Adolescent initiation of licit and illicit substance use: Impact of intrauterine
exposures and post-natal exposure to violence
Deborah A. Franka,⁎, Ruth Rose-Jacobsa, Denise Crooksb, Howard J. Cabralc, Jessie Gerteisb,
Karen A. Hackerd, Brett Martine, Zohar B. Weinsteinb, Timothy Heerene
aDepartment of Pediatrics, Boston University School of Medicine and Boston Medical Center, 725 Massachusetts Avenue, Mezzanine SW, Boston, MA 02118, United States
bDepartment of Pediatrics, Boston Medical Center, 725 Massachusetts Avenue, Mezzanine SW, Boston, MA 02118, United States
cDepartment of Biostatistics, Boston University School of Public Health, 801 Massachusetts Avenue, 3rd floor, Boston, MA 02118, United States
dInstitute for Community Health, 163 Gore Street, Cambridge, MA 02141, United States
eData Coordinating Center, Boston University School of Public Health, 801 Massachusetts Avenue, Boston, MA 02118, United States
a b s t r a c ta r t i c l ei n f o
Received 8 October 2009
Received in revised form 9 March 2010
Accepted 8 June 2010
Available online 23 June 2010
Prenatal drug exposure
Adolescent substance use
Whether intrauterine exposures to alcohol, tobacco, marijuana, or cocaine predispose offspring to substance
use in adolescence has not been established. We followed a sample of 149 primarily African American/
African Caribbean, urban adolescents, recruited at term birth, until age 16 to investigate intrauterine cocaine
exposure (IUCE). We found that in Kaplan–Meier analyses higher levels of IUCE were associated with a
greater likelihood of initiation of any substance (licit or illicit), as well as marijuana and alcohol specifically.
Adolescent initiation of other illicit drugs and cigarettes were analyzed only in the “any” summary variable
since they were used too infrequently to analyze as individual outcomes. In Cox proportional hazard models
controlling for intrauterine exposure to alcohol, tobacco, and marijuana and demographic and post-natal
covariates, those who experienced heavier IUCE had a greater likelihood of initiation of any substance, and
those with lighter intrauterine marijuana exposure had a greater likelihood of initiation of any substance as
well as of marijuana specifically. Time-dependent higher levels of exposure to violence between ages of 8
and 16 were also robustly associated with initiation of any licit or illicit substance, and of marijuana, and
© 2010 Elsevier Inc. All rights reserved.
By definition, women who continue to use licit or illicit
psychoactive substances in pregnancy, even at low levels, may be
considered to have problem substance use since they are unable to
discontinue use in spite of potential adverse consequences to
themselves and their children. Children whose parents have a
history of problem licit or illicit substance use may themselves be at
an elevated risk for becoming substance users and abusers in
adolescence and adulthood , a risk ascribed both to environ-
mental and genetic factors . However, after acknowledging
familial risk including risks of living with substance using parents or
parental figures in school age or adolescence, [3,19] surprisingly
little is known about whether there is an association of intrauterine
exposures to psychoactive substances, licit or illicit, with the
offspring's own age of initiation of substance use.
Substance use and pubertal maturation are linked . The
maturing of the brain in adolescence ushers in an epoch when
previously undetectable effects of intrauterine insults may manifest
and a time of unusual vulnerability to negative effects of the
adolescent's own substance use, particularly because of the immatu-
rity of the frontal cortex and subcortical monoaminergic systems .
The earlier the initiation of psychoactive substances, the greater the
risk of neuropsychological perturbations and of eventual problem
substance use [16,42,56–58,61,63], and other negative psychosocial
outcomes such as incarceration .
The animal data on intrauterine substance exposure and later
substance use are inconclusive, conflicting, and focus on adult rather
than adolescent animals. Animal studies of intrauterine cocaine
exposure (IUCE) provide an instructive example. One study 
suggested that in male mice intrauterine exposure to higher dosages
of cocaine increased the probability of the mouse acquiring cocaine
self-administration in adulthood, but a second study  suggested
that the reinforcing efficacy of cocaine was reduced in adult male
rats with IUCE. Mice with IUCE, compared to unexposed, also
showed reduced cocaine conditioned place-preference for high
doses of cocaine, though the a preference for low and medium
doses was not altered .
Published findings on whether intrauterine exposures affect
human offsprings' future licit and illicit substance use are also
Neurotoxicology and Teratology 33 (2011) 100–109
⁎ Corresponding author. Tel.: +1 617 414 5252; fax: +1 617 414 7047.
E-mail address: email@example.com (D.A. Frank).
0892-0362/$ – see front matter © 2010 Elsevier Inc. All rights reserved.
Contents lists available at ScienceDirect
Neurotoxicology and Teratology
journal homepage: www.elsevier.com/locate/neutera
contradictory and vary with the demographic characteristics of the
sample, and document substance use or problem use rather than age
at first use. Two studies in a predominantly European American,
middle-class United States sample [8,9] do not report age of alcohol
initiation, but suggest that intrauterine alcohol exposure confers
incremental risks for offspring to develop alcohol “problems,” but a
similar study from Australia  found that maternal alcohol use
during the youngster's early adolescence, but not the use in
pregnancy, predicted adolescent alcohol use. Data from the National
Collaborative Perinatal Project (NCPP)  suggested that intra-
uterine tobacco exposure increases the offspring's risk of developing
nicotine addiction, a finding replicated in Australia . Gender
specific effects were found in a retrospective study of treatment-
seeking smokers suggesting intrauterine tobacco exposure was
significantly associated with earlier age of tobacco initiation in
males and accelerated daily use in females . Intrauterine
marijuana exposure has been found to predict offspring's marijuana
use at age 14 in a sample which was 50% African American . In
the same sample, in contrast to the studies of European American
children, it was found that intrauterine tobacco exposure did not
predict offspring use once current maternal use was considered. A
single study to date in a predominantly African American low
income cohort suggests that boys, but not girls, with IUCE were more
likely than unexposed peers at age 11 years to engage in high-risk
behaviors, including tobacco use . These findings were not
replicated in a multi-site multiethnic sample by other researchers
 where risk behaviors correlated with the child's post-natal
violence exposure, but not with history of intrauterine exposures.
Additionally, there is not always a homology between the
intrauterine exposure and the substance that is used by the
offspring. For example, a longitudinal study following European–
Canadian middle-class children reported that intrauterine marijuana
exposure appears to predict the offspring's tobacco use , and a
retrospective study of adopted adults reported that intrauterine
alcohol exposure, independent of home environment, may increase
the risk of dependence not only on alcohol but on tobacco and other
Previous studies suggest it is important to perform analyses that
evaluate multiple intrauterine exposures in the context of post-natal
environmental factors which have also been linked in epidemiologic
studies to increased risk of early adolescent substance initiation and
later substance use disorders. For example, the retrospective
Adverse Childhood Exposures (ACE) study identified parental
substance use, parental incarceration, household dysfunction and
various forms of exposure to violence as factors likely to increase an
individual's own risk of developing a substance use disorder .
Conversely, multiple environmental factors in adolescents' lives
have been shown potentially to protect against substance abuse. A
longitudinal study of urban African American adolescents identified
an inverse relationship between having a higher number of
protective factors (such as increased religiosity) and drug abuse
. An inverse relationship between religiosity and substance use
was also identified in a retrospective study of a nationally
representative sample of adolescents as part of the National
Comorbidity Study . A longitudinal study of adolescents from
their senior year of high school through their first year after high
school found parental monitoring to be a protective factor
decreasing substance abuse as adolescents enter young adulthood
We hypothesize that adolescents who experienced intrauterine
cocaine exposure (IUCE) and intrauterine exposures to other
substances are more likely, after confound control, to become
themselves early initiators of licit and illicit psychoactive substances,
compared to demographically similar adolescents without such
exposures. Although we are aware that other child level factors such
as lower IQ  and childhood externalizing behavior and other
psychiatric symptoms [6,11] may be on the causal pathway to
substance misuse, we focus on potential familial and environmental
confounds rather than these possible mediators, which are
addressed only in secondary analyses.
