Self-Reported and Biologic Assessments of Prenatal Cannabis Use: Ancillary Analysis of a Prospective Observational Cohort

Abstract

Background:

We evaluated agreement between self-reports of prenatal cannabis use and biologic assays, and identified factors associated with discordance.

Methods:

Ancillary analysis of the “Nulliparous Pregnancy Outcomes Study: Monitoring Mothers to Be” (nuMoM2b) study, a United States multicenter prospective cohort of pregnant individuals from October 2010 to September 2013. Past 30-day self-reported prenatal cannabis use was assessed at nuMoM2b study visit 2 (16 weeks and 0 days-21 weeks and 6 days), visit 3 (22 weeks and 0 days-29 weeks and 6 days), and at delivery. Biologic detection of cannabis metabolite (11-nor-9-carboxy-delta-9-tetrahydrocannabinol) was performed with urine samples from visit 2 and visit 3 and the umbilical cord at delivery. The agreement between self-report and biologic assay was evaluated with a Kappa statistic. Factors associated with discordance were evaluated with multivariable logistic regression.

Findings:

Nine thousand one hundred sixteen participants were included, and 434 (4.76%) used cannabis by either self-report or biologic assay. Among participants who used cannabis, 74.5% of prenatal cannabis use at visit 2 and 79.6% of prenatal cannabis use at visit 3 was detected via urine assay and not self-report. Eighty-nine point seven percent of prenatal cannabis at delivery use was detected via cord assay and not self-report. The agreement between self-report and biospecimen assay was fair at visit 2 (Kappa 0.30, 95% confidence interval [CI] 0.24-0.36), fair at visit 3 (Kappa 0.23, 95% CI 0.17-0.29), and slight at delivery (Kappa 0.11, 95% CI 0.01-0.21). Age <30 years, being unmarried, government insurance, and moderate or high perceived stress were associated with discordance between self-reported use and biologic assay.

Conclusions:

Nearly threefold more prenatal cannabis use was detected by biologic assay compared with self-report. The agreement was fair between self-report and urine assay during pregnancy and slight between self-report and umbilical cord assay at delivery. Several sociodemographic factors were associated with discordance between self-reported use and biologic assay results.