In this semi-naturalistic, open-label laboratory study, we investigated cognitive performance before and after self-administration of a prescribed medical cannabis product. We found no evidence for impaired cognitive function when comparing baseline with post-treatment scores on a comprehensive neuropsychological test battery, nor did we observe any change in performance on the Druid test battery over time. The absence of evidence for cognitive impairment following medical cannabis self-administration was surprising, given prior and substantive evidence that non-medical (‘recreational’) cannabis use reliably impairs a range of cognitive functions [3]. At the same time, these findings are consistent with two systematic reviews published in the last year that suggest that medical cannabis, when used regularly and consistently for a chronic health condition, may have little if any impact on cognitive function [10, 11].

Wieghorst et al. [10] conducted a systematic review of 23 studies (total participant pool of N = 917) that administered any medical cannabis product to patients and compared cognitive performance under treatment and control (placebo or baseline) conditions. Participants were mostly female (n = 448) and the mean age ranged from 33 to 65 years. Treatment periods varied greatly from 1 day to 12 months. Fifteen studies found no impact of medical cannabis on cognitive function, while one reported an improvement and six reported impairments. Two of the studies that reported no cognitive impairment involved treatment with a CBD-only product that would not be expected to impact cognitive function [17, 18]. Of the six studies that reported impairment, four were characterized by administration of a single dose of cannabis via smoking or vaporization, and a requirement that patients avoid cannabis for 30 days prior to participating [19–22]. The fact that these patients were not taking a stable dose of medical cannabis might explain why we found no evidence for acute cognitive impairment when comparing baseline with post-treatment scores among patients who self-administered cannabis via vaporization.

We found little evidence in this study to suggest that magnitude of change in cognitive function differed with product type, apart from one measure (probability of a hit) on the RVP, where only patients who self-administered flower exhibited a significant improvement over time. Given the mean THC dose was higher for flower products (37.00 mg vs. 9.61 mg for oils), the observed improvement in performance, although minimal, is somewhat perplexing and might reflect a practice effect rather than true improvement following acute medical cannabis administration. It is also possible that any potential impairment had subsided by 3 h among patients who self-administered flower, noting that subjective drug effects were strongest at 1 h. Our rationale for having a second CANTAB assessment at 3 h and not sooner was based on two factors: (1) government data at the time indicated that most prescriptions were for orally administered products rather than flower; and (2) cognitive impairment following cannabis inhalation can still persist for up to 3–5 h despite the fact that self-reported intoxication tends to be strongest within the first 1–2 h [17, 23].

The recent systematic review by Motaghi et al. [11] examined studies that used an oromucosal spray containing an equal amount of THC and CBD, and compared cognitive performance under treatment and control (placebo or baseline) conditions. The mean age of patients ranged from 29 to 51 years, and again there was considerable heterogeneity in treatment duration, which ranged from 1 day to 12 months. The total number of sprays used by patients ranged from 4 to 16 per day, delivering approximately 10–40 mg THC and CBD, which is comparable with the range of doses used by patients in this study. Of the 10 studies that were included (total N = 510), seven included patients with multiple sclerosis. No evidence was found for cognitive impairment when comparing performance under treatment and control conditions; only one study reported impairment, and this was specific to a measure of long-term memory storage [24]. There has been speculation that CBD may attenuate THC effects [25, 26], which could explain the lack of cognitive impairment observed in the studies included in the review, but recent evidence suggests that this is unlikely and that coadministration of CBD may even increase blood THC concentrations [17, 27–29]. A wide range of cognitive tests were administered across the included studies, as was the case in the previous review [11] and in the study reported here, suggesting the lack of effects of medical cannabis on cognitive function is not specific to particular cognitive domains.

With increasing medical cannabis use in Australia and other international jurisdictions, there is considerable interest in the development of novel methods that might allow patients to assess their cognitive function in relation to their own baseline before performing safety-sensitive tasks such as driving [30]. The Druid app holds promise as it can be completed using a smartphone or tablet and takes only minutes to complete. Recent studies have indicated that it is indeed sensitive to cannabis intoxication in healthy volunteers [15, 16], but representative patient data are lacking. The changes in performance from baseline to +3 h (0.87 points) and +5.5 h (0.67) reported here are markedly smaller than the peak changes previously seen in healthy volunteers, with THC doses ranging from 5 mg (vaporized; +1.8 points) to 20 mg (vaporized; +9.0 points) to 25 mg (oral; +10.1 points) [15]. The fact that participants exhibited no impairment on either the Druid or the CANTAB tasks suggests concordance between these two test batteries. The Druid app may therefore be a useful tool for impairment screening, although further validation work is clearly needed. One notable caveat with the Druid app that has relevance for real-world use is the inclusion of a balance component that patients in the current study often found difficult due to age or condition-related balance issues, or due to exacerbation of pain.

