Cannabis Use During Pregnancy Is Associated with the Suppression of Circulating Maternal Cytokines

Abstract

Introduction:

The prevalence of prenatal cannabis use has nearly doubled in the United States. Cannabinoid 2 receptors are predominately expressed in cells of the human immune system, and delta-9 tetrahydrocannabinol (THC), the primary active component of cannabis, has been shown to suppress immune responses. Despite these findings, there is very little evidence on the impact of cannabis use on maternal immune system. Here, we evaluate the association between urine-verified cannabis use and the levels of T helper cytokines in the maternal circulation.

Methods:

This was an ancillary study of a prospective cohort of pregnant women who participated in the Michigan Archive for Research on Child Health study. Pregnant women (age ≥18 years) were recruited from 22 prenatal clinics in Michigan and matched on age, race, and tobacco smoking (n = 144). The urinary metabolite of delta-9 THC, 11-nor-9-carboxy-delta-9-THC (THC-COOH), was used to define cannabis use status. A bead-based assay was used for the simultaneous detection of maternal cytokines associated with cannabis use and pregnancy outcomes in previous studies.

Results:

Repeated-measures linear mixed models indicated that urine-verified cannabis use was associated with the suppression of maternal pro-inflammatory cytokines including interferon gamma (β = −0.5; 95% confidence interval [CI] = −0.8, −0.1) and interleukin (IL)-12 (β = −0.3; 95% CI = −0.6, −0.05), as well as the anti-inflammatory IL-4 (β = −0.7; 95% CI = −1.3, −0.2) and IL-10 (β = −0.4; 95% CI = −0.7, −0.03). Similar results were observed when heavy cannabis use was defined using the top tertile of urinary THC-COOH at each trimester.

Conclusions:

Urine-verified cannabis use was associated with the suppression of pro- and anti-inflammatory T helper cytokines in a cohort of pregnant women, suggesting that cannabis use can lead to modest dysregulation of the maternal immune system. Additional studies are needed to investigate the role of maternal immune responses in explaining birth outcomes linked to cannabis use.

Introduction

From 2002 to 2019, the prevalence of prenatal cannabis use has nearly doubled in the United States, with 1 in 14 pregnant women using cannabis in 2021–2023.^1,2 Likewise, cannabis potency, measured by the concentration of the main psychoactive constituent, delta-9 tetrahydrocannabinol (THC), has increased from 9% in 2009 to 14% in 2019. ³ Despite the increase in cannabis use and potency, there is very little evidence on the impact of cannabis use on maternal health, including on the immune system. ⁴

Pregnancy is characterized by an altered immune profile to enhance adaptation to the semi-allogeneic fetus. ⁵ Exposures affecting the immune system can disrupt the balance between maternal defenses against pathogens and tolerance to the growing fetus and negatively impact maternal health and pregnancy outcomes.^6,7 To illustrate, maternal immune dysregulation was linked to several adverse outcomes in previous studies including preeclampsia, fetal growth restriction, and impaired offspring neurodevelopment.^8,9

Cannabis exerts its effects primarily via interacting with the cannabinoid receptors (i.e., CB1R and CB2R). ¹⁰ The CB2R is one of the most abundant G protein-coupled receptors in the human immune system. ¹¹ In previous studies, activation of CB2R was found to suppress immune responses in disease models of infection and inflammation. ¹² Delta-9 THC was also found to polarize immune responses toward the anti-inflammatory T helper (Th) 2 profile by suppressing the levels of circulating Th1 pro-inflammatory cytokines via CB2R-dependent and -independent pathways.^13–16

Given the biological plausibility linking cannabis use and the immune system, it is important to elucidate the effects of prenatal cannabis use on immune homeostasis. This is important as fetal and maternal outcomes of suspected immune mechanisms (e.g., preeclampsia and fetal growth restriction) were linked to prenatal cannabis use in previous studies.^17–20 Human studies on prenatal cannabis use and immunomodulation, however, are almost nonexistent. In a study of 300 mother–child dyads, self-reported maternal cannabis use was associated with the suppression of immune gene networks in placenta. ²¹ In another study of 138 first-trimester pregnant women, maternal inflammation measured by serum tumor necrosis factor alpha (TNFα) was lower among participants who co-used cannabis and tobacco compared with participants who used tobacco only, suggesting an impact of cannabis use on suppressing pro-inflammatory immune responses. ²²

