Introduction

Campylobacter

spp. are recognised worldwide as oneof the commonest causes of acute bacterial diarrhoeain man [1, 2]. In England and Wales during 1994,campylobacters accounted for about 3,600 cases ofgastrointestinal infection reported per month to thePHLS Communicable Disease Surveillance Centre[3]. The majority of such infections, which are widelyregarded as food-borne [4], are apparently sporadic,and the exact sources of the infections remain largelyundetermined. Campylobacters, in particular C.jejuni, pose a significant public health problem yetlittle is known about the prevalence and distributionof specific strain types in England [5, 6]. A latespring peak is a consistent and distinctive feature ofcampylobacter infection epidemiology in the UK [6]but there is little detailed information about the con-tribution made by different strain types of C. jejunito that seasonal peak.

The purpose of the present study was to survey thestrain subtypes of C. jejuni isolated from diarrhoealstools in three geographically separate laboratories inEngland over a twelve month period. Strains weretyped according to the Penner heat stable antigenscheme (HS serotyping) and selected strains subtypedby pulsed field gel electrophoretic (PFGE) analysisof DNA digest profiles. Associations between straintypes causing infections at different locations were

investigated, and related to the late spring peak. Thevalue of serotyping as a general method for thesurveillance of campylobacters was assessed.

Methods

Bacterial strains and growth conditions. The 398strains of C. jejuni examined in the study comprised:159 isolates from Royal Lancaster Infirmary obtainedover a 12-month period from January to December1993; 137 isolates from Ashford PHL obtained overthe same 12-month period; and 102 isolates fromHarrogate General Hospital obtained over the 12-month period from July 1993 to June 1994. Theisolates were from faeces and were all associatedwith human gastrointestinal infections. All isolatesreceived from the three regions were included apartfrom those known or suspected to be associated withforeign travel. All strains were grown at 37 °C for 48hours on 5% v/v defibrinated horse blood agar undermicroaerobic conditions (5% O2, 5% CO2, 2% H2,88% N2) in a variable atmosphere incubator (DonWhitely Scientific Ltd, Shipley, UK). Stocks ofstrains were preserved in 10% v/v glycerol inNutrient Broth No. 2 (Oxoid; CM67) over liquidnitrogen.

Identification and biotyping. The range of tests used

European Journal of Epidemiology 13: 837–840, 1997. 1997 Kluwer Academic Publishers. Printed in the Netherlands.

Subtypes of

Campylobacter jejuni from sporadic cases of diarrhoealdisease at different locations in England are highly diverse

R.J. Owen1, E. Slater1, D. Telford2, T. Donovan3 & M. Barnham4

1 Laboratory of Enteric Pathogens, Central Public Health Laboratory, London, UK; 2 Microbiology Department,Royal Lancaster Infirmary, Lancaster, UK; 3 Public Health Laboratory, The William Harvey Hospital, Ashford, Kent,UK; 4 Harrogate General Hospital, Harrogate, Yorkshire, UK

Accepted in revised form 6 June 1997

Abstract. Serotyping (heat stable antigens) wasperformed on 398 strains of Campylobacter jejunifrom faeces of human enteritis cases in England.Strains isolated over 12 months at three locationswere heterogeneous with 33 HS serotypes repre-sented. HS1 and HS4 complex were the predominanttypes (34% of all strains). The monthly strain fre-quency distributions were similar at the three loca-tions. The late spring peak appeared to be associatedwith a rise in miscellaneous serotypes rather thanwith the emergence of any characteristic predominant

Key words: Campylobacter jejuni, Pulse field gel electrophoretic (PFGE) profiling, Serotyping

serotype. PFGE DNA restriction profiling providedevidence of a high degree of genomic diversitywithin the common HS1 and HS4 complex serotypes,irrespective of the geographical source, yet somesubtypes were common to more than one location.The study showed that C. jejuni strain subtypes fromhuman enteric infections in England were highlydiverse, and that HS serotyping must be combinedwith a more discriminatory subtyping method suchas PFGE DNA profiling to provide an accurate basisfor epidemiological surveillance.

to identify and biotype the strains comprised theGram stain, growth at 42 °C, hippurate hydrolysis,rapid H2S production, indoxyl acetate hydrolysis andDNA hydrolysis. The tests were performed bystandard methods as previously described [7].Biotypes were assigned according to the Liorextended scheme [8].

