Birth, Cognitive, and Behavioral Effects of Intrauterine Cannabis Exposure in Infants and Children: A Systematic Review and Meta-Analysis

Abstract

Background and Aims:

Δ9-tetrahydrocannabinol (THC), the principal psychoactive component of cannabis, has been implicated in affecting fetal neurodevelopment by readily crossing the placenta. However, little is known regarding the long-term effects of intrauterine cannabis exposure. This systematic review and meta-analysis synthesized prospective and cross-sectional human studies to measure the effects of intrauterine cannabis exposure on birth, behavioral, psychological, and cognitive outcomes in infancy until early childhood.

Methods:

Reporting according to the PRISMA statement, cross-sectional and prospective studies published from database inception until June 2023, investigating developmental outcomes of infants, toddlers, and young children with intrauterine cannabis exposure were considered. All articles were obtained from PubMed or PsycINFO databases.

Results:

The literature search resulted in 932 studies, in which 57 articles met eligibility criteria. The meta-analysis revealed that intrauterine cannabis exposure increases the risk of preterm delivery (odds ratio [OR]=1.68, 95% confidence interval [CI]=1.05– 2.71, p=.03), low birth weight (OR=2.60, CI=1.71–3.94, p<.001), and requirement for neonatal-intensive care unit (NICU) admission (OR=2.51, CI=1.46 – 4.31; p<.001). Our qualitative synthesis suggests that intrauterine cannabis exposure may be associated with poorer attention and externalizing problems in infancy and early childhood. We found no evidence for impairments in other cognitive domains or internalizing behaviors.

Conclusions:

Prenatal cannabis use appears to be associated with lower birth weight, preterm birth, and neonatal-intensive care unit admission in newborns, but there is little evidence that prenatal cannabis exposure adversely impacts behavioral or cognitive outcomes in early childhood, with the exception of attention and externalizing problems.

Introduction

The acceptability of cannabis use has dramatically changed over the previous two decades, accompanied by expanding legalization, decriminalization, and medicalization (1). Subsequently, cannabis use has increased, while perceived risk of harms has decreased (2). Perinatal women in Canada and the United States are not an exception, where prevalence of cannabis use among pregnant women has almost doubled within the past two decades, with approximately 10% of pregnancies being associated with maternal cannabis exposure (3, 4). A concerning trend involves pregnant women using cannabis to treat pregnancy-related side effects, including nausea (5). In one survey of women reporting cannabis use during pregnancy, the majority of respondents (77%) reported use to treat nausea, while over half of respondents reported use to treat either poor appetite, chronic pain, insomnia, anxiety, depression, or fatigue(6). While clinical evidence supports the use of medical cannabis for treatment of certain physical health disorders (7), there is considerable debate regarding its use among pregnant women (8). Specifically, because Δ9-tetrahydrocannabinol (THC)—the main psychoactive constituent of cannabis—readily crosses the placenta and enters the fetus (9), cannabis may significantly alter child development by impacting fetal endocannabinoid systems. Moreover, the early developmental years represent a distinctive phase where trajectories are being established and children exposed to substances are most likely to benefit from interventions (10). Consequently, there is a pressing need for research that broadens current knowledge on the early developmental outcomes related to prenatal cannabis exposure. We therefore aimed to synthesize the available research on developmental, physical, psychological, and behavioral consequences of prenatal cannabis exposure from birth through early childhood.

Intrauterine Exposure to Cannabis: Fetal Effects

Endogenous cannabinoids (e.g. anandamide) are naturally occurring arachidonic acid metabolites that regulate numerous processes, including movement, memory, sleep, appetite, body temperature, and pain perception (11). Two cannabinoid receptors have been identified, where cannabinoid receptors 1 (CB1) are predominantly located in the central nervous system (CNS), while cannabinoid receptors 2 (CB2) are primarily confined to immune cells (12). THC mimics endocannabinoid actions, exerting its effects via the CB1 receptors, which are also expressed early in the developing fetal brain. CB1 receptors are identifiable during the first trimester, and are localized in white matter and cell proliferative regions, including the hippocampus and prefrontal cortex (13). In the developing fetus, these receptors are involved in critical neurodevelopmental events, including neural progenitor proliferation, migration, and synaptogenesis. In preclinical models, downregulation of CB1 receptors during fetal development reduces neural progenitor proliferation, synapse formation, and fetal cortical and hippocampal connectivity, which coincides with impairments in memory and learning (14). Other evidence employing knockout models has indicated that mice lacking CB1 receptors are susceptible to numerous behavioral problems, including depression and anxiety (15).

Intrauterine cannabis exposure may also adversely impact the maturation of other neurotransmitter systems (16). Like endocannabinoids, dopaminergic neurons are expressed very early in neurodevelopment, and within the adult brain, dopamine is critical in regulating motor function, cognition, and emotional processes (17). Prenatal cannabis exposure impacts tyrosine hydroxylase activity, the rate-limiting enzyme in dopamine synthesis, which then alters the densities of dopaminergic neurons (18). Alterations in dopaminergic receptor densities is highly implicated in psychiatric disorders, including schizophrenia (19) and substance use disorders (20). Intrauterine cannabis exposure additionally impacts the expression and activity of opioid receptors, which play an influential role in motivation, mood, and pain perception (21). In adult rats, prenatal THC exposure decreased the expression of proenkephalin, an endogenous opioid polypeptide, in the caudate-putamen (22). These findings suggest that cannabis exposure during fetal development may disturb components of the endogenous opioid system which then persist into adulthood.

Glutamatergic systems, which is critical for neural proliferation, migration, and survival during neurodevelopment (16), also relies upon the endocannabinoid system (23). In rats, prenatal THC exposure downregulates certain glutamate receptors with changes emerging in fetal development, and persisting until adulthood (24). Additionally, postnatal glutamate release is lower in the frontal cortex and hippocampus of rats with intrauterine exposure to CB1 receptor agonists, where these glutamatergic alterations correlate with severity of cognitive impairments and stress reactivity in adulthood (25). Taken together, the evidence suggests that intrauterine cannabis exposure may adversely impact multiple neurotransmitter systems and fetal development, with effects persisting beyond infancy.

Study Aims

There is strong evidence from preclinical studies that intrauterine exposure to cannabis considerably impacts neurodevelopment. However, the effects of prenatal cannabis use in humans remains unclear, with some studies determining adverse effects (4, 26), while others detecting no impairments (27, 28). Moreover, information on long-term outcomes in children prenatally exposed to cannabis is sparse, with few longitudinal cohorts extending beyond infancy. Therefore, the aim of this systematic review and meta-analysis is to fill a significant gap in our scientific understanding by investigating the potential physical, behavioral, cognitive, and psychiatric consequences of prenatal cannabis exposure on offspring until early childhood (i.e., until the age of 8 (30)). Synthesizing the research on long-term consequences of intrauterine cannabis exposure up until this age group would provide further guidance to policymakers, clinicians, and patients, who have reported limited knowledge regarding cannabis use in pregnancy (29).

Previous systematic reviews examining offspring outcomes who were prenatally exposed to cannabis have focused on specific domains, including cognition (30, 31), neonatal development (32), and psychiatric outcomes (33). However, there has been little focus on qualitatively and quantitatively synthesizing the developmental manifestations of prenatal cannabis exposure across the duration of the early childhood period. A recent article by Grant et al (34) qualitatively examined the effects of prenatal cannabis exposure on neural, cognitive, and behavioral development from infancy through adolescence. However, the present systematic review and meta-analysis complements and extends Grant et al’s scoping review (34) in several important ways. First, we conducted a comprehensive and systematic search to identify all relevant longitudinal and cross-sectional human studies examining intrauterine cannabis exposure and physical, behavioral, psychological, and cognitive outcomes of infants and children, to understand cannabis-induced influences on early childhood development. Second, we conducted a quality assessment to provide a critical appraisal of the included studies. Third, we conducted a quantitative assessment for eligible outcomes to provide a numeric synthesis of the effects of intrauterine cannabis exposure on childhood development. Finally, our review captured studies published since the review by Grant et al (34), which covered articles published until 2019.

Methods

Search Strategy

The review was conducted in accordance with the PRISMA statement guidelines (Figure 1) (35). Original peer-reviewed research articles were searched for using the PubMed and PsycINFO databases. Articles available online in English from inception through June 2023, were considered. Search terms (found in the title or abstract) utilized to obtain relevant articles were: “cannabis” OR “tetrahydrocannabinol” OR “cannabidiol” OR “marijuana” AND “fetus”, OR “newborn”, OR “neonate”, OR “infant”, OR “toddler”, OR “child*,” AND “birth outcome”, OR “pregnancy outcome”, OR “birth weight”, OR “cognition”, OR “motor development”, OR “neurodevelopment” OR “behav*” OR “depression” OR “anxiety”. The protocol of this review was not prospectively registered in the International Prospective Register of Systematic Reviews (PROSPERO). Title and abstract screening in addition to full-text review were conducted by MS.

