Abstract
The D-zone test detects inducible clindamycin resistance in Staphylococcus spp. Two other methods not described by the Clinical and Laboratory Standards Institute (CLSI) are available to test for this resistance mechanism: an agar dilution method and new Vitek 2 cards. This study evaluated the performance of both methods in detecting inducible clindamycin resistance. Nonduplicate clinical strains of Staphylococcus spp. (111 Staphylococcus aureus and 52 coagulase-negative staphylococcus strains), intermediate or resistant to erythromycin but susceptible to clindamycin, were obtained from three hospitals in Montreal, Quebec, Canada. Molecular analysis to detect resistance genes was conducted on all strains. A Mueller-Hinton agar containing 1 mg of erythromycin and 0.5 mg of clindamycin/liter was used for the dilution method, and two inocula were tested: 104 and 105 CFU per spot. Plates were read at 24 and 48 h. The Vitek 2 AST-P580 card was used according to the manufacturer's recommendations. The results were compared to those of the D-zone test. The D-zone test was positive in 134 of 163 (82%) strains. With the 104 CFU inoculum, the sensitivities were 84 and 99% at 24 and 48 h, respectively. The 105 CFU inoculum increased the sensitivities at 24 and 48 h to 91 and 100%, respectively. The specificity was 100% for the 104 CFU inoculum at 24 h and 97% for the other combinations. The sensitivity and specificity for the Vitek 2 card were 93 and 100%, respectively. The performance of both the agar dilution method and the Vitek 2 card was good, but these methods were not as sensitive as the D-zone test at 24 h.
Inducible clindamycin resistance is caused by a ribosomal methylase encoded by erm genes; it results in resistance to macrolides, lincosamides, and streptogramin B antibiotics (MLSB phenotype). This phenotype can be either inducible (iMLSB) or constitutively expressed (cMLSB) (10, 11, 19). The inducible phenotype is expressed only in the presence of macrolides, but not lincosamides (10). In vitro, isolates harboring the iMLSB phenotype appear resistant to macrolides but susceptible to clindamycin. Treatment failures with the use of clindamycin have been described with iMLSB phenotype and raise concern about the selection of a cMLSB phenotype, especially in infections that are deep seated or with a large bacterial burden (11).
The Clinical and Laboratory Standards Institute (CLSI) recommends testing of erythromycin-resistant and clindamycin-susceptible isolates of Staphylococcus spp. by either D-zone test or broth microdilution to detect inducible clindamycin resistance (3). In some clinical microbiology laboratories, this increases the turnaround time by 24 h since isolates that meet the CLSI criteria have to be further tested by the D-zone test. One way to circumvent this problem would be to report all macrolide-resistant Staphylococcus spp. resistant to lincosamides but, because the prevalence of the iMLSB phenotype varies between different geographic locations and populations (1, 2, 6, 7, 14, 15, 18, 20, 22), it would prevent the use of clindamycin for patients that would likely respond to clindamycin therapy (20). Another option is to test all isolates of Staphylococcus spp. for inducible clindamycin resistance by performing the D-zone test on purity plates used with automated systems (9) but, for laboratories with a high workload, this can be time-consuming.
Two other methods not described in CLSI's M100 document allow inducible clindamycin resistance to be detected while conducting the other antimicrobial susceptibility tests. Fernandes et al. (5) recently described an agar dilution method to detect this resistance mechanism. A Mueller-Hinton agar with 3.3% defibrinated horse blood containing 0.5 mg of clindamycin and 1 mg of erythromycin/liter resulted in a sensitivity and specificity of 100%. The automated system Vitek 2 (bioMérieux, Marcy l'Étoile, France) now offers a panel that detects inducible clindamycin resistance directly.
The objectives of the present study were to evaluate the performance of the agar dilution method, using Mueller-Hinton agar unsupplemented with horse blood, and a Vitek 2 card in detecting inducible clindamycin resistance compared to the D-zone test.
