부산대학교 도서관

Observational studies have shown an association between the use of anticholinergic drugs and various negative health outcomes. However, when studying cognitive outcomes, there is great heterogeneity in previous results. The objectives of the present thesis are threefold. First, to explore the longitudinal patterns of anticholinergic prescribing in the UK. Second, to examine the association between anticholinergic burden and dementia. Third, to probe the relationship between anticholinergic burden, general cognitive ability, and brain structural MRI in relatively healthy participants. Chapter 1 provides an overview of the role of acetylcholine as a neurotransmitter in the human body. It begins with a description of its molecular characteristics and continues with a summary of anatomical and cellular features of cholinergic pathways in the brain. The chapter concludes with a description of the relevance of cholinergic processing in cognition and Alzheimer's disease. Chapter 2 gives a summary of anticholinergic drugs. It describes the history of anticholinergic compounds and their present use in medicine. It then appraises the tools used to gauge the anticholinergic potency of drugs. I conclude the Chapter by evaluating the available evidence on the effects of anticholinergic drugs on various important health outcomes. Chapter 3 focuses on UK Biobank, the sample used in all analyses presented in this thesis. The chapter briefly describes the conception of the study, the timeline of assessments, and the available variables. I focus in my descriptions on the variables that were used in the present thesis, especially cognitive tests, brain imaging, and linked health data. Chapters 4 to 6 present the empirical work conducted as part of this thesis. Chapter 4 presents an analysis of anticholinergic prescribing trends in UK primary care from the year 1990 to 2015. I first calculate an anticholinergic burden (AChB) according to 13 different anticholinergic scales and an average to derive a "Meta-scale". I then describe the prevalence of anticholinergic prescribing and its longitudinal trend for all scales. I use different plots of age-, period- and cohort effects on the AChB according to the Meta-scale to evaluate the contributions of these effects to the linear longitudinal trend. The study finds AChB to have increased 9-fold over 25 years and that this effect was attributable to both age- and cohort/period-related changes. In other words, ageing of the sample is not sufficient to explain the increase in anticholinergic prescribing; cohort- or period-effects must have contributed to the observed changes. Chapter 5 explores the relationship between anticholinergic prescribing and dementia. Previous studies on this topic had provided varied results. One of the goals of the present study was to probe potential factors for this heterogeneity. We find that greater AChB according to most of the studied anticholinergic scales (standardised HRs range: 1.027-1.125), as well as the slope of anticholinergic change (HR=1.094; 95% CI: 1.068-1.119), are associated with dementia. However, we find that not all drug classes are associated with dementia. Antidepressants (HR=1.11, 95% CI=1.07-1.17), antiepileptics (HR=1.07, 95% CI=1.04-1.11), and the antidiuretic furosemide (HR=1.06, 95% CI=1.02-1.10) exhibit the strongest effects. Interestingly, when exploring the effects of groups of anticholinergic drugs with different anticholinergic potencies, only the moderate potency group shows significant associations with dementia (HR=1.10, 95% CI=1.05-1.15). Chapter 6 examines the association between AChB, general cognitive ability, and brain structural MRI. It aims both to explore the potential sources of heterogeneity in previous work, as well as to expand on it by studying relatively healthy community-dwelling adults. We study brain structural MRI in a much bigger sample (at least 5x bigger) and use many more outcomes than previous studies. We find weak, but significant associations between AChB and general cognitive ability, and with 7/9 individual cognitive tests (standardised betas (β) range: -0.039, -0.003). Again, AChB in only some drug classes is associated with lower general cognitive ability, especially β-lactam antibiotics (β=-0.035, pFDR. Finally, chapter 7 summarizes the findings presented in chapters 4 to 6. The chapter also provides a critique of the sample and of my approach when conducting the analyses presented in the present thesis. The chapter concludes by discussing suggestions for future work on this topic.