RESEARCH ARTICLE

Maternal Age and Prevalence of Isolated Congenital

Heart Defects in an Urban Area of the United States
Assia Miller,1,2 Tiffany Riehle-Colarusso,1 Csaba Siffel,1,3 Jaime L. Fr
as,1,4 and Adolfo Correa1*
1
Division of Birth Defects and Developmental Disabilities, National Center on Birth Defects and Developmental Disabilities, Centers for Disease
Control and Prevention, Atlanta, Georgia
2
Research Triangle Institute International, Atlanta, Georgia
3
Computer Sciences Corporation, Atlanta, Georgia
4
McKing Consulting Corporation, Fairfax, Virginia

Received 2 September 2010; Accepted 1 May 2011

Although maternal age has been associated with a number of

birth defects in several reports, the literature on the association
of maternal age with isolated congenital heart defect (CHD)
phenotypes has been limited. We evaluated CHD prevalence
based on a cohort of 5,289 infants and fetuses with isolated
CHDs born during the period 19682005 and ascertained by the
Metropolitan Atlanta Congenital Defects Program (MACDP)
among residents of ve central counties in Atlanta. For our
denominator, we obtained information on births to residents
of the same counties from vital records (n 1,301,143). We
calculated prevalence ratios for 23 CHD phenotypes by several
maternal age categories, using the group 2529 years of age as a
reference group. We used Poisson regression models to estimate
adjusted prevalence ratios (aPRs) and 95% condence intervals
(CIs), controlling for maternal race, infant sex, and birth cohort.
A maternal age of 35 years or older was associated with an
increased prevalence for several CHD phenotypes: laterality
defects (aPR 2.06; CI 1.223.48), all conotruncal defects
(aPR 1.30; CI 1.031.65), and specically for dextro-trans-
position of the great arteries (aPR 1.65; CI 1.102.48), coarc-
tation of the aorta (aPR 1.54; CI 1.102.16), ventricular septal
defects (aPR 1.20; CI 1.061.36), and atrial septal defects
(aPR 1.36; CI 1.051.77). Our ndings suggest that the birth
prevalence of specic isolated CHDs varies with maternal age.
Further studies are warranted to corroborate these observations,
taking into account potential confounding by known modiable
risk factors. Published 2011 Wiley-Liss, Inc.
Key words: congenital heart defects; maternal age; epidemiology
of congenital heart defects
INTRODUCTION
Congenital heart defects (CHDs) are the most common type of
birth defects, occurring in approximately 39 of every 1,000 live
births [Wilson et al., 1993; Wren et al., 2000; Botto et al., 2001;
Dastgiri et al., 2002; Pradat et al., 2003; Correa et al., 2007; Reller
et al., 2008], and account for a signicant proportion of infant
mortality due to birth defects [Yang et al., 1997; Lee et al., 2001].
How to Cite this Article:
Miller A, Riehle-Colarusso T, Siffel C, Fr
as JL,
Correa A. 2011. Maternal age and prevalence
of isolated congenital heart defects in an
urban area of the United States.
Am J Med Genet Part A 9999:19.
Most CHDs have a multifactorial etiology, involving interactions
between genetic and nongenetic factors. Previous studies have
identied possible risk factors, including maternal pregestational
diabetes and rubella [Jenkins et al., 2007], gestational age
[Rasmussen et al., 2001; Tanner et al., 2005] and intrauterine
growth restriction [Malik et al., 2007], infant sex [Rothman and
Fyler, 1976; Pradat, 1992; OMalley et al., 1996; Ferencz et al., 1997;
Pradat et al., 2003], and maternal race [Francannet et al., 1993;
Ferencz et al., 1997]. In addition some investigators have shown an
increased prevalence of CHD among children born to women
35 years and older [Rothman and Fyler, 1976; Croen and Shaw,
1995; Hollier et al., 2000; Pradat et al., 2003; Reefhuis and Honein,
2004; Garg et al., 2008]. Evaluating the effect of maternal age on
CHD prevalence is of particular importance because during the past
decade the proportion of women in the United States who gave
Additional supporting information may be found in the online version of
this article.
The ndings and conclusions in this report are those of the authors and do
not necessarily represent the ofcial position of the Centers for Disease
Control and Prevention.
*Correspondence to:
Adolfo Correa, M.D., Ph.D., National Center on Birth Defects and
Developmental Disabilities, Centers for Disease Control and Prevention,
Mailstop E-86, 1600 Clifton Road, Atlanta, GA 30333
E-mail: acorrea@cdc.gov
Published online 00 Month 2011 in Wiley Online Library
(wileyonlinelibrary.com).
DOI 10.1002/ajmg.a.34130

Published 2011 Wiley-Liss, Inc.

This article is a US Government work and, as such, is in the public domain in the United States of America.

