To the Editor—The need for studies on coronavirus disease 2019 (COVID-19) superinfections that can inform rational antimicrobial treatment and stewardship strategies has been recognized. ^{Reference Clancy and Nguyen1} In a recent review from our institution, ^{Reference Nori, Cowman and Chen2} we found that up to 71% of patients admitted with COVID-19 received antibiotics. Anti–methicillin-resistant Staphylococcus aureus (anti-MRSA) agents, particularly vancomycin, are important stewardship targets, and they are included in the 2019 World Health Organization (WHO) Watch List of Antibiotics.

Recently, guidance was published on the treatment of possible concomitant community-acquired bacterial pneumonia (CAP) for patients admitted with COVID-19. ^{Reference Metlay and Waterer3} The authors recommend selective use of anti-MRSA therapy, as for any other patient with CAP, and the utility of the MRSA nares PCR is not addressed. However, the empiric use of these agents for admitted COVID-19 patients remains prevalent, driven by several factors. For one, staphylococcal superinfection is a common complication of other viral pneumonias, such as influenza, ^{Reference Vardakas, Matthaiou and Falagas4} and the 2018 Infectious Diseases Society of America (IDSA) guidelines on influenza treatment state that “agents with activity against MRSA should be included in the empiric regimen for critically ill patients.” ^{Reference Uyeki, Bernstein and Bradley5} Additionally, the real-world treatment of patients with COVID-19 is complicated by recurrent fevers, fluctuating oxygen requirements, prolonged hospitalization and/or ventilation, blurring the line between community-acquired versus hospital-acquired pneumonia. Therefore, we sought to determine the prevalence of MRSA in respiratory cultures of patients admitted with COVID-19, and we evaluated the diagnostic performance of the MRSA nares PCR test, a valuable stewardship tool for ruling out MRSA pneumonia in low-prevalence settings, ^{Reference Carr, Daley, Givens Merkel and Rose6} in this cohort.

Results

During the study period, 4,221 adult patients were admitted with COVID-19; only 472 patients (11.1%) had a respiratory culture. Patients with respiratory cultures were more severely ill than the patients without respiratory cultures, 78% of patients were intubated versus only 12% of those without respiratory cultures. Additionally, patients with respiratory cultures had longer lengths of stay (median, 19.5 days vs 6 days) and higher mortality (56% vs 21.5%) versus those who had no respiratory culture. Overall, 904 (21.4%) received empiric vancomycin within 48 hours of admission, and this was more commonly seen in the group that had respiratory cultures obtained (33.4%). Patient characteristics are summarized in Supplementary Table 1 (online).

Among the 4,221 patients in the entire cohort, 158 patients had respiratory cultures obtained by day 3 of hospitalization, 285 by day 7, 405 by day 14, and 472 by day 28. The prevalence of MRSA in respiratory cultures ranged from a low of 0.6% on day 3, to 5.7% on day 28, cumulatively (Table 1).

Table 1.

Prevalence of Methicillin-Resistant Staphylococcus aureus (MRSA) in Respiratory Cultures at Different Time Points of Hospital Stay

Also, 369 MRSA nares PCR tests were performed among the patients in the cohort; of these patients, 122 had corresponding respiratory cultures within 5 days of the PCR test. Of these 122 MRSA PCR tests, 12 were positive, of which 2 patients had a corresponding positive respiratory culture for MRSA. Of the 110 negative nasal MRSA PCR tests, none of the corresponding respiratory cultures had MRSA isolated, yielding a negative predictive value (NPV) of 100% (Supplementary Table 2 online).