Role of Poultry Meat in Sporadic Campylobacter Infections in Bosnia and Herzegovina: Laboratory-based Study

Abstract

Aim

To investigate genetic diversity and specificity of Campylobacter jejuni and Campylobacter coli strains isolated from humans, retail poultry meat, and live farm chickens in Zenica-Doboj Canton, Bosnia and Herzegovina, and identify the role of poultry meat in sporadic Campylobacter infections.

Methods

We determined the type of Campylobacter species using standard microbiological methods and multiplex polymerase chain reaction (PCR), and performed pulsed field gel-electrophoresis (PFGE) and restriction fragment length polymorphism (RFLP) typing of the flaA gene to investigate genetic diversity among the isolates.

Results

We isolated C jejuni and C coli from 75 (5.2%) of 1453 samples of consecutive outpatients with sporadic diarrhea; from 51 (34.7%) of 147 samples of poultry meat; and from 15 out of 23 farm chicken samples. The proportion of C coli found among human (30.1%), poultry meat (56.9%), and farm chicken isolates (53.3%), was greater than the proportion of C jejuni. Fourteen and 24 PFGE genotypes were identified among 20 C coli and 37 C jejuni isolates, respectively. Identical PFGE genotypes were found in two cases of human and poultry meat isolates and two cases of poultry meat and farm chicken isolates.

Conclusion

Only a minority of human Campylobacter isolates shared identical PFGE type with poultry meat isolates. Although poultry is the source of a certain number of human infections, there may be other more important sources. Further research is required to identify the environmental reservoir of Campylobacter spp responsible for causing human disease and the reason for the high prevalence of C coli human infections in this region.

Campylobacter jejuni and Campylobacter coli are still some of the most important enteropathogens worldwide (1-3). Understanding of their epidemiology is complicated by the sporadic nature of the disease, a lack of representative population sampling (4), wide distribution in the environment (5), and a high level of genetic diversity (6,7). The major route of infection in humans is through consumption of contaminated poultry meat, probably because of contamination of chicken carcasses with Campylobacter and the frequency of poultry consumption (3,8,9). Transmission to humans also occurs via other types of food, drinking water, and pets (8,10).

For studying the epidemiology of Campylobacter infections, several genetic typing methods have been developed in order to differentiate isolates below species level (7,9,11,12). Pulsed field gel electrophoresis (PFGE) typing analysis is a highly reproducible and discriminatory technique allowing a comparison between PFGE patterns in human isolates and isolates obtained from other sources (7,10,11,13).

Identical genotypes found in poultry products and humans might indicate common sources of infections and provide data on the genotype stability (11,14). Although most reports based on molecular typing have shown the role of poultry consumption in human Campylobacter infection, its epidemiology is still not completely defined (15).

An interesting epidemiological feature of human Campylobacter infections in Bosnia and Herzegovina, which has been noted since 1991, is the high prevalence of C coli in both asymptomatic carriers and diarrheic patients, comprising 36% and 26% of thermotolerant Campylobacter isolates, respectively, when the identification of species was based on the hippurate test (16).

The aim of this study was to determine genetic diversity and specificity of Campylobacter strains isolated during 2001 and 2002 according to the isolation source (human stool samples, retail poultry meat, and farm chickens) and species (C jejuni, C coli), and to identify the role of poultry meat in sporadic human Campylobacter infections.

Methods

Sampling patterns and isolation procedures

Human isolates. The Laboratory for Sanitary and Clinical Microbiology of the Canton Public Health Institution in Zenica has a total of 331 229 outpatients from Zenica-Doboj Canton, which consists of 149 053 inhabitants from urban areas and 182 176 inhabitants from rural areas. During 2001-2002, stool samples from 1453 consecutive outpatients with sporadic diarrhea were analyzed. The specimens originated from children younger than 6 years (n = 914), elementary school students (n = 188), high school students (n = 116), adults 20-64 years old (n = 177), and adults older than 65 years (n = 58). All stool samples were cultured on modified Preston medium (Oxoid, Basingstoke, United Kingdom) and incubated in a microaerophilic atmosphere (CampyGen, Oxoid) at 42°C for 48 hours.

