < 60% N. »» < 0.005 versus patients with LV mass < 60% N. 1<: P < 0.01 versus preoperative state. * p < 0.05 versus 2month postoperative state.

with time will be expected in such patients with a large LVEDV (Fig. 6). Myocyte replication is completed by 3 to 6 months of age in humans.P' Nutrition studies suggest that a genetically programmed number of mitoses does not exist; rather, the process is a response to hemodynamic and other stimuli and, once lost, the capacity for mitosis will never retum.I? Thus ideal timing for preparatory PAB is thought to be around 2 months of age. The optimum circumference of the PAB to get the intended LV pressure can be predicted from the infant's LVEDV (Fig. 4, B), simplifying the banding procedure. LVEDV reflects pulmonary blood flow,27 and the pressure gradient across the PAB is decided by the degree of banding and pulmonary blood flow. Great care should be taken in the patient whose LVEDV is small and atrial septal defect is large (Fig. 6). The LV may not respond to PAB since preload to the LV is not sufficient. A typical patient had simple TGA with anomalous pulmonary venous drainage into the right atrium. PAB did not increase the LV mass in 10 months, and the Mustard operation was performed. Another patient had simple TGA associated with a large atrial septal defect and persistent left superior vena cava. The LVEDV was 95% N. PAB at 5 months of age raised LV pressure to 56 mm Hg. Six months after operation, the LV pressure had decreased to 44 mm Hg and LV mass was 34% N. Pulmonary blood flow and LVEDV in TGA vary-? and the determinant

factors are still unknown. There may be patients in whom the LV cannot be trained by PAB. At present, if preparatory PAB is to be performed, care must be taken not to create an excessively large atrial septal defect, as suggested by Ilbawi and colleagues.i? Hypoxemia and tachycardia continued for 24 to 48 hours after banding and gradually improved thereafter. This mechanism is still unknown, but the compensatory reaction seems to begin early according to the findings of Nair and co-workers.l" which indicate that the heart weight and the ribonucleic acid content of the myocardium increased within 24 hours after aortic banding, reaching a maximal level in 2 days. There is no clear-cut study on how much LV mass is necessary to support the systemic circulation. In our series no patients died of LV failure except as a result of myocardial infarction caused by kinking of the coronary arteries. However, the smaller the LV mass, the more doses of dopamine were required postoperatively (Fig. 7). Patients with an LV mass less than 60% N required a large amount of dopamine, and patients with an LV mass more than 60% N required none or a minimum amount of dopamine. Moreover, the LVEDV of patients with an LV mass less than 60% N increased remarkably within 2 months, though in patients with an LV mass more than 80% N it decreased significantly (Fig. 8). In patients with an LV mass 60% to 80% N, the postoperative decrease in

