Authors: Ayesha C. Sujan, Anish Pal, Lyndsay A. Avalos, Kelly C. Young-Wolff
Categories: Pediatrics, cannabis, exposure, marijuana, pregnancy, prenatal exposure
Source: Frontiers in Pediatrics
Cannabis use among pregnant women has increased over time. Therefore, there is a great public health need to understand the consequences of in utero cannabis exposure. While several meta-analyses and reviews have summarized the evidence of in utero cannabis exposure on adverse obstetric outcomes (e.g., low birth weight and preterm birth) and long-term offspring development, there has not been a focus on in utero cannabis exposure and risk for structural birth defects.
We conducted a systematic review using PRISMA guidelines to evaluate the association between in utero cannabis exposure and structural birth defects.
We identified 20 articles to include in our review and focused on interpreting findings from the 12 that adjusted for potential confounders. We report findings by seven organ systems. Within the 12 articles, four reported on cardiac malformations, three reported on central nervous system malformations, one reported on eye malformations, three reported on gastrointestinal malformations, one reported on genitourinary malformations, one reported on musculoskeletal malformations, and two reported on orofacial malformations.
Findings on associations between in utero cannabis exposure and birth defects reported in more than two articles were mixed (i.e., findings for cardiac, gastrointestinal, central nervous system malformations). Findings for associations between in utero cannabis exposure and birth defects reported in two articles (i.e., orofacial malformations) or in a single article (eye, genitourinary, and musculoskeletal) suggested that cannabis exposure was not associated with these types of malformations, but strong conclusions cannot be drawn from such sparce research. We review the limitations and gaps in the existing literature and call for more research to rigorously evaluate associations between in utero cannabis exposure and structural birth defects.
identifier CRD42022308130.
Keywords: pregnancy, prenatal exposure, in utero exposure, cannabis, marijuana
Research has documented an increase in rates of cannabis use among pregnant people over time. Among a nationally representative sample of pregnant individuals in the United States, the prevalence of self-reported prenatal cannabis use in the past month increased from 3.4% in 2002–2003, to 7.0% in 2016–2017 (1). Prenatal cannabis use may increase even more rapidly as more US states legalize cannabis for recreational use (2–7). Moreover, cannabis use in pregnancy could impact fetal development because cannabis is lipid soluble and is able to cross the placenta and blood-brain barrier (8), and some previous studies have suggested a potential link between in-utero cannabis exposure and adverse offspring outcomes [e.g., (9)]. Therefore, there is a great public health need to understand the consequences of in utero cannabis exposure on offspring development. Several meta-analyses and reviews have summarized the evidence of in utero cannabis exposure on adverse obstetric outcomes (e.g., low birth weight and preterm birth) and long-term offspring development (8, 10–15). However, reviews to date have not focused on research regarding in utero cannabis exposure and risk for structural birth defects. The causes and risk factors for many structural birth defects remains unknown, and understanding preventable causes and risk factors for structural birth defects is particularly important given the strong association between birth defects and morbidity/mortality (16). Given this need, we conducted a systematic review using Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines to evaluate whether in utero cannabis exposure is associated with structural birth defects compared to pregnancies with no cannabis exposure (Prospective Register of Systematic Reviews [PROSPERO] registration CRD42022308130; (17)].
Web of Science and PubMed databases were searched for English language articles published before February 1, 2022 utilizing the following key “(Pregnancy OR Prenatal OR In utero OR Perinatal) AND (Cannabis OR Marijuana) AND (Birth defects OR Congenital malformations OR Congenital anomalies OR Central nervous system defect OR Neural tube defects OR Holoprosencephaly OR Microcephaly OR Ear defect OR Eye defect OR Gastrointestinal defect OR Biliary atresia OR Esophageal atresia OR Tracheoesophageal fistula OR Intestinal atresia OR Intestinal stenosis OR Pyloric stenosis OR Hypospadias OR Renal agenesis OR Renal hypoplasia OR Renal dysplasia OR Cardiac defect OR Musculoskeletal defect OR Congenital diaphragmatic hernia OR Gastroschisis OR Limb deficiency OR Omphalocele OR Orofacial defect OR Respiratory defect OR Choanal atresia OR Cleft lip OR Cleft palate).” The inclusion criteria were English-language articles and epidemiological studies. Animal studies and review articles were excluded as the focus of our review was strictly on human outcomes.
