|Home | About | Journals | Submit | Contact Us | Français|
MRSASelect agar (Bio-Rad, Redmond, WA) was evaluated for its performance in detecting MRSA directly from positive blood cultures containing Gram-positive cocci in clusters. Agar plates were evaluated for the presence of pink colonies at 18 to 24 h. Results were compared to organism identification by using standard laboratory methods. Confirming coagulase on pink isolates, the sensitivity and specificity were both 99%.
Staphylococcus aureus is an important human pathogen that causes a wide variety of nosocomial and community-acquired infections, ranging from skin and soft tissue infections to life-threatening bacteremia and sepsis. During the past 20 years, the incidence of both health care- and community-acquired Staphylococcus aureus bloodstream infections has significantly increased, and more than 55% of the isolates are now reported to be methicillin-resistant Staphylococcus aureus (MRSA) (1, 26, 32). Bloodstream infections due to MRSA have been shown to place patients at higher risk for additional health problems, including acute renal failure, prolonged hospitalization and intensive care needs, and development of ventilator dependency (5, 13). Mortality rates for patients with MRSA bacteremia are reported to be between 23% and 54% (5, 20, 30). Early identification of MRSA in blood cultures has an important role in patient care management (3); patients with S. aureus bacteremia often have significant delays in receiving the appropriate antimicrobial therapy or initially receive inappropriate antimicrobial therapy associated with higher in-hospital mortality rates (22, 23). Following the detection of Gram-positive cocci in clusters in clinical blood cultures, confirmatory identification of MRSA takes 24 to 48 h by the use of routine conventional and automated identification methods.
Because conventional methods of isolate identification are slow and because rapid determination of susceptibility among staphylococci has been shown to have an impact on clinical outcomes (2, 3), clinical microbiology laboratories have sought more-direct methods of determining oxacillin resistance. Recent introduction of molecularly based technology (PCR) for detection of MRSA has provided opportunities for more rapid detection of MRSA in clinical specimens (28, 31, 33). However, these methods often require expensive equipment and reagents in addition to a support infrastructure for appropriate utilization and work flow. Particularly for microbiology laboratories with fewer resources, molecularly based technologies may not be the solution to rapid MRSA detection from clinical specimens.
Various selective and differential chromogenic media have been approved for the detection of MRSA in pure cultures as well as from nasal swab specimens. The current chromogenic media allow for a turnaround time for reporting of MRSA that ranges from 18 to 48 h, depending on the type of agar used (7, 14, 15, 24). Various studies have demonstrated excellent performance of these media in detecting MRSA from naris swab specimens (17, 19, 21). Fewer studies demonstrate the utility of chromogenic agar media in the detection of MRSA from clinical specimens other than naris swabs (12, 18, 27).
(Part of this research was presented at the 109th General Meeting of the American Society for Microbiology, Philadelphia, PA [28a].)
In the current study, we evaluated the performance of the Bio-Rad MRSASelect agar for its ability to rapidly and correctly identify MRSA from 652 positive blood culture bottles (n = 652 patients) containing Gram-positive cocci in clusters. The study was approved by the Institutional Review Board of the Johns Hopkins University, School of Medicine. The investigations were performed at two microbiology laboratory sites, the Johns Hopkins Bayview Medical Center (JHBMC) and the Johns Hopkins Hospital (JHH). Between November 2008 and March 2009, a total of 652 consecutive clinical blood cultures from the Versa Trek system (Trek Diagnostic Systems, Cleveland, OH) (n = 249 from JHBMC) and the BacT/Alert system (bioMérieux, Durham, NC) (n = 403 from JHH), positive for “Gram-positive cocci in clusters” identified on Gram stain, were inoculated onto MRSASelect agar (Bio-Rad Diagnostics, Redmond, WA) for detection of MRSA (4). In addition to MRSASelect agar, blood from all blood culture bottles was inoculated onto sheep blood agar (SBA) plates (Trypticase soy agar with 5% sheep blood; BD Diagnostics Inc., Sparks, MD) by following standard laboratory procedures at each participating site. Both aerobic and anaerobic blood culture bottles in both systems were screened. All types of blood culture bottles, aerobic and anaerobic, positive for “Gram-positive cocci in clusters” were used in this study. However, only one type of blood culture bottle, aerobic or