Central Line Sepsis in a Child Due to a Previously Unidentified Mycobacterium

Abstract

A rapidly growing mycobacterium similar to strains in the present Mycobacterium fortuitum complex (M. fortuitum, M. peregrinum, and M. fortuitum third biovariant complex [sorbitol positive and sorbitol negative]) was isolated from a surgically placed central venous catheter tip and three cultures of blood from a 2-year-old child diagnosed with metastatic hepatoblastoma. The organism’s unique phenotypic profile and ribotype patterns differed from those of the type and reference strains of the M. fortuitum complex and indicate that this organism may represent a new pathogenic taxon.

Rapidly growing mycobacteria have been the cause of several forms of clinical disease. The most common types of infection involve skin and soft tissue; however, skeletal, pulmonary, and disseminated infections have often been described (11). Mycobacterium fortuitum, M. peregrinum, and M. fortuitum third biovariant complex are closely related, rapidly growing mycobacteria that are commonly grouped together as the M. fortuitum complex. Members of the M. fortuitum complex are ubiquitous in the environment and are increasingly associated with human disease (2, 13). During the last decade, coinciding with the increased use of indwelling catheters, strains of the M. fortuitum complex have been recognized more frequently as a cause of central venous catheter-related bacteremia in severely immunocompromised patients, in particular, patients with underlying malignancies (9).

The taxonomy of the M. fortuitum complex has recently been reviewed and revised (5, 6, 12). On the basis of DNA-DNA hybridization studies, Kusunoki and Ezaki (6) established M. fortuitum and M. peregrinum as separate species. The microorganisms of the unnamed third biovariant complex of M. fortuitum were not included in their study (6); however, recent 16S rRNA gene sequencing analysis suggested that M. fortuitum and the M. fortuitum third biovariant complex were separate species (5).

We describe a patient with catheter-related bacteremia in whom a previously unidentified mycobacterium was the infecting pathogen.

Case report.

A 2-year-old male was diagnosed with metastatic hepatoblastoma in June 1988, at which time his alpha-fetoprotein level was massively elevated. He began receiving chemotherapy, which was complicated by the development of Klebsiella pneumoniae bacteremia. This infection responded to therapy with ceftazidime and removal of the central catheter. He had a good initial response to chemotherapy, and on 8 August 1988, he underwent a right hemihepatectomy with intent to cure. In December 1989, he experienced pulmonary relapse of his malignancy, and chemotherapy with carboplatin, 5-fluorouracil, and cyclophosphamide was reinstituted. His last chemotherapy was given in July 1990; at follow-up his serum alpha-fetoprotein level was 60 ng ml⁻¹. On 1 August 1990, computed tomographic scans of his chest and abdomen showed no evidence of residual hepatic, pulmonary, or metastatic disease.

On 21 September 1990, following flushing of his central line he complained of feeling unwell; he was febrile, and on examination he had evidence of mild pharyngitis. Blood collected through the patient’s central venous catheter for culture was positive for the growth of a gram-positive coccobacillus (blood isolate W4962). He was seen again on 5 October 1990 for follow-up of his positive blood culture. At that time he was well and afebrile. The central line skin site was not inflamed, and results of physical examination were unremarkable. Blood samples for culture were collected from the central venous catheter, and penicillin and gentamicin therapy was started; subsequently, his repeat cultures were sterile. He remained well in the hospital, and on 9 October 1990, his antibiotic therapy was discontinued and he was discharged to home.

On a follow-up visit at the clinic on 31 October 1990, he complained of myalgias in his legs and it was noted that he had an intermittent cough. Nocturnal fever (to 40°C) had also been noted. Results of physical examination were unremarkable. A repeat blood culture collected via the central line grew the same microorganism (blood isolate W4963) that had grown before. Subsequently, a culture of another blood sample (which grew isolate W5064) obtained on 8 November 1990 and a culture of the central line tip (which grew isolate W4964), obtained after its removal on 9 November 1990, also grew this microorganism; a course of oral trimethoprim-sulfamethoxazole (TMP-SMX) was started, and he has remained well since.

