Hospital-Acquired Pneumonia (HAP) caused by Staphylococcus aureus (SA): Treatment, Resource Use, and Mortality

Background: HAP confers substantial mortality, morbidity and cost. SA has been associated with poor outcomes. The importance of appropriate empiric and definitive therapy for SA, both methicillin-resistant (MRSA) and methicillin-sensitive (MSSA), is a controversial topic and growing area of research.

Objective: To describe treatment patterns in SA-positive HAP and to determine the impact of type of SA infection (MRSA and MSSA) and antibiotic treatment (empiric and definitive) on length of hospital stay (LOS) and mortality. 

Methods: We conducted a retrospective analysis (2005-2008) of the Health Facts^® database (Cerner Corp., Kansas City, MO), which contains comprehensive clinical records from 70 US hospitals. To our knowledge, this is the largest retrospective study of HAP in the US. Adult patients with a secondary discharge diagnosis consistent with pneumonia and a positive respiratory or blood culture (Cx) obtained >48 hours after admission were included. Appropriate therapy was defined as orders for antibiotics active against MRSA or MSSA for each group based on antibiotic susceptibilities within 24 hours after index Cx (empiric) or 72-96 hours after Cx (definitive). Univariate analyses were performed to assess raw mean differences between MRSA and MSSA groups. Multivariable models were developed to determine the main effects of MRSA vs. MSSA, appropriate definitive therapy, and appropriate empiric therapy on post-culture LOS of survivors and in-hospital mortality. Covariates included variables representing severity of illness, time at risk, and time to receiving treatment. LOS was analyzed using a GLM accounting for hospital as a random effect. Mortality risk was examined using logistic regression.

Results: 3,041 patients had HAP based on our criteria; 434 (14.3%) had an index Cx positive for MRSA only, 598 (19.7%) for MSSA only, and 59 (2%) for both. Only 36% of MRSA-positive patients received appropriate empiric therapy; this improved to 55% for appropriate definitive therapy.  Vancomycin and linezolid were the most commonly used anti-MRSA agents. Among MSSA-positive patients, 62% received appropriate empiric and 66% appropriate definitive antibiotic therapy. The finding of a shorter adjusted LOS in the MRSA vs the MSSA group was unexpected (15.6 vs 18.4 days, p = 0.004). Appropriate empiric antibiotic therapy was associated with a shorter LOS (by 2.4 days, p = 0.018), while appropriate definitive therapy was associated with a mean increase of 3.2 days (p = 0.003). Adjusted mortality was similar in the two study groups. Mortality was lower in patients with appropriate definitive therapy (odds ratio 0.72, p = 0.047).

Conclusions: In a large group of HAP patients, more than one third had SA infection.  Inappropriate antibiotic therapy is common for patients infected with MRSA and MSSA. Choice of definitive and empiric therapy affects outcomes.  Patients with MRSA HAP had an unexpected shorter LOS.