Long-Term Survival of Infants with Atrioventricular Septal Defects

Assia Miller, MD, MPH, Csaba Siffel, MD, PhD, Chengxing Lu, PhD, Tiffany Riehle-Colarusso, MD, Jaime L. Frı́as, MD,
and Adolfo Correa, MD, PhD

Objective To examine the variation in survival in infants with atrioventricular septal defects (AVSD) with demo-
graphic factors and clinical characteristics, including the presence of Down syndrome.
Study design We selected infants with all types of AVSD with Down syndrome (n = 177) and without Down syn-
drome (n = 161), born between Jan 1, 1979, and Dec 31, 2003 and identified through the Metropolitan Atlanta Con-
genital Defects Program (MACDP). Infants were classified by the complexity of their cardiac defects and presence
of major non-cardiac malformations. Deaths (n = 111) were identified through 2004 with linkage with state vital re-
cords and the National Death Index. Kaplan-Meier survival probabilities and adjusted hazard ratios (HRs) were cal-
culated in relation to demographic and clinical characteristics.
Results Children with AVSD and Down syndrome had a similar overall survival probability (70%) as those without
Down syndrome (69%). Mortality was higher in children with a complex AVSD (adjusted HR = 7.0; 95% CI, 3.1-15.5)
and in children with $2 major non-cardiac malformations (adjusted HR = 3.4; 95% CI, 1.8-6.5) and was lower in
children in the 1992 to 2003 birth cohort (adjusted HR = 0.6; 95% CI, 0.4-0.998).
Conclusions Down syndrome was not a prognostic factor. Our findings might be helpful in assessing the long-
term prognosis of infants with AVSD. (J Pediatr 2010;156:994-1000).

A
trioventricular septal defects (AVSD) comprise a spectrum of congenital cardiac defects involving incomplete septa-
tion of the atrioventricular canal as a result of impaired fusion of the anterior and posterior endocardial cushions.
AVSDs usually include ostium primum atrial septal defects, inlet ventricular septal defects, and complete AVSD.
In complete AVSD, there is a single, common atrioventricular valve with large anterior and posterior bridging leaflets,
and defects of the atrial (ostium primum) and ventricular septi, allowing communication across all 4 cardiac chambers.1
Infants with AVSD also may have other cardiac and non-cardiac defects, comprising several clinical groups and exhibiting
genetic locus heterogeneity.2-4 In infants with AVSD, Down syndrome is the most frequently associated chromosomal anom-
aly (50%-74%).1,5,6
Without surgery, only 35% to 50% of infants with AVSD survive to 12 months of age.7-8 In recent years, there has been
a major reduction in mortality in children born with AVSD because of advances in surgical repair. The 10-year survival
probability of infants with AVSD increased from 78% during 1975 to 19959 to 86% to 94% during 1992 to 2002.10,11
Although current practice is to treat children with AVSD aggressively, whether or not they also have Down syndrome,12
the effect of Down syndrome and of the severity of heart defects on survival is unclear, because recent survival studies have
been few and inconsistent.6,12 One epidemiologic study reported that infants with AVSD and Down syndrome had higher
overall mortality rate (20%) compared with infants with AVSD without Down syndrome (5%).12 Conversely, other
epidemiologic studies found a higher survival probability for infants with AVSD and Down syndrome compared with
those with AVSD without Down syndrome.6,10,11 Evidence-based counseling on the survival of infants with AVSD would
benefit from more information on the prognostic significance of Down syndrome and other clinical and sociodemographic
factors.
In this study, we examined the long-term survival and identified selected demographic and clinical characteristics as possible
prognostic factors for survival of infants with AVSD, either with or without Down syndrome.

Methods
From the Division of Birth Defects and Developmental
Disabilities, National Center on Birth Defects and
The Metropolitan Atlanta Congenital Defects Program (MACDP) was granted Developmental Disabilities, Centers for Disease Control
authority to conduct birth defect surveillance in the five central counties of met- and Prevention, Atlanta, GA (A.M., C.S., C.L., T.R-C.,
J.F., A.C.); Oak Ridge Institute for Science and
Education, Atlanta, GA (A.M.); Computer Sciences
Corporation, Atlanta, GA (C.S.); Merck Research
Laboratories, Upper Gwynedd, PA (C.L.); and McKing
AVSD Atrioventricular septal defect Consulting Corporation, Fairfax, VA (J.F.)
CHD Congenital heart defect The authors declare no conflicts of interest. The findings
HR Hazard ratio and conclusions in this report are those of the authors
MACDP Metropolitan Atlanta Congenital Defects Program and do not necessarily represent the official position of
the Centers for Disease Control and Prevention.
NDI National Death Index
SES Socioeconomic status 0022-3476/$ - see front matter. Copyright Ó 2010 Mosby Inc.
All rights reserved. 10.1016/j.jpeds.2009.12.013

