Abstract
Background:
Research assessing the effects of marijuana use on preterm birth have found mixed results, in part, due to lack of attention to the role of maternal tobacco smoking during pregnancy.
Objectives:
The study objective was to investigate if maternal marijuana use was independently associated with gestational age, preterm birth and two preterm birth subtypes (spontaneous vs. clinician-initiated).
Methods:
Participants included 8261 mother-newborn pairs from the Boston Birth Cohort. Information on gestational age was collected from electronic medical records. Marijuana use and tobacco smoking during pregnancy were assessed through a standard questionnaire after birth. Linear and log-linear regression models were used to assess associations between marijuana use with and without tobacco smoking during pregnancy and the outcomes of interest.
Results:
Of the 8261 mothers, 27.5% had preterm births. About 3.5% of mothers with term deliveries and 5.2% of mothers with preterm births used marijuana during pregnancy. Marijuana use and cigarette smoking were independently associated with a decrease in gestational age by 0.50 weeks (95% CI −0.87, −0.13) and 0.52 weeks (95% CI −0.76, −0.28), respectively. Marijuana use during early or late pregnancy was associated with a similar decrease in gestational age by 0.50 weeks. When we examined the effects on the preterm birth subtypes, simultaneous marijuana use and tobacco smoking was associated with higher risk of spontaneous preterm birth (RR 1.64, 95% CI 1.23, 2.18). The elevated risk was not observed with clinician-initiated preterm birth.
Conclusions:
In this high-risk US population, maternal marijuana use and cigarette smoking during pregnancy were independently associated with shorter gestational age. When we examined the effects on preterm birth subtypes, the elevated risk was only observed with spontaneous preterm birth.
Keywords: cannabis, smoking, tobacco, preterm birth, minority health
Social media quote
Maternal marijuana use and cigarette smoking during pregnancy were independently associated with shorter gestational age. When we examined the effects on preterm birth subtypes, the elevated risk was only observed with spontaneous preterm birth.
Background
In the United States, the prevalence of marijuana use more than doubled between 2001–2002 to 2012–2013.1 From 2002 to 2014 the prevalence of marijuana use in pregnant women also rose from 2.4% to 3.9%.2 Since the main ingredient of marijuana, delta-9-tetrahydrocannabinol can cross the placenta,3–5 marijuana use by pregnant women potentially affects not only the mother but also the foetus.6, 7
Preterm birth increases the risk of mortality in the short term and is associated with chronic diseases in adulthood.8, 9 Research assessing the effects of marijuana use during pregnancy on preterm birth has yielded mixed results.3, 10, 11 One challenge is that a large proportion of marijuana users also smoke tobacco,12 which has well-documented associations with preterm birth.13 Given this relatively common combined risk profile, it is important to tease out the effects of the concomitant use of marijuana and tobacco. To date studies aiming to identify the effects of marijuana use during pregnancy on preterm birth have typically treated maternal smoking as a confounder. However, few have examined the joint and separate effects of tobacco smoking and marijuana, and those existing studies have been inconclusive,3, 11, 14 partly due to sample size constraints.
Using one of the largest US birth cohorts of predominantly urban low-income ethnic minorities, we sought to study the effects of marijuana during pregnancy on preterm birth. The large number of preterm births allowed us to further address whether marijuana use has differential effects on the two subtypes of preterm birth: spontaneous vs. clinician-initiated preterm birth.
Methods
Sample
Participants included 8261 mother-baby pairs from the Boston Birth Cohort (BBC). Between 1998 through 2018, the BBC enrolled mothers who delivered at the Boston Medical Center (BMC), as described previously.15, 16 The BBC was initiated to study environmental and genetic determinants of preterm birth, which was the primary outcome; subsequently, it was extended to study children’s postnatal health outcomes. The BBC enrolled preterm and low birthweight mother-infant pairs (cases) and term mother-infant pairs (controls) in a ~1:2 ratio within a few days postpartum. There were no specific matching criteria for non-cases, enabling us to study all potential risk factors. Women with multiple pregnancies, who became pregnant through in-vitro fertilization, or who had children with major congenital abnormalities were excluded. There were 2316 cases and 6193 term birth controls in the original sample. The final analytic sample included 8261 mother-baby pairs after excluding 248, due to missing data on gestational age, smoking or marijuana use, or major covariates (education and race). Questionnaires were administered to mothers at the time of enrolment and sociodemographic characteristics and substance use during pregnancy were assessed.
Birth outcomes
Information on gestational age was collected from electronic medical records. Preterm birth was defined as gestational age <37 weeks at delivery. Preterm birth was categorized as spontaneous preterm birth and clinician-initiated preterm birth. Spontaneous preterm birth was defined as a birth that occurred secondary to documented active preterm labour (uterine contractions with cervical effacement and dilation at <37 weeks), or premature rupture of the membranes at <37 weeks without uterine contractions, or both. Clinician-initiated preterm birth was defined as a birth delivered by medical induction or caesarean section at <37 weeks without uterine contractions or rupture of the membranes.
Exposures
Marijuana use during pregnancy
Mothers were asked about marijuana use during pregnancy which was defined as use during any trimester of gestation. We also investigated marijuana use by trimester given prior evidence of temporal associations, i.e. use in early- and late-pregnancy have been differentially associated with pregnancy outcomes.17–20 Early-pregnancy use was defined as reported marijuana use during the first trimester only. Late-pregnancy use was defined as reported marijuana use during the second trimester or third trimester (irrespective of reported first-trimester use).
