JAC Advance Access originally published online on March 5, 2007
Journal of Antimicrobial Chemotherapy 2007 59(4):763-766; doi:10.1093/jac/dkl548
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Comparison of cefoxitin and moxalactam 30 µg disc diffusion methods for detection of methicillin resistance in coagulase-negative staphylococci
1 Microbiology Laboratory 2 Biostatistics Department, Necker-Enfants Malades Hospital, Faculté de Médecine Paris—Descartes, 149 rue de Sèvres, 75015, Paris, France
* Corresponding author. Tel: +33-1-44-49-49-61; Fax: +33-1-44-49-49-60; E-mail: join{at}necker.fr
Received 5 October 2006; returned 2 November 2006; revised 1 December 2006; accepted 17 December 2006
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResults and discussionTransparency declarationsReferences |
|---|
Objectives: To compare cefoxitin and/or moxalactam 30 µg disc diffusion (DD) methods to detect methicillin resistance in coagulase-negative staphylococci (CoNS) using both high- and low-density (HD/LD) inoculum techniques.
Methods: A challenge set of 192 CoNS was tested. DD test results were compared with PBP2a detection.
Results: With the LD inoculum, the sensitivity/specificity of cefoxitin and moxalactam were 94.4%/100% and 100%/92.4%, respectively, using the DD breakpoints of the Comité de l'Antibiogramme de la Société Française de Microbiologie. With the HD inoculum, the sensitivity/specificity of cefoxitin and moxalactam were 93.7%/100% and 100%/96.9%, using the cefoxitin DD breakpoints of the CLSI and a resistant/susceptible breakpoint of < 20 mm/
20 mm for moxalactam. Comparison of receiver operating characteristic AUCs did not show significant difference between studied assays, but the overlapping zone where both PBP2a-positive and PBP2a-negative isolates were observed concerned a lower number of strains with moxalactam than with cefoxitin (P < 0.001). Combination of cefoxitin and moxalactam DD methods demonstrated that all isolates with a concordant cefoxitin/moxalactam phenotype were correctly classified. Interestingly, all isolates misclassified by each DD method used alone were cefoxitin-susceptible and moxalactam-resistant.
Conclusions: Although all DD methods studied here performed well for detecting methicillin resistance in CoNS, moxalactam had a higher accuracy than cefoxitin to differentiate heteroresistant isolates from PBP2a-negative strains. Identification of isolates that should be submitted to a confirmatory test to conclude on methicillin resistance can be easily obtained by combining cefoxitin and moxalactam DD methods.
Keywords: heteroresistance , PBP2a , conditional logistic regression
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionTransparency declarationsReferences |
|---|
The phenotypic detection of methicillin resistance in coagulase-negative staphylococci (CoNS) has been long considered difficult partly due to the heterogeneous expression of mecA that is more common in CoNS than in S. aureus.1,2 The Cefoxitin Study Group recently showed that cefoxitin disc diffusion (DD) test with a high-density inoculum (HD) gave better results than other phenotypic methods to detect methicillin resistance in CoNS. Nevertheless, it has been reported that this method failed to detect resistance in some strains of Staphylococcus epidermidis, Staphylococcus hominis and Staphylococcus simulans.3–5 This lack of sensitivity of cefoxitin has been related to CoNS strains that weakly express methicillin resistance (heteroresistant isolates).3,5
The Comité de l'Antibiogramme de la Société Française de Microbiologie (CASFM) recommends cefoxitin and/or moxalactam 30 µg discs to assess methicillin resistance both in Staphylococcus aureus and CoNS, using a low-density (LD) inoculum.6 To our knowledge however, neither the efficacy of cefoxitin 30 µg discs using an LD inoculum nor the efficacy of moxalactam to detect methicillin resistance in CoNS have been reported. We conducted this study to compare the accuracy of both molecules to detect methicillin resistance in CoNS, using either an LD or an HD inoculum and to investigate the interest of combining the two methods.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionTransparency declarationsReferences |
|---|
Bacterial strains
We studied 192 of 613 CoNS clinical isolates obtained from various specimens sent to the Microbiology Laboratory of the Necker-Enfants Malades Hospital during a 6 month period. Staphylococcus saprophyticus isolates were excluded from the study.4 We focused on heteroresistant strains and thus included all isolates with a cefoxitin inhibition zone diameter range of 19–31 mm (n = 65) as routinely determined using cefoxitin 30 µg discs with an LD inoculum.6 A panel of 72 and 55 randomly selected strains with a cefoxitin inhibition zone diameter > 31 mm and < 19 mm, respectively, was also studied. All selected strains were identified with the ID 32 Staph system (API-bioMérieux, Marcy-l'Etoile, France). Identification was confirmed in some cases by sequencing assays characterizing an internal fragment of the sodA gene.7 The methicillin-susceptible S. aureus strain ATCC 25923 and a methicillin-resistant clinical S. aureus strain (NEMSAMR 502422) were included as controls for disc diffusion tests, PBP2a agglutination and mecA PCR.
