Etiology of Ventricular Septal Defects: An

Epidemiologic Approach

Thomas B. Newman, MD, MPH

From the Institute for Health Policy Studies and Departments of Pediatrics and
Epidemiology and International Health, University of California at
San Francisco, San Francisco

ABSTRACT. To investigate the recent 150% increase in The epidemiologic approach to disease etiology
the reported incidence of ventricular septal defects involves studying the distribution of the disease in
(VSDs) in the United States, the epidemiology of yen-
different times, places, and people to discover risk
tricular septal defects was examined. The apparent mci-
dence of VSDs is highly dependent on case finding meth- factors and synthesis of these risk factors with
ods, and more complete diagnosis and reporting probably medical knowledge to yield a model of causation.
account for the increase in reported incidence. Variations Because inference from epidemiologic studies re-
in case ascertainment also account for the small differ- quires comparison of incidence rates, it is necessary
ences in incidence in studies from different places. The
to begin with the problems of case definition and
several known risk factors for VSD, including a family
history of congenital heart disease and exposure to cer- ascertainment involved with estimating VSD mci-
tam drugs, infectious agents, and maternal metabolic dence.
disturbances, explain few cases. Incidence rates are sim- VSDs can occur alone or in combination with
ilar in different races and seasons and are unrelated to other cardiac defects. The majority of VSDs are
maternal age, birth order, sex, and socioeconomic status.
isolated or simple23; these are probably etiolog-
VSDs occur naturally in a wide range of mammals and
in birds, which also have four-chambered hearts. Despite ically distinct from complex VSDs2 and are the
identical genes and similar prenatal environments, the focus of this paper. Most studies distinguish be-
concordance rate in identical twins is only about 10%. tween simple and complex VSDs; for those that do
The consistency of incidence among individuals with not do so explicitly, it will be assumed that the
widely differing genes and environments and the fre-
VSDs are simple.
quency of discordance in identical twins suggest that
VSDs often occur as random errors in development, at a VSDs usually are asymptomatic and often close
frequency largely determined by the complexity of normal spontaneously, making accurate estimation of in-
cardiac morphogenesis. This hypothesis has two major cidence difficult. The steps involved in counting
implications: many VSDs are not preventable and par- VSD cases to estimate the incidence and the many
ents need not feel responsible for VSDs in their children.
opportunities to lose cases are shown in Fig. 1.
Pediatrics 1985; 76:741-749; congenital heart defect, yen-
tricular septal defect, epidemiology.
Variations in inclusion criteria, surveillance level,
diagnostic accuracy, and referral patterns all must
be considered when incidence proportions are corn-
pared.

Recently, the Centers for Disease Control re-

DESCRIPTIVE EPIDEMIOLOGY
ported that the incidence of ventricular septal de-
fects (VSDs) in two US birth defect surveillance Secular Trends
systems had increased by 150% from 1968 to 1980.1
The first increase in the incidence of VSD, from
In view of current widespread concern about envi-
1.9/1,000 to 3.2/1,000 was reported by Carlgren4 in
ronmental contamination with teratogenic chemi-
Gothenburg, Sweden, from 1941 to 1960. The in-
cals, this increase is alarming and warrants a review
crease was entirely due to small, spontaneously
of what is known about the etiology of VSDs.
closing VSDs, which were not diagnosed in the early
years of the study. A similar increase, from 1.1/
Received for publication Nov 5, 1985; accepted Feb 12, 1985.
1,000 to 2.2/1,000, was reported by Bound and
Reprint requests to (T.B.N.) Bldg 1, Room 201, San Francisco
General Hospital, San Francisco, CA 94110.
Logan5 in Blackpool, England, from 1957 to 1971,
PEDIATRICS (ISSN 0031 4005). Copyright 1985 by the although no data were available on the size or
American Academy of Pediatrics. natural history of the VSDs.

PEDIATRICS Vol. 76 No. 5 November 1985 741

Downloaded from by guest on July 13, 2017
 STEPS TO DIAGNOSIS OF VSD CASES MISSED

I VSD occurs

E:::iiiii:: ... 5pnfoncus

Sponfonsous
Abortions
Still births
I Th.rap.uf b.for. Ic birth
closure
VSD present
L in live birth
. Spontanous Closure or loss
fo follow-up botor. firsf .xom

3
Murmur present on
first examination . Some uncoop.rotivs pO$i.nts
. Subfl#{149}
murmurs
. Examin.r in#{149}xprinc#{149}
. Sponton.ous closure or loss

to follow-up b.for.
5, murmur board
Murmur heard on
examination
. Misdiagnos.d cas.s
(.sp.cially atypical murmurs
in osympfomo$ic infants)
1
Murmur recognized
as pathologic
. Infants who ore nof r#{149}fsrrd

Referral
I to a
(e.g. b.caus.
osymptornotic)
th.y or.

cardiologist . Cos.s r.f.rr.d

non-participating cordiologisfs
. Sponfon.ous closur. or loss
fo follow-up b.for. examination
by investigators
Case counted by]
investigators

Fig 1. Steps involved in ascertaining ventricular septal defect (VSD) incidence and ways
for cases to be lost.