(The results of the present study were presented in part at the 109th Annual Meeting of the American Society for Microbiology in Philadelphia, PA, May 2009 [poster C-010].)
MATERIALS AND METHODS
Collection and characterization of isolates.
Consecutive, nonduplicate clinical isolates of Staphylococcus spp. resistant (MIC, ≥8 mg/liter) or intermediate (MIC, 1 to 4 mg/liter) to erythromycin but susceptible to clindamycin (MIC, ≤0.5 mg/liter) were identified and collected from two teaching hospitals in Montreal, Quebec, Canada (35 isolates from Hôpital Saint-Luc du Centre Hospitalier Universitaire de Montréal [CHUM] and 55 isolates from Hôpital Maisonneuve-Rosemont [HMR]), between May and July 2008. A collection of 73 isolates that met the same selection criteria from a third teaching hospital, but collected from August 2006 to March 2007 (Hôpital Notre-Dame du CHUM), was also included. Routine antimicrobial susceptibility testing was done by the Vitek 2 system in both CHUM hospitals and by agar dilution in HMR. The D-zone test was performed according to the CLSI recommendations using a disk separation distance of 15 mm (3).
Molecular analysis of all isolates was conducted. The presence of ermA or ermC, the two most common genes implicated in inducible clindamycin resistance in Staphylococcus spp., was detected by a multiplex PCR according to previously outlined methods (12).
Agar dilution.
The induction agar was prepared with Mueller-Hinton agar (BBL, Sparks, MD) containing 0.5 mg of clindamycin (Nucro-Technics, Scarborough, Canada)/liter and 1 mg of erythromycin (Sigma-Aldrich, St. Louis, MO)/liter and poured in 100-mm plates to a depth of 5 mm. The decision not to supplement the agar with horse blood was made because all other agars used in the laboratory are unsupplemented with blood according to CLSI recommendations. Plates were stored at 4°C, and a new lot was produced each month. Plates with 0.5 mg of erythromycin/liter and plates without antibiotics were used as growth controls. Three to five colonies of an 18- to 24-h-old culture of Staphylococcus spp. were inoculated in Mueller-Hinton broth (BBL) and adjusted to a concentration equivalent to a 0.5 McFarland standard (∼108 CFU/ml). A 500-μl volume of the suspension was placed in a 37-well seed block. A manual inoculum replicator was then used. Pins of 1 mm, delivering ∼0.1 μl of suspension (∼104 CFU) per spot, were selected. Because of the possible inoculum effect (5), a second induction agar plate was also inoculated with 3-mm pins that delivered ∼1 μl (∼105 CFU) per spot. Plates were incubated at 35°C and read at 24 and 48 h. Growth of more than one colony on the induction agar plate was considered a positive result for inducible clindamycin resistance.
Vitek 2.
The Vitek 2 AST-P580 card (bioMérieux, Marcy l'Étoile, France) was used according to the manufacturer's recommendations. Briefly, three to five colonies of an 18- to 24-h-old culture of Staphylococcus spp. were inoculated in a 0.45% NaCl solution and adjusted to a concentration equivalent to a 0.5 to 0.63 McFarland standard. The solution was then loaded with the card in the Vitek 2 system. The incubation period was determined by the Vitek 2 system. Two wells are used to detect inducible clindamycin resistance in the Vitek 2 card: one with 0.5 mg of clindamycin/liter and another one with a combination of 0.25 and 0.5 mg of clindamycin and erythromycin/liter, respectively. Both the instrument and the Advanced Expert System (AES) results were considered.
Quality control.
Positive (ATCC BAA-977, ermA) and negative (ATCC BAA-976, efflux pump msrA) strains were used for the D-zone test, the agar dilution, and the Vitek 2 card. Because the PCR used was a multiplex PCR, two in-house isolates—one with an ermC gene and one with both ermA and ermC—were added to the ATCC BAA-976 and the ATCC BAA-977 strains with each run of PCR. These additional quality control strains made us confident that ermC could be detected in each run and that there was no competition between the two sets of primers.
