Introduction

In the last decades, cannabis for recreational and more so for medical use has become more common than ever. United Nations (UN) reports estimate that over 192 million people globally were exposed to cannabis, reflecting a 4% annual prevalence of the worldwide population [1]. Based on the 2018 Annual Report Questionnaire, the annual prevalence of cannabis use in the United States is estimated as 19.4% in the adult population [2].

The immediate and prolonged effects of cannabis use are varied as it depends on the user’s characteristics such as age, medical history comorbidities, exposure to other drugs, cause for consumption [3,4,5]. Furthermore, cannabis products are different in tetrahydrocannabinol (THC) and cannabidiol (CBD) composition, other cannabinoid and non-cannabinoid content (such as terpenes), and method of consumption.

Cannabis may have beneficial effects on certain disorders. Many countries formally approved cannabinoids as a treatment for conditions such as epilepsy, nausea and vomiting, and spasticity in multiple sclerosis (Epidiolex, Dronabinol, and Sativex, respectively). Despite limited safety and efficacy data, cannabis is also used for conditions such as post-traumatic stress disorder (PTSD), inflammatory bowel diseases, and pain, designated as "medical cannabis" programs in a growing number of countries [6, 7]

In addition to the attempt to introduce cannabis into the bona fide medical pharmacopeia, cannabis is widely marketed for a wide array of conditions. Thus, online cannabidiol retailers in Canada claimed efficacy and effectiveness for cannabis in 171 conditions. Some of these are for medical conditions such as multiple sclerosis, anxiety, epilepsy, and cancer, while others are less for medical entities and primarily serve to relieve mood, fatigue, discomfort, and mental unclarity [8].

Although numerous cannabis products are marketed for emotional, mental, or mental-related conditions, one of the main concerns of cannabis use is its potential to cause or exacerbate mental and psychiatric disorders. Various reports and studies found an association between the use of cannabis and the new onset or exacerbation of mental conditions, specifically psychotic disorders [9,10,11]. Di-Forti et al. reported that daily exposure to cannabis was associated with higher rates of psychotic disorders compared with those who reportedly never used cannabis (OR = 3.2, 95%CI 2.2–4.1) [12]. In addition, cannabis was also reportedly associated with events of anxiety, dissociation, and mood changes [13]. However, it is still unclear whether one-time cannabis or chronic cannabis use may trigger substantial long-term mental deterioration rather than temporary toxification [12, 14,15,16,17]. Contrarily, others argue that cannabis may alleviate mental distress and even psychotic symptoms [18], including improvement of psychotic symptoms in schizophrenia [19].

Milder cases of adverse reactions to cannabis exposure are characterized by negative mental emotions, which are unpleasant, and often transient, sensations or thoughts, such as a state of confusion, disorganized thoughts, anxiety, slow responsiveness, and even paranoia [20]. Negative mental emotions are associated with commonly used medications as well such as antiepileptics [21], antiviral agents [22], and even antidepressants [23]. In the settings of these triggering treatments, while mild cases of negative mental emotion inflict the patient’s quality of life, more severe cases lead to therapy discontinuations [24], or even cause hospitalizations and death [Secondary analysis outcome

For most participants, the available data included a ranking of the symptom severity of the reason provided for cannabis use before and after cannabis exposure. Based on this data, we evaluated whether reporting negative mental emotion affects the change in severity of symptoms. The analysis was conducted in the whole population and in the two subgroups of users who utilized cannabis for pain or mental purposes.

Statistical analysis

Descriptive statistics were utilized for presenting the Sociodemographic characteristic traits of cannabis users and products included in the study: mean and standard deviation were calculated for continuous variables, and distributions were created for categorical variables. We used the “pandas” package in Python via the platform Jupyter Notebook [26, 27].