Mycobacterial strains.

The sources of the four patient isolates are given in the case report. Type strain M. fortuitum ATCC 6841, Type strain M. peregrinum ATCC 14467, and M. fortuitum third biovariant complex reference strains (sorbitol-positive strain ATCC 49403 and sorbitol-negative strain ATCC 49935) were generously provided by Richard J. Wallace, Jr., Department of Microbiology, The University of Texas Health Sciences Center at Tyler.

Phenotypic analysis.

The four patient isolates were inoculated onto heart infusion agar with rabbit blood (BBL, Becton Dickinson Microbiology Systems, Cockeysville, Md.) and were incubated at 35°C for 2 days for morphologic studies. Both Gram and modified Kinyoun’s acid-fast stains were used to study their microscopic morphologies and acid fastness, respectively (1). These isolates were examined at low power (magnification, ×10) under a stereomicroscope for the presence of aerial mycelium. Biochemical tests routinely used in the Centers for Disease Control and Prevention’s Actinomycete and Mycobacteriology Laboratories were performed (1, 10). In addition, we tested for hydrolysis of acetamide (BBL), utilization of citrate, and growth at 42°C.

Antimicrobial susceptibility tests were performed for all patient isolates and the type and reference strains of the M. fortuitum complex by a disk diffusion method for cefamandole (30 μg), tobramycin (10 μg), streptomycin (10 μg), gentamicin (10 μg), neomycin (30 μg), and kanamycin (30 μg), as described previously (4). The zone sizes for resistance used were those of Grange and Stanford (4). In addition, antimicrobial susceptibilities were determined by a previously described broth microdilution method with cation-supplemented Mueller-Hinton broth (11). For the latter method, the drugs tested were amikacin, amoxicillin-clavulanate, ampicillin, cefotaxime, ceftriaxone, ciprofloxacin, doxycycline, erythromycin, imipenem, minocycline, sulfamethoxazole, TMP-SMX, and vancomycin. The plates were incubated for 72 h at 35°C. Since the methods of testing and the breakpoints for resistance for rapidly growing mycobacteria have not been standardized or approved by the National Committee for Clinical Laboratory Standards (NCCLS), the breakpoints for resistance used were those of NCCLS for organisms that grow aerobically (8).

HPLC.

Inocula grown on Lowenstein-Jensen medium (BBL) for 2 days at 35°C were saponified and derivatized. Mycolic acids were assayed by high-performance liquid chromatography (HPLC) as described previously (3).

Genetic analysis.

All isolates were subcultured from Lowenstein-Jensen slants into 1.5 liters of Mueller-Hinton broth containing 22.5 g of glycine (Sigma Chemical Company, St. Louis, Mo.) and 7.5 ml of Tween 80 (Sigma) and were grown for 1 to 2 days at 33°C before harvesting by centrifugation. Genomic DNA was purified from lysed protoplasts as described previously (7).

One microgram of genomic DNA was digested with 10 U of PvuII, PstI, SmaI, and SalI (Boehringer Mannheim, Indianapolis, Ind.) for 8 h at 35°C in the buffer recommended by the manufacturer. The DNA fragments were separated by electrophoresis in a 0.85% (wt/vol) agarose gel (GIBCO BRL, Gaithersburg, Md.). The DNA fragments were then transferred to a nylon membrane (Nytran; Schleicher & Schuell, Keene, N.H.) and were hybridized with a digoxigenin-labeled plasmid pKK3535 DNA probe at 68°C as described previously (7). Plasmid pkk3535 contains the entire Escherichia coli 16S and 23S rRNA and is capable of hybridizing to a broad range of heterologous bacteria. Ribosomal DNA-containing fragments were visualized according to the Genius kit (Boehringer-Mannheim) protocol.

Results.

