The Association of Anticholinergic Drugs and Delirium in Nursing Home Patients With Dementia: Results From the SHELTER Study.

Objectives: Drugs with anticholinergic properties are associated with an increased prevalence of delirium, especially in older persons. The aim of this study was to evaluate the association between the use of this class of drugs in nursing home (NH) patients and prevalence of delirium, particularly in people with dementia.

Design: Cross-sectional multicenter study.

Setting And Participants: 3924 nursing home patients of 57 nursing homes in 7 European countries participating in the Services and Health for Elderly in Long TERmcare (SHELTER) project.

Methods: Descriptive statistics, calculation of percentage, and multivariable logistic analysis were applied to describe the relationship between anticholinergic drug use and prevalence of delirium in NH patients. The Anticholinergic Risk Scale (ARS) and the Anticholinergic Burden Scale (ACB) were used to calculate the anticholinergic load.

Results: 54% of patients with dementia and 60% without dementia received at least 1 anticholinergic drug according to the ACB. The prevalence of delirium was higher in the dementia group (21%) compared with the nondementia group (11%). Overall, anticholinergic burden according to the ACB and ARS was associated with delirium both in patients with and without dementia, with odds ratios ranging from 1.07 [95% confidence interval (CI) 0.94-1.21] to 1.26 (95% CI 1.11-1.44). These associations reached statistical significance only in the group of patients with dementia. Among patients with dementia, delirium prevalence increased only modestly with increasing anticholinergic burden according to the ACB, from 20% (with none or minimal anticholinergic burden) to 25% (with moderate burden) and 27% delirium (with strong burden scores).

Conclusions And Implications: The ACB scale is relatively capable to detect anticholinergic side effects, which are positively associated with prevalence of delirium in NH patients. Given the modest nature of this association, strong recommendations are currently not warranted, and more longitudinal studies are needed.