1
2 AMERICAN JOURNAL OF MEDICAL GENETICS PART A
birth after age 35 increased signicantly [Hamilton and Ventura,
2006].
However, there has been little consistency in the literature
regarding the strength and direction of the association with mater-
nal age [Hollier et al., 2000; Cedergren et al., 2002; Reefhuis and
Honein, 2004; Amorim et al., 2008; Garg et al., 2008] (see Supple-
mentary Appendix). While some authors have reported a U-shaped
effect [Cedergren et al., 2002; Reefhuis and Honein, 2004], others
have described a linear relationship with increasing strength of the
association with increasing maternal age [Hollier et al., 2000;
Amorim et al., 2008; Garg et al., 2008]. Furthermore, information
on the variation in prevalence of specic CHD phenotypes with
maternal age is limited by the inconsistency in the type of CHDs
examined and the mixed results observed across studies. For
example, while one investigator [Pradat et al., 2003] found a
signicant association between advanced maternal age and tetral-
ogy of Fallot and atrioventricular septal defects (AVSDs), other
investigators did not [Baird et al., 1991; OMalley et al., 1996; Loane
et al., 2009]. Furthermore, very few studies have been population-
based and have accounted for possible effect modiers such as low
birth weight (LBW), preterm delivery, maternal race, and infant sex.
In a previous analysis Reefhuis and Honein [2004] reported sig-
nicant associations between maternal age 3540 and tricuspid
atresia and right ow tract defects, but, even though they excluded
case infants with chromosomal abnormalities, they did not exclude
those with nonchromosomal syndromes or with nonsyndromic
associated extra-cardiac anomalies.
Our study expands on previous work by utilizing an extended
birth cohort and a current, standard CHD classication system, and
by examining isolated CHDs without associated major noncardiac
malformations. In this study, we used population-based surveil-
lance data to: (1) examine the effect of increasing maternal age on
the prevalence for several isolated (nonsyndromic) CHD pheno-
types, while accounting for selected maternal and infant character-
istics; and (2) explore potential effect modication of the maternal
age group-specic CHD prevalence estimates by preterm birth,
LBW, maternal race, and infant sex.
METHODS
Study Population
We identied singleton live born and stillborn infants and fetuses
diagnosed with CHDs during the period January 1, 1968
December 31, 2005 using data from the Metropolitan Atlanta
Congenital Defects Program (MACDP). MACDP is an ongoing
population-based birth defects surveillance system established in
1968 to actively ascertain birth defects among live born and stillborn
infants of
20 weeks gestation or >500 g birth weight born to
residents of ve counties in metropolitan Atlanta, Georgia. This
program routinely collects demographic characteristics and clinical
information of structural birth defects, chromosomal abnormal-
ities, and syndromes from several data sources. The defects are
coded using a modied 6-digit code from the International Clas-
sication of Diseases (ICD), Ninth Revision, Clinical Modication,
and the British Paediatric Association Classication of Diseases
developed for MACDP. MACDP was granted authority to conduct
birth defect surveillance in the ve central counties of metropolitan
Atlanta on behalf of the Georgia Division of Public Health, Georgia
Department of Human Resources, and has CDCs Institutional
Review Board approval. Further details about this system have been
published elsewhere [Correa et al., 2007].
Case Classication
All MACDP records on case infants with nonchromosomal CHDs
(ICD9-CM codes 745.0747.9) were reviewed and their CHDs were
classied by experts in pediatric cardiology according to a standard
clinical nomenclature from the Society of Thoracic Surgeons (STS)
and a presumed embryological developmental schema described
elsewhere [Riehle-Colarusso et al., 2007]. The records reviewed on
each CHD case consisted of information abstracted from medical
records by trained medical abstractors. Such information included
cardiac diagnoses, presence of other defects, methods of diagnosis
(e.g., echocardiography, catheterization, surgery or autopsy), and
results of laboratory tests (e.g., genetic tests). Normal physiologic
ndings in premature or newborn infants less than 6 weeks of age
such as patent ductus arteriosus (PDA), patent foramen ovale, or
valve insufciency unrelated to structural valve abnormality were
not considered structural abnormalities and were excluded from
analysis (n 613). PDA was only considered a CHD if occurring in
term infants, persisting beyond 6 weeks of life, and not maintained
patent to ensure survival due to another cardiac condition. For this
study, 23 CHD phenotypes were grouped into nine broad categories
(cardiac looping, conotruncal defects, AVSDs, left or right ven-
tricular tract obstruction (LVOTO or RVOTO) defects, ventricular
septal defects (VSDs), atrial septal defects (ASDs), laterality defects,
and cell growth defects). Inlet-type VSDs were included in the
AVSD category and malaligned or conotruncal type VSDs in the
conotruncal group. The cell growth defects category comprised
total and partial anomalous pulmonary venous return, ASD sinus
venosus, and cor triatrium. Phenotypes with fewer than ve cases
per maternal age category were excluded.
We rst identied all cases with CHDs in the MACDP (ICD-9
codes) born between January 1, 1968 and December 31, 2005. We
used ICD-9 codes starting with 758 to exclude cases with chromo-
somal abnormalities within this set of cases with CHDs. To further
exclude cases with multiple malformations, we used codes 756.4
and 756.5 that identify chondrodystrophies and osteodystrophies,
respectively. In addition, a clinical geneticist (JLF) manually
reviewed the records of all remaining cases with multiple codes
to identify and exclude cases with multiple congenital anomalies
and specic syndromes. Noncardiac defects were dened as
major if they had surgical, medical, or serious cosmetic impor-
tance (Rasmussen et al., 2003).
Denominator
For the denominator of our prevalence estimates, we obtained vital
records information on all infants born in ve counties in metro-
politan Atlanta from the same birth cohort that gave rise to the CHD
cases (n 1,301,143). Our denominator was limited to infants who
were singleton and live born at
20 weeks of gestation. Since we had
a limited ability to link individual cases to their underlying birth
records, we applied rate-based methodology for analyzing MACDP
MILLER ET AL.
population-based design. Our analyses employed grouped case and
live birth counts to produce both crude and model adjusted
prevalence estimates and prevalence ratio (PR) estimates.
Covariates
We used MACDP data to obtain information on maternal age,
parity, gestational age, birth weight, maternal race or ethnicity, and
infant sex for case infants and used birth certicate data to obtain
information on maternal age, parity, gestational age, birth weight,
maternal race or ethnicity, and infant sex for the denominator. We
grouped maternal age at birth into the following categories:
younger than 20, 2024, 2529 (reference), and 3034 years of
age, and 35 years of age or older. Infants of mothers with missing age
information were excluded from analysis (n 21). We created
categories for birth cohort (19681980, 19811992, and