Although these studies suggest that cannabis use in pregnancy alters immune responses, cannabis use often has been measured by self-report. The sensitivity of self-report for measuring cannabis use is low (34%), and many pregnant women will not disclose use owing to social stigma. ²³ Moreover, many studies used a single time point during pregnancy to assess cannabis use status. The prevalence of cannabis use is at its highest during the first trimester, likely because some pregnant women do not realize they are pregnant but declines significantly and steadily afterwards. ²⁴ Thus, a single time point measurement is likely insufficient to confirm cannabis use status throughout pregnancy. Therefore, the aim of the current study is to examine the relationship between cannabis use, measured by the levels of trimester-specific urinary THC metabolites, and circulating cytokines as an indicator of immune homeostasis, among pregnant women recruited and assessed for the prospective Michigan Archive for Research on Child Health (MARCH) study. We hypothesize that prenatal cannabis use decreases the ratio of Th1 to Th2 cytokines by suppressing the levels of circulating Th1 pro-inflammatory cytokines as observed in studies of non-pregnant individuals.

Methods

Study sample

This was an ancillary study of a prospective cohort of pregnant women who participated in MARCH, an ongoing population-based study of pregnant women recruited from 22 prenatal clinics in Michigan.^25,26 Exclusion criteria include age <18 years and the inability to be interviewed in English. During their first prenatal visit, information about MARCH was conveyed to pregnant women at any stage of pregnancy, and if they agreed to participate, they were consented and enrolled. Participants were administered one or two prenatal surveys in-person or by phone. In addition, up to two blood samples and three urine samples were collected from each participant. The MARCH study protocol was approved by the institutional review boards of Michigan State University (MSU IRB) and the participating sites. This ancillary study of de-identified data and biospecimens has been determined by MSU IRB to be exempt under 45 CFR 46.104(d) 4(ii).

To be selected for the current study, pregnant women enrolled in MARCH needed to complete at least one prenatal survey and had at least one urine and one blood sample stored (n = 573; 2017–2020). During both MARCH surveys, participants were asked if they used cannabis at any time during pregnancy. We identified cannabis users as participants who answered “yes” to this question at any of the two prenatal surveys (n = 89; 15.5%). We selected a comparison group of nonusers, that is, participants who answered “no” to the question on prenatal cannabis. To prevent misclassifying participants based on residual pre-pregnancy cannabis use, we excluded 164 individuals who reported use only in the three months before pregnancy. This is because THC metabolites can persist in urine for several weeks and could lead to false-positive results early in pregnancy. Compared with nonusers (n = 320), participants who used cannabis were more likely to be younger (26.7 vs. 30.0 years), non-White (63.8 vs. 39.3%), and active tobacco cigarette smokers (40.5 vs. 10.8%). We then used Stata “ccmatch” program to randomly match cannabis users and nonusers and created a subsample with complete data on covariates and balanced by age, race, and tobacco smoking (n = 78 users and 66 nonusers). ²⁷ The final sample (n = 144), based on the availability of survey data, urine samples, and blood samples, and matching results, was recruited from prenatal clinics affiliated with the following: Hurley Medical Center, Flint; Hutzel Women’s Hospital, Detroit; Beaumont Hospital, Dearborn; Sinai Grace Hospital, Detroit; St. John Hospital, Novi (categorized as Southeast Michigan); Covenant HealthCare, Saginaw; Munson Medical Center, Traverse City; Saint Joseph Mercy Hospital, Ann Arbor; and University of Michigan Hospital, Ann Arbor (categorized as otherwise).

