Comparative Epidemiology of Bacteremia due to Methicillin-Resistant Staphylococcus aureus between Older and Younger Adults: A Propensity Score Analysis

OBJECTIVE. We evaluated the clinical and molecular epidemiology of bloodstream infections (BSIs) due to methicillin-resistant Staphylococcus aureus (MRSA) in older versus younger patients treated with vancomycin, determining the independent effect of increased age on outcomes. DESIGN. Observational retrospective cohort study. SETTING. Detroit Medical Center, level I trauma center. PATIENTS. Adult older (65 years and older) and younger (younger than 65 years) patients with documented BSIs due to MRSA treated with vancomycin (2005-2010). METHODS. Collected demographics, comorbidities, microbiology, treatment, outcomes. Multivariable model used to generate propensity score for each patient on the basis of the probability of being 65 years of age or older. RESULTS. Three hundred twenty patients were eligible (69 patients 65 years and older; 251 patients younger than 65 years). Catheter-related infections and endocarditis were the most common sites of infection for older (20.3%) and younger (19.1%) adults, respectively. Median first total 24-hour vancomycin dose (1,000 vs 2,000 mg; P < .001) and initial trough (13.1 vs 15.0 mg/L; P = .043) was significantly lower in older versus younger patients. Vancomycin treatment failure rates were similar among older and younger patients (49.3% vs 53.4%; P = .545). In multivariable analysis of outcomes, after controlling for predictors of older age, there was no difference in clinical outcomes between older and younger adults. CONCLUSIONS. After accounting for confounders associated with increased age, failure rate of patients with BSIs due to MRSA treated with vancomycin was similar between older and younger patients. Older adults were less likely to have optimal vancomycin dosing and initial trough levels than younger patients. Efforts should be made to optimize dosing of medications such as vancomycin in older adults. Infect Control Hosp Epidemiol 2013;34(4):400-406