Coryneform Gram-Positive Rods


TAXONOMY Back to top

The bacteria discussed in this chapter all belong to the class Actinobacteria, the genera of

which are characterized by specific 16S rRNA gene signature nucleotides (144, 168)

belonging to the lineage of gram-positive bacteria with high guanine-plus-cytosine (G+C)

contents, with the exception of the genusExiguobacterium, which belongs to the

class Firmicutes, the members of which have low G+C contents. The coryneform bacteria are

most diverse and are differentiated by chemotaxonomic features (Table 1). Phylogenetic

investigations, in particular, 16S rRNA gene sequencing, have in general confirmed the

framework set by chemotaxonomic investigations. The 16S rRNA gene sequencing data

demonstrate that the genera Corynebacterium and Turicella are more closely related to the

partially acid-fast bacteria and to the genus Mycobacterium than to the other coryneform

organisms covered in this chapter (71, 108, 133). The genus Arthrobacter, which contains

rods, is phylogenetically intermixed with the genus Micrococcus (and genera formerly

called Micrococcus), which contains cocci (see chapter 19) (54, 84). The

genus Rothia contains both rod-forming organisms, represented by Rothia dentocariosa, and

a coccus-forming species, Rothia mucilaginosa, formerly Stomatococcus mucilaginosus (24).

Other genera which are phylogenetically closely related include Oerskovia,

Cellulosimicrobium, and Cellulomonas (14, 64, 137, 143), as well

as Arcanobacteriumand Actinomyces (109).



DESCRIPTIONS OF THE GENERA Back to top

Genus Corynebacterium

The number of species belonging to the genus Corynebacterium has dramatically increased

from 22 in 1990 to 81 species (and one taxon group) at the time of writing, 50 of which (and

one taxon group) are medically relevant. Clinically relevant Corynebacterium species validly

described since the last edition of this Manualinclude C. canis (50), C. freiburgense (53), C.

hansenii (119), C. massiliense (96), C. pilbarense (3), C. pyruviciproducens (151), C.

sputi (164), C. stationis (9), C. timonense (96), and C. ureicelerivorans (163).

The cell walls of corynebacteria contain meso- diaminopimelic acid (m-DAP) as the diamino

acid as well as short-chain mycolic acids with 22 to 36 carbon atoms (22). The medically

relevant Corynebacterium species C. amycolatum, C. atypicum, and C. kroppenstedtii as well

as C. caspium (from seals) and C. ciconiae (from storks) lack mycolates (21, 23, 77). The

corynebacterial cell wall contains arabinose and galactose (22). Palmitic (C16:0), oleic

(C18:1ω9c), and stearic (C18:0) acids are the main cellular fatty acids (CFAs) in all

corynebacteria, and tuberculostearic acid (TBSA) can be detected in some medically

relevant Corynebacterium species (e.g., C. urealyticum, C. kroppenstedtii, C. confusum, C.

appendicis, and C. minutissimum) (7, 23, 61, 165). The G+C contents

of Corynebacterium spp. vary from 46 to 74 mol%, indicating the enormous diversity within

this genus. The phylogenetic relationships within the genus Corynebacterium have been

outlined previously (108,133), creating an extensive and reliable database for future

comparative 16S rRNA gene studies, e.g., for the delineation of new species. Complete

genome sequences for Corynebacterium diphtheriae (17),Corynebacterium

jeikeium (148), C. urealyticum (150), and C. kroppenstedtii (149) have been described

previously.

Gram staining of corynebacteria shows slightly curved, gram-positive rods with sides not

parallel and sometimes slightly wider ends, giving some of the bacteria a typical club shape

(Fig. 1a). Corynebacteria whose morphologies differ from this morphology include C. canis,

C. durum, C. matruchotii, and C. sundsvallense(see below under each species’ name). Cells

generally stain evenly. If Corynebacterium cells are taken from fluid media, they are

arranged as single cells, in pairs, in V forms, in palisades, or in clusters with a so-called

Chinese-letter appearance. It is again emphasized that the club-shaped form of the rods is

observed only for true Corynebacterium spp.



Genus Turicella

The genus Turicella is phylogenetically closely related to the genus Corynebacterium but

contains Turicella otitidis as the only species. The cell wall contains m-DAP, but mycolic acids

are not present (71). The main CFAs for T. otitidis are the same as those

for Corynebacterium spp., but all T. otitidis strains also contain significant amounts of TBSA

(2 to 10% of all CFAs) (71). T. otitidis is the only coryneform bacterium that has a polar lipid

profile without glycolipids. The G+C content varies between 65 and 72 mol% (71). Gram

staining shows relatively long gram-positive rods (Fig. 1b). T. otitidis is catalase positive,

nonmotile, and an oxidizer.

Genus Arthrobacter

The genera Arthrobacter and Micrococcus are so closely related phylogenetically that it has

been stated that micrococci are, in fact, arthrobacters which are unable to express rod forms

(84). Presently, the genusArthrobacter contains over 50 species, of which only a few have

been recovered from human clinical specimens (92). Lysine is the diamino acid of the cell

wall, and C15:0ai is the overall dominating CFA; it represents more than 50% of all CFAs in

most Arthrobacter species. The G+C contents vary between 59 and 70 mol%, indicating the

diversity within this genus.

Gram staining may demonstrate a rod-coccus cycle (i.e., rod forms in younger cultures and

cocci in older colonies) when cells are grown on rich media (e.g., Columbia base agar).

Jointed rods (i.e., rods in a rectangular form that contributed to the designation of this genus

as “arthros,” which means “joint” in ancient Greek) may also be observed in younger

cultures (i.e., at 24 h) but may not be demonstrable for every species. Arthrobacters are

catalase positive, their motility is variable, and they are always oxidizers.

Genus Brevibacterium

The genus Brevibacterium presently comprises 22 species, of which 9 species are medically

relevant. m-DAP is the diamino acid type found in the cell wall. C15:0ai and C17:0ai usually

represent more than 75% of all CFAs (48). The G+C contents vary between 60 and 70

mol%.

Gram staining demonstrates relatively short rods, which may develop into cocci when

cultures become older (after 3 days). Brevibacteria are catalase positive, nonmotile, and

oxidizers.

Genus Dermabacter

The genus Dermabacter presently comprises only one species, D. hominis. m-DAP is the

diamino acid of the cell wall, and C15:0ai and C17:0ai usually account for 40 to 60% of all CFAs.

The G+C content range is between 60 and 62 mol% (81). Gram staining shows very short

rods (Fig. 1c), which are often initially misinterpreted as cocci. D. hominis strains are

catalase positive, nonmotile, and glucose fermenters.

Genus Helcobacillus

The genus Helcobacillus, part of the family Dermabacteraceae, has been described for an

isolate recovered from a patient with a cutaneous discharge presenting with an erythrasma.

Interestingly, phenotypic features of the bacterium were initially thought to be suggestive

of C. minutissimum but later, by a polyphasic approach, of D. hominis. The Gram stain

shows gram-positive, straight, short (0.7 to 1.0 μm in length by 0.4 to 0.7 μm in diameter),

irregular rods. Colonies are apigmented. The type strain is catalase positive, nonmotile, and

a glucose fermenter. This taxon is amycolated, with a G+C content of 68.6 mol%; m-DAP is

the cell wall diamino acid, with the CFAs C15:0ai , C17:0ai , and C16:0i (together, ~70% of the

total) predominating (122).

Genus Rothia

The genus Rothia is also included in this chapter because some species are rod-like. The

genus Rothia belongs to the family Micrococcaceae. Collins and colleagues

reclassified Stomatococcus mucilaginosus as Rothia mucilaginosa (24). Since Rothia

mucilaginosa exhibits coccoid forms in Gram stains, the genus Rothia is also covered

in chapter 19 (on the catalase-positive gram-positive cocci). However, the species Rothia

dentocariosaclearly exhibits mainly rod forms and is, therefore, covered in this chapter on

coryneform bacteria.

Lysine is the diamino acid of the cell wall, and C15:0ai and C17:0ai usually represent 40 to 60%

of all CFAs. The G+C contents range between 47 and 56 mol%. Rothia strains can be quite

pleomorphic by Gram staining, but filamentous forms are normally not observed. They have

a variable catalase reaction, are nonmotile, and exhibit a fermentative metabolism.

Genus Exiguobacterium

The genus Exiguobacterium is phylogenetically related to the so-called “group 2 bacilli” (39).

Presently, 13 species are included in this genus, of which only E. acetylicum and E.

aurantiacum have been mentioned in publications as being isolated from human clinical

material. Lysine is the diamino acid of the cell wall, and C15:0ai and C17:0ai comprise only about

30 to 40% of the total CFAs. E. acetylicum contains significant amounts of C13:0 and C13:0ai,

which are not found in any other coryneform taxon (7). The G+C content is about 47 mol%.

Exiguobacteria present as relatively short rods in young cultures. Strains are catalase

positive and motile and have a fermentative metabolism.

Genus Oerskovia

In older textbooks, oerskoviae were assigned to the nocardioform group of organisms due to

their morphological features. This includes branching vegetative substrate hyphae and

penetration into agar but no aerial hyphae. However, there is now phylogenetic and

chemotaxonomic evidence that Oerskovia, including the reclassified

genus Promicromonospora (143), is more closely related to genera like Cellulomonas than to

the mycolic acid-containing genera like Nocardia. Representatives of the type

species, Oerskovia turbata, were recovered from soil, but human pathogens originally

identified as O. turbata have now been placed inCellulosimicrobium funkei (14). Lysine is the

diamino acid of the cell wall, and C15:0ai is the main CFA in oerskoviae. The G+C content is 70

to 75 mol%.

Gram staining shows coccoid-to-rod-shaped bacteria which originate from the breaking up of

mycelia.Oerskovia strains are catalase positive, their motility is variable, and they are

fermentative.

Genus Cellulomonas

The genus Cellulomonas presently comprises 17 species, of which only C. hominis and C.

denverensis have been described as being isolated from humans (13, 64, 105). Ornithine is

the diamino acid of the cell wall, and C15:0ai and C16:0 are the main CFAs. The G+C content is

71 to 76 mol%.

Gram staining shows small, thin rods. All Cellulomonas spp., except C. fermentans and C.

humilata, are catalase positive, their motility is variable, the

environmental Cellulomonas strains are cellulolytic (whereas C. hominisdid not hydrolyze

cellulose in the test system used) (64), and they have a fermentative metabolism.

Genus Cellulosimicrobium

The Cellulosimicrobium genus presently comprises three species. The medically relevant

species C. cellulanshad been designated Cellulomonas cellulans or Oerskovia

xanthineolytica in the past (137). The reason for removing C. cellulans from the

genus Cellulomonas was that the topology of the 16S rRNA gene dendrogram indicated that

the branching point of this taxon was outside Cellulomonas proper. In addition, the

chemotaxonomic characteristic of lysine as diamino acid supported reclassification.

Predominant CFAs include C15:0ai, C15:0i, C16:0i, and C16:0. The major menaquinone (MK) is MK-

9(H4), and the G+C content is 74 mol%. It should be noted that the

genus Cellulosimicrobium is related to the genus Oerskovia but is nevertheless distinct.

In young cultures, a mycelium that fragments later into irregular, curved, and club-shaped

rods is produced. Catalase activity is detected, and strains are nonmotile. All strains have a

fermentative metabolism.

Genus Microbacterium

It has been known since the mid-1990s that the

genera Microbacterium and Aureobacterium are phylogenetically intermixed (114), and the

diamino acid in the third position of the tetrapeptide of the peptidoglycan was considered a

most important chemotaxonomical marker. L-Lysine is present in microbacteria and Dornithine

in the former aureobacteria, a phenomenon noted for some genera

(e.g.,Propionibacterium and Bifidobacterium), where there is not a good correlation between

the type of the diamino acid in the peptidoglycan layer and their phylogenetic positioning.

Because a set of signature nucleotides within the 16S rRNA genes of both microbacteria and

aureobacteria could be demonstrated, it was proposed that both genera be unified in a

redefined genus, Microbacterium (145).

Now, over 50 Microbacterium species have been validly named, but only a minority of them

has been demonstrated to be of clinical importance (74). Microbacteria are most frequently

encountered in environmental specimens (e.g., soil). C15:0ai and C17:0ai are the two main

CFAs, often representing up to 75% of the total CFAs (51, 73, 145). The G+C content

of Microbacterium spp. is 65 to 76 mol%, indicating the diversity within the genus.

Gram staining often shows thin or short rods with no branching. Catalase activity and

motility are variable. Microbacteria can be either fermenters or oxidizers.

Genus Curtobacterium

Curtobacterium spp., like microbacteria, belong to the peptidoglycan type B actinomycetes

(i.e., cross-linkage exists between positions 2 and 4 of the two peptide subunits). Ornithine

is the diamino acid and the only amino acid composing the interpeptide bridge. Curtobacteria

have an acetyl peptidoglycan acyl type and MK-9 as the major MK, whereas microbacteria

possess a glycolyl type and MK-11,12 (Table 1). For most curtobacteria, C15:0ai and

C17:0ai represent more than 75% of all CFAs (46). The G+C contents range from 68 to 75

mol%. Presently, nine Curtobacterium species are validly described.