Biospecimen-confirmed cannabis use

In MARCH, urine collection was matched to the timing of clinical collections that did not occur at the same gestational age for all participants. This strategy yielded a high frequency of sampling, with a mean of 2.5 urine samples per participant. Nearly all participants (n = 138; 96%) provided samples across two distinct trimesters, and a substantial subset (n = 64; 44%) contributed samples in all three trimesters. Urine samples were aliquoted and stored in −80 $\hspace{0.17em}\mathrm{°C}$ freezers. In the current study, urine samples were used to screen for and quantify THC metabolites. After cannabis use, THC is metabolized to the inactive compound 11-nor-9-carboxy-delta-9-THC (THC-COOH) and excreted in urine. ²⁸ Urinary THC-COOH is very stable, and cannabis use can be detected for several weeks. ²⁸ All urine samples were screened for THC-COOH using qualitative enzyme-linked immunosorbent assay (Neogen; I-50 sensitivity = 0.5 ng/mL). Positively screened samples were then quantified for THC-COOH and creatinine (to control for urine dilution; expressed in ng/mg creatinine) using liquid chromatography with tandem mass spectrometry (LC-MS/MS), a sensitive and specific technique that allows analysis of both major and minor cannabinoids at a low lower limit of quantitation, and it is the method of choice for the analysis of cannabinoids. ²⁹ Analysis was performed at MSU Mass Spectrometry Core, an open-access resource where researchers perform compound identification or quantitative analysis using 12 mass spectrometers.

Cytokines

Blood samples were collected during the first prenatal visit and glucose tolerance test (starting the second trimester) to minimize participant burden and integrate with clinical protocols. This approach yielded a mean of 1.5 samples per participant. Most participants (n = 131; 91%) provided samples in two distinct trimesters, ensuring longitudinal assessment. Only 13 participants had both samples collected within the same trimester (mean interval: 6 weeks). Peripheral blood was drawn in EDTA vacutainer tubes and processed to obtain plasma, which was then aliquoted and stored in −80°C freezers. To measure circulating cytokine levels, a bead-based high-sensitivity human cytokine kit (Millipore) was used, which enabled simultaneous detection of multiple analytes in a single reaction. Data were obtained using the Luminex® platform at MSU Flow Cytometry Core Facility.

The levels of several cytokines associated with cannabis use and pregnancy outcomes in previous studies (Table 1) were measured, including granulocyte-macrophage colony-stimulating factor (GM-CSF), interferon gamma (IFNγ), interleukin (IL) 1β, IL2, IL4, IL6, IL7, IL8, IL10, IL12 (p70), IL13, and TNFα. The percentage of plasma samples in which a given cytokine level fell below the lower limit of quantification was small (≤5%).

Table 1.

Prior Evidence Linking Circulating Cytokines, Pregnancy Outcomes, and Cannabis Use

