Abstract
Background and Aims
Cannabis use is increasingly common among pregnant individuals and might be a risk factor for severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2) infection. We aimed to test whether prenatal cannabis use is associated with increased risk of SARS‐CoV‐2 infection during pregnancy.
Design
This is a retrospective cohort study.
Setting
The study was conducted in California, USA.
Participants
A total of 58 114 pregnancies (with outcomes from 5 March 2020 to 30 September 2021) among 57 287 unique pregnant women aged 14–54 years who were screened for prenatal substance use, enrolled in Kaiser Permanente Northern California (KPNC) (a health‐care system) and had not tested positive for COVID‐19 prior to pregnancy onset.
Measurements
We utilized data from the KPNC electronic health record. Cannabis use status (current, recently quit and non‐user) was based on universal screenings during prenatal care (including urine toxicology testing and self‐reported use on a self‐administered questionnaire). SARS‐CoV‐2 infection [based on polymerase chain reaction (PCR) tests] was estimated in time‐to‐event analyses using Cox proportional hazard regression models adjusting for covariates. Secondary analyses examined differences in (a) SARS‐CoV‐2 testing rates and (b) SARS‐CoV‐2 infection rates among those tested.
Findings
We observed 348 810 person‐months of follow‐up time in our cohort with 41 064 SARS‐CoV‐2 PCR tests and 6% (n = 2414) of tests being positive. At the start of follow‐up, 7% of pregnant individuals had current use, 12% had recently quit and 81% did not use cannabis. Adjusting for covariates, current use was associated with lower rates of SARS‐CoV‐2 infection [adjusted hazard ratio (aHR) = 0.60, 95% confidence interval (CI) = 0.49–0.74 than non‐use. Those who had recently quit did not differ from non‐cannabis users in infection rates (aHR = 0.96, 95% CI = 0.86–1.08). Sensitivity analyses among patients who received a SARS‐CoV‐2 test also found lower odds of infection associated with current versus no cannabis use (aOR = 0.76, CI = 0.61–0.93).
Conclusions
Current cannabis use appears to be associated with a reduced risk of SARS‐CoV‐2 infection among pregnant individuals.
Keywords: Cannabis, COVID‐19, marijuana, pregnancy, prenatal, SARS‐CoV‐2
INTRODUCTION
Pregnant individuals infected with severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2) have an increased risk of becoming seriously ill compared to individuals not pregnant [1, 2, 3]. Further, SARS‐CoV‐2 infection during pregnancy is associated with increased morbidity and mortality for both mothers and fetuses [1, 2, 4, 5, 6]. Understanding factors associated with SARS‐CoV‐2 risk among pregnant individuals is critical for providing them with better guidance and protection.
Among pregnant and reproductive‐aged women in the United States, rates of cannabis use are rising [7, 8, 9, 10], corresponding with increases in its general acceptance and accessibility [11, 12]. Cannabis use during pregnancy is associated with health risks, including low infant birth weight and effects on offspring neurodevelopment [13, 14, 15, 16], but its effect on risk of SARS‐CoV‐2 infection is unknown. Cannabis smoke has toxicity similar to that of cigarette smoke [17, 18]. Cigarette smoking substantially increases the risk of some pulmonary infectious diseases, such as tuberculosis, and might increase the risk of viral infections SARS‐CoV‐2 [19]. In addition, it is possible that individuals who smoke or vape cannabis are more likely to become infected because of associated risky behaviors, including increased contact between the fingers and the mouth, removal of one's mask while vaping or smoking and/or sharing cannabis products with others who are infected. Conversely, cannabis products may have pharmacological effects that might offer some protection against SARS‐CoV‐2 infection [20, 21, 22, 23]. For example, some cannabinoids can bind to the SARS‐CoV‐2 spike protein that is critical for viral entry into cells [24]. Cannabidiol (CBD) may down‐regulate angiotensin converting enzyme 2 (ACE2), which is the site of SARS‐CoV‐2 viral attachment on cell surfaces, and inhibits viral replication in pulmonary epithelial cells [20, 22].
Findings from general population studies are limited and equivocal. A large, retrospective case–control study of electronic health record (EHR) data found that individuals with a past‐year (versus no past‐year) cannabis use disorder diagnosis had significantly increased odds of SARS‐CoV‐2 infection, but those with a life‐time (versus never) cannabis use disorder diagnosis had significantly lower odds of SARS‐CoV‐2 infection [25]. A population‐based cohort study found that having a cannabis use disorder was associated with an increased risk for SARS‐CoV‐2 breakthrough infection in fully vaccinated patients [26]. Conversely, several in‐vitro studies provide initial evidence that cannabis compounds may offer some protection against SARS‐CoV‐2 infection. Pre‐clinical studies show that cannabis compounds bind to the spike protein of SARS‐CoV‐2, preventing it from entering cells and causing infection [24], decrease ACE2 protein levels to which the spike protein binds [27], prevent SARS‐CoV‐2 replication [20] and reduce COVID‐19 related inflammation [21].
Well‐designed population‐based studies that compare SARS‐CoV‐2 infection risk among current, former and never‐cannabis users could improve understanding of the mechanisms that affect susceptibility to SARS‐CoV‐2 and inform public health strategies to mitigate risk in future outbreaks. Further, given that rates of prenatal cannabis use are rising [7, 8, 9] and pregnancy is a unique time in which the immune system is modulated and risks associated with SARS‐CoV‐2 infection are heightened, research among pregnant women is of particular importance. Leveraging data from a large, diverse sample of pregnant patients with universal screening for pre‐conception and prenatal cannabis use during standard prenatal care, we performed a retrospective cohort study to evaluate whether cannabis use status is associated with risk of SARS‐CoV‐2 infection during pregnancy.
METHODS
Setting and study population
Kaiser Permanente Northern California (KPNC) is an integrated health‐care system with 21 hospital‐based medical centers serving approximately 4.4 million racially and socio‐demographically diverse patients who are representative of Northern California [28]. On 4 March 2020, following the state's first COVID‐19 death, California's Governor declared a state of emergency [29].