2.1. Sample recruitment
The IRB of Boston Medical Center (then called Boston City
Hospital) approved this study at inception and yearly thereafter.
Prior to initiation of the study a Certificate of Confidentiality was
obtained from the federal government to protect researchers from
being compelled by subpoena to release data regarding study
participants. Soon after delivery, all birth mothers gave informed
consent to study participation. If the child changed caregivers, similar
informed consent was obtained for each new caregiver. Beginning at
the 8-year visit children also provided informed assent. The primary
goal of the study was to explore the potential impact of intrauterine
cocaine exposure (IUCE).
Sample recruitment on the post-partum floor of Boston Medical
Center from October 1990 to March 1993 has been published in
detail . All mother–infant dyads met the following criteria based
on review of mother and infant medical records and confirmed by
interviews, biological markers, and infant physical examinations
obtained by study personnel: 1) Infant gestational age greater than
or equal to 36 weeks; 2) No requirement for neonatal intensive care;
3) No obvious major congenital malformations; 4) No diagnosis of
fetal alcohol syndrome in the neonatal record; 5) No history of
human immunodeficiency virus seropositivity noted in the mother's
or infant's medical record; 6) Mother's ability to communicate
fluently in English; 7) No indication by neonatal or maternal urine
toxicology screen or history in medical record of mother's use during
pregnancy of illegal opiates, methadone, amphetamines, phencycli-
dine, barbiturates, or hallucinogens; and 8) Mother aged 18 years or
older. These criteria were established to exclude infants with known
major risk factors that might confound or obscure the effects, if any,
2.2. Method of intrauterine cocaine exposure classification
Mothers participating in the study were identified as either
heavier, lighter, or non-users of cocaine soon after delivery of the
index child by interview and by biological markers obtained by
clinicians and study personnel. At intake during the maternal post-
partum hospitalization, research assistants used the Addiction
Severity Index , supplemented by study-specific questions, to
interview the mothers about pregnancy and lifetime use of
cigarettes, alcohol, and illicit drugs.
During the period of study recruitment at Boston Medical Center,
urine testing for metabolites of illicit drugs was performed for
clinical indications at the discretion of health care personnel, but
was not universal. Results of the urine drug Enzyme Multiplied
Immunoassay Technique (EMIT) assays, which were obtained for
clinical purposes during prenatal care or labor and delivery from the
mother or from the newborn after birth, were recorded for the
present study (when available in the medical record). Exposed
newborns were targeted for recruitment on the basis of maternal
self-report or positive clinical urine assays obtained from either
mother or newborn.
nursery population as a whole, most of whom did not have maternal
urine assays performed for drug metabolites for clinical purposes.
Therefore, after recruitment and informed consent, additional urine
samples were collected from all study mothers for analysis for
benzoylecognine, opiates, amphetamines, benzodiazepines, and
D.A. Frank et al. / Neurotoxicology and Teratology 33 (2011) 100–109
RIA, Roche Diagnostics Systems, Inc, Montclair, NJ). Meconium speci-
mens were also sought from all enrolled newborns to be analyzed by
radioimmunoassay for benzoylecognine (a cocaine metabolite),
opiates, amphetamines, benzodiazepines, and cannabinoids. The
radioimmunoassay used was a modification of the method of Ostrea
et al. published in detail elsewhere [44,50]. All mother–infant dyads
provided at least one biological marker, either urine from mother or
infant or meconium that confirmed their exposure or lack of exposure
to cocaine during pregnancy. To be classified in the “unexposed”
group, mothers had to deny use of cocaine on interview and all
available biological markers needed to be negative for cocaine use. In
this sample, the mean days of self-reported cocaine use during
pregnancy was 20.6 days, with a range from 0 to 264. The mean
meconium concentration of benzoylecognine/g was 1143 ng with a
rangefrom0 ngto17950 ng/g.Beforedatawereanalyzed,acomposite
measure of “heavier” use was a priori defined as the top quartile of
meconium concentration for cocaine metabolites (≥3314 ng of
benzoylecognine/g meconium) and/or top quartile days of self-
reported use (N61 days) during the entire pregnancy. All other use
was classified as “lighter” . If benzoylecognine levels in meconium
were not in the top quartile, self-report took priority in classifying
levels of exposure comparable to the approach of other investigators,
where use of cocaine more than twice a week during pregnancy is
considered “heavier” use [4,33,59]. Not all infants with IUCE have
positive meconium assays . Moreover, meconium samples could
not be obtained from 14% of study infants whose exposure status was
confirmed by maternal or infant urine assay. Therefore, whichever
indicator (self-report or meconium assay) demonstrated higher
exposure was used to define exposure category.
2.3. Other intrauterine drug exposure classifications determined during
the neonatal period
Identification of prenatal marijuana exposure was based on
positive results of urine assay, meconium assay, or maternal self-
report. In our previous reports we have analyzed marijuana
categorically as exposed or unexposed, since meconium concentra-
tion is not entirely valid due to the storage of metabolites in the
mother's body fat  and because self-reported use was denied by a
third of the marijuana users in this cohort who were identified solely
on the basis of meconium or urine assay. However, recently we
constructed an a priori index of no marijuana use (negative: self-
report, urine, and meconium assays), heavier use (positive: self-
report and/or meconium or urine assay and positive urine at delivery
or the top quartile of self-reported days of use (N8 days during
pregnancy among users), or lighter use (positive: self-report or
assay who did not meet criteria for inclusion in the heavier group).
This classification index was significantly associated (pb0.0001)
with level of use of alcohol, cigarettes, and cocaine during the
pregnancy and with the infants' mean birth weight — Marijuana
Unexposed 3210 g (s.d. 477), Lighter 3069 g (s.d. 464), and Heavier
2943 g (s.d. 511), p=0.02). Therefore, in this analysis, unlike our
earlier developmental analyses [27,55], we are able to classify
intrauterine marijuana exposure as a three-level ordinal exposure.
At the time the study was initiated there was no established
biologic marker for gestational alcohol exposure, and cotinine assays
for tobacco metabolites were prohibitively expensive. Therefore, the
ascertainment of alcohol and tobacco use in pregnancy only by self-
report was state of the art at the time the current sample was
recruited. We quantified intrauterine alcohol exposure by mother's
self-reported average daily volume of alcohol in drinks per day (in
the 30 days prior to delivery, which was highly correlated with use
through pregnancy in our sample). During the post-partum
interview, mothers reported the average number of cigarettes that
they consumed per day while pregnant. For descriptive purposes we
categorized alcohol use as none, Lighter (b1 drink/day) and Heavier
(≥1 drinks/day) in the 30 days prior to delivery and cigarettes as
none, Lighter (b10 cigarettes a day), or Heavier (≥10 cigarettes a
day) during pregnancy.
2.4. Sample maintenance and retention
As described in previous publications [10,26,28,29,55], caregiver/
child dyads were assessed post-natally at ages 6 months, 1, 2, 4, 6, 8.5,
9.5, 11, 14 (early adolescence) and 16 (middle adolescence) years.
After each study visit, the caregiver and child were given store
vouchers and/or age appropriate gifts for completion of the interview
One hundred forty-nine of the original 252 study offspring (60%)
were assessed during early and mid-adolescence waves of follow up.
