Acinetobacter & Others

TAXONOMY Back to top

The organisms covered in this chapter belong to a group of taxonomically and

phylogenetically diverse, gram-negative nonfermentative rods and coccobacilli. Still, several

of the genera dealt with belong to the same family;

i.e., Acinetobacter, Moraxella, Oligella, and Psychrobacter belong to the family Moraxellaceae

(Gammaproteobacteria) (182), and Balneatrix, Bergeyella,

Chryseobacterium, Elizabethkingia, Empedobacter,Myroides, Sphingobacterium,

Wautersiella, and Weeksella belong to the family Flavobacteriaceae (Bacteroidetes) (11).

DESCRIPTION OF THE AGENTS Back to top

The species dealt with in this chapter all share the common phenotypic features of being

catalase positive and failing to acidify the butt of Kligler iron agar (KIA) or triple sugar iron

(TSI) agar or of oxidative-fermentative media, indicating their inability to metabolize

carbohydrates by the fermentative pathway. These organisms grow significantly better under

aerobic than under anaerobic conditions, and many, i.e., those species that can use only

oxygen as the final electron acceptor in the respiratory pathway, fail to grow anaerobically at

all.

EPIDEMIOLOGY AND TRANSMISSION Back to top

Most of the organisms described in this chapter are found in the environment, i.e., soil and

water. For methylobacteria, tap water has been implicated as a possible agent of

transmission in hospital environments, and methods for monitoring water systems for

methylobacteria have been described previously (178). No person-to-person spread has

been documented for the species covered in this chapter.

CLINICAL SIGNIFICANCE Back to top

Although for almost each of the species in this chapter, as for most other species in other

chapters, case reports of, e.g., meningitis and endocarditis can be found, their clinical

importance is mostly restricted to that of opportunistic pathogens, except, e.g.,

for Elizabethkingia meningoseptica, Moraxella lacunata (eye infections), or Moraxella

catarrhalis (respiratory tract infections).

The clinical role of Acinetobacter species has been re viewed previously (58, 115, 164).

These organisms are ty pical opportunistic pathogens that usually only form a threat to

critically ill, hospitalized patients. Hospital-acquired Acinetobacter infections comprise

ventilator- associated pneumonia, bloodstream infections, urinary tract infections, wound

infections, skin and soft tissue infections, and secondary meningitis. Acinetobacter

baumannii is the species most commonly implicated in hospital-acquired infections. The

clinical role of the closely related Acinetobacter genomic species 3 and 13TU resembles that

of A. baumannii (18, 22). For the purpose of this review, we consider A. baumannii to

comprise these two species as well, unless stated otherwise. A. baumannii ventilatorassociated

pneumonia and bloodstream infections have been documented to be associated

with a high degree of mortality and morbidity (41, 188). Particular manifestations of A.

baumannii are its implication in severely war-wounded soldiers (32, 52), from which stems

its popular designation “Iraqibacter,” and in victims of natural disasters (161).

The clinical impact of infections with A. baumannii is a continuous source of debate

(58, 164). Indeed, although severe infections with A. baumannii have been documented,

colonization is much more frequent than infection, and differentiation between these

conditions can be difficult.

Although uncommon, community-acquired infections with A. baumannii occur. In particular,

community-acquired pneumonia with A. baumannii is increasingly reported from tropical

areas, like Southeast Asia and tropical Australia (3, 134).

Other Acinetobacter species occasionally implicated in nosocomial infections are listed

in Table 1. A. johnsonii,A. lwoffii, and A. radioresistens seem to be natural inhabitants of

human skin (186). A. johnsonii, which has also been found frequently in feces of

nonhospitalized individuals (59), has been implicated in cases of meningitis (189). A.

lwoffii was a frequent species in clinical specimens during an 8-year study in a university

hospital, where it was isolated mainly from blood or intravascular lines (220). A.

ursingii and A. junii have been found to cause bloodstream infections in hospitalized patients

(63, 107, 139, 210), while A. junii has also been implicated in outbreaks of infection in

neonates (55, 121) and ocular infections (172). A. parvus is regularly isolated from blood

cultures (150, 210) but is misidentified by API 20NE as A. lwoffii (M. Vaneechoutte,

unpublished data). Many of the infections with these species are related to intravascular

catheters or have another iatrogenic origin (9, 63, 194, 238, 248), and their course is

generally benign. For various other named or yet-unnamed Acinetobacter species, although

recovered from clinical specimens (21, 206), a possible role in infection has not been

documented.

Moraxella species are rare agents of infections (conjunctivitis, keratitis, meningitis,

septicemia, endocarditis, arthritis, and otolaryngologic infections) (54, 122, 191, 223),

but M. catarrhalis has been reported to cause sinusitis and otitis media by contiguous spread

of the organisms from a colonizing focus in the respiratory tract (122). However, isolation

of M. catarrhalis from the upper respiratory tract (i.e., a throat culture) of children with otitis

media or sinusitis does not provide evidence that the isolate is the cause of these infections,

because M. catarrhalis is present frequently as a commensal of the upper respiratory tract in

children (232). Isolates from sinus aspirates and middle ear specimens obtained by

tympanocentesis should be identified and reported. Similarly, little is known about the

pathogenesis of lower respiratory tract infection in adults with chronic lung diseases,

although a clear pathogenic role may be assigned to this species because M. catarrhalis is

not a frequent commensal of the upper respiratory tract in adults (232) and because

examination of Gram-stained smears of sputum specimens from patients with exacerbations

of bronchitis and pneumonia due to M. catarrhalis usually reveals an abundance of

leukocytes, the presence of many gram-negative diplococci as the exclusive or predominant

bacterial cell type, and the presence of intracellular gram-negative diplococci. Such

specimens may yield M. catarrhalis in virtually pure culture, and the organism should be

identified and reported. M. lincolnii is not frequently isolated from clinical samples. M.

nonliquefaciens and M. osloensis are the two species most frequently isolated, approximately

in equal numbers, from nonrespiratory clinical material, especially blood cultures from

patients at risk. M. canis has been isolated from dog bite wounds (111) and from debilitated

patients (223). M. lacunata has been involved in eye infections (180).

COLLECTION, TRANSPORT, AND STORAGE OF

SPECIMENS Back to top

Standard methods for collection, transport, and storage of specimens as detailed in

chapters 9 and 16 are satisfactory for this group of organisms. The only fastidious species

handled in this chapter are Asaia species,Granulibacter bethesdensis,

Methylobacterium species, and some Moraxella species.

DIRECT EXAMINATION Back to top

There are no characteristics available that can help to recognize the species dealt with in this

chapter by means of direct microscopic examination of the samples. On Gram stain,

organisms appear as gram-negative rods, coccobacilli, or diplococci. Neither direct antigen

tests nor molecular genetic tests to use directly on clinical materials have been developed.

ISOLATION PROCEDURES Back to top

Initial incubation should be at 35 to 37°C, although some strains, among them many of the

pink-pigmented species, grow better at or below 30°C and may be detected only on plates

left at room temperature. In such cases, all tests should be carried out at room temperature.

In fact, some of the commercial kits, such as the API 20NE, are designed to be incubated at

30°C.

Growth on certain selective primary media, e.g., MacConkey agar, is variable and may be

influenced by lot-to-lot variations in the composition of media. Nonfermenters that grow on

MacConkey agar generally form colorless colonies, although some form lavender or purple

colonies due to uptake of crystal violet contained in the agar medium. Selective media have

been described for Acinetobacter spp. (8, 113) and for Moraxella spp. (231), but their

usefulness remains to be assessed.

