Cannabidiol (CBD) May Improve Craving and Anxiety in Opioid Addiction

The opioid abuse epidemic is expanding in the United States (US), and increasingly affects other nations, spurring the search for therapeutic agents that are not themselves μ-opioid agonists. The authors found that cannabidiol (CBD; from cannabis [Cannabis sativa, Cannabaceae]) reduced craving and anxiety associated with heroin-seeking behavior triggered by drug cues in vivo. Drug cues contribute to relapse in drug-abstinent individuals and are strong triggers for craving, a core component of opioid use disorder (OUD). Preclinical findings show that CBD’s reduction of heroin-seeking behavior lasts for “weeks.” Other in vivo studies report similar results for CBD in reducing contextual memories associated with drug-seeking behavior for other agents. A clinical trial found CBD to be associated with a less attentional bias to cigarette cues in tobacco (Nicotiana tabacum, Solanaceae) users. Clinical trials found high-dose CBD to be safe and without adverse effects (AEs), even when co-exposed with a potent opioid agonist. CBD is reported to reduce anxiety, which is expected to increase when people with OUD are given drug cues.

The authors conducted a randomized, double-blind, placebo-controlled trial of CBD (Epidiolex; GW Pharmaceuticals; Salisbury, Wiltshire, United Kingdom) at 400 and 800 mg/d vs. placebo. Study agents, identical in appearance and taste, included ethanol (79.0 mg/mL), sucralose (0.5 mg/mL), strawberry (Fragaria spp., Rosaceae) flavoring (0.2 mg/mL), and refined sesame (Sesamum indicum, Pedaliaceae) oil (to a volume of 1 mL), with CBD included in the two active formulas. Abstinent individuals with heroin use disorder between 21 and 65 years of age were recruited via advertisements in New York (New York, NY) and at Mount Sinai Health System (NY, NY) addiction clinics. Individuals who met criteria for opioid dependence and did not test positive for any psychoactive drug besides nicotine were included; those maintained on buprenorphine or methadone, or who showed signs of acute withdrawal, were allergic to cannabis, or had a significant medical history/condition were not included.* Of the 50 randomized participants, 42 completed the study. Seven (16.7%) were women. Forty-one (97.6%) had been daily heroin users, with 35 (83.3%) using ≥10 g heroin or other opioids per day. 27 participants (64.3%) had been abstinent for ≤1 month. On average, participants had used heroin for an average of 13.2 years. There were no significant between-group differences at screening (baseline).

Study agents were given daily at three test sessions on consecutive days; a fourth session took place one week after the last dose of the study agent was given. Before each session, participants were tested for alcohol and drug use; women were screened for pregnancy. The Heroin Craving Questionnaire (HCQ) and a visual analog scale for anxiety (VAS-A) were completed at the start of each session. The Clinical Opiate Withdrawal Scale was completed at screening and during session 4. In sessions 1, 2, and 4, participants were shown neutral and drug cues at two points, and salivary cortisol levels (high cortisol levels are an indication of anxiety and stress) were measured. Cues were videos of nature scenes (neutral cue) or of intravenous or intranasal heroin use (drug cue), the latter depending on the individual’s preferred method. Most (33 [78.6%]) participants were intranasal users. In session 2, after the cue presentation, participants were shown neutral objects or objects associated with heroin use. The order in which cues were given was balanced and randomized across participants. Tests were timed to detect effects on craving, anxiety, and other measures one hour after the first CBD/P dose, 24 hours after the second dose, and one week after the third dose. A VAS-C assessing heroin craving** and the Positive and Negative Affect Schedule (PANAS) were completed at various times. Vital signs and salivary cortisol levels were measured.

At baseline and session 3, cognitive tests were completed; in session 3, these began 10 minutes after study agent intake. CBD reaches peak plasma levels in 3-4 hours, with a half-life of 18-32 hours. At the end of sessions, vital signs and well-being were assessed and an AE questionnaire administered. Participants were guided in relaxation exercises and the VAS-C and VAS-A were again completed to assess craving and anxiety that might have arisen in the session. Participants completed the HCQ at home after sessions 1, 2, and 3. Exit interviews assessed participants’ study experience and offered appropriate treatment resources.

When participants were shown drug cues, VAS-C scores rose significantly (mean difference 1.47); neutral cues did not alter VAS-C scores (mean difference -0.51). Those in the placebo group had significantly higher VAS-C scores after drug cues than those in either the CBD group. There were no significant differences in VAS-C scores between the 400 and 800 mg CBD groups. The effects of CBD on VAS-C were most evident in session 1, with a significant group effect (P=0.0105). Lower VAS-C scores 24 hours after the first CBD intake did not reach significance. Apparent habituation of placebo group members to cues resulted in a 40% drop in their VAS-C scores in session 2, while CBD group scores were relatively stable across sessions. At session 4, those who had received a placebo had significantly higher VAS-C scores (mean difference 0.94) vs. those who received 800 mg CBD (mean difference 0.39) for a significant main group effect (P=0.0167). VAS-C scores at session 4 for those who had taken 400 mg CBD did not differ significantly from either the 800 mg CBD or P group. No significant effects of CBD were seen in HCQ scores; in HCQ scores, the craving was strongest at baseline and decreased significantly at each measured time point. Like VAS-C scores, VAS-A scores were significantly associated with drug cues with the main effect of drug group (P=0.0079). Women’s baseline VAS-A scores were higher than men’s. At session 4, the placebo group had significantly higher VAS-A scores than either the CBD group. Among physiological measures, heart rate, temperature, and salivary cortisol rose only in the placebo group with drug cue presentation, not with neutral cue presentation. Neither PANAS nor cognitive test scores were affected by CBD. No serious AEs were reported in any group.

The authors conclude that CBD may offer a new solution to fill the gap in nonopioid medication options for individuals with OUD considering CBD’s ability to reduce cue-induced craving and anxiety. The authors do acknowledge that the two primary outcomes, craving, and anxiety, are subjective, but the results of two objective metrics, heart rate, and cortisol levels, support the findings of the subjective measurements. Overall, the potential of CBD in heroin or opioid cessation continues to gain support with this well-designed study.

*However, 71.0% of participants were hypertensive; 25.3%, HIV positive, while 17.8% had hepatitis C.

**In subgroup analysis, women had significantly higher baseline VAS-C scores (P=0.0476) than men. There was no gender difference in HCQ scores.


Hurd YL, Spriggs S, Alishayev J, et al. Cannabidiol for the reduction of cue-induced craving and anxiety in drug-abstinent individuals with heroin use disorder: a double-blind randomized placebo-controlled trial. [published online on May 21, 2019]. Am J Psychiatrydoi: 10.1176/appi.ajp.2019.18101191.