ORIGINAL

ARTICLES
Prenatal Exposure to Tobacco and Offspring Neurocognitive
Development in the Healthy Start Study
Brianna F. Moore, PhD1,2, Allison L. Shapiro, PhD3, Greta Wilkening, PsyD, PhD4, Sheryl Magzamen, PhD5,
Anne P. Starling, PhD1,2, William B. Allshouse, PhD6, John L. Adgate, PhD6, and Dana Dabelea, MD, PhD1,2,4

Objective To explore the associations between prenatal exposure to tobacco and neurocognitive development,
in the absence of prematurity or low birth weight.
Study design We followed mother-child pairs within Healthy Start through 6 years of age. Children were born at
³37 weeks of gestation with a birth weight of ³2500 g. Parents completed the Third Edition Ages and Stages Ques-
tionnaire (n = 246) and children completed a subset of the National Institutes of Health Toolbox Cognition Battery
(n = 200). The Ages and Stages Questionnaire domains were dichotomized as fail/monitor and pass. Maternal uri-
nary cotinine was measured at approximately 27 weeks of gestation. Separate logistic regression models estimated
associations between prenatal exposure to tobacco (cotinine below vs above the limit of detection) and the Ages
and Stages Questionnaire domains. Separate linear regression models estimated associations between prenatal
exposure to tobacco and fully corrected T-scores for inhibitory control, cognitive flexibility, and receptive language,
as assessed by the National Institutes of Health Toolbox. A priori covariates included sex, maternal age, maternal
education, daily caloric intake during pregnancy, race/ethnicity, household income, maternal psychiatric disorders,
and, in secondary models, postnatal exposure to tobacco.
Results Compared with unexposed offspring, exposed offspring were more likely to receive a fail/monitor score
for fine motor skills (OR, 3.9; 95% CI, 1.5-10.3) and decreased inhibitory control (B: 3.0; 95% CI, 6.1 to 0.7).
After adjusting for postnatal exposure, only the association with fine motor skills persisted.
Conclusions Prenatal and postnatal exposures to tobacco may influence neurocognitive development, in the
absence of preterm delivery or low birth weight. Increased developmental screening may be warranted for exposed
children. (J Pediatr 2019;-:1-7).

A
lthough tobacco use in the US has declined,1 approximately 7% of women actively smoke during pregnancy2 25% are
exposed to secondhand smoke.3 This is concerning because prenatal exposure to tobacco has been linked to impaired
neurocognitive development in offspring. Children born to mothers who smoked during pregnancy are less likely to
meet age-appropriate developmental milestones and may have delays in fine or gross motor function.4-8 Additionally, exposed
offspring may exhibit impaired cognitive abilities in focused attention and response inhibition.9-15 These findings also have
been shown in animal models.16
Prenatal exposure to tobacco is a well-established risk factor for preterm delivery and low birth weight.17,18 Because children
born preterm or at a low birth weight are at greater risk for cognitive and motor impairment, these studies may confound risks
of prematurity and low birth weight with the risks of tobacco exposure.19,20 To date, only 1 published study restricted their
analyses to offspring with normal birth histories.21
Women who smoke during pregnancy, even those who attempt to quit, often smoke in the postpartum period.22 Continued
exposure to tobacco during early childhood may influence childhood neurocog-
nitive development.20 However, it is unclear whether the association between
prenatal exposure to tobacco and neurocognitive development is independent 1
From the Department of Epidemiology, and Lifecourse 2

of postnatal exposure to tobacco. Epidemiology of Adiposity and Diabetes (LEAD) Center,

Colorado School of Public Health; Department of 3

Finally, there is a need to evaluate the relationship between prenatal expo- 4

Psychiatry, and Department of Pediatrics, University of
5
Colorado School of Medicine, Aurora, CO; Department
sure to tobacco and offspring neurocognitive development using an objective of Environmental and Radiological Health Sciences,
Colorado State University, Fort Collins, CO; and the
measure of tobacco exposure. Self-report of smoking during pregnancy may 6
Department of Environmental and Occupational Health,
23
result in exposure misclassification. Cotinine, the major metabolite of nico- Colorado School of Public Health, Aurora, CO

tine,24 is a more accurate indicator of exposure and may reduce exposure Supported by the National Institutes of Health
(R01DK076648, 5UG3OD023248, R01ES022934,
misclassification. K99ES028711). The authors declare no conflicts of in-
terest.
Portions of this study were presented as a poster at the
31st Annual Conference for the International Society of
ASQ-3 Ages and Stages Questionnaire, Third Edition Environmental Epidemiology (ISEE 2019), August 25-29,
2019, Utrecht, The Netherlands.
DCCS Dimensional change card sort test
NIH National Institutes of Health 0022-3476/$ - see front matter. ª 2019 Elsevier Inc. All rights reserved.
https://doi.org/10.1016/j.jpeds.2019.10.056