Statistical analysis.
Confidence intervals (CI) were calculated using Stata 10.0 (Stata Corp., College Station, TX).
RESULTS
A total of 163 erythromycin-resistant or -intermediate, but clindamycin-susceptible Staphylococcus clinical isolates were included in the present study (Table 1). The iMLSB phenotype was detected by the D-zone test in 134 (82%) isolates. Of the D-zone test-positive isolates, all but one isolate were positive for ermA and/or ermC. All D-zone test-negative isolates were PCR negative for ermA and ermC genes.
TABLE 1.
Characteristics of clinical isolates of Staphylococcus spp.
| Identification | No. of isolates |
|||||
|---|---|---|---|---|---|---|
| D-zone test negativea | D-zone test positive |
Total | ||||
| ermA positive | ermC positive | ermA and ermC positive | ermA and ermC negative | |||
| MSSA | 8 | 64 | 21 | 0 | 1 | 94 |
| MRSA | 6 | 11 | 0 | 0 | 0 | 17 |
| CNS | 15 | 1 | 35 | 1 | 0 | 52 |
| Total | 29 | 76 | 56 | 1 | 1 | 163 |
All D-zone test-negative isolates were PCR negative for ermA and ermC genes.
Results of inducible clindamycin resistance by agar dilution, for both inocula at both reading times, compared to the D-zone test, are shown in Table 2. At 24 h, increasing the inoculum from 104 to 105 CFU/spot allowed the detection of inducible clindamycin resistance in 9 more strains, raising the sensitivity from 84 to 91%. Only the 105 CFU/spot inoculum, read at 48 h, permitted the detection of inducible resistance in all strains. Only one Staphylococcus isolate that was negative for inducible clindamycin resistance by D-zone test and negative for ermA or ermC grew on the induction agar plates with the 104 CFU/spot inoculum at 48 h and with the 105 CFU/spot inoculum at both 24 and 48 h. When only methicillin-sensitive Staphylococcus aureus (MSSA) and methicillin-resistant S. aureus (MRSA) were considered, the sensitivity and specificity were similar (data not shown).
TABLE 2.
Performance of the agar dilution method in detecting inducible clindamycin resistance in Staphylococcus spp. compared to the D-zone test
| Inoculum (CFU/spot) and time point | Agar dilution method result | No. of isolates with: |
% (95% CI) |
||
|---|---|---|---|---|---|
| Positive D-zone test | Negative D-zone test | Sensitivity | Specificity | ||
| 104 | |||||
| 24 h | Positive | 113 | 0 | 84 (77-90) | 100 (85-100) |
| Negative | 21 | 29 | |||
| 48 h | Positive | 133 | 1 | 99 (95-100) | 97 (80-100) |
| Negative | 1 | 28 | |||
| 105 | |||||
| 24 h | Positive | 122 | 1 | 91 (85-95) | 97 (80-100) |
| Negative | 12 | 28 | |||
| 48 h | Positive | 134 | 1 | 100 (97-100) | 97 (80-100) |
| Negative | 0 | 28 | |||
The performance of the Vitek 2 card is showed in Table 3. There was no difference between the results obtained from the instrument and the AES. Inducible clindamycin resistance was not detected in 10 strains (i.e., 5 MSSA and 5 coagulase-negative staphylococcus [CNS] strains) that were D-zone test positive and ermA or ermC positive by PCR (sensitivity of 93%). The specificity was 100%. When these strains were retested, 6 of the 10 were still negative with the Vitek 2 card the second time. Again, the results were similar when considering MSSA and MRSA separately (data not shown).
TABLE 3.