994
 Vol. 156, No. 6  June 2010

ropolitan Atlanta in collaboration with and on behalf of the the National Death Index (NDI) from Jan 1, 1979, to Dec
Georgia Division of Public Health by the Georgia Department 31, 2004. Some of the very early deaths were not captured
of Human Resources and has CDC’s Institutional Review with the NDI, so the 3 sources complement each other, and
Board approval. We identified infants born during the period we were able to confirm information on deaths matched in
from 1979 to 2003 in whom AVSD was diagnosed from the all sources. Details of the NDI matching process have been
MACDP. This is an ongoing, population-based birth defects described elsewhere.16,17 For infants with AVSD for whom
surveillance system established in 1967 to actively monitor no death records were available, survival was censored at
birth defects in infants born to women residing in the 5 central the end of the follow-up period (Dec 31, 2004).
counties of metropolitan Atlanta, Georgia. The program rou- We assigned a socioeconomic status (SES) indicator to
tinely collects data on clinical information and demographic case mothers on the basis of the percentage of people
characteristics of cases of structural and genetic birth defects living below the federally defined poverty line (as defined
in liveborn infants, stillbirths, and pregnancies terminated in the 1980, 1990, and 2000 US Census data18) in the
at or after 20 weeks of gestation. The defects are coded by us- mother’s census tract at the time of conception and
ing a modified 6-digit code from the International Classifica- according to the Public Health Disparities Geocoding
tion of Diseases, Ninth Revision, Clinical Modification and the Project19: (1) when <5% of people living in a given census
British Paediatric Association Classification of Diseases devel- tract were living in poverty, the tract was considered to
oped for MACDP. Further details about this system have been have a high SES; (2) when 5.0% to 9.9% were living in
published elsewhere.13 poverty, the tract was considered to have a middle to
All MACDP records of infants with cardiac defects high SES; (3) when 10.0% to 19.9% were living in poverty,
(745.000-747.999) were reviewed and classified by experts the tract was considered to have a middle to low SES; and
in pediatric cardiology, according to a standard clinical no- (4) when $20% were living in poverty, the tract was
menclature adapted from the Society of Thoracic Surgeons considered to have a low SES.
and a morphogenetic classification system described previ- For the cohorts of case infants with AVSD alone and for
ously.14 For this study, all records with at least 1 of 8 AVSD the cohorts of infants with Down syndrome, we estimated
subtypes were selected and then grouped by complexity of their survival probability from 1 to 5 years of age, from 5
defect. ‘‘Partial’’ defects were inlet ventricular septal defects, to 10 years of age, and from 10 to 25 years of age, and during
primum atrial septal defects, and transitional AVSD. ‘‘Com- the overall study period by using the Kaplan–Meier product-
plete’’ defects were complete and unspecified types of AVSD. limit method.20 We estimated Kaplan-Meier survival curves
‘‘Complex’’ defects were AVSD with unequal left or right ven- for children born during the periods 1979 to 1991 and
tricular sizes (unbalanced AVSD) and AVSD associated with 1992 to 2003, and for the entire 25-year period. We also esti-
tetralogy of Fallot or any other complex congenital heart mated the Kaplan-Meier survival probability stratified with
defect (CHD). ‘‘Heterotaxy’’ was AVSD associated with later- demographic characteristics, including maternal race (white,
ality defects. In our study, heterotaxy was present only in in- black or African-American, and other), maternal age, pater-
fants without Down syndrome; therefore, infants with nal age, sex, SES (high, mid-high, mid-low, and low), and
heterotaxy were analyzed separately. clinical covariates such as presence of Down syndrome, pres-
We used the MACDP codes 758.000 to 758.099 to identify ence of heterotaxy syndrome, birthweight (<2500 g and
infants with Down syndrome and reviewed cytogenetic data $2500 g), preterm birth (<37 weeks and $37 weeks gesta-
on all liveborn infants with Down syndrome. Only those in- tion), multiple defects (absent, 1 additional non-cardiac
fants with AVSD and a cytogenetically confirmed diagnosis of major defect, and $2 non-cardiac major defects), and com-
trisomy 21 were analyzed, including infants with full trisomy plexity of AVSD (partial, complete, complex, and hetero-
(n = 175) and Robertsonian translocations (n = 2). The group taxy). Cutoff points for maternal age (<29 years and $29
of infants with AVSD without Down syndrome included all years) and paternal age (<31 years and $31 years) were based
other infants except those with chromosomal abnormalities on mean values of maternal and paternal age. Greenwood’s
other than Down syndrome (eg, trisomy 13 or trisomy 18). method was used to obtain the 95% CIs for the estimates
Infants were classified by a clinical geneticist (J.F.) on the of the survival probability,21 and the log-rank test was used
basis of the number of major structural non-cardiac congen- to examine the variations in survival probabilities by period
ital defects or the presence of laterality defects (heterotaxy). of birth and possible risk factors. We estimated hazard ratios
Non-cardiac defects were defined as ‘‘major’’ when they (HRs) and 95% CIs in relation to possible risk factors, taking
had surgical, medical, or serious cosmetic importance.15 into account other covariates with Cox proportional hazards
We studied infants born from Jan 1, 1979, through Dec 31, regression models.22, 23 We checked the assumption of pro-
2003, and observed them through Dec 31, 2004. Deaths and portionality by plotting estimated log-cumulative hazard
death dates in liveborn infants with AVSD initially were iden- versus log of survival time to examine when they are parallel
tified by using MACDP hospital records and vital records for different categories of the risk factors and by testing pos-
from the state of Georgia. To obtain information on deaths sible time-dependent trends. We examined possible interac-
and death dates not captured with these two sources, such tions between risk factors and variables associated with
as deaths that occurred outside the state of Georgia, all P values <.05 in the univariate analysis in the final Cox
MACDP liveborn infants with AVSD were also linked with proportional hazard model.
995
THE JOURNAL OF PEDIATRICS  www.jpeds.com Vol. 156, No. 6

