Campylobacter and Arcobacter

TAXONOMY Back to top

Three closely related genera, Campylobacter, Arcobacter, and Sulfurospirillum, are included

in the familyCampylobacteraceae (99, 126). The family Campylobacteraceae includes 22

species within the genusCampylobacter, 7 species in the genus Arcobacter, and 6 species in

the genus Sulfurospirillum. Since the completion of the first genome sequence

of Campylobacter jejuni subsp. jejuni (103), additional complete genome sequences of C.

jejuni, C. coli, C. lari, and C. upsaliensis have been published (42, 54, 107) with genome

sizes varying from 1.59 to 1.85 Mb. Two reviews on comparative genomics

of Campylobacter have been published (18, 80). Since the last edition of this Manual, several

new species and subspecies ofCampylobacter have been proposed including C.

canadensis (61), isolated from whooping cranes at the Calgary Zoo; C. avium from poultry

(113), C. peloridis (26) isolated from human feces, dialysis fluid, and shellfish; two

subspecies of C. lari (26); and C. cuniculorum from rabbits (140); as well as

two Arcobacter species, Arcobacter mytili (23), isolated from mussels and A. thereius, from

pigs and ducks (55). Bacteroides ureolyticus was reclassified as Campylobacter

ureolyticus (129). A detailed review on the taxonomy of Campylobacteraceae was recently

published (25).

DESCRIPTION OF THE AGENTS Back to top

Campylobacter spp. are curved, S-shaped, or spiral rods that are 0.2 to 0.9 μm wide and 0.5

to 5 μm long. Occasional species such as C. hominis form straight

rods. Campylobacter species are gram-negative, non-spore-forming rods that may form

spherical or coccoid bodies in old cultures or cultures exposed to air for prolonged periods.

Organisms are usually motile by means of a single polar unsheathed flagellum at one or both

ends, but some may lack flagella. Species are generally microaerobic with a respiratory type

of metabolism; however, some strains grow aerobically or anaerobically. An atmosphere

containing increased hydrogen is required by some species for microaerobic growth (128).

Arcobacter spp. are gram-negative, slightly curved, curved, S-shaped, or helical non-sporeforming

rods that are 0.2 to 0.9 μm wide and 1 to 3 μm long. Organisms are motile with a

single polar unsheathed flagellum.Arcobacter spp. grow microaerobically at 15, 25, and 30°C

but have variable growth at 37 and 42°C. Organisms are microaerobic and do not require

increased hydrogen for growth. Arcobacter spp. may grow aerobically at 30°C and

anaerobically at 35 to 37°C. Most strains are nonhemolytic. A. skirrowii may be alphahemolytic

(130). A. halophilus is an obligate halophile and grows poorly on media containing

less than 2% NaCl (31).

Originally classified as free-living Campylobacter species, Sulfurospirillum spp. are slender,

curved gram-negative rods, 0.1 to 0.5 μm wide and 1 to 3 μm long. All of the species are

sulfur reducers and exhibit variable metabolic activity. S. deleyianum is the type species of

the genus. These species have no known pathogenicity for humans or animals, are

environmental organisms isolated from water sediments, and are not further discussed in

this chapter chapter 126.

EPIDEMIOLOGY AND TRANSMISSION Back to top

Campylobacter species are primarily zoonotic, with a variety of animals implicated as

reservoirs for human infection (Table 1). In addition to food animals such as poultry, cattle,

sheep, and pigs, Campylobacter species may be present in domestic pets. Humans appear to

be the only recognized reservoirs for the periodontal-disease-related species C. concisus, C.

rectus, C. curvus, and C. showae.

Campylobacter infections are common both in the developed and developing worlds. The

reported incidence of culture-confirmed infections varies considerably from country to

country, and as culturing practices and reporting requirements can vary, direct comparison

of the reported incidences can be complex. In the United States, where reporting practices

vary from state to state, the foodborne diseases active surveillance program FoodNet

(www.cdc.gov/foodnet) provides uniform reporting from a panel of sentinel sites, giving an

accurate incidence of diagnosed infections. Since FoodNet surveillance began in 1996, the

incidence of culture-confirmed Campylobacter infections in FoodNet sites has been found to

have declined 30% when 2006 data are compared to the 1996–1998 baseline (2). Most of

the decline occurred from 1996 to 1999; there was a more moderate decrease between 1999

and 2006. In 2008, the incidence of laboratory-confirmedCampylobacter infections in

FoodNet was 12.68 per 100,000 persons, ranging from 30.23 in California to 6.66 per

100,000 in Maryland (17); this estimated incidence of infections caused

by Campylobacter did not change significantly compared with the incidence for the preceding

3 years (2005 to 2007). However, C. jejuni subsp.jejuni (referred to as C. jejuni) continues

to be the most common enteric pathogen isolated from patients reported from some states

in FoodNet with 1.4 million cases estimated in the United States annually (115). Because of

underdiagnosis and underreporting, the actual incidence in any country is substantially

greater than the reported incidence. For example, in the United States, it was estimated that

the true incidence was 35-fold higher than reported incidence, or 515/100,000 in 1999

(115). In Europe, campylobacter infection is quite common, with an incidence rate of

approximately 50/100,000 population (96).

The epidemiology of campylobacteriosis in the United States does not appear to have

changed over the last 20 years (96). Campylobacter infections are usually sporadic; the

incidence starts to rise in March with a peak in the summer months and declines in early fall

(96). Infection usually follows ingestion of improperly handled or cooked food, primarily

poultry products. Case-control studies both in the United States and Europe continue to find

eating poultry to be a significant risk factor for developing campylobacteriosis (96).

Outbreaks usually occur in the spring and fall months, and in recent years, most outbreaks

have been associated with food (poultry or unpasteurized dairy products) or water.