Gram staining shows small and short rods with no branching. Catalase activity is positive,

motility is observed in most strains, and all strains show a respiratory metabolism which

proceeds slowly in oxidizing carbohydrates.

Genus Leifsonia

The former “Corynebacterium aquaticum” was transferred into the

genus Leifsonia as Leifsonia aquatica in 2000 (38) and is the only medically relevant species

in this genus. L. aquatica strains belong to the peptidoglycan B-type actinomycetes and,

therefore, cannot be true corynebacteria, which actually possess an A-type peptidoglycan

(i.e., cross-linkage exists between positions 3 and 4 of the two peptide subunits).

Diaminobutyric acid is the diamino acid of the cell wall peptidoglycan, and C15:0ai and

C17:0ai are the main CFAs, as with microbacteria, but represent <75% of all CFAs (73). The

G+C content is about 70 mol%.

Gram staining shows thin rods. The strains are catalase and oxidase positive (the latter

feature is atypical for coryneform bacteria), always motile, and oxidizers.

Other Unusual Coryneforms

Recent examination of coryneform bacteria using primarily sequence-based identification

approaches have shown that additional genera can be recovered from human clinical

materials. These are briefly mentioned below.

Genus Janibacter

Strains of the genus Janibacter (94), the first medically relevant coryneform reported from

the familyIntrasporangiaceae, were found to be associated with bacteremia (36, 91; K. A.

Bernard, unpublished observation). Janibacter strains can have Gram stain-variable or grampositive

coccoidal-to-rod-like forms in singles, pairs, or irregular clumps. Their DNA base

composition is 69 to 73 mol% G+C, with an unusual CFA profile consisting of significant

volumes of the CFAs C16:0i, C17:1, and C17:0. These bacteria are described as oxidizers, with

white, creamy, or yellowish pigments. They are nonmotile, and optimal growth may occur at

25 to 30°C.

Genus Pseudoclavibacter

The Pseudoclavibacter genus was first described in 2004 to accommodate the misidentified

type strain of “ Brevibacterium helvolum” (93). Shortly afterwards, the novel

genus Zimmermannella was proposed but was found to be an illegitimate, later, homotypic

synonym of Pseudoclavibacter(http://www.bacterio.cict.fr/xz/zimmermannella.html).

Pseudoclavibacter alba strains were recovered from urine, and Pseudoclavibacter

bifida strains were recovered from blood cultures and wounds. By Gram staining, these

species were short or medium-length gram-positive rods, with P. bifida demonstrating some

rudimentary branching. Their DNA base composition is 62 to 68 mol% G+C. Major CFAs are

C15:0ai, C16:0, C16:0i, and C17:0ai. These strains are oxidizers, with white or yellowish colonies

(93). Optimal growth occurs at 30°C.

Genera Brachybacterium and Knoellia

Isolates from the genus Brachybacterium and one strain from the genus Knoellia, all

recovered from blood cultures, have been characterized (K. A. Bernard, unpublished

observations). There are currently 12 species in the genus Brachybacterium, which is part of

the family Dermabacter aceae, and so they are most closely related to the

genus Dermabacter. Members of this genus grow at 37°C, exhibit gram-positive coccoidal

and rod-like forms, and have a G+C content of 68 to 73 mol%. The Brachybacterium blood

culture isolates are metabolically fermentative and have branched-chain-type CFAs. The

genus Knoellia, like that of the genus Janibacter, is a member of the

family Intrasporangiaceae. Cells are irregular gram-positive rods or cocci with major CFAs of

the branched-chain type, and their G+C content is 68 to 69 mol%. The single Knoellia blood

culture isolate is capnophilic, growing best in 5% CO2 at 37°C.

Genus Arcanobacterium

The Arcanobacterium genus presently contains nine species, of which A. haemolyticum, A.

bernardiae, and A. pyogenes have been recovered from human clinical specimens. Lysine is

the diamino acid of the cell wall, whereas in the phylogenetically closely

related Actinomyces spp., lysine or ornithine is found. Arcanobacteria contain MKs of the MK-

9(H4) type, whereas the Actinomyces spp. examined so far have MK-10(H4). The main CFAs

of arcanobacteria are C16:0, C18:1ω9c, and C18:0 (as in Corynebacterium spp. and T. otitidis),

but unlike with corynebacteria, significant amounts of C10:0, C12:0, and C14:0 may also be

detected (7). The G+C content is 48 to 52 mol%.

Gram staining of arcanobacteria shows irregular, gram-positive rods. All clinically relevant

arcanobacteria are catalase negative; they are nonmotile and fermenters.

Genus Gardnerella

The genus Gardnerella does not have a particular phylogenetic relationship to any of the

established genera described in this chapter. It is remotely related to the

genus Bifidobacterium (97), and these genera share some important features, such as

production of acetic and lactic acids as fermentation products. G. vaginalis is the only species

belonging to the genus Gardnerella. Studies of the ultrastructure of the cell wall of G.

vaginalishave demonstrated that it has a cell wall similar to but much thinner than the cell

walls of other gram-positive bacteria (i.e., there is a smaller peptidoglycan layer) (134).

Lysine is the diamino acid of the cell wall, and CFAs are similar to those detected

in Actinomyces spp., Arcanobacterium spp., and Corynebacterium spp., with C16:0and

C18:1ω9c predominating. The G+C content of 42 to 44 mol% is lower than that of every other

genus described in this chapter.

Gram stains show thin Gram stain-variable rods or coccobacilli (Fig. 1d). Catalase is not

produced, and cells are nonmotile and have a slow fermentative metabolism.

EPIDEMIOLOGY AND TRANSMISSION Back to top

Many species of the corynebacteria are part of the normal biota of the skin and mucous

membranes in humans and other mammals. The habitat for some medically irrelevant

corynebacteria (e.g., C. terpenotabidumand C. halotolerans) is the environment. It is

noteworthy that not all corynebacteria are equally distributed over skin and mucous

membranes; many of them occupy a specific niche. C. diphtheriae can be isolated from the

nasopharynx as well as from skin lesions, which actually represent a reservoir for the spread

of diphtheria. Important opportunistic pathogens like C. amycolatum, Corynebacterium

striatum, and D. hominisare part of the normal human skin biota but have thus far not been

recovered from throat swabs from healthy individuals (158). Coryneform bacteria prominent

in the oropharynx include Corynebacterium durum and R.

dentocariosa (158). Corynebacterium auris and T. otitidis seem to have an almost exclusive

preference for the external auditory canal. In nearly every instance that Corynebacterium

macginleyi has been isolated, eye specimens have been the source (62).

Another Corynebacterium species with a distinctive niche isCorynebacterium

glucuronolyticum, which is almost exclusively isolated from genitourinary specimens from

humans (47) and from animals (28). C. urealyticum, another genitourinary pathogen, has,

like C. jeikeium,been cultured from inanimate hospital environments.

The natural habitat of arcanobacteria is not fully understood, but A. haemolyticum is

recovered from throat as well as from wound swabs, whereas A. bernardiae has been found

mainly in abscesses adjacent to skin (G. Funke and K. Bernard, unpublished observations). It

is unclear whether the two species are part of the normal skin and/or the gastrointestinal

biota. A. pyogenes is found on mucous membranes of cattle, sheep, and swine. Brevibacteria

can be found on dairy products (e.g., cheese) but are also inhabitants of the human skin

(48). Arthrobacters are some of the most frequently isolated bacteria when soil samples are

cultured, but at least Arthrobacter cumminsii also seems to be present on human skin

(63, 92). Members of the generaExiguobacterium, Oerskovia, Cellulomonas,

Cellulosimicrobium, and Microbacterium have their habitats in the inanimate environment

(e.g., soil and activated sludge). Microbacterium spp. have also been recovered from hospital

environments (51). Curtobacteria are primarily plant pathogens (46).

G. vaginalis can be found in the anorectal biota of healthy adults of both sexes as well as in

that of children (16). It is also part of the endogenous vaginal biota in women of

reproductive age. The optimal pH for the growth of G. vaginalis is between 6 and 7, i.e., at

elevated pH in the vagina. The organism can also be recovered from the urethras of the male

partners of women with bacterial vaginosis (BV) (16).

CLINICAL SIGNIFICANCE Back to top

Estimating the clinical significance of coryneform bacteria isolated from clinical specimens is

often confusing for clinical microbiologists. This is in part due to the natural habitat of

coryneform bacteria, which may lead to their recovery if specimens were not taken correctly.

The reader is referred to the guidelines on minimal microbiological requirements in

publications on disease associations of coryneform bacteria (72).

Coryneform bacteria should be identified to the species level if they are isolated (i) from

normally sterile body sites, e.g., blood (but not if only one of multiple specimens becomes

positive), (ii) from adequately collected clinical material if they are the predominant

organisms, and (iii) from urine specimens if they are the only bacteria encountered and the

bacterial count is >104/ml or if they are the predominant organisms and the total bacterial

count is >105/ml.

The clinical significance of coryneform bacteria is strengthened by the following findings: (i)

multiple specimens are positive for the same coryneform bacteria; (ii) coryneform bacteria

are seen in direct Gram stains, and a strong leukocyte reaction is also observed; and (iii)

other organisms recovered from the same material are of low pathogenicity.

For a comprehensive summary of case reports on individual coryneform bacteria, the reader

is referred to review articles (6, 72).



Historically, diphtheria caused by C. diphtheriae (or Corynebacterium ulcerans) has been the

most prominent infectious disease for which coryneform bacteria are responsible. Therefore,

special attention is given to that disease in this chapter. Due to immunization programs, the

disease has nearly disappeared in countries with high socioeconomic standards. However,

the disease is still endemic in some subtropical and tropical countries as well as among

individuals of certain ethnic groups (e.g., indigenous peoples in the Americas and Australia).

In the 1990s, diphtheria reemerged in the states of the former Soviet Union. However,

despite increased global travel activities, only a few imported cases have been reported by

countries with well-developed health care systems.

The main manifestation of diphtheria is an upper respiratory tract illness accompanied by a

sore throat, dysphagia, lymphadenitis, low-grade fever, malaise, and headache. A

nasopharynx-adherent membrane which may occasionally lead to obstruction is

characteristic. The severe systemic effects of diphtheria include myocarditis, neuritis, and

kidney damage caused by the C. diphtheriae exotoxin, which is encoded by a bacteriophage

carrying the tox gene. C. diphtheriae may also cause cutaneous diphtheria or endocarditis

(with either toxin-positive or toxin-negative strains). Some people with poor hygienic

standards (e.g., drug and alcohol abusers) are prone to colonization (on the skin more often

than in the pharynx) by C. diphtheriaestrains, which are often nontoxigenic.

G. vaginalis is associated with BV. Its causative role in the syndrome is controversial

(141, 142), as it is certainly not the sole cause; other bacteria, like Atopobium vaginae,

Leptotrichia/Sneathia species, andMegasphaera-like bacteria are also involved in BV

(44, 155). Recurrent BV is due to reinfection rather than to relapse (i.e., overgrowth of the

previously colonizing biotype). In pregnant women, BV may lead to preterm birth, premature

rupture of membranes, and chorioamnionitis (16). G. vaginalis may also be recovered from

cultures of blood from patients with postpartum or postabortal fevers and may also cause

infections in newborns. Although it might be recovered from the urethras of males, its

disease association in males is usually questionable. Cases involving serious infections

(septicemia, wound infections) in sites other than those associated with the genital tract or

obstetrics are rare but have been reported, including in men (87).

COLLECTION, TRANSPORT, AND STORAGE OF

SPECIMENS Back to top

In general, coryneform bacteria do not need special handling when samples are collected.

The general principles outlined in chapters 9 and 16 in this Manual apply to this group of

organisms as well.

C. diphtheriae

The diagnosis of diphtheria is primarily a clinical one. The physician should notify the

receiving laboratory immediately of suspected diphtheria. In case of respiratory diphtheria,

material for culture should be obtained on a swab (either a cotton- or a polyester-tipped

swab) from the inflamed areas in the nasopharynx. Multisite sampling (nasopharynx) is

thought to increase sensitivity. If membranes are present and can be removed (swabs from

beneath the membrane are most valuable), they should also be sent to the microbiology

laboratory (although C. diphtheriae might not be culturable from those in every instance).

Nasopharyngeal swabs should be obtained from suspected carriers. It is preferable that the

swabs are immediately transferred to the microbiology laboratory for culturing. If the swabs

must be sent to the laboratory, semisolid transport media (e.g., Amies) ensure the

maintenance of the bacteria. All coryneform bacteria are relatively resistant to drying and

moderate temperature changes. Material from patients with suspected cases of wound

diphtheria can be obtained by swab or aspiration.