These adolescents did not differ by level of intrauterine exposures to
cocaine, tobacco, alcohol, or marijuana, birth weight, gender, or their
mother's age, ethnicity, or education at the time of their birth
(pN0.05) from the 103 in the birth sample who did not participate in
2.5.1. Outcome measures
Adolescents took an audio computer assisted self-interview
(ACASI) to assess their substance use. The software was
programmed and the computer screen positioned so that the
research assistant could not see or access the respondent's
confidential answers. The items on the survey were synthesized
from a number of different instruments to address the issues that
seemed most relevant to young adolescents from urban environ-
ments. In order to assess the participants' tobacco use, the ACASI
included all 10 items of the Hooked on Nicotine Checklist (HONC),
developed to assess adolescent nicotine dependence . Additional
questions about tobacco, alcohol, and other substance use were
taken from several components of the CDC's 2005 Youth Risk
Behavior Surveillance System (YRBSS) . The ACASI also included
questions from the Wisconsin YRBS Middle School Questionnaire,
the State and Local YRBS, and the Wisconsin YRBS High School
Questionnaire. These questions asked about the participants' past
and current use of alcohol and other licit or illicit substances. Using
the YRBS format, questions were also added about specific drugs,
such as illegally diverted prescription opiates, known to be a
problem in our geographic region.
For this analysis of substance use initiation, questions were framed
as follows for legal substances which do not require a prescription for
adults: “How old were you when you smoked a whole cigarette for the
first time?” “How old were you when you had your 1st drink of
alcohol other than a few sips?” It was specified that, “a drink of alcohol
is equal to having a can of beer (the same size as a soda can), a glass of
wine, a wine cooler, or a shot of liquor such as rum, gin, vodka, or
whiskey.” For substances that are legal with a prescription but not
without (such as amphetamines, steroids, oxycodone and other pain
killers, or benzodiazepines) the question was “How old were you
when you first tried taking (substance of interest) without a doctor or
nurse telling you to take them?” In the case of illicit substances
(marijuana, heroin, cocaine, “club drugs”) quantity was not specified,
with the question framed “How old were you when you first tried
(substance) for the first time?”
Participants selected from a forced choice list of “never” or of age
in years beginning at age “8 or younger” for each substance. The
same questions were asked at early and middle adolescence. In case
of discrepancy between the two ages for a given substance, the
middle adolescent answer was chosen for analysis, since answers
differed by more than a year primarily for age of alcohol initiation,
and there was concern that some early adolescents may not have
D.A. Frank et al. / Neurotoxicology and Teratology 33 (2011) 100–109
understood the definition of a “drink.” Adolescent participants also
furnished urine to be tested by the United States Drug Testing
Laboratories, Inc. using the No-Excuse Urine Panel, a limit of
detection panel that screens using EMIT at the lowest validated
concentrations that can be achieved with the reagent set and the
instrumentation. The GC/MS confirmations used are either at ½ or 1/
5 of the SAMHSA screening concentrations depending on the drug
class for cannabinoids, opiates, amphetamines and cocaine metabo-
lites and ELISA for cotinine. We classified adolescents as having used
a particular substance if either self-report or urine assay was
positive. In the 16 instances where the respondent denied initiation
of a given substance but the urine assay was positive for that
substance, the respondent's age at the date of the assay was taken as
the age of initiation.
Potential covariates were ascertained from interviews of care-
givers at each research contact since birth, and from the participants
themselves by face-to-face interview during school age or by ACASI
and face-to-face interview and cognitive assessments during
184.108.40.206. Caregiver measures. Trained research interviewers ques-
tioned caregivers at each study contact regarding their own
substance use in the last 12 months. For each substance, caregivers
were asked “Have you used…?” For each substance for which the
caregiver answered affirmatively they were also asked “How many
days have you used…during the past 30 days?”, “In the last
12 months, about how many total days were you using…?” and
“How many days ago was your most recent use of…?” Caregivers
were also asked if child/adolescent participants spent time with
household members or with anyone else who used tobacco, alcohol,
or any illegal substances; “How many people that your child spends
time with use the following drugs?” and the relationship between
the adolescent and the substance user. Urine samples were obtained
from caregivers and analyzed using the same No-Excuse Urine Panel
described above for the adolescents.
The participants' birth mothers or other caregivers were also
interviewed to assess familial risk of substance use. The biological
father's substance use was ascertained by birth mother's report on a
study-specific question immediately following delivery . Later
interviews asked using study-specific questions assessed whether
the child's biological father, biological grandmothers, biological
grandfathers, biological aunts, biological uncles or biological siblings
had problems with alcohol or drugs. The sum of affirmative answers
for alcohol and drug problems was the control variable for familial
The identity of the child's caregiver (birth mother vs. other) and
number of changes in caregiver since last contact were ascertained
at each study visit and summarized as whether or not the child had
lived with the birth mother from birth to mid-adolescence. Whether
the child had experienced the incarceration of a parent was also
ascertained from the accompanying caregiver at each visit .
Caregivers were also asked to complete the Child Behavior Problem
Checklist [1,2] at each study contact from age 2 to 11 years.
Household food insecurity as a measure of relative material hardship
was reported by the caregivers at assessments when the children
were ages 8–16  using the United States Department of
Agriculture (USDA) Food Security Scale (FSS) , which deter-
mines whether or not there was sufficient quantity and quality of
food in the preceding twelve months for all households members to
lead an active and healthy life.
220.127.116.11. Adolescent measures. Trained research assistants masked to
intrauterine exposure status administered the Wechsler Abbreviated
Scale of Intelligence (WASI),  to adolescents at age 14. Participants
beginning at age 9.5 years were asked at each study contact to
indicate their level of pubertal development on a series of schematic
drawing derived from the work of Tanner . ACASI questions about
peer use were the four questions from Peer Chemical Environment
Subscale, taken from the Personal Experience Inventory .
Children's self-report of exposure to violence was ascertained
using the Violence Exposure Scale for Children-Revised (VEX-R) 
designed to assess children's exposure to violence through self-report
using a 21 item, cartoon-based interview. The VEX-R uses a 4-point
Likert scale (0=never, 1=once, 2=a few times, 3=lots of times) to
determine how many times the child has witnessed different types of
violence (e.g. someone being yelled at, being beaten up, or stabbed
to the child him/herself. The VEX-R was administered face-to-face by
trained examiners when subjects visited the testing laboratory at ages
8.5, 9.5 and 11 years.
For the early and middle adolescence research visits we modified
the VEX-R to make it more appropriate for adolescent subjects by
removing the cartoons and the questions related to spanking,
administering the text of other questions by ACASI, and making
questions more time-specific to determine if the event took place
within the past year or at any point in the child's lifetime.
There is no standard method for scoring the VEX-R [5,14,38,41].
In our previous work with preadolescents  the VEX-R was
analyzed using a total score, calculated by simply adding up how
many times the subject reported experiencing each violent event,
without weighting for severity. In this analysis of adolescents
without attempting a priori weighting of violent events, to address
the non-interval scaling and the skewed nature of the VEX-R total
score, we created quartiles by ranking each subject's VEX-R score at
each of the 8.5, 9.5,11, 14 and 16 year old time points, and then
subdividing those scores into 4 approximately equal groups. The
highest quartile obtained by each subject at each time point was
then used to create the time-dependent variable used for our
analyses, with the 4th quartile representing the highest level of
violence exposure for age and the 1st quartile representing the
lowest level of exposure. When data were missing at any protocol
point, the last value was carried forward. If no such value was
available the closest later value was substituted.
Level of supervision by parents or other primary caregivers was
ascertained with the measure of Lamborn et al.  which evaluates
the adolescent's perception of parenting by summing Likert scoring
(1–4) of responses ranging from “never” to “frequently” of questions
such as, “My parents know exactly where I am most afternoons after
school.” The adolescent's religiosity was assessed using four questions
from the Longitudinal Study of Adolescent Health .