19932005), infant sex (male and female), and maternal race
(White and Nonwhite). The race category Nonwhite comprised
all other racial and ethnic categories (as dened by the U.S. Ofce of
Management and Budget). We dened parity as number of prior
live births, and grouped women without previous live births as
nulliparous and women with one or more live births as
multiparous.
Statistical Analysis
We calculated prevalence rates per 10,000 live births within each
maternal age group relative to those within the reference group. We
applied Pearson Chi-squared statistics and Poisson regression
models to estimate adjusted prevalence ratios (aPRs) and 95%
condence intervals (CIs) for each of the different age categories,
comparing them with the reference category [Cameron and Triv-
edi, 1998], adjusting for birth cohort (19681980, 19811992, or
19932005), infant sex (male or female), and maternal race (White
or Nonwhite). We used Cochran-Armitage trend test for analysis of
trend.
Maternal age was grouped into larger categories [1530
(reference), 3035, and 3555 years of age] to increase the sample
size of the maternal age groups, and stratied by maternal race,
infant sex, preterm delivery (<37 weeks or
37 weeks), and LBW
(<2,500 g or
2,500 g) to examine variations in PRs among strata
of these covariates. For examination of statistical interaction
between covariates and maternal age categories, we used a Chi-
squared test in Poisson regression. In our study of statistical
interactions between maternal age and covariates, we used multi-
plicative terms in Poisson models [Rothman et al., 2008]. We
present both crude and stratied PRs, and P-values for signicant
interactions (P < .05). We used SAS-PC for computation (9.01 v,
SAS Institute, Inc, Raleigh, NC).
RESULTS
We identied 8,277 live born and stillborn infants and fetuses with
nonchromosomal CHDs born during the period January 1, 1968-
December 31, 2005, among 1,301,143 live births in metropolitan
Atlanta. Exclusion of infants with syndromes or undened patterns
of associated major noncardiac congenital anomalies (n 2,375),
3
and prematurity-associated heart conditions (e.g., PDA, patent
foramen ovale, peripheral pulmonic stenosis; n 613) resulted in a
total of 5,289 infants with isolated CHDs and a prevalence of 40.6
per 10,000 live births. Table I shows the number of case infants with
isolated CHDs and their prevalence rates per 10,000 live births
among maternal age categories and for all births by birth cohort,
maternal race or ethnicity, infant sex, gestational age, and birth
weight. The prevalence of CHDs increased among all groups of
women, regardless of age, across successive birth cohorts, partic-
ularly among infants born to women 35 years of age or older from
prevalence per 10,000 live births of 32.5 (CI 23.843.3) for birth
cohort 19681980 to 63.0 (CI 57.868.4) for birth cohort
19932005.
Prevalence rate ratios for all selected CHDs in the aggregate
increased monotonically with maternal age, from the group
younger than 20 years of age (aPR 0.93; CI 0.811.06) to the
group 35 years of age or older (aPR 1.17; CI 1.051.31; Table II).
Table II shows the aPRs for CHDs across maternal age categories,
adjusted for infant sex, maternal race or ethnicity, and birth
period. Compared with infants born to mothers 2529 years of
age, infants born to mothers 35 years of age or older, had an
increased prevalence for several broad categories of CHD, including
laterality defects (aPR 2.03; CI 1.014.10); conotruncal defects
(aPR 1.30; CI 1.031.65), VSDs (aPR 1.20; CI 1.061.36);
ASDs (aPR 1.36; CI 1.051.77); as well as for specic phenotypes
such as D-TGA (aPR 1.65; CI 1.102.48), and coarctation of the
aorta (aPR 1.54; CI 1.102.16). Infants born to younger mothers
(<20 years of age) had a decreased prevalence of valvar pulmonic
stenosis (aPR 0.66; CI 0.440.98), and an increased prevalence of
laterality defects (aPR 2.06; CI 1.223.48). Additionally, infants
born to mothers 2024 years of age had a decreased prevalence of
aortic stenosis (aPR 0.50; CI 0.280.89), membranous VSDs
(aPR 0.78; CI 0.610.99), and muscular VSDs (aPR 0.83; CI
0.700.99). We examined the association of CHD and paternal
age and found no difference in specic associations compared
to the analysis of CHD and maternal age. We performed addi-
tional analyses comparing PRs before folate supplementation
(19681996) and after mandatory folate supplementation
(19992005) for each specic CHD selected for the study and
for CHD in the aggregate and we found no differences between
the two cohorts.
There were several signicant ndings for CHDs among infants
born to mothers in the older age groups, stratied by various
covariates (Table III). Compared with infants with normal birth
weight, infants with LBW born to mothers 3034 years of age had
signicantly increased prevalence of all selected CHDs
(P 0.0005), LVOTOs (P 0.0320), and VSDs (P 0.0064),
and those born to mothers 35 years of age or older had a signicantly
increased prevalence of RVOTO (P 0.0079; Table III). Preterm
infants born to mothers 30 years of age or older were more likely to
have an increased risk for total CHDs (P 0.0007 for the maternal
group 3034 years of age; P < 0.0000 for the group 35 years of age or
older); specically, premature infants born to mothers 3034 years
of age had a higher risk of all VSDs (P 0.0013). There were
signicant variations in prevalence estimates among infants born
to White women and infants born to Nonwhite women, all of whom
were 35 years of age or older, for the broad categories of cardiac
 TABLE I. Number of Cases and Prevalence of Isolated Congenital Heart Defects per 10,000 Livebirths by Demographic and Clinical Descriptive Characteristics, Metropolitan
Atlanta, 19682005
Cochran-
Maternal age <20 years Maternal age 2024 years Maternal age 2529 years Maternal age 3034 years Maternal age
35 years Total Armitage
trend test
Cases Prevalence Cases Prevalence Cases Prevalence Cases Prevalence Cases Prevalence Cases Prevalence
no. (95% CIa) no. (95% CI) no. (95% CI) no. (95% CI) no. (95% CI) no. (95% CI) P-Values
Birth cohort
19681980 191 29.5 (25.534.0) 296 27.2 (24.230.5) 282 29.5 (26.133.1) 116 26.8 (22.132.1) 46 32.5 (23.843.3) 931 28.5 (26.730.4) 0.8875
19811992 161 31.8 (27.137.2) 336 32.9 (29.536.6) 421 35.0 (31.738.5) 305 35.6 (31.739.8) 143 44.0 (37.151.8) 1,366 34.9 (33.136.8) 0.0060
19932005 261 43.2 (38.148.8) 564 44.3 (40.748.1) 798 51.7 (48.155.4) 819 53.2 (49.757.0) 550 63.0 (57.868.4) 2,992 51.3 (49.553.1) <0.0001
Race or ethnicity
White 273 36.5 (32.341.1) 666 36.3 (33.639.2) 971 41.7 (39.144.4) 849 45.0 (42.048.1) 500 56.9 (52.062.1) 3,259 42.4 (41.043.9) <0.0001
Nonwhite 339 33.7 (30.237.5) 528 34.3 (31.537.4) 523 38.4 (35.241.9) 388 41.7 (37.646.1) 238 52.2 (45.859.3) 2,016 38.1 (36.539.8) <0.0001
Infant sex
Male 308 34.2 (30.538.3) 590 34.2 (31.537.1) 765 40.4 (37.643.3) 635 43.9 (40.647.5) 380 55.4 (50.061.3) 2,678 40.3 (38.741.8) <0.0001
Female 304 35.5 (31.639.7) 606 36.6 (33.839.7) 736 40.7 (37.843.7) 604 43.7 (40.347.3) 359 54.8 (49.360.8) 2,609 41.0 (39.542.6) <0.0001
Gestational age
Term 484 39.6 (36.243.3) 963 38.9 (36.541.5) 1,239 43.9 (41.546.4) 995 44.0 (41.346.8) 571 51.8 (47.656.2) 4,252 43.0 (41.744.3) <0.0001
Preterm 118 49.7 (41.159.5) 217 64.5 (56.273.7) 245 77.3 (68.087.7) 220 93.6 (81.7106.8) 160 126.3 (107.5147.4) 960 76.6 (71.981.7) <0.0001
Birth weight

