Scand J Thor Cardiovasc Surg 25: 101-105, 1991

REPAIR OF COMPLETE ATRIOVENTRICULAR CANAL 15 Years’ Experience

Christina Frid, Claes T h o r h , Kim Book and Viking Olov Bjork

From the Pediatric Department, St G6ran ’s Children’s Hospital, and the Thoracic Clinic, Karolinska Hospital, Stockholm, Sweden

(Accepted for publication July 14, 1989)

.4bstract. Corrective surgery for complete atrioven- tricular canal was performed over a 15-year period on 72 patients, 49 of whom had Down’s syndrome: 46 were without major associated cardiac anomalies and 15 had previously undergone pulmonary artery band- ing. The pressures in the right and left ventricles equilibrated in 77% of the patients. The early mortal- ity rate was 18% and the late mortality 7%. At follow- up 4 % of the patients had severe mitral incompetence and 6 % had severe tricuspid incompetence. The mean reduction of pulmonary artery pressure was 40 mmHg. The mean peak systolic pressure ratio be- tween pulmonary artery and aorta was 0.73 before repair and had fallen to 0.38 at follow-up catheteriza- tion, when 88Yo of the patients were asymptomatic. Early repair in the first year of life is nowadays pre- ferred in order to avoid progressive pulmonary vascu- lar disease. Although the alternative of first-stage pul- monary artery banding gave lower ( I 3 Yo) mortality at the corrective operation, it cannot be recommended if atrioventricular valvular incompetence is significant. Key ulords: complete atrioventricular canal, pulmo- nary artery banding, Down’s syndrome.

Surgical correction of complete atrioventricular canal initially carried a high complication rate, mainly due to atrioventricular valve dysfunc- tion and conduction disturbances. Understand- ing of this complicated malformation was great- ly increased by the detailed description of its anatomic features published in 1966 by Rastelli and co-workers (6), whose subsequent report (7) also clarified its repair. The improved tech- nique dramatically reduced the operative mor- tality. In the past two decades standard man- agement of the defect has changed from a two- stage procedure, with initial pulmonary band- ing, to one-stage repair during the first year of life.

In this paper we review total repair of com- plete atrioventricular canal in 72 children at the

Karolinska Hospital in Stockholm in regard to early and late mortality, complications and late results, with special reference to competence of the atrioventricular valves and pulmonary vas- cular resistance.

MATERIAL AND METHODS Patients The 72 patients consecutively underwent repair of complete atrioventricular canal in the period January 1969-March 1983. The age of the 32 boys and 40 girls at operation ranged from 7 weeks to 15 years, with median 2 years (Fig. 1): of the 23 who were younger than 1 year at operation, 14 were less than 6 months old. Down’s syndrome was present in 49 cases (68%). There was no major associated cardiac anomaly in 45 cases (64%). Those with such anomaly are listed in Table I .

Pulmonary artery banding had been performed as a first-stage operation in 15 children from the early part of the study period, including six who were younger than 1 month at the time of banding. Equilibration of pressure between right and left ventricle was found in 53 (77%) of the 69 preoperatively evaluated patients.

Operatice procedure Cardiopulmonary bypass was instituted with system- ic hypothermia to 20-25°C and cardioplegic solution, usually cold blood, was infused into the aortic root. A wide incision in the right atrium provided good expo- sure of the atrioventricular valves. The anatomy was carefully assessed, particularly regarding valve struc- ture. As a rule the mitral and tricuspid valves had a common anterior and a common posterior leaflet. In single-path repair, the common leaflet was divided, with attention to the extreme importance of priority to the mitral side when attempting to ensure compe- tence of that valve (Fig. 2A).

A dacron patch was sutured to the ventricular sep- tal ridge, keeping to the right side of the ridge and using a 5-0 monofilament running suture or interrupt- ed sutures. In the vicinity of the bundle of His, near the coronary sinus, the sutures were placed in the base of the mitral valve, so as to avoid the conduction

&and J Thor Cardiovasc 25

Scan

d C

ardi

ovas

c J

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

McM

aste

r U

nive

rsity

on

11/2

6/14

For

pers

onal

use

onl

y.

