surgery for congenital heart disease · anomalies of right ventricutar out- flow bicuspid pulmonary...

SURGERY FOR CONGENITAL HEART DISEASE

[ I

INTERMEDIATE SURVIVAL IN NEONATES WITH AORTIC ATRESIA: A MULTI-INSTITUTIONAL STUDY

Marshall L. Jacobs, MD Eugene H. Blackstone, MD Leonard L. Bailey, MD The Congenital Heart Surgeons Society

Objective: Controversy persis ts with regard to the t rea tment of pat ients with aort ic a tresia . Staged reconstructive operat ions and p r imary t ransp lan ta - t ion have been advocated as t rea tment strategies, but in many instances no t rea tment is under taken. A mult i - ins t i tu t ional study was under taken fi~r the purpose of character iz ing this challenging pat ient group, comparing the prevalence and outcomes of the various t rea tment strategies, and identify- ing potent ial predic tors of success or fai lure with each. Methods and results: A total of 323 neonates with aort ic a t res ia were entered into a 21-insti tution prospective, nonrandomized study between January 1, 1994, and January 1, 1997. Three protocols were used, nonexclusively in many inst i tut ions: (1) staged reconstruct ive surgery with ini t ial pal l ia t ion by a Norwood procedure and eventual Fon tan operation, (2) hear t t r ansp lan ta t ion as initial definitive therapy, and (3) nonsurgical management . Analysis was based on init ial protocol ass ignment : staged reconstructive surgery in 253 patients , hear t t r ansp lan ta t ion in 49 patients, and nonsurgical management in 21 patients. For all pa t ients initially entered into the 2 surgical t rea tment protocols, survival at 1, 3, 12, 24, and 36 months af ter entry was 67%, 59%, 52%, 51%, and 50%, respectively. A mult ivar iable analysis found incremental r isk factors for death at any time after entry to be lower bi r th weight (P = .04), associated noncardiac anomaly (P = .007), and entry into the nonsurgical protocol (P < .0001) or the staged reconstruct ive surgery protocol (P = .03). Four inst i tut ions had higher survival statist ics; 2 used a hear t t r ansp lan ta t ion protocol and 2 used a staged reconstruct ive surgery protocol. Fo r the 113 pat ients t reated at these 4 inst i tut ions, survival at 1, 3, 12, 24, and 36 months after entry was 77%, 70%, 64%, 62%, and 61%, respectively. Survival among the 4 inst i tut ions was s imi lar (P = 0.1). Conclusions: Among pat ients with aort ic atresia, o ther features of cardiac s tructure including aort ic size, degree of left vent r icular hypoplasia , and degree of mi t ra l hypoplas ia or atresia are not predictive of survival from 2 surgical protocols. The highest survival was achieved with ei ther t rea tment strategy at institutions strongly committed to the use of one or the other surgical management protocol. (J Thorac Cardiovasc Surg 1998;116:417-31)

From the Division of Cardiothoracic Surgery, Department of Surgery, Deborah Heart and Lung Center, Browns Mills, NJ, The Department of Surgery, University of Alabama at Bir- mingham, The Division of Cardiac Surgery, Department of Surgery, Loma Linda University, Loma Linda, and the Con- genital Heart Surgeons Society.

Presented at the Seventieth Scientific Sessions, American Heart Association, Orlando, Fin, Nov 11, t997.

Received for publication May 14, 1998; revisions requested May 19, 1998; revisions received May 26, t998; accepted for publication May 26, 1998.

Address for reprints: Marshall L. ,lacobs, MD, Deborah Heart and Lung Center, 200 Trenton Rd, Browns Mills, NJ 08015.

Copyright © 1998 by Mosby~ lnc. 0022-5223/98 $5,00 + 0 12/1/91981

4 1 7

4 1 8 Jacobs et aL The Journal of Thoracic and

Cardiovascular Surgery

September 1998

A ort ic valve a t res ia is the ana tomic ha l lmark of the largest g roup of n e o n a t e s born with a card iac

m a l f o r m a t i o n with single vent r ic le physiology. Of all babies cons ide red to have hypoplas t ic left hea r t syndrome, two th i rds or m o r e have aor t ic valve a t r e s i a J ' 2 In the absence of specific therapy, physi-

ologic cons t r ic t ion of the ductus a r te r iosus af ter b i r th leads invar iab ly to dea th f rom systemic hypo- perfus ion, co rona ry insufficiency, or congest ive hea r t failure. T r e a t m e n t s t ra tegies have evolved on the basis of e i ther recons t ruc t ive surgery, a long the l ines deve loped by N o r w o o d and others, 3' 4 or car-

d iac r ep l acemen t , as p i o n e e r e d by Bai ley and others.S, 6 The re have been n u m e r o u s repor t s of en-

couraging and improv ing ou tcomes with one or the o the r t r e a t m e n t s t ra tegy f rom several inst i tut ions. 7-9

Nonethe less , the re la t ive efficacy of these 2 t rea t - men t s t ra tegies r ema ins a m a t t e r of specula t ion , and in many ins tances bab ies with aor t ic a t res ia are a l lowed to die of the p red i c t ab l e na tu ra l his tory of the d isease wi thou t t he r apeu t i c in tervent ion.