Using EHR data, we conducted a retrospective cohort study of pregnancies that began between 1 May 2019 and 1 December 2020 and ended between 5 March 2020 and 30 September 2021. This criterion ensured that our cohort did not over‐represent women with short pregnancies due to pregnancy loss. Non‐KPNC members at the time of pregnancy onset, patients who discontinued membership before the pandemic began or who tested positive for SARS‐CoV‐2 by polymerase chain reaction (PCR) prior to pregnancy onset were excluded. To obtain measures for all pregnancies, we excluded patients who did not complete the question about self‐reported cannabis use during the year before pregnancy and patients who had no urine toxicology test during pregnancy. The KPNC IRB approved the project with a waiver of informed consent.
Outcomes
The primary outcome was SARS‐CoV‐2 infection, defined as first positive SARS‐CoV‐2 PCR test recorded by KPNC during pregnancy (Supporting information). Our study included the time‐period before and during the Delta variant. A secondary outcome was receiving a SARS‐CoV‐2 PCR test in KPNC, regardless of test outcome, to assess whether current or former cannabis users were more or less likely to be tested for SARS‐CoV‐2. Once a patient tested positive via PCR test, no subsequent tests were included in any analyses.
Exposure
The primary exposure of interest was cannabis use status (‘current’, ‘recently quit’ and ‘non‐user’) during pregnancy, which we treated as time‐varying. Cannabis use status was derived primarily from urine toxicology tests performed during pregnancy. All patients had initial urine toxicology tests around their first prenatal visit (~8 weeks gestation) with possible subsequent tests during pregnancy (Supporting information). Positive toxicology tests were confirmed with a laboratory test. Cannabis use status was further established using responses to a substance use screening questionnaire self‐administered around the first prenatal visit, which included a question about any cannabis use during the year prior to pregnancy. A patient with a positive initial urine toxicology test was considered a ‘current’ cannabis user from the start of pregnancy until a subsequent negative prenatal toxicology test, if any. A negative test changed the patient's cannabis use status to ‘recently quit’. A patient with a negative initial urine toxicology test was considered a ‘non‐user’ from the start of pregnancy unless the self‐report questionnaire reported use of cannabis during the year before pregnancy, which changed their cannabis status to ‘recently quit’. A subsequent positive toxicology test changed their status to ‘current’.
Covariates
From the EHR documentation, we extracted the following baseline information as of the pregnancy onset date: age, self‐reported race/ethnicity (white, Hispanic, Asian/Pacific Islander, black and other/unknown/multi‐racial), neighborhood deprivation index (NDI; categorized into quartiles) [30], insurance payor [subsidized (Medicare, Medicaid or other) versus non‐subsidized], primary KPNC facility at which the patient received care (categorical, 25 facilities), parity, pre‐pregnancy body mass index [BMI (kg/m2); underweight < 18.5, normal 18.5–24.9, overweight 25.0–29.9, obese ≥ 30.0] [31], pre‐existing diabetes (based on the KPNC Diabetes Registry [32, 33, 34] within 2 years prior to pregnancy onset or a recorded A1C ≥ 6.5 in the first trimester) and pre‐existing hypertension classified by ICD‐10 codes (Supporting information). We also assessed tobacco smoking status during pregnancy, which we treated as a time‐changing variable (‘current’, ‘former’ and ‘never’). Self‐reported tobacco smoking status is routinely assessed and documented during KPNC primary and specialty care visits, including throughout prenatal care, and we used these assessments to create a time‐changing variable.
Analysis
We examined patient characteristics overall and by initial cannabis use status as measured by responses to the self‐administered substance use screening questionnaire and results of the first prenatal urine toxicology test. We calculated the observed rate (per 100 person‐months) of SARS‐CoV‐2 tests and SARS‐CoV‐2 infections during pregnancy among all pregnancies and for each stratum of cannabis use status. One patient could have more than one SARS‐CoV‐2 test (until a first positive test). Cannabis use status was treated as time‐varying and could change during follow‐up. Therefore, when calculating rates of SARS‐CoV‐2 tests and infections, one patient could contribute denominator time to more than one cannabis use status level.
We used extended multivariable counting‐process Cox proportional hazards [35] models that allow for time‐updated covariates to examine the association between cannabis use status (at time of outcome) and risk of SARS‐CoV‐2 infection. In all models, we used calendar time as the time‐scale to account for changes over time in SARS‐CoV‐2 circulation. Patients were followed from pregnancy onset (or from 5 March 2020 if onset was prior to 5 March 2020) until a positive PCR test, with censoring due to health plan disenrollment, first COVID‐19 vaccine administration or pregnancy outcome. Models were fitted with varying levels of adjustment to understand the potential impact of confounding: model 1 included age (23 categories:< 19, one for each year of age up to 39, 40+), race/ethnicity, NDI, insurance payor and facility; model 2 added pre‐pregnancy BMI, diabetes and hypertension, parity and tobacco smoking status. The outcome was the first positive SARS‐CoV‐2 test. To account for the correlation between multiple pregnancies in the same individual, we fitted marginal Cox proportional hazards models using robust sandwich covariance estimates.
We performed three sensitivity analyses where the outcome was SARS‐CoV‐2 infection. First, we performed a sensitivity analysis when the predictor of interest was the patient's initial cannabis use status rather than allowing cannabis use status to change. Secondly, given prior research indicating that current cigarette smoking might be related to lower risk of SARS‐CoV‐2 infection [36, 37, 38, 39, 40], we performed a sensitivity analysis where we excluded current or former tobacco smokers. In a third sensitivity analysis, we included only pregnancies in which the patient received at least one SARS‐CoV‐2 PCR test, and each test was a record in a logistic regression. A positive test result was the outcome and cannabis use status at the time of the test was the predictor of interest. The logistic regression was conditioned (‘stratified’) on the calendar date of the test so that patients testing positive on a certain day were compared to patients testing negative on that same day. This ‘case‐positive, control‐test‐negative’ design has often been used in studies of vaccine effectiveness, and may reduce bias associated with health‐care‐seeking behavior [41, 42, 43] and propensity to be tested.
In a secondary analysis, we examined the association between cannabis use status and receipt of SARS‐CoV‐2 PCR testing. In these analyses the outcome was a test, and testing was treated as a recurring event. The extended multivariable counting‐process Cox proportional hazards models used for these analyses were identical in all other respects to the models used for the primary analyses where SARS‐CoV‐2 infection was the outcome. Analyses were conducted using SAS software, version 9.4. Statistical significance was assessed at two‐sided P ≤ 0.05.