102 C. Frid et a1

early death late death

1 1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 1 5 Age years

Fig. 1. Age at surgical repair of complete atrioventric- ular canal in 72 consecutive patients.

system. The divided valve leaflets were attached to both sides of the patch with mattress sutures tied on the right atrial side (Fig. 2 B). The resulting cleft of the mitral valve's septal leaflet was closed at the base with a couple of single sutures to ensure valve competence, but with care taken to avoid stenosis. Finally the upper part of the patch was fixed to the rim of the atrial septal defect with a running suture. This tech- nique was employed in all patients except three in- fants with extremely thin valve tissue, for whom we used two separate patches.

The repair was carried out entirely through the right atrium.

Follou~-up Data on the 54 surviving children were retrospective- ly reviewed up to September 1985. All 54 underwent physical examination, with echocardiography in most cases. Postoperative cardiac catheterization was per- formed once or more in 41 cases. The interval from operation to catheterization avcraged 2.7 years, and the mean time from operation to latest physical ex- amination was 5.2 years. Cardiac symptoms were classified as absent, mild or severe. Follow-up find- ings were evaluated from comparisons with preoper-

Table I. Major cardiac anomalies associated with complete atrioventricular canal

Associated anomalies No. of patients

Patent ductus arteriosus 18 (25%) Tetralogy of Fallot Coarctation of aorta 1 Subvalvular aortic stenosis 1 Subvalvular aortic stcnosis +

I

4 (6 010)

bicuspid aortic valves + coarctation Pulmonary stenosis I

ascending aorta 1 Hypoplasia of left ventricle and

ative values in regard to cardiac volume, pulmonary artery mean and systolic pressures and relationship between pulmonary and aortic systolic pressures. Mi- tral or tricuspid incompetence was classified as none, mild, moderate or severe.

RESULTS Of the 72 patients, altogether 18 (25Yo) have died. Table 11 presents the mortality in relation to age at operation. Of the 15 patients withfirst- stage banding, two (1 3 O/o) died in association with the corrective procedure.

Mortality in relation to treatment period. In 1969-1 976 there were 2 1 treated patients, with mean age 4 years, three (14%) of whom died. I n 1977-1983 the mean age of the 5 1 patients was 1 year and 15 (29%) died. Repair on a child younger than 1 year was first performed in 1976.

Early mortality. There were 13 deaths ( 1 8 O/o)

within 30 days of operation. Of the 23 patients younger than 1 year (7 weeks- 10 months, mean 4.6 months) at operation, seven died. Three of these deaths occurred during surgery, and au- topsy showed primary pulmonary hypertension in one case (3-month-old baby) and atrial fi- broelastosis in the other two cases. Two pa- tients had fatal bleeding, and a 3-month-old girl died 2 hours postoperatively, as a large ductus arteriosus had been only partially closed. The seventh death, in a 4-month-old baby, was due to left ventricular hypoplasia.

Among the 49 patients who were older than I year at operation there were six ( 1 2 O/o) perioper- ative deaths. Three occurred in heart failure with fibroelastosis 2, 5 and 30 days postopera- tively. Another patient died of cardiac tampon- ade on day 2. A 2 1/2-year-old child with severe

Table 11. Total mortality in relation to uge at operation

Age at No. of operation patients Deaths

< I year 23 I0 (43 'Yo) 1-2 years 7 1 (1 4 "k)) 2-4 years 17 2 ( l 1 %I)

2 > 4 years 25 5 (20 940) Total 72 18 (25 (Yo)

Scund J Thor Chrdiouusc 25

Scan

d C

ardi

ovas

c J

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

McM

aste

r U

nive

rsity

on

11/2

6/14

For

pers

onal

use

onl

y.

Repair of' complete atrioventricular canal 103

a

Fig. 2. (A) Repair of complete atrioventricular canal viewed from the right atrium. The common leaflet has been divided and the cleft of the mitral valve closed with isolated sutures. (B) The single patch is first anchored to the right side of the ventricular

septum with isolated sutures. The base of the mitral valve is then sutured to the patch and to the base of the tricuspid valve with mattress sutures, patched over the mitral valve.

pulmonary hypertension died during surgery. The sixth death was due to cerebral complica- tions with renal failure after 18 days.