1
THE JOURNAL OF PEDIATRICS  www.jpeds.com Volume -  - 2019

In this analysis, we explored the association between pre- failing score (n = 9 for fine motor, n = 7 for gross motor,
natal exposure to tobacco (measured by maternal urinary co- n = 1 for communication, n = 3 for problem solving, and
tinine at 27 weeks of gestation) and offspring neurocognitive n = 5 for personal/social skills). Therefore, the ASQ-3
development at age 54 months among mother-child pairs domains were dichotomized as fail/monitor and pass.
enrolled in the longitudinal Healthy Start study. We hypoth- The National Institutes of Health (NIH) Toolbox Cogni-
esized that prenatal exposure to tobacco would be associated tion Battery is a series of tests designed to measure executive
with cognitive and motor impairment in early childhood, function across the lifespan (ages 3-85 years).28 Three tests in
in the absence of preterm delivery or low birth weight and the cognition battery were relevant for our study population:
independent of exposure to secondhand smoke in early the Flanker test (inhibitory control),28 the dimensional
childhood. change card sort test (DCCS, cognitive flexibility),29 and
the picture vocabulary test (receptive language). During the
in-person research visit, children completed the tests on a
Methods tablet computer while a trained professional research assis-
tant supervised. Raw scores were based on accuracy and
The Healthy Start study enrolled 1410 women ³16 years of response time (Flanker and DCCS) or accuracy (picture vo-
age and before 24 weeks of gestation with singleton pregnan- cabulary). Fully corrected T-scores were used,30 with a
cies from the obstetrics clinics at the University of Colorado normative mean fully corrected T-score as 50 for all tests
Hospital from 2010 to 2014. Participants completed 2 and a SD of 10.30
research visits during pregnancy (median, 17 and 27 weeks Mothers were asked to report the number of adults in the
of gestation), and another at delivery (median, 1 day after de- household (including themselves) who were regular smokers
livery). Women were not eligible to participate in the Healthy when their child was 5, 18, and 54 months of age. Responses
Start study if they had a previous stillbirth or preterm birth at to this question ranged from 0 to 6. We dichotomized these
<25 weeks of gestation or had preexisting diabetes, asthma, data into no household smokers vs any household smokers
cancer, or psychiatric illness. (³1 household smoker at 5, 18, or 54 months of age). This
Mother-child pairs were eligible for the current analysis if variable was used as the indicator for postnatal exposure to
they had exposure (urinary cotinine) and developmental tobacco.
outcome data. Mother-child pairs were excluded if born at Covariates were offspring sex, birth weight, and gestational
<37 weeks of gestation or the offspring was low birth weight age (obtained from medical records). Maternal race/
(<2500 g). The Healthy Start study protocol was approved by ethnicity, maternal education, and annual household income
the Colorado Multiple Institutional Review Board. All were self-reported. Although maternal psychiatric illness was
women provided written informed consent before the first part of the initial exclusion criteria, some women did not self-
study visit. The Healthy Start study was registered as an report a condition or were diagnosed after recruitment into
observational study to explore the fuel-mediated program- our study. Therefore, we obtained information about
ming of neonatal adiposity (NCT02273297), but has maternal psychiatric disorders (nonspecified) via medical re-
expanded its scope to explore how exposures in early life in- cords. Maternal daily caloric intake was measured using the
fluence childhood growth and development. Automated Self-Administered 24-hour Dietary Recall, an on-
Cotinine was measured in a subsample of women with line platform developed and hosted by the National Cancer
stored urine samples collected at approximately 27 weeks Institute (Automated Self-Administered 24-hour Dietary
of gestation. Cotinine was measured via solid phase compet- Recall-Beta and ASA24-2011, Bethesda, Maryland). The
itive ELISA, with a sensitivity of 1 ng/mL (Calbiotech Cotin- duration of exclusive breastfeeding was ascertained via
ine ELISA CO096D, Calbiotech, El Cajon, California). The questionnaire at age 5 months and was dichotomized as
limit of detection was 0.05 ng/mL. Urinary cotinine was cate- <5 months and >5 months.
gorized as no exposure (below the limit of detection), expo-
sure to secondhand smoke (above the limit of detection to Statistical Analyses
550 ng/mL; the established cut point for active smoking25), Logistic regression models were used to estimate associations
and active smoking (³550 ng/mL). Because only 15 mothers between prenatal cotinine categories (no exposure vs any
were classified as active smokers during pregnancy, prenatal exposure) and the 5 dichotomized domains of the ASQ-3
exposure to tobacco was defined as maternal urinary cotinine as separate outcomes. Our base models adjusted for covari-
levels above the limit of detection (indicating active and ates that are related to both childhood development
secondhand smokers). and maternal smoking during pregnancy (listed in Table I;
Development was assessed at 48, 54, or 60 months available at www.jpeds.com), as well as maternal
using the Ages and Stages Questionnaire, Third Edition psychiatric disorder (yes, no), and maternal daily caloric
(ASQ-3).26,27 The ASQ-3 assesses fine motor, gross motor, intake during pregnancy (kCal).31,32 In addition to these
communication, problem-solving, and personal/social devel- covariates, our secondary models also adjusted for
opmental domains. Scores for each domain were categorized postnatal exposure to tobacco (none, any).
as either fail, monitor, or pass based on the cut-offs provided Linear regression models estimated associations between
in the ASQ-3 User’s Guide.26 Very few children received a the prenatal cotinine categories and fully corrected T-scores
2 Moore et al
- 2019 ORIGINAL ARTICLES

