Other studies suggested the substance might have benefits, but these studies were not as well designed as this trial.
The study was small, which reduces its ability to detect differences between groups. But the trend was for a greater reduction of symptoms in the placebo group than the THC group, suggesting that THC would not be expected to be better, even with a larger group.
People taking the THC pills did not show more of the typical side effects expected, such as sleepiness or dizziness. This led researchers to suggest the dose of THC may need to be higher to be effective. Further studies would be needed to determine whether a higher dose would be effective, safe and tolerable.
What did the research involve?
The researchers enrolled 50 people with dementia and neuropsychiatric symptoms. They randomly assigned them to take either a THC pill or an identical-looking inactive placebo pill for three weeks. They assessed symptoms over that time and looked at whether these differed in the two groups.
The trial initially intended to assess people who also had pain, but the researchers could not find enough people with both symptoms to participate, so they focused on neuropsychiatric symptoms. It also intended to recruit 130 people, but did not reach this number because of delays in getting approval for the trial in some centres.
About two-thirds (68%) of the participants had Alzheimer's disease and the rest had vascular dementia or mixed dementia. They all had experienced neuropsychiatric symptoms for at least a month. Both groups were taking similar neuropsychiatric drugs, including benzodiazepines, and continued to take these drugs during the study period.
People with a major psychiatric disorder or severe aggressive behaviour were excluded. Just over half (52%) lived in a special dementia unit or nursing home. The participants were aged about 78 years on average.
The pills contained 1.5mg of THC (or none in the case of placebo) and were taken three times a day for three weeks. Neither the participants nor the researchers assessing them knew which pills they were taking, which stops them influencing the results.
The researchers assessed the participants' symptoms at the start of the study, two weeks later, and then at the end of the study. They used a standard questionnaire, which asked the caregiver about symptoms in 12 areas, including agitation or aggression and unusual movement behaviour, such as pacing, fidgeting or repeating actions such as opening and closing drawers.
The researchers used a second method to measure agitated behaviour and aggression, and also measured quality of life and the participants' ability to carry out daily activities. They also assessed whether the participants experienced any side effects from treatment. The researchers then compared the results of the two groups.
What were the basic results?
Three participants did not complete the study: one in each group stopped treatment because they experienced side effects, and one in the placebo withdrew their consent for taking part.
Both the placebo and the THC pill groups had a reduction in neuropsychiatric symptoms during the trial. There was no difference in the reduction between the groups. The groups also did not differ in a separate measure of agitation and anxiety, quality of life, or ability to carry out daily activities.
Two-thirds of people (66.7%) taking THC experienced at least one side effect, and over half of those taking placebo (53.8%). The sorts of side effects that have been previously reported with THC, such as sleepiness, dizziness and falls, were actually more common with placebo.
How did the researchers interpret the results?
The researchers concluded they found no benefit of 4.5mg oral THC for neuropsychiatric symptoms in
people with dementia after three weeks of treatment.
They suggested the dose of THC used may be too low because the participants did not experience the expected side effects of THC, such as sleepiness.
Conclusion
This small phase II trial has found no benefit of taking THC pills (4.5mg a day) for neuropsychiatric symptoms in people with dementia in the short term.
The authors say this contrasts with previous studies, which have found some benefit. However, they note the previous studies were also limited in that they were even smaller, did not have control groups, or did not collect data prospectively.
The study was small, which reduces its ability to detect differences between groups. However, the non-significant trend was for a greater reduction of symptoms in the placebo group than the THC group.
The researchers note the improvement in the placebo group was "striking" and may be the result of factors such as the attention and support received from the study team, expectations of the participants on the effects of THC leading to perceived improvement, and training of the nursing personnel in the study.
While the authors suggest the dose of THC may need to be higher, further studies are needed to determine whether a higher dose would be effective and safe.