without Down syndrome mainly had laterality defects

(23.0%), musculoskeletal malformations (13.0%), defects
of the renal (11.2%), digestive (6.8%), respiratory (6.2%)
systems, and facial clefts (5.0%).
During the study period, the overall prevalence of all types
of AVSD was 3.7 per 10 000 livebirths. The prevalence of
AVSD with Down syndrome was 1.9 per 10 000 livebirths.
The prevalence of AVSD without Down syndrome was 1.7
per 10 000 livebirths.
During the follow-up period, we identified 111 deaths in
Figure 1. Distribution of complexity of AVSD in infants by the 338 case infants with AVSD, for a 33% overall case fatality
presence of Down syndrome, Atlanta, Georgia, 1979 to 2003. rate. The frequency distribution of causes of death for infants
with AVSD without Down syndrome (congenital cardiac mal-
We used SAS-PC software (SAS Inc., Cary, North Caro- formations, 49.6%; respiratory system malformations, 3.6%;
lina) for computation23 and S-PLUS software (S-PLUS, Seat- gastrointestinal system malformations, 4.5%; central nervous
tle, Washington) for plotting survival curves.24 system malformations, 0.9%; and other causes, 4.5%) was sig-
nificantly different from that for infants with AVSD and
Down syndrome (CHD, 19.8%; respiratory system malforma-
Results tions, 2.7%; gastrointestinal system malformations, 1.8%; in-
fections, 4.5%; central nervous system malformations, 1.8%;
We identified 338 infants with AVSD born between Jan 31, infection, 4.5%; Down syndrome, 1.8%; and other causes,
1979, and Dec 31, 2003, of whom 177 (52.4%) had Down 4.5%; c2 test, P = .0072).
syndrome and 165 (48.8%) had complete AVSD. The com- The overall survival probability of infants with AVSD and
plexity pattern of AVSD varied between infants with Down Down syndrome was 66%, compared with 47% for infants
Syndrome and infants without Down syndrome (Figure 1). with AVSD alone. Specifically, for those with Down syn-
The partial subgroup was represented similarly in the 2 infant drome, the survival probabilities at 1, 5, and 10 years of age
groups. However, the complete subgroup was predominant were 84.2% (95% CI, 77.9%-88.8%), 81.1% (95% CI,
in infants with Down syndrome, and the complex subgroup 74.5%-86.2%), and 78.0% (95% CI, 70.8%-83.6%), respec-
and heterotaxy syndrome were prevalent in infants without tively. For children without Down syndrome, the corre-
Down syndrome. sponding survival probabilities were 69.6% (95% CI,
Infants with AVSD and Down syndrome had different 61.8%-76.0%), 60.4% (95% CI, 52.3%-67.5%), and 57.9%
types of associated major non-cardiac malformations com- (95% CI, 49.7%-65.3%), respectively.
pared with infants who did not have Down syndrome, in- Figure 2 shows the survival probability of infants with dif-
cluding musculoskeletal malformations (24.9%), eye and ferent subtypes of AVSD stratified by the presence of Down
ear (9.6%), central nervous system (9.0%), renal (5.1%), syndrome. For infants with AVSD and Down syndrome
and digestive system (4.5%) defects. Infants with AVSDs and infants with AVSD without Down syndrome, infants

Figure 2. Survival probability of infants with AVSD. A, Infants with Down syndrome and B, Infants without Down syndrome by
complexity of heart defects, Atlanta, Georgia, 1979 to 2003.

