ALCOHOLISM: CLINICAL AND EXPERIMENTAL RESEARCH

Vol. 38, No. 1

January 2014

The Association of Mild, Moderate, and Binge Prenatal

Alcohol Exposure and Child Neuropsychological Outcomes:
A Meta-Analysis
Audrey L. Flak, Su Su, Jacquelyn Bertrand, Clark H. Denny, Ulrik S. Kesmodel, and
Mary E. Cogswell

Background: The objective of this review is to evaluate the literature on the association between
mild, moderate, and binge prenatal alcohol exposure and child neurodevelopment.
Methods: Meta-analysis with systematic searches of MEDLINE (1970 through August 2012),
EMBASE (1988 through August 2012), and PsycINFO (1970 through August 2012) and examination
of selected references.
Results: From 1,593 articles, we identied 34 presenting data from cohort studies that met our
inclusion criteria. Information on study population, outcomes, measurement instruments, timing and
quantication of alcohol exposure, covariates, and results was abstracted. Outcomes included academic
performance, attention, behavior, cognition, language skills, memory, and visual and motor development. The quality of each article was assessed by 2 researchers using the NewcastleOttawa Scale.
Based on 8 studies of 10,000 children aged 6 months through 14 years, we observed a signicant detrimental association between any binge prenatal alcohol exposure and child cognition (Cohens d [a
standardized mean dierence score] 0.13; 95% condence interval [CI], 0.21, 0.05). Based on 3
high-quality studies of 11,900 children aged 9 months to 5 years, we observed a statistically signicant
detrimental association between moderate prenatal alcohol exposure and child behavior (Cohens d
0.15; 95% CI, 0.28, 0.03). We observed a signicant, albeit small, positive association between
mild-to-moderate prenatal alcohol exposure and child cognition (Cohens d 0.04; 95% CI, 0.00, 0.08),
but the association was not signicant after post hoc exclusion of 1 large study that assessed mild consumption nor was it signicant when including only studies that assessed moderate alcohol consumption. None of the other completed meta-analyses resulted in statistically signicant associations
between mild, moderate, or binge prenatal alcohol exposure and child neuropsychological outcomes.
Conclusions: Our ndings support previous ndings suggesting the detrimental eects of prenatal
binge drinking on child cognition. Prenatal alcohol exposure at levels less than daily drinking might be
detrimentally associated with child behavior. The results of this review highlight the importance of
abstaining from binge drinking during pregnancy and provide evidence that there is no known safe
amount of alcohol to consume while pregnant.
Key Words: Prenatal Alcohol Exposure, Child Neurodevelopment, Systematic Review,
Meta-Analysis.

From the Department of Epidemiology (ALF), Rollins School of

Public Health, Emory University, Atlanta, Georgia; Oak Ridge Institute
for Science and Education (ALF, SS), Oak Ridge, Tennessee; National
Center on Birth Defects and Developmental Disabilities (ALF, JB,
CHD), Centers for Disease Control and Prevention, Atlanta, Georgia;
National Center for Immunization and Respiratory Diseases (SS), Centers for Disease Control and Prevention, Atlanta, Georgia; Department of
Epidemiology (USK), School of Public Health, Aarhus University,
Aarhus, Denmark; Department of Obstetrics and Gynecology (USK),
Aarhus University Hospital, Aarhus, Denmark; and National Center for
Chronic Disease Prevention and Health Promotion (MEC), Centers for
Disease Control and Prevention, Atlanta, Georgia.
Received for publication July 16, 2012; accepted May 15, 2013.
Reprint requests: Audrey L. Flak, MPH, Department of Epidemiology, Emory University, CNR 3rd Floor, 1518 Clifton Road, Mailstop
1518-002-3BB, Atlanta, GA 30332; Tel.: 404-727-8710; Fax: 404-7278737; E-mail: aak@emory.edu
The ndings and conclusions in this report are those of the authors and
do not necessarily represent the ocial position of the Centers for Disease
Control and Prevention.
Published 2013. This article is a U.S. Government work and is in the
public domain in the U.S.A.
DOI: 10.1111/acer.12214
214

N THE UNITED States, from 1991 through 2005, 54%

of women aged 18 to 44 years reported consuming at least
1 drink of alcohol during the past 30 days. During the same
time period, 12% of pregnant women reported consuming at
least 1 drink of alcohol during the past 30 days (Centers for
Disease Control and Prevention, 2009). Despite the known
consequences of heavy prenatal alcohol exposure (often
dened as 1 drink per day) (Bailey and Sokol, 2008; U.S.
Department of Health and Human Services, 2005), including
fetal alcohol syndrome and other fetal alcohol spectrum
disorders (FASDs), the eects of mild-to-moderate (>0 to
6 drinks per week) and binge (usually dened as 4 or 5
drinks per occasion) prenatal alcohol exposure on neurodevelopment are inconsistent. Due to the large population of
women of childbearing age in the United States who consume alcohol and the high rates of unintended pregnancy,
even small eects of prenatal alcohol use could have detrimental repercussions to overall child neurodevelopment at
the population level (Centers for Disease Control and
Prevention, 2009; U.S. Department of Health and Human
Alcohol Clin Exp Res, Vol 38, No 1, 2014: pp 214226

PRENATAL ALCOHOL AND CHILD OUTCOMES

Services, 2000). Neurodevelopment refers to development of

the nervous system that includes a number of functional
domains, such as academic achievement, attention, behavior,
cognition, language development, memory, and motor
development.
In a previous meta-analysis, Testa and colleagues (2003)
found no consistent association between light (<1 drink per
day) and moderate (1 to 1.99 drinks per day) prenatal
alcohol consumption and infant mental development. While
their results were not consistent across child ages, they did
nd signicant associations between both levels of alcohol
consumption and infant mental development in 12- to 13month-old children. Testa and colleagues (2003) focused on
general mental development among infants aged 6 to
26 months and did not examine the eects of binge prenatal
alcohol consumption. In a 2007 systematic review, Henderson and colleagues (2007) concluded prenatal binge drinking
might be associated with neurodevelopment, but did not
attempt a meta-analysis due to the limited number and heterogeneity of studies. The 4 studies included by this review
examined neurodevelopment in children aged 18 months
through 14 years. More recently, in a 2011 systematic review,
Bay and Kesmodel concluded that light levels of prenatal
alcohol exposure (1 to 2 drinks per day) were not associated
with motor dysfunction in individuals aged 3 days through
26 years. Another recent review aimed at informing advice
for pregnant women concluded that recent studies indicate
an association between light and moderate prenatal alcohol
exposure and neurodevelopmental problems (OLeary and
Bower, 2012). This review summarized the eects of previously published systematic reviews as well as research studies
in 2009 to 2010, but did not focus on neurodevelopmental
outcomes nor did it include a meta-analysis.
A systematic review including older children (but not
adults), several functional domains of neurodevelopment,
and various levels of prenatal alcohol exposure with appropriate meta-analysis would allow a synthesis of the most
recent research on this topic and quantify summary eects
(Borenstein et al., 2009; Stroup et al., 2000). One of the primary advantages of a meta-analysis is the improvement in
sample size and statistical power by combining like studies
and outcomes. In addition, it allows for the assessment of
homogeneity of results between dierent studies and the
evaluation of evidence of publication bias.
The 3 main objectives of this review were to: (i) evaluate
evidence for an association between mild-to-moderate prenatal alcohol exposure (>0 to 6 drinks per week) and child neuropsychological outcomes; (ii) examine the evidence for an
association between binge prenatal alcohol exposure (usually
dened as 4 or 5 drinks per occasion) and child neuropsychological outcomes; and (iii) identify gaps in our knowledge
and directions for further research. We hypothesized that the
literature on mild-to-moderate prenatal alcohol exposure
and child neuropsychological outcomes would not provide a
discernible or consistent eect (positive or negative). For prenatal binge drinking, we hypothesized that the literature

