NOMURA, GILMAN, AND BUKA 199
Maternal Smoking During Pregnancy and Risk of Alcohol
Use Disorders Among Adult Offspring*
YOKO NOMURA, PH.D., M.P.H.,† STEPHEN E. GILMAN, SC.D.,† AND STEPHEN L. BUKA, SC.D.†
Department of Psychology, Queens College, The City University of New York, 65-30 Kissena Boulevard, Flushing, New York, 11367
ABSTRACT. Objective: The aim of this study was to evaluate the asso-
ciation between maternal smoking during pregnancy (MSP) and lifetime
risk for alcohol use disorder (AUD) and to explore possible mechanisms
through which MSP may be related to neurobehavioral conditions during
infancy and childhood, which could, in turn, lead to increased risk for
AUD. Method: A sample of 1,625 individuals was followed from preg-
nancy for more than 40 years. Capitalizing on the long follow-up time,
we used survival analysis to examine lifetime risks of AUD (diagnosed
according to the Diagnostic and Statistical Manual of Mental Disorders,
Fourth Edition) in relation to levels of MSP (none, <20 cigarettes/day,
and ≥20 cigarettes/day). We then used structural equation modeling to
test hypotheses regarding potential mechanisms, including lower birth
weight, neurological abnormalities, poorer academic functioning, and
behavioral dysregulation. Results: Relative to unexposed offspring,
offspring of mothers who smoked 20 cigarettes per day or more exhib-
ited greater risks for AUD (hazard ratio = 1.31, 95% CI [1.08, 1.59]).
However, no differences were observed among offspring exposed to
fewer than 20 cigarettes per day. In structural equation models, MSP was
associated with neurobehavioral problems during infancy and childhood,
which, in turn, were associated with an increased risk for adult AUD.
Conclusions: MSP was associated with an increased lifetime risk for
AUD. Adverse consequences were evident from birth to adulthood. A
two-pronged remedial intervention targeted at both the mother (to reduce
smoking during pregnancy) and child (to improve academic functioning)
may reduce the risk for subsequent AUD. (J. Stud. Alcohol Drugs, 72,
plinary Tobacco Use Research Center award P50 CA084719; an American
Beverage and Medical Research Grant award; and grants from the National
Cancer Institute, the National Institute on Drug Abuse, the Robert Wood
Johnson Foundation, the National Institute of Aging (AG023397), and the
National Institute of Mental Health grant (K01MH080062).
via email at: firstname.lastname@example.org. Yoko Nomura is also with the De-
partment of Psychiatry, Division of Child and Adolescent Psychiatry, Mount
Sinai School of Medicine, New York, NY. Stephen E. Gilman is with the
Department of Society, Human Development, and Health, and Department
of Epidemiology, Harvard School of Public Health, Boston, MA. Stephen
L. Buka is with the Department of Community Health, Brown University,
Received: September 11, 2009. Revision: August 19, 2010.
*This work was supported by National Institutes of Health Transdisci-
†Correspondence may be sent to Yoko Nomura at the above address or
cause of death in the United States, with approximately
79,000 deaths annually attributable to excessive alcohol use
(Centers for Disease Control and Prevention, 2004). Chronic
drinking causes a variety of functional impairments and has
many harmful health consequences. These include liver dis-
ease (Heron, 2007; Schiff, 1997); cancer (Baan et al., 2007);
cardiovascular disease (Rehm et al., 2003); neurological
damage (Corrao et al., 2002, 2004); HIV infection (Rosen-
bloom et al., 2007; Windle, 1997); and psychiatric problems
such as depression, anxiety, and antisocial personality disor-
der (Castaneda et al., 1996; Kessler et al., 1997; Rosenthal
and Westreich, 1999). Chronic drinking is also associated
with an elevated risk for unintended accidents such as falls
(Goodman et al., 1991), drowning (Cummings and Quan,
1999), burns (Hingson and Howland, 1993; McGill et al.,
CCORDING TO THE Centers for Disease Control and
Prevention, alcohol is the third highest lifestyle-related
1995), fatal motor vehicle crashes (National Highway Traf-
fi c Safety Administration, 2008), and interpersonal violence
(Bushman, 1997; Caetano et al., 2000).
Various biological and psychological risk factors for
alcohol use disorder (AUD) have been identifi ed, including
alcohol metabolism, genetic risks, and psychosocial risks
(i.e., lack of parental monitoring, severe and recurrent fam-
ily confl ict, and poor parent–child relationships) (Quete-
mont, 2004; Wall et al., 2007). However, to the best of our
knowledge, maternal smoking during pregnancy (MSP)
has never been found to be a potential distal risk factor
for alcohol use problems and AUD despite its association
with a variety of adverse outcomes, including substance
use disorders, in later life. For instance, the adverse effect
of MSP on birth outcomes is well established, including an
approximately 150-250 g decrement in birth weight (Sub-
stance Abuse and Mental Health Services Administration,
2005; Visscher et al., 2003) and a higher neonatal mortality
rate (Duncan et al., 2008; Fleming and Blair, 2007). Infants
exposed to MSP also displayed an elevated risk for sud-
den infant death syndrome (Edner et al., 2007; Markowitz,
2007; Weese-Mayer et al., 2007), neurological and lan-
guage problems (Fried, 1993; Fried et al., 1992a, 1992b),
diffi cult temperaments (Brook et al., 1998), aggression
(Tremblay et al., 2004), behavioral problems (Maughan et
al., 2004; Monuteaux et al., 2006; Orlebeke et al., 1997),
cognitive function defi cits (Keeping et al., 1989; Naeye and
Peters, 1984; Nomura et al., 2008), attention defi cits and
hyperactivity (Button et al., 2005; Linnet et al., 2005; Wak-
schlag et al., 1997), early onset of delinquency and antiso-
cial behavioral problems (Nomura et al., 2009; Piquero et
200 JOURNAL OF STUDIES ON ALCOHOL AND DRUGS / MARCH 2011
al., 2002; Wakschlag et al., 2003; Weitzman et al., 1992),
cigarette smoking in adolescence and adulthood (Buka
et al., 2003; Cornelius et al., 2000, 2005; Griesler et al.,
1998), and drug use (Ekblad et al., 2010) Fergusson et al.,
1998; Weissman et al., 1999). These fi ndings all indicate
that MSP may indeed impair the development of the fetal
central nervous system in a fashion that may predispose
the offspring to a wide array of neurobehavioral problems.
