Table 1 Key characteristics of included studies (n = 12)

Beatty et al. 2012 [30]

Large urban hospital in MI

To examine prenatal marijuana and tobacco use measured by self-report and compare marijuana use prevalence across self-report, urine drug assay, and hair sample testing

100 Women ages 18 years of age or older, English speaking, and had no postpartum administration of narcotic pain medication

Self-reported prevalence of any marijuana use was 11%, However, objectively defined marijuana use was more prevalent than self-reported tobacco use: 14% tested positive for marijuana by urinalysis, and 28% by hair analysis. A total of 14 participants were positive for past 3-week marijuana use: 10 by urine toxicology results only and 4 by both self-report and urine toxicology results.

Objective measures (urine and hair toxicology results) of recent and longer-term marijuana use revealed rates of marijuana use to be three times higher than was indicated by self-report. A broader public health response to address prenatal marijuana use is needed.

Chang et al. 2017 [31]

Five outpatient obstetrics and gynecology clinics in Pittsburg, PA

To examine audio-recorded 1st obstetric visits to assess rates of screening, disclosure, and use of prenatal marijuana and illicit drug use, comparing disclosure to urine screening

422 pregnant women 18 years of age or older, English speaking, and attending their first obstetric visit

Screening: OCPs asked about illicit drug use in 81%; 29% disclosed any current or past illicit drug use to their OCP. Among women who disclosed illicit drug use, 11% (n = 48) disclosed current marijuana use and another 7% (n = 30) disclosed past marijuana use.

Although marijuana is illegal in Pennsylvania, a high proportion of pregnant patients used marijuana, with many not disclosing use to their obstetric care providers. This highlights the limitations of perinatal illicit drug use studies that rely solely on self-report or medical record data.

Garg et al. 2016 [17]

University of New Mexico hospital-affiliated specialty prenatal clinic; Biomarkers in Pregnancy Study

To assess validity of self-reported drug use for major classes of illicit drugs and opioid-maintenance therapy among Hispanic and Native American pregnant women

83 pregnant women 18 years of age or older, fluent-English speaking, with a singleton pregnancy 35 weeks’ gestation or less

Prevalence of marijuana was (25.3%). Sensitivity of self-report for marijuana was 57.9%. Sensitivity of self-report for marijuana was higher among occasional users compared to regular users (42.9% vs. 38.5%). Specificity of self-report for all drug classes was high (≥ 90%), indicating that self-reported non-users were confirmed by negative urine drug screens in most cases.

Findings suggest that prenatal drug use is highly underreported, including among women who regularly participate in urine drug screenings. Though underreported, marijuana was more accurately reported than other drug classes. Future studies should be cautious about exclusive reliance on self-report.

Klawans et al. 2019 [32]

Urban, university-affiliated obstetric clinics in TX

To compare rates of prenatal marijuana and other illicit substance use via three identification and screening methods: self-report; urine drug screening ordered by clinician via routine practice; and universal urine drug screening

116 pregnant women presenting for first prenatal visit to the obstetric clinic between Aug and Dec 2015

Via self-report and universal urine drug screen results, 11.6% (n = 27) of the sample were current marijuana users. 80% of women with a marijuana-positive urine screen denied current use on survey; 75% of women who tested positive for marijuana via universal screening were not selected by clinicians for a urine screen. 90% of women who reported current marijuana use did not receive a clinician-ordered drug test.

Prenatal marijuana use was high, with limitations of patient self-report and selective, non-routine screening to identify prenatal substance use. Clinician-ordered tests increased identification by only 0.5% beyond self-report for marijuana. Effective, standardized, clinic-wide strategies are needed to support providers in identifying pregnant women who use substances to increase the frequency of education and intervention.

Metz et al. 2019 [21]

Two urban medical centers in Aurora and Denver, CO

To compare maternal marijuana use prevalence via self-reported to prevalence via umbilical cord sampling in a state with legalized marijuana. Secondary objective was to evaluate if reported frequency of use in the month prior to delivery was correlated with THC-COOH detection in the cord

116 women with a viable singleton pregnancy with 24 weeks or greater gestation who were admitted for delivery and delivered across 12 consecutive weekdays in Nov 2016

In the sample, 2.6% of patients reported marijuana use to health care providers; 14.7% reported past year use and 6.0% reported past-month use via survey.

There was moderate agreement between 30-day use on the survey and umbilical cord homogenate above the limit of detection for THC-COOH (kappa 0.52, 95% CI 0.32–0.72). Agreement between disclosure to health care providers and self-reported use in the past year on the survey was fair (kappa 0.27, 95% CI 0.02–0.51). The agreement between medical record review and umbilical cord homogenate above the limit of detection for THC-COOH was slight (kappa 0.17, 95% CI 0.0–0.34).