IDENTIFICATION Back to top

This chapter starts with an overview in Fig.1, which provides a key to the five large groups

that can be distinguished among the species described in this chapter. In the previous

edition, the simplified scheme for identification of this group of organisms in the clinical

laboratory was based on microscopic morphology, oxidase reaction, motility, acidification of

carbohydrates, indole production, and the production of pink-pigmented colonies. The

identification scheme presented here (Fig. 1) is further simplified and based only on colony

color (pink or not) and the presence or absence of oxidase, of benzyl arginine

aminopeptidase (trypsin) activity, and of the production of indole.

For each group of closely related species, we present their taxonomic history (explaining the

use of other names in the past and the taxonomic changes introduced since the previous

edition), address the clinical importance of the species, and describe the phenotypic data

that are useful to differentiate this group from other groups and to differentiate the species

within this group (emphasizing the major differences from the previous edition). When

relevant, antibiotic susceptibility characteristics and treatment options are discussed

immediately; otherwise, they are discussed at the end of each section for the five large

groups in this chapter.

The chapter on miscellaneous nonfermentative gram-negative bacteria in this edition of

the Manual differs from chapter 50 of the previous edition in several aspects. Some of the

species described in this chapter in the ninth edition are dealt with in other, more

appropriate chapters: Alcaligenes faecalis (chapter 43),Achromobacter

denitrificans and Achromobacter xylosoxidans (chapter 43), Advenella incenata (chapter

43),Bordetella (chapter 43), Herbaspirillum (including EF-1 isolates) (chapter 41), Kerstersia

gyiorum (chapter 43),Neisseria weaveri and N. elongata (chapter 32), and CDC groups EF-4a

and EF-4b (described here as Neisseria animaloris and Neisseria zoodegmatis [218])

(chapter 32). Halomonas venusta, Laribacter hongkongensis, andMassilia timonae are no

longer included.

In addition, we have included only species with validated names and no longer deal with the

following groups: Achromobacter group F (87); Agrobacterium yellow group (247); CDC

halophilic nonfermenter group 1 (close to Halomonos venusta) (242); CDC groups Ic (247),

IIg (85), EO-3 (45), EO-4, EO-5, NO-1 (86), O-1 (close to Hydrogenophaga palleronii) (174),

O-2 (close to Caulobacter vibrioides), O-3 (47), OFBA-1 (244),Pseudomonas-like group 2

(formerly included in the heterogeneous CDC group IVd and close to Herbaspirillum

rubrisubalbicans), and a group of thermophilic bacteria, some classified as “Tepidimonas

arfidensis” (128).Pedobacter species (199) are also no longer mentioned because they have

little or no clinical relevance.

Former unnamed groups have been described as species with validated names in the

meantime and are discussed under their appropriate names in this

chapter: Achromobacter groups B and E have been described as Pannonibacter

phragmitetus (94), CDC groups IIh and IIc as Chryseobacterium hominis (228), part of the

CDC group IIe strains as Chryseobacterium anthropi (120), and CDC group EO-2 strains

as Paracoccus yeei (48). Additional new species have been described in the meantime, and

those that are included are Granulibacter bethesdensis (79) and Wautersiella falsenii (119).

Although genera like Acinetobacter and Chryseobacteriumcomprise many more species than

the ones addressed here, we focus on those species that can be isolated from clinical

samples.

Classical Biochemical Identification Schemes Presented in This

Chapter

For all the species that remain in this chapter, except those of the genus Acinetobacter, the

biochemical tests listed have been carried out by one of us (G. Wauters), according to

standardized protocols, described in detail in chapter 31. This means that for most species

the number of strains tested is smaller than the number of strains tested in the previous

edition, but that the data listed are not compiled from the literature, whereby different

authors may have used different media and protocols. The limited number of tests that have

been used to discriminate between the species dealt with in this chapter have been selected

because they can be carried out easily and quickly, because they mostly yield uniform results

per group or species, and because they are highly discriminatory. For the

genus Acinetobacter, data based on standardized physiological and nutritional tests were

adapted from the literature or were provided by one of the authors (A. Nemec) (see

footnotes to Table 1).

Automated, Commercially Available Phenotypic Identification

Systems

Traditional diagnostic systems, e.g., those based on oxidation- fermentation media, aerobic

low-peptone media, or buffered single substrates, have now been replaced in many

laboratories by commercial kits or automated systems like the Vitek 2 (bioMerieux, Marcy

L’Etoile, France) and the Phoenix (BD Diagnostic Systems, Sparks, MD). The ability of

commercial kits to identify this group of nonfermenters is variable and often results in

identification to the genus or group level only, necessitating the use of supplemental

biochemical testing for species identification. O’Hara and Miller (160), using the Vitek 2 IDGNB

identification card, reported that of 103 glucose-fermenting and nonfermenting

nonenteric strains, 88 (85.4%) were correctly identified at probability levels ranging from

excellent to good and that 10 (9.7%) were correctly identified at a low level of

discrimination, for a total of 95.1% accuracy within this group. Bosshard et al. (19)

compared 16S rRNA gene sequencing for the identification of clinically relevant isolates of

nonfermenting gram-negative bacteria (non-Pseudomonas aeruginosa) with two

commercially available identification systems (API 20NE and Vitek 2 fluorescent card;

bioMerieux). By 16S rRNA gene sequence analysis, 92% of the isolates were assigned to

species level and 8% to genus level. Using API 20NE, 54% of the isolates were assigned to

species level, 7% were assigned to genus level, and 39% of the isolates could not be

discriminated at any taxonomic level. The respective numbers for Vitek 2 were 53, 1, and

46%. Fifteen percent and 43% of the isolates corresponded to species not included in the

API 20NE and Vitek 2 databases, respectively. Altogether, commercial identification systems

can be useful for identification of organisms commonly found in clinical specimens,

like Enterobacteriacaeae. However, for rare organisms the performance of these systems can

be poor. This is also illustrated by the performance of API 20NE and Vitek 2 for clinical

isolates of Acinetobacter(reference 12 and below).

Chemotaxonomic Methods

Identification of nonfermenters by automated cellular fatty acid analysis has also been

attempted (237). In view of the difficulties inherent in this approach (162), it is

recommended that fatty acid profiles be used only in conjunction with traditional or

commercial diagnostic systems. The fatty acid profiles for the most common species of

nonfermenting bacteria have been published (247). Unless specifically relevant, we have

omitted fatty acid composition data, which were presented in the tables of the previous

edition.

A recently developed method of bacterial identification is matrix-assisted laser desorption

ionization–time-of-flight mass spectrometry, for which commercial systems, with bacterial

mass spectrum databases, have become available recently (Autoflex II mass spectrometer

[Bruker Daltonics, Billerica, MA] and Axima [Shimadzu, Kyoto, Japan]). A recent evaluation

showed that 84.1% of 1,660 bacterial isolates analyzed were correctly identified to the

species level (190). However, few of the species dealt with in this chapter were included.

Another recent application of this technology deals with the Burkholderia cepacia complex,

indicating its applicability for gram-negative nonfermenters (GNF) (233).

DNA Sequence-Based Methods

Sequence-based methods involving rRNA (16S, 16S-23S spacer, or 23S) and housekeeping

genes, such as those encoding RNA polymerase subunit B (rpoB), gyrase subunit

B (gyrB), or the RecA protein (recA), have become standard techniques to identify bacteria

in general (167) and have contributed to the better delineation of several of these groups

and the discovery and description of new species. Because these are generally applicable

methods, their application for species of this chapter is not outlined in detail. Other

sequence-based methods, based on DNA array hybridization, have been used for some

species of these groups (129, 201). DNA sequence-based fingerprinting methods like

amplified ribosomal DNA (rDNA) restriction analysis (227, 230), amplified fragment length

polymorphism (AFLP) (112), and tDNA PCR (31, 67) have been applied for the identification

of species of several groups as well. These fingerprinting approaches are also generally

applicable, but they require reference fingerprint libraries and are often poorly exchangeable

between different electrophoresis platforms and laboratories.

