Burkholderia, Stenotrophomonas & Others

TAXONOMY Back to top

In 1973, the taxonomic heterogeneity of the genus Pseudomonas was revealed by the work

of Palleroni and coworkers, who identified five major species clusters (referred to as rRNA

homology groups) among the pseudomonads (187). DNA-rRNA hybridization experiments led

to the gradual dissection of the genus during the following decades (144). The

name Pseudomonas was confined to rRNA homology group I organisms because they

comprised the type species, Pseudomonas aeruginosa (see chapter 33).

The nomenclatural rearrangements of the genus Pseudomonas entailed the creation of

several new genera. Some of these encompassed complete rRNA homology groups (e.g.,

both rRNA homology group IV species were reclassified in the genus Brevundimonas),

whereas others encompassed only partial groups. rRNA group II pseudomonads belong to

the class of the Gammaproteobacteria and were reclassified into the

generaBurkholderia and Ralstonia (262, 263). rRNA group II pseudomonads form a

remarkable group of primary and opportunistic human, animal, and plant pathogens, as well

as environmental species with a considerable potential for biological control, remediation,

and plant growth promotion. During the past decade, the interest in several peculiar

characteristics of these organisms led to the discovery and description of a multitude of

novel species. The genus Burkholderia now contains about 60 validly named species, many

of which have been isolated from soil and water samples. Some other novel Burkholderia-like

species were found to represent a distinct phylogenetic lineage with a position intermediate

between those of the genera Burkholderia andRalstonia and were classified in the novel

genus Pandoraea (52).

Comparative 16S rRNA gene sequence analysis, further supported by phenotypic differences,

indicated that two distinct sublineages existed within the genus Ralstonia (229). It was

proposed that species of theRalstonia eutropha lineage be classified in a novel genus

named Wautersia, whereas the name Ralstonia was preserved for the lineage

comprising Ralstonia pickettii, the type species. Shortly thereafter, it became known (225)

that Wautersia eutropha, the type species of the genus Wautersia, was a junior synonym

of Cupriavidus necator, the type (and only) species of the genus Cupriavidus, an

environmental organism which was validly named in 1987, i.e., long before 16S rRNA gene

sequence studies were performed routinely (168). To conform to the International Code of

Nomenclature of Bacteria (212), the name Wautersia had to be replaced byCupriavidus and

all species of the genus Wautersia became species of the genus Cupriavidus.

Several Burkholderia species have been isolated from human clinical samples, but

only Burkholderia cepaciacomplex, B. gladioli (including strains previously classified as B.

cocovenenans [56]), B. mallei, and B. pseudomallei are generally recognized as human or

animal pathogens. Recent polyphasic taxonomic studies including traditional taxonomic

methods, but also multilocus sequence typing (MLST) and whole-genome studies, revealed

that B. cepacia-like bacteria belong to at least 17 distinct genomic species (genomovars),

referred to collectively as the B. cepacia complex (66, 228, 230, 231). Ongoing surveys of

the diversity of B. cepacia-like bacteria recovered from specimens from cystic fibrosis (CF)

patients and other specimens revealed the presence of several additional groups in the B.

cepacia complex that cannot be assigned to one of the established species within this

complex by using traditional or molecular identification approaches (200). Further polyphasic

taxonomic analyses are needed to determine if these groups represent additional novel

species within the B. cepacia complex or if they represent new variants of established

species. With the exception of B. ubonensis, all B. cepacia complex species have been

recovered from human clinical samples. Within this group, however, B. multivorans and B.

cenocepacia are the most common opportunistic pathogens in CF patients (112, 200).

Apart from the B. cepacia complex species, B. gladioli, B. mallei, and B. pseudomallei, the

genus Burkholderia now comprises an additional 39 validly named species. Most of these

organisms are not associated with human disease and are not discussed further here.

Species that rarely have been reported as associated with human infections include B.

fungorum, B. glumae, and B. thailandensis (21, 57, 240). A complete overview of validly

named species can be obtained through Internet sites such

as http://www.bacterio.cict.fr/ orhttp://www.dsmz.de/bactnom/genera1.htm.

There are now five species in the genus Ralstonia. The human pathogens include Ralstonia

pickettii, R. mannitolilytica (previously known as R. pickettii biovar 3/‘thomasii’) (85), and R.

insidiosa (54). Ralstonia paucula(previously known as Centers for Disease Control group IVc-

2) (226), Ralstonia gilardii (53), R. respiraculi (68),R. taiwanensis (35), and seven additional

species, which occur primarily in environmental samples, are now all classified

as Cupriavidus species (225).

Five distinct species of Pandoraea, Pandoraea apista (the type species), P. pulmonicola, P.

pnomenusa, P. sputorum, and P. norimbergensis, and four presently unnamed strains, each

representing a distinct additionalPandoraea species, have been reported (52, 84). Most of

these occur in human clinical specimens.

Organisms in the Pseudomonas rRNA homology group III also belong to

the Betaproteobacteria and are now classified in the family Comamonadaceae, which

includes the genera Comamonas, Delftia, and Acidovorax (242,249). The

genus Comamonas was originally created in 1985 and included a single

species, Comamonasterrigena. Two years later, Pseudomonas acidovorans and Pseudomonas

testosteroni were reclassified as members of the

genus Comamonas. Comamonas acidovorans was subsequently again reclassified as Delftia

acidovorans (242). Comamonas terrigena encompassed three strain clusters with human

clinical isolates (251), which are now known as C. terrigena,

Comamonas aquatica, and Comamonas kerstersii (239). Additional novel species have been

isolated from environmental samples (265).

Originally, Acidovorax facilis was classified as Hydrogenomonas facilis based on its ability to

oxidize hydrogen. Poly-β-hydroxybutyrate metabolism studies resulted in the transfer of this

species to the genus Pseudomonas,along with a new species called Pseudomonas delafieldii.

A new genus, Acidovorax, was proposed, which included three species, A. facilis, A.

delafieldii, and A. temperans, all members of rRNA homology group III (250). An additional

five plant- pathogenic pseudomonads and novel environmental species have now been

classified as Acidovorax species (100).

The genus Brevundimonas, consisting of the species Brevundimonas

diminuta and Brevundimonas vesicularis,was proposed for bacteria originally classified as

members of Pseudomonas rRNA homology group IV (206) and is a member of

the Alphaproteobacteria. This genus currently comprises 14 validly named species, most of

which are of environmental origin.

Finally, Pseudomonas maltophilia represented Pseudomonas rRNA homology group V (128).

Based on genotypic and phenotypic characteristics, its transfer to the genus Xanthomonas, a

member of the Gammaproteobacteria,was proposed (217). However, many differences were

also noted, including number of flagella, nitrate reduction characteristics, fimbriation, and

plant pathogenicity. Therefore, the organism was once again reclassified in a novel

genus, Stenotrophomonas (186). More recently, a novel species, Stenotrophomonas

africana, was proposed (94). However, Coenye et al. (69) demonstrated that S.

maltophilia and S. africanarepresented the same species, and nomenclatural priority was

given to the former. Seven additional environmental Stenotrophomonas species were

described recently (124).

DESCRIPTION OF THE AGENTS Back to top

Burkholderia, Ralstonia, Cupriavidus, Pandoraea, Brevundimonas, Comamonas,

Delftia, and Acidovorax spp. are aerobic, non-spore-forming, straight or slightly curved

gram-negative rods. They are 1 to 5 μm in length and 0.5 to 1.0 μm in width

(126). Stenotrophomonas spp. are straight rods and tend to be slightly smaller than

members of the other genera (0.7 to 1.8 μm in length and 0.4 to 0.7 μm in width) (126).

With the exception ofB. mallei, these organisms are motile due to the presence of one or

more polar flagella (185). These bacteria are catalase positive, and most, with the exception

of Stenotrophomonas and B. gladioli, are either weakly or strongly oxidase positive. All grow

on MacConkey agar, except for certain strains of B. vesicularis, and appear to be

nonfermenters. The majority of species degrade glucose oxidatively, and most degrade

nitrate to either nitrite or nitrogen gas. Certain species have distinctive colony morphologies

or pigmentation. They are nutritionally quite versatile, with different species being able to

utilize a variety of simple and complex carbohydrates, alcohols, and amino acids as carbon

sources. Certain species can multiply at 4°C, but most are mesophilic, with optimal growth

temperatures between 30 and 37°C (185). For some genera, growth at higher temperatures

(i.e., 42°C) can be useful for species identification.

