ABSTRACT
Non-diphtheriae Corynebacterium-associated disease has been increasingly observed and often presents a conundrum to the treating physician. Analysis of antibiotic susceptibility testing data for 1,970 clinical Corynebacterium isolates received between 2011 and 2016 revealed that empirical drug treatment options are limited to vancomycin and linezolid. Corynebacterium striatum was the most frequently observed species during this study period, along with C. amycolatum and C. pseudodiphtheriticum/C. propinquum. Low levels of susceptibility to penicillin (14.5%), erythromycin (15.1%), and clindamycin (8.7%) were observed for non-diphtheriae Corynebacterium species, while 3.0% of isolates were not susceptible to daptomycin. Similarly, 26.9% and 38.1% of Corynebacterium isolates were susceptible to ciprofloxacin and trimethoprim-sulfamethoxazole, respectively. Our data show much lower susceptibility to penicillin than previously reported in the literature and an increasing number of isolates resistant to daptomycin, highlighting the need for continued antibiotic surveillance studies for appropriate patient management and treatment success.
INTRODUCTION
Corynebacterium species are Gram-positive catalase-positive rod-shaped bacteria often referred to as “diphtheroids.” Effective vaccination programs have resulted in a decrease in Corynebacterium diphtheriae-related infection cases in the last 50 years. However, in the past 2 decades, there has been a rise in disease due to non-diphtheriae Corynebacterium species, with a variety of infections reported, including skin and soft tissue infections, prosthetic joint infections, bacteremia, respiratory infections, pneumonia, meningitis, surgical site infections, urinary tract infections, peritonitis, and endocarditis (1–3). Of additional concern, some of the non-diphtheriae Corynebacterium species have been shown to be resistant to multiple classes of antibiotics, thus potentially limiting effective empirical treatment (1–4).
This study presents the in vitro susceptibility profiles of 1,970 isolates of non-diphtheriae Corynebacterium species to 19 different antibiotics that were tested at the Public Health Ontario Laboratory (PHOL; Ontario, Canada) from 2011 to 2016.
RESULTS AND DISCUSSION
This study evaluated the antibiotic susceptibility test results of 1,970 isolates of nondiphtherial corynebacteria that were submitted to the PHOL from 2011 to 2016. Within this group, a total of 42 different Corynebacterium species were represented, as well as 24 isolates that could only be identified to the genus level and were reported as Corynebacterium species. The majority of the isolates belonged to C. striatum (47%), followed by C. pseudodiphtheriticum/C. propinquum (10.96%), C. amycolatum (9.64%), C. afermentans subsp. afermentans (5%), C. minutissimum (4%), C. jeikeium (4%), C. coyleae (3%), and C. urealyticum (3%). Other Corynebacterium species tested during this time period included C. resistens, Corynebacterium CDC group G, C. accolens, C. glucuronolyticum, and others (see Table S1 in the supplemental material). In our study, C. striatum was the most commonly submitted species for identification and susceptibility testing, whereas other reports have noted C. amycolatum as the most prevalent nondiphtherial corynebacteria reported (1, 5). Despite being the most commonly submitted species, there was no significant difference in the proportions of C. striatum isolates submitted over the years (P = 0.45) (data not shown). Additionally, no significant trend was observed for the top five commonly encountered corynebacterial species in terms of the proportion submitted for identification and susceptibility testing over time.
Isolates were predominantly cultured from blood (n = 479 [24.3%]), followed by sputum (n = 382 [19.4%]) (Table S2). A number of isolates were also recovered from bone (n = 97 [4.9%]) and synovial fluid (n = 41 [2.0%]) (Table S2).
