Anxiety and depressive disorders are highly prevalent (1), recurrent (2, 3), and can have a substantial negative impact on quality of life (4). Outcomes are worsened in the likely incidence of comorbidity (5–7), and both depression and comorbid anxiety/depression are associated with increased risk of mortality (8), particularly in people with co-occurring chronic physical illnesses (9). Yet, treatment is often not pursued despite the availability of multiple treatment options (6, 10–13).

Several pharmacotherapeutic interventions show efficacy in the treatment of anxiety and depression (14, 15). However, many patients are skeptical about the use of medication (16, 17), and antidepressants, the most frequently prescribed of these medications (18, 19), are not without contraindications. Adverse events are fairly common across antidepressant drug classes, and can disincentivize initiation and contribute to discontinuation of pharmacotherapy (14, 20–24). Further, although antidepressants are demonstrably superior to placebo at alleviating symptoms of both anxiety and depression, effect sizes are small (14, 25), and, thus, may not always be perceptible at the patient level. Finally, discontinuation of antidepressant treatment after sustained use is associated with a withdrawal syndrome in most patients that ranges in severity and can last for several months (26). Taken together, even though there is clear evidence of efficacy for antidepressants at the population level, perceived variability in cost-benefit ratio at the patient level means many people with anxiety or depression are interested in alternative options.

In this vein, an increasing number of people struggling with anxiety and/or depression are trying cannabis products for symptom management (27–29). Cannabis products can generally be separated into three “chemotypes” based on the predominant chemical constituents: (1) Δ9-tetrahydrocannabinol (THC) dominant products, (2) cannabidiol (CBD) dominant products, and (3) products that contain roughly equal amounts of both THC and CBD. Published studies on the impact of cannabis use on anxiety and depression have shown mixed results, and often vary based on the chemotype of the product under investigation and the duration of the dosing regimen. For example, the two studies in which THC was acutely administered to people with clinical anxiety showed limited evidence of anxiolysis (30, 31), but chronic nabilone (oral synthetic THC analog) administration over 4 weeks was associated with a significant reduction in anxiety in a placebo-controlled trial (32). Differential effects of cannabinoid type 1 receptor (CB1R) agonists, like THC, observed in long-term vs. acute dosing studies may be a product of increased CB1R binding on cortical glutamatergic neurons due to CB1R downregulation on GABAergic terminals (33, 34); a similar mechanism is implicated in the dose-dependent effects of acute THC exposure on anxiety (33). Neuroplasticity following extended exposure may also explain anxiogenesis frequently reported in cannabis withdrawal (35).

Effects of THC treatment on depression also appear mixed, though no clinical trials have been conducted to examine a direct effect of THC on depressive symptoms. Epidemiological studies suggest that non-medicinal (“recreational”) use of cannabis, which is typically THC-dominant, may be associated with increased risk of developing a depressive disorder (36) and greater depressive symptom severity (37), an association not observed for anxiety disorders (38). Rather than being causative, however, non-medicinal cannabis use may instead represent an attempt at self-medication during a prodromal period. Indeed, the CB1R agonist activity of THC mimics endogenous cannabinoid signaling, which is notably downregulated in women with clinical depression (39), and endogenous cannabinoids appear to regulate neural serotonergic signaling (40). THC itself can produce feelings of euphoria (41), and clinical trials of a CB1R antagonist were discontinued following reported increases in depression and suicidality (42). However, THC has shown no evidence of antidepressant efficacy when assessed as a secondary outcome in treatment trials for chronic pain, though self-reported depression scores in these trials were already low at baseline (43–45).

CBD, a phytocannabinoid that lacks the abuse liability of THC (46), has potential for therapeutic use in psychiatry. CBD has shown anxiolytic efficacy both acutely [(47, 48) but see (31)] and following chronic treatment in people with clinical anxiety (49). Preclinical evidence suggests that anxiolytic effects are produced via 5-HT1A receptor agonism in both acute (50–52) and chronic dosing models, without impacting 5-HT1A receptor expression (53). This lack of neural remodeling may explain why CBD discontinuation does not appear to produce a THC-like withdrawal syndrome (54). Additionally, the proposed serotonergic mechanism of CBD is distinct from that used by most common antidepressant medications, which selectively inhibit cellular reuptake of serotonin and/or norepinephrine (SSRIs, SNRIs), and is instead more comparable to the anxiolytic medication buspirone (55). Like CBD, buspirone does not appear to produce a withdrawal syndrome (56). Antidepressant effects of CBD have also been consistently demonstrated preclinically following both acute and chronic administration (57–60), though no clinical trials have yet been published. Antidepressant effects appear to be a product of the same serotonergic mechanism that drives anxiolysis (57, 58), and have been shown to synergize with other serotonergic medications (61). This again draws comparison with buspirone, which shows evidence of efficacy both as a depression monotherapy (62) or as an adjunct treatment to SSRIs (63).

