For two of the prespecified primary feasibility thresholds the STAND trial demonstrated the safety and feasibility of conducting a randomised, placebo-controlled study of cannabinoid oral spray for agitation in patients with AD. Despite under-recruitment, most screened patients participated, with high retention and adherence throughout the trial (to the treatment and for completing outcome measures). No AEs attributable to the intervention were identified, and the safety profile appeared to be similar between the nabiximols arm and the placebo arm. Unfortunately, it is not possible to determine whether the target recruitment would have been possible due to pandemic related recruitment difficulties, and therefore, the estimated effect size calculation was also compromised. A future study should incorporate interim analyses to assess these criteria.

The recruitment challenges should be considered in the context of recruiting in care homes during the COVID pandemic. Critically, the initial planned 12-month recruitment window was reduced to 7.5-months, resulting in 4.5 fewer months than originally planned to reach the recruitment target of 60 participants (see Table 2 for precise dates of first and last randomisations). Even within these 7.5-months, recruitment had been paused because of the Omicron wave in the earlier months, and for ~6-weeks to resupply our stock of drug and placebo (which had expired since first receiving in early 2020 when the trial was initially planned to start). Further, the active recruitment window could be argued to be closer to ~6 months. Accounting for only having potentially ~50% of the planned recruitment window, intermittent trial pauses, and with our recruitment rate accelerating in the later months, we cautiously argue that it may be that the sample target would have been reached if the trial was operational for the full planned 12 months. The unprecedented burnout among care home staff globally limited the capacity for clinical research [16], even as residents likely experienced escalating BPSD [17]. Additionally, the pandemic contributed to attrition through increased mortality, participant ineligibility due to presence of covid symptoms and care home closures across the UK—all of which impacted the recruitment and randomisation of the study. Despite challenges, retention reached 100% and data completion for key neuropsychiatric outcomes was robust, supporting the feasibility of administering this intervention.

With regard to the safety profile, although all the participants in this study were classified as either moderately severe or severe dementia, and the majority of the participants (100% in placebo arm and 73% in nabiximols arm) had moderate to severe frailty, there were very few reports of AEs. A similar study conducted by our team, applying pure CBD of a much higher dosage (up to 600 mg daily) in a modest sample, found that 12.5% of participants in the CBD arm had treatment-related dizziness [18]. Another small-sized study done in Australia using higher dosage THC/CBD mixture (up to 50-mg THC and 34-mg CBD daily) also found a significantly higher rate of overall AEs in their treatment arm [19]. Compared to these studies, ours applying a mixture of low dose CBD and medium dose THC appeared to be safer and better tolerated, even in a cohort of frail, late-stage ad patients.

Safety remains the foremost concern in designing interventions for BPSD. Antipsychotics, benzodiazepines and antidepressants are commonly used empirically for these symptoms. However, prior to the approval of brexpiprazole—a partial agonist at dopamine D2 and D3 receptors—for agitation in AD, no medication was considered safe for long-term use [20]. Even brexpiprazole carries a black box warning, highlighting the increased risk of mortality associated with antipsychotic use in older adult patients with dementia-related psychosis [7]. Benzodiazepines are known for their negative impact on the cognition and postural balance of older adults [21]. Antidepressants like citalopram and mirtazapine have limitations too: citalopram may reduce agitation but can impair cognition and prolong QT interval [22], while mirtazapine showed no benefit for agitation in a large RCT and might increase mortality risk [23]. Given the favourable safety and tolerability profile shown at the current dose, we would recommend a further dose-finding study to establish the minimal effective dose before initiating a larger phase-III trial. The STAND trial focused on a nursing home population, allowing for continuous safety and compliance monitoring. However, based on our positive safety results, future studies should consider including community-dwelling participants to enhance recruitment and reach a broader patient population in need.

While the adjusted analyses indicated little evidence of a signal from nabiximols, this study was not powered to demonstrate a clinical effect. The small sample size and lack of precision should be considered when interpreting the effectiveness of the intervention, along with the finding that only half of participants completed the full titration schedule as planned. In addition, agitation is likely a heterogeneous and multifactorial syndrome that needs to be defined and measured with more specificity, as recommended previously by the SYMBAD study [23]. Although using the questionnaires such as the NPI and the CMAI for agitation assessment have been widely validated and implemented, wearable device has been proposed to complement these traditional instruments [24, 25]. These device-based measurements might have the potential to reveal information associated with the outcome not easily captured by the questionnaires and therefore stratify the patients and outcomes more specifically. For example, the International Psychogeriatric Association recently suggested the recognition and separation of nocturnal/circadian agitation [3]; wearable devices capable of continuous behavioural measurement might have the potential to distinguish symptoms happen in daytime and at night and therefore facilitate the outcome measurement.

This trial introduces two key innovations relevant to the feasibility, safety and efficacy to be considered for future studies: the use of an oromucosal spray for drug delivery and the application of cannabinoids in dementia care in a dementia cohort with predefined nonpain-related agitation.

