Fly transmission of Campylobacter.

An annual increase in Campylobacter infection in England and Wales begins in May and reaches a maximum in early June. This increase occurs in all age groups and is seen in all geographic areas. Examination of risk factors that might explain this seasonal increase identifies flies as a potential source of infection. The observed pattern of infection is hypothesized to reflect an annual epidemic caused by direct or indirect contamination of people by small quantities of infected material carried by flies that have been in contact with feces. The local pattern of human illness appears random, while having a defined geographic and temporal distribution that is a function of the growth kinetics of one or more fly species. The hypothesis provides an explanation for the seasonal distribution of Campylobacter infections seen around the world.

**********

Campylobacter spp. are the most common bacterial causes of diarrhea in England and Wales (1). The epidemiologic features of Campylobacter infection have proved difficult to discover, and extensive strain typing has failed to clarify the main transmission routes. Testable hypotheses must be established to explain available evidence, particularly the reason for the observed seasonality. Relatively few outbreaks of Campylobacter gastroenteritis occur (2), and most cases are sporadic. In case-control and case-case studies of sporadic Campylobacter infections, most cases remain unexplained by recognized risk factors (3,4).

The annual increase in Campylobacter infections in England and Wales begins at approximately day 130 (May 9) and reaches a maximum at approximately day 160 (June 8) (Figure 1). Although this seasonal rise is seen in all ages, it is more marked in children (5). Cases in towns and cities across England and Wales show broadly similar seasonal changes in distribution (Figure 2). The relative geographic uniformity of the increase seen in May of most years has the temporal appearance of an annual national epidemic. Because person-to-person infection within the community is uncommon, it is likely that the epidemic is caused by a single main driver for human Campylobacter infection. The possible seasonal drivers were examined, and only vector transmission by flies appears to provide a convincing explanation for the observed seasonal trends (Table).

The seasonal increase in Campylobacter infections in May and June in England and Wales is hypothesized to reflect an annual epidemic caused by direct or indirect exposure of humans to contaminated material carried by several fly species that have been in contact with human, bird, or animal feces or contaminated raw foods. Flies have been shown to carry Campylobacter and can infect both humans and animals (6-8). Intervention studies have demonstrated diarrheal disease reduction linked to control of flies (9-11), and deaths from diarrheal diseases have been linked to measurements of fly abundance (12). The local pattern of human Campylobacter infection appears random, while having a defined geographic and temporal distribution. This distribution is predicted to be linked to the growth kinetics of 1 or more fly species and their access to environmental sources of Campylobacter in feces or food. The seasonal increase in fly populations results from rainy weather and an increase in temperature that causes the development from egg to fly to occur in days rather than months. Individual flies can lay hundreds of eggs, which can result in a large increase in fly numbers in a short period. Fly numbers fluctuate through the summer and decline in October, but the decline is less dramatic and defined than the spring increase.

Disease transmission is hypothesized to occur through small quantities of contaminated material carried on the feet, proboscis, legs, and body hairs or from material regurgitated or defecated by flies. The variety, numbers, virulence and viability of organisms in the contaminated material will differ, and some contamination will include Campylobacter while others will not. Contamination will be distributed over a variety of food types. Contamination of food by flies could occur at any stage of the food supply chain, but Campylobacter counts within the contaminated material on foods will decrease over time; consequently, most infection will result from contamination close to consumption (e.g., in the domestic or catering environment). Because whether a fly has visited contaminated feces is unknown and how a person becomes infected is uncertain, epidemiologic investigation is difficult.

A number of synanthropic fly species could be involved, including houseflies (e.g., Musca spp., Fannia spp.), blowflies (e.g., Calliphora spp., Lucilia spp.), and other dung-related flies (e.g., Sarcophaga spp., Drosophila spp.) (13). These flies have individual behavioral patterns, ecology, physiology, and temporal and geographic distributions that will influence the likelihood of their being in kitchens, on human or animal feces, and on food. Although Musca domestica is the species most likely to be involved because it is commonly found in houses and food-processing establishments, larger flies (e.g., Calliphora spp.) may be able to transmit larger numbers of Campylobacter.