IDENTIFICATION OF THE FIVE GENOTYPIC GROUPS Back to

top

Oxidase-Negative GNF

Acinetobacter

The taxonomy of the genus Acinetobacter (21, 23, 57, 206, 224) has recently been updated

with extended descriptions and formal species names for three species previously designated

with provisional designations. These include A. venetianus (229) and A. bereziniae and A.

guillouiae (previously designated genospecies 10 and 11, respectively) (21, 154). New

species, comprising strains of clinical origin, have been described as well, i.e., A.

beijerinckii (154), A. gyllenbergii (154), A. parvus (150), A. schindleri (149), and A.

ursingii (149). Some of the species that were described recently have been shown to be

synonymous to already existing species: A. grimontii (29) was shown to be synonymous

to A. junii (225), and “A. septicus” is synonymous to A. ursingii(156). At present, the genus

comprises 21 validly named species and 11 species with provisional names. The G+C content

of the genus ranges from 38 to 47 mol%. The genomes of seven Acinetobacter strains have

been sequenced (NCBI, July

2009; http://www.ncbi.nlm.nih.gov/genome?term=acinetobacter).

Members of the genus Acinetobacter are widespread in nature and have been cultured from

soil, water, sewage, and food and from human and animal specimens. The ecology of most

species is unknown. Species of clinical importance are listed in Table 1.

Bacteria belonging to the genus Acinetobacter are strictly aerobic, nonfermenting gramnegative

coccobacillary microorganisms with a negative oxidase reaction and a positive

catalase reaction. Tween 80 esterase activity is frequently present, hemolysis and gelatinase

production vary, and nitrate reductase is mostly absent. Motility (hanging drop) is negative,

but twitching motility on soft agar occurs occasionally. Individual cell sizes are 0.9 to 1.6 μm

in diameter and 1.5 to 2.5 μm in length. In the stationary phase, the organisms are usually

coccoid. Cells frequently occur in pairs, resembling Neisseria species, but this may be strain

or species dependent. In the Gram stain, the organisms can be slightly gram positive.

Growth temperature varies, but most species grow between 20 and 35°C. Clinically

important species commonly grow well at 37°C or at higher temperatures.

The organisms can form a pellicle on the surface of fluid media. They grow well on complex

media, including blood agar, nutrient agar, and MacConkey agar. Colonies are 1 to 2 mm in

diameter (sometimes pinpoint), colorless to beige, domed, and smooth to mucoid (Fig. 2).

Colonies on MacConkey agar can become pink. Many strains can use a wide variety of carbon

sources for growth. Selective enrichment can be obtained in mineral media with acetate as

the carbon source and ammonium salt as the nitrogen source with shaking incubation at

30°C (8, 56, 61). General features of Acinetobacter species have been reviewed previously

(57, 116).

For genus level identification of Acinetobacter isolates, the following characters can be used:

gram-negative coccobacilli, oxidase negative, aerobic (nonfermenting), and nonmotile.

Phenotypic identification ofAcinetobacter species in the clinical microbiology laboratory by

commercial identification systems is problematic (12). This results from the small number of

relevant characters tested in these systems and/or from the insufficient quality of reference

data in the identification matrices. A. baumannii and the closely related

speciesAcinetobacter genomic species 3 and 13TU, which are clinically the most important

species, and A. calcoaceticus,an environmental species, together referred to as the A.

calcoaceticus-A. baumannii complex, are generally not differentiated by these systems.

Nonetheless, these systems can be useful for genus level identification and, when

supplemented with aerobic acidification of glucose (oxidation-fermentation test), hemolysis,

and growth at 44°C, also for presumptive identification of A. baumannii (Table 1). We

compared Vitek 2 and Phoenix for the ability to identify 76 isolates of 16

clinical Acinetobacter species and found that only 19 isolates were correctly identified by

Vitek 2 and 5 by Phoenix (M. Vaneechoutte, unpublished data). Phenotypic identification

of Acinetobacter species can be achieved using physiological, i.e., biochemical and growth

temperature, characteristics, and nutritional, i.e., assimilation, characteristics, based on the

system of Bouvet and Grimont (22). Table 1 presents a recent update of this system aimed

to differentiate all validly named species of clinical importance. Assimilation tests were

carried out using the minimal medium of Cruze et al. (44), dispensed into tubes (12-mm

inner diameter) in 3-ml volumes inoculated with a small inoculum. Growth on carbon sources

was evaluated after 2, 4, 6, and 10 days by means of visual comparison between inoculated

tubes containing carbon sources and control tubes containing only inoculated basal medium.

Unfortunately, the species of theA. calcoaceticus-A. baumannii complex are not clearly

distinguished from each other by this approach. In addition, the need for in-house

preparation of most of the tests precludes the use of this identification scheme in most

diagnostic laboratories.

Therefore, genotypic methods are indispensable for unambiguous identification

of Acinetobacter species. Well- validated methods are amplified rDNA restriction analysis

(60, 227) and whole genomic fingerprinting by AFLP, based on the selective amplification of

chromosomal restriction fragments (57, 112). Currently, sequence-based species

identification is becoming more and more the standard. Targets for this purpose are the 16S

rDNA sequence (224), the rpoB gene sequence (81), and the 16S-23S rRNA gene spacer

region (34), which has also been used for oligonucleotide array-based identification of

species of the A. calcoaceticus-A. baumannii complex (129). PCR detection of the bla oxa-51-

like gene has been shown to be a rapid method for identification of A. baumannii isolates

(212).

The ecology of most Acinetobacter species is still poorly resolved. A.

baumannii and Acinetobacter genomic species 3 and 13TU have been mainly recovered from

clinical specimens in hospitals. Human skin carrier rates of A. baumannii outside hospitals

have been shown to be as low as 0.5 to 3% (10, 186), but higher rates (also

for Acinetobacter genomic species 3 and 13TU) have been found in tropical areas (40). A.

baumannii has been isolated from sick animals (15, 226), but an animal or environmental

reservoir has not been found. A. baumannii is, due to its role as a prominent nosocomial

pathogen, the species for which the epidemiology has been studied most intensively.

Epidemic strains of this species can survive well in the environment, as they have been

found on equipment and on environmental surfaces and materials (219), usually in the

vicinity of colonized patients. Multiple sites of the skin and mucosae of patients can be

colonized, and colonization may last days to weeks (61, 142).

Genotyping and Epidemiology

A variety of genotyping methods have been described for differentiation between isolates of

the same species and study of the epidemiology of acinetobacters, in particular that of A.

baumannii. Standardized random amplification PCR-fingerprinting was useful for local typing,

but its (interlaboratory) reproducibility was limited (80). Macrorestriction analysis with

pulsed-field gel electrophoresis allowed for 95% intra- and 89% interlaboratory

reproducibility (187). AFLP fingerprinting also enables genotyping of strains (57, 62, 153),

and its robustness makes it suited for setting up a local database for longitudinal studies.

Genotyping based on the variable number of tandem repeat loci has allowed for additional

subtyping in conjunction with pulsed-field gel electrophoresis analysis (209).

With the introduction of sequence-based methods, it has become possible to set up Internetbased

databases to study the global epidemiology of organisms. Three multilocus sequence

typing systems, mainly aimed at studying the population biology of A. baumannii, have been

developed (6, 65; S. Brisse et al., unpublished data [http://www.pasteur.fr/mlst]). Further

to typing in the strict sense, specific antibiotic resistance genes like the OXA genes, which

confer resistance to carbapenems, are frequently used for additional characterization

of Acinetobacter isolates (39, 130).