Poultry meat isolates. A total of 147 samples of fresh or frozen retail poultry meat products (25 of the liver and 122 of the skin from the legs) from 53 markets in the Zenica-Doboj Canton were examined for the presence of C jejuni and C coli during 13 sampling visits between May and October 2002. Poultry meat samples came from 14 national and 7 international chicken meat producers. For the isolation of Campylobacter from poultry meat, the standardized procedure recommended by ISO 10272 was followed (17). Chicken liver or skin from legs (5 g) were enriched in 45 mL of selective Preston broth (Oxoid), containing 5% of horse blood (SR 048C, Oxoid) and incubated in a microaerophilic atmosphere for 18 hours at 42°C (CampyGen, Oxoid). One loopful of enrichment broth was streaked on modified Preston medium (Oxoid) and incubated in a microaerophilic atmosphere at 42°C for 3 days. C jejuni and C coli were identified using standard microbiological methods (18).

Farm chicken isolates. We also isolated C jejuni and C coli strains from cloacal samples of chicken from the biggest local farm. The isolation took place during 2001. Isolates were stored at -70°C in a medium consisting of nutrient broth No. 2 (CM0271 Oxoid) (32 g), agar (1.2 g), glycerol (150 mL), and distilled water (up to 1000 mL), supplemented with two vials of Campylobacter growth supplement (SR0232E, Oxoid).

DNA preparation and species identification

Isolated strains of C jejuni and C coli were sent to the Biotechnical Faculty of University Ljubljana, Slovenia, for further analysis. DNA was isolated from the colonies using two methods (9): a) a classical procedure of DNA extraction with guanidine thiocyanate (19) and b) a simple boiling procedure (12). C jejuni and C coli species identification was performed using polymerase chain reaction (PCR), based on amplification of flagellar genes (12), cytolethal distending toxin genes (20), hippuricase gene in C jejuni, and aspartokinase gene in C coli (21).

Fla-RFLP typing and PFGE typing

A flaA gene was amplified using the consensus pair of primers (7). A 1.7-kb amplicon was digested using CfoI restriction endonuclease (22).

The base for PFGE typing was the procedure for Salmonella strains (23,24). which was optimized for Campylobacter strains (22). For PFGE typing, the cells were cultured on Mueller Hinton agar and agarose blocks, prepared using INCERT agarose (FMC Bio Products, Rockland, ME, USA). After washing the blocks, the DNA was digested using SmaI (Boehringer, Mannheim, Germany) restriction enzyme. The fragments were separated by electrophoresis in CHEF Mapper XA System (Bio-Rad Laboratories, Hercules, CA, USA) for 22 hours at 200 V and 14°C (22).

Data analysis

Negative images of macrorestriction profiles were optimized and assimilated for numerical analysis by GelCompar II software (Applied Maths, Kortrijk, Belgium). Inter-gel variation was corrected by a reference pattern (Golden standard). The position of each band in the digested profiles was defined with a reference to the original photograph and to the densitometric curve of the scanned image. Optimal band alignment was performed automatically, using a maximum position tolerance of 1.1%. Interstrain relationships were evaluated by calculating pattern similarities with the Dice correlation coefficient and subsequent clustering by the unweighted pair group method, using arithmetic averages clustering algorithm (UPGMA) (24).

Results

Distribution of C jejuni and C coli in different sources is shown in Table 1.

Table 1.

The prevalence of Campylobacter jejuni and Campylobacter coli from different sources

Origin of samples	No. (%) of Campylobacter strains isolated
	Total	C jejuni	C coli	C jejuni/C coli coinfection
Human stool (n = 1453)	75 (5.2)	51 (68.0)	23 (30.1)	1 (1.3)
Retail poultry meat (n = 147)	51 (34.7)	21 (41.2)	29 (56.9)	1 (2.0)
Domestic (n = 66)	30 (45.5)	10 (33.3)	19 (63.3)	1 (3.3)
Imported (n = 81)	21 (25.9)	11 (52.4)	10 (47.6)
Farm chickens (n = 23)	15 (65.2)	7 (46.7)	8 (53.3)