Serotyping. Strains were serotyped with heat-stable(HS) antigens according to the passive haemaggluti-nation method of Penner & Hennessy [9] using avian(chicken) erythrocytes and a panel of 47 antisera forC. jejuni provided by Dr A. Lastovica (Red CrossWar Memorial Children’s Hospital, Cape Town,South Africa).

Pulsed field gel electrophoresis. The preparation ofbacterial DNA for PFGE and the separation of SmaIrestriction fragments using a CHEF-DR II system(BioRad Laboratories, USA) were as previouslydescribed [10]. DNA degradation in Lior biotype II(DNase positive) strains was prevented by formalde-hyde fixation of the bacterial cells prior to their incor-poration into agarose [10].

Results

Frequency of C. jejuni isolations. The number ofstrains of C. jejuni isolated per month from Januaryto December 1993 at Ashford and Lancaster areshown in Figure 1. Most strains were either Liorbiotype I or II. The main feature of the distributionswas their lack of uniformity with the characteristiclate-spring (May/June) peak of isolates in both lab-oratories. Isolates from Harrogate were obtained overan overlapping period from July 1993 to June 1994and the results also show the characteristic peak inJune, although it coincided with the end of thesurvey. A secondary peak was observed in January1994 (data not shown).

Frequency of serotypes over 12 months. Overall, 33

different HS serotypes were identified amongst the398 isolates of C. jejuni examined. For each labora-tory, the numbers of different serotypes in each strainset were as follows: 26 serotypes for both Ashfordand Lancaster, and 24 serotypes for Harrogate. Thedetailed serotype results for the three laboratories(Table 1), show that the five predominant serotypeswere the HS4 complex (18%), HS1 (16%), HS11(9%), HS6 complex (8%) and HS2 (6%). Those fivemain serotypes in total accounted for 56% of all theC. jejuni isolates examined. A detailed breakdown ofthe frequency of occurrence of strains of C. jejuniof the two numerically most predominant serotypes(HS4 and HS1) compared with the other serotypesover the 12-month period is shown in Figure 1. Theresults on the annual distribution of strain typesindicated that the two main serotypes occurredthroughout the 12 months in all three laboratories.The number of different serotypes identified eachmonth (Ashford and Lancaster) are presented inFigure 1. The Harrogate strains were isolated over aslightly different 12-month period (July 1993 to June1994) so were excluded from this figure.

838 EJEP ART NO. 1127 PIPS NO. 142896

Figure 1. Monthly distribution of the main HS serotypesof C. jejuni isolated at Ashford and Lancaster, during1993, and monthly variation in the numbers of differentHS serotypes of C. jejuni from Ashford and Lancaster (

d).

Table 1. Summary of the main serotypes of C. jejuni isolated over a 12-month period for three geographically separatelaboratories, Ashford (1993), Lancaster (1993) and Harrogate (July 1993 to June 1994)

Laboratory Number of isolates

HS4a HS1 HS11 HS6a HS2 Othersb Total

Lancaster 36 (23%) 31 (20%) 09 (6%) 09 (6%) 08 (5%) 066 (42%) 159Ashford 24 (18%) 26 (19%) 10 (7%) 08 (6%) 09 (7%) 060 (44%) 137Harrogate 12 (12%) 07 (7%) 15 (15%) 13 (13%) 08 (8%) 047 (46%) 102

Totals 72 (18%) 64 (16%) 34 (9%) 30 (8%) 25 (6%) 173 (43%) 398

a Associated serogroups. HS4 complex comprised: 4, 13, 16, 43, 50 and 64 serogroups; HS6 complex comprised: 6, 7,and 27 serogroups. b Includes non-typables.

Frequency of serotypes during seasonal peak. Theresults on isolation of C. jejuni at Lancaster andAshford showed a characteristic peak in numbersduring May and June 1993 (the late-spring peak),which accounted for 13% and 21%, respectively, ofthe total annual number of faecal isolates in thosetwo laboratories. No single HS serotype was uniquelyassociated with the late-spring peak and the increasednumbers over the period were attributed mainly to arise in occurrence of strains with other less commonserotypes (Figure 1). Similar results for serotypefrequency were obtained for the C. jejuni isolatesfrom Harrogate during the spring peak as well as asecondary peak in January 1994. The pooled data forLancaster and Ashford during 1993 was used tocompare the increases in HS4 complex and HS1against the ‘other’ serotypes. The variation over theyear in these pooled serotypes is shown in Figure 1.The number of isolates classed as other serotypesvaried more over the year than serotypes HS4complex and HS1. The differences were seenthroughout the year, but there is some suggestion thatthere may be a more pronounced late-spring peak inthe reports of isolates in the aggregated ‘other’ groupof serotypes.

