Abstract
BACKGROUND AND OBJECTIVE: Anticholinergic medications are known to affect the prognosis of older nursing home residents. Various anticholinergic scales were developed to measure the cumulative anticholinergic burden; among them, the CRIDECO Anticholinergic Load Scale (CALS) has recently emerged as a new tool to identify patients with cognitive impairment due to anticholinergic burden. This study aimed to externally validate the CALS and to evaluate the association of CALS and the anticholinergic cognitive burden (ACB) scales with baseline cognitive and functional impairment, as well as with 3-year mortality rates. METHODS: A prospective cohort of 600 nursing home residents (mean age 80.4 ± 8.0 years; 69.8% women) underwent a comprehensive geriatric assessment. Anticholinergic burden was assessed at baseline using both CALS and ACB scales. Cognitive impairment (Mini-Mental State Examination < 24) and physical disability (one or more impaired activities of daily living) were evaluated cross-sectionally using a logistic regression model. Cox proportional hazards models were used to estimate the association between anticholinergic burden and 3-year mortality, adjusting for age, sex, multimorbidity, nutritional status, and cognitive and functional status. RESULTS: Among 600 nursing home residents included in the study, 72.0% had cognitive impairment and 56.3% had at least one activity of daily living limitation. The CALS and ACB scores were significantly correlated ( ρ = 0.76), but CALS identified a higher number of residents with moderate-to-high anticholinergic burden. Multivariate logistic regression showed that CALS ≥ 2 was independently associated with cognitive impairment (odds ratio 1.84, 95% confidence interval 1.02-3.34), whereas ACB ≥ 2 was not. Both scales were associated with activities of daily living disability, with a stronger gradient and better goodness of fit for CALS than ACB. During the 3-year follow-up, 25.3% of residents died. Cox regression analyses showed that residents with CALS or ACB ≥ 2 had significantly lower survival over 3 years. In fully adjusted Cox models, both CALS ≥ 2 (hazard ratio 1.93, 95% confidence interval 1.07-3.46) and ACB ≥ 2 (hazard ratio 1.69, 95% confidence interval 1.02-2.83) remained associated with increased mortality. Prognostic performance was similar (CALS C-index: 0.783; ACB: 0.781), but the model fit favored CALS. CONCLUSIONS: In this cohort of nursing home residents, anticholinergic burden as measured by both CALS and ACB was associated with baseline physical impairment and 3-year mortality, but CALS showed a better goodness of fit. Between the two scales, CALS only was independently associated with baseline cognitive impairment. These findings support the clinical utility of CALS in assessing anticholinergic-related risk among frail older adults in institutional settings.