The analysis was not pre‐registered, and the results should be considered exploratory.
RESULTS
We initially identified 69 569 pregnancies during the study period. After excluding 6080 pregnancies not meeting membership criteria, 352 pregnancies testing positive for SARS‐CoV‐2 by PCR prior to pregnancy onset, 463 pregnancies because the patient did not complete the question about cannabis use during the year before pregnancy and 4375 pregnancies lacking ≥ 1 urine toxicology test during pregnancy, the final study sample included 58 114 pregnancies from 57 287 unique pregnant women (Supporting information).
The patient population had a mean age of 31 years at pregnancy onset and was 35% white, 27% Hispanic, 26% Asian or Pacific Islander, 6% black and 6% other/unknown/multi‐racial (Table 1). At the start of follow‐up, cannabis use status was ‘current’ for 7% of the pregnancies, ‘recently quit’ for 12% and ‘non‐user’ for 81%. Patients with ‘current’ initial cannabis use were younger, more likely to be Hispanic or black, more likely to live in more deprived neighborhoods, have subsidized medical insurance, have hypertension and lower parity than patients classified as ‘recently quit’ or as ‘non‐user’ (all P‐values < 0.01). Among all pregnancies, 11% received ≥ 1 COVID‐19 vaccine dose during pregnancy and 54% had ≥ 1 SARS‐CoV‐2 test recorded in the EHR. Most pregnancies (88.63%) were followed until a positive PCR test or the pregnancy outcome; 0.64% of pregnancies were censored due to membership disenrollment; 10.73% of pregnancies were censored due to vaccination.
TABLE 1.
Pregnancies during the COVID‐19 pandemic, Kaiser Permanente Northern California
| Characteristic | Cannabis use status at start of follow‐up a | |||
|---|---|---|---|---|
| All (n = 58 114) | Current (n = 4053) | Recently quit (n = 7002) | Non‐user (n = 47 059) | |
| Age group, n (%) b , c , d | ||||
| < 25 years | 6447 (11.09) | 1347 (33.23) | 1139 (16.27) | 3961 (8.42) |
| 25–< 35 years | 36 742 (63.22) | 2152 (53.10) | 4402 (62.87) | 30 188 (64.15) |
| 35+ years | 14 925 (25.68) | 554 (13.67) | 1461 (20.87) | 12 910 (27.43) |
| Age, mean (median) b , c , d | 31.08 (31.00) | 27.69 (27.00) | 30.18 (31.00) | 31.50 (32.00) |
| Race/ethnicity, n (%) b , c , d | ||||
| Asian or Pacific Islander | 15 317 (26.36) | 233 (5.75) | 1136 (16.22) | 13 948 (29.64) |
| Black | 3536 (6.08) | 925 (22.82) | 525 (7.50) | 2086 (4.43) |
| Hispanic | 15 528 (26.72) | 1234 (30.45) | 1796 (25.65) | 12 498 (26.56) |
| Other/unknown/multi‐racial | 3509 (6.04) | 291 (7.18) | 506 (7.23) | 2712 (5.76) |
| White | 20 224 (34.80) | 1370 (33.80) | 3039 (43.40) | 15 815 (33.61) |
| Neighborhood deprivation, n (%) b , c | ||||
| Quartile 1, least deprived | 13 201 (22.72) | 390 (9.62) | 1670 (23.85) | 11 141 (23.67) |
| Quartile 2 | 14 461 (24.88) | 756 (18.65) | 1775 (25.35) | 11 930 (25.35) |
| Quartile 3 | 14 027 (24.14) | 1075 (26.52) | 1660 (23.71) | 11 292 (24.00) |
| Quartile 4, most deprived | 16 344 (28.12) | 1820 (44.91) | 1881 (26.86) | 12 643 (26.87) |
| Missing | 81 (0.14) | 12 (0.30) | 16 (0.23) | 53 (0.11) |
| Subsidized insurance—Medicare, Medicaid or other, n (%) b , c , d | 6016 (10.35) | 1033 (25.49) | 718 (10.25) | 4265 (9.06) |
| Body mass index, n (%) b , c , d | ||||
| Underweight | 1355 (2.33) | 104 (2.57) | 136 (1.94) | 1115 (2.37) |
| Normal | 23 070 (39.70) | 1142 (28.18) | 2861 (40.86) | 19 067 (40.52) |
| Overweight | 16 154 (27.80) | 1035 (25.54) | 1910 (27.28) | 13 209 (28.07) |
| Obese | 15 738 (27.08) | 1651 (40.74) | 1905 (27.21) | 12 182 (25.89) |
| Unknown | 1797 (3.09) | 121 (2.99) | 190 (2.71) | 1486 (3.16) |
| Pre‐existing diabetes, n (%) e | 1080 (1.86) | 75 (1.85) | 107 (1.53) | 898 (1.91) |
| Pre‐existing hypertension, n (%) b , c , e | 4061 (6.99) | 412 (10.17) | 513 (7.33) | 3136 (6.66) |
| Parity, n (%) b , c , d | ||||
| 0 | 23 027 (39.62) | 1770 (43.67) | 4203 (60.03) | 17 054 (36.24) |
| 1 | 20 803 (35.80) | 1204 (29.71) | 1554 (22.19) | 18 045 (38.35) |
| 2 | 7959 (13.70) | 506 (12.48) | 578 (8.25) | 6875 (14.61) |
| 3 | 2546 (4.38) | 205 (5.06) | 131 (1.87) | 2210 (4.70) |
| 4+ | 1247 (2.15) | 75 (1.85) | 48 (0.69) | 1124 (2.39) |
| Tobacco smoking status at start of follow‐up, n (%) a , b , c , d | ||||
| Current | 1503 (2.59) | 522 (12.88) | 294 (4.20) | 687 (1.46) |
| Former | 6761 (11.63) | 1094 (26.99) | 1244 (17.77) | 4423 (9.40) |
| Unknown | 7 (0.01) | 0 (0.00) | 1 (0.01) | 6 (0.01) |
| Never | 49 843 (85.77) | 2437 (60.13) | 5463 (78.02) | 41 943 (89.13) |
The COVID‐19 pandemic was considered to have begun on 5 March 2020. Pregnancies were included if they began prior to 1 December 2020 and ended between 5 March 2020 and 30 September 2021.
