30 seconds summary
- Dementia affects more than memory; it can also cause agitation, aggression and emotional distress.
- A recent clinical trial found that a carefully controlled combination of THC and CBD significantly reduced agitation in people with late-stage dementia, with nearly 90% showing improvement after 12 weeks.
- While promising, cannabis is not a cure, and researchers warn that ordinary commercial products are not equivalent to the medically formulated treatment used under professional supervision.
Dementia is often described as a condition that affects memory, but memory loss is only one part of the experience. A person may also become anxious, withdrawn, suspicious, restless or frightened. They may have disrupted sleep, lose interest in familiar activities or become distressed during personal care. In some cases, agitation or aggression can place both the person and their family carers at risk.
These emotional and behavioural changes are among the most difficult aspects of providing dementia care in home settings. They can be exhausting for relatives, confusing for visiting care workers and deeply distressing for the person experiencing them. When reassurance, environmental changes and established treatments do not provide enough relief, families may begin looking for alternatives. Cannabis and cannabis-derived medicines are increasingly part of that conversation.
Early research suggests that certain cannabinoids may reduce agitation in some people with dementia. However, the evidence remains limited, the products studied are not interchangeable, and side effects such as sedation, dizziness, low blood pressure, confusion and paranoia may be especially dangerous for an older person living at home. No cannabis product has been approved by the US Food and Drug Administration specifically for dementia, and unregulated products have not undergone the same testing for consistency, safety and effectiveness as licensed medicines.
The most balanced conclusion is therefore neither that cannabis is a cure nor that it has nothing to offer. It is a developing area of research that may eventually provide another option for carefully selected patients—but only within a person-centred, medically supervised care plan.
Dementia affects feelings as well as thinking
Emotional changes in dementia are sometimes grouped under the clinical term “behavioural and psychological symptoms of dementia”. These may include anxiety, depression, irritability, apathy, agitation, aggression, sleep disturbance, hallucinations and delusions.
Although these symptoms may be caused partly by changes in the brain, it is important not to assume that every difficult behaviour is simply an unavoidable feature of dementia. Behaviour can be a form of communication, particularly when someone can no longer explain what is wrong.
A person who repeatedly tries to stand may be in pain, need the toilet or be looking for someone. A person who strikes out during washing may feel cold, exposed or frightened because they do not understand what is happening. Someone who calls for a deceased relative may be expressing loneliness or a need for security rather than asking for a factual correction.
A sudden increase in distress can also be caused by constipation, infection, dehydration, medication side effects, poor vision, hearing difficulties, hunger, fatigue or delirium. Changes in careers, household noise, unfamiliar visitors and disrupted routines may add to the person’s confusion.
For these reasons, current NICE guidance recommends a structured assessment before treating distress. This should look for clinical causes, environmental factors and unmet needs. Psychosocial and environmental approaches should normally be tried before medication, and personalised activities should be offered to people experiencing agitation or aggression.
Cannabis should not be used to silence behaviour that is communicating pain, fear or discomfort. The first question must be, “What might this person be experiencing?” rather than, “How can we make this behaviour stop?”
Why are cannabinoids being considered?
Cannabis contains many biologically active compounds. The two most widely discussed are tetrahydrocannabinol, or THC, and cannabidiol, or CBD. THC is the main intoxicating component and can alter mood, perception, memory and thinking. CBD does not generally produce the euphoria or “high” associated with THC, although it is not risk-free and can still cause drowsiness, mood changes, digestive problems, liver injury and drug interactions.
Researchers are interested in cannabinoids because the body’s endocannabinoid system is involved in processes connected with mood, stress, appetite, pain, sleep and memory. In theory, carefully adjusting cannabinoid activity might reduce agitation, improve sleep or ease pain that is contributing to emotional distress.
However, a biological theory is not the same as clinical proof. Substances that produce promising results in laboratory or animal studies frequently fail to provide meaningful benefits in people. It is also misleading to speak of “cannabis” as though it were one standard treatment.
A purified CBD medicine, a low-dose THC–CBD extract, a synthetic cannabinoid such as nabilone and a high-THC product purchased from a dispensary can have very different effects. Their strength, absorption, duration and side-effect profiles may differ considerably. Evidence from a clinical trial of one formulation cannot automatically be applied to another.
What has the research found?
One of the most frequently cited studies investigated nabilone, a synthetic medicine with effects similar to THC. The 2019 randomised, double-blind crossover trial included 39 people with moderate-to-severe Alzheimer’s disease and clinically significant agitation. Participants received nabilone and placebo during separate six-week periods.
On average, agitation scores, overall neuropsychiatric symptoms and caregiver distress were better during nabilone treatment than during placebo treatment. The findings suggested a possible therapeutic effect rather than sedation alone. Nevertheless, the study was small, short and conducted mainly among people with advanced impairment. Its average participant was approximately 87 years old, and the results may not apply to everyone with dementia.