Vaginal and extravaginal specimens for culturing G. vaginalis can be collected with cottontipped

swabs. It is best to take one swab for direct examination and to take another swab for

culture if necessary, such as for epidemiologic studies. If culture media cannot be directly

inoculated, then the swab should be placed in a transport medium (e.g., Amies) and culture

should be done within 24 h. It is noteworthy that G. vaginalis is susceptible to sodium

polyanethol sulfonate (SPS), so an SPS-free medium (or an SPS medium supplemented with

gelatin) should be used in order to achieve optimal recovery of G. vaginalis from blood

culture systems whenever G. vaginalis is suspected.

Long-term preservation in skim milk at −70°C is applicable to all coryneform bacteria.

Except with samples containing lipophilic corynebacteria, the same skim milk tube can be

repeatedly thawed and frozen (G. Funke, unpublished observation). For nonlipophilic

coryneforms, good results were also observed with Microbank tubes (Pro Lab Diagnostics,

Austin, TX) (G. Funke, unpublished observation). The advantage of using these tubes is that

individual beads can be taken out of the tube. Coryneform bacteria can also be stored for

decades when they are kept lyophilized in an appropriate medium (e.g., 0.9% NaCl

containing 2% bovine serum albumin).

DIRECT EXAMINATION Back to top

After the appropriate isolation media have been inoculated (see below under “Isolation

Procedures”), the swabs taken from diphtheritic membranes may be subjected to Neisser or

Loeffler methylene blue staining. A positive stain is characterized by metachromatic granules

(polar bodies). However, it is noteworthy that the sensitivity of the microscopic examination

is limited. Antigen assays for the direct detection of coryneform bacteria are not

recommended.

As described previously, C. diphtheriae, C. ulcerans, and C. pseudotuberculosis are the only

species able to harbor the bacteriophage which carries the diphtheria tox gene and

potentially produce diphtheria toxin (72). PCR-based, direct-detection systems for the

diphtheria tox gene have been described, using methods to detect fragment A and/or the

entire tox gene (32, 107) or fragment A and B subunits of the tox gene (101). The system

described by Nakao and Popovic had the highest sensitivity when Dacron polyester-tipped

swabs were used and when silica gel packages were stored at 4°C rather than at room

temperature (101). Conventional PCR detection of the regulatory dtxR gene has been

evaluated (111). Detection of the tox gene using a real-time platform has been outlined

(15, 99, 136). Real-time detection using primers targeting the C. diphtheriae tox gene for C.

ulcerans strains have required modifications (15, 135, 136). Direct detection of the

diphtheria tox or dtxR gene as the sole test of clinical specimens has not been

recommended, as expression of diphtheria toxin must be demonstrated (some strains may

harbor but not express the gene), and so microbiological culture is essential for confirming

diphtheria (33).

The “gold standard” for the diagnosis of BV is direct examination of vaginal secretions and

not the culture ofG. vaginalis, since G. vaginalis can also be recovered from healthy women.

A bedside test for BV is examination of the vaginal discharge to detect the typical “fishy”

trimethylamine odor, which is enhanced after alkalinization with 10% KOH (but G.

vaginalis is not responsible for the amine production). The typical smear of vaginal discharge

from BV patients shows “clue cells” (bacteria covering epithelial cell margins) together with a

mixed biota consisting of large numbers of small gram-negative

(predominantly Prevotella and Porphyromonasspp.) and gram-variable (G. vaginalis) rods

and coccobacilli, whereas lactobacilli are almost always absent. It is recommended that a

standardized Gram staining interpretative scheme be used in order to improve the

reproducibility of this method (104, 142). Although not recommended for routine laboratory

procedures, the isolation of G. vaginalis can support the diagnosis of BV.

ISOLATION PROCEDURES Back to top

Coryneform bacteria, including C. diphtheriae, can be readily isolated from a 5% sheep blood

agar (SBA)-based selective medium containing 100 μg of fosfomycin per ml (plus 12.5 μg of

glucose-6-phosphate per ml), since nearly all coryneforms (except Actinomyces spp. and D.

hominis) are highly resistant to this compound (152). It is also possible to put disks

containing 50 μg of fosfomycin (plus 50 μg of glucose-6-phosphate [already incorporated in

the disk]) (BD Diagnostics, Sparks, MD) on an SBA plate and then examine the colonies

which grow around the disk. Selective media for coryneform bacteria containing 50 to 100

μg/ml furazolidone (Sigma, St. Louis, MO) have also been described. If lipophilic

corynebacteria like C. jeikeium or C. urealyticum are sought, then 0.1 to 1.0% Tween 80

(Merck, Darmstadt, Germany) could be added to an SBA plate (add Tween 80 before pouring

the medium). Additional methods to demonstrate lipophilia are described below. Medically

relevant coryneforms described to date do not grow on MacConkey agar. However, if

“coryneform” bacteria are recovered from this medium, they should be examined carefully to

rule out rapidly growing mycobacteria.

With very few exceptions (some arthrobacters, microbacteria, and curtobacteria, which have

optimal growth temperatures of between 30 and 35°C), the medically relevant coryneform

bacteria grow at 37°C. It is desirable to culture specimens for coryneform bacteria in a CO2-

enriched atmosphere since some taxa, e.g.,Rothia and Arcanobacterium spp., grow much

better under those conditions. Nearly all medically relevant coryneform bacteria grow within

48 h, so that primary culture plates should not be incubated longer than that. However, if

liquid media are used (e.g., for specimens from normally sterile body sites), specimens

should be held for 5 days before the culture is declared negative. Final Gram staining and

subculture should be performed only with turbid broths.

It is recommended that urine specimens be incubated for longer than 24 h to check for the

presence of C. urealyticum but only when patients are symptomatic or have alkaline urine or

struvite crystals in their urine sediment.

C. diphtheriae

The primary plating media for the cultivation of C. diphtheriae should be SBA plus one

selective medium (e.g., cystine-tellurite blood agar [CTBA] or freshly prepared Tinsdale

medium) (33, 34). If silica gel is used as a transport medium, the desiccated swabs need to

be additionally incubated overnight in broth (supplemented with either plasma or blood),

which should then be streaked onto the primary plating medium. The plates are read after

18 to 24 h of incubation at 37°C, preferably in a 5% CO2-enriched atmosphere. Tellurite

inhibits the growth of many noncoryneform bacteria, but even a few C. diphtheriae strains

are sensitive to potassium tellurite and will therefore not grow on CTBA but may grow on

SBA. It is noteworthy that growth on CTBA and tellurite reduction are not specific for C.

diphtheriae, since many other coryneforms may also produce black (albeit smaller) colonies.

The best medium for direct culturing of C. diphtheriae is Tinsdale medium (34). However, the

limitations of Tinsdale medium are its relatively short shelf life (<4 weeks) and the necessity

to add horse serum to it. On Tinsdale plates, both tellurite reductase activity (as shown by

black colonies) and cystinase activity (as shown by a brown halo around the colonies) can be

observed. If neither CTBA nor Tinsdale medium is available, colistin-nalidixic acid blood agar

plates are recommended for the isolation of C. diphtheriae or any other coryneform

bacterium. It is necessary to pick multiple colonies from colistin-nalidixic acid blood agar

plates to rule out C. diphtheriae (first by Gram staining and then by subculturing, with

subsequent biochemical testing). Nonselective Loeffler serum slants are no longer

recommended for the primary isolation of C. diphtheriae because of overgrowth by other

bacteria.

G. vaginalis

Vaginal swabs are cultured on vaginalis agar (see chapter 17 in this Manual for the

preparation) and should be semiquantitatively streaked out with a loop. Incubation is at 35

to 37°C in a 5% CO2-enriched atmosphere or in a candle jar. Slight beta-hemolysis is

observed on human or rabbit blood-containing media but not on SBA (on which G.

vaginalis can also grow, exhibiting alpha-hemolysis). Plates may be checked for the growth

of diffuse beta-hemolytic colonies of <0.5 mm in diameter after 24 h, but very often G.

vaginalis is best observed after 48 h. Gram staining of the suspected colonies confirms the

diagnosis of G. vaginalis.

IDENTIFICATION Back to top

Basic tests available in every microbiology laboratory are of great value for the identification

of coryneform bacteria. The Gram staining morphology of the cells can exclude the

assignment to many genera and may even lead to the assignment to the correct genus (e.g.,

to the genus Corynebacterium, Turicella, orDermabacter) (Fig. 1). Morphology, size,

pigment, odor, and hemolysis of colonies are also valuable criteria in the differential

diagnosis of coryneform bacteria.

von Graevenitz and Funke (157) outlined a biochemical identification system for coryneform

bacteria which was based on previous results from the Centers for Disease Control and

Prevention’s (CDC’s) Special Bacteriology Reference Laboratory (79). This system includes

testing for catalase; fermentation or oxidation (which is best observed in semisolid cystine-

Trypticase agar medium rather than on triple sugar iron or oxidation-fermentation media,

with fermentation indicated by acid or alkali production in the entire tube and oxidation

found at the surface of the tube); motility; nitrate reduction (24 h of incubation); urea

hydrolysis (24 h of incubation); esculin hydrolysis (up to 48 h of incubation); acid production

from glucose, maltose, sucrose, mannitol, and xylose (48 h of incubation); CAMP reaction

(24 h of incubation) with a beta-hemolysin-producing strain of Staphylococcus aureus (e.g.,

strain ATCC 25923), i.e., with a positive reaction indicated by an augmentation of the effect

of S. aureus beta-hemolysin on erythrocytes, resulting in a complete hemolysis in an

arrowhead configuration (Fig. 2); and lipophilia (24 h of incubation), the test for which is

performed only for catalase-positive colonies of <0.5 mm in diameter. For the test for

lipophilia, colonies are subcultured onto ordinary SBA and onto a 0.1%- to 1%-Tween 80-

containing SBA plate. Lipophilic corynebacteria develop colonies up to 2 mm in diameter

after 24 h on Tween 80-supplemented agar. It has also been suggested that levels of growth

in brain heart infusion broth with and without supplementation of 1% (vol/vol) sterile Tween

80 be compared and that strains which grow only in the supplemented broth can be called

lipophilic. The identification protocols given in this chapter are, in principle, based on the

identification system of von Graevenitz and Funke (157) (Tables 3 and 4).



Manually performed identification panels include the API (formerly RapID) Coryne system

(bioMerieux, Marcy l’Etoile, France) and the RapID CB Plus system (Remel, Lenexa, KS),

which are widely used. The API Coryne system contains 50 taxa in its present database

(version 3.0), and comparison to an online database, APIWEB, is available

(https://apiweb.biomerieux.com). In a comprehensive multicenter study, it was found that

90.5% of the strains belonging to the taxa included were correctly identified, with additional

tests needed for correct identification for 55.1% of all strains tested (69). Reproducible

results are best obtained if the manufacturer’s recommendations for use are rigorously

followed. It was concluded that the system is a useful tool for the identification of the diverse

group of coryneform bacteria encountered in routine clinical laboratories. The RapID CB Plus

system correctly identified 80.9% of the strains to the genus and species levels and an

additional 12.2% to the genus level, but with less accurate species designations; it was also

concluded that this system may perform well under the conditions of a routine clinical

laboratory (67). An updated version of a panel for the automated Vitek system (bioMerieux)

has been described (120). However, it is always important to question critically the

identifications provided by any commercial identification system and to correlate the results

with simple basic characteristics, such as macroscopic morphology and Gram staining

results. Furthermore, it is important to note that for both commercially available manual

identification systems, the databases have not been updated since the end of the1990s, and

therefore, the recently described taxa are not covered.

For some identifications, the commercial API 50CH system (bioMerieux) has been found to

be useful. For example, when the AUX medium (usually attached to the kit for gram-negative

nonfermenters [bioMerieux]) is applied to the API 50CH system, utilization reactions which

allow the differentiation of Brevibacterium spp. or some Arthrobacter spp. can be observed

(48, 54).

A reference laboratory also uses chromatographic techniques for further characterization of

coryneform bacteria. The presence of mycolic acids and their chain lengths can be detected

by thin-layer chromatography (TLC), gas chromatography, and mass spectrometry or highperformance

liquid chromatography (27). These methods can be useful for the differentiation

of Corynebacterium spp. (mycolic acids of 22 to 36 carbon atoms) from the partially acid-fast

bacteria (mycolic acids of 30 to 78 carbon atoms) but may also provide evidence that a

coryneform bacterium is not a Corynebacterium (exceptions include C. amycolatum, C.

atypicum, and C. kroppenstedtii) if mycolic acids are not detected. The detection of the

diamino acid of the peptidoglycan by one-dimensional TLC is of certain value for determining

the genus to which a particular strain belongs (Table 1). In some cases, partial hydrolysates

of the peptidoglycan are separated by two-dimensional TLC to reveal the interpeptide bridge

of the peptidoglycan in order to distinguish between genera having the same diamino acid in

the peptide moiety. For example, some of the yellow-pigmented microbacteria and all

curtobacteria have ornithine as their diamino acids, but microbacteria have (glycine)-

ornithine as the interpeptide bridge, whereas curtobacteria possess ornithine only.