Follow‐up began at the earliest of pregnancy onset or 5 March 2020. In statistical models, cannabis use and tobacco smoking status were treated as time‐changing variables during the course of the pregnancy.
Differences between people whose cannabis use status at start of follow‐up was ‘current’ and people whose cannabis use status was ‘recently quit’ were significant at P < 0.01.
Differences between people whose cannabis use status at start of follow‐up was ‘current’ and people whose cannabis use status was ‘non‐user’ were significant at P < 0.01.
Differences between people whose cannabis use status at start of follow‐up was ‘recently quit’ and people whose cannabis use status was ‘non‐user’ were significant at P < 0.01.
Differences between people whose cannabis use status at start of follow‐up was ‘recently quit’ and people whose cannabis use status was ‘non‐user’ were significant at P < 0.05.
We observed 348 810 person‐months of follow‐up time in our cohort and 41 064 SARS‐CoV‐2 PCR tests, with 6% (n = 2414) of tests being positive (Table 2). The overall rate of testing was 11.77 per 100 person‐months; the overall rate of infection was 0.692 per 100 person‐months. The unadjusted rate of testing was lower among those with current cannabis use (9.33 per 100 person‐months) than among those who had recently quit (13.58) or non‐users (11.61). The unadjusted rate of infection was also lower among those with current cannabis use (0.647 per 100 person‐months) than among those who had recently quit (0.730) or non‐users (0.688).
TABLE 2.
Unadjusted rates of SARS‐CoV‐2 PCR testing and infections among pregnant women
| Pregnancy time classification | Pregnancies b | Person‐months | SARS‐CoV‐2 PCR tests a | Positive SARS‐CoV‐2 PCR tests a | ||
|---|---|---|---|---|---|---|
| Number of tests | Tests per 100 person‐months | Number of positive tests | Positive tests per 100 person‐months | |||
| All pregnancy time | 58 114 | 348 810 | 41 064 | 11.77 | 2414 | 0.692 |
| By cannabis use status c | ||||||
| Current | 3626 | 18 088 | 1688 | 9.33 | 117 | 0.647 |
| Recently quit | 9256 | 49 745 | 6756 | 13.58 | 363 | 0.730 |
| Non‐user | 47 059 | 280 978 | 32 620 | 11.61 | 1934 | 0.688 |
Person‐time began at the latest of pregnancy onset or 5 March 2020. Person‐time was censored at the earliest of the first positive SARS‐CoV‐2 polymerase chain reaction (PCR) test, disenrollment from Kaiser Permanente Northern California (KPNC), any COVID‐19 vaccination or pregnancy outcome.
People could contribute time to more than one cannabis use category. Therefore, the sum of the number of pregnancies across the cannabis use categories totals more than the number of unique pregnancies in the study cohort.
Cannabis use status was a time‐changing variable. The same person can contribute time to more than one category.
In the fully adjusted models, patients with current cannabis use had a lower hazard of SARS‐CoV‐2 infection [Table 3; adjusted hazard ratio (aHR) = 0.60, 95% confidence interval (CI) = 0.49–0.74] compared to non‐users. Those who had recently quit cannabis use did not differ from non‐users in risk of infection (aHR = 0.96, 95% CI = 0.86–1.08).
TABLE 3.
Relationship of cannabis use status to SARS‐CoV‐2 infection
| Time‐changing cannabis use status | Adjusted hazard ratio (95% confidence intervals) | |
|---|---|---|
| Model 1 a | Model 2 b | |
| Current | 0.61 (0.50, 0.74)* | 0.60 (0.49, 0.74)* |
| Recently quit | 0.92 (0.82, 1.03) | 0.96 (0.86, 1.08) |
| Non‐user | 1.00 (reference) | 1.00 (reference) |
Outcome was the first positive SARS‐CoV‐2 polymerase chain reaction test during pregnancy.
Model 1 is adjusted for age (21 categories: < 19, one for each year of age up to 39, 40+), race/ethnicity, neighborhood deprivation index, insurance payor and primary Kaiser Permanente Northern California (KPNC) facility.
Model 2 is adjusted for all covariables in model 1 one plus body mass index, pre‐existing diabetes and hypertension, parity and tobacco smoking status.
Significant at P ≤ 0.05.
Results of sensitivity analyses were consistent with our overall finding of a lower relative risk of SARS‐CoV‐2 infection among subjects with current cannabis use compared to those with non‐use. In the sensitivity analysis using only the initial cannabis use status as the predictor of interest, current cannabis use was associated with a lower hazard of infection (aHR = 0.71, 95% CI = 0.60–0.83) compared to non‐use (Table 4). In the sensitivity analysis that excluded those with current or former cigarette smoking status, current cannabis use was associated with a lower hazard of infection (aHR = 0.59, 95% CI = 0.46–0.76) compared to non‐use (Table 4). Similarly, in the ‘case‐positive, control‐test‐negative’ sensitivity analysis restricted to patients who received SARS‐CoV‐2 testing, current cannabis use was associated with lower odds of infection [adjusted odds ratio (aOR) = 0.76, 95% CI = 0.61–0.9] compared to non‐use, as was recently quit status (aOR = 0.83, 95% CI = 0.73–0.94) compared to non‐use (Table 4).
TABLE 4.
Sensitivity analyses, relationship of cannabis use status to SARS‐CoV‐2 infection
| Cannabis use status | Sensitivity analysis 1: using initial cannabis use status rather than time‐varying status | Sensitivity analysis 2: excluding time when people were current or former smokers | Sensitivity analysis 3: case‐positive, control‐test‐negative approach a | |||
|---|---|---|---|---|---|---|
| Adjusted hazard ratio (95% confidence intervals) | Adjusted odds ratio (95% confidence intervals) | |||||
| Model 1 b | Model 2 c | Model 1 b | Model 2 c | Model 1 b | Model 2 c | |
| Current | 0.71 (0.60, 0.83)* | 0.71 (0.60, 0.83)* | 0.59 (0.46, 0.76)* | 0.59 (0.46, 0.77)* | 0.73 (0.60, 0.90)* | 0.76 (0.61, 0.93)* |
| Recently quit | 0.92 (0.81, 1.04) | 0.97 (0.85, 1.10) | 0.92 (0.81, 1.05) | 0.96 (0.84, 1.09) | 0.77 (0.68, 0.87)* | 0.83 (0.73, 0.94)* |
| Non‐user | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) |
Logistic regression model that includes all COVID tests during follow‐up, stratified by (conditioned on) calendar date of the test. This analysis only includes people who received at least one test. The cannabis status of people who tested positive on a given day is compared to the cannabis status of people who tested negative on the same day.