Flies contaminated through fecal contact will carry heterogeneous mixtures of organisms, including any pathogens that are present within the feces, and may be able to cause a variety of human infections, including infection by different Campylobacter species and types. This fact partially explains the lack of a clear epidemiologic picture arising from Campylobacter typing work. Gastrointestinal disease caused by flies is more likely to involve pathogens with a low infectious dose (e.g., Shigella, Campylobacter, Cryptosporidium, Giardia, Cyclospora, Escherichia coli O157), and some of these could have a seasonal component related to flies. Where high fly populations and poor hygiene conditions prevail, as in disasters or famines, or where pathogens can grow within fly-contaminated food, the potential exists for transmitting pathogens with a high infectious dose (e.g., Vibrio cholerae, Salmonella spp.). The access that flies have to human and animal feces will influence the degree to which they are contaminated with different enteric pathogens.

Contamination of a range of foods by flies will result in a pattern of infection that will not be amenable to identifying specific vehicles through standard case-control, case-case, or cohort studies, unless specific objective or subjective assessments of fly numbers can be obtained. Fly monitoring will need to be undertaken. An alternative approach could use estimates of fly population numbers based on climatic conditions to compare with data on human Campylobacter infections. This approach has the advantage of being able to use historical climatic and disease surveillance data. The broad relationship between Campylobacter cases and ambient temperature has not been explained in terms of disease causation. The time taken for the larvae of M. domestica to develop (13) was applied to temperature data for England and Wales and has been used to show a strong relationship between Campylobacter cases per week and M. domestica larval development time for 1989 to 1999 (Figure 3). Periods when Campylobacter cases exceed a 7-day average of 170 cases per day occurred when M. domestica larval development time was <3 weeks.

[FIGURES 1-3 OMITTED]

The hypothesis predicts that the Campylobacter infection rates will be higher in persons living close to animal production and lower in urban settings because fly numbers will be lower. Some evidence from the United Kingdom (1,14) and Norway (15) supports this hypothesis. Seasonal changes in Campylobacter incidence that are seen around the world may result from changes in fly populations and flies' access to human and animal feces. Much emphasis on foodborne disease reduction has rightly been on kitchen hygiene, since the low infectious dose of Campylobacter makes cross-transmission from raw meats to ready-to-eat foods a substantial risk in domestic and catering environments. Fly transmission may be the most important source of infection in kitchen transmission routes, and establishments that sell ready-to-eat foods may be sources of Campylobacter, if effective fly control is not in operation. Flies may also be important in transmitting Campylobacter in poultry flocks (16) and between other agricultural animals.

While flies are regarded as important mechanical vectors of diarrheal disease in developing countries, control has largely concentrated on improving drinking water and sewage disposal. In the industrialized world, flies are thought to play a minor role in the transmission of human diarrheal diseases. Immediately intervening in the transmission of Campylobacter gastroenteritis should be possible through increased public awareness and more effective fly control.

Table. Risk factors that might affect Campylobacter seasonality *

Risk factor                        Outbreaks    Evidence of seasonality

Barbecuing                            Yes               Medium
Birds                                 Yes               Strong
Bottled water                         No                 None
Chicken                               Yes               Medium
Cross-contamination                   Yes                None
Domestic catering                     No                 None
Farm visit                            Yes                None
Farm animals                          Yes                Weak
Flies                                 No                Strong
Food handlers                         Yes                None
Food packaging                        No                 None
Immunologic response                  No                 Weak
Mains supply drinking water           Yes                None
Nosocomial                            Yes                None
Pets                                  No                 Weak
Pools, lakes, streams                 No                 None
Private drinking water supplies       Yes                Weak
Protozoa                              No                 None
Salads and fruit                      Yes                Weak
Stir-fried food                       Yes                None
The countryside                       No                 Weak
Transmission in families              Yes                None
Travel abroad                         No                 None
Unpasteurized milk                    Yes                Weak
Weather/climate                       No                Medium

                                   Credibility as the main seasonal
Risk factor                                     driver

Barbecuing                                       Low
Birds                                            Low
Bottled water                                    Low
Chicken                                         Medium
Cross-contamination                              None
Domestic catering                                None
Farm visit                                       None
Farm animals                                     Low
Flies                                            High
Food handlers                                    None
Food packaging                                   None
Immunologic response                             None
Mains supply drinking water                      None
Nosocomial                                       None
Pets                                             Low
Pools, lakes, streams                            None
Private drinking water supplies                  None
Protozoa                                         Low
Salads and fruit                                 Low
Stir-fried food                                  None
The countryside                                 Medium
Transmission in families                         None
Travel abroad                                    None
Unpasteurized milk                               None
Weather/climate                                 Medium

* Evidence base provided in online Appendix (available from
http://www.cdc.gov/ncidod/EID/vol11no03/04-0460_app.htm).