A study of o u t c o m e s in a large group of neona te s with aor t ic a t res ia could be of value to pe r ina to lo - gists, neona to log is t s , ped ia t r i c cardiologis ts and surgeons, and o the r s involved in the suppor t and m a n a g e m e n t of famil ies with fetuses and neona te s with aor t ic a t res ia . A mul t i - ins t i tu t iona l s tudy was begun to p rov ide such informat ion . Analysis is cen- t e r ed a round the in t en t ion to t rea t (or not t rea t ) with a p ro toco l based on (1) s taged recons t ruc t ive surgery ( N o r w o o d p r o c e d u r e fol lowed u l t imate ly by F o n t a n - t y p e recons t ruc t ion) , (2) hear t t r ansp lan ta - t ion as initial defini t ive therapy, or (3) no surgical t r ea tment . A m o n g the goals of the s tudy were to def ine the spec t rum of ca rd iac morpho logy associ- a t ed with aor t ic valve atresia , to define the spec t rum of surgical a p p r o a c h e s cur ren t ly used and thei r respect ive ou tcomes , and to u n d e r t a k e analyses lead ing to in fe rences concern ing the impac t of pa- t ient charac ter i s t ics and t r ea tmen t s t ra tegies on ou tcome. In this ini t ial r e p o r t of the Congen i ta l H e a r t Surgeons Socie ty Mul t i - Ins t i tu t iona l Study, analysis is str ictly l imi ted to those pa t ien ts with aor t ic valve a t res ia (a da t abase has been s imul ta- neously, g e n e r a t e d for pa t ien t s with aor t ic valve stenosis and assoc ia ted left hear t obst ruct ive lesions), and cons ide ra t ion of ou tcomes is l imited to d e t e r m i n a t i o n of i n t e r m e d i a t e - t e r m survival.

Patients and m e t h o d s

Patients. Between January 1, 1994, and January 1, 1997, 323 neonates (age < 30 days) with aortic valve

Table I. The spectrum of" cardiac anomalies" associated with aortic valve atresia in 323 patients' (CHSS: 1994-1997)

Associated cardiac anomaly

Total deaths

~ % CL

No. % (%)

Position cardiac anomaly Dextrocardia 1 0 0 0-85

Atrial anomalies Intact atrial septum 13 8 62 43-77 Left juxtaposition of atrial ap- 1 1 100 15-100

pendages Cortriatriatum dexter 1 1 100 15-100

Anomalies of right ventricutar out- flow

Bicuspid pulmonary valve 4 2 50 18-82 Pulmonary valve stenosis (mild) 3 1 33 4-76

Anomalies of coronary circulation LV-coronary artery fistula 8 5 62 38-83 Single coronary orifice 4 4 100 62-100 LAD arising from RCA 1 (/ 0 0-85 Origin of CCX from RPA 1 1 100 15-100

Cardiac rhythm disturbance Wolff-Parkinson-White syndrome 1 0 0 0-85

Anomalies of the aorta Aberrant origin of right subclavian 11 1 9 1-28

artery interrupted aortic arch 1 0 0 0-85 Unusual aortic branching pattern 1 0 0 0-85 Aneurysm of transverse arch and 1 1 100 I5-100

descending thoracic aorta Right aortic arch passing retro- 1 1 100 15-100

esophageally to left-sided descending thoracic aorta

Anomalies of systemic venous connection

Bilateral SVC 28 11 39 29-51 Unroofed coronary sinus* 5 4 80 47-97 Left SVC to left atrium 1 0 0 0-85 Left atrium to systemic venous 4 2 50 18-82

channels? Anomalies of pulmonary venous

connection/drainage Partial:) 9 3 33 15-56 Total§ 3 2 67 24-96

Aortopulmonary collaterals 1 1 100 15-100

LV,, Left ventricle; LAD, left anterior descending; RCA, right coronary artmy; CCX,, circumflex coronary artery; RPA, right pulmonary artery; SVC, superior vena cava; CL, confidence limits. *Intact atrial septum in 3. ?Intact atrial septuln in all. ~Intact atrial septum in 1. §Intact atrial scptum in 1.

atresia were entered into a prospective study in 21 institutions (Appendix A). Two thirds of the patients were aged 2 days or less at admission and 90% were aged 7 days or less. The diagnosis of aortic valve atresia was made in all instances by 2-dimensional echocardiography.

Follow-up. The physician, family, or guardian of each patient not known to be dead was contacted each year since the beginning of the study. The most recent fol-

The Journal of Thoracic and Cardiovascular Surgery Volume 116, Number 3

Jacobs et al. 4 1 9

E © o

o_

A

1 0 ~ Months %

90 1 63%

8 3 55% I 12 50% I I 24 48%

7 0 ~ 36 47%

80r.-_~_ (176) 154 ( 26) (81) (68)

. . . . . . . . . . . . . . . . . . . . . . . . . . .

I [ . . . . " . . . . . . . . . . . . _ ' ~ . . . . . L . . . . . . . . . . . . . . . . .

40

30

20

10

, , ! , , I , , I • , I , , | , • I , • I i i i i i I , i I i , I , , I

3 6 9 12 15 18 21 24 27 30 33 36

Interval (Months) After Entry

0.20

0.18

0.1E

0.1~

T 0.12

C o 2; 0.10 t -

~ 0.08 CI

0.0t

0.0,

B

0.02

0.00 0

Hazard Months (x 1000)

1 150 3 34

12 4.6 24 1.6 36 .88

I

i

3 6 9 12 15 18 21 24 27 30 33 36


Fig. 1. Non-risk-adjusted survival and hazard function for death for all 323 patients in the study. A, Survival after entry (at time zero). Each circle" represents an actual death, positioned at the time of death along thc horizonal axis and actuarially along the vertical axis. The vertical bars depict _+ 1 standard error. The numbers indicate the number of patients remaining at risk at the time of the estimate. The solid line is the parametric estimate of survival, and the clashed lines enclose the 70% confidence intervals. B, Hazard function tor death.

tow-up was conducted from January through March 1997. At that time, 165 patients were known to be dead. During the 1997 follow-up, 149 of the 158 patients not known to be dead were successfully traced. Seven

additional patients have follow-up through 1996. Only 2 patients not known to be dead have no follow-up since the time of hospital discharge. Median follow-up time for survivors is 20 months, range 1.1 to 38 months;

4 2 0 Jacobs et al. The Journal of Thoracic and

Cardiovascular Surgery September 1998

T a b l e II . The spectrum of noncardiac anomalies associated with aortic valve atresia in 323 patients (CHSS; 1994-1997)

