Coryneform bacteria in infectious diseases: clinical and laboratory aspects

Antimicrobial susceptibility pattern of Corynebacterium striatum

The in vitro activities of 16 antimicrobial agents against 86 strains of Corynebacterium striatum were evaluated by microdilution using cation-adjusted Mueller-Hinton broth. MICs at which 90% of strains were inhibited were 0.06 microgram/ml for teicoplanin, 1 microgram/ml for vancomycin, 0.03 to 8 micrograms/ml for beta-lactams, 8 micrograms/ml for sparfloxacin, 16 micrograms/ml for ciprofloxacin, 16/304 micrograms/ml for co-trimoxazole (trimethoprim-sulfamethoxazole), 64 micrograms/ml for tetracycline, 128 micrograms/ml for gentamicin, and > 128 micrograms/ml for amikacin, erythromycin, and rifampin.

Patient-to-patient spread of a single strain of Corynebacterium striatum causing infections in a surgical intensive care unit

Over a 12-month period, Corynebacterium striatum strains were isolated from clinical specimens from 14 patients admitted to a surgical intensive care unit. These isolates were identical by morphology and biotype and displayed the same antibiogram. Ten isolates were found to be the sole possible pathogen. These 10 isolates were from six patients, three of whom had signs of infection at the time of positive culture. Further typing was performed by random amplification of polymorphic DNA analysis, by which all strains were identical and were found to differ to various degrees from reference strains and from isolates found in clinical samples from other wards. In a case-control study the only independent risk factor for acquiring the strain was intubation for longer than 24 h (odds ratio, 20.09; 95% confidence interval, 2.29 to 176.09). The same strain was isolated from surfaces and from air sampled in the direct vicinity of infected patients but never from surfaces or air in other places of the ward. The strain was not isolated from the ventilators. The strain was cultured from the hands of personnel attending to infected patients, but no long-term carriers were found among members of the hospital personnel, suggesting transient carriage only. We conclude that C. striatum can cause serious nosocomial infections in surgical intensive care unit patients and may spread from patient to patient via the hands of attending personnel.

Species identities and antimicrobial susceptibilities of corynebacteria isolated from various clinical sources

Over a 14-month period, 415 clinical isolates of coryneform gram-positive rods were recovered from various sources and identified to the species level according to recent identification schemes. Corynebacterium urealyticum, Corynebacterium striatum, Corynebacterium amycolatum, and Corynebacterium jeikeium predominated, accounting for 63% of all isolates. Corynebacterium accolens, Corynebacterium striatum, Corynebacterium argentoratense, Corynebacterium propinquum and Corynebacterium pseudodiphtheriticum were mostly recovered from the respiratory tract, whereas Corynebacterium afermentans, CDC group G, and Corynebacterium jeikeium were mainly isolated from blood. None of the isolates was identified as Corynebacterium diphtheriae or Corynebacterium xerosis. Ampicillin resistance was detected in Corynebacterium jeikeium (96%) and Corynebacterium urealyticum (99%) and varied among Corynebacterium amycolatum (56%) and CDC group G (26%). These data emphasize the need for an accurate identification of coryneform organisms at the species level and for antimicrobial susceptibility testing of these organisms.

Comparison of antimicrobial susceptibilities of Corynebacterium species by broth microdilution and disk diffusion methods

Corynebacterium species are increasingly being implicated in foreign-body infections and in immunocompromised-host infections. However, there are no specific recommendations on the method or the criteria to use in order to determine the in vitro activities of the antibiotics commonly used to treat Corynebacterium infections. The first aim of our study was to compare the susceptibilities of various species of Corynebacterium to vancomycin, erythromycin, and penicillin by using a broth microdilution method and a disk diffusion method. Second, the activity of penicillin against our isolates was assessed by using the interpretative criteria recommended by the National Committee for Clinical Laboratory Standards for the determination of the susceptibility of streptococci and Listeria monocytogenes to penicillin. Overall, 100% of the isolates were susceptible to vancomycin, while considerable variations in the activities of erythromycin and penicillin were noted for the different species tested, including the non-Corynebacterium jeikeium species. A good correlation in the susceptibilities of vancomycin and erythromycin between the disk diffusion and the microdilution methods was observed. However, a 5% rate of major or very major errors was detected with the Listeria criteria, while a high rate of minor errors (18%) was noted when the streptococcus criteria were used. Our findings indicate considerable variations in the activities of erythromycin and penicillin against the various species of Corynebacterium. Because of the absence of definite recommendations, important discrepancies were observed between the methods and the interpretations of the penicillin activity.