EPIDEMIOLOGY AND TRANSMISSION Back to top

Burkholderia, Ralstonia, Cupriavidus, Pandoraea, Comamonas, Delftia,

Acidovorax, Brevundimonas, andStenotrophomonas spp. are environmental organisms found

in water, soil, the rhizosphere, and in and on plants including fruits and vegetables. They

have a worldwide distribution. Members of these genera are widely recognized as

phytopathogens, and many species were first described in that context. Because of their

ability to survive in aqueous environments, these organisms have become particularly

problematic as opportunistic nosocomial pathogens in hospitals and health care settings.

The natural distribution of B. cepacia complex species is being intensively studied because of

interest in their biotechnological properties and their pathogenicity in persons with CF

(112, 157). B. cepacia complex bacteria often have antifungal, antinematodal, or plant

growth-promoting properties, which makes them attractive as biological pesticides and

fertilizers (188). Because of their nutritional versatility, B. cepacia complex bacteria also

have applications for bioremediation of contaminated soils. Unlike P. aeruginosa, B.

cepacia complex bacteria are rarely recovered from environmental sites such as sinks,

swimming pools, showers, and salad bars (112, 174). However, some species within the

complex, especially B. ambifaria, B. anthina, and B. pyrrocinia, are frequently recovered

from soil and environmental water samples (13, 197), provided that appropriate growth

conditions are used to inhibit the growth of vast numbers of other environmental bacteria. Of

particular note, B. multivorans, which is among the most common B. cepacia complex

species recovered from CF patients, has been relatively infrequently recovered from

environmental sources. Studies of a variety of foodstuffs and bottled water have shown

that B. cepacia complex bacteria have been found in unpasteurized dairy products (18, 173).

Due to their intrinsic resistance to antibiotics and disinfectants, B. cepacia complex bacteria

are also notorious contaminants of pharmaceutical preparations and medical equipment such

as nebulizers, which may be sterilized with contaminated anti-infectives (129, 148, 182).

Genotypic and conventional epidemiologic investigations provide compelling evidence for

interpatient transmission of common or epidemic B. cepacia complex strains among persons

with CF (155). One such strain, referred to as the ET12 (for electrophoretic type 12) lineage,

is common among CF patients in eastern Canada and the United Kingdom (137, 195). This

organism is a B. cenocepacia strain that is characterized by the presence of a distinctive

cablelike pilus and an associated adhesin that mediates adherence to respiratory epithelium

(203). B. cenocepacia strain PHDC dominates among infected CF patients in the mid-Atlantic

region of the United States and has recently been identified in agricultural soil as well as in

CF patients in several European countries (34, 65, 161). B. cenocepacia‘Midwest clone’ is

common among CF patients in the midwestern United States (58).

B. pseudomallei and B. thailandensis are found in soil and surface water primarily in tropical

and subtropical areas. Both species have been isolated in the rice-growing regions of

northeast Thailand, western Cambodia, Laos, and southern and central Vietnam

(30, 189, 260). In northern Australia, associations have been made between B.

pseudomallei and native grasses in undisturbed land and the presence of livestock animals,

lower soil pH, and different combinations of soil texture and color in environmentally

disturbed sites (140). Recent environmental studies have identified B. thailandensis-like

organisms from Australia (102).

The known endemic distribution of B. pseudomallei is being expanded beyond the traditional

regions of endemicity for melioidosis in Southeast Asia and northern Australia, with recent

case reports of the disease from the Americas, Madagascar, Mauritius, India, and elsewhere

in south Asia, China, and Taiwan. To what extent this reflects a true expansion of endemicity

rather than unmasking of the long-standing presence of the bacterium remains unclear (73).

What is apparent is that B. pseudomallei can occasionally persist in temperate environments

after introduction via animals infected with melioidosis (82).

Because of the increasing frequency of nosocomial infections due to S. maltophilia, its

presence in hospital environments is being more closely examined. Like P. aeruginosa, S.

maltophilia is ubiquitous in aqueous environments and can be readily cultured from water

sources in homes and hospitals (91).

Unlike that of certain B. cepacia complex strains, evidence for person-to-person transmission

of B. gladioli, B. pseudomallei, B. mallei, S. maltophilia, and the other species discussed in

this chapter is lacking.

CLINICAL SIGNIFICANCE Back to top

B. cepacia Complex and B. gladioli

B. cepacia has long been recognized as an occasional opportunistic human pathogen, capable

of causing a variety of infections, including bacteremia, urinary tract infection, septic

arthritis, peritonitis, and pneumonia in persons with underlying illness (157). Persons with

chronic granulomatous disease (CGD) and CF are particularly susceptible to infection

(252). B. cepacia also has a history as a nosocomial pathogen, causing infections associated

with contaminated hospital equipment, medications, and disinfectants including povidoneiodine

and benzalkonium chloride (156). Nosocomial outbreaks of respiratory tract infections

in patients on mechanical ventilation in intensive care units have been attributed to

contamination of nebulizers or nebulized medications such as albuterol (199). Contamination

of blood culture systems or disinfectants resulting in pseudobacteremia has been described

following the isolation of B. cepacia from the blood of multiple patients over a short period

(72). Early reports of infection in CF described patients with acute pulmonary deterioration

and sepsis (referred to as cepacia syndrome) or chronic respiratory tract infections

associated with an accelerated decline in lung function (134, 221). Clinical outcome studies

consistently identified B. cepacia infection as a significant independent risk factor for

morbidity and mortality in CF (71, 153).

The recognition that several closely related species can be distinguished from among

organisms previously identified as B. cepacia has stimulated interest in the clinical

significance of each of these species (154). Approximately 3% of CF patients in the United

States are infected with B. cepacia complex species, although rates of infection vary from 0

to 20% among CF treatment centers (80). Rates of infection increase with increasing patient

age; approximately 5% to 7% of adults with CF are infected (80). Most strains are inherently

resistant to currently available antimicrobial agents (see below), and pulmonary infection is

generally refractory to therapy. Furthermore, due to the poor postoperative prognosis

associated with B. cepacia complex, most CF treatment centers consider infection to be an

absolute contraindication for lung transplantation, which at present offers the only

therapeutic option for successful intermediate-term survival of persons with end-stage

pulmonary disease (155). Thus, respiratory tract infection by these species is a cause of

great concern to CF patients and their caregivers.

Although 16 of the 17 species of the B. cepacia complex have been recovered from persons

with CF (the exception is B. ubonensis), the distribution of species in this patient population

is quite disproportionate. In the United States, B. multivorans and B. cenocepacia together

account for approximately 80% of B. cepaciacomplex infection, with B. vietnamiensis, B.

cepacia, and B. dolosa accounting for approximately 7%, 3%, and 3% of infection,

respectively (200; J. J. LiPuma, unpublished data). In Canada and some European

countries,B. cenocepacia alone accounts for as much as 80% of infection (3, 213). Some B.

cepacia complex species are recovered only rarely. B. stabilis, B. ambifaria, B. anthina, B.

pyrrocinia, B. contaminans, B. seminalis, B. diffusa, B. metallidurans, B. arboris, B.

latens, and B. lata each account for less than 1% of B. cepacia complex infections among

infected CF patients (200; LiPuma, unpublished).

Emerging data suggest that B. cepacia complex species also vary with respect to their

virulence levels and clinical impacts in CF. Studies involving lung transplant recipients, for

example, indicate that rates of postoperative mortality are greater for persons infected

preoperatively with B. cenocepacia than for patients infected with other B. cepacia complex

species (4, 8, 93). Carefully conducted multivariate analyses of posttransplant outcomes are

less definitive, however (177). Thus, although it is almost certainly true that B.

cenocepacia is the species most frequently associated with cepacia syndrome, it remains to

be shown whether this species, in general, is disproportionately associated with poor

outcome; case reports document fatal infection associated with other B. cepacia complex

species, including B. multivorans, B. stabilis, and B. dolosa(19, 86, 183). Thus, although a

positive correlation between species frequency and poor clinical outcome seems likely, firm

conclusions regarding the relative virulence of B. cepacia complex species must await more

definitive study. Evidence also suggests that certain strains are relatively more virulent in

human infection. The B. cenocepacia ET12 epidemic strain, in particular, appears to be

relatively more virulent in CF patients (149). Again, however, further comparative outcome

studies are needed before firm conclusions about the relative virulence of specific strains can

be drawn.