2.6. Data analyses
First, we generated descriptive statistics for each variable of
interest, with means and standard deviations for continuous
variables and counts and percentages for categorical variables, in
quartiles (e.g., for VEX-R measured at each protocol age as
previously described). Next, in bivariate analyses we compared the
three IUCE groups on each of our dependent variables of interest:
age at first use of any licit or illicit substance; age at first use of each
category of substances — alcohol, marijuana, tobacco and other
drugs. For each, we used Kaplan–Meier analyses together with Cox
proportional hazards regression without additional covariates. We
generated unadjusted (crude) hazard ratios (HR) and their 95%
confidence intervals (CI) from these Cox models as well as from the
multivariable models to be described. The IUCE variables in these
models were modeled as dummy variables: heavier compared to
unexposed; lighter compared to unexposed. We then added a series
of potential covariates in a one-at-a-time fashion to the model in
order to assess their confounding effects by applying a 10% change-
D.A. Frank et al. / Neurotoxicology and Teratology 33 (2011) 100–109
in-estimates criterion for the adjusted hazard ratios compared to the
unadjusted. Those variables for which inclusion in the model
changed either the hazard ratio for lighter IUCE compared to
unexposed or for heavier IUCE compared to unexposed were
retained in the full, main effects-only Cox regression model.
Variables assessing the intrauterine exposure to marijuana, cigar-
ettes, and alcohol were included in all final models based on a priori
theoretical considerations to avoid misattribution of one intrauter-
ine substance exposure to another. We assessed potential collinear-
ity among the independent variables in the final model by examining
bivariate correlations and through examination of principal compo-
nents analysis-based methods as described for linear models by
Belsley et al. . We found no collinearity in these data among the
predictors used in our final statistical models.
We used SAS version 9.1.2 for all of our analyses. Results were
deemed statistically significant where two-tailed pb0.05.
3.1. Sample characteristics
Table 1 summarizes sample characteristics by the level of IUCE.
Compared to those with no IUCE, children with IUCE were more likely
to experience higher levels of marijuana, alcohol, and tobacco
intrauterine exposure, and to be of lower birth weight, but there
were no sex differences by level of exposure. Mothers who used
cocaine during pregnancy were more likely to be born in the United
States, and to be older at the child's birth, but the groups did not differ
in education level or in the percent who described themselves as
African American or African Caribbean (“Black”). The groups did not
differ significantly in the number of first or second-degree relatives
who were reported by birth mother or other caregivers as having a
drug or alcohol problem.
Adolescents with IUCE were more likely than adolescents
without IUCE to have experienced multiple changes in caregivers
up to the middle adolescent assessment, and less likely to have
always lived with their birth mother. However, we did not find any
statistically significant difference in percent who reported the
strictest (top quartile) level of parental supervision, or of violence
exposure in the top quartile between adolescents with and without
IUCE. The three groups of adolescents also did not differ in self-
reported religiosity, measured IQ, or average caregiver reported
externalizing score on the Child Behavior Problem Checklist
between ages 2 and 11, average lead level or age at menarche for
girls, although the heavily exposed adolescents were less likely than
the other groups to rate themselves as Tanner 1 V/V at age 14 and 16.
There was a dose related difference in the Mean Peer Use Score at
16 years, with the most heavily IUCE adolescents reporting the
highest levels of peer use. In preadolescence, level of IUCE was
associated with the percent of children who experienced tobacco,
alcohol, cocaine, and other drug use in their household, but no
difference in household marijuana use.
3.2. Bivariate associations of intrauterine cocaine exposure with
3.2.1. Initiation of any substance
As shown by the bivariate Kaplan–Meier survival curves in Fig. 1,
earliest age of initiation of any licit or illicit substance varied by IUCE
group. In these unadjusted Kaplan–Meier analyses, 74% of the
heavier cocaine exposure group had initiated any licit or illicit
substance use by age 16, compared to 58% of the lighter cocaine
exposure group, and 48% of the unexposed, p=0.02 via a global log
ranktest(Fig. 1).ACoxregression model includingonlythetwo IUCE
dummy variables showed hazard ratios (HR) of 2.10 for heavier IUCE
exposure vs. unexposed (p=0.009) and 1.29 for lighter IUCE
exposure vs. unexposed (p=0.31). The most commonly used
substances were alcohol and marijuana, as described in detail in
Sections 3.2.2 and 3.2.3. Cigarette use was less common (23% of
unexposed, 21% of the lighter IUCE, 30% of the heavier IUCE, p=0.85
by log rank test). Use of other substances was relatively rare (3.4%
cocaine, 3% glue, 1% prescription opiates, 1% amphetamines). No
participant acknowledged cocaine use by self-report on the ACASI;
the urine levels of cocaine metabolites using the No-Excuses panel
were all below the level that could rule out passive exposure, so the
cocaine use finding is tentative. Although these substances were all
considered in the analyses of any initiation of substance use reported
above, the number of users of these as individual substances in each
group was too small to allow for appropriate multivariable analyses.
Excluding the adolescent cocaine urine data did not change the
analyses for age of initiation of any substance, since all those with
positive urines for cocaine had also initiated other substances.
3.2.2. Initiation of any marijuana use
As shown in Fig. 2, in Kaplan–Meier analysis 56% of the heavier,
38% of the lighter, and 35% of the unexposed IUCE groups had used
marijuana up to age 16 (global log rank test p=0.06; Fig. 1). The
unadjusted hazard ratio for the heavier vs. unexposed contrast was
2.07 (p=0.03) and for lighter vs. unexposed was 1.17 (p=0.61).
3.2.3. Initiation of any alcohol use
As seen in Fig. 3, in Kaplan–Meier analysis, 59% of the heavier,
45% of the lighter, and 33% of the unexposed IUCE groups had used
alcohol up to age 16 (global log rank test p=0.05), with a hazard
ratio of 2.10 for the heavier IUCE group compared to unexposed
(p=0.02) and 1.38 for the lighter IUCE group compared to
3.3. Multivariable analyses
Because of the relatively low prevalence of cigarette and other
drug use, we restricted the multivariable analyses which follow age
at first use of any licit or illicit substance as a composite outcome and
to marijuana, and alcohol as individual outcomes.
To arrive at the final models, we tested and excluded the following
variables (because they did not change the relationship of IUCE with
outcome by more than 10%): 1st degree relative (father, sister,
brother) with history of drug or alcohol problem, 2nd degree relative
(grandmother, grandfather, aunt, uncle) with history of drug or
alcohol problem, history of having an incarcerated parent, birth
mother's education, and whether she was born in the United States
and the child's birth weight, child's gender, adolescent religiosity,
experience of strict supervision by caregivers, time-dependent
household food security from age 8.5 to 16 years, time-dependent
ascertainment of household members' use of alcohol, cigarettes, and
cocaine from the participant's age 8.5–16 years, number of caretaker
changes from birth through middle adolescence, and peer substance
use. Of all the variables retained in the multivariable analyses which
follow, only VEX-R scores in the 3rd and 4th quartiles had hazard
ratios of comparable or greater magnitude than those found for
intrauterine exposure to cocaine or marijuana and so only these are
discussed in detail.
