Table. Prenatal Cannabis Exposure and Child Psychopathology as Children Enter Adolescence.
| Variable | Mean (SD)a | Mixed-model ANCOVA | P value for post hoc mixed-model, uncorrected | |||||
|---|---|---|---|---|---|---|---|---|
| Exposed pre- and postknowledge (n = 208) | Exposed preknowledge only (n = 391) | No exposure (n = 10 033) | χ2 | P value (FDR) | Pre- and postknowledge vs no | Preknowledge only vs no | Pre- and postknowledge vs preknowledge only | |
| CBCL | ||||||||
| Total problems | 31.47 (23.41) | 23.78 (19.66) | 16.7 (15.55) | 13.59 | .004 | <.001 | .29 | <.001 |
| Externalizing factor | 8.83 (7.8) | 6.05 (6.44) | 3.95 (4.93) | 16.29 | .002 | <.001 | .96 | <.001 |
| Rule-breaking behavior | 2.76 (2.66) | 1.8 (2.1) | 1.04 (1.52) | 33.08 | <.001 | <.001 | .37 | <.001 |
| Aggressive behavior | 6.07 (5.53) | 4.25 (4.66) | 2.91 (3.66) | 12.34 | .007 | <.001 | .78 | <.001 |
| Internalizing factor | 7.77 (7) | 6.36 (5.6) | 4.9 (4.9) | 5.06 | .14 | .04 | .23 | .08 |
| Withdrawn/depression | 1.96 (2.3) | 1.54 (1.92) | 1.06 (1.51) | 6.83 | .06 | .02 | .20 | .17 |
| Somatic complaints | 2.12 (2.1) | 1.87 (1.83) | 1.42 (1.65) | 1.23 | .58 | .51 | .34 | .67 |
| Anxious/depressed | 3.69 (3.49) | 2.95 (2.9) | 2.43 (2.69) | 7.01 | .06 | .01 | .26 | .04 |
| Social problems | 3.04 (2.86) | 2.13 (2.31) | 1.43 (1.91) | 18.09 | <.001 | <.001 | .46 | <.001 |
| Thought problems | 2.9 (2.89) | 2.19 (2.58) | 1.5 (1.86) | 17.85 | <.001 | <.001 | .06 | .02 |
| Attention problems | 4.99 (3.83) | 3.98 (3.69) | 2.71 (3.05) | 19.17 | <.001 | <.001 | .11 | .002 |
| Sluggish cognitive tempo | 1.1 (1.27) | 0.83 (1.13) | 0.48 (0.81) | 42.87 | <.001 | <.001 | <.001 | .05 |
| Stress problems | 5.06 (4.23) | 3.76 (3.49) | 2.74 (2.92) | 14.44 | .003 | <.001 | .44 | .005 |
| Obsessive-compulsive problems | 2.17 (2.13) | 1.72 (1.88) | 1.29 (1.56) | 13.16 | .005 | <.001 | .13 | .07 |
| ADHD problemsb | 4.24 (3.23) | 3.4 (2.98) | 2.35 (2.59) | 15.28 | .002 | <.001 | .10 | .005 |
| Anxiety problemsb | 3.13 (2.88) | 2.46 (2.37) | 1.96 (2.12) | 7.74 | .06 | .01 | .20 | .13 |
| Conduct problemsb | 3.14 (3.55) | 1.89 (2.59) | 1.1 (1.91) | 27.22 | <.001 | <.001 | .86 | <.001 |
| Depression problemsb | 2.23 (2.61) | 1.75 (2.12) | 1.31 (1.83) | 3.97 | .19 | .06 | .45 | .04 |
| Oppositional defiant problemsb | 2.82 (2.19) | 2.22 (2.13) | 1.61 (1.76) | 7.17 | .06 | .01 | .90 | .01 |
| Somatic problemsb | 1.51 (1.49) | 1.36 (1.39) | 1.03 (1.25) | 1.04 | .61 | .58 | .36 | .70 |
| PQ-BC | ||||||||
| Psychoticlike experiences | 3.14 (3.32) | 2.95 (2.97) | 2.01 (2.54) | 10.51 | .02 | .03 | .009 | .26 |
Abbreviations: ADHD, attention-deficit/hyperactivity disorder; ANCOVA, analysis of covariance; CBCL, Child Behavior Checklist; FDR, false discovery rate; PQ-BC, Prodromal Questionnaire–Brief Child Version.
Group means (SDs) of each raw CBCL3 subscale score and the total number of psychoticlike experiences reported, regardless of distress, on the PQ-BC.4 Twenty CBCL subscales including those based on factor analyses (n = 11; ie, total problems [includes externalizing, internalizing, social, thought, and attention problems subscales], externalizing problems [includes rule-breaking and aggressive behaviors], internalizing problems [includes withdrawn/depression, somatic complaints, and anxiety/depressed problems), other scales (n = 3; ie, sluggish cognitive tempo, stress problems, and obsessive-compulsive problems), and those judged by experts to be consistent with Diagnostic and Statistical Manual of Mental Disorders (Fifth Edition) categories (n = 6; ie, depressive, anxiety, somatic, ADHD, oppositional defiant, and conduct problems). Mean values were calculated within-person before computing the group means (mean values were not calculated in analyses). ANCOVA χ2 and P values are from mixed-model comparing models with PCEs (ie, exposure pre- and postknowledge of pregnancy, exposure preknowledge of pregnancy only) to 1 without these terms, Benjamini-Hochberg FDR-corrected for multiple comparisons (n = 42 comparisons, df = 2; Figure). Linear mixed-model random intercepts included (1) family, (2) research sites, and (3) participant identification number. Fixed-effect covariates included (1) child age at each visit, (2) child sex (0 = male, 1 = female), (3-8) self-reported child race and ethnicity (African American, Asian, Hispanic, Native American, Pacific Islander, White), (9-12) current household income (<$50 000, <$75 000, <$100 000, >$100 000), (13) current maternal education, (14) pubertal status at baseline, (15-19) first-degree familial history of mental illness (depression, psychosis, anxiety, mania, antisocial behavior), (20-21) first-degree familial history of drug or alcohol problems, (22-24) prenatal exposure to alcohol, tobacco, or other drugs before maternal knowledge of pregnancy, (25-28) prenatal exposure to alcohol, tobacco, or other drugs after maternal knowledge of pregnancy, (29) child substance use (tobacco puff or alcohol sip), and (30) twin or triplet status. Follow-up analyses included pregnancy-associated variables, which were excluded from primary analyses because of high missingness (remaining n = 9367, k = 26 585 observations), including (1) length of time pregnant before maternal knowledge of pregnancy, (2) whether the pregnancy was unplanned (0 = planned, 1 = unplanned), (3) maternal age at birth, (4) prenatal vitamin usage, and (5) birth weight. Further follow-up analyses were restricted to participants of gnomically confirmed European ancestry and included covariates 1-10 (10 ancestry principal components) and 11-17 (polygenic risk for cannabis use disorder, cross-disorder risk, schizophrenia, major depressive disorder, ADHD, generalized anxiety disorder, and risk taking). For additional details regarding genetic analyses please refer to Paul et al.2
Using Diagnostic and Statistical Manual of Mental Disorders (Fifth Edition) codes.