The analysis of CFAs by means of gas-liquid chromatography with the Sherlock system (MIDI

Inc., Newark, DE) is a useful method for the identification of coryneform bacteria. This

system is, in general, able to correctly identify coryneform bacteria to the genus level, but

identification to the species level is, in most cases, impossible, although the commercial

database suggests that it is possible. This is due to the very closely similar CFA profiles

obtained for coryneform bacteria belonging to the same genus (7) and because the

quantitative profiles observed strongly depend on the incubation conditions. When a

laboratory creates an individual database based on its own entries, species identification

becomes more likely (K. A. Bernard, unpublished observation). The mycolic acids of some

corynebacteria (e.g., C. auris) are cleaved at the temperature (300°C) produced in the

injection port of the system, resulting in peaks being misidentified as fatty acids, e.g.,

C17:1v6c to Cv9c, by the Sherlock system (7, 57).

Molecular genetics-based identification systems for coryneform bacteria have been outlined

in recent years. Restriction fragment length polymorphism analysis of the partly amplified

and digested 16S rRNA gene has been demonstrated to be of use for the identification of

species, e.g., within the genus Corynebacterium (154). Some corynebacteria may also be

identified to the species level by examination of the length of the 16S-23S rRNA intergenic

spacer region (4). A very useful approach for the identification of true corynebacteria is the

sequencing of a 434- to 452-bp fragment of the rpoB gene (using primers designated

C2700F and C3130R), since this particular region of the gene displays a high degree of

polymorphism within the genusCorynebacterium (82, 83). A divergence of >5% within this

particular part of the rpoB genes of two compared strains suggests that they belong to two

different species. For more definitive taxonomic investigations of coryneforms, and in cases

of the growth of coryneform bacteria from difficult-to-obtain clinical material (132), fulllength

16S rRNA gene sequencing might be indicated. Determination of the complete 16S

rRNA gene sequence is a rational approach for identifying corynebacteria, since most

established species exhibit 3% or greater divergence, except for C. afermentans, C. coyleae,

C. mucifaciens (<2% divergent); C. aurimucosum, C. minutissimum, and C.

singulare (<2%); C. sundsvallense and C. thomssenii (<1.5%); C. ulcerans and C.

pseudotuberculosis (<1% divergent from each other and both <2% divergent from C.

diphtheriae); C. propinquum and C. pseudodiphtheriticum (<2%); C. xerosis, C.

freneyi, and C. hansenii (<2%); and C. macginleyiand C. accolens (<2%).

In very few selected cases (i.e., if the divergence of the 16S rRNA genes is ≤1.3%),

quantitative DNA-DNA hybridizations might be necessary, with sequencing of the

complete rpoB gene recently being suggested as a substitute for that approach (2). Because

of the ever-growing number of coryneform taxa encountered in clinical specimens, it has

become difficult to readily differentiate these taxa by biochemical means alone, so

sequencing studies are likely to replace some of the biochemical testing in the near future. It

is emphasized that unidentifiable, clinically significant coryneform bacteria should be sent to

an established reference laboratory experienced in corynebacterial identification for

characterization which includes sequence-based analyses.

IDENTIFICATION: DESCRIPTIONS OF GENERA AND

SPECIES Back to top

Genus Corynebacterium

C. accolens

C. accolens (103) is found in specimens from the eyes, ears, nose, and oropharynx.

Endocarditis of native aortic and mitral valves due to this agent has been described. Colonies

are, as for all other lipophilic corynebacteria, convex, smooth, and <0.5 mm in diameter on

SBA. C. accolens strains had initially been described to exhibit satellitism in the vicinity of S.

aureus strains, attributable to its lipophilism (for the method recommended to demonstrate

lipophilism, see above under “Identification”). C. accolens has a variable pyrazinamidase

reaction but is negative for alkaline phosphatase, which differentiates it from the

morphologically and biochemically closely related species C. tuberculostearicum (Table 3).

The API Coryne and RapID CB Plus systems correctly identify C. accolens (67, 69). C.

accolens strains are susceptible to a broad spectrum of antibiotics.



C. afermentans subsp. afermentans

C. afermentans subsp. afermentans (125) is part of the normal human skin biota and has so

far been isolated mainly from blood cultures. Colonies are whitish, convex with regular

edges, creamy, and about 1 to 1.5 mm in diameter after 24 h of incubation. C.

afermentans subsp. afermentans has an oxidative metabolism. The API Coryne system

provides the numerical code 2100004 for this species. About 60% of all strains of this taxon

are CAMP reaction positive. C. afermentans subsp. afermentans can be differentiated from C.

auris and T. otitidis (both of which give the same API Coryne numerical code) by the

consistency of its colonies (C. auris is slightly adherent to agar) and morphology upon Gram

staining (T. otitidis has longer cells). By chemo taxonomic means, both C.

afermentans subspecies and C. auris contain mycolates, whereas T. otitidis lacks them, but

this technique is not applicable in routine clinical laboratories. C.

afermentans subsp. afermentans is generally susceptible to β-lactam antibiotics, but

resistance to erythromycin, clarithromycin, azithromycin, and clindamycin have been

reported (42).

C. afermentans subsp. lipophilum

Strains belonging to the species C. afermentans subsp. lipophilum (125) have been isolated

mainly from blood cultures but also from superficial wounds. Colonies are, typically for

lipophilic corynebacteria, convex, smooth, and <0.5 mm in diameter after 24 h. C.

afermentans subsp. lipophilum has an oxidative metabolism and does not produce acid from

any of the carbohydrates usually tested (Table 3). It is the only species of nonfermenting,

lipophilic corynebacteria which may exhibit a positive CAMP reaction. C.

afermentans subsp.lipophilum is not included in the API Coryne database. The numerical

profile observed for the species is 2100004 and so by that method cannot be discerned from

more robustly growing C. afermentans subsp. afermentans, C. auris, or T. otitidis, which

have the same code as well. Strains are usually susceptible to β-lactam antibiotics.

C. amycolatum

C. amycolatum is part of the normal human skin biota but was not recovered from throat

swabs from healthy persons (158). C. amycolatum is the most frequently

encountered Corynebacterium species in human clinical material (56). It is also the most

frequently isolated nonlipophilic Corynebacterium in dairy cows with mastitis (80). C.

amycolatum strains are nearly always multidrug resistant (68). Colonies are very typically

dry, waxy, and grayish white with irregular edges and are 1 to 2 mm in diameter after 24 h

of incubation (Fig. 3a). C. amycolatum actually has a fermentative metabolism, but when

cystine-Trypticase agar media are used for the observation of acid production from

carbohydrates, C. amycolatum appears to resemble an oxidizer (i.e., to have main acid

production at the surface of the medium). Strains of C. amycolatum are remarkable for their

variability in basic biochemical reactions (Table 3) and are often misidentified as the

biochemically similar species C. xerosis, C. striatum, or C. minutissimum (56, 161, 170).

These species can be differentiated by the following means. C. amycolatum and C.

minutissimum do not grow at 20°C, but C. xerosis and C. striatum do; in addition, C.

xerosis does not ferment glucose at 42°C, whereas the other three species do; and C.

minutissimum and C. striatum produce alkali from formate but C. amycolatum and C.

xerosis do not (161). When tested on Mueller-Hinton agar supplemented with 5% sheep

blood, nearly all C. amycolatum strains were resistant to the vibriocidal compound O/129

(150-μg disks) (Oxoid, Basingstoke, United Kingdom), as indicated by there being no zone of

inhibition around the disk (56). In contrast, only 4% of all C. amycolatumstrains were

resistant to O/129 when tested on Mueller-Hinton agar with 5% horse blood (80). The API

Coryne system identifies this species well, but in every case additional reactions must be

carried out in order to confirm the identification of C. amycolatum (69). All C.

amycolatum strains produce propionic acid as the major end product of glucose metabolism.

In contrast to many other corynebacteria, C. amycolatum exhibits only weak or no leucine

arylamidase activity. The identification may also be suggested by the absence of mycolic

acids. In addition, it may be shown that acyl phosphatidylglycerol is a major phospholipid

in C. amycolatum, unlike in other Corynebacterium spp., in which other phospholipids are

predominant (153).



C. appendicis

The one strain of C. appendicis described in the literature was isolated from a patient with

appendicitis accompanied with abscess formation (165). This lipophilic species contains large

amounts of TBSA (up to 50% of all CFAs), which is not seen in any

other Corynebacterium species. It is differentiated from CDC coryneform group F-1 bacteria

by a positive alkaline phosphatase reaction but negative reactions for nitrate reduction and

sucrose fermentation.

C. argentoratense

C. argentoratense (127) has been isolated from the human throat and was once recovered

from a blood culture (8). Colonies are cream colored, nonhemolytic, slightly rough, and 2

mm in diameter after 48 h of incubation. Phenotypically, C. argentoratense may appear to be

very similar to (rare) ribose-negative strains ofC. coyleae. However, glucose fermentation

by C. argentoratense is quite rapid, compared to that of the slowly fermenting species C.

coyleae. In addition, CAMP reaction-negative C. argentoratense produces propionic acid as a

fermentation product, but CAMP reaction-positive C. coyleae does not (8). C.

argentoratense is the only medically relevant Corynebacterium species expressing α-

chymotrypsin activity, which can be observed in the API ZYM (bioMerieux) system; however,

the blood culture isolate was not observed to produce that enzyme (8). Although C.

argentoratense is phylogenetically closely related to C. diphtheriae, it does not harbor

the toxgene, coding for the diphtheria toxin.

C. atypicum

Although C. atypicum clearly belongs to the genus Corynebacterium, corynomycolic acids, as

in C. amycolatumand C. kroppenstedtii, are not detected (77). C. atypicum is not lipophilic

but shows only pinpoint-sized colonies after 48 h of incubation. It is the only medically

relevant Corynebacterium not expressing pyrazinamidase but expressing β- glucuronidase

activity.

C. aurimucosum

The initial description of C. aurimucosum was based on a single strain which exhibited

slightly yellow and sticky colonies on 5% SBA plates but had colorless and slimy colonies on

Trypticase soy agar without blood (166). Biochemically, this particular C. aurimucosum strain

was similar to C. minutissimum. The number of C. aurimucosum strains was significantly

enhanced when it was demonstrated that some former CDC coryneform group 4 bacteria

actually belong to C. aurimucosum (26). It is important to note that strains of C.

aurimucosumexhibit a grayish-black pigment which is not seen in any other

true Corynebacterium. Strains originally designated “Corynebacterium nigricans” were later

shown to be C. aurimucosum (26). Some strains of R. dentocariosa can also exhibit a

charcoal-black pigment (26); these strains are differentiated from C. aurimucosum by being

constantly nitrate reductase positive, having a possibly negative catalase reaction, and

having branched-chained CFAs, as opposed to straight-chain-type CFAs in C. aurimucosum.

API Coryne codes for pigmented C. aurimucosum strains include 0000125, 2000125, and

2100327. Phylogenetically, C. aurimucosum is closely related (>98.8% identity) to both C.

singulare and C. minutissimum by 16S rRNA gene sequencing but can readily be discerned

by partial rpoB gene sequence analysis.

C. auris

C. auris (57) has almost exclusively been isolated from the ear region. Colonies are dryish

and slightly adherent to agar but do not penetrate agar; they become slightly yellowish with

time, and they have diameters ranging from 1 to 2 mm after 48 h of incubation. C.

auris does not produce acid from any carbohydrates usually tested. All C. auris strains are

strongly CAMP test positive. The API Coryne system provides the numerical code 2100004

for this species. Abundant degradation products of mycolic acids are indirectly observed

when CFA patterns are determined with the Sherlock system (57). It is noteworthy that the

MICs of β-lactam antibiotics for C. auris strains are elevated, but the molecular mechanism

for this is not known at present (68).

C. bovis

Occasionally, human infections have been attributed to the lipophilic bovine species C. bovis.

Characterization of lipophilic corynebacteria solely on the basis of the results of phenotypic

tests, in the absence of modern polyphasic methods or identification schemes, was probably

incorrect (Table 3). This species had not been definitively recovered for many years from

human clinical material (72). Recently, an oxidase-positive, human blood culture isolate of C.

bovis was identified based on a polyphasic approach, with an API Coryne code of 0101104

(8), as was an isolate from a prosthetic joint infection (1).

C. canis

This nonlipophilic Corynebacterium was isolated from a wound infection after a dog bite (50).

It has some unusual microscopic features in that it exhibits very filamentous rods (>15 μm

in length), and some cells even show branching. C. canis is esculinase and α-glucosidase

positive and is the only Corynebacterium expressing trypsin activity.