Cytokine	Role	Pregnancy	Cannabis
GM-CSF	A glycoprotein that stimulates monocytes and macrophages to produce pro-inflammatory cytokines 30	Blood levels of GM-CSF decreased in spontaneous abortion31,32 and increased in preterm birth and preeclampsia.33,34	GM-CSF-expressing T cells were inhibited in peripheral blood of healthy volunteers after THC treatment. 35
IFNγ	Potent pro-inflammatory Th1 cytokine 36	Mid-pregnancy maternal blood levels of IFNγ increased in preterm birth. 37 Blood levels also increased in preeclampsia. 38	THC reduced levels of IFNγ in preclinical models of Legionella pneumophila infection. 39
IL1β	Pro-inflammatory cytokine 40	IL1β increased at the maternal-fetal interface in spontaneous abortion 41 and preeclampsia. 42	THC reduced levels of IL1β in rodent models of rheumatoid arthritis. 43
IL2	Th1 cytokine that has both pro- and anti-inflammatory functions 44	Infusion of IL2 decreased maternal blood pressure but reduced fetal weight and survival.45,46	THC suppressed IL2-expressing T cells as well as several important functions of IL2.35,47
IL4	Anti-inflammatory Th2 cytokine that is often co-expressed with IL13 48	Low circulating levels of IL4 have been linked with spontaneous abortions, preeclampsia, and preterm birth.49,50	THC induced a modest nonsignificant increase in the release of IL4 from peripheral blood mononuclear cells of healthy human donors. 14 Conversely, a decrease in plasma levels of IL4 was observed in chronic cannabis users. 51
IL6	A cytokine that has both pro- and anti-inflammatory functions 52	Elevated IL6 levels in mid-pregnancy were associated with preterm birth. 37 Elevated maternal IL6 levels were also associated with preeclampsia. 53	Lifetime but not recent cannabis users had lower IL6 levels than nonusers.54,55
IL7	Pro-inflammatory cytokine 56	IL7 overexpression in the maternal-fetal interface contributed to recurrent spontaneous abortion. 57	In a model of inflammatory autoimmune pathologies, THC induced a dose-dependent suppression in IL7 production and secretion. 58
IL8	Pro-inflammatory cytokine 59	Levels of IL8 significantly increased in preterm birth and preeclampsia.60,61	In in vitro studies of cytokine storm, administration of THC diminished levels of IL8. 62 Patients with cannabis use disorder, however, had increased serum levels of IL8. 63
IL10	Anti-inflammatory cytokine that is secreted by both Th1 and Th2 cells 64	Circulating levels of IL10 were reduced in pregnant women at the time of active preeclampsia. 65	A modest attenuation of IL10 production after THC administration was detected in in vitro studies of cytokine storm. 62
IL12	Pro-inflammatory cytokine that induces IFNγ production 66	Circulating levels of IL12 increased in preeclampsia and preterm birth. 67–69	THC administration reduced IL12 levels in preclinical models, but no changes were detected in patients with multiple sclerosis.39,70
IL13	Anti-inflammatory Th2 cytokine 71	Elevated gestational IL13 was associated with asthma and behavioral sequelae in the offspring.72,73	THC induced a modest reduction of IL13 production in a rodent model of inflammation. 74 Cannabis smoke exposure also reduced IL13 levels in a rodent model of influenza infection. 75
TNFα	Pro-inflammatory Th1 cytokine 76	Elevated maternal blood levels of TNFα were associated with an increased risk of preterm birth and preeclamsia.34,77	Administration of THC decreased levels of TNFα in a rodent model of staphylococcal infection. 78 Cannabis use was also associated with lower TNFα levels than nonuse.79,80

GM-CSF, granulocyte-macrophage colony-stimulating factor; IFNγ, interferon gamma; IL, interleukin; THC, tetrahydrocannabinol; Th, T helper; TNFα, tumor necrosis factor alpha.

Other variables

Potential confounding variables were selected based on previous prenatal cannabis use research. ⁸¹ Data on age, race, education, cigarette smoking, alcohol drinking, pre-pregnancy body mass index (BMI; self-reported), prenatal use of pain relief medications (prescription opioids and/or over-the-counter medications), and history of medically diagnosed depression during pregnancy were collected during the prenatal surveys. Age and BMI were analyzed as continuous variables. Race, a variable that is associated with both cannabis use and variations in cytokine levels in previous studies likely due to socioeconomic and lifestyle factors, was categorized into White and all others. Education was categorized into high school or less and above high school. Cigarette smoking, alcohol drinking, and pain medication use were categorized into yes (any use during pregnancy) and no (former and never users).

Statistical analysis

Descriptive statistics were used to characterize the study sample by survey-reported cannabis use status and by urinary THC-COOH levels. Mean, standard error, and absolute and relative frequencies are reported. To test for differences between groups, Student’s t, χ²), and analysis of variance (ANOVA) tests were used when applicable. A p- value of 0.10 was used to assess matching and estimate the differences between the two groups.

Repeated-measures linear mixed models (Stata command mixed) were used to compare cytokine levels by cannabis use status. Random effects at the subject level were used to account for the correlations between repeated measurements within each participant. In the current study, participants had different numbers of available measurements. Compared with repeated-measures ANOVA, mixed models can accommodate unbalanced data patterns and include all available data in the analysis. ⁸² All models included cannabis use status, trimester of pregnancy (categorical), and cannabis use status × trimester interactions. The appropriate contrasts from the model were used to calculate the overall pregnancy, as well as the trimester-specific estimates. Crude models are presented first, and then models (adjusted model 1) were adjusted for variables that were significantly different between the two groups after matching (race [White/all others], education [high school or less/above high school], and prenatal tobacco smoking [yes/no]). Adjusted model 2 was additionally adjusted for age (years), pre-pregnancy BMI (kg/m²), prenatal alcohol drinking (yes/no), recruitment site (Southeast Michigan/otherwise), pain medication use (yes/no), depression (yes/no), and urinary tract infection (UTI; yes/no). Stata software (v15; StataCorp) was used for the analysis, and p values <0.05 were considered statistically significant for the regression models. A Bonferroni adjustment was not done, given that the cytokine outcomes were a priori selected based on the previous research.