Each of the four patient isolates produced the same phenotypic profile. All were gram-positive pleomorphic coccobacilli, with longer filamentous forms frequently observed. These isolates were acid fast by the modified Kinyoun method. The stereomicroscopic (magnification, ×10) appearance of the cultures showed them to have slightly beige, cerebriform, and irregularly edged colonies that did not demonstrate aerial hyphae. All isolates grew in less than 7 days, grew at 28°C in 5% (wt/vol) NaCl and on MacConkey agar without crystal violet, were positive for iron uptake and nitrate and arylsulfatase reduction, and were negative for growth in lysozyme, utilization of citrate, and hydrolysis of acetamide. The patient isolates produced acid from d-glucose, d-mannitol, salicin, glycerol, d-xylose, i-myo-inositol, d-trehalose, d-mannose, and d-fructose. They did not produce acid from lactose, sucrose, maltose, starch, l-arabinose, l-rhamnose, galactose, d-sorbitol, dulcitol, raffinose, adonitol, i-erythritol, cellobiose, or melibiose. All patient isolates were resistant to ampicillin, cefotaxime, ceftriaxone, cefamandole, and streptomycin. They were susceptible to the 14 remaining antimicrobial agents tested. The distinguishing phenotypic characteristics between the patient isolates and the type and reference strains of the M. fortuitum complex are given in Table 1.

TABLE 1.

Phenotypic differences between patient isolates and type and reference strains of M. fortuitum complex^a

Characteristic	M. fortuitum ATCC 6841T	M. peregrinum ATCC 14467T	M. fortuitum third biovariant complex		Patient isolates (n = 4)
			d-sorbitol-positive strain ATCC 49403	d-sorbitol-negative strain ATCC 49935
Growth at 42°C	+	−	+	−	−
Hydrolysis of acetamide	+	−	+	+	−
Acid production from:
d-Mannitol	−	+	+	+	+
i-myo-Inositol	−	−	+	+	+
d-Sorbitol	−	−	+	−	−
Resistance tob:
Erythromycin (MIC, ≥8 μg/ml)	+	+	+	+	−
Minocycline (MIC, ≥16 μg/ml)	−	+	+	+	−
Doxycycline (MIC, ≥16 μg/ml)	−	+	+	+	−
Vancomycin (MIC, ≥32 μg/ml)	+	+	+	+	−
Kanamycin (zone size, <15 mm)	−	−	+	−	−
Tobramycin (zone size, <20 mm)	+	+	+	+	−
Neomycin (zone size, <20 mm)	−	−	+	−	−

^aSymbols: +, positive; −, negative. 

^bThe MIC breakpoints for resistance that were used are those of NCCLS (8). Disk contents were as follows: kanamycin, 30 μg; tobramycin, 10 μg; and neomycin, 30 μg. Zone sizes for resistance are those of Grange and Stanford (4). 

With the present HPLC technique, the mycolic acid profiles of the patient isolates and the type and reference strains of M. fortuitum, M. peregrinum, and M. fortuitum third biovariant complex were too similar to be separated.

Ribotype analysis was performed to determine the epidemiologic and genetic relatedness between the isolate from the central line tip, the isolates from the patient’s blood, and the type and reference strains of the M. fortuitum complex. SalI-digested genomic DNA hybridized to the digoxigenin-labeled rDNA probe, as shown in Fig. 1, gave four identical hybridization bands among the isolates from the catheter line tip and the patient’s blood; this pattern of bands differed from those for the type and reference strains of the M. fortuitum complex. Similar results were obtained for PvuII, PstI, and SmaI (results not shown).

FIG. 1.

Ribotype patterns from SalI-digested genomic DNAs of patient isolates and the type and reference strains of the M. fortuitum complex. Lane 1, bacteriophage λ DNA molecular size marker digested with EcoRI and HindIII; lane 2, type strain M. fortuitum ATCC 6841; lane 3, type strain M. peregrinum ATCC 14467; lane 4, M. fortuitum third biovariant complex (sorbitol-positive) reference strain ATCC 49403; lane 5, M. fortuitum third biovariant complex (sorbitol-negative) reference strain ATCC 49935; lanes 6 to 9, patient isolates W4962, W4963, W4964, and W5064, respectively.

Discussion.