The search revealed 299 potentially relevant articles of which 48 were duplicates. We created an EndNote library of 251 non-duplicate articles. Two authors then independently reviewed the titles and abstracts of the articles in the EndNote library to exclude articles that did not meet the inclusion criteria. After their independent reviews, the two authors discussed disagreements and together decided to include 37 articles for a full text review. During the full text review, 17 additional articles were excluded for the following study design was a case study (18), a comparable study was conducted by the same authors using the same dataset (19–23), and the study did not specifically evaluate associations between cannabis exposure in pregnancy and birth defects [e.g., cannabis was included in a general substance use exposure variable or the outcome studied was not a birth defect; (24–34)]. Therefore, the final review included 20 articles (9, 22, 35–53). See Figure 1 for a PRISMA flow diagram illustrating our identification process of articles for our final review.
Figure 1 PRISMA flow diagram.
Of the 20 included articles, 8 were from prospective studies using recruited samples (35–42), and 12 were from retrospective cohort or case-control studies using health care records (9, 43–53). Samples sizes varied from 50 to 3,067,069. Earliest birth years for cohorts varied from 1968 to 1980. Only 3 of the articles reported on studies using urine toxicology tests (46–48); the rest reported on studies that relied on self-report to measure prenatal cannabis use. Of the 17 articles that reported on studies using self-reports to measure cannabis use, 16 had measures of self-reported cannabis use, and 1 had a measure of self-reported cannabis-related diagnoses (43). The outcome definitions varied across studies with some investigating associations with specific malformations and other studies investigating associations with any malformation. While 8 studies did not adjust for any potential confounders (38, 40, 41, 44, 46, 47, 49, 53), the rest adjusted for confounding, though the specific factors adjusted for varied across studies. Table 1 provides information about the characteristics of each individual study.
Table 2 includes information on adjusted associations between in utero cannabis exposure and specific birth defects. When examining associations, we only considered the 12 studies that adjusted for confounding, given the importance in doing so in assessing epidemiologic relationships (54). We included information about associations with specific malformations whenever available. However, given the rarity of specific malformations, most studies evaluated associations with organ specific malformations grouped together.
Results were inconsistent across the four articles reporting findings from studies assessing associations between in utero cannabis exposure and cardiac malformations (9, 43, 45, 52). One article (9) indicated a dose-response relationship between self-reported cannabis use three months before pregnancy through the first trimester and ventral septal defect [any use OR: 1.9, 95% CI: 1.3, 2.8; use <2 days/week OR: 2.20, 95% CI: 1.2, 3.9; use >3 days/week OR: 3.7, 95% CI: 1.6, 9.0; (9)]. Another study found increased odds of Ebstein anomaly associated with maternal self-reported first-trimester cannabis use, though the confidence interval around the estimate was wide and included the null [OR: 1.8, 95% CI: 0.9, 3.8; (45)]. Additionally, two other articles (43, 52) reported no elevated risk of any cardiac malformation among infants born to individuals with a cannabis-related diagnosis made during pregnancy or delivery [RR: 1.0, 95% CI: 0.8, 1.2; (43)] and no associations between maternal self-reported cannabis use in the month before pregnancy or during the first trimester of pregnancy and eight specific cardiac malformations [Table 2; (52)].
Three articles reported findings from studies assessing in utero cannabis exposure and central nervous system (CNS) malformations, and results were conflicting (43, 50, 52). Two studies focused on neural tube defects [NTD; (50, 52)]—Van Gelder et al. reported on two subtypes of NTD [anencephaly and spina bifida; (52)], while Shaw et al. focused on any NTD (50). Van Gelder et al. reported increased odds of anencephaly [odds ratio [OR]: 2.2, 95% CI: 1.3–3.7; (52)] but not spina bifida [OR: 0.9, 95% CI: 0.6–1.4; (52)] among infants born to individuals who self-reported cannabis use in the month before pregnancy or during the first trimester of pregnancy (52); and, Shaw et al. failed to find an association between self-reported cannabis use three months before pregnancy through pregnancy and any NTD [OR: 0.7, 95% CI: 0.5–1.2; (50)]. The third study (43) found an increased risk of any CNS malformations among infants born to individuals with a self-reported cannabis-related diagnosis [relative risk [RR]: 1.2, 95% CI: 1.0, 1.5; (43)].