Anderson et al6 reported an increase in VSD 3.0

incidence from 1968 to 1975 in the Birth Defects
in
Monitoring Program (BDMP) and the Metro- w
-C

politan Atlanta Congenital Defects Program

(MACDP). The trend continued through 19801 (Fig
z-
wo
2.0
2). Is it real, or might it be due only to more
U- 1.5
complete case finding?
The BDMP is a registry based on discharge sum- 1.0
manes of newborns from about 1,500 US hospitals,
>;
whereas the MACDP includes cases diagnosed up 0. .5
to 1 year of age in the Atlanta area.7 VSD incidence
in the BDMP is only 30% to 40% of that in the 0

MACDP, because most VSDs are asymptomatic in 68 70 72 74 76 78 80

the newborn period. What changes could have led YEAR

to a greater number of VSDs being recorded in the Fig 2. Incidence of ventricular septal defects (VSDs) in
two surveillance systems? During the period of two birth defect surveillance systems, 1968 to 1980.
increasing incidence, use of two new technologies, (MACDP, Metropolitan Atlanta Congenital Defects Pro-
gram; BDMP, Birth Defect Monitoring Program.)
neonatal intensive care and pediatric echocardiog-
raphy, increased rapidly.
VSDs occur more frequently in premature and led to longer hospital stays or more frequent ex-
low-birth-weight infants, who account for one third aminations before death or discharge for these in-
of all VSDs.8 Neonatal intensive care could have fants. For example, if the proportion of low-birth-

742 VENTRICULAR SEPTAL DEFECTS

Downloaded from by guest on July 13, 2017
weight infants with VSDs diagnosed before hospital time of year, seasonal variation in incidence of
discharge increased from 20% to 60%, more than VSDs may yield clues to environmental causes. The
half of the observed increase in VSD incidence in incidence rates of VSDs by month and quarter in
the BDMP would be explained. (This calculation is six studies are presented in Fig 3. None of the
based on data from Mitchell et a!8, including a 10% studies found statistically significant seasonal var-
frequency of babies less than 2,500 g and a 0.8% iation, nor does an analysis of the pooled data (P
incidence of VSD in these infants.) > .5 using Edwardss2 method). Although it is
Increased use of echocardiography could also possible that teratogens occurring at different sea-
have led to more complete ascertainment of VSDs. sons are balancing each other, that high sponta-
Pediatric echocardiography tends to be a hospital- neous abortion rates are masking the seasonal var-
based technology, and both the BDMP and the iation, or that teratogens appear at different sea-
MACDP are more likely to count patients diag- sons in different times and places,22 it seems most
nosed in hospitals.7 Thus, in addition to the help it likely that teratogens with seasonal variation do
may provide in diagnosing nonspecific murmurs, not cause many VSDs.
echocardiography might result in increased ascer-
tainment simply by virtue of its location in referral Geographic Differences
centers.
Congenital heart disease has been reported in all
Whether due to these or other factors, more
parts of the world, but the most complete popula-
complete diagnosis of minor VSDs seems a plausi-
tion-based studies have been done in the United
ble explanation for the increase in incidence in the
States and Northwestern Europe (Table 1). VSD
MACDP. To test whether this occurred, Layde et
incidence rates are similar in all of these studies.
a!9 assumed that these minor VSDs were more
The higher incidence in the studies by Mitchell
likely to close spontaneously, and they compared
et al1#{176}
and Hoffman and Christianson4 reflects the
the rate of spontaneous closure of the VSDs diag-
close surveillance of the infants born into these two
nosed in 1970 to 1972 (when the incidence was
prospective studies of pregnancy outcome. The
about 1/1,000) with the rate for those diagnosed in
highest rate was obtained by Kerrebijn,5 who
1975 to 1976 (when the incidence was about 2.5/
screened most of the patients himself, thus avoiding
1,000). They found no difference (23.8% v 23.9%)
losses due to lack of recognition or referral by
and concluded that the increase in incidence of
primary care physicians. The true incidence of VSD
VSDs was real. However, their assumption that
apparently is close to 3/1,000 live births; lower rates
minor defects are appreciably more likely to close
probably reflect incomplete case finding rather than
spontaneously within 1 year may have been incor-
geographic variation in incidence.
rect. Mitchell et aP#{176}
reported similar rates of spon-
Because detection of other types of congenital
taneous closure for symptomatic and asymptornatic
heart disease (CHD) also increases with closer sur-
VSDs: five of 17 (29%) and 32 of 92 (34%), respec-
veillance, the proportion of total CHD made up by
tively. Although the spontaneous closure rate for
VSD may be less dependent on case ascertainment
small VSDs is probably higher overall,2 the dif-
ference at 1 year of age may be small, because the
majority of asymptomatic VSDs are still patent at
that age.#{176}3
U)
Comparison of the proportion of symptomatic Lu
U)
VSDs rather than the proportion closing sponta- U
0
neously for the periods of low and high incidence Lu
I-
would provide better evidence, because the diagno- 0
a.
Lu
sis of symptomatic VSDs is likely to have been
U.
more complete in both time periods. However, even 0
if the increase reported through 1980 is artifactual, Lu

continued surveillance is essential. Because the in-

z
cidence in the MACDP is now close to the level
reported in the most intensive incidence sur-
veys,445 it will be hard to attribute any further JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
(1) (2) (3) (4)
increases to better diagnosis and reporting.
MONTH (quarter)

Seasonal Trends Fig 3. Seasonal incidence of ventricular septal defects

in six studies (Kenna et al,6 Slater et al,tT Rothman and
Because exposure to many teratogens (such as Fyler8 Campbell and Goodwin,9 Rosenberg and Heino-
infectious agents and air pollution) varies by the nen,3 and Edwards20).