2,500 g 476 31.8 (29.034.8) 988 32.5 (30.534.6) 1,261 36.9 (34.939.0) 1,026 39.0 (36.641.4) 600 48.7 (44.952.8) 4,351 36.8 (35.737.9) <0.0001
<2,500 g 135 64.6 (54.276.5) 205 72.4 (62.883.0) 235 97.1 (85.1110.3) 212 121.1 (105.4138.6) 134 133.1 (111.5157.7) 921 91.2 (85.497.3) <0.0001
Parity
Nulliparous 460 34.7 (31.638.0) 584 35.3 (32.538.3) 648 42.6 (39.446.0) 391 42.8 (38.747.3) 192 58. 0(50.166.8) 2,275 39.6 (38.041.2) <0.0001
Multiparous 140 35.8: (30.142.3) 581 34.9 (32.137.8) 831 39.0 (36.441.8) 833 44.1 (41.247.2) 541 54.2 (49.759.0) 2,926 41.4 (39.942.9) <0.0001

Total 613 34.9 (32.237.8) 1,196 35.4 (33.437.4) 1,501 40.5 (38.542.6) 1,240 43.8 (41.446.4) 739 55.1 (51.259.3) 5,289 40.6 (39.641.8) <0.0001
a
Condence interval.
 MILLER ET AL.

TABLE II. Adjusted Prevalence Ratios for Isolated Congenital Heart Defects Among Maternal Age Groups, Adjusted for Sex, Time Cohort, and Maternal Race or Ethnicity, Metropolitan
Atlanta Congenital Defects Program, 19682005

Maternal age <20 years Maternal age 2024 years Maternal age 3034 years Maternal age
35 years
Congenital heart defects Cases no. aPR (95% CIa) Cases no. aPR (95% CI) Cases no. aPR (95% CI) Cases no. aPR (95% CI)
Laterality defect 31 2.06 (1.223.48)b 37 1.41 (0.872.30) 27 1.51 (0.872.62) 15 2.03 (1.014.10)
Cardiac looping 13 1.12 (0.552.28) 32 1.29 (0.762.19) 34 1.69 (1.002.86) 12 1.54 (0.763.11)
Any single ventricle 10 1.18 (0.542.60) 23 1.12 (0.612.06) 24 1.64 (0.893.01) 9 1.57 (0.703.55)
Conotruncal defect 105 0.96 (0.761.22) 200 0.94 (0.771.13) 164 0.89 (0.731.09) 109 1.30 (1.031.65)
Tetralogy of Fallot 48 0.95 (0.661.37) 92 0.91 (0.681.21) 77 0.98 (0.721.32) 46 1.30 (0.901.86)
D-TGA 32 0.83 (0.551.26) 58 0.80 (0.571.12) 48 0.76 (0.531.10) 39 1.65 (1.102.48)
Vascular ring 9 1.24 (0.562.76) 14 0.87 (0.441.71) 22 1.61 (0.872.98) 8 1.11 (0.472.60)
AVSDs 11 1.00 (0.492.03) 24 0.67 (0.401.14) 20 0.67 (0.381.18) 14 1.30 (0.662.54)
Complete AVSDs 9 1.54 (0.594.07) 15 0.91 (0.431.93) 14 1.27 (0.572.82) 6 1.73 (0.456.73)
LVOTOs 76 0.77 (0.591.01) 171 0.84 (0.691.03) 179 1.00 (0.821.22) 94 1.12 (0.871.44)
HLHS 34 1.00 (0.661.54) 66 1.05 (0.741.48) 53 1.09 (0.751.57) 21 1.17 (0.691.98)
Coarctation of the aorta 25 0.66 (0.421.03) 83 0.86 (0.651.15) 91 1.09 (0.821.45) 57 1.54 (1.102.16)
Aortic stenosis 13 0.88 (0.451.72) 16 0.50 (0.280.89) 29 0.88 (0.541.44) 12 1.01 (0.521.96)
RVOTOs 57 0.79 (0.581.08) 123 0.91 (0.711.15) 105 0.84 (0.651.08) 74 1.15 (0.861.53)
Valvar pulmonic stenosis 31 0.66 (0.440.98) 83 0.81 (0.611.08) 82 0.86 (0.651.15) 66 1.33 (0.971.82)
VSDs 283 0.91 (0.801.05) 560 0.89 (0.801.00) 638 1.02 (0.921.14) 379 1.20 (1.061.36)
Membranous VSDs 55 0.81 (0.601.11) 111 0.78 (0.610.99) 124 0.86 (0.681.09) 91 1.21 (0.931.58)
Muscular VSDs 95 0.82 (0.651.03) 218 0.83 (0.700.99) 351 1.03 (0.881.19) 212 1.11 (0.941.32)
ASDs 72 0.97 (0.731.29) 117 0.81 (0.641.03) 118 0.85 (0.671.08) 91 1.36 (1.051.77)
Secundum ASDs 39 0.89 (0.611.31) 76 0.90 (0.671.21) 82 0.87 (0.651.17) 64 1.32 (0.961.82)
Ebsteins anomaly 6 1.79 (0.684.73) 13 1.22 (0.582.54) 12 1.20 (0.552.60) 10 1.94 (0.834.53)
Cell growth defects 13 1.23 (0.642.37) 29 0.86 (0.531.40) 30 1.06 (0.641.73) 12 1.20 (0.592.43)
Total and partial anomalous 11 1.25 (0.622.56) 22 0.88 (0.511.51) 24 1.42 (0.812.52) 10 1.51 (0.693.29)
pulmonary venous return
Total heart defects 613 0.93 (0.811.06) 1,196 0.91 (0.821.00) 1,240 0.98 (0.891.07) 739 1.17 (1.051.3)
aPR, adjusted prevalence ratio (reference group is maternal age 2529 years); CI, condence interval; D-TGA, dextro-transposition of the great arteries; AVSDs, atrioventricular septal defects; HLHS, hypoplastic left heart syndrome; LVOTOs,
left ventricular outow tract obstructions; RVOTOs, right ventricular outow tract obstructions; VSDs, ventricular septal defects; ASDs, atrial septal defects; TOF, tetralogy of Fallot; TAPVR, total anomalous pulmonary venous return.
a
Condence interval.
b
Bolded results indicate statistically signicant values (P < 0.05).
5
6 AMERICAN JOURNAL OF MEDICAL GENETICS PART A

