Size of the associations between anticholinergic burden tool scores and adverse outcomes in older patients.

Background Several anticholinergic scales and equations to evaluate the anticholinergic burden have been previously created. Association of these instruments with the anticholinergic outcomes are usually estimated by means of hypothesis contrast tests, which ignore the size of the association effect. Objective To evaluate the effect size of the associations between the scores on cumulative anticholinergic burden instruments with peripheral or central anticholinergic adverse outcomes in older patients. Setting Internal medicine ward of a Tertiary University Hospital. Methods A case-control study was conducted in patients over 65 years who were admitted to two internal medicine wards of a Portuguese university hospital. The Anticholinergic Drug Scale, Anticholinergic Risk Scale, Anticholinergic Cognitive Burden scale and Drug Burden Index were used to calculate the patients' anticholinergic burden. Peripheral (dry mouth-swab technique; dry eye-Schirmer test) and central (falls and cognitive impairment-Mini-Mental State Examination) anticholinergic adverse outcomes were investigated. The Barthel Index was used to assess overall physical functionality. The Mann-Whitney test was used to evaluate probabilistic differences in the anticholinergic scores between case and control individuals. To establish the effect size of the associations, the area under the curve of the receiver operating characteristics curve was calculated. Main outcome measure Anticholinergic adverse effects. Results A total of 250 patients (mean age 81.67 years, standard deviation 7.768; 50% females) were included. In total, 148 patients (59.2%) presented with dry mouth, 85 (34%) with dry eye, 141 (56.4%) with impaired functionality, 44 (17.6%) with a history of falls and 219 (87.6%) with cognitive impairment. Significant differences (p < 0.05) were obtained for the majority of the associations between Anticholinergic Drug Scale, Anticholinergic Risk Scale, Anticholinergic Cognitive Burden and Drug Burden Index and adverse effects. Conversely, the effect sizes of these associations ranged from "fail" (area under the curve 0.5 to 0.6) to "fair" (area under the curve 0.7 to 0.8). Conclusion Although significant differences in the scores of anticholinergic burden instruments and adverse outcomes may exist, the effect sizes of these associations ranged from 'fail' to 'fair', which limits their utility in preventing anticholinergic adverse outcomes with medication review interventions.