C. confusum

C. confusum has been isolated from patients with foot infections, a blood culture (61), and a

breast abscess (8). Colonies are whitish, glistening, convex, creamy, and up to 1.5 mm in

diameter after 48 h. Acid from glucose is produced only very weakly, becoming visible in the

API Coryne or the API 50CH gallery only after 48 to 72 h. Weak growth under anaerobic

conditions corresponds to slow fermentative acid production. It is advisable to incubate the

API Coryne system after 24 h for another day in those cases in which the results for acid

production are ambiguous (i.e., with only a slight change in the color of the indicator). After

48 h of incubation, the API Coryne system provides the numerical code 3100304 for this

species; the breast abscess strain had a code of 3100104. Interestingly, the breast abscess

strain was also CAMP reaction positive, making it potentially more difficult to discern from C.

coyleae isolates (8). C. confusum is correctly identified by the RapID CB Plus system (67). If

glucose fermentation is judged to be negative, C. confusum strains can be misidentified as C.

propinquum. However, in contrast to that species, C. confusum does not hydrolyze tyrosine

and contains small amounts of TBSA (1 to 3%), whereas C. propinquum hydrolyzes tyrosine

but does not contain TBSA. C. confusum is differentiated from C. coyleae and C.

argentoratense by its ability to reduce nitrate.

C. coyleae

C. coyleae (65) has been isolated mainly from cultures of blood and other normally sterile

body fluids, but it may also be recovered from a variety of sterile body fluids, abscesses, and

urogenital specimens (8, 40). Colonies are whitish and slightly glistening, with entire edges,

and are about 1 mm in diameter after 24 h. The consistency of the colonies is either creamy

or sticky. A slow fermentative acid production from glucose and a strongly positive CAMP

reaction are the most significant phenotypic characteristics. C. coyleae is positive for cystine

arylamidase, which is not observed for many other corynebacteria. Various API Coryne

numerical codes have been observed, especially 2100304 and 6100304. C. coyleae is always

positive for ribose fermentation, whereas the biochemically similar species C.

argentoratense varies in this reaction. The API Coryne database lists only 6% of glucosefermenting

C. coyleae strains, and therefore, when applying this commercial identification

system, the clinical microbiologist may not receive a correct identification (69). However, the

two numerical profiles given above combined with a positive CAMP reaction are highly

indicative of C. coyleae. This species is correctly identified by the RapID CB Plus system (67).

Macrolide resistance has been reported (42).

CDC Group F-1 Bacteria

The lipophilic CDC group F-1 bacteria (128) have not been given a species name. Although

genetically distinct, no distinguishing phenotypic markers which clearly allow their separation

from other defined Corynebacteriumspp. have been found. The characteristics of the CDC

group F-1 bacteria are consistent with the definition of the genus Corynebacterium in all

respects. Of note is the negative alkaline phosphatase reaction (Table 3). CDC group F-1

strains are usually susceptible to penicillin but are often resistant to macrolides.

C. diphtheriae

In 2003, the complete genome sequence of a Corynebacterium diphtheriae strain

representative of the diphtheria outbreak in the former Soviet Union states in the 1990s was

determined (17). The genome consists of a single circular chromosome of 2,488,635 bp with

no plasmids. A complete set of enzymes for the glycolysis, gluconeogenesis, and pentosephosphate

pathways is present, as are all the de novo amino acid biosynthesis pathways.

Fimbrial and fimbria-related genes, sialidase (neuraminidase) genes, and iron uptake

systems have been detected as pathogenicity factors.

C. diphtheriae is commonly divided into four biotypes, gravis, mitis, belfanti, and

intermedius; biotype differentiation is recommended by the WHO (33, 34), although biotypes

cannot be assigned separate subspecies status, and biotyping is not satisfactory for

epidemiologic tracking. Initially, these biotypes were defined by differences in colony

morphology and biochemical reactions (Table 3). However, only C. diphtheriaebiotype

intermedius can be identified on the basis of colonial morphology (small, gray, or translucent

lipophilic colonies) as well as positivity for dextrin fermentation. Other C.

diphtheriae biotypes produce larger (up to 2 mm after 24 h) white or opaque colonies (Fig.

3b) which are indistinguishable from each other. The lipophilic C. diphtheriae biotype

intermedius occurs only rarely in clinical infections, and C. diphtheriae biotype belfanti

strains rarely express diphtheria toxin.

Presumptive identification of C. diphtheriae (as well as of C. pseudotuberculosis and C.

ulcerans) may be made by testing suspicious gram-positive rods for the presence of

cystinase (as detected by using freshly prepared Tinsdale medium or diagnostic tablets

[Rosco, Taastrup, Denmark]) and the absence of pyrazinamidase (diagnostic tablets are

available from Key Scientific Products, Stamford, TX). The API Coryne system identifiesC.

diphtheriae strains, but additional tests are needed for the differentiation of C.

diphtheriae biotype mitis, C. diphtheriae biotype belfanti, and C. diphtheriae biotype

intermedius (69). Usually, C. diphtheriae strains do not ferment sucrose, but in Brazil,

sucrose-positive strains have been described. Large amounts of propionic acid are produced

as the end product of glucose metabolism. C. diphtheriae strains are distinct from all other

coryneform bacteria (except C. pseudotuberculosis and C. ulcerans) in their CFA patterns by

the presence of a large volume of C16:1v7c (7).

Diphtheria Toxin Testing

It is recommended that at least 10 colonies of C. diphtheriae and related species be tested

for diphtheria toxin by the Elek method, modified as described by Engler et al. (37), in a

laboratory with skill in performing the test and in interpreting the test results. The modified

Elek method as described by the WHO Diphtheria Reference Unit was initially used to

characterize strains from the epidemic in the 1990s in Russia and Ukraine and was found to

be faster and less technically problematic than the original version. Antitoxin from various

suppliers (Berna Biotech AG [Switzerland], Mikrogen [Russia], Biomed [Russia], Pasteur

Merieux/Aventis Pasteur [France], BulBio-NCIPD [Bulgaria], Instituto Butantan [Brazil], or

Refik Saydam National Hygiene Centre [Turkey]) (102) applied to blank filter disks at 10

IU/disk have been successfully used with the modified Elek test.

PCR-based methods for the detection of the diphtheria toxin gene (tox) in isolated bacteria

have been developed and validated (32, 78, 107). Conventional PCR detection of the

regulatory dtxR gene has been evaluated (111). Detection of the tox gene using a real-time

platform has been described (15, 99, 136). Primers targeting the C. diphtheriae tox gene

for C. ulcerans strains perform better with modifications (15, 135,136). tox PCR assays

applied directly to clinical specimens are acceptable, particularly because isolation is not

always possible for patients already receiving antibiotics. However, a PCR-positive patient

from whom bacteria are not isolated or who lacks a histopathologic diagnosis and an

epidemiologic linkage to a patient with a laboratory-confirmed case of diphtheria should be

classified as a “probable case” of diphtheria, since to date there are insufficient data to

conclude that a PCR-positive result always infers diphtheria. Also, detection of the toxin gene

in samples by PCR cannot automatically be attributed to one species because C.

diphtheriae as well as C. ulcerans and C. pseudotuberculosis may harbor the bacteriophage

which carries the diphtheria toxin gene. Furthermore, tox-containing, nontoxigenic isolates

have been described and characterized further. Difficulties in identifying C. diphtheriae and in

correctly performing toxigenicity tests have recently been demonstrated by an external

quality control program observing 23 national diphtheria reference centers in Europe; in

these centers, 21% of specimens were misidentified and 13% of toxigenicity reports were

determined to be unacceptable (102).

Nontoxigenic strains of C. diphtheriae, i.e., those which do not express toxin in the Elek test

or those which lack a detectable diphtheria toxin gene by PCR, have also caused serious

disease, such as cases or outbreaks of skin disease, endocarditis, and occasional mortality

among homeless people, alcoholics, and intravenous drug abusers (45, 116, 131). For

nontoxigenic C. diphtheriae strains circulating in the United Kingdom, it has been shown that

the diphtheria toxin repressor (dtxR) genes are functional, so that if these strains are

lysogenized by a bacteriophage, they could represent a reservoir for toxigenic C.

diphtheriae (29).

Antibiotic treatment, with antibiotics of choice being penicillins or macrolides, is required to

eliminate C. diphtheriae and prevent its spread; however, it is not a substitute for antitoxin

prevention. Sporadic isolates ofC. diphtheriae resistant to erythromycin or rifampin have

been reported.

C. durum

C. durum (126) was originally described as being exclusively isolated from respiratory tract

specimens. Well- characterized isolates have now been recovered from additional sites,

including the gingiva, blood cultures, and abscesses (115). C. durum strains were originally

isolated after 2 to 3 days from nonselective buffered charcoal-yeast extract plates inoculated

with sputa or bronchial washings, but isolates grow well on other standard laboratory

media. C. durum is the most frequent Corynebacterium isolated from throat swabs of healthy

persons (158). Its pathogenic potential remains unclear. C. durum is a peculiar nonlipophilic

organism that forms colonies of only 0.5 to 1 mm in diameter after aerobic incubation for 72

h. The original description of this bacterium cited beige and rough colonies with convolutions,

an irregular margin, and strong adherence to agar if grown under aerobic conditions (126).

However, strains were later described to be sometimes smoother and not necessarily

adherent to agar (115). Gram staining of aerobic cultures shows long and filamentous rods,

with occasional “bulges,” but true C. durum isolates do not have C. matruchotii-like “whip

handles” (Fig. 1e and f). Long forms are not otherwise found among

other Corynebacterium species (except C. canis), nor are they observed for C. durum when

cells are grown in a 10% CO2-enriched atmosphere (126). Strains grow only weakly under

anaerobic conditions. They always reduce nitrate, and some may exhibit weak and delayed

urease and esculinase activities. The majority (but not all) of C. durum strains ferment

mannitol, which is another very unusual feature for true corynebacteria (Table 3). API

Coryne codes observed for C. durum include 3000135, 3001135, 3040135, 3400115,

3400135, 3400305, 3400325, and 3400335 (126), as well as 3040325, 3040335, 3440335,

and 3441335 (115). This suggests that most strains are negative for alkaline phosphatase,

and all appear to be negative for pyrrolidonyl arylamidase. Only a small number of C.

durum strains have been tested with the RapID CB Plus system, and all were correctly

identified (67). It is most likely that some strains identified as C. matruchotii in the past may

actually have been C. durum strains and that differentiation can be difficult if phenotypic

methods alone are used. Both species produce propionic acid as a fermentation product

(6). C. durum usually ferments galactose and very often mannitol, whereas C. matruchotii is

usually negative for those sugars. The C. matruchotii type strain exhibits α-glucosidase

activity, which is not observed in C. durum (126). It has been shown that some C.

durum strains also express β-galactosidase activity and ferment ribose (158).

C. falsenii

C. falsenii strains (139) have been isolated mainly from sterile body fluids. Colonies are

whitish, glistening, and smooth, with entire edges, and are 1 to 2 mm in diameter after 24 h.

After 72 h, most strains described to date exhibit a yellowish pigment which becomes even

more intense after 120 h. The most characteristic biochemical features of C. falsenii are a

slow but fermentative acid production from glucose, a weak pyrazinamidase reaction, and a

weak urease activity which becomes visible in either Christensen’s urea broth or the API

Coryne system after overnight incubation only. API Coryne codes observed for C.

falsenii have been 2101104 and 2101304 (8, 139).

C. freiburgense

This nonlipophilic species has probably been transmitted to a human by a dog bite (53). The

5-day-old colonies exhibit a very peculiar “spoked-wheel” macroscopic morphology not

observed in other true corynebacteria but observed in some R. dentocariosa strains. Colonies

are also strongly adherent to blood agar. Distinct biochemical features are the lack of

pyrazinamidase activity and a positive β- galactosidase reaction.

C. freneyi

This species was initially outlined on the basis of a study of three strains (117) which came

from skin-related material. There is now evidence that C. freneyi is also isolated from

genitourinary specimens (52) and bacteremic patients (5). C. freneyi is phylogenetically

closely related to C. xerosis. Colonies are typically wrinkled, whitish, dry, and rough, have

irregular edges, and are 0.5 to 1 mm in diameter after 48 h incubation, but C. freneyi strains

are nonlipophilic. The basic biochemical profile (Table 3) is similar to that of C. xerosis. AllC.

freneyi strains studied so far exhibit α-glucosidase activity, which is not frequently observed

in otherCorynebacterium species (very few C. amycolatum and all C. xerosis strains express

this enzyme). C. freneyi can be furthermore differentiated from C. xerosis by glucose

fermentation at 42°C and growth at 20°C, whereas C. xerosis is negative for these two

reactions. This species is also closely related to C. hansenii.

C. glucuronolyticum

C. seminale is a later synonym of C. glucuronolyticum (28, 47). This species is probably part

of the normal genitourinary biota of males, while its presence in females is uncertain.

