Surgical Management of Hypoplastic Right Ventricle with Pulmonary Atresia or Critical Pulmonary Stenosis and Intact Ventricular Septum

ORIGINAL ARTICLES

Surgical Management of Hypoplastic Right Ventricle with Pulmonarv Atresia or Critical Pulmonarv Stenosis and Intact Venbicular Septum . I

Clarence S. Weldon, M.D., Alexis F. Hartmann, Jr., M.D., and Robert C. McKnight, M.D. ABSTRACT Our experience with the surgical management of hypoplastic right ventricle with intact ventricular septum includes 26 patients with pulmonary atresia and 4 with critical pulmonary stenosis. Group 1 consisted of 8 neonates managed initially by transventricular valvotomy; 6 later required a secondary procedure, with 100% survival. Group 2 had 11 neonates managed by aortapulmonary artery shunting without operative death. However, only 3 have survived over the long term and l has required an additional shunt procedure. Group 3 had 9 infants who underwent concomitant valvotomy and shunting. There were 4 operative deaths and 1 late death. Finally, Group 4 included 2 infants managed by primary repair at 3 days and 6 days old with prosthetic enlargement of the right ventricle; 1 required the addition of a shunt. Both are alive. Seven of the 15 patients in Groups 1,2, and 3 who survived neonatal palliative procedures have undergone reparative operations. Two had no growth of the right ventricle and underwent repair after conversion to hicuspid atresia, by a Fontan procedure. Five had prosthetic enlargement of the right ventricle in childhood with 1 late death. Findings of this review were as follows: (1) effective palliation of pulmonary atresia and intact ventricular septum or critical pulmonary stenosis with cavitary hypoplasia of the right ventricle is rare unless transventricular flow can be established; (2) establishment of transventricular flow produces a high incidence of cavitary ”growth,” which permits later repair; (3) the Fontan operation is available for repair in patients who have no cavitary growth; and (4) when all three portions of the right ventricular cavity can be identified by angiography, a primary repair can be performed in the neonatal period with a good long-term prognosis.

Pulmonary atresia with intact ventricular septum is associated with a high neonatal mortality even though the

From the Division of Cardiothoracic Surgery, Department of Surgery, Washington University School of Medicine, St. Louis, MO. Presented at the Nineteenth Annual Meeting of The Society of Thoracic Surgeons, San Francisco, CA, Jan 17-19, 1983. Address reprint requests to Dr.Weldon, Suite 3108 Queeny Tower, 4989 Barnes Hospital Plaza, St. Louis, MO 63110.

12

diagnosis is established and palliative operation performed before closure of the ductus arteriosus [l-31. Cavitary hypoplasia of the right ventricle accompanying pulmonary atresia with intact ventricular septum prevents diastolic filling and thereby constitutes an obstruction to pulmonary blood flow that cannot be relieved by either pulmonary valvotomy or valvectomy [4]. Systemic-pulmonary artery shunts to reestablish pulmonary blood flow have improved neonatal survival, but have failed to provide long-term satisfactory palliation [5]. Reestablishment of transventricular flow is sometimes followed by “growth” of a hypoplastic right ventricle and by restoration of ventricular compliance [61. The finding of such change provides an opportunity for eventual repair by surgical prosthetic enlargement of the ventricular cavity. Some improvement in survival, palliation, and ventricular “growth” is achieved by a combination of valvotomy and systemic-pulmonary artery shunting [7].The introduction of techniques for bypassing the right ventricle [8] provided a method for repairing persistent severe hypoplasia of that ventricle. However, successful use of the Fontan operation requires that palliative procedures be performed in such a way that the compliance and resistance characteristics of the peripheral pulmonary circulation remain normal. Our recent success with the prosthetic enlargement of two hypoplastic right ventricles in newborn infants suggests that a more aggressive approach may be warranted in order to establish flow through the right ventricle prior to closure of the ductus arteriosus. This report examines the experience with pulmonary atresia with intact ventricular septum at the St. Louis Children’s Hospital and Barnes Hospital between 1969 and 1982. Review of all records in combination with a retrospective analysis of angiograms was undertaken to determine the success and failure of a variety of palliative and reparative operations in order to develop more effective management of this still highly lethal malformation.

Material and Methods The records of all patients classified as having a “hypoplastic right heart syndrome” and managed by surgical intervention were reviewed. Retrospective analysis of angiograms revealed that 26 infants had pulmonary atresia with intact ventricular septum and 4 had pinpoint openings at the pulmonary valve associated with right ventricular hypoplasia. These 4 patients, properly classified as having critical pulmonary stenosis (Fig l),

13 Weldon, Hartmann, and McKnight: Hypoplastic Right Ventricle with Pulmonary Atresia

hours of birth in 10 patients, within the first week of life in 27 patients, and at age 2 months, 3%months, and 11 months in 3 patients with a large patent ductus. The retrospective review of 28 angiograms analyzed right ventricular morphology according to the tripartite approach developed by Goor and Lillehei (91 and recommended as applicable to pulmonary atresia with intact ventricular septum by Bull and colleagues [lo] (Table). According to this analytic scheme, three types of hypoplastic right ventricles are identified: (1) those in which there is generalized hypoplasia, but in which an inlet portion of the right ventricle, a trabecular portion, and an infundibulum are distinguished; (2) those in which the trabecular portion of the right ventricle is so overgrown by hypertrophied muscle as to be effectively absent; and (3) those in which both the trabecular and infundibular portions of the right ventricle are overgrown by hypertrophied muscle (Fig 2).