Results

Table 2 Panel A presents the characteristics of the study sample by survey-reported cannabis use status. After matching, there were no differences between the two groups by age, race, education, cigarette smoking, alcohol drinking, pre-pregnancy BMI, or recruitment site (p > 0.10).

Table 2.

Characteristics of the Study Sample by Prenatal Cannabis Use Status (n = 144)

	Mean ± SE or n (column %)
	Yes (n = 78)	No (n = 66)	p Value a
Panel A: Survey-reported cannabis use status
Age (years)	26.6 ± 0.6	26.6 ± 0.7	0.99
Race			0.50
White	30 (38.5)	29 (43.9)
All others	48 (61.5)	37 (56.1)
Education			0.33
High school or less	43 (55.1)	31 (47.0)
Above high school	35 (44.9)	35 (53.0)
Cigarette smokingb
Yes	30 (38.5)	17 (25.8)	0.11
No	48 (61.5)	49 (74.2)
Alcohol drinkingc			0.68
Yes	10 (12.8)	7 (10.6)
No	68 (87.2)	59 (89.4)
Pre-pregnancy BMI (kg/m2)	29.2 ± 1.0	27.9 ± 0.9	0.38
Recruitment site			0.55
Southeast Michigan	51 (65.4)	40 (60.6)
Otherwise	27 (34.6)	26 (39.4)

^aStudent’s t test for continuous variables and χ² test for categorical variables.

^bCigarette smoking and alcohol drinking were categorized into yes (any use during pregnancy) and no (former and never users).

^cUrine-verified cannabis use was defined as THC-COOH levels ≥15 ng/mg creatinine in one or more urine samples.

BMI, body mass index; SE, standard error; THC-COOH, 11-nor-9-carboxy-delta-9-delta tetrahydrocannabinol.

Out of 66 participants who did not report cannabis use before or during pregnancy, only 4 participants (6.1%) had significant levels of THC-COOH (≥15 ng/mg creatinine) in 1 or more urine samples, ranging from 68.2 to 959.47 ng/mg creatinine, indicating active use. Eighteen participants reported cannabis use during pregnancy but had THC-COOH levels <15 ng/mg creatinine in all collected urine samples. After accounting for the urinalysis results, there were no differences between the two groups by age, alcohol drinking, pre-pregnancy BMI, or recruitment site (p > 0.10), yet participants with urinary THC-COOH ≥15 ng/mg creatinine were more likely to be non-White, smoke during pregnancy, and have lower educational attainment than participants with urinary THC-COOH <15 ng/mg creatinine (Table 2 Panel B). There were no statistically significant differences in THC-COOH concentrations by pregnancy trimester (Fig. 1).

FIG. 1.

Urinary THC-COOH concentrations across trimesters of pregnancy. Mean urinary THC-COOH concentrations (ng/mg creatinine) quantified by liquid chromatography-tandem mass spectrometry across trimesters of pregnancy in the MARCH study. Error bars represent the standard error of the mean. No significant differences were found between trimesters using a one-way ANOVA, F(2, 322) = 0.6; p = 0.6. ANOVA, analysis of variance; MARCH, Michigan Archive for Research on Child Health; THC-COOH, 11-nor-9-carboxy-delta-9-delta tetrahydrocannabinol.