Tobacco smoking
Mothers were asked about tobacco smoking status. Mothers who had smoked during pregnancy after the first trimester were considered continuous smokers, whereas mothers who had smoked during the first trimester only or before pregnancy were considered non-continuous smokers. Mothers who had never smoked tobacco were considered never smokers. Tobacco smoking status was dichotomized as continuous smoker vs. non-continuous smoker and never smoker.
Covariates
For covariate adjustment, maternal age and BMI were used as continuous variables, while parity, education, income, race/ethnicity, marital status, child sex, tobacco smoking during pregnancy, alcohol use during pregnancy and year of child’s birth were categorical variables. Maternal age, race/ethnicity, marital status, parity, education, income, and alcohol use during pregnancy were self-reported at time of enrolment (See Table 1). These covariates were chosen based on previous studies and by using a directed acyclic graph (DAG) (Figure 1).14, 21–23 Based on the DAG, even after adjustment for socio-demographic factors, we still needed to adjust for maternal age and race/ethnicity. After adjusting for socio-demographic factors, BMI was not associated with marijuana use but was still associated with the outcome. Child sex was not a confounder of the exposure-outcome relationship but was a variable associated with the outcome. Adjustment for variables associated with only the outcome will decrease variability of the outcome and lead to increased precision in analyses with linear regression and log-linear regression.23, 24 This was the case for BMI and child sex. In contrast, in logistic regression analyses, adjustment for variables associated with only the outcome will lead to decreased precision.23, 24 Thus, in the linear regression model and the log-linear regression model, covariate adjustment was made for maternal age, BMI, race/ethnicity, marital status, parity, education, income, alcohol use during pregnancy, child sex, and year of child’s birth. In the logistic regression model presented in supplementary tables, covariate adjustment was made for maternal age, race/ethnicity, marital status, parity, education, income, alcohol use during pregnancy, and year of child’s birth without adjustment for BMI and child sex.
Table 1.
Demographic characteristics of the study participants, stratified by term and preterm birth in the original unimputed dataset.
| Term N=5990 |
Preterm birth N=2271 |
Spontaneous N=1496 |
Clinician-initiated N=775 |
|
|---|---|---|---|---|
| N (%) | N (%) | |||
| Maternal age (Mean (IQR)) | 28.0 (22.9, 32.5) | 28.8 (23.3, 34.0) | 28.2 (22.6, 33.4) | 30.0 (24.6, 35.3) |
| BMI (Mean (IQR)) | 25.9 (21.6, 28.5) | 26.3 (21.5, 29.2) | 25.6 (21.3, 28.5) | 27.5 (22.4, 30.9) |
| Race/Ethnicity | ||||
| Black | 2763 (46.1%) | 1137 (50.1%) | 725 (48.5%) | 412 (53.2%) |
| White | 683 (11.4%) | 295 (13.0%) | 209 (14.0%) | 86 (11.1%) |
| Hispanic | 1785 (29.8%) | 573 (25.2%) | 391 (26.1%) | 182 (23.5%) |
| Other | 759 (12.7%) | 266 (11.7%) | 171 (11.4%) | 95 (12.3%) |
| Marital status | ||||
| Single | 3624 (60.5%) | 1470 (64.7%) | 1003 (67.0%) | 467 (60.3%) |
| Married | 2091 (34.9%) | 694 (30.6%) | 425 (28.4%) | 269 (34.7%) |
| Other (Widow, divorced, etc.) | 162 (2.7%) | 77 (3.4%) | 52 (3.5%) | 25 (3.2%) |
| Missing | 113 (1.9%) | 30 (1.3%) | 16 (1.1%) | 14 (1.8%) |
| Parity | ||||
| Nulliparous | 2589 (43.2%) | 972 (42.8%) | 651 (43.5%) | 321 (41.4%) |
| Multiparous | 3387 (56.5%) | 1295 (57.0%) | 842 (56.3%) | 453 (58.5%) |
| Missing | 14 (0.2%) | 4 (0.2%) | 3 (0.2%) | 1 (0.1%) |
| Education | ||||
| Completed high school | 1900 (31.7%) | 698 (30.7%) | 479 (32.0%) | 219 (28.3%) |
| Above high school | 4090 (68.3%) | 1573 (69.3%) | 1017 (68.0%) | 556 (71.7%) |
| Yearly Income | ||||
| <30,000 | 2652 (44.3%) | 1010 (44.5%) | 680 (45.5%) | 330 (42.6%) |
| >=30,000 | 2649 (44.2%) | 1098 (48.3%) | 703 (47.0%) | 395 (51.0%) |
| Missing | 689 (11.5%) | 163 (7.2%) | 113 (7.6%) | 50 (6.5%) |
| Alcohol (during pregnancy) | ||||
| Any | 493 (8.2%) | 229 (10.1%) | 150 (10.0%) | 79 (10.2%) |
| Never | 5316 (88.7%) | 1972 (86.8%) | 1303 (87.1%) | 669 (86.3%) |
| Missing | 181 (3.0%) | 70 (3.1%) | 43 (2.9%) | 27 (3.5%) |
| Smokinga | ||||
| non-continuous smoker | 880 (14.7%) | 389 (17.1%) | 255 (17.0%) | 134 (17.3%) |
| continuous smoker | 647 (10.8%) | 364 (16.0%) | 277 (18.5%) | 87 (11.2%) |
| Never smoker | 4463 (74.5%) | 1518 (66.8%) | 964 (64.4%) | 554 (71.5%) |
| Marijuana use during pregnancy | ||||
| Any | 209 (3.5%) | 119 (5.2%) | 86 (5.7%) | 33 (4.3%) |
| Never | 5781 (96.5%) | 2152 (94.8%) | 1410 (94.3%) | 742 (95.7%) |
| Child sex | ||||
| Male | 2975 (49.7%) | 1154 (50.8%) | 787 (52.6%) | 408 (52.6%) |
| Female | 3015 (50.3%) | 1117 (49.2%) | 709 (47.4%) | 367 (47.4%) |
| Child gestational age (Mean (IQR)) | 39.5 (38.6, 40.4) | 33.9 (32.6, 36.2) | 33.9 (32.7, 36.3) | 33.8 (32.3, 36.1) |
| Year of child’s birth (Mean (IQR)) | 2006 (2003, 2010) | 2006 (2003, 2010) | 2006 (2003, 2010) | 2007 (2004, 2010) |
Continuous smoker refers to a woman who smoked during pregnancy after the first trimester, whereas non-continuous smoker refers to a woman who had smoked during the first trimester only or before pregnancy. Never smoker was defined as a woman who had never smoked tobacco.