To date, few studies have tried to systematically elucidate
potential mechanisms by which MSP may impair the child’s
central nervous system early in life and the subsequent in-
creased risk for AUD. Furthermore, we are unaware of prior
studies that have examined whether there is a direct link
between MSP and AUD in adulthood or if their association
can, at least in part, be explained by problems in the preced-
ing stages of life. Approximately 10-20% of women smokers
who become pregnant continue to smoke during pregnancy
(DiFranza and Lew, 1995; Maughan et al., 2004), and MSP
is one of the early modifi able risk factors that could poten-
tially reduce the incidence of adverse outcomes throughout
the life course (Ringel and Evans, 2001; Shiono and Beh-
In this study, we used data from a population-based
sample of children who have been followed more than 40
years to address two aims: (a) to evaluate the lifetime risk
for Diagnostic and Statistical Manual of Mental Disorders,
Fourth Edition (DSM-IV; American Psychiatric Associa-
tion, 1994), AUD associated with exposure to MSP and (b)
to elucidate possible mechanisms through which MSP af-
fects neurobehavioral conditions during early infancy and
childhood (i.e., birth weight, neurological abnormalities,
academic functioning, and behavioral regulation), which
may lead to an increased risk for AUD. We hypothesized
that there will be a greater risk for AUD among offspring
exposed to MSP than those unexposed. We also hypothesized
that MSP is associated with increased risk for AUD, and that
this association will be mediated in part through problems in
infancy and childhood.
ect (CPP), consisting of prospective data collected from
a representative sample of pregnant women who received
prenatal care and delivered their babies during 1960-1966
(Niswander and Gordon, 1972). The CPP used a single study
design across all 12 sites. The sites participated in a system-
atic data collection from pregnancy to the fi rst 7 years to
identify perinatal and early childhood factors that adversely
affect subsequent child development (Buka et al., 1993).
Reports from the CPP have been summarized elsewhere
(Nicholas and Chan, 1981).
Data were derived from the Collaborative Perinatal Proj-
was established to locate and interview a sample of the adult
CPP offspring at the Providence, RI, and Boston, MA, sites.
Participants were selected through a multistage sampling
procedure, which involved a core assessment interview and
three component studies. Screening questionnaires were
mailed to 4,579 of the 15,721 Providence and Boston CPP
offspring who lived beyond 7 years of age. Of the 3,121
questionnaires that were returned (68.2%), 2,271 were eligi-
ble for participation based on the combined inclusion criteria
of the three component studies. Those who enrolled had a
somewhat higher level of education (e.g., 64.1% with at least
some college education) than participants who were eligible
but not enrolled (e.g., 51.8% with at least some college
education). Data from 49 of the individuals were excluded
from the fi nal sample either because of their participation in
a pilot version of the survey (n = 4) or because of problems
with the interview administration (n = 45). This resulted in
1,625 completed adult assessments (Gilman et al., 2008a,
2008b). As part of the study design, siblings were over-
sampled. The fi nal sample included these 1,625 offspring
of 1,254 mothers; analyses were conducted to account for
these sibling sets.
Between 2001 and 2004, the New England Family Study
visit, women reported whether they were currently smoking
and, if so, the number of cigarettes they smoked per day.
These questions were repeated at each subsequent prena-
tal visit up until the time of delivery. From these repeated
measurements, we determined the maximum number of
cigarettes smoked per day at any point during pregnancy.
Women were then classifi ed into three levels of smoking:
never smoked during any pregnancy day (coded 0), smoked
fewer than 20 cigarettes during any pregnancy day (coded
1), and smoked 20 cigarettes or more during any pregnancy
day (coded 2). This categorical smoking variable was used
in this article.
Birth weight. Birth weight was recorded in grams by a
nurse observer at the time of delivery.
Neurological abnormality at age 1. A trained pediatrician
or pediatric neurologist performed a neurological evaluation
of the child and screened for a variety of potential develop-
mental abnormalities when the child was approximately 1
year old (50-56 weeks). There were 116 items that were used
to characterize the child’s neurological status. Neurological
abnormality at age 1 was defi ned as the number of abnor-
malities coded either abnormal or suspect.
Academic functioning at age 7. The Wide Range Achieve-
ment Test measured learning (dis)abilities (i.e., reading,
arithmetic, and spelling; Jastak and Jastak, 1965). We used
the standardized scores for this analysis. The mean (SD)
scores for the three areas were the following: reading, 104.75
Maternal smoking during pregnancy. At the fi rst prenatal
NOMURA, GILMAN, AND BUKA 201
(16.39); arithmetic, 99.44 (9.80); and spelling, 100.87
(13.51). The ranges were 70-165, 50-165, and 63-165, re-
spectively. The three separate Wide Range Achievement Test
scores formed a single factor (eigenvalue = 2.43), namely
Behavioral regulation. Trained child psychologists evalu-
ated the child’s behavioral functioning in 15 domains using
a 5-point Likert scale. Those 15 areas include shyness,
fearfulness, dysinhibition, self-confi dence, emotional reac-
tivity, degree of cooperation, frustration tolerance, degree
of dependency, attention span, goal orientation, activity
level, nature of communication, impulsivity, assertiveness,
and hostility. Because MSP has been more closely linked to
self-regulatory control defi cits (Huijbregts et al., 2008; Vuijk
et al., 2006), we chose dysinhibition, emotional reactivity,
activity level, and impulsivity as our behavioral measures.
These four observed scores formed a single factor (eigen-
value = 2.17), namely “behavioral regulation.”
Clinical diagnosis of alcohol use. The lifetime occurrence
of three distinctive DSM-IV symptoms clusters—namely
alcohol abuse, alcohol dependence, and alcohol withdraw-
al—were assessed with an expanded version of the AUD
module of the Composite International Diagnostic Interview
(World Health Organization, 1993, 1997). There were four
symptoms of alcohol abuse, seven symptoms of alcohol de-
pendence, and eight symptoms of alcohol withdrawal. Diag-
nosis of AUD was generated based on DSM-IV criteria. Age
at onset for AUD (abuse or dependence) was defi ned as the
age reported for the fi rst symptom of AUD (for details, see
Dierker et al., 2007). AUD is our primary outcome in both
survival analysis and structural equation modeling (SEM).
In survival analysis, additionally, risks for alcohol abuse and
alcohol dependence were examined.
Potential confounders and missing values. We chose
socioeconomic status, child gender and race, and mother’s
self-reported mental health status during pregnancy as the
most pertinent potential demographic and maternal charac-
teristic confounders, because they are known to be associated
with birth weight, neurological abnormality, and childhood
problems with behavior and learning (Gilman et al., 2008a).
They were included in all analyses for statistical adjustment.
Information on parental race/ethnicity and socioeconomic
status was collected during the fi rst prenatal visit. A compos-
ite index was calculated on the basis of methods developed
by the U.S. Census Bureau. The index refl ects the education
and occupation of the head of household, along with house-
hold income (Myrianthopoulos and French, 1968). Mother’s
self-reported mental health status was recoded as none,
hospitalized, receiving outpatient treatment, or alcohol/drug
Rates of missing data were 6.6% for AUD diagnostic
outcome, 2.0% for age at fi rst alcohol-abuse symptom, and
0.2% for age at fi rst alcohol-dependence symptom. Missing
data for predictors were approximately 3.0% for Wide Range
Achievement Test, Wechsler Intelligence Scale for Children,
and behavioral observation scores at age 7. Missing data for
other covariates were 4.2% for family socioeconomic status
and 6.4% for mother’s self-reported mental health status dur-
ing pregnancy. There were no missing data for gender, race,
birth weight, MSP, or age at the time of interview.