TABLE III. Crude and Stratied Prevalence Ratios for Isolated Congenital Heart Defects Among Maternal Age Groups by Maternal Race or
Ethnicity, Sex, Birth weight, and Preterm Delivery, Metropolitan Atlanta Congenital Defects Program, 19682005
Maternal age
Defect (years) PR crude (95% CI) Cofactor PR stratied (95% CI) P-Valuesa
Laterality defects
35 1.01 (0.591.74) Male 1.74 (0.873.48) 0.0496
Female 0.55 (0.221.37)
Cardiac looping defects 3035 1.46 (0.972.19) White 0.95 (0.561.61) 0.0136
Nonwhite 2.80 (1.425.50)b
Conotruncal defects 3035 0.95 (0.801.13) Preterm 1.52 (0.962.39) 0.0205
Term 0.85 (0.701.02)
AVSDs 3035 0.92 (0.561.51) Preterm 2.71 (0.868.53) 0.0454
Term 0.73 (0.421.29)
LVOTOs 3035 1.18 (0.991.40) LBW 1.94 (1.213.12) 0.0320
Normal BW 1.11 (0.921.34)
RVOTOs
35 1.46 (1.131.87) LBW 2.57 (1.624.10) 0.0079
Normal BW 1.21 (0.901.64)

35 1.46 (1.131.87) Preterm 2.28 (1.443.61) 0.0117
Term 1.12 (0.821.52)
VSDs 3035 1.27 (1.161.39) LBW 1.71 (1.372.13) 0.0064
Normal BW 1.22 (1.101.35)
3035 1.27 (1.161.39) Preterm 1.68 (1.352.10) 0.0013
Term 1.13 (1.021.25)

35 1.59 (1.421.78) Preterm 2.01 (1.552.60) 0.0113
Term 1.39 (1.221.57)

35 1.59 (1.421.78) White 1.69 (1.481.93) 0.0458
Nonwhite 1.31 (1.051.62)
Total heart defects 3035 1.17 (1.101.25) LBW 1.55 (1.321.81) 0.0005
Normal BW 1.14 (1.061.22)
3035 1.17 (1.101.25) Preterm 1.44 (1.231.68) 0.0007
Term 1.07 (0.991.15)

35 1.47 (1.361.60) Preterm 1.94 (1.632.31) 0.0000
PR, prevalence ratios (reference group is maternal age <30 years); CI, condence interval; LBW, low birthweight; D-TGA, dextro-transposition of the great arteries; AVSDs, atrioventricular
septal defects; HLHS, hypoplastic left heart syndrome; LVOTOs, left ventricular outow tract obstructions; RVOTOs, right ventricular outow tract obstructions; VSDs, ventricular septal defects; ASDs,
atrial septal defects; PDA, patent ductus arteriosus; TOF, tetralogy of Fallot.
a
P < 0.05 of Chi-squared test for interaction term of maternal age category and cofactor for specic CHDs in Poisson regression.
b
Bolded results indicate statistically signicant values P < 0.05.