Various methods, often in combination, of genotyping A. baumannii isolates from different

institutes and countries have identified three major groups of genetically highly related

strains, the so-called European clones I to III (151, 221). Many of the strains allocated to

these clones are multidrug resistant and have been implicated in outbreaks. Clone I

prevailed in the 1980s, but recent studies indicate that subclones of clone II have emerged

in the United Kingdom, the Czech Republic, and Portugal (50, 152, 208). Identification of

isolates of these clones can be obtained by comparing them to reference sets of the three

clones by AFLP analysis (62, 151). Comparative typing of isolates to only one reference

strain of each clone may lead to under identification of the clones, since one reference strain

does not cover the intraclonal variation. Multilocus sequence typing with seven genes

(http://www.pasteur.fr/mlst) is expected to be the most reliable method for identification of

strains of clones I to III (Brisse et al., unpublished). Rapid assignment to the clones by a

multiplex PCR targeting the ompA, csuE, and bla OXA-51-like gene sequences is promising

(208).

Antimicrobial Susceptibilities

Acinetobacter species are increasingly resistant to multiple antibiotics (108, 134). With the

emergence of carbapenem resistance, a last option for treatment of infections with these

organisms is disappearing. Multidrug resistance is mainly confined to A. baumannii, but

strains of the closely related species Acinetobactergenomic species 3 can also be multidrug

resistant (18, 234). The rates of resistance to different antibiotics can vary among hospitals

and regions, depending on the endemic or epidemic presence of multidrug-resistant A.

baumannii. Resistance mechanisms in A. baumannii comprise all currently known

mechanisms, including enzymatic breakdown, modification of target sites, active efflux, and

decreased influx of antibiotics. The known mechanisms have been reviewed previously (58),

and new mechanisms have been discovered since (2, 171).

Recent genomic studies have shed new light on the genetic organization of resistance

determinants and their transmission. For example, a resistance island integrated within the

ATPase gene has been found in differentA. baumannii strains for which the genome has been

sequenced (1, 71, 109). Among these strains, a variable composition of resistance

determinants interspersed with transposons, integrons, and other mobile elements has been

identified. Other elements, like insertion sequence elements (211), distributed throughout

the genome, are also important for the overall resistance (1).

In vitro determination of antimicrobial susceptibility can be achieved by disk diffusion, agar

dilution, or broth microdilution, as recommended by the Clinical and Laboratory Standards

Institute (CLSI) (203), or by Etest. The panel of tested antibiotics should cover the spectrum

of agents with potential action against A. baumannii,including third- or fourth-generation

cephalosporins, sulbactam, ureidopenicillins, carbapenems, aminoglycosides,

fluoroquinolones, and tetracyclines. Of note, susceptibility to polymyxins, a current last

option for treating pandrug- resistant A. baumannii, should not be tested by disk diffusion

due to poor diffusion of these compounds in agar. Etest and broth microdilution for

determination of the MIC for colistin have been compared and showed a good concordance in

the MIC range of 0.25 to 1 mg/liter (4). In case of carbapenem resistance, the genes

encoding beta-lactamases with carbapenemase activity can be determined by specific PCR

(68), to provide better insight into the epidemiology of the resistance.

Granulibacter bethesdensis

Granulibacter bethesdensis (Acetobacteraceae, Alphaproteobacteria) (79) is a gramnegative,

aerobic, coccobacillary to rod-shaped bacterium, the only species of a new

sublineage within the acetic acid bacteria in the family Acetobacteraceae. This fastidious

organism grows poorly and slowly on sheep blood agar (SBA) at an optimum temperature of

35 to 37°C and an optimum pH of 5.0 to 6.5. It produces a yellow pigment, oxidizes lactate

and weakly acetate to carbon dioxide and water, acidifies ethanol, and can use methanol as

a sole carbon source, all characteristics that distinguish it from other acetic acid bacteria.

The two major fatty acids are C18:1ω7c and C16:0. The DNA base composition is 59.1 mol%

G+C. It was first isolated from three patients with chronic granulomatous disease (79) and

from an additional patient with chronic granulomatous disease more recently (138).

Oxidase-Positive, Indole-Negative, Trypsin-Negative

Nonfermenters

Haematobacter

Three Haematobacter species (Rhodobacteraceae, Alphaproteobacteria) have been

described, i.e., H. massiliensis(former Rhodobacter massiliensis), H.

missouriensis, and Haematobacter genomospecies 1 (Table 2) (84). These species cannot

easily be differentiated phenotypically, and even the 16S rRNA gene sequences are closely

related. Haematobacter species were described as asaccharolytic, but using low-peptone

phenol red agar (seechapter 31), H. missouriensis is clearly saccharolytic, producing acid

from glucose and xylose and sometimes from mannitol, whereas H. massiliensis strains do

not acidify carbohydrates. Acid is produced from ethylene glycol by all species. All the

species are strongly urease and phenylalanine deaminase positive. Arginine dihydrolase is

also positive but sometimes delayed. Asaccharolytic Haematobacter strains

resemblePsychrobacter phenylpyruvicus but can be differentiated by the lack of tributyrine

esterase, the lack of growth improvement by Tween 80, and the presence of arginine

dihydrolase. Differences from Psychrobacter faecalis,Psychrobacter pulmonis, and related

species are the lack of tributyrine and Tween 80 esterase, the lack of nitrate reductase, and

a positive arginine dihydrolase test.

Moraxella

The genus Moraxella comprises approximately 20 species that have been validly named. M.

catarrhalis, M. osloensis, M. nonliquefaciens, and M. lincolnii are part of the normal

microbiota of the human respiratory tract. Animal species include M. bovis, isolated from

healthy cattle and other animals, including horses; M. boevreiand M. caprae (goats); M.

canis (dogs, cats, and camels); M. caviae (guinea pigs); M. cuniculi (rabbits); and M.

ovis and M. oblonga (sheep). The clinical importance of the different species is addressed

below.

Both M. catarrhalis and M. canis grow well on sheep blood agar (SBA) and even on tryptic

soy agar (TSA), and their colonies may reach more than 1 mm in diameter after 24 h of

incubation. Colonies of M. catarrhalis grow well on both blood and chocolate agars, and some

strains also grow well on modified Thayer-Martin and other selective media. Colonies are

generally gray to white, opaque, and smooth and measure about 1 to 3 mm after 24 h of

incubation. Characteristically, the colonies may be nudged intact across the plate with a

bacteriological loop like a “hockey puck” and can be removed from the agar entirely, being

very consistent. Most M. canis colonies resemble those of the Enterobacteriaceae (large,

smooth colonies) and may produce a brown pigment when grown on starch-containing

Mueller-Hinton agar (111). Some strains may also produce very slimy colonies resembling

colonies of Klebsiella pneumoniae (111). M. nonliquefaciens forms smooth, translucent to

semiopaque colonies 0.1 to 0.5 mm in diameter after 24 h and 1 mm in diameter after 48 h

of growth on SBA plates. Occasionally, these colonies spread and pit the agar. The colonial

morphologies of M. lincolnii (217), M. osloensis, and Psychrobacter

phenylpyruvicus (formerly M. phenylpyruvica) are similar, but pitting is rare. On the other

hand, pitting is common with M. lacunata, whose colonies are smaller and form dark haloes

on chocolate agar. Rod-shaped Moraxella species, especially M. atlantae and M. lincolnii, are

more fastidious and display smaller colonies on SBA, less than 1 mm in diameter after 24 h.

