A Three-Year Study of Campylobacter jejuni Genotypes in Humans with Domestically Acquired Infections and in Chicken Samples from the Helsinki Area

Abstract

Campylobacter jejuni isolates from stool samples of patients with domestically acquired sporadic infections and from chicken from retail shops were studied during seasonal peaks from June to September over a 3-year period from 1996 to 1998. A large number of pulsed-field gel electrophoresis (PFGE) genotypes (a combined SmaI-SacII pattern) were identified each year. Certain genotypes persisted for the whole study period, and predominant genotypes represented 28 to 52% of the strains during a restricted period of time. The peak level of positive chicken samples was between July and August of each study year, when 10 to 33% of the samples were positive for campylobacter. The same PFGE genotypes found in humans were also detected in the chicken samples. This suggests that common genotypes were circulating in the area.

Campylobacter jejuni is the most common bacterial enteropathogen in developed countries. The sources and transmission routes of human campylobacteriosis are not fully understood, but handling and eating poultry have been shown to be important risk factors (10, 12, 14, 15). Several molecular typing methods have been used to support studies of the epidemiology of campylobacter infections during the 1990s, and pulsed-field gel electrophoresis (PFGE) pattern analysis has been shown to be a highly discriminatory method (3, 5). We have also recently used PFGE typing for studies of the epidemiology of campylobacter infections in humans in Finland (6). Little is known about the persistence or fluctuation of the genotypes of C. jejuni through the years. Identical genotypes might indicate common sources of infection and provide data on the stability of the genotypes. This paper describes PFGE genotypes of C. jejuni isolates gathered over a three year period from fecal samples of enteritis patients from the Helsinki, Finland, area with infections acquired in Finland during the summer months. In addition, PFGE patterns of C. jejuni strains isolated from chicken sold at the retail level or from chicken fecal samples obtained at slaughterhouses during the same study period were included for comparison.

Bacterial isolates.

C. jejuni isolates from fecal samples of enteritis patients were collected from June through September during each year of a 3-year period from 1996 to 1998. The isolates were regarded as domestic in origin if the patients had not travelled abroad prior to their illness. The fecal samples were inoculated on campylobacter blood-free selective medium (charcoal-cefoperazone-desoxycholate agar; Oxoid, Ltd., Basingstoke, Hampshire, England) and incubated at 42°C in a microaerobic atmosphere. Chicken samples were cultured after enrichment in Lab M enrichment broth (Bury, United Kingdom) on charcoal-cefoperazone-desoxycholate agar (6, 7) or according to the method outlined by the International Standardization Organization (8). Fresh chicken pieces (legs and breasts) sold at retail that were studied represented three major chicken producers in Finland and were collected from shops in the Helsinki area. For the year 1996, strains isolated from chicken cecal samples taken during the slaughtering process were also included for PFGE genotype analysis. Only C. jejuni isolates were included in the study (6). After the original isolation, the isolates were stored at −70°C.

Typing C. jejuni isolates by PFGE.

For PFGE analysis, the isolates were grown on brucella blood agar (Oxoid) for 2 days at 37°C in a microaerobic atmosphere. The bacterial cells were harvested and treated with formaldehyde to inactivate endogeneous nucleases (4). Otherwise, DNA plugs were prepared, and electrophoresis conditions were as described earlier (6, 7, 11). A combined SmaI-SacII pattern was designated a genotype. Certain strains were also analyzed by digestion with KpnI. If the strains had one- to three-band differences in their SmaI, SacII, and KpnI patterns, they were designated subtypes and marked with a lowercase letter (for example, VIa, VIb, and VIc).

