Determinants of outcome in hospitalized infants with congenital heart disease

CONGENITAL

HEART DISEASE

Determinants of Outcome in Hospitalized Infants with Congenital Heart Disease Gail D. Pearson, MD, ScD, Catherine A. Neil, MD, Timothy M. Beittel, and Langford Kidd, MD

To identify fbtermhumts of adverse outcome in this era of early, definitive treatment, retrospective data were analyzed for 1966 on infants aged
(81%)underwentl16surgicalproeedures,37% (40 of 166) were hospitalized for >26 days and 29% (31 of 166) dbd dving the first year. Univariate analysis showed that complex disease (i.e., severe ventricular hypopksia, ventriwlar septal malatlgnment or outfbw atresia), extracardiac anomalies, early initial presentation, and prdonged stay in the intensive care unit were significantly assodated with infant death, whereas surgery was assodated with a significantly increased rate of survival. The findings for compbx disease and surgery persisted In multlpb b&tic regresslon analysis. It is conch&d that outcome In most infants with congenital cardiac defects ls now extremely favorable, and that major research and preventive efforts should focus on complex WngeLilital cardiac defects. (Am JCardld 1991;68:1665-1669)

nfants with symptomatic heart diseasearc being treated at younger agesthan ever before. At least 50% of definitive pediatric cardiac surgical proccdures are now being performed before I year of age, with sharp declinesin operative mortality since the early 1970s.’ The newborn with transposition of the great arteries that is repaired early and who is discharged <3 weeks after surgery is rapidly becoming the paradigm for newbornswith serious heart disease.Unfortunately, however, a small proportion of hospitalized cardiac infants continue to have poor outcomes. Cardiovascular malformations still account for 44% of the premature mortality associatedwith all congenital anomalies, which arc themselvesthe fifth ranking causeof prcmaturc death in the United States.2Many published studies have analyzed determinants of survival for specific defects and operative procedures,3-7as well as the effect of extracardiac anomalieson morta1ity.s However, sincethe report from the New England Regional Infant Cardiac Program on determinants of 5-year survival,9 there has been little work on factors affecting survival in a cohort of infants requiring early hospitalization, or on the spectrum of cardiac diseaseseverity in such a cohort. This study, therefore, analyzesthe spectrum 01 diseaseseverity and determinants of outcome among infants hospitalized during the first year of life for cardiac disease.

I

METHODS

From the Division of Pediatric Cardiology, Johns Hopkins Hospital, Baltimore, Maryland. Manuscript received April 3, 1991; revised manuscript received June 5, 199 I, and accepted June 6. Address for reprints: Catherine Neill, MD, Division of Pediatric Cardiology, Johns Hopkins Hospital, Brady 516, 600 N. Wolfe Street, Baltimore, Maryland 21205.

Patientsr All infants
lies included chromosomal abnormalities (e.g., trisomy 21) Mendelian and other syndromes (e.g., Noonan), other major anomalies (e.g., encephalocele),12and significant problems associatedwith preterm birth. Analytic methods: Univariate analyses were performed using chi-square techniques.13A multiple logistic regressionmodel was then generatedto examine the effect of each independent variable on outcome while holding the remaining covariates constant.14,15 The dependent variable was death in the first year of life. The independent variables were: complex lesion, extracardisc anomaly, surgery, medical admission, intensive care unit stay, hospital stay, time to diagnosis,race, sex, birthweight and type of insurance.All testsof statistical significance were 2-tailed, and based on rejecting the null hypothesis at the p = 0.05 level.

TABLE I NoncomplexCardiacDefects:AssociatedExtracardiac Anomalies and Mortality No. of Pts.

Definitive Surgery* (no. of pts.)

Conotruncal & major septation Tetralogy with ECA d-Transposition with ECA TAPVR with ECA IAA with ECA ECD with ECA Valvar & stenotic Aortic stenosis with ECA Pulmonary stenosis with ECA Coarctation with ECA Septal & PDA VSD with ECA PDA with ECA

32

24

5 1

4

0

1 7

0

Total

76

Category

7 0

No. of Deaths

2

5

1

-

22 11 44 2

Mortality Rate

16%

0 -

-

0 0 0 1

1

21

-

8

3

18

14

4

3

0

Palliative Surgery (no. of pts.)