Colonies of M. atlantaeare small (usually 0.5 mm in diameter) and show pitting and

spreading (24). The growth of M. atlantae is stimulated by bile salts, which explains its

growth on MacConkey agar. M. nonliquefaciens and M. osloensisproduce colonies that are

somewhat larger than those of M. atlantae and that are rarely pitting. Colonies of M.

nonliquefaciens may be mucoid. A selective medium, acetazolamide agar, inhibiting growth

of neisseriae when incubated in ambient atmosphere, has been described for M.

catarrhalis (231).

Moraxella species are coccoid or coccobacillary organisms (plump rods), occurring

predominantly in pairs and sometimes in short chains, that tend to resist decolorization in

the Gram stain (49). M. canis and M. catarrhalisare Neisseria-like diplococci, and they can

easily be distinguished from other moraxellae or other coccoid species by performing a Gram

stain on cells cultured in the vicinity of a penicillin disk: cells of M. canis and M.

catarrhalis remain spherical diplococci of 0.5 to 1.5 μm in diameter, although of irregular

size, whereas coccobacilli show obviously rod-shaped and filamentous cells.

Moraxella species are asaccharolytic and strongly oxidase positive. M. catarrhalis and M.

canis are also strongly catalase positive, and most strains reduce nitrate and nitrite. M.

catarrhalis and M. canis may be easily distinguished from the commensal Neisseria species,

which are also frequently isolated from respiratory clinical specimens, by the ability of the

former to produce DNase and butyrate esterase (tributyrine test). Rapid butyrate esterase

tests have been described (198), and the indoxyl-butyrate hydrolysis spot test is

commercially available (Remel, Inc., Lenexa, KS). Butyrate esterase is, however, also

present in some otherMoraxella species. M. canis acidifies ethylene glycol and alkalinizes

acetate, in contrast to M. catarrhalis. There are few biochemical differences between M.

catarrhalis and M. nonliquefaciens, which are differentiated from each other mainly on the

basis of morphological characteristics and by nitrite reductase and DNase activity ofM.

catarrhalis.

M. atlantae is the only Moraxella species to be pyrrolidonyl aminopeptidase (17) positive. M.

lacunata is the only proteolytic species with gelatinase activity. Using the plate method

(see chapter 31), gelatin hydrolysis occurs usually within 2 to 4 days. A more rapid and

almost equally specific test to differentiate M. lacunata from other moraxellae is the

detection of Tween 80 esterase activity, which is often positive within 2 days, whereas all

other species, except for very rare M. osloensis strains, remain negative. This species should

also be distinguished from Psychrobacter species, which are also Tween 80 esterase positive,

but P. phenylpyruvicus is urease positive and P. immobilis and related species exhibit

luxuriant growth on plain agar, like TSA, even at 25°C.

M. lincolnii is biochemically quite inactive.

M. osloensis is acetate alkalinization positive, acidifies ethylene glycol, and is resistant to

desferrioxamine (250-μg disk). M. nonliquefaciens has opposite properties to those of M.

osloensis and is, in addition, always nitrate positive.

Antimicrobial Susceptibilities

Most Moraxella species are susceptible to penicillin and its derivatives, cephalosporins,

tetracyclines, quin olones, and aminoglycosides (70, 197). Production of beta- lactamase has

been only rarely reported forMoraxella species other than M. catarrhalis, of which most

isolates produce an inducible, cell-associated beta-lactamase (231). Isolates of M.

catarrhalis are generally susceptible to amoxicillin-clavulanate, expanded-spectrum and

broad-spectrum cephalosporins (i.e., cefuroxime, cefotaxime, ceftriaxone, cefpodoxime,

ceftibuten, and the oral agents cefixime and cefaclor), macrolides (e.g., azithromycin,

clarithromycin, and erythromycin), tetracyclines, rifampin, and fluoroquinolones.

Oligella urethralis and O. ureolytica

The genus Oligella comprises two species, O. ureolytica (formerly CDC group IVe) and O.

urethralis (formerlyMoraxella urethralis and CDC group M-4) (181), which have both been

isolated chiefly from the human urinary tract and have been reported to cause urosepsis

(173). A case of septic arthritis due to O. urethralis has also been reported (144).

Colonies of O. urethralis are smaller than those of M. osloensis and are opaque to whitish.

Colonies of O. ureolytica are slow growing on blood agar, appearing as pinpoint colonies after

24 h but large colonies after 3 days of incubation. Colonies are white, opaque, entire, and

nonhemolytic.

O. ureolytica and O. urethralis are small asaccharolytic coccobacilli that rapidly acidify

ethylene glycol and are susceptible to desferrioxamine. Most strains of O. ureolytica are

motile by peritrichous flagella; all are strongly urease positive (with the urease reaction often

turning positive within minutes after inoculation) and reduce nitrate. Oligella

urethralis strains are nonmotile and urease and nitrate reductase negative, but they reduce

nitrite and are weakly phenylalanine deaminase positive. Bordetella

bronchiseptica and Cupriavidus pauculus are also rapidly urease positive but are

desferrioxamine resistant.

O. urethralis and M. osloensis have biochemical similarities, e.g., accumulation of poly-β-

hydroxybutyric acid and failure to hydrolyze urea, but can be differentiated on the basis of

nitrite reduction and alkalinization of formate, itaconate, proline, and threonine, all positive

for O. urethralis (169). Moreover, O. urethralis is susceptible to desferrioxamine and

tributyrate esterase is negative, in contrast to M. osloensis.

O. urethralis is generally susceptible to most antibiotics, including penicillin, while O.

ureolytica exhibits variable susceptibility patterns (70).

Psychrobacter

The genus Psychrobacter (117) comprises more than 30 species, of which only a few are

clinically important. Apart from Psychrobacter phenylpyruvicus, the Psychrobacter strains

isolated from clinical material were considered until recently as belonging to the

species Psychrobacter immobilis. In a recent study, 16Psychrobacter isolates of clinical origin

were analyzed. Ten were identified as P. faecalis, four were identified asP. pulmonis, and two

could not be identified but clustered close to Psychrobacter when the 16S rRNA gene

sequence was determined (G. Wauters, unpublished data). These findings suggest that the

majority of the clinical isolates belong to P. faecalis and P. pulmonis, both first described to

occur in animals (pigeons [118] and lambs [235], respectively). P. immobilis itself is

apparently rarely isolated, if at all, from humans.

P. faecalis and P. pulmonis are coccoid gram-negative rods growing on TSA with large,

creamy colonies. P. faecalis is saccharolytic and acidifies glucose and xylose, while P.

pulmonis is asaccharolytic. Both species produce acid from ethylene glycol. They are Tween

80 esterase and tributyrate esterase positive. They are nitrate reductase positive and, unlike

the type strain of P. immobilis, are urease negative and nitrite reductase positive. Colonies

may resemble those of Haematobacter, but the latter lack nitrate reductase, Tween 80

esterase, and tributyrin esterase and are strongly urease positive, arginine dihydrolase

positive, and phenylalanine deaminase positive.

One case of ocular infection (76) and one case of infant meningitis (137) have been reported

to be caused byP. immobilis, but in light of the data reported here, this might concern

infection with one of the otherPsychrobacter species.

P. phenylpyruvicus, formerly Moraxella phenylpyruvica (25), has the morphological and

cultural appearance of moraxellae but is urease and phenylalanine deaminase positive. A

unique feature of the species is its marked growth improvement by Tween 80. Colonies on

TSA with 1% Tween 80 have a size two to three times larger than on SBA. The

other Psychrobacter species, in contrast to P. phenylpyruvicus, grow abundantly on ordinary

media such as TSA, and their growth is not promoted by Tween 80. They

resemble Haematobacter species.Psychrobacter species are resistant to penicillin but

susceptible to most other antibiotics (76, 137).