In the primary analysis, multivariable logistic regression models with mixed effects were constructed to study the associations between the users’ attributes (age and the reason provided for cannabis use), and cannabis product attributes (product type and product composition), and reported negative mental emotions using the “glmmTMB” package in R programming language [28]. Each model corresponded with its relevant confounders based on DAG to avoid the “Table 2 fallacy”. The “Table 2 fallacy” occurs when multiple effects are estimated and interpreted by a single multivariable analysis, equating the primary and secondary effects in the analysis and giving the reader the misimpression of parity across the effect estimates [29, 30]. As offered by Westreich et al. DAGs were used to draw plausible causal association in the model variables while focusing on the analyzed primary effects [30]. The user was defined as a random effect, and other covariables were defined as fixed effects. Changes in the reason provided for cannabis use, the product type, and composition of products are plausible in different sessions retrieved from the same patient. All these variables were included as fixed effects in the mixed effects model, as they can explain changes in the response variable in the same user.

In the secondary analysis, a linear regression model with mixed effects was constructed to assess the change in the severity of mental and pain symptoms, which were the reason for cannabis use, using the “lme4” package in R programming language [31], between those who reported on a negative mental emotion and those who did not report on such emotion. Outputs are presented as adjusted Odds Ratios (OR) for the primary analysis, and adjusted change in the rank of severity symptom and their correspondent 95% confidence interval (CI) [32, 33].

Results

In the primary analysis, the study population comprised 4,435 users, who reported 34,279 sessions different in date of use and containing complete data regarding their cannabis experience (Fig. 1). 55,452 sessions were included in the analysis for the secondary outcome assessing symptom relief in case of reporting negative mental emotions (Fig. 2).

Fig. 1
figure 1

Data extraction of Releaf database – Flow chart of the primary analysis.*Duplicated sessions were omitted since they were identical sessions in terms of user’s identity, date of use, reason for use, and could not contribute more than one time to the analysis

Full size image
Fig. 2
figure 2

Data extraction of Releaf database – flow chart of secondary analysis. *Duplicated sessions were omitted since they were identical sessions in terms of user’s identity, date of use, reason for use and ranking symptom relief and could not contribute more than one time to the analysis (for more explanation about the differences between the primary and secondary flow charts see Appendix 5)

Full size image

Table 1 presents the characterization of the sessions in the primary analysis. Users aged 18–30 reported less than 25% of the sessions, those aged 30–40 reported about a third, the age group of 40–50 years reported about 25% of the sessions, and older users reported about 20%. Almost 60% of the sessions were reported by those who defined themselves as beginners or with little experience with cannabis while the remaining roughly 40% of the sessions were reported by those who defined themselves as experts or have a lot of experience. Approximately 84% reported that they used a product administrated via the lungs (smoking or va** with concentrates and flowers), whereas the remaining 16% used an oral product (edibles, pills and tinctures). In 77.1% of the sessions, the users consumed THC-dominant products, while the rest were split roughly equally between balanced and CBD-dominant products. Similar observations are found in user distribution (Table S1). Of 34,279 sessions, 10,664 negative mental emotions were reported (31.1%). About 27% of the negative mental emotions were reported by users under 30 years of age, representing a slightly higher portion than its share in the total number of sessions in the study population (Table S2).

Table 1 Characterization of included sessions in the study
Full size table

The association between users’ attributes age and reason for use and negative mental emotions: Reporting on negative mental emotions was associated with younger ages, as the odd ratios of 18–30 years were higher than most other age groups (OR = 0.82, 95%CI [0.65–1.02], OR = 0.48, 95%CI [0.37–0.62], OR = 0.48, 95%CI [0.35–0.67], OR = 0.49, 95%CI [0.30–0.80], OR = 0.66, 95%CI [0.25–1.73],in the age groups 30–40 years, 40–50 years, 50–60 years, and 70–90 years compared with 18–30 years, respectively) (Fig. 3a). No interactions were found between age and gender for reporting negative mental emotions (Appendix 6).