our hypothesized models using survival analysis and SEM
techniques. First, using survival analysis techniques, the Wil-
coxon test showed the overall differences for proportion of
subjects free from AUD, alcohol abuse, and dependence over
time, for the three MSP groups. Cumulative lifetime rates of
AUD, as well as alcohol abuse and dependence separately,
were then evaluated by the Kaplan-Meier method (Williams,
1995). To estimate the risk of AUD, and alcohol abuse and
dependence separately, among offspring exposed to MSP
(<20 cigarettes/day and ≥20 cigarettes/day), compared with
that among offspring unexposed to MSP, Cox proportional
hazards regression models (Binder, 1992; Cox, 1972) were
fi tted using SUDAAN (Shah et al., 1997) to account for po-
tential nonindependence of outcomes for offspring from the
Second, using SEM we attempted to explain our fi nd-
ings regarding the association between MSP (none, <20
cigarettes/day, and ≥20 cigarettes/day) and the risk for AUD
through childhood neurological, cognitive, and behavioral
problems, as well as the direct effect of MSP on the in-
creased risk for AUD. SEM allows simultaneous testing of
all of the associations among the different risk factors stud-
ied and hence the assessment of direct and indirect associa-
tions of all predictors, while taking into account a variety of
control variables (Linver et al., 2002). We used the software
Mplus (van Horn et al., 2009; Muthén and Muthén, 1998-
2007) to adjust for potential nonindependence of outcomes
for offspring from the same family and to normalize our
nonlinear (i.e., dichotomous) outcome. The transformation
used in Mplus is the logit function, which is the natural
log of the odds. As we have done in the survival analysis,
potential nonindependence of outcomes for offspring from
the same family has been adjusted, using Mplus (van Horn
et al., 2009). We assessed two childhood risk constructs (i.e.,
academic functioning and behavior regulation) as forms of
latent variables, which are hypothetical underlying constructs
that cannot be measured directly (MacCallum and Austin,
2000). Behavioral regulation was measured by four observed
variables, and academic functioning was measured by three.
Our fi nal outcome measure of AUD in adulthood was a
dichotomous diagnostic outcome. SEM with latent variable
modeling has been shown to be useful in situations in which
(a) measurement error is an issue, (b) the phenomena under
study are not directly observed, and (c) multiple indicators
After an initial descriptive univariate analysis, we tested
202 JOURNAL OF STUDIES ON ALCOHOL AND DRUGS / MARCH 2011
are needed to describe various aspects of a phenomenon
(Muthen, 1992). Our two hypothesized latent constructs
fi t all of the previously described circumstances, thus war-
ranting the use of SEM with latent variables. Because the
constructs to be examined are based on latent variables,
to maintain the validity of the two latent variables, we ran
exploratory factor analysis to determine if the number of
factors (i.e., single factor) and the loadings of observed
variables on the factor conformed to what is expected. If
the factor analysis suggested a single latent construct, we
mapped these observed variables onto a single latent vari-
able. Cronbach’s alpha was obtained as an index of reliability
associated with the variation accounted for by the true score
of the underlying latent construct (Hatcher, 1994).
We used the full information maximum-likelihood
method for estimation. This method uses all available ob-
served data without deleting records from participants with
missing values on covariates and, therefore, provides a more
effi cient and less biased analysis than complete case methods
(Arbuckle, 1996; Enders and Bandalos, 2001; McArdle and
Hamagami, 1996). Before the analysis, the data were evalu-
ated for normality by examining the univariate indices of
skewness. If normal assumption was found to be violated,
normalization through log transformation was applied. Path
coeffi cients (standardized beta weights) can be interpreted
both in terms of their signifi cance and magnitude. The over-
all fi t of the hypothesized model was evaluated by various in-
dices: A nonsignifi cant chi-square, Tucker-Lewis index (TLI)
greater than .95, a normed fi t index (NFI) greater than .95,
and root mean square error of approximation (RMSEA) less
than .06 were used to indicate a good fi t (Joreskog and Sor-
bom, 1986; Tucker and Lewis, 1973). Potential confounders,
including race, gender, socioeconomic status, and mother’s
self-reported mental health status during pregnancy, were in-
cluded in the SEM for statistical adjustment. Specifi cally, the
model included paths from race, socioeconomic status, and
mother’s self-reported mental health status during pregnancy
to MSP; paths from race, gender, socioeconomic status, and
mother’s self-reported mental health status during pregnancy
to birth weight, neurological abnormality, behavioral regula-
tion, and academic achievement; and paths from gender and
socioeconomic status to AUD.
We tested two structural equation models: (a) a model
with a direct path only from MSP to AUD and (b) a model
with a direct path and an indirect association between MSP
and AUD through lower birth weight, neurological abnormal-
ity, and lower academic functioning.
the mean age was 39.1 years (SD = 1.9; range: 34-44). The
The sample was 59.2% female, 61.1% were married, and
racial/ethnic composition was 83.5% non-Hispanic White,
9.6% Black/African American, 1.1% Hispanic/Latino, and
5.9% other. Six percent of the participants had not completed
high school, 19.1% had received a high school diploma or
general educational development credential only, 46.4%
had had some postsecondary education, 18.9% had college
degrees, and 9.6% had graduate degrees. The mean for the
maximum number of cigarettes smoked among mothers,
birth weight, and number of neurological abnormalities
were 11.04 (SD = 0.16), 3,257 g (SD = 538), and 1.32 (SD
= 1.92), respectively. The mean for spelling, reading and
arithmetic scores were 100.9 (SD = 13.5), 104.7 (SD = 16.4),
and 99.4 (SD = 9.8), respectively.
Cumulative risk for AUD among offspring as affected by
the amount of maternal smoking during pregnancy
spring had lifetime AUD (39.6% offspring had lifetime
abuse, and 11.6% had lifetime alcohol dependence). The
amount of MSP (none, <20 cigarettes/day, and ≥20 ciga-
rettes/day) was linearly associated with an increased rate of
AUD 37.0% for the offspring of nonsmoker mothers; 39.7%
for the offspring of mothers who smoked fewer than 20 ciga-
rettes per day at any time during pregnancy, and 44.2% for
the offspring of mothers who smoked 20 cigarettes or more
per day at any time during pregnancy (p = .01). This pattern
is the same for alcohol abuse (35.9% for none; 39.2% for
<20 cigarettes/day; and 44.0% for ≥20 cigarettes/day, p =
.005) and for dependence (9.8% for none; 11.1% for <20
cigarettes/day; and 13.8% for ≥20 cigarettes/day, p = .04).