996 Miller et al
June 2010 ORIGINAL ARTICLES

Table I. Survival probabilities for infants with atrioventricular septal defects by demographic and clinical characteristics,
Atlanta, Georgia, born 1979 to 2003
Characteristics Births, n Deaths, n (%) Overall survival, % (95% CI) Log-rank test P value
Birth period
1979-1991 140 58 (41.43) 55.56 (46.05-64.06) .2168
1992-2003 198 53 (26.77) 72.61 (65.68-78.38)
Sex
Male 151 53 (35.10) 60.15 (48.46-70.00) .4911
Female 187 58 (31.02) 58.33 (46.92-68.12)
Birthweight, g
<2500 80 30 (37.50) 52.62 (30.42-70.72) .1974
$2500 258 81 (31.40) 61.25 (52.61-68.79)
Gestational age, weeks .0046
Preterm (<37) 71 32 (45.07) 46.08 (26.52-63.62)
Term ($37) 258 76 (29.46) 63.46 (54.75-70.94)
Race
White (not Hispanic) 169 51 (30.18) 64.09 (53.80-72.67)
Black or African-American (not Hispanic) 143 47 (32.87) 57.27 (43.73-68.69) .0079
Others 25 13 (52.00) 45.25 (24.38-64.05)
Maternal age, years
<29 167 60 (35.93) 54.69 (42.96-64.98) .3802
>29 170 51 (30.00) 65.13 (54.89-73.60)
Paternal age, years
<31 80 25 (31.25) 63.44 (47.76-75.56) .9321
>31 100 31 (31.00) 61.87 (46.95-73.71)
SES
0.0%-4.9% 108 34 (31.48) 62.25 (48.89-73.05)
5.0%-9.9% 86 28 (32.56) 60.41 (44.01-73.38) .5059
10.0%-19.9% 58 23 (39.66) 57.90 (43.24-70.03)
$20% 79 25 (31.65) 56.88 (38.63-71.56)
Down syndrome*
Absent 111 28 (25.23) 70.34 (56.28-80.62) .6156
Present 177 41 (23.16) 69.54 (58.14-78.40)
Complexity of AVSD
Partial 92 10 (10.87) 90.20 (82.02-94.78)
Complete 165 43 (26.06) 67.22 (55.96-76.19) <.0001
Complex 31 16 (51.61) 47.64 (29.23-63.96)
Heterotaxy 50 42 (84.00) 22.36 (09.58-38.73)
Multiple non-cardiac defects* <.0004
Absent 226 46 (20.35) 74.44 (64.46-82.01)
1 major defect 36 11 (30.56) 65.71 (42.69-81.29)
$2 major defects 25 12 (48.00) 52.00 (31.25-69.24)
*Case infants with heterotaxy excluded from analysis.

with the less complex type of AVSD had better survival rates exclusion of heterotaxy case infants, the complexity of
than infants with complex types. The risk of mortality was AVSD and the presence of multiple non-cardiac malforma-
6-fold higher in infants with heterotaxy and AVSD than in tions showed elevated adjusted HR, and the birth period
infants with AVSD who did not have heterotaxy (adjusted showed decreased adjusted HR (Table II). There appeared
HR = 6.20; 95% CI, 4.2-9.2). The presence of Down syn-
drome was not a significant predictor of survival (P = .6156).
Other factors that appeared to be associated with the sur- Table II. Adjusted hazard ratios and 95% CI for infants
vival of infants with AVSD were maternal race or ethnicity, with atrioventricular septal defects, Atlanta, Georgia,
gestational age, the complexity of the AVSD, and the presence born 1979 to 2003*
of associated major non-cardiac malformations (Table I). Adjusted full Adjusted final
Characteristics HR (95% CI) HR (95% CI)
However, after excluding case infants with heterotaxy, in uni-
variate analysis only the complexity of the AVSD and the Birth period (1992-2003 vs1979-1991) 0.59 (0.3-0.98) 0.60 (0.4-0.998)
Preterm (<37 vs $37) 1.65 (0.96-2.8)
presence of associated major non-cardiac malformations Race (black or African-American 0.87 (0.5-1.5)
seemed to have affected significantly the survival of infants vs white)
with AVSD. We examined statistically significant variables Race (others vs white) 2.35 (0.95-5.8)
Presence of Down syndrome 0.87 (0.5-1.5)
(P < .05) in the univariate analysis (heart complexity and Complete AVSD vs partial AVSD 2.34 (1.1-4.8) 2.24 (1.1-4.5)
presence of multiple defects) and variables of interest (birth Complex AVSD vs partial AVSD 7.11 (3.2-15.9) 6.97 (3.1-15.5)
period, maternal race or ethnicity, gestational age, and pres- 1 other defect vs no other defect 1.28 (0.6-2.5) 1.26 (0.6-2.5)
2 other defects vs no other defect 3.32 (1.7-6.3) 3.42 (1.8-6.5)
ence of Down syndrome) in the multivariate full Cox propor-
tional hazard model. In the multivariate model, with the *Case infants with heterotaxy excluded from analysis.