Using cannabis for a mental purpose such as depression or anxiety was associated with a small increase in reporting on negative mental emotions, whereas using cannabis for pain was not associated with reporting negative mental emotions (OR = 1.10, 95%CI [1.03–1.19], OR = 0.99, 95%CI [0.91–1.08], compared to other symptoms) (Fig. 3b).

Fig. 3
figure 3

a The association between age and reports of negative mental emotions. *Adjusted for gender, experience, and the reason provided. b The association between the reason provided for cannabis use and reports of negative mental emotions. *Adjusted age, gender, experience, dominance, product type

Full size image

Odds ratios for negative mental emotions – the association with cannabis product attributes: Product types were associated with negative mental emotions, including concentrates and oral products (OR = 1.10, 95%CI [0.99–1.23], OR = 1.16, 95%CI [0.97–1.40], OR = 1.65, 95%CI [1.39–1.95], OR = 1.32, 95%CI [1.00–1.75] in concentrates, edible products, pills, and tinctures respectively compared to flowers) (Fig. 4a). THC-dominant products were associated with reporting more negative mental emotions compared to balanced products (OR = 1.21, 95%CI [1.06–1.38]) (Fig. 4b).

Fig. 4
figure 4

a Product types in users who experience negative mental emotions. Adjusted to reason provided for use, product composition, and experience. b Cannabis composition in users who experience negative mental emotions: *Adjusted to reason provided for use, product type, and experience

Full size image

Changes in symptoms in sessions with negative mental emotion: We also evaluated the change in symptoms for which cannabis was used in those who reported a negative mental emotion while adjusting to other covariables with linear regression. Based on the analysis, after adjusting to covariables in a mixed effect model, reporting on negative mental emotion was not associated with worsening or improvement in the symptom average (Fig. 5). In the subgroup analysis in pain or mental users, no such effect was observed as well (0.02, 95%CI [ – 0.02 to 0.07]; 0.02, 95%CI [ – 0.05 to 0.09], 0.02, 95%CI [ – 0.03 to 0.08], respectively).

Fig. 5
figure 5

Changes in symptoms in users who reported negative mental emotions compared to users with no such reports

Full size image

Discussion

To our knowledge, it is the first observational study on large scales assessing the characteristics of cannabis users who experience a negative mental or emotional effect, based on real-time reporting data. In the last decades more countries allow cannabis use for various medical indications and legalize cannabis recreational use, and thus the putative risk of cannabis use must be determined. If cannabis use is indeed found hazardous, it is imperative to learn whether certain characteristics predispose individuals to the negative attributes of this substance [12, 34].

Although most of the data was collected from North America, it represents general trends of cannabis utility in Western societies, including Europe, and focused on cannabis utility attributes rather than the consumption of specific products marketed only in North America. Many markets around the world look at the developments in the United States to better inform how to approach medical cannabis legalization and access. Given this focus, the findings have merit. They can help inform patients, healthcare providers, regulators, and researchers at a global level, specifically in Europe, where approximately 8% of European adults (22.6 million aged 15 to 64) have used cannabis in the last year. These results are heterogeneous at the national level, probably due to different approaches for the medical and legalized cannabis use [3, 35, 36].

In this study, most of the participants were females. This distribution was seen in previous studies based on this database [37], but conflicts with known data suggesting that cannabis consumption among men is more frequent [38]. Nevertheless, cannabis use among women is increasing, and recent studies based on validated registries demonstrated a majority of females for cannabis use [39, 40]. Nevertheless, the high prevalence of women in our database most likely reflects differential reporting habits between genders.

This study suggested that specific characteristics of cannabis users are associated with more reports of negative mental emotions, such as young age (< 30 years). Regarding the cannabis product, the oral route of consumption was associated with more reports of negative mental emotions than the inhaled route of consumption, as well as mental reasons for cannabis use and THC-dominant products. This work did not include the negative mental emotions in children, although previous work showed that cannabis treatment was associated with negative cognitive and mental emotions in this population [41].