"Corynebacterium aquaticum" wound infection after high-pressure water injection into the foot

The organisms presently named "Corynebacterium aquaticum" have their natural habitat in water and are increasingly often isolated in clinical specimens, but are very seldom the proven cause of infection. A case of a 24-year-old man with a "C aquaticum" wound infection secondary to a high-pressure water injection injury in the foot is described. Cefadroxil and cefuroxime were used for treatment.

Recurrent bacteremia due to Brevibacterium casei in an immunocompromised patient

A case of an immunocompromised patient who experienced two episodes of septicemia caused by a coryneform bacterium is reported. Biochemical characteristics and analysis of cellular fatty acids and of cell wall components showed two identical strains of Brevibacterium casei to be responsible for these infections. The lack of easy-to-perform methods for identification may have led, in the past, to an underestimation of the role of this bacterium, especially in immunocompromised patients.

Evaluation of the applicability of amplified rDNA-restriction analysis (ARDRA) to identification of species of the genus Corynebacterium

The 16S rRNA genes (rDNA) of 50 strains belonging to 26 different coryneform bacterial species and genomospecies and of the type strain of Rhodococcus equi were enzymatically amplified. Amplified rDNA restriction analysis (ARDRA) with the enzymes AluI, CfoI and RsaI was carried out. The combination of the ARDRA patterns obtained after restriction with these three different enzymes enabled the differentiation between the following species: Corynebacterium accolens (number of strains = 2), C. afermentans subsp. afermentans (2), C. afermentans subsp. lipophilum (2), C. amycolatum (3), CDC coryneform group ANF-1-like (1), CDC coryneform group ANF-3-like (1), C. cystitidis (1), C. diphtheriae (4), C. jeikeium (3), C. macginleyi (2), C. minutissimum (1), C. pilosum (1), C. pseudotuberculosis (2), C. renale (2), C. striatum (2), C. urealyticum (3), C. xerosis (1), CDC coryneform groups B-1 (2), B-3 (2), F-1, genomospecies 1 and 2 (6), G, genomospecies 1 (1) and G, genomospecies 2 (2). The following strains or species could not be differentiated from each other: C. pseudodiphtheriticum (2) from C. propinquum (former CDC coryneform group ANF-3) (2), CDC coryneform group F-1, genomospecies 1 (4) from genomospecies 2 (2) and C. jeikeium genomospecies A (1) from genomospecies C (2). ARDRA may represent a possible alternative for identification of coryneforms, since this technique enabled the identification of most coryneforms tested and since DNA extraction (i.e. cell lysis by boiling), amplification, restriction and electrophoresis can be carried out within 8 hours. This might allow quick identification of C. diphtheriae and other possible pathogens of the genus Corynebacterium.

Corynebacterium seminale sp. nov., a new species associated with genital infections in male patients

We studied 12 coryneform isolates having similar biochemical profiles which did not permit their assignment to any recognized taxa. Human semen was the source for seven of these strains, whereas the other strains were isolated from urethra, urine, and blood specimens of adult male patients. These bacteria were found in significant quantities (10(4) to 10(5) CFU/ml) in semen specimens from infertile male patients with the diagnosis of prostatitis. These strains had characteristics of the genus Corynebacterium, such as 60 mol% G + C in the DNA and corynemycolic acids, meso-diaminopimelic acid, arabinose, and galactose in the cell wall. Quantitative DNA-DNA hybridizations (S1 nuclease procedure) and phylogenies based on comparisons of almost-complete small-subunit ribosomal DNA sequences confirmed that these strains constitute a single new species within the genus Corynebacterium. All 12 strains showed similar phenotypic features, i.e., good growth on sheep blood agar in contrast with poor growth on the same medium supplemented with 1% Tween 80, a positive CAMP test in the presence of Staphylococcus aureus, glucose and sucrose fermentation, and the presence of beta-glucuronidase. Some strains reduced nitrate and hydrolyzed urea or esculin. These features allowed us to distinguish these strains from members of any other coryneform taxon, and the proposed name is Corynebacterium seminale with strain IBS B12915 (CIP 104297) as the type strain. The description and delineation of these strains as a new species should be useful for further studies, including evaluations of their prevalence among the normal flora and their clinical implications.