Long-Term Survival of Infants with Atrioventricular Septal Defects 997

THE JOURNAL OF PEDIATRICS  www.jpeds.com Vol. 156, No. 6

to be a decrease in survival rate with the increasing complex- risk would presumably be the same for all infants with
ity of the AVSD and with the increasing number of non- Down syndrome and cardiac defects; however, the study
cardiac defects. For Down syndrome versus non-Down syn- found mortality for infants with Down syndrome and other
drome, the 95% CI for the overall adjusted HR of AVSD cardiac lesions was comparable to that for infants without
with heterotaxy included the null value (adjusted HR = Down syndrome. The mortality difference was primarily in
0.79; 95% CI, 0.5-1.3); the same was true for AVSD without earlier years, perhaps when infants with AVSD and Down syn-
heterotaxy (adjusted HR = 0.87; 95% CI, 0.5-1.5), for partial drome did not receive adequate cardiac treatment.29 Addi-
subtype (adjusted HR = 1.9; 95% CI, 0.5-8.0), complete sub- tionally, in this study the presence of $2 major non-cardiac
type (adjusted HR = 0.7; 95% CI, 0.4-1.4), and severe subtype defects was significantly associated with poorer survival. These
of AVSD (adjusted HR = 1.3; 95% CI, 0.4-4.0). findings were consistent with those of earlier reports.1,30,31
The survival of children with AVSD was no longer influ- The presence of heterotaxy was associated significantly with
enced by the presence of Down syndrome or demographic poor survival rate (22.4%) in our study and contributed to the
characteristics. We did not detect statistically significant decreased survival rate of infants with AVSD without Down
2-way interactions (P < .05) among any of these variables: syndrome. This result was consistent with those of earlier
birth period, gestational age, maternal race or ethnicity, pres- publications reporting common atrioventricular canal as 1
ence of Down syndrome, complexity of AVSD, and presence of of the independent predictors of mortality in children with
multiple defects. We did evaluate the interactions in the sub- heterotaxy.2,32,33 Although a few studies have shown the im-
groups with and those without exclusion of heterotaxy case in- proving surgical outcomes of using the Fontan operation in
fants and the results did not change. In infants with heterotaxy, children with heterotaxy syndromes,34-36 our results suggested
the only significant predictor of mortality was preterm birth; that mortality did not improve with time and remained high
the risk was almost 3-fold higher in preterm infants (adjusted in these patients. Multivariate Cox proportional hazards re-
HR = 2.8; 95% CI, 1.4-5.6) than in term infants. gression analysis in cases with AVSD and heterotaxy showed
that only preterm birth was associated significantly with in-
creased mortality. Because mortality was highest for this
Discussion group, it is possible that the main determinant of mortality
was the heterotaxy syndrome itself, perhaps because of the
Our findings highlight significant clinical prognostic factors complexity of the syndrome and multi-organ effects. More
for survival, namely the complexity of AVSD and presence studies with larger samples of case infants are warranted to ex-
of associated major non-cardiac defects. This study uniquely amine any possible variations in survival with sociodemo-
distinguishes between less and more complex subtypes of graphic and other characteristics. Several authors have
AVSD, rather than analyzing them as an entire group or identified preterm birth as a risk factor for poor survival1,37-39
only focusing on complete AVSD. The differences in AVSD and have shown that the mortality rate for cardiac surgery
complexity between infants with Down syndrome and infants was twice as high in preterm infants compared with term
without Down syndrome were potentiallly the result of ge- infants.40
netic heterogeneity1-4,25,26 and were reflected in the survival Earlier studies have demonstrated a significant improve-
analyses. Infants with Down syndrome had a significantly ment in the survival of infants with AVSD with time and
higher survival probability (69.5%) than infants without for infants with AVSD and Down syndrome.6,27,41,42 Our
Down syndrome (48%). However, after excluding infants findings were consistent with those of earlier publications
with heterotaxy, the survival probability of infants with and and, after accounting for heterotaxy, showed a moderate
of infants without Down syndrome became similar (70.3% but statistically significant increase in the survival rate for in-
and 69.5%, respectively). These findings also corresponded fants with AVSD with time. By the mid-1980s, in most med-
with those of earlier publications on the survival of infants ical centers in the United States the surgical repair rate for
with Down syndrome and cardiac malformations1,6,27 and infants with AVSD and Down syndrome was comparable
supported results of recent studies on the postoperative prog- with that of infants with AVSD without Down syndrome.12
nosis of infants with AVSD and Down syndrome. Several au- Thus, our ability to observe an improvement in survival
thors have noted that, in infants with AVSD, the surgical with time during recent years suggests that better care has re-
prognosis of those with Down syndrome was better compared sulted in an increased survival of infants with AVSD.
with those without Down syndrome.9,27,28 Therefore, the In univariate analysis, white infants had a higher survival
presence of Down syndrome does not seem to deteriorate probability than did black or African-American infants and
long-term results when the defect is repaired during the first infants with race designated as ‘‘other’’; however, controlling
year of life.11 In contrast, only one population-based study12 for the presence of Down syndrome, the presence of multiple
of a cohort from the period 1958 to 1997 reported that infants non-cardiac malformations and the complexity of heart de-
with complete AVSD and Down syndrome actually had fects attenuated the race and ethnic variation in survival.
a higher mortality rate (20%) than did infants without Some authors have reported a racial or ethnic disparity in
Down syndrome (5%; P =0.04) and attributed the differences survival of infants with AVSD,43 whereas other authors
in mortality to a presumed higher risk of pulmonary hyper- have not found such variations.1 The extent to which these
tension in infants with Down syndrome. This attributable differences in the reported association of survival with race
998 Miller et al
June 2010 ORIGINAL ARTICLES