ARTICLES 743
Downloaded from by guest on July 13, 2017
TABLE 1. Incidence of Ventricular Septal Defect (VSD) in Population-Based Studies*
First Author of Place (Years) Total No. of VSD/1,000 VSD/CHD Case Ascertainment Method
Reference Live Births Live
Births

Pinkley USA (1970-1980) 1,000,000/yr 0.4-1.2 BDMP: Newborn discharge sum-

manes
Pinkley Atlanta (1968-1980) 27,000/yr 1.02.6 MACDP: Birth defects registry
Bound5 Blackpool, England
(1957-1961 and 18,073t 1.1 18% Author saw all local referrals
1967-1971) 19,112 2.2 35%
Laursen23 Denmark 86,049 1.5 24% Death certificates and records of
(1963-1973) pediatric cardiology depart-
ments
Dickinson24 Liverpool, England 160,480 1.8 33% Congenital anomalies registry
(1960-1969)
Feldt25 Olmsted City, MN 32,393 1.9 33% Computerized record retrieval
(1950-1969)
Carlgren426 Gothenburg, Sweden
(1941-1950 and 58,105 1.9 28% Author saw all local referrals
1951-60) 58,314 3.2 41%
Mitchell#{176} 12 centers in US 54,765 2.3 30% Collaborative peninatal study
(1959-1966)
Hoffman4 Oakland, CA 19,044 2.8 31% Record review and examination
(1959-1966) of Kaiser health plan patients
Kerrebijn5 Leiden, Netherlands 1,817 3.3 40% Author screened most patients;
(1958) also some referrals.
* Abbreviations used are: CHD, congenital heart disease; BDMP, Birth Defects Monitoring Program; MACDP,
Metropolitan Atlanta Congenital Defects Program.
t Secular changes; see text.
:1:Includes stillbirths.

methods than is the absolute incidence of VSD. TABLE 2. Ventricular Septal Defect (VSD) as a Pro-
This can be seen in Table 2: the relative incidence portion of Total Congenital Heart Disease in Pediatric
Case Series
of VSD is consistently about 30% in case series
from around the world, just as it is in the popula- First Author of Place (Ethnic Group) Total No. % VSD
Reference of Cases
tion-based studies summarized in Table 1.
Ta?7 Singapore 100 22
Schnire28 Cape Town, S Africa
Genetic Factors (white) 420 26
A genetic contribution to a birth defect may be (Colored) 262 27
(Bantu) 43 26
suspected when the defect runs in families. How- Wallooppi1lai Colombo, Ceylon 259 27
ever, such clustering does not prove a genetic etiol- Anderson3#{176} Northern US (Chip- 68 28
ogy because environmental factors, such as mater- pewa and Sioux)
nal diabetes or alcoholism, can also lead to familial Pailt* Manipal, India 200 29
Shann32 Taipei, Taiwan 586 32
clustering. In addition, although most VSDs occur
Caddell33 Uganda, Africa 44 36
as isolated defects, there are dozens of malforma- (8 native tribes)
tion syndromes that include VSD. These syn- Caddell Nigeria, Africa 67 39
dromes often have 25% to 50% recurrence risks, so (Yoruba)
that inclusion of even a few such patients in the Tofler35 Perth, Australia 41 39
(Aboniginals)
index group will lead to overestimation of the re-
* Case series includes some adults.
currence risk of isolated VSD.
Several studies have estimated the incidence of
CHD in first degree relatives of patients with
VSDs incidence of VSD, about 0.6% to 1% is also about
(Table 3). About one third of affected relatives have three times the population VSD risk (0.2% to 0.3%).
VSDs, one third have related types of CHD, and Children and parents of patients with VSD also
the remainder have diverse, seemingly unrelated, appear to be at increased risk, but there are fewer
types of CHD. Siblings of patients with VSD have data and results are less consistent. Of particular
an overall incidence of CHD of about 2% to 3%, interest is the study by Whittemore et al,44 which
compared with the population CHD incidence of found a 22% incidence of CHD in 78 offspring of
0.6% to 1%, for a risk ratio of about 3. Their mothers with VSD. Although a high incidence was

744 VENTRICULAR SEPTAL DEFECTS

Downloaded from by guest on July 13, 2017
 TABLE 3. Reported Incidence of Congenital Heart Disease in First Degree Relatives of
Patients with Ventricular Septal Defect (VSD)
First Author of No. of Siblings With No. of Children With No. of Parents With
Reference VSD/Total No. of VSD/Total No. of VSD/Total No. of
Siblings (%) Children (%) Parents (%)
Anderson 29/1,396 3/1,144
(2.1) (0.26)
Ando37 18/1,237 0/75
(1.5) (0)
Campbelll9* 6/359
(1.7)
Dennis 2/104
(1.9)
LamyS* 1/213
(0.5)
Laursen2l* 23/1,237
(1.9)
Mitchell40 0/18 1/98
(0) (1)
Nora443* 28/672 7/174 9/424
(4.2) (4.0) (2.1)
Whittemore44 17/78
(22)
105/5,114 26/449 13/1,666
(2.1) (5.2) (0.8)
* Studies in which probands with malformation syndromes were not clearly excluded.

