Understanding multiple medication use in people with dementia. A pharmacoepidemiology study of prevalence, intervention and associated harms.

Background: Around 70% of people with dementia live with comorbidities and are subsequently prescribed multiple medications (polypharmacy). Some medicines are considered ‘potentially inappropriate’ when prescribed to older and cognitively impaired adults. Cognitive impairment, dementia and frailty may influence the prevalence and associated effects of polypharmacy and potentially inappropriate medicines (PIMs). Aims: To estimate prevalence trends of polypharmacy and PIMs (antipsychotics, antidepressants, benzodiazepines, anticholinergics and proton pump inhibitors) among people with dementia. To identify associated factors, including the impact of care and medication reviews and to understand the impact of polypharmacy, PIMs and frailty on health in people with cognitive impairment. Methods: Two cohorts were analysed, including primary care electronic medical records (Clinical Practice Research Datalink, 2015-2017) (n=22,448) and the Cognitive Function and Ageing Study (2008-2011) (n=1,154). Results: Polypharmacy and PIMs were prevalent in people with dementia and cognitive impairment. On average, people with dementia were prescribed 8 medications, 30% were prescribed inappropriate PPIs, 17% anticholinergics, 8% antipsychotics, 7% tricyclics and 4% of people with cognitive impairment were prescribed benzodiazepines. Dementia annual reviews and medication reviews were associated with medicines optimisation. Prevalence of PIM was greater in care homes and a medication review in a care home was associated with reduced use of PIMs. Polypharmacy was associated with worse survival. PIMs were not associated with worse survival, with the exception of antipsychotics (adjusted HR=3.24, 95% CI=1.83-5.73). Being cognitively impaired and frail was associated with worse survival overall but frailty was not found to moderate the relationship between polypharmacy, PIMs and survival. Conclusions: Few prescribing guidelines specifically address medicines use in people with cognitive impairment or dementia, despite the prevalence of polypharmacy and PIMs. The number of medicines prescribed should be carefully monitored to reduce harm. Incorporating medication reviews into annual dementia care reviews may optimise prescribing and identify people at increased risk of adverse effects. The findings from this thesis will improve understanding and support the optimisation of medicines for people living with dementia