Figure 1.
Directed acyclic graph for the association between maternal marijuana use during pregnancy and preterm birth. Solid lines signify biasing paths whereas dashed lines signify risk factors for the outcome. Socio-demographic factors include education, income, marital status, and parity. Other measured and unmeasured confounders were: year of child’s birth (measured confounders), and maternal and paternal mental health (unmeasured confounders). BMI: body mass index.
Maternal Body Mass Index (BMI) was based on self-reported pre-pregnancy weight and height. BMI was calculated by dividing weight in kilograms by height in meters squared. Information on child sex was extracted from electronic medical records. To control for period effects, year of child’s birth was also included as a categorical variable grouped in five-year increments. This was done to allow for non-linear effects of time on the relationship of interest.
Statistical analysis
We fit linear regression and log-linear regression models with robust error variance25 to assess associations between exposures and outcomes. We also fit logistic regression models (See Supplementary Tables) to account for oversampling of preterm births and low birthweight births. Categorical risk profiles were defined to examine the combined use of marijuana and tobacco during pregnancy (using neither, only one, or using both substances) on outcomes. In addition, the relative excess risk due to each interaction (RERI) was also generated to describe departures from additive risks.
Missing data
To estimate missing values on covariates, we used multiple imputation. Variables included in the imputation included BMI, parity, income, gestational age, preterm birth, child sex, maternal age, smoking, alcohol use during pregnancy, race/ethnicity, marital status, number of PNC visits, pre-eclampsia, and self-reported stress (one year before the pregnancy and by trimester of pregnancy). Fifty imputed datasets were created using the “mice” package, version 2.46.0.26 The results using all imputed datasets were combined using Rubin’s rules for multiple imputation.27, 28
Sensitivity analyses
As a sensitivity analysis, we calculated E-values for all statistically significant associations in the log-linear, linear, and logistic models, using the “EValue” package, version 2.0.029 in R, version 3.5.030 and methods outlined in VanderWeele & Ding, 2017.29
Ethics approval
This study was approved by the Institutional Review Boards of Boston University Medical Center and Johns Hopkins Bloomberg School of Public Health (IRB Number: 00003960; date of approval: 2/27/2019). We obtained written informed consent from all participating pregnant women.
Results
Of the 8261 mothers enrolled, 72.5% had term deliveries, and 27.5% had preterm births. About 3.5% of mothers with term deliveries used marijuana during pregnancy, while 5.7% of mothers with spontaneous preterm births and 4.3% of mothers with clinician-initiated preterm births used marijuana during pregnancy (Table 1).
Table 2 shows associations between marijuana use during pregnancy and preterm birth without adjustment for smoking. We have also displayed E-values to estimate the impact of unmeasured confounding on the observed associations between exposures and outcomes (eTable1). After covariate adjustment, marijuana use was associated with 24% higher risk of preterm birth (95% CI 1.03, 1.50, E-value 1.79) and decreases in gestational age by 0.68 weeks (95% CI −1.04, −0.32, E-value 1.72) without adjustment for tobacco smoking. Regarding the timing of marijuana use during pregnancy, marijuana use during only the 1st trimester was associated with 10% higher risk of preterm birth (95% CI 0.84, 1.45), and was associated with a 0.59 decrease in gestational age in weeks (95% CI −1.08, −0.11, E-value 1.65) without adjustment for smoking. Use during the 2nd or 3rd trimesters was associated with 38% higher risk of preterm birth (95% CI 1.07, 1.77, E-value 2.10) and a decrease in gestational age by 0.76 weeks (95% CI −1.27, −0.26, E-value 1.79) without adjustment for smoking. Regarding marijuana use and preterm birth risk with adjustment for smoking (Table 2), marijuana use was associated with 15% higher risk of preterm birth (95% CI 0.95, 1.40) and a decrease in gestational age by 0.50 weeks (95% CI −0.87, −0.13, E-value 1.58) after adjusting for other covariates including tobacco smoking. When examining the timing of marijuana use during pregnancy, the use of marijuana only during 1st trimester was not associated with preterm birth, but was associated with a 0.49 decrease in gestational age in weeks (95% CI −0.98, −0.01, E-value 1.57) with adjustment for smoking. Marijuana use during the 2nd or 3rd trimester was associated with 24% higher risk of preterm birth (95% CI 0.96, 1.60), and a decrease in gestational age by 0.50 weeks (95% CI −1.02, 0.02) with adjustment for smoking. We obtained similar results when using a logistic regression model, presented in eTable 2.
Table 2.