B. gladioli is most notable as a plant pathogen but is also well recognized to be capable of

causing infection in persons with CF or CGD and, occasionally, other immunocompromised

patients (114, 202). Anecdotal reports describe acute pulmonary deterioration and recurrent

soft tissue abscesses, as well as severe post-lung transplantation infections due to B.

gladioli in CF patients (16, 138, 145, 177). A more complete appreciation of the

epidemiology and clinical significance of B. gladioli infection in CF has been confounded by

difficulty with accurate identification of this species, which typically is capable of growth on

selective media used to isolate B. cepacia complex species (49) and is frequently

misidentified as a member of the B. cepacia complex by commercial test systems (51, 209).

Genetic analysis of Burkholderia isolates recovered from CF patients indicates that B.

gladioli is much more commonly involved in infection in this patient population than are

most B. cepacia complex species, with the exception of B. multivorans and B. cenocepacia.

B. pseudomallei and B. mallei

B. pseudomallei is the causative agent of the human and animal disease melioidosis, which is

endemic in Southeast Asia and tropical northern Australia and is being increasingly

recognized on the Indian subcontinent and in Central and South America (37, 73). In

locations where the disease is endemic, infection is seasonal, with up to 85% of cases

occurring during the monsoon wet season. Severe weather events and environmental

disturbances have been associated with melioidosis clusters in Australia, and the Asian

tsunami of 2004 resulted in cases across the affected region (12, 48). As travel to Southeast

Asia and northern Australia has become more frequent, reports of melioidosis in travelers

returning to Europe and the United States are becoming more common (81). Melioidosis is

an especially important potential travel-related illness for those with CF, and persistent

colonization of airways with B. pseudomallei can occur in CF despite prolonged therapy, with

repeated disease flares and deteriorating lung function (181, 237). Infection with this

organism should be considered in the differential diagnosis of any individual with a fever of

unknown origin or a tuberculosis-like illness who has a history of travel to a region where B.

pseudomallei infection is endemic.

B. pseudomallei is acquired from the environment by inoculation through cut or abraded

skin, inhalation, or ingestion. Zoonotic disease is described but is exceedingly uncommon, as

are person-to-person transmission and laboratory-acquired infection (37). The association of

severe weather events with respiratory infection and high mortality rates has been attributed

to a shift from percutaneous inoculation to inhalation (78). This idea supports the potential

of B. pseudomallei as a bioterrorism agent; its isolation from patients who do not give a

history of travel to an area where melioidosis is endemic should be immediately reported to

local or state public health authorities. For further details, see chapter

12 or www.bt.cdc.gov.

The majority of persons exposed to B. pseudomallei do not develop clinical infection, with

rates of seropositivity in the general population as high as 80% in some locations

(141, 256). Latent infection with subsequent reactivation is well recognized, with a recent

description of disease onset in the United States, where melioidosis is not endemic, 62 years

after presumed infection in Thailand (178). Such cases often achieve notoriety but are rare,

and the vast majority of cases of melioidosis are from recent infection, with an incubation

period of 1 to 21 days (mean, 9 days). Risk factors for clinical disease following infection

with B. pseudomallei include diabetes, excessive alcohol consumption, chronic renal disease,

and chronic lung disease (79, 152). Twenty to 36% of cases have no identified risk factor,

and mortality in this group is usually low. Disease in children is also uncommon, although

parotid abscesses are well recognized as an important manifestation of melioidosis in

children in Thailand. Overall rates of mortality from melioidosis vary from 15% in centers

where state-of-the-art intensive care therapy is available to over 50% in locations with poor

resources (37, 248). Fifty percent of cases present with pneumonia, which can be part of a

fatal septicemia or a less severe unilateral infection indistinguishable from other communityacquired

pneumonias or a chronic illness mimicking tuberculosis (33). Chronic melioidosis,

defined as illness present for over 2 months, occurs in only 10% of cases. Overall, 50% of

cases are bacteremic; the presence of >100 CFU/ml of blood and a blood culture showing

growth in the first 24 h of incubation are markers for high mortality (222). At the mild end of

the clinical spectrum of melioidosis is presentation with a skin ulcer or abscess without

systemic illness (103). Other common presentations with or without bacteremia are

genitourinary infections, septic arthritis, and osteomyelitis (37, 75, 248). Prostatic abscesses

are especially common (75). Abscesses can also occur in spleen, liver, kidneys, and adrenal

glands. Parotitis, lymphadenitis, sinusitis, orchitis, myositis (especially psoas abscess),

mycotic aneurysms, and pericardial and mediastinal collections have all been described.

Lesions can be frankly purulent and may include microabscesses or granulomas or a

combination of these features. Clinical meningitis is rare, but melioidosis encephalomyelitis

syndrome (74, 75) and brain abscesses have also been reported. The one presentation that

has yet to be described is B. pseudomallei endocarditis.

B. mallei is the etiologic agent of glanders, a highly communicable disease of livestock,

particularly horses, mules, and donkeys. It can be transmitted to humans and is also

identified as a potential agent of bioterrorism. Unlike B. pseudomallei, B. mallei is a hostadapted

pathogen that does not persist in the environment outside its host. Glanders has

been eradicated from most countries, but enzootic foci persist in the Middle East, Asia,

Africa, and South America. The only human case of glanders in the past 50 years in the

United States was a recent laboratory-acquired case in a biodefense scientist (27). Like

melioidosis, human glanders can be acute or chronic, with the clinical presentation and

course depending on the mode of infection, the inoculation dose, and host risk factors.

Respiratory inoculation can result in pneumonia with potential for dissemination to internal

organs and septicemia. Cutaneous inoculation can result in skin nodules and regional

lymphadenitis, also with potential for disseminated disease. Involvement of nodes, both

mediastinal and peripheral, is much more common in glanders than in melioidosis, often with

suppurative abscesses in untreated cases.

S. maltophilia

S. maltophilia, although typically not pathogenic for healthy persons, is a well-known

opportunistic human pathogen. It is among the most common causes of wound infection due

to trauma involving agricultural machinery (2). It is also an important nosocomial pathogen

associated with substantial morbidity and mortality, particularly in debilitated or

immunocompromised patients and patients requiring ventilatory support in intensive care

units (5, 99, 119, 176). The incidence of human infection appears to have increased in

recent years, and a variety of clinical syndromes have been described, including bacteremia,

pneumonia, urinary tract infection, ocular infection, endocarditis, meningitis, soft tissue and

wound infection, mastoiditis, epididymitis, cholangitis, osteochondritis, bursitis, and

peritonitis (9, 205). Septicemia can be accompanied by ecthyma gangrenosa, a skin lesion

more commonly associated with Pseudomonas aeruginosa and Vibrio spp. (232).

The incidence of S. maltophilia respiratory tract infection in persons with CF also appears to

be increasing (88,198, 218); however, the unreliability of historical data limits firm

conclusions. Approximately 12% of CF patients included in the CF Foundation’s patient

registry were culture positive for S. maltophilia in 2005 (80). In large multicenter clinical

trials, however, S. maltophilia was found in a larger proportion of CF patients, being second

only to P. aeruginosa in frequency of isolation from study subjects (26). Infection or

colonization was most frequently transient, with 30% of subjects having at least one sputum

culture positive for S. maltophiliaduring the course of 6 months (113). Several case-control

studies have drawn conflicting conclusions regarding the role that S. maltophilia plays in

contributing to pulmonary decline in CF (88, 111).

Ralstonia and Cupriavidus spp.

