Presence Of Antimicrobial-Resistant Pathogens In Retail Poultry Products


An Information
Paper Presented by
Consumers International
to the Codex Committee on Residues of Veterinary Drugs in Foods
14th Session, 4 to 7 March, 2003
Washington, DC, USA


Bacterial resistance to antimicrobial
drugs is a worldwide problem that has also emerged among common foodborne pathogens.
Three CI members, each an independent consu-mer organization, the Australian
Consumers’ Association (ACA), Consumers Union of United States (CU), and the
Institutue for Agriculture and Trade Policy (IATP), have recently tested poultry
products purchased at retail for the presence of antimicrobial-resistant pathogens.

ACA surveyed chickens purchased in
Brisbane and Sydney, testing enterococci isolated from these birds for resistance
to vancomycin. CU purchased a nationally representative sample of chickens from
around the US, and tested campylobacter isolates for resistance to six antibiotics,
and salmonella isolates for resistance to seven antibiotics. IATP, working with
the Sierra Club (SC), a US environmental NGO that is not a CI member, purchased
whole chickens and ground turkey meat in two midwestern US cities; isolated
salmonella, campylobacter and enterococci from their samples; and tested selected
bacterial isolates of each type for resistance to up to 17 different antimicrobial

Materials and Methods

During January 2002, ACA purchased
141 fresh or frozen whole chickens from major supermarkets in Brisbane and Sydney.
Samples were shipped in chilled containers to a contract laboratory, where they
were processed with a bag rinse method. Each sample was aseptically removed
from its packaging and placed in a sterile bag. A sterile rinse fluid (0.1%
buffered peptone water with 0.5% tween 80) was added, the bag was massaged,
and the rinse liquid was transferred to a sterile container. 0.1 ml aliquots
of the rinse fluid were streaked onto plates of vancomycin-resistant enterococci
(VRE) agar (Aesculin bile agar with vancomycin supplement), S&Bv agar (Slanetz
and Bartley agar with vancomycin supplement), and plain S&B agar. The agar
plates were incubated at 37º C, and colonies were examined.

Colonies from samples that exhibited
typical morphology in all media were screened for susceptibility to vancomycin
using the Calibrated Dichotomous Sensitivity (CDS) method. Colonies were simultaneously
confirmed by growth at 44.5º C and in 6.5% salt, and biochemical profiling
using ß haemolysis and d-xylose fermentation. Additional tests were carried
out to identify isolates as Enterococcus faecalis.

CU tested 484 fresh whole broilers
representing 29 different brands purchased from over 100 retail outlets in 25
metropolitan areas, nationwide, in a 3-week period in April and May, 2002. The
birds were packed in coolers and shipped overnight to a laboratory for processing.
Samples were processed using a bag rinse method similar to that described above.
After aseptic removal of packaging, necks and giblets, each chicken was placed
in a sterile bag with 400 ml of sterile Butterfield’s Phosphate Buffer. The
bag was then shaken for one minute and the rinse fluid was transferred to a
sterile container.

Portions of the rinsate from each
chicken sample were transferred to separate enrichment broths for salmonella
and campylobacter, and analyzed using standard methods.(1)
CU’s tests had two objectives. The first was to determine the prevalence of
these two pathogens in retail chicken samples. Overall, 42 percent of the chickens
contained campylobacter; 12 percent contained salmonella; 5 percent had both;
and 51 percent had neither. Pathogens from each sample that contained them were
then isolated and tested for antibiotic resistance. Isolates were maintained
at 5º C until testing.

Salmonella isolates were transferred
to trypticase soy agar (TSA) slants and analyzed manually using custom micro-dilution
Sensititre® panels (TREK Diagnostics), following the manufacturer’s instructions.
Various concentrations of seven antibiotics (ampicillin, ceftriaxone, ciprofloxacin,
gentamicin, nalidixic acid, tetracycline and trimethoprim-sulfamethoxazole)
were inoculated with a McFarland’s standardized bacterial suspension. Microtitre
plates were read following manufacturer’s instructions and isolates were classified
as susceptible, intermediate, or resistant based on available NCCLS standards
established for human medicine. A pathogen isolate was rated "susceptible"
if colony growth was inhibited by the lower tested doses of the antibiotic and
"resistant" if colonies grew even in the presence of the highest tested
doses. The "intermediate" classification means some colonies grew
at lower doses of the tested antibiotics but were inhibited by higher doses;
this result suggests that resistance is developing but higher doses of the tested
drug may still be effective against the pathogen.

