Multi-locus sequence types of Campylobacter carried by flies and slugs acquired from local ruminant faeces


Emma L. Sproston, The Macaulay Land Use Research Institute, Craigiebuckler, Aberdeen, AB15 8QH, UK.


Aims:  To assess whether flies and slugs acquire strains of Campylobacter jejuni and Campylobacter coli present in local ruminant faeces.

Methods and Results: Campylobacter was cultured from flies, slugs and ruminant faeces that were collected from a single farm in Scotland over a 19-week period. The isolates were typed using multi-locus sequence typing (MLST) and compared with isolates from cattle and sheep faeces. Campylobacter jejuni and Camp. coli were isolated from 5·8% (n = 155, average of 75 flies per pool) and 13·3% (n = 15, average of 8·5 slugs per pool) of pooled fly and slug samples, respectively. The most common sequence type (ST) in flies was Camp. coli ST-962 (approx. 40%) regardless of the prevalence in local cattle (2·3%) or sheep (25·0%) faeces. Two positive slug pools generated the same ST that has not been reported elsewhere.

Conclusions:  Despite their low carriage rate, flies are able to acquire Campylobacter STs that are locally present, although the subset carried may be biased when compared to local source. Slugs were shown to carry a previously unreported Campylobacter ST.

Significance and Impact of the Study:  This study has demonstrated that flies carry viable Campylobacter and may contribute to the transfer of STs within and between groups of animals on farms. Further, they may therefore present a risk to human health via their contact with ready-to-eat foods or surfaces.


Campylobacter is the major cause of identified bacterial gastrointestinal infections worldwide with an incidence of 5–6 per 100 000 in the USA (Blaser 1997) and 88 per 100 000 in the UK (ECDC 2007). Almost all human Campylobacter reports are classified as sporadic (Pebody et al. 1997; Frost et al. 2002) with only 0·2% being recognized as part of an outbreak (Frost et al. 2002). Most (approx. 93%) human infections are because of Camp. jejuni with the majority of the remainder attributed to Camp. coli (Gillespie et al. 2002). Case–control studies indicate that human infection is frequently associated with the consumption and handling of poultry (Pebody et al. 1997; Rodrigues et al. 2001; Frost et al. 2002; Evans et al. 2003; ECDC 2007). Other reported food vehicles for transmission are milk and dairy products (Rodrigues et al. 2001), salad (Pebody et al. 1997; Rodrigues et al. 2001; Frost et al. 2002; Evans et al. 2003) and water (Pebody et al. 1997; Evans et al. 2003).

Farm ruminants have been recognized as asymptomatic carriers of Campylobacter, where it is shed transiently in their faeces (Jones 2001; Stanley and Jones 2003). Campylobacter prevalence in cattle varies between studies, but generally ranges from 23 to 56% (Nielsen 2002; Brown et al. 2004) with Camp. jejuni generally predominating (55·4 and 100%) (Nielsen 2002; Brown et al. 2004). Prevalence in sheep is approx. 30% (Stanley et al. 1998; Jones et al. 1999; Jones 2001) with Camp. coli being more common (33–39%) than in cattle (Brown et al. 2004; Acik and Cetinkaya 2006). The local contamination of fields from grazing animals offers the opportunity for flies and slugs to acquire Campylobacter. Flies have the potential to transfer Campylobacter to humans via contact with ready-to-eat foods and surfaces and slugs by contact with salad crops or fruit.

Flies classed as ‘filth flies’ require a faecal/decomposing substrate necessary to complete their life cycle (Olsen 1998) and can be further termed as ‘disease causing flies’ when they exhibit certain behavioural traits. Potential ‘disease causing flies’ need to be synanthrophic (associated with food, carrion and waste), endophilic (attracted to mans environment) and will oscillate between filth and food. In addition, they need to have previously been associated with carriage of enteric pathogens (Salmonella (Khalil et al. 1994; Barro et al. 2006), Streptococci (Barro et al. 2006), Shigella spp. (Echeverria et al. 1983; Khalil et al. 1994; Barro et al. 2006), enterotoxigenic Escherichia coli (Echeverria et al. 1983; Khalil et al. 1994), Escherichia coli O157:H7 (Rahn et al.1997; Alam and Zurek 2004; Szalanski et al. 2004), Vibrio spp. (Echeverria et al. 1983) Campylobacter spp. (Khalil et al. 1994; Szalanski et al. 2004) and Camp. jejuni (Rosef and Kapperud 1983; Wright 1983)). These traits (classed as filth flies, ‘disease causing flies’ and associated with enteric pathogens) apply to three families; Muscidae, Calliphoridae and Sarcophagidae (Olsen 1998).