Approximately one-half of the outbreaks in the United States from 1998 to 2004 were

associated with dairy products or water; the remaining outbreaks were mostly foodborne,

and 44% were attributed to poultry (96). The number of foodborne Campylobacter outbreaks

in the United States appears to be increasing; from 1998 to 2002 there were 64

foodborne Campylobacter outbreaks causing 1,628 illnesses, compared to 92 outbreaks and

1,431 illnesses during 2003 to 2007.

(http://www.cdc.gov/foodborneoutbreaks/outbreak_data.htm). Outbreaks in other

developed countries are also associated with food, water, or dairy contamination (96). The

incidence of Campylobacter infection in developing countries such as Mexico and Thailand is

much higher than in the United States. In developing countries, Campylobacter is frequently

isolated from individuals who may or may not have diarrheal disease. Most symptomatic

infections occur in infancy and early childhood, and incidence decreases with age. Travelers

to developing countries are at risk for developing Campylobacter infection, with isolation

rates from 0 to 39% reported in different studies. The incidence of infection follows a

bimodal age distribution with the highest incidence in infants and young children followed by

a second peak in young adults 20 to 40 years old (115). Secondary transmission

of Campylobacter from ill persons to other individuals is rare, even though the infectious

dose for developing illness is not particularly high (96).

CLINICAL SIGNIFICANCE Back to top

C. jejuni and C. coli

C. jejuni and C. coli have been recognized since the early 1970s as agents of gastrointestinal

infection. C. jejuni is one of the most common causes of bacterial enteritis in the United

States. C. jejuni and C. colicontinue to be the most common Campylobacter species

associated with diarrheal illness and produce clinically indistinguishable infections. Most

laboratories do not routinely distinguish between these organisms. In patients with

gastroenteritis caused by C. jejuni/C. coli, patients’ symptoms range from none to severe,

including fever, abdominal cramping, and diarrhea (with or without blood/fecal white cells)

that lasts several days to more than 1 week (10). The usual incubation period is about 3

days with a general range of 1 to 7 days. Symptomatic infections are usually self-limited, but

relapses may occur in 5 to 10% of untreated patients (10). Campylobacter infection may

mimic acute appendicitis and result in unnecessary surgery. Extraintestinal infections have

been reported following Campylobacter enteritis and include bacteremia, hepatitis,

cholecystitis, pancreatitis, abortion and neonatal sepsis, nephritis, prostatitis, urinary tract

infection, peritonitis, myocarditis, and focal infections including meningitis, septic arthritis,

and abscess formation (10). Bacteremia has been reported to occur at a rate of 1.5 per

1,000 intestinal infections with the highest rate in the elderly (118). Persistent diarrheal

illness and bacteremia may occur in immunocompromised hosts, such as patients with

human immunodeficiency virus infection, and may be difficult to treat (10). Deaths

attributable toC. jejuni infection are uncommon (10). The health burden of

campylobacteriosis appears to be substantial and may be underrecognized (84).

C. jejuni is the most often recognized infection preceding the development of Guillain-Barre

syndrome (GBS), an acute paralytic disease of the peripheral nervous system (89). Certain

heat-stable (HS) serotypes appear to be overrepresented in some GBS cases, such as HS:19

and HS:41; however, other more common serotypes are frequently reported (89). The

pathogenesis of Campylobacter-induced GBS involves host immune responses to

gangliosidelike epitopes present in the core region of the lipooligosaccharide (44), which in

the susceptible host mediate damage to the peripheral nerves, where ganglioside targets are

highly enriched (138).

Reactive arthritis sometimes follows Campylobacter infection, with the onset of pain and joint

swelling averaging 2 weeks, with an average range lasting from a few weeks to nearly a

year. Reiter’s syndrome may also occur in some patients (10). The literature is mixed on the

role of HLA B27 as a risk factor for reactive arthritis (10).

The pathogenesis of Campylobacter enteric infection is still not well understood. The infective

dose ofCampylobacter is not well defined, but as few as 500 organisms may be capable of

causing illness (10). The signs and symptoms of infection suggest an invasive mechanism of

disease. A variety of determinants may be important in the virulence of C. jejuni infection,

including adherence to the intestinal mucosa, bacterial effects on the cell, and inflammatory

responses by the host (68). Campylobacter does not produce a classic, choleralike

enterotoxin (137).

Campylobacter Species Other than C. jejuni and C. coli

Campylobacter species other than C. jejuni and C. coli are increasingly isolated from human

infections by improved culture methods that are more optimal for the non-C. jejuni and non-

C. coli species.

C. fetus subsp. fetus is primarily associated with bacteremia and extraintestinal infections

during pregnancy or in the compromised host (11). Although gastroenteritis does occur with

this species, the incidence is probably underestimated because the organism may not grow

well at 42°C and is usually susceptible to cephalothin (cefalotin), an antimicrobial agent used

in some common selective media for stool culture (132). C. fetus subsp.fetus produces a

surface protein microcapsule composed of a high-molecular-weight surface layer protein that

is essential for virulence (11). C. fetus subsp. venerealis causes bovine venereal

campylobacteriosis and is a cause of bovine infertility but is rarely the cause of human

infection (11).

C. upsaliensis is a thermotolerant species that causes diarrhea and bacteremia in humans

and is also associated with canine and feline gastroenteritis (69). Over a 10-year period, C.

upsaliensis was the most common non-C. jejuni/coli species isolated from stool samples

submitted to the laboratory for culture (133). C. upsaliensis is susceptible to many

antimicrobial agents present in C. jejuni selective media and thus is usually not isolated on

routine primary isolation media; it can be recovered using the filtration technique described

below.