Campylobacter isolates were transferred
to Brucella broth and maintained under micro-aerophilic conditions (10% CO2
+ 5% O2). Pure cultures were isolated using TSA (+5% sheep blood) plates, and
then Campy-Line agar. Isolates were analyzed using the E-Test® (AB BIODISK)
agar dilution method. Plastic strips impregnated with a concentration gradient
of one of six antibiotics (ciprofloxacin, clindamycin, erythro-mycin, gentamicin,
ofloxacin, and tetracycline) were applied (4/plate) to Mueller-Hinton agar (containing
5% sheep blood) plates that had been uniformly inoculated with a McFarland’s
standardized suspension of a campylobacter isolate. Isolates were classified
as susceptible, intermediate, or resistant, as above.

For both pathogens, Minimum Inhibitory
Concentrations were interpreted according to NCCLS Table 2A and the US National
Antimicrobial Resistance Monitoring System (NARMS) 2000 Annual Report.

IATP/SC bought 100 fresh whole chickens
and 100 packages of ground turkey from stores of two leading local grocery chains
in each of its two study cities, Des Moines, Iowa and Minneapolis, Minnesota,
during September and October of 2002. Samples in each city represented one leading
regional or national brand of chicken and one brand of turkey; i.e., only two
brands of each meat were tested. Samples were immediately placed in self-sealing,
coded plastic bags at the point of purchase, then refrigerated and shipped in
chilled, insulated containers by overnight carrier to a contract laboratory.

Each whole chicken sample was aseptically
transferred from the store package to a sterile stomacher bag and rinse sampled
with 400 ml of sterile Butterfield’s Phosphate Diluent (BPD), using USDA methods.(2)
25 grams of each ground turkey sample was aseptically transferred
from the store package to a sterile stomacher bag, and massaged with 275 g of
BPD. The rinsates from chicken and homogenates from turkey samples were analyzed
for salmonella, campylobacter and enterococcus.

All 200 chickens and 200 turkey samples
were tested for the presence of salmonella; to reduce costs, only every other
sample was tested for campylobacter and enterococcus. In all, 50 chickens from
each city (100 total), and 50 samples of Minneapolis turkey and 51 samples of
Des Moines turkey (101 total) were tested for these additional pathogens. No
attempt was made to determine individual strains of bacteria. The microbial
testing followed standard USDA, AOAC, CDC and NARMS methods.(3)
The USDA method for detection of campylobacter was slightly modified to enhance
recovery and isolation of this pathogen — the Hunt’s enrichment broth cultures
were streaked onto Campy-Line agar (CLA) rather than MCCD agar.

Bacterial isolates cultured from
positive samples were then tested for antimicrobial resistance (AMR). All 35
salmonella isolates and half (47) of the campylobacter isolates from chicken
were tested for AMR. Half (45) of the salmonella isolates and all (2) campylobacter
isolates from turkey were tested. Half of the enterococci isolates from both
chicken and turkey (101 isolates in all) were tested for AMR. Antibiotic resistance/
sensitivity testing was carried out basically according to CDC/NARMS procedures
(see CDC/NARMS publication cited in Footnote 3).

Salmonella and enterococci isolates
from positive poultry samples were assayed for resistance to multiple antibiotics
using the Sensititre® broth micro-dilution method (TREK Diagnostic Systems).
Salmonella isolates were transferred to TSA (Trypticase Soy agar) slants and
maintained at 5°C until tested for antibiotic resistance. The Sensititre®
method was conducted manually with custom micro-dilution (MIC) panels according
to the TREK Diagnostics instructions. McFarland suspensions were made in sterile
0.85% saline and the MIC custom panels were inoculated manually with a micropipetter.
The MIC panels were incubated at 35°C for 20-24 h and read visually according
to NCCLS M7 and TREK Diagnostics procedures. Campylobacter isolates from positive
poultry samples were transferred to Brucella broth and maintained at 5°C
under microaerophilic conditions (10% CO2 + 5% O2) until they were tested for
antibiotic resistance. Campylobacter isolates were tested for resistance to
multiple antibiotics using the E-test® agar dilution method. For AMR testing,
the Brucella broth cultures were initially streaked onto TSA (+ 5% sheep blood)
plates for isolation and scrutinized for purity. If mixed cultures were observed,
the culture was streaked onto Campy-Line agar (CLA) to isolate a pure culture.
Mueller-Hinton agar containing 5% sheep blood was used to make the lawns for
application of the E-strips. Test methods followed NCCLS standards where applicable.