215

would reveal a detrimental eect on at least some of the

neuropsychological outcomes examined.
MATERIALS AND METHODS
Search Strategy
We searched 3 databases using OvidSP: MEDLINE (1970
through August 2012), EMBASE (1988 through August 2012), and
PsycINFO (1970 through August 2012). Search strategies included
the keywords alcohol, drinking behavior, and fetal development, and a list of outcomes of interest such as cognition disorders (see inclusion criteria that follow). Final search results
excluded editorials, letters, reviews, and articles in languages other
than English. In collaboration with a medical librarian, we modied
the search strategies used by Henderson and colleagues (2007) to
include mild and moderate alcohol exposure as well as additional
terms on neurodevelopment and to exclude terms for outcomes not
of interest (e.g., preterm delivery). The nal search strategy is provided in Table S1. Bibliographies of included articles and systematic
reviews relating to this subject were searched manually for articles
missed by electronic searches.
Study Selection
An article was included in this review if it presented data from a
cohort or casecontrol study on the relation between mild, moderate, or binge prenatal alcohol exposure and 1 or more child neuropsychological outcomes (Table 1). These outcomes included
cognition, motor skills, language, behavior, vision, hearing, development, information processing, academic achievement, attention,
memory, executive function, mental health, social skills, and hand
eye coordination. For the purposes of this review, mild, mild-tomoderate, moderate, and heavy drinking were dened as up to 3
drinks per week, up to 6 drinks per week, up to 6 drinks per week
including some individuals who consumed at least 3 drinks per
week, and more than 6 drinks per week, respectively, with 13.7 g of
alcohol equaling 1 drink. There are no internationally acknowledged standard denitions of mild, moderate, and heavy alcohol
consumption during pregnancy. The categories used in this review
were dened by the authors to focus on drinking at less than daily
levels and after careful review of exposure categories used in the literature. Binge drinking was dened as 4 or more drinks on 1 occasion (binge drinking dened as 5 or more drinks on 1 occasion were
included as a subset; National Institute on Alcohol Abuse and
Alcoholism, 2004; Wechsler et al., 1995).
Articles that met the initial inclusion criteria were examined
against a list of exclusion criteria (Table 1) such as including children with an explicit diagnosis of an FASD or lack of a comparison
group of children with no prenatal alcohol exposure. Individuals
with a diagnosed FASD were excluded because current evidence has
indicated that these diagnoses are associated with levels of alcohol
exposure above the mild-to-moderate level examined in this review
and meta-analysis (Bertrand et al., 2004). Our aim was to investigate the neuropsychological eects of prenatal alcohol exposure
below levels previously associated with these clinical diagnoses.
Initially, we screened the title and abstract of each identied article using our inclusion criteria and full list of exclusion criteria. An
article was excluded with no further review if it clearly did not meet
our article requirements. If the abstract contained any indication of
relevance to the review, 1 author also screened the full article
(Fig. 1).
Data Abstraction
A study researcher trained in epidemiology abstracted data
from the articles marked for inclusion using a standard form.

FLAK ET AL.

216

Table 1. Systematic Review Inclusion and Exclusion Criteria

Inclusion criteria
1. Cohort or casecontrol design
2. Includes data on the relation between mild, moderate, or binge prenatal
alcohol exposure and at least 1 child neuropsychological outcome of
interest
Exclusion criteria
1. Includes only the following outcomes: infant and child growth and
anthropometric measures, congenital anomalies, outcomes at birth, or
child or adolescent alcohol and substance abuse
2. Does not use a standardized scale to measure outcome
3. Includes children with diagnoses of fetal alcohol syndrome or fetal
alcohol eects
4. Lacks a comparison group of mothers with no alcohol exposure
5. Focuses primarily on other drug use with alcohol included only as a
confounder or adjustment variable
6. Examines only children whose mothers were exposed to alcohol in
combination with other drug exposures (e.g., tobacco, cocaine
[i.e., does not include exposure to alcohol alone])
7. Alcohol abuse or alcoholism is the only exposure during the
prenatal period
8. Alcohol exposure measured as a continuous variable, all exposure categories include women who drink >7 drinks per week, and investigators
assume a linear association between alcohol intake and neurodevelopment
(also applies if alcohol exposure transformed in the analysis in attempts to
compensate for skewness or outliers, or both)
9. Contains information only on the same cohort and outcome(s) as
already included articles

Information extracted from articles included study design, population, outcomes, outcome measurement instruments, timing and
quantication of alcohol exposure, measured covariates, and
results. To assess reliability, about half of the included articles were
abstracted by 1 or more additional researchers. The research team
collectively resolved any discrepancies in screening or abstraction. If
articles contained information on the same children and outcomes,
the outcome measurement at the participants oldest age was
selected. We chose to include outcomes at older ages as in general,
clinical issues are more likely to be identied as development
advances and for children with prenatal alcohol exposure, abnormalities in subtle mental health issues or higher-order neurocognition (e.g., executive function) may only emerge or be assessed at
older ages. If more than 1 measurement on the same outcome was
completed for the same child at the same age, we chose the measurement based on maternal report (over paternal or teacher report).
For full data selection criteria, see Table 2. If an article presented
results stratied by a characteristic (e.g., child sex) and results could
not be combined, then data in each strata were abstracted and
included in the meta-analysis.
Authors of included articles were contacted for additional data if
the data presented on 1 or more associations in their papers were
incomplete and could not be used in the meta-analysis (e.g., data
did not provide a measure of variance). Author-supplied data were
included in the meta-analysis. In cases for which authors no longer
had access to the data and information was incomplete, the data
were excluded from this review.
Outcome Classication
For the purposes of this review, we divided neuropsychological
outcomes into 8 functional domains: academic performance,
attention, behavior, cognition, language and verbal development,
memory, executive function, and visual and motor development.
Other neurodevelopmental outcomes (e.g., mental health) were
examined separately. This grouping prevented instruments measuring dierent constructs from being included in the same analysis.

Two psychologists reviewed each instrument independently and

classied them into 1 of these domains. All discrepancies in
outcome classication were discussed, and a consensus was reached.
Quality Assessment
The quality of each article meeting the review criteria was
assessed by 2 authors using an adapted NewcastleOttawa Scale
(NOS) for assessing the quality of nonrandomized studies in metaanalyses (Wells et al., 2013). Dierences were settled by discussion.
The adapted scale is provided in Table 3 and includes an assessment
of the following potential confounders: socioeconomic status (SES),
cigarette smoking, and maternal age and intelligence. We specically noted whether studies of high quality controlled for SES (i.e.,
income or education), given the strong associations between SES
and both prenatal alcohol use and child neuropsychological outcomes (Bradley and Corwyn, 2002; Centers for Disease Control and
Prevention, 2009). Detailed NOS score ratings for each of the
papers may be obtained from the authors.
Statistical Analysis
Measures of eect from binary and correlational data were converted to Cohens d values, a standardized mean dierence score,
using the methods outlined by Borenstein and colleagues (2009).
These conversions prevented studies from being excluded from this
review based on their use of measures of association (such as odds
ratios [ORs]). Random eects meta-analyses were completed separately for each of the neurodevelopmental domains and exposure
quantity combinations using Cohens d as the summary measure.
For each exposureoutcome meta-analysis, a measure of overall
eect and measures of heterogeneity (i.e., Q statistic, I2) were calculated. A sensitivity analysis based on study quality and a publication
bias assessment using funnel plots and the Egger test were completed (results available from the authors upon request; Egger et al.,
1997). Comprehensive Meta-Analysis 2.0 software was used for all
meta-analyses (Borenstein et al., 2005).