3.3.1. Initiation of any licit or illicit substance use
As seen in Table 2, in an adjusted model that included the level of
intrauterine exposures to alcohol, marijuana, and tobacco, VEX-R
quartiles at the time of substance initiation or censoring (i.e. the last
available observation up to which time there was no substance use),
African American/Caribbean maternal ethnicity, maternal age at
delivery, and having had the birth mother as the sole caretaker from
birth, we found a moderate-to-strong positive adjusted association
between IUCE and earlier age at initiation, with hazard ratios of 2.19
D.A. Frank et al. / Neurotoxicology and Teratology 33 (2011) 100–109
for heavier IUCE exposure vs. unexposed (95% C.I.: 1.10, 4.36,
p=0.03) and 1.69 for lighter IUCE exposure vs. unexposed (95% C.I.:
0.95, 3.03, p=0.07). There was also a statistically significant
association of younger maternal age at delivery with an increased
likelihood of substance initiation by age 16 (HR=0.94 per year of
mother's age; p=0.02) as well as of having had VEX-R scores in the
third or fourth quartiles compared to the first: fourth quartile:
HR=3.44 (95% C.I.: 1.70, 6.94, p=0.0006); third quartile: HR=2.29
(95% C.I. 1.10, 4.75 p=0.03). Levels of prenatal alcohol or cigarette
exposure were not statistically significant predictors of age at
initiation of any licit or illicit substance use. There was suggestive
evidence, however, of a positive association with lighter but not
heavier intrauterine marijuana exposure with substance initiation
(HR of 2.14, 95% C.I.: 1.07, 4.28; p=0.03).
3.3.2. Initiation of marijuana use
As showninTable 2,after adjustmentforthesamecovariates asin
the Cox model for any substance use by age 16 plus an additional
covariate of having a member of the adolescent's household who
used marijuana (which changed the relationship between predictor
and outcome more than 10% as described above), the heavier IUCE
group had only 1.40 times the hazard for marijuana use compared to
Sample characteristics — age of drug use initiation (N=149).
Perinatal and family covariates Intrauterine cocaine exposure group
Unexposed (N=69) Lighter (N=53) Heavier (N=27)
Maternal characteristics at birth
Mean maternal age at child's birth (yrs)
US born (%)
Black-US or other (%)
Mean mother's education (yrs)
Marijuana exposure (%)
Alcohol (≥.5 drinks per day) (%)
Cigarettes (≥1/2 pack per day) (%)
Child characteristics at birth
Male child (%)
Mean birthweight (grams)
Family substance use at birth
Mean number of 1st degree relatives with drug or alcohol problem at intake
Mean number of 2nd degree relatives with drug or alcohol problem at intake
Mean number of 1st degree relatives with drug or alcohol problem at age14 or 16
Mean number of 2nd degree relatives with drug or alcohol problem at age14 or 16
Strict supervision ever 4th quartile (%)
Violence exposure revised (VEX-R) ever 4th quartile (%)
Mean religiosity scale
Always in birth mother care up to age 16 (%)
Mean caregiver changes to age 16
Food insecure ever (%)
Incarcerated parent ever (%)
Mean peer use score at age 16
Mean CBCL ages 2–10 years
Mean WASI IQ
Mean maximum lead level (N=117)
Mean age at menarche (N=69)
Tanner stage-age 9.5 (N=114)
Tanner stage-age 11 (N=111)
Tanner stage-age 14 (N=135)
Tanner stage-age 16 (N=139)
Household/regular contact up to age 6 — cigarettes (%)
Ever household/regular contact ages 8–16 cigarettes (%)
Household/regular contact up to age 8–16 alcohol (%)
Ever household/regular contact ages 8–16 alcohol (%)
Household/regular contact up to age 8–16 marijuana (%)
Ever household/regular contact ages 8–16 marijuana (%)
Household/regular contact up to age 8–16 cocaine (%)
Ever household/regular contact ages 8–16 cocaine (%)
Household/regular contact up to age 8–16 other drugs (%)
Ever household/regular contact ages 8–16 other drugs (%)
53.6 50.9 44.4 0.72b
3350.0 (513.9) 3034.1 (373.7)2864.6 (348.0)
1st degree relatives include father and siblings. 2nd degree relatives include grandparents, aunts and uncles.
aMeans (S.D.) P-value via one-way ANOVA.
bP-value via chi-square test.
D.A. Frank et al. / Neurotoxicology and Teratology 33 (2011) 100–109
those unexposed to IUCE (95% C.I.: 0.62, 3.15, p=0.42). Likewise, for
those in the lighter IUCE group compared to the unexposed, there
was no significant association with an increased likelihood of
initiation of marijuana use by age 16 (HR=1.29, 95% C.I.: 0.64,
2.60, p=0.47). There was significant evidence of a positive
association with lighter marijuana exposure (HR=2.86, 95% C.I.:
1.20, 6.77; p=0.02), as well as for having had a VEX-R score in the
fourth quartile versus the first quartile (HR=5.05 95% C.I.: 1.95,
13.07, p=0.0008) and for VEX-R scores in the third quartile
compared to the first (HR=3.60, 95% C.I.: 1.40, 9.29, p=0.008).
3.3.3. Initiation of alcohol use
In a Cox regression analysis adjusting for the same covariates as in
the Cox model for any substance use by age 16, since no other
variables met our criteria for inclusion in this analysis, (Table 2), we
found similar, elevated, but non-significant levels of association with
initiation of alcohol use by age among the heavier IUCE (HR=1.86,
95% C.I: 0.96, 3.63, p=0.17) and lighterIUCE (HR=1.73, 95%C.I.: 0.79,
3.82, p=0.07) groups versus the unexposed. As in the Cox analyses
for any substance use and marijuana use by age 16, we found that
those who had VEX-R scores in the fourth quartile had a markedly and
statistically significantly increased likelihood of alcohol use by age 16
compared to those in the first quartile (HR=3.96, 95% C.I.: 1.74, 8.99,
p=0.001), as did those in the third quartile versus the first quartile
(HR=2.82, 95% C.I.: 1.20, 6.62, p=0.02).
3.4. Interaction terms and secondary analyses
From the preceding models, we examined potential effect
modification of IUCE and VEX-R effects by including interaction
terms of level of prenatal marijuana, alcohol and cigarette exposures
in the model and assessing their statistical significance at the 0.05
level (those not meeting this criterion were not included in
subsequent models). Likewise, we verified that the proportional
hazards assumption was met by including interaction terms of the
IUCE and VEX-R variables with age at first use and found them not to
be statistically significant at the 0.05 level. In secondary analyses
results were not substantially changed after including adolescent IQ
and average externalizing CBCL score through age 11 years as fixed
covariates, and the participant's self-reported Tanner pubertal stage
as a time-dependent covariate beginning at age 9.5 years. (Analyses
available from the authors on request)
This is one of only a few published studies in humans which
evaluate a possible association between prospectively identified
intrauterine exposure to cocaine or marijuana with initiation of
adolescent substance use, after considering intrauterine exposure to
tobacco and alcohol, and potentially relevant covariates. Consistent
with neonatal  and adolescent neuropsychological findings 
in this cohort, there appears to be an ordered relationship after
confound control between the heavier IUCE and the initiation of any
licit or illicit psychoactive substance, and for alcohol but not
marijuana initiation. The relationship between intrauterine mari-
juana exposure and the adolescent's own use of any licit or illicit
substance or of marijuana is not as clearly ordered as that for IUCE.
We do not have a definitive explanation for this. One criterion for
“heavier” marijuana use in our sample required the mother or
infant to have positive urine toxic screen at delivery indicating
exposure in the last month of pregnancy. It may be that exposure to
marijuana earlier rather than later in gestation confers greater risk
for adolescent substance initiation or there may be unmeasured
covariates which explain these findings.
While our findings are intriguing from a neuroteratologic
perspective, and seem to indicate that heavier IUCE increases the
risk for early substance use initiation, it is crucial not to over-interpret
them from a clinical or public policy perspective. It is important to
stress that this analysis does not evaluate problem or heavy substance
use, but only age at first use which, while concerning, by no means
indicates that an adolescent will inevitably develop a substance use
disorder. The substances most commonly used in our cohort were
those which are widely available and commonly used by other
adolescents nationally and in our region: alcohol and marijuana.
Alcohol is currently used by approximately 45%of adolescents age 12–
17 in Massachusetts and 46% nationally while approximately 41% of
adolescents in Massachusetts and 38% of adolescents nationally have
ever used marijuana.