Results of the tests for antimicrobial
resistance are summarized in Table 1 (for ACA and CU) and Table 2 (for IATP/SC).

In ACA’s tests, vancomycin-resistant
strains of Enterococcus faecalis (VRE) were isolated from 11 percent of the
chickens purchased in Brisbane and 14 percent of those purchased in Sydney (13
percent, overall).

In CU’s tests, combining the overall
prevalence of pathogens with the prevalence of resistance in tested isolates,
37 percent of the chicken samples contained one or both pathogens resistant
or intermediate in resistance to one or more antibiotics, and 29 percent of
the samples had at least one pathogen that was fully resistant to at least one

Campylobacter isolates from 48 of
203 CU positive samples for that pathogen (24 percent) had insufficient growth
on the Mueller-Hinton agar to permit accurate testing for antimicrobial resistance.
Isolates from the rest of the campylobacter-positive chickens (155 samples)
were tested for resistance. Of these, 66 percent were resistant to tetracycline;
20 percent were resistant and 45 percent intermediate in resistance to erythromycin;
26 percent were resistant to ciprofloxacin; 26 percent were resistant to ofloxacin;
21 percent were resistant and 8 percent intermediate in resistance to clindamycin;
and 8 percent were resistant to gentamicin. Overall, 90 percent were resistant
or intermediate in resistance to at least one antibiotic, 43 percent were resistant
to two or more antibiotics, and nearly 20 percent showed at least some resistance
to three different classes of antibiotics (clindamycin, tetracycline, and erythromycin).

Of 58 CU salmonella isolates, 19
percent were resistant to ampicillin, 17 percent were resistant to tetracycline,
5 percent were resistant to gentamicin, 3 percent to nalidixic acid, and 10
percent showed intermediate resistance to ceftriaxone. Overall, 34 percent were
resistant to at least one tested antibiotic, and 9 percent were resistant to
two or more tested antibiotics.

In the IATP/SC tests, overall, 17.5
percent of the whole chickens and 45 percent of ground turkey samples were contaminated
with salmonella; 95 percent of the chicken but only 2 percent of the turkey
samples contained campylobacter; and 100 percent of samples of both kinds of
poultry product contained enterococci.

IATP/SC AMR results (see Table 2)
found 6 percent of salmonella isolates from chicken each resistant to streptomycin,
sulfamethoxazole and tetracycline, and 3 percent each resistant to ampicillin,
amoxicillin-clavulinic acid, chloramphenicol and gentamicin. Salmonella isolates
from turkey samples showed much greater resistance, with 49 percent each resistant
to streptomycin and tetracycline, 42 percent to sulfamethoxazole, 36 percent
to gentamicin, and 16 percent to kanamycin.

Since there were just two campylobacter
isolates from turkey, results for this pathogen from chicken and turkey were
combined. Overall, 61 percent of isolates were resistant to tetracycline, and
8 percent each were resistant to ciprofloxacin and to nalidixic acid. Intermediate
resistance to three antibiotics was observed in campylobacter as well, with
22 percent of isolates showing some resistance to erythromycin, 14 percent to
clindamycin, and 12 percent to azithromycin. Similarly, combined results for
enterococci isolates from chicken and turkey found 96 percent overall were resistant
to quinupristin/dalfopristin, 81 percent to tetracycline, 20 percent to erythromycin,
11 percent to gentamicin, and 9 percent to streptomycin. In contrast to ACA’s
results from Australia, where VRE is a problem, none of the enterococci isolates
tested by IATP/SC were resistant to vancomycin.