PFGE DNA profiles of HS1 and HS4 complex strains.The HS1 and HS4 complex strains, which were thepredominant serotypes, were investigated in moredetail by PFGE analysis of DNA digests. The SmaIdigested DNA profiles comprised either nine frag-ments (HS1 strains) or eight fragments (HS4 complexstrains) ranging in size between 40–480 kbp. Therewere 36 different PFGE profiles amongst the HSlstrains and these showed considerable heterogeneityboth within and between the sets from differentlaboratories (Table 2). Profile types HS1 A and B (theletters indicate previously described designations[11]) were the commonest with five strains of eachidentified in Ashford and Lancaster, respectively.Likewise, the HS4 strains were diverse with 32 dif-ferent PFGE profiles. The commonest DNA profilewas HS4 type A (the letter refers to a previouslydescribed designation [11]), which was a feature ofl5 strains from the three laboratories.

Discussion

The recognition of campylobacter enteritis as adisease dates back almost 20 years [12] yet still onlylimited information is available on the distributionand frequency of different strain types in England.Our survey shows that the overall monthly variationin the number of C. jejuni isolated at three quiteseparate locations in England was similar to previ-ously reported data for the seasonal distribution ofCampylobacter spp. in England and Wales [5] andalso in the USA [13]. The results conform to theconsistent UK pattern for notifications of campy-lobacters of a peak in late May or early June followedby a steady decline until the end of the year [5].

A novel aspect of the present study was to monitorthe strain serotypes of C. jejuni arising throughout a12-month period in clinical laboratories at three dif-ferent locations in England that were geographicallywell separated. The study provided new informationon several important aspects of Campylobacter sppstrain incidence in England. The strains of C. jejuniassociated with enteric disease encountered in thesurvey were a heterogenous group overall withrespect to HS serotypes. Most serotypes occurredrelatively infrequently, whereas 56% of the strainswere accounted for by just five serotypes (or com-plexes of associated serotypes) namely HS1, 2, 4, 6and 11. The predominance of those serotypes, inparticular HS1 and HS4 complex, which togetherrepresented 34% of strains, has been observed pre-viously in the UK and also in many other industri-alized countries [13]. The constancy of the mainserotypes in the UK was borne out by a recent surveyof campylobacter bacteraemia in which strains of theHS4 complex were the predominant isolates fromblood (33%) and faeces (22%), although relativelymore frequent in the former [14]. Likewise, a two-year study of campylobacters in human faeces fromthe Reading area reported a wide distribution ofserotypes and that the commonest were HS1, HS2and HS4 (together they comprised 48% of the isolates[15]). The comparable figure of 40% was obtainedin the present study for those three serotypes.Significantly, we found that only about 8% of allisolates of C. jejuni were untypable, which contrastedwith a previously reported figure of l5% thoughbased on a slightly smaller panel of 43 HS antisera[l4]. There was no evidence from the present surveyof any significant differences between the threeregions either in the range of different serotypesisolated or in the predominant serotypes encountered.

Our analysis of individual HS serotype variationof C. jejuni on a monthly basis over a 12-monthperiod was of particular interest because it showed aconstancy in predominant serotypes with littlemarked variation either within or between differentlaboratory strain sets. Similarly, the late-spring peakwas associated with a general increase in miscella-

EJEP ART NO. 1127 PIPS NO. 142896 839

Table 2. PFGE DNA subtypes within the main serotypesof C. jejuni from three geographically separate laboratories

Lancaster Ashford Harrogate Total1993 1993 Jul ’93 to

Jun ’94

HS1 StrainsNo. strain 31 26 07 64No. types 22 18 07 36

HS4 StrainsNo. strains 36 24 12 72No. types 15 18 09 32

neous strain serotypes, rather than with a significantincrease in one or more of the usually predominantserotypes. The trend was the same at all three loca-tions in England and there was no indication that asingle serotype was the main contributing factor tothe annual spring peak at any single location.