Oxidase-Positive, Indole-Negative, Trypsin-Positive

Nonfermenters

Alishewanella fetalis

Alishewanella fetalis (Alteromonadaceae, Gammaproteobacteria) (Table 3) is a gramnegative

rod that grows at temperatures between 25 and 42°C, with optimum growth at

37°C. A. fetalis can withstand NaCl concentrations of up to 8% but not 10%, which helps

differentiate this species from Shewanella algae, which can grow in 10% NaCl (240). Also, in

contrast to Shewanella species, it does not produce H2S in the butt of TSI and KIA. The type

strain tested by us acidifies glucose and does not hydrolyze esculin.

Shewanella algae and S. putrefaciens

The organisms formerly called Pseudomonas putrefaciens, Alteromonas

putrefaciens, Achromobacter putrefaciens,and CDC group Ib have been placed in the

genus Shewanella (140), which comprises over 50 species. S. putrefaciens was described

with two CDC biotypes. CDC biotype 1 was later described as S. putrefaciens sensu stricto,

whereas CDC biotype 2 was subsequently assigned to a new species, S. alga (158), later

corrected toS. algae.

Colonies of Shewanella species on SBA are convex, circular, smooth, and occasionally

mucoid, produce a brown to tan soluble pigment, and cause green discoloration of the

medium. Cells are long, short, or filamentous, reminiscent of Myroides. Motility is due to a

single polar flagellum.

Most strains of both Shewanella species produce H2S in KIA and TSI agar, a unique feature

among clinically relevant nonfermenters. Both are also ornithine decarboxylase positive and

have strong alkaline phosphatase, strong trypsin, and strong pyrrolidonyl aminopeptidase

activities. S. algae is halophilic, asaccharolytic, and requires NaCl for growth, with growth

occurring already on TSA plus 0.5% NaCl. S. putrefaciens does not require NaCl for growth

and is saccharolytic, producing acid from maltose and sucrose, and irregularly and weakly

from glucose.

Khashe and Janda (126) have reported that S. algae is the predominant human clinical

isolate (77%), while S. putrefaciens represents the majority of nonhuman isolates (89%).

Although infrequently isolated in the clinical laboratory, S. putrefaciens and S. algae have

been recovered from a wide variety of clinical specimens and are associated with a broad

range of human infections, including skin and soft tissue infections (36), otitis media (99),

ocular infection (28), osteomyelitis (20), peritonitis (46), and septicemia (110). The habitat

for S. algae is saline, whereas S. putrefaciens has been isolated mostly from fish, poultry,

and meats as well as from freshwater and marine samples.

Shewanella species are generally susceptible to most antimicrobial agents effective against

gram-negative rods, except penicillin and cephalothin (70, 241). The mean MICs of S.

algae for penicillin, ampicillin, and tetracycline are higher than the corresponding MICs of S.

putrefaciens (126, 239).

Sphingobacterium

A total of 15 species have been described as belonging to the genus Sphingobacterium.

Based on 16S rRNA gene sequence data, the indole-producing Flavobacterium

mizutaii belongs to the genus Sphingobacterium (G. Wauters and M. Vaneechoutte,

unpublished observation) and should be transferred to the

genusSphingobacterium as Sphingobacterium mizutaii. As a consequence, the description of

the genusSphingobacterium as indole negative will have to be emended.

The species of the genus Sphingobacterium encountered in clinical material include S.

multivorum (formerlyFlavobacterium multivorum and CDC group IIk-2), S.

spiritivorum (including the species formerly designated asFlavobacterium spiritivorum, F.

yabuuchiae, and CDC group IIk-3), S. thalpophilum, and Flavobacterium mizutaii(205, 249).

Colonies are yellowish. Sphingobacterium species are middle-sized, nonmotile gram-negative

rods. Species of this genus do not produce indole, but Flavobacterium mizutaii is indole

positive and is therefore dealt with among the indole-positive nonfermenters in Table 4. All

species are strongly saccharolytic; i.e., glucose, xylose, and other sugars are acidified. No

acid is formed from mannitol, except by S. spiritivorum, which is also the only species to

produce acid from ethylene glycol. S. thalpophilum can be distinguished from

otherSphingobacterium species by its nitrate reductase and its growth at 42°C.

S. multivorum is the most common human species. It has been isolated from various clinical

specimens but has only rarely been associated with serious infections (peritonitis and sepsis)

(73, 91). Blood and urine have been the most common sources for the isolation of S.

spiritivorum (90). F. mizutaii has been isolated from blood, cerebrospinal fluid (CSF), and

wound specimens (247). S. thalpophilum has been recovered from wounds, blood, eyes,

abscesses, and an abdominal incision (247).

Sphingobacterium species are generally resistant to amino glycosides and polymyxin B while

susceptible in vitro to the quinolones and trimethoprim-sulfamethoxazole. Susceptibility to

beta-lactam antibiotics is variable, requiring testing of individual isolates (197).

Sphingomonas Species

On the basis of 16S rRNA gene sequence and the presence of unique sphingoglycolipid and

ubiquinone types, the genus Sphingomonas (Sphingomoadaceae, Alphaproteobacteria) was

created for organisms formerly known as Pseudomonas paucimobilis and CDC group IIk-1

(89, 250). Since the original proposal, a total of almost 60 novel species, originating from

various environments, have been added to the genus Sphingomonas. The former

genus Sphingomonas can be divided into four phylogenetic groups, each representing a

different genus (204), whereby the emended genus Sphingomonas contains at least 12

species, of which only S. paucimobilisand S. parapaucimobilis are thought to be clinically

important. However, recent 16S rRNA gene sequencing of 12 strains of clinical origin

(Wauters, unpublished) revealed that several named and unnamed Sphingomonasspecies

were present, but no S. paucimobilis and only two S. parapaucimobilis isolates. Because

many phenotypic characteristics are shared by these species, routine laboratories best report

them asSphingomonas species.

Sphingomonas colonies are slow growing on blood agar medium, with small colonies

appearing after 24 h of incubation. Growth occurs at 37°C but not at 42°C, with optimum

growth at 30°C. Almost all strains produce a yellow insoluble pigment, different from

flexirubin pigments, as can be established by the KOH test (11). Few strains are

nonpigmented or develop a pale yellow color after several days. Older colonies demonstrate

a deep yellow (mustard color) pigment.

Sphingomonas species are medium to long motile rods with a single polar flagellum. Motility

occurs at 18 to 22°C but not at 37°C. However, few cells are actively motile in broth culture,

thus making motility a difficult characteristic to demonstrate.

Oxidase is only weakly positive or even absent. All the strains are saccharolytic, but some

acidify glucose only weakly and slowly. Urease is always negative, and nitrate reduction is

only very rarely positive. Esculin is hydrolyzed, and beta-galactosidase and alkaline

phosphatase are positive. The yellow pigment of some strains may hamper a correct reading

of the yellow color shift when nitrophenyl compounds of the latter substrates are used.

Members of this genus are known as decomposers of aromatic compounds and are being

developed for use in bioremediation.

Sphingomonas species are widely distributed in the environment, including water, and have

been isolated from a variety of clinical specimens, including blood, CSF, peritoneal fluid,

urine, wounds, the vagina, and the cervix, as well as from the hospital environment

(103, 148, 175). S. parapaucimobilis clinical isolates have been obtained from sputum, urine,

and the vagina (250).

Most strains are resistant to colistin, but all are susceptible to vancomycin, which is

exceptional for gram- negative nonfermenting rods. This is elsewhere only found

in Chryseobacterium and related genera

likeElizabethkingia and Empedobacter. Most Sphingomonas strains are susceptible to

tetracycline, chloramphenicol, trimethoprim-sulfamethoxazole, and aminoglycosides.

Susceptibility to other antimicrobial agents, including fluoroquinolones, varies

(70, 103, 175).