C. lari is a nalidixic acid-resistant, thermophilic species first isolated from gulls of the

genus Larus and from other avian species, dogs, cats, and chickens. C. lari has been

infrequently reported from humans with bacteremia and gastrointestinal and urinary tract

infections (69). Recent phylogenetic studies have described two subspecies, C.

lari subsp. concheus and C. lari subsp. lari (26).

Other Campylobacter species have been isolated from clinical specimens of patients with a

variety of diseases, but their pathogenic role has not been determined (69). C.

jejuni subsp. doylei is a nitrate-negative subspecies of C. jejuni rarely recovered from

patients with upper gastrointestinal tract infections and gastroenteritis (69).C.

hyointestinalis has been occasionally associated with proctitis and diarrhea in human

infection. C. concisus is associated primarily with periodontal disease but has also been

isolated from patients with bacteremia, foot ulcer, and upper and lower gastrointestinal tract

infections (69). Although C. concisus has been isolated from many patients with diarrheal

illness, it can also be isolated from the feces of healthy individuals, and there is no

convincing evidence to date that it causes diarrhea (35). C. sputorum has been associated

with lung, axillary, scrotal, and groin abscesses (70). C. sputorum bv. paraureolyticus,

formerly referred to as catalase-negative urease-positive campylobacter, has been isolated

from patients with diarrhea, but the significance of this finding is unknown (101). C.

mucosalis was reported to have been isolated from two children with enteritis, but

subsequent testing showed that the isolates were actually C. concisus (100). C.

helveticus (119) has been recovered from domestic cats and dogs and has not been reported

from human sources. C. rectus is primarily isolated from patients with active periodontal

infections but has also been isolated from a patient with pulmonary infection (69, 111) and

breast abscess (49). There is some suggestion that C. curvus may be an etiologic agent in

diarrheal illness (1), but it was rarely isolated from stool samples in another large study

(35). C. curvus is also isolated from patients with periodontal infections and in patients with

a liver abscess and pneumonia (49). C. showae has been isolated from the human gingival

crevice (36). C. gracilis has been isolated from patients with appendicitis/peritonitis,

bacteremia, soft-tissue abscesses, and pulmonary infections (85). C. hominis has been

isolated from fecal samples of healthy individuals and may be a commensal of the oral cavity

(71). C. lanienae was isolated from two asymptomatic abattoir workers, but its clinical

significance is unknown (73). C. canadensis (61) has been isolated from whooping cranes at

the Calgary Zoo and C. peloridis (26) from human feces, dialysis fluid, and shellfish. C.

cuniculorum was isolated from the cecum of rabbits but not reported from humans (140). C.

avium is a hippurate hydrolase-positive species that was isolated from broiler chickens and

turkeys but has not been reported from human samples (113). A review on the clinical

significance of non-C. jejuni/C. coli species was published previously (69).

Arcobacter

Arcobacter spp. are aerotolerant, Campylobacter-like organisms frequently isolated from

bovine and porcine products of abortion and feces of animals with enteritis (39). Two of the

seven Arcobacter species have been associated with human infection. A. butzleri has been

isolated from patients with bacteremia, endocarditis, peritonitis, and diarrhea (70, 132). A.

cryaerophilus has been previously characterized into two DNA related groups, 1A and 1B

(65). A. cryaerophilus group 1B has been isolated from patients with bacteremia and

diarrhea (70, 132) and also from healthy individuals (57), suggesting a commensal role for

this species. Group 1A has been isolated from animal sources (65). Arcobacter butzleri was

reported to be the fourth most commonCampylobacter-like organism isolated from patients

with diarrhea by Vandenberg et al. (132) and was also one of the most common non C.

jejuni/coli species isolated over a 10-year period from over 73,000 stool samples (133).

Thus, A. butzleri may be underrecognized if appropriate culture conditions are not used. In a

survey of 2,855 Campylobacter-like isolates submitted for characterization from laboratories

in France, A. butzleri was identified in 1%, primarily from fecal samples of patients with a

diarrheal illness (109). A. skirrowii was reported to be isolated from a human stool culture in

a patient with chronic diarrhea, but the role of this species in human disease is unknown

(139). A. nitrofigilis, a nitrogen-fixing bacterium found on the roots of a small marsh plant in

Nova Scotia, is not associated with human disease (39). Arcobacter cibarius has been

isolated only from poultry carcasses; the medical significance of this species is unknown

(56). Arcobacter halophilus requires increased salt for growth in culture; however, the

medical significance of this species is unknown (31). Arcobacter thereius has been isolated

from liver and kidney of spontaneous porcine abortions and from the cloacae of ducks but

has not been reported from human samples (55). Arcobacter mytili was recently isolated

from shellfish from northeastern Spain and has not been reported from human samples

(23).

COLLECTION, TRANSPORT, AND STORAGE OF

SPECIMENS Back to top

Fecal Samples

Fecal specimens are preferred for isolating Campylobacter species from patients with

gastrointestinal infections; however, rectal swabs are acceptable for culture. For hospitalized

patients, the “3-day” rule (rejection of specimens collected >72 h after admission) should be

used as a criterion for acceptability of routine culture requests (48, 53). For routine

purposes, a single stool sample has high sensitivity for common enteric pathogens, but two

samples may be desirable, depending upon clinical circumstances such as a >2-h delay in

transport of the first sample that could affect recovery (39). A transport medium should be

used when a delay of more than 2 h is anticipated and for transporting rectal swabs. Several

types of transport media are useful for Campylobacter including alkaline peptone water with

thioglycolate and cystine, modified Stuart medium, and Cary-Blair medium (39). Transport

media such as commercial Stuart medium and buffered glycerol saline do not appear to

perform well. Modified Cary-Blair medium containing reduced agar (1.6 g/liter) appears to be

the most suitable single transport medium for Campylobacter as well as other enteric

pathogens. Specimens received in Cary-Blair medium should be stored at 4°C if processing is

not performed immediately. Use of Cary-Blair medium supplemented with laked sheep blood

may be useful for prolonged storage of stool samples and recovery of C. jejuni (136).

