Is anxiety a risk factor in cognitive ageing?

Abstract

Background: The modifiability of anxiety, combined with the extraordinary worldwide growth in the prevalence of dementia, have motivated previous research which suggests anxiety may be a predictor of cognitive ageing. The aim of this PhD investigation is to extend the published research with new data, pool results with a fresh meta-analysis, and examine methods with a view to recommending strategies for future research. Method: The two, primary research questions, are: (1) Is anxiety a risk factor for the rate of age-associated, cognitive decline; and, (2) Is anxiety a risk factor for age-associated, incident, cognitive impairment? From published evidence on neuropsychological mechanisms, I developed "The Diathesis-Anxiety Heuristic of Cognitive Ageing". This model suggested a causal relationship between long-term anxiety and cognitive ageing and introduced the possibility of neuropsychological feedback loops which may serve as a control mechanism. My systematic review and meta-analysis updated previously published, pooled results. This statistical investigation was extended by drawing on new data from the Personality And Total Health (PATH) Through Life, dataset. PATH is an Australian, population based, prospective cohort study over four waves of data, at four-yearly intervals. Participants were aged 60 to 64 years at baseline, with sample size of 2,390. Analyses included multilevel modelling with stratifications and alternative temporal treatments, and testing for current and delayed effects of anxiety. Results: For anxiety as primary predictor, the only significant, meta-analysis result was for dementia as outcome, based on five studies: relative risk ratio (RR) = 1.81 (95% confidence Interval (CI): 1.22-2.70), p = 0.003, dispersion (I2) = 78.6%. From PATH, the only fully adjusted association found was for participants who consumed anxiolytics at baseline (n = 126). Anxiety symptoms were associated with working memory, with coefficient: 0.215 (CI: 0.001-0.429), p = .049. Discussion: For the meta-analysis, the dispersion percentage reflected high levels of methodological or sample differences between studies, and the result was, therefore, inconclusive. The result from PATH, for the anxiolytics stratification, was examined for the meaning of the direction of change, and for effect size among other criteria, and was found to be of marginal credibility. Analytical methods adopted in past research and in the operationalisation of anxiety, were likely to have contributed to the inconclusive nature of these results. Recommended future developments of methods are discussed to resolve these limitations. Additionally, all previous studies, including PATH, were observational. To establish causation, randomised control studies would be necessary, using treatment interventions, to determine if reversal of the risk factor is protective. Conclusions: A predictive association between anxiety and cognitive ageing has not been established. A strategic approach is recommended for future research which should include: (A) development of a more valid operationalisation of anxiety; (B) Statistical analysis methods which account for long term effects of anxiety; (C) Further investigation of the biological mechanisms and the possibility of neurological feedback loops; and, (D) placebo controlled, randomised anxiety treatment intervention trials, establishing whether there is a causal link between anxiety and cognitive ageing.