RESULTS
Systematic Review Results
Of the 1,593 articles reviewed, data from 34 met all criteria
and were used in meta-analyses (Fig. 1). The majority of
articles were excluded after review of the titles and abstracts
(n = 1,289), with a smaller group excluded after review of the
study methods and data in the full articles (n = 270). The
most common reasons for exclusion were lack of information
on prenatal alcohol exposure quantity (n = 588) and measuring outcomes other than neuropsychological outcomes
(n = 466). Authors were contacted to request additional
information on 23 reviewed articles. We obtained the desired
information for 16 of these articles. The authors for the
remaining 7 articles either no longer had access to the desired
information or did not reply to our requests.
Study data on 1 or more of the associations between neuropsychological outcomes and mild or moderate prenatal alcohol exposure were in 22 articles (Alati et al., 2008; Bay et al.,
2012; Brown et al., 2010; Forrest et al., 1991; Jacobson et al.,
1993a,b, 2011; Kaplan-Estrin et al., 1999; Kelly et al., 2009,
2012; Kesmodel et al., 2012; Larkby et al., 2011; Larroque
et al., 2000; OCallaghan et al., 2007; OLeary et al., 2010;

PRENATAL ALCOHOL AND CHILD OUTCOMES

217

Articles pulled from systematic searches (n = 1,835)

MEDLINE 1970-2012 (n = 720)
EMBASE 1988-2012 (n = 815)
PsycINFO 1970-2012 (n = 300)
Additional articles from bibliographies screened (n = 30)
Articles identified by outside experts (n = 10)
Duplicates excluded (n = 282)
Total nonduplicates (n = 1,593)
Articles excluded after title and abstract review (n = 1,289):
No prenatal alcohol exposure measure (n = 412)
No outcome of interest (n = 433)
No alcohol exposure of interest (e.g. only examined the effect of daily drinking
during pregnancy) (n = 57)
Not original research (n = 60)
Only includes children diagnosed with a fetal alcohol spectrum disorder (n = 144)
Not a human study (n = 98)
Other (e.g. no control group without any alcohol exposure) (n = 85)
Articles retained for full article review (n = 304)
Articles excluded after full article review due to (n = 270):
No prenatal alcohol exposure measure (n = 48)
No outcome of interest (n = 33)
No alcohol exposure of interest (e.g. only examined the effect of daily drinking
during pregnancy) (n = 71)
Not original research (n = 22)
No control group without any alcohol exposure (n = 32)
Primarily focuses on other drug use with alcohol included only as a confounder
or adjustment variable (n = 9)
Does not include a measure of association that can be included in this review
(n = 4)
Other (e.g. only includes children diagnosed with fetal alcohol spectrum disorder)
(n = 51)
Articles retained for meta-analysis (n = 34)
Fig. 1. Summary of article review process including primary reasons for article exclusion.

Parry and Ogston, 1992; Richardson et al., 1995; Robinson

et al., 2010; Rodriguez et al., 2009; Sayal et al., 2007; Sood
et al., 2001; Willford et al., 2006; Table 4). Data on 1 or more
associations between neuropsychological outcomes and
binge prenatal alcohol exposure were in 15 articles (Alati
et al., 2008; Alvik et al., 2011; Bailey et al., 2004; Coles et al.,
2000; Fraser et al., 2012; Goldschmidt et al., 2004; Kesmodel
et al., 2012; Lemola et al., 2009; Nulman et al., 2004;
OCallaghan et al., 2007; Olsen, 1994; Streissguth et al.,
1989, 1994a,b; Willford et al., 2004; Table 5). The most

common outcomes examined were cognition, behavior, and

visual and motor development (data in 16, 14, and 13 articles,
respectively). Academic development, attention, language,
memory, and executive function each were in 6 or fewer articles. Outcomes not tting in these domains, such as mental
health, were in 4 articles. The scarcity of data on these additional outcomes precluded further analysis of their results.
Study quality scores, assessed using the adapted NOS scale,
ranged from 2 to 8 on an 8-point scale (with 8 representing a
study of the highest quality; Tables 4 and 5). A study earning

FLAK ET AL.

218

Table 2. Data Selection Criteria

Data selection criteria applied when: (i) A given outcome was assessed
multiple times in the same cohort, or (ii) Multiple cohorts prenatally
exposed to alcohol were compared with the same nonexposed cohort
1. When a given outcome was assessed by multiple scales or at multiple time
points in the same children, the following rules were applied:
a.
b.
c.
d.

Parental assessment was included instead of teacher assessment

Maternal assessment was included instead of paternal assessment
Results measuring the outcome at the oldest age were included
The most standardized, comprehensive measures were included

2. When multiple cohorts of children prenatally exposed to alcohol were

compared with the same nonexposed cohort, only 1 comparison could be
included in each analysis. This prevented including the same control
group more than once in a given meta-analysis. The following rules were
applied in this situation:
a. If the exposed children were all exposed prenatally to binge drinking,
the cohort of children with the lightest exposure was included (e.g.,
children whose mothers binge drank once a week during pregnancy
were included instead of children whose mothers binge drank 3 times
a week during pregnancy)
b. If the cohorts of exposed children were prenatally exposed to alcohol
during dierent trimesters, the cohort of children with the earliest
exposure was included

6 or more points was deemed of high quality for the purposes of sensitivity analyses. This cut-point was decided upon
after examining the distribution of quality scores.

Table 3. Adapted NewcastleOttawa Quality Assessment Scale: Cohort

Studies
Maximum
points
Selection
1. Representativeness of the exposed cohort
a. The researchers attempted to select participants that
were reasonably representative of the average low,
moderate, or binge drinkers in the community (1 point)
b. The researchers and the reader have reason
to believe that their participants are reasonably
representative of the average low, moderate, or
binge drinkers in the community (1 point)
c. Either no data on representativeness, or no
reason to believe the participants were
representative of the average low, moderate,
or binge drinkers in the community, or
over-sampled heavy drinkers (0 points)

2. Selection of the nonexposed cohort

a. Drawn from the same community as the
exposed cohort (1 point)
b. Drawn from a dierent source than the
exposed cohort (0 points)
c. No description of the derivation of the
nonexposed cohort (0 points)

3. Ascertainment of exposure
a. Structured interview (1 point)
b. Self-administered questionnaire (0 points)
c. No description (0 points)

4. Demonstration that outcome of interest

was not present at start of study (omitted)

Meta-Analysis Results
We conducted meta-analyses on all exposureoutcome
associations, with data from 2 or more separate populations
resulting in 21 separate meta-analyses. By exposure quantity,
our meta-analyses were as follows: (i) mild exposure and 4
outcomes (behavior, cognition, language and verbal, and
visual and motor), (ii) moderate exposure and 5 outcomes
(attention, behavior, cognition, language and verbal, and
visual and motor), (iii) mild-to-moderate exposure and 3 outcomes (behavior, cognition, and visual and motor), and (iv)
binge exposure and 9 outcomes (academic reading performance, academic math performance, attention, behavior,
cognition, language and verbal, memory, visual and motor,
and executive function).
When we used data from all studies without accounting
for quality (i.e., NOS scores), we did not nd any signicant
associations between mild, moderate, or mild-to-moderate
prenatal alcohol exposure and neuropsychological outcomes
(i.e., attention, behavior, cognition, visual and motor development, and language skills; Fig. S1 presents all calculated
meta-analyses of mild and moderate alcohol exposure with
nonsignicant results). When meta-analyses were limited to
studies of high quality as determined by NOS scores, 2 of the
observed associations were statistically signicant. Based on
3 studies with approximately 11,900 children aged 9 months
to 5 years, we observed a statistically signicant detrimental
association between moderate prenatal alcohol exposure and
child behavior (Cohens d 0.15; 95% condence interval
[CI], 0.28, 0.03; p = 0.01; Fig. 2). The associations of all