Blood

Campylobacter species, primarily C. fetus, C. jejuni, and C. upsaliensis, have been isolated

from blood; however, in only a few studies have optimal conditions for

isolating Campylobacter from blood culture systems been evaluated. Both the Bactec system

(BD, Sparks, MD) (aerobic bottles) and Septi-Chek system (BD, Sparks, MD) appear to

support the growth of the common Campylobacter species (39). The BacT/Alert system

(bioMerieux, Inc.) also supports the growth of Campylobacter fetus (22). Other systems such

as anaerobic broth or lysis centrifugation may not be as sensitive (39).

DIRECT EXAMINATION Back to top

Microscopy

Clinical microbiologists do not normally consider performing Gram stain analysis of stool

samples for diagnosis of bacterial gastroenteritis; however, this is a rapid and sensitive

method for presumptive diagnosis ofCampylobacter enteritis. Campylobacter spp. are not

easily visualized with the safranin counterstain commonly used in the Gram stain procedure

and are somewhat thinner than other enteric gram-negative bacteria; carbol-fuchsin or 0.1%

aqueous basic fuchsin should be used as the counterstain for smears of stools or pure

cultures (39). Because of their characteristic morphology, Campylobacter spp. may be

detected by direct Gram stain examination of stools obtained from patients with acute

enteritis, with sensitivity ranging from 66% to 94% and specificity above 95%. Phasecontrast

and dark-field microscopy have also been used to directly detect motile

campylobacters in fresh stool samples; however, the sensitivity of these approaches has not

been studied widely, and in our opinion these methods require significant microscopic

expertise (39).

Fecal white cells may be present during Campylobacter infection and have been reported in

25% to 80% of culture-proven cases (53). There is no known correlation between the

number of cells present and infection. While the likelihood of infection with Campylobacter or

other enteroinvasive pathogens may be higher in the qualitative presence of fecal

leukocytes, the absence of fecal leukocytes does not rule out the diagnosis. Thus, routine

examination of stool samples for fecal leukocytes is not recommended as a test for predicting

bacterial infection or for selective culturing for Campylobacter or other stool pathogens

(39, 48).

Antigen Detection

Several commercially available antigen detection systems for Campylobacter in stool samples

are now currently available; the ProSpecT Campylobacter assay (Alexon-Trend, Inc.,

distributed through Remel), the Premier Campy Campylobacter assay (Meridian Biosciences),

and the ImmunoCard Stat! Campy assay (Meridian Biosciences). When compared with

culture, the ProSpecT immunoassay has been shown to vary in sensitivity from 80 to 96%

and has a specificity of >97% (28, 52, 124). This enzyme immunoassay (EIA) was found to

cross-react with C. upsaliensis (28). Antigen may be detected in stored stool samples at 4°C

for several days (32). The Premier Campylobacter assay is a microtiter plate-based EIA,

while the ImmunoCard STAT! Campy assay is a one-step lateral flow immunoassay; both are

reported to be specific for the detection ofCampylobacter jejuni and C. coli, but there are

limited available data on their performance characteristics. Other EIAs available outside the

United States have variable performance (123). Given that a Campylobacter infection is a

low-incidence disease, the specificity values described to date for the Campylobacter antigen

detection assays mentioned above suggest that these tests can lead to poor positive

predictive values.

Nucleic Acid Detection Techniques

Amplification techniques have been used directly to detect Campylobacter in stool samples

(105, 116). Molecular approaches to detecting Campylobacter directly may improve the time

to detection, identification to the species level, and identification of the less

common Campylobacter species often missed by conventional culture. A commercially

available molecular test for detection of Campylobacter spp. in fecal samples is not currently

available. This approach is also more expensive than culture and does not provide an isolate

for further characterization.

ISOLATION PROCEDURES Back to top

Most Campylobacter species require a microaerobic atmosphere containing approximately

5% O2, 10% CO2, and 85% N2 for optimal recovery. Several manufacturers produce

microaerobic gas generator packs that are suitable for routine use. A trigas incubator or

evacuation and replacement of an anaerobic jar with the approximate gas mixture may also

be used for routine cultures (39). The anoxomat (Mart Microbiology, distributed through

Spiral BioTech) is a convenient automated system for the evacuation and gas replacement of

jars used for generating different atmospheric conditions, including microaerobic conditions

(13). The concentration of oxygen generated in candle jars is suboptimal for the isolation

of Campylobacter and should not be used for routine laboratory isolation procedures (39).

Some species of Campylobacter, such as C. sputorum, C. concisus, C. mucosalis, C.

curvus, C. rectus, and C. hyointestinalis, require increased hydrogen for primary isolation

and growth. These species will usually not be recovered under the conventional microaerobic

conditions, since the amount of hydrogen generated in properly used commercial gas-packs

is <2%. A gas mixture of 10% CO2, 6% H2, and the balance N2 used in an evacuationreplacement

jar is sufficient for isolating hydrogen-requiring species. A study by Vandenberg

and colleagues reemphasized the requirement of increased hydrogen for isolating

certain Campylobacter spp. (132).

A number of selective media have been recommended for isolating C. jejuni and C. coli.

These include two blood-free media, charcoal cefoperazone deoxycholate agar (CCDA) and

charcoal-based selective medium (CSM); and two blood-containing media, Campy-CVA

(cefoperazone, vancomycin, amphotericin) medium and Skirrow medium (39). Although CVA

medium is commonly used in the United States for isolating Campylobacterfrom clinical stool

specimens, there are limited data available to assess the ability of CVA to

recoverCampylobacter species from stool specimens, when compared to

other Campylobacter-selective media; additional evaluation studies are warranted. Charcoalbased

media containing cefoperazone, amphotericin, and teicoplanin (CAT media) are

selective media for the primary isolation of C. upsaliensis (7). Two studies, however, did not

isolate C. upsaliensis from any stool samples by use of this medium (35, 52). C.

upsaliensismay occasionally be recovered on some other selective media. C.

upsaliensis isolates can also be recovered by using the filtration method, and some strains

may grow better in a hydrogen-enriched atmosphere (46, 70).