As described above, the taxonomy of the genus Ralstonia has been recently revised, with

several species being assigned to the genus Cupriavidus (225). Among the species in these

two genera, R. pickettii is best known with respect to human infection. Older reports describe

this species as being recovered from a variety of clinical specimens (201) and as causing

various infections including bacteremia, meningitis, endocarditis, and osteomyelitis (243). R.

pickettii also has been identified in pseudobacteremias and nosocomial outbreaks due to

contamination of intravenous medications, “sterile” water, saline, chlorhexidine solutions,

respiratory therapy solutions, and intravenous catheters (20, 45, 96). This species has also

been recovered from the respiratory tract of persons with CF (26). However, R. pickettii is

easily confused with Pseudomonas fluorescensand members of the B. cepacia complex based

on phenotype (26, 85, 123). Furthermore, several newly recognized R. pickettii-like species

are also now known to be involved in human infection, particularly in CF (64). Thus, the role

of R. pickettii as a human pathogen is difficult to assess based on historical data.

R. mannitolilytica (formerly known as R. pickettii biovar 3/‘thomasii’) was recently described

as causing nosocomial outbreaks and a case of recurrent meningitis (85). This species

accounts for the majority ofRalstonia infection in CF patients, being found in more than twice

as many CF patients as those infected withR. pickettii (64). R. insidiosa and Cupriavidus

respiraculi are recently described species recovered from persons with CF

(54, 68). Cupriavidus gilardii has been recovered from cerebrospinal fluid (53), and cases

of Cupriavidus pauculus bacteremia, peritonitis, and tenosynovitis have been reported (226).

Both of these species may be found in sputa from patients with CF (64).

Although Cupriavidus metallidurans and Cupriavidus basilensis are not known to cause other

human infection, they too have been recovered recently from sputum cultures from patients

with CF (64). Despite these observations, the roles of Ralstonia and Cupriavidus species in

human infection, particularly in persons with CF, require further elucidation.

Other Genera

In general, Brevundimonas, Comamonas, Delftia, Acidovorax, and Pandoraea spp.

infrequently cause human infection. Interest in these species focuses primarily on their roles

as plant pathogens or in studies of microbial biodiversity and biodegradation.

Brevundimonas spp. are occasionally recovered from clinical specimens

(46). Brevundimonas vesicularisbacteremia in patients with various underlying illnesses has

been reported (104), and the organism has been recognized in cervical specimens because

of its ability to produce bright orange colonies on Thayer-Martin agar (184). Brevundimonas

diminuta has been recovered from blood, urine, and pleural fluid from patients with cancer

(117).

Among the Comamonas species, Comamonas testosteroni has been implicated most often in

human infection, with reports describing endocarditis, meningitis, and catheter-associated

bacteremia due to this species (7,70, 150). D. acidovorans has similarly been reported to

cause infection, being identified in cases of bacteremia, endocarditis, ocular infection, and

suppurative otitis (95). Acidovorax spp. have been isolated from a variety of clinical sources

(250), including blood from a patient with hematological malignancy

(261). Acidovorax spp.,Comamonas testosteroni, and D. acidovorans have also been

recovered from sputa of persons with CF (55; LiPuma, unpublished); however, the roles of

these species in contributing to lung disease in CF have not been established.

In addition to causing infection in CF patients (136, 139), Pandoraea spp. have been

recovered from blood and from patients with chronic obstructive pulmonary disease or CGD

(83). Although the roles of these species in contributing to poor outcomes in persons with

underlying diseases are unclear, a recent report describes sepsis, multiple organ failure, and

death in a patient who underwent lung transplantation due to sarcoidosis (216).

COLLECTION, TRANSPORT, AND STORAGE Back to top

The genera described in this chapter include organisms that can survive in a variety of

hostile environments and at temperatures found in clinical settings. Therefore, standard

collection, transport, and storage techniques as outlined in chapters 9 and 16 are sufficient

to ensure the recovery of these organisms from clinical specimens. Recovery of B.

pseudomallei for the diagnosis of melioidosis is increased by the additional collection of

throat and rectal swabs into selective media (see “Culture and Isolation” below) and by

collecting larger than standard volumes of cerebrospinal fluid for culture in suspected

neurological melioidosis (44).

DIRECT EXAMINATION Back to top

Members of the genera discussed here have similar morphologies and, with the exception

of B. pseudomallei,are not easily distinguished from each other on the basis of Gram

staining. B. pseudomallei may appear as small gram-negative bacilli with bipolar staining,

making the cells resemble safety pins (Fig. 1). This may increase the index of suspicion for

the presence of B. pseudomallei, but the sensitivity and specificity of this appearance are not

high enough to be relied upon for a presumptive clinical diagnosis.

Although PCR-based assays have been described for the identification of B. cepacia complex

species, B. pseudomallei, B.

gladioli, several Ralstonia and Cupriavidus species, Pandoraea species, and S.

maltophiliafollowing culture and isolation (see “Identification” below), the use of PCR for

direct detection of these species in clinical specimens remains a research tool (84, 169, 246).

Studies of CF sputum samples have indicated that some specimens may be PCR positive but

culture negative for certain B. cepacia complex species, raising important questions about

the natural history of infection in CF. However, the sensitivities and specificities of such PCR

assays for the intended target species are difficult to determine in the absence of reliable

“gold standards.” The development of assays employing real-time PCR technology may yield

reliable approaches to direct detection of these species in clinical specimens in the near

future.

Because septicemia with B. pseudomallei is frequently fatal and death often occurs in the

first few days after presentation to the hospital prior to the availability of culture results,

several rapid direct-detection methods have been developed including urinary antigen

detection using latex agglutination (LA) and enzyme immunoassay (EIA), and direct

fluorescent-antibody (DFA) staining (92, 211, 258). The EIA for the detection of urinary

antigen is more sensitive than LA, with an overall sensitivity of 71% for patients with

melioidosis compared with an LA sensitivity of 62% (with concentrated urine) or only 17.5%

(with unconcentrated urine). The EIA has an even higher sensitivity (84%) for samples from

septicemic patients. Cross-reactions with other urinary tract pathogens including Klebsiella

pneumoniae and Escherichia coli have been reported with EIA but not LA; therefore, EIA

results must be interpreted cautiously (92, 211). Antibodies raised against heat-killed whole

cells of B. pseudomallei have been used to prepare a reagent for DFA staining. When this

DFA reagent was used to stain clinical specimens from patients with suspected melioidosis,

using a rapid assay that took 10 minutes to prepare and read, it showed a sensitivity of 66%

and specificity of 99.5% (258). The reagents described in the literature are largely prepared

in-house and are not available commercially.

The evaluation of a conventional PCR assay targeting the 16s rRNA gene for the detection

of B. pseudomalleiin clinical specimens indicated that the assay was sensitive but lacked

specificity, resulting in positive predictive values of only 70% (115). Real-time PCR assays

targeting B. pseudomallei genes encoding 16S rRNA, flagellin (fliC), ribosomal protein

subunit S21 (rpsU), or type III secretion system (TTS) genes have been developed and been

shown to have high sensitivities and specificities with pure bacterial cultures (179, 220,223).

Two clinical evaluations of real-time PCR have met with mixed results, with a sensitivity of

91% in one study using an assay targeting a gene in the TSS1 cluster (171) and a

considerably lower sensitivity of 61% in a second study that used an assay targeting the 16s

rRNA gene (29). Sensitivity of PCR is highest for pus and other body fluids but is low for

blood, most likely reflecting a bacterial copy number effect. Loop-mediated isothermal

amplification (LAMP) is easy and quick to perform and needs minimal equipment, with

amplification being achieved at a fixed temperature in a water bath or heating block. LAMP

has been developed and evaluated for the detection of B. pseudomallei and diagnosis of

melioidosis (28). The assay was sensitive and specific for the laboratory detection of B.

pseudomallei and had 100% specificity when applied to clinical samples but had a very low

diagnostic sensitivity (44%). At present, molecular assays are not sufficiently sensitive to

replace conventional culture.

CULTURE AND ISOLATION Back to top

The species discussed in this chapter grow well on standard laboratory media such as 5%

sheep blood and chocolate agars. Such media can be used to recover the organisms from

sterile fluid or tissue where a mixed biota is not anticipated (see chapter 16). All species that

have been reported to be recovered from blood, including B. pseudomallei (222), grow in

broth-based blood culture systems within the standard 5-day incubation period, so that

special blood culture techniques such as lysis-centrifugation and extended incubation periods

are not required. The use of selective media facilitates the isolation of these organisms from

specimens with mixed microbiota. With the exception of Brevundimonas

vesicularis, MacConkey agar can be used to isolate most species of these genera.