Comparability
1. Comparability of cohorts on the basis
of the design or analysis
a. Study controls for SES (could assess using
proxy measures such as education or income) (1 point)
b. Study controls for any of: cigarette smoking,
maternal age, maternal IQ (1 point)
Outcome
1. Assessment of outcome
a. Independent blind assessment (1 point)
b. Not a blind assessment (0 points)
c. No description of outcome assessment (0 points)

2. Follow-up long enough for outcomes to occur

a. Yes (1 point)
b. No (0 points)

3. Adequacy of follow-up cohorts

a. Complete follow-up or subjects lost to follow-up
unlikely to introduce bias: 70% follow-up
and description provided of those lost indicates
they are comparable to those kept on SES
and prenatal alcohol exposure (1 point)
b. Follow-up rate <70% and description of
those lost does not indicate comparability (0 points)
c. Does not fall under (a) or (b) or no statement
describing the adequacy of follow-up cohorts (0 points)

Total points possible

SES, socioeconomics status.

studies in this subanalysis were in the same direction and

adjusted for SES, but only 1 was statistically signicant by
itself. This study, by Brown and colleagues (2010), was conducted among 9-month-old infants and assessed behavior

Richardson/1995, United States

Robinson/2010, Australia

645 (84.5)q

Parry/1992, Scotland, Denmark,

Germany

1,952 (68.1)
2,127 (74.2)j
2,037 (71.0)k
1,977 (68.9)r
1,744 (60.8)s

1995 to 1997d

1,890 (85)i
1,624 (73)j
1,357 (61)k
1,361 (63.9)l
1,360 (63.8)m

OLeary/2009, Australia

1989 to 1991

1986 to 1986n
1988 to 1989o
1987 to 1988p
1983 to 1986

1982 to NR
1985 to 1986
1981 to 1984

592 (71.4)
156 (47.9)
5,139 (71.1)g
3,731 (51.7)h

Larkby/2011, United States

Larroque/2000, France
OCallaghan/2007, Australia

2003 to 2008

1,628 (51.1)

NR
2000 to 2002d
2000 to 2002d

NR
1986 to 1989
1986 to 1989

1985 to 1986

1991 to 1992d
2003 to 2008
2001d

Enrollment period

Kesmodel/2012, Denmark

92 (24.1)
9,460 (75.9)
11,513 (93.6)

382 (NR)
262 (NR)
310 (76.9)

Jacobson/1993a, United States

Jacobson/2011, United States
Jacobson/1993b, United States

Kaplan-Estrin/1999, United States

Kelly/2009, United Kingdom
Kelly/2010, United Kingdom

592 (70.0)

4,332 (32.5)
685 (86.9)
10,500e (NR)

Alati/2008, United Kingdom

Bay/2012, Denmark
Brown/2010, United States

Forrest/1991, Scotland

Sample size at follow-up

n (% of baseline)

First author/year, country

Cognition
Visual and Motor
Behavior

Visual and Motor

Cognition
Behavior
Language
Visual and Motor
Attention
Cognition
Other
Behavior
Other
Academic
Attention
Cognition
Attention
Behavior
Cognition
Cognition
Visual and Motor

Cognition
Visual and Motor
Behavior
Cognition
Visual and Motor
Cognition
Visual and Motor
Cognition
Attention
Behavior

Outcome categories
assesseda

2y
5y
8y
10 y
14 y

18 m

2y
5y
8y
18 m

16 y
4.5 y
14 y

5y

13 m
7.5 y
6.5 m
12 m
26 m
3y
5y

18 m

8y
5y
9m

Age at
follow-up

1st

1st, 2nd, 3rd

NR

1st, 2nd, 3rd

1st, 3rd
1st
1st, 3rd

1st, 2nd

1st, 2nd, 3rd

1st, 2nd, 3rd
1st, 2nd, 3rd

1st, 2nd, 3rd

1st, 2nd, 3rd
1st, 2nd, 3rd

1st, 2ndf

1st
1st, 2nd
3rd

Trimester(s)
assessed

70

Continued.

13.7

27.4 to 82.2

5
5
6

42.5
13.7 to 82.2
48.0

30 to 59

6
6
6

5
3
7

6
8
6

NOS
scorec

12.0 to 48.0

40.0 to 78.3
82.2
82.2

0.2 to 78.3
1.6 to 78.3
0.2 to 78.3

1.0 to 49.0

12.0 to 48.0
13.7 to 41.1

Moderate
(g/wk)

<38.4

29

0.2 to 39.8
27.4
27.4

<13.7

<13.7

Mild
(g/wk)

Prenatal alcohol exposure assessmentb

Table 4. Characteristics of Included Studies that Measured Mild or Moderate Levels of Prenatal Alcohol Exposure by Article

PRENATAL ALCOHOL AND CHILD OUTCOMES

219

Attention

Behavior
Behavior
Cognition

1990 to 1992t
1984 to 1987u
1986v
1991 to 1992d
1989 to 1991d
1983 to 1985

4,968 (60.3)t
7,844 (70.4)u
8,525 (91.1)v
8,046 (63.5)
506 (76.1)
611 (73.7)

Rodriguez/2009, Denmark,
Finland

Sayal/2007, United Kingdom

Sood/2001, United States
Willford/2006, United States

10 yt
12 yt
15 yu
7 to 8 yv
81 m
6 to 7 y
10 y

Age at
follow-up

1st
1st, 2nd, 3rd
1st

1st, 2nd, 3rdt,u

1st, 2ndv

Trimester(s)
assessed

<38.4

<13.7

Mild
(g/wk)

<48.0

13.7 to 54.8

Moderate
(g/wk)

Prenatal alcohol exposure assessmentb

5
3
3

NOS
scorec

g/wk, grams per week; NOS, NewcastleOttawa Scale; y, year; m, month; NR, not reported.
a
Visual and Motor = Visual and Motor Development, Academic = Academic Performance, Language = Language and Verbal.
b
Alcohol Exposure Quantification: All alcohol exposure categories were converted to grams per week using the conversion 13.7 g alcohol = 0.6 oz = 1 drink (U.S. Centers for Disease Control
and Prevention. Alcohol and Public Health FAQs. http://www.cdc.gov/alcohol/faqs.htm). The exposure categories reported by studies in grams were kept intact regardless of the conversions used
by the researchers. Alcohol categories were rounded to the nearest tenth. Mild exposure: any exposure up to 3 drinks per week (41.1 g/wk). Moderate exposure: any exposure up to 6 drinks per
week (82.2 g/wk) which included some individuals with exposure of at least 3 drinks per week (41.1 g/wk). In the event that a study had multiple exposure categories that fit this classification, the category that covered the largest range was chosen.
c
NOS Quality Assessment Score. Possible values: 1 (lowest quality) to 8 (highest quality).
d
Range of birth years instead of enrollment period.
e
Rounded to the nearest 50.
f
Sulaiman and colleagues (1988).
g
Attentional and learning questionnaires.
h
Psychometric assessment.
i
2 years.
j
5 years.
k
8 years.
l
Bayley Scales of Infant Development: Mental Development Index.
m
Bayley Scales of Infant Development: Psychomotor Development Index.
n
Dundee [UK] cohort (Bolumar, 1992).
o
Odense [Denmark] cohort (Bolumar, 1992).
p
Berlin [Germany] cohort (Bolumar, 1992).
q
18 months.
r
10 years.
s
14 years.
t
Aarhus Birth Cohort (ABC) [Denmark].
u
Habits for Two (HHT) [Denmark].
v
Northern Finland Birth Cohort (NFBC) [Finland].
< and grams of alcohol do not include 0. Outcomes and ages assessed by articles, but not included in meta-analyses, are not listed in this table.