Neonatal Operations 1: VALVOTOMY. Eight infants were managed initially by transventricular valvotomy. Instruments to open all atretic or critically stenotic valves were introduced through a felt-buttressed mattress suture placed in the outflow tract of the right ventricle. Ophthalmic knives, a Potts-Riker valvulotome, Brock knives, and small arterial forceps were used singly or in combination to open the pulmonary valves. All patients were operated on within 24 hours of birth. All three portions of the right ventricle were demonstrated angiographically in all 8 infants. GROUP 2: AORTA-PULMONARY ARTERY SHUNTS. Eleven infants were managed by initial shunt procedures without concomitant or secondary valvotomies. This group represented the most severe form of right ventricular hypoplasia in the entire series. Only 3 patients demonstrated all three portions of the right ventricle. Four patients were operated on within 24 hours of birth and 5 within two to seven days of birth. Two patients with a persistently patent ductus had shunt procedures at 2 months and 3% months of age. Waterston shunts were used in 8 patients and Blalock-Taussig shunts in 2. A Gore-Tex shunt was constructed between the main pulmonary artery and the aorta in l patient. GROUP 3: CONCOMITANT VALVOTOMY A N D SHUNT OPERATIONS. Concomitant valvotomies and shunt operations were performed in 9 patients. Severe hypoplasia of the right ventricle and pulmonary atresia were present in all. One patient had no recognizable trabecular portion of the right ventricle, and 1 had an associated Ebstein's malformation. All operations were performed through a median sternotomy. A right lateral extension was made in 2 of the 8 patients who had a Waterston shunt. Valvotomy was performed first in 8 patients. A central Gore-Tex shunt was constructed in 1 patient. Operations were done within the first 24 hours of life in 4 patients, between the second and fourth day of life in 4 patients, and at 11months of age in 1patient with a large, persistently patent ductus. GROUP

Fig 1 . Right ventriculogram from a newborn with cavita y hypoplasia of the right ventricle and a pinpoint opening in the pulmona y valve. The trabecular portion of the ventricle is nearly obliterated. The infundibulum is hypoplastic. There is severe tricuspid insufficiency. The tricuspid valve measured 20 mm in diameter. This infant underwent transventricular valvotomy when she was 2 days old. A right-sided Blalock-Taussig shunt was constructed when she was 7 days old and a left-sided Blalock-Taussig shunt at 20 days old. Now 4% years old, the patient is well.

had suprasystemic pressures in the hypoplastic right ventricle, tricuspid valve insufficiency, right-to-left shunting at the atrial level, and a persistently patent ductus arteriosus. There were two justifications for including 4 infants with critical pulmonary stenosis in a series of patients with pulmonary atresia with intact ventricular septum. First, these patients represented identical problems in management since there was obstruction to pulmonary blood flow at the level of the small noncompliant right ventricle, making restoration of normal pulmonary blood flow simply by valvotomy impossible. Second was the finding that 3 of the 4 patients required shunting operations to effect a rise in the arterial oxygen content. No infant with pulmonary atresia and a normal or dilated right ventricle has been seen at our institution. Infants with critical or severe pulmonary stenosis and a normal-sized right ventricle are not included in this series. Prior to 1977, all operations were performed immediately following the establishment of a diagnosis by cardiac catheterization. Since 1977, all patients have been managed with prostaglandin El to maintain ductal patency until an elective procedure could be performed within 24 hours of the time of cardiac catheterization. Initial palliative operations were performed within 24

14 The Annals of Thoracic Surgery Vol 37 No 1 January 1984

Data from Retrospective Analysis of Angiograms Performed in 28 Neonates

Palliative Procedure”

Diagnosisb

Valvotomy

Pulmonary atresia

Valvotomy

Pulmonary stenosis

Valvotomy, delayed shunt Valvotomy, delayed shunt Valvotomy, delayed shunt Valvotomy, delayed shunt Valvotomy, delayed shunt Valvotomy + shunt

Pulmonary stenosis Pulmonary stenosis Pulmonary atresia Pulmonary atresia Pulmonary atresia Pulmonary atresia

Valvotomy

+ shunt

Pulmonary atresia

Valvotomy

+ shunt

Pulmonary atresia

Valvotomy

+ shunt

Pulmonary atresia

Valvotomy

+ shunt

Pulmonary atresia

Valvotomy

+

shunt

Pulmonary atresia

Valvotomy Valvotomy

+ shunt shunt

Pulmonary atresia Pulmonary atresia

Valvotomy

+ shunt

Pulmonary atresia

+

Shunt

Pulmonary stenosis

Shunt Shunt

Pulmonary atresia Pulmonary atresia

Shunt Shunt Shunt Shunt Shunt Shunt Shunt Primary repair

Pulmonary atresia Pulmonary atresia Pulmonary atresia Pulmonary atresia Pulmonary atresia Pulmonary atresia Pulmonary atresia Pulmonary atresia

Primary repair

Pulmonary atresia

Right Ventricular Components Identified‘ Inflow, trabecular, infundibular Inflow, trabecular, infundibular Inflow, trabecular, infundibular Inflow, trabecular, infundibular Inflow, trabecular, infundibular Inflow, trabecular, infundibular Inflow, trabecular, infundibular Inflow, trabecular, infundibular Inflow, trabecular, infundibular Inflow, trabecular, infundibular Inflow, trabecular, infundibular Inflow, trabecular, infundibular Inflow, trabecular, infundibular Inflow, infundibular Inflow, trabecular, infundibular Inflow, trabecular, infundibular Inflow, trabecular, infundibular Inflow, infundibular Inflow, trabecular, infundibular lnfundibular lnfundibular Inflow, infundibular Inflow, infundibular Inflow, infundibular Inflow, infundibular Inflow, infundibular Inflow, trabecular, infundibular Inflow, trabecular, infundibular

Tricuspid Insufficiency

Tricuspid Valve Diameter (mm)

Outcome

4+

12

Alive

1+

12

Alive

2+

20

Alive

0

16

Alive (repaired)

2+

12

Alive (repaired)

4+

12

Repaired (late death)

3+

17

Alive

4+

17

Operative death

3+

12

Operative death

3+

13

Hospital death

3+

20

Alive

...