or ethnicity reflect differences in the control of potential con- We thank the Metropolitan Atlanta Congenital Defects Program ab-
founders remains to be determined. CHD was the most fre- stractors for their conscientious and skilled data collection efforts,
and Don Gambrell (Computer Sciences Corporation; contractor for
quently reported cause of death in infants with AVSD with
CDC) for his technical assistance with data linkage.
Down syndrome and infants with AVSD without Down syn-
drome, which is consistent with earlier observations.44 Submitted for publication Jun 17, 2009; last revision received Oct 26, 2009;
This study had several strengths. First, it was population- accepted Dec 8, 2009.
based, including liveborn infants from a defined popula- Reprint requests: Assia Miller, MD, MPH, National Center on Birth Defects and
tion, and ascertained by a well-established birth defects Developmental Disabilities, Centers for Disease Control and Prevention,
Mailstop E-86, 1600 Clifton Road, Atlanta, GA 30333. E-mail: [email protected].
surveillance system with active and multiple source
ascertainment methods.13 Second, the 25-year follow-up al-
lowed for estimation of an overall survival probability of
AVSD through 24 years of age, which is longer than has References
been previously possible. The follow-up period was not
1. Ferenz C, Loffredo C, Correa-Villasenor A, Wilson PD. Genetic and
the same for all infants, being more limited for infants
environmental risk factors of major cardiovascular malformations.
born in more recent years. This limited our ability to exam- The Baltimore-Washington Infant Study 1981-1989. Armonk, NY: Fu-
ine changes in survival beyond infancy with time. However, tura Publishing Company; 1997.
determination of vital status was based on vital records and 2. Marino B, Digilio MC. Congenital heart disease and genetic syndromes:
NDI linkages, which previously have been shown to be specific correlation between cardiac phenotype and genotype. Cardiovas
Pathol 2000;9:303-15.
accurate in identifying a high proportion of deaths.45,46
3. Maslen CL. Molecular genetics of atrioventricular septal defects. Curr
Third, the availability of cytogenetic data allowed us to Opin Cardiol 2004;19:205-10.
exclude cases without a cytogenetic diagnosis of trisomy 4. Maslen CL, Babcock D, Robinson SW, Bean LJ, Dooley KJ, Willour VL,
21. Fourth, a clinical, standard nomenclature used to clas- et al. CRELD1 mutations contribute to the occurrence of cardiac atrio-
sify heart defects enabled us to examine the complexity of ventricular septal defects in Down syndrome. Am J Med Genet Part A
AVSD as a prognostic factor.14 Finally, we were able to ex- 2006;140:2501-5.
5. Craig B. Atrioventricular septal defect: from fetus to adult. Heart 2006;
amine a wide range of potential demographic and clinical 92:1879-85.
predictors of survival. 6. Frid C, Björkhem G, Jonzon A, Sunnegårdh J, Annerén G, Lundell B.
Our study had some limitations. We did not trace people Long-term survival in children with atrioventricular septal defect and
with AVSD individually through the end of the follow-up pe- common atrioventricular valvar orifice in Sweden. Cardiol Young
riod to determine whether they were alive. Instead, people 2004;14:24-31.
7. Berger TJ, Blackstone EH, Kirklin JW, Bargeron LM, Hazelrig JB,
with AVSD for whom no dates of death were available in Turner ME. Survival and probability of cure without and with operation
