rethinkTHC Search
Menu
Study breakdown

Heavy Drinking, Not Cannabis, Was Linked to Unintended Pregnancy Among Those Wanting to Avoid It

Prospective CohortModerate evidence
The takeaway

Among 2,015 individuals, heavy drinking but not cannabis use was associated with higher pregnancy risk among those most wanting to avoid pregnancy.

Reproductive health providers, substance use counselors, public health researchers studying unintended pregnancy.

51% higher pregnancy risk with heavy drinking

What the researchers found

In a longitudinal cohort of 2,015 individuals capable of pregnancy, heavy drinking and daily cannabis use were associated with higher desire to avoid pregnancy. Among those with high desire to avoid pregnancy, heavy drinking was associated with 51% higher pregnancy risk (aHR 1.51, 95% CI 1.12-2.04). Cannabis use was not associated with elevated pregnancy risk among those wanting to avoid it. Other drug use was not associated with pregnancy regardless of preferences.

Why it matters

This study addresses a long-standing assumption that substance use leads to unintended pregnancy. The nuanced finding that heavy drinking specifically, and not cannabis or other drug use, is associated with unintended pregnancy in those wanting to avoid it suggests targeted interventions should focus on alcohol rather than substance use broadly.

The numbers in context

2,015 participants. 40% heavy drinking. 16% cannabis use. 3% other drug use. 282 incident pregnancies. Heavy drinking + high DAP: aHR 1.51 (p < 0.01). Cannabis not associated with pregnancy in high DAP group.

How the study worked

Longitudinal cohort study of 2,015 individuals ages 15-34 from 23 healthcare facilities in five southwestern U.S. states (2019-2022). Past-month substance use self-reported at baseline. Pregnancy preferences measured quarterly using the Desire to Avoid Pregnancy scale. Incident pregnancy reported every 6 weeks. Adjusted mixed effects regression and Cox proportional hazard models used.

What this study cannot tell us

Self-reported substance use may underestimate actual consumption. Only baseline substance use measured. Five southwestern states may not represent national patterns. Cannot determine mechanisms (e.g., whether alcohol affects contraceptive use or judgment).

How to read the evidence

Moderate: well-designed longitudinal cohort with prospective pregnancy preference measurement, though self-reported substance use and regional sample limit generalizability.

When this study was published

2025 study (data from 2019-2022)

The bigger picture

Substance use and unintended pregnancy are often discussed together as a generalized risk, but this study shows the relationship is substance-specific. The finding that cannabis use did not increase unintended pregnancy risk challenges assumptions that may drive stigmatizing clinical practices.

Questions still open

  • Does alcohol specifically impair contraceptive use or decision-making? Would the findings replicate with biomarker-confirmed substance use? Should reproductive health counseling differentiate substance-specific risks more clearly?

Common questions

Does cannabis use increase unintended pregnancy risk?
In this study, cannabis use was not associated with higher pregnancy risk among those wanting to avoid pregnancy. Heavy alcohol use was the only substance associated with elevated unintended pregnancy risk.
Why does heavy drinking increase unintended pregnancy risk?
The study found the association but did not determine the mechanism. Possible explanations include alcohol impairing judgment about contraceptive use or reducing the likelihood of consistent contraception.

Read the original research

Alcohol and drug use and attainment of pregnancy preferences in the southwestern United States: A longitudinal cohort study.

Addiction (Abingdon, England), 120(12), 2527-2537

Citation

Raifman, Sarah; Roberts, Sarah C M; Rocca, Corinne H. (2025). Alcohol and drug use and attainment of pregnancy preferences in the southwestern United States: A longitudinal cohort study.. Addiction (Abingdon, England), 120(12), 2527-2537. https://doi.org/10.1111/add.70135

Explore the wider topic