10

Alive (repaired)

3+

5

3+ 1+

8 15

Operative death Late death

4+

13

Alive (repaired)

1+

15

Alive (repaired)

3+ 3+

14 9

Late death Late death

0 3+ 4+ 1+ 3+ 3+ 2+

4 5 8 5 11 6 8 18

Late death Alive Late death Alive Hospital death Hospital death Late death Alive

4+

9

2+

Alive

Alive

‘Valvotomy = transventricular valvotomy. bPulmonary atresia = pulmonary atresia with intact ventricular septum and pulmonary stenosis = critical pulmonary stenosis. Inflow, trabecular, or infundibular portion of right ventricle.

15 Weldon, Hartrnann, and McKnight: Hypoplastic Right Ventricle with Pulmonary Atresia

Fig 2. Right ventriculograms from 3 newborns with pulmonary atresia with intact ventricular septum. Each had different degrees of cavitary hypoplasia, and each was managed differently. ( A ) Only the inlet portion of the ventricle can be identified. Contrast medium fills sinusoids. The trabecular portion and infundibulum are obliterated. A Waterston shunt was constructed on the first day of life. The infant died afrer discharge from the hospital. (B)All three portions are identifiable. The infundibulum is widely patent. The tricuspid valve is near normal size. There is gross tricuspid insufficiency. A transventricular valvotomy was done on the first day of life, and a delayed Waterston shunt was constructed on the second day of life. Two later operations were done (see Seconda ry Corrective Procedures under Results). The child died of complications of hemorrhagic diathesis 24 hours after the fourth operation. (C) There is severe cavitary hypoplasia. All three parts of the ventricle are identifiable, although the trabecular portion is nearly obliterated and the infundibulum is very small. A combined transventricular valvotomy and Waterston shunt was done at age 3 days. The child is now 8 years old. The right ventricular cavity remains too small and noncompliant to permit prosthetic enlargement. Pulmonary artery pressure is too high to permit a Fontan procedure.

16 The Annals of Thoracic Surgery Vol 37 No 1 January 1984

GROUP

4:

PRIMARY REPAIR I N THE NEONATAL PERIOD.

Two infants with pulmonary atresia and severe right ventricular hypoplasia had all three portions of the right ventricle recognizable and were treated by primary repair at age 3 days and 6 days. Operations were performed with circulatory arrest at profound hypothermia. In both patients, the pulmonary artery, pulmonary valve annulus, infundibulum, and trabecular portion of the right ventricle were enlarged by the insertion of a GoreTex patch, and the foramen ovale was sutured closed (Fig 3). The addition of a central Gore-Tex shunt was required in 1 patient with a small-volume right ventricle and severe tricuspid insufficiency.

Fig 3 . Right ventriculograms from a patient with pulmona y atresia with intact ventricular septum managed by primary repair in the neonatal period. ( A ) Appearance of ventricle 48 hours after birth. All three portions are recognizable. The trabecular portion is nearly obliterated, and there is marked hypoplasia of the infundibulum. (B) Appearance of ventricle at age 19 days following prosthetic enlargement at age 3 days. The Gore-Tex shunt placed at the time of repair is obliterated by a balloon catheter. The ventricle fills and empties well. The shunt was ligated the day after this ventriculogram was obtained.

Secondary Repairs in Childhood Seven of the 15 patients in Groups 1, 2, and 3 who survived neonatal palliative procedures underwent secondary repairs. Five children had sufficient development of the right ventricle to permit prosthetic enlargement of that ventricle (Fig 4). The size of the right ventricle was estimated to be 50% of normal in 4 and normal in 1. Among the 4 patients achieving 50% growth, there were 3 with pulmonary atresia and 1 with critical pulmonary stenosis and hypoplastic right ventricle. The patient with critical pulmonary stenosis had survived to age 6 years with a Waterston shunt only. The palliative procedure that had permitted survival in the 4 patients with pulmonary atresia with intact ventricular septum were a Waterston anastomosis performed concomitantly with a valvotomy on the first day of life in 2 patients, a valvotomy performed on the first day of life followed by a shunt on the second day of life in 1 patient, and a valvotomy performed at 3 days of life followed by a secondary valvotomy at 6 months of life in 1 patient. Reparative operations were performed at the ages of 3 years (1 patient), 4% years (l),6 years (2), and 12 years (1). All operations consisted of a myomectomy to enlarge the sinus portion of the right ventricle followed by enlargement of the anterior right ventricular wall including the distal trabecular portion, the infundibular portion, the pulmonary valve annulus, and the proximal pulmonary artery (Fig 5). Waterston shunts were dismantled in 3 patients, and plastic reconstruction of the scarred right pulmonary arteries was performed with the insertion of gussets of pericardium. In 1 patient, a small, functioning Waterston anastomosis was allowed to remain. The enlargement of the right ventricular outflow tract was made with a patch of Teflon felt covered with autogenous pericardium in 2 patients, with Gore-Tex in 2, and with a segment of irradiated and frozen preserved homograft of ascending aorta, aortic valve, and anterior leaflet of mitral valve in 2 patients. Two patients with no growth of the right ventricle (Fig 6) were managed by removal of functioning Gore-Tex central shunts, closure of the tricuspid valve, and the interposition of a valved conduit between the right atrium and the pulmonary artery (Fontan procedure) (Fig 7). The first patient had survived to the age of 15 months with a valvotomy and a central Gore-Tex shunt that had heen mnstntllrted nn the fniirfh dav nf life The

