Patterns of anti-inflammatory drug use and risk of dementia: a matched case-control study.

Abstract

There is limited primary-care-based evidence about a potential association between anti-inflammatory therapy and dementia subtypes. The present study addressed this limitation by using electronic health records from a large primary care database. A case-control study was implemented using electronic medical records. Cases had a diagnosis of dementia between 1992 and 2014. Up to four controls matched on age, gender, family practice and index date were selected for each case. Use of non-steroidal anti-inflammatory drugs (NSAIDs) and glucocorticoid drugs represented the exposure variables. Primary outcome measures included all-cause dementia and main dementia subtypes, including Alzheimer disease (AD), vascular dementia (VaD) and Lewy body dementia (LBD). Data were analysed using conditional logistic regression. The study identified 31,083 patients with AD, 23,465 with VaD and 1694 with LBD. Ever-used NSAIDs were associated with a modest increase in the risk of all-cause dementia (odds ratio 1.04, 95% confidence interval 1.02-1.05, P < 0.006), whilst no association was apparent for ever-used glucocorticoids (0.98, 0.96-1.01, P = 0.152). There was no evidence for an association between NSAIDs and AD (1.03, 0.99-1.06, P = 0.07) or LBD (1.13, 0.99-1.29, P = 0.08). However, a significant increase in the risk for VaD (1.33, 1.29-1.38, P < 0.001) was observed. Similar patterns emerged for glucocorticoid therapy. In a large primary care population, there was no robust evidence for a potential association between anti-inflammatory drugs and risk of AD or LBD. NSAIDs and glucocorticoid drugs were associated with higher risk of VaD.