Between 2011 and 2016, 28 different corynebacterial species were isolated and identified from blood, with C. striatum as the most frequent (26.2%), followed by C. afermentans (19.8%), C. coyleae (9.8%), and C. minutissimum (7.7%). Blood culture was the most common source for C. mucifaciens (92% [n = 13]), C. afermentans (83% [n = 89]), and C. coyleae (81% [n = 58]). The most frequently recovered species from sputum specimens were C. striatum (63%) and C. pseudodiphtheriticum/C. propinquum (35%). For C. striatum, the chief specimen sources were sputum (26%), blood (14%), tissue (11%), bone (9%), wounds (6%), bronchoalveolar lavage (BAL) and/or pleural fluid (3%), and peritoneal fluid (2%).
Of the 1,970 isolates submitted for susceptibility testing, 1,010 isolates were recovered from male patients (51.3%), 739 isolates were from female patients (37.5%), and 221 specimens were from a patient of unknown sex. While 15 (0.7%) specimens were recovered from patients less than 1 year of age, 24 (1.2%) specimens were from patients age 1 to 17 years, 844 (42.8%) specimens were from patients age 18 to 64 years, and 1,070 (54.3%) specimens were from patients age 65 or older; the remaining 17 isolates did not have the patient's age on the submission form.
The susceptibility profiles of the 1,970 nondiphtherial corynebacterial isolates showed that they were universally susceptible to vancomycin and linezolid (Table 1; see also Data Set S1 in the supplemental material). The majority of isolates (>90%) were susceptible to rifampin, gentamicin, and quinupristin-dalfopristin. Interestingly, 45 out of the 1,959 isolates tested for daptomycin had an MIC of ≥2 mg/liter, which is considered “nonsusceptible” based on the current CLSI interpretative criteria (6). Overall as a group, the nondiphtherial corynebacteria in this study demonstrated low rates of susceptibility to penicillin (14.5%), erythromycin (15.12%), clindamycin (8.7%), and ciprofloxacin (26.9%). Based on the EUCAST clinical breakpoint (30), 27.7% of nondiphtherial corynebacterial isolates were susceptible to moxifloxacin.
In vitro susceptibilities of the Corynebacterium isolates received at the PHOL from 2011 to 2016
Among the nondiphtherial corynebacteria in this study, C. pseudodiphtheriticum/C. propinquum isolates showed the highest susceptibility to penicillin (95.8%), whereas C. amycolatum (10.5%), C. afermentans subsp. afermentans (1.1%), C. minutissimum (3.9%), C. coyleae (1.7%), C. urealyticum (3.8%), and C. aurimucosum (5.9%) isolates were the least susceptible to penicillin (Table 1). Only a single isolate of C. striatum (1/931 isolates) and none of the C. jeikeium (n = 76) or C. resistens (n = 19) isolates were susceptible to penicillin. Susceptibility to penicillin and other beta-lactams has reduced in the last 2 decades among certain nondiphtherial corynebacteria, including C. striatum (7–10). Consistent with our findings, a study from Japan showed that none of the C. striatum (n = 22) isolates recovered from blood specimens were susceptible to penicillin (9). However, a recent Canadian study (4) reported the susceptibility rate of penicillin for the Corynebacterium genus to be 77%, which is much higher than the susceptibility rate (14.5%) found in our study. The difference in the susceptibility rate may be partially explained because Bernard et al. (4) used CLSI 2010 interpretative criterion (susceptibility [S], ≤1 mg/liter) versus the revised CLSI 2015 interpretative criterion (S ≤ 0.12 mg/liter) for penicillin that was used in our study. Our study also did not include any C. diphtheriae isolates, which constituted 35% of the 595 isolates tested in the Bernard et al. study. Another explanation could be differences in geographical distribution of various clones of nondiphtherial corynebacteria. For example, our data demonstrate that C. pseudodiphtheriticum/C. propinquum was mostly susceptible to beta-lactams (95%); however, this differs from a report from Brazil in which high rates of penicillin resistance were observed (11). In our study, the MIC50 and MIC90 for ampicillin among C. striatum were 4 mg/liter and >8 mg/liter, respectively. On the other hand, Soriano et al. (8) reported the MIC50 and MIC90 values of ampicillin for C. striatum to be 0.5 mg/liter and 2 mg/liter, respectively, in a study conducted on isolates collected in Spain. Similarly, Martínez-Martínez et al. (12) reported the MIC50 and MIC90 of ampicillin for C. striatum (n = 86) of 1 mg/liter and 2 mg/liter, respectively, on isolates collected in Spain.