Most mental disorders appear before age 35 and reckless cannabis consumption also reduces with age [42, 43]. The notion that older individuals are less mentally triggered by cannabis, can be linked to the mixed evidence of the changes in the endocannabinoid system during aging. Some in-vivo studies suggest a reduction in the endocannabinoid tone in aging, expressed mainly by the decline of CB1 receptors in several regions in the brain, namely the cortex, the cerebellum, the hippocampus, and the limbic and hypothalamic structures [44]. On the other hand, other works did not find a low density of CB1 in the older murine CNS but indeed reported diminished coupling between CB1 receptors and Gi proteins, leading to diminished cannabinoid signal transduction [45]

Most oral cannabis products were associated with a substantially increased risk for negative mental emotions compared to smoking flowers. THC oral ingestion leads to extensive metabolism in the liver of the prominent psychoactive metabolite of THC, 11-OH-THC, which may be responsible for some of the cannabis effects, and one of them might be the induction of negative mental emotions [46]. Oral administration of THC was also reported to induce anxiety specifically in naïve subjects rather than experienced users who are more familiar with cannabis [46]. On the other hand, the study cannot exclude a reverse causal where users with a predisposition to negative mental emotions after cannabis exposure tend to consume oral cannabis.

THC-dominant products were reported to be associated with the risk for negative mental emotions, compared to balanced and CBD-dominant products. These results corroborate the results of previous studies that reported an elevated risk of psychosis after THC exposure, which is known for its varied psychoactive effects [12]. Nevertheless, it is noteworthy that the obtained data only contained information regarding the THC and CBD levels in the product, while the dose, the total THC concentration and the contents minor cannabinoids were not introduced to the analysis due to lack of data.

The study has several limitations. First, it is an observational study, and it is hard to determine whether the app users represent the population who consume cannabis. The outcomes are based on self-reports and are not considered medically validated tests for psychiatric assessment. Nevertheless, the negative mental emotions reported by users indicate a specific state of mind, which may be unpleasant in itself. In addition to reflecting the momentary experience of the cannabis user, whether this is an indicator or a risk factor for the evolution of a more severe and enduring mental condition remains open. While cannabis is gradually more accepted as a medical therapy, its association with various emotional and psychiatric effects should be considered, since it may affect the adherence and safety profile of the treatment, like in other psycho-affective therapies [21, 22, 47].

Another area for improvement relates to the knowledge and skills of users. Although the use of the app is intuitive, it requires the following essential skills and knowledge, which may introduce a bias: minimal familiarity with cannabis products, to identify the active ingredients THC and CBD, basic technology orientation to operate the app, and basic knowledge of English, since it is only available in English.

However, these requirements are trivial for most users. Medicinal and legalized cannabis programs require full label details on the product at licensed dispensaries as required by law. Therefore, as long as cannabis use is within the law, this info is always available to the users who can fill this data manually into the Releaf App. In some cases, the product label detail is pre-populated. In addition, the study utilized only basic data about the THC:CBD ratio, and therefore, it is highly conceivable that the Releaf app users are familiar with this consumption level. In addition, the app is currently only available in English, so there might be a selection bias.

In addition, several personal and use characteristics were unavailable, such as whether the cannabis use is medical or recreational, the exact dose taken by the user, comedications, and comorbidities. In addition, the mental state of the users at baseline is unknown. Although some users state that the reason they consume cannabis is self-medication, their actual mental condition is not known. It is essential to investigate further whether those in compromised mental states are at more risk for negative mental emotions after cannabis consumption.

Conclusions

Albeit its limitations, this is the first study based on a large-scale, real-time database acquired during cannabis consumption examining an association between cannabis use and negative emotional symptoms. It suggests that some characteristics of cannabis use, such as oral consumption and young age, are associated with negative emotion symptoms that can inflict cannabis users and cannabis patients. Cannabis use for mental reasons was associated with more negative emotional experiences than use for pain user. Further studies should examine the potential interface between cannabis consumption, characteristics of consumers, and negative emotional experience or perhaps even long-term mental disorders.