Phenotypic characteristics of 31 strains of Corynebacterium striatum isolated from clinical samples

During a 34-month period (January 1991 to October 1993), 31 Corynebacterium striatum stains recovered from clinical samples from 24 patients were characterized. Twenty (64%) strains were isolated from wound exudates, 5 (16%) were isolated from bronchial aspirates, 2 (7%) were isolated from urine, 2 (7%) were isolated from endotracheal tubes, 1 (3%) was isolated from a catheter, and 1 (3%) was isolated from empyema. The organisms were identified by conventional culture and phenotypic characterization, the API CORYNE system, and cellular fatty acid composition analyses. The colonies of C. striatum could be confused with those of coagulase-negative staphylococci upon primary isolation from clinical material. A consistent phenotypic pattern was observed: all strains reduced nitrate, hydrolyzed tyrosine, and produced acid from glucose, fructose, and sucrose but not from maltose. API CORYNE profile numbers were 3100105 (28 strains) and 3000105 (3 strains). Susceptibility testing of C. striatum was performed by disk diffusion. All strains were susceptible to both imipenem and vancomycin and resistant to fosfomycin; most strains were susceptible to ampicillin and cephalosporins and resistant to clindamycin, erythromycin, and tetracycline. Performing a Gram stain of fosfomycin-resistant "Staphylococcus-like" colonies was critical in order to identify C. striatum.

Comparison of E-test with broth microdilution and disk diffusion for susceptibility testing of coryneform bacteria

The susceptibilities of 135 coryneform bacteria isolated from clinical samples to ampicillin (AMP), cephalothin (CR), cefoxitin (FOX), cefotaxime (CTX), erythromycin (E), ciprofloxacin (CIP), tetracycline (TE), amikacin (AK), vancomycin (VA), and rifampin (R) were determined by disk diffusion, broth microdilution, and the E-test. The following species (number of isolates in parentheses) were included: Corynebacterium urealyticum (30), Corynebacterium minutissimum (20), coryneform CDC group ANF-1 (20), Corynebacterium striatum (20), Corynebacterium jeikeium (15), coryneform CDC group I2 (8), Listeria monocytogenes (7), Corynebacterium xerosis (5), and other coryneform bacteria (10). Agreement within one twofold dilution between the E-test and broth microdilution was 31% (VA), 64% (AK), 71% (CTX), 77% (FOX and CIP), 79% (TE), 84% (AMP), 87% (E), and 88% (CR and R). For the 1,350 combinations of microorganisms and antimicrobial agents, 85 (6.3%) discrepancies in interpretive category were found (4.2% minor, 1.2% major, and 0.9% very major). Seventy (5.1%) disagreements in interpretive category were found between disk diffusion and the E-test (3.8% minor, 0.4% major, and 0.9% very major), and 85 (6.3%) disagreements were found between microdilution (reference method) and disk diffusion (4.2% minor, 0.5% major, and 1.5% very major). MICs obtained with the E-test were highly reproducible. No category discrepancy was observed for VA, despite quantitative results. Considering interpretive categories, there is a good overall agreement between the three methods studied here, but further evaluation of current methodologies for susceptibility testing is required when considering coryneform bacteria and determination of quantitative activity of antimicrobial agents.

Additional tests to differentiate Arcanobacterium haemolyticum and Actinomyces pyogenes

A commercially available biochemical test panel, commercially available diagnostic tablets and gas-liquid chromatography (GLC) of cellular fatty acids were used to find out whether Arcanobacterium haemolyticum and Actinomyces pyogenes could be further differentiated from each other. Xylitol and alpha-methyl-D-glucoside fermentation, Voges-Proskauer reaction and tributyrate hydrolysis were found to be useful additional tests which differentiated Arc. haemolyticum and A. pyogenes. GLC analysis revealed major differences in the cellular 16:0, 18:2(9,12) and 18:1(9) fatty acid composition of the two species. Especially the Voges-Proskauer test available as diagnostic tablets can be easily performed in clinical microbiology laboratories, in addition to the tests now used to differentiate Arc. haemolyticum from A. pyogenes.

Taxonomy of Corynebacterium diphtheriae and related taxa, with recognition of Corynebacterium ulcerans sp. nov. nom. rev

Levels of genomic DNA relatedness were determined using a S1 nuclease procedure for reference bacteria identified as biotypes of Corynebacterium diphtheriae, biovars of Corynebacterium pseudotuberculosis, and 'Corynebacterium ulcerans'. These results showed that the three species are separate taxa at the genomospecies level whereas biotypes and biovars are closely related genomically within each species. Phylogenetic analyses of small-subunit rDNA sequences revealed that 'Corynebacterium ulcerans' forms a tight cluster with Corynebacterium pseudotuberculosis within the robust branch that groups all Corynebacterium sequenced to date. Therefore, we propose that the species incertae sedis 'C. ulcerans' should be conclusively recognized as a distinct species within the genus Corynebacterium with strain CCUG 2708 = NCTC 7910 as type strain. This species is characterized by urease production and fermentation of glycogen.