Research investigating flies as carriers of Campylobacter has shown large differences within and between studies. For example, Camp. jejuni was isolated from 43·2% of flies collected from a piggery and 50·7% from a chicken farm, but no flies were positive at either a cattle or turkey farm (Rosef and Kapperud 1983). Other reports show that Camp. jejuni was isolated from 8·2% of flies sampled from areas surrounding broiler houses (Hald et al. 2004), 13·4% from a turkey farm (Szalanski et al. 2004), 8·9% from a cattle farm (Adhikari et al. 2004) and 2·4% from suburban areas (Wright 1983).

Slugs ingest soil bacteria (Walker et al. 1999), are widespread pests in agriculture and frequently contaminate leafy crops (Port and Ester 2002). Escherichia coli O157 has previously been isolated from slugs and the strain was matched to that in local sheep faeces using multi-locus variable number tandem repeats analysis (Sproston et al. 2006). In addition, commensal E. coli was present in high concentrations in slugs (Sproston et al. 2006). No reports have described the potential for slugs to act as a vector for Campylobacter and reports on other invertebrate vectors are limited. The darkling beetle (Alphitobus diaperinus) and their larvae (the lesser mealworm) have been shown to carry Camp. jejuni at a prevalence of 35% in broiler houses (Jacobs-Reitsma et al. 1995) and has been proposed to be responsible for the dissemination of this pathogen among broiler flocks (Strother et al. 2005).

Multi-locus sequence typing (MLST) is a sequence-based typing method increasingly used for typing Campylobacter (Dingle et al. 2002; Colles et al. 2003; Manning et al. 2003). The seven MLST housekeeping genes show sufficient genetic variation to enable discrimination of Campylobacter sequence types (STs) isolated from cattle faeces, wildlife, water and soils where particular STs have shown host association (French et al. 2005; McCarthy et al. 2007). MLST will, therefore, help in ascertaining whether Campylobacter carried by flies and slugs are likely to have been acquired from local sources (e.g. cattle or sheep), as well as indicate any host association with these vectors.

The aim of this study was firstly to assess the prevalence of Campylobacter in pooled samples of flies and slugs. Secondly to determine whether flies carry the same STs as those present in local ruminant faeces, thus demonstrate their potential to act as vectors between animals and of human infection.

Materials and methods

Collection of field samples

Flies and slugs were collected from a mixed arable and livestock farm in NE Scotland containing approx. 200 sheep and 48 cattle, between 16th May 2006 and 19th September 2006 (described as weeks of the year, 20–38 incl.). Fly samples were collected using a fine mesh (1·4 × 1·4 mm) sweep-net, by sweeping above the foliage for approx. 10 min. Areas sampled for flies were from fields containing sheep, cattle or both (Table 1). Each collection was transferred to a sterile plastic container and the net sterilized by thorough spraying with 70% ethanol and dried in air before reuse. An average of 10·3 (±3·96 SD) pooled fly samples were collected each week over 15 weeks (no flies were collected during weeks 23, 24, 28 and 38) totalling 155 pooled fly samples, each containing an approx. average of 75 (6–304) flies. Slug samples were collected depending on their availability. In total, 15 pooled samples were obtained with a mean of 8·5 (5–12) slugs per pool. Samples were chilled (∼4°C) in a cool box with icepacks and immediately processed on return to the laboratory. In the areas of fly and slug collection, individual faecal samples from cattle and sheep were collected from the ground. In total, 338 cattle and 214 sheep faecal samples were also collected during 18 of 19-week sampling period. The cattle herd and sheep flock studied did not remain static on the farm. Sometimes the cattle herd moved within the farm or was split between two different fields. The flock of sheep was also frequently moved to different fields both internally and externally. During some of the sampling times, cattle and sheep were mixed on the same field. Fly samples were therefore collected from areas containing livestock. Table 1 shows the location of livestock, the number of cattle/sheep present and the number of faecal and fly samples taken from that particular area on a weekly basis.