Recent studies using umbilical cord testing have suggested an association between marijuana use and adverse outcomes including stillbirth and neonatal morbidity. Umbilical cord sampling results in higher estimates of prenatal marijuana use than self-report even in the setting of legalization. Thus, umbilical cord assays for THC-COOH demonstrate promise for quantifying use.

Yonkers et al. 2011 [33]

Integrated obstetrical/substance use treatment program, NR

To determine the relationship between self-report and urine toxicology tests of drug use over many time intervals prior to assessment of marijuana or cocaine use

168 pregnant women at least 16 years of age, English- or Spanish-speaking, not yet completing their 29th week of pregnancy, reported alcohol or illicit drug use other than opiates, during 28 days prior to screening or scored > 3 on the TWEAK

Mean reported frequency of past 30-day use for use of any hazardous substance other than nicotine was 5.5 days (SD = 8.3). Of those indicating past or current problems with marijuana, mean frequency of past 28-day use was 8.7 (SD = 9.3). Of the 69 women who tested positive for marijuana, 64% (n = 44) reported use within 1–10 days of the test, with 78% (n = 54) reporting some use within 28 days of the test. Of the 47 women who reported using marijuana 1–10 days before the test, 44 (94%) tested positive. Marijuana agreement between self-report and toxicology results for the prior month was k = 0.74 (95% CI = 0.63, 0.84).

Analysis of pregnant substance users found good agreement with one-month self-report of marijuana use and urine toxicology reports. Many women who screened positive reported use later than suggested by the toxicology screening test, particularly for cocaine. A question about use of marijuana or cocaine during the preceding month rather than the prior few days may be a better indicator of use. Positive past-month reports of drug use may indicate current drug use.

Young-Wolff et al. 2020 [23]

Two Kaiser Permanente Northern California (KPNC) medical centers, CA

To determine validity of self-reported prenatal cannabis use via comparison to positive urine toxicology testing, and predictors of nondisclosure using data from a large integrated healthcare delivery system with universal screening for prenatal cannabis use.

281,025 KPNC pregnant women who were screened for self-reported cannabis use during pregnancy between 2009 and 2017, and had a urine toxicology test for cannabis 2 weeks from the date they completed the self-reported screening questionnaire

Urine toxicology testing identified more instances of prenatal cannabis use than self-report (4.9% vs 2.5%). Older women, those of Hispanic race/ethnicity, and those with lower median neighborhood incomes were most likely to be misclassified as not using cannabis by self-reported screening. In our sample, self-reported screening correctly identified only 34% of those who had a positive urine toxicology test. About 2/3 of women tested positive for prenatal cannabis use by toxicology testing. We validated self-reported prenatal cannabis use using the urine toxicology test as the criterion standard; sensitivity of self-reported use was very low (33.9%) and the PPV of self-reported prenatal cannabis use was moderate (65.8%). Specificity of the urine toxicology test (99.1%) and NPV (96.7%) were excellent. Sensitivity of the toxicology test was higher (65.8%), with greater detection of self-reported daily (83.9%) and weekly (77.4%) than monthly or less use (54.1%).

Results from this study indicate that sensitivity of self-reported prenatal cannabis use during prenatal care is low and misclassification of use by self-report may vary with sociodemographic characteristics. Sensitivity of the urine toxicology test is higher, with greater detection of self-reported daily and weekly use versus monthly or less use. Given that many women chose not to disclose prenatal cannabis use in healthcare settings, it is important that clinicians educate all patients of reproductive age about the potential risks of prenatal cannabis use and advise prenatal patients to avoid using cannabis during pregnancy.

Bessa et al. 2010 [34]

Labor and delivery unit of Mario Moraes Altenfelder Silva Maternity Hospital, São Paulo, Brazil

To check the validity of self-report of drug of pregnant adolescents, by comparing interview responses about cocaine and marijuana use with hair samples

1000 pregnant teenage inpatients ages 11 to 19

Hair analysis detected the use of cocaine and/or marijuana in the third trimester of the pregnancy in 6% (n = 60) of patients, with 4% (n = 40) using only marijuana, 1.5 (n = 17) used only cocaine and 0.3% (n = 3) used both drugs. None of the patients had reported the use of these substances in their interviews. 0% disclosure (0/957) reported cannabis use, but 43 tested positive.

Drug abuse during teenage pregnancy is a major health problem and the identification of infants born from these mothers should be done using sensitive methods of detection right after birth so that appropriate intervention can be performed.

