Except for Corynebacterium diphtheriae, all Corynebacterium spp. are ubiquitous skin commensals and are usually considered to be blood culture contaminants.1 However, some reports have highlighted the importance of C. striatum and C. jeikeium as causes of catheter-related bloodstream infections in immunocompromised patients, and as causes of various infections in immunocompetent patients, such as arthritis and endocarditis.2–5 It is difficult to distinguish contamination from bacteremia when Corynebacterium spp. or coagulase-negative staphylococci are isolated from blood samples.6–9 In clinical practice, it is important to distinguish contamination from bacteremia to prevent unnecessary prescription of antimicrobial agents, which can lead to selection of antimicrobial-resistant organisms, longer duration of hospitalization, and increased costs.10

There are few studies on Corynebacterium spp. infection, and the clinical significance of these organisms remains unclear. Thus, in this study, we aimed to evaluate the clinical characteristics of bloodstream infections caused by Corynebacterium spp. by analyzing clinical cases of blood cultures positive for Corynebacterium spp.

During the study period, 66 patients had positive blood cultures for Corynebacterium spp. Three patients were excluded because of a second episode of serial infection. Of the remaining 63 patients, 28 (44%) met the criteria for bacteremia and 35 (56%) met the criteria for contamination. Patients’ baseline characteristics are summarized in Table 1. The overall median age was 73 years, and the contamination group was significantly older than the bacteremia group (p=0.023). Men constituted 74.6% of cases. Fourteen cases were identified in the emergency department and critical care medicine; this department had a significantly higher proportion of contamination cases than did the other departments (p=0.049). Overall, 39 patients had a central venous catheter in place, five had a peripheral venous catheter in place, and 19 did not have a venous catheter; bacteremia was diagnosed more frequently in patients who had peripheral or central venous inserted catheters (p=0.002). The median interval from admission to blood culture sampling was 25 days (IQR, 6–49); this interval was significantly longer in the bacteremia group than in the contamination group (p=0.006). The median time to culture positivity was significantly shorter in the bacteremia group than in the contamination group (p=0.009). There were no significant differences in median leukocyte count or C-reactive protein level between the two groups. In 51 of the 63 cases, the patient had fever (temperature >38°C) at the time the blood culture sample was taken.

Final identification of the Corynebacterium spp. isolated were C. striatum (n=38), C. jeikeium (n=6), and C. argentoratense (n=2). The species was reported only as “Corynebacterium sp.” in 19 cases, either because the physician in charge did not request further identification or the laboratory was unable to identify the species. Two species (C. striatum and Corynebacterium sp.) were identified simultaneously in two cases (Table 2). There was a significant difference between the different species regarding rates of bacteremia versus contamination (p=0.001). Two sets of blood cultures were obtained from 60 patients, and one set of blood cultures was obtained from three patients. Multiple blood culture specimens were positive in 38% (23/60) of these patients. The same organism isolated from blood samples was isolated from venous catheter tip specimens (n=11), followed by sputum, wound site, urine, and drain specimens. In 45 patients, only the blood culture specimen was positive for Corynebacterium spp. A central or peripheral venous catheter was considered the source of infection in 20 of the 28 patients with bacteremia; the catheter tip was culture positive for Corynebacterium spp. in 12 cases and culture negative in three cases; the catheter was not removed and sent for culture in the remaining five cases.

The results of the antimicrobial susceptibility tests are shown in Table 3. All isolates were susceptible to minocycline, vancomycin, and teicoplanin. Most isolates were resistant to penicillin, imipenem/cilastatin, erythromycin, clindamycin, and levofloxacin. Table 4 shows the antimicrobial agents administered after the final report was issued by the microbiology laboratory. Appropriate antimicrobials were administered in all except one case in which the patient died before the final report was issued. Of the 27 patients who received appropriate antimicrobial therapy, 24 were cured while three died as a result of Corynebacterium bacteremia.