The results of survival analyses of AUD by MSP are pre-
sented in Figure 1. The test of equality of strata (Wilcoxon
test) shows that there was a signifi cant difference in the onset
of AUD among the three groups by MSP, χ2(1) = 5.78, p =
.016. Table 1 further shows the cumulative risks for lifetime
AUD (alcohol abuse or alcohol dependence), as measured
by the hazard ratio (HR) in offspring exposed to MSP (<20
cigarettes/day and ≥20 cigarettes/day), relative to offspring
unexposed to MSP, the reference group. Offspring exposed
to 20 cigarettes or more per day had a greater cumulative
lifetime risk for AUD (HR = 1.24, 95% CI [1.03, 1.49], p
= .02) relative to the reference group. The increased risk
remained signifi cant after adjustment for potential confound-
ers (adjusted HR = 1.31, 95% CI [1.08, 1.59], p = .009), but
offspring exposed to fewer than 20 cigarettes per day had
no signifi cant increase in the risk of AUD (adjusted HR =
1.13, 95% CI [0.91, 1.41], p = .26) relative to the reference
group. Similarly, offspring exposed to 20 cigarettes or more
per day had a greater cumulative lifetime risk for alcohol
abuse (adjusted HR = 1.34, 95% CI [1.10, 1.62], p = .004),
but offspring exposed to fewer than 20 cigarettes per day had
no signifi cant increase in the risk of AUD (adjusted HR =
1.12, 95% CI [0.94, 1.46], p = .16) relative to the reference
Initial univariate analysis shows that 40.3% of the off-
NOMURA, GILMAN, AND BUKA 203
group. However, there was only a marginal difference in
the increased risk for alcohol dependence among offspring
exposed to 20 cigarettes or more per day relative to the refer-
Descriptives for latent measures used in the SEM
areas were as follows: reading = 104.75 (16.39), arithmetic =
99.44 (9.80), and spelling = 100.87 (13.51). The ranges were
70-165, 50-165, and 63-165, respectively. The three separate
Wide Range Achievement Test scores had excellent internal
consistency (Cronbach’s α = .87).
Behavioral regulation. The mean (SD) scores for the four
measures were as follows: dysinhibition = 2.91 (0.65), emo-
tional reactivity = 2.89 (0.42), activity level = 2.96 (0.48),
and impulsivity = 2.98 (0.28). All ranges were from 1 to 5.
The four behavioral characteristics scores had good internal
consistency (Cronbach’s α = .72).
Academic functioning. The mean (SD) scores for the three
Model fi t for structural equation model
associations between MSP, birth weight, neurological ab-
normality within a year after birth, academic functioning
(reading, spelling, and arithmetic scores), and behavioral
regulation (inhibition, emotional regulation, activity level,
and impulsivity). The correlation matrix of the variables used
for testing the model is presented in Table 2. We found that
correlations between the hypothesized relationships were
signifi cant but modest.
We tested two models: one (simple model) with only one
path from MSP to AUD and another (expanded model) based
on the hypothesis that the effect of MSP is both directly and
indirectly associated with increased risk for AUD through
elevated problems in infancy and childhood. The simple
The use of SEM allowed for a simultaneous test of the
model demonstrated good fi t, including an NFI of .98, a TLI
of .97, and a RMSEA of .043. The expanded model (see
Figure 2) also demonstrated good fi t, including an NFI of
.98, a TLI of .96, and a RMSEA of .035. We reached this
conclusion despite the signifi cant chi-square test of model
fi t, χ2(89) = 229.31, p < .01, because this statistic is known
to be especially sensitive to large sample sizes and captures
even small deviations from the causal model (Byrne, 2001).
The magnitude of associations in the hypothesized models
between MSP and AUD
fi t (fi gure not shown). There was a strong association be-
tween MSP and AUD (β = .07, p = .01). The unstandardized
path coeffi cient for this standardized path coeffi cient of .07
in a logit scale was .18, which represents the odds ratio of
1.2. This estimation is very similar to what we estimated in
our survival analysis. The expanded model, which includes
childhood outcomes as possible preceding conditions, is
presented in Figure 2 with the standardized path coeffi cients.
MSP was inversely associated with birth weight (β = -.20,
p < .0001), which in turn was inversely associated with a
greater number of neurological abnormalities at age 1 year
(β = -.16, p < .0001). Although MSP was not directly asso-
ciated with neurological abnormality, it was associated with
subsequent lower academic functioning at age 7 (β = -.23, p
< .0001). However, there was no notable association between
neurological abnormality and behavioral regulation at age
7. MSP was also associated with academic functioning (β
= -.10, p = .02) but not behavioral regulation (β = -.02, p =
.60) at age 7. Greater academic functioning was associated
with a decreased risk of AUD (β = -.13, p = .002). In addi-
tion, although only marginally signifi cant, MSP was found
to be directly associated with a greater risk for AUD (β =
.04, p = .07). Of note, there was a reduction in the stan-
Our initial model with only MSP and AUD showed good
TABLE 1. Cumulative risk and 95% confi dence interval (CI) for lifetime alcohol use disorder among offspring
exposed to varying degrees of smoking in utero
HR [95% CI]
HR [95% CI]
Alcohol use disorder
1.10 [0.89, 1.35]
1.24 [1.03, 1.49]
1.13 [0.91, 1.41]
1.31 [1.08, 1.59]
1.12 [0.91, 1.38]
1.28 [1.06, 1.54]
1.12 [0.94, 1.46]
1.34 [1.10, 1.62]
1.15 [0.77, 1.71]
1.44 [1.02, 2.04]
1.10 [0.73, 1.65]
1.38 [0.97, 1.95]
Notes: Adjusted for gender, family socioeconomic status at birth, mother’s self-reported mental health status
during pregnancy, and child’s race. Potential cluster effect due to multiple children from the same family has
been adjusted. There are eight missing cases. HR = hazard ratio.
204 JOURNAL OF STUDIES ON ALCOHOL AND DRUGS / MARCH 2011
FIGURE 1. Survival curves for alcohol use disorders among offspring exposed to maternal smoking during pregnancy in varying degrees (none, <20 cigarettes/
day, and ≥20 cigarettes/day during any pregnancy day). Bold solid line = offspring of mothers who smoked ≥20 cigarettes/day during pregnancy; dashed
line = offspring of mothers who smoked <20 cigarettes//day during pregnancy; dotted line = offspring of mothers who did not smoke during pregnancy.
Signifi cant difference among groups by amount of maternal smoking (none; <20 cigarettes/day; and ≥20 cigarettes/day) was found in the test of equality of
strata (Wilcoxon test), χ2(1) = 5.78, p = .016. The cumulative risks of lifetime alcohol use disorder in offspring exposed to smoking (<20 cigarettes/day and
≥20 cigarettes/day), relative to those unexposed, were estimated after adjusting for gender, race, social economic status, and mother’s self-reported mental
health status during pregnancy.
TABLE 2. Inter-correlations among study variables
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.
1. Maximum no. cigarettes
smoked during pregnancya
2. Birth weight (g)
3. Neurological abnormality (1 year)
4. Spelling (7 year)
5. Reading (7 year)
6. Arithmetic (7 year)
7. Emotional reactivity (7 year)
8. Dysinhibition level (7 year)
9. Activity level (7 year)
10. Impulsivity (7 year)
11. Alcohol use disorder
14. Socioeconomic status
15. Mother’s self-reported mental
.– -.21** .004 -.10** -.10** -.08
. – -.20** .07** .08** .09** -.04
. – -.20** .20** -.23** .007 -.08** .08** .002
. – .88** .64** .05*
. – .62** .08** .14** -.07*
-.13** -.07** .08** -.07*
-.02 .07** -.16** .09**
.15** -.16** .32** -.03
.12** -.15** .34** -.02
.05* -.19** .30** -.03
.39** -.007 -.07** .04§ -.06*
. – -.04 -.05*
. – -.19** -.01
-.03* -.13** .06**
.10** -.08** .02
.07** -.007 .08** .13** -.05*
. – .44** .43** .35** -.001 -.05
. – .39** .33** -.04
aThe maximum number of cigarettes smoked per day at any point during pregnancy was classifi ed into three levels of smoking: never smoked during any
pregnancy day (coded 0), smoked fewer than 20 cigarettes during any pregnancy day (coded 1), and smoked 20 cigarettes or more during any pregnancy
day (coded 2).