Model 1 is adjusted for age (21 categories: < 19, one for each year of age up to 39, 40+), race/ethnicity, neighborhood deprivation index, insurance payor and primary Kaiser Permanente Northern California (KPNC) facility.
Model 2 is adjusted for all covariables in model 1 plus body mass index, pre‐existing diabetes and hypertension, parity and tobacco smoking status.
Significant at P ≤ 0.05.
In the fully adjusted models, compared to non‐users, patients with current cannabis use had a lower hazard of SARS‐CoV‐2 testing (Table 5; aHR = 0.80, 95% CI = 0.76–0.85), and those who had recently quit cannabis use had a higher hazard of SARS‐CoV‐2 testing (aHR = 1.10, CI = 1.06–1.13).
TABLE 5.
Relationship of cannabis use status to SARS‐CoV‐2 testing
| Time‐changing cannabis use status | Adjusted hazard ratio (95% confidence intervals) | |
|---|---|---|
| Model 1 a | Model 2 b | |
| Current | 0.83 (0.79, 0.88)* | 0.80 (0.76, 0.85)* |
| Recently quit | 1.11 (1.08, 1.15)* | 1.10 (1.06, 1.13)* |
| Non‐user | 1.00 (reference) | 1.00 (reference) |
Outcome was receipt of SARS‐CoV‐2 polymerase chain reaction (PCR) tests during pregnancy. Testing was treated as a recurring event outcome. All SARS‐CoV‐2 PCR tests prior to a patient's first positive test were included.
Model 1 is adjusted for age (21 categories: < 19, one for each year of age up to 39, 40+), race/ethnicity, neighborhood deprivation index, insurance payor and primary Kaiser Permanente Northern California (KPNC) facility.
Model 2 is adjusted for all covariables in model 1 plus body mass index, pre‐existing diabetes and hypertension, parity and tobacco smoking status.
Significant at P ≤ 0.05.
DISCUSSION
In a large multi‐specialty health‐care system with routine screening for prenatal cannabis use during prenatal care, pregnant patients with current cannabis use had a lower risk of SARS‐CoV‐2 infection relative to those not using cannabis. This finding is consistent with several studies that found a lower risk of SARS‐CoV‐2 infection associated with current versus never‐cigarette smoking [36, 37, 38, 39, 40]. In our study, those with current cannabis use were more likely than recently quit or non‐cannabis users to smoke cigarettes; however, results remained similar in sensitivity analyses that excluded current and former cigarette smokers, indicating that differences in cigarette smoking were not driving results. The reduced risk of SARS‐CoV‐2 infection among those with current cannabis use is notable as they were more likely than non‐cannabis users to have other risk factors associated with increased risk of SARS‐CoV‐2 infection (e.g. higher neighborhood deprivation, higher percentage subsidized insurance and greater proportion of black and Hispanic patients).
Our main analysis leaves open the possibility of unmeasured confounding due to lower rates of testing among those with current cannabis use and possibly different propensity to be tested given the same symptoms and same infection status. Our ‘case‐positive, control‐test‐negative’ sensitivity analysis was limited to patients who received a PCR test and may be less sensitive to biases associated with the propensity to be tested. Although potential sources of bias and confounding remain with the case–positive, control test‐negative approach [42, 43, 44, 45], the results of that sensitivity analysis were consistent with our main analysis that included all patients, indicating that current cannabis use was associated with a lower hazard of SARS‐CoV‐2 infection despite potential differences in testing propensity.
From a public health perspective, our findings suggest that it is unlikely that current cannabis use during pregnancy increases risk of SARS‐CoV‐2 infection. The protective effect is provocative, but not determinative, and worth further study. Pre‐clinical research provides biological support for a protective effect of cannabis use on SARS‐CoV‐2 infection. The initial steps of viral infection involve attachment of the SARS‐CoV‐2 virus spike protein to the enzyme ACE2 in the lung, oral cavity or nose. Transmembrane serine protease 2 (TMPRSS2) on a host cell then activates the spike protein producing conformation changes that allow the virus to enter the cell. The SARS‐CoV‐2 Mpro enzyme is a viral enzyme that plays a critical role in viral replication. There is evidence that cannabinoids may have protective effects on each of the pathways.
The cannabis plant (Cannabis sativa) contains numerous pharmacologically active chemicals, including more than 100 phytocannabinoids [the most prevalent being delta 9‐tetrahydrocannabiol (THC) and CBD]. Some cannabinoids bind to the spike protein and prevent infection of human lung epithelial cells [24]. Cannabis cultivars containing high levels of CBD were shown in three‐dimensional human tissue models to down‐regulate the ACE2 receptor, which is the initial binding site for SARS‐CoV‐2 [27]. CBD inhibits SARS‐CoV infection in lung epithelial cells and in mice by inhibiting viral replication [20]. Structural‐based screening of cannabinoids found that THC and CBD bind stably to SARS‐CoV‐2 Mpro, a critical enzyme in the life cycle of the virus; and these cannabinoids inhibit viral replication in vitro [22]. Notably, these authors found that the antiviral potencies of THC and CBD effects were similar to that of remdesivir, a Food and Drug Administration (FDA)‐approved drug for treating COVID‐19 infection. While a mechanism by which cannabis use might reduce the likelihood of SARS‐CoV‐2 infection has not been determined, the evidence to date suggests that such an effect is biologically plausible.