Research evaluating the anxiolytic or antidepressant effects of products with a more balanced THC:CBD ratio is limited. Some studies in humans indicate that concurrent CBD/THC administration attenuates anxiogenic effects produced by THC (64, 65), but this has not been observed consistently (66, 67) and may be dose-dependent (68). This inconsistency is mirrored in the preclinical literature (69–71), making it difficult to determine a responsible mechanism given the diverse pharmacological activity of CBD (72, 73). Balanced THC:CBD products have also not been assessed for efficacy in psychiatric populations, though effects on anxiety and depression have been reported as secondary outcomes in clinical trials for other conditions. Nabiximols produced no effect on symptoms of anxiety or depression in people with multiple sclerosis (74) or in people with chronic pain due to cancer (75). Notably, both of these studies listed current psychiatric diagnosis as exclusion criteria, making it difficult to extrapolate these outcomes to people with clinical anxiety or depression.

Given these conflicting outcomes, the impact of medicinal cannabis use on anxiety and depression remains an open question. Our group previously found that medicinal cannabis users reported reduced anxiety and depression when compared to a control group that was considering, but had not yet initiated medicinal cannabis use (76). This impact of cannabis use was observed both cross-sectionally and longitudinally. However, effect sizes were only modest, and this was likely a product of the diverse array of clinical conditions represented in the sample. The purpose of the current study was to extend prior findings by limiting our focus to only participants that reported having anxiety and/or depression. We also expand on other epidemiological work that has principally focused on the impact of non-medicinal cannabis use on symptoms of anxiety and depression by providing insight into the effects of medicinal cannabis use on these symptoms.

Despite the high incidence of anxiety and depressive disorders, particularly among women (1), many people suffering with these conditions do not pursue treatment (6, 10–13) and pharmacotherapeutic options remain subpar (14, 23, 25). Here, we evaluated symptoms of anxiety and depression, as well as other general health metrics, in a convenience sample of medicinal Cannabis Users with anxiety and/or depression in comparison with a non-using Control group in a longitudinal web-based survey study. In the cross-sectional comparison at baseline, Cannabis Users self-reported less severe depression, but not anxiety. This effect was strongest among participants using CBD-dominant products, and was not impacted by participant gender or concurrent use of serotonergic antidepressants. Cannabis Users also reported superior sleep, quality of life, and lower average pain relative to Controls at baseline. Adverse effects attributed by participants to cannabis product use were infrequent, were more associated with THC-dominant product use, and, with the exception of nausea, were distinct from those typically associated with antidepressants (21, 23). In longitudinal analyses, participants who initiated medicinal cannabis use during the follow-up period showed a significant reduction in both depression and anxiety symptoms. A similar, albeit smaller magnitude, effect was observed in participants that sustained medicinal cannabis use throughout the study, suggesting an improvement in anxiety and depression symptoms with both the onset of cannabis use and with extended use.

A handful of studies have previously examined the anxiolytic effects of THC (30–32) and CBD (31, 47–49) in clinical populations, and most of these have found a positive effect. While we observed no effect of medicinal cannabis use on anxiety at baseline, participants that initiated cannabis use during the follow-up period reported a significant reduction in anxiety that was not mirrored in Non-initiators. This discrepancy between cross-sectional and longitudinal anxiety outcomes may reflect on the therapeutic window for CBD. As participants may have been using medicinal cannabis for any period of time prior to baseline assessment to be considered a Cannabis User, differences in outcomes may be attributable to the development of tolerance to anxiolytic effects of cannabinoids, although positive effects observed in the Sustainer group suggest otherwise. Alternatively, this discrepancy may reflect a latency period during which the clinical benefits of cannabinoids for anxiety are not yet observed. Controlled clinical studies and dense sampling data (e.g., ecological momentary assessment) are needed to clarify these early-in-treatment effects with greater precision. It is also important to note that the CBD doses used in trials that found an anxiolytic effect were far greater than the average reported by participants in our study, and the time course for anxiolytic efficacy may differ by dose (47–49). Finally, it is possible that anxiolytic effects of CBD may be condition-specific, as symptom improvement has been consistently observed in studies of social anxiety disorder (47–49), but not in obsessive compulsive disorder (31). The heterogeneity of our sample and reliance on self-report for psychiatric condition(s) of a given participant may have precluded observation of anxiolysis at baseline.