Table 1.   The weekly locations of livestock during the 19-weeks sampling (weeks 20–38) and the number of fly pools collected from each field
WeekField 1NField 2NField 3NField 4NField 5NField 6*NField 7†N
  1. S (x/n), number of faecal samples/number of sheep present; C (x/n), number of faecal samples/number of cattle present; N, number of fly samples collected from that field.

  2. *off site approx. 1·0 km away.

  3. †off site approx. 1·5 km away.

20C (12/33)0C (0/15) S (10/200)6          
21C (8/33)4C (0/15) S (11/200)5          
22C (2/33)0C (2/15) S (9/200)6          
23C (53/33)0C (8/15) S (20/200)0          
24No Sampling
25C (12/33)5C (2/15) S (11/200)5          
26C (13/48)5           S (11/200) 5
27C (12/48)6  S (9/200)5        
28C (10/33)0C (0/15) S (14/200)0          
29C (10/33)10C (0/15) S (11/200)12          
30 3C (1/15) S (10/200)5C (10/33)2        
31C (7/30) S (9/200)7C (8/18)2          
32C (10/30)5C (5/18)3      S (10/200)2  
33C (15/48)5        S (10/200)5  
34  C (15/48)5      S (10/200)6  
35C (15/48)5        S (10/200)5  
36C (15/48)5        S (10/200)1  
37C (18/48)10    S (14/200)5      
38C (75/48)0      S (25/200)0    

Fly identification

At each fly sampling date, one collection was reserved for identification. Here, flies were killed with chloroform, viewed using a hand lens (×10 magnification) and placed into 6-ml plastic-labelled containers. Flies of similar appearance were allocated together. A representative selection was taken from each group and identified under a dissecting microscope (×40 magnification). For example, a small sample would be taken from a group that contained individuals that initially appeared identical. In comparison, a larger sample would be selected for identification from a group that contained nonidentical flies but were grouped together because of common matching features. If flies within a single group belonged to more than one family, an additional portion would be selected for further identification. The representative sample size was also dependent on group size. The proportions of families within a single group were presumed to be representative.

Isolation of Campylobacter

Containers of flies were placed at −20°C for 1 min to immobilize and then quickly transferred to a sterile plastic bag 9 × 11·5 cm. Approx. numbers were recorded prior to being manually crushed in 5-ml nutrient broth (DM180D; Mast, Bootle, UK) with 5% defibrinated horse blood (DHB100; E & O Laboratories, Bonnybridge, UK), growth supplement (SRO232E; Oxoid, UK), amphotericin (2 μg ml−1), trimethoprim (10 μg ml−1), cefoperazone (15 μg ml−1), polymixin B (2–5 IU l−1) and rifampicin (5 μg ml−1). The fly mix was then transferred to sterile 6-ml plastic containers. Slugs were aseptically cut into 2-mm pieces and diluted ninefold with enrichment broth plus supplement. All samples were incubated under microaerobic conditions at 37°C and 100 μl volumes plated after 2 and 5 days onto charcoal cefoperozone deoxycholate agar (CCDA; CM0739; Oxoid) with selective supplement (SR155E, Oxoid). Presumptive Campylobacter colonies were confirmed using latex agglutination (M46CE; Microgen, Camberley, UK).

Often the number of Campylobacter colonies present from a positive pooled sample was low; the number of colonies that were picked, cultured and stored (−80°C in nutrient broth plus 15% glycerol) ranged from 1–5 colonies from each positive sample.

Ten grams of faecal material from either cattle or sheep was diluted ninefold in the enrichment broth and further serially diluted in PBS and plated directly onto supplemented CCDA. The plates and the broth were incubated for 2 days after which presumptive Campylobacter colonies were recorded from direct plates. To determine the presence/absence when concentrations were below the direct plating threshold, 100 μl of enrichment was spread onto CCDA and incubated as stated previously.