Fig. 1. Age of initiation of any substance use by level of intrauterine cocaine exposure.
Fig. 2. Age of initiation of marijuana use by level of intrauterine cocaine exposure.
Fig. 3. Age of initiation of alcohol use by level of intrauterine cocaine exposure.
D.A. Frank et al. / Neurotoxicology and Teratology 33 (2011) 100–109
In contrast, use is less prevalent of “hard drugs” such as cocaine
(9% in Massachusetts and 7% nationally) or illegal amphetamines (4%
both in Massachusetts and nationally) . Questions about whether
early initiation of alcohol and marijuana for these adolescents
presages future “hard drug” use will only be answered through
further follow up study. We have no evidence to suggest whether
children with intrauterine cocaine or marijuana exposure will or will
not go on to become adult addicts. Additionally, it is important to
remember that even at age 16 years, even at the highest levels of
intrauterine cocaine exposure, 26% of our sample were complete
abstainers from any licit or illicit psychoactive substance use while
nearly half of the intrauterine unexposed adolescents had initiated
As far as we know, the strong prospective relationship we
identified between the highest time-dependent levels of childhood
and adolescent exposure to violence and licit or illicit substance
initiation, has been noted only in one other prospective study which
extended just to age 11  years, although this relationship has
been reported in a large retrospective study . High levels of
exposure to violence as reported through childhood and adolescence
by the adolescents were associated with the initiation of any licit or
illicit substance, of marijuana, and of alcohol with hazard ratios as
high, or higher, than those associated with any prenatal exposure,
including cocaine (see Table 2). Notably high levels of violence
exposure were robustly associated with initiation of alcohol and
marijuana use even when heavier intrauterine exposures to cocaine
or marijuana were not.
The study has a number of limitations including sample size. In
a larger sample certain effects that are statistically marginal in our
analyses (for example, the association of heavier IUCE with early
alcohol initiation) might be more robust and multivariable analysis
of initiation of tobacco and illicit substances other than marijuana
would be possible. Our sample is also by design quite homogeneous,
sofindingsshouldnotbegeneralized toadolescents bornpremature-
ly, from non Black ethnic groups, or different levels of socioeconomic
privilege. Undoubtedly some of the 17 cocaine using women for
whom neonatal meconium samples were not available minimized
self-reported use, and therefore some heavier users may have been
misclassified as lighter. Misclassifying heavier users as lighter would
decrease the likelihood of finding ordinal dose effect, so that any
found in this study are robust. Our characterization of familial risk is
quite crude without confirmation by detailed genogram or genetic
markers. Therefore, we cannot rule out that the apparent impact of
intrauterine exposures on offspring's substance initiation is a marker
for unmeasured genetic factors, rather than a neuroteratologic
mechanism. While we attempted to control for numerous psycho-
social variables, we do not know the impact of unmeasured or
unanalyzed covariates, such as post-natal lead levels which did not
differ between exposure groups but for which we do not have a
complete data set.
Comprehensive treatment of pregnant women who are strug-
gling with use of any psychoactive substance, whether licit or illicit,
should be a clinical and public health priority. Effective prevention
of early substance use initiation also requires focus on the post-natal
quality of life for impoverished children and adolescents, particu-
larly protection from high levels of exposure to violence, which
correlate in this cohort with as great or greater odds of early
substance use initiation than intrauterine exposures to cocaine or
In the United States during pregnancy approximately 1% [31,45]
of women use cocaine and between 3%  and 6%  use
marijuana. However, 17% of children age 12–17 witness violence
and 39% are the victims of violence . The role that violence plays
on the initiation of licit or illicit substance use even after controlling
for multiple intrauterine exposures and other factors thus has far
broader implications for public health than intrauterine exposure to
illicit drugs, but receives less attention in the public discourse.
Cox proportional hazards regression models of initiation of any substance, marijuana, and alcohol N=149.
Any substance usea
Any marijuana useb
Any alcohol usec
HR (95% C.I.)*
HR (95% C.I.)*
HR (95% C.I.)*
Intrauterine cocaine exposured
Intrauterine marijuana exposured
Nat. log, av. daily vol. alcohol, 30 days pre-delivery
Nat. log, prenatal daily cigarettes
Birth mother African American/Caribbeane
Birth mother continuous caretakerf
Birth mother age at delivery (years)
Violence exposure revised (VEX-R)g
Marijuana user in householdh
2.19 (1.10, 4.36)
1.69 (0.95, 3.03)
1.40 (0.62, 3.15)
1.29 (0.64, 2.60)
1.73 (0.79, 3.82)
1.86 (0.96, 3.63)
0.84 (0.39, 1.83)
2.14 (1.07, 4.28)
1.16 (0.63, 2.14)
0.97 (0.79, 1.19)
0.66 (0.29, 1.48)
0.83 (0.51, 1.34)
0.94 (0.90, 0.99)
1.36 (0.56, 3.28)
2.86 (1.20, 6.77)
1.42 (0.74, 2.75)
1.15 (0.89, 1.48)
0.89 (0.34, 2.31)
0.76 (0.42, 1.38)
0.94 (0.88, 0.99)
0.69 (0.26, 1.84)
2.12 (0.94, 4.74)
1.54 (0.76, 3.11)
1.07 (0.85, 1.36)
0.38 (0.12, 1.27)
0.88 (0.49, 1.56)
0.95 (0.90, 1.00)
3.44 (1.70, 6.94)
2.29 (1.10, 4.75)
1.81 (0.89, 3.68)
5.05 (1.95, 13.07)
3.60 (1.40, 9.29)
1.40 (0.51, 3.82)
2.02 (0.92, 4.41)
3.96 (1.74, 8.99)
2.82 (1.20, 6.62)
1.09 (0.44, 2.69)
*HR: hazard ratio; CI: confidence interval.
an=84 events, n=65 censored; model p=0.001.
bn=59 events, 90 censored; model p=0.0014.
cn=63 events, n=86 censored; p=0.0002.
evs. non African American/Caribbean.
fvs. not in continuous care of birth mother.
gvs. 1st quartile.
D.A. Frank et al. / Neurotoxicology and Teratology 33 (2011) 100–109
Conflict of interest statement
The authorshave no financial, personal, orother relationships with
people or organizations that may inappropriately influence the
authors' submitted work.
This research was supported by the National Institutes on Drug
Abuse (NIDA) grant number DA06532 to Dr. Frank and by grant MO1
RR00533 and RR025771 from the National Institutes of Health/
National Center for Research Resources, a component of the National
Institutes of Health (NIH). This work would not have been completed
without the wise guidance and unfailing support of Dr. Vincent
Smeriglio to us and all the other longitudinal projects. We are
everlastingly grateful to him. We would also like to thank Shayna
Soenksen, Heather Baldwin, Laura Anatale, and Mattia Chason for
their diligence in performing study assessments and to the
participants and their caregivers who have worked with us for
more than 19 years.
 T.M. Achenbach, Manual for Child Behavior Checklist 4–18, 1991 Profile,
Department of Psychiatry University of Vermont, Burlington, VT, 1991.
 T.M. Achenbach, Manual for the Child Behavior Checklist/2–3 and 1992 Profile,
University of Vermont, Burlington, VT, Department of Psychiatry, 1992.
 R. Alati, J.M. Najman, S.A. Kinner, A.A. Mamun, G.M. Williams, M. O'Callaghan, W.
Bor, Early predictors of adult drinking: a birth cohort study, American Journal of
Epidemiology 162 (2005) 1098–1107.
 S.M. Alessandri, M. Bendersky, M. Lewis, Cognitive functioning in 8- to 18-month-
old drug-exposed infants, Developmental Psychology 34 (1998) 565–573.