The cognitive findings were also inconsistent. A brief mental-status measure favoured nabilone, while another assessment used among participants with severe impairment favoured placebo. Most importantly for home care, sedation occurred during 45% of nabilone treatment periods compared with 16% of placebo periods.
Sedation may appear beneficial when a person has been pacing, shouting or resisting care. Yet becoming quieter is not necessarily the same as becoming less frightened or enjoying a better quality of life. Excessive sleepiness can reduce conversation, mobility, eating and participation in meaningful activities. It can also increase the risk of falls and make it harder to recognise illness.
A 2023 retrospective study reviewed the records of 50 people with dementia for whom clinicians had recommended dispensary-obtained THC. Thirty-eight carers obtained and administered it. Among those 38 patients, carers reported improvement in neuropsychiatric symptoms for 30, or 79%. THC was most commonly used for agitation, aggression, anxiety, irritability and insomnia. Of 20 patients taking antipsychotic medication before using THC, 12 later reduced or discontinued the antipsychotic.
These figures sound encouraging, but the study was not a randomised trial. There was no placebo comparison, products and doses were not necessarily uniform, and improvement was based mainly on caregiver reports. Families who obtained and continued THC may also have been more likely to believe that it was helping. Reported adverse reactions included dizziness, worsening agitation and worsening paranoia.
A small phase-two trial published in 2025 examined a very low-dose, balanced THC–CBD extract in people aged 60 to 80 with Alzheimer’s-related dementia. After 26 weeks, the cannabis group performed better than the placebo group on a general cognitive screening test. However, there were no significant differences in the trial’s secondary outcomes. The results therefore do not show that the extract reliably treated anxiety, agitation or other emotional symptoms. The researchers themselves called for larger and longer trials.
A larger phase-three trial of nabilone for agitation in Alzheimer’s disease is still under way. The NAB-IT study plans to enrol 112 participants and, as of July 2026, is expected to finish in 2027. Until its results and those of other substantial trials are available, conclusions will remain uncertain.
Taken together, the evidence suggests that cannabinoids may reduce agitation for some people. It does not yet establish which patients are most likely to benefit, which cannabinoid is best, how long treatment should continue or whether the improvement outweighs the risks.
Which emotional symptoms might improve?
Agitation currently has the strongest—although still preliminary—human evidence. This can include pacing, repetitive movements, shouting, verbal hostility, resisting assistance or physical aggression.
Some carers have also reported improvements in anxiety, irritability and sleep. Cannabinoids might indirectly reduce distress when untreated pain, poor appetite or insomnia is contributing to the person’s behaviour. A person who is more comfortable may naturally appear calmer.
The difficulty is separating genuine emotional relief from intoxication or sedation. A successful treatment should ideally do more than reduce visible activity. It should improve comfort, engagement, sleep quality or the person’s ability to accept necessary care without causing an unacceptable loss of alertness and independence.
Evidence for treating depression, apathy, hallucinations or delusions is particularly weak. THC can itself provoke anxiety, suspiciousness, perceptual changes or hallucinations. A product that calms one person may make another more frightened or disorientated.
CBD is sometimes marketed as a gentler alternative, but there is not enough evidence to conclude that readily available CBD oils treat emotional symptoms in dementia. Commercial products may contain different quantities from those on their labels, and some may contain THC even when buyers are not expecting it. The FDA warns that non-approved cannabis products can have uncertain potency, purity and labelling.
Why the risks may be greater at home
Older people with dementia are particularly vulnerable to adverse drug effects. Many already have impaired balance, frailty, low blood pressure or difficulties judging hazards. They may also be taking several medicines for sleep, anxiety, pain, blood pressure, seizures or heart problems.
Nabilone’s official prescribing information lists possible effects including drowsiness, dizziness, hypotension, anxiety, panic, paranoia, hallucinations, changes in mood and reductions in cognitive performance and memory. It can also have additive sedating effects when combined with alcohol or other drugs that depress the central nervous system.
In a supervised research unit, staff can observe subtle changes and respond quickly. In dementia care at home, the risks may be harder to manage. A mildly dizzy person may fall while going to the bathroom. Someone who is unexpectedly confused may leave the house, attempt to use a cooker or become distressed by a familiar carer. Excessive sleepiness can make swallowing, transferring and personal care more difficult.
Because the person may be unable to describe side effects, carers must watch for changes in walking, facial expression, appetite, wakefulness, speech and social interaction. A reduction in agitation accompanied by vacant staring, unsteady movement or sleeping through meals would not represent an uncomplicated success.
Product variability creates another problem. Most research has involved measured oral preparations or prescription medicines. Results cannot safely be generalised to smoked cannabis, vaping products, homemade foods or unregulated oils. Edible products may take time to have an effect, making accidental repeat dosing possible. They must also be stored securely so that the person with dementia, children, visitors or pets cannot consume them unintentionally.