DNA extraction

Campylobacter isolates were re-plated from glycerol storage medium onto CCDA and grown as described previously for isolation. DNA was extracted from colony growth by lysis in the presence of Chelex 100 resin (142–1253; Bio-Rad Laboratories, CA, USA) as described previously (Gormley et al. 2008).

Molecular typing

Isolates were initially screened for likely identity as Camp. jejuni or Camp. coli using the primers pgmF1 and pgmR1 (Miller et al. 2005), which are reported to be diagnostic for these species with distinction from Campylobacter lari, Campylobacter upsaliensis and Campylobacter helveticus. Isolates yielding pgm amplicons of the expected size (720 bp) were further typed by MLST using seven loci (aspA, glnA, gltA, glyA, pgm, tkt, uncA) as described previously (Gormley et al. 2008). Sequencing traces were aligned and assigned to allele numbers using stars software ( (accessed August 2009)) and the 7-locus allelic profile was assigned a ST and grouped into a clonal complex (CC) when alleles at 5 or more loci were shared, using the public Campylobacter MLST profile database ( (accessed August 2009)).

Statistical analysis

Values of prevalence with binominal 95% confidence intervals were calculated in Excel. A bootstrap method using Poptools ( (accessed August 2009)) was used to determine 95 percentiles from counts obtained from cattle and sheep datasets. This involves sampling (10 000 times) with replacement of the original count data and determining the average counts as well as the percentiles.

Fisher’s exact test of differentiation was used to statistically compare either the proportions of Camp. coli or a particular ST between flies, cattle and sheep. Odds ratios were calculated and P-values recorded (SISA-Binomial (accessed August 2009)). The exact test of differentiation using the Markov Chain analysis, which is analogous to the Fisher’s test but extends the 2 × 2 contingency table into a multiple contingency table (Raymond and Rousset 1995; Goudet et al. 1996) was used to analyse whether the range of STs isolated from cattle, flies and sheep showed significant differences. The number of iterations for this test was 100 000 and performed on Arlequin software (Excoffier et al. 2005).

The binomial distribution test with perfect sensitivity (Vose 2000) was used to estimate the prevalence value required to be >95% confident of a particular ST being identified.

A single consensus tree to summarize the clonal relationship between Campylobacter STs isolated from cattle, sheep, flies and slugs (Fig. 1) was constructed using a model-based approach. For each ST identified, the DNA sequence for each of the 7 loci was analysed using the ClonalFrame software (Didelot and Falush 2007) ( (accessed January 2010).

Figure 1.

 The clonal relationship of Campylobacter STs isolated from cattle (□), sheep (▪), flies from cattle areas (⋄), flies from sheep areas (♦) and from slugs (+).



Overall 21% (95% CI, 17·5–26·3) of cattle were positive for Campylobacter (74/338 faecal samples) with an average of 6·0 × 102 CFU g−1 (95% CI, 277–984) shed in their faeces (inc. nonshedders). Sheep had a significantly lower prevalence than cattle at 14% (95% CI, 9·4–18·7) (30/214 faecal samples) (P < 0·01). The concentrations shed by sheep at 8·2 × 102 CFU g−1 (95% CI, 271–1455) were not significantly different to cattle (P > 0·05).

A total of 21 different STs were isolated from cattle and nine different STs were isolated from sheep (Fig. 2). The frequency of STs identified in cattle was significantly different to those found in sheep (< 0·001) Camp. jejuni were the majority of Campylobacter isolated (97·7% (95% CI, 94·6–100·9)) from cattle with the remainder being Camp. coli. Campylobacter coli contributed to a significantly higher proportion in sheep (59·4% (95% CI, 42·4–76·4)) than in cattle (< 0·00001).

Figure 2.

 The relative proportion (RP) of sequence types (ST) isolated from (a) flies collected from areas containing cattle, (b) flies collected from areas containing sheep, (c) cattle faeces and (d) sheep faeces.


Campylobacter prevalence in flies was low with 9/155 pooled samples testing positive (5·8% (95% CI, 2·1–9·5)). Campylobacter were not isolated from flies collected in fields containing mixed livestock (0/47), therefore all positive samples were from fields containing either cattle (4/73) or sheep (5/35). The number of positive fly samples showed no significant difference in Campylobacter prevalence whether collected in areas containing cattle or sheep.