Total deaths

70% Associated noncardiac anomaly n No. % CL (%)

Turner's syndrome 7 6 86 59-98 Atagille's syndrome 1 1 100 15-100 Chromosome 15 abnormality 1 1 100 15-100 Charge association 1 1 100 15-100 Langer-Giedion syndrome 1 1 100 15-100 DiGeorge syndrome 1 1 100 15-100 Cystic fibrosis 1 1 100 15-100 Microcephaly 3 2 67 24-96 Severe diffuse CNS dysfunction 1 1 100 15-100 Unspecified CNS abnormality 1 1 100 15-100 Tracheoesophageal fistula with 2 2 100 39-100

esophageal atresia Imperforate anus 1 1 100 15-100

CNS, Central nervous system; CL, confidence limits.

mean fol low-up t ime for survivors is 21 + 9.6 months (s tandard deviat ion) .

Morphology. Estimates of left ventricular size and of the structure and patency of the mitral valve were obtained from echocardiographic reports from the participating institutions. Information regarding associated cardiac anomalies and size of the ascending aorta were obtained from echocardiographic and operative reports.

Protocols. Each patient was entered into 1 of 3 protocols based on initial intention to treat at the participating institution. Thus all patients listed for heart transplantation were considered to have entered the heart transplantation protocol. All patients managed with intent to perform an initial palliative reconstructive operat ion were considered to have entered the staged reconstructive surgery protocol. Patients who entered into neither surgical t reatment protocol were considered to have entered a protocol of nonsurgical management . Analysis was undertaken in relation to intention to treat (initial protocol assignment). Inasmuch as there were a few instances of crossover (from nonsurgical management to either of the t reatment protocols or from one surgical t reatment protocol to the other), analysis was subsequently undertaken with respect to the event of either Norwood procedure or heart transplant operation.

Data collection and analyses. Copies of all hospital documents were sent to the Data and Analysis Center at the University of Alabama at Birmingham. The data were abstracted into computer files, and these and the copies of the hospital documents were retained in confidential storage. Numerous tabulations, contingency tables, and life table analyses (Kaptan-Meier method) were made. Timc-related f reedom from death or other outcome events, and the hazard function, were also computed parametrically. Numerous multivariable analyses were made in the t ime-related hazard function regression do- main1°; other regression models were used when indi-

T a b l e I I I . Distribution of patients (n = 323) by institution and by initial protocol assignment (CHS& 1994-1997)

Institution n

Protocol

No Staged treatment reconstruction Transplant

No. % No. % No. %

R 36 3 8 33 92 H 20 1 5 19 95 C 11 8 73 3 27 N 56 56 100 J 51 51 100 P 12 12 100 F 11 11 100 D 9 9 100 K 5 5 100 S 3 3 100 U 4 4 100 G 3 3 100 A 2 2 100 T 2 2 100 B 1 1 100 O 15 2 13 E 14 3 21 I 24 23 96 M 5 4 80 L 25 3 12 1 4 Q 14 1 7 11 79

Total 323 21 7 253 78

13 87 11 79

1 4 1 20

21 84 2 14

49 16

cated. All continuous variables were examined as continuous variables (not polytomized). Interaction terms were sought in all multivariable analyses: in equations of this paper, these may be thought of as terms (variables) that are active or inactive only in certain specified relations with other variables. Variables with a P value < .1 were retained in the final equation. Variables used in the multivariable analyses and incremental risk factors for death at any time after entry are listed in Appendix B.

To estimate the proport ion of neonates entering each of the two surgical protocols that will achieve definitive repair, a Competing Risks of Events Analysis was per- formed. 11 The mutually exclusive definitive states after entry into the staged reconstructive surgery protocol were as follows: Fontan operation, 2-ventricle repair, crossover to heart transplantation, and death before definitive repair. Those for the heart transplantation protocol were heart transplantation, crossover to a Norwood operation, and death while awaiting transplantation.

R e s u l t s

M o r p h o l o g y . O f t he 323 pa t i en t s wi th ao r t i c va lve

a t res ia , 309 pa t i en t s ( 9 6 % ) h a d s e v e r e lef t ven t r i c -

u la r hypop las i a . O f these , 175 p a t i e n t s h a d mi t r a l

a t res ia , 117 h a d a p a t e n t bu t s t eno t i c o r h y p o p l a s t i c


Jacobs et al. 421

(1)

4 ~

-[

(323)

[ Entry into 1 ~ s t i t 7

R Staged .. ] [ Transplant I I . o Treatment e co,!~truc[ive / I Protocol I I Protocol vrotocol j ~ ~ .~,

tu)

L---4 1 - -7- ~ - ~[ Trans.-piant ] iv, _ ueam

- - (120)_ I. l J (7) *' i (121) ~ " ?

SVC- PA 1 I 2-Ventricle Connection Repair

(37)[ - [

a Operation ]

(7)

Fig. 2. Schematic diagram of the treatment pathway and fate of the 323 patients after initial protocol assignment. SVC-PA, Superior vena cava-pulmonary artery.

10C

9C

80

70

._> ~> 60 -1 Or) .~ 50 C (1) o 40 13.

30

20

"•" \.(83) (72) 43

. . . . . . . . . . . . . . . . ( ) 31

Months % 1 77% 3 70% 12 64% 24 62% 36 61%

lO Staged Reconstructive Protocol (institutions N and R) Heart Transplant Protocol (Institutions L and O)

0 . • I . . I • • ! . • I . . I • • ! • • I • • I • • II ~ ° I . , |

0 3 6 9 12 15 18 21 24 27 30 33 36


Fig. 3. Non-risk-adjusted survival for patients treated at the 4 low-risk institutions.