Figure 1:

PRISMA Flow Diagram

Inclusion and Exclusion Criteria

We performed a systematic literature review according to the PICOS strategy (36). Studies were included in this review if they satisfied the following criteria—population (P): studies recruiting infants or children between the ages of 0 days old to 8 years of age (i.e., until early childhood); intervention (I): studies recruiting infants or children between the ages of 0 days old to 8 years of age with prenatal exposure to cannabis; comparison (C): studies including a control group of children between the ages of 0 days to 8 year of age that were not prenatally exposed to cannabis; outcomes (O): studies utilizing a validated or objective measurement to evaluate either birth outcomes (e.g., hospital records), cognition (e.g., scores from the Wechsler Preschool and Primary Scale of Intelligence), psychological development (e.g., scores from the Revised Child Anxiety and Depression Scale), or behavioral development (e.g., scores from the Child Behavior Checklist [CBCL]); and study design (S): studies employing either a cross-sectional or prospective design.

Accordingly, exclusion criteria for this review are as follows: (1) studies examining fetuses or children who are older than 8-years of age (i.e., beyond the age of early childhood (37)); (2) animal studies; and (3) literature reviews, meta-analyses, dissertations, commentaries, conference presentations, abstracts, and case studies.

Quality Assessment

The Newcastle-Ottawa Scale (NOS) was administered to evaluate the quality of included studies by two researchers (MS and CW) (38). The NOS evaluates non-randomized studies on a scale from 0–9 based on three main criteria: (1) selection of cases and controls (e.g., adequate definition, representativeness, and selection of cases); (2) comparability of cases and controls (e.g., adequate controlling or adjustment for confounding variables); and (3) ascertainment of exposure (e.g., adequate assurance that the cases were exposed to the variable of interest). Studies receiving a score ≥7 on the NOS are considered high quality, while studies receiving a score ≤4 exhibit low methodological quality (38). Disagreements between assessors were resolved with consensus achieved through discussion.

Data Extraction

Two researchers (MS and CW) extracted study characteristics, including name of first author, year of publication, study design, location of study sample, whether use of cannabis was self-reported or biologically measured, sample size, variables of interest related to infant or childhood outcomes, and study findings. Study authors were contacted for missing values to achieve maximal results for the meta-analysis. For studies where data could not be extracted, a qualitative synthesis of findings was undertaken.

Data Analysis

We employed Revman 5.3 to conduct a random-effects meta-analysis to investigate the relationship between intrauterine cannabis exposure and subsequent infant and childhood outcomes. In this study, a random-effects model was utilized to estimate pooled odds ratios (ORs), in order to take into account for the heterogeneity observed among the included studies. Outcomes with at least three studies containing extractable data were pooled together for a random-effects meta-analysis. For dichotomous outcomes such as low birth weight, ORs and 95% confidence intervals (CIs) were extracted from the available data. Heterogeneity was assessed using the I² statistic. Respective cut-off-values of 25%, 50%, and 75% denoted whether heterogeneity was low, moderate, or high [33].

Results

Characteristics of Included Studies and Quality Assessment

The initial literature search retrieved 932 studies. After removing 133 duplicates, 799 records were retained for title and abstract screening. There were 158 full-text articles reviewed for eligibility and 57 papers (12,901,376 participants; n=102,835 cannabis-exposed) met inclusion criteria (Figure 1). Studies were published between 1984 and 2023 and were conducted in a broad range of countries, including the USA, Australia, Canada, the Netherlands, Norway, France, and Italy. The number of cannabis-exposed children per paper ranged between 14 to 29,101. Table 1 presents an overview of each of the included studies. Methods to ascertain prenatal cannabis use included self-report questionnaires, salivary samples, placental or meconium analyses, urine testing, and clinical interviews. Thirty-six studies examined adverse birth outcomes (e.g., stillbirth, small for gestational age [SGA; <10^th percentile body weight]), eight studies examined infant or toddler cognition or neurodevelopment, four studies examined infant or toddler behavioral or psychiatric outcomes, eleven studies examined early childhood cognition or neurodevelopment, and four studies examined early childhood psychiatric or behavioral outcomes.

Table 1:

Characteristics of Included Studies

Author, Year	Sample Characteristics	Study Design	Cannabis Use Measure	Was Other Substance Use Controlled?	Outcome Variable	Outcome Measures	Effect Sizes+	Results	Newcastle Ottawa Scale* (NOS) Score
Birth Outcomes (N=12,697,935; n= 99,728 CE ; NOS Summary Score = 8.45)
Bada et al., 2005(48)	N=11,811 maternal-infant dyads in US (n=812 CE)	Cross-Sectional	Self-report	Yes	Birthweight, preterm birth	Hospital records	Low birth weight OR = .1.21 Preterm birth OR = .9	Prenatal cannabis use was not associated with increased likelihood of low birthweight or preterm birth.	9
Baer et al., 2019(92)	N-2,890,555 live births in California (n= 15,739 CE)	Cross-sectional	ICD-9 interview	Yes	Preterm birth	Hospital records	Preterm birth RR = 1.21	Mothers using cannabis were not significantly more likely to give birth preterm than mothers without any substance use.	9
Bandoli et al., 2021a(93)	N=484,905 live births in California with sudden unexpected infant death (n=3,993 CE)	Cross-sectional	ICD-9 interview	Yes	Sudden unexpected infant death	Hospital records	Sudden unexpected infant death OR = 2.7+	Mothers with a current CUD increased the odds ratio of sudden unexpected infant death by 2.7 in comparison to mothers without CUD.	8
Bandoli et al., 2021b(94)	N=3,067,069 live births in Californ,ia (n=29,101 CE)	Cross-sectional	ICD-9 Interview	Yes	Preterm birth, birth weight, NICU admission, birth defects	Hospital records	Preterm birth aRR =2.3+ Low birth weight aRR = 1.9+ NICU admission aRR = 1.1 Birth defect aRR = 1.2+	CUD increased the relative risk of preterm birth by 2.3 relative to no cannabis-related diagnosis. Co-morbid substance use increased the risk further. Current CUD increased the relative risk of low birth weight and birth defects by 1.9 and 1.2, respectively.	8
Brown et al., 2016(95)	N=344 live births in Australia(n=71 CE)	Cross-sectional	Self-report	Yes	Birthweight, preterm birth	Hospital records	Low birth weight OR = 6.5+ Preterm birth OR = 1.9+	Compared against mothers not using cannabis or cigarettes, mothers using cannabis were significantly more likely to give birth to preterm babies or babies with a low birth weight.	8
Burns et al., 2006(96)	N=416,834 live births in Australia (n=2,171 CE)	Cross-sectional	ICD-10 Interview	Yes	NICU admission, premature, birth weight	Hospital records	NICU admission OR = 2.0+ Preterm birth OR = 2.2+ Low birth weight OR = 2.0+	Mothers with CUD were significantly more likely to give birth to preterm babies, babies with a low birth weight, and in need of NICU admission in comparison to mothers without a substance use disorder.	8
Chabarria et al., 2016(39)	N=12,069 live births in US (n=106 CE)	Cross-sectional	Self-report	No	Preterm birth, birth weight	Hospital records	Preterm birth OR =2.6+ Low birthweight OR =2.8+	Cannabis use alone was not associated with significant adverse neonatal outcomes. However, co-use of tobacco was associated with increased risk of preterm birth and low birth weight	7
Coleman-Cowger et al., 2018(40)	N=500 live births in USA (n=106 CE)	Cross-sectional	Urine testing	Yes	Birthweight, preterm birth, NICU admission, stillbirths, birth defects	Hospital records	Low birth weight OR = 2.4 Preterm birth OR = 2.2 Stillbirth OR =12.1+ Birth defects OR = 1.2 NICU admission OR = 1.5	Co-use of tobacco and cannabis among mothers was associated with a significantly greater risk of birth defects in comparison to mothers without any prenatal substance use or mothers using only cannabis or tobacco. Cannabis use only was associated with a 12x greater odds of stillbirth.	9
Conner et al., 2015(28)	N=8,138 live births in USA (n=680 CE)	Cross-sectional	Self-report (single-item question), urine testing	Yes	Birthweight, NICU admission,	Hospital records	Low birthweight aOR = 1.3 NICU admission aOR = 1.6	After adjusting for relevant covariates, including other substance use and demographic variables, no significant differences in neonatal morbidities was observed between the cannabis-using group and non-cannabis using mothers.	9
Corsi et al., 2019(4)	N=661,617 live births in Ontario (n=15,066 CE)	Cross-sectional	Self-report (single-item question)	Yes	Preterm birth, SGA, stillbirths	Hospital records	Preterm birth RR = 1.96+ Stillbirths = 1.60+ SGA = 2.60+	Self-reported cannabis use during pregnancy was associated with an increased risk of preterm birth, stillbirth, and SGA.	8
Crume et al., 2018(97)	N=3,207 live births in Colorado (n=183 CE)	Cross-sectional	Self-report (single-item question)	Yes	SGA, preterm birth, NICU admission, birth weight	Hospital records	SGA OR = 1.7+ Preterm birth OR = 1.3 NICU admission = 1.0 Low birth weight = 1.8+	Cannabis use was associated with only an increased risk of SGA after adjusting for significant covariates.	8
Dotters-Katz et al., 2017(50)	N=1,867 live births in US (2yrs at follow-up n=135 CE)	Prospective	Urine testing	No	SGA	Hospital Records	SGA OR = .83	Prenatal cannabis use was not associated with SGA.	9
Gabrhelik et al., 2020(98)	N=10,101 live births in Norway (n=272 CE)	Cross-sectional	Self-report (frequency of use)	Yes	Birth weight	Hospital records	Birth weight B = −.334+	Self-reported cannabis use in at least two trimesters of pregnancy was associated with an increased risk of low birth weight.	8
Gray et al., 2010(41)	N=86 live births in US (n=38 CE)	Cross-sectional	Self-report, oral fluid samples	Yes	Birth weight	Hospital records	Low birth weight B = −.299+	Prenatal cannabis use was a significant risk factor for low birth weight among newborns.	9
Greiner et al., 2020(99)	N=1,217 live births to mothers with hypertension in US (n=117 CE)	Cross-sectional	Self-report, urine testing	Yes	Birthweight, preterm birth, NICU admission, stillbirths, birth defects	Hospital records	Low birth weight OR = 1.36 Preterm birth OR = .93 NICU admission OR = 1.30 Stillbirth OR = .90 Birth defect OR =.90	After controlling for confounds, cannabis use was not associated with any adverse neonatal outcomes.	9
Haight et al., 2021(100)	N=5,548 live births in US (n=812 CE)	Cross-sectional	Self-report (frequency of use)	Yes	Birthweight, SGA, preterm birth	Hospital records	Low Birth weight aOR =9.7+ SGA aOR = 10.2 Preterm birth aOR = 6.1	>Weekly cannabis use was associated with low birthweight. No effects of cannabis on SGA or preterm birth were found.	8
Hayatbakhsh et al., 2012(101)	N=28,874 live births in Australia (n=751 CE)	Cross-sectional	Self-report	Yes	Birthweight, preterm birth, birth length, NICU admission	Hospital records	Low Birth weight aOR = 1.7+ Preterm birth aOR = 1.5+ NICU Admission aOR = 2.2+	Newborns of women who smoked cannabis during pregnancy were more likely to be of lower birth weight, born preterm, and require NICU admission.	8
Howard et al., 2019(42)	N=2,173 live births in US (n=565 CE)	Cross-sectional	Urine testing	Yes	Birthweight	Hospital records	Unextractable Data	Cannabis use was associated with lower birth weight after controlling for tobacco use. Concomitant tobacco use did not impact birth weight.	9
Kharbanda et al., 2020(102)	N =3,435 live births in US (n=283 CE)	Cross-sectional	Urine testing	Yes	Preterm birth, SGA, birth defects	Hospital records	Preterm birth RR = 1.06 SGA RR = 1.69+ Birth defect RR = .58	Intrauterine cannabis exposure was associated with increased risk of SGA only.	9
Koto et al., 2022(103)	N=143,405 live births in Nova Scotia (n=3,144 CE)	Cross-sectional	Self-report	Yes	NICU admission, SGA,	Hospital records	NICU admission RR = 1.13+ SGA RR =1.52+	Infants with cannabis-using mothers were significantly more likely to require NICU admission and be of low birth weight.	8
Lee et al., 2020(43)	N=343 live births in California (n=45 CE)	Cross-sectional	Urine testing	Yes	SGA, NICU admission, preterm birth	Hospital records	Unextractable Data	Infants with cannabis-using mothers were not more likely to require NICU admission, be SGA, or be born prematurely than infants with non-drug-using mothers.	9
Mark et al., 2016(104)	N=170 live births in US (n=116 CE)	Cross-sectional	Self-report, urine testing	Yes	NICU admission, birthweight	Hospital records	Low birth weight OR = .87 NICU admission data is unextractable	After controlling for confounds, infants with intrauterine cannabis exposure were not more likely to require NICU admission or have a low birth weight in comparison to babies without substance exposure.	9
Michalski et al., 2020(44)	N=1,778 live births in Canada(n= 216 CE)	Cross-sectional	Self-report (single-item)	Yes	Birth weight, preterm birth, SGA	Hospital records	Unextractable Data	Intrauterine cannabis exposure was associated with increased risk for lower birth weight and SGA, but not preterm birth.	8
Nawa et al.,(45)	N= 8,261 mother-newborn pairs in US (n=447)	Cross-sectional	Self-report	Yes	SGA, preterm birth,	Hospital Records	Preterm birth RR = 1.34+ SGA B = −.83+	Cannabis use and cigarette smoking were independently associated with a decrease in gestational age. Simultaneous cannabis use and cigarette smoking was associated with higher risk of preterm birth.	8
Nguyen et al., 2022(105)	N=32,583 live births in California (n=1,597 CE)	Cross-sectional	Self-report (frequency of use)	Yes	SGA, bbirth weight, irthweight, preterm birth	Hospital records	Low birth weight OR = 1.89+ Preterm birth OR = 1.16 SGA OR =1.35+	Weekly≤ cannabis use was associated with significantly greater odds of low birth weight and SGA. Comorbid cigarette use increased the odds further.	8
Ostrea et al., 1997(106)	N=2,964 mother-child dyads in US (11-month follow-up; n=338)	Prospective	Meconium analysis	No	Mortality, birthweight, birth length	Hospital records	Mortality OR = .74 Other data is unextractable	There was no significant difference in risk of mortality in cannabis-exposed infants and controls. Cannabis-exposed infants were significantly more likely to weigh less and have a smaller length than controls.	7
Prunet et al., 2017(46)	N=27,211 live births in France (n=156 CE)	Cross-sectional	Self-report (single item)	No	Preterm birth	Hospital records	Preterm birth OR = 1.9+	Intrauterine cannabis exposure was a significant risk factor for preterm birth.	6
Roca et al., 2020(107)	N=372 live births in Italy (n=24 CE)	Cross-sectional	Meconium analysis	No	Preterm birth, SGA,	Hospital records	Unextractable Data	Intrauterine cannabis exposure was significantly associated with preterm birth, and low birth weight.	7
Rodriquez et al., 2019(108)	N=1,206 livebirths in Colorado (n=211 CE)	Cross-sectional	Self-report (frequency of use), urine testing	No	Preterm birth, stillbirth, SGA, NICU admission	Hospital records	Unextractable Data	Intrauterine cannabis exposure was significantly associated with low birth weight only.	8
Sasso et al., 2021(109)	N=343 live births in California (n=151 CE)	Cross-sectional	Self-report (frequency of use), urine testing	No	Preterm birth, NICU admission, SGA	Hospital records	NICU admission OR= 2.12 Preterm delivery OR = .93 SGA OR = 4.24+	In comparison to non-exposed infants, those with prenatal cannabis exposure had an increased risk of SGA, but no elevated risk for other adverse fetal outcomes.	7
Saurel-Cubizolles et al., 2014(110)	N=13,545 live births in France (n=156 CE)	Cross-sectional	Self-report	Yes	SGA, preterm birth	Hospital records	SGA OR = 1.98+ Preterm birth OR = 2.68+	In comparison to no substance use, prenatal cannabis use increased the risk of preterm birth and SGA	8
Shi et al., 2020(47)	N=4,830,239 live births in California (n=20,237 CE)	Prospective	ICD-9 (CUD only)	Yes	SGA, birth weight, preterm birth, mortality by year 1,	Hospital records	SGA OR = 1.13+ Preterm birth OR = 1.06+ Mortality by year 1 OR = 1.35+	In comparison to no substance use, prenatal cannabis use increased the risk of low birth weight, death within a year, and SGA.	9
Shiono et al., 1995(111)	N=7,470 live births in US (n=75)	Cross-sectional	Self-report (single-item), blood test	No	Preterm birth, birth weight	Hospital records	Preterm birth OR =1.1 Low birth weight OR = 1.1	After controlling for confounds, intrauterine cannabis exposure was not a risk factor for low birth weight or premature birth.	8