anaerobic, from one set, i.e., patient, was included for the subsequent data analysis. Positive blood culture bottles were gently mixed, and 2 ml of broth was then aseptically removed with a syringe. All SBA plates and MRSASelect agar plates were inoculated by streaking 1 to 2 drops of broth from the blood culture bottle. Blood agar plates were incubated by following standard laboratory procedures. MRSASelect agar plates were incubated at 35°C without CO2 in an inverted position and protected from light exposure, according to manufacturer's instructions (4). MRSASelect agar plates were examined for bacterial growth, and the presence of small pink colonies (positive for MRSA) at 18 to 24 h. If no pink colonies were identified at that time, the specimen was considered negative for MRSA. To further validate the manufacturer's claim of using the 18- to 24-h interval for final evaluation of the MRSASelect agar, all MRSA-negative plates were incubated for an additional 24 h. However, no additional pink colonies indicative of MRSA were identified after the additional 24 h of incubation period, verifying in our laboratories the manufacturer's claim for the need of 18 to 24 h of incubation before the final result report. Colonies that appeared pink on MRSASelect agar were compared to the bacterial growth on SBA, and bacterial colonies were further tested using the catalase test, the Staphaurex test (Remel, Lenexa, KS), or the tube coagulase test to confirm the presence of S. aureus. Coagulase and catalase testing was performed on colonies from SBA and MRSASelect agar, and the results were 100% concordant. All bacterial growth identified on blood agar plates was further examined using each participating laboratory's algorithms and test methods for isolate identification and antimicrobial susceptibility testing (AST). For AST, the MicroScan (Siemens, Sacramento, CA) was used at JHBMC and the Phoenix microbial identification and AST system (Becton Dickinson [BD], Sparks MD) was used at JHH. All MRSA detected by both MRSASelect agar and conventional methods was also confirmed by the cefoxitin disk method (30 μg on Mueller-Hinton agar) by following CLSI standards and criteria (11, 16, 29). Results for detection of MRSA by use of the MRSASelect agar were compared to the culture and identification of organisms by the use of conventional methods and cefoxitin screen. PCR testing for mecA was not performed in this study; the phenotypic detection methods in conjunction with AST were considered the definitive result. Statistical analyses, measures of association, and descriptive statistics were performed using Stata 9 (Stata Corporation, TX).
A combined total of 652 blood culture bottles from separate blood culture sets and unique patients, positive for “Gram-positive cocci in clusters” on Gram stain, were included in this study (Table (Table1).1). All organisms grew in pure culture without evidence of contamination. A total of 103 isolates of MRSA were identified in this study. Of those, one isolate was not detected by the MRSASelect agar (JHH study site). No differences in detection of MRSA with regard to the blood culture bottle type were seen. A total of 549 Gram-positive cocci/bacterial organisms, other than MRSA, were identified and confirmed by conventional, standard laboratory methods. Of these isolates, 539 isolates were negative for MRSA on screening with MRSASelect agar and confirmed by other, conventional standard laboratory methods as non-MRSA isolates. These isolates did not grow on MRSASelect agar or grew as white, nonpink colonies. Ten isolates presented as faintly pink colonies on MRSASelect agar at the 18- to 24-h reading of plates. Of the 10 non-MRSA isolates with pink colony appearance on MRSASelect agar, one isolate was confirmed as a Micrococcus species, another isolate was identified as methicillin-susceptible Staphylococcus aureus (MSSA) by the cefoxitin disk method, and the remaining eight isolates were identified as coagulase-negative staphylococci (CoNS), using the Staphaurex test. All 10 “false-positive” isolates occurred at the same testing site (JHBMC), and that site did not report any “false-negative” isolates. Detailed results are summarized in Table Table2.2. The initial performance of the MRSASelect agar without use of ancillary laboratory tests demonstrated a sensitivity of 99%, specificity of 98%, a positive predictive value (PPV) of 91%, and a negative predictive value (NPV) of 100%. A secondary analysis was performed considering the use of ancillary laboratory tests (i.e., coagulase detection) on the pink isolates from MRSASelect agar to differentiate between Staphylococcus aureus and CoNS. Considering the eight CoNS species and one Micrococcus as being true negatives, the revised calculations were as follows: sensitivity, 99%; specificity, 99%; PPV, 99%; and NPV, 100%.