Resistance to multiple antibiotics
was common in IATP/SC’s results. Only two of 35 salmonella isolates from chicken
showed any resistance, but one of those isolates was resistant to four drugs,
while the other was resistant to six. Among salmonella isolates from ground
turkey samples, 62 percent were resistant to one or more antibiotics, 49 percent
to three or more, 31 percent to four or more, and 7 percent were resistant to
five or more drugs. Among campylobacter isolates from chicken, 62 percent were
resistant to at least one drug, 6 percent to two or more drugs. Among enterococci
isolates from chicken, 98 percent were resistant to at least one drug, 74 percent
to at least two drugs, 38 percent to at least three drugs, 18 percent to at
least four drugs, and 2 percent to 5 or more drugs. Of enterococci isolates
from ground turkey, 100 percent were resistant to at least one drug, 94 percent
to two or more, and 18 percent to three drugs.

Overall, IATP/SC’s results suggest
that consumers of the ground turkey brands and chicken brands tested in the
cities where products were purchased are routinely exposed to at least one type
of bacteria that is resistant to at least one tested antibiotic.


These three studies by independent
consumer organizations provide evidence that public exposure to antibiotic-resistant
bacteria in poultry is widespread in both the United States and Australia. ACA’s
tests were narrow in scope, looking for a single bacterial strain resistant
to one major antibiotic, but they nonetheless found that an Australian consumer
living in one of the metropolitan areas sampled stands a 13 percent chance,
on average, of buying a chicken that harbors vancomycin-resistant Enterococcus
faecalis. While CU found some regional differences in the prevalence of pathogen
contamination, overall, CU’s data suggest that US consumers who buy fresh, whole
chickens have about a 1-in-3 chance of encountering a salmonella or a campylobacter
strain resistant to at least one major antibiotic. The IATP/SC testing was limited
geographically to just two US cities, and sampled just two brands each of chicken
and turkey. But by testing for enterococci as well as for salmonella and campylobacter,
and by testing for resistance to a much wider range of antibiotics than CU did,
IATP/SC demonstrated that exposure to antimicrobial-resistant bacteria is essentially
unavoidable for US consumers of poultry products. As both CU’s and IATP/SC’s
tests show, many American poultry consumers are routinely exposed to bacteria
resistant to multiple antibiotics.

The presence of antimicrobial-resistant
pathogens in chicken is one aspect of the much larger public health problem
associated with antibiotic-resistant bacteria. Fresh poultry is a relatively
commonplace vehicle for food borne exposure to pathogens. Campylobacter is a
leading cause of bacterial food poisoning in the US, often associated with poultry;
salmonella causes fewer cases but has a higher risk of mortality. Enterococci
are not food poisoning bacteria, and normally live in the human intestine with
no ill effects. However, enterococci can cause serious infections if they colonize
the urinary tract or get into the bloodstream during surgery, for example, or
infect immunocompromised individuals. Antibiotic-resistant strains of these
pathogens may not respond to initial treatment, which can be life-threatening
in vulnerable patients, and at minimum can greatly extend the length and costs
of medical treatment.

A recent statement from the US Centers
for Disease Control reports that campylobacter isolated from humans with food
poisoning is resistant to ciprofloxacin in 19 percent of cases, compared to
14 percent reported in 2000, and no detected resistance reported in 1990. CDC
officials have associated the increasing occurrence of ciprofloxacin-resistant
human campylobacter infections with the increasing prevalence of bacteria resistant
to this antibiotic in poultry.

Bacteria can also pass on resistance
traits to other bacteria that have not been exposed directly to antibiotics,
even across species. The widespread occurrence of campylobacter and enterococci
in poultry products and the prevalence in both bacteria of resistance to multiple
antibiotics suggest that consumption of poultry products could be a significant
medium for transmission of resistance from poultry-borne microbes to other bacteria
in the human digestive tract.

Few comparable studies of resistant
pathogens in chicken purchased at retail have been published, so it is not possible
to assess trends in this problem definitively. However, NARMS data, obtained
from chicken slaughterhouses and/or veterinary isolates, provide a basis for
some general comparisons.

Table 3 compares CU’s results (the
only national sample reported here) with NARMS data from the year 2000. CU’s
tests of salmonella found more resistance to ampicillin and nalidixic acid,
and less resistance to tetracycline and gentamicin, than NARMS has reported
for slaughterhouse isolates. CU’s campylobacter results and NARMS data show
a similar prevalence of resistance to tetracycline and erythromycin, but CU
found more resistance than the NARMS survey to ciprofloxacin, clindamycin, and

Overall, the picture seen from these
different data sets is similar, but CU’s results appear to suggest more resistance
to certain antibiotics in both salmonella and campylobacter on chicken at retail
than NARMS has found at other points in the food chain. Resistance to ciprofloxacin,
clindamycin and gentamicin is especially prevalent in campylobacter, a pathogen
known for its rapid development of drug resistance. Ten percent of CU’s salmonella
isolates also displayed intermediate resistance to ceftriaxone, suggesting emerging
resistance to this important cephalosporin drug. The NARMS survey found 0.2
percent of salmonella from veterinary isolates to be resistant, and 5 percent
intermediate in resistance, to ceftriaxone in its 2000 testing.