for inhibitory control (Flanker task), cognitive flexibility lower birth weight than offspring born to nonexposed
(DCCS), and receptive language (picture vocabulary test) women (P = .02). There was a statistically significant
as separate outcomes. The fully corrected T-scores adjust difference in the duration of exclusive breastfeeding among
for age, sex, race/ethnicity, and maternal education. Base exposed and nonexposed women. One-half of the
models adjusted for maternal age, annual household income, nonexposed women exclusively breastfed their infants
nonspecified maternal psychiatric disorder (yes, no), and ³5 months, whereas only 25% of the exposed women did
maternal daily caloric intake during pregnancy (kCal). In so. Exposed and nonexposed participants did not differ in
addition to these covariates, our secondary models also maternal daily caloric intake during pregnancy (P = .07),
adjusted for postnatal exposure to tobacco (none, any). offspring sex (P = .46), gestational age (among term births)
We calculated aORs or means and beta coefficients with (P = .32), child age at NIH Toolbox Cognition Battery
corresponding 95% CIs for each separate model. An alpha assessment (P = .46), child age at ASQ-3 assessment
level of 0.05 was used to determine statistical significance. (P = .38), or the ASQ-3 version completed (P = .14).
All analyses were performed using Stata, version 14 (Stata-
Corp LP, College Station, Texas). NIH Toolbox Cognition Battery and ASQ-3
Secondary analyses explored the association between pre- Response inhibition (Flanker) was lower among children
natal exposure to tobacco and ASQ-3 categories at 18 months with an ASQ-3 fail/monitor score for fine motor (Table III
of age. We also repeated our base model analyses among [available at www.jpeds.com]; P = .02), gross motor
those with ASQ-3 results at 18 and 54 months of age (P = .03), personal-social (P = .02), and problem-solving
(n = 133). skills (P = .02). There were no differences in response
inhibition across the categories of the ASQ-3
Results communication domain (P = .28). There were no
differences in cognitive flexibility (DCCS) or receptive
Healthy Start initially enrolled 1410 participants (Figure; language (picture vocabulary) across the five ASQ-3
available at www.jpeds.com). Of these, 689 did not have domains.
cotinine measured in stored urinary samples collected Table IV presents the association between prenatal
during pregnancy and 19 were missing information on exposure to tobacco and the ASQ-3 domains. Compared
maternal psychiatric disorders. An additional 55 with nonexposed offspring, exposed offspring had 3.9 times
participants were excluded owing to birth before 37 weeks the odds of a fail/monitor score for fine motor skills (95%
(n = 30) or a birth weight of <2500 g (n = 25). Of the CI, 1.5-10.3). The association between prenatal exposure
eligible sample (n = 647), 401 did not complete the ASQ-3 to tobacco and fine motor skills remained statistically
at approximately 54 months of age; the ASQ-3 is valid significant after adjusting for postnatal exposure to tobacco
through 66 months of age and some children were no (aOR, 3.3; 95% CI, 1.1-10.2). No significant associations
longer eligible when they attended the visit. Therefore, the were observed between prenatal exposure to tobacco and
final analytic sample for the ASQ-3 analyses was 246. Of the other ASQ-3 domains (gross motor, personal-social,
the eligible sample (n = 647), 447 did not complete the communication, or problem solving skills).
NIH Toolbox Cognition Battery because this assessment Table V presents the association between prenatal
was introduced later during the study. Therefore, the final exposure to tobacco and fully corrected T-scores for the
analytic sample for the NIH Toolbox analyses was 200. NIH Toolbox assessments. Compared with nonexposed
There were no differences in maternal age, race/ethnicity, offspring, exposed offspring exhibited decreased inhibitory
education, or offspring sex between the entire cohort control on the Flanker test (adjusted beta coefficient, 3.0;
(n = 1410) and the ASQ-3 sample (n = 246) (Table I). 95% CI, 6.1 to 0.7). This association was no longer
Women whose children completed the NIH Toolbox statistically significant after adjustment for postnatal
Cognition Battery were more likely to be non-Hispanic exposure to tobacco (adjusted beta coefficient, 2.5;
White, to have an annual household income >$70 000, and 95% CI, 5.9 to 1.0). No significant associations were
to have at least some college education. Maternal age or observed between prenatal exposure to tobacco and
offspring sex did not differ between participants from the cognitive flexibility (DCCS) or receptive language (picture
entire cohort (n = 1410) and the NIH toolbox sample vocabulary test).
(n = 200).
Characteristics of the analytic sample are summarized in Secondary Analyses
Table II. Based on maternal urinary cotinine, a majority of When we restricted our analyses to the subsample of children
the women were classified as having no exposure (n = 181, with ASQ-3 assessed at 18 and 54 months of age (n = 133), we
74%). Women with any exposure to tobacco during did not detect an association between prenatal exposure to
pregnancy were younger than those with no exposure tobacco and a fail/monitor score for fine motor skills at
(P<.01). Nonexposed women were more likely to be non- 18 months of age (Table VI; available at www.jpeds.com).
Hispanic white (P<.01), to have an annual household Consistent with our main analyses, no significant
income above $70 000 (P<.01), and to have attended associations were observed between prenatal exposure to
college (P<.01). Offspring born to exposed women had a tobacco and the other ASQ-3 domains (gross motor,
Prenatal Exposure to Tobacco and Offspring Neurocognitive Development in the Healthy Start Study 3
THE JOURNAL OF PEDIATRICS  www.jpeds.com Volume -

Table II. Characteristics of eligible mothers and children in the Healthy Start study
Prenatal cotinine categories
No exposure Any exposure
Characteristics All (n = 246) (cotinine < LOD) (n = 181) (cotinine ‡ LOD) (n = 65) P value
Mother characteristics
Age, y 29  6 30  6 27  6 <.01
Race/ethnicity
Non-Hispanic white 55 61 37
Non-Hispanic black 12 4 34
Hispanic 28 31 20
Other 5 4 9 <.01
Household income
<$40 000 26 20 42
$40 001-$70 000 13 14 12
>$70 000 39 48 17
Don’t know 21% 18 30 <.01
Highest level of education
<High school 15 11 25
High school degree 15 12 25
Some college or more 70 77 51 <.01
Maternal psychiatric disorder (nonspecified)
No 85 89 74
Yes 15 11 26 <.01
Maternal daily caloric intake during pregnancy (kCal) 2009  648 1966  574 2133  818 .07
Child characteristics
Male 52 53 48 .46
Birth weight (g) 3333  420 3364  421 3245  408 .02
Gestational age at birth (weeks) 40  1 40  1 40  1 .32
Age at NIH Toolbox Cognition Battery assessment, mo 55  4 54  4 54  3 .46
Age at ASQ-3 assessment, mo 54  3 54  3 54  4 .38
ASQ-3 version, mo
48 22 22 27
54 66 67 54
60 12 11 18 .14
Duration of exclusive breastfeeding, mo
<5 55 49 75
³5 45 51 25 <.01

LOD, limit of detection.