Late mortality. Five patients (7%) died 2-6 months postoperatively, three because of infec- tion. One of the three was a 9-month-old boy who succumbed after 2 months to sepsis with disseminated intravascular coagulation and en- docarditis at the base of the mitral valve, where three stitches had cut through, causing mitral incompetence. A 15-year-old boy with sepsis and hemolytic anemia died in tricuspid incom- petence after 3 months, and cardiac failure due to infection was fatal 6 months postoperatively in a 9-month-old boy (autopsy showed also se- vere aortic stenosis and asplenia syndrome). Valvular malfunction with severe heart failure was the cause of the other two late deaths, one 3 months and one 6 months postoperatively. In the first case (1 0-month-old boy) tricuspid ste- nosis was caused by suture insufficiency and septal hematoma, and the second patient (1 8- month-old boy) had tricuspid regurgitation: He was not examined post mortem.

Complications. In a girl aged 2 years at opera- tion, angiography 1 year postoperatively re- vealed a small, narrow-necked aneurysm at the site of aortic cannulation. She remained free from symptoms 3 years later. In a girl with pulmonary artery banding at 8 months and re-

pair at 4 years, bradyarrhythmia appeared 6 years later, necessitating pacemaker treatment. After 1 month she had permanent AV-block 11, allowing pacemaking to be discontinued.

Valve function. Mitral and tricuspid valve function in the surviving 54 patients was evalu- ated at physical examination and angiography or echocardiography. Mitral incompetence was absent in 31 cases, mild in 14, moderate in seven and severe in two cases (57, 26, 13 and 4%). One of the two patients with severe mitral incompetence underwent mitral valve replace- ment in adolescence. In this body, reconstruc- tion after banding had resulted in stenosis of the right pulmonary artery, which also was sur- gically corrected (Fig. 3, Case B). The second of these two patients also had incompetence of the tricuspid and aortic valves and a small residual ventricular septal defect.

Tricuspid incompetence was absent in 38 cases, mild in six, moderate in two and severe in three cases (77, 12, 4 and 6%). All three patients with severe tricuspid incompetence died, as reported in connection with late mor- tality.

Pulmonary hypertension. Preoperative heart catheterization and angiography were per- formed on all 72 children, and in 41 cases the postoperative evaluation also included cath- eterization. Significant reduction of the systolic

Scand J Thor Cardiovasc 25

Scan

d C

ardi

ovas

c J

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

McM

aste

r U

nive

rsity

on

11/2

6/14

For

pers

onal

use

onl

y.

104 C. Frid et al.

1 1 -

1 .o-

Q.9-

: 0.8-

& 0.7-

Lo 0)

U

0

r

._ - E 0.6-

._ U r 0.5- 0 -. m 0 E

3 Q

F 0 4-

0 3 -

0 2-

0 1 -

pre op

0 Pa band

0 follow up

onstrated (Fig. 3). Of the 43 patients (60°/o) with systolic pulmonary hypertension > 42 mmHg preoperatively. 13 (30%) died, three (7 O/O)

showed pressure fall to <20 mmHg and 21 (49 O/o) to 20-40 mm. In five other cases (1 2 O/o)

there was also fall, but the values remained above 40 mmHg. One patient (Fig. 3 , Case A) had much higher pressure postoperatively, due to severe subaortic stenosis.

In six patients with severe pulmonary hyper- tension preoperatively, the systolic pressure re- mained above 40 mmHg. One of them (Fig. 3 ,

which was later corrected. A mitral valve pros- thesis was also inserted, after which the pres-

Case B) had right pulmonary artery stenosis, X 0.38

sure fell to a normal level, and he is now well and working. A third patient (Fig. 3 , Case C) had associated subaortic stenosis and severe mitral incompetence. Three patients were asymptomatic several years after the postoper-

I

2 4 6 8 10 12 14 16 18 20 22 Age years

Fig. 3. Individual mean pulmonary artery pressure before repair of complete atrioventricular canal and during follow-up. --- denotes patients with initial pulmonary artery banding. I n Cases A and C severe

subaortic stenosis and mitral regurgitation developed. Case B had persistent stenosis of the pulmonary trunk in combination with severe mitral incompetence. The normal level of 20 mmHg is indicated.

Scan

d C

ardi

ovas

c J

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

McM

aste

r U

nive

rsity

on

11/2

6/14

For

pers

onal

use

onl

y.

Repair yf complete atrioventricular canal 105

The mean ratio of peak systolic pulmonary artery : aortic pressure was 0.73 before repair of atrioventricular canal and 0.38 at follow-up in- vestigation. The individual values are shown in Fig. 4.