Associations between self-reported marijuana use during pregnancy and risk of preterm birth without and with adjustment for tobacco smoking.
| Preterm birth | Gestational age (weeks) | |||
|---|---|---|---|---|
| RR (95% CI) | Coefficient (95% CI) | N | % Prevalence preterm birth |
|
| CRUDE ESTIMATES | ||||
| Marijuana (crude) | 1.34 (1.12, 1.61) | −0.83 (−1.18, −0.47) | 328 | 31.6% |
| Frequency of marijuana use | ||||
| None | 1.00 (Reference) | 0.00 (Reference) | 7933 | 27.1% |
| Only during 1st trimester | 1.17 (0.89, 1.53) | −0.71 (−1.19, −0.22) | 171 | 31.6% |
| During 2nd or 3rd trimester | 1.52 (1.19, 1.95) | −0.96 (−1.46, −0.45) | 157 | 41.4% |
| WITHOUT ADJUSTMENT FOR SMOKING | ||||
| Marijuana use | ||||
| Marijuanaa | 1.24 (1.03, 1.50) | −0.68 (−1.04, −0.32) | 328 | 31.6% |
| Frequency of marijuana use | ||||
| None | 1.00 (Reference) | 0.00 (Reference) | 7933 | 27.1% |
| Only during 1st trimestera | 1.10 (0.84, 1.45) | −0.59 (−1.08, −0.11) | 171 | 31.6% |
| During 2nd or 3rd trimestera | 1.38 (1.07, 1.77) | −0.76 (−1.27, −0.26) | 157 | 41.4% |
| WITH ADJUSTMENT FOR SMOKING | ||||
| Marijuana use | ||||
| Marijuanab | 1.15 (0.95, 1.40) | −0.50 (−0.87, −0.13) | 328 | 31.6% |
| Smokingb | 1.23 (1.08, 1.41) | −0.52 (−0.76, −0.28) | 1011 | 36.0% |
| Frequency of marijuana use | ||||
| None | 1.00 (Reference) | 0.00 (Reference) | 7933 | 27.1% |
| Only during 1st trimesterb | 1.06 (0.80, 1.39) | −0.49 (−0.98, −0.01) | 171 | 31.6% |
| During 2nd or 3rd trimesterb | 1.24 (0.96, 1.60) | −0.50 (−1.02, 0.02) | 157 | 41.4% |
Adjusted for maternal age, BMI, race/ethnicity, marital status, parity, education, income, alcohol use during pregnancy, child sex, and year of child’s birth.
Adjusted for maternal age, BMI, race/ethnicity, marital status, parity, education, income, alcohol use during pregnancy, tobacco smoking during pregnancy (reported in the table) child sex, and year of child’s birth.
In Table 3 and eTable 3, we present categorical risk profiles of using marijuana or tobacco smoking during pregnancy and their e-values, respectively. Marijuana use in early pregnancy without smoking was associated with 15% higher risk of preterm birth (95% CI 0.82, 1.63), and was associated with a decrease in gestational age by 0.65 weeks (95% CI −1.27, −0.02, E-value 1.69). In addition, smoking without marijuana use in early pregnancy was associated with 22% higher risk of preterm birth (95% CI 1.06, 1.39, E-value 1.74) and a decrease in gestational age by 0.49 weeks (95% CI −0.74, −0.23, E-value 1.56). RERIs for preterm birth and gestational weeks were −0.30 (95% CI −0.84, 0.24) and 0.41 (95% CI −0.10, 0.92), respectively, suggesting no additive interaction was observed. Regarding late pregnancy, marijuana use without smoking was associated with 22% higher risk of preterm birth (95% CI 0.76, 1.97), and was associated with a decrease in gestational age by 0.25 weeks (95% CI −1.14, 0.64). Smoking without marijuana use in late pregnancy was associated with 24% higher risk of preterm birth (95% CI 1.08, 1.42, E-value 1.79) and a decrease in gestational age by 0.50 weeks (95% CI −0.75, −0.25, E-value 1.57). Marijuana use with tobacco smoking was associated with higher risk of preterm birth (RR 1.42, 95% CI 1.14, 2.04, E-value 2.19), and a greater decrease in gestational age by 1.10 weeks (95% CI −1.72, −0.49, E-value 2.08), compared with no use of either substance. RERIs for preterm birth and gestational weeks were 0.08 (95% CI −0.51, 0.66) and −0.01 (95% CI −0.72, 0.69), respectively, suggesting that no additive interaction was observed. We obtained similar results from a logistic regression model, presented in eTable 4.
Table 3.
Assessment of categorical risk profiles of using marijuana or tobacco smoking during pregnancy and the adjusted association with preterm birth and gestational age in weeks.
| Preterm birth | Gestational Age (weeks) | |||
|---|---|---|---|---|
| Adjusted RR (95% CI) | Coefficient (95% CI) | N | Preterm birth prevalence | |
| No marijuana in early pregnancy, no smoking | 1.00 (Reference) | 0.00 (Reference) | 7094 | 26.2% |
| Marijuana in early pregnancy, no smoking | 1.15 (0.82, 1.63) | −0.65 (−1.27, −0.02) | 106 | 32.1% |
| No marijuana in early pregnancy, smoking | 1.22 (1.06, 1.39) | −0.49 (−0.74, −0.23) | 839 | 36.4% |
| Marijuana in early pregnancy, smoking | 1.08 (0.69, 1.69) | −0.67 (−1.44, 0.12) | 65 | 30.8% |
| Relative Excess Risk due to Interaction | −0.30 (−0.84, 0.24) | 0.41 (−0.10, 0.92) | ||
| Preterm birth | Gestational Age (weeks) | |||
| Adjusted RR (95% CI) | Coefficient (95% CI) | N | Preterm birth prevalence | |
| No marijuana in late pregnancy, no smoking | 1.00 (Reference) | 0.00 (Reference) | 7094 | 26.3% |
| Marijuana in late pregnancy, no smoking | 1.22 (0.76, 1.97) | −0.25 (−1.14, 0.64) | 50 | 34.0% |
| No marijuana in late pregnancy, smoking | 1.24 (1.08, 1.42) | −0.50 (−0.75, −0.25) | 839 | 35.0% |
| Marijuana in late pregnancy, smoking | 1.42 (1.14, 2.04) | −1.10 (−1.72, −0.49) | 107 | 45.9% |
| Relative Excess Risk due to Interaction | 0.08 (−0.51, 0.66) | −0.01 (−0.72, 0.69) |
All models were adjusted for maternal age, BMI, race/ethnicity, marital status, parity, education, income, alcohol use during pregnancy, child sex, and year of child’s birth.