In this retrospective study, we analyzed the clinical data of patients who had Corynebacterium spp. identified on blood culture. More than 40% of patients were determined to have bacteremia based on their clinical characteristics and course. C. striatum was the species most frequently identified in patients with bacteremia. The route of infection was via venous catheters in >50% of patients with bacteremia; this infection route has been reported in several previous studies.2,7,12 Regarding intravascular catheter-related infections, there is broad consensus that growth of the same organism in paired blood culture samples, drawn from a peripheral vein and the suspected source, indicates a true or relevant infection.13Corynebacterium spp. can colonize prostheses, catheter tips, and ventilator and feeding tubes; however, Corynebacterium spp. isolated from these prostheses are usually considered contaminants. In general, bacterial skin commensals have relatively low virulence, but De Souza et al.14 recently reported that biofilm production by antimicrobial-resistant C. striatum is a new virulence factor and is related to nosocomial outbreaks. In addition, this organism has been reported to cause pneumonia and urinary tract infections.15,16

As mentioned above, for the purpose of this study, bacteremia was defined as two blood culture sets testing positive or one blood culture set testing positive plus a positive culture result for a specimen from another site. There is, however, evidence that bacterial skin commensals isolated from even one blood culture specimen might reflect clinically relevant infection.17,18 Several studies have reported that the time to positivity can be used to discriminate between contamination and bacteremia.19–21 Zhang et al.21 reported that the time to positivity was <36h in 98% of cases of bacteremia caused by Gram-positive bacteria. In our study, although the exact time to positivity (in hours) was not recorded, the time to positivity (in days) was significantly longer in the contamination group and was within three days in all cases of bacteremia. In the present study, a time to positivity of four days or more suggested contamination.

Past studies have generally reported on the susceptibility of C. striatum to the β-lactam group of antimicrobials.1,22,23 However, with increasing use of broad-spectrum antibiotics, recent studies have shown the emergence of multidrug resistant strains.24,25 In our study, most isolated strains were multidrug resistant, and all tested strains were susceptible only to glycopeptides, such as vancomycin and teicoplanin. Based on our results (and in accordance with those of previous studies), when infection with Corynebacterium spp. is suspected, initial therapy should include vancomycin because in vitro resistance to vancomycin has not been reported.2,25

All but one patient in this study received prompt treatment with an antimicrobial agent to which the isolated Corynebacterium spp. was susceptible. Nonetheless, three patients died; their comorbidities were malignancy (n=2) and immunosuppression (n=1). All three patients had central venous catheters in situ; these catheters were not removed because of the severe clinical condition of the patients. Previous studies have reported that while Corynebacterium spp. infection can be associated with death, most cases are relatively easy to manage with antimicrobial therapy and catheter removal, if applicable.26–28 Kimura et al.29 also reported that catheter removal within a week tended to be associated with better outcomes in patients with a central venous catheter.

This study has some limitations. First, it was a retrospective study. Second, we could not clearly distinguish between the true infection group and contamination group because of the absence of established clinical and bacteriological markers of infection with Corynebacterium spp.18,30 Therefore, some patients with contamination may have been included in the bacteremia group, and vice versa. However, all patients in the bacteremia group were considered to have bacteremia and all received treatment for bloodstream infection with Corynebacterium spp., based on their clinical course, culture results, and the recommendation of infectious disease specialists. Third, species identification was carried out in only a small number of patients because it is not generally performed in Japan, especially if contamination is strongly suspected. Thus, we could not evaluate the impact of different species on clinical outcomes. Lastly, data on the total duration of therapy were not collected; thus, an important dimension in the treatment and course of bacteremia was neglected.

In conclusion, our analysis showed that Corynebacterium spp. isolated from blood culture specimens can cause bloodstream infections in a considerable proportion of patients. Younger age, shorter time to culture positivity, and the presence of an indwelling catheter are associated with bacteremia. Because multidrug-resistant Corynebacterium spp. have become common, administration of appropriate antimicrobial agents, such as vancomycin, should be initiated pending the results of antimicrobial susceptibility testing in patients with suspected bacteremia caused by Corynebacterium spp.

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

The authors declare no conflicts of interest.