Performance of the Vitek 2 card in detecting inducible clindamycin resistance in Staphylococcus spp. compared to the D-zone test
| Vitek 2 result | No. of isolates with: |
% (95% CI) |
||
|---|---|---|---|---|
| Positive D-zone test | Negative D-zone test | Sensitivity | Specificity | |
| Positive | 124 | 0 | 93 (86-96) | 100 (85-100) |
| Negative | 10 | 29 | ||
DISCUSSION
In the context of increasing prevalence of community-acquired MRSA, alternative drugs to treat skin and soft tissue infections are needed. Clindamycin appears to be an interesting option because of the availability of an oral formulation, good bioavailability, and distribution in skin and abscesses (4). Resistance to clindamycin is highly variable in different patient populations (15, 20) and, if this drug is to be used, rapid susceptibility testing for inducible clindamycin resistance must be available. In order to choose an appropriate method for each laboratory and patient populations, the performance has to be evaluated.
For the agar dilution method, our study confirmed the importance of the inoculum effect. Unfortunately, the two inocula tested were not 100% sensitive at 24 h. Fernandes et al. (5) mentioned that horse blood was added in their agar because the opaque red background facilitates the detection of scanty growth. The absence of horse blood in the agar might explain why, even in the hands of experienced technicians, some positive results were missed at 24 h. It is also possible that without horse blood supplementation, other antibiotic concentrations should be used to reach 100% sensitivity at 24 h (5). Laboratories that choose to use unsupplemented agar should be aware of this difference in sensitivity and should probably keep the plates for 48 h. This will affect turnaround time, but with our isolates, 91% of iMLSb phenotype strains would have been reported resistant at 24 h (24 h earlier than with the use of the D-zone test when it is done only after the susceptibilities to erythromycin and clindamycin are known). Moreover, a second reading at 48 h is less labor-intensive in a laboratory using agar dilution than performing the D-zone test on isolates that meet the screening criteria.
The Vitek 2 system was reported to be 98% sensitive in detecting inducible clindamycin resistance in a study that tested 62 strains of Staphylococcus spp. (16). In two other studies that evaluated the new cards, the sensitivity for inducible clindamycin resistance detection was 99% (8, 21). In the present study, the Vitek 2 card failed to detect inducible clindamycin resistance in 10 strains (negative predictive value of 74%; 95% CI = 58 to 87%). All of these isolates were from clinical specimens of different patients and were found in the three participating hospitals. Half were MSSA harboring the ermA gene and the others were CNS harboring the ermC genes so we know it is not a problem associated with one single clone that would be present in the three hospitals. Moreover, these results were reproducible in more than half of the isolates. There were no false-positive results for inducible clindamycin resistance with the Vitek 2 card, so positive results can be reported without further confirmation with D-zone test. Our findings indicate that the Vitek 2 card would allow 93% of isolates to be correctly reported as resistant concomitantly with the other antimicrobial susceptibility results. Laboratories that want to reach 100% sensitivity would still have to test erythromycin resistant/clindamycin susceptible isolates that showed negative results for inducible clindamycin resistance with the Vitek 2 card with the D-zone test.
Interestingly, one strain of MSSA in our study was D-zone test positive and also positive for inducible clindamycin resistance by the agar dilution method and by the Vitek 2 card but was negative for ermA or ermC. This strain was tested in another laboratory for the presence of ermB by PCR (13) and was negative for this gene. At least two other similar strains have been reported (5, 17). The reason why this strain has an iMLSB phenotype but is not harboring a ermA, ermB, or ermC gene is still unclear. ermY genes have been described in Staphylococcus spp. (19). It is also possible that mutations at the target sites of primers for ermA, ermB, and ermC could be responsible for this observation.
This is the first study comparing these two methods of detection of inducible clindamycin resistance to D-zone test. We raise the concern that the agar dilution method is probably not as sensitive as previously described at 24 h if not supplemented by horse blood. We also raise the concern that some strains of Staphylococcus spp. with the iMSLB are not detected with the Vitek 2 card. Although the clinical significance of the iMSLB phenotype is still uncertain, testing of inducible clindamycin is recommended. Our results will help clinical microbiologists to decide which technique to use depending on their local epidemiology and the techniques already in place in their institution.