To account for the dependence of repeated-measure, multiple mixed models were used to regress cytokine levels on cannabis use status. Survey-reported cannabis use was not associated with any of the cytokines studied here. However, when LC-MS/MS was used to define cannabis use status (Table 3), urinary THC-COOH ≥15 ng/mg creatinine in one or more samples was inversely associated with several cytokines, regardless of their inflammation profile. These associations were statistically significant for the pro-inflammatory IFNγ and IL12, the anti-inflammatory IL4 and IL10 cytokines, in addition to IL2. When levels of IFNγ, TNFα, and IL2 cytokines were used to define Th1 responses and levels of IL4, IL10, and IL13 cytokines were used to define Th2 responses, the ratio of Th1/Th2 cytokines was positively associated with urinary THC-COOH levels indicating greater reduction of anti-inflammatory cytokine levels in women who used cannabis during pregnancy (β = 0.5; 95% CI = 0.1, 1.0). Adjustment for age, BMI, alcohol drinking, and recruitment site, however, attenuated the association between cannabis use and Th1/Th2 ratio toward the null (β = 0.4; 95% CI = 0.0, 0.8).

Table 3.

Results from Repeated-Measures Linear Mixed Models Estimating the Relationship Between Urine-Verified Cannabis Use (Defined as THC-COOH ≥15 ng/mg Creatinine in One or More Samples) and Maternal Plasma Cytokines

	β (95% CI)
Cytokine (n of participants)	Crude model b	Adjusted model 1 c	Adjusted model 2 d
GM-CSF a (n = 140)	−0.2 (−0.5, 0.2)	−0.1 (−0.5, 0.2)	−0.3 (−0.6, 0.1)
IFNγa (n = 143)	−0.5 (−0.8, −0.2)	−0.4 (−0.7, −0.1)	−0.5 (−0.8, −0.1)
IL1βa (n = 143)	−0.2 (−0.4, 0.0)	−0.2 (−0.4, 0.0)	−0.2 (−0.5, 0.0)
IL2a (n = 143)	−0.2 (−0.5, 0.0)	−0.2 (−0.5, 0.0)	−0.3 (−0.6, −0.02)
IL4 a (n = 140)	−0.8 (−1.4, −0.3)	−0.7 (−1.3, −0.2)	−0.7 (−1.3, −0.1)
IL6 a (n = 143)	−0.8 (−1.4, −0.2)	−0.7 (−1.3, −0.04)	−0.5 (−1.2, 0.2)
IL7 a (n = 143)	−0.2 (−0.4, 0.1)	−0.2 (−0.4, 0.1)	−0.2 (−0.5, 0.0)
IL8a (n = 143)	−0.2 (−0.9, 0.4)	−0.1 (−0.8, 0.5)	0.1 (−0.6, 0.7)
IL10a (n = 143)	−0.4 (−0.8, −0.1)	−0.3 (−0.6, 0.0)	−0.4 (−0.7, −0.03)
IL12a (n = 143)	−0.4 (−0.7, −0.1)	−0.3 (−0.6, −0.01)	−0.4 (−0.7, −0.1)
IL13a (n = 141)	−0.9 (−1.6, −0.2)	−0.6 (−1.3, 0.1)	−0.6 (−1.3, 0.1)
TNFα a (n = 143)	0.0 (−0.1, 0.1)	0.0 (−0.1, 0.1)	0.0 (−0.1, 0.1)
Th1/Th2 ratioa,e (n = 136)	0.5 (0.1, 1.0)	0.5 (0.01, 0.9)	0.4 (0.1, 0.8)

Data are from the MARCH study.

Bold estimates indicate statistical significance at p < 0.05 level.

^aThe outcome is log-transformed due to its skewed distribution.

^bModel 1 is unadjusted and includes cannabis use status defined as THC-COOH ≥15 ng/mg creatinine in one or more urine samples and trimester of pregnancy as fixed covariates, as well as cannabis use status × trimester interaction. Random intercepts for pregnant women accounted for the dependence of repeated measures.

^cModel is additionally adjusted for variables that were significantly different between the two groups after matching (i.e., race [White/all others], education [high school or less/above high school], and prenatal tobacco smoking [yes/no]).

^dModel 3 is additionally adjusted for age (years), pre-pregnancy BMI (kg/m²), prenatal alcohol drinking (yes/no), recruitment site (Southeast Michigan/otherwise), pain relief medication use (yes/no), history of depression (yes/no), and history of urinary tract infection (yes/no).