Straub et al., 2021(112)	N=5,343 live births in US (n=1,268 CE)	Cross-sectional	Urine testing	Yes	Birth weight, SGA	Hospital records	Low birth weight OR = 1.42+ SGA OR = 1.51+	Prenatal cannabis use dose-dependently increased the risk of low birth weight and SGA.	9
Van Gelder et al., 2010 (113)	N=5,871 live births in US (n=185)	Cross-sectional	Self-report	Yes	Birth weight, preterm birth	Hospital records	Low birth weight OR =.8 Preterm birth OR = 1.8	After adjustment for confounding variables, cannabis use was not associated with low birth weight or preterm birth.	9
Warshak et al., 2015 (27)	N=6,481 live births in US (n=361 CE)	Cross-sectional	Urine testing	Yes	Preterm birth, NICU admission, stillbirth, SGA	Hospital records	Preterm birth OR = 1.09 SGA OR =1.3+ NICU admission OR =1.54+	After controlling for confounds, there was an increased risk for SGA and NICU admission only among infants with intrauterine cannabis exposure.	9
Infant - Toddler Cognitive and Neurodevelopmental Outcomes (N=5,116; n=801 CE; mean NOS score = 8.14)
Day et al., 1994(56)	N=829 mother-child dyads in US (3yrs at follow-up; n=266 CE)	Prospective	Self-report (retrospective recall)	Yes	IQ	SBIS	Unextractable Data	There were no significant effects of cannabis use during any trimester of pregnancy on the composite score of the SBIS. However, intrauterine cannabis use was associated with poorer short-term memory.	9
Dotters-Katz et al., 2017(50)	N=1,867 live births in US (2yrs at follow-up n=135 CE)	Prospective	Urine testing	No	Cognition, language, and motor development	BSID-II	Unextractable Data	After controlling for confounds, intrauterine cannabis exposure was not associated with the presence of any developmental deficits.	7
Fried & Watkinson, 1988	N= 217 children at 12 months and 153 at 24 months (n=32 CE)	Prospective	Self-report (frequency of use)	Yes	Cognition, language, and motor development	BSID	Unextractable Data	At 12-months, only maternal nicotine use was associated with lower cognition scores. Alcohol and cannabis use were not associated with BSID scores at 12 or 24 months.	7
Fried et al., 1990(51)	N=263 mother-child dyads in Canada (36–48 months at follow-up; n =133 CE)	Prospective	Self-report (frequency of use)	Yes	Verbal memory, motor development, cognition	MSCA	Unextractable Data	There was a dose-dependent relationship between frequency of cannabis use during pregnancy and infants’ verbal memory, motor development, and general cognition.	7
Grewen et al., 2015(57)	N=63 infants in US (n=20 CE)	Prospective	TLFB	Yes	Neural functional connectivity	fMRI; Whole-brain functional connectivity	Unextractable Data	Infants with intrauterine cannabis exposure demonstrated reductions in bilateral caudate and left anterior insula functional connectivity with cerebellum, and right caudate functional connectivity with occipital/fusiform regions in comparison to both drug-free and non-cannabis drug-exposed groups.	9
Richardson et al., 1995	N=569 mother-child dyads in US (9- and 19-month median follow-up; n= 104 CE)	Prospective	Self-report (frequency of use)	Yes	Cognition, language, and motor development	BSID	BSID B = −4.7	Third trimester cannabis use was associated with decreased BSID mental scores at 9 months only. Alcohol or tobacco use was not associated with decreased BSID scores.	9
Smid et al., 2022(52)	N=1,197 mother-child dyads in US (annual follow-up until 48 months; n=47 CE)	Prospective	Urine testing	Yes	FS-IQ, cognitive, motor and language development	BSID-III, WPPSI-R	Unextractable Data	After controlling for covariates, intrauterine cannabis exposure was not associated with any cognitive impairments at either 12, 24, or 36 months of age.	9
Stroud et al., 2018(53)	N=111 mother-child dyads in US (30-day follow-up; n=64 CE)	Prospective	TLFB, meconium analysis	Yes	Attention	NICU Network Neurobehavioral Scale	Attention B = −.185+	Cannabis and tobacco-exposed infants showed decreased ability to attend to stimuli than single-substance exposed infants or healthy controls.	9
Infant – Toddler Behavioral and Psychiatric Outcomes (N=4,682; n=346 CE; mean NOS Score = 8.25)
Eiden et al., 2018(114)	N=247 mother-child dyads in US (16-month at follow-up; n=97 CE)	Prospective	Self-report, saliva, meconium analysis	Yes	Emotion regulation	Gentle arm restraint test	Unextractable Data	Intrauterine cannabis exposure and frequency of maternal cannabis use was not associated with emotion dysregulation in infants.	9
El Marroun et al., 2011(61)	N=4,077 infants in Netherlands (18-month at follow-up ; n=88 CE)	Prospective	Self-report (single-item)	No	Internalizing and externalizing problems	CBCL	Aggression in Girls B =.91+ Attention in Girls B = .36+ Aggression in Boys B = −.15 Attention in Boys B = .36	Intrauterine cannabis exposure was a significant risk factor for aggression and attention problems in toddlerhood but only among girls.	7
Schuetze et al., 2019(115)	N=247 mother-child dyads in US (16-month at follow-up; n=97 CE)	Prospective	Self-report, saliva, meconium analysis	Yes	Autonomic reactivity and regulation	Gentle arm restraint test	Autonomic reactivity and regulation B = −.05	After controlling for covariates, intrauterine cannabis exposure was not a significant risk factor for deficits in autonomic reactivity and regulation in infants.	8
Stroud et al., 2018(53)	N=111 mother-child dyads in US (30-days at follow-up; n=64 CE)	Prospective	TLFB, meconium analysis	Yes	Self-regulation, handling (need for soothing), motor development	NICU Network Neurobehavioral Scale	Self-regulation B = −.36+ Handling B = .28 Motor development B =.14+	Cannabis and tobacco-exposed infants showed decreased ability to self-regulate and reduced motor activity than single-substance exposed infants or controls. Co-exposed infants also required additional soothing and all effects were stronger for girls than boys.	9
Early Childhood Cognitive and Neurodevelopmental Outcomes (N=193,432 n=1,953 CE; mean NOS Score = 8.36)
Betts et al., 2022(64)	N=189,558 children in Australia (8-yrs at follow-up; n=887 CE)	Prospective	ICD-10	Yes	Reading, writing, numeracy, spelling, and grammar	National Assessment Program – Literacy and Numeracy	Reading OR = 1.11 Spelling OE = 1.35+ Grammar OR = 1.40+ Writing OR = 1.31+ Numeracy OR = 1.25	After controlling for confounds, children with intrauterine cannabis exposure were significantly more likely to not meet the minimum national standards for writing, spelling, and grammar.	9
El Marroun et al., 2016(63)	N=263 children in Netherlands (6–8yrs at follow-up; n=96 CE)	Prospective	TLFB, urine testing	No	Brain morphology	3-Tesla MRI	Superior frontal area thickness B = .18+ Frontal pole thickness B = .20+	Prenatal cannabis exposure was not associated with differences in global brain volumes, such as total brain volume, gray matter volume, or white matter volume; however, cannabis-exposed children had thicker frontal cortices than controls and tobacco-exposed children.	7
Goldschmidt et al., 2007(65)	N=648 mother-child dyads in US (6yrs at follow-up; n=284 CE)	Prospective	Semi-structured interview	Yes	Verbal reasoning, quantitative reasoning, abstract reasoning, short-term memory	SBIS	Unextractable Data	Daily cannabis use in the first trimester of pregnancy was associated with lower verbal reasoning scores on the SBIS. Daily use in the second trimester predicted deficits in short-term memory and quantitative scores. Third-trimester daily use was associated with poorer quantitative scores. Daily use in at least one trimester was associated with lower intelligence test scores.	8
Leech et al., 1999(66)	N=608 mother-child dyads in US (6yr at follow-up; n=248 CE)	Prospective	Semi-structured interview	Yes	Attention	Continuous-Performance Test	CPT Omission Errors B =1.21+ CPT Commission Errors B = −.56+	Prenatal cannabis exposure in the second trimester of pregnancy was associated with poorer CPT performance.	9
Moore et al., 2023(70)	N=81 mother-child dyads in US (5-yr at follow-up; n=6 CE)	Prospective	Urine testing	Yes	Cognitive flexibility, inhibitory control, language	NIH Toolbox	Unextractable Data	There was no difference in the cognitive scores among those with and without prenatal exposure to cannabis.	9
Murnern et al. 2021(67)	N=99 children in US (3.5yrs at follow-up; n=15 CE)	Prospective	Self-report, urine testing	Yes	Attention, inhibitory control, executive functioning, episodic memory, processing speed	BSID-III, Block Design Test, NIH Toolbox for Children,	Attention B =1.80 Inhibitory Control B = 1.80 Executive functioning B = 5.16 Episodic memory B = 1.80 Processing speed B = −.67	After controlling for confounds, intrauterine cannabis exposure did not increase the odds of poorer performance in any cognitive task.	9
Noland et al., 2003(68)	N=316 mother-child dyads in US (4yrs at follow-up; n=85 CE )	Prospective	Self-report, meconium analysis	Yes	Executive Functioning	WPPSI-R	Executive functioning B = .16	After controlling for confounds, intrauterine cannabis exposure did not increase the odds of poorer performance on a task assessing executive functioning.	9
Noland et al., 2005(26)	N=330 mother-child dyads in US (4yrs at follow-up; n=85 CE )	Prospective	Self-report, meconium analysis	Yes	Attention	CPT	CPT Omission errors B = .32+	Severity of maternal cannabis use was positively correlated with omission errors, suggesting impaired sustained attention in exposed children.	9
Fried et al., 1992 (60)	N=272 Canadian mother-child dyads (60–72-month at follow-up) (n=137 CE)	Prospective	Self-report (frequency of use)	Yes	Verbal memory, motor development, cognition	MSCA	Unextractable Data	There was no relationship between frequency of prenatal cannabis use and performance on any of the cognitive outcomes.	7
O’Connell & Fried, 1991(69)	N=56 mother-child dyads in Canada (6–9yrs at follow-up; n = 28 CE)	Prospective	Self-report (frequency of use)	Yes	IQ	WISC-R	IQ Cohen’s d = .35 (p >.05)	After controlling for sociodemographic variables, there was no association between prenatal cannabis exposure and IQ scores.	7
Smid et al., 2021(52)	N=1,197 mother-child dyads in US (4yrs at follow-up; n=47 CE)	Prospective	Urine Testing	Yes	Cognitive, motor, and language development, attention	BSID-III, WPPSI-R, Conners’ Rating Scales–Revised	Unextractable Data	Intrauterine cannabis exposure was associated with impaired attention at 48 months, but was not associated with impairments in any other cognitive domain.	9
Early Childhood Behavioral and Psychiatric Outcomes (N=1,699 n=139; mean NOS Score = 8.75)
Moore et al., 2023(70)	N=81 mother-child dyads in US (5-yr at follow-up; n=6 CE)	Prospective	Urine testing	Yes	Internalizing and externalizing problems	CBCL	Unextractable Data	prenatal exposure to cannabis was associated with fewer somatic complaints and withdrawal behaviors than non-exposed controls	9
Murnern et al. 2021(67)	N=99 children in US (3.5yrs at follow-up; n=15 CE)	Prospective	Self-report, Urine Testing	Yes	Internalizing and externalizing problems	Bobo Test, CBCL	Unextractable Data	Cannabis-exposed children were significantly more aggressive and demonstrated more externalizing problem behaviors than children without any intrauterine substance use exposure.	9
Rompala et al., 2021(71)	N=322 mother-child dyads in US (3–6yrs at follow-up; n=71 CE)	Prospective	Self-report, placenta autopsy	Yes	Behavioral and psychological functioning	BASC-2	Aggression Cohen’s d =.54+ Anxiety Cohen’s d =.42+ Depression Cohen’s d =.26 Hyperactivity Cohen’s d =.39+ Withdrawn behavior Cohen’s d =.07	Prenatal cannabis use was a significant risk factor for aggression, anxiety, and hyperactivity in children but was not a significant risk factor for other behavioral/psychiatric problems.	8
Smid et al., 2021(52)	N=1,197 mother-child dyads in US (4yrs at follow-up; n=47 CE)	Prospective	Urine testing	Yes	Internalizing and externalizing problems	CBCL	Unextractable Data	Intrauterine cannabis exposure was not associated with any adverse behavioral or social outcomes in children.	9

