Attributable Mortality after S. aureus Bacteremia

Background: S. aureus is responsible for approximately 20% of nosocomial bloodstream infections, second only to coagulase-negative staphylococci.  Here, we describe the long term attributable mortality rates in a large single center cohort of S. aureus bacteremia (SAB).

Objective: To determine the 1 year attributable mortality in SAB.

Methods: Blood cultures growing S. aureus from 1/1/2006 to 12/31/2008 at the Cleveland Clinic were identified. Gender, age at positive blood culture, and social security were extracted from the electronic medical record (EMR). Survivor status was evaluated by the Social Security Death Index and EMR. 2 sets of up to 4 controls, matched first for either length of stay (LOS) or all patient refined diagnosis related groups-severity code (APR-DRG), followed by matching for age, gender, date of admission were selected. Day 0 for cases was defined as the date of the first positive blood culture. The corresponding hospital day was used for matched controls as day 0. Statistics were performed with JMP software (SAS, Cary, NC).

Results: 914 SAB cases were identified; 58% had methicillin-resistant S. aureus (MRSA), 61% were male, and the median age was 60 years (IQR 49-71). Median age was higher (61 vs. 58 years, p=0.04) and median LOS was longer (18 vs. 13 days, p<0.001) in patients with MRSA, compared to MSSA. 2655 APR-DRG matched controls and 2579 LOS matched controls were included (61% male, median age 60 years [IQR 49-71], median LOS 13). All-cause 30 day mortality in patients with MRSA, MSSA, the APR-DRG controls, and LOS-controls was 18%, 16%, 9%, and 7%. Therefore, the attributable 30-day mortality is estimated to be 9-11% for MRSA, and 7-9% for MSSA. All-cause 1 year mortality in patients with MRSA, MSSA, the APR-DRG controls, and LOS-controls was 43%, 34%, 28%, and 25%. The attributable 1 year mortality is estimated to be 15-18% for MRSA, and 6-9% for MSSA. On Cox proportional hazard analysis, factors associated with 1 year mortality were LOS, age, and SAB. When stratified by age group, the burden of attributable mortality increased with age; in patients older than 70 years, the 1 year attributable mortality for MRSA and MSSA was 22-23%, and 14-15%, respectively. In patients 70 years old or younger, these numbers were 13-16%, and 4-7%, respectively (p<0.01 for both MRSA and MSSA).  In a multivariate model within the SAB group, age (OR 1.03 per year, p<0.001), days from admission to SAB (OR1.03 per day, p<0.001), and number of positive blood cultures (OR 1.04 per additional positive blood culture) were predictive of 1 year mortality. In addition, a trend towards increased 1 year mortality was seen in MRSA (OR 1.3, p=0.06) and with increased LOS (OR 1.008 per day, p=0.07).

Conclusions: Even when comparing patients with SAB to patients with similar acute and chronic severity of illness, substantial excess mortality is observed. This effect is largest in patients over the age of 70 years, and in those patients who develop SAB later during their hospital stay.