~

~~~~

17 Weldon, Hartmann, a n d McKnight: Hypoplastic Right Ventricle with Pulmonary Atresia

Fig 4. Right ventriculograms from a patient with pulmonary atresia with intact ventricular septum treated with concomitant transventricular valvotomy and Waterston shunt within 24 hours of birth. (A) Appearance of right ventricle (RV) on first day of life, There is generalized cavitary hypoplasia. All three portions are identifiable. There is severe tricuspid insufficiency. (B)Appearance of ventricle at age 6

years just before the shunt was obliterated and prosthetic enlargement was performed. There is excellent transventricular pow. The infundibulum is widely patent. The cavity is about 50% of normal size. Tricuspid insufficiency is mild. (RA = right atrium; PA = pulmonary artery.)

Fig 5. Technique used for both primary and secondary repair of cavitary hypoplasia of the right ventricle. (A) The pulmona ry artery, pulmonaru valve, infundibulum. and anterosuverior trabecular nortion

superior trabecular portion. Extensive myectomy is also done to enlarge the ventricular infundibulum and sinus. ( B ) An effort is made to vrovide some vulmnnaru n a l r ~s i~n m funrtion hu ottorhino thr Ai-

18 The Annals of Thoracic Surgery Vol 37 No 1 January 1984

Fig 6. Right ventriculograms from a patient with pulmonay atresia with intact ventricular septum managed initially at age 4 days by a transventricular valvotomy and a central Gore-Tex shunt, and by a Fontan procedure at 15 months of age. ( A ) Appearance of ventricle on first day of life. There is generalized cavita y hypoplasia, and all three

A

B

portions are identifiable. ( B ) Appearance of ventricle at 15 months of age. The cavity has failed to "grow." The pulmona y arteries are greatly enlarged following the central shunting. A Fontan procedure was done after this ventriculogram was obtained.

\

Fig 7. Technique used to convert right ventricular cavitay hypoplasia previously treated by transventricular valvotomy and central GoreTex shunting to a right ventricular bypass (Fontan). (A)The operation is done with circulatory arrest at profound hypothermia (18°C). The cannula has been removed from the right atrial appendage. The

tricuspid valve has been closed with a Teflon felt patch and the foramen ovale, by suture or patch. The Gore-Tex shunt has been divided, and the aorta is cannulated through it. This shunt is excisedfrom the pulmonary artery. ( B ) A prosthetic valved conduit is interposed between the right atrium and the pulmonay artery.

19 Weldon, Hartmann, and McKnight: Hypoplastic Right Ventricle with Pulmonary Atresia

Neonatal Valvotomy 8

Neonatal Shunt I1

Neonatal Shunt + Flvotomy

Neonatal Repair

2

I\

t Shunt

Survive Survive

Repir

.1.

~

t21

’

141

t41

0 Shunt

1

[21

-

(14) = Deaths [I61 Survivors

J

Fig 8. Tabulation of the fate of 30 infants who had pulmonary atresia (26) or criticnl pulmonary stenosis (4) with intact ventricular septum. (PI = pulmonary insufficiency.)

second patient had had a series of palliative operations including a valvotomy at birth which produced no communication between the right ventricle and the pulmonary artery, a Blalock-Taussig operation at 7 months of age which failed, placement of a central Gore-Tex shunt also at 7 months of age, and a narrowing of this GoreTex shunt, which had produced pulmonary overcirculation at 8 months of age. No growth of the right ventricle was demonstrated, and severe hypoxemia had reappeared. A right ventricular bypass was performed when the patient was 3% years old.

Results Mortality Sixteen patients survived (53%).Three of the 4 patients who had critical pulmonary stenosis and 12 of the 26 patients with pulmonary atresia are alive (Fig 8). GROUP 1. All 8 infants with recognizable tripartite right ventricles survived valvotomy, but only 2 maintained satisfactory pulmonary blood flow and arterial oxygen saturation. The remainder (75%) required a secondary procedure to relieve pulmonary oligemia. Secondary palliation included a Waterston shunt within 24 hours of valvotomy (2 patients), a Blalock-Taussig shunt within one week of valvotomy (2 patients), bilateral BlalockTaussig shunts seven and twenty days following valvotomy (1 patient), a Blalock-Taussigshunt and tertiary Gore-Tex shunt 7 months following valvotomy (1 patient), and a secondary valvotomy 6 months following original valvotomy (1 patient). All 6 infants survived the secondary procedures. Survival of infants treated by initial transventricular valvotomy was 100%.Four patients in the group underwent secondary repairs. GROUP 2. This group of 11 patients included only patients with the most unfavorable forms of right ventricu-

lar anatomy that precluded procedures other than a shunt. A shunt was done without valvotomy because the infundibulum was severely hypoplastic or not demonstrable at all by angiography. Only 3 of the ll patients survived. There were no deaths in the operating room, and only 2 died within thirty days. Six late deaths occurred between 3 and 6 months after the shunt procedure. One of the survivors is a patient who had critical pulmonary stenosis and hypoplasia of the right ventricle and who, by the age of 6 years, had attained growth of the right ventricle to approximately 50% of normal size and successfully underwent a corrective operation with prosthetic enlargement of the anterior right ventricular wall. Another survivor is a patient with a large Waterston shunt who underwent construction of a secondary Blalock-Taussig shunt at 10 years old to relieve profound cyanosis, hypoxemia, and polycythemia. The third survivor is an infant who was operated on only 2 months ago. GROUP 3. There were 4 operative deaths in this group, all of which occurred before the introduction of prostaglandin therapy. One patient died within thirty days of the operation, and 1 died 3 months following operation. Four of the 9 patients survive (44%),2 of whom are alive seven and eight years after the combined procedure. The other 2 patients have had successful repair procedures. GROUP 4. Two patients, both with pulmonary atresia with intact ventricular septum, were managed by primary prosthetic enlargement of the right ventricle performed when they were 6 days and 3 days old. The first patient had a Gore-Tex shunt added because of low arterial oxygen saturations. This provided pulmonary overcirculation two weeks later and was ligated. The 3-dayold child with severe right ventricular hypoplasia had sustained normal oxygen saturations following the repair. Both patients survive and are asymptomatic 18 months and 7% months after operation.

20 The Annals of Thoracic surgery Vol 37 No 1 January 1984