looping (P 0.0136) and all VSDs (P 0.0458). We found a
predominance of laterality defects among males born to women
35 years of age or older (Table III).
DISCUSSION
Our ndings suggest that the birth prevalence of isolated CHDs in
the aggregate might be associated with advanced maternal age,
especially 35 years of age or older. Infants born to mothers older
than 35 years of age seemed to be at 20% increased risk of CHDs,
while infants born to younger mothers tended to be at decreased risk
of CHDs. Our results also suggest an effect of advanced maternal age
on the prevalence of specic groups of CHDs, such as laterality
defects, conotruncal defects, LVOTOs, RVOTOs, AVSDs, and
VSDs.
To our knowledge, our study is the rst population-based study
to assess the associations between various maternal age groups and
different phenotypic categories of isolated CHDs, and to investigate
possible variation of these associations by potential effect modiers
such as birth weight, preterm birth, maternal race, and infant sex. In
addition, our study included cases of CHDs among liveborns,
stillborns, and fetuses from a dened population ascertained by
an active surveillance system with high sensitivity; thus, it repre-
sents a more inclusive group of case infants than reported by most
other surveillance programs. Furthermore, a clinical, standardized
nomenclature and developmental schema were used to classify
heart defects and eliminate nonstructural and newborn or
prematurity-associated cardiac conditions [Riehle-Colarusso
et al., 2007], and cases with associated noncardiac anomalies
were carefully reviewed by a dysmorphologist to exclude those
with major anomalies and syndromes. Comparing to the study by
Reefhuis and Honein [2004] we included only cases of CHD
without associated major noncardiac malformations, used a differ-
ent CHD classication and groups of CHD [Riehle-Colarusso et al.,
2007], had an extended birth cohort, and also accounted for
possible effect modiers such as LBW, preterm delivery, maternal
race, and infant sex. Syndromic cases and cases with additional
noncardiac major malformations were excluded from our analysis.
Our study had several limitations. First, the sample size for some
of the CHD phenotypes was relatively small. Consequently, we
MILLER ET AL.
might not have had an adequate statistical power to detect real
associations or evaluate the cumulative effect of birth weight and
gestational age on specic CHD phenotypes among maternal
categories 3539 and 4045 years of age. Second, our reference
population did not include stillbirths and elective terminations.
Also, we were able to stratify maternal race or ethnicity only as
White versus Nonwhite because we did not have more detailed
information on race or ethnicity categories for the earlier years of
the study data. Thus, we could not determine if racial or ethnic
disparity in our results reected disparities for Hispanics/Latinos,
Black or African-American, or for Asian or other ethnic subgroups.
Although advanced paternal age may be an additional confounder,
as it is associated with increased mutation rate for autosomal
dominant diseases, we were not able to study it because MACDP
did not start collecting paternal age information until 1984. Con-
sequently, paternal age information was missing in 44% of cases.
Furthermore, this information is often missing on vital records,
which are the main source of data on paternal age. Lastly, our
ascertainment of CHDs relied heavily on hospital charts and their
related diagnostic evaluations (roentgenograms, color Doppler
echocardiography, or autopsies when available). We have opted
for not using any of the methods to account for multiple compar-
isons because such methods have not proven to be widely accepted
or used, particularly in situations where the inferential process may
be more complex and involve a sequential exploratory approach
(e.g., going from the aggregate to subgroups to more specic
phenotypes) in an effort to determine possible sources of variation.
Our results suggest that the prevalence of isolated CHDs as a
group increases with increasing maternal age, as has also been noted
by other authors [Croen and Shaw, 1995; Hollier et al., 2000;
Reefhuis and Honein, 2004; Forrester and Merz, 2008]. In the
maternal age category 35 years of age or older, our estimates were
within the ranges previously reported by others [Hollier et al., 2000;
Reefhuis and Honein, 2004; Reller et al., 2008]. However, Reefhuis
and Honein [2004] found a U-shaped relationship, resulting from a
slightly increased PR for all CHDs among younger maternal age
categories. This discrepancy with our ndings could be explained in
part by the fact that we used a different cohort of CHD case infants
and a different CHD classication system. While four previous
studies did not nd an association of CHDs in the aggregate with
maternal age [Baird et al., 1991; Ferencz et al., 1997; Cedergren et al.,
2002; Amorim et al., 2008], Baird et al. [1991] and Ferencz et al.
[1997] did report associations of maternal age with specic CHD
phenotypes. However, Ferencz et al. [1997] relied on data from
1981 to 1987, in which the proportion of births to older mothers was
low, and the oldest maternal age category in their study was 30 years
of age or older. Thus, the maternal age effect among older age
groups might have been underestimated. This also might have been
true for the population-based study by Baird that covered the
period 19661981 [Baird et al., 1991]. Cedergren et al. [2002], in
a hospital-based study with small sample size found a nonsigni-
cant U-shaped curve without clear cut dependency.
With regard to young maternal age and specic CHD pheno-
types, a previous study had shown a decreased aPR for RVOTO
among infants of women younger than 20 years of age group
[Reefhuis and Honein, 2004], while in our study we found a
signicant association only for one major type of RVOTO, valvar
7
pulmonary stenosis. We also observed a decreased aPR for mem-
branous and muscular VSDs among offspring of mothers in the
2024 years of age group, which is consistent with a previous
publication [Pradat et al., 2003]. Further studies are warranted
to corroborate these ndings and, if conrmed, to elucidate pos-
sible reasons for these associations.
Our nding of an association of ASDs with advanced maternal
age is consistent with several previous reports [Rothman and Fyler,
1976; Ferencz et al., 1997; Pradat et al., 2003; Forrester and Merz,
2008]. Our ndings also corroborated other studies showing an
effect of advanced maternal age on VSDs [Rothman and Fyler, 1976;
Forrester and Merz, 2008], coarctation of the aorta [McBride et al.,
2005], and D-TGA [Rothman and Fyler, 1976; Ferencz et al., 1997;
Reefhuis and Honein, 2004]. Some discrepancies between our study
and previous reports regarding the association between maternal
age and increased risk of selected CHD phenotypes might be due in
part to differences in study populations, methods, and unaccounted
confounding or effect modication [Correa-Villase~ nor et al., 1994;
Pradat et al., 2003; Reefhuis and Honein, 2004; McBride et al., 2005;
Forrester and Merz, 2008]. For example, a casecontrol study