Outcome categories
assesseda

Enrollment period

First author/year, country

Sample size at follow-up

n (% of baseline)

Table 4. (Continued)

220
FLAK ET AL.

PRENATAL ALCOHOL AND CHILD OUTCOMES

221

Table 5. Characteristics of Included Studies that Measured Binge Levels of Prenatal Alcohol Exposure by Article
Prenatal alcohol exposure
assessment
Sample size at follow-up
n (% of baseline)

Enrollment
period

Outcome categories
assesseda

Age at
follow-up

Trimester(s)
assessed

Binge (drinks/
occasion)

NOS
scoreb

Alati/2008, United Kingdom

Alvik/2011, Norway
Bailey/2004, United States

4,332 (32.5)
1,303 (69.6)
499 (75.0)d
537 (80.8)e

1991 to 1992c
2000 to 2001
1989 to 1991c

8y
6m
7y

2nd, 3rd
1st
1st, 2nd, 3rd

4+
5+
5+f

6
6
7

Coles/2000, United States

136 (41.7)

1993 to 1994g

12 m

1st, 2nd, 3rd

5+

Fraser/2012, Canada

195 (81.3)

NR

6m

1st, 2nd

5+

Goldschmidt/2004,
United States
Kesmodel/2012, Denmark

606 (79.4)

1983 to 1986h

10 y

1st, 2nd, 3rd

4+

1,628 (51.1)

2003 to 2008

5y

1st, 2nd

5+

Lemola/2009, Switzerland
Nulman/2004, Canada

323 (70.5)
102 (71.3)

NR
1987 to 1997

17 m
7 y

1st, 2nd, 3rd

1st

4+
5+

3
6

5,139 (71.1)i
3,731 (51.7)j

1981 to 1984

14 y

1st, 3rd

5+

276 (84.2)k
251 (76.5)l
486 (86.0)

1988 to 1989

18 m
3.5 y
7.5 y

1st, 2nd, 3rd

8+

1974 to 1975

1st, 2nd

5+

Streissguth/1994a, United States

Streissguth/1994b, United States

464 (82.0)
462 (82.0)

1974 to 1975m
1974 to 1975m

14 y
14 y

1st, 2nd
1st, 2nd

5+
5+

3
7

Willford/2004, United States

580 (70.0)

1983 to 1985n

Cognition
Behavior
Behavior
Language
Visual and Motor
Behavior
Cognition
Visual and Motor
Visual and Motor
Cognition
Academic
Other
Attention
Cognition
Executive Function
Behavior
Academic
Cognition
Language
Visual and Motor
Other
Academic
Attention
Cognition
Cognition
Visual and Motor
Behavior
Cognition
Language
Visual and Motor
Other
Academic
Attention
Executive Function
Memory
Memory

14 y

1st

4+

First author/year, country

OCallaghan/2007, Australia
Olsen/1994, Denmark
Streissguth/1989,
United States

NOS, NewcastleOttawa Scale; y, year; m, month; NR, not reported.

a
Visual and Motor = Visual and Motor Development, Academic = Academic Performance, Language = Language and Verbal.
b
NOS Quality Assessment Score. Possible values: 1 (lowest quality) to 8 (highest quality).
c
Range of birth years instead of enrollment period.
d
Behavior.
e
Language and Verbal, Visual and Motor Development.
f
At least once every 2 weeks during pregnancy.
g
Drews and colleagues (2003).
h
Goldschmidt and colleagues (1996).
i
Attention, Academic Performance.
j
Cognition.
k
18 months.
l
3.5 years.
m
Streissguth and colleagues (1981).
n
Willford and colleagues (2006).
Outcomes and ages assessed by articles, but not included in meta-analyses, are not listed in this table.

using the behavior rating scale of the Bayley Scales of Infant

Development (Bayley, 1993).
In a separate meta-analysis based on 7 studies with NOS
scores of 6 or more including approximately 26,100 children,
we observed a statistically signicant, albeit small, benecial
association between mild-to-moderate prenatal alcohol
exposure and cognition (Cohens d 0.04; 95% CI, 0.00, 0.08;
p = 0.03; Fig. 2). None of the associations observed in the

individual studies in this subanalysis were statistically signicant. Although not statistically signicant, the direction of 1
study, by Brown and colleagues (2010), was opposite to the
others and another, by Kelly and colleagues (2009), was centered at zero among girls, but not boys.
This association between mild-to-moderate prenatal alcohol exposure and cognition was of similar magnitude, but no
longer signicant in a post hoc analysis considering only the

FLAK ET AL.

222

6 studies that controlled for SES (excluding a study by Alati

et al. [2008] that accounted for SES in published analyses
including individuals with heavy prenatal alcohol exposure,
but not in published analyses that met our inclusion criteria;
Cohens d 0.04; 95% CI, 0.01, 0.09; p = 0.08). The association with cognition also was not signicant when including
only studies that assessed moderate alcohol consumption
(Fig. S1), even when limiting these studies to those that were
of high quality.
When including studies of all quality scores, we observed a
signicant detrimental association between binge prenatal
alcohol exposure and child cognition (Cohens d 0.13; 95%
CI, 0.21, 0.05; p < 0.01; Fig. 2). This analysis used data
on children aged 6 months to 14 years from 8 studies
(n  10,000). The results of this meta-analysis were borderline signicant when limited to data from studies of high
quality (n  9,000, p = 0.054).
None of the meta-analyses resulting in signicant associations between mild, moderate, or binge prenatal alcohol exposure and neuropsychological outcomes showed any evidence
of heterogeneity (determined by p-value > 0.05 for the Q statistic) or publication bias (Egger test p-value > 0.05).
All remaining analyses failed to show statistically signicant associations between binge prenatal alcohol exposure
and neuropsychological outcomes (Fig. S2 presents all calculated meta-analyses of binge alcohol exposure with nonsignicant results). The majority of these analyses included data
from more than 5 populations and showed no indication of
heterogeneity or publication bias.
DISCUSSION
This meta-analysis is the rst we know of to suggest that
moderate prenatal alcohol consumption at levels less than
daily drinking might aect child behavior. The studies used
in our meta-analysis on child behavior examined behavioral
aspects such as social engagement, aect, and conduct. While
in our meta-analysis, there were statistically signicant ndings showing that moderate levels of alcohol exposure were
associated with behavior, the clinical, or functional signicance of the results from the original studies varied from
subtle to moderate. However, across these studies, some
children with the lowest levels of moderate prenatal alcohol
exposure demonstrated behaviors of concern, including:
increased demand for attention, behavior regulation problems, and poorer interactive play skills.
Our systematic review and meta-analyses corroborated
other meta-analyses and systematic reviews and have provided further evidence for a strong association between binge
prenatal alcohol use and cognition (where binge alcohol use
is dened as drinking 4 or more drinks on 1 occasion). We
observed a robust detrimental association between such
exposure and diverse aspects of cognition in children aged
6 months to 14 years. Among 6 study populations for whom
this relation was assessed, children of mothers who engaged
in binge drinking during pregnancy scored lower on tests of