Continuous variables are expressed as means  SD. Independent samples t tests were used to examine the differences in means by cotinine categories. Categorical variables are expressed as
proportions of column totals. c2 tests were used to examine differences in proportions by cotinine categories.

personal-social, communication, or problem-solving skills) motor development and reduced inhibitory control in
at age 18 months. children born at ³37 weeks of gestation and birth weights
of ³2500 g.
Discussion Tobacco smoke is a complex mixture of >5000 chemicals
and compounds.33 Although many of these individual con-
Our results confirm earlier findings of less optimal neurocog- stituents may contribute to neurocognitive delays, neuroi-
nitive development among children exposed to tobacco in maging studies suggest that nicotine is especially toxic to
utero. In addition, we provide evidence that early life expo- the developing brain. Nicotine is a vasoconstrictor and
sure to tobacco smoke is associated with less optimal fine decreases uterine blood flow to the placenta.34 This decrease

Table IV. aORs for a fail/monitor score on separate ASQ-3 domains, by maternal urinary cotinine categories
aORs for fail/monitor score for separate ASQ-3 domains (95% CIs)
Cotinine categories n Fine motor Gross motor Personal-social Communication Problem solving
Model 1*
<0.5 ng/mL (LOD) 181 1 [Reference] 1 [Reference] 1 [Reference] 1 [Reference] 1 [Reference]
³0.5 ng/mL 65 3.9 (1.5-10.3); P < .01 1.6 (0.4-5.9); P = .46 0.9 (0.2-4.0); P = .96 1.2 (0.2-6.1); P = .80 1.2 (0.4-4.0); P = .72
Model 2†
<0.5 ng/mL (LOD) 172 1 [Reference] 1 [Reference] 1 [Reference] 1 [Reference] 1 [Reference]
³0.5 ng/mL 62 3.3 (1.1-10.2); P = .03 1.8 (0.4-7.5); P = .45 0.8 (0.2-3.9); P = .80 0.7 (0.1-3.9); P = .65 0.9 (0.2-3.1); P = .82

*Model 1 adjusted for maternal age (years), maternal education (<high school, high school diploma, some college), maternal race/ethnicity (non-Hispanic white, non-Hispanic black, Hispanic, other),
offspring sex, a nonspecified maternal psychiatric diagnosis (yes, no), and maternal daily caloric intake during pregnancy (kCal).
†Model 2 adjusted for model 1 covariates, as well as self-report of a household smoker in early childhood (none, any).

4 Moore et al
- 2019 ORIGINAL ARTICLES

Table V. Adjusted means and beta coefficients for fully adjusted T-scores for NIH Toolbox Cognition Battery tests,
according to maternal urinary cotinine categories
Adjusted means and beta coefficients (95% CIs)
Cotinine categories n Flanker test Dimensional change card sort Picture vocabulary test
Model 1*
<0.5 ng/mL (LOD) 152 51.5 (50.2 to 52.9) 49.6 (47.9 to 51.2) 50.8 (48.6 to 53.0)
³0.5 ng/mL 48 3.0 ( 6.1 to 0.7); P = .04 1.8 ( 1.8 to 5.5); P = .32 1.8 ( 3.0 to 6.6); P = .83
Model 2†
<0.5 ng/mL (LOD) 149 51.3 (49.9 to 52.7) 49.4 (47.7 to 51.0) 51.6 (49.3 to 53.8)
³0.5 ng/mL 48 2.5 ( 5.9 to 1.0); P = .16 2.8 ( 1.3 to 6.8); P = .18 0.1 ( 5.5 to 5.2); P = .96
*Model 1 adjusted for maternal age (years), annual household income (<$40 000, $40 001-$70 000, >$70 000, missing or do not know), a nonspecified maternal psychiatric diagnosis (yes, no), and
maternal daily caloric intake during pregnancy (kCal).
†Model 2 adjusted for model 1 covariates, as well as self-report of a household smoker in early childhood (none, any).