To achieve the highest yield of Campylobacter from stool samples, a combination of media

that includes either CCDA or CSM appears to be the optimal method (33) and may increase

the recovery of Campylobacter by as much as 10 to 15% over the use of a single medium. If

only a single medium is used, we suggest using Campy-CVA, CCDA, or CSM. In a

comparative study, CCDA medium was found to be the most sensitive for detecting C.

jejuni and C. coli compared with Skirrow’s medium, CAT agar, and filtration technique (35).

If Campylobacter infection is suspected at the time blood specimens are drawn, broth media

should be subcultured after 24 to 48 h to a nonselective blood agar medium and plates

incubated under microaerobic conditions at 37°C, preferably with increased hydrogen. This

allows for the isolation of thermophilic and nonthermophilic species. While commonly used

blood culture systems should support the growth ofCampylobacter and give appropriate

signals if positive, it may be prudent to perform a blind subculture. Similarly, blood drawn in

Isolator (Wampole Laboratories, Cranbury, NJ) tubes for bacterial culture should include a

nonselective blood agar plate incubated under microaerobic conditions at 37°C

if Campylobacterinfection is suspected. If a curved, gram-negative rod is observed upon

Gram stain examination of a positive blood culture bottle, an aliquot should be cultured on a

nonselective blood agar plate and incubated under microaerobic conditions at 37°C. An

alternative staining method such as acridine orange may also be useful for detecting

campylobacters in blood culture bottles if the Gram stain is negative.

Optimal conditions for recovery of Arcobacter from clinical specimens have not been

determined. Arcobacterspp. were first isolated on semisolid media designed to

isolate Leptospira spp. (39). Arcobacter species are aerotolerant and have been recovered on

certain selective media such as Campy-CVA (5) incubated under microaerobic conditions at

37°C and on nonselective media used in the filtration method (132). Selective media for

isolation of Arcobacter spp. from human stool samples were evaluated by Houf and Stephan

(57). Both selective plates and enrichment broth containing selective supplements with 5-

fluorouracil, amphotericin B, cefoperazone, novobiocin, and trimethoprim showed good

recovery of Arcobacter sp. (57). Several other media have been reported to

recover Arcobacter species but have not been studied in clinical settings (5, 24,130).

Enrichment Cultures

Enrichment broths formulated to enhance the recovery of Campylobacter from stool include

Preston enrichment, Campy-thio, and Campylobacter enrichment broth (39). Enrichment

cultures may be beneficial in instances where low numbers of organisms may be expected

due to delayed transport to the laboratory or after the acute stage of disease when the

concentration of organisms may be low, such as in the investigation of GBS following

acute Campylobacter infection (87). The clinical advantage and cost-effectiveness of using

enrichment cultures as part of the routine stool culture setup have not been studied

adequately.

Filtration

Filtration techniques designed to isolate C. jejuni and C. coli as well as

other Campylobacter species (35, 39,132) and Arcobacter spp. (65, 132, 133) that are

susceptible to antibiotics present in most selective media should be used to complement

direct culture to selective plating media. As only stool samples containing ~105CFU/ml

of Campylobacter will be detected with filtration, it should not be used as a replacement for

direct culture, because the filtration method is not as sensitive as primary culture with

selective media (46).

The method is based on the principle that campylobacters can pass through membrane filters

(0.45-μm to 0.65-μm pore size) with relative ease (because organisms are thin and highly

motile) while other elements of the stool microbiota are retained during the short processing

time. Cellulose acetate membrane filters with a 0.65-μm pore size are recommended for

routine use and available from a number of suppliers (39). Filtration is performed by placing

a sterile 0.65-μm-pore-size cellulose acetate filter onto the surface of an agar medium such

as antibiotic-free CCDA, CSM, or blood-containing medium. Ten to 15 drops of fecal

suspension are placed on the filter, and the plate is incubated at 37°C for 45 to 60 min. The

filter is then removed, and the plate incubated at 37°C under microaerobic conditions,

preferably with an atmosphere containing increased hydrogen (for the hydrogen-requiring

species).

Species within the genus Campylobacter and Arcobacter have different optimal temperatures

for growth. The choice of incubation temperature for routine stool cultures is critical in

determining the spectrum of species that will be isolated. By convention, most laboratories

use 42°C as the primary incubation temperature forCampylobacter. This temperature allows

growth of C. jejuni and C. coli on selective media while inhibiting other fecal microbiota. C.

upsaliensis also grows well at 42°C but usually is not recovered on selective media. C.

fetus exhibits variable growth at this temperature and may not be

recovered. Arcobacter species will generally not be recovered at 42°C.

In contrast, most Campylobacter and Arcobacter species grow well at 37°C. Selective media,

such as Skirrow medium, were devised for use at 42°C and have poor selective properties at

37°C, whereas CCDA and CSM show good selective properties at 37°C (33). Plates should be

incubated a minimum of 72 h before being reported as negative. It has been reported that

incubation of CCDA medium for 5 to 6 days increased the yield of C. jejuni and C.

coli compared with 2 days of incubation (35).