to be expected from the study design, which in- TABLE 4. Reported Concordance Rates for Ventricular
volved sequential examinations beginning at birth Septal Defect (VSD) and Congenital Heart Disease
by a suspicious cardiologist, the observed incidence
(CHD) in Twins
is still considerably higher than would be expected First VSD CHD
Author
from other studies. Additional studies of offspring MZ DZ ?Z MZ DZ ?Z
of patients with CHD will be needed to explain this Reference
discrepancy. Anderson46 0/11 0/9 0/2 4/59 1/43 0/5
Studies of twins can be helpful in estimating the Ando3T 11/47 3/39 2/27
relative importance of genetic and environmental Campbell47 0/2 0/4 0/2 0/10
risk factors for disease, especially if zygosity is Dennis 0/3 0/1 0/2 0/10 0/3
Hay48 1/30 6/99
accurately determined.45 However, it is difficult to
Kenna6 1/5 0/2t 2/15 2/14t
accumulate an adequate number of twins with a Lam?9 0/2 0/2 0/8 1/6
particular birth defect, even one as common as Nora49 1/1 0/4 6/13 1/24 1/6
VSD. Furthermore, because twinning itself may be Ross#{176} 2/11 2/26
associated with CHD, conclusions drawn from con- Uchida5t PLa 0/13 0/13

cordance rates in twins may not be generalizable to

2/24 0/20 #{246}7
25/172 11/207 9/152
(8%) (0%) (0%) (15%) (5%) (6%)
singletons. * For probands with either VSD or any CHD, the pair is
The results of studies of VSD and CHD in twins
counted as concordant if the cotwin has any type of CHD.
are listed in Table 4. Only tentative conclusions are Abbreviations used are: MZ, monozygotic twins; DZ, di-
possible because the studies had differing inclusion zygotic twins; ?Z, unknown zygosity.
criteria and methods of case ascertainment. Iden- t A tnizygotic set of triplets (one VSD, two normal) is
tical twins of patients with CHD are apparently at counted as two discordant dizygotic twin pairs.
increased risk of CHD (25/156 [16%] of pairs con-
cordant), even compared with dizygotic twins or due to environmental factors to which only the
siblings. Although the numbers are small, twins of affected twin was exposed, but for all of the com-
patients with VSD also seem to be at increased risk, monly mentioned environmental factors,43 such as
with 2/24 (8%) of reported pairs concordant. maternal exposure to chemicals or infectious
The most striking feature of Table 4 is the fre- agents, both twins would be exposed. Thus, some-
quency of discordance, even in identical twins. Such thing more must be involved than genetic factors
discordance cannot be due to genetic differences, and the types of environmental factors that twins
because the twins have the same genes. It could be share.

Downloaded from by guest on July 13, 2017 ARTICLES 745

 The degree to which genetic and familial environ- fants with VSD were an average of 580 g smaller
mental factors contribute to a disease can be esti- and 2 weeks less mature than normal infants. They
mated. Falconer52 and Smith53 have described a suggested that some VSDs may close late in gesta-
model in which liability to a disease is a continuous tion and that these would be revealed by premature
inheritable trait that is normally distributed in the birth.
general population and in the relatives of patients Although some congenital heart lesions have dif-
with the disease but with different means in the ferent incidences in males and females, VSD ap-
two groups. If the incidence in relatives is compared pears to be equally distributed by sex. In five studies
with that in the general population, an estimate of of 820 infants,46193955 the percentage of males
the heritability of liability to the disease can be ranged from 465%16 to 62.3%, averaging 52.7%
made. For VSD, the population incidence of 0.3% (not significantly different from 50%: z = 1.27;
and incidence in siblings of 1% (one third of recur- P> .2).
rences) gives h2 of 30% to 35%, similar to that of
CHD.54 Compared with other diseases, such as Animal Studies
schizophrenia (85%), pyloric stenosis (75%), and
VSDs have been found to occur naturally in
club foot (68%), the heritability of VSD is low.54
birds and in many mammals including dogs,69
Maternal Factors cattle,7#{176}
rats,7 goats,72 cats,73 and horses.74 In some
animals, certain strains have been shown to be
The associations between VSD and maternal age
genetically predisposed to particular cardiac mal-
and birth order have been investigated with nega-
formations.7175 Many chemicals can cause VSDs in
tive results.69395556 Although the power of the
animals, if they are injected at the proper time of
individual studies to detect an association was low,
gestation. However, even among littermates in
taken together they suggest that when patients with
highly inbred strains, only some will have the de-
Down Syndrome are excluded, there is no clinically
fect, although all were exposed.76 Thus, the obser-
significant association between VSD and birth or-
vations in animals mirror those in humans: VSDs
der or maternal age.
occur in very different populations, they are partly
Neither race nor socioeconomic status is associ-
genetic, they may be caused by teratogens, and
ated with VSDs. Schrirem found nearly identical
fetuses with the same genes exposed at the same
relative incidence rates for VSD among white, col-
time to the same environment may be differentially
ored, and Bantu patients in Cape Town (Table 2),
affected.
and Heinonen57 found similar incidence rates of
VSD and CHD in whites, blacks, and Puerto Ricans
DISCUSSION
and no association with socioeconomic status in the
Collaborative Perinatal Study. Multifactonal Inheritance Hypothesis
Many agents and conditions are known or sus-
According to the multifactorial inheritance hy-
pected causes of birth defects, but only a few envi-
pothesis proposed by Nora,77 genetic predisposition
ronmental causes of CHD were identified in a re-
interacts with an environmental trigger to cause
cent review by Nora and Nora,43 and not all of these
VSDs. The model accounts for the familial nature
have been shown to cause VSD. Thalidomide,58
of VSD and explains the difficulty of identifying
rubella infection,59 and maternal alcoholism60 and
environmental factors, because only a small portion
phenylketonuria6 result in an increased frequency
of the population (those with the genetic predispo-
of VSDs, often accompanied by other defects. Cig-
sition) would be susceptible to each such factor.
arette smoking was found to be weakly associated
However, the multifactonial inheritance hypoth-
with congenital heart defects (RR = 1.5) in one
esis leaves some epidemiologic observations unex-
large study62 but not in others.5763 Exogenous sex
plained. The incidence of VSD is remarkably uni-
hormones have also been associated with CHD
form by sex and race, despite genetic differences.
(primarily with defects other than simple VSD65),
Most environmental factors vary in different set-
but studies of this association have also been con-
tings, but the incidence of VSD does not. Yet mon-
flicting.7 If environmental factors are important
ozygotic twins, whose environments in the same 4-
causes of VSDs, most of them have yet to be dis-
to 6-week uterus are nearly identical, need somehow
covered.
to be differentially exposed to environmental trig-
gers for this model to explain their lack of concord-
Fetal Factors
ance.
Both prematurity and low birth weight are asso- As evidence for the multifactorial inheritance
ciated with VSDs. Mitchell et al,8 using data from hypothesis, Nora and co-workers note that, for each
the Collaborative Perinatal Study, found that in- type of heart defect, the recurrence risks for CHD