results in fetal hypoxia with sustained deprivation of nutri- control, in the absence of low birth weight or preterm deliv-
ents and oxygen. Nicotine-induced fetal hypoxia may lead ery. Consistent with our results, no clear relationships have
to significant changes in brain structures involved in been established for the association between self-report of
learning and memory, such as decreased volume in the maternal active smoking during pregnancy with receptive
cortical areas35 and the amygdala.36 Additionally, nicotine language or cognitive flexibility.14,43-46
may act as a neuro teratogen by overstimulating nicotinic Our results may have implications for the impact of prena-
acetylcholine receptors.37 These receptors are abundant in tal exposure to tobacco smoke on overall neurocognitive
the cerebellum, which plays an important role in motor development. Fine motor skills are essential for early
control and coordination,38 and the hippocampus, which learning. A majority of classroom activities in kindergarten
is responsible for memory and learning.39 Nicotine expo- involve fine motor skills, such as coloring, copying, cutting,
sure may contribute to damage of the nicotinic cholinergic and drawing.47 Early life fine motor function is often associ-
system in the offspring cerebellum, resulting in subsequent ated with later academic achievement, especially in mathe-
motor dysfunction.38 Dysfunction of nicotinic acetylcholine matics.48 Furthermore, some studies, including data from
receptors in the offspring hippocampus has been linked to the present study, have shown that fine motor coordination
cognitive deficits.39 is associated with inhibitory control.49 Inhibitory control is
Consistent, positive associations between prenatal expo- an essential first step in solving complex problems. As chil-
sure to tobacco and delayed motor development have been dren become capable of inhibiting responses to distractions,
described in the literature.4-8,21 Most of these studies were other executive functions (such as working memory and
conducted among preschool-aged children. Only three cognitive flexibility) can develop to allow them to negotiate
studies have examined these associations among tod- increasingly complex problems.50 Furthermore, both fine
dlers.4,6,21 Contrary to our findings, Gusella and Fried4 re- motor skills and inhibitory control have been linked to fluid
ported weaker fine motor skills at age 13 months, and intelligence, or the ability to think logically and solve prob-
Evlampidou et al reported weaker gross motor skills at age lems in novel situations.51
18 months.6 These studies included infants born at <37 weeks It is difficult to disentangle the interplay between prenatal
of gestation or with a low birth weight, which may have and postnatal exposure to tobacco in their relationship with
contributed to the positive results. Among a population of offspring neurocognitive development. Prenatal and post-
children in Korea with normal birth histories, Lee et al failed natal exposures to tobacco may act synergistically to influ-
to detect an association between prenatal exposure to to- ence neurodevelopment. However, due to the low sample
bacco and motor development at age 24 months.21 Thus, sizes within exposure subgroups, we were unable to specif-
the impact of prenatal exposure to tobacco on delayed motor ically test for an interaction. In secondary analyses, we
development may not emerge until later in life among chil- included postnatal exposure to tobacco as a covariate. After
dren who are not typically considered to be at risk for devel- adjusting for postnatal exposure to tobacco, the association
opmental delays. between prenatal exposure to tobacco and fine motor skills
Research has established an association between self- remained significant, but the association with inhibitory con-
report of maternal active smoking during pregnancy and trol was attenuated. Fine motor skills are well-established by
poorer offspring performance with inhibitory tasks at age preschool age, whereas cognition continues to develop
4-18 years.9-15 Additionally, neuroimaging studies suggest rapidly throughout adolescent.52 This finding suggest that
that prenatal exposure to tobacco is associated with increased pregnancy may be the most susceptible developmental period
activation in brain regions related to response inhibition dur- for offspring motor development, whereas postnatal expo-
ing a Flanker/NoGo task among adolescents and young sures continue to influence offspring cognitive development.
adults.40-42 Our data demonstrated that prenatal exposure Prospective cohorts with sufficiently large subgroups of chil-
to tobacco alone was associated with impaired inhibitory dren with objective assessment of both prenatal and postnatal
Prenatal Exposure to Tobacco and Offspring Neurocognitive Development in the Healthy Start Study 5
THE JOURNAL OF PEDIATRICS  www.jpeds.com Volume -