Burkholderia species grow on MacConkey agar (Fig. 2a), but the use of specific selective

media with the ability to inhibit P. aeruginosa is preferred for the isolation of B.

cepacia complex and B. pseudomallei. Several selective media have been described, and

some are commercially available. A multicenter comparison of three media, PC

(for Pseudomonas cepacia) agar (BD Diagnostics, Franklin Lakes, NJ) (106), OFPBL (for

oxidation-fermentation base- polymyxin B-bacitracin-lactose) agar (BD Diagnostics) (241),

and BCSA (for B. cepaciaselective agar; Hardy Diagnostics, Santa Maria, CA) (121), showed

that BCSA was superior, being both more sensitive (more B. cepacia isolates were

recovered) and more specific (fewer other types of organisms grew) than PC or OFPBL agar

(121, 122). The sensitivities of TB-T (for trypan blue-tetracycline) (116), PC-AT

(forPseudomonas cepacia azelaic acid) (24), and BCSA (121, 122) were also compared with

those of three commercial media, i.e., B. cepacia media from MAST Diagnostics (Bootle,

Merseyside, United Kingdom), LAB M Ltd. (Bury, United Kingdom), and Oxoid Ltd.

(Basingstoke, United Kingdom), through the analysis of 142 clinical and environmental

isolates representing all species within the B. cepacia complex (235). BCSA and Mast B.

cepacia medium supported the growth of B. cepacia complex isolates most efficiently. The

latter two media were also compared in a study to evaluate the sensitivities and specificities

for the isolation of B. cepaciacomplex species from sputum specimens from CF patients

Ashdown medium is effective for the isolation of B. pseudomallei (Fig. 2); crystal violet and

gentamicin act as selective agents. It has been shown to be superior to MacConkey agar or

MacConkey agar supplemented with colistin for the recovery of B. pseudomallei from clinical

specimens containing mixed bacterial microbiota, such as throat, rectal, and sputum

specimens (257). A more recently described selective agar, BPSA (for B.

pseudomallei selective agar), was reported to improve the recovery of B. pseudomallei over

that with other media (127); however, this medium is not yet commercially available. BPSA

was more inhibitory to P. aeruginosa and B. cepacia complex species and made recognition

of Burkholderia species easier due to their distinctive colony morphology. A clinical

comparison of BPSA, Ashdown medium, and Burkholderia cepaciamedium demonstrated

equivalent sensitivity for all three media, but the selectivity of BPSA was lower than that for

the other two media (191). Burkholderia cepacia medium is widely used and represents a

good alternative when Ashdown medium is not available. An enrichment broth consisting of

Ashdown medium supplemented with 50 mg of colistin is superior to standard enrichment

broth such as tryptic soy broth and increases recovery of B. pseudomallei from clinical

specimens taken from colonized sites compared with plating on Ashdown medium alone

(44, 238). Selective broth cultures should be subcultured to Ashdown medium after 48 hours

of incubation in air at 37°C, and all inoculated plates should be incubated at 37°C in air and

examined daily for 4 days before being discarded, since some colonies become apparent to

the naked eye only after extended incubation.

The use of selective media (143) increases the isolation rates of S. maltophilia from clinical

and environmental samples (89). Denton et al. (89) studied the sensitivity of a selective

medium incorporating vancomycin, imipenem, and amphotericin B as selective agents (VIA

medium) for isolating S. maltophilia from sputum samples collected from children with CF.

This study compared the use of VIA medium to an existing in-house method that utilized an

imipenem disk placed upon bacitracin-chocolate agar (BC medium) and reported an

improved detection using VIA as a selective medium.

IDENTIFICATION Back to top

B. cepacia Complex and B. gladioli

Accurate identification of B. cepacia complex species presents a challenge (170). Commercial

bacterial identification systems are not able to reliably distinguish among the species of

the B. cepacia complex and often fail to differentiate these species from other closely related

species such as B. gladioli and Ralstonia, Cupriavidus, and Pandoraea spp.

(22, 123, 146, 209). This failure presents a serious problem for CF patients and their

caregivers as detailed in “Clinical Significance” above. The identification of B.

cepacia complex species from CF sputum culture has a dramatic impact on patient

management and is a cause of considerable anxiety for patients with CF (155, 156).

Consequently, when Burkholderia, Ralstonia, Cupriavidus, or Pandoraea species are

tentatively identified in a patient with CF by using a commercial system, the identity of the

isolate should be confirmed by conventional biochemical testing (123) and, if necessary,

molecular techniques. To aid clinical microbiologists in the United States, the CF Foundation

has established a B. cepacia reference laboratory, which uses a combination of phenotypic

and genotypic methods (described below) to confirm the identity of suspected B.

cepacia complex isolates (159). Further information concerning the B. cepacia reference

laboratory can be found on the CF Foundation website (http://www.cff.org).

B. cepacia complex species may require 3 days of incubation before colonies are seen on

selective media. On MacConkey or Mueller-Hinton agar, these colonies may be punctate and

tenacious, and on blood agar or selective medium such as BCSA, PC agar, or OFPBL agar,

the colonies are smooth and slightly raised; occasional isolates are mucoid. On MacConkey

agar, colonies of the B. cepacia complex frequently become dark pink to red due to oxidation

of lactose after extended incubation (4 to 7 days). Most clinical isolates are nonpigmented,

but on iron-containing media such as a triple sugar iron slant, many strains produce a bright

yellow pigment. B. cepacia complex species have a characteristic dirtlike odor.

The species of the B. cepacia complex are phenotypically very similar, making their

differentiation, even with an extended panel of biochemical tests, rather difficult (Table 1)

(123). Further, isolates within these species show considerable phenotypic variability, which

is likely due to their unusually large genomes rich in insertion sequences and mobile

elements such as plasmids, transposons, and bacteriophages (167). These features can

contribute to genetic plasticity and diversity, which, when differentially expressed in isolates,

results in variable biochemical phenotypic profiles. Most strains are weakly oxidase positive,

although some strains of B. contaminans, B. lata, and B. pyrrocinia are oxidase negative. B.

multivorans, B. stabilis, and B. dolosa rarely oxidize sucrose. B. stabilis is ornithine

decarboxylase positive, as are most B. cenocepacia strains, but is distinctive in that more

than two-thirds of strains are o-nitrophenyl-β-D-galactopyranoside (ONPG) negative. B.

stabilis, B. lata, and most B. ambifaria strains show poor growth at 42°C. B. dolosa is usually

lysine decarboxylase negative, whereas only approximately one-half of B. multi

vorans strains are negative. Other B. cepacia complex species are usually lysine

decarboxylase positive. B. vietnamiensis and most B. anthina strains do not oxidize

adonitol. B. anthina strains show a distinctive creamy morphology on BCSA, which also turns

pink (i.e., alkaline) despite the ability of this species to utilize sucrose (227).

Phenotypic differentiation of B. cepacia complex species from B. gladioli and Pandoraea spp.

is also difficult (Table 1). Cellular fatty acid analysis is unable to differentiate B.

cepacia complex species from B. gladioli (215). However, in contrast to B. cepacia complex

species, most B. gladioli strains are oxidase negative, and whereas most B. cepacia complex

strains oxidize maltose and lactose, B. gladioli typically oxidizes neither. Pandoraea spp. do

not oxidize maltose, lactose, xylose, sucrose, or adonitol, and most are ONPG negative. B.

cepacia complex species also may be difficult to differentiate

from Ralstonia and Cupriavidus species. However, several of the latter species show a fast

and strong oxidase reaction whereas B. cepacia complex species produce a slow, weakpositive

oxidase test. Further, in contrast to most B. cepacia complex

species, Ralstonia and Cupriavidusare lysine decarboxylase negative and most often ONPG

negative.

The difficulty in differentiating B. cepacia complex species has prompted the development of

molecular genetic diagnostic tests capable of identifying these species individually and

distinguishing them (as a group) from biochemically similar species. DNA sequence

differences in 16S and 23S rRNA genes have been used to develop species-specific PCR

assays for the identification of several B. cepacia complex species (17, 159, 233), as well

as B. gladioli (247). B. multivorans, B. vietnamiensis, and B. dolosa can be reliably identified

with 16S rRNA-targeted assays, but insufficient sequence variation in rRNA genes exists to

enable reliable separation of the remaining B. cepacia complex species. Fortunately, speciesspecific

sequence variation does exist in therecA gene, and PCR assays targeting this locus

enable the reliable identification of the B. cepacia complex species most commonly recovered

from human specimens (61, 165, 230, 231, 234). Other 16S rRNA- and recA-based PCR

assays identify all Burkholderia spp. (i.e., at the genus level) or all species within the B.

cepaciacomplex (i.e., as a group) (159, 165, 190).