During the study period, C jejuni and C coli were isolated from 75 (5.2%) of 1453 single outpatient stool samples: 53 isolates during 2002 and 22 isolates during 2001, with the incidence rates of 16.0 and 6.6 cases per 100 000/y, respectively. Most of the isolates (n = 47, 62.7%) were obtained from children younger than 6 years: 30 isolates during 2002 and 17 isolates during 2001, with the incidence rate of 88.7 and 50.3 per 100 000/y, respectively. Corresponding age-specific incidence rates for other age groups in both years were in the range of 0-14.5 per 100 000/y. Half of the isolates (22 out of 53 in 2002 and 10 out of 22 in 2001) were obtained from urban population. The incidence rates per 100 000/y for 2002 and 2001 in urban population were 19.5 and 6.7, respectively, and 13.2 and 6.6 in rural population, respectively. C coli was isolated from 23 (30.1%), and C jejuni from 51 (68.0%) of 75 stool samples, while both species were isolated from a single stool sample.

Poultry meat of domestic origin was obtained from 37 (69.8%) of 53 markets in this region. Imported poultry meat originated from markets in Croatia (n = 11), Germany (n = 11), Slovenia (n = 9), the Netherlands (n = 2), and Turkey (n = 2). Isolation rates of Campylobacter were 30/66 for domestic poultry, and 7/20, 7/24, and 5/20 for poultry imported from Croatia, Slovenia, and Germany, respectively. C jejuni and C coli were isolated from 51 out of 147 (34.7%) poultry meat samples, ie, 31 out of 106 (29.2%) frozen samples and 20 out of 41 (48.8%) fresh samples. C coli was isolated from 29 (56.9%) and C jejuni from 21 (41.2%) samples. In a single sample, both Campylobacter species were confirmed. C jejuni was isolated more frequently from frozen samples (76.2%), while C coli was isolated more frequently from fresh samples (54.2%).

Fifteen out of 23 farm chicken cloacal samples, obtained from the biggest local farm, were positive for Campylobacter. Eight out of 23 samples contained C coli and 7 out of 23 samples contained C jejuni.

Analysis of PFGE clusters according to the species and origin of strains

Of 131 C jejuni and C coli isolates, 75 originated from human stool samples (53 isolates from 2002 and 22 from 2001), 51 were isolated from retail poultry meat, and 15 originated from farm chickens. PFGE typing was performed for a convenience sample of 57 isolates uncontaminated after defrosting (11 and 8 human isolates, respectively; 26 retail poultry meat, and 12 farm chicken isolates), yielding 38 different PFGE types (Table 2).

Table 2.

FlaA-pulsed field gel-electrophoresis (PFGE) and restriction fragment length polymorphism (RFLP) fingerprints comparison of 57 Campylobacter jejuni and Campylobacter coli human, poultry meat, and farm chicken isolates