exposure to tobacco are needed to explore these questions Reprint requests: Brianna F. Moore, PhD, Postdoctoral Fellow, Colorado
School of Public Health, Epidemiology, 12474 E 19th Ave, Aurora, CO 80045.
more conclusively. E-mail: [email protected]
A strength of our approach was the use of a biomarker to
objectively characterize prenatal exposure to tobacco.
Compared with self-report of exposure during pregnancy,
References
maternal urinary cotinine is more likely to capture prenatal 1. Wang TW, Asman K, Gentzke AS, Cullen KA, Holder-Hayes E, Reyes-
Guzman C, et al. Tobacco product use among adults - United States,
exposure to tobacco. This factor was especially true for 2017. MMWR Morbid Mortal Wkly Rep 2018;67:1225-32.
secondhand exposures, which are more likely to be under- 2. Drake P, Driscoll AK, Mathews TJ. Cigarette smoking during pregnancy:
reported than active smoking among pregnant women.23 United States, 2016. NCHS Data Brief. Hyattsville (MD): National Cen-
However, cotinine cannot differentiate the source and type ter for Health Statistics; 2018. p. 1-8.
of exposure. In addition to tobacco products, nicotine expo- 3. Homa DM, Neff LJ, King BA, Caraballo RS, Bunnell RE, Babb SD,
et al. Vital signs: disparities in nonsmokers’ exposure to secondhand
sure can arise from nicotine replacement therapy, as well as
smoke–United States, 1999-2012. MMWR Morb Mortal Wkly Rep
consumption of certain foods, such as tomatoes, potatoes, 2015;64:103-8.
and black tea.53 Although cotinine is not tobacco specific, co- 4. Gusella JL, Fried PA. Effects of maternal social drinking and smoking on
tinine tends to agree with tobacco-specific nitrosamines, such offspring at 13 months. Neurobehav Toxicol Teratol 1984;6:13-7.
as 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol.54 There- 5. Polanska K, Hanke W, Sobala W, Trzcinka-Ochocka M, Ligocka D,
Brzeznicki S, et al. Developmental effects of exposures to environmental
fore, it is likely that the potential for exposure misclassifica-
factors: the Polish Mother and Child Cohort Study. BioMed Res Int
tion is low. The decrease in exposure measurement error 2013;2013:629716.
may have provided a more accurate representation of the 6. Evlampidou I, Bagkeris M, Vardavas C, Koutra K, Patelarou E, Koutis A,
true association between prenatal exposure to tobacco and et al. Prenatal second-hand smoke exposure measured with urine cotin-
childhood neurocognitive development. ine may reduce gross motor development at 18 months of age. J Pediatr
2015;167:246-52.e2.
A limitation of our study is the diminished ability to estab- 7. Hsieh CJ, Liao HF, Wu KY, Hsieh WS, Su YN, Jeng SF, et al. CYP1A1
lish causality, given the observational nature of this study. Ile462Val and GSTT1 modify the effect of cord blood cotinine on neuro-
Another limitation is the relatively small number of partici- development at 2 years of age. Neurotoxicology 2008;29:839-45.
pants with the ASQ-3 and NIH Toolbox assessments. 8. Trasti N, Vik T, Jacobsen G, Bakketeig LS. Smoking in pregnancy and
Maternal smoking during pregnancy is associated with lower children’s mental and motor development at age 1 and 5 years. Early Hu-
man Dev 1999;55:137-47.
socioeconomic status, lower educational attainment,
Jacobson SW,
9. Boucher O, Jacobson JL, Burden MJ, Dewailly E,
maternal psychopathology, malnutrition of the mother dur- Muckle G. Prenatal tobacco exposure and response inhibition in
ing pregnancy, and a shorter duration of breastfeeding. school-aged children: an event-related potential study. Neurotoxicol
Although we adjusted for many of these covariates, it remains Teratol 2014;44:81-8.
difficult to causally attribute the impaired developmental 10. Cornelius MD, De Genna NM, Leech SL, Willford JA, Goldschmidt L,
Day NL. Effects of prenatal cigarette smoke exposure on neurobehavioral
outcomes to smoking itself. Co-use of tobacco and other sub-
outcomes in ten-year-old children of teenage mothers. Neurotoxicol
stances, such as cannabis, cocaine, or opioids, during preg- Teratol 2011;33:137-44.
nancy is common and may contribute to neurocognitive 11. Huijbregts SCJ, Warren AJ, de Sonneville LMJ, Swaab-Barneveld H. Hot
delays through similar mechanisms.55,56 However, these and cool forms of inhibitory control and externalizing behavior in chil-
data were not available in the present study. Therefore, we dren of mothers who smoked during pregnancy: an exploratory study. J
Abnormal Child Psychol 2008;36:323-33.
cannot rule out the possibility of confounding by co-use of
12. Julvez J, Ribas-Fito N, Torrent M, Forns M, Garcia-Esteban R, Sunyer J.
other substances during pregnancy. Maternal smoking habits and cognitive development of children at age 4
Early life exposure to tobacco continues to be an important years in a population-based birth cohort. Int J Epidemiol 2007;36:825-32.
public health concern. Although many women attempt to 13. Micalizzi L, Marceau K, Brick LA, Palmer RH, Todorov AA, Heath AC,
quit smoking during pregnancy, ³1 in 3 children in the US et al. Inhibitory control in siblings discordant for exposure to maternal
smoking during pregnancy. Dev Psychol 2018;54:199-208.
are exposed to some level of tobacco in utero.2,3 After birth,
14. Daseking M, Petermann F, Tischler T, Waldmann H. Smoking during
exposure to tobacco becomes more prevalent. Parents who pregnancy is a risk factor for executive function deficits in preschool-
smoke during pregnancy continue to smoke after delivery.22 aged children. Geburtshilfe und Frauenheilkunde 2015;75:64-71.
Among women who quit smoking during pregnancy, relapse 15. Piper BJ, Corbett SM. Executive function profile in the offspring of
in the early postpartum period is common.22 Furthermore, women that smoked during pregnancy. Nicotine Tobacco Res 2012;14:
191-9.
there is concern that the prevalence of this exposure may in- 16. Abbott LC, Winzer-Serhan UH. Smoking during pregnancy: lessons
crease, as more youth, the future generation of parents, adopt learned from epidemiological studies and experimental studies using an-
the use of e-cigarettes.57 The results of our study, coupled imal models. Crit Rev Toxicol 2012;42:279-303.
with recent trends in smoking prevalence and market shifts 17. Kallen K. Maternal smoking during pregnancy and infant head circum-
to different nicotine products, suggest that it is important ference at birth. Early Human Dev 2000;58:197-204.
18. Jaddoe VW, Troe EJ, Hofman A, Mackenbach JP, Moll HA, Steegers EA,
to encourage parents to quit smoking and limit their chil-
et al. Active and passive maternal smoking during pregnancy and the
dren’s exposure to nicotine and tobacco during and after risks of low birthweight and preterm birth: the Generation R Study. Pae-
pregnancy. n diatr Perinat Epidemiol 2008;22:162-71.
19. Blencowe H, Lee AC, Cousens S, Bahalim A, Narwal R, Zhong N, et al.
Submitted for publication Jun 18, 2019; last revision received Sep 23, 2019; Preterm birth-associated neurodevelopmental impairment estimates at
accepted Oct 22, 2019. regional and global levels for 2010. Pediatr Res 2013;74(Suppl 1):17-34.

