Accuracy of a Negative Methicillin Resistant Staphylococcus aureus (MRSA) Screen in the Intensive Care Unit (ICU)

Accuracy of a Negative Methicillin Resistant Staphylococcus aureus (MRSA) Screen in the Intensive Care Unit (ICU)

 

Background: MRSA nasal colonization has been recognized as an independent risk factor for the development of MRSA infection during hospitalization and up to one year after discharge. However, the accuracy of a negative nasal swab (NS) for colonization in relation to MRSA infection and the length of time before becoming colonized in a hospital setting is not known.

 

Objective: Determine if a negative MRSA screen for nasal colonization can accurately exclude MRSA infection.

Methods: Retrospective cohort analysis of ICU patients screened using a NS for MRSA over 6 months. Detection of MRSA colonization was performed using real-time polymerase chain reaction (RT-PCR). NS were collected on admission, on transfer between units and on discharge (standard culture methods on agar plates). Data included patient demographics, NS results, risk factors (RF) for colonization, acquisition of MRSA infection, empiric vancomycin use, length of stay in the ICU, and readmission rates. We utilized chi-squared analysis for categorical variables. Statistically significant variables have a P value < 0.05.

Results: A total of 111 patient charts were reviewed. The baseline nasal MRSA colonization rate on admission was 10%. Forty-eight percent of patients had no RF for colonization and 19% had 2 or more RF. Only 1 of 111 patients (0.9%) with a negative NS converted to positive during hospitalization but had 2 RF. Two patients of 37 (6%) had MRSA infection on readmission. The sensitivity, specificity, negative (NPV) and positive predictive (PPV) value of a negative NS and no RF was 100%, 100%, 100% and 100% respectively. For a negative NS with ≥ 2 RF, it was 95%, 100%, 100% and 98%. A negative nasal MRSA screen on admission in the setting of zero RF for MRSA colonization correlated with lower rates of invasive MRSA infection (p value .007, OR 10.8, 99% CI 1.3-87). This relationship disappeared when analyzed with 2 or more RF. Empiric vancomycin was used in 26% of patients with a negative NS. Duration of therapy with vancomycin was longer in patients with zero RF and a negative NS. Please see Table 1 for detailed data.

Conclusions: A negative NS combined with zero RF for MRSA colonization on initial ICU admission and through hospital stay was highly predictive of no MRSA infection. However, the test was less sensitive when there were > 2 RF for MRSA colonization. Duration of therapy with vancomycin was longer in patients with a negative NS on admission. However, this was not a statistically significant finding. Further prospective evaluation needs to be performed.