Prosthetic valve endocarditis caused by Corynebacterium afermentans subsp. lipophilum (CDC coryneform group ANF-1)

Corynebacteria are important causes of endocarditis in individuals with valvular prostheses. We report the first published case of prosthetic valve endocarditis caused by the newly defined species Corynebacterium afermentans subsp. lipophilum (former CDC coryneform group ANF-1). The isolate was recovered from a perivalvular abscess specimen and 5 of 15 Bactec blood cultures after 7 to 15 days of incubation. The isolation, identification, and susceptibility testing of Corynebacterium species are discussed.

Assignment of Actinomyces pyogenes-like (CDC coryneform group E) bacteria to the genus Actinomyces as Actinomyces radingae sp. nov. and Actinomyces turicensis sp. nov

In a previous study the authors reported the characterization of some facultatively anaerobic, Gram-positive, non-sporeforming rods which were found in mixed cultures from various infectious processes, including patients with otitis, empyema, perianal abscesses and decubitus ulcers. Phenotypically these organisms closely resembled Actinomyces pyogenes although their precise taxonomic position remained unknown. In the present investigation the authors have determined the 16S rRNA gene sequences of some representative strains of the Actinomyces pyogenes-like bacteria and report the results of a comparative sequence analysis. On the basis of the results of the present and earlier findings two new Actinomyces species, Actinomyces radingae sp. nov. and Actinomyces turicensis sp. nov. are proposed. The type strains are DSM 9169T and DSM 9168T, respectively.

Comparison of one-day versus two-day incubation of urine cultures

The value of incubating urine cultures for 1 versus 2 days was evaluated prospectively for 1526 consecutive specimens. A total of 507 cultures (33.2%) were positive after 1 day; 41 (2.7%) showed different results after 2 days. Only yeasts and corynebacteria were detected more often with longer incubation. Patient charts were available for review from 27 of 41 late positives; in only three instances (11.1%) was action taken by physicians based on these results.

Corynebacterium afermentans subsp. lipophilum: multiple abscess formation in brain and liver

Corynebacterium afermentans, a species recently identified, has previously been isolated from human blood cultures. We report the case of a previously healthy 39-year-old man who developed a brain abscess and a liver abscess due to Corynebacterium afermentans subsp. lipophilum. The liver abscess penetrated through the diaphragm to cause pleural effusion and periostitis of the ribs. We believe this is the first reported case of disseminated infection with abscess formation due to this organism.

Antimicrobial susceptibilities of Corynebacterium species and other non-spore-forming gram-positive bacilli to 18 antimicrobial agents

The susceptibilities of 265 strains of Corynebacterium species and other non-spore-forming gram-positive bacilli to 18 antimicrobial agents were tested. Most strains were susceptible to vancomycin, doxycycline, and fusidic acid. Corynebacterium jeikeium and Corynebacterium urealyticum were the most resistant organisms tested. Resistance to beta-lactams, clindamycin, erythromycin, azythromycin, ciprofloxacin and gentamicin was common among strains of Corynebacterium xerosis and Corynebacterium minutissimum. Ampicillin resistance among Listeria monocytogenes was more prevalent than previously reported. Optochin, fosfomycin, and nitrofurantoin showed very little activity against most organisms tested, but the use of nitrofurantoin as a selective agent in culture medium may prevent the recovery of some isolates. Except for the unvarying activity of vancomycin against Corynebacterium species, the antimicrobial susceptibilities of the latter to other antibiotics are usually unpredictable, such that susceptibility tests are necessary for selecting the best antimicrobial treatment.

Evaluation of API Coryne system for identifying coryneform bacteria

AIM

To identify rapidly and accurately coryneform bacteria, using a commercial strip system.

METHODS

Ninety eight strains of Corynebacterium species and 62 additional strains belonging to genera Erysipelothrix, Oerskovia, Rhodococcus, Actinomyces, Archanobacterium, Gardnerella and Listeria were studied. Bacteria were identified using conventional biochemical tests and a commercial system (API-Coryne, BioMèrieux, France). Fresh rabbit serum was added to fermentation tubes for Gardnerella vaginalis isolates.

RESULTS

One hundred and five out of the 160 (65.7%) organisms studied were correctly and completely identified by the API Coryne system. Thirty five (21.8%) more were correctly identified with additional tests. Seventeen (10.6%) organisms were not identified by the system and three (1.9%) were misidentified.

CONCLUSIONS

The system was a good alternative for identification of coryneform organisms. When occasionally performed with some additional tests, this method permits reliable and rapid identification of coryneform organisms compared with conventional methods.