746 VENTRICULAR SEPTAL DEFECTS

Downloaded from by guest on July 13, 2017
approximate the square root of the population in- esis. For VSDs, this probability may be fairly high.
cidence,427778 as predicted by Edwards.79 However, As Jackson#{176} wrote,
Edwardss approximation assumes a heritability of perhaps it is the necessity of normal growth in three
100%-an assumption clearly violated for CHD. separate areas of the heart, and the proper timing of such
Furthermore, Anderson36 has pointed out that it is growth for final division of the ventricle that results in
such frequent occurrence of ventricular septal defects.
inconsistent to compare (as Nora et al do) recur-
rence risk for CHD, which includes the incidence From a preventive medicine viewpoint, this hy-
for all types of CHD combined, with the population pothesis is discouraging, but it has a certain intui-
incidence of one specific defect. When recurrence tive appeal. The construction of a four-chambered
rates and incidence rates for either one specific heart is a difficult task, and occasional mistakes
lesion or for all types of CHD combined are com- are inevitable. The frequency with which mistakes
pared, the model no longer fits. (Actually, even the are made depends on the quality of the instructions
recurrence risk data given by Nora provide no evi- (genetic factors) and the conditions under which
dence to contradict the null hypothesis that all the task needs to be accomplished (environmental
types of congenital heart disease have the same factors). During millions of years, mammals have
recurrence risk. Analysis of Nora and Noras437 evolved the best possible set of instructions, but
Table 4 gives x2 = 13.79, df = 14, P > .4.) Thus, it there are still occasional errors, especially in the
is not surprising that the correspondence between difficult parts, like forming ventricular and atrial
the recurrence risk and the square root of the septa and connecting the great vessels. In some
population incidence reported by Nora et al. has families, the instructions are of lower quality, but
not been observed by other investigators.3637 not enough so that overall survival has been im-
paired, because only a few members of such families
Genetic Variant Hypothesis are unlucky enough to get CHD.
The chance + genes + environment model ex-
Taussig74 proposed another model that accounts plains the epidemiologic observations of VSD. VSD
for the lack of variation in CHD incidence. She incidence rates are uniform because the basic task
noted the similarity of CHD incidence among dif- of forming the ventricular septum is the same in all
ferent peoples around the world, and in six orders people (and similar in other mammals and in birds).
of mammals, and hypothesized that cardiac malfor- Patients with a family history of CHD and patients
mations are genetic variants rather than develop- exposed to teratogens are at higher risk, because
mental errors. Because mammals evolved from rep- adverse genes and environment can increase the
tiles approximately 60 to 70 million years ago, she baseline risk of VSD. Identical twins are generally
proposed that the genes for primeval hearts (that discordant, however, because, even with genetic and
we now call CHD) have remained in the genetic environmental risk factors, most individuals are
pool since then. lucky enough to escape a VSD.
Although Taussigs model addresses the striking
similarity of CHD observed in different types of IMPLICATIONS
animals with four-chambered hearts, it leaves two
important questions unanswered. Why are mono- Prevention
zygotic twins not concordant for this genetic var-
If the proposed model is correct, a portion (which
iant? Why have the genes that code for such hearts
may be substantial for VSD) of CHD is not pre-
persisted in a similar fraction of the population in
ventable-the price we pay for having a four-cham-
each of several orders of mammals, in spite of the
bered heart. However, we should not diminish our
obvious selective disadvantage?
vigilance for changes in CHD incidence rates. Car-
diovascular teratogens definitely exist and new ter-
Chance + Genes + Environment Hypothesis atogens could enter the environment at any time.
Even if we are unable to reduce the rate of VSD,
Normal embryologic events depend on a long
we can try to prevent an increase in incidence. The
sequence of steps that begins at conception. Each
incidence trend for VSD reported by the Centers
of these steps has an extremely high probability of
for Disease Control thus deserves close attention,
occurring, but development is so complicated that
particularly if it continues.
the number of steps required is also extremely high.
Without implicating any etiologic factor other than
Counseling
the complexity inherent in normal development, we
can predict that there will be a finite probability of Parents of children with VSD should continue to
any given defect, which depends on the particular be counseled that CHD runs in families, that sub-
sequence of steps necessary for normal morphogen- sequent children will be at increased risk, and that

