Objective: To describe the epidemiology of and to develop a simple 30-day mortality clinical decision rule among critically ill patients ≥65 yrs. Increasing incidence of hospitalizations with and emergence of hypervirulent epidemic strains have made Clostridium difficile-associated disease an important public health concern. Advanced age is a risk factor for development of and death from Clostridium difficile-associated disease. Intensive care unit patients with Clostridium difficile-associated disease have a high mortality, but neither the burden of nor risk factors for death among the elderly intensive care unit patients with Clostridium difficile-associated disease are well understood.
Design: Secondary analysis of a retrospective cohort study.
Setting: All intensive care units at a single academic institution.
Patients: A total of 278 critically ill patients with Clostridium difficile-associated disease; n = 148 aged ≥65 yrs.
Interventions: None in addition to routine intensive care unit care.
Measurements and Main Results: Univariate analyses were performed to compare characteristics and outcomes of the elderly vs. the younger groups, and elderly 30-day survivors with nonsurvivors. Multivariable logistic regression model was developed with 30-day mortality as a dependent variable. Covariates retained in the model were assigned weighted points to develop a 30-day mortality prediction score. Area under the receiver operating characteristics curve and cross-validation analyses evaluated the score characteristics. Elderly patients were 68% more likely to experience 30-day mortality than the younger group. Absence of chronic respiratory disease (R), age 75+ yrs (A), septic shock (S), and Acute Physiology and Chronic Health Evaluation II score 20+ (A) comprised the RASA score, whose receiver operating characteristics was 0.740; 95% Confidence Interval was 0.663–0.817.
Conclusions: Elderly patients represent approximately 50% of intensive care unit patients with Clostridium difficile-associated disease and have a higher 30-day mortality than younger patients. A simple prediction rule incorporating determinants of 30-day mortality easily available at the bedside may aid in optimizing treatment decisions in this growing population.