Perhaps most interesting among CU’s
results are the data on resistance to nalidixic acid in salmonella. Great caution
is needed in interpreting this result; just two out of 58 tested isolates (3
percent) were resistant. However, this rate is six times as high as that reported
by the NARMS survey in 2000, based on 1173 isolates from slaughterhouse chicken,
and 15 times as high as what NARMS reported in 1999, 0.2 percent of 1438 isolates.
Salmonella becomes resistant to the entire class of quinolone drugs, including
the important human drug ciprofloxacin, by acquiring two mutations. The first
mutation confers resistance to nalidixic acid, and the second, to fluoroquinolones.
If salmonella resistant to nalidixic acid are becoming more widespread in chicken,
as CU’s data hint may be occurring, the stage could be set for chicken to become
an additional reservoir of ciprofloxacin-resistant salmonella.

In CU’s campylobacter samples, widespread
resistance to tetracycline was notable and in line with previous data. Also
noteworthy was the comparatively high prevalence of resistance to ciprofloxacin
and ofloxacin, two members of the fluoroquinolone family of antibiotics. The
same isolates, 26 percent of the total, were resistant to both drugs.

IATP/SC also found ciprofloxacin
and nalidixic acid resistance in campylobacter, although in just 8 percent of
their tested isolates. In this respect (prevalence of specific resistance traits),
CU’s data, which are more representative of the national chicken market, deserve
more weight, but it is noteworthy that fluoroquinolone resistance was a clear
finding of both test regimes. One of the two chicken producers whose broilers
were tested by IATP/SC has said that it no longer uses fluoroquinolones in its

IATP/SC’s tests of enterococci found
a very high prevalence (96 percent overall, in 101 tested isolates) of resistance
to quinupristin/dalfopristin (Synercid). Like ciprofloxacin (Cipro), Synercid
is an important new antibiotic, used since 1999 in human medicine, and is likely
to be a drug of choice for treating infections resistant to other antibiotics.
The almost universal resistance to Synercid among poultry-borne enterococci
in this study is therefore a particularly significant public health concern.
Synercid resistance appears to be inherent to some strains of enterococci (e.g.,
E. faecalis), but can be acquired in others (E. faecium, for example). It seems
likely that both kinds of resistance are represented here. In studies where
enterococci on retail chickens have been speciated, E. faecium have been found
to constitute up to 60 percent of all enterococci.

The relationship between antibiotic
use in food animal production, and in particular the addition of sub-therapeutic
antibiotic doses to feeds, and the development of antibiotic resistance in bacteria
has been extensively debated. Many US poultry producers claim to have reduced
or eliminated use of fluoroquinolones in recent years, especially those used
in human medicine. However, Baytril, a member of the fluoroquinolone family,
is still being marketed for use in poultry flocks, and many experts believe
its continued use is prolonging or exacerbating resistance to important human
drugs like ciprofloxacin. The test results reported here suggest that whatever
steps have been taken have not prevented widespread resistance to these drugs
among poultry-borne campylobacter. This problem requires further investigation
and monitoring.

Enterococci are particularly adept
at acquiring and disseminating resistance genes. Studies in Europe have suggested
a link between vancomycin-resistant enterococci in animals and in humans, although
the vast majority of human VRE infections reported in Australia are not of the
type associated with food animals. VRE has been observed in chickens and pigs
fed avoparcin, which is similar to vancomycin. Avoparcin was banned in Europe
in 1997 and was withdrawn from sale worldwide in 1999; it is no longer used
in Australia. It has been largely replaced by virginiamycin, which has been
used widely as a sub-therapeutic growth promoter in poultry production since
1974. Virginiamycin is a close analog of quinupristin/dalfopristin (Synercid).
The virtually universal resistance to Synercid found in IATP/SC’s tests of enterococci
isolates from poultry suggests that possible links between virginiamycin use
in poultry and Synercid-resistant enterococci infections in humans also warrant
continued monitoring and further research.