Another molecular genetic approach to identifying B. cepacia complex species involves

restriction fragment length polymorphism (RFLP) analysis of either 16S rRNA or recA genes

(165, 207). Again, insufficient sequence variation in the 16S rRNA gene limits the use of

RFLP analysis of this locus, even when multiple restriction enzymes are used (98, 207, 236).

In contrast, recA RFLP analysis has proved quite useful in reliably distinguishing all species

within the B. cepacia complex (165, 190, 230, 231, 234).

Although an MLST scheme for the B. cepacia complex was developed primarily as a tool to

study the epidemiology and population genetic structure of these species (14), it offers

another approach to differentiate species within this group. This scheme was recently

modified to allow typing of all species within the genus Burkholderia (214). Other genomic

approaches, including amplified fragment length polymorphism typing, ribotyping, and

whole-cell protein profiling, have been proposed for the differentiation of B. cepaciacomplex

species (23, 62, 228). However, these methods are time-consuming and expensive and

require an extensive validated database before isolates can be reliably identified. These

limitations render them impractical for use in a routine diagnostic laboratory. Cellular fatty

acid methyl ester analysis is useful for identification of Burkholderia strains at the genus level

but is not reliable for identification of individual B. cepacia complex species and does not

differentiate B. gladioli (51, 226).

B. pseudomallei and B. mallei

B. pseudomallei colonies on blood agar are typically small, smooth, and creamy in the first

48 hours. On further incubation, this appearance changes to give dry, wrinkled colonies. B.

pseudomallei on Ashdown agar grows as very small (pinpoint) colonies by 18 h, which are

usually purple, flat, dry, and wrinkled (“cornflower head”) after 48 hours of incubation (Fig.

2c). The organism is motile, indole negative, oxidase positive, and resistant to colistin and

gentamicin, features that aid identification (82). Other typical biochemical reactions are

shown in Table 2. B. pseudomallei produces a distinctive musty or earthy odor, but sniffing

of open plates should never be undertaken on safety grounds (11). B. thailandensis may be

indistinguishable from B. pseudomallei by these simple criteria but can be distinguished

based on arabinose assimilation, since B. thailandensis can utilize L-arabinose as the sole

carbon source but neither B. pseudomallei nor B. mallei can do so (21, 175, 210).B.

thailandensis is found in clinical samples extremely rarely, and reported cases of human B.

thailandensisinfection in which the bacterial species was fully verified amount to a single

patient (108). B. mallei is also an extreme rarity in clinical specimens from humans but could

be confused with B. pseudomallei and B. thailandensis. Two differentiating tests in the event

that B. mallei is suspected or needs to be ruled out are (i) that B. mallei is nonmotile and B.

pseudomallei and B. thailandensis are motile and (ii) that B. mallei is susceptible to

gentamicin (142) while the other two species are inherently resistant. The latter represents a

catch for the inexperienced microbiologist, since Ashdown medium normally contains

gentamicin and thus fails to support the growth of B. mallei. Cellular fatty acid profiles may

be useful for differentiating B. pseudomalleifrom other genera, but reports vary on its utility

in differentiating B. thailandensis and B. pseudomallei (130) or other

pathogenic Burkholderia species including B. mallei, B. cepacia complex species, and B.

gladioli.

Multiple evaluations have been performed to determine the accuracy of API 20NE for the

identification of B. pseudomallei, with the reported percentage of isolates identified correctly

ranging from 37% to 99% (6, 82, 87,108, 131, 132, 162). One possible reason for this

interstudy variability is that B. pseudomallei is phenotypically distinct in different

geographical areas and/or between clinical isolates and those from the environment. A

recent study using a large collection (n= 800) of B. pseudomallei isolates obtained from

clinical cases, the environment, and animals from seven Asian countries and northern

Australia reported that the API 20NE correctly identified 99% of B. pseudomallei isolates (6).

This supports the use of API 20NE for the identification of B. pseudomallei. API 20NE is

unable to identify B. mallei or B. thailandensis (6). The automated Vitek 1 system provides

accurate identification of B. pseudomallei (162). Two evaluations of the Vitek 2 colorimetric

GN card system have reported an accuracy of around 80% for B. pseudomallei, the most

common incorrect identification being the Burkholderia cepacia group (87, 163). The

accuracy of Vitek 2 has been reported to be affected by the medium on which B.

pseudomallei is grown prior to testing, with culture on Columbia horse blood agar being

associated with the highest rate of accuracy (163). Because of the difficulty associated with

accurate laboratory identification, referral to a reference laboratory is advised when isolation

of B. pseudomallei or B. mallei is suspected. This is especially important with the advent of

emerging bioterrorism legislation in many countries (see chapter 12).

B. pseudomallei must be differentiated from Pseudomonas stutzeri and B. cepacia complex

species in clinical specimens. Pseudomonas stutzeri appears very similar to B.

pseudomallei after a few days of incubation, and both B. pseudomallei and B.

cepacia complex species may be isolated from persons with CF (181, 237). Whereas B.

pseudomallei produces gas from nitrate and is arginine dihydrolase positive, most B.

cepaciacomplex isolates are negative for both characteristics. Pseudomonas stutzeri is

negative for arginine dihydrolase, oxidation-fermentation glucose, and gelatin hydrolysis.

Also, Pseudomonas stutzeri has only one flagellum, and B. pseudomallei has more than one.

Ralstonia and Cupriavidus spp.

Although R. pickettii was considered to be the Ralstonia species most frequently isolated

from clinical specimens (201), the recent recognition that several

other Ralstonia and Cupriavidus species can be identified from among R. pickettii-like

isolates limits previous observations. As is the case with B. cepacia complex

species, Ralstonia and Cupriavidus species are phenotypically similar, requiring rather

extensive biochemical testing to reliably differentiate them; species level identification with

standard biochemical testing is difficult (Table 3). These species may grow slowly on primary

isolation media, requiring ≥72 h of incubation before colonies are visible. They are lysine

decarboxylase negative and generally catalase positive, although catalase-negative R.

pickettii strains have been described (53, 226). Most species show a fast and strong oxidase

reaction; however, the intensity of the oxidase reaction varies for R. mannitolilytica, R.

pickettii, andCupriavidus gilardii, with some strains showing a weakly positive reaction

(53, 226; LiPuma, unpublished). R. pickettii, R. mannitolilytica, and R. insidiosa grow on

BCSA; most Cupriavidus strains do not, but growth is strain dependent. These species do not

produce acid from sucrose. Most R. mannitolilytica strains acidify lactose, whereas most

strains from other species do not. R. insidiosa, C. respiraculi, Cupriavidus

gilardii, and Cupriavidus pauculus are differentiated from R. pickettii and R. mannitolilytica in

failing to acidify glucose. Cupriavidus species have a characteristic cellular fatty acid profile

different from that of other Ralstonia species (53, 85). The main fatty acid components of

these species are C16:0, C16:1 w7c, and C18:1 w7c (each accounting for 20 to 30% of the

overall fatty acid content); in addition C14:0, C14:0 3OH, and C17:0 cyclo are always

present (each accounting for 5 to 10% of the overall fatty acid content) (68, 110).

Pandoraea spp.

Overall, the biochemical profiles of Pandoraea strains are similar to those

of Burkholderia and Ralstonia strains isolated from clinical specimens (Table 1)

(52, 83, 123). The lack of saccharolytic activity is indicative ofPandoraea but is also seen

with some Ralstonia species. Definitive identification of putative Pandoraea isolates requires

molecular confirmation. Coenye et al. (60) described 16S rRNA gene-based PCR assays for

the identification of these bacteria. A PCR assay was developed for the identification

of Pandoraea isolates to the genus level. PCR assays for the identification of P. apista and P.

pulmonicola (as a group), P. pnomenusa, P. sputorum, and P. norimbergensis were also

developed. Pandoraea strains can be differentiated fromBurkholderia and Ralstonia strains by

their specific 16S rRNA gene restriction profile (123, 208) and can be identified at the

species level through MspI restriction analysis of the gyrB gene (59). A quantitative

comparison of the whole-cell fatty acid profiles of the members of these three genera allows

the differentiation of Pandoraea strains from the others (53, 83). However, with the use of

the commercially available microbial identification system database (Microbial ID, Inc.,

Newark, DE), these organisms are mostly identified with low identification scores

as Burkholderia or Ralstonia species (53, 123) due to a lack of discriminatory fatty acids.

