Discussion

In this large retrospective study of immunocompromised patients, we showed that negative MRSA nasal swab culture results had consistently high NPV for MRSA infection, regardless of the culture site. These findings support the results from Perreault and colleagues, who retrospectively evaluated the NPV of MRSA nasal swab screenings using PCR and/or culture in 194 AML patients with neutropenic fever. ^{Reference Perreault, Binks, McManus and Topal10} Their results showed an overall sensitivity of 62%, specificity of 98%, PPV of 38%, and a NPV of 99%. Similarly, the retrospective study by Talagtag and colleagues evaluated 98 AML patients with pneumonia and demonstrated a sensitivity of 75%, specificity of 100%, PPV of 100%, and a NPV of 98.9% using MRSA nasal swab screening. ^{Reference Talagtag, Patel and Scappaticci11} The high NPV for respiratory cultures is consistent with our study, which showed a NPV of 98.5% and specificity of 98.7% for the utility of MRSA swab screening for detecting MRSA in a clinical culture.

In immunocompetent patients, the literature on the NPV of MRSA nasal swab screenings is predominantly focused on pneumonia, although there is some data from other sources as well. ^{Reference Carr, Daley, Givens Merkel and Rose13} A meta-analysis by Parente and colleagues included 5163 patients with pneumonia found a NPV of 96.5% with a 10% prevalence of potential MRSA pneumonia. ^{Reference Parente, Cunha, Mylonakis and Timbrook8} A large retrospective study conducted by Mergenhagen and colleagues reported an overall specificity of 81.2% and NPV of 96.5% for MRSA nasal screening using PCR and/or culture for any clinical culture site. ^{Reference Mergenhagen, Starr, Wattengel, Lesse, Sumon and Sellick12} The NPV for bloodstream infections was 96.5%, for intra-abdominal cultures 98.6%, for respiratory cultures 96.1%, for wound cultures 93.1%, and for cultures from the urinary system 99.2%. Their NPV for wound and urine cultures (93.1% and 99.2%) were similar to our study (92.6% and 99.7%). Another retrospective study by Noeldner and colleagues evaluated 1989 immunocompetent patients and yielded a NPV of 99.8% for blood, and 92.7% for bone and soft-tissue cultures using PCR assay. ^{Reference Noeldner, Bliek and Jones14} Overall, our current study had higher NPVs compared to the current literature in the immunocompetent population. ^{Reference Carr, Daley, Givens Merkel and Rose13} The lower MRSA infection rate at our institution and the fact that the majority of cultures obtained were from blood or urine sources were major contributors to the NPV. Additionally, immunocompromised patients often have many cultures drawn at hospital admission due to conditions such as neutropenic and tumor fever. This could result in a falsely elevated NPV; however, we were able to minimize this by only analyzing the first culture taken at each culture site during the hospital admission.

Patients with active cancer carry an even greater risk of infection due to immune suppression from chemo-radiation and frequent healthcare exposure, making it important to understand local MRSA prevalence and epidemiology when deciding on antibiotic therapy. A meta-analysis by Li and colleagues evaluated the pooled prevalence of MRSA bacteremia in cancer patients and showed the prevalence of MRSA among bloodstream infections was 1% in the Americas. ^{Reference Li, Zhuang, Wang, Wang and Dong15} Similarly, studies evaluating MRSA nasal swab screenings found a low prevalence rate of MRSA ranging from 4.1% to 5% among patients with febrile neutropenia and AML. ^{Reference Perreault, Binks, McManus and Topal10,Reference Talagtag, Patel and Scappaticci11} In our study, we had a lower prevalence rate of MRSA infections at 2.6%. Despite the low prevalence rate of MRSA infections in our patient population, the use of empiric anti-MRSA antibiotics remains high. In our study, 82% of patients received at least one anti-MRSA agent prior to MRSA nasal swab culture. The routine use of nasal swab screening may be useful for more rapid de-escalation and avoidance of long-term empirical anti-MRSA therapy. However, delaying anti-MRSA therapy de-escalation is much more common in immunocompromised patients due to the increased risk of organisms such as Streptococcus spp, Enterococcus spp, and coagulase-negative Staphylococci.

This study is the largest to date analyzing the predictive values of nasal swab culture screenings in the cancer population. Our study included a large number of patients and cultures, which increased the validity of the predictive value results. Limitations of our study include the retrospective study design potentially introducing confounding variables resulting from missing or incomplete data. Additionally, only the first hospital encounter per patient was included for analysis, which impacted the total number of clinical cultures and could have excluded positive MRSA cultures. Positive MRSA cultures were based on microbiology culture findings only, which could make it difficult to interpret a true infection versus colonization in respiratory and wound cultures. This limitation could potentially explain the higher PPV seen with the wound and respiratory clinical cultures. Furthermore, 82% of the patients received anti-MRSA therapy before the MRSA nasal swab culture. Prior MRSA therapy may reduce the sensitivity of the MRSA nasal swabs culture however, when antibiotics are administered within the first 48 hours, the reduction in sensitivity is limited. ^{Reference Shenoy, Noubary and Kim9} Lastly, we did not determine if the patients underwent decolonization which would impact the nasal swab culture results.

In conclusion, this retrospective study in patients with cancer showed a high specificity and negative predictive value of MRSA swab culture screenings, which is consistent with current literature for both immunocompromised and immunocompetent patients. These results may aid in antimicrobial stewardship activities and help guide the early discontinuation of empiric anti-MRSA antibiotics in patients with cancer.