In the upper half of the table, categorical risk profiles of using marijuana in early pregnancy (only the first trimester) or tobacco smoking during pregnancy (dichotomized as continuous smoker vs. non-continuous smoker and never smoker) were shown.
In the lower half of the table, categorical risk profiles of using marijuana in late pregnancy (either second or the third trimester) or tobacco smoking during pregnancy (dichotomized as continuous smoker vs. non-continuous smoker and never smoker) were shown.
Table 4 and eTable 5 show categorical risk profiles of using marijuana or tobacco smoking on spontaneous preterm birth and clinician-initiated preterm birth and their E-values. Smoking without marijuana use was associated with 45% higher risk of spontaneous preterm birth (95% CI 1.23, 1.71, E-value 2.26). Marijuana without smoking was associated with 21% higher risk of spontaneous preterm birth (95% CI 0.85, 1.72). Marijuana use with tobacco smoking was associated with the highest risk of spontaneous preterm birth (RR 1.64, 95% CI 1.23, 2.18, E-value 2.66). No associations between clinician-initiated preterm birth and tobacco smoking or/and marijuana use were observed. The results of the logistic regression model are also presented in eTable 6. Regarding the adjusted ORs for the two sub-types of preterm birth and RERI for clinician-initiated preterm birth, similar results were obtained for log-linear regression and with logistic regression. Regarding RERI for spontaneous preterm birth, the analysis with log-linear regression and logistic regression yielded similar small RERIs (for the log-linear regression model the RERI was −0.01 (95% CI −0.54, 0.52); for the logistic regression model the RERI was 0.04 (95% CI −0.77, 0.86)).
Table 4.
Categorical profiles of smoking and marijuana use reported with adjusted risk ratio for preterm birth, stratified by clinician-initiated pre-term birth, and spontaneous pre-term birth.
| Type of Preterm Birth | |||
|---|---|---|---|
| Adjusted RR (95% CI) | N | Preterm birth Prevalence | |
| Spontaneous preterm birth (N=1496) | |||
| No smoking, No marijuana | 1.00 (Reference) | 1185 | 16.7% |
| No smoking, Marijuana | 1.21 (0.85, 1.72) | 34 | 21.8% |
| Smoking, No marijuana | 1.45 (1.23, 1.71) | 225 | 26.8% |
| Smoking, Marijuana | 1.64 (1.23, 2.18) | 52 | 30.2% |
| Relative excess risk due to effect modification | −0.01 (−0.54, 0.52) | ||
| Clinician-initiated preterm birth (N=775) | |||
| No smoking, No marijuana | 1.00 (Reference) | 671 | 9.5% |
| No smoking, Marijuana | 1.10 (0.64, 1.91) | 17 | 10.9% |
| Smoking, No marijuana | 0.92 (0.70, 1.23) | 71 | 8.5% |
| Smoking, Marijuana | 1.01 (0.59, 1.73) | 16 | 9.3% |
| Relative excess risk due to effect modification | −0.11 (−0.85, 0.63) |
All models were adjusted for maternal age, BMI, race/ethnicity, marital status, parity, education, income, alcohol use during pregnancy, child sex, and year of child’s birth.
Comment
Principal findings
Marijuana use and tobacco smoking during pregnancy were associated with decreased gestational age at birth. We also found marijuana use to be associated with 24% increased likelihood of preterm birth, in models adjusted for covariates except for tobacco use. After additional adjustment for smoking, the association remained positive with a 15% higher likelihood of preterm birth. Regarding the timing of marijuana use in pregnancy, the risk ratio was higher for marijuana use during the 2nd and 3rd trimesters, however similar decreases by 0.50 gestational weeks were observed. Given the well-established association between tobacco smoking and preterm birth,13 importantly, the positive association between marijuana use and the decrease in gestational age remained after adjustment for simultaneous use of tobacco.
Strengths of the study
Study strengths include a large sample size and comprehensive clinical and epidemiological information. Few studies assessing the effect of marijuana use during pregnancy on preterm birth have distinguished between smokers and non-smokers and between clinician-initiated and spontaneous preterm birth. By analysing data from a relatively large cohort of primarily urban low-income US ethnic minorities, we found that even among non-smokers, marijuana use was associated with decreased gestational age.
Limitations of the data
There are several study limitations. First, tobacco smoking and marijuana use were based on maternal self-report after delivery, instead of biological indicators.31 Both differential and non-differential biases may co-exist. As a result of nondifferential measurement error, the associations of interest may have been biased toward the null. On the other hand, because information on substance use -- such as tobacco smoking and marijuana use -- was collected retrospectively, recall may have been influenced by birth outcomes, leading to differential misclassification. However, due to potential effects of social desirability on women’s responses, we expect that under-reporting would be more likely, leading to an underestimate of the true effect. Second, because we did not have information on the extent of cigarette usage, this could be subject to residual confounding. Given these limitations, more studies are needed to conclude that the effect of marijuana use on the risk of preterm delivery is causal. Third, since the BBC oversampled preterm and low birthweight deliveries in a predominantly urban low-income US sample, our findings may not be generalizable to other populations and thus, caution is needed when generalizing our findings.