El Marroun et al. 2011 [24]

Generation R study in Rotterdam, Netherlands

To verify self-reported information on prenatal drug use in urine

8880 women from prenatally enrolled population-based birth cohort within the Generation R study

Of the 3997 with urine samples available, 92 (2.3%) reported having used cannabis during pregnancy and 71 (1.8%) had positive urine screens. 35% of the 92 women with self-reported cannabis use also had a positive urine screen. Positive urines were frequent in women reporting cannabis use before pregnancy only (7.6%) and in women with missing information (2.6%).

Sensitivity and specificity of urinalysis compared to self-report were 0.46 and 0.98. Sensitivity and specificity of self-report compared to urinalysis were 0.36 and 0.99. Yule’s Y amounted to 0.77, indicating substantial agreement between the two measures. Compared to women that did not report use, paternal cannabis use was more common in women disclosing use (78.9 vs. 8.5%; p < 0.001). Paternal cannabis use was also more common when maternal urine samples were positive (71.7 vs. 9.2%; p < 0.001)

These findings indicate that both approaches perform very well in the identification of non-cannabis users, but that both measures seem to identify partially different subpopulations of cannabis users during pregnancy. In conclusion, researchers and clinicians should acknowledge that pregnant women may underreport current cannabis use, a situation that seems most prevalent in women admitting past cannabis use and in women refusing to provide information on prenatal cannabis use. Findings illustrate the difficulties in obtaining valid information on prenatal cannabis use; self-report seems to be an acceptable single method to determine cannabis use during pregnancy in epidemiological studies.

Lamy et al. 2017 [35]

Maternity hospitals in Rouen, Normandy, France

To compare self-reported prevalence of alcohol, tobacco and/or cannabis use during the third trimester of pregnancy with results of meconium testing of their metabolites in newborns

724 pregnant women aged 18 or over, living in our catchment area that delivered a child in one of these maternity hospitals

Cannabis use prevalence was low via self-report (0.8%, n = 6); and meconium sampling (1.1%, n = 7). We found a low level of concordance (Kappa = 0.30) between cannabinoid metabolites in meconium samples and self-reports of cannabis use during the 3rd trimester.

Only 2/7 positive meconium samples were concordant with self-reports (1–2 joints/day during the 3rd trimester). In all 7 cases, cotinine was also positive. In 3 women, including 2 women reporting daily use in 3rd trimester reporting prenatal cannabis use, cannabinoid metabolites were negative in meconium.

One pair of dizygotic twins had positive cotinine, EtG and THC-COOH meconium samples, with EtG and THC-COOH concentrations 6 and 2 times higher in the female twin compared to the male twin, respectively.

Maternal psychoactive substance use is an ongoing concern; detecting prenatal use is a crucial component of early diagnosis of fetal alcohol syndrome and neonatal care. There was almost no concordance between maternal self-reports of cannabis use and THC-COOH quantitative measures. Assessment of prenatal cannabis exposure, using meconium testing needs to be improved.

Williams et al. 2020 [36]

Community-based clinics called midwife obstetric units in Greater Cape Town (Metropole), South Africa

To examine agreement among simple dichotomous self-report, validated screening results, and biochemical screening results of prenatal alcohol and other drug use

684 pregnant women 16 years or older, presenting for prenatal care

The weighted sensitivity for ASSIST self-report of cannabis use compared to urine screening biomarkers was 51.4% (95% CI: 27.8–74.9), and the specificity was 98.4% (95% CI: 97.7–99.2) with the PPV being 37.6% and the NPV being 99.1%.

Self-reported prevalence of illicit drug use was underreported. Combined use of urine screenings and self-report can be recommended especially for identifying underreported substances to accurately detect AOD use in pregnancy, to enable identification and referral to intervention(s) can occur.

Chiandetti et al. 2017 [37]

Global: US, Spain, Canada, Sweden, Italy, Denmark, Uruguay

To compare reported rates of prenatal alcohol and drugs of abuse exposure with biomarkers of exposure by a comprehensive review of available literature

Studies published in English between 1992 and 2015. Inclusion criteria was “diagnosis/identification/detection of prenatal exposure to drugs of abuse or alcohol”.

Studies agreed that either meconium or hair analysis proved more sensitive than maternal interview for drugs of abuse. Garcia-Serra et al. found more sensitivity in hair analysis than maternal meconium to detect cannabis. The percentage of women who admitted to using THC was 2.9%. Positive results in biomarkers were up to 4% for THC, whilst Lendorio et al. found up to 12.4% positives for THC.

We propose using biomarkers as the main screening tool in patients in environments with high prevalence of AOD of abuse, along with questionnaires. Studies with biomarkers may not be available in all services but should be considered in cases with suspected use even if denied in questionnaires.