Recoveries from blood cultures have been documented (8). Colonies are whitish yellow,

convex, and creamy, and they are 1 to 1.5 mm in diameter after 24 h. The fermentative

species C. glucuronolyticum is remarkable for its variability in basic biochemical reactions

(Table 3). It is the only medically relevant, large-colony Corynebacterium species exhibiting

β-glucuronidase activity. When urease activity is present, it is abundant in Christensen’s urea

broth, becoming positive after only 5 min of incubation at room temperature (47). C.

glucuronolyticum is also one of the very few corynebacteria which are able to hydrolyze

esculin. All C. glucuronolyticum strains are CAMP reaction positive (Fig. 2). The API Coryne

strip identifies C. glucuronolyticum well but not strains which are alkaline phosphatase

positive (69), although profiles obtained from human strains may differ from those of animal

isolates (28). Propionic acid is one of the major end products of glucose metabolism. C.

glucuronolyticumstrains are often tetracycline resistant and may also exhibit resistance to

macrolides and lincosamides (68). 16S rRNA gene sequences derived from fluids of patients

with prostatitis have been found to be homologous with sequences derived for this species,

indicating that C. glucuronolyticum might be involved in selected cases of prostatitis (146).

C. hansenii

This nonlipophilic species exhibits yellow and dry colonies (119). Biochemical identification

reactions are similar to those of C. freneyi and C. xerosis. However, C. hansenii can be

distinguished from these species by being negative for alkaline phosphatase and α-

glucosidase reactions. These three species are not well discerned by either 16S rRNA or

partial rpoB gene sequencing but were found to represent different taxa by DNA-DNA

hybridization (119).

C. imitans

C. imitans was originally isolated from a nasopharyngeal specimen of a child suspected of

having throat diphtheria, as well as from three adult contacts (49). This was the first welldocumented

case of the person-to-person transmission of a Corynebacterium species other

than C. diphtheriae in a nonhospital setting. Additional strains of C. imitans have been

recovered from blood cultures (8). Colonies are creamy, whitish gray, and glistening, with

entire edges, and they are 1 to 2 mm in diameter. The strain does not produce a brown halo

on Tinsdale medium but is tellurite reductase positive. Interestingly, it is positive for polar

bodies by Neisser staining. Pyrazinamidase activity is only weak, as is fermentation of

sucrose, which may lead to an initial misidentification as an atypical C. diphtheriae strain. It

is possible that C. imitans may have been misidentified as C. minutissimum in the past, since

the biochemical reactions of both taxa are similar (Table 3). However, C. imitans is CAMP

reaction positive and does not hydrolyze tyrosine, whereas the opposite reactions are

observed for C. minutissimum. The API Coryne system provided the numerical codes

1100325, 2100324, and 3100325 for C. imitans, indicating a negative α-glucosidase

reaction, whereas all C. diphtheriaestrains express this enzyme. C. imitans strains do not

produce propionic acid as a fermentation product, unlikeC. diphtheriae (8), and the CFA

composition profiles for the species qualitatively differ, as C. diphtheriae and closely related

species have a unique pattern among Corynebacterium species. Diphtheria toxin assessment

by Elek testing or by assaying for the tox gene by PCR are negative for C. imitans strains

(8, 49). C. imitans is resistant to O/129, while C. diphtheriae is not.

C. jeikeium

C. jeikeium is a frequently encountered Corynebacterium in clinical specimens (72).

Nosocomial transmission has been described. The complete genome sequence of a C.

jeikeium strain has been determined (147, 148), indicating that the lipophilic phenotype

of C. jeikeium originates from the absence of fatty acid synthase. C. jeikeium is often

resistant to multiple antibiotics (including penicillin and gentamicin) (42, 75), but this cannot

be used as a taxonomic characteristic because the phenotypically closely related C.

tuberculostearicum may also demonstrate multidrug resistance. Quantitative DNA-DNA

hybridization experiments have shown that C. jeikeium includes two genomospecies for

which penicillin and gentamicin MICs are low, but as they could not otherwise be

differentiated phenotypically from the resistant C. jeikeium strains, they were not proposed

as independent species (124). Colonies of C. jeikeium are tiny, low, entire, and grayish

white. C. jeikeium is a strict aerobe which may oxidatively produce acid from glucose and

sometimes from maltose but not from fructose (C. tuberculostearicum is positive for acid

production from fructose). The RapID CB Plus system correctly identifies C. jeikeium, as does

the API Coryne system if ancillary tests are used (67, 69).

C. kroppenstedtii

C. kroppenstedtii (23), a rather rarely recovered species, was originally recovered from the

sputum of a patient with pulmonary disease. Additional strains have been isolated from lung

biopsy specimens, sputum, a breast abscess, and patients with granulomatous lobular

mastitis (8, 110). Colonies are grayish, translucent, slightly dry, and less than 0.5 mm in

diameter after 24 h of incubation at 37°C. C. kroppenstedtii is lipophilic and is one of the few

medically relevant Corynebacterium species exhibiting esculinase activity. Other biochemical

characteristics are given in Table 3. API Codes of C. kroppenstedtii include 0101104,

2040104, and 2040105 (8). It can be separated from C. durum, C. matruchotii, and C.

glucuronolyticum by colony and Gram stain morphologies and from C. glucuronolyticum by

its negative CAMP reaction. The determination of the whole-genome sequence revealed that

lipophilism is the dominant feature involved in the pathogenicity of C. kroppenstedtii (149).

C. lipophiloflavum

Initially, C. lipophiloflavum (55) was represented by a single strain which had been isolated

from vaginal discharge from a patient with BV. Additional strains have now been described

from blood cultures (G. Funke, R. Frodl, E. Falsen, C. Sproer, H.-P. Klenk, and S. Stenger,

submitted for publication). This species contains lipophilic strains, but the majority of strains

are nonlipophilic (Funke et al., submitted). The lipophilic strains have the same biochemical

screening pattern as C. urealyticum, except that they exhibit a strong yellow pigment and

weaker urease activity and slowly produce acid from glucose (Table 3). In contrast to

most C. urealyticum strains, the C. lipophiloflavum strains isolated were not multidrug

resistant. The nonlipophilic strains have a biochemical profile similar to that of C. falsenii but

do not ferment galactose and trehalose (Funke et al., submitted).

C. macginleyi

C. macginleyi (128) has almost exclusively been isolated from eye specimens, whether from

diseased (62) or healthy conjunctiva. Colonies are typical for lipophilic corynebacteria (see

above). When grown on Tween 80-SBA plates (better growth is usually found on plates

supplemented with 0.1% Tween 80 than on those supplemented with 1.0% Tween 80),

some C. macginleyi strains exhibit a rose pigment which is not seen for any other

lipophilic Corynebacterium species. C. macginleyi is one of the very

few Corynebacterium species not expressing pyrazinamidase activity (Table 3). Most strains

ferment mannitol, while the majority of other corynebacteria are unable to do so. The API

Coryne system correctly identifies C. macginleyi (69). Strains belonging to this species are

susceptible to a broad spectrum of antibiotics (62), but high-level fluoroquinolone resistance

has been reported (35).

C. massiliense

This species was isolated from synovial fluid (96). It does not produce acid from any of the

carbohydrates tested in the differentiation of true corynebacteria (Table 3), but its cell wall

contains TBSA.

C. mastitidis-Like Organism

Bacteria recovered from ocular specimens (patients with cataracts, diabetic retinopathy, or

dry eyes) were by 16S rRNA gene sequencing found to be closest (98.2% identity) to C.

mastitidis, which to date has otherwise been found to cause sheep mastitis. In contrast to C

macginleyi, these strains are susceptible to the fluoroquinolones, but no other information

about these strains is extant (35).

C. matruchotii

C. matruchotii is thought to be a natural inhabitant of the oral cavity, particularly on calculus

and plaque deposits, and so has been much studied by oral microbiologists (115). Otherwise,

it is a very rare human pathogen. Microcolonies appear flat, filamentous, and spider-like, but

macrocolonies have a variable appearance (19). C. matruchotii demonstrates a very unusual

appearance by Gram staining in that so-called whip handles (i.e., filamentous bacteria with a

single short bacillus adjacent to the end of the filament that creates the illusion of a whip)

are observed (Fig. 1f). This microscopic presentation is consistent even when isolates that

have been preserved for many years in a culture collection are stained. It has been

demonstrated that heterogeneity exists among C. matruchotii strains obtained from

international culture collections and that some strains represented were misidentified C.

durum isolates (115). C. matruchotiistrains are consistently negative for galactose,

whereas C. durum strains can be positive. The API Coryne system database does not

contain C. matruchotii; the numerical codes observed for C. matruchotii include 7000325,

7010325, and 7050325.

C. matruchotii-Like Strain

A C. matruchotii-like strain is represented by a single strain, ATCC 43833 (115). It was

deposited in the ATCC as C. matruchotii, but it is a distinct species, as revealed by dot blot

hybridization and 16S rRNA gene sequencing data (GenBank accession no. AF260434).

Colonies are the size of pinpoints to 0.1 mm in diameter and grayish white, with a smooth,

nonadherent texture. Biochemical screening reactions are similar to those of C.

minutissimum except that strain ATCC 43833 exhibits esculinase activity in the API Coryne

system, with an API Coryne code of 2140325.

C. minutissimum

C. minutissimum is part of the normal human skin biota, and its historical association with

erythrasma is highly questionable (72). Colonies of C. minutissimum are whitish gray, shiny,

moist, convex, and circular; they have entire edges and are about 1 to 1.5 mm in diameter

after 24 h. Most of the colonies are creamy, but some may also have a sticky consistency. C.

minutissimum strains have a fermentative metabolism and produce acid from sucrose

variably. Very few C. minutissimum strains are also able to produce acid from mannitol. The

API Coryne system identifies C. minutissimum, with additional tests being necessary for most

strains (69). Many C. minutissimum strains are pyrrolidonyl arylamidase positive. C.

minutissimum strains exhibit DNase activity, nearly all strains hydrolyze tyrosine, and a very

few strains exhibit a positive CAMP reaction. Lactic and succinic acids are major end products

of glucose metabolism. Some isolates possess TBSA in their cell membranes. Nearly all C.

minutissimum strains are susceptible to O/129 (150-μg disk); i.e., they exhibit an inhibition

zone around the disk (usually between 20 and 35 mm in diameter). Phylogenetically, C.

minutissimum is closely related (>98.8% identity) to both C. singulare and C.

aurimucosum by 16S rRNA gene sequencing, but these are readily discerned by

partial rpoB gene sequence analysis.

C. mucifaciens

C. mucifaciens (58) has been isolated mainly from blood cultures and other sterile body

fluids but has also been recovered from abscesses, soft tissue, and dialysate (8). Colonies

are very distinct because they are slightly to overtly yellow and very mucoid (Fig. 3c) (very

few strains are not mucoid [G. Funke, unpublished observation]). C. mucifaciens is the only

presently known Corynebacterium species exhibiting such mucoid colonies; this characteristic

strongly reminds bacteriologists of Rhodococcus equi colonies. An extracellular substance

(probably polysaccharides) causing connective filaments between the cells has been

demonstrated as the ultrastructural correlate of the mucoid colonies. Colonies are about 1 to

1.5 mm after 24 h of incubation and have entire edges. They appear less mucoid after

extended incubation for 96 h. C. mucifaciens has an oxidative metabolism. It consistently

produces acid from glucose, but acid production from sucrose is variable. The API Coryne

numerical codes 2000004, 2000104, 2000105, 2100104, 2100105, 6000004, 6100104, and

6100105 have been observed for C. mucifaciens, suggesting that occasionally glucose

oxidation may be too slow to be observed by that method. C. mucifaciens is enzymatically

less active than R. equi, which exhibits α- and β-glucosidase activities not observed for C.

mucifaciens. In addition, C. mucifaciens produces acid from fructose and may produce acid

from glycerol and mannose, but acid production from these sugars is not seen in R.

equi strains. TBSA can be detected in amounts of 1 to 2% of the total CFAs. β-Lactam

antibiotics and aminoglycosides show very good activities against C. mucifaciens.

C. pilbarense

This nonlipophilic species was isolated from an ankle aspirate of a male thought to be

suffering from gout (3). It can be differentiated from C. striatum and C. simulans by being

nitrate reduction negative and from C. minutissimum by not fermenting maltose.

Phylogenetically, it is closest (98.7 to 99.0% identity) to C. ureicelerivorans, C. coyleae, C

afermentans, and C. mucifaciens but was distinguished from those species by DNA-DNA

hybridization and biochemically (Table 3) (3).

C. propinquum

C. propinquum is the closest phylogenetic relative of C. pseudodiphtheriticum (108, 133) and

shares the same niche (i.e., the oropharynx) as C. pseudodiphtheriticum. Colonies are

whitish and somewhat dryish, with entire edges, and are 1 to 2 mm in diameter after 24 h of

incubation. This species reduces nitrate and hydrolyzes tyrosine but does not hydrolyze urea

(Table 3). The API Coryne system and the RapID CB Plus system correctly identify C.

propinquum strains (67, 69).