Two different Campylobacter STs were isolated from a single fly pool collected from cattle areas (totalling five isolates from four positive pools). Two pooled samples of flies collected from areas containing sheep yielded two different STs (totalling seven isolates from five positive pools). In total, five different STs of Camp. jejuni were identified from flies and one Camp. coli (ST 962). The proportions of different STs isolated from flies collected from areas containing cattle (a) or sheep (b) are shown in Fig. 2. Flies collected in areas containing cattle or sheep showed similar proportions of Camp. coli (ST-962) at 40·0% (95% CI, 2·9–82·9) and 42·9% (95% CI, 6·2–79·5), respectively. Flies collected in areas containing cattle or sheep had no significant differences in the prevalence of Camp. coli/ST-962 (Table 2). Despite the flies (collected from cattle areas) showing significant higher proportions of Camp. coli/ST-962 than cattle, there was no overall significant difference in the frequencies of the STs identified from flies when compared to those from cattle faeces (< 0·05) (Tables 2 and 3).

Table 2.   The statistical significance of the Fishers test for analysis of parameters recorded between cattle, flies and sheep
Parameter testedOdds ratioSignificance (P)
Flies (C) vs Flies (sh)
 Campylobacter coli/ST-9620·89ns
 Campylobacter prevalence0·35ns
Cattle vs Sheep
 ST 9620·07< 0·001
 Camp. coli0·02< 0·001
 Campylobacter Prevalence1·72< 0·01
Flies (Sh) vs Sheep
 ST 9622·25ns
 Camp. coli0·51ns
Flies (C) vs Cattle
 ST 96228·33< 0·05
 Camp. coli28·33< 0·05
Table 3.   Comparisons of the frequencies of STs between cattle, flies and sheep using the exact test of differentiation
 Cattle STsSheep STsFlies (C)All ruminants
  1. Ns, not significant.

Sheep STs< 0·001   
Flies (C)nsns  
Flies (Sh)< 0·001< 0·05ns 
All flies< 0·05

All STs isolated from flies (excepting ST-3217) collected in fields holding cattle were also isolated from cattle faeces during the sampling period. Therefore, flies collected from cattle areas did show some similarities to source. It is possible that ST-3217 was absent in cattle (ST-3217 would only need to have been present at prevalence 1·1% for one quarter of the sampling time so as to be 95% confident of its detection by using the binomial distribution test (Vose 2000)).

The frequencies of STs isolated from flies collected in areas holding sheep were significantly different to those isolated from sheep faeces (< 0·05) (contrary to what was found in cattle areas), (Table 3). This is regardless of no significant difference in the prevalence of Camp. coli or ST-962 between flies and sheep. The lack of similarity between fly and sheep STs is because of two STs (ST-52 and ST-206) present in flies failing to be identified from sheep. The absence of these two STs in sheep may be because of sampling limitations. STs would need to be at a prevalence of >5% for half of the sampling time so as to be 95% confident in their detection, using the binominal distribution test with perfect sensitivity (Vose 2000).

When all flies were grouped together (irrespective of collection area) and all faeces were grouped together (regardless of origin), the frequencies of STs obtained from flies were significantly different to STs isolated from all local ruminant faeces (< 0·05) (Table 3).

Figure 1 shows the genetic relationships of all STs identified from cattle, sheep, flies and slugs. Because of the low number of positive fly samples, the STs isolated from flies collected in both cattle and sheep areas show no obvious distribution within the consensus tree. The STs isolated from cattle show a continuous array of genetically related STs. The STs identified from sheep appear to show more distantly related Campylobacter STs with a distinct group dominated by the Camp. coli CC 828 at the base of the consensus tree.

The majority (77%) of flies were identified as ‘filth flies’ and belonged to 12 of the 16 families listed in Table 4. However, only three of these families (Muscidae, Calliphoridae and Sarcophagidae) could be classified as ‘disease causing flies’ (Olsen 1998) which here contributed to an estimated 5%. Approx. 40% were minute (≤2 mm) or too damaged to be identified to family and were assigned to the subgroup small Acalyptratae. Approx., 5% were unidentified (Table 4).