Consent and dignity matter
A person with dementia should be involved in treatment decisions as fully as possible. Capacity is specific to the decision and may fluctuate. Some people can understand a simple explanation about trying a medicine for distress even if they need support with more complex choices.
Where a person cannot make the decision independently, families and clinicians must follow the relevant legal process for making decisions in the person’s best interests. The purpose should be relief of distress and improvement in wellbeing—not making care more convenient for others.
Secretly adding a cannabis product to food or drink raises serious ethical and legal concerns. Covert administration should never be treated as an informal family decision. It requires an appropriate capacity assessment, a best-interests process and professional guidance under the laws of the country concerned.
Legal access also varies considerably between jurisdictions. A product being available from a shop or dispensary does not mean that it has been proven safe for dementia. In the United States, the FDA has approved certain cannabinoid medicines for specific seizure disorders, chemotherapy-related nausea and AIDS-associated weight loss, but none for dementia.
A safer framework for home dementia care
Before considering any cannabinoid, the care team should identify one or two clearly defined target problems. “Difficult behaviour” is too broad. A more useful description might be, “Paces and calls for her mother for two hours each evening,” or, “Becomes frightened and strikes out during bathing three times a week.”
The next step is to look for causes. Sudden or severe changes require medical assessment, particularly when accompanied by fever, pain, reduced fluid intake, urinary symptoms, constipation, weakness or a change in consciousness. Delirium can resemble a rapid worsening of dementia but may be caused by an urgent and treatable illness.
The home environment should then be reviewed. Predictable routines, gentle lighting, reduced background noise, familiar objects, comfortable clothing and consistent carers can lower confusion. Regular meals, hydration, toilet visits, movement and pain treatment can prevent avoidable distress.
Personalised activity is equally important. Music connected with the person’s past, folding laundry, gardening, walking, baking, looking through photographs or holding familiar objects may provide reassurance and purpose. NICE recommends activities that reflect the individual’s life experiences, preferences, abilities and needs. It also recommends multicomponent sleep support involving sleep hygiene, daylight exposure, exercise and personalised daytime activities.
A medication review should follow. Sedatives, anticholinergic drugs, opioid painkillers and some bladder, allergy or sleep medicines can worsen confusion. Starting another psychoactive substance without reviewing the existing combination may make the situation more dangerous.
When severe distress continues despite these measures, a clinician with expertise in older adults or dementia can discuss the recognised treatment options, their evidence and their risks. Cannabis should not be tried casually alongside several simultaneous changes because it will become impossible to know what helped or caused harm.
Where a qualified clinician considers a cannabinoid appropriate and legally available, treatment should use a defined product with reliable concentrations. The principle should be cautious initiation and slow adjustment rather than pursuing rapid sedation. Families should not copy a dose from another patient or use trial dosages as self-treatment instructions.
A daily record can help the clinician judge the result. It might document the target behaviour, its duration and severity, sleep, appetite, alertness, walking, falls, hallucinations and the amount of assistance required. Carer wellbeing should also be recorded, but a reduction in caregiver burden must be considered alongside the person’s alertness, comfort and participation.
The plan should include stopping criteria. Treatment needs urgent review if there is marked drowsiness, fainting, repeated falls, severe confusion, new hallucinations, worsening aggression, inability to eat or drink, or a major change in mobility. Emergency help may be needed for collapse, difficulty waking, breathing problems, chest pain, seizure-like activity or sudden neurological symptoms.
Cannabis should not replace compassionate care
Medication can sometimes reduce severe symptoms, but emotional security in dementia is created largely through relationships and surroundings. A calm voice, unhurried care, familiar routines and validation of the person’s feelings may be more effective than repeated attempts to correct their beliefs.
Instead of saying, “Your mother died years ago,” a carer might respond, “You miss your mother. Tell me about her.” Instead of insisting on a bath at a fixed time, they might offer a wash later, use warmed towels or allow the person to choose between two options. Such adjustments protect dignity while reducing confrontation.
Family carers also need practical and emotional support. Exhaustion can make even the most compassionate person less able to respond calmly. Respite care, dementia-care training, support groups and help from community or home-care services are not optional extras; they are part of effective treatment.
Conclusion
Could cannabis help with the emotional challenges of dementia? Possibly for a limited group of people, under careful clinical supervision. The strongest early signal concerns agitation, particularly with measured THC-like medicines such as nabilone. Yet the research involves small numbers, short treatment periods and inconsistent outcomes. Sedation is common enough to be a serious concern.
Cannabis does not cure dementia, restore lost memories or replace an investigation into pain, illness, loneliness and environmental stress. CBD oils and other commercial products should not be assumed safe simply because they are described as natural or non-intoxicating.
For families providing dementia care in home environments, the goal should never be quietness at any cost. It should be greater comfort, dignity, connection and safety. Cannabinoids may eventually become one carefully controlled tool within that broader approach. At present, they remain a promising but unproven option—one that requires professional oversight, honest monitoring and a willingness to stop when the harms outweigh the benefits.
About Author