M. fortuitum third biovariant complex isolates usually cause sporadic cases of community-acquired cutaneous disease, and the majority of infected patients have had a history of some preceding injury at the site of infection. These infections have been associated with diverse clinical syndromes, including pulmonary infections, skin and soft-tissue infections, meningitis, osteomyelitis, keratitis, and disseminated disease (11).

Successful therapy for M. fortuitum-infected patients has involved surgical debridement, removal of foreign bodies, and appropriate antimicrobial therapy (11). The patient described here was initially treated with penicillin and gentamicin, but only when his central line was removed and his therapy was changed to oral TMP-SMX did he achieve a lasting clinical response. Results of follow-up antimicrobial susceptibility tests showed that the patient isolates were resistant to penicillin but susceptible to gentamicin and TMP-SMX. Since clinical isolates of M. fortuitum-related species have variable antimicrobial drug susceptibilities, the data suggest that antimicrobial susceptibilities should be determined for all clinical isolates that are related to M. fortuitum to assist the clinician when choosing an appropriate drug regimen for the treatment of infected patients. For example, when compared to the type and reference strains of the M. fortuitum complex, this patient’s isolates are uniquely susceptible to erythromycin, vancomycin, and tobramycin. Had the patient not tolerated TMP-SMX, the drug of choice for the treatment of M. fortuitum complex infections (11), any of these antimicrobial agents might have been used as an effective alternative monotherapy or combination therapy.

The patient isolates were initially identified as M. fortuitum third biovariant complex. Upon examination of these isolates for the utilization of d-sorbitol, as recommended by Wallace et al. (11), to further subdivide the third biovariant complex into two groups, they were grouped as M. fortuitum third biovariant complex (d-sorbitol negative). To further characterize these isolates, additional phenotypic characteristics not routinely used in laboratories were compared, namely, growth at 42°C, hydrolysis of acetamide, and resistance to cefamandole (30 μg), tobramycin (10 μg), streptomycin (10 μg), gentamicin (10 μg), neomycin (30 μg), and kanamycin (30 μg) by the disk diffusion method (4). Compared with the four type and reference strains of the M. fortuitum complex, the patient isolates most closely resembled M. peregrinum according to the results of these additional tests; the only differences were acid production from i-myo-inositol and susceptibility to tobramycin for the patient isolates.

The lack of detection of specific mycolic acid patterns by HPLC for the four patient isolates was not unexpected. Previously, Butler and Kilburn (3) reported that peak height ratios of mycolic acids were useful for the subidentification of M. chelonae strains but not for the subidentification of M. fortuitum strains in an extensive study of 100 clinical isolates and type strains of M. fortuitum. However, as the taxonomy of the M. fortuitum complex becomes more defined and HPLC technology develops, mycolic acid profiles may become useful in differentiating between organisms of the complex. Until such time, it is clinically important to keep in mind that the current members of the complex as well as this potentially new species have various drug susceptibilities, and further studies should be performed to properly identify and treat the infecting pathogen.

Species-specific differences in ribotype patterns are used to successfully classify new species previously misidentified due to the similarity of their phenotypic characteristics with those of other formerly recognized microorganisms (7). When the results of ribotype analysis with four different restriction endonucleases were examined, it was found that the ribotype patterns of the four patient isolates were identical, suggesting that the strain colonizing the catheter tip was most likely responsible for the patient’s bacteremia, and that they clearly differed from the patterns observed for the type and reference strains of the M. fortuitum complex. When Kirschner et al. (5) compared the partial 16S rRNA nucleotide sequences of each of the biovars of M. fortuitum, they suggested that each biovar was represented as a genetically distinct taxon. The ribotype analysis presented in this report is in agreement with the findings of Kirschner et al. (5). The ribotype pattern and phenotypic differences noted between the catheter tip isolate and the patient’s blood isolates and the members of the M. fortuitum complex cannot exclude the existence of another d-sorbitol-negative biovariant of the M. fortuitum third biovariant complex. The data suggest that the infecting pathogen causing the catheter-related bacteremia in this child with metastatic hepatoblastoma may represent a novel species of the genus Mycobacterium.