Downloaded from by guest on July 13, 2017 ARTICLES 747

the risk to offspring of patients with VSD may be disease in Liverpool: 1960-69. Q J Med 1975;44:17-44
17. Slater CS, Watson GI, McDonald JC: Seasonal variation in
considerably higher. A careful prenatal history congenital abnormalities: Preliminary report of a survey
should be obtained for all infants with birth defects, conducted by the research committee of the council of
because identification of previously undiscovered general practitioners. Br J Prey Soc Med 1964;18:1-7
18. Rothman KJ, Fyler DC: Association of congenital defects
teratogens depends on the continued vigilance of with season and population density. Teratology 1975;13:29-
parents and clinicians. However, if nothing suspi- 34
cious is found, we no longer need to tell parents 19. Campbell M, Goodwin JF: Some factors in the etiology of
ventricular septal defect. Prog Cardiovasc Dis 1965;7:417-
that, along with the familial factors, an unknown 433
environmental trigger caused the disease. Telling 20. Edwards JH: Seasonal incidence of congenital disease in
parents about the uniform incidence around the Birmingham. Ann Hum Genet 1961;25:89-93
21. Edwards JH: The recognition and estimation of cyclic
world, as well as the occurrence in animals, and trends. Ann Hum Genet 1961;25:83-87
about the lack of concordance in twins may help 22. Rose V, Hewitt D, Milner J: Seasonal influences on the risk
them to reduce the guilt-induced search of their of cardiac malformation: Nature of the problem and some
memories for something they ate, drank, or
results from a study of 10,077 cases. mt Epidemiol
1972;1:235-244
breathed that might have caused the defect. 23. Laursen HB: Some epidemiologic aspects ofcongenital heart
disease in Denmark. Acta Paediatr Scand 1980;69:619-624
24. Dickinson DF, Arnold R, Wilkinson JL: Ventricular septal
ACKNOWLEDGMENTS defect in children born in Liverpool 1960-1969: Evaluation
of natural course and surgical implications in an unselected
This research was funded by the A. W. Mellon Foun- population. Br Heart J 1981;46:47-54
dation. I wish to thank Dr Julien Hoffman, the faculty 25. Feldt RH, Avasthey P, Yoshimasu F, et al: Incidence of
and students of the Division of Epidemiology at the congenital heart disease in children born to residents of
School of Public Health of the University of California, Olmsted County, Minnesota, 1950-1969. Mayo Clin Proc
Berkeley, and my co-workers in the A. W. Mellon Clinical 1971;46:794-799
26. Carlgren L-E: The incidence of congenital heart disease in
Epidemiology Program, especially Drs Warren S.
children born in Gothenburg 1941-1950. Br Heart J
Browner and Stephen B. Hulley for their helpful sugges- 1959;21:40-50
tions. 27. Tay JSH: Parental age and birth order in Chinese children
with congenital heart disease. J Med Genet 1982;19:441-443
REFERENCES 28. Schrire V: Experience with congenital heart disease at
GrooteSchuur Hospital, Capetown: An analysis of 439 pa-
1. Pinkley K, Stoesz PA: Current trends: Ventricular septal tients over an 11 year period. S Afr Med J 1963;37:1175-
defect. MMWR 1981;30:609-610 1180
2. Rothman KJ, Fyler DC: Seasonal occurrence of complex 29. Wallooppillai NJ, Jayasinghe M de 5: Congenital heart
ventricular septal defect. Lancet 1974;2:193-197 disease in Ceylon. Br Heart J 1970;32:304-306
3. Rosenberg LA, Heinonen OP: Seasonal occurrence of yen- 30. Anderson RC: Congenital heart malformations in North