Oxidase-Positive, Indole-Positive Nonfermenters

The natural habitats of most oxidase-positive, indole-positive nonfermenters (Table 4) are

soil, plants, and food and water sources, including those in hospitals. Clinically relevant

species include Chryseobacteriumspecies, Elizabethkingia meningoseptica, Empedobacter

brevis, Wautersiella falsenii, Flavobacterium mizutaii,Weeksella virosa, Bergeyella

zoohelcum, and Balneatrix alpica. All are indole, trypsin, pyrrolidonyl aminopeptidase, and

alkaline phosphatase positive, except for B. zoohelcum, which is pyrrolidonyl aminopeptidase

negative, and B. alpica, which is both trypsin and pyrrolidonyl aminopeptidase

negative. Table 4 presents an overview of the characteristics useful to differentiate among

these species.

Balneatrix alpica

B. alpica was first isolated in 1987 during an outbreak of pneumonia and meningitis among

persons who attended a hot (37°C) spring spa in southern France (51). Isolates from eight

patients were recovered from blood, CSF, and sputum, and one was recovered from water.

This species is only rarely isolated from human clinical specimens.

B. alpica produces colonies that are 2 to 3 mm in diameter, convex, and smooth. The center

of the colonies is pale yellow after 2 to 3 days and pale brown after 4 days. B. alpica is a

straight or curved gram-negative rod. It is the only motile species among the clinically

relevant indole-positive nonfermenters. Cells have one or two polar flagella.

The species is strictly aerobic and saccharolytic. Both trypsin and pyrrolidonyl

aminopeptidase are negative, unlike with other indole-positive nonfermenters. Growth occurs

at 20 to 46°C on ordinary media such as TSA but not on MacConkey agar. It acidifies

glucose, mannose, fructose, maltose, sorbitol, mannitol, glycerol, inositol, and xylose. B.

alpica is nitrate reductase and weakly gelatinase positive. It is similar to E.

meningoseptica but can be differentiated from this species by its motility and nitrate

reductase and by the absence of beta-galactosidase.

B. alpica has been reported to be susceptible to penicillin G and all other beta-lactam

antibiotics and to all aminoglycosides, chloramphenicol, tetracycline, erythromycin,

sulfonamides, trimethoprim, ofloxacin, and nalidixic acid. It is resistant to clindamycin and

vancomycin (30).

Bergeyella zoohelcum

Bergeyella zoohelcum and Weeksella virosa are morphologically and biochemically similar

organisms with cells measuring 0.6 by 2 to 3 μm, with parallel sides and rounded ends. B.

zoohelcum colonies are sticky and tan to yellow.

Both species fail to grow on MacConkey agar and are nonsaccharolytic. Both species are

susceptible to desferrioxamine and have the unusual feature of being susceptible to

penicillin, a feature that allows them to be easily differentiated from the related

genera Chryseobacterium and Sphingobacterium. B. zoohelcum can be differentiated

from W. virosa because it is pyrrolidonyl aminopeptidase negative, strongly urease positive,

and resistant to colistin. B. zoohelcum comprises formerly CDC group IIj strains (97).

B. zoohelcum is isolated mainly from wounds caused by animal (mostly dog) bites (97, 176).

Meningitis or septicemia due to B. zoohelcum has occurred in patients either bitten by a dog

(146) or with continuous contact with cats (157).

Both B. zoohelcum and W. virosa are susceptible to most antibiotics. However, at present no

specific antibiotic treatment is recommended, and antimicrobial susceptibility testing should

be performed on significant clinical isolates.

Chryseobacterium

CDC group IIb comprises the species Chryseobacterium indologenes, C. gleum, and other

strains, which probably represent several unnamed taxa.

Strains included in CDC group IIb are nonmotile rods. Cells of C. indologenes are similar to

those of E. meningoseptica, C. anthropi, C. hominis, and F. mizutaii; i.e., they are thinner in

their central than in their peripheral portions and include filamentous forms.

CDC group IIb strains are oxidase and catalase positive, produce flexirubin pigments

(11, 168), are moderately saccharolytic, and are esculin and gelatin hydrolysis positive. C.

indologenes and C. gleum can easily be differentiated from each other by four

characteristics: C. indologenes displays a broad beta-hemolysis area within 3 days of

incubation at 37°C on SBA, is always arabinose negative, does not acidify ethylene glycol,

and does not grow at 42°C (214). C. gleum exhibits pronounced alpha-hemolysis,

resembling viridans discoloration; always acidifies ethylene glycol; is arabinose positive or

delayed positive; and grows at 42°C.

Beta-hemolysis is absent or very rare in other strains of CDC group IIb and is therefore

almost specific for the identification of C. indologenes, while the profile of C. gleum may be

shared by other strains of this group. It should be noted that some C. indologenes strains do

not produce flexirubin.

Among CDC group IIb species, C. indologenes is usually considered most frequently isolated

from clinical samples, although it rarely has clinical significance (241). It causes bacteremia

in hospitalized patients with severe underlying disease, although the mortality rate is

relatively low even among patients who were administered antibiotics without activity

against C. indologenes (195). Nosocomial infections due to C. indologenes have been linked

to the use of indwelling devices during hospital stays (7, 102, 159).

Still, the frequency of C. indologenes as reported in the literature should be interpreted with

caution, because until recently and without molecular biology, C. indologenes could almost

not be distinguished routinely from other CDC group IIb strains. We have recently examined

21 CDC group IIb strains both phenotypically and by 16S rDNA sequencing and found 9 C.

indologenes isolates, 5 C. gleum isolates, and 7 isolates belonging to

unnamed Chryseobacterium species.

The production of novel types of metallo-beta-lactamases from C. indologenes has been

studied in detail (136,166).

Chryseobacterium anthropi represents part of the strains formerly designated as CDC group

IIe (120). Most strains display very sticky colonies, which are nonpigmented but may

develop a slightly salmon-pinkish, rarely yellowish color after a few days. In contrast to C.

hominis, the species is negative for esculin hydrolysis and acidification of ethylene glycol. In

addition, many strains are susceptible to desferrioxamine. One case of meningitis caused by

CDC group IIe has been reported (245). Most clinical isolates used for the description of the

species were from wounds and blood cultures (120).

Chryseobacterium hominis includes the strains formerly included in CDC group IIc and most

of the strains of CDC group IIh (228). This species does not produce flexirubin pigments, but

some strains exhibit a slightly yellowish pigmentation. Colonies are often mucoid. C.

hominis can be differentiated from C. gleum by the absence of flexirubin pigments and the

lack of acid production from arabinose. C. indologenes strains lacking flexirubin pigments

may resemble C. hominis, but the latter is never beta-hemolytic and always acidifies

ethylene glycol.

Many strains have been isolated from blood. Others have been isolated from dialysis fluid,

pus, the eye, infraorbital drain, and aortic valve, but their clinical significance remains to be

assessed (228).

Elizabethkingia meningoseptica and E. miricola

Colonies of Elizabethkingia meningoseptica, formerly Chryseobacterium

meningosepticum (127), are smooth and fairly large, either nonpigmented or producing a

pale yellow or slightly salmon-pinkish pigment after 2 or 3 days. Characteristic features are

acid production from mannitol and beta-galactosidase activity. Gelatin and esculin hydrolysis

are positive. Elizabethkingia and Chryseobacterium species can be differentiated as well on

the basis of 16S rRNA sequence analysis (120, 127).