Interpretation
Two systematic reviews and meta-analyses have reported mixed results regarding the association between marijuana use during pregnancy and preterm birth or gestational age when not adjusting for tobacco smoking.3, 10 One meta-analysis reported no association between marijuana use during pregnancy and preterm birth or gestational age after adjustment for tobacco smoking.3 The difference in results between our study and the meta-analysis could be due to study population characteristics.3 For example, US Blacks are at greater risk of preterm birth than non-Hispanic whites and several biological mechanisms have been implicated.32, 33 Exposure to chronic stress in US blacks may lead to HPA axis dysfunction, which plays a role in regulation and timing of delivery.33 In addition, bacterial vaginosis is a biological risk factor for preterm birth and may be a contributing factor for racial disparities in preterm birth risk since bacterial vaginosis is more prevalent in blacks compared to whites.33 Although more research is needed, marijuana may also affect the HPA axis and suppress immune functions that are necessary for clearance of infectious agents.34, 35 Thus, the biological effects of marijuana might increase the risk of preterm birth especially among high risk populations. Given inconsistencies in results and the different rates of preterm birth and risk factors for preterm birth among US Blacks,32 more studies are needed to evaluate reasons for different findings across subgroups.
Stratified analyses to calculate the risk of each combination of marijuana and tobacco use separately showed that marijuana use in late pregnancy without smoking was associated with 22% higher risk of preterm birth, and was associated with a decrease in gestational age by 0.25 weeks. The combined use of marijuana in late pregnancy with smoking was associated with 42% higher risk of preterm birth, and a decrease in gestational age by 1.10 weeks, compared with use of neither substance (RERI for preterm birth was 0.08, while RERI for gestational age was −0.01). Prior studies have also examined the effects of marijuana use on preterm birth while distinguishing between smokers and non-smokers. Saurel-Cubizolles et al. found that a positive relation between marijuana use and preterm birth was observed only among smokers, although the association was still positive among non-smokers.14 In contrast, a meta-analysis reported that marijuana use during pregnancy was not associated with increased risk of preterm birth among non-smokers.3 In the meta-analysis, when compared with mothers who used neither marijuana and tobacco, those who used both marijuana and tobacco had an increased risk.3 Authors suggested that the association between marijuana use during pregnancy and adverse birth outcomes including preterm birth may originate from the concomitant use of tobacco and other factors. However, their pooled analysis with stratification by tobacco smoking (i.e., a secondary analysis that showed that marijuana use during pregnancy was not associated with increased risk of preterm birth among non-smokers) was based on only two studies that were available for the secondary analysis, and the association between marijuana use and gestational age using stratification by tobacco smoking was not presented. Based on our findings that marijuana use both with and without smoking was associated with a decrease in gestational age, smokers and non-smokers may need to be encouraged to discontinue marijuana use during pregnancy.
We also examined differences in spontaneous and clinician-initiated preterm birth. While we observed no associations between marijuana use with/without tobacco smoking and clinician-initiated preterm birth, marijuana without tobacco smoking was associated with 21% higher risk of spontaneous preterm birth and marijuana use with tobacco smoking was associated with even higher risk (64%) of spontaneous preterm birth, though no additive interaction was observed, given the RERI was −0.01. Few studies have distinguished between spontaneous preterm birth and clinician-initiated preterm birth.14, 21, 36 Metz et al. reported that women who used marijuana had spontaneous preterm birth more frequently than those who did not use marijuana.21 They did not present the effects of marijuana use on clinician-initiated preterm birth.21 Saurel-Cubizolles et al. showed that marijuana use during pregnancy was associated with higher odds of spontaneous preterm birth.14 They also revealed that the association between marijuana use during pregnancy and spontaneous preterm birth was only detected among smokers, while among non-smokers, the association was less evident. Consistent with our study, they did not observe an association between clinician-initiated preterm birth and marijuana use during pregnancy.14 The distinction between spontaneous and clinician-initiated preterm birth is important since, although the two subtypes may share some underlying biological pathways, the clinical conditions that often necessitate medical induction (e.g., pre-eclampsia, foetal growth restriction) and the conditions that lead to spontaneous preterm birth (e.g. spontaneous labour onset) are different.37–39
For sensitivity analyses, E-values were calculated. The results suggest that if possible unmeasured confounders including maternal and paternal mental health are both associated with the exposure and the outcome at the level of the calculated E-values, adjustment for the unmeasured confounders would bias the reported results towards the null. Since this may not be implausible with the E-values we calculated, future studies are needed to confirm our findings.
In the light of the legalization of medical marijuana in many US states over the last two decades and more recently in the UK,12, 40 these findings are timely. Among US states that have legalized marijuana, none have listed pregnancy as a contraindication for medical use of marijuana.41 The American College of Obstetricians and Gynecologists42 and the American Academy of Pediatrics (AAP)41 have both expressed concerns regarding the current use of marijuana during pregnancy. Although they did not provide conclusions regarding the risk of marijuana use on preterm birth, the authors of an AAP report cited two systematic reviews and meta-analyses that showed no associations between marijuana use and preterm birth. Because our study showed associations between marijuana use and preterm birth in a relatively under-studied high-risk US population, more research may be needed before such conclusions are made.