422 Jacobs et al. The Journal of Thoracic and


"E 09

"E o ,o o D-

A

1° I 90 Staged Reconstructive Protocol, n = 253

8

~ ~ , ' N ~ - d10 ,9 ) " (94) (62) (46)

0t- 4O

Months % 30 1 64%

3 55% 20" 12 50%

24 48% 10 36 47%

0 I ! II ' " I . . I . • I . • I . . I . • I . • I . II I . • I • m I , , I

0 3 6 9 12 15 18 21 24 27 30 33 36


0.20

0.18

0.16

0.14 T

0.12 o .

0 .10 &

"~ o.o8

0.06

0.04

0.02

O.OC 0

B

Staged Reconstructive Protocol, n = 253


1 150

i~ 3 34 12 4.6 24 1.6

\ ~ 36 .90

'~'~.. 3 6 9 12 15 18 21 24 27 30 33 36

interval (Months) After Entry

Fig. 4. Non-risk-adjusted survival and hazard function for patients (n = 253) initially entered in the staged reconstructive surgery protocol. A, Survival aftcr entry (at time zero). B, Hazard function for death.

mitral valve, and 17 patients had unknown mitral valve morphology. Six patients (2%) had a moderate degree of left ventricular hypoplasia, and 8 patients (2.5%) had a normal-sized left ventricle. Of these 8 patients, all but 1 had a nonrestrictive ventricular

septal defect. In total, there were 13 patients (4%) with ventricular septal defects of various sizes. The spectrum of cardiac anomalies associated with aortic valve atresia is shown in Table I and associated noncardiac anomalies in Table II.

The Journal of Thoracic and


Volume 116, Number 3

Jacobs et al. 423

0 9 ¢-

O

c -

O

10(]

90

80

Competing Risks Analysis Staged Reconstructive Protocol,

n = 253

70 i~ilL Alive, Awaiting Definitive Repair

50 c ©

4OifD He~rlf::~:plep ~ eath~Befo air30" 20

10~ . 2,VTntric:e Re~air ~ . ~ T V ~ . F o n t : n Oprrati2n ~. 0 ~ , • I , ~, I r ~ n I I ' I ~ . . : • • : . . ; . . I . . I . . I

0 3 6 9 12 15 18 21 24 27 30 33 36

Interval (Months) After Entry Fig. 5. Competing Risks Analysis for patients in the staged reconstructive surgery protocol (n - 253). The number of patients (expressed as a percent of total) in each of 5 categories across the time of follow-up is depicted. All patients begin alive, awaiting definitive repair. Thereafter, patients migrate to death, the Fontan operation, heart transplantation, and 2-ventricle repair according to the respective hazard functions operating on those remaining alive, awaiting definitive repair. At every point in time, the number of patients in all categories adds to 100% (so-called conservation of patients). Symbols represent each event, positioned on the vertical axis by multiple decrement analysis. The solid curves represent mathematical estimates of the accumulative effect of the 4 hazard functions operating on the original group of patients.

Survival. For all 323 patients, survival at 1, 3, 12, 24, and 36 months after entry was 63%, 55%, 50%, 48%, and 47%, respectively (Fig. 1, A). The single phase of hazard was a rapidly declining one, but an appreciable instantaneous risk of death was still present 12 months after entry (Fig. 1, B). As stated, numerous patient factors, protocol or procedure factors, and experience factors were used in multivariable analyses to identify risk factors for (corre- lates of) death at any time after entry. Among demographic and morphologic factors, those identified as risk factors for death at any time after entry include lower birth weight (/9 = .04), nonwhite race (P - .02), and the presence of an associated noncardiac anomaly (P = .007). None of the features of cardiac morphology analyzed, including ascending aortic size, degree of left ventricular hypoplasia, and degree of mitral hypoplasia or atresia, were predictive of survival°

Of the 21 participating institutions, 9 institutions entered patients into more than 1 protocol, including 4 institutions that used both surgical treatment protocols (Table ITI). A total of 253 patients (78%) were entered into a staged reconstructive surgery protocol, 49 patients (16%) into a heart transplantation protocol, and 21 patients (7%) into a nonsurgical protocol. There were a total of 10 crossovers, including 3 patients from the nonsurgical protocol into either the staged reconstructive surgery (n = 2) or transplantation (n -- 1) protocol. All 3 of these patients were alive at most recent follow-up. Fig. 2 depicts schematically the fate of the 323 patients after initial protocol assignment.

In comparing protocols, the likelihood of death at any time after entry was higher for patients in the no treatment protocol (P < .0001) or the staged reconstructive surgery protocol (P = .03) than for those in the transplantation protocol. Four institutions were

4 2 4 Jacobs el al. The Journal of Thoracic and


September 1998

I0C

9C

80

70 t~ >

"~ 60

O3 .~ 50 r-- ID

o 40 O_

30

2O

10

0

A

71 ) Heart Transplant Protocol, n = 4 9

- . (32)

""" i "16)

Months % 1 85% 3 71% 12 64% 24 63% 36 62%

• , I , . I , , I . . I , . I . • I . • I . . I , ,

3 6 9 12 15 18 21 24


I . , I , , I . , I

27 30 33 36

0.20

0.18

0.16

0.14 7

0.12 t'- o

0.10 ~b t "

*~ 0.08 £3

0.06

B

0.04

0.02

0.00 0

Heart Transplant Protocol, n-=49


1 160 ;~1 3 42

I I ~/I 12 3.4 I1~1 24 .85 i ~ 36 .38

I I

|

3 6 9 12 15 18 2"1 24 27 30 33 36


Fig. 6. Non-risk-adjusted survival and hazard function for death for patients (n = 49) initially entered in the heart transplant protocol. A, Survival after entry (at time zero). B, Hazard function for death.

identified as having higher survival than the rest (lower risk of death at any time after entry), P < .05. Two of these institutions used the staged reconstructive surgery protocol and 2 the heart transplantation protocol. For the 113 patients treated at these 4

institutions, survival at 1, 3, 12, 24, and 37 months after entry was 77%, 70%, 64%, 62%, and 61%, respectively (Fig. 3).

