Catheter-related bloodstream infection caused by Gordonia terrae in a bone-marrow transplant patient : case report and review of the literature

Members of the genus Gordonia are aerobic, mycolic-acid containing actinomycetes that are widely distributed in nature. Most Gordonia spp. have been isolated from environmental sources; however, a few have been reported as a rare cause of human infections. Gordoniae are difficult to identify with routine laboratory tests, which often leads to the misidentification as other Gram-positive bacilli such as Rhodococcus or Nocardia. Among Gordonia spp. of clinical significance, G. terrae, G. sputi, G. bronchialis, G. rubripertincta, G. polysoprenivorans, G. otitidis, G. araii and G. aichiensis have been found to cause infections such as cutaneous infections (Richet et al., 1991), mediastinitis (Kuwabara et al., 1999), granulomatous mastitis (Zardawi et al., 2004), abscesses (Werno et al., 2005, Blanc et al., 2007), osteomyelitis (Siddiqui et al., 2012) and bacteraemia (Buchman et al., 1992).

Catheter-related bacteraemia is the most commonly reported systemic infection caused by Gordonia spp.However, in most reported cases, the diagnosis was based only on clinical findings.Thus, it is difficult to be certain that the patients had true bloodstream infections (BSIs) and that the catheter was the likely source.Definitive diagnosis by laboratory methods is important to establish optimal management of catheter-related BSIs (CR-BSIs) caused by unusual pathogens.We present here a case of confirmed CR-BSI by G. terrae in a bone-marrow transplant patient.We have also reviewed the cases published in the literature.

Case report
A 19-year-old man with acute lymphoblastic leukaemia underwent an allogenic bone-marrow transplant from an unrelated donor 3 months prior to admission.He was admitted to the hospital because of acute renal failure secondary to cyclosporine prophylaxis to prevent graftversus-host disease (GVHD).Cyclosporine was discontinued, and 16 days after hospitalization he developed acute GVHD.A long-term central venous catheter (Hickman) was inserted, and treatment with methylprednisolone and pentostatin was started.Surveillance blood cultures drawn through the catheter were obtained weekly and incubated onto BACTEC 9240 instruments (Becton Dickinson).Four months after admission, the Hickman catheter was still in place and a surveillance blood-culture sample was positive (anaerobic bottle) for Gram-positive cocci identified as Enterococcus faecium and, after 4 days of incubation, the aerobic bottle was positive for Grampositive coryneform rods.Subcultures on 5 % sheep blood agar and chocolate plates yielded smooth, orange, catalasepositive and oxidase-negative colonies after 48 h of Abbreviations: BSI, bloodstream infection; CR-BSI, catheter-related bloodstream infection; GVHD, graft-versus-host disease incubation.At that time, the patient was afebrile and had no clinical signs or symptoms of infection.In addition, there were no signs of infection at the insertion site of the catheter.To determine the clinical significance of the isolate, two sets of blood cultures collected from a percutaneous venipuncture (one set) and from the central catheter (one set) were taken and treatment with teicoplanin was started.Both peripheral and catheter blood cultures were positive for Gram-positive coryneform rods.The differential time to positivity between catheter and peripheral blood cultures was .120min.
The catheter was not removed because the patient remained afebrile and had no clinical signs of sepsis.Repeated blood cultures after 1 week of treatment were still positive.Finally, blood cultures obtained after 14 days of intravenous therapy with teicoplanin were negative.Despite this, the patient died due to steroid-refractory GVHD 1 month after the first isolation of the Gordonia sp.

