By Matthew Brignall, ND
Link Community Clinic, Tacoma, WA
SYNOPSIS: Cannabis use disorder is associated with small but significant increased risks measured in multiple neonatal outcomes.
SOURCE: Shi Y, Zhu B, Liang D. The associations between prenatal cannabis use disorder and neonatal outcomes. Addiction 2021;116:3069-3079.
Cannabis use is common among young adults worldwide. There are an estimated 193 million cannabis users worldwide, and 10% of these have cannabis use disorder (CUD).1 However, the medical community has very little understanding of its safety during pregnancy. Given the longstanding and well-documented problems with alcohol, tobacco, and stimulant use during pregnancy, it should seem possible that cannabis use has important risks. Restrictions on research and incomplete reporting because of legal concerns have made studying this subject particularly difficult. It has been only in the past decade that credible research on the topic has begun to emerge.
This trial analyzed data from all singleton births in a hospital setting recorded in the state of California from Jan. 1, 2001, to Dec. 31, 2012. There were data available for 5.68 million live births during this period, representing 96% of live births during the period in California. After exclusions for multiple births, age younger than 9 years or older than 49 years, gestation longer than 44 weeks, and missing information on outcomes or covariates, a total of 4,830,239 mother-infant pairs were available for study. Infants were considered exposed if an International Classification of Diseases (ICD), Ninth Revision code for CUD appeared in the mother’s discharge records at the time of delivery.
CUD is a condition defined in the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) for the first time. At the time these data were collected, the DSM-4 included the current diagnostic criteria, and the condition was called “cannabis abuse” or “cannabis dependence.” The authors do not clarify how this affects the dataset.
Features of CUD include increasing use over time and/or tolerance to cannabis effect; unsuccessful attempts to quit; craving for cannabis; problems with work, school, home, or social activities from cannabis use; and continued use even with the knowledge that it causes or contributes to problems with physical or mental health outcomes.
Study outcome measures included gestation length, size at birth, admission to neonatal intensive care unit (NICU), and death or hospitalization in the first year of life. The researchers controlled for a number of important variables, including maternal age; education; race and ethnicity; insurance; maternal health history (including mental health); and the use of tobacco, alcohol, and opiates.
There were 20,237 mothers with documented CUD in the dataset. CUD more than doubled over the course of the study, from 2.8 to 6.9 per 1,000 births in 2001 and 2012, respectively. Babies born to mothers with CUD weighed 40 grams less than unexposed neonates (95% confidence interval [CI] = -50.34, -29.77), and the odds of low birth weight were higher with cannabis exposure (odds ratio [OR] = 1.13; 95% CI = 1.07, 1.20). CUD was associated with increased risk of preterm birth (OR = 1.06; 95% CI = 1.01, 1.12) and higher odds of death within a year of birth (OR = 1.35; 95% CI = 1.12, 1.62).
CUD was not associated with a significant difference in the chance of infants being admitted into neonatal intensive care. Exposed infants were less likely to be hospitalized within one year of birth (OR = 0.91; 95% CI = 0.86-0.96). The combination of prematurity and low birth weights implies more at-risk pregnancies, which is consistent with higher mortality in the first year. It is unclear why the hospitalization data point in the other direction, but it could be possible that CUD is associated with poor care access or some other non-health-related reason for this finding.
COMMENTARY
The distinction between CUD and occasional cannabis use is potentially important to this study. CUD is a DSM-5 diagnosis with complex criteria described previously. By limiting this study to people with CUD, the authors are possibly studying pregnant individuals with different patterns of use vs. earlier studies, which generally relied on self-reported use of any amount as their criteria for inclusion. The authors did not clarify how hospitals chose to screen for CUD, nor did they review whether this diagnosis is applied consistently across different socioeconomic or racial categories.
This study used propensity score matching, a statistical modeling methodology. This approach balanced the distribution of covariates in the exposed and the control groups to correct for variables that likely were not randomly distributed between the two populations. Proponents of this method claim that it can reduce confounding bias in this type of observational study.2
This California retrospective analysis follows a 2019 Canadian study that analyzed most of the same birth outcomes among 670,000 pregnancies that included more than 9,000 self-reported cannabis users.3 The 2019 study also showed significantly increased risk of low birth weight and preterm birth. Unlike the newer study, the Canadian study found a higher risk of NICU admission with cannabis use as well. The 2019 analysis did not include any longer-term follow-up on survival or developmental milestones. This earlier study had a rate of self-reported cannabis use of 1.4% in the cohort, which is much smaller than other estimates of use during pregnancy and calls into question the use of self-reporting as the screening metric.
The new California study only followed a year of developmental data, so it does not provide any information about intellectual development of these infants into school age. Studies on this relationship exist, but they are too small to be definitive.4-6 Each of these small studies shows a significant detrimental effect on at least one outcome variable. This study does not tell us anything important about intellectual development related to cannabis exposure.
There were some other problems with the data analysis in this trial. Exclusion for missing data was more common in infants exposed to CUD in utero vs. the control group in this study. The authors did not appear to take into account whether the use of medical marijuana — which was legal in California throughout the study — was recorded as CUD in medical charts, a question that may introduce more bias.
The safety of cannabis use during pregnancy likely is going to remain a very difficult subject to assess accurately. Taboos and legal restrictions make self-reporting a questionable way to screen participants. Using diagnostic codes may eliminate the concerns about inaccurate reporting, but it is not clear that the ICD-10 codes are used uniformly across the population. One previous meta-analysis found that the associations of prenatal cannabis exposure with low birth weight and preterm delivery were no longer significant when correcting for maternal tobacco use.7 Note that this 2016 meta-analysis was published before either of the larger cohort retrospective trials reviewed here were available. The California trial strongly suggests that the effects of heavy cannabis use are an independent risk factor from tobacco.
Given the uncertainty introduced by both the methods used in the large-scale Canadian and Californian trials, the fact that the results are congruent between the two models gives greater weight to either finding alone. At this time, it appears that maternal use of cannabis — at least if it is heavy use — is associated with small but significant increases in premature birth and low birth weights. The newer trial also suggests higher infant mortality in the first year after birth. As such, counseling pregnant women to avoid cannabis should be a part of routine prenatal care.
REFERENCES
- Connor JP, Stjepanovic D, Le Foll B, et al. Cannabis use and cannabis use disorder. Nat Rev Dis Primers 2021;7:16-68.
- Wang J. To use or not to use propensity score matching. Pharm Stat 2021;20:15-24.
- Corsi DJ, Walsh L, Weiss D, et al. Association between self-reported prenatal cannabis use and maternal, perinatal, and neonatal outcomes. JAMA 2019;322:145-152.
- Goldschmidt L, Richardson GA, Willford J, Day NL. Prenatal marijuana exposure and intelligence test performance at age 6. J Am Acad Child Adolesc Psychiatry 2008;47:254-263.
- Smith AM, Mioduszewski O, Hatchard T, et al. Prenatal marijuana exposure impacts executive functioning into young adulthood: An fMRI study. Neurotoxicol Teratol 2016;58:53-59.
- Moore BF, Salmons KA, Hoyt AT, et al. Associations between prenatal and postnatal exposure to cannabis with cognition and behavior at age 5 years: The Healthy Start Study. Int J Environ Res Public Health 2023;20:4880.
- Conner SN, Bedell V, Lipsey K, et al. Maternal marijuana use and adverse neonatal outcomes: A systematic review and meta-analysis. Obstet Gynecol 2016;128:713-723.