Standard culture media together with conventional laboratory tests and AST are the most commonly used media and techniques for the identification of MRSA from clinical blood culture specimens. Various chromogenic media have been shown to reliably identify MRSA with sufficient sensitivity and specificity for routine use (14, 15, 17). Other studies have described the use of various phenotypic and molecular methods for the rapid identification of MRSA from various different clinical specimens, including blood cultures (18, 25, 27). Considering the longer turnaround time for conventional culture and AST methods, and the technical requirements and costs associated with molecular tests, a simple, culture-based approach for rapid detection of MRSA in blood cultures, such as chromogenic media, is desirable.
One method to detect the altered gene product of mecA, namely penicillin binding protein (PBP2a), in MRSA isolates is the PBP2a latex agglutination test (Oxoid, Hampshire, United Kingdom). When applied to isolates from primary culture media, this test reliably differentiates between MRSA and MSSA at a cost that is reasonable to most laboratories (6, 8). The PBP2a latex agglutination test, however, has shown greater variability in its performance when used for the direct detection of MRSA from blood cultures (10).
In this study, we demonstrated the utility of another chromogenic agar type for the rapid detection of MRSA from blood culture bottles. While initially 10 bacterial, not-MRSA organisms grew as faintly pink colonies on MRSASelect agar, the additional use of coagulase tests (e.g., tube coagulase testing or Staphaurex) rapidly identified nine isolates as non-MRSA. CoNS have been previously described as giving false-positive results on MRSA chromogenic media (27). Tube coagulase testing has been found to be highly specific for S. aureus, despite occasional false-negative results having been reported (9, 34). In our study, 2% of non-MRSA isolates presented with a faint-pink colony morphology; however, we found that with the use of coagulase testing in conjunction with the MRSA chromogenic media, CoNS can be rapidly excluded. We suggest using a coagulase test method as an ancillary step in MRSA identification on MRSASelect agar from blood cultures, to avoid the misidentification of CoNS as MRSA. In our study, the use of additional testing for confirmation of S. aureus resulted in improved sensitivity, specificity, and predictive values of the MRSASelect agar. We did not investigate the performance of the MRSASelect agar in mixed blood cultures. Laboratories considering the use of MRSASelect agar for direct detection of MRSA in blood cultures may want to include mixed blood cultures for their in-house validation. For the purpose of this study, we used only one positive blood culture bottle per set, as we wanted to include as many organisms from different patients as possible. In clinical settings, investigators could use one MRSASelect agar plate, divided in half, for inoculation of broth from both aerobic and anaerobic bottles from the same set, i.e., same patient.
After initial detection of Gram-positive cocci in clusters in blood cultures, the turnaround time for final result reporting of MRSA from blood cultures ranges from 36 to 48 h in our laboratories, using standard conventional laboratory test methods. This average turnaround time is based on the time necessary to have sufficient growth on agar media and the subsequent time necessary for organism identification and AST by automated systems. In this study, all MRSA isolates were identified within 18 to 24 h after the initial identification of a blood culture with growth of Gram-positive cocci in clusters.
Cost considerations are often the reason why more rapid detection methods are not used in microbiology laboratories with fewer financial resources. The average cost for one MRSASelect agar plate is $5, while the average cost for molecular test methods is substantially higher. The list price for the BD GeneOhm MRSA assay is $29.50 per test, and the list price for the Cepheid Xpert MRSA assay is $42.00. Additional costs for the analyzers necessary to perform these assays range between $35,000 and $78,000. The use of MRSASelect agar combined with Gram staining and coagulase testing for confirmation of suspect colonies afforded us the possibility to detect MRSA from clinical blood cultures 12 to 24 h sooner than and yet as cost-effectively as would conventional laboratory methods.
Bio-Rad Laboratories, Redmond, WA, provided the MRSASelect agar plates for use in this study at the Johns Hopkins Bayview Medical Center microbiology laboratory site.
Published ahead of print on 14 April 2010.