One article reported on the association between in utero cannabis exposure and eye malformation (43). The study failed to find an association between a cannabis-related diagnosis made during pregnancy or delivery and eye malformation [RR: 1.1, 95% CI: 0.7, 1.7; (43)].
Three articles reported findings from studies assessing associations between in utero cannabis exposure and gastrointestinal malformations (43, 51, 52). The findings from these three studies were mixed. Two articles suggested in utero cannabis exposure was associated with increased risk of gastrointestinal malformations. Specifically, one article (43) reported an association between cannabis-related diagnoses made during pregnancy or delivery and any gastrointestinal malformation [RR: 1.3, 95% CI: 1.1, 1.5; (43)]; and, one article (51) reported an association between self-reported cannabis use in the first trimester and gastroschisis [OR: 4.5, 95% CI 2.1–9.8; (51)]. However, another article (52) did not find any significant associations between self-reported cannabis use in the month before pregnancy or during the first trimester and several specific gastrointestinal birth defects [Table 2; (51)], including gastroschisis [OR: 1.2, 95% CI: 0.9, 1.7 (52)].
One article reported on associations between in utero cannabis exposure and genitourinary malformations (52). This study failed to find an association between self-reported cannabis use in the month before pregnancy or during the first trimester and hypospadias [OR: 0.8, 95% CI: 0.5–1.2; (52)].
One article reported on association between in utero cannabis exposure and musculoskeletal malformations (52). The study failed to find associations between self-reported cannabis use in the month before pregnancy or during the first trimester and (a) craniosynostosis (OR: 0.8, 95% CI: 0.5–1.3) or (b) transverse limb deficiency [OR: 1.0, 95% CI: 0.6–1.7; (52)].
Associations between in utero cannabis exposure and specific orofacial malformations were reported on in two articles (43, 52). Both articles reported associations close to the null for each malformation [Table 2; (43, 52)]. Specifically, Van Gelder et al. reported associations close to the null for anotia/microtia (OR: 0.9, 95% CI: 0.5–1.7), cleft lip with or without cleft palate (OR: 1.0, 95% CI: 0.8–1.3), and cleft palate [OR: 1.0, 95% CI: 0.7–1.5; (52)]; and Bandoli et al. reported an association close to the null for oral cleft [RR: 1.1, 95% CI: 0.9, 1.5; (43)].
This systematic review found mixed and inconclusive associations between in utero cannabis exposure and risk for structural birth defects. Results were mixed among (a) the four articles reporting on adjusted associations with cardiac malformations (9, 43, 45, 52), (b) the three articles reporting on adjusted associations with central nervous system malformations (43, 50, 52), and (c) the three articles reporting on adjusted associations with gastrointestinal malformations (43, 51, 52). Some studies suggested in utero cannabis exposure was not associated with these types of birth defects; and, other articles suggesting that in utero cannabis exposure was associated with increased risk of these types of birth defects. Only two articles reported on adjusted associations with orofacial malformations (43, 52); and, only single articles reported on adjusted associations with eye malformation (43), genitourinary malformations (52), and musculoskeletal malformations (52). Though the articles reporting on associations with orofacial, eye, genitourinary, and musculoskeletal malformations all suggested that in utero cannabis exposure was not associated with these types of malformations (43, 52), strong conclusions cannot be drawn from these few studies that all had limitations.