Table 4.   Estimations of the proportions of fly families collected during the study and their association with faeces and/or decaying matter
FamilyApprox. %Associated faeces/decaying matter
  1. *Fly families containing species classed as ‘disease causing flies’ (i.e. preference towards human settlements) (Olsen 1998).

  2. †Larvae feed and develop on ruminant flesh.

Small Acalypterates43·0Yes
Not done5·3


The typical slug species collected were Derocerous reticulatum, Arion subfuscus, Arion circumscriptus, Limax marginatus, Arion ater, Limax tenellus and Limax maximus with the former two being the most common. From 15 pooled slug samples (average of 8·5 per pool), two tested positive for Campylobacter (13·3% (95% CI, 30·54–3·87)) each with a unique and identical allele number combination assigned ST-3218. There was no matching ST in ruminant faeces collected in this study. The two positive samples were collected 7 weeks apart (week 20 and 27), the first from areas containing sheep and cattle, and the second from a field holding cattle alone and the collection sites were approx. 100 m apart.


The advantage of using a single sampling location over a relatively long period of time (19 weeks) has provided an overview of the STs locally present confirming that flies are able to carry locally acquired Campylobacter STs.

If prevalence at source was a function of Campylobacter carriage by flies, it may be expected that flies would, to some degree, reflect the highly significant differences shown to exist between cattle and sheep (prevalence, species composition, ST frequencies). Firstly, it may be expected that flies collected from areas containing cattle would have higher carriage rates than those collected from areas holding sheep, (because of the significant difference in Campylobacter prevalence between cattle and sheep) (21·9 and 14%, respectively). However, the number of positive fly samples collected from areas of cattle and sheep showed no significant difference at 4/73 and 5/35, respectively (Table 2). Indeed, when expressed as a percentage (5·48 and 14·29 %), sheep (with the lower prevalence) appears to have a higher percentage of positive flies collected from holding areas. Secondly, it may be presumed that flies sampled from sheep areas would carry a higher load of Camp. coli and ST-962 than those collected in cattle areas, where the prevalence of ST-962 in cattle and sheep was 2·3 and 25%, respectively. Despite the significant difference in the prevalence of Camp. coli/ST-962 between cattle and sheep, flies carried a similar proportion of Camp. coli/ST-962 (40·0 and 42·9% respectively), regardless of the livestock present (Table 2). Thirdly, some of the most common STs isolated from sheep (ST-825) and cattle (ST-48, ST-42) were not isolated from flies. One plausible suggestion is that Campylobacter may exhibit some degree of host association; following the transfer of Campylobacter to flies from a faecal source, particular STs may be better adapted to survive, such as possessing greater resistance strategies. However, it is important to emphasize that because of the low number of positive fly samples in this work, interpretation is limited and cannot be verified at this stage.

Some similarities between ST in flies and source were observed; the flies collected from areas containing cattle failed to show significant differences in the frequencies of STs isolated, despite the differences evident in the proportions of Camp. coli or ST-962. This suggests a degree of similarity between source and carriage by flies and agrees with the work by Adhikari et al. (2004) and Hald et al. (2004) who both found a strong clonal association of isolates from flies to those of local sources using pulsed-field gel electrophoresis. The differences between the frequencies of STs isolated from flies and sheep faeces was likely to be because of the two apparently absent STs (ST-52 and ST-206) in sheep. These STs may have been unidentified in sheep because of sampling limitations or alternatively, ST-52 and ST-206 may have originated from cattle, present in an adjacent field during the time of sampling. Only a single ST isolated from flies collected from cattle (ST-3217) was not identified from either cattle or sheep here. This ST was likely to have been absent in the ruminant groups sampled. ST-3217 may have originated from a different environmental source as Campylobacter has commonly been reported to be carried and excreted by other animals including, rabbits, rodents’, birds, domestic pets and other wildlife (Cabrita et al. 1992; Adhikari et al. 2003; Colles et al. 2003), or possibly from other nearby cattle; ST-3217 has been recorded on a single occasion from cattle faeces in Aberdeenshire during 2004 ( (accessed August 2009)) from a farm that was approx. 10 miles north-east of the present site.