Staged reconstructive surgery protocol. The 253 patients entered into a staged reconstructive surgery



Volume 116, Number 3 Jacobs et al. 425

'°° l 90i

of) 80,

0 • t - 70 o

w 6o r-

5O c-

._m

13-

"~ 3o

20 ID

IX. 10

0 0

Competing Risks Analysis Heart Transplant Protocol, n =49

Alive, Awaiting Transplant

" " " Heart Transplant

o I

Death Awaiting Transplant

u ,../ . .-Norwood Operation

1 2 3 4 5 6


Fig. 7. Competing Risks Analysis for patients (n = 49) initially entered into the heart transplant protocol. The number of patients (expresscd as a percent of total) in each of 4 categories across the time of follow-up is depicted. All patients begin alive, awaiting transplantation. Thereafter, patients migrate to heart transplantation, death awaiting transplantation, and Norwood operations according to the respective hazard function. The depiction is otherwise as in Fig. 5.

protocol were distributed among 19 different institutions, with 3 institutions accounting for 140 (55%) of the patients and with half of the institutions each accounting for 5 patients or fewer. Time-related survival among the 253 patients entered into the staged reconstructive surgery protocol is depicted in Fig. 4, A. The hazard function for death has only a single declining phase with a sharp change in the rate of decline (inflection) at about 4 months after entry (Fig. 4, B). For patients entered into the staged reconstructive surgery protocol, incremental risk factors for death at any time after entry (multivariable analysis) include lower birth weight and associated noncardiac anomalies. More recent date of entry tended to be associated with a higher risk of mortality (P = .09). Two institutions were identified as negative risk factors for death at any time after entry. However at each of these 2 low-risk institutions, more recent date of admission was associated with a higher risk of mortality (P < .05).

Of patients managed initially with a first-stage Norwood procedure, 121 patients have subsequently undergone a superior vena cava-pulmonary artery anastomosis (bidirectional Glenn or hemi-Fontan

procedure) at a median age of 6.4 months. Thirty- eight have thus far undergone a completion Fontan operation at a median age of 20 months. Fig. 5 depicts a Competing Risks Analysis of the various events after initial enrollment in the staged reconstructive surgery protocol.

Transplantation protocol. The 49 patients entered into a heart transplantation protocol were distributed among 6 different institutions, with 3 institutions accounting for 45 (92%) of the patients. Time-related survival among the 49 patients entered into a heart transplantation protocol is depicted in Fig. 6, A. The hazard function for death has only a single declining phase with a change in the rate of decline (inflection) at 4 to 5 months (Fig. 6, B). For patients entered into the heart transplantation protocol, incremental risk factors for death at any time after entry included lower birth weight and nonwhite race. Of the 3 institutions accounting for most of the patients in the heart transplantation protocol, 2 were identified as having higher survivals for patients in the heart transplantation protocol than the others. These 2 centers together accounted for 34 pa-

426 Jacobs et al.

The Journal of Thoracic and Cardiovascular Surgery

September 1998

100

90

80

70

> "~ 60 "-I

O9 50

P 4o 0..

3O

20

(25)

(3)

(14)

. ~ ) " " " . (4 ) Awaiting Transplant (n =49)

No Treatment (n = 21)

lO

o o 1 2 3 4 5 6


Fig. 8. Non-risk-adjusted survival for patients awaiting transplantation in the heart transplantation protocol, and for patients in the no surgical treatment protocol.

tients in the heart transplantation protocol and for 5 patients in the nonsurgical protocol.

Of the total of 49 patients initially entered into the heart transplantation protocol, 36 underwent heart transplantation. Four patients crossed over to reconstructive surgery and underwent a first-stage Norwood procedure (with 2 survivors) and 1 was removed from the transplant list and died. Eight others died awaiting transplantation. Fig. 7 depicts a Competing Risks Analysis of the various events after initial enrollment in the heart transplantation protocol. That the altered natural history of patients receiving supportive care in anticipation of eventual transplantation is markedly different from the natural history of the disease itself is illustrated in Fig. 8, depicting non-risk-adjusted survival of patients awaiting transplantation and of those in the nonsurgical treatment protocol.

Protocol comparisons. Non-risk-adjusted survival at any time after entlT was higher for patients in the heart transplantation protocol than for those in the staged reconstructive surgery protocol (Fig. 9). Time-related survival was by far lowest in the nonsurgical protocol, with all survival accounted for by protocol crossovers.

Each of the 4 low-risk institutions made virtually exclusive use of one or the other of the 2 surgical

treatment protocols. Three of the 4 institutions also entered a total of 5 patients into a nonsurgical treatment protocol. To characterize the outcome of surgical treatment at a low-risk institution, and to determine whether an optimal treatment protocol could be identified, the survival of patients managed in the respective surgical treatment protocols at these 4 institutions was compared (Fig. 10, A). Despite the use of the 2 different surgical protocols, the difference in survival statistics among the 4 institutions was not more than could be due to chance alone (P - .1 ), with 70% confidence limits converging at an interval of approximately 1 week after entry. Sur- vival at 1, 3, 12, 24, and 36 months after entry was 77%, 70%, 64%, 62%, and 62%, respectively, for patients treated at the 4 tow-risk institutions. These results using the 2 treatment protocols at the 4 low-risk institutions are also compared with corresponding results at the remainder of the institutions (Fig. 10, B).

Discussion

Critique of the study. This study has the advan- tage of having enrolled 323 neonates with a single diagnosis in 3 years (enrollment and annual follow-up continue). Most of the previous studies



Volume 116, Number 3 Jacobs et aL 4 2 7

100 9o ~, Heart Transplant Protocol

7 ; . . . . . . . . . . . . . . . . .