our 3 sources of vital status were assumed to be alive at the in complete atrioventricular canal. Ann Thorac Surg 1978;27:104-11.
end of the follow-up period. Although it is possible that 8. Samanek M. Children with congenital heart disease: probability of nat-
some deaths might have been missed because of missing ural survival. Pediatr Cardiol 1992;13:152-8.
data on a linkage variable (eg, month of birth), it is likely 9. Boening A, Scheewe J, Heine K, Hedderich J, Regensburger D,
Kramer HH, et al. Long-term results after surgical correction of atrio-
that the number of missed deaths was small, because the ventricular septal defects. Eur J Cardiohorac Surg 2002;22:167-73.
NDI uses several factors for linkage.17 By limiting our study 10. Formigari R, Di Donato RM, Gargiulo G, Di Carlo D, Feltri C,
to cytogenetically confirmed Down syndrome, we might Picchio FM, et al. Better surgical prognosis for patients with complete
have missed infants who died before cytogenetic specimens atrioventricular septal defect and Down’s syndrome. Ann Thorac Surg
2004;78:666-72.
were obtained or whose specimens were sent to a laboratory
11. Masuda M, Kado H, Tanoue Y, Fukae K, Onzuka T, Shiokawa Y, et al.
to which MACDP did not have access. Cytogenetic confirma- Does Down syndrome affect the long-term results of complete atrioven-
tion was available less frequently during the earliest period tricular septal defect when the defect is repaired during the first year of
(1979-1991) than during the more recent periods; however, life? Eur J Cardiothorac Surg 2005;27:405-9.
the total number of infants for whom cytogenetic data were 12. Reller MD, Morris CD. Is Down syndrome a risk factor for poor out-
unavailable was small (n = 2). We used an SES indicator come after repair of congenital heart defects? J Pediatr 1998;132:738-41.
13. Correa A, Cragan JD, Kucik ME, Alverson CJ, Gilboa SM,
on the basis of the mother’s census tract at the time of deliv- Balakrishnan R, et al. Reporting birth defects surveillance data 1968-
ery as a proxy for SES; we did not have information on direct 2003. Birth Defects Res Part A Clin Mol Teratol 2007;79:65-186.
family income or source of payment for medical services. We 14. Riehle-Colarusso T, Strickland MJ, Reller MD, Mahle WT, Botto LD,
also did not have information on such possible demographic Siffel C, et al. Improving the quality of surveillance data on congenital
and clinical covariates of survival as maternal and paternal heart defects in the metropolitan Atlanta congenital defects program.
Birth Defects Res Part A Clin Mol Teratol 2007;79:743-53.
education and occupation, environmental factors, and type 15. Rasmussen SA, Olney RS, Holmes LB, Lin AE, Keppler-Noreuil KM,
of cardiac surgery. Moore CA. Guidelines for case classification for the National Birth De-
Further work is warranted to elucidate the possible reasons fects Prevention Study. Birth Defects Res Part A Clin Mol Teratol 2003;
for the variations in survival of infants with AVSD by type of 67:193-201.
cardiac and major non-cardiac defects. Our findings suggest 16. Center for Disease Control and Prevention, National Center for Health
Statistics. National Death Index user manual. Hyattsville, MD: National
that future studies of outcomes in infants with AVSD might Center for Health Statistics; 2000.
focus on morbidity associated with late cardiac complica- 17. Wentworth DN, Neaton JD, Rasmussen WL. An evaluation of the Social
tions and other major unrelated defects. n Security Administration master beneficiary record file and the National