Susceptibility to erythromycin varied among Corynebacterium spp., with none of the C. afermentans subsp. afermentans isolates displaying susceptibility (n = 89), while only 1.9% of C. urealyticum and 21.2% of C. pseudodiphtheriticum/C. propinquum isolates were found to be susceptible (Table 1). Similarly, clindamycin susceptibility ranged from 0% to 20.8% among nondiphtherial corynebacteria. Resistance to erythromycin among nondiphtherial corynebacteria has been reported previously (1, 4, 7, 8) and is mainly attributed to the presence of the ermX gene (3). The erythromycin and clindamycin susceptibilities in this study were lower than those recently reported by Bernard et al. (4).
Among non-diphtheriae Corynebacterium species in this study, 26.9% of isolates were susceptible to ciprofloxacin. Ciprofloxacin had the lowest activity against C. striatum (4.6%) and C. urealyticum (3.8%) and the highest activity against C. pseudodiphtheriticum/C. propinquum (Table 1). The activities of moxifloxacin and levofloxacin were similar to that of ciprofloxacin for all nondiphtherial corynebacteria. C. urealyticum has been known to cause acute and chronic urinary tract infections (UTIs) and may lead to bacteremia. Consistent with our findings, a previous study had reported that 96.1% of C. urealyticum isolates were found to be nonsusceptible to ciprofloxacin (13). Fluoroquinolones are extensively used to treat UTIs (14); however, isolates of C. urealyticum in this study displayed resistance to several major drug classes, including ciprofloxacin. Ciprofloxacin resistance has also been reported among C. glucuronolyticum (15). In our study, 73% of the C. glucuronolyticum (n = 15) isolates were susceptible to ciprofloxacin.
The recent study by Bernard et al. (4) reported the rates of susceptibility to ciprofloxacin, tetracycline, and trimethoprim-sulfamethoxazole to be 69.9%, 87.1%, and 73.1%, respectively, for isolates identified as Corynebacterium since the 1990s (4). However, in our study, the susceptibilities among all the Corynebacterium isolates are lower for ciprofloxacin, tetracycline, and trimethoprim, with susceptibility rates of 27.5%, 64%, and 38%, respectively. The differences in the percentage of susceptible isolates to ciprofloxacin seen here are in line with what has been observed in other recent studies by Hahn et al. (2) and Soriano et al. (16) and may due to differences in the number of isolates and, in particular, the years of collection.
Resistance to gentamicin has been reported previously in C. striatum and is associated with the presence of the aac(3)-XI gene encoding AAC(3)-XI, a new aminoglycoside 3-N-acetyltransferase (17). In our study, 92.8% of C. striatum isolates were found to be susceptible to gentamicin. Among all species of Corynebacterium, C. urealyticum (55.8%), C. jeikeium (81.2%), and Corynebacterium CDC group G (85.7%) showed the lowest susceptibility against gentamicin.
As stated earlier, 45 (3.0%) out of 1,959 Corynebacterium isolates tested for daptomycin showed an MIC of ≥2 mg/liter and are considered nonsusceptible isolates. Daptomycin has been increasingly suggested as an important therapeutic alternative for multidrug-resistant Corynebacterium species, including C. striatum and C. jeikeium (18). However, daptomycin-nonsusceptible isolates of C. striatum (n = 16), C. minutissimum (n = 5), C. jeikeium (n = 7), C. coyleae (n = 3), and a single isolate each of C. amycolatum, C. aurimucosum, Corynebacterium CDC group G, C. durum, C. imitans, C. pseudodiphtheriticum, C. singular, C. simulans, C. riegelii, C. resistens, and other Corynebacterium species were identified in this study. Daptomycin resistance has been previously described in C. jeikeium (19), as well as in patients with C. striatum infection receiving daptomycin treatment (20–23). Significantly, this study shows that daptomycin nonsusceptibility is not limited to C. striatum and C. jeikeium but is present in other nondiphtherial corynebacteria as well. Details regarding treatment were not available to us; therefore, it cannot be concluded if these isolates developed resistance while the patients had received treatment with daptomycin.