^eTh1/Th2 ratio is calculated as (IFNγ + TNFα + IL2)/(IL4 + IL10 + IL13).

CI, confidence interval; GM-CSF, granulocyte-macrophage colony-stimulating factor; IFNγ, interferon gamma; IL, interleukin; TNFα, tumor necrosis factor alpha; MARCH, Michigan Archive for Research on Child Health; THC-COOH, 11-nor-9-carboxy-delta-9-delta tetrahydrocannabinol.

Given the lack of a validated urinary THC-COOH cutoff for heavy cannabis use, particularly in pregnant populations with known increases in renal clearance, we employed a statistical approach, classifying participants in the top tertile of THC-COOH concentration at each trimester as heavy users (Fig. 2). Results indicated that urinary THC-COOH concentration in the top tertile was inversely associated with several cytokines, including IFNγ, IL2, IL6, IL7, IL12, and IL13. These associations did not differ by the trimester of pregnancy. No association was detected between urinary THC-COOH concentrations in the highest tertile and Th1/TH2 ratio.

FIG. 2.

Association between prenatal cannabis use (highest urinary THC-COOH tertile) and maternal cytokine levels across pregnancy. Parameter estimates from repeated-measures linear mixed models examining the relationship between urine-verified cannabis use (defined as the highest tertile of urinary THC-COOH concentration at each trimester) and maternal plasma cytokine levels. Data are from the MARCH cohort study. Outcomes (i.e., cytokines) were log-transformed to adjust for skewed distributions. Models were adjusted for trimester, a cannabis use × trimester interaction term, age, race, education, prenatal tobacco use, pre-pregnancy BMI, prenatal alcohol use, recruitment sites, pain medications, depression, and prenatal urinary tract infection. Estimates in red indicate a statistically significant association (p < 0.05). BMI, body mass index; MARCH, Michigan Archive for Research on Child Health; THC-COOH, 11-nor-9-carboxy-delta-9-delta tetrahydrocannabinol.

Discussion

Using data from the prospectively conducted MARCH study, we found evidence that urine-verified maternal cannabis use is associated with the suppression of several circulating cytokines during pregnancy. This suppression was observed for both pro-inflammatory Th1 and anti-inflammatory Th2 cytokines. The association between cannabis use and maternal cytokines did not differ by the trimester of pregnancy and was independent of sociodemographic characteristics, cigarette smoking, alcohol drinking, use of pain medications, and history of depression or prenatal UTI. We did not observe similar associations when cannabis use was measured via self-report, highlighting the importance of objectively measuring cannabis use to identify active consumption and false reporting.

Despite adjusting for conditions that frequently motivate cannabis use (e.g., pain relief and depression), it remains possible that reverse causation or unmeasured confounders (e.g., poor sleep and anxiety) contributes to the observed associations. However, these conditions are often linked to a pro-inflammatory state, characterized by elevated levels of Th1 cytokines. Our findings, however, demonstrate reductions in these cytokines among cannabis users. This opposing direction makes it unlikely that reverse causation fully explains our results. Future studies with repeated measures of cannabis use and immune markers beginning pre-conception are needed to fully disentangle these temporal relationships.

There are growing concerns regarding the health consequences of maternal cannabis use. Accumulating evidence indicates a potential link between prenatal cannabis exposure and low birthweight, small for gestational age, and preterm birth. ⁸³ In addition, the incidence of childhood neurodevelopmental disorders was found to be higher in pregnancies exposed to cannabis than in unexposed pregnancies. ⁸⁴ To better understand the mechanistic pathways via which maternal cannabis use impacts birth outcomes, Rompala et al. processed placental specimens collected at birth for RNA sequencing and found that in utero exposure to cannabis was associated with greater anxiety, aggression, and hyperactivity in offspring that correlated with reduced expression of many placental genes involved in immune functions including type I interferon, neutrophil, and cytokine signaling. ²¹ These findings highlight the importance of investigating the immunomodulatory effects of cannabis use in pregnant women, as aberrant maternal immune activity is implicated in a plethora of birth outcomes and child neurodevelopment. ⁸⁵