Conclusions
Marijuana use both with and without tobacco smoking during pregnancy was associated with decreases in gestational age at birth. Further, the combination of smoking and marijuana use was associated with increased risk for spontaneous preterm birth. Given the increasing trend of marijuana use during pregnancy,2 future studies need to confirm our findings before conclusions can be made regarding the risk of marijuana use and preterm birth.
Supplementary Material
Synopsis.
Study questions:
What are the effects of marijuana during pregnancy on preterm birth, and on the two subtypes of preterm birth: spontaneous vs. clinician-initiated preterm birth?
What’s already known:
Studies assessing effects of marijuana use on preterm birth have found mixed results, in part, due to lack of attention to the role of maternal tobacco smoking during pregnancy.
What this study adds:
Within this high-risk US population, maternal marijuana use and cigarette smoking during pregnancy were independently associated with shorter gestational age. When we examined the effects on preterm birth subtypes, the elevated risk was only observed with spontaneous preterm birth.
Acknowledgements:
We thank the study participants for taking part in the study. We are also grateful for the dedication of the field team at the Department of Pediatrics, Boston University School of Medicine, and for the help of the obstetric nursing staff at Boston Medical Center. Dr. Hong is partially supported by Hopkins Population Center (NICHD R24HD042854). Dr. Ji is partially supported by the NIH grant R01HG006282.
Funding:
The Boston Birth Cohort (the parent study) is supported in part by the National Institutes of Health (NIH) under grants number R21HD085556, R21ES011666, R21HD066471, R01HD086013 and 2R01HD041702; and by the March of Dimes PERI under grants 20-FY02-56 and 21-FY07-605; and by the Health Resources and Services Administration (HRSA) of the U.S. Department of Health and Human Services (HHS) under grant number R40MC27443 and cooperative agreement UJ2MC31074. This information or content and conclusions are those of the authors and should not be construed as the official position or policy of, nor should any endorsements be inferred by NIH, HRSA, or HHS the U.S. Government. The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Footnotes
Conflicts of Interest: The authors have no conflicts of interest relevant to this article to disclose.
Reference
- 1.Hasin DS, Saha TD, Kerridge BT, et al. Prevalence of marijuana use disorders in the United States between 2001–2002 and 2012–2013. JAMA Psychiatry. 2015; 72:1235–1242. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Brown QL, Sarvet AL, Shmulewitz D, Martins SS, Wall MM, Hasin DS. Trends in marijuana use among pregnant and nonpregnant reproductive-aged women, 2002–2014. JAMA. 2017; 317:207–209. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Conner SN, Bedell V, Lipsey K, Macones GA, Cahill AG, Tuuli MG. Maternal marijuana use and adverse neonatal outcomes: a systematic review and meta-analysis. Obstet Gynecol. 2016; 128:713–723. [DOI] [PubMed] [Google Scholar]
- 4.Behnke M, Smith VC, Committee on Substance Abuse, Committee on Fetus and Newborn. Prenatal substance abuse: short- and long-term effects on the exposed fetus. Pediatrics. 2013; 131:e1009–1024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Bailey JR, Cunny HC, Paule MG, Slikker W Jr. Fetal disposition of delta 9-tetrahydrocannabinol (THC) during late pregnancy in the rhesus monkey. Toxicol Appl Pharmacol. 1987; 90:315–321. [DOI] [PubMed] [Google Scholar]
- 6.Zuckerman B, Frank DA, Hingson R, et al. Effects of maternal marijuana and cocaine use on fetal growth. N Engl J Med. 1989; 320:762–768. [DOI] [PubMed] [Google Scholar]
- 7.Hingson R, Alpert JJ, Day N, et al. Effects of maternal drinking and marijuana use on fetal growth and development. Pediatrics. 1982; 70:539–546. [PubMed] [Google Scholar]
- 8.Blencowe H, Cousens S, Oestergaard MZ, et al. National, regional, and worldwide estimates of preterm birth rates in the year 2010 with time trends since 1990 for selected countries: a systematic analysis and implications. Lancet. 2012; 379:2162–2172. [DOI] [PubMed] [Google Scholar]
- 9.D’Onofrio BM, Class QA, Rickert ME, Larsson H, Langstrom N, Lichtenstein P. Preterm birth and mortality and morbidity: a population-based quasi-experimental study. JAMA Psychiatry. 2013; 70:1231–1240. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Gunn JK, Rosales CB, Center KE, et al. Prenatal exposure to cannabis and maternal and child health outcomes: a systematic review and meta-analysis. BMJ Open. 2016; 6:e009986. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Metz TD, Stickrath EH. Marijuana use in pregnancy and lactation: a review of the evidence. Am J Obstet Gynecol. 2015; 213:761–778. [DOI] [PubMed] [Google Scholar]
- 12.Volkow ND, Compton WM, Wargo EM. The risks of marijuana use during pregnancy. JAMA. 2017; 317:129–130. [DOI] [PubMed] [Google Scholar]
- 13.Hammoud AO, Bujold E, Sorokin Y, Schild C, Krapp M, Baumann P. Smoking in pregnancy revisited: findings from a large population-based study. Am J Obstet Gynecol. 2005; 192:1856–1862; discussion 1862–1853. [DOI] [PubMed] [Google Scholar]
- 14.Saurel-Cubizolles MJ, Prunet C, Blondel B. Cannabis use during pregnancy in France in 2010. BJOG. 2014; 121:971–977. [DOI] [PubMed] [Google Scholar]