Abbreviations: CE, cannabis exposure; NICU, neonatal intensive care unit; ICD, international classification of diseases; BSID, Bayley Scales of Infant and Toddler Development; SBIS, Stanford-Binet Intelligence Scale; MSCA, McCarthy Scales of Children’s Abilities; CBCL, Child Behavior Checklist; NIH, National Institute of Health; WPPSI, Wechsler Preschool and Primary Scale of Intelligence; CPT, Continuous Performance Test; WISC-R, Wechsler Intelligence Scale for Children-Revised; BASC, Behavior Assessment System for Children

^*:Newcastle Ottawa Scale (NOS) Scores of 7–9 indicate high methodological quality, 5–6 indicate medium methodological quality, and 1–4 indicate low methodological quality

^+:significant p-values (p <.05)

Quality assessments of included studies are reported in Table 2. Studies scored between 6 and 9 on the NOS, indicative of high methodological quality (38). The average NOS score was 8.17, with one study indicating a moderate risk of bias. A summary of our findings are depicted in Table 3.

Table 2:

Quality Assessment of Included Studies Using the NOS

Author, Year	Selection				Comparability		Outcome or Exposure			Total Points+
	1	2	3	4	1a*	1b	1**	2	3
Bada et al., 2005(48)	✔	✔	✔	✔	✔	✔	✔	✔	✔	9
Baer et al., 2019(97)	✔	✔	✔	✔	✔	✔	✔	✔	✔	9
Bandoli et al., 2021a(93)		✔	✔	✔	✔	✔	✔	✔	✔	8
Bandoli et al., 2021b(94)		✔	✔	✔	✔	✔	✔	✔	✔	8
Betts et al., 2022(64)	✔	✔	✔	✔	✔	✔		✔	✔	8
Brown et al., 2016(95)		✔	✔	✔	✔	✔	✔	✔	✔	8
Burns et al., 2006(96)		✔	✔	✔	✔	✔	✔	✔	✔	8
Chabarria et al., 2016(99)		✔	✔	✔		✔	✔	✔	✔	7
Coleman-Cowger et al., 2018(40)	✔	✔	✔	✔	✔	✔	✔	✔	✔	9
Conner et al., 2015(4, 28)	✔	✔	✔	✔	✔	✔	✔	✔	✔	9
Corsi et al., 2019(4)	✔	✔	✔	✔	✔	✔		✔	✔	8
Crume et al., 2018(97)	✔	✔	✔	✔	✔	✔		✔	✔	8
Day et al.,1994	✔	✔	✔	✔	✔	✔	✔	✔	✔	9
Dotters-Katz et al., 2017(50)		✔	✔	✔		✔	✔	✔	✔	7
Eiden et al., 2018(114)	✔	✔	✔	✔	✔	✔	✔	✔	✔	9
El Marroun et al., 2011(61)	✔	✔		✔		✔	✔	✔	✔	7
El Marroun et al., 2016(63)	✔	✔		✔		✔	✔	✔	✔	7
Fried et al., 1988(55)	✔	✔		✔	✔	✔		✔	✔	7
Fried et al., 1990(51)	✔	✔		✔	✔	✔		✔	✔	7
Fried et al., 1992(60)	✔	✔		✔	✔	✔		✔	✔	7
Gabrhelik et al., 2020(98)	✔	✔	✔	✔	✔	✔		✔	✔	8
Goldschmidt et al., 2007(65)	✔	✔	✔	✔	✔	✔		✔	✔	8
Gray et al., 2010(41)	✔	✔	✔	✔	✔	✔	✔	✔	✔	9
Greiner et al., 2020(99)	✔	✔	✔	✔	✔	✔	✔	✔	✔	9
Grewen et al.,2015(57)	✔	✔	✔	✔	✔	✔	✔	✔	✔	9
Haight et al., 2021(100)	✔	✔	✔	✔	✔	✔		✔	✔	8
Hayatbakhsh et al., 2012(101)		✔	✔	✔	✔	✔	✔	✔	✔	8
Howard et al., 2019(42)	✔	✔	✔	✔	✔	✔	✔	✔	✔	9
Kharbanda et al., 2020(102)	✔	✔	✔	✔	✔	✔	✔	✔	✔	9
Koto et al., 2022(103)	✔	✔	✔	✔	✔	✔		✔	✔	8
Lee et al., 2020(43)	✔	✔	✔	✔	✔	✔	✔	✔	✔	9
Leech et al., 1999(66)	✔	✔	✔	✔	✔	✔	✔	✔	✔	9
Mark et al., 2016(104)	✔	✔	✔	✔	✔	✔	✔	✔	✔	9
Michalski et al., 2020(103)	✔	✔		✔	✔	✔	✔	✔	✔	8
Moore et al., 2023(70)	✔	✔	✔	✔	✔	✔	✔	✔	✔	9
Murnern et al. 2021(67)	✔	✔	✔	✔	✔	✔	✔	✔	✔	9
Nawa et al.,(104)		✔	✔	✔	✔	✔	✔	✔	✔	8
Nguyen et al., 2022(105)	✔	✔	✔	✔	✔	✔		✔	✔	8
Noland et al., 2003(68)	✔	✔	✔	✔	✔	✔	✔	✔	✔	9
Noland et al., 2005(26)	✔	✔	✔	✔	✔	✔	✔	✔	✔	9
O’Connell & Fried, 1991(51)	✔	✔	✔	✔	✔	✔		✔	✔	8
Ostrea et al., 1997(106)	✔	✔	✔	✔			✔	✔	✔	7
Prunet et al., 2017(46)	✔	✔	✔	✔				✔	✔	6
Richardson et al., 1995(54)	✔	✔	✔	✔	✔	✔	✔	✔	✔	9
Roca et al., 2020(107)	✔	✔	✔	✔			✔	✔	✔	7
Rodriquez et al., 2019(108)	✔	✔	✔	✔		✔	✔	✔	✔	8
Rompala et al., 2021(71)	✔	✔	✔	✔	✔	✔		✔	✔	8
Sasso et al., 2021(109)	✔	✔	✔	✔		✔		✔	✔	7
Saurel-Cubizolles et al., 2014(110)	✔	✔	✔	✔	✔	✔		✔	✔	6
Scuhetze et al., 2019(115)	✔	✔	✔	✔		✔	✔	✔	✔	8
Shi et al., 2020(47)	✔	✔	✔	✔	✔	✔	✔	✔	✔	8
Shiono et al., 1995(111)	✔	✔	✔	✔		✔	✔	✔	✔	8
Smid et al., 2021(52)	✔	✔	✔	✔	✔	✔	✔	✔	✔	9
Straub et al., 2021(112)	✔	✔	✔	✔	✔	✔	✔	✔	✔	9
Stroud et al., 2018(53)	✔	✔	✔	✔	✔	✔	✔	✔	✔	9
Van Gelder et al., 2010(113)	✔	✔	✔	✔	✔	✔	✔	✔	✔	9
Warshak et al., 2015(27)	✔	✔	✔	✔	✔	✔	✔	✔	✔	9