Isolate	Month/y of isolation	Source	No of samples from supplier (for poultry meat)*	Market place	Multiplex PCR	FlaA PCR	PFGE type	RFLP-FlaA type
9249/02	April 2002	human			C coli	+	1	7
11146/02	April 2002	human			C jejuni	+	2	3
134/P	October 2002	poultry meat	20 (Germany)	53	C coli	+	3	1
52/P	July2002	poultry meat	8 (BH)	21	C coli	+	4	1
56/P	July2002	poultry meat	23 (Turkey)	42	C coli	+	4	1
91/F	July2001	farm chicken			C coli	-	5
4442/02	February 002	human			C coli	+	6	8
5815/02	Maerch 2002	human			C coli	+	7	8
20/F	July 2001	farm chicken			C coli	+	8	5
15/F	July 2001	farm chicken			C coli	+	8	5
11272/02	April 2002	human			C coli	-	9
17/F	July 2001	farm			C coli	+	10	5
69/P	July 2002	poultry meat	17 (Slovenia)	31	C coli	+	11	5
59/P	July 2002	poultry meat	6 (BH)	42	C coli	+	12	2
60/P	July 2002	poultry meat	6 (BH)	42	C coli	+	12	2
21697/02	August 2002	human			C coli	+	12	2
114/P	September 2002	poultry meat	1(BH)	33	C coli	+	13	2
115/P	September 2002	poultry meat	1(BH)	33	C coli	+	13	2
141/P	October 2002	poultry liver	18 (BH)	16	C coli	+	13	2
16/F	July 2001	farm chicken			C coli	+	14	2
49/F	July 2001	farm chicken			C coli	+	14	4
26/F	July 2001	farm chicken			C coli	+	14	2
100/P	September 2002	poultry meat	17 (Slovenia)	36	C coli	+	14	2
21/F	July 2001	farm chicken			C coli	+	14	4
36/F	July 2001	farm chicken			C coli	+	14	2
28833/01	December 2001	human			C coli	+	15	5
28841/01	December 2001	human			C coli	+	15	5
21/P	June 2002	poultry meat	18 (BH)		C coli	-	16
70/P	July 2002	poultry meat	17 (Slovenia)	31	C coli	+	17
35/P	July 2002	poultry meat	18 (BH)	16	C coli	-	18
20579/ 01	August 2001	human			C coli	-	19
131/P	September 2002	poultry liver	18 (BH)	34	C coli	+	20	2
25657/02	October 2002	human			C coli	+	21	3
11/P	June 2001	poultry meat	2 (Croatia)	10	C coli	+	22	5
71/P	July 2002	Poultry meat	17 (Slovenia)	31	C coli	+	23	3
72/P	July 2002	poultry meat	17 (Slovenia)	31	C coli	+	24	1
27/F	July 2002	farm chicken			C coli	+	25	8
125/P	September 2002	poultry liver	1(BH)	49	C coli	+	25	8
25862/01	December 2001	human			C jejuni	+	26	5
28432/01	December 2001	human			C jejuni	+	26	5
96/P	July 2002	poultry meat	3 (BH)	44	C jejuni	+	26	5
28264/01	December 2001	human			C jejuni	+	27	7
28233/01	December 2001	human			C jejuni	+	28	1
28439/01	December 2001	human			C jejuni	+	28	1
21823/02	August 2002	human			C jejuni	+	29
23416/02	September 2002	human			C jejuni	+	30
124/P	September 2002	poultry meat	1 (BH)	49	C jejuni	-	31
121/P	September 2002	poultry liver	2 (Croatia)	14	C jejuni	+	32	6
130/P	September 2002	poultry meat	18 (BH)	34	C jejuni	+	32	6
19809/02	June 2002	human			C jejuni	-	33
31/F	June 2001	farm chicken			C jejuni	+	34	4
32/F	June 2001	farm chicken			C jejuni	+	34	4
25475/02	September 2002	human			C jejuni	+	35
116/P	September 2002	poultry meat	2 (Croatia)	20	C jejuni	+	36	1
128/P	September 2002	poultry meat	17 (Slovenia)	22	C jejuni	+	37	2
132/P	September 2002	poultry meat	18 (BH)	34	C jejuni	-	38
133/P	September 2002	poultry meat	18 (BH)	34	C jejuni	-	38

*Abbreviations: BH – Bosnia and Herzegovina; PCR –polymerase chain reaction.

PFGE patterns, consisting of 5 to 12 fragments, ranging in size from less than 50 kb to 580 kb, are shown in Figure 1. The clusters were delineated at similarity value of 90%. According to this, we identified five clusters (CL1-CL5) of selected isolates (Figure 1).

Figure 1.

SmaI pulsed field gel electrophoresis (PFGE) profiles of 57 human, poultry and farm chicken isolates of Campylobacter jejuni and Campylobacter coli.

Respecting 90% delineation level for a subcluster designation (PFGE type, present in at least two strains), we found 12 subclusters containing 31 (54.4%) of all isolates studied. These subclusters contained 2-6 isolates sharing identical or highly related PFGE types and originating from the same or different sources (Table 2). Seven subclusters were identified among C coli isolates, including 54.1% of typed C coli strains, and 5 subclusters were identified among C jejuni isolates, including 55.0% of the typed C jejuni strains.

Analysis of PFGE clusters for C coli

Within 2 subclusters of C coli containing PFGE type 12 and 13, (isolates 59/P and 60/P and 114/P and 115/P, respectively), identical strains were isolated from meat samples from the same market place and the same producer, indicating common origin of poultry meat isolates. However, in one subcluster, two C coli poultry meat isolates (52/P and 56/P) with identical PFGE type 4 originated from distinct geographical locations (from Bosnia and Herzegovina and Turkey). Similarly, C coli isolate 141/P could not be epidemiologically connected with the other two isolates of PFGE type 13 (Table 2), because it originated from a different market place.