Conversely, compared to non‐cannabis users, those who had recently quit using cannabis were significantly more likely to be tested for SARS‐CoV‐2, but they did not differ in their risk of infection. This is consistent with active but not recent cannabis use having a potentially protective effect. In our case‐positive, control‐test‐negative sensitivity analysis that was limited to patients who were tested for SARS‐CoV‐2 infection, those who had recently quit had a lower risk of infection relative to non‐cannabis users. This sensitivity analysis provides some evidence that pregnant patients who had recently quit using cannabis may also have a lower risk of infection. If this is confirmed, it could mean that a less strong protective effect of cannabis use persists for a period of time after patients stop using cannabis, and the higher rates of SARS‐CoV‐2 testing among patients who had recently quit use in the primary analysis may have masked this effect. It is also possible that SARS‐CoV‐2 infection rates do not truly differ among those who recently quit cannabis use versus non‐users. The higher rates of testing among the recently quit group may have created the appearance of a lower risk because the sensitivity analysis includes more negative tests in the denominator. Notably, those who had recently quit using cannabis (versus current users) tended to be more similar to non‐cannabis users in socio‐demographics and clinical characteristics. Additional research is needed to replicate these findings and to more clearly understand the factors that might be related to differences in risk of SARS‐CoV‐2 infection during pregnancy among current versus recently quit cannabis users.
It is critical to note that cannabis use during pregnancy is associated with adverse perinatal outcomes, including low offspring birth weight and pre‐term delivery [46, 47], and may increase risk for offspring psychopathology in childhood, including externalizing, attention, thought and social problems [48]. Because of concerns regarding these health risks, national organizations recommend that pregnant women abstain from prenatal cannabis use [49]. While cannabis use may not be a risk factor for SARS‐CoV‐2 infection, having a past‐year cannabis use disorder is associated with elevated risk for SARS‐CoV‐2 [25] and having a current or in‐remission cannabis use disorder is associated with elevated risk for a breakthrough infection among fully vaccinated individuals [26]. Given documented health risks associated with prenatal cannabis use, clinicians should continue to encourage pregnant patients to avoid prenatal cannabis use and offer support and resources to assist with quitting.
Limitations and strengths
This study has several limitations and results should be considered preliminary. Our sample is limited to KPNC pregnant patients screened for prenatal substance use and results may not generalize to all pregnant women or other populations. Our measure of prenatal cannabis use does not account for frequency, potency or mode of cannabis use. Although population‐based surveillance data indicate that most (91%) women who use cannabis during pregnancy report smoking it [50], the individuals in our study may have differed in their patterns of use (e.g. frequency, mode of administration and potency of products used). We were also unable to determine whether there is a dose–response relation between prenatal cannabis use and SARS‐CoV‐2 infection. Additional studies are needed that examine how these important aspects of cannabis use relate to SARS‐CoV‐2 infection during pregnancy. Further, because we only assessed cannabis use during the year before and during pregnancy, individuals who quit using cannabis more than a year prior to pregnancy cannot be differentiated from those who never used cannabis and are combined in our ‘non‐cannabis’ use group. Self‐reported data on cannabis use during the year before pregnancy may be under‐reported, and some women with former cannabis use may have been classified as never‐users. Although a study strength was accounting for changes in cannabis use status during pregnancy, not all patients had subsequent toxicology tests, so some misclassification of exposure is possible. However, we conducted a sensitivity analysis that limited the exposure to cannabis use status at baseline, and results were very similar to results from our models where cannabis use was time‐varying. PCR tests were limited to those conducted in KPNC; positive cases tested elsewhere or not tested at all were not included. Our study was limited to first SARS‐CoV‐2 infection, and additional studies are needed to examine the relation between prenatal cannabis use and risk of recurrent infection. The Cox proportional hazards model assumes non‐informative censoring, and bias may occur if patients have a censoring time that is dependent upon their failure time. While no study can unequivocally rule out this potential for bias, we do not expect that our censoring events (vaccination and health plan disenrollment) would have a strong dependency on the unobserved outcomes of those patients. Finally, while we adjusted for many confounders, some possible confounders were missed, such as exposure to the virus, number and frequency of social contacts and gatherings, occupation and essential worker designation and mask‐wearing. The study is observational and has a potential bias related to PCR testing behaviors. To address potential differences between cannabis users and non‐users in the propensity to be tested, we performed a sensitivity analysis using only people who were tested and found similar results to our main analysis. Further, while we adjusted for several important covariates, unmeasured biases related to who is tested or potential differences in behaviors or risk (e.g. willingness to wear a mask or socially distance from others and essential worker designation) may remain. Finally, causal interpretations should not be based on one observational study.
This study also has important strengths. We included a large sample of pregnant patients at‐risk for COVID‐19 and followed them from testing to infection. Our study design allows us to estimate risk properly over time using semi‐parametric Cox proportional hazards models and is more rigorous than studies using convenience samples. Our models account for changes in underlying baseline risks and variations in risks and exposures during the course of the pandemic.
CONCLUSIONS
In this retrospective cohort study, prenatal cannabis use was associated with a reduced risk for SARS‐CoV‐2 infection during pregnancy. Additional research is needed to replicate this finding and to understand possible factors driving lower infection risk during pregnancy among individuals with prenatal cannabis use. Regardless of its association with COVID‐19, because of other health risks associated with prenatal cannabis use, clinicians should continue to advise pregnant patients to abstain from using cannabis.
DECLARATION OF INTERESTS
All authors declare no conflicts of interest.
AUTHOR CONTRIBUTIONS
Kelly Young‐Wolff: Conceptualization; funding acquisition; investigation; methodology; supervision. G. Thomas Ray: Data curation; formal analysis; methodology. Stacey Alexeeff: Conceptualization; formal analysis; investigation; methodology; supervision. Neal Benowitz: Investigation; methodology. Sara Adams: Data curation; investigation; methodology. Monique Does: Investigation; methodology; project administration. Nancy Goler: Investigation; methodology. Deborah Ansley: Investigation; methodology. Amy Conway: Investigation; methodology. Lyndsay Avalos: Conceptualization; funding acquisition; investigation; methodology; supervision.
Supporting information
Data S1. Supporting Information
ACKNOWLEDGEMENTS
This study was supported by the National Institute on Drug Abuse Grants DA047405, DA043604 and DA048033. 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.