While we found no evidence for cognitive impairment in this study, we did observe a change in subjective drug effects over time, with participants rating themselves as significantly more ‘stoned’ and ‘sedated’ and significantly less ‘anxious’ following medical cannabis self-administration. While significant, the overall magnitude of change in perceived intoxication (i.e., ‘stoned’) was considerably lower than the peak change seen in healthy volunteers administered 13.75 mg THC via vaporization [17], and closer to the peak change associated with a 13.75 mg dose of CBD (also vaporized), which is considered to be non-intoxicating [17]. There were however some differences in subjective drug effects depending on the product type. Inhalation of flower via vaporization was associated with feeling more stoned and sedated, likely due to the higher mean dose of THC in vaporized products relative to oils (37.00 mg vs. 9.61 mg) and a more rapid onset of effects relative to rapid absorption of THC through the lungs into the blood stream [31]. Interestingly, medical cannabis oils were not associated with a significant increase in sedation relative to baseline at any timepoint. By way of comparison, a 10 mg oral dose of THC (comparable with the mean THC dose of 9.61 mg here) elicited a significant ‘drug effect’ rating in healthy volunteers who had not used cannabis within the previous 3 months [13].

The absence of any ostensible cognitive impairment, despite reports of mild intoxication for up to 4 h, and particularly with vaporized products, might reflect acute symptom alleviation leading to an improvement in physical and/or psychological function, or potentially tolerance to the impairing effects of cannabis [3, 32]. Tolerance, which arises with repeated and long-term cannabis use, has been demonstrated in previous laboratory studies with non-medical cannabis [33–35] and has been comprehensively documented in recent reviews [36, 37]. In a neuroimaging study by Mason et al. [38], occasional cannabis users administered 300 μg/kg THC exhibited significant alterations in reward circuitry, including reduced functional connectivity and increased striatal glutamate, as well as impaired performance on a sustained attention task. Chronic cannabis users administered the same dose of THC did not exhibit these same neurometabolic alterations or performance decrements, despite reporting a significant increase in intoxication relative to placebo.

Considering this emerging evidence for pharmacodynamic tolerance to the effects of cannabis, stable dosing with THC (and gradual dose titration up until, and not beyond, the point that effective symptom relief is achieved) is likely critical for any potential mitigation of impairment. So long as medical cannabis is prescribed within a controlled, medical framework, this is something that can be effectively managed by the prescribing physician. Performing safety-sensitive tasks, such as driving, is therefore not advised until patients are taking a stable dose of THC, and should also be temporarily avoided following any increase in THC dose [39, 40]. Future studies might consider tracking cognitive performance over a longer period in patients commencing medical cannabis treatment, focusing on specific health indications or other cannabinoid preparations, or recruiting patients from different age groups.

While this study benefits from a semi-naturalistic design that allowed us to quantify acute changes in the cognitive function of patients prescribed medical cannabis using a comprehensive neuropsychological test battery, the lack of a placebo control means we were unable to disentangle baseline performance from possible residual impairment resulting from medical cannabis use the evening prior. The use of a semi-naturalistic design gives this study high ecological validity and also introduced considerable variability in the type of medical cannabis products used, THC/CBD dose, route of administration, and health indication. We aimed to capture potential impairment resulting from either inhaled or oral medical cannabis by having an assessment at 3 h, but we recognize that impairment could have already subsided by this point in patients who were using flower. This means we cannot rule out the possibility of cognitive impairment 1–2 h after inhaling medical cannabis. It is also possible that participants improved on the CANTAB over time due to a practice effect, which may have masked potential performance decrements. Future studies might consider assessing cognitive function at different timepoints based on the type of product that patients are using and its expected peak effects, as well as assessment of baseline cognitive performance prior to initiation of medical cannabis treatment.

Accompanying blood and oral fluid THC data will be reported separately along with driving performance outcomes. The decision to present these data separately was based on the number of outcomes already being reported here, and the greater relevance of blood/oral fluid THC levels for driving laws. We also note that participants were not required to undergo urine or oral fluid screening for recent drug use prior to study initiation. This was because we expected some patients to be using prescribed medications (e.g., opioids and benzodiazepines) that can be detected in standard urine and oral fluid drug screens. Finally, we note that patients’ use of medical cannabis in this study may not perfectly reflect their real-world use. For instance, patients may typically only use medical cannabis in the evening before sleeping. As patients had been using medical cannabis for more than 10 months on average, these findings cannot be generalized to patients who are just commencing treatment with medical cannabis or changing dose/product type. Given our small sample size and open-label naturalistic design, larger and controlled trials are needed to confirm these findings.