Secondary Corrective Procedures Seven patients underwent secondary corrective procedures. There were no operative deaths. Normal right ventricular pressures and normal arterial saturations were recorded at the conclusion of five operations done to enlarge the small right ventricles by myectomy and prosthetic or homograft enlargement at the right ventricular wall. Four children remain without symptoms or exercise intolerance three, five, seven and one-half, and eight years following repair. There has been 1 late death 10 months following repair done when the patient was 4% years old. The repair involved dismantling a Waterston shunt, pulmonary arterioplasty, and prosthetic enlargement of the right ventricle. Sepsis, pulmonary embolic occlusion of the right pulmonary artery, and pulmonary insufficiency with severe congestive failure developed. A fourth operation was done to insert a prosthetic pulmonary valve. The patient died 24 hours postoperatively of a bleeding diathesis causing pulmonary hemorrhage. This represents the only death among 9 patients having either primary or secondary repairs. Two patients with no growth of the right ventricle underwent interruption of previously placed central Gore-Tex shunts, patch closure of the tricuspid valve, suture closure of the foramen ovale, and the interposition of a valved conduit between the right atrial appendage and the pulmonary artery (Fontan procedure). Operations were performed when the patients were 15 months and 3%years old. The tricuspid valve was closed because the asynchrony of tricuspid regurgitant flow and atrial systole would provide atrial overload rather than atrial assistance. Normal arterial saturations were produced in both patients. Approximately 3 months following operation, the younger patient was readmitted to the hospital because of pleural effusions, which were drained and responded to a single intrapleural installation of tetracycline. Both patients remain asymptomatic with normal exercise tolerance at 15 months and 8 months after operation.

Comment Pulmonary atresia with intact ventricular septum is an uncommon condition. Keith and co-authors [ l l ] reported an incidence of 1%of the total population of patients with congenital cardiac malformations. Postnatal survival is dependent on patency of the ductus arteriosus. Therefore, both diagnosis and surgical therapy must be instituted before closure of the ductus arteriosus. Before 1977, it frequently was necessary to carry out diagnostic angiography and palliative.surgica1 operations in infants with profound hypoxemia and acidosis. Since the introduction of prostaglandin therapy to maintain ductal patency, operations can be performed following metabolic correction. This has contributed to substantial improvement in neonatal surgical mortality. In 1956, Greenwold and colleagues (121 provided a pathological classification for pulmonary atresia with intact ventricular septum that divided it into two types. In

type 1, which was common, there was a small right ventricular cavity. In type 2, which was uncommon, there was a normal or dilated cavity. We have never encountered a patient with a type 2 malformation at our institution. The implication of this classification system was that infants with large, normal, or dilated cavities would be excellent candidates for transventricular valvotomy, whereas those who had small cavities were less likely to survive and the lesion ultimately was irreparable. Investigators from a number of centers reported generally poor results with the use of transventricular valvotomy in patients with pulmonary atresia with intact ventricular septum [l, 2, 6, 7, 10). Several authors [13, 141, however, were able to demonstrate survival figures approaching 50%. Others recommended excision of the pulmonary valve either on cardiopulmonary bypass [15] or with inflow occlusion [16]. A more recent view, one we share, is that the dismal results in the management of this condition by transventricular valvotomy can be attributed to the associated right ventricular cavitary hypoplasia [4]. According to this view, the cavitary hypoplasia resulting from muscular hypertrophic overgrowth of the trabecular and infundibular portions of the right ventricle produces a noncompliant ventricle that cannot fill and therefore becomes an effective obstruction to pulmonary blood flow. The role of diminutive, hypoplastic, thickened tricuspid valves, which frequently accompany this malformation and provide additional obstruction to pulmonary blood flow, has been debated [17]. Arguing that the obstruction to blood flow therefore was irremediable in most instances and that the primary problem with these infants was persistent hypoxia, a number of investigators [l, 12, 181 recommended the use of aortopulmonary shunts. Early reports from the centers adopting this policy demonstrated improved survival in patients with pulmonary atresia with intact ventricular septum. In 1971, Bowman and colleagues [7] recommended the combination of concomitantly performed valvotomy and shunting, and demonstrated improved results with this palliative approach. Furthermore, they reported that a patient who had undergone this combined operation had substantial right ventricular growth, which permitted later repair by prosthetic enlargement of the anterior right ventricular wall. These authors postulated that only by the establishment of transventricular flow could "growth" of hypoplastic right ventricles be achieved sufficient to permit a later repair. They suggested an anterior thoracotomy for construction of a Waterston shunt, with extension across the sternum to perform the transventricular valvotomy. Others have recommended a Potts shunt [19] or a modified Blalock shunt [20], along with a valvotomy performed through a left lateral thoracotomy. Others have shown, from serial angiography, that growth of the hypoplastic right ventricular cavity was not uncommon following the performance of a successful transventricular valvotomy [21, 221. Bull and associates [lo] in 1982 adopted a method of classification for right ventricular anatomy originally de-

21 Weldon, Hartmann, and McKnight: Hypoplastic Right Ventricle with Pulmonary Atresia

scribed by Goor and Lillehei [9]. This classification recognizes the tripartite anatomy of the right ventriclethe inflow area, the trabecular portion, and the infundibulum. Using this classification, de Leva1 and colleagues [20] reported the results in 60 patients with pulmonary atresia with intact ventricular septum. The highest neonatal mortality (50%)was, curiously, among the patients with all three portions of the right ventricle present. This finding was explained by the fact that almost half of that group underwent valvotomy alone without a systemic shunt. Those patients without infundibular portions were treated by single or multiple shunts. Overall, final survival in this series of 60 patients was 32%.Among the surviving patients in whom palliation was achieved, 3 have undergone patch enlargement of the right ventricle with 2 survivors, and 4 have undergone Fontan procedures with 2 survivors. Our series of 30 patients managed between 1969 and 1982 reflects, to a minor degree, this evolution of surgical thinking. However, during this time, our therapeutic approach has been guided by two firmly held principles. The