The largest subcluster in our collection, consisting of PFGE type 14, contained C coli farm chicken isolates (Table 2), and included 5 out of 10 C coli strains isolated from the same farm. However, this subcluster also included one C coli isolate (100/P) from poultry meat imported from Slovenia. PFGE type 25 belonged to another subcluster consisting of a farm chicken and a poultry meat C coli isolate, but isolated in 2001 and 2002, respectively. A subcluster with PFGE type 8 included two identical C coli farm chicken isolates. Another two C coli farm chicken isolates (27/F, 91/F) showed significantly different PFGE types (PFGE type 25 and PFGE type 5).

Among other subclusters with different sources of isolates, there was one subcluster with both human and poultry meat isolates (PFGE type 12), which consisted of two C coli isolates from two different poultry meat samples originating from the same producer, sampled in the same market place, and of a C coli isolate from a human stool sample, all sampled in July/August 2002. One subcluster of C coli isolates, consisting of PFGE type 15, contained only human isolates (28833/01 and 28841/01).

Analysis of PFGE clusters for C jejuni

Only C jejuni poultry meat isolates from the subcluster that contained PFGE type 38 (132/P and 133/P) were of common origin. Additionally, one subcluster of C jejuni poultry meat isolates contained two isolates with identical PFGE type 32 (121/P and 130/P), but they were from distinct geographical locations (from Bosnia and Herzegovina and Croatia). Only one subcluster of C jejuni farm chicken isolates, with PFGE type 34, was detected, and contained two identical isolates. Among C jejuni subclusters with isolates from different sources, the PFGE type 26 contained two C jejuni human isolates from 2001 and one retail poultry meat isolate of domestic origin from 2002. One subcluster with PFGE type 28 contained only C jejuni human isolates (28233/01 and 28439/01), originating from the same patient (Table 2, Figure 1).

Analysis of PFGE clusters according to the origin of strains

Among 11 human C jejuni and 8 human C coli isolates from 2002 and 2001, we identified 11 and 5 PFGE types, respectively. Moreover, 10 (91%) of these 11 PFGE types from 2002 (except PFGE type 12: strain 21697) and 4 (80%) out of 5 PFGE types from 2001 (except PFGE type 26) were unique human genotypes. None of the 11 and 5 PFGE types of human isolates were identical in the two investigated periods of human infection. A large variation in genotypes was also demonstrated for retail poultry meat and farm chicken isolates. Among 26 poultry meat isolates, we identified 20 PFGE types, while among 12 farm chicken isolates we identified 6 PFGE types

RFLP analysis of flaA gene

Forty-eight of the 57 tested strains produced a flaA-PCR amplicon of the expected 1.7-kb size. The others remained untypable by this method, although all of them were confirmed as C coli or C jejuni by multiplex PCR (Table 2). Restriction analysis with CfoI did not yield a reliable result for additional 4 isolates. Finally, fla-RFLP typing distributed 44 strains into 8 different fla-RFLP types. Five fla-RFLP types (1-3,5, and 8,) were found in both, human and poultry meat isolates (Table 2). Only 4 animal isolates showed fla-RFLP type 4, 2 showed meat isolates type 6, and 2 showed human isolates type 7. Fla-RFLP types were also not species-specific – all types except 6 and 8 were found in isolates of both species.

Discussion

Our study demonstrated a significant proportion of C coli among human, poultry meat, and farm chicken isolates. Large heterogeneity of genotypes among C coli and C jejuni isolates were identified. Identical PFGE genotypes were found only in two cases of human and poultry meat isolates and two cases of poultry meat and farm chicken isolates.