~, e0 ~ 2 ..ii;...._...._..~....._..!.t_..:.[_.~ "-..T."2- = 2 . - -.-.-.-.'. -..-.'..- - " - . 2 . ¢J) 50 " " . . . . . . .

.......... ii ........ &i °l/ ,.- 40

a.. 5) Staged Reconstruction Protocol No Treatment Protocol 30 _'x,...

10

0 0 3 6 9 12 15 18 21 24 27 30 33 36


Fig. 9. Actuarial (life-table) survival of patients stratified according to the 3 treatment protocols. Each symbol represents an actual death, positioned at the time of death along the horizontal axis and actuariaUy along the vertical axis. The vertical bars depict ± 1 standard error. The numbers indicate the number of patients remaining at risk at the time of estimate. The solid lines are the parametric estimates of survival, and the dashed lines enclose the 70% confidence intervals.

originating from individual institutions report the results of t rea tment of considerably smaller patient groups. Among those institutions having repor ted on experience with large volumes of patients, virtually all have contr ibuted to this multi-institutional study. 12-16 The majority of previous reports focus on the outcome of a particular operative strategy (ie, either the Norwood operation or or thotopic hear t transplantation) and, in most instances, are reflective of the management of a he terogeneous patient group in that they describe the management and outcomes of patients with hypoplastic left heart syndrome or lesions resembling this syndrome. By limiting the analysis to those patients with aortic valve atresia, we have hoped to shed light on the influence of t rea tment strategy on a more anatomically and physiologically homogeneous patient population. Nevertheless, the study has limitations. No at- tempt was made to prospectively randomize management protocols. As such, the selection of protocols and therefore the inferences that can be

made concerning results may to some extent still be influenced by individual institutional biases.

Study design did not enable us to ascertain whether there are anatomic, morphologic, or other patient characteristics which for this specific cardiac anomaly would mitigate for a higher likelihood of survival with one or the other of the surgical treatment strategies. It has been speculated, for example, that this may be true with regard to either severe tricuspid valve incompetence or pulmonary venous obstruction resulting from intact atrial septum or anomalies of pulmonary venous return. Although these last 2 were among the associated cardiac anomalies included in the multivariable analyses, their relatively low prevalence together with the overall mortality of approximately 50% may have led to an underestimate of their potential impor- tance.

The study is also limited by our current inability to identify and analyze any late-phase events associated with either of the treatment strategies. Cer- tainly a consideration of the late-phase events after

428 Jacobs et aL The Journal of Thoracic and


I0C

9G

80

70

:> '~ 60

09 5O

o.) o 40

30

20

10

A

• Heart Transplant Protocol ,~'., ~ Institutions L and O

~ ] % % % . ° .

', "-....~. - . . . . . . . . . . . . , - ~ " . . . . . . . . . ~_ . . . . . . . . . . . . . .

, . . . . . . . . . . . . . , . . . . . . . . . . . . . . .

" - . . . . . . . . ( . . , . . . . . ....~.t~u) 1 (24)

P = .1 ~ Staged Reconstructive Protocol Institutions N and R

0 . • I . . I • . . . , .....) . . I . . l . • I • • I . . I ! i | i , I I • • I , I , i i 0 3 6 9 12 15 18 21 24 27 30 33 36


l°o°t. 90 Treatment Protocol Patients; n = 3 0 2

8

) - ' < 7 2 ) . _ .

¢.~ .~ 50

9 40 (D (3_

3O

20

B

10

0 0

P < .0001

Interval Institutions Other (Months) L,N,O, or R Institutions

1 77% 60% 3 70% 51%

12 64% 45% 24 62% 43% 36 61% 42%

3 6 9 12 15 18 21 24 27 30 33 36


Fig. 10. A, Non-risk-adjusted survival of patients entered into treatment protocols at the 4 low-risk institutions. The 70% confidence intervals converge at a point approximately 1 week after entry and continue to overlap thereafter, B, Survival of patients initially entering either surgical treatment protocol at the 4 low-risk institutions is shown, together with results at the remaining institutions. Patients for whom the nonsurgical protocol was initially designated are excluded from this analysis.

creation of the Fontan-type circulation (eg, ventricular dysfunction, arrhythmias, protein-losing enter- opathy) and those after transplantation (eg, late rejection, lymphoproliferative disease, graft athero- sclerosis) is essential to the overall comparisml of

surgical treatment strategies for aortic atresia. Ad- ditionally, the majority of patients enrolled in the reconstructive surgical protocol have yet to undergo a definitive repair (Fontan operation), although to date mortality for the second- and third-stage pro*


Jacobs et al. 4 2 9

cedures has been low. It should be possible, after a longer period of follow-up of this patient cohort, to know more about the relative efficacy of treatment strategies.

Inferences as to therapy. For neonates with aortic atresia, supportive medical therapy (as applied in anticipation of t ransplantat ion as initial and definitive therapy) results in a drastic al teration of the adverse natural history associated with the lesion itself. As such, it is possible that the application of the same principles of supportive therapy to all patients in anticipation of either t ransplantat ion or palliative reconstructive surgery might favorably affect overall survival. Al though the interval be tween entry and procedure was among those factors analyzed in the multivariable analyses and was not found to be predictive of survival, 75% of the patients who underwent a Norwood opera t ion did so by 8 days of age, whereas the median age for heart t ransplantat ion was 1.9 months. The experience of the transplant centers is supportive of the inference that the majori ty of patients can be successfully stabilized and suppor ted for a period of time, obviating the necessity for early emergency surgery in mos t instances. It would be of interest to explore the hypothesis that an addit ional per iod of medical stabilization and supportive therapy might favorably affect the ou tcome of initial palliative reconstructive surgery.