first is that valvular obstruction should be relieved whenever possible so that some degree of transventricular flow can be established. This maneuver should be part of the therapy whether or not it provides relief of hypoxemia. Persistent hypoxemia requires an augmentation of pulmonary blood flow with a systemic pulmonary shunt. The second principle is that analysis of angiograms prior to operation provides information about whether a transventricular valvotomy can be performed successfully. Patients who are shown to have no infundibular portion of the right ventricle or who have severe obstruction of the infundibular portion cannot be managed by transventricular valvotomy, and the definitive palliative maneuver is a systemic-pulmonary shunt. We believe that all therapy must be planned on an analysis of right ventricular anatomy and not on the condition of the pulmonary valve, which can always be opened. It is for this reason that our series includes 4 patients with critical pulmonary stenosis and hypoplastic right ventricles but does not include patients with a normal right ventricle and critical pulmonary stenosis. An analysis of the results in our 30 patients indicates that if transventricular flow cannot be or is not established, systemic-pulmonary shunts will encourage survival in the neonatal period, but will rarely provide longterm palliation. Therefore, we believe this approach has little value and should be abandoned for a more effective method to achieve long-term palliation. Ideas deserving a trial include prosthetic enlargement of the pulmonary artery, pulmonary valve annulus, infundibulum, and anterior trabecular portion of the right ventricle in the neonate in combination with a centrally placed shunt, and the adaptation of the Fontan procedure for use in neonates whenever the pulmonary arteries are sufficiently well developed. In contrast to the reports of others, our experience with pulmonary valvotomy performed as an initial procedure in infants with a demonstrable infundibulum has

been uniformly successful, even though 75% of patients so managed required an additional shunting procedure to restore pulmonary blood flow and to relieve hypoxemia. This approach has not only provided uniformly satisfactory palliation but also has permitted a high incidence of later reparative procedures. The analysis of our results does not provide evidence against the combined approach of concomitant valvotomy and shunting as recommended by Bowman and colleagues [7], but does suggest that before the use of prostaglandin, this approach was difficult and associated with a high operative mortality. Finally, our data show that the Waterston lateral anastomosis performed during the first few days of life as part of the palliative management of hypoplastic right heart malformations is associated with a subsequent high incidence of pulmonary arterial destruction from scamng and kinking. This arterial destruction complicates later repairs to enlarge the hypoplastic right ventricle and invalidates the use of a Fontan procedure to manage those patients who have insufficient right ventricular growth. In view of the finding that of the 15 infants who survived neonatal palliative operations, 7 were suitable for reparative operations in childhood, palliative procedures should be designed to promote and facilitate an eventual long-term repair. Success with two reparative operations performed primarily in the neonatal period warrants further exploration of that approach. The following are our current suggestions for the management of pulmonary arterial atresia with intact interventricular septum or with critical pulmonary stenosis and hypoplastic right ventricle. Whenever preoperative analysis of angiograms reveals the persistence of a patent infundibulum, two approaches have merit: transventricular pulmonary valvotomy followed by a systemic-pulmonary artery shunt in the form of a central Gore-Tex shunt or a Blalock-Taussig shunt if the initial maneuver does not produce a substantial and sustained rise in arterial oxygen tension or a definitive repair in the neonatal period followed by placement of a central shunt if the initial repair does not produce a sustained rise in the arterial oxygen tension. Whenever angiographic analysis reveals no infundibular portion or severe hypoplasia of the infundibulum, a shunt should be performed with plans to convert the malformation to tricuspid atresia and manage it with a Fontan operation after a brief interval. An alternative approach, not yet tried, is an open operation in which the entire outflow tract portion of the right ventricle, the pulmonary valve annulus, and the pulmonary artery are splayed open and enlarged with a gusset of prosthetic material. A central prosthetic shunt is added. Theoretically, success of such a procedure would permit later ligation of the shunt if there was adequate growth of the right ventricle. If growth was inadequate, a modified Fontan procedure with a valved conduit inserted between the right atrium and the prosthetic outflow tract portion of the right ventricle could be performed.

22 The Annals of Thoracic Surgery Vol 37 No 1 January 1984

Addendum Since the submission of this manuscript, a third child with critical pulmonary stenosis and cavitary hypoplasia has been successfully managed with the Fontan procedure. A diagnosis of single ventricle with pulmonary stenosis was made at birth. Palliation with a central Gore-Tex shunt was performed in the neonatal period. Recatheterization established the correct diagnosis. During the corrective operation, performed at age 21 months, the pulmonary valve was opened. The tricuspid valve measured 0.5 cm in diameter. A 15 mm valved conduit was inserted between the right atrium and the pulmonary artery. A third study showed more than 90% of pulmonary flow traversing the conduit with less than 10% traversing the hypoplastic right ventricle.

18. Edmunds LH Jr, Fishman NH, Heymann MA, Rudolph

19.

20.

21.

22.

References 1. Chanavaravibul S, Nora JJ, McNamara DG: Pulmonary

2. 3.

4.

5.

6.

7.

atresia with intact ventricular septum: problems in diagnosis and results of treatment. J Pediatr 77:1010, 1970 Cole RB, Muster AJ, Leu M, Paul MH: Pulmonary atresia with intact ventricular septum. Am J Cardiol21:23, 1968 Davignon AL, Greenwold WE, DuShane JW, Edwards JE: Congenital pulmonary atresia with intact ventricular septum: clinicopathological correlation of two anatomic types. Am Heart J 62:591, 1961 Patel RG, Freedom MB, Moes CAF, et al: Right ventricular volume determinations in 18 patients with pulmonary atresia and intact ventricular septum. Circulation 61:428, 1980 Subramanian S: Surgical treatment of complex cyanotic anomalies in infants: pulmonary atresia with intact ventricular septum. In Davila JC (ed): Second Henry Ford Hospital International Symposium on Cardiac Surgery. New York, Appleton-Century-Crofts, 1977, p 316 Moller JH, Girod D, Amplatz K, Varco RL: Pulmonary valvotomy in pulmonary atresia with hypoplastic right ventricle. Surgery 68630, 1970 Bowman FO, Malm JR, Hayes CJ, et al: Pulmonary atresia with an intact ventricular septum. J Thorac Cardiovasc Surg