§p < .10. *p < .05; **p < .01.
NOMURA, GILMAN, AND BUKA 205
dardized path coeffi cient for the path from MSP to AUD in
the expanded model, compared with the one in the simple
model. This indicates that some of the effect between MSP
and AUD was explained by childhood factors. Furthermore,
although behavioral regulation did not play a role in our
hypothesized mechanism, poorer behavioral regulation was
associated with an increased AUD risk (β = -.07, p = .02).
40 years), the study evaluated the lifetime risk of AUD
among adults according to prior exposure to MSP. We further
investigated whether increased risk for AUD can, at least in
part, be explained by conditions at birth and in infancy and
in childhood to test hypotheses regarding potential mecha-
nisms by which MSP leads to adverse consequences. The
study has three main fi ndings and provides a foundation for
future studies on the pathways through which MSP can infl u-
ence risk for AUD in adulthood. First, adults whose mothers
smoked 20 or more cigarettes per day during pregnancy, but
not fewer than 20 cigarettes, had a signifi cantly increased
risk for AUD, compared with adults whose mothers did not
Capitalizing on the extended follow-up time (more than
(p < .0001)
at age 1
(p < .0001)
(p = .65)
(p < .0001)
(p = .002)
(p = .40)
(p = .02)
(p = .60)
(p = .07)
(p = .13)
FIGURE 2. Hypothesized model of adult alcohol use disorder (abuse or dependence) directly from maternal smoking during pregnancy and indirectly through
various problems in infancy and childhood Parameter estimates are standardized. Model-fi t indices: χ2(89) = 229.31, p < .001, normed fi t index = .98, Tucker-
Lewis index = .96, root mean square error of approximation = .035.
***p < .0001.
smoke during pregnancy. Second, in SEM, we also found
that MSP is associated with a signifi cant increased risk for
AUD, which is equivalent to the estimate in the survival
analysis (1.3 times increase). Third, this increase was par-
tially explained by problems in infancy (e.g., lower birth
weight) and in childhood (e.g., lower academic functioning).
Our study demonstrates that MSP is associated with a
signifi cant increased risk for AUD (adjusted HR = 1.31, p =
.009) among offspring exposed to 20 cigarettes or more in
utero, relative to offspring not exposed to any cigarettes in
utero. Although the magnitude of the effect is modest, even
a small increase in risk at one point in time could lead to a
change in trajectory that would lead to signifi cant implica-
tions at a different time.
Results from SEM also show a direct link from MSP to
AUD in adulthood in our simple model, which examined
only direct links between MSP and AUD. The model has
been expanded with problem factors in infancy and child-
hood. We found that MSP was indirectly associated with
an increased risk for AUD through problems in infancy and
childhood. Those results suggest that MSP may be indirectly
associated with an increased risk for AUD through problems
in infancy and childhood as well. These results provide
206 JOURNAL OF STUDIES ON ALCOHOL AND DRUGS / MARCH 2011
important clues regarding how MSP may increase the AUD
later in life through developmental deviations as a result
of neurocognitive impairment in early childhood. Specifi -
cally, the current study found that MSP was associated with
decreased birth weight, which in turn led to increased fre-
quency in neurological abnormality. Neurological abnormal-
ity was strongly associated with lower academic functioning
(β = -.23, p < .0001), and lower academic functioning was
associated with a subsequent AUD (β = -.13, p = .002).
However, (lower) behavioral regulation was not infl uenced
by neurological abnormality in the preceding stage. Of note,
the association between the two was signifi cant and showed
that neurological abnormality was negatively associated
with lower behavioral regulation before adjustments for
confounder effects. This may suggest that the link between
neurological abnormality and lower behavioral regulation is
explained, at least in part, by psychosocial and demographic
risk factors such as maternal poverty, low socioeconomic
status, and race, all of which are known to correlate with
The expanded model in our SEM analysis found that,
although the direct path from MSP to AUD was marginally
signifi cant, the mediating paths leading to AUD was sig-
nifi cant. If these associations represent causal effects, they
suggest that interventions directly addressing MSP reduc-
tion—as well as early identifi cations of problems at birth
and during infancy and childhood—may contribute to the
reduction of the risk for AUD in adulthood. Some evidence
for a direct path from MSP to increased subsequent sub-
stance use has been found in animal studies. For example,
after being exposed to nicotine in utero, nicotine receptors
(Lv et al., 2008) and altered catecholamine systems (Dwyer
et al., 2008) have been found in the brains of neonates. These
infl uences on the developing brain could make offspring
susceptible to substance use during adolescence and adult-
hood (Azam et al., 2007; Dwyer et al., 2008). However, we
should also note that the effect of MSP through childhood
problems is small. Specifi cally, the standardized path co-
effi cient for MSP and AUD in the simple model was just a
little greater (.07 vs. .04) than in the extended model with
childhood outcomes such as birth weight, neurological prob-
lems, behavioral regulation, and academic functioning. The
direct path may be explained, in part, by pertinent mediating
conditions such as adolescent delinquency, peer infl uences,
and family confl ict, which were not measured. If we are
able to identify the factors that mediate between childhood
academic function impairment and AUD, it might provide
additional options for potential intervention. Although our
current study did not include factors in adolescence, these
may be of particular signifi cance given that adolescence has
the highest risk for substance use initiation.
This study needs to be evaluated in light of its various
methodological strengths. Notably, this was a population-
based cohort that was systematically followed from pregnan-
cy and studied longitudinally more than 40 years. Maternal
smoking histories were ascertained at every prenatal visit
prospectively. Birth weight was recorded by a nurse observer
at the time of delivery, a method far superior to mothers’
retrospective reports. Research pediatricians or pediatric
neurologists prospectively evaluated potential neurological
problems at 4, 8, and 12 months. Latent variable academic
functioning was based on Wide Range Achievement Test
scores, and behavioral profi les at age 7 were ascertained
by child psychologists who were trained for high reliability
and validity. AUD was assessed using structured interviews
administered by trained interviewers blind to the mother’s
Despite these strengths, there are also several limitations.
First, the sample was not designed to be representative of the
broader population, which potentially limits external validity.
Second, the level of obstetric care in the 1960s is different
from the current standard of care. Our sample was born in
the pre-neonatal intensive care unit era, and the mortality
rate for those born prematurely with very low birth weight
was higher than it is currently. Also, since the late 1960s,
there has been considerable progress in the development of
more extensive special or remedial education along with the
introduction of early intervention programs. It is, therefore,
likely that many of the children who had problems with aca-
demic functioning would have been offered more effective
and targeted remedial assistance had they been born today.