The sample of strains of C. jejuni examined rep-resented just under 1% of the total number ofcampylobacter infection currently notified to CDSC,so it is not feasible to draw any general conclusionshere about variations in other regions of England.However, it appears reasonable to assume, given theconcordance with previously published results [11,14] that the predominant serotypes throughout theyear in most parts of England are currently probablyHS1, HS2 and the HS4 complex. Recent studies,based on high resolution molecular typing methods,including pulsed field gel electrophoresis of DNArestriction digests, have shown considerable genomicdiversity within individual HS serotypes of C. jejuni[16]. Similarly, a high level of diversity was foundamongst the HS1 and HS4 strains examined in thepresent study. Although no particular genomic typepredominated and there was no evidence for regionaldifferences, our study of the genomic types in thepredominant HS serotypes pin-pointed some strainsthat had conserved genotypes at different locationse.g. HS1/PFG-A and HS4/PFG-A. Such strains mayprovide useful markers for comparing populationsof C. jejuni in different hosts and for monitoring theirprogression through the food chain.

The main conclusions from this study are that C.jejuni is a diverse species and that HS serotypingprovides a valuable initial approach to typingisolates. There was general uniformity between thethree locations in the range and frequency of indi-vidual serotypes of C. jejuni occurring throughout theyear. For the predominant HS serotypes our resultsshow further discrimination can be obtained, ifnecessary, by the use of PFGE as a supplementarysubtyping method, should this be needed for sur-veillance purposes in the future.

Acknowledgments

The work was supported by a grant from theDepartment of Health, London. The PHL StatisticsUnit is thanked for advice.

References

01. Skirrow MB, Blaser MJ. Clinical and epidemiologicconsiderations. In: Nachamkin I, Blaser MJ, TompkinsLS (eds), Campylobacter jejuni: Current Status and

Future Trends. Washington, DC: ASM Press, l992:3–8.

02. Healing TD, Greenwood MH, Pearson AD. Campy-lobacter and enteritis. Rev Microbiol 1992; 3:159–167.

03. CDSC. Other gastrointestinal tract infections. CommDis Rep l995; 5: 6.

04. Cowden JM. Campylobacter: Epidemiological para-doxes. Br Med J 1992; 305: 132–133.

05. Pearson AD, Healing TD. The surveillance andcontrol of campylobacter infection. CDR Rev 1992;12: R133–139.

06. Advisory Committee on the Microbiological Safety ofFood. Interim report on Campylobacter. London:HMSO, 1992.

07. Barrow GI, Feltham RKA (eds). Cowan and Steel’smanual for the identification of medical bacteria.Cambridge: Cambridge University Press, 1993.

08. Lior H. New extended biotyping scheme forCampylobacter jejuni, Campylobacter coli and‘Campylobacter laridis’. J Clin Microbiol 1984; 20:636–640.

09. Penner JL, Hennessy JN. Passive haemagglutinationtechnique for serotyping Campylobacter fetus subsp.jejuni on the basis of soluble heat stable antigens. JClin Microbiol 1980; 12: 732–737.

10. Gibson JR, Sutherland K, Owen RJ. Inhibition ofDNase activity in PFGE analysis of DNA fromCampylobacter jejuni. Lett Appl Microbiol 1994; 19:357–358.

11. Owen RJ, et al. A molecular subtyping scheme forserotypes HS1 and HS4 of Campylobacter jejuni. JClin Microbiol 1995; 33: 872–877.

12. Skirrow MB. Campylobacter enteritis, a ‘new’disease. Br Med J 1977 (ii): 9–11.

13. Tauxe RV. Epidemiology of Campylobacter jejuniinfection in the United States and other industrialisednations. In: Nachamkin I, Blaser MJ, Tompkins LS(eds), Campylobacter jejuni: Current Status andFuture Trends. Washington, DC: American Society forMicrobiology, 1992; 9–19.

14. Skirrow MB, Jones DM, Sutcliffe E, Benjamin J.Campylobacter bacteraemia in England and Wales,1981–91. Epidemiol Infect 1993; 110: 567–573.

15. Fricker CR, Park RWA. A two-year study of thedistribution of thermophilic campylobacter in human,environmental and food samples from the Readingarea with particular reference to toxin production andheat-stable serotype. J Appl Bacteriol 1989; 66:477–490.

16. Gibson JR, Fitzgerald C, Owen RJ. Comparison ofPFGE, ribotyping and phage-typing in the epidemio-logical analysis of Campylobacter jejuni serotype HS2infections. Epidemiol Infect 1995; 115: 215–225.

Address for correspondence: Dr Robert Owen, Laboratoryof Enteric Pathogens, Central Public Health Laboratory,London NW9 5HT, UKPhone: +44 1 81 200 4400; Fax: +44 1 81 200 7874

840 EJEP ART NO. 1127 PIPS NO. 142896