It was not uniformly the case that every high- volume center was among the low-risk institutions. At the same time, the nearly exclusive use of one or the o ther of the 2 surgical t rea tment protocols at each of the 4 low-risk institutions suggests that a strong institutional commi tment to a particular ther- apeutic strategy is a principal ingredient of intermediate-term success. At the present time, survival of neonates with aortic atresia managed by either of the 2 surgical t rea tment protocols at a low-risk institution is slightly greater than 60% at 2 years. This is not important ly lower than the likelihood of survival for patients with some of the o ther challenging congenital hear t malformat ions that require initial surgical interventions during the newborn period.17, 18 This study was intentionally limited to an examination of survival after initial neonata l managemen t of aortic atresia. Addi t ional follow-up and fur ther analyses of this relatively large pat ient group may fur ther enhance our ability to optimize the medical and surgical managemen t of patients with this challenging cardiac malformation.

R E F E R E N C E S 1. Murdison KA, Baffa JM, Farrell PE, Chang AC, Barber G,

Norwood WI, et al. Hypoplastic left heart syndrome: outcome after initial reconstruction and before modified Fontan procedure. Circulation 1990;82:199-207.

2. Jonas RA, Hansen DD, Cook N, Wessel D. Anatomic subtype and survival after reconstructive operation for hypoplastic left heart syndrome. J Thorac Cardiovasc Surg 1994; 107:1121-8.

3. Norwood WI, Lang P, Hansen DD. Physiologic repair of aortic atresia: hypoplastic left heart syndrome. N Engl J Med 1983;308:23-6.

4. Duty DB, Knott HW. Hypoplastic left heart syndrome: experience with an operation to establish functionally normal circulation. J Thorac Cardiovasc Surg 1977;75:624-30.

5. Bailey LL, Nelsen-Cannarella SL, Doroshow RW, Jacobson JG, Mastin RD, Allard MW, et at. Cardiac allotransplanta- tion in newborns as therapy for hypoplastic left heart syndrome. N Engl J Med 1986;315:949-51.

6. Chiavarelli M, Gundry SR, Razzouk AJ, Bailey LL. Cardiac transplantation for infants with hypoplastic left heart syndrome. JAMA 1993;270:2944-7.

7. Norwood WI, Jacobs ML, Murphy JD. Fontan procedure for hypoplastic left heart syndrome. Ann Thorac Surg 1992;54: 1025-30.

8. Iannettoni MD, Bove EL, Mosca RS, Lupinetti FM, Dorost- kar PC, Ludomirsky A, et al. Improving results with first- stage palliation for hypoplastic left heart syndrome. J Thorac Cardiovasc Surg 1994;107:934-40.

9. Bailey LL, Gundry SR, Razzouk AJ, Wang N, Sciolaro CM, Chiavarelli M. Bless the babies: one hundred fifteen late SUlaAvors of heart transplantation during the first year of life. J Thorac Cardiovasc Surg 1993;105:805-15.

10. Blackstone EH, Naftel DC, Turner ME. The decomposition of time-varying hazard into phases, each incorporating a separate stream of concomitant information. J Am Stat Assoc 1986;81:615-24.

11. David HA, Moeschberger ML. The theory of competing risks. New York: Macmiltian; 1978. p. 45-56.

12. Forbess JM, Cook N. Ruth SJ, Serraf A, Mayer JE, Jonas RA. Ten-year institutional experience with palliative surgery for hypoplastic left heart syndrome. Circulation 1995;92: (Suppl):II262-6.

13. Pigott JD, Murphy JD, Barber G, Norwood WI. Palliative reconstructive surgery for hypoplastic left heart syndrome. Ann Thorac Surg 1998;45:122-8.

14. Mosca RS, Bore EL, Crowley DC, Sandhu SK, Schork MA, Kulik TJ. Hemodynamic characteristics of neonates following first stage palliation of hypoplastic left heart syndrome. Circulation 1995;92:(Suppl)II267-71.

15. Kern JH, Hayes C J, Michler RE, Gersony WM, Quaegebeur JM. Survival and risk factor analysis for the Norwood procedure for hypoplastic left heart syndrome. Am J Cardiol 1997;80:170-4.

16. Bailey LL, Conception W, Shattuck H, Huang L Method of heart transplantation for hypoplastic left heart syndrome. J Thorac Cardiovasc Surg 1986;92:105.

17. Jonas RA, Quaegebeur JM, Kirklin JW, Blackstone EH, Daicoff G. Outcomes in patients with interrupted aortic arch and ventricular septal defect. J Thorac Cardiovasc Surg 1994; t07:1099-113.

18. Hanley FL, Sade RM, Blackstone EH, Kirklin JW, Freedom

430 Jacobs et al. The Journal of Thoracic and


RM, Nanda NC. Outcomes in neonatal pulmonary atresia with intact ventricular septum. J Thorac Cardiovasc Surg 1993;105:406-27.

Appendix A

The 21 institutions, in randomly determined order, participating in this study of the Congenital Heart Sur- geons Society.

Mott Children's Hospital, Ann Arbor, MI; The Univer- sity of Alabama at Birmingham, Birmingham, AL; The Children's Hospital, Boston, MA; The Children's Hospital of Buffalo, Buffalo, NY; Children's Memorial Hospital, Chicago, IL; The University of Chicago, Chicago, IL; Children's Hospital Medical Center, Cincinnati, OH; The Children's Hospital, Denver, CO; The Children's Hospital of Michigan, Detroit, MI; Duke University Medical Cen- ter, Durham, NC; The Milton S. Hershey Medical Center, Hershey, PA; Loma Linda University Medical Center, Loma Linda, CA; The Children's Hospital of Los Angeles, Los Angeles, CA; Jackson Memorial Hospital, Miami, FL; Columbia-Presbyterian Medical Center, New York, NY; The Children's Memorial Hospital, Omaha, NB; The Children's Hospital of Philadelphia, Philadelphia, PA;

Appendix B2. Aortic valve atresia (CHSS; 1994 to 1997; n = 323): Incremental risk ,factors for death at any time after entry, entering patient-specific variables" only

Incremental risk factors for death after entry: Patient Jactors only

Single hazard phase

Coefficient ++_ SD P value

Patient Birth weight* (lower) -0.0526 _+ 0.025 .04 African American or Hispanic 0.494 _+ 0.20 .02

ethnicity Left superior vena cava 0.541 + 0.31 .08 Associated noncardiac anomaly 0.728 _+ 0.27 .007

Intercept 0.270

6 = 0, p 0.390, v - 1.83, in - - 0. *(Birth weight [kg]) 2 squared transformation.