Assessment of the antidepressant effects of medicinal cannabis use has thus far been more limited. In our study, we found that Cannabis Users reported reduced depression relative to non-using Controls at baseline. Consistent with prior research in people with comorbid chronic pain (43–45), we found that use of THC-dominant products was not superior to use of non-THC-dominant products in alleviating depression symptoms. In contrast, participants that reported use of CBD-dominant products provided significantly lower depression scores relative to those that did not, consistent with preclinical findings (57, 85, 86). Cannabis Users also reported superior sleep, average pain, and quality of life relative to Controls. This is unsurprising given the interrelated nature of these constructs with depression (87–89), and both pain and quality of life have been shown to be improved with traditional antidepressant treatment (90, 91). Consistent with baseline outcomes, we found that initiation of medicinal cannabis use was associated with a significant reduction in depressive symptoms, sustained use was associated with a modest reduction, and participants that did not use cannabis at all showed no difference in symptom expression between baseline and follow-up. These combined cross-sectional and longitudinal findings show a consistent antidepressant effect of medicinal cannabis.

This study has several limitations. Notably, these outcomes rely entirely on participant self-report. It is possible that some of the symptom improvement reported by medicinal Cannabis Users can be attributable to an expectancy effect, especially in consideration of the low daily CBD doses reported by participants relative to those used in previous clinical studies (31, 47–49). However, as research in this area is still in its infancy, it is presently unknown what constitutes an optimal dose to maximize antidepressant/anxiolytic efficacy and minimize incidence of side effects, particularly when dosing over an extended period. It is also unclear if minor cannabinoids or terpenes present in whole-plant or “full spectrum” products used in this study, compared with isolated pure chemical substances used in previous research, confer additional therapeutic benefits via an entourage effect (92). Similarly, Controls were people that were considering the use of medicinal cannabis to treat their condition. Therefore, it is also not surprising that people who believed medicinal cannabis might assist with their condition enough to initiate use would perceive a substantial benefit. This study additionally represents a convenience sample of people registered with the Realm of Caring foundation willing to complete lengthy online assessments for only modest incentives. People with mild symptoms, with more responsibilities in daily life, or who did not see a substantial improvement from medicinal cannabis treatment may have been less likely to complete the survey. Finally, dosing data could only be obtained for about half of the medicinal Cannabis Users given poor or non-existent labeling information on some retail and all black market cannabis products, and most participants did not complete a follow-up assessment. Supplemental analyses showed no significant differences in baseline health behavior between participants who did and did not provide follow-up data, p > 0.39, d < 0.08.

In sum, this study suggests that use of CBD-dominant cannabis products is associated with reduced depression in a sample of mostly female, Caucasian adults. The study extends prior research by including a control group, and through a study design that includes both cross-sectional and within-subject, longitudinal comparisons. Though antidepressant effects of CBD have been consistently reported preclinically (57, 85, 86), our work contributes to the literature by showing a potential for translation across species without many of the negative side effects associated with traditional antidepressants (21, 23). It is recommended that this antidepressant effect of CBD be evaluated further in placebo-controlled clinical trials, and that participants remain under observation following treatment completion to confirm the absence of a discontinuation syndrome. Additionally, considering the average daily CBD dose reported in our study was quite low relative to previous clinical work (31, 47–49), future research is necessary to determine best dosing practices to achieve optimal antidepressant effects. Medicinal cannabis products may also alleviate anxiety, but it is unclear if this effect is gated by duration of use. Placebo-controlled clinical trials are necessary to further explore the potential efficacy of CBD in the treatment of anxiety and depression.

The datasets presented in this article are not readily available because the corresponding author does not own this data. Requests to access the datasets should be directed to Ryan Vandrey (rvandrey@jhmi.edu).

The studies involving human participants were reviewed and approved by Institutional Review Board of Johns Hopkins University. The patients/participants provided their written informed consent to participate in this study.

HJ, RV, MB-M, NS, and JM conceptualized and executed the parent survey. EM conceptualized this secondary analysis, curated the data, and drafted the original outline. EM and JS conducted statistical analyses and created the figures and tables. EM, JS, and RV wrote the first and final manuscript drafts. All authors reviewed and edited both manuscript drafts and approved the final manuscript.

This work was supported by the Realm of Caring Foundation and by National Institute on Drug Abuse (NIDA) grants T32-DA007288 (EM) and T32-DA007209 (NS and JS).

RV has received financial compensation as a consultant or advisory board member from Canopy Growth Corporation, MyMD Pharmaceuticals, and Syqe Medical Ltd. MB-M is an employee of Canopy Growth Corporation and past board of directors member for AusCann Group Holdings Ltd. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.