Predictors of Clinical Success Among a National Veterans Affairs Cohort With Methicillin-Resistant Staphylococcus aureus Pneumonia

Abstract

BackgroundThe treatment of methicillin-resistant Staphylococcus aureus (MRSA) pneumonia is exceedingly complicated, which is concerning because of the high mortality rate associated with the infection. Identification of independent predictors of clinical success can optimize patient care by assisting clinicians in treatment decisions. ObjectivesOur goal was to identify independent predictors of clinical success in a national Veterans Affairs (VA) cohort of patients with MRSA pneumonia. MethodsA nested case-control study was conducted among a cohort of VA patients with MRSA pneumonia receiving linezolid or vancomycin between January 2002 and September 2010. Cases included those demonstrating clinical success, defined as discharge from the hospital or intensive care unit by day 14 after treatment initiation, in the absence of death, therapy change, or intubation by day 14. Control subjects represented nonsuccess, defined as therapy change, intubation, intensive care unit admission, readmission, or death between treatment initiation and day 14. The potential predictors assessed included treatment, patient demographic and admission characteristics, previous health care and medication exposures, comorbidities, and medical history. Odds ratios (ORs) and 95% CIs were calculated from logistic regression. ResultsOur study included 2442 cases of clinical success and 1290 control subjects. Demographic characteristics varied between the clinical success and nonsuccess groups, including age, race, and region of facility. A current diagnosis of chronic respiratory disease (46% vs 42%) and diagnosis of pneumonia in the year before the MRSA pneumonia admission (37% vs 32%) were both more common in the clinical success group. Despite these significant differences, only 2 predictors of clinical success were identified in our study: previous complication of an implant or graft, including mechanical complications and infections, in the year before the MRSA pneumonia admission (adjusted OR, 1.55 [95% CI, 1.17–2.06]) and treatment with linezolid (adjusted OR, 1.53 [95% CI, 1.12–2.10]). Predictors of nonsuccess (adjusted OR [95% CI) included diagnosis of concomitant urinary tract infection (0.82 [0.70–0.96]), intravenous line (0.76 [0.66–0.89]), previous coagulopathy (0.74 [0.56–0.96]), previous amputation procedure (0.72 [0.53–0.98]), current coagulopathy diagnosis (0.71 [0.53–0.96]), dialysis (0.54 [0.38–0.76]), multiple inpatient procedures (0.53 [0.45–0.62]), inpatient surgery (0.48 [0.41–0.57]), and previous endocarditis (0.24 [0.07–0.81]). ConclusionsMRSA pneumonia tends to affect patients with complex care, and identification of the predictors of clinical success is useful when considering different therapeutic approaches. In this national cohort of VA patients with MRSA pneumonia, treatment was the only modifiable variable predicting clinical success.