Volume 98 Number 4 October 1989

LVEDV was slight and insignificant. The echocardiographic follow-up of LVEDV showed similar findings (Fig. 9). It appears that the LV with a small muscle mass has to enlarge to increase preload so that it can overcome the sudden increase in afterload. In fact, pulmonary congestion and cardiomegaly were encountered in those patients for about 2 months after correction but disappeared spontaneously thereafter. These facts suggest that the ability of the LV to support systemic circulation is excellent if the LV mass is more than 80% N; fair if the LV mas is 80% to 60% N, which requires an intrinsic compensatory mechanism; and poor if the LV mass is less than 60% N, which requires extrinsic inotropic support in addition to intrinsic compensatory mechanism. Thus, in estimating the safety of this correction, we believe the LV mass should be more than 60% N. In several patients, our PAB was too loose to get the LV mass greater than 60% N. Although Danford, Huhta, and Gutgesell-? suggested that preparatory PAB should load the LV greater than half systemic pressure, our experience indicates that a greater load is necessary. Since the efficacy of PAB for LV training is affected by many factors such as the LVEDV, the duration after PAB (Fig. 6), age at PAB, and growth rate, it will be difficult to show uniform criteria for the degree of PAB. However, the linear correlation between LV pressure and LV mass shown in Fig. 5 is appreciable. Thus, as a general rule, a PAB that increases the LV pressure to more than 65 mm Hg and/or the LV /RV pressure ratio to more than 0.8 will be satisfactory. Patients with TGA plus ventricular septal defect are excellent candidates for this correction, as Jatene and associates I indicated. The LV mass of this group was the largest among the four groups (Table IV). Patients with simple TGA with a patent ductus arteriosus or dynamic LV outflow tract obstruction are candidates for primary repair if the LV mass is satisfactory. Neonates with simple TGA are also good candidates who do not require preparatory PAB. With the ductus arteriosus widely patent as a result of the infusion of prostaglandin E" a decrease in LV mass is prevented and correction can be performed with low risk during the first 10 to 20 days of life. The mean value of the LV mass was 70% Nand postoperative dopamine requirement was similar to that in the other groups (Table IV). In conclusion, an arterial switch operation can be performed at low risk but is safest in patients with an LV mass larger than 60% N. Most of the patients with simple TGA can undergo correction safely in the first 10 to 20 days oflife if prostaglandin E 1 is used to keep the ductus arteriosus patent and thereby to load the LV. Preparatory PAB, when necessary, will be satisfactory if the LV

Arterial switch operationfor TGA

609

load is greater than 65 mm Hg and/or the LV/RV pressure ratio is greater than 0.8. The optimal circumference of the PAB can then be estimated from the patient's LVEDV. Wethank WilfredG. Bigelow, MD,and WilliamG. Williams, MD, for their suggestions and reviewof this manuscript. REFERENCES I. Jatene AD, Fontes YF, Souza LCB, Paulista PP, Neto CA, Sousa JEMR. Anatomic correction of transposition of the great arteries. J THORAC CARDIOVASC SURG 1982;83: 20-6. 2. Yacoub MH, Bernhard A, Lange P, et al. Clinical and hemodynamic results of the two-stage anatomic correction of simple transposition of the great arteries. Circulation 1980;62(Pt 2):1190-6. 3. Bical 0, Hazan E, Lecompte Y, et al. Anatomic correction of transposition of the great arteries associated with ventricular septal defect: midterm results in 50 patients. Circulation 1984;70:891-7. 4. Castaneda AR, Norwood WI, Jonas RA, Colon SO, Sanders SP, Lang P. Transpositionof the great arteries and intact ventricular septum: anatomical repair in the neonate. Ann Thorac Surg 1984;38:438-43. 5. Kanter KR, Anderson RH, Lincoln C, Rigby ML, Shinebourne EA. Anatomic correction for complete transposition and double-outlet right ventricle. J THORAC CARDIOVASC SURG 1985;90:690-9. 6. Quaegebeur JM, Rohmer J, Ottenkamp J, et al. The arterial switch operation: an eight-year experience. J THORAC CARDIOVASC SURG 1986;92:361-84. 7. Yacoub MH, Radley-Smith R. Anatomy of the coronary arteries in transpositionof the great arteries and method for their transfer in anatomical correction. Thorax 1978; 33:418-24. 8. Lecompte Y, Zannini L, Hazan E, et al. Anatomic correction of transposition of the great arteries: new technique without useofa prostheticconduit. J THORAC CARDIOVASC SURG 1981 ;82:629-31. 9. Yacoub MH, Arensman FW, Keck E, Radley-Smith R. Fate of dynamic left ventricular outflow tract obstruction after anatomic correction of transposition of the great arteries. Circulation 1983;68(Pt 2):1156-62. 10. Chapman CB, Baker 0, Reynolds J, Bonte FJ. Use of biplane cinefluorography for measurement of ventricular volume. Circulation 1958;18:1105-17. 11. Rackley EC, Dodge HT, Coble YO, Robert EH. A method for determining left ventricular mass in man. Circulation 1964;29:666-71. 12. Nakazawa M, Marks RA. Isabel-Jones J, Jarmakani JM. Right and left ventricular volumecharacteristics inchildren with pulmonary stenosis and intact ventricular septum. Circulation 1976;53:884-90. 13. Graham TP, Jarmakani JM, Canent RY JR, Morrow MN. Left heart volume estimation in infancy and childhood: re-

The Journal of

6I0

14. 15.

