Changes in opioid use outcomes after passage of medical marijuana laws

By the end of 2019, 32 states had adopted medical marijuana laws (MCL) alone. All states that went on to adopt recreational marijuana laws (RCL) had previously adopted an MCL.

Using National Survey on Drug Use and Health (NSDUH) data from 2015-2019, the researchers estimated cannabis law associations with opioid (prescription opioid misuse and/or heroin use) misuse and use disorder.

The study adds to the literature for investigating effects of MCL alone and RCL+MCL on opioid misuse and DSM-IV opioid use disorder overall and by prior cannabis use history. The findings also build on several recent analyses of NSDUH data.

"Our study is the first to investigate the effects of cannabis laws on opioid outcomes among people that used cannabis in the past year and that initiated cannabis use prior to cannabis law adoption in their state," noted Silvia Martins, MD, PhD, professor of Epidemiology at Columbia Mailman School. "Much has been speculated on whether the enactment of medical and recreational cannabis laws can have a role in decreasing prescriptions for opioid pain relievers, opioid use and misuse, opioid use disorder, and overdose in the U.S."

Martins continues, "Comparing individuals in states with medical cannabis laws alone to those in states without such laws, we found an inconsistent pattern of decreased odds of opioid outcomes, which were more pronounced among people reporting cannabis use. The pattern did not hold for individuals in states with recreational cannabis laws, suggesting that MCLs may be associated with reductions in opioid use among people using cannabis but additional work to replicate and expand on these findings is needed."

Overall, MCL and RCL adoption were not associated with changes in the odds of any opioid outcome. After restricting to respondents reporting past-year cannabis use, decreased odds of past year opioid misuse were observed among individuals in states with MCL compared to those in states without cannabis laws. RCLs were not associated with changes in the odds of any opioid outcome beyond MCL adoption.

From 2015 to 2019, approximately 282,768 respondents participated in the NSDUH study. The prevalence of past year and past month opioid misuse was reported by 4 percent and 1.3 percent of participants, respectively. Approximately 3 percent of the participants met criteria for past-year DSM-IV opioid use disorder (OUD) -- abuse and dependence. Prevalence of past year and past month opioid misuse, as well as opioid use disorder, were all higher among respondents reporting past year cannabis use (15 percent) compared to those reporting no use.

Reductions in some measures of opioid misuse and opioid use disorder were reported in states with MCLs alone, but only among individuals reporting past-year cannabis use who had initiated cannabis use prior to cannabis law adoption in their state. However, the same relationships did not hold for combined RCLs and MCLs. "Our findings shed light on the potential beneficial effects of MCLs alone and future studies should continue to monitor changes in opioid-related harm outcomes, as the cannabis legalization landscape evolves," said Emilie Bruzelius, PhD, post-doctoral fellow in the Department of Epidemiology at Columbia Mailman School and co-author of the study.

"Our findings generally support the premise that state adoption of RCLs+MCLs has few benefits in terms of reducing substance misuse and indicates that continued monitoring of RCL+MCL-related trends is warranted," said Martins. "One important issue to consider is that compared to MCLs, relatively fewer states have adopted RCLs+MCLs, and most laws have been adopted within the past decade. Therefore, the impact of RCLs+MCLs may become clearer as more states adopt these laws and as post-law observation time accumulates. More studies are needed to examine opioid use among individuals who receive cannabis from medical and recreational dispensaries."

Co-authors are Christine Mauro, and Kara Rudolph,Columbia Mailman School; Julian Santaella-Tenorio, Katherine Wheeler-Martin, and Magdalena Cerda, NYU Grossman School of Medicine; Anne Boustead, University of Arizona, Hillary Samplesand Stephen Crystal, Rutgers University; Deborah Hasin, Columbia Mailman School and Columbia Department of Psychiatry; David Fink^, New York State Psychiatric Institute; Corey Davis, NYU Grossman School of Medicine and Network for Public Health Law, MN.

The study was supported by NIH-NIDA grants R01DA045872, T32DA031099, K01DA049950, and K99DA055724.