Discussion
Gordonia spp.have been identified as an unusual cause of different infections, mainly catheter-associated bacteraemia in immunocompromised patients.It is believed that their role as a cause of human infections could be underestimated, as their identification by conventional microbiological methods is difficult.Moreover, the microorganism is often misidentified as other actinomycetes, as initially occurred in this case, and requires the use of molecular techniques to secure an accurate identification to species level (Sng et al., 2004;Shen et al., 2006).
Table 1 shows the relevant clinical information for the 18 CR-BSI adult cases and our case.Most cases were due to G. terrae (n59) and G. sputi (n57).Almost all isolates (15 isolates) were initially identified as Rhodococcus spp.by conventional biochemical tests.In one case, the isolate was identified as Corynebacterium pseudodiphtheriticum, and two case reports did not include biochemical identification.
Most of these BSIs occurred in cancer patients and were associated with long-term central venous catheter use.Gordonia spp., like many other actinomycetes, are able to colonize the human skin.In addition, they have the ability to produce extracellular polysaccharides and thereby to form biofilms (Arensko ¨tter et al., 2004).These are important features that strongly contribute to the development of catheter-related infections.
All the reported cases of CR-BSI caused by Gordonia spp.were considered true episodes of bacteraemia.However, catheter-related bacteraemia was confirmed in only two cases (cases 2 and 18).In case 13, there was no information concerning the laboratory diagnosis.In seven cases (patients 1, 4, 5, 6, 10, 11 and 12), the organism was isolated from a single blood culture obtained through the catheter.Therefore, it is questionable whether the Gordonia isolates were true pathogens or whether they represented contaminants or colonizers.For the eight remaining cases (patients 3, 7, 8, 9, 14, 15, 16 and 17), two sets of blood cultures (one obtained from the catheter and the other from a peripheral vein) were positive and six of these episodes were persistent bacteraemia or complicated infections.Therefore, it must be assumed that all these episodes represented true infections.Although the source of bacteraemia was not confirmed in any of these cases, the catheter was the most probable source of infection.The definitive diagnosis of CR-BSI was obtained by culture of the catheter tip (case 2), and paired quantitative blood cultures (case 18).In the present case, CR-BSI was confirmed by using the differential time to positivity of blood cultures drawn simultaneously from both the catheter and a peripheral vein.
Definitive diagnosis of CR-BSIs caused by unusual pathogens is crucial for guiding decisions regarding the management of these infections, particularly those regarding removal of the catheter.Overall, 10 of the 18 previous reported cases were treated with catheter removal, whilst in nine patients (including the present case), the catheter was left in situ.
The catheter was retained in most cases (6/9) in which the Gordonia sp.might represent a colonizer or a contaminant.By contrast, removal of catheter was the strategy of treatment in the majority of cases (8/10) corresponding to those considered true BSI episodes.In addition, in five of these cases, catheter removal was necessary because of persistent bacteraemia (four cases) or complicated infection (infective endocarditis).Despite this strategy of treatment, two patients died.
Clinical management of G. terrae CR-BSIs is far from clear.Analysis of the published experiences suggests that prompt catheter removal is not always necessary for haemodynamically patients.In addition, blood-culture contamination or colonization must be ruled out before the decision to remove the catheter is made.
The data reviewed here emphasize that, for catheter management, it is essential to ensure that a positive blood culture reflects a true BSI and that the catheter is the source.In view of the above, it could be suggested that removal of the catheter should be recommended for the treatment of these infections.Indeed, Blaschke et al. (2007) recommend catheter removal for Gordonia infections in children.However, in two previously reported cases of true BSIs (patients 3 and 17) and in the present case, the catheter was retained and the infection cleared despite the presence of persistent bacteraemia and complicated infection.One of these patients died, but the death was not related to the infection.This observation suggests that, although Gordonia spp.may be difficult to eradicate, in some cases, proven BSIs by these organisms may be cured while leaving the catheter in place.
Antibiotic therapy varied widely among the reports.The antimicrobial susceptibility data of Gordonia isolates that have been published previously (Blaschke et al., 2007;Aoyama et al., 2009;Moser et al., 2012) show that the most active in vitro antimicrobials were carbapenems (99 %, n566 isolates tested), fluoroquinolones (92 %, n565 isolates tested) and aminglycosides (100 %, n553 isolates tested against amikacin; 98 %, n547 isolates tested against gentamicin).Although vancomycin is used as an initial therapy for infections caused by Gram-positive rods, it is of note that only 81 % (51/63) of isolates were susceptible.Therefore, therapy with this agent should be based on the isolate's in vitro antimicrobial susceptibility test results.
The overall mortality was 21 % (4 of 19 patients died).In the present case, the death was not related to the CR-BSI.
In general, Gordonia spp.are thought to be of low virulence; in our patient, as in other cases, the infection was unapparent.It has to be noted, however, that persistent bacteraemia was common (six cases) and two patients developed infective endocarditis (mitral and aortic native valve).Thus, we believe that CR-BSIs caused by Gordonia spp.must be taken seriously and should prompt initiation of antibiotic therapy.In addition, the patients should be monitored to detect complicated infections, mainly persistent bacteraemia despite 3 days of appropriate therapy and infective endocarditis.
In conclusion, Gordonia spp.are a very rare cause of CR-BSIs, mostly in cancer patients with long-term indwelling catheters.Although the present case report describes a case of confirmed infection in which persistent bacteraemia cleared without catheter removal, further experience is required to confirm that these infections can, in some cases, be successfully treated whilst retaining the catheter.