There was a remarkable fluctuation in the number of domestically acquired campylobacter cases in humans during the study. The summer months of June through September were chosen for the follow-up period, as there is an isolation peak from July to August in Finland, and outside this season, domestic campylobacter isolates are uncommon (6). In 1996, 89 domestic campylobacter cases were identified, whereas the respective number in 1997 was only 36 and in 1998, 69. The reason for this fluctuation is not known, but the summer in 1996 was hot and dry, whereas the summers of 1996 and 1998 were especially rainy. During the same time period, campylobacter was also isolated from chicken pieces sold at the retail level (Table 1). The isolation rates were low in the beginning of the summer (0 to 10% in May; unpublished results), increased to 10 to 30% from July to August, and dropped again in September (8 to 14%). In comparison with other countries, the isolation rate of campylobacter in chicken pieces was low. In a recent study from the United Kingdom, approximately 80% of chicken pieces sampled from retail shops were positive for campylobacter (F. J. Bolton, J. K. Williamson, G. Allen, D. R. Wareing, and J. A. Frost, Abstr. 10th Intl. Workshop on Campylobacter, Helicobacter, and Rel. Org., abstr. CF2, 1999). In a Danish study, 30 to 40% of chicken pieces sampled in one summer were positive (H. Rosenquist and N. L. Nielsen, Abstr. 10th Intl. Workshop on Campylobacter, Helicobacter, and Rel. Org., abstr. CF20, 1999). Similar seasonal variation as noted in infections in humans has also been verified in the carriage rates of C. jejuni in live chickens (2). This finding was confirmed in our study as well. However, the annual variation in the incidence of campylobacter cases in humans could not be explained only by the variable contamination level of retail chicken pieces, because approximately the same percentage of chicken pieces were found to be positive for campylobacter each year. Most chickens in Finland are sold in fresh cut-up pieces, which means that the time between slaughtering and consumption is short. Although the peaks in infections in humans and contamination rates of chicken meat are both seen from July to August, it does not necessarily indicate that chicken would be the most important direct source of infections in humans in Finland. Recent case-control studies from other countries actually showed that eating or handling chicken meat could explain less than 10% of the human campylobacter cases (13, 14).

TABLE 1.

C. jejuni strains isolated from samples of chicken meat collected in the Helsinki area, 1996 to 1998

Month of collection	Characteristics of C. jejuni strains collected in:
	1996		1997		1998
	No. positive/ no. studieda	PFGE genotypesb	No. positive/ no. studied	PFGE genotype(s)	No. positive/ no. studied	PFGE genotypes
June	0/15		5/49	I/137	0/25
July	2/20	I/E, V, VIa, VIb	8/55	I/K, VIb, II/A	21/82	T101b, T101a, T131, T131b
August	6/20	XI/XIc, I/B, V, I/E, -/P, II/A, VIa, VIb	10/30	IV, I/K, I/B, II/A	22/80	I/K, VIb, I/137, VIc, T131a, T131b, I/B
September	2/25	I/E, V, -/P	6/72	T101b	8/56	I/K, -/P, VIc

^aThe number of chicken meat samples positive for C. jejuni versus the number of samples tested. 

^bSmaI-SacII PFGE genotypes found in human strains of C. jejuni are presented; also included are genotypes identified from slaughterhouse chicken cecal samples (1996 only). 

Along with the two restriction enzymes, SmaI and SacII, 30 PFGE genotypes were found among the human isolates in 1996, 19 genotypes were found in 1997, and 28 were found in 1998. The overall diversity of genotypes was similar each year. Seven genotypes (including variants) (I/E, I/J, I/K, II/A, IV, VIa, VIb, VIc, and -/P) were identical for all three seasons (1996 to 1998), and genotypes I/B, I/L, and H were identified at two isolation periods (Tables 2 and 3). A total of 41 genotypes were unique and typical for a certain time period only. PFGE patterns of five persistent genotypes (I/B, I/K, IV, VI, and T101) and variants of genotypes VI (a, b, and c) and T101 (a and b) are shown in Fig. 1. A large variation in genotypes was also demonstrated for the chicken isolates, as 15 genotypes were identified in 1996, 19 genotypes were found in 1997, and 27 were found in 1998 (Table 1). The same genotypes seen in human strains (for example, I/B, I/K, IV, VIa, VIb, VIc, T101a, T101b, and -/P) were identified among chicken strains as well (Table 1; Fig. 1).

TABLE 2.

Common C. jejuni PFGE genotypes (SmaI-SacII) isolated from patients in the Helsinki area, 1996 to 1998

Month of collection	Characteristics of genotypes collected in:
	1996		1997		1998
	No.a	Major PFGE genotypesb	No.	Major PFGE genotypes	No.	Major PFGE genotypes
June	10	XI/S, I/B, I/J, I/J, I/L, VIb	2		11	T101b, I/K, II/A, IV
July	45	I/B (7), VIa (2), VIb (6), IVc, XI/R (3), IV (2), V(3), I/E, II/A	13	T101a (3), I/K, I/L, VIc, -/P, I/J	22	T101b (3), I/K (4), VIb (2), VIc (3), I/J, IV
August	23	VII (4), V (5), I/B (3), VIII (2), I/104, II/A	16	IV (4), VIa, VIb, VIc, I/K, I/L, I/E, II/A	24	I/K (6), VIa (4), -/P (4), VIa (2), I/B, I/J, I/E
September	11	IV, I/E, I/B, -/P	5	T101a (3), I/E	12	T101b (5), I/B

^aNumber of C. jejuni-positive patients. 

