The number of flies transported by means of ventilation air into a broiler house in Denmark was counted from July 22, 2003, to July 28, 2003, and the Campylobacter carriage rate of flies captured in the environment of the broiler house was estimated. The study period was chosen because flies generally peak in activity and abundance in July to August in Denmark. In addition, chickens in the broiler house and the animals in the area around the broiler house (5 sheep, 4 horses, and 1 dog with 10 puppies) were tested for carriage of Campylobacter. DNA from all C. jejuni isolates was analyzed by pulsed-field gel electrophoresis (PFGE) to determine if strains from different animals were similar.

The broiler house (80 m x 15 m) was located at Universal Transverse Mercator Grid zone 32, East 564,137 m, North 6,294,759 m (http://mac.usgs.gov/mac/isb/pubs/factsheets/fs07701.html). The facility was negative-pressure ventilated through a total of 84 wall valves for air intake (16.5 cm x 52.5 cm) and 12 round chimneys (diameter 62 cm) for active air outlet through the roof. The house was emptied, and the chickens (n = 28,235) were slaughtered on July 29, 2003. Reports of local weather data from the Danish Meteorological Institute (www.dmi.dk) for that week were a maximum day temperature of 25.4°C, a minimum night temperature of 11.9°C, and days with bright sunshine with no wind or rain.

Flies were collected in polyester nets equipped at two wall inlets (one net at each end of the house) in the dynamic air flow measured (7) at a pressure of 21 Pa of influx ventilation air (speed 3.6 m/s, volume 1,213 m³/h per inlet valve). After the nets were harvested, flies were visually sorted from other insects and counted. The flies, identified primarily to the order Diptera and the families Muscidae and Calliphoridae, were counted; the count showed that 917 ± standard deviation (SD) (843.5–990.5) flies (a + b)/2 x 84, with a and b representing the number of flies in the two nets) had entered the broiler house per day through the ventilation system, or approximately 1 fly per 2,700 m³ of ventilation air (1,213 m³/h x 84 x 24 h/917). For this specific broiler house, this amount equals approximately 30,000 flies per broiler cycle in the summer season. To estimate the possibility of roof inlets as entrance route for flies into houses with an air inlet in the roof, the fan of one chimney was switched off. This step generated inlet air in this chimney. The result showed 167 flies per day entered through this one chimney, or an average of 45 flies per 2,700 m³ air. The flies captured in the ventilation system were used for counting only, since they were dead and dried out at the time of harvest.

To determine the prevalence of Campylobacter in flies around the broiler house, a total of 96 flies were captured singly within a distance of 50 m from the house by rackets equipped with small disposable plastic bags. Captured flies were narcotized with CO₂. After species or genus was determined, each fly was transferred with a pair of tweezers to live storage in a sterile plastic tube with 1 mL of saline. The tubes were kept in an insulated container, which was transported to the laboratory within 24 hours after capture.

For Campylobacter detection, each fly was macerated in a sterile mortar, suspended in 2 mL of 0.9% saline. The mixture was centrifuged at 15,870 x g for 7 min. The pellet was resuspended in 2 mL of Bolton broth (CM0983 with SR0183 and SR0048) (Oxoid, Basingstoke, UK) and vortexed before incubation for enrichment at 37°C for 24 h. After enrichment, the tube was again centrifuged, and 100 μL of the sample from each of 49 tubes was streaked onto modified cefoperazone charcoal deoxycholate agar (mCCDA) (blood-free agar base supplemented with CM739 + SR155) (Oxoid). These plates were incubated at 42°C for 48 h in a microaerobic atmosphere (6% O₂, 6% CO₂, 4% H₂ in N₂). The tubes from the remaining 47 samples were subjected to DNA analysis by a nested polymerase chain reaction (PCR) (8).

Figure 1

Figure 1. Campylobacter-infected flies captured around a broiler house in Denmark. Flies are grouped according to species. Campylobacter-positive and -negative by A, Campylobacter bacteriologic culture, and B, nested polymerase chain reaction.

The results of conventional culture showed that, of 49 flies tested, C. jejuni was isolated from 4 (8.2%) (Figure 1A); when a nested PCR was used, 33 (70.2%) of the 47 flies were Campylobacter-positive (Figure 1B). The species distribution according to the nested PCR when species-specific primers were used showed that, of 47 samples, 56.4% were positive with C. jejuni primers, 18.0% were positive with C. coli primers, and 25.6% were positive with Campylobacter species primers. Co-infection with C. jejuni and C. coli or Campylobacter spp. was found in six flies. However, PCR also detects nonculturable and dead Campylobacter. The reason for dividing the flies into two equal portions, one for conventional culture and the other for PCR, was to avoid reducing the assumed low number of Campylobacter on each fly.

Cloacal or rectal swab samples from 20 broilers, 5 sheep, and 4 horses were cultured on mCCDA agar, as described above. Rectal swabs from 11 dogs were streaked onto cefoperazone amphoricin teicoplanin (CAT) agar plates (blood-free agar base with CM739 and SR174) (Oxoid), and incubated at 42°C for 96 h. Subsequently, the swabs were tested for Campylobacter by our laboratory’s routine PCR for feces (9). C. jejuni was isolated from 20 broilers and 4 sheep and C. upsaliensis from 11 dogs. One sheep and the four horses were culture-negative. However, all fecal swab samples were Campylobacter-positive by the routine PCR.

Figure 2

Figure 2. Dendrogram of pulsed-field gel electrophoresis types with SmaI of Campylobacter jejuni isolates from chickens in a Danish broiler house and from animals outside (flies and sheep).

A total of 28 C. jejuni isolates from 20 broilers, 4 flies, and 4 sheep were fingerprinted with PFGE with two different restriction enzymes, SmaI and KpnI. With both enzymes, 27 of the isolates had an identical PFGE pattern (S6), whereas a single broiler isolate had a slightly different, but closely related pattern (S7), which probably was derived from the more prevalent pattern. The SmaI patterns are shown in Figure 2. Twenty-seven of 28 isolates from three animal sources, broilers, sheep, and flies, and from both inside and outside the broiler house, belonged to the same clone.

Under experimental conditions (10), flies are able to transmit Campylobacter among chickens. Moreover, a high prevalence of Campylobacter-infected flies captured in a broiler house has been found (11). However, no study has yet been able to demonstrate a significant role of flies captured in the houses for transmitting infection from flock to flock (5). Our results suggest that the potential of flies to transmit infection depends upon a current supply to the broiler house of Campylobacter-infected flies from the outside. Furthermore, the number of flies entering the broiler house must increase as the need for ventilation air increases as a consequence of the growth of chickens. Thus, the risk of introducing Campylobacter to the house increases with the age of the chickens.

This study has demonstrated that flies pose a threat of Campylobacter infection, from which chickens currently are unprotected from April to October, when insects are in season in the Northern Hemisphere. We found that in July hundreds of flies per day passed through the ventilation system into a broiler house and that 8.2% of flies captured in the environment had the potential to transmit C. jejuni from outside animals to chickens in the broiler house. These results warrant further research on how to combat the summer peak of Campylobacter in broilers to improve the safety of the human food supply.