Cefoxitin and moxalactam susceptibility tests were performed using both high- and low-density (HD/LD) inoculum techniques, according to the guidelines of the CLSI and CASFM, respectively.4,6 Mueller–Hinton agar plates were inoculated by swabbing and allowed to dry for five minutes. Moxalactam (30 µg) and cefoxitin (30 µg) discs (Bio-Rad, Marnes la Coquette, France) were then laid on the surface. Plates were incubated for 18 h at 35°C. Susceptibility to cefoxitin and moxalactam with an LD inoculum was interpreted according to the DD breakpoints of the CASFM (resistant < 25 mm/susceptible 27 mm and resistant < 23 mm/susceptible 24 mm, respectively).6 Susceptibility to cefoxitin with an HD inoculum was interpreted according to CLSI guidelines (resistant < 25 mm/susceptible 25 mm).4
All isolates were screened for the presence of PBP2a with the Oxoid PBP2' latex agglutination test, following the manufacturer's instructions but using a high-density inoculum as previously published.8 PBP2a agglutination was considered as gold standard for detecting methicillin resistance. We confirmed this test by detection of mecA using real-time PCR (5'–exonuclease qPCR, TaqMan),8 on all strains that showed discrepant results between DD tests and PBP2a or an intermediate susceptibility to cefoxitin or moxalactam.6
Phenotypic DD assays were first analysed by determining their receiver operating characteristic (ROC) AUCs. ROC AUCs were compared using the DeLong method.9 AUC 95% confidence intervals were estimated from 100 000 bootstrap samples of the original dataset.10 An exact conditional logistic regression approach was also used to specifically assess the relationship between the assay characteristics and the probability for an isolate to belong to the assay's overlapping region. Computations were done with the SAS system V9.1.3 (SAS Institute Inc., Cary, NC, USA).11 The sensitivity and specificity of DD tests, defined as the percentage of PBP2a-positive strains determined to be methicillin-resistant and the percentage of PBP2a-negative strains determined to be methicillin-susceptible, respectively, were then calculated according to CASFM and CLSI DD breakpoints.4,6 Strains displaying an intermediate susceptibility to cefoxitin or moxalactam were excluded from the analysis since the CASFM considers that an additional test (PBP2a or mecA PCR) should be performed to conclude on methicillin resistance in this case.
|
Results and discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionTransparency declarationsReferences |
|---|
The selected studied isolates included 122 S. epidermidis strains, 21 S. hominis, 16 Staphylococcus capitis, 9 Staphylococcus warneri, 9 Staphylococcus haemolyticus, 6 S. simulans, 5 Staphylococcus lugdunensis, 2 Staphylococcus cohnii, 1 Staphylococcus xylosus and 1 Staphylococcus gallinarum. PBP2a agglutination tests were positive for 95 of the 192 strains (49.5%).
Whatever the inoculum used, a strong correlation was observed between cefoxitin and moxalactam inhibition zone diameters (R2 = 0.92 and 0.94 with the LD and the HD inoculum, respectively, P < 0.0001). Inhibition zone diameters obtained with the HD inoculum were 1.8 ± 1.9 mm and 1.9 ± 2.1 mm (mean ± SD) smaller than with the LD inoculum for cefoxitin and moxalactam, respectively, demonstrating an inoculum effect for both molecules. As previously shown for other phenotypic DD methods,2 an overlapping zone containing both PBP2a-positive and PBP2a-negative strains was always observed (Figure 1). Global analysis showed that ROC AUCs were not significantly higher for moxalactam (AUCMOX LD = 0.9995, 95% confidence interval (CI): 0.9982–1; AUCMOX HD = 0.9998, CI: 0.9993–1) than for cefoxitin (AUCFOX LD = 0.9987, CI: 0.9960–1; AUCFOX HD 0.9986, CI: 0.9959–1). Nevertheless, the number of isolates within the overlapping region (Figure 1) was lower with moxalactam (MOX-HD n = 4 and MOX-LD n = 7) than with cefoxitin (FOX-HD n = 10, FOX-LD n = 15). Conditional logistic regression confirmed that the probability for an isolate to be in the overlapping region was significantly lower for moxalactam assays compared with cefoxitin (P < 0.001), but not for HD inoculum assays compared with LD ones (P = 0.15). These data demonstrate that although all DD methods tested here displayed a similar accuracy to detect methicillin resistance in CoNS, moxalactam better differentiated heteroresistant isolates from PBP2a-negative strains.
|
Whatever the inoculum used and using CASFM and CLSI guidelines, the specificity of cefoxitin was 100%. The sensitivity of cefoxitin was 94.4% with the LD inoculum and 93.7% with the HD inoculum, with 5 and 6 PBP2a-positive isolates misidentified as methicillin-susceptible, respectively (Table 1). These data confirm that some CoNS isolates are misclassified as methicillin-susceptible by the cefoxitin-HD inoculum DD method,3,5 but also using an LD inoculum as recommended by the CASFM. The five cefoxitin-intermediate isolates (LD inoculum technique) were PBP2a-positive, suggesting that they should rather be considered as cefoxitin resistant (Table 1).