Long-Term Survival of Infants with Atrioventricular Septal Defects 999

THE JOURNAL OF PEDIATRICS  www.jpeds.com Vol. 156, No. 6

Death Index in the ascertainment of vital status. Am J Public Health 32. Foerster SR, Gauvreau K, McElhinney DB, Geva T. Importance of totally
1983;73:1270-4. anomalous pulmonary venous connection and postoperative pulmonary
18. US Census Bureau. Census 2000 summary file 2—technical documenta- vein stenosis in outcomes of heterotaxy syndrome. Pediatr Cardiol 2008;
tion, prepared by the U.S. Census Bureau. Washington, DC: Department 29:536-44.
of Commerce; 2001. 33. Phoon CK, Neill CA. Asplenia syndrome—risk factors for early unfavor-
19. Krieger N, Chen JT, Waterman PD, Soobader MJ, Subramanian SV, able outcome. Am J Cardiol 1994;73:1235-7.
Carson R. Choosing area based socioeconomic measures to monitor so- 34. Stamm C, Friehs I, Duebener LF, Zurakowski D, Mayer JE, Jonas RA,
cial inequalities in low birth weight and childhood lead poisoning: The et al. Improving results of the modified Fontan operation in patients
Public Health Disparities Geocoding Project (US). J Epidemiol Commu- with heterotaxy syndrome. Ann Thorac Surg 2002;74:1967-78.
nity Health 2003;57:186-99. 35. Azakie A, Merklinger SL, Williams WG, Van Arsdell GS, Coles JG,
20. Kaplan EL, Meier P. Nonparametric estimation from incomplete obser- Adatia I. Improving outcomes of the Fontan operation in children
vations. J Am Stat Assoc 1958;53:457-81. with atrial isomerism and heterotaxy syndromes. Ann Thorac Surg
21. Greenwood GM. The natural duration of cancer. Reports on the Public 2001;72:1636-40.
Health and Medical Subjects. Rep Public Health Med Subj (Lond) 1926; 36. Morales DL, Braud BE, Booth JH, Graves DE, Heinle JS, McKenzie ED,
33:1-26. et al. Heterotaxy patients with total anomalous pulmonary venous
22. Cox DR, Oakes D. Analysis of survival data. London, England: Chapman return: improving surgical results. Ann Thorac Surg 2006;82:1621-7.
& Hall; 1984. 37. Alexander GR, Kogan M, Bader D, Carlo W, Allen M, Mor J. US birth
23. SAS Institute, 2002-2004. Cary, NC. Available at http://www.sas.com/ weight/gestational age-specific neonatal mortality: 1995-1997 rates for
technologies/analytics/statistics/stat/index.html; Accessed at January whites, Hispanics, and blacks. Pediatrics 2003;111:6.
12, 2010. 38. Kramer MS, Demissie K, Yang H, Platt RW, Sauvé R, Liston R. The con-
24. S-PLUS 6.2. Seattle, WA Insightful Corporation, 2003. Available at tribution of mild and moderate preterm birth to infant mortality. Fetal
http://spotfire.tibco.com/Products/S-Plus-Overview.aspx; Accessed at and Infant Health Study Group of the Canadian Perinatal Surveillance
January 12, 2010. System. JAMA 2000;284:843-9.
25. Digilio MC, Marino B, Toscano A, Giannotti A, Dallapiccola B. Atrio- 39. Tomashek KM, Shapiro-Mendoza CK, Davidoff MJ, Petrini JR. Differ-
ventricular canal defect without Down syndrome: a heterogeneous mal- ences in mortality between late-preterm and term singleton infants in
formation. Am J Med Genet 1999;85:140-6. the United States, 1995-2002. J Pediatr 2007;151:450-6, 6.e1.
26. Robinson SW, Morris CD, Goldmuntz E, Reller MD, Jones MA, 40. Tanner K, Sabrine N, Wren C. Cardiovascular malformations among
Steiner RD, et al. Missense mutations in CRELD1 are associated with preterm infants. Pediatrics 2005;116:e833-8.
cardiac atrioventricular septal defects. Am J Hum Genet 2003;72: 41. Yang Q, Rasmussen SA, Friedman JM. Mortality associated with Down’s
1047-52. syndrome in the USA from 1983 to 1997: a population-based study. Lan-
27. Marino B, Diociaiuti L, Calcagni G, Mastroiacovo P. Outcome in Down cet 2002;359:1019-25.
syndrome fetuses with cardiac anomalies. Am J Med Genet Part A 2004; 42. Leonard S, Bower C, Petterson B, Leonard H. Survival of infants born with
128:101-3. Down’s syndrome: 1980-96. Paediatr Perinat Epidemiol 2000;14:163-71.
28. Al-Hay AA, MacNeill SJ, Yacoub M, Shore DF, Shinebourne EA. Com- 43. Correa-Villaseñor A, McCarter R, Downing J, Ferencz C. White-black
plete atrioventricular septal defect, Down syndrome, and surgical out- differences in cardiovascular malformations in infancy and socioeco-
come: risk factors. Ann Thorac Surg 2003;75:412-21. nomic factors. The Baltimore-Washington Infant Study Group. Am J
29. Sondheimer HM, Byrum CJ, Blackman MS. Unequal cardiac care for Epidemiol 1991;134:393-402.
children with Down’s syndrome. Am J Dis Child 1985;139:68-70. 44. Shin M, Kucik JE, Correa A. Causes of death and case fatality rates
30. Agha MM, Williams JI, Marrett L, To T, Dodds L. Determinants of sur- among infants with down syndrome in metropolitan Atlanta. Birth
vival in children with congenital abnormalities: a long-term population- Defects Res Part A Clin Mol Teratol 2007;79:775-80.
based cohort study. Birth Defects Res Part A Clin Mol Teratol 2006;76: 45. Lash TL, Silliman RA. A comparison of the National Death Index and
46-54. Social Security Administration databases to ascertain vital status. Epide-
31. Cleves MA, Ghaffar S, Zhao W, Mosley BS, Hobbs CA. First-year survival miology 2001;12:259-61.
of infants born with congenital heart defects in Arkansas (1993-1998): 46. Rich-Edwards JW, Corsano KA, Stampfer MJ. Test of the National Death
a survival analysis using registry data. Birth Defects Res Part A Clin Index and Equifax Nationwide Death Search. Am J Epidemiol 1994;140:
Mol Teratol 2003;67:662-8. 1016-9.

1000 Miller et al