It has been noted in a target product profile developed by a consortium of experts that an ideal therapeutic for agitation in ad would address difficulty swallowing (dysphagia) and refusal to swallow [26]. Administering cannabinoid-based medicine (CBM) in dementia patients through oral mucosae bypassing the first-pass effect of liver and the gastrointestinal tract would theoretically result in higher bioavailability, more direct neuropsychiatric response and less adverse effects from the gastrointestinal tract (particularly associated with THC), in addition to preventing swallowing-related issues. To our knowledge, only five RCTs have been published previously examining the effect of cannabinoids on BPSD [8, 19], and ours is the first one testing the oromucosal administration of medication. Addressing common challenges like dysphagia and refusal to swallow, and bypassing the gastrointestinal tract likely contributed to the high tolerability and adherence observed, as well as the very low adverse effect recorded, despite a moderately high dose of THC used in our study.

In addition, our study design recruiting patients with predefined clinical agitation with little or no pain symptom, was crafted to make more precise inference on agitation. However, it potentially nullified the effect, as previous studies show that agitation associated with unpleasant sensory input such as pain might be more responsive to CBM [8]. Future studies might compare nabiximols in agitated dementia patients with and without pain to understand its true effect. Nabiximol might also be considered as an adjunct to sensory interventions—such as music, light and aroma therapy [27, 28]—leveraging CBM’s potential to enhance sensory experiences and regulate pain [11, 12] and circadian rhythm [29, 30]. This combined CBM-sensory approach could enable gradual deprescription, supporting long-term, nonpharmacological management of agitation in ad.

The primary limitations of this study include under-recruitment due to the COVID-19 pandemic, which reduced the statistical power needed to determine effect size, under-dosing, racially homogenous sample, a failure of stratified block randomisation to ensure equal baseline severity of agitation between groups, inadequately tracking rescue medication administration and an over-reliance on proxy-informant measures as a result of pandemic restrictions. Although our sample was slightly larger than those in comparable studies and featured a more balanced sex distribution, the participant numbers remain insufficient to robustly estimate the intervention’s effect size. Additionally, while the relatively low doses of THC and CBD likely contributed to fewer AEs compared with prior studies, these dosages may also be less effective in reducing agitation and aggressive behaviours. The limited sample size warrants cautious interpretation of the safety and efficacy profile in the nabiximols group. The ethnic homogeneity of our sample may limit the generalisability of these findings beyond the UK and Ireland, particularly given the high proportion of white British participants. Also, as our study had significant between-group differences in agitation severity at baseline, a future study should consider a more precise stratification variable other than the FAST to balance baseline agitation severity between groups. Additionally, while we recorded the medications that participants were on at the start the trial, and communicated with the care home that participants must remain on the same dose of their antidepressants, antipsychotics, anti-epileptics, benzodiazepine, lithium and hypnotics dosage, we did not adequately monitor whether rescue medications were administered ‘as needed’ in response to acute episodes of agitation or aggression during the trial period, and thus we do not know if the two groups differed in the administration of ‘as needed’ medications during the study. This level of data should be included for future studies trying to estimate the true effect of the medicinal compound. Finally, an over-reliance on proxy measures meant that we did not collect adequate direct participant clinical information. This is particularly prominent for the sMMSE, a cognitive assessment that we included at the outset as a precautionary exploratory measure, just in case there was an effect (positive or negative) on cognition. Due to pandemic restrictions, it was extremely difficult to complete face–face measures, rendering it uninformative.

This study has three main strengths: a robust design, a favourable safety profile and a high retention rate. Strict inclusion and exclusion criteria, while reducing participant numbers, safeguarded this vulnerable population from SAEs and minimised infection risk among participants, caregivers and researchers. The inclusion of participants with no or minimal pain symptoms also enhances our understanding of the true effect of CBM on agitation in dementia patients. The observed safety profile in this frail, late-stage dementia cohort likely contributed to the 100% retention rate, underscoring the intervention’s feasibility even amid pandemic-related pressures on care home staff. Although, it should be noted that the threshold of at least one dose per week should be updated to be more stringent (perhaps one dose per day).

In conclusion, this study demonstrates the feasibility of a pilot randomised, placebo-controlled trial of nabiximols oral spray for agitation in AD patients in care homes, with no safety concerns observed. The study addresses a novel gap in the literature and clinical practice in the treatment of agitation in AD as (i) it was the first to investigate the social acceptability and tolerability of CBM within care homes in the UK, providing valuable insights into practical administration aspects; (ii) it was the first to evaluate an oral spray formulation of a CBM containing a moderately high dose of THC and a low-dose CBD specifically for dementia patients; and (iii) the clearly predefined. The findings provide insights that can inform future trials in refining dosing schedules and minimising AE risks. However, establishing the intervention’s effect size will require further investigation to assess therapeutic efficacy more definitively. We recommend a larger confirmatory trial be undertaken, incorporating insights from this feasibility study (such as ascertaining a maximum-tolerated dose in this specific patient population), to evaluate nabiximols therapeutic efficacy to treat agitation in AD.