Genomic diversity among Corynebacterium jeikeium strains and comparison with biochemical characteristics and antimicrobial susceptibilities

Levels of DNA relatedness were determined by performing DNA-DNA hybridization experiments (S1 nuclease procedure) with 13 human isolates exhibiting various antimicrobial susceptibility patterns which had been identified as Corynebacterium jeikeium by classical tests and the API Coryne system and with reference strains of C. jeikeium and related taxa. Twelve of 13 isolates which formed three genomic groups showed between 22 and 75% relatedness with the type strain of C. jeikeium. One of these genomic groups included all the strains resistant to penicillin and gentamicin and is genomically related to the C. jeikeium type strain at the species level. In addition, the reference strain of "Corynebacterium genitalium" biotype II was found to belong to this genospecies and therefore can be considered as a synonym of C. jeikeium. In contrast, one isolate and the reference strains of "Corynebacterium pseudogenitalium" biotypes C-3 and C-4 which were assigned to C. jeikeium by the API Coryne system were less than 10% related to the C. jeikeium type strain. These nongenomically related strains can be differentiated from the jeikeium-related strains on the basis of positive acidification from fructose and growth under anaerobic conditions. Furthermore, these strains exhibited full susceptibility to penicillin whereas the strains related to the C. jeikeium type strain are resistant to or only moderately susceptible to penicillin. No genomic relationship was found between C. jeikeium-related strains and other lipophilic coryneforms, identified as Corynebacterium accolens or Corynebacterium group G or F. Our study demonstrates the necessity to perform the fructose fermentation test or respiratory-type test for the correct identification of lipophilic coryneforms as C. jeikeium. Although these strains show genomic diversity at the species level, in a practical aspect, biochemical properties as well as antimicrobial susceptibility may allow the classification of such isolates in this single taxon.

Seven patients with respiratory infections due to Corynebacterium pseudodiphtheriticum

Seven cases of lower respiratory tract infection due to Corynebacterium pseudodiphtheriticum are described. Lower respiratory tract infections with C. pseudodiphtheriticum in immunocompetent patients are usually associated with pre-existing chronic pulmonary disease, and are sometimes associated with endotracheal intubation. Antimicrobial susceptibility testing of these isolates showed uniform sensitivity to penicillin and variable results to erythromycin.

Differentiation of Brevibacterium spp. encountered in clinical specimens

Forty-three strains belonging to the genus Brevibacterium which were encountered in clinical materials over 2 decades were compared with reference strains, including the type strains, of B. casei, B. epidermidis, B. mcbrellneri, B. iodinum, and B. linens. By means of carbohydrate assimilation tests (CATs) the 43 clinical isolates could be assigned to the species B. casei (n = 41) and B. epidermidis (n = 2). DNA-DNA hybridizations were performed for 20 clinical isolates and confirmed the species identification of the isolates. Cellular fatty acid profiles of all strains were determined and found to have less discriminative power than CATs. This is the first report indicating that most clinical Brevibacterium isolates are B. casei and that CATs provide an easy-to-perform method for species determination within the genus, thus avoiding nucleic acid techniques.

Problems in minimum inhibitory concentration determinations in coryneform organisms. Comparison of an agar dilution and the Etest

The antimicrobial susceptibility of 159 coryneform organisms was determined by an agar dilution and Etest methods. Overall, the correlation between minimum inhibitory concentrations obtained by both techniques was good (> or = 0.09) for most antibiotics and organisms although the essential agreement ranged from 59% to 88.3%. Most organisms were equally categorized (sensitive, intermediate, or resistant) by both methods with only 0.2%, 0.4%, and 3.5% of very major, major or minor discrepancies, respectively. Such percentages dropped significantly when discrepant strains were retested. The correlation was specially good for Corynebacterium jeikeium and Corynebacterium urealyticum.

Comparative in vitro activities of new quinolones against coryneform bacteria

The in vitro activities of eight quinolones against 115 coryneform bacteria (20 Corynebacterium jeikeium, 15 Corynebacterium minutissimum, 15 Corynebacterium striatum, 25 Corynebacterium urealyticum, 10 Corynebacterium xerosis, 10 Corynebacterium group ANF-1, 10 Corynebacterium group 12, and 10 Listeria monocytogenes) were determined. The MICs of ciprofloxacin, ofloxacin, and sparfloxacin for 90% of C. jeikeium, C. urealyticum, and C. xerosis isolates tested were > 16 micrograms/ml. Those of BAY Y 3118 and clinafloxacin against these species were 0.5 and 1 to 2 micrograms/ml, respectively. The MICs for 90% of all 115 strains tested were 0.5 microgram/ml for BAY Y 3118, 1 microgram/ml for clinafloxacin, 2 micrograms/ml for E-5068, 4 micrograms/ml for E-5065, and > 16 micrograms/ml for ciprofloxacin, ofloxacin, sparfloxacin, and E-4868.