Tigecycline is another antibiotic that has been suggested in recent reports to be a therapeutic alternative for the treatment of complicated infections caused by these corynebacteria (24, 25). There are no breakpoints for resistance as defined by CLSI or EUCAST for tigecycline. The MIC50 and MIC90 were 0.12 mg/liter and 0.25 mg/liter, respectively, for all the corynebacteria in this study.
Antibiotic susceptibility testing data showed that 24 Corynebacterium species within the genus were completely susceptible to quinupristin-dalfopristin, and more than 90% of C. striatum, C. riegelii, C. accolens, C. jeikeium, C. minutissimum, C. imitans, C. aurimucosum, Corynebacterium CDC group G, and C. amycolatum isolates were susceptible. However, none of the C. glycinophilum (n = 3) isolates were susceptible to quinupristin-dalfopristin, and lower rates of susceptible isolates were found among C. coyleae (43%), C. kroppenstedtii (50%), C. afermentans (50%), C. auris (67%), and C. tuberculostearicum (67%).
C. minutissimum was previously considered to be susceptible to many different classes of antibiotics (1, 7, 9, 26), but most of the isolates in this study were resistant to more than four drug classes, including penicillin, erythromycin, gentamicin, and ciprofloxacin. Similarly, C. striatum was found to be highly multidrug resistant, with varied percentages of resistance to penicillin, erythromycin, gentamicin, ciprofloxacin, and daptomycin. Other frequently reported Corynebacterium species that were resistant to more than four drug classes included C. amycolatum, C. aurimucosum, C. afermentans, C. jeikeium, C. urealyticum, and C. coyleae. C. pseudodiphtheriticum/C. propinquum isolates were mostly susceptible to beta-lactams (95%), vancomycin, linezolid, tetracycline, and gentamicin, though the susceptibility to macrolides and lincosamides was greatly reduced.
Our findings demonstrate that there are differences in antibiotic susceptibility among Corynebacterium species and that there is considerable antibiotic resistance in some species. Laboratories should identify clinically relevant Corynebacterium isolates to the species level, which may provide additional clues regarding antibiotic susceptibility. The increased use of matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS) identification systems in clinical laboratories will greatly enhance front-line laboratory capacity and capability in identifying Corynebacterium organisms to the species level.
One of the limitations of our study is that it is not known whether Corynebacterium isolates recovered from specimens were truly the causative agents of the infection. The clinical relevance of nondiphtherial corynebacteria, including C. striatum, has been well documented since their identification (27), and Corynebacterium species, such as C. striatum, have been linked to multiple outbreaks in hospitals and nosocomial environments (28, 29). Therefore, although not often considered pathogenic, it is now clear that under the right circumstances, nondiphtherial corynebacteria can be the causative agents of infection. Therefore, in the absence of other organisms, the isolation of Corynebacterium spp. from sterile sites and/or from immunocompromised patients could be clinically relevant, and clinicians will need to decide whether to treat the infection based on clinical presentation as well as other findings. As nondiphtherial corynebacteria exhibit varied susceptibilities to many antibiotics, empirical treatment options are limited to vancomycin and linezolid. Multidrug resistance is on the rise among the nondiphtherial corynebacteria, with a decrease in susceptibility rates among species, such as C. striatum and C. minutissimum, which were previously reported to have low or no resistance (10, 26). As susceptibility varies among different nondiphtherial corynebacteria, and many commonly encountered Corynebacterium spp. are multidrug resistant, it is important to perform susceptibility testing particularly for those isolates that are recovered from sterile sites in order to manage patients appropriately. However, identification and susceptibility testing on these isolates can be time-consuming; therefore, regular surveillance studies may be useful in monitoring changes in trends of resistance and providing better empirical treatment options to patients.