0

1 4

4 3

17%

1

-

2

1

1

0

1 8 3

1 6 3

1 2

21 12

16 10

7

5

3 1 1

0 1 1 5 1 3

11 l-

-

54

RESULTS

24%

0

1

1

9

13

18%

*Patient categorized as having definitive repair if ultimate procedure was definitive. ECA = extracardiac anomaly; ECD = endocardial cushion defect; IM = interrupted aorticarch; PDA = persistent ductusarteriosus; TAPVR = total anomalous pulmonary venous return; VSO = ventricular septal defect.

sis, age at first admission, age at surgery, type of surgery, number of admissionsin first year, total hospital days in first year,.total neonatal and pediatric intensive care unit days in first year, reasonfor admission (medical management, surgery), cardiac diagnoses,extracardisc defect(s), outcome at 1 year (alive or dead) and age at death. Data were coded on standard forms by a single reviewer (TMB) using the RBase statistical package,and were checkedextensively for errors. Complete data were available for 98% (106 of 108) of the infants identified who met the study criteria. The primary classification of heart defectsused the hierarchical systemadopted by the Baltimore Washington Infant Study. lo Each infant thus had 1 primary cardiac diagnosis based on the most recent cardiac record. In addition, the concept of a spectrum of severity, as delineated by Clark” to characterize the range of cardiac diseaseresulting from specific embryologic mechanisms,was used as the basisfor classifying cardiac disease as complex or noncomplex. In our study, complex cardiac defectswere those that representedextreme outliers on the spectrum of severity: defectswith marked ventricular hypoplasia, septal malalignment or outflow atresia. All other defectswere classifiedas noncomplex (Tables I and II). Major extracardiac anoma1056

THE AMERICAN

JOURNAL

OF CARDIOLOGY

A total of 132 infants
VOLUME

68

OCTOBER

15, 1991

of study population:

Median time to diagnosis was 7 days. A third of caseswere diagnosedwithin 1 day after birth (32%; 35 of 108) and over half (54%; 58 of 108) were diagnosed by 1 week of life. Consistent with this finding, 68% of the infants (73 of 108) were hospitalized for their cardiac diseasein the neonatal period (<29 days of life). A total of 116 surgical procedures were performed on 88 patients. Of these, 76% (67 of 88) required only 1 procedure in the first year of life, 17% ( 15 of 88) had 2, and 7% (6 of 88) had 23. The majority (59%; 52 of 88) of first surgical procedures were definitive. Open heart procedures were performed in 59 infants; 29 underwent closed heart procedures only. Median age at first surgery was 46 days (range 1 to 334 days), with 22% (19 of 88) during the first week of life and 44% (39 of 88) within the neonatal period. Examining surgical procedures by complexity, we found that 60 infants with noncomplex diseasehad 68 procedures, or 1.1 per patient, and only 7% (6 of 88) had >2 procedures. Conversely, 28 infants with complex disease underwent 48 procedures, an average of 1.7 per infant, with 22 procedures performed in 54% (15 of 28). Of these 48 procedures, 37 were palliative (including 14 shunts, 5 bands, 3 coarctation repairs with band, and 4 Nor-woodstage I procedures),and 11 were intended as definitive (including 1 transplant and 2 repeat open repairs). The 20 infants who did not have surgery were a diverse group, including infants with hypoplastic left ventricle and lethal chromosomal anomalies (trisomy 18) and those whose surgical repairs were postponeduntil after the first year of life. The 108 study patients had a total of 239 hospital admissionsduring the first year of life (mean 2, range 1 to 10). Many infants (46%; 50 of 108) had only 1 admission, but 17% (18 of 108) had >4 admissions.The 239 admissionsresulted in 4,166 hospital days, with an averagelength of stay of 17 days. Hospital days totaled >28 for 37% (40 of 108) of patients. Infants with complex disease were more likely to have >28 inpatient days than infants with noncomplex disease,but this difference was not statistically significant (49 vs 3 1%; p = 0.11). Intensive care unit days totaled 2,676, or 64% of all hospital days; 54% (58 of 108) had an intensive care unit stay of >7 days. Outcome: Overall mortality in the first year of life was 29% (3 1 of 108). Mortality data for noncomplex and complex diagnosesare presentedin Tables I and II; the difference in mortality by complexity in the tables was significant (chi-square 9.5, p = 0.002). The highest mortality rate (58%; 7 of 12) was experienced by infants with both complex disease and extracardisc anomalies; the lowest (6%; 3 of 47) occurred in the absence of either condition (chi-square 14.8, p = 0.0001). The mortality rate among infants with