6 Moore et al
- 2019 ORIGINAL ARTICLES

20. Herrmann M, King K, Weitzman M. Prenatal tobacco smoke and post- 39. Yakel JL. Cholinergic receptors: functional role of nicotinic ACh recep-
natal secondhand smoke exposure and child neurodevelopment. Curr tors in brain circuits and disease. Pflugers Arch 2013;465:441-50.
Opinion Pediatr 2008;20:184-90. 40. Bennett DS, Mohamed FB, Carmody DP, Bendersky M, Patel S,
21. Lee M, Ha M, Hong Y-C, Park H, Kim Y, Kim E-J, et al. Exposure to pre- Khorrami M, et al. Response inhibition among early adolescents prena-
natal secondhand smoke and early neurodevelopment: Mothers and tally exposed to tobacco: an fMRI study. Neurotoxicol Teratol 2009;31:
Children’s Environmental Health (MOCEH) study. Environ Health 283-90.
2019;18:22. 41. Holz NE, Boecker R, Baumeister S, Hohm E, Zohsel K, Buchmann AF,
22. Tong VT, Jones JR, Dietz PM, D’Angelo D, Bombard JM. Trends in et al. Effect of prenatal exposure to tobacco smoke on inhibitory control:
smoking before, during, and after pregnancy - Pregnancy Risk Assess- neuroimaging results from a 25-year prospective study. JAMA Psychia-
ment Monitoring System (PRAMS), United States, 31 sites, 2000-2005. try 2014;71:786-96.
Morbid Mortal Wkly Rep Surveill Summ 2009;58:1-29. 42. Longo CA, Fried PA, Cameron I, Smith AM. The long-term effects of
23. Moore BF, Starling AP, Magzamen S, Harrod CS, Allshouse WB, prenatal nicotine exposure on response inhibition: an fMRI study of
Adgate JL, et al. Fetal exposure to maternal active and secondhand smok- young adults. Neurotoxicol Teratol 2013;39:9-18.
ing with offspring early-life growth in the Healthy Start study. Int J Obes 43. Huijbregts SCJ, S
eguin JR, Zelazo PD, Parent S, Japel C, Tremblay RE.
2019;43:652-62. Interrelations between maternal smoking during pregnancy, birth
24. Benowitz N, Goniewicz ML, Eisner MD, Lazcano-Ponce E, Zielinska- weight and sociodemographic factors in the prediction of early cognitive
Danch W, Koszowski B, et al. Urine cotinine underestimates exposure abilities. Infant Child Dev 2006;15:593-606.
to the tobacco-derived lung carcinogen 4-(methylnitrosamino)-1-(3- 44. Eskenazi B, Trupin LS. Passive and active maternal smoking during preg-
pyridyl)-1-butanone in passive compared with active smokers. Cancer nancy, as measured by serum cotinine, and postnatal smoke exposure. II.
Epidemiol Biomarkers Prev 2010;19:2795-800. Effects on neurodevelopment at age 5 years. Am J Epidemiol 1995;142:
25. Zielinska-Danch W, Wardas W, Sobczak A, Szoltysek-Boldys I. Estima- S19-29.
tion of urinary cotinine cut-off points distinguishing non-smokers, pas- 45. Makin J, Fried PA, Watkinson B. A comparison of active and passive
sive and active smokers. Biomarkers 2007;12:484-96. smoking during pregnancy: long-term effects. Neurotoxicol Teratol
26. Squires J, Twombly E, Bricker D, Potter L. ASQ-3 user’s guide. Baltimore 1991;13:5-12.
(MD): Brookes Publishing; 2009. 46. Ramsay H, Barnett JH, Murray GK, M€aki P, Hurtig T, Nordstr€ om T,
27. Schonhaut L, Armijo I, Schonstedt M, Alvarez J, Cordero M. Validity of et al. Smoking in pregnancy, adolescent mental health and cognitive per-
the Ages and Stages Questionnaires in term and preterm infants. Pediat- formance in young adult offspring: results from a matched sample
rics 2013;131:e1468-74. within a Finnish cohort. BMC Psychiatry 2016;16.
28. Weintraub S, Dikmen SS, Heaton RK, Tulsky DS, Zelazo PD, Bauer PJ, 47. Marr D, Cermak S, Cohn ES, Henderson A. Fine motor activities in Head
et al. Cognition assessment using the NIH Toolbox. Neurology 2013;80: Start and kindergarten classrooms. Am J Occup Ther 2003;57:550-7.
S54-64. 48. Pitchford NJ, Papini C, Outhwaite LA, Gulliford A. Fine motor skills
29. Beck DM, Schaefer C, Pang K, Carlson SM. Executive function in pre- predict maths ability better than they predict reading ability in the early
school children: test-retest reliability. J Cogn Dev 2011;12:169-93. primary school years. Front Psychol 2016;7:783.
30. Casaletto KB, Umlauf A, Beaumont J, Gershon R, Slotkin J, 49. Rigoli D, Piek JP, Kane R, Oosterlaan J. An examination of the relation-
Akshoomoff N, et al. Demographically corrected normative standards ship between motor coordination and executive functions in adoles-
for the English version of the NIH Toolbox cognition battery. J Int Neu- cents. Dev Med Child Neurol 2012;54:1025-31.
ropsychol Soc 2015;21:378-91. 50. Pennequin V, Sorel O, Fontaine R. Motor planning between 4 and 7
31. Alvik A. Variables predicting low infant developmental scores: maternal years of age: changes linked to executive functions. Brain Cogn
age above 30 years is a main predictor. Scand J Public Health 2014;42: 2010;74:107-11.
113-9. 51. van der Fels IM, Te Wierike SC, Hartman E, Elferink-Gemser MT,
32. Kerstjens JM, Bos AF, ten Vergert EM, de Meer G, Butcher PR, Smith J, Visscher C. The relationship between motor skills and cognitive
Reijneveld SA. Support for the global feasibility of the Ages and Stages skills in 4-16 year old typically developing children: a systematic review. J
Questionnaire as developmental screener. Early Human Dev 2009;85: Sci Med Sport 2015;18:697-703.
443-7. 52. Zysset AE, Kakebeeke TH, Messerli-B€ urgy N, Meyer AH, St€ ulb K,
33. Borgerding M, Klus H. Analysis of complex mixtures–cigarette smoke. Leeger-Aschmann CS, et al. Predictors of executive functions in pre-
Exp Toxicol Pathol 2005;57(Suppl 1):43-73. schoolers: findings from the SPLASHY Study. Front Psychol 2018;9:
34. Walsh RA. Effects of maternal smoking on adverse pregnancy outcomes: 2060.
examination of the criteria of causation. Hum Biol 1994;66:1059-92. 53. Davis RA, Stiles MF, deBethizy JD, Reynolds JH. Dietary nicotine: a
35. El Marroun H, Schmidt MN, Franken IH, Jaddoe VW, Hofman A, van source of urinary cotinine. Food Chem Toxicol 1991;29:821-7.
der Lugt A, et al. Prenatal tobacco exposure and brain morphology: a 54. Moore BF, Clark ML, Bachand A, Reynolds SJ, Nelson TL, Peel JL. Inter-
prospective study in young children. Neuropsychopharmacology actions Between diet and exposure to secondhand smoke on metabolic
2014;39:792-800. syndrome among children: NHANES 2007-2010. J Clin Endocrinol
36. Haghighi A, Schwartz DH, Abrahamowicz M, Leonard GT, Perron M, Metab 2016;101:52-8.
Richer L, et al. Prenatal exposure to maternal cigarette smoking, amygdala 55. Oga EA, Mark K, Coleman-Cowger VH. Cigarette Smoking Status and
volume, and fat intake in adolescence. JAMA Psychiatr 2013;70:98-105. Substance Use in Pregnancy. Matern Child Health J 2018;22:1477-83.
37. Slotkin TA, Southard MC, Adam SJ, Cousins MM, Seidler FJ. Alpha7 56. Morris CV, DiNieri JA, Szutorisz H, Hurd YL. Molecular mechanisms of
nicotinic acetylcholine receptors targeted by cholinergic developmental maternal cannabis and cigarette use on human neurodevelopment. Eur J
neurotoxicants: nicotine and chlorpyrifos. Brain Res Bull 2004;64:227-35. Neurosci 2011;34:1574-83.
38. Blood-Siegfried J, Rende EK. The long-term effects of prenatal nicotine 57. Arrazola RA, Singh T, Corey CG, Husten CG, Neff LJ, Apelberg BJ, et al.
exposure on neurologic development. J Midwife Women’s Health Tobacco use among middle and high school students - United States,
2010;55:143-52. 2011-2014. MMWR Morbid Mortal Wkly Rep 2015;64:381-5.