Because of the expense of including several types of media and the filtration method in the

initial workup forCampylobacter, a practical approach is to use a single medium for isolation

of thermophilic Campylobacter spp. in the workup of acute bacterial gastroenteritis, such as

Campy-CVA, CCDA, or CSM incubated at 42°C. If the primary culture workup is unrevealing

and for patients with persistent diarrhea, cultures for non-C. jejuni/C. coli species may be

appropriate. Additional stool samples should be plated on multiple selective media (e.g.,

CCDA or CVA), processed by the filtration method as well, and incubated at 37°C under

microaerobic conditions with increased hydrogen.

IDENTIFICATION Back to top

The identification of Campylobacter species is made difficult because of their complex and

rapidly evolving taxonomy, fastidious growth requirements, and biochemical inertness (Table

2). These problems have resulted in a proliferation of phenotypic and genotypic methods for

identifying members of this group (39).

Campylobacter spp. and Arcobacter spp.

Depending on the growth medium used, Campylobacter colonies may have different

appearances. In general,Campylobacter spp. produce gray, flat, irregular, and spreading

colonies. Spreading along the streak line is commonly seen, particularly on freshly prepared

media. As the moisture content decreases, colonies may form round, convex, and glistening

colonies with little spreading observed. Thus, proper storage of media to ensure moisture

content is important for optimal isolation and recognition of Campylobacter spp. Hemolysis

on blood agar is not observed. Arcobacter colonies are morphologically similar to those

of Campylobacter (126, 130).

The Gram stain appearance of Arcobacter may differ from that of typical Campylobacter. A.

butzleri is only slightly curved, while A. cryaerophilus tends to be much more helical in

appearance than Campylobacter. Commercial systems for identification

of Campylobacter species were not found to be more accurate than conventional tests (60).

Unfortunately, Campylobacter species are difficult to differentiate from Arcobacterspecies

based on phenotypic tests. However, an aerotolerant species (i.e., exhibiting growth under

aerobic conditions) that grows on MacConkey agar under microaerobic conditions could be

presumptively identified asArcobacter. The failure to grow on MacConkey, however, does not

rule out Arcobacter species.

C. jejuni and C. coli

For initial analysis, a Gram stain examination of the colony should be performed along with

an oxidase test. Oxidase-positive colonies exhibiting a characteristic Gram stain appearance

(e.g., gram-negative, curved to S-shaped rods) isolated from selective media incubated at

42°C under microaerobic conditions can be reliably reported as Campylobacter spp. The most

common species, C. jejuni, is relatively easy to identify phenotypically; hydrolysis of sodium

hippurate is the major test for distinguishing C. jejuni (and also C. jejunisubsp. doylei) from

other Campylobacter species. Strains isolated on selective media that grow at 42°C, are

oxidase positive, show characteristic microscopic morphology, and are positive for hippurate

hydrolysis should be reported as C. jejuni, and for routine clinical purposes no other tests

need to be performed (Fig. 1). Methods for this test are described elsewhere (74).

Occasional strains of C. jejuni are hippurate negative, making them more difficult to identify.

Gas-liquid chromatography for detecting benzoic acid (liberated from hydrolysis of sodium

hippurate) is the most sensitive assay for hippurate hydrolysis and can be used for more

definitive determination. Molecular detection of the hipO (hippuricase) gene or other C.

jejuni-specific markers (Table 3) by PCR may be useful for identifying phenotypically

negative isolates (50) and weakly positive isolates (15) and to clarify false-positive results

for non-C. jejuni species (27, 29). A comparison of different PCR targets to differentiate C.

jejuni from C. coli has been published (31, 112). Evaluations of these C. jejuni-specific PCR

tests have shown no one test to be entirely specific or sensitive; therefore, the use of more

than one target for molecular identification of C. jejuni is recommended. In addition, false

negatives or nonspecifically amplified product(s) have been noted for some of the C. jejunispecific

assays; therefore, a second PCR, targeting another C. jejuni-specific gene, may be

necessary in some instances. The use of heated lysates rather than purified DNA may not

always be a suitable reaction template for these PCR assays (83,97).

With the exception of hippuricase activity, which C. coli is lacking, C. coli and C. jejuni are

similar biochemically (Table 2). Therefore, molecular methods are needed to accurately

identify C. coli and differentiate it from hippurate-negative C. jejuni; most have proved both

accurate and sensitive (97). If molecular testing is not available, strains isolated on selective

media that grow at 42°C, are oxidase positive, show characteristic microscopic morphology,

and are hippurate negative and indoxyl acetate positive should be reported as hippuratenegative

C. jejuni/C. coli (Fig. 1). Susceptibility (inhibition) or resistance

of Campylobacter spp. to nalidixic acid and cephalothin were historically used as an aid for

species identification. However, with the increasing prevalence of fluoroquinolone resistance

in these species, the use of these disk identification assays can no longer be relied upon.

For species other than C. jejuni, phenotypic characterization is more problematic. An

algorithm for identification of the thermophilic Campylobacter spp. is shown in Fig. 1. The

most useful tests for initial identification include growth at 25°C, 37°C, and 42°C, catalase

production, hippurate hydrolysis (74), indoxyl acetate hydrolysis (75), and production of H2S

(8).

Additional tests can be performed to aid in the identification of Campylobacter spp.(Fig. 1).

To obtain consistent and reproducible results, a standardized suspension and inoculum

should be used for performing phenotypic tests. For growth temperature and oxygen

tolerance studies, a suspension of the organism in heart infusion broth or tryptic soy broth

with turbidity at a McFarland standard of 1 should be used. A fiber-tipped swab dipped in the

broth suspension should be used to make a single streak across the plate (Mueller-Hinton

agar with 5% sheep blood is a suitable medium), and the plates should be incubated at the

desired temperature and/or atmospheric conditions (8, 86).