^bValues in parentheses indicate the number of patients, marked only if more than one patient was sampled. 

TABLE 3.

Distribution of five predominant domestic human C. jejuni genotypes from June to September of 1996, 1997, and 1998

Genotype	No. of isolates (% of the total no. of isolates) collected in:
	1996c	1997d	1998e
I/K	1	2	12 (17)
T101aa	0	6 (17)	0
T101ba	1	0	12 (17)
VIab	2	1	5 (7.2)
VIbb	11 (12)	1	4 (5.7)
VIcb	3	1	3 (4.3)
IV	3	4 (11)	2
I/B	11 (12)	0	4 (5.7)

^aSubtypes of genotype T101. 

^bSubtypes of genotype VI. 

^cA total of 89 isolates were tested. 

^dA total of 36 isolates were tested. 

^eA total of 69 isolates were tested. 

FIG. 1.

SmaI, SacII, and KpnI patterns of certain common Finnish C. jejuni genotypes. M, molecular size marker.

Information on the PFGE patterns over a longer time span is limited. In the present study, two predominant genotypes (VIb and I/B in 1996 and IV and T101a in 1997) represented 25 and 28% of the human strains isolated in 1996 and 1997, and three genotypes (I/K, T101b, and the VI variants VIa, VIb, and VIc) represented 52% of the strains isolated in 1998 (Table 3). Predominant PFGE genotypes varied by year and also from one month to another during the same year (Table 2 and 3). For instance, PFGE genotypes I/B and VIb were the most common genotypes among the C. jejuni strains isolated in 1996, but no I/B isolates and only one VIb isolate were found in 1997. However, other variants of genotype VI (VIa and VIc), were common among the isolates each year (Table 3). Similarly, PFGE genotype T101a was only found among the strains isolated in 1997, when it represented over 16% of all isolates, whereas another highly related genotype (T101b) represented over 17% of the isolates in 1998. Predominant genotypes persisted through the 3-year follow-up, although their relative proportion in patients markedly varied from one year to another. This phenomenon suggests genomic stability.

Most of the predominant human PFGE genotypes were found among chicken isolates (Table 3), although not always among strains isolated during the same month. However, this does not prove that the direct source for the human infections was chicken but could indicate that during summer months campylobacters are common in many sources and that certain predominant genotypes circulate and colonize various host animals. Humans are probably infected from many sources, direct contact with chicken being only one of them. Some chicken PFGE genotypes were never found among human isolates, and vice versa, which also suggests that there are other important sources than chicken for human campylobacteriosis. Chickens are raised in large units of approximately 10,000 animals and slaughtered at the age of 6 weeks. Typing of C. jejuni strains from several subsequent breeding units has shown that each unit is usually contaminated with a different genotype of C. jejuni (1, 2, 9). This explains the high number of unique genotypes identified from chicken samples.

In many countries, as in Finland, the number of campylobacter cases in humans has increased significantly in recent years (12, 14). In Finland in 1998, the annual number of reported campylobacter cases (2,938) exceeded that of salmonella (2,735) for the first time, according to the National Infectious Disease Registry, National Public Health Institute, Helsinki, Finland). Except for large outbreaks, however, the source of human campylobacteriosis is rarely identified, and therefore the exact transmission routes are poorly understood. Precise molecular typing methods are necessary epidemiological tools, and PFGE has been shown to be a discriminatory method with the possibility of digitizing the pattern analysis (http://www.svs.dk/campynet/). Our present survey showed that there was a great fluctuation not only in the number of domestically acquired campylobacter cases in humans but also in the PFGE genotypes identified during the 3-year follow-up period. Some PFGE genotypes were found throughout the study period, although the predominant genotypes usually differed each year. At the same time, the prevalence of campylobacter among chicken meat samples at the retail level was quite low. Although most of the predominant PFGE genotypes among human isolates were the same as those seen among chicken isolates during the same time period, other sources for Campylobacter infections in humans should also be considered.