Evaluation of a commercial kit in the identification of Arcanobacterium haemolyticum and Actinomyces pyogenes

Listeria-Zym is a commercial kit designed to identify Listeria spp. within 4 h. Its ability to identify two clinically important aerobic catalase-negative, gram-positive rods, Arcanobacterium haemolyticum and Actinomyces pyogenes, was evaluated. Xylose fermentation and alpha-mannosidase tests differentiated Arcanobacterium haemolyticum (n = 49) from Actinomyces pyogenes (n = 24) strains.

Characteristics of CDC group 3 and group 5 coryneform bacteria isolated from clinical specimens and assignment to the genus Dermabacter

Over a 1-year period, 11 isolates (including 5 from blood cultures) of the recently described CDC group 3 and group 5 coryneform bacteria were derived from clinical specimens and compared with reference strains. Biochemical characteristics indicated a very close relationship between CDC group 3 and group 5 coryneform bacteria. The ability of CDC group 3 and the inability of CDC group 5 coryneform bacteria to ferment xylose were the only reactions that were different for the two taxa. Chemotaxonomic features of the two groups included the presence of meso-diaminopimelic acid, a lack of mycolic acids, and the presence of predominantly branched cellular fatty acids, a combination found among gram-positive rods only in Brevibacterium spp., Brachybacterium faecium, and Dermabacter hominis. 16S rRNA gene sequence analysis revealed that CDC group 3 and group 5 coryneform bacteria are members of the genus Dermabacter, which to date has been isolated exclusively from human skin.

Prevalence of Corynebacterium urealyticum in urine specimens collected at a university-affiliated medical center

Corynebacterium urealyticum (formerly Corynebacterium group D2) has been implicated as a cause of alkaline-encrusted cystitis and urinary tract struvite calculi. Despite preselecting urine specimens with neutral and alkaline pHs and using prolonged incubation on a selective medium, isolation of this organism was rarely observed in a population of hospitalized patients. We do not recommend routine cultures for this organism unless the urine is alkaline and struvite crystals, leukocytes, and erythrocytes are present.

Identification of coryneform and other gram-positive rods with several methods

The identification of 202 isolates of aerobically growing Gram-positive rods from clinical material was attempted by using a combination of "traditional" morphological and biochemical tests (Hollis & Weaver (20)) plus patterns of cellular and metabolic fatty acids. This system served as the "gold standard" for three others, i.e. API Coryne (Rapid Coryne), MIDI TSBA and MIDI CLIN Aerobic. In addition, several growth, biochemical and susceptibility tests (growth on cystine-tellurite blood agar, DNase, hippurate and starch hydrolysis, methanethiol formation, API ZYM, CAMP reaction, susceptibility to O/129 and to six antimicrobials) were done in order to check their usefulness for the identification of this group of bacteria. Our system, with the help of chemotaxonomic tests (m-DAP and mycolic acids), was able to identify 154/202 (76%) of the isolates by species and an additional 41/202 (21%) by genus only; 7 (3%) could not be identified. The API Coryne system identified to species or genus level 140/195 isolates (72%). Corresponding figures for the MIDI TSBA and CLIN systems were 63/195 (32%) and 88/195 (45%); further details of species and genus identification are presented in the text. The main drawback of the commercial systems is the extent and probably the numerical depth of the data base. We recommend the use of our multisystem approach for the identification of Gram-positive rods until commercial systems are based on a broader and numerically more extensive data base. The additional tests did not prove species- or genus-specific.

Oral microflora of the ferret at the gingival sulcus and mucosa membrane in relation to ligature-induced periodontitis

The aim of this study was to describe, in the domestic ferret, the composition of the normal gingival sulcus and the near related mucosa microflora and the shift in the composition during an acute periodontal breakdown induced by the placement of ligatures. Five animals were used. At the start of the experiment (day 0) and before the placement of the ligatures, 2 microbiological samples were obtained from the gingival sulcus and the related mucosa at the mid-buccal surface of the right upper P4. The same sites were sampled on the left side, serving as control sites. On day 28, new microbiological samples were taken from the same 4 sites. The samples were analyzed with darkfield microscopy and with viable counts. In the darkfield analysis, on day 0, the predominant bacterial morphotype was cocci, ranging from 88-92%. The viable counts showed that nearly 100% of the total cultivable flora was composed of facultative anaerobic gram-negative and gram-positive rods. Pasteurella spp., Corynebacterium spp. and Rothia spp. were the major components of these bacterial groups. No anaerobic bacteria were detected. On day 28, cocci decreased to approximately 75% in the experimental gingival sulcus sample, and filaments and rods increased significantly. Spirochetes remained largely unchanged during the experiment. The composition of the experimental site samples changed. There was a decrease in the mean numbers of facultative anaerobic gram-positive rods, and approximately one third of the microflora comprised anaerobic gram-negative rods, mainly Porphyromonas gingivalis and Fusobacterium spp. Although the composition of the microflora of the ferret presented few similarities to that described in dogs and monkeys, the anaerobic portion of the gingival sulcus and oral mucosa microflora related to ligature-induced periodontitis presented microorganisms observed in experimental periodontitis both in dogs and monkeys as well as in human periodontal diseases.