⁺Newcastle Ottawa Scale (NOS) Scores of 7–9 indicate high methodological quality, 5–6 indicate medium methodological quality, and 1–4 indicate low methodological quality

^*,study controls for other substance use

^**,follow-up time of >30 days (for prospective studies only)

Table 3:

Summary of Qualitative Evidence Concerning Intrauterine Cannabis Exposure and Infant/Childhood Development

Developmental Outcome	Number of Studies Finding a Negative Impact of Cannabis Use	Number of Studies Finding No Impact of Cannabis Use	Percentage of Studies Obtaining an Adverse Effect of Cannabis Use
Birthweight	6	4	60%
NICU Admission	0	1	0%
Infant Mortality	1	0	100%
Preterm Birth	2	4	33.3%
Infant/Toddler Neurodevelopment	1	0	100%
Infant/Toddler Attention	2	0	100%
Infant/Toddler Language Development	0	2	16.7%
Infant/Toddler Cognition (General)	2	5	28.6%
Infant/Toddler Motor Development	2	2	50%
Infant/Toddler Emotion Regulation	0	2	0%
Infant/Toddler Internalizing Problems	0	1	0%
Infant/Toddler Externalizing Problems	1	0	100%
Early Childhood Neurodevelopment	1	0	100%
Early Childhood Attention	3	1	75%
Early Childhood Verbal Reasoning	2	1	66.7%
Early Childhood Memory	1	2	33.3%
Early Childhood Quantitative Reasoning	1	0	100%
Early Childhood Executive Functioning	0	1	0%
Early Childhood IQ	0	1	0%
Early Childhood Internalizing Problems	1	2	33.3%
Early Childhood Externalizing Problems	2	1	66.7%

Meta-Analytical Findings

Birth Outcomes

In our meta-analysis of 20 studies exploring the association between intrauterine cannabis exposure and risk of preterm delivery, mothers using cannabis were 1.68x more likely to spontaneously deliver their newborn preterm in comparison to mothers without cannabis use (OR=1.68, 95% CI=1.05– 2.71, p=.03, I²=99%, Figure 2). Concerning the risk of infant mortality, our analysis of five publications did not reveal any statistically significant difference in the risk of infant death between newborns exposed to cannabis in-utero and those who were not exposed (OR=2.31, CI=.96 – 5.59, p>.05, I²=89%, Figure 3). Regarding the risk of low birth weight, the meta-analysis of 18 studies revealed that mothers using cannabis during pregnancy were 2.60x more likely to give birth to a baby with low birth weight compared to mothers who were not using cannabis during pregnancy (OR=2.60, CI=1.71–3.94, p<.001; I²=99%; Figure 4). To determine the risk of requiring neonatal intensive care unit (NICU) admission, our random-effects meta-analysis of 10 studies revealed that newborns with intrauterine cannabis exposure were 2.51x more likely to require NICU admission than nonexposed newborns (OR=2.51, CI=1.46 – 4.31; p<.001; I²=100%; Figure 5). However, we found no elevated risk of birth defects (Figure 6) among newborns exposed to cannabis in-utero when compared to those without cannabis exposure across three studies exploring this outcome (OR=2.11, CI=.91–4.86; p<.05; I²=98%).

Figure 2.

Meta-analysis of preterm birth after cannabis use during pregnancy.

Figure 3.

Meta-analysis of infant mortality after cannabis use during pregnancy.

Figure 4.

Meta-analysis of low birth weight after cannabis use during pregnancy.

Figure 5.

Meta-analysis of neonatal intensive care unit (NICU) admission after cannabis use during pregnancy.

Figure 6.

Meta-analysis of birth defects after cannabis use during pregnancy.

Qualitative Findings

Birth Outcomes

An additional twelve studies exploring the relationship between intrauterine cannabis exposure and various birth outcomes were identified; however, data were unextractable, and therefore could not be included in the meta-analysis. Of these publications, eight determined at least one adverse neonatal outcome associated with intrauterine cannabis exposure(39–47), with the most commonly reported adverse outcome being low birth weight (Table 3).

In a large retrospective study examining over 4 million hospital records in California, Shi et al.(47) observed that newborns with intrauterine cannabis exposure exhibited a significantly higher likelihood of being born with low birth weight compared to those without intrauterine substance exposure. Similar findings were obtained in four additional research groups, where prenatal cannabis use was associated with an elevated risk of low birth weight among cannabis-exposed newborns (41, 42, 44, 45). In contrast, in a retrospective analysis examining 11,811 maternal-infant dyads, Bada et al. (48) did not observe an elevated risk for low birth weight or preterm birth amongst mothers self-reporting prenatal cannabis use.

Coleman-Cowger examined whether cannabis and tobacco co-use during pregnancy corresponded with adverse birth outcomes, including NICU admission, stillbirth, presence of a birth defect, and spontaneous preterm birth (41). They observed that that the adjusted odds ratio of stillbirth rose to 12.1 (CI: 1.0–141.8) when mothers yielded positive tests for cannabis. Furthermore, co-use of tobacco and cannabis was associated with an elevated risk of delivering a baby with a birth defect in comparison to mothers without any substance use, mothers using tobacco only, or mothers using cannabis only. In contrast, Conner et al. did not observe a significant difference in neonatal morbidities, including low birthweight and NICU admission among infants with intrauterine cannabis exposure in comparison to unexposed infants, after adjusting for other substance use and demographic variables (49).