University of Pittsburgh Children's Hospital, Pittsburgh, PA; The Mayo Clinic, Rochester, MN; The Hospital for Sick Children, Toronto, ON, Canada; Georgetown Uni- versity Medical Center, Washington, DC.

Appendix B1. Aortic valve atresia (CHSS; 1994 to 1997; n = 323): Variables used in the multivariable analyses

Patient Dern%q'aphic: Age (days) at entry, gestation (weeks), birth weight (kg), gender (male, female), race (white, African- American, Hispanic, other), weight at ent131 (kg), height at entry (cm), body surface area at entry (mr), prenatal diagnosis Molphologic: Mitral stenosis (vs atresia), size (Z-value) of ascending aorta, left ventricular-coronary artery fistulas, presence and size of ventricular septal defect, anomalous origin of right subclavian artery, anomalous pulmonary venous connection, intact atrial septum, important noncardiac congenital anomaly Management: Treatment protocol determined at entry (intent-to-treat)

Experience Date of entry, individual institution identifier, individual surgeon identifier

Procedure (when time zero is the time of an operation) Demographic: Age at procedure, interval between entry and procedure, interval between listing and transplantation (for heart transplantation) Norwood operation: More-or-less classic Norwood, modified Norwood (augmentation with descending aorta, end-to-end aorta-pulmonary artery anastomosis with ascending aortic implant), size of systemic-puhnonary shunt (in mm, and nor- malized to body surface area and weight) Support: Total circulatory arrest time (rain), myocardial isch- emic time (min) (includes storage time for transplanted hearts)

Surgical experience (when time zero is the time of an operation)

Date of procedure, individual institution identifier, individual surgeon identifier

Appendix B3. Aortic valve atresia (CHSS," 1994 to 1997," n = 323)." Incremental risk.factors ,for death at any time after entry, entering protocol factors only (above) and patient and protocol factors (below)

Incremental risk factors for death after entry: Protocols only

Single hazard phase

Coefficient ± SD P value

Protocol No treamaent protocol 1.87 ± 0.34 <.000I Staged reconstructive surgery 0.539 ± 0.25 .03

protocol Intercept -0.740

6 - 0, p - 0.397, v - 1.81, m - 0.

Incremental risk factors for death after entry: Patient and

protocol fi4ctors

Single hazard phase

Coefficient ++_ SD P value

Patient Birth weight* (lower) 0.0577 _+. 0.025 African American or His- 0.483 ± 0.21

panic ethnicity Left ventricle-coronary ar- 0.876 ± 0.46

tery fistulas Left superior vena cava 0.538 ± 0.32 Associated noncardiac anom- 0.533 _+_ 0.27

aty Protocol

No treatmcnt protocol 1.83 + 0.34 Staged reconstructive surgery 0.629 + 0.26

protocol Intercept - 0.234

6 0, p 0.436, v = t.85, m - 0. *(Birth weight [kg]) 2 squared transformation.

.02

.02

.06

.09

.05

<.0001 .01

The Journal of Thoracic and Cardiovascular Surgery

Volume 116, Number 3 Jacobs et al. 43 1

A p p e n d i x B 4 . Aortic valve atresia (CHSS; 1994 to 1997," n = 323)." Incremental risk factors for death at any time after entry, entering patient-related factors, protocol factors, and experience factors

incremental risk .factors .for death after entry: Patient, protocol, and experience .factors

Single hazard phase

Coejficient +_ SD P value

Patient Birth weight* (lower) -0.0595 _+ 0.026 .02 African American or Hispanic ethnicity 0.707 + 0.21 .0008 Left ventricle-coronary artery fistulas 0.984 + 0.47 .04 Left superior vena cava 0.603 -+ 0.32 .06 Associated noncardiac anomaly 0.933 -+ 0.28 .0009

Protocol No treatment protocol 2.06 + 0.36 <.0001 Staged reconstructive surgery protocol 1.12 + 0.30 .0002 Experience Institution E 0.995 +_ 0.40 .01 Institution N -0 .937 _+ 0.36 .009

And (later) date of admissiont 0.154 ± 0.074 .04 Institution R -0 .946 ± 0.53 .07

And (later) date of admission 0.603 _+ 0.26 .02 Intercept - 0.590

6 - 0, p - 0.502, ~ - 1.91, m - 0 . *(Birth weight [kg]) 2 squared transformation. ?(Years since January 1, 1994) 2 squared transformation.

Availability of Journal back issues

As a service to our subscribers, copies of back issues of The .Ioumal of Thoracic and Cardiovascular Surgery for the preceding 5 years are maintained and are available for purchase from Mosby at a cost of $16.00 per issue until inventory is depleted. The following quantity discounts are available: 25% off on quantities of 12 to 23, and one third off on quantities of 24 or more. Please write to Mosby, Inc, Subscription Services, 11830 Westline Industrial Drive, St Louis, M O 63146-3318, or call 800-453-4351 or 314-453-4351 for information on availability of particular issues. If unavailable from the publisher, photocopies of complete issues may be purchased from UMI, 300 N Zeeb Rd, Ann Arbor, MI 48106, 313-761o4700.

surgery for congenital heart disease · anomalies of right ventricutar out- flow bicuspid pulmonary...

Documents