MATERIALS AND METHODS
In Ontario, hospital and community laboratories generally identify Corynebacterium to the genus level and refer isolates that are deemed clinically significant for species-level identification and susceptibility testing to the PHOL. All isolates of nondiphtherial corynebacteria recovered from specimens received by the PHOL between 2011 and 2016 were included in the study. Since the data were deidentified prior to analysis, it is not known what the proportion of isolates is from the same patients. From 2011 to 2014, bacterial identification was primarily done by biochemical tests (3, 5), and when a definitive identification could not be made, 16S rRNA gene sequence analysis was performed. From 2015 onward, MALDI-TOF MS (MALDI BioTyper; Bruker) was primarily used for identification combined with select biochemical tests and 16S rRNA sequencing for definitive identification when needed. Prior to the implementation of MALDI-TOF MS for the identification of Corynebacterium spp., biochemical testing and 16S rRNA gene sequencing were unable to reliably differentiate between C. pseudodiphtheriticum and C. propinquum. As a result, these isolates were reported as C. pseudodiphtheriticum/C. propinquum. When a definitive species-level identification could not be made, organisms were reported out as Corynebacterium species (not Corynebacterium diphtheriae).
Antibiotic susceptibility testing was performed using commercial broth microdilution Sensititre GPALL1F plates (Trek Diagnostic Systems, Thermo Fisher Scientific), and the susceptibility results for each antibiotic were interpreted as per the Clinical and Laboratory Standards Institute (CLSI) guidelines (6). MIC testing was performed for the following antibiotics: erythromycin (0.25 to 4 mg/liter), penicillin (0.06 to 8 mg/liter), vancomycin (0.25 to 32 mg/liter), gentamicin (2 to 16 mg/liter), daptomycin (0.5 to 4 mg/liter), ciprofloxacin (1 to 2 mg/liter), moxifloxacin (0.25 to 4 mg/liter), levofloxacin (0.25 to 4 mg/liter), ampicillin (0.125 to 8 mg/liter), oxacillin (0.25 to 4 mg/liter), nitrofurantoin (32 to 64 mg/liter), rifampin (0.5 to 4 mg/liter), tigecycline (0.03 to 0.5 mg/liter), clindamycin (0.5 to 2 mg/liter), tetracycline (2 to 16 mg/liter), chloramphenicol (2 to 16 mg/liter), trimethoprim-sulfamethoxazole (0.5 to 4 mg/liter), quinupristin-dalfopristin (0.5 to 4 mg/liter), and linezolid (1 to 8 mg/liter). The CLSI does not have interpretative criteria for moxifloxacin, and therefore, clinical breakpoints for moxifloxacin were used from the European Committee for Antimicrobial Susceptibility Testing (EUCAST) (30).
The number of isolates susceptible to each antibiotic was determined per year. The Cochran-Armitage test was used to measure trends in antimicrobial resistance over time, with a P value of less than 0.05 considered statistically significant. The Mann-Kendall test at a 5% significance level was used to analyze trends of Corynebacterium species submitted over time. Statistical analyses were done using R version 3.3.2.
ACKNOWLEDGMENTS
We thank the staff of the reference identification and susceptibility sections of the PHOL for performing identification and susceptibility testing as part of routine clinical testing.
FOOTNOTES
- Received 24 August 2017.
- Returned for modification 18 October 2017.
- Accepted 4 January 2018.
- Accepted manuscript posted online 16 January 2018.
Supplemental material for this article may be found at https://doi.org/10.1128/AAC.01776-17.
- © Crown copyright 2018.
The government of Australia, Canada, or the UK (“the Crown”) owns the copyright interests of authors who are government employees. The Crown Copyright is not transferable.