TABLE II Complex Cardiac Defects: Associated Extracardiac Anomalies and Mortality Definitive No. Surgery* of (no. of Pts. pts.) Looping l-Transposition + Ebstein with ECA lvemark Complex conotruncal &major septation Tetralogy& PA with ECA

4 2 o217 3 0-

d-transposition & 1 pulmonary stenosis with ECA 0 Double-outlet 5 ventricle with ECA lSingle ventricle 2with ECA 2ECD with mal2 alignment with ECA IAtresias & 16 hypoplasias Hypoplastic LVf 5 with ECA 3Hypoplastic RV$ 4 with ECA lShonevariant 1 with ECA 2 Total

37

*Patient categorized

as having

Palliative Surgery (no. of pts.)

Total MorProce- No. of tality dures Deaths Rate

1 1

1 1

2 2

1 0

5

11

24

1 8

2

1

2

0

1

1 -

-

2

-

-

3

7

5

1 2 2

1 3 3 5

1 0 0 1

3

1 12

1 22

0 9

1

4

7

2

4 1 1 2

6 3 1 5

4 3 0 1 0 1

48

18

2

9 definitive

47%

-1

-

25%

24 repair

56%

49%

in first year of life if ultimate

procedure was definitive. Kategory includes 5 patients with aortic valve atresia, 2 with mitral valve atresia and 1 with extreme mitral valvestenosis. Kktegory includes 4 patients with pulmonary valve atresia and 1 with tricuspid valve atresia. ECA = extracardiac anomaly; ECD = endocardial cushion defect; LV = left ventricle; RV = right ventricle; PA = pulmonary atresia.

major extracardiac anomalies was 47% (17 of 36) compared with 19% (14 of 72) for infants with minor or no anomalies (chi-square 7.7, p = 0.005). Median age at death was 75 days (range 2 to 361). Infants who underwent 2 1 surgical procedure had a mortality rate of 20% (18 of 88) whereas those who did not undergo surgery in infancy had a mortality rate of 65% (13 of 20; chi-square 13.7, p = 0.0002). Both early diagnosis(57 days) and a prolonged intensive care unit stay (>7 days) were significantly associated with a higher risk of mortality (chi-square 8.5, p = 0.004; chi-square 4.3, p = 0.04). The multiple logistic regressionmodel did not differ significantly from a “perfect” model, as measured by log likelihood (chi-square 86.8, degrees of difference 94, p = 0.69) and goodnessof fit (chi-square 89.4, degreesof difference 94, p = 0.62) statistics, and demonstrated a correct classification rate of 81%. Complex disease was significantly associated with death by 1 year of age (p = 0.003, odds ratio = 6, 95% confidence OUTCOME IN CONGENITAL HEART DISEASE 1057

interval 2 to 19), whereas surgery (L 1 procedure) was significantly associated with survival to 1 year (p = 0.02, odds ratio = 8, 95% confidence interval 1 to 42). Extracardiac anomalies, highly significant in the univariate analyses,only approachedsignificance in the multivariate model (p = 0.12, odds ratio 3, 95% confidenceinterval 0.8 to 8). The remaining variables (race, sex, birthweight, age at diagnosis,insurance type, medical admission, prolonged intensive care unit stay [>7 days], prolonged hospital stay [>28 days]) were not significantly associatedwith outcome.