However, we know that access to health care is substantially
underused or unrecognized, especially among the most vul-
nerable children with biological and social risk factors (Rob-
erts et al., 2008). We can only speculate whether children
with poorer academic functioning and behavioral regulation
would qualify for and receive such services. It is hoped
that more specifi c and accurate delineations of underlying
risk mechanisms in future birth cohort studies will enable
clarifi cation of the extent to which the risk for AUD can be
modifi ed through early cognitive remediation and interven-
tion for behavioral regulation. Third, although we included
key confounders in our model, given that maternal smoking
occurs within a broad constellation of social and behavioral
factors that also may infl uence developmental trajectory,
our fi ndings could have been subject to bias in an unknown
fashion. Gilman et al. (2008b), for example, evaluated as-
sociations with MSP and various childhood outcomes using
conditional fi xed-effects models that controlled for shared
familial vulnerability to adverse outcomes of MSP; in these
analyses, there were no detectable effects of MSP on child-
hood neurological and psychological outcomes. Therefore,
unmeasured confounding factors may have resulted in an
overestimation of the causal effects of MSP on AUDs and
a similar overestimation of the strength of the mediating
In a similar vein, as noted above, beyond childhood
strength/problems at age 7, we do not have measures of
NOMURA, GILMAN, AND BUKA 207
adolescent social behavior, such as delinquent offending, in
our model. If we had a measure of delinquency or familial
confl icts for adolescence, we could have evaluated whether
they further mediated the associations between childhood
problems (i.e., lower academic functioning and behavioral
dysregulation) and alcohol use in adulthood or whether those
additional psychosocial burdens at the time of adolescence
would only moderate the magnitude of the association be-
tween MSP and alcohol use.
Results obtained from the current study could potentially
guide research into effective intervention programs that pro-
mote abstinence from drinking, especially in vulnerable
populations such as those who had been exposed to MSP.
Vulnerabilities as a result of MSP may be reversible if iden-
tifi ed early (Bergh, 1990; Fisher et al., 2000) and could be
targeted to limit or prevent alcohol use or to delay the time
of initiation. Furthermore, enhanced awareness of the poten-
tial adverse consequences of MSP at different times in life
could lead to developing and testing targeted interventions,
especially for high-risk children, which would presumably
modify the trajectory of subsequent alcohol use problems.
American Psychiatric Association. (1994). Diagnostic and statistical manual
of mental disorders (4th ed.). Washington, DC: Author.
Arbuckle, J. L. Full information estimation in the presence of incomplete
data. In G. A. Marcoulides & R. E. Schumacker (Eds.), Advanced struc-
tural equation modeling (pp. 243-277). Mahwah, NJ: Lawrence Erlbaum
Associates, Inc., 1996.
Azam, L., Chen, Y., & Leslie, F. M. (2007). Developmental regulation of
nicotinic acetylcholine receptors within midbrain dopamine neurons.
Neuroscience, 144, 1347-1360.
Baan, R., Straif, K., Grosse, Y., Secretan, B., El Ghissassi, F., Bouvard, V.,
Altieri, A., & Cogliano, V. (2007). Carcinogenicity of alcoholic bever-
ages. Lancet Oncology, 8, 292-293.
Binder, D. A. (1992). Fitting Cox’s proportional hazards models from survey
data. Biometrika, 79, 139-147.
Brook, J. S., Cohen, P., & Brook, D. W. (1998). Longitudinal study of co-
occurring psychiatric disorders and substance use. Journal of American
Academy of Child & Adolescent Psychiatry, 37, 322-323.
Buka, S. L., Shenassa, E., & Niaura, R. (2003). Elevated risk of tobacco
dependence among offspring of mothers who smoked during pregnancy:
A 30-year prospective study. American Journal of Psychiatry, 160,
Buka, S. L., Tsuang, M. T., & Lipsitt, L. P. (1993). Pregnancy delivery
complications and psychiatric-diagnosis: A prospective study. Archives
of General Psychiatry, 50, 151-156.
Bushman, B. J. (1997). Effects of alcohol on human aggression: Validity
of proposed explanation. In M. Galanter (Ed.), Recent developments
in alcoholism: Vol. 13. Alcohol and Violence (pp. 227-243). New York:
Button, T., Thapar, A., & McGuffi n, P. (2005). Relationship between anti-
social behaviour, attention-defi cit hyperactivity disorder and maternal
prenatal smoking. British Journal of Psychiatry, 187, 155-160.
Byrne, B. M. (2001). Structural equation modeling with AMOS: Basic con-
cepts, applications, and programming. Mahwah, NJ: Lawrence Erlbaum.
Caetano, R., Schafer, J., Clark, C. L., Cunradi, C. B., & Raspberry, K.
(2000). Intimate partner violence, acculturation, and alcohol consump-
tion among Hispanic couples in the United States. Journal of Interper-
sonal Violence, 15, 30-45.
Castaneda, R., Sussman, N., Westreich, L., Levy, R., & O’Malley, M.
(1996). A review of the effects of moderate alcohol intake on the treat-
ment of anxiety and mood disorders. Journal of Clinical Psychiatry,
Centers for Disease Control and Prevention. (2004). Alcohol-Related
Disease Impact (ARDI). Retrieved from http://apps.nccd.cdc.gov/ardi/
Corrao, G., Rubbiati, L., Zambon, A., & Arico, S. (2002). Alcohol-attribut-
able and alcohol-preventable mortality in Italy. A balance in 1983 and
1996. European Journal of Public Health, 12, 214-223.
Corrao, G., Bagnardi, V., Zambon, A., & La Vecchia, C. (2004). A meta-
analysis of alcohol consumption and the risk of 15 diseases. Preventive
Medicine, 38, 613-619.
Cornelius, M. D., Leech, S. L., Goldschmidt, L., & Day, N. L. (2000).
Prenatal tobacco exposure: Is it a risk factor for early tobacco experi-
mentation? Nicotine and Tobacco Research, 2, 45-52.
Cornelius, M. D., Leech, S. L., Goldschmidt, L., & Day, N. L. (2005). Is
prenatal tobacco exposure a risk factor for early adolescent smoking? A
follow-up study. Neurotoxicology and Teratology, 27, 667-676.
Cox, D. R. (1972). Regression models and life tables. Applied Statistics,
Cummings, P., & Quan, L. (1999). Trends in unintentional drowning: The
role of alcohol and medical care. Journal of American Medical Associa-
tion, 281, 2198-2202.
Dierker, L. C., Donny, E., Tiffany, S., Colby, S. M., Perrine, N., & Clayton,
R. R. (2007). The association between cigarette smoking and DSM-IV
nicotine dependence among fi rst year college students. Drug and Alco-
hol Dependence, 86, 106-114.
DiFranza, J., & Lew, R. (1995). Effect of maternal cigarette smoking on
pregnancy complications and sudden infant death syndrome. The Jour-
nal of Family Practice, 40, 385-394.
Duncan, J. R., Randall, L. L., Belliveau, R. A., Trachtenberg, F. L., Randall,
B. R., Habbe, D., . . . Kinney, H. (2008). The effect of maternal smoking
and drinking during pregnancy upon 3H-nicotine receptor brainstem
binding in infants dying of the sudden infant death syndrome: Initial
observations in a high risk population. Brain Pathology, 18, 21-31.
Dwyer, J. B., Broide, R. S., & Leslie, F. M. (2008). Nicotine and brain
development. Birth Defect Research, 84, 30-44.
Ekblad, M., Gissler, M., & Korkeila, J. (2010). Prenatal smoking exposure
and the risk of psychiatric morbidity into young adulthood. Archives of
General Psychiatry, 67, 841-849.