Several commercial systems have been developed as an aid to

identifying Campylobacter spp. to the genus level. Two immunologic reagents are currently

available in the United States for culture identification: Campy-JCL (Scimedx Corp., Denville,

NJ) and Dryspot Campylobacter Test Kit (Remel). Campy-JCL was previously evaluated and

does not differentiate between C. jejuni and C. coli (90). The Dryspot Campylobacter latex

test is reported by the manufacturer to identify but not differentiate C. jejuni, C. coli, C.

lari, and C. upsaliensis and to yield variable results for C. fetus subsp. fetus (Oxoid USA,

www.oxoid.com). A DNA probe (Accuprobe; Gen-Probe Inc., San Diego, CA) directed

against Campylobacter rRNA sequences identifies Campylobacter to the genus level and

detects C. jejuni subsp. jejuni, C. jejuni subsp. doylei, C. coli, and C. lari (108, 122).

However, the probe also hybridized with 2 of 17 C. hyointestinalis strains (108). Thus, these

methods may be useful for confirming Campylobacter to the genus level if other tests are not

conclusive. However, they cannot be used to rule out Campylobacter, and the crossreactivity

of these tests with closely related taxa and/or more newly described species needs

to be determined.

Because many species of Campylobacter and Arcobacter are difficult to identify by

phenotypic testing alone, tests for detection of species-specific sequences via PCR have been

developed. The 16S rRNA gene and 23S rRNA gene are widely used for genus- and speciesspecific

tests; PCR assays based on these targets have been described for 12

different Campylobacter species (98) and three Arcobacter species (14, 58). Broad-range

molecular identification schemes involving restriction fragment analysis of PCR-amplified

regions of the 16S or 23S rRNA genes have also been described for identification

of Campylobacter and Arcobacter species (16, 37,38).

Many other gene targets have been used in species-specific PCR assays,

including gyrA (79, 134), glyA (4),ceuE gene (45), asp (72), lpxA (67), and a GTPase gene

(134). Subspecies identification by PCR within C. fetus(59, 125) and C. jejuni (81) has also

been described. While the use of such PCR tests combines the advantages of being quick and

easy to perform with low cost and high-throughput capability, amenable to automation, it is

important to validate PCR tests to fully determine their specificity and sensitivity before use.

More recently, species-specific microarrays have been described for identification of

several Campylobacterspecies, including C. jejuni, C. coli, C. lari, and C.

upsaliensis (64, 110, 135). While these methods are promising tools for both identification

and further genetic characterization of Campylobacter spp., the cost and limited availability

of the technology in the clinical laboratory make this approach not currently practical for

routine application in the clinical setting.

Comparison of 16S rRNA gene sequences is also a useful tool for differentiation

of Campylobacter spp. from closely related taxa, such

as Arcobacter and Helicobacter. However, it is important to note that species level

identification based on 16S is much more difficult, particularly for the common species

of Campylobacter. At or above 97% identity, some groups of closely related species such

as C. jejuni, C. coli, and C. lari; C. upsaliensisand C. helveticus; and C. fetus, C.

hyointestinalis, and C. lanienae cannot be confidently distinguished from each other based on

16S rRNA gene sequences (98). Conversely, intraspecies 16S rRNA gene diversity is seen in

other species, such as C. hyointestinalis (51). A commercial 16S rRNA gene microbial

identification system, MicroSEQ (Applied Biosystems), is available to identify and classify

unidentified bacteria, includingCampylobacter and Arcobacter species, by comparing either

full- or partial-length (500-bp)16S rRNA gene sequences to a validated microbial 16S rRNA

gene sequence library. However, partial gene sequencing using the MicroSEQ 500 system is

not recommended for accurate identification of some of the commonly

encountered Campylobacter species, as sequencing of the first 500 bp of the gene only

examines two of the four variable regions within the 16S rRNA gene

of Campylobacter species (47); this can lead to misidentification of the groups of closely

related Campylobacter species described above.

TYPING SYSTEMS Back to top

Typing systems for Campylobacter epidemiologic studies vary in complexity and ability to

discriminate between strains. Common phenotypic methods that have been applied include

biotyping, phage typing, and serotyping (94, 104). The heat-labile serotyping scheme,

originally described by Lior, can detect over 100 serotypes of C. jejuni, C. coli, and C.

lari (104). Uncharacterized bacterial surface antigens and, in some serotypes, flagellar

antigens are the serodeterminants for this serotyping system (3). The heat-stable Penner

(HS) serotyping scheme detects 60 types of C. jejuni and C. coli (104). Initially thought to

detect lipopolysaccharide antigenic determinants, the HS system has been shown to detect

a Campylobacter capsular polysaccharide (62). Serotyping (HS) is performed in only a few

reference laboratories worldwide because of the time and expense needed to maintain

quality antisera. A serotyping reagent kit is also commercially available (Denka Seiken USA

Inc., Campbell, CA).

The limitations of phenotypic subtyping methods and the rapid growth of molecular biology

techniques in the 1990s led to the development of a range of molecular subtyping methods

such as restriction endonuclease analysis, ribotyping, PCR-based techniques, pulsed-field gel

electrophoresis of macrorestricted chromosomal DNA (PFGE), and amplified fragment length

polymorphism (94, 104). The development of a rapid 1-day standardized PFGE protocol

(112), which is used by participants of the PulseNet national surveillance network for

foodborne pathogens (www.cdc.gov/pulsenet), has facilitated the use of this approach for

outbreak investigations of campylobacteriosis (95). However, interpretation of data can be

difficult since genomic rearrangements can lead to changes in PFGE profiles (40). Advances

in DNA sequencing technology have provided a means to investigate strain variation at the

nucleotide level and have led to the emergence of DNA-sequencing-based subtyping systems

such as multilocus sequence typing (MLST) (30). MLST is useful for studies of the population

structure and molecular epidemiology of C. jejuni (30), although commonly circulating

sequence types can make recognition of outbreaks caused by these strains problematic

(19, 114). In addition, the generation of Campylobacter whole-genome sequences has led

the way for the development of a new technique, genomotyping based on microarray

technology (66). At present, no method alone is adequate for all applications, and a

combination of methods such as serotyping and molecular methods should be used for

reliable determination of strain relatedness and for studying the epidemiology

of Campylobacter infections (94). A more detailed discussion of molecular typing methods

and their application for use in epidemiological studies of Campylobacter species has been

recently published elsewhere (102).