C. pseudodiphtheriticum

C. pseudodiphtheriticum is part of the normal oropharyngeal biota. As described in Table 2,

this species has been well documented to cause pneumonia in various patient populations.

Colonies are whitish and slightly dry, with entire edges, and they are 1 to 2 mm in diameter

after 48 h of incubation. This nonfermenting species reduces nitrate and hydrolyzes urea but

does not produce acid from any of the commonly tested carbohydrates (Table 3). Some

strains hydrolyze tyrosine. The API Coryne and the RapID CB Plus systems correctly

identify C. pseudodiphtheriticum strains (67, 69). Imperfectly cleaved mycolic acids coelute

with CFAs (7). C. pseudodiphtheriticum strains are susceptible to β-lactam antibiotics, but

resistance to macrolides and lincosamides has been observed.

C. pseudotuberculosis

C. pseudotuberculosis is phylogenetically closest to C. ulcerans and to C.

diphtheriae (108, 133); like those species, it may harbor the diphtheria toxin gene and

produce propionic acid as a fermentation product, and its cell wall contains large amounts of

the CFA C16:1v7c (7). Colonies are yellowish white, opaque, convex, and about 1 mm in

diameter after 24 h. Like C. ulcerans, C. pseudotuberculosis is positive for urease and

positive by the CAMP inhibition test. In this test, complete inhibition of the effect of S.

aureus beta-hemolysin on sheep erythrocytes is achieved by streaking the presumed C.

pseudotuberculosis strain in a right angle toward S. aureus and incubating overnight; a betahemolysin

inhibition zone in the form of a triangle is observed, as is true for A.

haemolyticum (Fig. 2). C. pseudotuberculosis is not susceptible to O/129, whereas C.

ulcerans strains are. C. pseudotuberculosis has varied results for both nitrate reduction and

sucrose fermentation. The API Coryne system and the RapID CB Plus panel correctly identify

this species (67, 69). Human disease to date has been acquired by handling of infected

sheep.

C. pyruviciproducens

This species was independently isolated from a groin abscess (151) and a urethral swab (G.

Funke, unpublished observation) and was initially believed to be a lipophilic variant of C.

glucuronolyticum, which was β-glucuronidase and CAMP test positive. However, 16S rRNA

gene sequence analysis, quantitative DNA-DNA hybridization, and partial rpoB gene

sequence analysis clearly demonstrated that this species is distinct fromC.

glucuronolyticum (151).

C. resistens

This species has entire, grayish-white, and glistening colonies and is lipophilic. It is unusual

in having a negative pyrazinamidase reaction, which separates it from the phenotypically

related C. jeikeium or C. tuberculostearicum. In addition, C. resistens grows slowly under

anaerobic conditions, whereas C. jeikeium is unable to do so. The C. resistens strains

reported in the literature are resistant to penicillin, cephalosporins, aminoglycosides,

clindamycin, and ciprofloxacin but remain susceptible to glycopeptides (106). It is presently

unknown whether true C. resistens strains are often misidentified as C. jeikeium in routine

laboratories.

C. riegelii

C. riegelii strains were originally described as being isolated from females with urinary tract

infections (59), but additional strains have been recovered from blood cultures, including

cord blood (8). Colonies are whitish, glistening, and convex, with entire margins, and are up

to 1.5 mm in diameter after 48 h of incubation. Some colonies are of a creamy consistency,

whereas others are sticky. C. riegelii strains exhibit a very strong urease activity with

Christen sen’s urea broth, becoming positive within 5 min at room temperature after

inoculation. A very peculiar characteristic of C. riegelii is the slow fermentation of maltose

but not glucose. No other definedCorynebacterium species exhibits this feature (Table 3).

The weak anaerobic growth of C. riegelii corresponds to the weak fermentative metabolism.

API Coryne system codes observed for C. riegelii include 0101224, 2001224, and 2101224.

C. simulans

This species was originally delineated from some C. s triatum-like strains (159). The three

strains described in the original publication came from skin-related specimens (foot abscess,

lymph node biopsy specimen, and boil). Two additional strains have been characterized, one

from bile and one from a blood culture (8). Colonies of C. simulans (grayish white, glistening,

creamy, 1 to 2 mm in diameter) are very similar to those of C. minutissimum, C.

singulare, and C. striatum, its closest phylogenetic neighbors. C. simulans is the only

validCorynebacterium species described to date which reduces nitrite. Further characteristics

which separate C. simulans from the closely related nonlipophilic, fermentative

corynebacteria are an inability to acidify ethylene glycol and to grow at 20°C (unlike C.

striatum). API Coryne profiles include 0100305, 2100105, 2100301, 2100305, and 3000125

(including a falsely negative nitrate reduction reaction because of the strong nitrite

reduction).

C. singulare

C. singulare colonies are circular and slightly convex, with entire margins, and are of a

creamy consistency, as observed for C. minutissimum and C. striatum (130). Key

biochemical reactions are like those for C. minutissimum except that urease activity is

observed (Table 3). The numerical API Coryne system profile is 6101125, which indicates

that pyrrolidonyl arylamidase activity is present. Like C. minutissimum and C. striatum, C.

singulare hydrolyzes tyrosine. C. singulare does not produce propionic acid as a fermentation

product, differentiating it from C. amycolatum. Phylogenetically, this species is closely

related (>98.8% identity) to both C. aurimucosum and C. minutissimum by 16S rRNA gene

sequencing but can readily be discerned by partial rpoB gene sequence analysis.

C. sputi

This species is the only true Corynebacterium that expresses α-glucosidase activity and is

positive for TBSA (164). It can be differentiated from C. ulcerans by being positive for

pyrazinamidase activity but negative for alkaline phosphatase and maltose fermentation.

Unlike C. ulcerans, this species cannot harbor the bacteriophage which bears the

diphtheria tox gene.

C. stationis

This species, originally called Brevibacterium stationis, has recently been reassigned to the

genusCorynebacterium (9). B. stationis ATCC 14403T was originally recovered from sea

water, but subsequent studies found that two human blood culture isolates plus “C.

ammoniagenes” ATCC 6872, recovered from an infant’s stools, formed a single taxon group

now designated C. stationis comb. nov. Strains of both C. stationisand C.

ammoniagenes sensu stricto demonstrate the ability to alkalinize citrate, using either

Simmon’s citrate or a heavy (not light) inoculum in the citrate reaction chamber found in an

API 20E strip, a feature not previously associated with members of the

genus Corynebacterium. Colonies of C. stationis grow well in 24 h (that is, they are not

lipophilic), are yellow or yellowish, and ferment glucose, fructose, ribose, and mannose. All

are positive for the alkalinization of citrate, produce urease, reduce nitrate, and hydrolyze

tyrosine (other features are shown in Table 3). API Coryne codes generated were 1001304

and 3001304 (that is, the enzyme PYZ is variably detected). Strains of this species are

susceptible to all antimicrobials; however, one blood culture isolate was shown to be

resistant to erythromycin.

C. striatum

C. striatum is part of the normal human skin biota. Nosocomial transmission of C.

striatum has been documented (89, 121). Colonies are convex, circular, shiny, moist, and

creamy, with entire edges, and are about 1 to 1.5 mm in diameter after 24 h of incubation.

Some investigators have described C. striatum colonies as being somewhat like those of

small coagulase-negative staphylococci. C. striatum has a fermentative metabolism, and acid

production from sucrose is variable. The API Coryne system identifies C. striatum, but

additional tests are needed in most cases (69). All C. striatum strains hydrolyze tyrosine,

and some strains are CAMP reaction positive; however, the CAMP reaction of C.

striatum strains is usually not as strong as that of other CAMP test-positive species (e.g., C.

auris or C. glucuronolyticum). Lactic and succinic acids are the major end products of glucose

metabolism. All C. striatum strains are susceptible to O/129. Resistance to macrolides and

lincosamides due to the presence of an rRNA methylase has been described. C. striatum may

also be resistant to quinolones and tetracyclines (42, 75), which has led to a renewed

interest in this agent as an emerging pathogen (10).

C. sundsvallense

C. sundsvallense (8, 20) has been isolated from blood cultures, a vaginal swab, and a sinus

drainage specimen from an infected groin. Colonies of this nonlipophilic species are buff or

slightly yellowish, they adhere to agar, and they have a sticky consistency. Gram staining

shows bulges or knobs at the ends of some rods, and these are not seen in any other

corynebacteria. Fermentation of glucose, lactose, and sucrose is slow (Table 3). C.

sundsvallense can be separated from C. durum by its positive α-glucosidase reaction and its

inability to ferment galactose. It is differentiated from C. matruchotii by expressing urease

but not nitrate reductase activity and by not producing propionic acid as an end product of

glucose metabolism (8, 20).

C. thomssenii

C. thomssenii (169) is a rarely found species. It was originally repeatedly isolated from a

patient with pleural effusion, and a second strain was recovered from the environment in

Canada (8). This species is fastidious and grows slowly, resulting in colonies of <0.5 mm

after 48 h, but it is not lipophilic. After 96 h, colonies are molar-tooth-like, very sticky, and

slightly adherent to agar. The clinical strain of C. thomssenii is the onlyCorynebacterium

species expressing N-acetyl-β-glucosaminidase activity, which can be observed in either the

API Coryne or the API ZYM system. Acid is slowly and fermentatively produced from glucose,

maltose, and sucrose, and the resulting API Coryne code for C. thomssenii is 2121125.

C. timonense

This species was isolated from blood cultures of a patient with endocarditis. Its most peculiar

feature is a weakly positive esculinase reaction. It can be differentiated from other

esculinase-positive corynebacteria by having negative reactions for maltose and sucrose

fermentation as well as a negative CAMP reaction (96).

C. tuberculostearicum

This species was revived for a never validly published taxon and also includes a strain of the

nonvalidated species “C. pseudogenitalium” (43). Nearly all “CDC group G” bacteria can be

assigned to C. tuberculostearicum(K. A. Bernard and G. Funke, unpublished observation).

Unfortunately, the effective publication on C. tuberculostearicum did not include any strains

of CDC group G bacteria (43). C. tuberculostearicum has some variable biochemical key

reactions and can be differentiated from C. accolens if TBSA is detected as one of its CFAs. It

can be separated from C. jeikeium by anaerobic growth and fermentative acid production

from fructose. C. tuberculostearicum can be multidrug resistant, but the most frequently

observed resistance is to macrolides and lincosamides.

C. tuscaniense

This species has been isolated from blood cultures of a patient suffering from endocarditis. C.

tuscaniensedoes not grow under anaerobic conditions, which distinguishes this species from

the phenotypically similar C. minutissimum. In addition, C. tuscaniense hydrolyzes hippurate

but not tyrosine, whereas C. minutissimumhas opposite reactions (Table 3). C.

tuscaniense colonies are rounded and regular, in contrast to the biochemically similar

species C. amycolatum, whose colonies exhibit irregular edges (123).

C. ulcerans

Phylogenetically, C. ulcerans (129) is closely related to C. pseudotuberculosis, and both

species are the closest relatives to C. diphtheriae (108, 133). Its cell membrane contains

significant amounts of the CFA C16:1v7c. As described previously, C. ulcerans can harbor the

diphtheria toxin gene, but differences in the receptor-binding and translocation domains

have been described, so a C. ulcerans diphtheria toxin-specific PCR has been developed for

conventional PCR (138) or real-time PCR (15, 135, 136). Disease associated with this

bacterium is rare, but if it is recovered from pseudomembranous material, it must be treated

like a case of diphtheria (33,34). Recent reports have described the transmission of C.

ulcerans from companion pets to humans (11, 31,88), and so expanded national cases of

diphtheria-like diseases that include cases involving toxigenic strains of C. ulcerans or C.

pseudotuberculosis have formally been defined in some countries (K. Bernard, unpublished

observation) (11). C. ulcerans colonies are somewhat dry, waxy, and grayish white, with

light hemolysis; they are 1 to 2 mm in diameter after 24 h. C. ulcerans may be differentiated

from C. diphtheriae by urease activity and a CAMP inhibition reaction and from C.

pseudotuberculosis by source, biochemically, and by partial rpoBgene but not 16S rRNA gene

sequencing. Strains of C. ulcerans are positive for glycogen, starch, and trehalose

fermentation. The API Coryne system and the RapID CB Plus identification strip correctly

identify C. ulcerans(67, 69).

C. urealyticum

C. urealyticum is one of the relatively frequently isolated clinically significant corynebacteria

in clinical specimens (72). C. urealyticum is strongly associated with urinary tract infections.

Recovery of this bacterium is often associated with urine with an alkaline pH, resulting in

struvite crystals. As for all other lipophilic corynebacteria, colonies are the size of a pinpoint,

convex, smooth, and whitish gray on regular SBA. C. urealyticum is a strict aerobe and has

very strong urease activity (Table 3). Commercial identification systems correctly identify C.

urealyticum. C. urealyticum is almost always multidrug resistant (42, 72, 140). Sequencing

of the whole C. urealyticum genome indicated that the multidrug resistance is mediated by

transposable elements (150).