Both the disproportionate occurrence and effect on mortality of extracardiac anomalies in children with cardiac diseaseare well known. The incidence of such anomalies in our study, 33%, is consistent with other studies.*J***l The combined effect of complex disease and major extracardiac anomalies has not been investigated systematically. In our study, infants with both conditions had a mortality rate of 58%, compared with a rate of just 6% for infants with neither condition. Determinants of mortality for infants between these extremes require further study. Like Fyler et al9 we found that surgery exerts a significant influence on mortality rates in a multivariate model. The positive relation of surgery to successful outcome is illustrated most clearly by the changes in mortality for tetralogy and transposition. Unlike Fyler et aL9 we found no significant association of birthweight with mortality. It is likely that improved management for low birthweight infants over the past 20 years is responsiblefor most of this difference. In addition, our study excluded preterm infants with the sole diagnosis of persistent ductus arteriosus, and only 2 study infants weighed <1,500 g, the most high-risk group given current neonatal management. In an era of increasing concern over the health status of poor children, we were interested in the effect of socioeconomicstatus on mortality.** Using insurance type as a proxy for socioeconomicstatus, we found no effect of this variable on outcome. Moreover, race had no significant influence on mortality. Although these data cannot be usedto comment on accessto care, they do suggestthat once children with cardiac diseaseare admitted to a tertiary care center, their socioeconomic status does not affect survival. A possible limitation of this study is its small size, compared, e.g., with the various regional, populationbased studies that have addressedsimilar questions.9,23 On the other hand, the strength of the study is that it includes a heterogeneousgroup of cases,defined and studied at 1 tertiary center. Moreover, we were gratified that the frequency of key variables was comparable to findings from other studies and that available statistical measuressupported the appropriatenessof our regression model for the data set. Further research as well as preventive efforts should focus on complex cardiac defects.

DISCUSSION Comparison of our findings with those of Fyler et al9 nearly 20 years ago reveals similarities as well as differences. The l-year mortality rate is lower (29 vs approximately 38%) and, when specific cardiac defects are compared, the mortality differences are even more striking. For example, l-year mortality rates for tetralogy of Fallot and transposition were 40 and 25% in their study, and almost none of these infants had successful open repair. In our study, there were no deaths among the 13 infants with these diagnoses,and all had successfulanatomic correction. Like Fyler et aL9we found that complexity of cardiac defect was a significant determinant of poor outcome. There has been no consensuson definition of complexity across a study group. The definition used here is derived from morphogenetic studies” and comprises severe ventricular hypoplasia, ventricular septal malalignment and outflow atresia. Application of this definition, which combines groups of defects analyzed separately in most other studies, revealeda remarkably worse l-year survival in the complex group, as well as more surgical proceduresand a higher incidence of prolonged hospitalization. Our findings are consistent with previous studies showing that outcome for individual cardiac defects is affected by complexity,499J6-‘9and may seemclinically intuitive, but there is increasing interest in rigorous statistical documentation of clinical intuition.*O Moreover, it is valuable to document that surgical and medical advanceshave now made possible an extremely favorable l-year outcome in hospitalized infants with cardiac disease, including those with tetralogy and transposition. Our findings also emphasizethat complex defects still present a major challenge. In particular, the extremely high mortality associatedwith hypoplastic left ventricle in our series (88%) highlights the continuing REFERENCES Benson DW. Changing profile of congenital heart disease. Pediatrics crucial importance of full discussion of the treatment 1. 1989;83:790-791. options with parents. If larger studies confirm our find- 2. Clark EB, Takao A. Overview: a focus for research in cardiovascular developings, the definition of complexity we have used could ment. In: Clark EB, Takao A, eds. Developmental Cardiology: Morphogenesis and Function. Mount Kisco, NY: Futura Publishing, 1990:3-12. prove useful in future outcome comparisonsof hospital- 3. Hammon JW, Graham TP, Boucek RJ, Bender HW. Operative repair of coarctation of the aorta in infancy: results with and without ventricular septal ized infants or in regional studies. 1058

THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 68

OCTOBER 15, 1991

defect. Am / Cardiol 1985;55:1555-1559. 4. Qureshi SA, Maruszewski B, McKay R, Arnold R, West CA, Hamilton DI. Determinants of survival following repair of interrupted aortic arch in infancy. Am J Cardiol 1990;26:303-312. 5. Pinsky WW, Arciniegas E. Tetralogy of Fallot. Pediatr Clin North Am 1990;37:179-192. 6. Norwood WI, Pigott JD. Hypoplastic left-sided heart syndrome. In: Grille HC, Austen WC, Wilkins EW, Mathisen DJ, Vlahakes GJ, eds. Current Therapy in Cardiothoracic Surgery. Toronto: BC Decker, 1989:473-480. 7. Kirklin JW, Colvin EV, McConnell ME, Bargeron LM. Complete transposition of the great arteries: treatment in the current era. Pediatr C/in North Am 1990;37:171-177. 8. Greenwood RD, Rosenthal A, Parisi L, Fyler DC, Nadas AS. Extracardiac abnormalities in infants with congenital heart disease. Pediatrics 1975; 55:485-492. 9. Fyler DC, Rothman KJ, Buckley LP, Cohn HE, Hellenbrand WE, Castaneda A. The determinants of five year survival of infants with critical congenital heart disease. Cardiovasc Clin 1981;11:393-405. 10. Ferencz C, Rubin JD, McCarter RJ, Brenner JI, Neil1 CA, Perry LW, Hepner SI, Downing JW. Congenital heart disease: prevalence at livebirth. Am J Epidemiol 1985;121:31-36. 11. Clark EB. Mechanisms in the pathogenesis of congenital cardiac malformations. In: Pierpont ME, Moller JH, eds. Genetics of Cardiovascular Disease. Hingham, MA: Martinus Nijhoff, 1986:3-l 1. 12. Ferencz C, Rubin JD, McCarter RJ, Boughman JA, Wilson PD, Brenner JI, Neil1 CA, Perry LW, Hepner SI, Downing JW. Cardiac and noncardiac malformations: observations in a population-based study. Teratology 1987;35:367-378. 13. Zar JH. Biostatistical Analysis. 2nd ed. Englewood Cliffs, NJ: Prentice-Hall, 1984:64.

14. Weisberg S. Applied Linear Regression. 2nd ed. New York: John Wiley & Sons, 1985:214-215x 15. Kleinbaum DG, Kupper LL, Morgenstern H. Epidemiologic Research: Principles and Quantitative Methods. New York: Van Nostrand Reinhold, 1982: 421-427. 16. Ward KE, Pryor RW, Matson JR, Razook JD, Thompson WM, Elkins RC. Delayed detection of coarctation in infancy: implications for timing of newborn follow-up. Pediatrics 1990;86:972-976. 17. Sondheimer HM, Freedom RM, Olley PM. Double outlet right ventricle: clinical spectrum and prognosis. Am J Cardiol 1977;39:709-714. 18. Roberson DA, Silverman NH. Malaligned outlet septum with subpulmonary ventricular septal defect and abnormal ventriculoarterial connection. J Am Coil Cardiol 1990;16:459-468. 19. Clapp SK, Perry BL, Farooki ZQ, Jackson WL, Karpawich PP, Hakimi M, Arciniegas E, Green EW. Surgical and medical results of complete atrioventricular canal: a ten year review. Am J Cardiol 1987;59:454-458. 20. Whitley R, Arvin A, Prober C, Corey L, Burchett S, Plotkin S, Starr S, Jacobs R, Powell D, Nahmias A, Sumaya C, Edwards K, Alford C, Caddell Ci, Soong S, and the National Institute of Allergy and Infectious Diseases Coilaborative Antiviral Study Group. Predictors of morbidity and mortality in neonates with herpes simplex virus infections. N Engl J Med 1990;324:450-454. 21. Hoffman JIE, Christianson R. Congenital heart disease in a cohort of 19,502 births with long-term follow-up. Am .I Cardiol 1978;42:641-647. 22. Allen HD, Taubert KA, Deckelbaum RJ, Driscoll D, Dunnigan A, Gidding SS, Herndon P, Kavey R, Mullins C, Snider AR, Strong WB, Washington R. Poverty and cardiac disease in children. Am J Dis Child 1991;145:550-553. 23. Moller JH, Hofschire PJ, Bessinger FB Jr, Rittet DG, Lauer RM, Nelson RP. Cardiac disease in infants: changes in mortality, 196991987 (abstr). Circulation 1990;82(supp III):III-154.

OUTCOME IN CONGENITAL HEART DISEASE 1059