Edner, A., Wennborg, M., Alm, B., & Lagercrantz, H. (2007). Why do ALTE
infants not die in SIDS? Acta Paediatrica, 96, 191-194.
Enders, C. K., & Bandalos, D. L. (2001). The relative performance of
full information maximum likelihood estimation for missing data in
structural equation models. Structural Equation Modeling: A Multidis-
ciplinary Journal, 8, 430-457.
Fergusson, D., Woodward, L., & Horwood, J. (1998). Maternal smoking dur-
ing pregnancy and psychiatric adjustment in late adolescence. Archives
of General Psychiatry, 55, 721-727.
Fleming, P., & Blair, P. S. (2007). Sudden infant death syndrome and paren-
tal smoking. Early Human Development, 83, 721-725.
Fried, P. (1993). Prenatal exposure to tobacco and marijuana: Effects dur-
ing pregnancy, infancy, and early childhood. Clinical Obstetrics and
Gynecology, 36, 319-336.
Fried, P. A., O’Connell, C. M., & Watkinson, B. (1992). 60- and 72-month
follow-up of children prenatally exposed to marijuana, cigarettes, and
alcohol: Cognitive and language assessment. Journal of Developmental
and Behavioral Pediatrics, 13, 383-391.
Fried, P. A., Watkinson, B., & Gray, R. A. (1992). Follow-up-study of atten-
tional behavior in 6-year-old children exposed prenatally to marijuana,
cigarettes, and alcohol. Neurotoxicology and Teratology, 14, 299-311.
208 JOURNAL OF STUDIES ON ALCOHOL AND DRUGS / MARCH 2011
Gilman, S. E., Martin, L. T., Abrams, D. B., Kawachi, I., Kubzansky, L.,
Loucks, E. B., . . . Buka, S. L. (2008a) Educational attainment and
cigarette smoking: A causal association? International Journal of Epi-
demiology, 37, 615-624.
Gilman, S. E., Gardener, H., & Buka, S. L. (2008b). Maternal smoking
during pregnancy and children’s cognitive and physical development:
A causal risk factor? American Journal of Epidemiology, 168, 522-531.
Goodman, R. A., Istre, G. R., Jordan, F. B., Herndon, J. L., & Kelaghan, J.
(1991). Alcohol and fatal injuries in Oklahoma. Journal of Studies on
Alcohol, 52, 156-161.
Griesler, P., Kandel, D., & Davies, M. (1998). Maternal smoking in preg-
nancy, child behavior problems, and adolescent smoking. Journal of
Research Adolescence, 8, 159-185.
Hatcher, L. (1994). A step-by-step approach to using the SAS(R) system
for factor analysis and structural equation modeling. Cary, NC: SAS
Heron, M. (2007). Death: Leading causes for 2004. National Vital Statistic
Reports, 56, 1-96.
Hingson, R., & Howland, J. (1993). Alcohol and non-traffi c unintended
injuries. Addiction, 88, 877-883.
Huijbregts, S. C., Warren, A. J., de Sonneville, L. M., & Swaab-Barneveld,
H. (2008). Hot and cool forms of inhibitory control and externalizing
behavior in children of mothers who smoked during pregnancy: An
exploratory study. Journal of Abnormal Child Psychology, 36, 323-333.
Jastak, J., & Jastak, S. (1965). Wide Range Achievement Test. Wilmington,
DE: Jastak Associates.
Joreskog, K. G., & Sorbom, D. (1986). LISREL V: Analysis of linear struc-
tural relationships by maximum likelihood and least squares methods.
Mooresville, IN: Scientifi c Software.
Keeping, J., Najman, J., Morrison, J., Western, J., Andersen, M., & Williams,
G. (1989). A prospective longitudinal study of social, psychological
and obstetric factors in pregnancy: Response rates and demographic
characteristics of the 8556 respondents. British Journal of Obstetrics &
Gynaecology, 96, 289-297.
Kessler, R. C., Crum, R. M., Warner, L. A., Nelson, C. B., Schulenberg, J.,
& Anthony, J. C. (1997). Lifetime co-occurrence of DSM-III-R alcohol
abuse and dependence with other psychiatric disorders in the National
Comorbidity Survey. Archives of General Psychiatry, 54, 313-321.
Linnet, K. M., Wisborg, K., Obel, C., Secher, N. J., Thomsen, P. H., Agerbo,
E., & Henriksen, T. B. (2005). Smoking during pregnancy and the risk
for hyperkinetic disorder in offspring. Pediatrics, 116, 462-467.
Linver, M. R., Brooks-Gunn, J., & Kohen, D. E. (2002). Family processes as
pathways from income to young children’s development. Developmental
Psychology, 38, 719-734.
Lv, J., Mao, C., Zhu, L., Zhang, H., Pengpeng, H., Xu, F., . . . Xu, Z. (2008).
The effect of prenatal nicotine on expression of nicotine receptor sub-
units in the fetal brain. Neurological Toxicology, 29, 711-726.
MacCallum, R. C., & Austin, J. T. (2000). Applications of structural equa-
tion modeling in psychological research. Annual Review of Psychology,
McArdle, J. J., & Hamagami F. (1996). Multilevel models from a mul-
tiple group structural equation perspective. In G. Marcoulides & R.
Schumacker (Eds.), Advanced Structural Equation Modeling: Issues and
Techniques. Mahwah, NJ: Lawrence Erlbaum, 89-124.
Markowitz, S. (2007). The effectiveness of cigarette regulations in reducing
cases of Sudden Infant Death Syndrome. Journal of Health Economics,
Maughan, B., Taylor, A., Caspi, A., & Moffi tt, T. (2004). Prenatal smoking
and early childhood conduct problems. Archives of General Psychiatry,
McArdle, J. J., & Hamagami F. (1996). Multilevel models from a mul-
tiple group structural equation perspective. In Marcoulides, G. &
Schumacker, R. (Eds.), Advanced Structural Equation Modeling: Issues
and Techniques. Hillsdale, N.J.: Erlbaum, 89-124.
McGill, V . M. J., Kowal-Vern, A., Fisher, S. G., Kahn, S., & Gamelli, R. L.
(1995). The impact of substance use on mortality and morbidity from
thermal injury. Journal of Trauma-Injury Infection & Critical Care, 38,
Monuteaux, M. C., Blacker, D., Biederman, J., Fitzmaurice, G., & Buka,
S. L. (2006). Maternal smoking during pregnancy and offspring overt
and covert conduct problems: A longitudinal study. Journal of Child
Psychology and Psychiatry, 47, 883-890.
Muthén, L.K., & Muthén, B.O. (1998-2007). Mplus user’s guide. Los
Muthén, B. O. (1992). Latent variable modeling in epidemiology. Alcohol
Health and Research World, 16, 286-292.
Myrianthopoulos, N. C., & French, K. S. (1968). An application of the U.S.
Bureau of the Census socioeconomic index to a large, diversifi ed patient
population. Social Sciences & Medicine, 2, 283-299.
Naeye, R., & Peters, E. (1984). Mental development of children whose
mothers smoked during pregnancy. Obstetrics & Gynecology, 64,
National Highway Traffi c Safety Administration. (2008). National Center
for Statistics and Analysis. Retrieved from http://www-nrd.nhtsa.dot.