Here, it was attempted to relate ST prevalence found in flies to that of local faeces. However, the concentration of Campylobacter at source is likely to be a stronger influencing factor of fly carriage. Flies need to acquire, via a small area of contact, sufficient Campylobacter for detection, but any Campylobacter transferred to the fly will show a rapid decline. For example, at 25°C, 93% of flies directly inoculated with 107 CFU ml−1 were negative by enrichment after 24 h (Hald et al. 2007). The rate of decline has been shown to be significantly associated with time, temperature and dosage (Hald et al. 2007). A fly is likely to visit several faecal sources, if all are Campylobacter positive, but at a low concentration, it is plausible that the fly will fail to acquire a high enough load for detection. Alternatively, if faeces of a high concentration is present among many negative samples, a detectable quantity of Campylobacter may be transferred to the fly on visits to this particular faeces. It is not possible, in this study, to fully assess concentration as a determining factor, as individual faeces with a high Campylobacter concentration present at the time of collection of positive flies may not have been sampled. However, the overall low prevalence of Campylobacter in flies and the lack of significant difference in the number of positive fly samples between those collected in areas holding cattle or sheep may be accounted for by the lack of significant differences in the overall low concentration of Campylobacter shed in the faeces of cattle and sheep, 6·0 × 102 CFU g−1 and 8·2 × 102 CFU g-1, respectively. This would also explain the reason for the higher Campylobacter prevalence reported for flies collected from poultry farms compared to this study (Hald et al. 2004) where poultry typically excrete 4·2 × 105 CFU g−1 (Bull et al. 2006).

The identical STs from the two positive slug samples were collected from two different locations, 7 weeks apart and failed to be identified elsewhere in this study. From the two sources, cattle are more likely to have been the source because of being present at the time of both positive samples. However, cattle would only need to have a prevalence of <0·1% so as to be >95% confident of its detection using the binomial distribution test assuming perfect sensitivity (Vose 2000), therefore ST-3218 was likely to have been absent in ruminants and acquired from another environmental source. The closest match to the slug isolate is Camp. jejuni ST-954. This was identified from a wild bird in Cheshire, UK during 2002 ( (accessed January 2010)). Five of the seven alleles were identical between ST-3218 and ST-954, the differences being in the glnA and tkt allele, which had 2-bp and 4-bp differences, respectively. It is feasible that wild bird faeces may have been the source of the Campylobacter isolate from the slugs. Using the presumption that it was a single slug within each of the two positive pools, this would give the lowest estimate of individual field prevalence of 1·6%. This is almost eight times greater than the percentage found to be carrying E.  coli O157:H7 (Sproston et al. 2006). Slug populations of up to 300 slugs per m2 have been reported for arable fields (Wilson et al. 1994). Using the individual prevalence found in this study (1·6%), the density of slugs positive for Campylobacter could therefore be up to 5 per m−2. The slug ST has not yet been reported from human cases thus further work is required to ascertain whether slugs have the potential to carry other STs.

The low number of positive fly samples limits the evaluation into the pattern of ST carriage by flies in relation to local sources. Therefore, the significant differences identified in the statistical analysis should be cautiously interpreted. However, flies did appear to carry a relatively high prevalence of Camp. coli/ST-962 despite the low prevalence in cattle. Significant differences were observed between the frequencies of STs in flies collected areas holding sheep and those identified in sheep faeces and between all the flies (regardless of collection site) when compared to all ruminant faecal samples. However, all the STs isolated from flies in this study were present in local ruminant faeces (except ST-3217). Further research is required to verify and quantify the relative importance of host specificity, concentration and ST variation. Extended studies need to statistically assess the potential risk of Camp. jejuni and Camp. coli transfer by flies and slugs between different farm animals and to humans via vegetable crops, ready-to eat foods and surfaces.

All of the STs isolated from flies (except ST-3217) have also been isolated Scottish clinical isolates (Food Standards Agency-Scotland 2009). Although this study has been limited by the low number of positive samples, this study emphasizes the potential of flies to act as vectors of locally acquired Campylobacter STs.


We thank The Biotechnology and Biological Sciences Research Council for funding this PhD Studentship, the Foods Standard Agency, Scotland for funding the typing of the Campylobacter isolates. We also thank Professor Hugh Pennington for assistance in fly identification.