PFGE patterns were remarkably diverse in this study. The genetic heterogeneity in the genotypes identified in many studies may in part reflect the diversity of sources, disease etiologies, and geographic origin of the Campylobacter strains, and indicate that sporadic cases of C jejuni and C coli enteritis are caused by a very diverse group of isolates (6,25). Predominant clonal groups of isolates have been associated with human infections in only a few reports (5,26). It is noteworthy, however, that 3 PFGE types from our study contained isolates of different origin, from geographically distinct locations, and were isolated over a two-year period. This suggests that, although the population of Campylobacter spp is highly diverse, it also contains strains that are a consistent feature of the epidemiology of this disease (5,27). Farm chickens exhibited colonization by multiple strains of C coli and a single strain of C jejuni. The presence of multiple genotypes on the same farm may be related to the multiple sources of infection, or persistent infection load to a genetic drift within the bacterial population (28).

In the period of 1999-2001, we recorded a surprisingly low number of isolates, and therefore, considerably lower Campylobacter infection incidence in this region (7.5, 1.8, and 6.6 per 100 000/y, respectively) (29), compared with the years 1998 and 2002 (18.3 and 16.0 per 100 000/y, respectively). This decrease could be associated with changes in the nutritional habits of the population during the 1992-1995 war: between 1990 and 1998, a 75%-90% decrease was recorded in livestock resources in Bosnia and Herzegovina, and consequently, far less milk, meat, and products derived from livestock were consumed during the war and in the postwar period. In addition, there was a substantial decrease in direct contact with farm animals. Some of the year-to-year variation in the incidence of food-borne diseases can be attributed to the prevalence of pathogens in their animal reservoirs and the foods derived from them (1).

Although C coli accounts for a minority of human Campylobacter infections, its health burden is greater than previously thought (3,30). The majority of C coli isolates analyzed in previous studies was obtained from pigs, suggesting that pigs were the most probable source of human infections with C coli, rather than chicken and cattle (31). This study, as well as previous studies in our region (4,16), showed a higher isolation rate of C coli in clinical material than in other countries (26,27,31). It is not likely that pigs were the source of human C coli infections, since post-war population of Zenica-Doboj Canton consist of 82.3% of Muslims, who do not consume pork meat because of religious reasons (32,33). This suggests that the primary source of C coli infections might be other than pigs. Some authors suggest that C coli chicken and porcine isolates represent host-specific populations (8,15,34), while others suggest that some C coli strains have a wider host range than others (35). One recent study has shown that the majority of C coli isolates from pigs were distinct from human isolates, and accordingly, that pork might be an infrequent source of human Campylobacter infections (35). Pigs, however, remain a potential source of C coli infection for persons who come into close contact with them (8). On the other hand, since some chicken C coli isolates have shown identical fingerprints with human C coli isolates, it has been proposed that not only chickens but other poultry sources such as turkeys, ostriches, and ducks may be a transmission vehicle of C coli to humans (35).

The question arises whether some clones are so common in the environment that they can infect farm chickens in different geographic areas (Bosnia and Herzegovina, Slovenia) by some, as yet unexplained, transmission route, such as through wild birds or by sharing a previously infected source of breeders for broiler production in these countries (36).

We are aware of the limitations of this study, such as the examination of only two possible sources of Campylobacter infections and the small number of human, poultry, and farm chicken isolates. Typing of a small number of isolates might be a possible reason why we did not observe more clusters containing the isolates of different origin. We demonstrated, however, that the interpretation of the data obtained by genotyping methods for epidemiological investigation of sporadic infections was still difficult (13). Epidemiological data about the food suspected to be involved in sporadic Campylobacter infections and the data about the contact with animals were available for only 28 patients, which was insufficient for a proper analysis.

PFGE showed that poultry meat may be the source of sporadic Campylobacter infections (2,11,22). However, PFGE- and RFLP-typing analysis of human and poultry meat isolates in our study, where only a minority of clinical and poultry meat isolates of Campylobacter shared identical PFGE types, showed that the sources other than poultry might be important.

All subclusters were unified according to the species, but not according to the isolation source or geographical affiliation, and therefore, the potential sources of the majority of clinical Campylobacter enteric infections were difficult to identify. High prevalence of C coli isolates from humans, poultry meat, and farm chickens suggests, however, that there may be a common source in the environment, which might be absent in other geographical regions. We have not yet been able to explain the high prevalence of C coli human infections in this region.

Further prospective research is required to examine potential reservoirs of Campylobacter species in the environment. A formal surveillance system supported by the characterization and typing of larger numbers of isolates from different sources may help to identify epidemiological relationships.