61:85, 1971 8. Fontan F, Baudet E: Surgical repair of tricuspid atresia. Thorax 26240, 1971 9. Goor DA, Ldlehei CW: Congenital Malformations of the Heart. New York, Grune & Stratton, 1975, p 11 10. Bull C, de Leval MC, Mercanti C, et al: Pulmonary atresia

and intact ventricular septum: a revised classification. Circulation 66:266, 1982 11. Keith JD, Rowe RD, Wad P: Heart Disease in Infancy and Childhood. Second edition. New York, Macmillan, 1967, pp 818-830 12. Greenwold WE, DuShane JW, Burchell HB, et al: Congenital pulmonary atresia with intact interventricular septum: two anatomic types (abstract). Circulation 14:945,1956 13. Celemajor JM, Bowder JD, Gengos DC, et al: Pulmonary valve function with intact ventricular septum. Am Heart J 76:452, 1968 14. Dobell ARC, Grignon A: Early and late results-in pulmonary atresia. Ann Thorac Surg 24:264, 1977 15. Gomez-Engler HE, Grunkmeier GL, Starr A: Critical pul-

monary valve stenosis with intact ventricular septum. Thorac Cardiovasc Surg 27160, 1979 16. Willis WH: Discussion of Dobell and Grignon [14] 17. Freedom RM, Dische MR, Rowe R D The tricuspid valve in pulmonary atresia with intact ventricular septum: a morphological study of 60 cases. Arch Pathol Lab Med 102:28, 1978

AM: Anastomosis between aorta and right pulmonary artery (Waterston) in neonates. N Engl J Med 284463, 1971 Trusler GA, Yamamoto N, Williams WG, et al: Surgical treatment of pulmonary atresia with intact ventricular septum. Br Heart J 38:957, 1976 de Leval M, Bull C, Stark J, et al: Pulmonary atresia and intact ventricular septum: surgical management based on a revised classification. Circulation 66:272, 1982 Graham TP, Bender HW, Atwood GF, et al: Increase in right ventricular volume following valvotomy for pulmonary atresia or stenosis with intact ventricular septum. Circulation 5OSuppl 2:69, 1974 Rao PS, Lebman J, Borkat G: Right ventricular growth in a case of pulmonic stenosis with intact ventricular septum and hypoplastic right ventricle. Circulation 53:389, 1976

Discussion DR. JOHN J. LAMBERTI (San Diego, CA): I congratulate Dr. Weldon on an excellent presentation and appreciate the opportunity to review the manuscript before this meeting. My colleagues and I agree in general with the conclusions of his group and concur that there can be no simple answer for the treatment of this complex condition. The degree of hypoplasia of the right ventricle will always dictate the surgical approach. We have been concerned about the management of the subset of infants with severe hypoplasia of the right ventricle and myocardial sinusoids. In such patients, there are always fewer than three components of the right ventricle. In a typical patient, right ventricular angiography will result in retrograde filling of the right and left coronary artery systems. In a patient seen recently, retrograde perfusion of contrast medium into the ascending aorta could be detected after injection into the right ventricle. This newborn infant initially was treated with a modified Blalock-Taussig shunt. At the age of 6 months, the patient underwent Dacron patch closure of the tricuspid valve using deep hypothermia and circulatory arrest. Four months later, follow-up angiography revealed no filling of the right ventricular cavity after injection into the right atrium. An aortogram showed complete filling of the right and left coronary artery systems without any evidence of retrograde perfusion through the previously existing sinusoids. We presume that this therapy eliminates the retrograde flow through the sinusoids and thereby protects the left ventricle from chronic hypoxia. In our series of patients with pulmonary atresia, intact ventricular septum, and right ventricular hypoplasia, we have seen 8 patients during the last five years. No infant with a hypoplastic right ventricle was managed by valvotomy only. Four patients were treated by shunt alone, 2 were treated by shunt plus valvotomy (inflow occlusion), and 2 underwent primary repair using cardiopulmonary bypass. Our neonatal operative mortality was 25%. One death occurred in a 1,700-gm premature infant treated by shunt only and was due to pulmonary overperfusion aggravated by persistent patency of a large ductus arteriosus. The second death occurred in a patient undergoing primary repair as an infant, and presumably was due to inadequate pulmonary blood flow. This death might have been prevented by early application of a shunt as suggested by Dr. Weldon and his colleagues. There was 1late death in a patient who, as a newborn, underwent successful placement of a shunt. This child was seen at 14 months of age with left ventricular dysfunction and tricuspid regurgitation. Because of severe myocardial sinusoid forma-

23 Weldon, Hartmann, and McKnight: Hypoplastic Right Ventricle with Pulmonary Atresia

tion, surgical repair including decompression of the right ventricle was attempted. A Fontan procedure was performed, and the tricuspid valve was excised. The child died early postoperatively of a low cardiac output state. Thus, in our series, 2 of the 3 deaths occurred in patients with myocardial sinusoids. We believe that the presence of myocardial sinusoids adversely affects the outcome in the treatment of this condition. Dr. Weldon, d o you have data regarding the influence of sinusoids on your survival statistics? In our limited experience, tricuspid valve closure appears to eliminate the physiological consequences of the myocardial sinusoids and should be considered in all patients with sinusoids. Dr. Weldon, when d o you advocate tricuspid valve operation for the treatment of myocardial sinusoids? DR. JAMES A. DUFF (Corpus Christi, TX): I compliment Dr. Weldon on his excellent presentation and also thank him for providing me with a copy of the manuscript in advance. From 1972 through 1982,my associates and 1 saw 16 patients with pulmonary atresia and intact ventricular septum. Three had critical pulmonary stenosis with hypoplastic right ventricle. Six patients were treated with a primary shunt procedure. Two of them died thirty-three days and forty days postoperatively: 1 of shunt failure and 1 of pneumonia and sepsis. Both patients had only a small inlet portion of the right ventricle with complete obliteration of the trabecular and outlet portions. Two late deaths occurred following secondary procedures to provide an outlet from the right ventricle. One patient with pulmonary atresia and intact ventricular septum and a dilated, enlarged right ventricle died 6 hours after transventricular valvotomy. Mortality for the entire group was 31%. All of the 11 survivors had small but morphologically intact right ventricles. Three patients, 1 who had primary right ventricular outlet reconstruction and 2 who had primary open valvotomy, are doing well and require no further operations. Four patients, 2 