Molecular evidence of person-to-person transmission of a pigmented strain of Corynebacterium striatum in intensive care units

During a 14-month period, a unique strain of Corynebacterium striatum that produces a diffusible brown pigment was isolated from purulent sputa of nine patients and from nonrespiratory sites of two additional patients. Seven nonpigmented clinical isolates from the same period and three reference strains of C. striatum were compared with the brown isolates. Most patients had multiple sputum cultures with no coryneforms before the brown strain emerged, suggesting that the organism was hospital acquired. DNA restriction fragment patterns and Southern hybridization with the att site probe of Corynebacterium diphtheriae indicated that the brown isolates were a single strain which was distinct from the heterogeneous nonpigmented strains. A common source for the brown C. striatum was not recognized, although all of these patients were located in two adjoining intensive care units. All of the brown isolates, three of the nonpigmented clinical isolates, and two reference strains had positive CAMP reactions with Staphylococcus aureus, which has not been reported for C. striatum prior to this study.

Catheter-related septicemia caused by a vancomycin-resistant Coryneform CDC group A-5

A case of catheter-related septicemia, due to Coryneform CDC group A-5, in an 11 yr old boy with acute myelomonocytic leukemia is discussed. The child failed to respond to initial antibiotic therapy, even following the addition of vancomycin. Laboratory studies later showed the organism to be vancomycin resistant but cefotaxime susceptible.

Rapid method for presumptive identification of Corynebacterium jeikeium

Corynebacterium jeikeium causes systemic infections, particularly in immunocompromised hosts. A minitube assay has been developed for the presumptive identification of C. jeikeium. With our rapid sucrose-urea test and conventional biochemical tests, sixty isolates of gram-positive, catalase-positive bacilli were identified in our laboratory. Results indicated that our assay has a sensitivity of 100% and a specificity of 90%.

Characteristics of CDC group 1 and group 1-like coryneform bacteria isolated from clinical specimens

Fifteen strains of CDC group 1 coryneform and biochemically similar bacteria were isolated from clinical specimens. Of the 15 strains isolated, 11 were derived from abscesses and purulent lesions, mostly from the upper part of the body, and 3 were grown from blood cultures. Nine strains were associated with mixed anaerobic but no other aerobic flora. Seven strains exhibited the classical biochemical profile of CDC coryneform group 1; however, eight strains were unable to reduce nitrate and were called "group 1-like." Other reactions to differentiate CDC group 1 and group 1-like coryneform rods include alpha-hemolysis on human blood agar, fermentation of adonitol, and the presence of alkaline phosphatase. Fifteen strains showed marked CAMP reactions on different erythrocyte agars. Gas-liquid chromatography of volatile and nonvolatile fatty acids as well as cellular fatty acid patterns and the composition of cell wall components suggest that CDC group 1 and group 1-like coryneform bacteria do not belong to the genus Corynebacterium but possibly to the genus Actinomyces or Arcanobacterium. DNA-DNA hybridization studies revealed that group 1 and group 1-like strains represent different species.

Identification of clinical isolates of nondiphtherial Corynebacterium species and their antibiotic susceptibility patterns

Starting in 1982, our laboratory has performed species identification of coryneform bacteria isolated from blood cultures, intravenous (i.v.) catheter tips and sites, urines with high colony counts, and other potentially significant cultures, using predefined criteria. Of 283 isolates identified, Corynebacterium jeikeium was the most common (47%), followed by CDC group G2 (12%) and C. minutissimum (8%). Blood cultures and i.v. catheter-related sources were the most frequent sources (58% of total). Certain species or groups, like CDC group G2, were most frequently isolated from blood or i.v. catheter sites. CDC group G2 showed a progression to greater multiple antibiotic resistance during this 9-year period. Occasional multiresistant strains of other species were also encountered. By in vitro testing, we note vancomycin remains the most active agent against corynebacterialike organisms, and is the most reliable antibiotic to use while awaiting susceptibility testing results.