 A. Ardila-Rey, M. Killen, A. Brenick, Moral reasoning in violent contexts: displaced
and non-displaced Colombian children's evaluations of moral transgressions,
retaliations and reconciliation, Social Development 18 (2008) 182–209.
 T.D. Armstrong, E.J. Costello, Community studies on adolescent substance use,
abuse, or dependence and psychiatric comorbidity, Journal of Consulting and
Clinical Psychology 70 (2002) 1224–1239.
 A. Arria, C. Derauf, L. Lagasse, P. Grant, R. Shah, L. Smith, W. Haning, M. Huestis, A.
Strauss, S.D. Grotta, et al., Methamphetamine and other substance use during
pregnancy: preliminary estimates from the Infant Development, Environment,
and Lifestyle (IDEAL) Study, Maternal and Child Health Journal 10 (2006)
 J.S. Baer, H.M. Barr, F.L. Bookstein, P.D. Sampson, A.P. Streissguth, Prenatal alcohol
exposure and family history of alcoholism in the etiology of adolescent alcohol
problems, Journal of Studies on Alcohol 59 (1998) 533–543.
 J.S. Baer, P.D. Sampson, H.M. Barr, P.D. Connor, A.P. Streissguth, A 21-year
longitudinal analysis of the effects of prenatal alcohol exposure on young adult
drinking, Archives of General Psychiatry 60 (2003) 377–385.
 M. Beeghly, B. Martin, R. Rose-Jacobs, H. Cabral, T. Heeren, M. Augustyn, D.
Bellinger, D.A. Frank, Prenatal cocaine exposure and children's language
functioning at 6 and 9.5 years: moderating effects of child age, birthweight, and
gender, Journal of Pediatric Psychology 31 (2006) 98–115.
 H.M.E. Belcher, H.E. Shinitzky, Substance abuse in children: prediction, protection,
and prevention, Archives of Pediatric and Adolescent Medicine 152 (1998)
 D.A. Belsley, E. Kuh, R.E. Welsch, Regression Diagnostics: Identifying Influential
Data and Sources of Collinearity, Wiley, New York, 1980.
 D. Bennett, M. Bendersky, M. Lewis, Preadolescent health risk behavior as a
function of prenatal cocaine exposure and gender, Journal of Developmental and
Behavioral Pediatrics 28 (2007) 467–472.
 R. Brandt, C. Ward, A. Dawes, A. Flisher, Epidemiological measurement of
children's and adolescents' exposure to community violence: working with the
current state of the science, Clinical Child and Family Psychology Review 8 (2005).
 S.L. Buka, E.D. Shenassa, R. Niaura, Elevated risk of tobacco dependence among
offspring of mothers who smoked during pregnancy: a 30-year prospective study,
The American Journal of Psychiatry 160 (2003) 1978–1984.
 A. Carlin, S. O'Malley, Neuropsychological consequences of drug abuse, in: I. Grant,
K. Adams (Eds.), Neuropsychological Assessment of Neuropsychiatric Disorders,
Oxford University Press, New York, 1996, pp. 486–503.
 Comparison Between Massachusetts Students and U.S. Students 2007 YRBS,
Centers for Disease Control and Prevention, 2007.
 J.T. Cook, D.A. Frank, Food security, poverty, and human development in the
United States, Annals of the New York Academy of Sciences 1136 (2008) 193–209.
 N.L. Day, L. Goldschmidt, C.A. Thomas, Prenatal marijuana exposure contributes to
the prediction of marijuana use at age 14, Addiction 101 (2006) 1313–1322.
 J.R. DiFranza, J.A. Savageau, K. Fletcher, J.K. Ockene, N.A. Rigotti, A.D. McNeill, M.
Coleman, C. Wood, Measuring the loss of autonomy over nicotine use in
adolescents: the DANDY (Development and Assessment of Nicotine Dependence
in Youths) study, Archives of Pediatrics & Adolescent Medicine 156 (2002)
 D.K. Eaton, L. Kann, S. Kinchen, J. Ross, J. Hawkins, W.A. Harris, R. Lowry, T.
McManus, D. Chyen, S. Shanklin, et al., Youth risk behavior surveillance — United
States, 2005, MMWR: Surveillance Summaries 55 (2006) 1–108.
 J. Estelles, M. Rodriguez-Arias, C. Maldonado, M.A. Aguilar, J. Minarro, Gestational
exposure to cocaine alters cocaine reward, Behavioural Pharmacology 17 (2006)
 V.J. Felitti, R.F. Anda, D. Nordenberg, D.F. Williamson, A.M. Spitz, V. Edwards, M.
P. Koss, J.S. Marks, Relationship of childhood abuse and household dysfunction
to many of the leading causes of death in adults. The Adverse Childhood
Experiences (ACE) Study, American Journal of Preventative Medicine 14 (1998)
 D. Fergusson, L. Horwood, E. Ridder, Show me the child at seven II: childhood
intelligence and later outcomes in adolescence and young adulthood, Child
Psychology and Psychiatry 46 (2005) 850–858.
 N.A. Fox, L.A. Leavitt, VEX-R, The Violence Exposure Scale for Children — Revised, ,
 D.A. Frank, K.M. McCarten, C.D. Robson, M. Mirochnick, H. Cabral, H. Park, B.
Zuckerman, Level of in utero cocaine exposure and neonatal ultrasound findings,
Pediatrics 104 (1999) 1101–1105.
 D.A. Frank, J. Brown, S. Johnson, H. Cabral, Forgotten fathers: an exploratory study
of mothers' report of drug and alcohol problems among fathers of urban
newborns, Neurotoxicology and Teratology 24 (2002) 339–347.
 D.A. Frank, R. Rose-Jacobs, M. Beeghly, M. Augustyn, D. Bellinger, H. Cabral, T.
Heeren, Level of prenatal cocaine exposure and scores on the Bayley Scales of
Infant Development: modifying effects of caregiver, early intervention, and birth
weight, Pediatrics 110 (2002) 1143–1152.
 D.A. Frank, R. Rose-Jacobs, M. Beeghly, M.B. Wilbur, D. Bellinger, H. Cabral, Level of
prenatal cocaine exposure and 48-month IQ: importance of preschool enrich-
ment, Neurotoxicology and Teratology 27 (2005) 15–28.
 J.N. Giedd, The teen brain: insights from neuroimaging, The Journal of Adolescent
Health 42 (2008) 335–343.
 J. Havens, L.A. Simmons, L.M. Shannon, W.F. Hansen, Factors associated with
substance use during pregnancy: results from a national sample, Drug and Alcohol
Dependence 99 (2008) 89–95.
 G.S. Hecht, N.E. Spear, L.P. Spear, Alterations in the reinforcing efficacy of cocaine
in adult rats following prenatal exposure to cocaine, Behavioral Neuroscience 112
 S.W. Jacobson, J.L. Jacobson, R.J. Sokol, S.S. Martier, L.M. Chiodo, New evidence for
neurobehavioral effects of in utero cocaine exposure, The Journal of Pediatrics 129
 J.L. Johnson, M. Leff, Children of substance abusers: overview of research findings,
Pediatrics 103 (1999) 1085–1099.
 N. Joseph, M. Augustyn, D.A. Frank, H. Cabral, Preadolescents' report of exposure
to violence: association with friends' and own substance use, The Journal of
Adolescent Health 38 (2006) 669–674.
 Y. Kaminer, O. Bukstein, R.E. Tarter, The Teen-Addiction Severity Index: rationale
and reliability, The International Journal of the Addictions 26 (1991) 219–226.
 D.G. Kilpatrick, B.E. Saunders, D.W. Smith, Youth Victimization: Prevalence and
Implications, National Institute of Justice, 2003.