cognitive ability than children whose mothers did not binge

drink during pregnancy. Our meta-analysis included 5 additional studies, whose populations were not considered in the
systematic review by Henderson and colleagues (2007).
Importantly, 2 of these studies were conducted among older
children (i.e., aged 8 to 14 years), suggesting associations
observed among younger children extended to older children
as well.
In contrast to binge prenatal alcohol exposure, mildto-moderate prenatal alcohol exposure was not associated
consistently with cognition, corroborating the meta-analysis
by Testa and colleagues (2003), which also did not nd a consistent association between less than daily drinking and mental development at all ages examined (6 to 26 months). Our
analysis expanded on the one by Testa by including children
aged 14 years and younger, as well as data not available in
2003 on more than 20,000 children. The results from this
analysis showed a small, benecial association between mildto-moderate prenatal alcohol exposure and child cognition.
We suspect that this association was due to residual confounding as the association was no longer signicant when
considering only studies that controlled for SES.
We detected no consistent evidence that mild or moderate
prenatal alcohol exposure was associated with attention,
cognition, language skills, and visual or motor development,
or that binge drinking was associated with outcomes other
than cognition. Overall, these results align with our initial
hypotheses that (i) we would not observe a consistent association between mild and moderate prenatal alcohol exposure
and child neuropsychological outcomes and (ii) prenatal
binge drinking would be detrimentally associated with at
least some of the outcomes examined. Although 21 metaanalyses used data from more than 20 studies, the number
of high-quality studies in which investigators controlled for
SES was much lower, suggesting the need for further cohort
studies designed to rule out or at least minimize potential
confounding.
While our results suggest underlying associations between
prenatal alcohol consumption and child neurodevelopment,
there are other explanations to consider. Due to the number
of analyses completed, random eects might have played a
role in these results. In choosing estimates to include in analyses, preference was given to those that accounted for potential confounders such as SES, maternal intelligence, and
home environment. However, studies were not excluded if
they did not adequately control for those factors. The results
we observed might have been a remnant of some of the
eects of these uncontrolled factors. Various types of selection were not accounted for, including self-selection by
authors who provided additional data to this analysis compared with those who did not, as well as potential selection
bias within each of the included studies. Additionally, potential variances in the ways individuals were aected by prenatal exposures might have obscured underlying true eects.
That is, for some children, attention might have been
aected adversely, while for other children the same prenatal

PRENATAL ALCOHOL AND CHILD OUTCOMES

Reference

223

Cohens d,
Weight (%) Random (95% CI)

Cohens d,
Random (95% CI)

Association between moderate prenatal alcohol exposure and child behavior (High Quality)
38.1
-0.25 (-0.45, -0.06)
Brown 2010a
b
16.4
-0.14 (-0.44, 0.16)
Jacobson SW 1993
25.5
Kelly 2010 - Boys
-0.05 (-0.29, 0.19)
Kelly 2010 - Girls
19.9
-0.11 (-0.38, 0.16)
Combined
-0.15 (-0.28, -0.03)
99.9
Heterogeneity: Chi2=1.79, df=3 (P=0.62); I2=0%
Test for overall effect: Z=-2.49 (P=0.01)

Scale

Age

BRS
Elicited Play
SDQ
SDQ

9m
12m
5y
5y

-1.00
0.00
1.00
Favors unexposed Favors exposed

Association between mild-to-moderate prenatal alcohol exposure and child cognition (High Quality)
WISC-III
33.6
Alati 2008
0.04 (-0.02, 0.11)
a
MDI
3.6
Brown 2010
-0.15 (-0.35, 0.04)
MDI
4.8
Forrest 1991
0.05 (-0.12, 0.21)
7.6
Kelly 2009 - Boys
BSRA
0.08 (-0.06, 0.22)
6.2
Kelly 2009 - Girls
BSRA
0.00 (-0.15, 0.15)
WPSSI-R
12.9
Kesmodel 2012
0.03 (-0.07, 0.14)
Raven's
28.0
O'Callaghan 2007
0.06 (-0.01, 0.13)
3.3
Richardson 1995b
MDI
0.10 (-0.11, 0.31)
100.0
Combined
0.04 (0.00, 0.08)

8y
9m
18m
3y
3y
5y
14y
18m

-1.00
0.00
1.00
Favors unexposed Favors exposed

Heterogeneity: Chi2=4.99, df=7 (P=0.66); I2=0%

Test for overall effect: Z=2.20 (P=0.03)

Association between binge prenatal alcohol exposure and child cognition (All estimates)
-0.16 (-0.26, -0.06)
25.7
Alati 2008
-0.23 (-0.65, 0.20)
3.6
Coles 2000

WISC-III
MDI

8y
12m
6m
5y

Fraser 2012
Kesmodel 2012

6.3
24.3

0.02 (-0.29, 0.33)

0.01 (-0.09, 0.12)

FTII - NP
WPPSI-R

Nulman 2004
O'Callaghan 2007

1.9
19.5

-0.28 (-0.87, 0.32)

-0.19 (-0.32, -0.05)

McCarthy GCI 7y
14y
Raven's

Olsen 1994c
Streissguth 1989d

6.3
12.5

-0.14 (-0.45, 0.17)

-0.27 (-0.47, -0.08)

MDI
WISC-R

100.1

-0.13 (-0.21, -0.05)

Combined

Heterogeneity: Chi2=11.11, df=7 (P=0.13); I2=37%

Test for overall effect: Z=-3.02 (P=0.003)

18m
7.5y

-1.00
0.00
1.00
Favors unexposed Favors exposed

Fig. 2. Meta-analysis results for the associations between: (1) moderate prenatal alcohol exposure and child behavior, (2) mild-to-moderate prenatal
alcohol exposure and child cognition, and (3) binge prenatal alcohol exposure and child cognition. CI, confidence interval; SDQ, Strengths and Difficulties
Questionnaire; MDI, Bayley Scales of Infant Development: Mental Development Index; WISC, Wechsler Intelligence Scale for Children (III = Third UK
Edition; R = Revised); BRS, Behavior Rating Scale Social Engagement Subscale, Bayley Scales of Infant Development, Second Edition; BSRA, Bracken
School Readiness Assessment; Elicited Play, Complexity of Play TestElicited Play Level; McCarthy GCI, McCarthy Scales of Childrens Abilities: General Cognitive Index; Ravens, Ravens Standard Progressive Matrices Test; WPPSI-R, Wechsler Preschool and Primary Scale of IntelligenceRevised;
FTII-NP, FTII Novelty Preference aSample size and standard error obtained from author. bMean, standard deviation, and sample size obtained from
author. cUsed standard deviation calculated from other articles using this scale. dUsed approximate sample sizes and standard deviation of the overall
group.

exposure levels might have produced an eect on motor

function. Because both of these eects would not be reected
in an analysis focusing on only 1 of these outcomes, sucient
statistical power to detect subtle eects might not have been
available. This study also could not account for dierences

in drinking patterns and the fact that some individuals may

be more genetically susceptible to the eects of mild and
moderate drinking than others (Lewis et al., 2012). Finally,
the functional domain categories used for this analysis did
not represent mutually exclusive areas of neurodevelopment.

FLAK ET AL.

224

Measures used to assess 1 area almost always evaluate

related areas; for example, measures of behavior frequently
also assess aspects of attention. This overlap might have
diluted eects when single measures or domains were
analyzed.
The primary strengths of this analysis were the longitudinal design of included studies and the systematic search and
review process. The longitudinal design of studies allowed
for examination of temporal associations between the exposures and outcomes of interest. Our systematic review of the
literature, followed by screening articles using explicit
inclusion criteria, aided us in capturing articles published in
English relevant to this topic. Published data from studies
were supplemented by additional information gained by
contacting study authors.
The primary limitation of this review was the inconsistency
of the methodologies of the included studies. Alcohol consumption was measured via self-report during dierent
trimesters. Studies such as that by Robinson and colleagues
(2010) focused on exposure during the rst trimester, while
studies such as the one by Kelly and colleagues (2009) were
concerned with any exposure during the entire prenatal period. This variation raises concerns about aggregating the
results of these studies, particularly because timing is an
important determinant of the eects of prenatal alcohol
exposure (at least with respect to short-term adverse outcomes). While self-report is the best available method to
obtain information about alcohol consumption, there continue to be concerns that reporting might be aected by the
time interval between consumption and recall, the mode of
data collection (Ekholm et al., 2011; Kesmodel and Frydenberg, 2004), and social stigma against drinking during pregnancy. In regard to outcome assessment, the
neuropsychological outcomes we considered were measured
using dierent scales on children ranging in age from
6 months to 15 years. Meta-analysis is a benecial tool for
summarizing study results, but can be aected by methodological study heterogeneity, as well as assumptions when converting between measures of eect. These eects can be
particularly worrisome when including individuals of a wide
age range to examine such a complex area as neurodevelopment. An additional limitation of this review is the potential
bias introduced by exclusion of articles not in English and
exclusion of results from unpublished studies.
This review highlights the importance of abstaining from
binge drinking during pregnancy. It provides evidence that
there is no known safe amount of alcohol to consume while
pregnant. Eects of prenatal alcohol exposure on neurodevelopment might start at levels <1 drink a day during the
prenatal period, and drinking at this level may have substantial implications for public health at the population level.
Research is needed in this eld to develop better methods for
prenatal alcohol exposure assessment. Future studies using
such assessment methods and standardized quantications
of exposure will aid us in combining results from multiple
studies. Such studies also should control for important

