Intra-individual changes from before the pandemic to during the pandemic are of significant interest in terms of understanding the dynamic relationship between mental health factors and cannabis consumption in the context of a global crisis. Here, we have utilized a retrospective approach with data sampled in the early stages of the first wave of the pandemic. Overall, our findings revealed a significant increase in stress, anxiety, and depression symptom severity among all cannabis groups during the first wave of the pandemic relative to self-reported retrospective pre-pandemic estimates. Only overall depression scores differed significantly among the cannabis groups. Specifically, we found that compared to non-users and those who maintained a stable frequency of cannabis use during the pandemic, those who decreased or increased (trend) their cannabis use experienced higher depression symptoms both before and during the pandemic (i.e. no significant interaction with pandemic-related changes in depression). Those who used cannabis but did not change their frequency of use during the pandemic maintained the lowest QIDS-SR16 scores across this period. Our analysis also revealed that clinically important worsening of depression (based on retrospective estimates) was related to increased cannabis use in individuals with a self-reported current diagnosis of PTSD; no such association was observed for stress or anxiety.

Our findings indicated that 72% of those who increased their cannabis use during the pandemic experienced clinically important worsening of depression symptoms (based on retrospective estimates), compared with 40–50% for all other groups. Since we did not observe any significant interaction between cannabis use and retrospectively estimated pandemic-related changes in continuous variables reflecting symptom severity in our first analyses, the relationship between depressive symptoms worsening and increased cannabis use during the pandemic may have been more pronounced in cases of prominent and clinically significant symptom worsening rather than in cases of sub-clinical symptom changes. Lastly, we did not observe any significant differences in the proportions of those who experienced clinically important worsening of stress or anxiety among the cannabis use groups. Thus, it appears that depression symptoms alone were chiefly associated with changes in cannabis use patterns in the face of this external stressor.

Our findings with regards to depression may be consistent with other research on cannabis-use patterns during the pandemic. Recent evidence has identified self-isolation and coping with depression as primary drivers behind cannabis buying practices during the pandemic in a non-PTSD sample [68]. Our work would therefore appear to support those results for individuals with PTSD who experienced clinically meaningful worsening in depression symptoms (based on retrospective estimates), suggesting that depression symptoms are an important factor associated with cannabis intake in PTSD during a global crisis (an observation that is particularly relevant given the high comorbidity of PTSD and depression [69, 70]).

Our findings regarding stress may also be somewhat consistent with past research. Previous work has demonstrated for example that cannabis use may be associated with a blunted stress response [71]. Regarding anxiety, however, the evidence is less clear-cut. In the short-term, high doses of tetrahydrocannabinol (THC, the major psychoactive constituent of cannabis) are anxiogenic [72] while low doses are anxiolytic [42, 73, 74]. Meta-analyses indicate that, in the long-term, there is a weak but positive relationship between cannabis use and anxiety [75] (although it is unclear whether anxiety is a main motivating factor for cannabis use in PTSD specifically [76]). Some work has also suggested that the largest increases in medical cannabis intake during the pandemic were observed in participants with pre-existing anxiety [77]. We did not observe this in our sample, however, as there were no significant differences between our cannabis groups in terms of either GAD-7 scores or comorbid anxiety disorders. Interestingly, multiple studies have indicated that the relationship between cannabis and anxiety may also be specific to social anxiety disorder [78, 79]. While our questionnaire measures did not specifically isolate social anxiety symptoms, we did not observe high levels of self-reported current diagnoses of social anxiety disorder in this sample among any of the cannabis-using groups.

It remains unclear whether cannabis use may increase as a result of therapeutic use/self-medication or whether it instead may be a driving factor behind symptoms changes. Given the evidence of the association between the pandemic and worsening of mental health [80–82], it is possible that cannabis consumption was increased as a coping strategy in response to mental health challenges imposed by this global external stressor. Previous evidence has highlighted coping as a motive of cannabis use [76, 83–85], and recent work has more specifically indicated that cannabis users with greater difficulty tolerating sad mood states use cannabis more frequently [85]. Thus, one possibility in the current study is that, among those with lower resiliency to sad mood states, worsening of depressive symptoms may have motivated increased cannabis use as a coping strategy during the onset of the pandemic. However, research also points to cannabis as a risk factor for depressive symptoms [86]. The relationship between cannabis use and depression may thus be bidirectional. It is also worth noting that comorbid substance use could also have played a role in these effects (since other studies have reported increased use of both alcohol and cannabis among substance-using individuals during the pandemic [30]).

Interestingly, those who reported not changing their levels of cannabis use during the onset of the pandemic also reported the lowest overall depression symptoms. Indeed, some positive effects of cannabinoids have been shown for PTSD (e.g., the reduction of PTSD-related nightmares by Nabilone [12]). Since PTSD may be characterized by reduced levels of the endocannabinoids anandamide and a compensatory upregulation of central CB1 receptors [87], it is possible that supplementation with stable consumption of THC (which, like anandamide, acts as a CB1 partial agonist [42]) could re-normalize the central endocannabinoids system. Conversely, acute changes in cannabis consumption may alter this balance. However, more work is needed in this area since there is a dearth of information on the effects of THC on CB1 receptor expression in humans.

The present study has several limitations. Firstly, the information about cannabis use was limited to weekly frequency of use, and no information was available about other important factors such as dosage, the ratio of CBD/THC concentrations, or medical versus recreational use which are speculated to play an important role in responses to cannabis use [88]. This limitation is particularly important given the known differences in the effects of dosage and CBD versus THC concentrations on behaviour and emotion as well as recent concerns raised about the lack of adequate standardization of cannabinoid-based products [10, 89, 90]. Secondly, this study relied on self-reported diagnosis without clinician-based confirmation, and it was also based on retrospective assessment of pre-pandemic mental health and substance use. Because of the spontaneous nature of the pandemic, obtaining a prospective sample from before the pandemic was not possible. Our results may therefore be affected by recall bias. In addition, the nature of self-report measures is such that participants could omit or misreport information pertaining to their diagnoses or cannabis intake (which is a concern in all studies relying on self-report measures). Thirdly, the generalizability of this research is also somewhat restricted by the high proportion of females. Mian et al. [91] recently reported that females are significantly more likely to participate in survey-based cannabis research than males. Although the current data is issued from a larger survey which was not focused on cannabis, it is possible that volunteer bias influenced participant’s responses to this portion of the survey. For future research, it may be important to achieve more representative samples by oversampling groups which are more difficult to reach. This sample was also biased towards middle age white individuals, many with high total family income levels. Fourthly, while our study had adequate sample sizes overall, a larger sample would have allowed for more in-depth characterization of the data across the groups (particularly with regards to comparing those with clinically meaningful worsening in mental health (based on retrospective estimates) versus those who did not in a between-groups manner). Fifthly, several statistical tests were conducted, increasing the risks of type I error, and the observational nature of the study precludes any inference of causality. Finally, while the data reported in this sample were restricted to individuals with self-reported PTSD, we cannot rule out the possibility that our findings could also be more generalized to other populations. Overall, our findings support and expand upon prior work regarding substance use in individuals with PTSD and during the COVID-19 pandemic.