Empirical Coverage of Methicillin-ResistantStaphylococcus aureusin Community-Acquired Pneumonia: Those Who Do Not Remember the Past Are Doomed to Repeat It

To the Editor—Previous Infectious Diseases Society of America and American Thoracic Society recommendations for empirical therapy [1] and the concept of healthcare-associated pneumonia have caused significant harm to antibiotic stewardship efforts by promoting overtreatment. This includes overuse of vancomycin and linezolid to cover methicillin-resistant Staphylococcus aureus (MRSA), as illustrated by the data of Self et al [2], published in this issue of Clinical Infection Diseases. These data show that despite a low prevalence of MRSA and lack of evidence of benefit, 30% of patients in these US centers are treated empirically with anti-MRSA therapies [2, 3]. The evidence supporting these prior recommendations have been widely criticized, with methodological issues including low sample sizes, patients with only culture-positive pneumonia, inconsistent microbiological testing, publication bias, and lack of generalizability [4]. In their discussion, Self et al [2] rightly highlight the need to avoid antibiotic overtreatment, but they also recommend empirical treatment of MRSA for all patients admitted to the intensive care unit with community-acquired pneumonia. We believe that we should learn from the previous experience of healthcare-associated pneumonia and be cautious in making such important recommendations without more generalizable data. The study reported by Self et al is limited by the inclusion of only 15 patients with MRSA; the exclusion of patients with prior hospitalization, nursing home residence, or other risk factors that may be very relevant in clinical practice; and the poor generalizability of the cohort, given the inclusion of patients with low disease severity and a low mortality rate, coming from only 3 hospitals in the United States [5]. A key finding of previous work in this area has been the huge variability in rates of resistance across different healthcare systems and even within the same country. We and other authors have highlighted the need to identify local prevalence data for resistant pathogens in order to “localize” recommendations for empirical therapy and avoid overtreatment in areas of low prevalence [4, 6]. We urge caution and highlight the need for a large, representative, international multicenter study specifically designed to assess the prevalence of MRSA and its risk factors in patients with pneumonia. The question of whether empirical MRSA coverage is needed and for which patients is crucial, but we have seen the consequences of recommending overtreatment on the basis of limited data. Those who do not remember the past are doomed to repeat it. Potential conflicts of interest.