Young‐Wolff KC, Ray GT, Alexeeff SE, Benowitz N, Adams SR, Does MB, et al. Association of cannabis use during pregnancy with severe acute respiratory syndrome coronavirus 2 infection: a retrospective cohort study. Addiction. 2023;118(2):317–326. 10.1111/add.16056
Funding information National Institute on Drug Abuse, Grant/Award Numbers: DA048033, DA043604, DA047405
REFERENCES
- 1. Walter K. COVID‐19 and pregnancy. JAMA. 2021;40:173–9. [Google Scholar]
- 2. Centers for Disease Control and Prevention, National Center for Immunization and Respiratory Diseases . Investigating the impact of COVID‐19 during pregnancy. 2021. Available at: https://www.cdc.gov/coronavirus/2019-ncov/downloads/cases-updates/covid-fs-Pregnancy.pdf (accessed 24 January 2022).
- 3. Metz TD, Clifton RG, Hughes BL, Sandoval G, Saade GR, Grobman WA, et al. Disease severity and perinatal outcomes of pregnant patients with coronavirus disease 2019 (COVID‐19). Obstet Gynecol. 2021;137:571–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Ferrara A, Hedderson MM, Zhu Y, Avalos LA, Kuzniewicz MW, Myers LC, et al. Perinatal complications in individuals in California with or without SARS‐CoV‐2 infection during pregnancy. JAMA Intern Med. 2022;182:503–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Ko JY, DeSisto CL, Simeone RM, Ellington S, Galang RR, Oduyebo T, et al. Adverse pregnancy outcomes, maternal complications, and severe illness among US delivery hospitalizations with and without a coronavirus disease 2019 (COVID‐19) diagnosis. Clin Infect Dis. 2021;73:S24–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Villar J, Ariff S, Gunier RB, Thiruvengadam R, Rauch S, Kholin A, et al. Maternal and neonatal morbidity and mortality among pregnant women with and without COVID‐19 infection: the INTERCOVID multinational cohort study. JAMA Pediatr. 2021;175:817–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. 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;317:207–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Young‐Wolff KC, Tucker LY, Alexeeff S, Armstrong MA, Conway A, Weisner C, et al. Trends in self‐reported and biochemically tested marijuana use among pregnant females in California from 2009–2016. JAMA. 2017;318:2490–1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Young‐Wolff KC, Sarovar V, Tucker LY, Conway A, Alexeeff S, Weisner C et al. Self‐reported daily, weekly, and monthly cannabis use among women before and during pregnancy. JAMA Netw Open 2019;2:e196471. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Volkow ND, Han B, Compton WM, Mccance‐Katz EF. Self‐reported medical and nonmedical cannabis use among pregnant women in the United States. JAMA. 2019;322:167–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Jarlenski M, Koma JW, Zank J, Bodnar LM, Bogen DL, Chang JC. Trends in perception of risk of regular marijuana use among US pregnant and nonpregnant reproductive‐aged women. Am J Obstet Gynecol. 2017;217:705–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Daniller A. Two‐thirds of Americans support marijuana legalization. 2019. Available at: https://www.pewresearch.org/fact-tank/2019/11/14/americans-support-marijuana-legalization/ (accessed 16 March 2022).
- 13. Metz TD, Allshouse AA, Hogue CJ, Goldenberg RL, Dudley DJ, Varner MW, et al. Maternal marijuana use, adverse pregnancy outcomes, and neonatal morbidity. Am J Obstet Gynecol. 2017;217:478.e1–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Sharapova SR, Phillips E, Sirocco K, Kaminski JW, Leeb RT, Rolle I. Effects of prenatal marijuana exposure on neuropsychological outcomes in children aged 1–11 years: a systematic review. Paediatr Perinat Epidemiol. 2018;32:512–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. National Academies of Sciences, Engineering, and Medicine, Health and Medicine Division . The Health Effects of Cannabis and Cannabinoids. The Current State of Evidence and Recommendations for Research. Washington, DC: National Academies Press; 2017. [PubMed] [Google Scholar]
- 16. 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:e20181889. [DOI] [PubMed] [Google Scholar]
- 17. Benowitz NL. Managing cannabis use in patients with cardiovascular disease. Can J Cardiol. 2019;35:138–41. [DOI] [PubMed] [Google Scholar]
- 18. Moir D, Rickert WS, Levasseur G, Larose Y, Maertens R, White P, et al. A comparison of mainstream and sidestream marijuana and tobacco cigarette smoke produced under two machine smoking conditions. Chem Res Toxicol. 2008;21:494–502. [DOI] [PubMed] [Google Scholar]
- 19. Lopez‐Hernandez Y, Rivas‐Santiago CE, Lopez JA, Mendoza‐Almanza G, Hernandez‐Pando R. Tuberculosis and cigarette smoke exposure: an update of in vitro and in vivo studies. Exp Lung Res. 2018;44:113–26. [DOI] [PubMed] [Google Scholar]
- 20. Nguyen LC, Yang D, Nicolaescu V, Best TJ, Ohtsuki T, Chen SN, et al. Cannabidiol inhibits SARS‐CoV‐2 replication and promotes the host innate immune response. bioRxiv. 2021. 10.1101/2021.03.10.432967 [DOI] [PubMed] [Google Scholar]
- 21. Anil SM, Shalev N, Vinayaka AC, Nadarajan S, Namdar D, Belausov E, et al. Cannabis compounds exhibit anti‐inflammatory activity in vitro in COVID‐19‐related inflammation in lung epithelial cells and pro‐inflammatory activity in macrophages. Sci Rep. 2021;11:1462. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Raj V, Park JG, Cho KH, Choi P, Kim T, Ham J, et al. Assessment of antiviral potencies of cannabinoids against SARS‐CoV‐2 using computational and in vitro approaches. Int J Biol Macromol. 2021;168:474–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Malinowska B, Baranowska‐Kuczko M, Kicman A, Schlicker E. Opportunities, challenges and pitfalls of using cannabidiol as an adjuvant drug in COVID‐19. Int J Mol Sci. 2021;22:1986. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. van Breemen RB, Muchiri RN, Bates TA, Weinstein JB, Leier HC, Farley S, et al. Cannabinoids block cellular entry of SARS‐CoV‐2 and the emerging variants. J Nat Prod. 2022;85:176–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Wang QQ, Kaelber DC, Xu R, Volkow ND. COVID‐19 risk and outcomes in patients with substance use disorders: analyses from electronic health records in the United States. Mol Psychiatry. 2021;26:30–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Wang L, Wang Q, Davis PB, Volkow ND, Xu R. Increased risk for COVID‐19 breakthrough infection in fully vaccinated patients with substance use disorders in the United States between December 2020 and August 2021. World Psychiatry. 2022;21:124–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Wang B, Kovalchuk A, Li D, Rodriguez‐Juarez R, Ilnytskyy Y, Kovalchuk I et al. In search of preventive strategies: novel high‐CBD Cannabis sativa extracts modulate ACE2 expression in COVID‐19 gateway tissues. Aging 2020;12:22425–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Gordon NP. Similarity of adult Kaiser Permanente members to the adult population in Kaiser Permanente's Northern California Service Area: comparisons based on the 2017/2018 cycle of the California Health Interview Survey. Oakland, CA: Report prepared for the Kaiser Permanente Division of Research. 2020. Available at https://divisionofresearch.kaiserpermanente.org/projects/memberhealthsurvey/SiteCollectionDocuments/compare_kp_ncal_chis2017-18.pdf (accessed 15 May 2022).