designed to evaluate the effects of exposure to trichloroethylene
(TCE) on the risk of CHDs found maternal age to be an effect
modier for the relationship between TCE and CHDs [Yauck et al.,
2004]. So differences in exposure to possible effect modiers among
studies could potentially result in apparent null effects when in fact
the level of association might differ across strata of the effect
modier of interest. Advanced maternal age has also been asso-
ciated with the use of assisted reproductive technology, which in
turn has been associated with septal heart defects [Reefhuis et al.,
2009]. So differences in exposure to assisted reproductive technol-
ogy or prevalence of infertility among study populations could also
explain some of the differences in ndings between studies. We had
no information on the use of assisted reproductive technology in
our study population, so we were not able to address this issue. We
did examine parity as a covariate by conducting analyses stratied
by parity and found no effect of parity (data not shown). Addi-
tionally, several maternal co-morbidities such as diabetes [Wren
et al., 2003; Correa et al., 2008], obesity [Gilboa et al., 2010],
hypertension [Caton et al., 2009], medication use (e.g., selective
serotonin-reuptake inhibitors [Louik et al., 2007]), and smoking
[Malik et al., 2008; Alverson et al., 2011] could be related to
advanced maternal age and explain the association between
advanced maternal age and CHD. We had no information on these
factors and, therefore, could not account for them in our analyses.
Although we observed a signicantly stronger maternal age
association among LBW and preterm infants for several groups
of CHD among offspring of women in the 3034 years and in the 35
years of age or older categories (Table III), we did not identify a
consistent pattern in the estimates of associations. The association
of CHD with LBW and prematurity has been reported previously
[Mili et al., 1991; Rosenthal et al., 1991; Rosenthal, 1996; Rasmussen
et al., 2001; Tanner et al., 2005; Malik et al., 2007], but the reasons
for such association are not known. One plausible explanation is
that the presence of CHD might impair normal fetal growth.
However, although we could suggest that impaired fetal circulation
may cause LBW, no consistent hemodynamic pattern with respect
to oxygen saturations or hemoperfusion can explain our sugges-
8 AMERICAN JOURNAL OF MEDICAL GENETICS PART A
tion. Another possibility is that both altered fetal growth and CHD
represent co-outcomes associated with a common but yet unde-
ned genetic susceptibility or prenatal factor (e.g., alcohol con-
sumption, infection, and nutritional status) and that the effect of
this factor could vary with maternal age and vice versa.
We found signicant racial or ethnic variations for some of the
maternal age and CHD associations (Table III). Our ndings for
LVOTOs, VSDs, and AVSDs, were consistent with previously
reported observations [Ferencz et al., 1997; Botto et al., 2001;
Pradat et al., 2003], but not for conotruncal defects [Ferencz
et al., 1997] and ASDs [Pradat et al., 2003]. Variability in rates
by race or ethnicity might reect variations in genetic susceptibility,
socioeconomic status, and/or access to health care. For example, a
study of CHD and race/ethnicity by Correa-Villasenor et al. [1991]
showed an excess of cases of L-TGA among White infants and an
excess of aortic stenosis (included in LVOTOs) among White
infants whose mothers were of low and middle socioeconomic
strata. These results suggest that some of the variations in the
associations of maternal age with CHD by race or ethnicity might be
associated with socioeconomic factors [Correa-Villasenor et al.,
1991]. Our results on specic phenotypes on variation by infant sex
are conrmed by some studies [Samanek, 1994; Correa et al., 2007],
but differ from others [Ferencz et al., 1997; Lary and Paulozzi,
2001]. Differences in prevalence rates by maternal age and infant sex
might represent differences of in utero survival of male and female
fetuses, or they might be indicative of toxic and endocrine environ-
mental exposures that can accumulate with advanced maternal age
and can affect infant sex ratios [Jarrell et al., 2002; Karmaus et al.,
2002; Weisskopf et al., 2003].
CONCLUSION
Our ndings suggest that the birth prevalence of specic isolated
CHD phenotypes, such as coarctation of the aorta, valvar pulmonic
stenosis, VSDs, and ASDs, might be associated with advanced
maternal age, especially among offspring of mothers in the 35 years
of age or older. Our results also suggest a modication by LBW and
preterm delivery of the variation in prevalence with advanced
maternal age for selected CHD phenotypes. Therefore, certain
isolated CHD phenotypes might need to be considered in future
studies assessing the risk of birth defects in relation to maternal age.
Further studies are warranted to corroborate these ndings, taking
into account other possible confounding factors, such as diabetes,
obesity, hypertension, and smoking.
REFERENCES
Alverson CJ, Strickland MJ, Gilboa SM, Correa A. 2011. Maternal smoking
and congenital heart defects in the Baltimore-Washington infant study.
Pediatrics 127:647653.
Amorim LFP, Pires CAB, Lana AMA, Campos AS, Aguiar RALP, Tib
urcio
JD, Siqueira AL, Mota CCC, Aguiar MJB. 2008. Presentation of con-
genital heart disease diagnosed at birth: Analysis of 29,770 newborn
infants. J Pediat (Rio J) 84:8390.
Baird PA, Sadovnick AD, Yee IM. 1991. Maternal age and birth defects:
A population study. Lancet 337:527530.
Botto LD, Correa A, Erickson JD. 2001. Racial and temporal variations in
the prevalence of heart defects. Pediatrics 107:E32.
Cameron AC, Trivedi PK. 1998. Regression analysis of count data.
Cambridge, UK: Cambridge University Press.
Caton AR, Bell EM, Druschel CM, Werler MM, Lin AE, Browne ML,
McNutt LA, Romitti PA, Mitchell AA, Olney RS, Correa A. 2009.
Antihypertensive medication use during pregnancy and the risk of
cardiovascular malformations. Hypertension 54:6370.
Cedergren MI, Selbing AJ, Kall
en BAJ. 2002. Risk factors for cardiovascular
malformationA study based on prospectively collected data. Scand J
Work Environ Health 28:1217.
Correa A, Cragan JD, Kucik JE, Alverson CJ, Gilboa SM, Balakrishnan R,
Strickland MJ, Duke CW, OLeary LA, Riehle-Colarusso T, Siffel C,
Gambrell D, Thompson D, Atkinson M, Chitra J. 2007. Reporting birth
defects surveillance data 19682003. Birth Defects Res A Clin Mol Teratol
79:65186.
Correa A, Gilboa SM, Besser LM, Botto LD, Moore CA, Hobbs CA, Cleves
MA, Riehle-Colarusso TJ, Waller DK, Reece EA. 2008. Diabetes mellitus
and birth defects. Am J Obstet Gynecol 199:237239.
Correa-Villasenor A, McCarter R, Downing J, Ferencz C. 1991. White-
black differences in cardiovascular malformations in infancy and socio-
economic factors. The Baltimore-Washington Infant Study Group. Am J
Epidemiol 134:393402.
Correa-Villase~
nor A, Ferencz C, Neill CA, Wilson PD, Boughman JA. 1994.
Ebsteins malformation of the tricuspid valve: Genetic and environ-
mental factors. The Baltimore-Washington Infant Study Group. Tera-
tology 50:137147.
Croen LA, Shaw GM. 1995. Young maternal age and congenital malfor-