Prenatal Exposure to Tobacco and Offspring Neurocognitive Development in the Healthy Start Study 7
THE JOURNAL OF PEDIATRICS  www.jpeds.com Volume -

Table I. Characteristics of the Healthy Start cohort, the

ASQ-3 sample, and the NIH Toolbox subsample
ASQ-3 NIH Toolbox
Healthy Start sample sample
Characteristics cohort (n = 1410) (n = 246) (n = 200)
Maternal age, y 28  6 29  6 30  5
Race/Ethnicity
Non-Hispanic white 53 55 69
Non-Hispanic black 15 12 11
Hispanic 25 28 16
Other 6 5 5
Household income
<$40 000 30 26 19
$40 001-$70 000 20 13 18
>$70 000 32 39 52
Don’t know 18 21 12
Highest level of education
<High school 14 15 7
High school degree 18 15 13
Some college 67 70 80
Male offspring 53 52 55

Continuous variables are expressed as means  SD. Categorical variables are expressed as
proportions of column totals.

Table III. Means and SD of fully adjusted T-scores of NIH Toolbox cognition tasks by ASQ-3 domains
Fine motor Gross motor Communication Personal-social Problem solving
P P P P P
Tasks Fail/monitor Pass value Fail/monitor Pass value Fail/monitor Pass value Fail/monitor Pass value Fail/monitor Pass value
Flanker 47  9 50  8 .02 45  13 50  8 .03 48  5 50  8 .28 45  10 50  8 .02 45  5 50  8 .02
DCCS 49  6 50  10 .34 50  10 52  10 .17 45  8 50  10 .11 46  10 50  10 .13 45  6 50  10 .07
Picture 51  13 52  11 .65 49  12 51  13 .23 47  10 51  13 .15 45  11 51  13 .06 48  15 51  13 .24
vocabulary

7.e1 Moore et al
- 2019 ORIGINAL ARTICLES

Table VI. aORs for a fail/monitor score on separate ASQ-3 domains, among a subset of children with ASQ-3 assessed
at 18 months of age (n = 133)
Adjusted* ORs for fail/monitor ASQ score (95% CI)
Cotinine categories n Fine motor Gross motor Personal-social Communication Problem solving
<0.5 ng/mL (LOD) 97 1 [Reference] 1 [Reference] 1 [Reference] 1 [Reference] 1 [Reference]
³0.5 ng/mL 36 0.9 (0.2-3.8) 1.4 (0.1-20.9) Empty 0.7 (0.2-2.9) 1.0 (0.1-4.6)

LOD, limit of detection.

*All models adjusted for maternal age (years), maternal education (<high school, high school diploma, some college), maternal race/ethnicity (non-Hispanic white, non-Hispanic black, Hispanic,
other), offspring sex, and a nonspecified maternal psychiatric diagnosis (yes, no).

Healthy Start cohort

n=1410

Missing cotinine: 689

Gestation weeks<37: 30
Birth weight<2500g: 25
Missing maternal psychiatric disorder: 19

Potentially eligible
n=647

Missing Ages and Stages Missing NIH Toolbox

Questionnaire (ASQ-3) at Cognition Battery at ~54
~54 months of age: 401 months of age: 447

Analytic sample for Analytic sample for

ASQ-3 analyses NIH Toolbox analyses
n=246 n=200

Figure. Flow of participants through the study.

Prenatal Exposure to Tobacco and Offspring Neurocognitive Development in the Healthy Start Study 7.e2