E. meningoseptica has been reported to be associated with (neonatal) meningitis and

nosocomial outbreaks (14, 33, 38, 106, 195, 207), endocarditis (16), cystic fibrosis airway

infections (131), retroperitoneal hematoma (133), community- acquired osteomyelitis (132),

adult pneumonia and septicemia (14, 135, 192, 241), respiratory colonization and infection

following aerosolized polymyxin B treatment (26), and infections reported in dialysis units

(135, 165). A clinical case of E. miricola was reported only once, in a case of sepsis (77).

Empedobacter brevis

Empedobacter brevis (216) colonies are yellowish pigmented but do not produce

flexirubin. E. brevis can be differentiated from C. indologenes, C. gleum, other CDC group IIb

strains, and C. hominis by its lack of esculin hydrolysis. Growth on MacConkey agar and a

stronger gelatinase activity are useful to distinguish it from C. anthropi. The species is rarely

recovered from clinical material.

Flavobacterium mizutaii

F. mizutaii is saccharolytic, producing acid from a large number of carbohydrates, including

xylose, similar toSphingobacterium species, from which it can be distinguished by its indole

production and by its failure to grow on MacConkey agar and its usual lack of urease activity

(247).

F. mizutaii can be distinguished from Chryseobacterium and Empedobacter species by its

lack of gelatin hydrolysis and of flexirubin production. F. mizutaii produces acid from xylose

but not from ethylene glycol, allowing differentiation from other indole-positive species. The

phenotypic profile of F. mizutaii is similar to that of the strains described

as Chryseobacterium CDC group IIi. Furthermore, 16S rRNA gene sequencing confirms that

most CDC group IIi strains actually belong to the species F. mizutaii.

F. mizutaii has been described as an indole-negative species (249), but in our hands all

strains tested, including the type strain, produce as much indole as

the Chryseobacterium strains. According to 16S rRNA gene sequencing, this species is closely

related to Sphingobacterium species, indicating that F. mizutaii—formerlySphingobacterium

mizutae (98)—should be transferred back to the genus Sphingobacterium as S. mizutaii. F.

mizutaii has been isolated from blood, CSF, and wound specimens (247).

Wautersiella falsenii

Wautersiella falsenii is closely related to E. brevis, from which it differs by its urease activity.

Two genomovars have been described (119): genomovar 1 is always esculin positive and

beta-galactosidase negative, whereas 90% of the genomovar 2 strains are esculin negative

and 63% are beta-galactosidase positive.

W. falsenii was described as belonging to a separate genus from Empedobacter, based on

comparison of its 16S rRNA gene sequence with an E. brevis EMBL sequence of poor quality.

A high-quality sequence of the rRNA gene of the type strain of E. brevis indicates that W.

falsenii probably has to be renamed as Empedobacter falsenii.

W. falsenii is much more frequently isolated from clinical samples than E. brevis (119). Its

clinical significance remains to be assessed.

Weeksella virosa

W. virosa colonies are mucoid and adherent to the agar, reminiscent of the sticky colonies

of B. zoohelcum. Colonies are not pigmented after 24 h of incubation but may become

yellowish, tan to brown, after 2 or 3 days. The cellular morphology of Weeksella virosa is

dealt with above in the discussion of Bergeyella zoohelcum. W. virosa can be differentiated

from B. zoohelcum because it is urease negative and polymyxin B and colistin susceptible,

whereas B. zoohelcum is rapid urease positive and polymyxin B and colistin resistant. W.

virosacomprises formerly CDC group IIf strains (96). W. virosa is isolated mainly from urine

and vaginal samples (96,177), in contrast to B. zoohelcum, which is isolated mostly from

animal bites.

The appropriate choice of effective antimicrobial agents for the treatment of chryseobacterial

infections is difficult (106). Chryseobacterium species and E. meningoseptica are inherently

resistant to many antimicrobial agents commonly used to treated infections caused by gramnegative

bacteria (aminoglycosides, beta-lactam antibiotics, tetracyclines, and

chloramphenicol) but are often susceptible to agents generally used for treating infections

caused by gram-positive bacteria (rifampin, clindamycin, erythromycin, trimethoprimsulfamethoxazole,

and vancomycin) (70, 197, 241). Although early investigators

recommended vancomycin for treating serious infection with E. meningoseptica (83),

subsequent studies showed greater in vitro activity of minocycline, rifampin, trimethoprimsulfamethoxazole,

and quin olones (14, 72, 197). Among the quinolones, levofloxacin is

more active than ciprofloxacin and ofloxacin (197). C. indologenes is reported to be

uniformly resistant to cephalothin, cefotaxime, ceftriaxone, aztreonam, aminoglycosides,

erythromycin, clindamycin, vancomycin, and teicoplanin, while susceptibility to piperacillin,

cefo perazone, ceftazidime, imipenem, quinolones, minocycline, and trimethoprimsulfamethoxazole

is variable, requiring testing of individual isolates (133, 197, 243). Several

studies reported that administration of quinolone, minocycline, trimethoprimsulfamethoxazole,

or rifampin, and treatment of local infection improve the clinical outcome

of patients with E. meningoseptica infections. The choice of appropriate antimicrobial therapy

is further complicated by the fact that MIC breakpoints for resistance and susceptibility of

chryseobacteria have not been established by the CLSI and the results of disk diffusion

testing are unreliable in predicting antimicrobial susceptibility ofChryseobacterium species

(35, 72, 243). The Etest is a possible alternative to the standard agar dilution method for

testing cefotaxime, ceftazidime, amikacin, minocycline, ofloxacin, and ciprofloxacin but not

piperacillin (101). Definitive therapy for clinically significant isolates should be guided by

individual susceptibility patterns determined by an MIC method.

ANTIMICROBIAL SUSCEPTIBILITIES Back to top

Decisions about performing susceptibility testing are complicated by the fact that the CLSI

interpretive guidelines for disk diffusion testing of the nonfermenting gram-negative bacteria

are limited to Pseudomonasspecies, Burkholderia cepacia, Stenotrophomonas

maltophilia, and Acinetobacter species and therefore, except for Acinetobacter species, do

not include the organisms covered in this chapter. Furthermore, results obtained with,

e.g., Acinetobacter species by using disk diffusion do not correlate with results obtained by

conventional MIC methods. In general, laboratories should try to avoid performing

susceptibility testing on the organisms included in this chapter. When clinical necessity

dictates that susceptibility testing be performed, an overnight MIC method, e.g., Etest

(bioMerieux) (101), is recommended.

EVALUATION, INTERPRETATION, AND REPORTING OF

RESULTS Back to top

Although certain nonfermenting bacteria can on occasion be frank pathogens,

e.g., Pseudomonas aeruginosa, Burkholderia pseudomallei, and Elizabethkingia

meningoseptica, they are generally considered to be of low virulence and often occur in

mixed cultures, making it difficult to determine when to work up cultures and when to

perform susceptibility studies. Elizabethkingia meningoseptica in neonatal

meningitis, Moraxella lacunata in eye infections, and M. catarrhalis in respiratory tract

infections should be reported as significant pathogens. Direct Gram stain interpretation of

clinical specimens may be of limited importance, because these organisms often occur in

mixed infections and because their clinical importance has to be interpreted taking into

account the considerations discussed below. Decisions regarding the significance of GNF in a

clinical specimen must take into account the clinical condition of the patient and the source

of the specimen submitted for culture. In general, the recovery of a GNF in pure culture from

a normally sterile site warrants identification and susceptibility testing, whereas predominant

growth of a GNF from a nonsterile specimen, such as an endotracheal culture from a patient

with no clinical signs or symptoms of pneumonia, would not be worked up further. Because

many GNF exhibit multiple-antibiotic resistance, patients who are on antibiotics often

become colonized with GNF. GNF species isolated in mixed cultures can usually be reported

by descriptive identification, e.g., “growth of P. aeruginosa and two varieties of

nonfermenting gram-negative rods not further identified.”