Predicting Risk for Death from MRSA Bacteremia¹

Mina Pastagia

, Lawrence C. Kleinman, Eliesel G. Lacerda de la Cruz, and Stephen G. Jenkins

Author affiliations: The Rockefeller University, New York, New York, USA (M. Pastagia); Mount Sinai School of Medicine, New York (L.C. Kleinman, E.G. Lacerda de la Cruz); and Weill Cornell School of Medicine, New York (S.G. Jenkins)

Abstract

Methicillin-resistant Staphylococcus aureus (MRSA) bacteremia is often fatal. To determine predictors of risk for death, we conducted a retrospective cohort study. We examined 699 episodes of MRSA bacteremia involving 603 patients admitted to an academic medical center in New York City during 2002–2007. Data came from chart reviews, hospital databases, and recultured frozen MRSA specimens. Among the 699 episodes, 55 were caused by vancomycin–intermediate resistant S. aureus strains, 55 by heteroresistant vancomycin-intermediate S. aureus strains, and 589 by non–vancomycin-resistant strains; 190 (31.5%) patients died. We used regression risk analysis to quantify the association between clinical correlates and death. We found that older age, residence in a nursing home, severe bacteremia, and organ impairment were independently associated with increased risk for death; consultation with an infectious disease specialist was associated with lower risk for death; and MRSA strain types were not associated with risk for death.

Methicillin-resistant Staphylococcus aureus (MRSA) is a worldwide concern; it colonizes and infects patients in the hospital and in the community (1). For the past 50 years in the United States, the standard therapy has been vancomycin. Recent vancomycin treatment failures have raised questions regarding optimal treatment (2). Although new antimicrobial drugs (e.g., linezolid, daptomycin, tigecycline) have been developed, none has been consistently superior to vancomycin for the treatment of MRSA (3,4), and MRSA resistance rapidly develops for many new drugs (5,6). Some studies have suggested MIC creep (increasing vancomycin MICs against MRSA over time), but others have not (7,8). In 2006, the upper limit of vancomycin susceptibility for S. aureus was redefined, lowered from 4 µg/mL to 2 µg/mL, first by the Clinical and Laboratory Standards Institute and soon thereafter by the US Food and Drug Administration and the European Committee on Antimicrobial Susceptibility (9).
Vancomycin treatment failures for MRSA occur even when MICs are within the range considered susceptible, especially 1–2 µg/mL (10–13). Among high-risk bacteremic patients, Sakoulas et al. documented treatment failure rates of 44% when vancomycin MICs were <0.5 µg/mL and of 90% when vancomycin MICs were 1–2 µg/mL (p = 0.01) (10). Hidayat et al. found that mortality rates were higher for patients infected with strains with higher vancomycin MICs (11)

Some apparently susceptible strains of MRSA might actually be heteroresistant vancomycin-intermediate S. aureus (hVISA) strains. That is, although the hVISA isolates seem to be susceptible to vancomycin according to conventional testing, the isolates contain subpopulations of colonies resistant to vancomycin. Testing for hVISA has not been standardized and is not routinely undertaken. hVISA strains are more common in strains with higher vancomycin MICs (14,15). hVISA might contribute to worse clinical outcomes, but this possibility has not been convincingly confirmed by large studies.

To determine predictors of risk for death among patients with MRSA bacteremia, we conducted a retrospective study that compared demographic and clinical characteristics of adult patients with MRSA bacteremia. MRSA strains from these patients were vancomycin susceptible, VISA, and hVISA. We analyzed a 5-year trend of vancomycin MICs among adult patients with MRSA bacteremia. We also analyzed the associations between host factors, organism factors, and death versus survival, and quantifed the marginal contribution of key factors to risk for death.