Association of Anticholinergic Burden and Risk of Pneumonia in Aged People

Background: Use of anticholinergics may lead to dry mouth, oropharyngeal and esophageal dysphagia, and eventually aspiration pneumonia. Although previous studies have suggested an increased risk of pneumonia with increasing anticholinergic burden (ACB), the effect was substantially confounded by indication. In addition, recent research has also suggested that the frailty index may be a strong predictor of the clinical outcome of pneumonia.

Objectives: We leveraged case-crossover design to minimize confounding by indication and to evaluate the association between ACB and pneumonia while applying the multi-morbidity frailty index (m-Fi) to test the effect modification by frailty

Methods: This case-crossover study used data from the National Health Insurance Database. We included patients over 65 years of age with hospitalized pneumonia from 2011 to 2020. The index date was the date of admission for pneumonia. Patients with immune dysfunction, pneumonia, ventilator dependence, or tuberculosis infection prior to the index date were further excluded. We defined the hazard period as 30 days prior to the index date with a 30-day washout period before the hazard period. The reference periods were randomly selected from the time horizon (i.e., 61 to 90, 91 to 120, 121 to 150, and 151 to 180 days) before the index date. We calculated the sum of ACB and the m-Fi score within each period of one year before the index date, and patients were classified into four groups (i.e., fit, mild frailty, moderate frailty, and severe frailty) according to the median and 99th percentile of m-Fi. The conditional logistic regression model was implemented to estimate the effect of ACB on the risk of pneumonia among different frailty groups.

Results: A total of 401,781 patients hospitalized with pneumonia were included, and patients were classified into different frailty groups based on the following delimiters of m-Fi < 0.125, 0.126-0.208, 0.209-0.292, and >0.292. We included 188,740 patients in the fit group, while there were 133,038, 61,805, and 18,198 patients in the mild, moderate, and severe frailty groups, respectively. For the risk of pneumonia, an increased risk was observed with each 1-point increase in ACB (Odds ratio (OR) 1.35; 95% CI 1.34-1.35) in the fit group; however, the OR was 1.24, 1.18, and 1.12 in the mild, moderate, and severe frailty groups, respectively. Stratified risk analysis based on the frailty subgroups showed an obvious pattern of effect modification in our study.

Conclusions: Our study confirmed the association between ACB and hospitalized pneumonia; furthermore, the baseline frailty index seems to modify the risk as patients with lower frailty were more susceptible to the anticholinergic burden.