- 29. Eby K. Coronavirus Timeline: tracking major moments of COVID‐19 pandemic in San Francisco Bay Area. 2022. Available at: https://abc7news.com/timeline-of-coronavirus-us-covid-19-bay-area-sf/6047519/ (accessed 24 January 2022).
- 30. Messer LC, Laraia BA, Kaufman JS, Eyster J, Holzman C, Culhane J, et al. The development of a standardized neighborhood deprivation index. J Urban Health. 2006;83:1041–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Centers for Disease Control and Prevention . Defining adult overweight and obesity. Available at: https://www.cdc.gov/obesity/adult/defining.html (accessed 15 March 2022.
- 32. Karter AJ, Ackerson LM, Darbinian JA, D'Agostino RB Jr, Ferrara A, Liu J, et al. Self‐monitoring of blood glucose levels and glycemic control: the northern California Kaiser Permanente diabetes registry. Am J Med. 2001;111:1–9. [DOI] [PubMed] [Google Scholar]
- 33. Karter AJ, Ferrara A, Liu JY, Moffet HH, Ackerson LM, Selby JV. Ethnic disparities in diabetic complications in an insured population. JAMA. 2002;287:2519–27. [DOI] [PubMed] [Google Scholar]
- 34. Selby JV, Ray GT, Zhang D, Colby CJ. Excess costs of medical care for patients with diabetes in a managed care population. Diabetes Care. 1997;20:1396–402. [DOI] [PubMed] [Google Scholar]
- 35. Andersen PK, Gill RD. Cox's regression model for counting processes: a large sample study. Ann Stat. 1982;10:1100–20. [Google Scholar]
- 36. Simons D, Shahab L, Brown J, Perski O. The association of smoking status with SARS‐CoV‐2 infection, hospitalization and mortality from COVID‐19: a living rapid evidence review with Bayesian meta‐analyses (version 7). Addiction. 2021;116:1319–68. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Lee SC, Son KJ, Kim DW, Han CH, Choi YJ, Kim SW, et al. Smoking and the risk of severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2) infection. Nicotine Tob Res. 2021;23:1787–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Tsigaris P, Teixeira Da Silva JA. Smoking prevalence and COVID‐19 in Europe. Nicotine Tob Res. 2020;22:1646–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Paleiron N, Mayet A, Marbac V, Perisse A, Barazzutti H, Brocq FX, et al. Impact of tobacco smoking on the risk of COVID‐19: a large scale retrospective cohort study. Nicotine Tob Res. 2021;23:1398–404. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Prinelli F, Bianchi F, Drago G, Ruggieri S, Sojic A, Jesuthasan N, et al. Association between smoking and SARS‐CoV‐2 infection: cross‐sectional study of the EPICOVID19 internet‐based survey. JMIR Public Health Surveill. 2021;7:e27091. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Foppa IM, Haber M, Ferdinands JM, Shay DK. The case test‐negative design for studies of the effectiveness of influenza vaccine. Vaccine. 2013;31:3104–9. [DOI] [PubMed] [Google Scholar]
- 42. Jackson ML, Nelson JC. The test‐negative design for estimating influenza vaccine effectiveness. Vaccine. 2013;31:2165–8. [DOI] [PubMed] [Google Scholar]
- 43. Sullivan SG, Tchetgen Tchetgen EJ, Cowling BJ. Theoretical basis of the test‐negative study design for assessment of influenza vaccine effectiveness. Am J Epidemiol. 2016;184:345–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Ferdinands JM, Foppa IM, Fry AM, Flannery BL, Belongia EA, Jackson ML. Re: ‘Invited commentary: Beware the test‐negative design’. Am J Epidemiol. 2017;185:613. [DOI] [PubMed] [Google Scholar]
- 45. Westreich D, Hudgens MG. Invited commentary: Beware the test‐negative design. Am J Epidemiol. 2016;184:354–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46. Marchand G, Masoud AT, Govindan M, Ware K, King A, Ruther S, et al. Birth outcomes of neonates exposed to marijuana in utero: a systematic review and meta‐analysis. JAMA Netw Open. 2022;5:e2145653. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Sujan A, Young‐Wolff KC, Avalos LA. In‐utero cannabis exposure and long‐term psychiatric and neurodevelopmental outcomes: a critical review and recommendations for future research. Birth Defects Res B Dev Reprod Toxicol. 2022;114:689–713. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48. Paul SE, Hatoum AS, Fine JD, Johnson EC, Hansen I, Karcher NR, et al. Associations between prenatal cannabis exposure and childhood outcomes: results from the ABCD study. JAMA Psychiatry. 2021;78:64–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49. Committee on Obstetric Practice . Committee opinion no. 722: Marijuana use during pregnancy and lactation. Obstet Gynecol. 2017;130:e205–9. [DOI] [PubMed] [Google Scholar]
- 50. Castillo‐Carniglia A, Keyes KM, Hasin DS, Cerda M. Psychiatric comorbidities in alcohol use disorder. Lancet Psychiatry. 2019;6:1068–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data S1. Supporting Information