Infant to Toddler (30 days–3 years) Cognitive and Neurodevelopmental Outcomes

Due to unextractable data, the effects of intrauterine cannabis exposure on infant cognitive and neurodevelopmental outcomes are discussed qualitatively. Seven studies evaluated cognition of infants or toddlers with intrauterine cannabis exposure (50–56), while one publication examined functional connectivity among cannabis exposed infants (57). Of these publications, four did not detect any cognitive impairments among infants or toddlers prenatally exposed to cannabis (50, 52, 56, 58). Utilising the Bayley Scales of Infant and Toddler Development-III (BSID-III), Dotters-Katz et al. (50) longitudinally assessed cognitive functioning of 135 cannabis-exposed infants and 1732 matched controls for 24 months. At endpoint, no group differences in cognition, language, or motor development emerged after controlling for confounds, including other substance use and demographics. Similarly, Fried and Watkinson (59) assessed the cognitive functioning of 217 24-month old children with substance exposure, including cannabis, tobacco, or alcohol, and found no difference on BSID-II total and composite scores across groups. However, in a prospective study following 9-month old infants with prenatal substance exposure, Richardson et al. (54) observed that infants exposed to cannabis in-utero had decreased BSID composite scores than infants with intrauterine alcohol or tobacco exposure.

The McCarthy Scales of Children’s Abilities (MSCA) was administered in two studies (51, 60) from the same research group to estimate the cognitive abilities of children with prenatal cannabis exposure at 36-, 48-, 60-, and 72-months of age. The researchers found a dose-dependent relationship between frequency of cannabis use during pregnancy and infants’ verbal memory, motor development, and IQ at 36 and 48 months (51), but not at 60 or 72 months(60).

In one study administering the NICU Network Neurobehavioral Scale to assess 30-day old infants’ attention, significant impairments were detected among cannabis- and tobacco-exposed infants after controlling for covariates (53). However, impairments were not detected in infants that were exposed to only cannabis or tobacco. These effects were stronger for females than males.

Grewen et al. (57) investigated the impact of prenatal cannabis exposure on functional neurocircuitry in infants with intrauterine cannabis exposure against infants exposed to other substances and drug-free controls. The results showed reduced connectivity patterns between the insula and three major brain regions, including the frontal cortex and striatum, between infants with prenatal cannabis exposure and those without.

Infant to Toddler (30 days – 3 years) Behavioral and Psychiatric Outcomes

In two studies, Eiden et al (2018; 2020) investigated autonomic reactivity, autonomic regulation, and emotion regulation in cannabis-exposed infants. Using the gentle arm restraint test to induce stress, infants at 16-months who were exposed to cannabis in-utero, demonstrated comparable scores in autonomic reactivity, autonomic regulation, and emotional regulation as matched controls without prenatal substance exposure. Instead, among both groups, maternal dysregulation, maternal tobacco use post-pregnancy, and low maternal sensitivity predicted poorer emotion regulation. El Marroun et al (61) examined behavioral problems among 18-month-old toddlers via the CBCL, a widely used instrument relying upon parental report to identify eight potential syndromes, including: social withdrawal, anxiety/depression, somatic complaints, delinquency, aggression, social problems, thought problems, and attention problems (62). After controlling for confounds, intrauterine cannabis exposure was associated with greater attention problems and aggression in girls only. Postnatal tobacco exposure further increased girls’ risk of attention and aggression problems. Stroud et al (53) examined self-regulation, arousal, and motor development of 111 30-day-old infants, where 64 infants were exposed to cannabis and tobacco prenatally, as indicated through meconium analysis. In comparison to healthy controls and controls exposed to only one substance, dual-exposed infants demonstrated reduced motor activity and poorer self-regulation abilities. Further, these effects in cannabis and tobacco-exposed infants were more pronounced for females versus males.

Early Childhood (4–8 years) Cognitive and Neurodevelopmental Outcomes

One study compared the brain morphology of children with intrauterine cannabis exposure against matched controls (63). Using structural MRI, the total, gray, and white matter brain matter volume of 96 6-year-old children with intrauterine cannabis exposure and 167 matched healthy controls were compared. Contrary to the researchers’ hypothesis, cannabis-exposed children demonstrated thicker frontal cortices than controls, and no significant differences in grey or white brain matter volume were obtained.

Ten studies examined differing cognitive outcomes of children with in-utero cannabis exposure (26, 52, 60, 64–70). There was substantial heterogeneity in the cognitive outcome assessed (e.g., attention, memory, overall IQ) and available data provided which precluded a quantitative synthesis of the findings. Of the ten studies, four detected adverse cognitive outcomes in children with intrauterine cannabis exposure (26, 64–66) with impairments predominantly occurring in attention and verbal reasoning (Table 3). Leech et al.(66) assessed 6-year-old children’s attention using the Continuous Performance Test (CPT), and found that even after controlling for confounds, cannabis use in the second or third trimester of pregnancy predicted impaired performance. Similarly, Noland et al. (26) employed the CPT to assess differences in attention between 6–9 year-old children with intrauterine cannabis exposure against controls without any prenatal substance exposure. After controlling for covariates, a dose-dependent relationship between maternal cannabis use and performance on the CPT emerged. Namely, only mothers using cannabis more than weekly as indicated through self-report and confirmed through meconium analysis, had children with impaired attention abilities. These findings parallel Goldschmidt et al (65) where daily cannabis use in the second and third trimester predicted poor performance on tests assessing memory and quantitative reasoning among 6-year-old children. Moreover, daily cannabis use in any trimester predicted lower childhood IQ. In one study comparing the scores of children prenatally exposed to cannabis on a national literacy and numeracy test against children without any intrauterine substance exposure, Betts et al (64) found that prenatal CUD was associated with poorer performance on writing, spelling, and grammar components of the test, even after adjusting for maternal confounds.

Concerning null findings, Noland and colleagues (68) found that after controlling for maternal confounds, fetal-related variables, and demographics, no differences in a test assessing executive functioning emerged between children with intrauterine cannabis exposure and matched controls. In a more recent study comparing fifteen 4-year-olds with intrauterine cannabis exposure against matched controls on multiple cognitive domains, including attention, inhibitory control, executive functioning, and memory, no cognitive differences emerged between groups after controlling for baseline demographics, maternal substance use, and birth outcomes (67). Similar findings were obtained in three additional studies (52, 60, 69, 70), where after controlling for maternal and sociodemographic variables, no differences in cognition emerged between cannabis-exposed children and matched controls.

Early Childhood (4–8 years) Behavioral and Psychiatric Outcomes

Four of the included studies assessed behavioral and/or psychiatric outcomes of young children with intrauterine cannabis exposure (52, 67, 71). However, the inability to extract data prevented a quantitative synthesis from being conducted. Utilizing the Behavioral Assessment System for Children (BASC) among a sample of 3.5–6 year-olds, Rompala et al. (71) observed that after controlling for maternal substance use and other baseline variables, prenatal cannabis use was a significant risk factor for increased aggression, anxiety, and hyperactivity in children. Similar findings were obtained by Murnern et al.(67), where prenatal cannabis use increased the odds of exhibiting aggressive, oppositional defiant, and externalizing problems in 3.5–4-year old children. In contrast, Moore et al. (70) found that in a sample of 5-year-old children, those exposed to cannabis in-utero were more likely to have fewer somatic complaints and withdrawal problems than non-exposed controls. However, in a different sample of 5-year-olds, Smid et al (52) found that prenatal cannabis use predicted greater severity of problems on all subscales of the CBCL in 5-year-old children, even after controlling for baseline and maternal substance use variables.

Discussion

Among pregnant women, cannabis use rates are rising as rapidly as they are amid non-pregnant women of reproductive age (72). Given that THC, the main psychoactive constituent in cannabis, readily crosses the placenta (9) and binds to endocannabinoid receptors in the fetal brain (71), there is concern for adverse birth and childhood effects. This article sought to systematically review and meta-analyze the immediate and long-term physical, cognitive, psychological, and behavioral outcomes of infants and young children (<8 years of age) with intrauterine cannabis exposure.

Our meta-analysis suggests that infants prenatally exposed to cannabis are more likely to be born preterm, with a low birth weight, and require NICU admission than infants without cannabis exposure. However, cannabis exposed infants are not at greater risk of a birth defect or death within a year, including sudden unexpected infant death. Regarding the qualitative findings, our analysis revealed a mixed body of evidence concerning the effects of intrauterine cannabis exposure on the global cognitive functioning of infants and toddlers. However, we observed disrupted neural activity in cannabis-exposed infants when compared to both drug-free and non-cannabis drug-exposed toddlers. Furthermore, no significant evidence supported the presence of internalizing problems among this age group; however, externalizing problems, such as aggression and hyperactivity, may be associated with prenatal cannabis exposure in toddlerhood. Extending beyond this age group, attentional impairments may appear in early childhood among those with intrauterine cannabis exposure. However, global cognitive functioning cognition remains intact. In terms of behavioral outcomes, intrauterine cannabis exposure seemed to heighten the risk of externalizing problems, but the evidence concerning its association with internalizing problems remained inconclusive.