 D.J. Kolko, M.S. Hurlburt, J. Zhang, R.P. Barth, L.K. Leslie, B.J. Burns, Posttraumatic
stress symptoms in children and adolescents referred for child welfare
investigation: a national sample of in-home and out-of-home care, child
maltreatreatment, , 2009.
 L. Lagasse, J. Hammond, L.I.U. Jing, B. Lester, S. Shankaran, H. Bada, C.R. Bauer, R.
Higgins, D.A.S. Abhik, Violence and delinquency, early onset drug use, and
psychopathology in drug-exposed youth at 11 years, Annals of the New York
Academy of Sciences 1094 (2006) 313–318.
 S. Lamborn, N. Mounts, L. Steinberg, S. Dornbusch, Patterns of competence and
adjustment among adolescents from authoritative, authoritarian, indulgent, and
neglectful families, Child Development 62 (1991) 1049–1065.
 W. Marie-Alsana, M. Haj-Yahia, Violence among Arab elementary school pupils in
Israel, Journal of Interpersonal Violence 21 (2006) 58–88.
 K.L. Medina, K.L. Hanson, A.D. Schweinsburg, M. Cohen-Zion, B.J. Nagel, S.F. Tapert,
Neuropsychological functioning in adolescent marijuana users: subtle deficits
detectable after a month of abstinence, Journal of the International Neuropsy-
chological Society 13 (2007) 807–820.
 L. Miller, M. Davies, S. Greenwald, Religiosity and substance use and abuse among
adolescents in the National Comorbidity Survey, Journalof the American Academy
of Child and Adolescent Psychology 39 (2000) 1190–1197.
 M. Mirochnick, D.A. Frank, H. Cabral, A. Turner, B. Zuckerman, Relation between
meconium concentration of the cocaine metabolite benzoylecgonine and fetal
growth, The Journal of Pediatrics 126 (1995) 636–638.
 National Pregnancy and Health Survey: Drug Use Among Women Delivering
Livebirths: 1992, National Institute on Drug Abuse, Rockville, MD, 1996.
 J.M. Nonnemaker, C.A. McNeely, R.W. Blum, Public and private domains of
religiosity and adolescent health risk behaviors: evidence from the National
Longitudinal Study of Adolescent Health, Social Science & Medicine 57 (2003)
 F.V. O'Callaghan, M. O'Callaghan, J.M. Najman, G.M. Williams, W. Bor, R. Alati,
Prediction of adolescent smoking from family and social risk factors at 5 years,
and maternal smoking in pregnancy and at 5 and 14 years, Addiction 101 (2006)
 C. Oncken, S. McKee, S. Krishnan-Sarin, S. O'Malley, C. Mazure, Gender effects of
reported in utero tobacco exposure on smoking initiation, progression and
D.A. Frank et al. / Neurotoxicology and Teratology 33 (2011) 100–109
nicotine dependence in adult offspring, Nicotine & Tobacco Research 6 (2004) Download full-text
 K. Ostaszewski, M.A. Zimmerman, The effects of cumulative risks and promotive
factors on urban adolescent alcohol and other drug use: a longitudinal study of
resiliency, American Journal of Community Psychology 38 (2006) 237–249.
 E.M. Ostrea Jr., M. Brady, S. Gause, A.L. Raymundo, M. Stevens, Drug screening of
newborns by meconium analysis: a large-scale, prospective, epidemiologic study,
Pediatrics 89 (1992) 107–113.
 G.C. Patton, B.J. McMorris, J.W. Toumbourou, S.A. Hemphill, S. Donath, R.F.
Catalano, Puberty and the onset of substance use and abuse, Pediatrics 114 (2004)
 A.J. Porath, P.A. Fried, Effects of prenatal cigarette and marijuana exposure on
drug use among offspring, Neurotoxicology and Teratology 27 (2005)
 C.A. Prescott, P.A.F. Madden, M.C. Stallings, Challenges in genetic studies of the
etiology of substance use and substance use disorders: introduction to the special
issue, Behavior Genetics 36 (2006) 473–482.
 B.A. Rocha, A.N. Mead, B.E. Kosofsky, Increased vulnerability to self-administer
cocaine in mice prenatally exposed to cocaine, Psychopharmacology 163 (2002)
 R. Rose-Jacobs, D. Waber, M. Beeghly, H. Cabral, D. Appugleise, T. Heeren, J.
Marani, D.A. Frank, Intrauterine cocaine exposure and executive functioning in
middle childhood, Neurotoxicology and Teratology 31 (2009) 159–168.
 M. Roselli, A. Ardila, M. Lubomski, S. Murray, K. King, Personality profile and
neuropsychological test performance in chronic cocaine-abusers, The Interna-
tional Journal of Neuroscience 110 (2001) 55–72.
 S.B. Rourke, T. Loberg, Neurobehavioral correlates of alcoholism, in: I. Grant, K.
Adams (Eds.), Neuropsychological Assessment of Neuropsychiatric Disorders,
Oxford University Press, New York, 1996, pp. 423–485.
 A.D. Schweinsburg, S.A. Brown, S.F. Tapert, The influence of marijuana use on
neurocognitive functioning in adolescents, Current Drug Abuse Review 1 (2008)
 L.T. Singer, R. Arendt, S. Minnes, A. Salvator, A.C. Siegel, B.A. Lewis, Developing
language skills of cocaine-exposed infants, Pediatrics 107 (2001) 1057–1064.
 E. Slade, E. Stuart, D. Salkever, M. Karakus, K. Green, N. Ialongo, Impacts of age of
onset of substance use disorders on risk of adult incarceration among
disadvantaged urban youth: a propensity score matching approach, Drug and
Alcohol Dependence 95 (2008) 1–13.
 N. Solowij, Cannabis and Cognitive Functioning, Cambridge University Press, New
 J.M. Tanner, Growth and endocrinology of the adolescent, in: L. Gardner (Ed.),
Growth and Endocrinology of the Adolescent, 2nd ed., W.B. Saunders, Philadel-
phia, PA, 1975, pp. 14–64.
 S.F. Tapert, E. Granholm, N.G. Leedy, S.A. Brown, Substance use and withdrawal:
neuropsychological functioning over 8 years in youth, Journal of the International
Neuropsychological Society 8 (2002) 873–883.
 E.Z. Tronick, D.A. Frank, H. Cabral, M. Mirochnick, B. Zuckerman, Late dose–
response effects of prenatal cocaine exposure on newborn neurobehavioral
performance, Pediatrics 98 (1996) 76–83.
 U. S. Economic Research Service, U.S. Household Food Security Module: Three-
stage Design, with Screeners, , 2008.
 D. Wechsler, Wechsler Abbreviated Scale of Intelligence (WASI), Psychological
Corporation, San Antonio, TX, 1999.
 H.R. White, B.J. McMorris, R. Catalano, C.B. Fleming, K.P. Haggerty, R.D. Abbott,
Increases in alcohol and marijuana use during the transition out of high school
Into emerging adulthood: The effects of leaving home, going to college, and high
school protective factors, Journal on the Study of Alcohol 67 (2006) 810–822.
 M.B. Wilbur, J.E. Marani, D. Appugliese, R.Woods, J.A. Siegel, H.J. Cabral, D.A. Frank,
Socioemotional effects of fathers' incarceration on low-income, urban, school-
aged children, Pediatrics 120 (2007) e678–e685.
 K.C. Winters, G.A. Henly, Western Psychological Services, Personal Experience
Inventory (PEI): Manual, Western Psychological Services, Los Angeles, Calif, 1989.
 W.R. Yates, R.J. Cadoret, E.P. Troughton, M. Stewart, T.S. Giunta, Effect of fetal
alcohol exposure on adult symptoms of nicotine, alcohol, and drug dependence,
Alcohol Clinical and Experimental Research 22 (1998) 914–920.
D.A. Frank et al. / Neurotoxicology and Teratology 33 (2011) 100–109