- 15.Wang X, Zuckerman B, Pearson C, et al. Maternal cigarette smoking, metabolic gene polymorphism, and infant birth weight. JAMA. 2002; 287:195–202. [DOI] [PubMed] [Google Scholar]
- 16.Wang G, Divall S, Radovick S, et al. Preterm birth and random plasma insulin levels at birth and in early childhood. JAMA. 2014; 311:587–596. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Richardson GA, Day NL, Goldschmidt L. Prenatal alcohol, marijuana, and tobacco use: infant mental and motor development. Neurotoxicol Teratol. 1995; 17:479–487. [DOI] [PubMed] [Google Scholar]
- 18.Day NL, Richardson GA, Goldschmidt L, et al. Effect of prenatal marijuana exposure on the cognitive development of offspring at age three. Neurotoxicol Teratol. 1994; 16:169–175. [DOI] [PubMed] [Google Scholar]
- 19.Day N, Sambamoorthi U, Taylor P, et al. Prenatal marijuana use and neonatal outcome. Neurotoxicol Teratol. 1991; 13:329–334. [DOI] [PubMed] [Google Scholar]
- 20.Tennes K, Avitable N, Blackard C, et al. Marijuana: prenatal and postnatal exposure in the human. NIDA Res Monogr. 1985; 59:48–60. [PubMed] [Google Scholar]
- 21.Metz TD, Allshouse AA, Hogue CJ, et al. Maternal marijuana use, adverse pregnancy outcomes, and neonatal morbidity. Am J Obstet Gynecol. 2017; 217:478 e471–478 e478. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Tronnes JN, Wood M, Lupattelli A, Ystrom E, Nordeng H. Prenatal paracetamol exposure and neurodevelopmental outcomes in preschool-aged children. Paediatr Perinat Epidemiol. 2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Bandoli G, Palmsten K, Flores KF, Chambers CD. Constructing Causal Diagrams for Common Perinatal Outcomes: Benefits, Limitations and Motivating Examples with Maternal Antidepressant Use in Pregnancy. Paediatr Perinat Epidemiol. 2016; 30:521–528. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Cologne J, Furukawa K, Grant EJ, Abbott RD. Effects of Omitting Non-confounding Predictors From General Relative-Risk Models for Binary Outcomes. J Epidemiol. 2019; 29:116–122. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Zou G A modified poisson regression approach to prospective studies with binary data. Am J Epidemiol. 2004; 159:702–706. [DOI] [PubMed] [Google Scholar]
- 26.van Buuren S, Groothuis-Oudshoorn K. mice: Multivariate Imputation by Chained Equations in R. Journal of Statistical Software. 2011; 45:1–67. [Google Scholar]
- 27.Sterne JA, White IR, Carlin JB, et al. Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls. BMJ. 2009; 338:b2393. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Rubin DB. Multiple imputation for nonresponse in surveys. New York. NY: John Wiley & Sons; 1987. [Google Scholar]
- 29.VanderWeele TJ, Ding P. Sensitivity Analysis in Observational Research: Introducing the E-Value. Ann Intern Med. 2017; 167:268–274. [DOI] [PubMed] [Google Scholar]
- 30.R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria: URL https://www.R-project.org/. 2017. [Google Scholar]
- 31.Hingson R, Zuckerman B, Amaro H, et al. Maternal marijuana use and neonatal outcome: uncertainty posed by self-reports. Am J Public Health. 1986; 76:667–669. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.McKinnon B, Yang S, Kramer MS, Bushnik T, Sheppard AJ, Kaufman JS. Comparison of black-white disparities in preterm birth between Canada and the United States. CMAJ. 2016; 188:E19–E26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Kramer MR, Hogue CR. What causes racial disparities in very preterm birth? A biosocial perspective. Epidemiol Rev. 2009; 31:84–98. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Cservenka A, Lahanas S, Dotson-Bossert J. Marijuana Use and Hypothalamic-Pituitary-Adrenal Axis Functioning in Humans. Front Psychiatry. 2018; 9:472. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Medicine NAoSEa. The Health Effects of Cannabis and Cannabinoids: The Current State of Evidence and Recommendations for Research. Washington, DC: The National Academies Press; 10.17226/24625. 2017. [DOI] [PubMed] [Google Scholar]
- 36.Dekker GA, Lee SY, North RA, McCowan LM, Simpson NA, Roberts CT. Risk factors for preterm birth in an international prospective cohort of nulliparous women. PLoS One. 2012; 7:e39154. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Savitz DA, Dole N, Herring AH, et al. Should spontaneous and medically indicated preterm births be separated for studying aetiology? Paediatr Perinat Epidemiol. 2005; 19:97–105. [DOI] [PubMed] [Google Scholar]
- 38.Ananth CV, Getahun D, Peltier MR, Salihu HM, Vintzileos AM. Recurrence of spontaneous versus medically indicated preterm birth. Am J Obstet Gynecol. 2006; 195:643–650. [DOI] [PubMed] [Google Scholar]
- 39.Ananth CV, Vintzileos AM. Epidemiology of preterm birth and its clinical subtypes. J Matern Fetal Neonatal Med. 2006; 19:773–782. [DOI] [PubMed] [Google Scholar]
- 40.Freeman TP, Hindocha C, Green SF, Bloomfield MAP. Medicinal use of cannabis based products and cannabinoids. BMJ. 2019; 365:l1141. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Ryan SA, Ammerman SD, O’Connor ME, Committee On Substance Use and Prevention, Section on Breastfeeding. Marijuana use during pregnancy and breastfeeding: implications for neonatal and childhood outcomes. Pediatrics. 2018; 142. [DOI] [PubMed] [Google Scholar]
- 42.Committee on Obstetric Practice. Committee opinion No. 722: marijuana use during pregnancy and lactation. Obstet Gynecol. 2017; 130:e205–e209. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.