In vitro and rodent studies have reported that the administration of THC induced a greater suppression of Th1 cytokines in models of infection and inflammation, thus polarizing the immune responses toward the anti-inflammatory Th2 profile. ¹⁴ Some laboratory studies have also indicated that the administration of THC can further increase the levels of Th2 cytokines, including IL4 and IL10, prompting the call to investigate the use of cannabinoids as anti-inflammatory agents. ⁵¹ Yet, results from human studies do not support the notion that cannabis use increases the levels of anti-inflammatory cytokines. For example, Luo et al. have documented lower levels of plasma IL4 cytokines in chronic cannabis users than nonusers. ⁸⁶ Similarly, we found that urine-verified cannabis use was associated with a reduction of both Th1 and Th2 cytokines in maternal plasma and not solely Th1 cytokines. It is yet to be determined how these reductions translate to clinical outcomes, especially as they relate to infections and conditions with inflammatory pathogenesis.

Strengths of the MARCH study include the prospective design and the collection of urine and blood samples at multiple timepoints throughout pregnancy. These strengths allowed us to screen for THC metabolites in urine samples collected at different trimesters compared with previous studies that relied solely on self-report or used a single specimen collection to confirm prenatal cannabis use status. In addition, we quantified THC-COOH and estimated its association with circulating maternal cytokines, an important research question that was not investigated in prior human studies.

Weakness of the study includes the relatively small sample size that did not allow for subgroup analysis and the lack of in-depth survey data on cannabis use including the timing and frequency of use as well as the method of administration. Several participants reported cannabis use during pregnancy but had insignificant levels of THC-COOH in all collected urine samples. It is possible that these participants used cannabis early in pregnancy but did not continue use after pregnancy recognition. The lack of data on the timing of prenatal cannabis use limited testing of this hypothesis. Also, most prenatal cannabis use is a continuation of pre-pregnancy behavior, and initiation of cannabis use during pregnancy was found to be very rare. ²⁴ This is also evidenced in our cohort as all, but two users reported pre-pregnancy use. Although cannabis exposure was defined in our study as use preceding the blood draw (confirmed by survey reporting and urine toxicology sensitive to ∼4 weeks), we cannot pinpoint use in the days/hours immediately prior to the blood draws. This is important given the short half-life of most cytokines measured. It is thus possible that our results underestimate the acute effects of cannabis use on cytokine levels.

Different methods of cannabis administration can also modulate the impact of cannabis use on immune functions. Cannabis smoking is the most frequent method of administration, whereas vaping and edibles are less frequent but increasing in popularity. ⁸⁷ Compared with smoking, cannabis vaping and edibles may limit exposure to toxic chemicals generated by combustion and implicated in inflammation and immune dysregulation. ⁸⁸

Conclusions

Urine-verified cannabis use was independently associated with modest suppression of circulating levels of Th1 and Th2 cytokines in a cohort of pregnant women. Future research is needed to understand the clinical implication of maternal cytokine suppression and how it relates to birth outcomes linked to cannabis use including low birth weight, preterm birth, and child neurodevelopment.

Authors’ Contributions

O.A., V.P., and L.K.O.: Conceptualization and methodology. M.S. and E.K.: Biospecimen analysis and data curation. All authors: Writing—review and editing.

Abbreviations

BMI

Body mass index

CB1R

Cannabinoid 1 receptor

CB2R

Cannabinoid 2 receptor

GM-CSF

Granulocyte-macrophage colony-stimulating factor

IFNγ

Interferon gamma

IL

Interleukin

LC-MS/MS

Liquid chromatography with tandem mass spectrometry

MARCH

Michigan Archive for Research on Child Health

Th1

T helper 1

Th2

T helper 2

THC

Delta-9 tetrahydrocannabinol

THC-COOH

11-nor-9-carboxy-delta-9-THC

TNFα

Tumor necrosis factor alpha

Cite this article as: Alshaarawy O, Sotzen M, Kurjan E, Padmanabhan V, Ruden DM, Olson LK (2026) Cannabis use during pregnancy is associated with the suppression of circulating maternal cytokines, Cannabis and Cannabinoid Research 11:1, 78–88, DOI: 10.1177/25785125251388764.