who underwent primary shunt operation and 2 who had primary open valvotomy, have had delayed right ventricular outlet reconstruction with good results. Four patients recently underwent successful open valvotomy and are doing well clinically. Recatheterization has yet to be done. We believe that in patients with morphologically intact though small right ventricles, primary open valvotomy without concomitant placement of a shunt can be employed successfully in many instances. Nine out of 10 patients who had primary valvotomy alone survived without an additional shunt. While these patients remained cyanotic in the early postoperative period, they did not become acidotic. Maintenance of ductal patency with infusion of prostaglandin during the postoperative period has been employed for as long as eleven days. In our experience, if these ventricles recover at all, they begin to function within the first week or two. Serial contrast twodimensional echocardiography and Doppler flow measurement of the right ventricular outflow tract in the postoperative period have recently been instituted as a means of assessing right ventricular recovery. The findings will form the basis of a future report. We agree with Dr. Weldon and his co-workers that in the patient with absent infundibular and trabecular portions of the right ventricle, a primary shunt procedure is the best alternative. However, we think that a right ventricular reconstruction procedure should be undertaken early. We have not had an opportunity to use the Fontan procedure for this particular defect, but certainly agree that a central prosthetic shunt is preferable to preserve the pulmonary arterial anatomy if such a procedure is to be done.

DR. ANTHONY L. MOULTON (Baltimore, MD): This excellent series again underscores how dramatically prostaglandin infusions have altered the early outlook for the newborn with pulmonary atresia and intact ventricular septum, and justifies a reevaluation of our initial approach to these patients. With this in mind, my associates and I updated the series from ColumbiaPresbyterian Medical Center and the University of Maryland Hospital by reviewing the records of 44 patients. Forty-two patients underwent initial palliation at either institution (Table); the other 2 patients underwent a second-stage procedure following Blalock shunts elsewhere. Eleven patients underwent valvotomy alone. Among the 7 patients undergoing transventricular valvotomy, there was only 1 survivor. More recently, 4 patients underwent a supravalvular valvotomy with perioperative prostaglandin infusions. The 2 early deaths among these 4 patients were in a patient with severe coronary artery abnormalities and in another with transmural bacterial myocarditis. The 2 survivors required palliative reoperation in the form of a shunt at 3 weeks and a right ventricular outflow patch at 20 months of age. Eight patients had placement of a shunt alone: five Waterston and three Blalock shunts. In many of these patients, of course, the right ventricle was so small that a closed valvotomy was not possible, as was the case in 7 of the 11 patients in the series of Dr. Weldon and his colleagues who underwent shunt alone. Nineteen patients underwent simultaneous valvotomy and shunt. Among the late deaths, 2 occurred at reparative openheart operations in patients 1year and 2 years old; 1 occurred 18 months following reparative open-heart operation; and another occurred during an attempted repeat valvotomy in a patient 5 months old. There are, however, 10 long-term survivors, 7 of whom have undergone reparative open-heart operation. A total of 14 babies have undergone "corrective" open-heart operation using a variety of techniques. There are 10 long-term survivors up to thirteen years postoperatively. One patient required reoperation at 10 years of age for right ventricular failure with a patent foramen ovale and progressive pulmonary insufficiency with an aortic homograft. Based on this experience, we agree with Dr. Weldon that the right ventricle can be salvaged in most of these patients and that every effort should be made to maximize that potential. It appears that this can be achieved by simultaneous valvotomy and shunt in most patients, although the approach must be individualized. Even in the patient with minimal right ventricular growth, we would incorporate the right ventricular chamber in a modified Fontan repair, and the technique described by Dr. Weldon and his co-workers might potentiate that. The patient who can undergo primary definitive repair is rare indeed.

Results of lnitial Palliation in 42 Patients Procedure Valvotomy alone Shunt alone Valvotomy + shunt RVOT patch on bypass No operation"

No. of Patients

Early Survivors

Late Survivors

11 8 19 1

3 (27%) 2 (25%) 15 (79%) 1 (100%)

3 (27%) 1 (12.5%) 10 (53%) 1(100%)

3

3 (100%)

2 (67%)

'Patients had large patent ductus arteriosus. RVOT = right ventricular outflow tract.

24 The Annals of Thoracic Surgery Vol 37 No 1 January 1984

I thank those who discussed the paper and praise their excellent results. Dr. Lamberti asked about patients with sinusoids connecting to the coronary circulation. I really d o not know the answer to this. We have had 3 patients with such a condition, and all of them are among the group of patients who received shunts only and who subsequently died. They are, in fact, patients with a terribly difficult condition to treat and are, I think, at special risk at the time of diagnosis, since the contrast medium gains immediate access to the coronary circulation and may induce arrhythmias or even ventricular fibrillation. In regard to closure of the tricuspid valve, our recommendation was in conjunction with the Fontan operation and our argument was that regurgitation through the tricuspid valve would be asynchronous with atrial systole and simply provide an atrial overload. However, there may be some value DR. WELDON:

Notice Regarding Manuscript Submission As of January 1, 1984, manuscripts submitted for publi-

cation in The Annals of Thoracic Surgery should be sent to Thomas 8. Ferguson, M. D., Editor, 3108 Queeny Tower, Barnes Hospital Plaza, St. Louis, MO 63110.

in excluding the right ventricle at an early time in these patients to prevent a runoff from the coronary circulation. In response to Dr. Duff's discussion, it seems to me that if you are going to do an open operation, it is probably best not to be contented with a simple valvotomy. Rather, some effort should be made to enlarge the hypoplastic infundibulum and the anterosuperior portion of the right ventricle to get as much transventricular flow as possible. Finally, in response to Dr. Moulton's comments, we are well aware of the excellent work that has been done by the Columbia group on this difficult problem and acknowledge that these investigators were the first to demonstrate that the early establishment of transventricular flow is essential for cavitary growth that will permit eventual repair. All of us are in their debt.