mations: A population-based study. Am J Public Health 85:710713.
Dastgiri S, Stone DH, Le-Ha C, Gilmour WH. 2002. Prevalence and secular
trend of congenital anomalies in Glasgow, UK. Arch Dis Child
86:257263.
Ferencz C, Loffredo C, Correa-Villase~
nor A, Wilson PD. 1997. Genetic And
Environmental Risk Factors of Major Cardiovascular Malformations.
The Baltimore-Washington Infant Study, 19811989. Armonk, NY:
Futura Publishing Company.
Forrester MB, Merz RD. 2008. Precurrence risk of birth defects in Hawaii.
Congenit Anom (Kyoto) 48:4044.
Francannet C, Lancaster PA, Pradat P, Cocchi G, Stoll C. 1993. The
epidemiology of three serious cardiac defects. A joint study between
ve centres. Eur J Epidemiol 9:607616.
Garg A, Connolly C, Hollier L. Maternal age malformations in singleton
births. Female Patient. 2008. Accessed January 29, 2010, from http://
www.femalepatient.com/html/cme/articles/033_07_038.asp.
Gilboa SM, Correa A, Botto LD, Rasmussen SA, Waller DK, Hobbs CA,
Cleves MA, Riehle-Colarusso TJ. 2010. Association between prepreg-
nancy body mass index and congenital heart defects. Am J Obstet Gynecol
202:51.e151.e10.
Hamilton B, Ventura S. 2006. Fertility and abortion rates in the United
States, 19602002. Int J Androl 29:3445.
Hollier LM, Leveno KJ, Kelly MA, McIntire DD, Cunningham FG. 2000.
Maternal age and malformations in singleton births. Obstet Gynecol
96:701706.
Jarrell JF, Gocmen A, Akyol D, Brant R. 2002. Hexachlorobenzene exposure
and the proportion of male births in Turkey 19351990. Reprod Toxicol
16:6570.
Jenkins K, Correa A, Feinstein J, Botto L, Britt A, Daniels S, Elixson M,
Warnes C, Webb C. 2007. Noninherited risk factors and congenital
cardiovascular defects: Current knowledge: A scientic statement from
MILLER ET AL.
the American Heart Association Council on Cardiovascular Disease in
the Young: Endorsed by the American Academy of Pediatrics. Circu-
lation 115:29953014.
Karmaus W, Huang S, Cameron L. 2002. Parental concentration of
dichlorodiphenyl dichloroethene and polychlorinated biphenyls in
Michigan sh eaters and sex ratio in offspring. J Occup Environ Med
44:813.
Lary JM, Paulozzi LJ. 2001. Sex differences in the prevalence of human birth
defects: A population-based study. Teratology 64:237251.
Lee K, Khoshnood B, Chen L, Wall SN, Cromie WJ, Mittendorf RL. 2001.
Infant mortality from congenital malformations in the United States,
19701997. Obstet Gynecol 98:620627.
Loane M, Dolk H, Morris JK. 2009. Maternal age-specic risk of non-
chromosomal anomalies. BJOG 116:11111119.
Louik C, Lin AE, Werler MM, Hern
andez-D
az S, Mitchell AA. 2007. First-
trimester use of selective serotonin-reuptake inhibitors and the risk of
birth defects. N Engl J Med 356:26752683.
Malik S, Cleves MA, Zhao W, Correa A, Hobbs CA. 2007. Association
between congenital heart defects and small for gestational age. Pediatrics
119:e976e982.
Malik S, Cleves MA, Honein MA, Romitti PA, Botto LD, Yang S, Hobbs CA.
2008. Maternal smoking and congenital heart defects. Pediatrics
121:810816.
McBride KL, Marengo L, Caneld M, Langlois P, Fixler D, Belmont JW.
2005. Epidemiology of noncomplex left ventricular outow tract
obstruction malformations (aortic valve stenosis, coarctation of the
aorta, hypoplastic left heart syndrome) in Texas, 19992001. Birth
Defects Res A Clin Mol Teratol 73:555561.
Mili F, Edmonds LD, Khoury MJ, McClearn AB. 1991. Prevalence of birth
defects among low-birth-weight infants. A population study. Am J Dis
Child 145:13131318.
OMalley CD, Shaw GM, Wasserman CR, Lammer EJ. 1996. Epidemiologic
characteristics of conotruncal heart defects in California, 19871988.
Teratology 53:374377.
Pradat P. 1992. Epidemiology of major congenital heart defects in Sweden,
19811986. J Epidemiol Community Health 46:211215.
Pradat P, Francannet C, Harris JA, Robert E. 2003. The epidemiology of
cardiovascular defects, part I: A study based on data from three large
registries of congenital malformations. Pediatr Cardiol 24:195221.
Rasmussen SA, Moore CA, Paulozzi LJ, Rhodenhiser EP. 2001. Risk for
birth defects among premature infants: A population-based study.
J Pediatr 138:668673.
Rasmussen SA, Olney RS, Holmes LB, Lin AE, Keppler-Noreuil KM, Moore
CA. 2003. Guidelines for case classication for the National Birth Defects
Prevention Study. Birth Defects Res A Clin Mol Teratol 67:193201.
Reefhuis J, Honein MA. 2004. Maternal age and non-chromosomal birth
defects, Atlanta19682000: Teenager or thirty-something, who is at
risk? Birth Defects Res A Clin Mol Teratol 70:572579.
9
Reefhuis J, Honein MA, Schieve LA, Correa A, Hobbs CA, Rasmussen SA.
2009. Assisted reproductive technology and major structural birth defects
in the United States. Hum Reprod 24:360366.
Reller MD, Strickland MJ, Riehle-Colarusso T, Mahle WT, Correa A. 2008.
Prevalence of congenital heart defects in metropolitan Atlanta,
19982005. J Pediatr 153:807813.
Riehle-Colarusso T, Strickland MJ, Reller MD, Mahle WT, Botto LD,
Siffel C, Atkinson M, Correa A. 2007. Improving the quality of surveil-
lance data on congenital heart defects in the metropolitan Atlanta
congenital defects program. Birth Defects Res A Clin Mol Teratol
79:743753.
Rosenthal GL. 1996. Patterns of prenatal growth among infants with
cardiovascular malformations: Possible fetal hemodynamic effects.
Am J Epidemiol 143:505513.
Rosenthal GL, Wilson PD, Permutt T, Boughman JA, Ferencz C. 1991. Birth
weight and cardiovascular malformations: A population-based study.
The Baltimore-Washington Infant Study. Am J Epidemiol 133:
12731281.
Rothman KJ, Fyler DC. 1976. Sex, birth order, and maternal age character-
istics of infants with congenital heart defects. Am J Epidemiol
104:527534.
Rothman KJ, Greenland S, Lash TL. editors. 2008. Modern epidemiology.
3rd edition, Philadelphia: Lippincott-Williams Wilkins.
Samanek M. 1994. Boy:girl ratio in children born with different forms of
cardiac malformation: A population-based study. Pediatr Cardiol
15:5357.
Tanner K, Sabrine N, Wren C. 2005. Cardiovascular malformations among
preterm infants. Pediatrics 116:e833e838.
Weisskopf M, Anderson H, Hanrahan L. 2003. Decreased sex ratio follow-
ing maternal exposure to polychlorinated biphenyls from contaminated
Great Lakes sport-caught sh: A retrospective cohort study. Environ
Health 2:2.
Wilson PD, Correa-Villase~nor A, Loffredo CA, Ferencz C. 1993. Temporal
trends in prevalence of cardiovascular malformations in Maryland and
the District of Columbia, 19811988. The Baltimore-Washington Infant
Study Group. Epidemiology 4:259265.
Wren C, Richmond S, Donaldson L. 2000. Temporal variability in birth
prevalence of cardiovascular malformations. Heart 83:414419.
Wren C, Birrell G, Hawthorne G. 2003. Cardiovascular malformations in
infants of diabetic mothers. Heart 89:12171220.
Yang Q, Khoury MJ, Mannino D. 1997. Trends and patterns of mortality
associated with birth defects and genetic diseases in the United States,
19791992: An analysis of multiple-cause mortality data. Genet
Epidemiol 14:493505.
Yauck J, Malloy M, Blair K, Simpson P, McCarver DG. 2004. Proximity of
residence to trichloroethylene-emitting sites and increased risk of off-
spring congenital heart defects among older women. Birth Defects Res A
Clin Mol Teratol 70:808814.