There were several limitations of the included studies that may have contributed to the mixed findings on in utero cannabis exposure and birth defects. These limitations are similar to those of studies on in utero cannabis exposure and other outcomes, such as long-term neurodevelopmental and psychiatric problems (15). First, many of the studies had samples that were relatively small (e.g., 6 of the 20 studies had samples under 500) and reported findings with wide confidence intervals. Therefore, these studies had poor precision and likely were underpowered to detect associations that truly exist. Second, several studies (i.e., 16 of 20) utilized birth cohorts with births occurring more than 20 years ago, which could be problematic given increasing cannabis potency in recent years (55–57) and the proliferation of newer modes of administration (e.g., vaping, edibles) with potentially different risk profiles (58). Third, most studies utilized self-report data, which may underestimate cannabis exposure (59, 60). Therefore, these studies may have mistakenly classified exposed offspring as unexposed, reducing the likelihood of detecting a true association. Fourth, many studies did not address timing of exposure, which is particularly problematic when studying birth defects given that exposures early in pregnancy may be particularly risky for the development of major structural birth defects (61). Fifth, most studies did not assess associations with dose of cannabis exposure. This is a major limitation given that some research has supported a dose-response relationship between in utero cannabis exposure and birth defects (9), and research has shown dose-dependent associations between in utero cannabis exposure and other outcomes (8). Sixth, most studies did not adequately account for potential confounders, such as co-exposure to other substances. Despite the high co-occurrence of cannabis use and use of other substances, particularly tobacco and alcohol, in pregnancy 12 of the 20 studies did not take this into consideration. Therefore, observed associations between in utero cannabis exposure and birth defects could be attributable to exposure to a substance other than cannabis or could be explained by an interactive effect of cannabis use plus use of another substance (62, 63). We note that one study did find similar associations with and without limiting the sample to pregnancies with substance use disorder diagnoses other than cannabis-related diagnoses (43). Nonetheless, more research is needed to parse apart the effects of in utero cannabis exposure from exposure to other substances.
It is important to recognize that the mixed and inconclusive results on associations between in utero cannabis exposure and structural birth defects should not be interpreted as evidence suggesting cannabis use in pregnancy is safe. Rather these results indicate that the relationship between in utero cannabis exposure and structural birth defects is unknown and point to a critical need for future research. This need is particularly pressing given the documented increasing rates in prenatal cannabis use (1).
There are several important avenues for future research. First, samples should be sufficiently large to have adequate statistical power to identify associations if they truly exist. Second, studies with large sample sizes should evaluate associations with specific malformations within organ-specific malformation groups. Third, studies would benefit from including samples comprised of recent birth cohorts given changes in cannabis potency and modes of administration that have occurred in recent years. Fourth, utilizing biological measures (e.g., urine toxicology tests) in addition to self-reported cannabis use would reduce measurement error related to in utero cannabis exposure. Fifth, assessing the influence of timing of exposure and particularly focusing on first-trimester exposure is important. Sixth, it is also important for future studies to quantify the amount of prenatal cannabis exposure by considering the dose, frequency, potency, mode of administration and duration of use during pregnancy. Seventh, studies should utilize methods that rigorously evaluate the potential influence of confounding factors. Using conceptual models based on previous literature, researchers can identify potential factors that may confound associations between in utero cannabis exposure and birth defects. Researchers could also consider using advanced epidemiological methods that have been utilized to study other in utero exposures to help adjust for confounding factors, such as propensity scores, cannabis use before but not during pregnancy as a comparator, and comparisons of differentially exposed siblings [see Sujan et al. for a review of methods that have been used to study antidepressant medications during pregnancy (64)].
Importantly, no single study can implement all of these recommendations, particularly given common obstacles faced by researchers, such as funding limitations restricting the scope of studies, challenges enrolling participants, difficulty obtaining biological samples, and loss to follow-up. However, future research should try to incorporate as many of these recommendations as possible to reduce biases and maximize the overall quality of the studies. Rigorous, high-quality information on the potential consequences of in utero cannabis exposure is vital for individuals to make informed choices about cannabis use in pregnancy, as well as for families and providers caring for infants exposed to cannabis in utero.
This study was supported by grants R01DA047405 funded by National Institute on Drug Abuse (NIDA) and R01DA48033 funded by NIDA and Office of the Director, NIH (OD). The funding organizations had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication.
The original contributions presented in the study are included in the article. Further inquiries can be directed to the corresponding author.
All authors conceptualized the study. AS and AP conducted the literature review and extracted information from the reviewed studies. All authors interpreted the findings. AS drafted the manuscript, and all authors provided critical revisions of the manuscript. KY and LA supervised AS. AS and KY supervised AP. All authors contributed to the article and approved the submitted version.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
The original contributions presented in the study are included in the article. Further inquiries can be directed to the corresponding author.