Nichols, P. L., & Chen, T. C. (1981). Minimal brain dysfunction: A prospec-
tive study. Mahwah, NJ: Lawrence Erlbaum.
Niswander, K. R., & Gordon, M. (1972). The women and their pregnancies:
The Collaborative Perinatal Study of the National Institute of Neurologi-
cal Diseases and Stroke. Washington, DC: National Institute of Health.
Nomura, Y., Halperin, J. M., Newcorn, J. H., Davey, C., Fifer, W. P., Savitz,
D. A., & Brooks-Gunn, J. (2009). The risk for impaired learning-related
abilities in childhood and educational attainment among adults born
near-term. Journal of Pediatric Psychology, 34, 1084-1089.
Nomura, Y., Rajendran, K., Brooks-Gunn, J., & Newcorn, J. H. (2008).
Roles of perinatal problems on childhood developmental problems and
adolescent delinquency among children born after 33 completed weeks
of gestation: A prospective investigation. Journal of Child Psychology
and Psychiatry, 49, 1108-1117.
Orlebeke, J. F., Knol, D. L., & Verhulst, F. C. (1997). Increase in child
behavior problems resulting from maternal smoking during pregnancy.
Archives of Environmental Health, 52, 317-321.
Piquero, A. R., Gibson, C. L., Tibbetts, S. G., Turner, M. G., & Katz, S. H.
(2002). Maternal cigarette smoking during pregnancy and life-course-
persistent offending. International Journal of Offender Therapy and
Comparative Criminology, 46, 231-248.
Quertemont, E. (2004). Genetic polymorphism in ethanol metabolism:
Acetaldehyde contribution to alcohol abuse and alcoholism. Molecular
Psychiatry, 9, 570-581.
Rehm, J., Gmel, G., Sempos, C. T., & Trevisan, M. (2003). Alcohol-related
morbidity and mortality. Alcohol Research & Health, 27, 39-51.
Ringel, J. S., & Evans, W. N. (2001). Cigarette taxes and smoking during
pregnancy. American Journal of Public Health, 91, 1851-1856.
Roberts, G., Howard, K., Spittle, A. J., Brown, N. C., Anderson, P. J., &
Doyle, L. W. (2008). Rates of early intervention services in very pre-
term children with developmental disabilities at age 2 years. Journal of
Paediatrics and Child Health, 44, 276-280.
Rosenbloom, M. J., Sullivan, E. V., Sassoon, S. A., O’Reilly, A., Fama, R.,
Kemper, C. A., . . . Pfefferbaum, A. (2007). Alcoholism, HIV infection,
and their comorbidity: Factors affecting self-rated health-related quality
of life. Journal of Studies on Alcohol and Drugs, 68, 115-125.
Rosenthal, R. N., & Westreich, L. (1999). Treatment of persons with dual
diagnoses of substance use disorder and other psychological problems.
In B. S. McCrady, & E. E. Epstein (Eds.), Addictions: A comprehensive
guidebook (pp. 439-476). New York: Oxford University Press.
Shah, B. V., Barnwell, B. G., & Gayle, G. S. (1997). SUDAAN User’s
Manual 7.5. Research Triangle Park, NC: Research Triangle Institute.
Schiff, E. R. (1997). Hepatitis C and alcohol. Hepatology, 26, 39s-42s.
NOMURA, GILMAN, AND BUKA 209
Shiono, P. H., & Behrman, R. (1995). Low birth weight: Analysis and rec-
ommendations. Future Child, 5, 35-51.
Substance Abuse and Mental Health Services Administration (Offi ce of
Applied Studies). (2005). Results from the 2004 National Survey on
Drug Use and Health: National fi ndings (NSDUH Series H-28, DHHS
Publication No. SMA 05-4062) Rockville, MD: Author.
Tremblay, R., Nagin, D., Séguin, J., Zoccolillo, M., Zelazo, P., Boivin, M.,
. . . Japel, C. (2004). Physical aggression during early childhood: Tra-
jectories and predictors. Pediatrics, 114, e43-e50.
Tucker, L. R., & Lewis, C. (1973). A reliability coeffi cient for maximum
likelihood factor analysis. Psychometrika, 38, 1-10.
Van Horn, M. L., Jaki, T., Masyn, K., Ramey, S. L., Smith, J. A., & An-
taramian, S. (2009). Assessing differential effects: Applying regres-
sion mixture models to identify variations in the infl uence of family
resources on academic achievement. Developmental Psychology, 45,
Visscher, W., Feder, M., Burns, A., Brady, T., & Bray, R. (2003). The impact
of smoking and other substance use by urban women on the birthweight
of their infants. Substance Use and Misuse, 38, 1063-1098.
Vuijk, P., van Lier, P. A., Huizink, A. C., Verhulst, F. C., & Crijnen, A. A.
(2006). Prenatal smoking predicts non-responsiveness to an intervention
targeting attention-defi cit/hyperactivity symptoms in elementary school
children. Journal of Child Psychology and Psychiatry, 47, 891-901.
Wakschlag, L., Lahey, B., Loeber, R., Green, S., Gordon, R., & Leventhal,
B. (1997). Maternal smoking during pregnancy and the risk of conduct
disorder in boys. Archives of General Psychiatry, 54, 579-580.
Wakschlag, L., Pickett, K., Middlecamp, M., Walton, L., Tenzer, P., &
Leventhal, B. (2003). Pregnant smokers who quit, pregnant smokers
who don’t: Does history of problem behavior make a difference? Social
Science & Medicine, 56, 2449-2460.
Wall, T. L., Schoedel, K., Ring, H. Z., Luczak, S. E., Katsuyoshi, D. M., &
Tyndale, R. F. (2007) Differences in pharmacogenetics of nicotine and
alcohol metabolism: Review and recommendations for future research.
Nicotine and Tobacco Research, 9 (Suppl. 3), S459-S474.
Weese-Mayer, D., Ackerman, M., Marazita, M., & Berry-Kravis, E. (2007).
Sudden Infant Death Syndrome: Review of implicated genetic factors.
American Journal of Medical Genetics, 143, 771-788.
Weissman, M. M., Warner, V ., Wickramaratne, P. J., & Kandel, D. B. (1999).
Maternal smoking during pregnancy and psychopathology in offspring
followed to adulthood. Journal of the American Academy of Child &
Adolescent Psychiatry, 38, 892-899.
Weitzman, M., Gortmaker, S., & Sobol, A. (1992). Maternal smoking and
behavior problems of children. Pediatrics, 90, 342-349.
World Health Organization. (1993). The ICD-10 classifi cation of mental and
behavioural disorders. Geneva, Switzerland: Author.
World Health Organization, (1997) Composite International Diagnostic
Interview, Version 2.1. Geneva, Switzerland: Author.
Williams, R. (1995). Product-limit survival functions with correlated sur-
vival times. Lifetime Data Analysis, 1, 171-186.
Windle, M. (1997). The trading of sex for money or drugs, sexually trans-
mitted diseases (STDs), and HIV-related risk behaviors among multi-
substance using alcoholic inpatients. Drug and Alcohol Dependence,