Disturbances in neurodevelopment may underlie impairments in certain behavioral and cognitive domains. A recent narrative review observed that brain regions with highest levels of CB1 receptor expression, including the prefrontal cortex, may function differently in children and adolescents with intrauterine cannabis exposure than non-exposed controls (73). For example, El Marroun et al (63) found atypical frontal cortex thickness among 6–8 year-old children with prenatal cannabis exposure. Similarly, Grewen et al. (57) identified a distinct pattern of hypo-connectivity between the insula and other major neural regions, including the cerebellum, prefrontal cortex, and striatum. These early differences may adversely impact subsequent development of inhibitory control networks, and contribute to behavioral and cognitive deficits reported in children with prenatal cannabis exposure. Relatedly, studies exploring the structural and molecular composition of human fetal brains with intrauterine cannabis exposure are scarce, but have produced significant results. Examining 42 mid-gestation human fetuses using in situ hybridization histochemistry, Wang et al (13) found maternal cannabis use as correlating with alterations in the endogenous opioid system, including, reduced mRNA expression levels for proenkephalin in the putamen, elevated m-opioid receptor expression in the amygdala, and decreased kappa receptor expression in the thalamus. Taken together, these findings suggest that the endocannabinoid system is essential in the ontogeny of the CNS during fetal neurodevelopment and that early fetal exposure to cannabis may induce subtle, but lasting, neurodevelopmental modifications. Moreover, preclinical models parallel the review’s findings, where prenatal administration of THC negatively impacts physical, cortical, cognitive, and behavioral development of offspring (24, 25). Rodents with prenatal THC exposure demonstrate reduced birthweight (74). Further, intrauterine THC exposure in rodents corresponds with increased ultrasonic vocalizations (cries) when isolated in infancy, indicative of increased anxiety and a more pronounced stress responses (75). Among other preclinical models of intrauterine cannabis exposure, impairments in learning and memory are among the most frequently demonstrated effects (76). Recent research has demonstrated that administering THC (.5 mg/kg/day) to pregnant rats, which is equivalent to estimates of moderate THC exposure in humans (77), leads to significant neural and memory impairments in offspring (25). Namely, prenatal THC exposure induced long-lasting alterations within cortical glutamatergic and noradrenergic genes, which corresponded with deficits in short- and long-term memory. Although findings from preclinical models support some of the review’s results, there is a clear need for further research investigating other outcomes in animals and humans to ascertain whether these deficits are directly attributed to intrauterine cannabis exposure.

Although a comprehensive systematic review and meta-analysis was performed, several limitations are noted. First, there was significant methodological heterogeneity across studies, including differences in sample characteristics (e.g., polysubstance-using mothers and mothers with comorbid physical and/or psychiatric disorders), variability among measurements of cannabis use and primary outcomes, and inconsistencies in controlled variables. For example, there was substantial heterogeneity in methods employed to quantify cannabis use among mothers, which included single-item measures, meconium analysis, urine testing, and saliva sampling. The use of a single-item question to ascertain cannabis use is especially problematic, as pregnant women are likely to conceal their substance use status with healthcare providers due to fears of prosecution or losing child custody, and stigma (78). Relatedly, meconium analysis was administered in only five studies, despite being considered the clinical gold standard for assessing prenatal drug exposure (79). However, this test is not without imperfections, as recent evidence indicates that meconium may be incapable of detecting cannabis use during the first or second trimester of pregnancy (41). Given this limitation, administering frequent urine toxicology tests in addition to reliable and valid self-report questionnaires such as the Timeline Follow Back (TLFB) is recommended to provide the most accurate estimates of prenatal cannabis use (80). Further, this approach would illuminate whether there is a dose-dependent relationship between frequency of use and offspring outcomes. A similar limitation entails the majority of studies measuring cannabis use as a dichotomous construct (yes/no), with few studies evaluating frequency of cannabis use and timing of use. Prior literature has suggested dose-dependent relationships between frequency of cannabis use and adverse outcomes in adolescent and adult populations, including development of psychosis, reduced occupational attainment, and impairments in select cognitive domains (81). Amongst the few studies assessing frequency of maternal cannabis use, there were clear dose-dependent relationships between levels of use and adverse birth and childhood outcomes. Moreover, timing of cannabis exposure may be an additional important factor to assess, where cannabis use in later trimesters may correspond with more severe impairments (65, 66). An additional major limitation is the significant heterogeneity in data reporting, which precluded a quantitative synthesis for numerous outcomes. A final limitation involves the discrepancy in controlled variables across the studies which inherently limit causal inferences between intrauterine cannabis exposure and behavioral, cognitive, psychiatric, and physical outcomes of infants and children. Cannabis use in pregnancy accompanies several additional adverse risk factors that negatively impact postnatal outcomes, including maternal polydrug exposure, presence of psychiatric disorders, impoverished nutritional status, elevated maternal stress, low socioeconomic status, and unstable home environments (82, 83). Although most of the included studies adjusted for many of these confounding factors, these effects remain difficult to isolate and may create interpretation bias. Furthermore, none of the included studies collected or controlled for genetic confounds, which raises concerns about the potential impact of genetic factors on the observed associations between cannabis use and infant or childhood outcomes. Without accounting for genetic predispositions, it is difficult to disentangle whether the reported effects are solely attributed to cannabis use or if they are influenced, at least in part, by preexisting genetic vulnerabilities.

Pregnancy provides a unique opportunity for providing care to women who may not otherwise see a physician (78). Additionally, pregnant women represent a captive audience who may be motivated to improve health behaviors for the benefit of the infant (84). Finally, intervening during the perinatal period is arguably the most opportune period of intervention because of short- and long-term intergenerational effects. Thus, healthcare providers are uniquely poised to provide clinical advice concerning the potential health risks of prenatal cannabis use. Moreover, healthcare providers should explore patients’ reasons for cannabis use, as many pregnant patients report use to provide relief from pre-existing conditions or pregnancy-related symptoms, including nausea, chronic pain, and psychiatric disorders (6). When cannabis is being used for these purposes, providers should discuss alternative solutions, including safer substitutions. Pregnant women using cannabis therapeutically will also require guidance on when and how to resume use postpartum.

In addition, our review demonstrates that intrauterine cannabis exposure may increase risk for certain cognitive deficits and externalizing problems. Further, the review indirectly showed that many pregnant women using cannabis are polysubstance users, which yields important implications for childhood outcomes. Young children prenatally exposed to substances may face significant obstacles in adulthood, as impairments occurring in early childhood are seldomly restricted to this developmental period but typically affect the individual throughout their lifespan (85). For example, cognitive impairments in preschool-aged children with intrauterine exposure to alcohol predicts poorer psychosocial functioning, development of a psychiatric disorder, and reduced life satisfaction in adulthood (85, 86). Unfortunately, there is a dearth of research evaluating the outcomes of adolescents and adults with intrauterine cannabis exposure; yet preclinical models show adverse effects. Adult rats perinatally exposed to THC exhibit abnormal social behaviors (87), while prenatal THC exposure leads to impaired reward-seeking behaviors (88), cognition (25), and altered stress sensitivity (89).

While environmental confounders may obscure the long-term impacts of prenatal cannabis exposure, supporting this population of children remains imperative to mitigate any potential negative consequences. There are a multitude of promising interventions that support families and children impacted by maternal substance use (90). Many of these programs aim to increase the physical and emotional wellbeing of mothers, provide healthcare for the child, and enhance the child-caregiver attachment and interactions. Implementation of these programs are associated with significant improvements in the child’s psychological, behavioral, and cognitive development, with enduring positive benefits (91).

We obtained evidence that prenatal cannabis use is associated with lower birth weight, preterm birth, and NICU admission in newborns, but little evidence that prenatal cannabis exposure adversely impacts behavioral or cognitive outcomes in early childhood, with the exception of attention, and externalizing problems. It should be noted that our conclusions are limited by significant methodological variations across included studies. Therefore, future studies utilizing well-controlled, prospective designs with sufficiently powered sample sizes are needed to fully elucidate the effects of intrauterine cannabis exposure and early childhood outcomes. Future research should also implement standardized measures to assess cannabis use, including frequency of use, modes of administration, and cannabis composition (THC:CBD ratios) to better understand the relationship between prenatal cannabis use and developmental outcomes. Finally, since human studies are invariably complex and cannot fully control for environmental variables, further preclinical research is essential in providing insight into which neurobiological mechanisms underlie the association between intrauterine cannabis exposure and various offspring outcomes. As shifts in the social and legal landscape is likely to foster the use of cannabis, understanding the effects on birth and childhood outcomes should become a global priority.