Acknowledgments

This hypothesis arose after a lecture by Professor Sandy Cairncross at the Centers for Disease Control and Prevention, Atlanta, in the spring of 2002. I thank Fay Burgess, Radha Patel, Chris Lane, Douglas Harding, and Erol Yousef for help in preparing the data; Jim McLauchlin, Barry Evans, Chris Little, and John Edmonds for critically commenting on versions of the paper; and Andre Charlett for statistical support.

References

(1.) Tam CC. Campylobacter reporting at its peak year of 1998: don't count your chickens yet. Commun Dis Public Health. 2001;4:194-9.

(2.) Frost JA, Gillespie IA, O'Brien SJ. Public health implications of campylobacter outbreaks in England and Wales, 1995-9: epidemiological and microbiological investigations. Epidemiol Infect. 2002; 128:111-8.

(3.) Adak GK, Cowden JM, Nicholas S, Evans HS. The Public Health Laboratory Service national case-control study of primary indigenous sporadic cases of campylobacter infection. Epidemiol Infect. 1995;115:15-22.

(4.) Gillespie IA, O'Brien S J, Frost JA, Adak GK, Horby P, Swan AV, et al. A case-case comparison of Campylobacter coli and Campylobacter jejuni infection: a tool for generating hypotheses. Emerg Infect Dis. 2002;8:937-42.

(5.) Louis VR, Gillespie IA, O'Brien S J, Russek-Cohen E, Pearson AD, Colwell RR. Temperature driven Campylobacter seasonality in England and Wales. Appl Environ Microbiol. 2005;71:85-92.

(6.) Khalil K, Lindblom GB, Mazhar K, Kaijser B. Flies and water as reservoirs for bacterial enteropathogens in urban and rural areas in and around Lahore, Pakistan. Epidemiol Infect. 1994; 113:435-44.

(7.) Rosef O, Kapperud G, House flies (Musca domestica) as possible vectors of Campylobacter fetus subsp, jejuni. Appl Environ Microbiol. 1983;45:381-3.

(8.) Shane SM, Montrose MS, Harrington KS. Transmission of Campylobacterjejuni by the housefly (Musca domestica). Avian Dis. 1985;29:384-91.

(9.) Chavasse DC, Shier RP, Murphy OA, Huttly SR, Cousens SN, Akhtar T. Impact of fly control on childhood diarrhoea in Pakistan: community-randomised trial. Lancet. 1999;353:22-5.

(10.) Cohen D, Green M, Block C, Slepon R, Ambar R, Wasserman SS, et al. Reduction of transmission of shigellosis by control of houseflies (Musca domestica). Lancet. 1991 ;337:993-7.

(11.) Emerson PM, Lindsay SW, Walraven GE, Faal H, Bogh C, Lowe K, et al. Effect of fly control on trachoma and diarrhoea. Lancet. 1999;353:1401-3.

(12.) Niven J. Summer diarrhoea and enteric fever. Proc R Soc Med. 1910;III(Epidem. Sect.):131-216.

(13.) Kettle DS. Medical and veterinary entomology. 2nd ed. Wallingford (UK): CABI Publishing; 2000.

(14.) Skirrow MB. A demographic survey of campylobacter, salmonella, and shigella infections in England. A Public Health Laboratory Service survey. Epidemiol Infect. 1987;99:647-57.

(15.) Kapperud G, Aasen S. Descriptive epidemiology of infections due to thermotolerant Campylobacter spp. in Norway, 1979-1988. APMIS. 1992;100:883 90.

(16.) Hald B, Skovgard H, Bang DD, Pedersen K, Dybdahl J, Jespersen JB, et al. Flies and Campylobacter infection of broiler flocks. Emerg Infect Dis. 2004;10:1490 2.

Dr. Nichols is an epidemiologist in the Communicable Disease Surveillance Centre, which is part of the Health Protection Agency Centre for Infections in London. His research interests include waterborne diseases, foodborne diseases, cryptosporidiosis, and enteric infections.

Address for correspondence: Gordon L. Nichols, Environmental and Enteric Diseases Department, Communicable Disease Surveillance Centre, Health Protection Agency Centre for Infections, 61 Colindale Ave, London NW9 5EQ, United Kingdom; fax: [phone]; email: [e-mail]

Gordon L. Nichols, Health Protection Agency Centre for Infections, London, United Kingdom