S. maltophilia

Key features for identifying S. maltophilia include oxidation of glucose and maltose with a

more intense reaction with the latter, positive reactions for DNase and lysine decarboxylase,

and a tuft of polar flagella (Table 4) (244). Although most strains were previously believed to

be oxidase negative, testing of large numbers of isolates recovered from human specimens

and referred to a reference laboratory indicates that as many as 20% may be oxidase

positive (LiPuma, unpublished). This proportion may be higher than expected since strains

with an “atypical” phenotype may be preferentially referred for analysis. Detection of

extracellular DNase activity by S. maltophilia is a key to differentiating this species from

most other glucose-oxidizing, gram-negative bacilli. It can be detected on tube-base or

plated DNase medium with a methyl green indicator. Care must be taken when interpreting

the DNase reaction, since one report documented the misidentification of S. maltophilia as B.

cepacia partially based on false-negative DNase reactions that were finalized with 48 h of

incubation rather than 72 h (25). Selected isolates of Flavobacterium and Shewanella spp.

may also be DNase positive. On sheep blood agar, colonies appear rough and lavender-green

and have an ammonialike odor. S. maltophilia has a characteristic cellular fatty acid profile

with large amounts (>30%) of 13-methyl tetradecanoic acid (C15:0 iso) and lesser amounts

(>10%) of 12-methyl tetradecanoic acid (C15:0 anteiso) and cis-9-hexadecenoic acid (C16:1 cis9)

(244). To overcome the problems associated with definitive identification of S.

maltophilia, Whitby et al. (245) developed a species-specific PCR assay targeting the 23S

rRNA gene and reported sensitivity and specificity of 100%. This PCR test was used as a

standard to evaluate the identification of S. maltophilia using the API 20NE strip and the

VITEK 2 ID-GNB card (107). Both systems showed good reliability compared to PCR. A

multiplex PCR assay to identify P. aeruginosa, B. cepacia complex species, and S.

maltophilia directly in sputum and oropharyngeal specimens from CF patients has been

reported, but only a very limited number of S. maltophilia isolates were examined (84).

Acidovorax, Brevundimonas, Delftia, and Comamonas spp.

Characteristics of Acidovorax, Brevundimonas, Delftia, and Comamonas are given in Table

4.

Acidovorax species, rarely encountered in clinical and environmental samples, are straight to

slightly curved gram-negative bacilli which occur either singly or in short chains. They are

oxidase positive and nonpigmented and have a single polar flagellum. Urease activity varies

among strains (244, 250).

Brevundimonas diminuta and Brevundimonas vesicularis, infrequently encountered in clinical

and environmental samples, have growth requirements for specific vitamins, including

pantothenate, biotin, and cyanocobalamin. An additional growth requirement

for Brevundimonas diminuta is cysteine. Most strains of Brevundimonas diminuta grow on

MacConkey agar, while only approximately 25% of Brevundimonas vesicularis strains do so.

On primary isolation media, Brevundimonas diminuta colonies are chalk white whereas many

strains ofBrevundimonas vesicularis are characterized by an orange intracellular pigment.

These organisms are oxidase positive, have a single polar flagellum, and weakly oxidize

glucose (Brevundimonas vesicularis more so thanBrevundimonas diminuta), and the vast

majority fail to reduce nitrate to nitrite. The most reliable method for differentiating these

two species is the test for esculin hydrolysis. Almost all strains of Brevundimonas

vesicularis (88%) are reported to hydrolyze this substrate, while Brevundimonas

diminuta strains rarely do (5%) (Table 4) (244).

Comamonas spp. are straight to slightly curved gram- negative bacilli that occur singly or in

pairs. The organisms are catalase and oxidase positive and have a single tuft of polar

flagella. All human clinicalComamonas species reduce nitrate to nitrite. Phenotypic

differentiation of Comamonas terrigena fromComamonas testosteroni is difficult, and as a

result isolates are typically reported as Comamonas spp. (Table 4).

D. acidovorans is phenotypically similar to Comamonas. Key characteristics of the species

include abilities to oxidize fructose and mannitol. One-quarter of the strains produce a

fluorescent pigment, while approximately one-half of the strains may produce a soluble

yellow to tan one (244, 249).

TYPING SYSTEMS Back to top

Several molecular genetic methods are available to assess the relatedness of isolates of

these genera during nosocomial or community outbreak investigations. These methods are

preferred over phenotypically based systems, which are less discriminatory and reproducible.

Analysis of whole-genome macrorestriction profiles with pulsed-field gel electrophoresis

(PFGE) has gained acceptance as a preferred genotyping method and has proved useful in

numerous studies of Burkholderia, Ralstonia, and S. maltophilia (45, 63, 224). The

endonucleases XbaI and SpeI are most frequently used and typically yield a dozen or more

DNA fragments for analysis. Care must be taken in interpreting PFGE profiles

of Burkholderia species, however. These species have unusually large and dynamic

multichromosome genomes that are prone to large-scale alterations in content and

arrangement (151). Consequently, epidemiologically irrelevant genomic polymorphisms may

arise in the short term and confound outbreak investigations (63). Ribotyping, which relies

on polymorphisms in and around rRNA operons, has been used to investigate the

epidemiology of B. cepacia complex and B. pseudomallei (133, 158, 160). Both PFGE and

ribotyping are relatively time-consuming and expensive to perform and are therefore not

particularly well suited for routine analysis by clinical microbiology laboratories. A variety of

PCR-based methods, including randomly amplified polymorphic DNA typing and repetitivesequence

PCR typing, offer attractive alternatives for genotyping S.

maltophilia and Burkholderia, Ralstonia, and Pandoraeaspp. (34, 55, 147, 166, 208). These

methods are inexpensive and can provide rapid, reliable results. MLST, which assesses DNA

sequence variation at several chromosomal loci, has been developed for numerous species,

including the B. cepacia complex, B. pseudomallei, and B. mallei (14, 39, 109). A recent

modification of the scheme developed for the B. cepacia complex enables MLST analysis of

all species within the genus (214). This genotyping strategy provides robust, reproducible,

and portable results and is quickly becoming the preferred method for investigating bacterial

epidemiology, evolution, and population structure. Both repetitive-sequence PCR using a

BOX A1R primer and multilocus variable-number tandem repeat analysis have been

developed for B. pseudomallei to exclude a clonal outbreak (76, 77). Typing methods have

not been reported for Brevundimonas, Delftia, Comamonas, or Acidovorax spp.

SEROLOGIC TESTS Back to top

Of the organisms discussed in this chapter, B. pseudomallei is the only one for which

serologic tests have been used clinically to diagnose the infection. The indirect

hemagglutination assay, although not available commercially, is the most widely used test

(10). It is performed by using a prepared antigen from strains of B. pseudomallei sensitized

to sheep cells and includes unsensitized cells as a control. This assay can be adapted to a

microtiter plate test system. The serologic tests currently in use have limited value for the

diagnosis of melioidosis in persons who have lived in regions where melioidosis is endemic

because the healthy indigenous population is often seropositive (256, 259, 260). Serologic

testing is potentially useful for persons who do not normally reside in regions endemic for

melioidosis, including returning travellers and laboratory workers following accidental

laboratory exposure to B. pseudomallei (192). The interpretation of the indirect

hemagglutination assay or other serologic assays is complicated by the fact that there are no

validated guidelines, and different cutoff points have been used to define seroconversion

following exposure and acute infection. Testing should be performed whenever possible on

paired samples. Seroconversion with the development of detectable antibodies to B.

pseudomallei in the second sample is supportive of exposure. A fourfold rise in titer is

commonly used to diagnose a range of infectious diseases, but this has not been validated

for melioidosis and any reproducible rise between two samples should be viewed as possible

evidence of exposure. A single high titer in persons from a nonendemic region with a

relevant travel history who presents late after a putative exposure event and for whom

paired sera may be less relevant is also suggestive of exposure. Some individuals with

culture-proven melioidosis do not have detectable antibodies (41), and so a negative

serologic test does not rule out exposure or melioidosis. Given the complexity of this

situation, experts in the field should be consulted when serology is used to diagnose

melioidosis. Several evaluations of a commercial rapid immunochromatographic test kit

(Pan-Bio, Windsor, Queensland, Australia) for the detection of immunoglobulin G and

immunoglobulin M antibodies to B. pseudomallei have been performed (42, 180, 254), but

this test is not currently available.