Coryneform bacteria isolated from middle ear fluid

Nineteen strains of facultatively anaerobic gram-positive rods isolated in pure culture from middle ear fluids were identified. All effusions were collected by tympanocentesis in children with acute otitis media. Identification of microorganisms to the genus level was done by studying the cell wall composition. Sixteen strains contain meso-diaminopimelic acid and arabinogalactan polymer but lack mycolic acids; therefore, these strains do not belong to a previously described taxon. Because of similarities with Corynebacterium afermentans (Centers for Disease Control group ANF-1), we temporarily classified these mycolateless strains ANF-1 like. Isolation of these microorganisms in pure culture from middle ear fluids collected by tympanocentesis is a strong argument for their involvement in acute otitis media.

Adherence of Corynebacterium urealyticum (CDC group D2) and Corynebacterium jeikeium to intravascular and urinary catheters

The ability of Corynebacterium urealyticum, Corynebacterium jeikeium and other control strains to adhere to two intravascular catheters (polyvinyl chloride and Teflon) and one urinary catheter (Teflon-coated rubber) was studied. Results demonstrated that the Corynebacterium species adhered to all catheter materials in greater numbers than a control strain of Micrococcus luteus (p < 0.001). There was not a clear difference in the ability of the strains of Corynebacterium jeikeium and Corynebacterium urealyticum to adhere to the catheters tested, so that differences other than this property could explain their different pathogenicity for humans.

Isolation of gram-positive rods that resemble but are clearly distinct from Actinomyces pyogenes from mixed wound infections

Beginning in 1990, gram-positive rods resembling Actinomyces pyogenes were found with increasing frequency in mixed cultures from various infectious processes, most of them from patients with otitis, empyema, pilonidal cysts, perianal abscesses, and decubitus ulcers. Ribotyping and hybridization showed that these gram-positive rods could be divided into five groups not related to known Actinomyces species. Biochemical markers for reliable differentiation into these groups, however, could not be found. Therefore, naming new species is not warranted unless parameters are discovered that allow identification without DNA hybridization. These gram-positive rods have been isolated only in mixed cultures with anaerobes, Staphylococcus aureus, Streptococcus "milleri," enterococci, and gram-negative rods. Their exact role in these possibly synergistic infections needs further investigation.

Chemotaxonomy and molecular taxonomy of some coryneform clinical isolates

Six reference strains of the genus Brevibacterium as well as fifteen clinical isolates tentatively assigned to the genus using conventional biochemical methods, were the subject of chemotaxonomic and DNA similarity studies. Five of these clinical isolates were assigned either to the genera Aureobacterium, Mycobacterium, Gordona or to Rhodococcus on the basis of their DNA mol% G+C, mycolic acid, amino-acid, sugar and menaquinone contents. Among the ten remaining strains, six were not brevibacteria and only four conformed to the description of the genus Brevibacterium sensu stricto. These strains showed low values of DNA relatedness with Brevibacterium epidermidis ATCC 35514T and Brevibacterium linens ATCC 9174, and could only be described as Brevibacterium spp. The results indicate that studies of chemical markers are essential for the correct identification of brevibacteria.

Two cases of Actinomyces pyogenes infection in humans

Few confirmed human cases of Actinomyces pyogenes infection have been reported partly because of inadequate identification of this bacterium. In this study, two new cases are reported with emphasis on the characteristics which are relevant for identification of the isolates. A review of previously reported cases is presented.

Urinary tract infection with Corynebacterium urealyticum in South Africa

Corynebacterium urealyticum has been well documented as a urinary pathogen in Europe. The purpose of this study was to investigate the incidence and clinical relevance of Corynebacterium urealyticum as a urinary pathogen in a predominantly Third World population (South Africa) and to attempt to increase the isolation rate by culturing urine specimens on a selective medium. Two methods were used to isolate Corynebacterium urealyticum from urine specimens. Blood agar plates from routine urine cultures of 7,912 urine specimens were incubated for 48 hours and 1,281 specimens were cultured on a selective medium as well as on routine media. The antimicrobial susceptibility of all isolates of Corynebacterium urealyticum was tested. The yield of Corynebacterium urealyticum on blood agar was three isolates in three patients (0.038% of 7,912), all of whom had pyuria, alkaline urine and risk factors for Corynebacterium urealyticum infection. Use of selective media increased the yield of Corynebacterium urealyticum to 15 of 1,281 specimens (1.17%), however 73% of these urine samples yielded other pathogens and none had an alkaline pH which could not be attributed to the presence of another urealytic pathogen. All isolates were susceptible to vancomycin and 92.6% susceptible to norfloxacin. The pathogenic potential of Corynebacterium urealyticum was confirmed in South African patients, but the incidence of infection was low. The use of a selective medium is therefore not cost-effective in all cases but could be used selectively on the basis of typical urine findings and patient criteria.