Acknowledgments
bioMérieux provided the Vitek 2 AST-P580 cards. We thank Ann Huletsky and the Centre de Recherche en Infectiologie (Centre Hospitalier Universitaire de Québec) for testing one of our isolates for the presence of the ermB gene. We also thank Nicole Bouvier, Michèle Gauthier, Hélène Guay, and Christiane Restieri for their technical support.
M.R. is supported by career award from the Fonds de la Recherche en Santé du Québec.
Footnotes
Published ahead of print on 17 February 2010.
REFERENCES
- 1.Ajantha, G. S., R. D. Kulkarni, J. Shetty, C. Shubhada, and P. Jain. 2008. Phenotypic detection of inducible clindamycin resistance among Staphylococcus aureus isolates by using the lower limit of recommended inter-disk distance. Indian J. Pathol. Microbiol. 51:376-378. [DOI] [PubMed] [Google Scholar]
- 2.Cetin, E. S., H. Gunes, S. Kaya, B. C. Aridogan, and M. Demirci. 2008. Macrolide-lincosamide-streptogramin B resistance phenotypes in clinical staphylococcal isolates. Int. J. Antimicrob. Agents 31:364-368. [DOI] [PubMed] [Google Scholar]
- 3.CLSI. 2008. Performance standards for antimicrobial susceptibility testing: 18th informational supplement. CLSI document M100-S18. Clinical Laboratory Standards Institute, Wayne, PA.
- 4.Daurel, C., C. Huet, A. Dhalluin, M. Bes, J. Etienne, and R. Leclercq. 2008. Differences in potential for selection of clindamycin-resistant mutants between inducible erm(A) and erm(C) Staphylococcus aureus genes. J. Clin. Microbiol. 46:546-550. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Fernandes, C. J., M. V. N. O'Sullivan, Y. Cai, F. Kong, X. Zeng, G. L. Gilbert, and G. Kotsiou. 2007. Agar dilution method for detection of inducible clindamycin resistance in Staphylococcus spp. J. Clin. Microbiol. 45:4018-4020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Fokas, S., M. Tsironi, M. Kalkani, and M. Dionysopouloy. 2005. Prevalence of inducible clindamycin resistance in macrolide-resistant Staphylococcus spp. Clin. Microbiol. Infect. 11:337-340. [DOI] [PubMed] [Google Scholar]
- 7.Goyal, R., N. P. Singh, V. Manchanda, and M. Mathur. 2004. Detection of clindamycin susceptibility in macrolide resistant phenotypes of Staphylococcus aureus. Indian J. Med. Microbiol. 22:251-254. [PubMed] [Google Scholar]
- 8.Griffith, R., S. Messina-Powell, D. Creely, M. Dante, M. Ullery, N. Ledeboer, and S. Shortridge. 2007. A new vancomycin test and inducible clindamycin resistance test for gram-positive organisms with the Vitek 2 systems, abstr. C-014. Abstr. 107th Gen. Meet. Am. Soc. Microbiol. American Society for Microbiology, Washington, DC.