potential confounders, such as SES and parental intelligence,

to dierentiate between the eects of prenatal alcohol exposure and other potential determinants of neuropsychological
outcomes. Studies that use more sophisticated measures of
neurodevelopment and assess domains omitted from previous studies, such as mental health and executive function,
are particularly important. As more studies on the neurodevelopment of children with mild, moderate, and binge prenatal alcohol exposure that incorporate these modications are
available, periodic systematic reviews and meta-analyses will
make important contributions to this eld.
ACKNOWLEDGMENTS
The authors thank Camille Smith, MS, EdS for her work
on outcome classications for this study and Gail Bang,
MLIS for her development and completion of the systematic
searches. This research was supported in part by an appointment to the Research Participation Program at the Centers
for Disease Control and Prevention (CDC) administered by
the Oak Ridge Institute for Science and Education through
an interagency agreement between the U.S. Department of
Energy and CDC.
REFERENCES
Alati R, Macleod J, Hickman M, Sayal K, May M, Smith G, Lawlor D
(2008) Intrauterine exposure to alcohol and tobacco use and childhood
IQ: ndings from a parental-ospring comparison within the Avon Longitudinal Study of Parents and Children. Pediatr Res 64:659666.
Alvik A, Torgesen AM, Aalen OO, Lindemann R (2011) Binge alcohol exposure once a week in early pregnancy predicts temperament and sleeping
problems in the infant. Early Hum Dev 87:827833.
Bailey BA, Sokol RJ (2008) Pregnancy and alcohol use: evidence and recommendations for prenatal care. Clin Obstet Gynecol 51:436444.
Bailey BN, Delaney-Black V, Covington CY, Ager J, Janisse J, Hannigan JH,
Sokol RJ (2004) Prenatal exposure to binge drinking and cognitive and
behavioral outcomes at age 7 years. Am J Obstet Gynecol 191:10371043.
Bay B, Kesmodel US (2011) Prenatal alcohol exposurea systematic review of
the eects on child motor function. Acta Obstet Gynecol Scand 90:210226.
Bay B, Stvring H, Wimberley T, Denny CH, Mortensen EL, Eriksen HLF,
Kesmodel US (2012) Low to moderate alcohol intake during pregnancy
and risk of psychomotor decits. Alcohol Clin Exp Res 36:807814.
Bayley N (1993) Bayley Scales of Infant Development. 2nd ed. Psychological
Corporation, San Antonio, TX.
Bertrand J, Floyd RL, Webber MK, OConnor M, Riley EP, Johnson KA,
Cohen DE, National Task Force on FAS/FAE (2004) Fetal Alcohol Syndrome: Guidelines for Referral and Diagnosis. CDC, Atlanta, GA.
Bolumar F (1992) Chapter 7 Methodscomparison between centres. Int J
Epidemiol 21:S40S44.
Borenstein M, Hedges L, Higgins J, Rothstein H (2009) Introduction to
Meta-Analysis. John Wiley & Sons, Ltd., Chichester, UK.
Borenstein M, Hedges LV, Higgins J, Rothstein HR (2005) Comprehensive
Meta-Analysis Version 2. Biostat Inc., Englewood, NJ.
Bradley RH, Corwyn RF (2002) Socioeconomic status and child development. Annu Rev Psychol 53:371399.
Brown CW, Olson HC, Croninger RG (2010) Maternal alcohol consumption
during pregnancy and infant social, mental, and motor development.
J Early Interv 32:110126.
Centers for Disease Control and Prevention (2009) Alcohol Use Among
Pregnant and Nonpregnant Women of Childbearing AgeUnited States,
19912005. MMWR CDC, Atlanta.

PRENATAL ALCOHOL AND CHILD OUTCOMES

Coles CD, Kable JA, Drews-Botsch C, Falek A (2000) Early identication of

risk for eects of prenatal alcohol exposure. J Stud Alcohol 61:607616.
Drews CD, Coles CD, Floyd RL, Falek A (2003) Prevalence of prenatal
drinking assessed at an urban public hospital and a suburban private
hospital. J Matern Fetal Neonatal Med 13:8593.
Egger M, Smith GD, Schneider M, Minder C (1997) Bias in meta-analysis
detected by a simple, graphical test. BMJ 315:629634.
Ekholm O, Strandberg-Larsen K, Gronbak M (2011) Inuence of the recall
period on a beverage-specic weekly drinking measure for alcohol intake.
Eur J Clin Nutr 65:520525.
Forrest F, Florey CD, Taylor D, McPherson F, Young JA (1991) Reported
social alcohol consumption during pregnancy and infants development at
18 months. BMJ 303:2226.
Fraser SL, Muckle G, Abdous BB, Jacobson JL, Jacobson SW (2012) Eects
of binge drinking on infant growth and development in an Inuit sample.
Alcohol 46:277283.
Goldschmidt L, Richardson GA, Cornelius MD, Day NL (2004) Prenatal
marijuana and alcohol exposure and academic achievement at age 10.
Neurotoxicol Teratol 26:521532.
Goldschmidt L, Richardson GA, Stofer DS, Geva D, Day NL (1996) Prenatal alcohol exposure and academic achievement at age six: a nonlinear t.
Alcohol Clin Exp Res 20:763770.
Henderson J, Kesmodel U, Gray R (2007) Systematic review of the fetal
eects of prenatal binge-drinking. J Epidemiol Community Health
61:10691073.
Jacobson JL, Dodge NC, Burden MJ, Klorman R, Jacobson SW (2011)
Number processing in adolescents with prenatal alcohol exposure and
ADHD: dierences in the neurobehavioral phenotype. Alcohol Clin Exp
Res 35:431442.
Jacobson JL, Jacobson SW, Sokol RJ, Martier SS, Ager JW, Kaplanestrin
MG (1993a) Teratogenic eects of alcohol on infant development. Alcohol
Clin Exp Res 17:174183.
Jacobson SW, Jacobson JL, Sokol RJ, Martier SS, Ager JW (1993b) Prenatal alcohol exposure and infant information-processing ability. Child Dev
64:17061721.
Kaplan-Estrin M, Jacobson SW, Jacobson JL (1999) Neurobehavioral
eects of prenatal alcohol exposure at 26 months. Neurotoxicol Teratol
21:503511.
Kelly Y, Sacker A, Gray R, Kelly J, Wolke D, Quigley MA (2009) Light
drinking in pregnancy, a risk for behavioural problems and cognitive decits at 3 years of age? Int J Epidemiol 38:129140.
Kelly YJ, Sacker A, Gray R, Kelly J, Wolke D, Head J, Quigley MA (2012)
Light drinking during pregnancy: still no increased risk for socioemotional
diculties or cognitive decits at 5 years of age? J Epidemiol Community
Health 66:4148.
Kesmodel U, Frydenberg M (2004) Binge drinking during pregnancyis it
possible to obtain valid information on a weekly basis? Am J Epidemiol
159:803808.
Kesmodel US, Bertrand J, Stvring H, Skarpness B, Denny CH, Mortensen
EL, The Lifestyle During Pregnancy Study Group (2012) The eect of different alcohol drinking patterns in early to mid pregnancy on the childs
intelligence, attention, and executive function. BJOG 119:11801190.
Larkby CA, Goldschmidt L, Hanusa BH, Day NL (2011) Prenatal alcohol
exposure is associated with conduct disorder in adolescence: ndings from
a birth cohort. J Am Acad Child Adolesc Psychiatry 50:262271.
Larroque B, Kaminski M, Dehaene P, Subtil D, Querleu D (2000) Prenatal
alcohol exposure and signs of minor neurological dysfunction at preschool
age. Dev Med Child Neurol 42:508514.
Lemola S, Stadlmayr W, Grob A (2009) Infant irritability: the impact of fetal
alcohol exposure, maternal depressive symptoms, and low emotional support from the husband. Infant Ment Health J 30:5781.
Lewis SJ, Zuccolo L, Davey Smith G, Macleod J, Rodriguez S, Draper ES,
Barrow M, Alati R, Sayal K, Ring S, Golding J, Gray R (2012) Fetal alcohol exposure and IQ at age 8: evidence from a population-based birthcohort study. PLoS ONE 7:e49407.
National Institute on Alcohol Abuse and Alcoholism (2004) NIAAA council
approves denition of binge drinking. NIAAA Newsletter 3(Winter):3.