16.

17.

18.

19.

20.

21.

22.

Thoracic and Cardiovascular Surgery

Yasui et al.

evaluation of methodology and normal values. Circulation 1971;43:895-904. Fukushige J. Echocardiography in normal infants and children. Jpn Circ J 1980;44(suppl):351-62. Suga H, Katabatake A, Sagawa K. End-systolic pressure determines stroke volume from fixed end-diastolic volume in the isolated canine left ventricle under a constant contractile state. Circ Res 1979;40:238-49. McKay RG, Aroesty JM, Heller GV, Royal HD, Warren SE, Grossman W. Assessment of the end-systolic pressurevolume relationship in human beings with the use of a time-varying elastance model. Circulation 1986;74:97-104. Yacoub MH, Radley-Smith R. Venous versus arterial switch for repair of simple transposition of the great arteries: a review. Pediatr Cardiol 1983;4(suppl):151-8. Goor DA, Shem-Tov A, Neufeld HN. Impeded coronary flow in anatomic correction of transposition of the great arteries: prevention, detection, and management. J THORAC CARDIOVASC SURG 1982;83:747-54. Doesburg NH, Bierman FZ, Williams RG. Left ventricular geometry in infants with d-transposition of the great arteries and intact ventriculuar septum. Circulation 1983; 68:733-9. Ilbawi MN, Idriss FS, DeLeon SY, et al. Preparation of the left ventricle for anatomic correction in patients with simple transposition of the great arteries: surgical guidelines. J THoRAc CARDIOVASC SURG 1987;94:87-94. Lange PE, Onnasch DGW, Stephan E, et al. Two-stage anatomic correction of complete transposition of the great arteries: ventricular volume and muscle mass. Herz 1981; 6:336-43. Danford DA, Huhta JC, Gutgesell HP. Left ventricular wall stress and thickness in complete transposition of the

23.

24.

25.

26.

27.

28. 29.

30.

great arteries: implications for surgical intervention. J THORAC CARDIOVASC SURG 1985;89:610-5. Maroto E, Fouron JC, Douste-Blazy MY, Carceller AM, Doesburg N, Kratz C. Influence of age on wall thickness, cavity dimensions and myocardial contractility of the left ventricle in simple transposition of the great arteries. Circulation 1983;67:1311-7. Smith A, Wilkinson JL, Arnold R, Dickinson DF, Anderson RH. Growth and development of ventricular walls in complete transposition of the great arteries with intact.septum (simple transposition). Am J Cardiol 1982;49:362-8. Huhta JC, Edwards WD, Feldt RH, Puga FJ. Left ventricular wall thickness in complete transposition of the great arteries. J THORAC CARDIOVASC SURG 1982;84:97101. Mair DD, Ritter DG. Factors influencing intercirculatory mixing in patients with complete transposition of the great arteries. Am J Cardiol 1972;30:653-8. Keane JF, Ellison RC, Rudd M, Nadas AS. Pulmonary blood flow and left ventricular volumes in transposition of the great arteries and intact ventricular septum. Br Heart J 1973;35:521-6. Zak R: Development and proliferative capacity of cardiac muscle cells. Circ Res 1974;17(Pt 2):1134-5. Zak R, Kizu A, Bugaisky L. Cardiac hypertrophy: its characteristic as a growth process. Am J Cardiol 1979; 44:941-6. Nair KG, Cutilletta AF, Zak R, Koide T, Rabinowitz M. Biochemical correlates of cardiac hypertrophy. I. Experimental model: changes in heart weight, RNA content, and nuclear RNA polymerase activity. Circ Res 1968;23:45162.