Consumers International offers the
following recommendations to the Codex Committee on Residues of Veterinary Drugs
in Foods, to FAO, WHO, OIE and other appropriate international organizations,
and to member governments:

· Data on consumer exposure
to antimicrobial-resistant bacteria carried on everyday foods such as poultry
(among others) should be gathered in additional countries. Data from developing
countries would be especially valuable for defining the international dimensions
of the problem.

· Risk assessments carried
out for Codex committees, including CCFH, CCRVDF and TFAF, should incorporate
sensitivity to the additional risk issues raised by the presence of antimicrobial
resistance among pathogens assessed (e.g., salmonella or campylobacter in
poultry). CCRVDF should pursue discussions with JECFA, WHO and FAO to clarify
ways that AMR issues can be more effectively addressed in risk assessments
performed by JECFA for CCRVDF.

· Risk assessment and risk
management recommendations for fluoroquinones, tetracyclines, and other antimicrobials
of immediate or future importance to human medicine should be re-examined
by CCRVDF, JECFA, and other bodies as appropriate, taking into consideration
the risks to public health from antimicrobial resistance to these and related
compounds. Consideration should be given not only to the direct health risks
of acute infection following ingestion of food products carrying antimicrobial-resistant
pathogens, but also indirect risks stemming from colonization with resistant
strains potentially leading to future infections, with the onset of illness
or compromise in the immune system.

· In the absence of data
to the contrary, risk assessments should reflect the increasing weight of
scientific evidence indicating the potential for even non-pathogens carrying
genetic determinants of antimicrobial resistance to carry the seeds of this
resistance to the human population — including the possible transfer of these
genes to other bacterial species in the human gut, which may include pathogenic

· Further research is needed
on the role of poultry-borne bacteria as vectors in transmitting resistance,
on the development of cross-resistance, and on the rates of emergence and
spread of resistance to additional drugs in common pathogens. Research also
should further explore potential links between resistance to critical antibiotics
in human medicine and the use of those or similar drugs in poultry feeds.
In particular, the persistence of resistance to fluoroquinolones in campylobacter
on chicken in the US despite the claimed phase-out of feed uses of these drugs,
and a possible link between use of virginiamycin as a feed additive and resistance
to quinupristin/dalfopristin (Synercid) in poultry-borne enterococci, are
among the important questions research should pursue. CI notes that awaiting
the results of such research should not delay actions in the interim to help
reduce the threat of spreading antimicrobial resistance.

· CI supports the timely
completion by CCRVDF of the Proposed Draft Code of Practice to Minimize and
Contain Antimicrobial Resistance. Guidelines spelled out in the current draft
clearly advise against the use of subtherapeutic antibiotic doses in animal
feeds to promote growth, particularly for drugs important in human medicine.
CI agrees with that advice and urges CCRVDF to make the language on this point
as strong and unequivocal as possible.

· Recognizing the role of
other Codex bodies in addressing AMR issues, and especially that of the Ad
Hoc Intergovernmental Task Force on Good Animal Feeding, CI plans to present
a similar paper and set of recommendations to the TFAF. We have already presented
an informational paper to CCFH.

We note that the Codex Executive
Committee has discussed convening an expert consultation and establishing an
Ad Hoc Task Force to address antimicrobial resistance issues, which cut across
the terms of reference of several Codex committees, in an integrated manner.
CI supports that proposed process and looks forward to participating in the
work of the new task force, should the Commission choose to establish one.

Results of Testing by the Australian Consumers Association (ACA) and Consumers
Union of United States (CU)


Results of Tests by the Institute for Agriculture and Trade Policy and the
Sierra Club (IATP/SC)


Results of Testing by the Institute for Agriculture and Trade Policy and the
Sierra Club (IATP/SC)


National Antimicrobial Resistance Monitoring System – Enteric
Bacteria Veterinary Isolates Final Report, 2000.
Campylobacter data are from veterinary isolates; salmonella data
are from slaughterhouse isolates. Caution must be exercised when
comparing these data with results from retail chicken samples.
* NARMS campylobacter data are reported for two species, C. jejuni
and C. coli. CU isolates were not identified as to species; to
compensate for uncertainty as to which species is the better comparison,
data for both species are displayed here.