225

Nulman I, Rovet J, Kennedy D, Wasson C, Gladstone J, Fried S, Koren G

(2004) Binge alcohol consumption by non-alcohol-dependent women
during pregnancy aects child behaviour, but not general intellectual
functioning; a prospective controlled study. Arch Womens Ment Health
7:173181.
OCallaghan FV, OCallaghan M, Najman JM, Williams GM, Bor W (2007)
Prenatal alcohol exposure and attention, learning and intellectual ability at
14 years: a prospective longitudinal study. Early Hum Dev 83:115123.
OLeary CM, Bower C (2012) Guidelines for pregnancy: whats an acceptable risk, and how is the evidence (nally) shaping up? Drug Alcohol Rev
31:170183.
OLeary CM, Nassar N, Zubrick SR, Kurinczuk JJ, Stanley F, Bower C
(2010) Evidence of a complex association between dose, pattern and timing
of prenatal alcohol exposure and child behaviour problems. Addiction
105:7486.
Olsen J (1994) Eects of moderate alcohol-consumption during pregnancy
on child-development at 18 and 42 months. Alcohol Clin Exp Res
18:11091113.
Parry GJ, Ogston SA (1992) EUROMAC. A European concerted action:
maternal alcohol consumption and its relation to the outcome of pregnancy and child development at 18 months. Resultschild development at
age 18 months. Int J Epidemiol 21:S72S78.
Richardson GA, Day NL, Goldschmidt L (1995) Prenatal alcohol, marijuana, and tobacco useinfant mental and motor development. Neurotoxicol Teratol 17:479487.
Robinson M, Oddy WH, McLean NJ, Jacoby P, Pennell CE, De Klerk NH,
Zubrick SR, Stanley FJ, Newnham JP (2010) Low-moderate prenatal
alcohol exposure and risk to child behavioural development: a prospective
cohort study. BJOG 117:11391150.
Rodriguez A, Olsen J, Kotimaa A, Kaakinen M, Moilanen I, Henriksen T,
Linnet K, Miettunen J, Obel C, Taanila A, Ebeling H, Jarvelin M (2009) Is
prenatal alcohol exposure related to inattention and hyperactivity
symptoms in children? Disentangling the eects of social adversity. J Child
Psychol Psychiatry 50:10731083.
Sayal K, Heron J, Golding J, Emond A (2007) Prenatal alcohol exposure
and gender dierences in childhood mental health problems: a longitudinal
population-based study. Pediatrics 119:e426e434.
Sood B, Delaney-Black V, Covington C, Nordstrom-Klee B, Ager J, Templin T, Janisse J, Martier S, Sokol RJ (2001) Prenatal alcohol exposure
and childhood behavior at age 6 to 7 years: I. Doseresponse eect.
Pediatrics 108:e34.
Streissguth AP, Barr HM, Olson HC, Sampson PD, Bookstein FL, Burgess
DM (1994a) Drinking during pregnancy decreases word attack and arithmetic scores on standardized testsadolescent data from a populationbased prospective-study. Alcohol Clin Exp Res 18:248254.
Streissguth AP, Barr HM, Sampson PD, Bookstein FL, Darby BL (1989)
Neurobehavioral eects of prenatal alcohol: Part I. Research strategy.
Neurotoxicol Teratol 11:461476.
Streissguth AP, Martin DC, Martin JC, Barr HM (1981) The Seattle longitudinal prospective study on alcohol and pregnancy. Neurobehav Toxicol
Teratol 3:223233.
Streissguth AP, Sampson PD, Olson HC, Bookstein FL, Barr HM, Scott M,
Feldman J, Mirsky AF (1994b) Maternal drinking during pregnancy
attention and short-term-memory in 14-year-old ospringa longitudinal
prospective-study. Alcohol Clin Exp Res 18:202218.
Stroup D, Morton S, Olkin I, Williamson G, Rennie D, Moher D, Becker B,
Sipe T (2000) Meta-analysis of observational studies in epidemiology: a
proposal for reporting. JAMA 283:20082012.
Sulaiman ND, Florey CD, Taylori DJ, Ogston SA (1988) Alcohol consumption in Dundee primigravidas and its eects on outcome of pregnancy.
Br Med J 296:15001503.
Testa M, Quigley BM, Eiden RD (2003) The eects of prenatal alcohol exposure on infant mental development: a meta-analytical review. Alcohol
Alcohol 38:295304.
U.S. Department of Health and Human Services (2000) Health People 2010:
Objective for Improving Health. 9. Family Planning. 2nd ed. U.S. Government Printing Oce, Washington, DC.

FLAK ET AL.

226

U.S. Department of Health and Human Services (2005) U.S. Surgeon General Releases Advisory on Alcohol Use in Pregnancy. US Department of
Health and Human Services, Washington, DC. Available at: http://www.
surgeongeneral.gov/pressreleases/sg02222005.html. Accessed June 1, 2011.
Wechsler H, Dowdall GW, Davenport A, Rimm EB (1995) A gender-specic
measure of binge drinking among college students. Am J Public Health
85:982985.
Wells G, Shea B, OConnell D, Peterson J, Welch V, Losos M, Tugwell P
(2013) The Newcastle-Ottawa Scale (NOS) for assessing the quality of
nonrandomised studies in meta-analyses. Available at: http://www.ohri.ca/
programs/clinical_epidemiology/oxford.htm. Accessed June 25, 2013.
Willford JA, Leech SL, Day NL (2006) Moderate prenatal alcohol exposure
and cognitive status of children at age 10. Alcohol Clin Exp Res 30:1051
1059.
Willford JA, Richardson GA, Leech SL, Day NL (2004) Verbal and visuospatial learning and memory function in children with moderate prenatal
alcohol exposure. Alcohol Clin Exp Res 28:497507.

SUPPORTING INFORMATION
Additional Supporting Information may be found in the
online version of this article:

Table S1. Detailed Systematic Search Criteria.

Fig. S1. Meta-analysis results for the associations
between: (1) moderate prenatal alcohol exposure and child
attention, (2) mild prenatal alcohol exposure and child behavior, (3) mild prenatal alcohol exposure and child cognition,
(4) moderate prenatal alcohol exposure and child cognition,
(5) mild prenatal alcohol exposure and child language and
verbal development, (6) moderate prenatal alcohol exposure
and child language and verbal development, (7) mild prenatal alcohol exposure and child visual and motor development, (8) moderate prenatal alcohol exposure and child
visual and motor development.
Fig. S2. Meta-analysis results for the associations between
binge prenatal alcohol consumption and: (1) academic math
performance, (2) academic reading performance, (3) attention, (4) behavior, (5) executive function, (6) language and
verbal development, (7) visual and motor development.