ANTIMICROBIAL SUSCEPTIBILITIES Back to top

Specific susceptibility testing interpretative criteria are not available for all of the species

discussed in this chapter. For some species, such as the B. cepacia complex and S.

maltophilia, interpretive criteria for disk diffusion testing are available for only a limited

number of antibiotics. In general, MIC broth microdilution tests or Etests are preferred for

this group of organisms.

B. cepacia complex species are among the most antimicrobial-resistant bacteria encountered

in the clinical laboratory. These species are intrinsically resistant to aminoglycoside and

polymyxin antibiotics and are often resistant to β- lactam antibiotics due to inducible

chromosomal β-lactamases and altered penicillin-binding proteins (118). Antibiotic efflux

pumps may mediate resistance to chloramphenicol, fluoroquinolones, and trimethoprim

(196). Clinical strains may be susceptible to only a handful of agents, including

trimethoprim-sulfamethoxazole (TMP-SMX), ceftazidime, chloramphenicol, minocycline,

imipenem, meropenem, and some fluoroquinolones (1, 125, 194). The glycylcycline antibiotic

tigecycline shows highly variable activity in vitro (172). The relatively high MIC observed for

some strains and the potential for discoloration of permanent teeth in children younger than

7 years of age limit the use of tigecycline in CF patients. Clinical and Laboratory Standards

Institute (CLSI; formerly NCCLS) interpretative criteria for disk diffusion susceptibility testing

are available for ceftazidime, meropenem, minocycline, and TMP-SMX (50). MIC broth

microdilution tests or Etests are preferred methodologies for susceptibility testing of these

species. Because isolates that are initially susceptible may become resistant during the

course of therapy, susceptibility testing of repeat isolates may be warranted. Furthermore,

strains recovered from patients with CF who have received repeated courses of antibiotic

therapy are frequently resistant to all currently available antimicrobial agents (1, 105).

Combinations of antimicrobial agents may provide synergistic activity against resistant

strains; however, antagonism with combinations is also observed in vitro (1).

B. pseudomallei is intrinsically resistant to penicillins, aminoglycosides, and macrolides.

Susceptibility testing should be performed to the antimicrobial agents commonly used to

treat melioidosis, which are ceftazidime, imipenem or meropenem, amoxicillin-clavulanate,

doxycycline, and trimethoprim-sulfamethoxazole (TMP-SMX).B. pseudomallei is usually

susceptible to all of these agents with the exception of TMP-SMX, reported rates of resistance

for which are in the order of 2% in Australia (135, 193) and 13 to 16% in northeast Thailand

(164,255). Disk diffusion testing of TMP-SMX overestimates resistance and is unreliable

(164, 193, 255); acceptable alternatives include Etest, broth microdilution, and agar dilution.

Fluoroquinolones are associated with a high rate of therapeutic failure (32) and should not be

included in the test panel. Recent studies indicate that tigecycline has good activity

against B. pseudomallei in vitro and is effective when combined with other agents in an

animal model of B. pseudomallei infection (97, 219).

Current trends in the management of melioidosis involve an initial 10- to 14-day intensive

therapy phase with ceftazidime or meropenem, followed by eradication therapy with TMPSMX

with or without doxycycline for at least 3 months (31, 37, 47, 248). In Australia, TMPSMX

is added to ceftazidime or meropenem during the intensive phase for neurological,

prostatic, cutaneous, and bone and joint melioidosis. Amoxicillin-clavulanate is recommended

for eradication therapy in pregnancy and is an alternative to TMP-SMX in children (36). In

critically ill patients requiring intensive care, meropenem or imipenem may be superior to

ceftazidime, and granulocyte colony-stimulating factor is being used in some centers,

although a study from Thailand showed no benefit (38, 40, 43, 135). From a molecular

genotyping study of cases of recurrent melioidosis, relapse following antimicrobial therapy

occurred in 9.7% of patients and a new infection occurred in 3.4% (152).

Because of the potential role of B. mallei as a bioterrorism agent, studies have been done

recently to determine the activities of a variety of agents against this species. B. mallei has a

susceptibility profile similar to that of B. pseudomallei, except that B. mallei is susceptible to

aminoglycosides and newer macrolides such as clarithromycin and azithromycin, whereas B.

pseudomallei is resistant (120, 142). Current recommended treatment and duration of

therapy for glanders are the same as those for melioidosis.

Guidelines on the management of accidental laboratory exposure to B. pseudomallei and B.

mallei have been recently published (192).

S. maltophilia is intrinsically resistant to many classes of antibiotics. Resistance can also

develop rapidly during infection (101). Resistance to β-lactam agents is mediated by at least

two β-lactamases, one of which is zinc dependent and resistant to β-lactamase inhibitors and

confers resistance to imipenem. Aminoglycoside and quinolone resistance results from

mutations in outer membrane proteins. In a study of isolates recovered from patients with

CF, doxycycline was the most active agent in vitro (204). TMP-SMX is usually active and is

often used in combination with ticarcillin-clavulanate, minocycline, or piperacillin-tazobactam

(204). Other combinations that may be effective include ciprofloxacin paired with ticarcillinclavulanate,

ciprofloxacin and piperacillin-tazobactam, or doxycycline and ticarcillinclavulanate.

Tigecycline is reported to have good activity in vitro (172). CLSI interpretive

criteria for disk diffusion susceptibility testing are available for minocycline, levofloxacin, and

TMP-SMX (50). However, broth microdilution, Etest, or agar dilution methods are the

preferred susceptibility testing methods (9, 264). Many U.S. laboratories comment only on

the activity of TMP-SMX but will test additional antibiotics such as minocycline, ceftazidime,

ticarcillin-clavulanate, and ciprofloxacin or levofloxacin upon request.

In general, Comamonas testosteroni is susceptible to extended- and broad-spectrum

cephalosporins, carbapenems, quinolones, and TMP-SMX (15). D. acidovorans is frequently

resistant to the aminoglycosides.

EVALUATION, INTERPRETATION, AND REPORTING OF

RESULTS Back to top

The species discussed in this chapter are found in the natural environment and may

occasionally contaminate clinical specimens. Nevertheless, they are increasingly recognized

as nosocomial and opportunistic pathogens, especially in certain patient populations, such as

persons with CF. They are also frequently misidentified by commercial microbial identification

systems. Therefore, their recovery in the clinical laboratory must be given careful

consideration. In particular, species of the B. cepacia complex are not reliably differentiated

by phenotypic analyses, and their recovery from persons with CF has serious consequences

with respect to patient management and psychosocial well-being (154). Identification of

these species should be confirmed by genotypic analyses at a reference laboratory and

should promptly be reported to the CF care team. Recovery of B. pseudomallei and B.

mallei in any context should always be considered to reflect clinical disease. Identification of

these species should be confirmed by a reference laboratory with experience with these

species. Care must be given to ensure that culture handling and shipping comply with

current biosafety regulations (see chapters 6 and 12). Identification of these species must be

reported to public health officials due to the potential of these species as agents of

bioterrorism (see chapter 7). The relevance of the recovery of the other genera described in

this chapter, outside the context of CF, is less clear and should be interpreted with caution.

Interpretive criteria for disk diffusion antimicrobial susceptibility testing of most of these

species are lacking; MIC broth microdilution and the Etest are therefore the preferred

methodologies for susceptibility testing. For multiresistant strains, consideration could be

given to testing for synergy with double or triple combinations of antimicrobial agents in

reference laboratories (1). It is important to note, however, that neither checkerboard MIC

broth microdilution testing nor multiple combination bactericidal antibiotic testing is

standardized at present.