C. ureicelerivorans

Lipophilic C. ureicelerivorans, first described from a blood culture, has very strong and

rapidly detected (~60 s) urease activity as its most prominent feature (163). Subsequently,

recovery of this organism from ascites fluid and multiple blood cultures from

immunocompromised patients or those with digestive disorders has been described (41). It

can be differentiated from CDC group F-1 bacteria by a positive alkaline phosphatase

reaction and by a negative reaction for acid production from maltose and sucrose. By 16S

rRNA and partialrpoB gene sequencing, this species is closest to C. mucifaciens; C.

ureicelerivorans can readily be discerned from that species by its smooth rather than mucoid

yellowish colonies as well as by observation of rapid urease activity.

C. xerosis

C. xerosis colonies are dry, granular, and yellowish, with irregular edges, and are 1 to 1.5

mm in diameter after 24 h. It must be emphasized that nearly all “C. xerosis” strains which

have been described in the literature before 1996 may have been misidentified C.

amycolatum strains (56). C. striatum strains were also misidentified as C. xerosis in the

past. C. xerosis has a fermentative metabolism and has variable results for the presence of

nitrate reductase but always expresses α-glucosidase as well as leucine arylamidase activity.

Because C. xerosis was thought to be rarely encountered in clinical specimens, it was not

included in the API Coryne system version 3.0 database. The numerical profiles observed

for C. xerosis strains, such as 2110325 and 3110325, provide API Web responses that

include “C. striatum/C. amycolatum.” The RapID CB Plus system correctly identifies C.

xerosis (67). Lactic acid is the major end product of glucose metabolism, and strains are

susceptible to O/129. As described previously, C. xerosis is closely phylogenetically related

to C. freneyi and C. hansenii (119), and so if definitive identification is required,

characterization must include genetic testing.

Genus Turicella

T. otitidis is almost exclusively isolated from clinical specimens from the ear region, but it

does not cause otitis media with effusion in children. Colonies are whitish, convex, and

creamy, with entire edges, and are 1 to 1.5 mm in diameter after 48 h of incubation. Some

young colonies show a greenish appearance when taken away from the plates with a swab.

The distinctive Gram stain morphology of T. otitidis is given in Fig.1b. Differentiation from C.

auris and C. afermentans subsp. afermentans is readily achieved by morphologic features,

but utilization reactions may also assist in the differentiation of these taxa (Tables 3 and 4)

(57,118). All T. otitidis strains are strongly CAMP reaction positive and give the numerical

code 2100004 in the API Coryne system. The MICs of β-lactam antibiotics for many strains

are very low; some strains might be resistant to macrolides and clindamycin (68, 75). The

mechanism of resistance to macrolides appears to be due in part to mutations in the 23S

rRNA (rrl) gene (12).

Genus Arthrobacter

Arthrobacter spp. might be part of the indigenous normal human biota, but their main

habitat is soil. A. cumminsii seems to be a normal commensal in humans and appears to be

the most frequently isolatedArthrobacter species in human clinical specimens (63, 92),

and A. oxydans is the second-most-frequently encountered species

(92). Arthrobacter colonies are usually whitish gray, slightly glistening, creamy, and 2 mm or

greater in diameter after 24 h. A. cumminsii is slightly smaller than the other arthrobacters

and may also exhibit a sticky consistency (63). Arthrobacter spp. usually do not oxidize any

of the carbohydrates routinely tested and do not emit a cheese-like smell, as is often found

for the phenotypically closely related brevibacteria. Some arthrobacters are motile, whereas

brevibacteria are always nonmotile. Like brevibacteria,Arthrobacter spp. express DNase and

have gelatinase activity (54). The identification of arthrobacters to the species level might be

achieved by carbohydrate utilization tests, but this is recommended for reference

laboratories only. A. albus (160) is phylogenetically most closely related to A. cumminsii but

might be differentiated phenotypically by being resistant to desferrioxamine, whereas A.

cumminsii is susceptible. A. cumminsii has a distinctive CFA pattern, with C14:0i and

C14:0 each representing 2 to 4% of all CFAs (54). Penicillin MICs for most Arthrobacter strains

are low, with quinolones showing only weak activities againstArthrobacter spp. (54, 92).

Genus Brevibacterium

Some Brevibacterium spp. are part of the normal human skin biota. Colonies are whitish

gray (or yellowish likeB. luteolum), convex, mostly creamy, and 2 mm or greater in diameter

after 24 h. B. mcbrellneri colonies have a more granular appearance and are dryer than

those of other brevibacteria. Some brevibacteria may develop a yellowish or greenish

pigment after prolonged incubation. Many Brevibacterium strains isolated from human

clinical material give off a distinctive cheese-like odor. Brevibacteria are nonmotile, are

halotolerant (6.5% NaCl), and form methanethiol from methionine, but this test is specific

for brevibacteria only when it is read within 2 h (48). Brevibacteria can be identified to the

species level by carbohydrate utilization tests. More than 90% of all

clinical Brevibacterium isolates are B. casei (48). Brevibacterium sanguinis is very similar

to B. caseiand can be differentiated from this species by susceptibility to thallium acetate.

The MICs of β-lactam antibiotics for brevibacteria are often elevated (68).

Genus Dermabacter

D. hominis strains are part of the normal skin biota. Colonies are whitish, convex, of a

creamy or sticky consistency, and 1 to 1.5 mm in diameter after 48 h (Fig. 3e). D.

hominis strains are sometimes mistaken for small-colony coagulase-negative staphylococci.

The Gram staining result is distinctive, with coccobacillary or coccoidal forms (Fig. 1c). The

key biochemical reactions are given in Table 4. D. hominis is one of the few coryneform

bacteria with a variable reaction for xylose fermentation. It is the only catalase-positive

coryneform bacterium (except Actinomyces neuii) that is able to decarboxylate lysine and

ornithine (70). The API Coryne system and the RapID CB Plus panel correctly identify this

species (67, 69). D. hominis strains may be resistant to aminoglycosides (68, 152).



TYPING SYSTEMS Back to top

Outbreaks of C. diphtheriae in the states of the former Soviet Union and other locations have

been studied by whole-cell peptide analysis, whole-genome restriction fragment length

polymorphism analysis, ribotyping, pulsed-field gel electrophoresis, PCR–single-strand

conformation polymorphism analysis, analysis of tox anddtxR as well as of the 16S-23S rRNA

spacer region, amplified fragment length polymorphism analysis, random amplification of

polymorphic DNA, and multilocus enzyme electrophoresis (30, 113). An international

database for C. diphtheriae ribotypes using the endonuclease BstEII has been established

(76). Ribotyping is regarded as being highly discriminatory and, based on a comprehensive

comparison of methods, was found to be a preferred typing method for C. diphtheriae (30).

Other methods, including a spoligotyping system (similar to the spacer oligonucleotide typing

for Mycobacterium tuberculosis), have been described (98). Sequencing studies with C.

diphtheriae strains from the epidemic in the former Soviet Union have shown that point

mutations within the tox gene were silent mutations and that multiple point mutations

(which even led to amino acid substitutions) were observed for the dtxR gene, corresponding

to the heterogeneity of outbreak strains as revealed by PCR–single-strand conformation

polymorphism analysis (100). Isolates derived from specific populations in the United States

and Canada and characterized by multilocus enzyme electrophoresis, ribotyping, and random

amplification of polymorphic DNA were found to be members of persistent endemic strains,

rather than being imported from other countries where diphtheria is endemic. More recently,

a standardized multilocus sequence typing method based on analysis of short sequences

derived from seven housekeeping alleles has been under development, with which

concatenated sequence data, sent by the Internet to the curator (initially the University of

Warwick and then the Institut Pasteur, Paris [relocation in 2009]), are compared to a large

database and given an existing or new sequence type (25). Due to the technical simplicity of

this approach and the ease of comparison of strain data internationally, this is being touted

as a typing method for the future. Multilocus sequence typing has also been effectively

applied to an outbreak investigation involving C. macginleyi and C mastitidis-like isolates

(35).

SEROLOGIC TESTS Back to top

Detection of antibodies directed against diphtheria toxin is the only established serologic test

for coryneform bacteria. Toxin neutralization assays using a Vero cell culture system have

been replaced mainly by enzyme immunoassays. Levels of ≥0.1 IU/ml serum are thought to

confer protection, whereas levels of <0.01 IU/ml indicate a susceptible host and levels of

0.01 to 0.1 IU/ml indicate partially immune individuals. It is believed that between 20 and

60% of adults in the United States lack protective antibodies to diphtheria toxin due to

declining antibody titers in immunized persons and in those persons who did not receive the

primary immunization series. This could pose a potentially significant public health risk and

could result in the reemergence of this disease. Booster doses of toxoid should be

administered at 10-year intervals.

ANTIMICROBIAL SUSCEPTIBILITIES Back to top

The susceptibility patterns for each taxon were given with the descriptions of each taxon (see

above). Since the antimicrobial susceptibility of coryneform bacteria is not predictable in

every case, susceptibility testing should always be performed with clinically significant

isolates (see “Clinical Significance” above). Due to the emergence of vancomycin-resistant

gram-positive organisms, it has become inappropriate to recommend glycopeptides as firstline

drugs for the treatment of infections caused by coryneform bacteria. It is also

noteworthy that some coryneform bacteria (e.g., Microbacterium resistens) are intrinsically

vancomycin resistant.

The Clinical and Laboratory Standards Institute has published testing conditions and

interpretive criteria for susceptibility testing of coryneform bacteria using a broth

microdilution method (18). Direct colony suspensions equivalent to a 0.5 McFarland standard

are prepared and strains are incubated in cation-adjusted Mueller-Hinton broth with 2 to 5%

(vol/vol) lysed horse blood at 35°C in ambient air for up to 48 h. Interpretative categories

for the MICs obtained are presently available for 16 antimicrobial agents. In summary, using

the broth microdilution method, it has recently been established that multidrug resistance

was found for 4 or more drug classes (out of 16 classes tested) for some or most strains

of C. amycolatum, C. jeikeium, C. urealyticum, C. resistens, C. tuberculostearicum, and C.

striatum, as well as representatives of C. afermentans and C. aurimucosum found in one

national culture collection (C. Singh, T. Burdz, and K. Bernard, unpublished observations).

Very few studies have been performed comparing broth microdilution and disk diffusion

results for susceptibility testing of coryneform bacteria (162).

In the past, MICs were determined by either the Etest or the agar dilution or broth

microdilution method. The results of the Etest have been shown to correlate reasonably well

with those of both the broth microdilution and the agar dilution method

for Corynebacterium spp. (95, 167). The Etest should be carried out on Mueller-Hinton agar

supplemented with 5% sheep blood. The same medium is used for the agar dilution method

(68), but this method is not applicable in routine laboratories and, rather, should be used in

studies with individual antimicrobial agents.

Metronidazole is the drug of choice both for local therapy of BV and for systemic therapy of

extravaginal infections caused by BV-associated biota. Systemic infections due to G.

vaginalis alone can be treated with ampicillin or amoxicillin, since β-lactamase-producing G.

vaginalis strains have not been observed so far. Susceptibility testing for G. vaginalis is not

recommended.

EVALUATION, INTERPRETATION, AND REPORTING OF

RESULTS Back to top

The guidelines related to when coryneform bacteria should be identified to the species level

(see “Clinical Significance” above) are also applicable for evaluating and interpreting culture

results; i.e., whenever coryneform bacteria are identified to the species level, the results

should be reported.

In the rare case of microscopically suspected C. diphtheriae (i.e., a positive Neisser staining

result), the physician in charge of the patient should be notified immediately, although

culture results and toxin testing results become available only later.

It is evident that repeated isolation of a predominant strain of a coryneform bacterium or a

coryneform bacterium growing in pure culture suggests an etiological relationship to the

patient’s disease. If coryneform bacteria are present in blood cultures, the physician in

charge should be notified immediately, and it should be emphasized when reporting that the

clinical significance of the coryneform bacteria must be carefully examined by cooperation

between the microbiology laboratory and the physician. In our experience, one positive blood

culture out of two or three aerobically and anaerobically incubated pairs of blood cultures is

hardly ever clinically significant (except in cases of treated endocarditis). Care must be taken

in the interpretation of the results for those patients for whom half or more of the blood

specimens taken for culture become positive for coryneform bacteria, in particular when

lipophilic corynebacteria are cultured, since not all blood samples taken from patients with

endocarditis due to lipophilic corynebacteria may eventually become positive.

On the other hand, coryneform bacteria should be reported as “normal flora” when they are

grown from nonsterile sites together with other resident biota in equal or smaller numbers. It

is suggested that the primary isolation plates be retained for at least 72 h before they are

discarded in order to have the opportunity to assess the bacterial population retrospectively.

No comments:

Post a Comment