tricular septal defect. Lancet 1974;2:903-904 American Indian children. Pediatrics 1977;59:121-123
4. Carigren L-E: The incidence of congenital heart disease in 31. Pai BV, Varkey CC: Spectrum of congenital heart disease
Gothenburg. Proc Assoc Europ Peed Cardiol 1969;5:2-8 in a general hospital: Study of 200 cases. Indian J Pediatr
5. Bound JP, Logan WFWE: Incidence of congenital heart 1974;41:317-321
disease in Blackpool 1957-71. Br Heart J 1977;39:445-450 32. Shann MKM: Congenital heart disease in Taiwan, Republic
6. Anderson CE, Edmonds LD, Erickson JD: Patent ductus of China. Circulation 1969;39:251-258
arteriosus and ventricular septal defect: Trends in reported 33. Caddell JL, Conner DH: Congenital heart disease in Ugan-
frequency. Am J Epidemiol 1978;107:281-289 dan children. Br Heart J 1966;28:766-767
7. Edmonds LD, Layde PM: Congenital malformation surveil- 34. Caddell JL, Morton P: Pattern of congenital heart disease
lance: Two American systems. mt Epidemiol 1981;10:247- in Yoruba children of Western Nigeria. Am Heart J
252 1967;73:431-432
8. Mitchell SC, Berendes HW, Clark WM: The normal closure 35. Tofler OB: Congenital heart disease in Aboriginals, letter.
of the ventricular septum. Am Heart J 1967;73:334-338 Med J Aust 1979;13:620-622
9. Layde PM, Dooley K, Erickson JD, et al: Is there an epi- 36. Anderson RC: Fetal and infant death, twinning and cardiac
demic of ventricular septal defects in the U.S.A.? Lancet malformations in families of 2000 children with and without
1980;1:407-408 cardiac defects. Am J Cardiol 1976;38:218
10. Mitchell SC, Korones SB, Berendes HW: Congenital heart 37. Ando M, Takao A, Mon K: Genetic and environmental
disease in 56,109 births: Incidence and natural history. factors in congential heart disease, in Inouye E, Nishimura
Circulation 1971;43:323-332 H (eds): Gene-Environment Interaction in Common Dis-
1 1. Hoffman JIE, Rudolph AM: The natural history of ventric- eases. Tokyo, University Park Press, 1977, pp 71-78
ular septal defect, with special reference to selection of 38. Dennis NR, Warren J: Risks to the offspring of patients
patients for surgery. Adv Pediatr 1970;17:57-79 with some common congenital heart defects. J Med Genet
12. Corone P, Doyon F, Gaudeau 5: Natural history of ventric- 1981;18:8-16
ular septal defect: A study of 790 cases. Circulation 39. Lamy M, DeGrouchy J, Schweisguth 0: Genetic and non-
1977;55:908-915 genetic factors in the etiology of congenital heart disease: A
13. Campbell M: The natural history ofventricular septal defect. study of 1188 cases. Am J Hum Genet 1957;9:17-41
Br Heart J 1971;33:246-257 40. Mitchell SC, Sellmann AH, Westphal MC, et al: Etiologic
14. Hoffman JIE, Christianson R: Congenital heart disease in a correlates in a study of congenital heart disease in 56109
cohort of 19,502 births with long term follow-up. Am J births. Am J Cardiol 1971;28:653-657
Cardiol 1978;42:641 41. Nora JJ, Dodd PF, McNamara DG, et al: Risk to offspring
15. Kerrebijn KF: Incidence in infants and mortality from con- of parents with congenital heart defects. JAMA
genital malformations of the circulatory system. Acta Pae- 1969;209:2052-2053
diatr Scand 1966;55:316-320 42. Nora JJ, McGill CW, McNamara DG: Empiric recurrence
16. Kenna AP, Smithells RW, Fielding DW: Congenital heart risks in common and uncommon congenital heart lesions.

