Editor’s Note: At the IDWeek 2023, Dr. Rebekah W. Moehring from the Duke Center for Antimicrobial Stewardship and Infection Control at Duke University presented a fascinating report on the use of MRSA surveillance in guiding empirical therapy for MRSA and de-escalation strategies to optimize vancomycin antibiotic management.

Infectious Diseases physician, hospital epidemiologist, and Associate Professor

the Duke Center for Antimicrobial Stewardship and Infection Control at Duke University

Guidelines recommend that in community-acquired pneumonia (CAP), risk assessment can be used to determine which patients should receive empirical treatment covering MRSA, while in hospital or ventilator-associated pneumonia (HAP or VAP), the decision is based on local epidemiological data (≥10% to 20%) and risk assessment.

Epidemiological studies have shown that most CAP pathogens are unknown (62%) or viral (22%), with less than 2% being Staphylococcus aureus, while most HAP cases are of unknown etiology (58%) or viral (12%).

MRSA nasal screening provides direct biological information about the relationship between nasal colonization and disease progression. However, there are issues with implementing active surveillance for infection prevention, such as heterogeneity and sampling issues, choice between PCR and selective media, other variables (e.g., exposure to antibiotics like fluoroquinolones), logistics, scalability, and cost.

So, can MRSA surveillance data’s negative predictive value (NPV) be used in clinical practice? What is the threshold for MRSA prevalence? A meta-analysis by Parente et al., which included 22 studies with 5,263 patients, found an overall MRSA pneumonia rate of 10% (range: 8%-13%). The diagnostic sensitivity of MRSA nasal screening in the general population was 70.9%, with specificity at 90.3% and NPV at 96.5%. In VAP, sensitivity was slightly lower (40.3%), but NPV was 94.8%. Many studies have also used MRSA nasal screening as an antimicrobial stewardship (ASP) tool for pneumonia de-escalation, contributing to reduced vancomycin treatment costs and duration of therapy (DOT).

Dr. Moehring points out that false-negative results with MRSA nasal screening cannot completely rule out MRSA infection. She provides examples of a 67-year-old male patient with COPD/emphysema, acute respiratory failure, pulmonary infiltrates, and bloody sputum. Despite a negative nasal screen, he was eventually diagnosed with MRSA necrotizing pneumonia. Similarly, a 37-year-old female with cough, fever, pleuritic chest pain, and multiple tissue abscesses was diagnosed with left foot MRSA SSTI/tendonitis and right foot MSSA IE despite a negative nasal screen.

The top ten indications for vancomycin include empirical/sepsis therapy, skin and soft tissue infections (SSTI), surgical prophylaxis, bloodstream infections (BSI), pneumonia, bone/joint infections, intra-abdominal infections, neutropenic fever, central nervous system infections, and cardiovascular infections.

Dr. Moehring then raises two questions: Can MRSA surveillance results be used to discontinue vancomycin treatment for infections other than pneumonia? Should MRSA surveillance results be used to estimate the likelihood of MRSA infection in patients?

To answer these questions, she first cites a retrospective cohort study by Mergenhagen et al., which analyzed over 240,000 patients and over 560,000 clinical cultures. The MRSA nasal screening positivity rate was 22.9%, while the clinical culture MRSA positivity rate was 8.3%. The sensitivity and specificity of MRSA clinical culture were 67.4% and 81.2%, respectively. The overall NPV of MRSA screening for excluding MRSA infection was 96.5%, with most sources having NPV of over 90%.

MRSA distribution in skin and soft tissue infections (SSTI) is heterogeneous. In monomicrobial infections, non-purulent infections are often difficult to culture, with MRSA being a very small portion, while purulent infections are around 40%-60% MRSA. In polymicrobial infections, diabetic foot and sacral ulcers have about 15%-30% MRSA, and necrotizing fasciitis typically requires surgical treatment.

A study by Acquisto et al. included 116 SSTI patients with wound cultures and MRSA nasal swabs. The MRSA prevalence was 44% (52/116), and the NPV of MRSA nasal swab was 72.8%. Another study by Mergenhagen et al. found an overall NPV of 89.6% for MRSA nasal screening, with NPV exceeding 89% for deep tissues (89.2%), feet (89.7%), and toes (89.4%).

The risk of MRSA intra-abdominal infection is very low. In intra-abdominal infections, Amp-R E. coli is the main pathogen, not Staphylococcus aureus. However, MRSA infection risk is relatively higher in patients with peritoneal dialysis, peritonitis, postoperative wound infections, and complications of bacteremia.

For patients in areas with very low MRSA prevalence (<2%), there is no indication for empirical vancomycin treatment. For patients in areas with low MRSA prevalence (<10%), there is an indication for empirical vancomycin treatment, but the likelihood of MRSA infection is smaller. For patients in areas with a high prevalence of MRSA or in cases with low prevalence but severe diseases, there is an indication for empirical vancomycin treatment, but culture data and clinical response assessment are required.

Q1. What are the usual syndromes in patients?

• A: This is the most difficult question to answer. If you cannot determine the syndrome of a localized infection, you should not rely on MRSA nasal screening. This includes undifferentiated shock or septic shock with an unknown source of infection.

Q2. What is the expected MRSA infection rate in syndromes?

• A: Refer to the MRSA prevalence estimates and antibiogram data for your specific region and the literature. The combination of syndrome characteristics and other epidemiological factors (e.g., community-acquired vs. nosocomial, purulent vs. non-purulent) should be considered. Direct microbial diagnosis is always preferred over alternative diagnostics.

Q3. Should other drug-resistant GP pathogens be considered in patients?

• A: Vancomycin is commonly used in the treatment of the following pathogens: Coagulase-negative Staphylococci (CoNS) are mainly associated with artificial joints, VP shunts, central venous catheters, and dialysis access. Ampicillin-resistant Enterococci are seen in settings such as intravascular devices, endocarditis, and nosocomial intra-abdominal infections with secondary bloodstream infections. Penicillin-resistant Streptococcus pneumoniae is associated with community-acquired meningitis with concurrent pneumococcal pneumonia, viridans group Streptococci (VGS) endocarditis, and neutropenic fever with VGS bacteremia, among others.

Dr. Moehring concludes by stating that whether MRSA can help reduce vancomycin use in patients with other syndromes is challenging to answer. In populations with low MRSA prevalence (<10%-15%), empirical vancomycin use is common. Ideally, withholding vancomycin and using expected management, understanding the NPV, and always considering vancomycin use in the context of individual clinical background is essential. Data for most syndromes outside of pneumonia are limited. Microbiological databases stratified by sample type are too coarse to fully assess utility in other (more heterogeneous) syndromes.