The association between medication or alcohol use and the incidence of frailty: a retrospective cohort study.

BACKGROUND: Understanding potentially modifiable factors that influence the risk of frailty is a key concern for the management of this urgent contemporary public health challenge. This study evaluates the association between the use of various medications or alcohol and the incidence of frailty among older adults. METHODS: This study was a retrospective cohort study on older adults (≥ 65 years) using data from the longitudinal Survey of Health, Ageing and Retirement in Europe (SHARE survey, 28 countries). Medication use was measured as taking several different groups of medications. Alcohol use was assessed with SHARE questions corresponding to AUDIT-C. The outcome measure was the incidence of frailty after two years, defined by frailty index (FI) and frailty phenotype (FP). A multiple logistic regression model was used to evaluate the association with adjustment for several potential confounding factors. RESULTS: Of the 14,665 FI-population participants, 1800 (12.3%) developed frailty within two years. Of the 8133 FP-population participants, 2798 (34.4%) developed pre-frailty and 247 (3.0%) developed frailty within two years of baseline. After adjustment for potential confounding variables, non-hazardous alcohol use (adjusted OR; 95% CI for the FI-population: 0.68; 0.60-0.77) and hazardous alcohol use (0.80; 0.68-0.93) are associated with lower incidence of frailty compared to no alcohol use. The odds of frailty are increased when taking medications; the largest effect size was observed in older adults taking medication for chronic bronchitis (adjusted OR; 95% CI for the FI-population: 2.45; 1.87-3.22), joint pain and other pain medication (2.26; 2.00-2.54), medication for coronary and other heart disease (1.72; 1.52-1.96), medication for diabetes (1.69; 1.46-1.96), and medication for anxiety, depression and sleep problems (1.56; 1.33-1.84). Additionally, the risk of frailty was increased with stroke, Parkinson's disease and dementia. CONCLUSIONS: Taking certain groups of medication was associated with increased incidence of frailty and pre-frailty, which might be due to either medication use or the underlying disease. Alcohol use was associated with a lower risk of pre-frailty and frailty compared to no alcohol use, which might be due to reverse causality or residual confounding. There was no significant interaction effect between medication groups and alcohol use on frailty incidence.