An MLST system for Arcobacter spp. was recently reported (82). Using a set of 374 isolates

comprising different species, from different sources and geographic locations, no association

of MLST type and location or source was observed.

SEROLOGIC TESTS Back to top

Serum immunoglobulin G (IgG), IgM, and IgA levels rise in response to infection, but serum

and fecal IgA levels appear during the first few weeks of infection and then fall rapidly

(9, 121). Serum antibody assays vary in both sensitivity and specificity for

detecting Campylobacter infection, and test performance appears to be population

dependent. Campylobacter antibody assays have been used to study patients with GBS and

reactive arthritis (6). Patients with Campylobacter infection may give falsepositive

Legionella antibody test results (12). Serologic testing appears to be useful for

epidemiologic investigations and is not recommended for routine diagnosis (120).

ANTIMICROBIAL SUSCEPTIBILITIES Back to top

C. jejuni and C. coli have variable susceptibility to a variety of antimicrobial agents, including

macrolides, fluoroquinolones, aminoglycosides, chloramphenicol, nitrofurantoin, and

tetracycline (www.cdc.gov/NARMS). Azithromycin and erythromycin are drugs of choice for

treating C. jejuni gastrointestinal infections, and for susceptible organisms, ciprofloxacin or

norfloxacin may also be used. Early therapy of susceptibleCampylobacter infection with

erythromycin or ciprofloxacin is effective in eliminating the organism from stool and may also

reduce the duration of symptoms associated with infection (10).

C. jejuni is generally susceptible to erythromycin, with resistance rates of less than 10%

(10, 34, 43). Rates of erythromycin resistance in C. coli are generally higher than in C.

jejuni and vary considerably, with up to 25 to 50% of strains showing resistance in some

studies (10, 43). Although ciprofloxacin has been effective in

treating Campylobacter infections, emergence of ciprofloxacin resistance during therapy has

been reported (106). Several in vitro studies show significant rates of resistance to

fluoroquinolones (34, 63, 91). Resistance to fluoroquinolones has ranged from <5% in

Australia to approximately 80% reported in Thailand (10, 41). In 2006, approximately 20%

of Campylobacter strains reported through the National Antimicrobial Resistance Monitoring

System at CDC were fluoroquinolone resistant (http://www.cdc.gov/NARMS/). Individuals

with fluoroquinolone-resistant C. jejuni have been shown to have a longer duration of

diarrhea, and thus, routine testing of isolates may be indicated (10, 93). C. jejuni and C.

coli are resistant to β-lactam antibiotics, generally penicillins and narrow-spectrum

cephalosporins, but imipenem has good anticampylobacter activity.

Parenteral therapy is used to treat systemic C. fetus infections; drugs used include ampicillin,

aminoglycosides, imipenem, and chloramphenicol, depending upon the type of infection. C.

lari is resistant to nalidixic acid but may be susceptible to fluoroquinolones, and resistance to

macrolides is generally low (41). C. upsaliensis is generally susceptible to a variety of

antimicrobial agents and shows low rates of resistance to macrolides and fluoroquinolones

(41). Arcobacter butzleri and A. cryaerophilus have variable resistance to macrolides and

fluoroquinolones (41).

Agar dilution is the method recognized by the Clinical Laboratory Standards Institute (CLSI)

for testingCampylobacter spp.; quality control ranges for several antimicrobial agents have

been published (21, 76). A broth microdilution method with published quality control ranges

for several antimicrobial agents and a disk diffusion method are also approved by CLSI

(20, 77). Studies testing Campylobacter with the Etest have been published (78).

EVALUATION, INTERPRETATION, AND REPORTING OF

RESULTS Back to top

Campylobacter species, including the common thermophilic species C. jejuni and C.

coli, should be sought in all diarrheic stools submitted to the laboratory for routine culture.

Except for epidemiological purposes, cultures of formed stools should not be performed.

Isolation of Campylobacter from a patient with acute diarrhea is usually significant, since the

carrier rate in developed countries is quite low; however, in developing countries, isolation

might be more difficult to interpret, especially in the presence of other enteric pathogens. In

acute infection, there are usually a high number of organisms in the stool, but the quantity

of organisms is not related to the severity of infection or indicative of a carrier state. Gram

stain analysis of fecal samples to look for organisms with typical Campylobacter morphology

is a highly sensitive and specific test that is currently underutilized; it should be performed

for rapid preliminary diagnosis of Campylobacter infection. Other species, such as C.

fetus subsp. fetus and C. upsaliensis, may be important causes of diarrhea and are not

isolated on routine selective media. Special methods including alternative incubation

techniques are required as described in this chapter and should be performed by special

request. Oxidase-positive, curved, gram-negative rods that are hippurate hydrolysis positive

should be reported as C. jejuni without further workup. The importance of identifying other

species depends upon the clinical circumstance, but identification tests should always be

performed with isolates from blood or other sterile sites, since this could influence

antimicrobial therapy decisions. Given that fluoroquinolone resistance is present in a

significant proportion of C. jejuni isolates, fluoroquinolone susceptibility testing is suggested

for patients that are receiving or being considered for therapy of gastroenteritis.

Susceptibility testing should be performed with all isolates from sterile clinical sites.

Use of trade names is for identification only and does not imply endorsement by the Public

Health Service or by the U.S. Department of Health and Human Services.