- 9.Jorgensen, J. H., S. A. Crawford, M. L. McElmeel, and K. R. Fiebelkorn. 2004. Detection of inducible clindamycin resistance of staphylococci in conjunction with performance of automated broth susceptibility testing. J. Clin. Microbiol. 42:1800-1802. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Leclercq, R. 2002. Mechanisms of resistance to macrolides and lincosamides: nature of the resistance elements and their clinical implications. Clin. Infect. Dis. 34:482-492. [DOI] [PubMed] [Google Scholar]
- 11.Lewis, J. S., 2nd, and J. H. Jorgensen. 2005. Inducible clindamycin resistance in staphylococci: should clinicians and microbiologists be concerned? Clin. Infect. Dis. 40:280-285. [DOI] [PubMed] [Google Scholar]
- 12.Lina, G., A. Quaglia, M. E. Reverdy, R. Leclercq, F. Vandenesch, and J. Etienne. 1999. Distribution of genes encoding resistance to macrolides, lincosamides, and streptogramins among staphylococci. Antimicrob. Agents Chemother. 43:1062-1066. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Martineau, F., F. J. Picard, N. Lansac, C. Menard, P. H. Roy, M. Ouellette, and M. G. Bergeron. 2000. Correlation between the resistance genotype determined by multiplex PCR assays and the antibiotic susceptibility patterns of Staphylococcus aureus and Staphylococcus epidermidis. Antimicrob. Agents Chemother. 44:231-238. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Mohapatra, T. M., B. Shrestha, and B. M. Pokhrel. 2009. Constitutive and inducible clindamycin resistance in Staphylococcus aureus and their association with methicillin-resistant S. aureus (MRSA): experience from a tertiary care hospital in Nepal. Int. J. Antimicrob. Agents 33:187-189. [DOI] [PubMed] [Google Scholar]
- 15.Moore, Z. S., R. C. Jerris, and J. A. Hilinski. 2008. High prevalence of inducible clindamycin resistance among Staphylococcus aureus isolates from patients with cystic fibrosis. J. Cystic Fibrosis 7:206-209. [DOI] [PubMed] [Google Scholar]
- 16.Nakasone, I., T. Kinjo, N. Yamane, K. Kisanuki, and C. M. Shiohira. 2007. Laboratory-based evaluation of the colorimetric VITEK-2 Compact system for species identification and of the Advanced Expert System for detection of antimicrobial resistances: VITEK-2 Compact system identification and antimicrobial susceptibility testing. Diagn. Microbiol. Infect. Dis. 58:191-198. [DOI] [PubMed] [Google Scholar]
- 17.O'Sullivan, M. V. N., Y. Cai, F. Kong, X. Zeng, and G. L. Gilbert. 2006. Influence of disk separation distance on accuracy of the disk approximation test for detection of inducible clindamycin resistance in Staphylococcus spp. J. Clin. Microbiol. 44:4072-4076. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Patel, M., K. B. Waites, S. A. Moser, G. A. Cloud, and C. J. Hoesley. 2006. Prevalence of inducible clindamycin resistance among community- and hospital-associated Staphylococcus aureus isolates. J. Clin. Microbiol. 44:2481-2484. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Roberts, M. C., J. Sutcliffe, P. Courvalin, L. B. Jensen, J. Rood, and H. Seppala. 1999. Nomenclature for macrolide and macrolide-lincosamide-streptogramin B resistance determinants. Antimicrob. Agents Chemother. 43:2823-2830. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Schreckenberger, P. C., E. Ilendo, and K. L. Ristow. 2004. Incidence of constitutive and inducible clindamycin resistance in Staphylococcus aureus and coagulase-negative staphylococci in a community and a tertiary care hospital. J. Clin. Microbiol. 42:2777-2779. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Sharp, S., K. Weber, and K. Bennett. 2009. Detection of inducible clindamycin resistance, mecA-mediated oxacillin resistance, and vancomycin resistance with the Vitek 2 GP67 card as compared to the D-test, and cefoxitin and vancomycin disk diffusion testing in clinically significant isolates of Staphylococcus aureus, abstr. C-007. Abstr. 109th Gen. Meet. Am. Soc. Microbiol. 2009. American Society for Microbiology, Washington, DC.
- 22.Yilmaz, G., K. Aydin, S. Iskender, R. Caylan, and I. Koksal. 2007. Detection and prevalence of inducible clindamycin resistance in staphylococci. J. Med. Microbiol. 56:342-345. [DOI] [PubMed] [Google Scholar]