748 VENTRICULAR SEPTAL DEFECTS

Downloaded from by guest on July 13, 2017
 Teratology 1970;3:325-329 63. Andrews J, McGarry JM: A community study of smoking
43. Nora JJ, Nora AH: Genetic epidemiology ofcongenital heart during pregnancy. J Obstet Gynaecol Br Commonw
diseases. Prog Med Genet 1983;5:91-137 1972;79:1057-1073
44. Whittemore R, Hobbins JC, Engle MA: Pregnancy and its 64. Nora JJ, Nora AH, Blu J, et al: Exogenous progestogen and
outcome in women with and without surgical treatment of estrogen implicated in birth defects. JAMA 1978;240:837-
congenital heart disease. Am J Cardiol 1982;50:641-651 842
45. Hrubec Z, Robinette CD: The study of human twins in 65. Janerich DT, Dugan JM, Standfest SJ, et al: Congenital
medical research. N EngI J Med 1984;310:435-441 heart disease and prenatal exposure to exogenous sex hor-
46. Anderson RC: Congenital cardiac malformations in 109 sets mones. Br Med J 1977;1:1058-1060
of twins and triplets. Am J Cardiol 1977;39:1045-1050 66. Rothman KJ, Fyler DC, Goldblatt A, et al: Exogenous
47. Campbell M: Twins and congenital heart disease. Acta Genet hormones and other drug exposures of children with and
Med 1961;10:443-456 without congenital heart disease. Am J Epidemiol
48. Hay 5, Wehrung DA: Congenital malformations in twins. 1979;109:433-439
Am J Hum Genet 1970;22:662-678 67. Ferencz C, Matanoski GW, Wilson PD, et al: Maternal
49. Nora JJ, Gilliland JC, Sommerville RI, et al: Congenital hormone therapy and congenital heart disease. Teratology
heart disease in twins. N EngI J Med 1967;277:568-571 1980;21:225-239
50. Ross U: Congenital cardiovascular anomalies in twins. Cir- 68. Siller WG: Congenital heart defects in fowl. Proc R Soc Med
culation 1959;20:327-342 1953;61:1289
51. Uchida IE, Rowe RD: Discordant heart anomalies in twins. 69. Mulvihill JJ, Priester WA: Congenital heart disease in dogs:
Am J Hum Genet 1957;9:133-140 Epidemiologic similarities to man. Teratology 1973;7:73-76
52. Falconer DS: The inheritance ofliability to certain diseases, 70. Fisher EW, Pine HM: Cardiovascular lesions in cattle. Ann
estimated from the incidence among relatives. Ann Hum NYAcad Sci 1965;127:606-622
Genet 1965;29:51-76 71. Fox MH: Genetic transmission of congenital membranous
53. Smith C: Heritability of liability and concordance in mon- ventricular septal defects in selectively inbred substrains of
ozygous twins. Ann Hum Genet 1970;34:85-91 rats. Circ Res 1967;20:422-433
54. Emery EH: Methodology in Medical Genetics: An Introduc- 72. Parry BW, Wrigley RH, Reuter RE: Ventricular septal
tion. Edinburgh, England, Churchill Livingston, 1976, p 54 defects in three familially related female Saanen goats. Aust
55. Rothman KJ, Fyler DC: Sex, birth order, and maternal age Vet J 1982;59:72-76
characteristics of infants with congenital heart defects. Am 73. Bolton GR, Liu S-K: Congenital heart diseases of the cat.
J Epidemiol 1976;104:527-534 Vet Clin North Am 1977;7:341-354
56. MacMahon B: Association of congenital malformation of 74. Taussig HB: World survey of the common cardiac malfor-
the heart with birth rank and maternal age. Br J Soc Med mations: Developmental error or genetic variant? Am J
1952;6:178-182 Cardiol 1982;50:544-559
57. Heinonen OP: Risk factors for congenital heart disease, in 75. Patterson DF, Pyle RL, Van Mierop L, et al: Hereditary
Kelly S (ed): Birth Defects: Risks and Consequences. New defects of the conotruncal septum in Keeshond dogs: Patho-
York, Academic Press, 1976, pp 221-264 logic and genetic studies. Am J Cardiol 1974;34:187-205
58. Warkany J: Congenital malformations: Notes and Comments. 76. Nora JJ, Sommerville RJ, Fraser FC: Homologies for con-
Chicago, Year Book Medical Publishers, 1971, pp 465-466 genital heart disease: Murine models, influenced by dextro-
59. Smith DW, Jones KL: Recognizable Patterns of Human amphetamine. Teratology 1968;1:413-416
Malformation. Genetic, Embryologic, and Clinical aspects, ed 77. Nora JJ: Multifactorial inheritance hypothesis for the etiol-
3. Philadelphia, WB Saunders Co, 1982, pp 424-425 ogy of congenital heart disease: The genetic environmental
60. Hanson JW, Jones KL, Smith DW: Fetal alcohol syndrome: interaction. Circulation 1968;38:604-617
Experience with 41 patients. JAMA 1976;235:1458 78. Nora JJ, Nora AH: Recurrence risks in children having one
61. Lenke RR, Levy HL: Maternal phenytlketonuria and hy- parent with congenital heart disease. Circulation
perphenylalaninemia: An international survey of the out- 1976;53:701-702
come of treated and untreated pregnancies. N EngI J Med 79. Edwards JH: The simulation of Mendelism. Acta Genet
1980;303:1202-1208 1960;10:63-70
62. Fredrick J, Alberman ED, Goldstein H: Possible teratogenic 80. Jackson BT: The pathogenesis of congenital cardiovascular
effect of cigarette smoking. Nature 1971;231:529-530 anomalies. N EngI J Med 1968;279:25-29, 80-89

ARTICLES 749
Downloaded from by guest on July 13, 2017
 Etiology of Ventricular Septal Defects: An Epidemiologic Approach
Thomas B. Newman
Pediatrics 1985;76;741
Updated Information & including high resolution figures, can be found at:
Services /content/76/5/741
Citations This article has been cited by 7 HighWire-hosted articles:
/content/76/5/741#related-urls
Permissions & Licensing Information about reproducing this article in parts (figures, tables)
or in its entirety can be found online at:
/site/misc/Permissions.xhtml
Reprints Information about ordering reprints can be found online:
/site/misc/reprints.xhtml

PEDIATRICS is the official journal of the American Academy of Pediatrics. A monthly publication, it
has been published continuously since 1948. PEDIATRICS is owned, published, and trademarked by the
American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk Grove Village, Illinois, 60007.
Copyright 1985 by the American Academy of Pediatrics. All rights reserved. Print ISSN: 0031-4005.
Online ISSN: 1098-4275.

Downloaded from by guest on July 13, 2017

 Etiology of Ventricular Septal Defects: An Epidemiologic Approach
Thomas B. Newman
Pediatrics 1985;76;741

The online version of this article, along with updated information and services, is located on
the World Wide Web at:
/content/76/5/741

PEDIATRICS is the official journal of the American Academy of Pediatrics. A monthly publication,
it has been published continuously since 1948. PEDIATRICS is owned, published, and trademarked
by the American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk Grove Village,
Illinois, 60007. Copyright 1985 by the American Academy of Pediatrics. All rights reserved. Print
ISSN: 0031-4005. Online ISSN: 1098-4275.

Downloaded from by guest on July 13, 2017