DISCUSSION

The complete form of atrioventricular canal still carries a considerable operative mortality. Our 18 O/o 30-day mortality compares favorably with reports ranging from 15 to 37% (2, 4, 8-10). The numerous patients with Down’s syn- drome (68 O/o) in our series showed no increased operative risk, and their hospital stay was not longer than in the other cases.

The single-patch method (3) is the preferred surgical approach. It provides the advantage of better exposure, especially in patients with a very large ventricular septal defect. The two- patch technique (1, 2) has proved advantageous only for a few operations in infants with very thin and delicate valve leaflets. To avoid dis- turbance of the conduction system, we sutured the patch to the base of the mitral valve in the area where the bundle of His runs. N o total AV block was seen in this series. The mitral ‘cleft’ was always partially closed, taking care to avoid stenosis. Reconstruction was sometimes imped- ed by lack of valve tissue after dividing the common valves, which can be an important causal factor in postoperative mitral incompe- tence. During 15-year follow-up only one pa- tient required a mitral valve replacement be- cause of left ventricular failure due to severe mitral incompetence. Presumably, however, some patients will require valve replacement in the future.

Our experience showed good results of recon- structive surgery after first-stage pulmonary banding, but other authors (5) reported signifi- cantly higher mortality. Nowadays, however, repair during the first year of life is preferred, before pulmonary vascular resistance increases. Early repair can be performed with 12 O/n opera- tive mortality (1). Although in our series the patients with pulmonary artery banding in early infancy had only 13% operative mortality at corrective repair, banding cannot be recom- mended if incompetence of the atrioventricular valves is significant. The importance of the grade of such incompetence for late results fol-

lowing repair of the malformation was empha- sized in an extensive study from the USA (10).

The superior possibilities for evaluating asso- ciated major cardiac anomalies in older chil- dren, as well as their stronger valve tissue facili- tating attachment of the valves to the patch, accounted for the lesser surgical mortality in this group than in patients younger than 12 months at operation. Consequently, the merits of the two-stage procedure, with initial pulmo- nary artery banding, should not be disregarded for cases without significant valvular incompe- tence.

REFERENCES I .

2.

3.

4.

5.

6.

7.

8.

9.

10.

Bender HW, Hammon JW, Hubbard SG, Muir- head RN. Graham TP. Repair of atrioventricular canal malformation in the first year of life. J Thorac Cardiovasc Surg 1982: 84: 5 15-522. Chin AJ, Keane JF, Norwood WI, Castaneda AR. Repair of complete common atrioventricu- lar canal in infancy. J Thorac Cardiovasc Surg

McGoon D, Rastelli G. Operation for persistent atrioventricular canal. Atrioventricular canal de- fects. Philadelphia: w. B. Saunders Company, 1976. McGrath LB, Gonzalez-Levin L. Actuarial sur- vival, freedom from reoperation and other events after repair of atrioventricular septal defects. J Thorac Cardiovasc Surg 1987; 94: 582-590. Midgley FM, Galioto FM, Shapiro SR, Perry LW, Scott LP. Experience with repair of com- plete atrioventricular canal. Ann Thorac Surg

Rastelli GC, Kirklin JW, Titus JL. Anatomic ob- servations on complete form of persistent com- mon atrioventricular canal with special reference to atrioventricular valves. Mayo Clin Proc 1966:

Rastelli GC, Ongley PA, Kirklin JW, McGoon DC. Surgical repair of the complete form of per- sistent common atrioventricular canal. J Thorac Cardiovasc Surg 1968; 55: 299-308. Rizzoli G, Mazzucco A, Brumana T, Valfre C, Rubino M, Rocco F, Daliento L, Frescura C, Gallucci V. Operative risk of correction of atrio- ventricular septal defects. Br Heart J 1984; 52:

Stewart S. Harris P, Manning J. Complete endo- cardial cushion defect. The late result of repair using the single-patch technique. Ann Thorac Surg 1985; 39: 234-237. Studer M. Blackstone EH, Kirklin JW, Pacifico AD, Soto B, Chung GKT, Kirklin JK, Bargeron LM. Determinants of early and late results of repair of atrioventricular septum (canal) defects. J Thorac Cardiovasc Surg 1982; 84: 523-542.

1982; 84: 437-445.

1980; 30: 151-159.

4 1 : 296-308.

258-265.

Scand J Tlior Cardiovasc 25

Scan

d C

ardi

ovas

c J

Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

McM

aste

r U

nive

rsity

on

11/2

6/14

For

pers

onal

use

onl

y.