|
Disc diffusion breakpoints for moxalactam are only available for the LD inoculum technique.6 In these conditions, the sensitivity of moxalactam was 100% and its specificity 92.4%, as seven PBP2a-negative CoNS were misclassified as methicillin-resistant. Most of these isolates (n = 6) were identified as non-S. epidermidis strains (Table 1). The five moxalactam-intermediate isolates were PBP2a-negative (Table 1) suggesting that they should rather be considered as moxalactam-susceptible (Table 1). With an HD inoculum, moxalactam resistant/susceptible DD breakpoints of < 20 mm/
20 mm gave a sensitivity/specificity of 100%/96.9%. Using resistant/susceptible DD breakpoints of < 19 mm/19 mm, the specificity of moxalactam increased to 100% but its sensitivity decreased to 98.9%. These data demonstrate that moxalactam discs can be used as well as cefoxitin with an HD inoculum to detect methicillin resistance in CoNS. The interest of using moxalactam 30 µg discs is also that inhibition zone diameters are smaller than those of cefoxitin 30 µg discs, and are therefore more adapted to be used together with other antibiotic discs on a single 9 cm agar plate. Because the CASFM recommends the use of cefoxitin and/or moxalactam to detect methicillin resistance in CoNS, we investigated whether a combination of these methods would be of interest. For this analysis, taking into account PBP2a agglutination results, all cefoxitin intermediate isolates were considered as cefoxitin-resistant and all moxalactam intermediate isolates were considered as moxalactam-susceptible. With the HD inoculum, the moxalactam resistant/susceptible DD breakpoint of 20 mm was used. If the results of cefoxitin and moxalactam tests were concordant, the phenotypic determination of methicillin resistance systematically correlated with the result of PBP2a, giving a sensitivity and specificity of 100% for this test. This result was expected since we thus combined the 100% specificity of cefoxitin with the 100% sensitivity of moxalactam DD techniques. Of interest, all strains that were misclassified by each DD method alone (Table 1) were easily identified, because they showed a cefoxitin-susceptible/moxalactam-resistant phenotype.
In conclusion, all variant DD methods tested showed a similar intrinsic performance to detect methicillin resistance in CoNS, although moxalactam better differentiated heteroresistant isolates from PBP2a-negative strains. Determination of methicillin resistance in CoNS with these phenotypic DD methods should always include an intermediate susceptibility zone where interpretation is not permitted and a supplementary test should be performed to conclude on methicillin resistance. The combined cefoxitin-moxalactam technique is an accurate alternative method to identify heteroresistant isolates.
|
Transparency declarations |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionTransparency declarationsReferences |
|---|
None to declare.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionTransparency declarationsReferences |
|---|
1 Archer GL and Climo MW. (1994) Antimicrobial susceptibility of coagulase-negative staphylococci. Antimicrob Agents Chemother 38:2231–7.
2
Skov R, Smyth R, Larsen AR, et al. (2005) Evaluation of cefoxitin 5 and 10 microg discs for the detection of methicillin resistance in staphylococci. J Antimicrob Chemother 55:157–61.
3
Swenson JM and Tenover FC. (2005) Results of disk diffusion testing with cefoxitin correlate with presence of mecA in Staphylococcus spp. J Clin Microbiol 43:3818–23.
4 National Committee for Clinical Laboratory Standards. (2005) Performance Standards for Antimicrobial Susceptibility Testing. Fifteenth Informational Supplement. M100-S15(NCCLS, Villanova, PA, USA).
5
Frigatto EA, Machado AM, Pignatari AC, et al. (2005) Is the cefoxitin disk test reliable enough to detect oxacillin resistance in coagulase-negative staphylococci? J Clin Microbiol 43:2028–9.
6 Communiqué 2006 du Comité de l'Antibiogramme de la Société Française de Microbiologie. http://www.sfm.asso.fr/nouv/general.php?pa=2 (29 November 2006, date last accessed).
7
Poyart C, Quesne G, Boumaila C, et al. (2001) Rapid and accurate species-level identification of coagulase-negative staphylococci by using the sodA gene as a target. J Clin Microbiol 39:4296–301.
8
Horstkotte MA, Knobloch JK, Rohde H, et al. (2001) Rapid detection of methicillin resistance in coagulase-negative staphylococci by a penicillin-binding protein 2a-specific latex agglutination test. J Clin Microbiol 39:3700–02.
9 DeLong ER, DeLong DM, Clarke-Pearson DL. (1988) Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 44:837–45.[CrossRef][ISI][Medline]
10 Efron B and Tibshirani RJ. (1993) An Introduction to the Bootstrap. Monographs on Statistics and Applied Probability No. 57(Chapman & Hall, New York).
11 Mehta CR and Patel NR. (1995) Exact logistic regression: theory and examples. Stat Med 14:2143–60.[ISI][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||