Contents. Preface 3. Introduction 4. General features of the Laboratory Surveillance of Infectious Diseases 6

Transcription

1 Zoonoses and zoonotic agents in humans, food, animals and feed in the Netherlands 2002

2 Contents Preface 3 Introduction 4 General 4 Demographic data 5 Number of animals slaughtered 5 Notifiable diseases 5 Chapter 1 General features of the Laboratory Surveillance of Infectious Diseases 6 Frame 1 Explosion project: a study of outbreaks of gastroenteritis in the Netherlands 6 Chapter 2 General features of surveillance, monitoring and control programmes in feed, animals and food Monitoring programme for Salmonella in feed Surveillance of zoonotic bacteria in farm animals Control programmes for Salmonella and Campylobacter in poultry Salmonella and Campylobacter monitoring of poultry meat at retail 11 Chapter 3 Zoonotic pathogens Brucella spp Campylobacter spp. 13 Frame 2 Study into the similarity of Campylobacter isolates from humans and their pets 17 Frame 3 Campylobacter risk management and assessment Chlamydophila psittaci Echinococcus spp Shiga toxin-producing Escherichia coli (STEC) 20 Frame 4 Escherichia coli O157 infections - direct and environment-mediated transmission 22 Frame 5 Zoonotic pathogens on children's farms Hepatitis E virus Influenza Listeria monocytogenes Mycobacterium spp. 29 Frame 6 Epidemiological and economic evaluation of detection methods for bovine tuberculosis 31 Frame 7 Mycobacterium avium subsp. paratuberculosis Rabies virus Salmonella spp. 34 Frame 8 Monitoring of feed ingredients and compound feed from outside the European Union 48 Frame 9 Salmonella control programme in the pig sector 49 Frame 10 Purchase of pig manure as a preventable risk factor for Salmonella Typhimurium DT-104 infections in dairy cattle 49 Frame 11 Explosive increase of Salmonella Paratyphi B variation Java in poultry Tick-transmittable zoonotic pathogens: Borrelia burgdorferi, Ehrlichia/Anaplasma spp., tick-borne encephalitis virus and Babesia spp. 52 Frame 12 Doubling of tick bites and erythema migrans between 1994 and 2001 in humans and related changes in risk factors Trichinella spp. 55 Chapter 4 Monitoring of antimicrobial resistance 58 Chapters 66

3 Abbreviations of institutes and organisations involved Preface This report presents a summary of data up to and including 2002 on the occurrence of zoonoses and zoonotic agents in humans, food, slaughter animals, wildlife animals and feed in the Netherlands. This report is based on data that, in accordance with the zoonoses directive 92/117/EEC, are reported annually to the European Commission, supplemented with data from Dutch surveillance, monitoring and control programmes and relevant research projects concerning zoonoses and zoonotic agents by the different institutions that have contributed to the preparation of this report. The report also includes information on recent research on the antibiotic resistance of microorganisms derived from human and animal material. Specific documentation and reports regarding the described programmes and research projects are available from the authors mentioned in the editorial list. AID ASG CBS CIDC EMC GD GGD IGZ KDD LCI LNV NRL PDV PHL PVE RIVM UM UU VMDC V&V VWA VWA/KvW VWA/RVV VWS General Inspectorate (Ministry of LNV) Animal Sciences Group (Lelystad) Central Statistical Office, Netherlands Central Institute for Animal Disease Control Erasmus Medical Centre Animal Health Service Municipal Health Service Inspectorate for Health Care (Ministry of VWS) Animal Feed Sector Inspection Services National Co-ordinating Centre for Infectious Diseases (Ministry of) Agriculture, Nature and Food Quality National Reference Laboratory Product Board Animal Feed Regional Public Health Laboratory Product Boards for Livestock, Meat and Eggs National Institute for Public Health and the Environment University of Maastricht University of Utrecht Veterinary Microbiological Diagnostic Centre (UU) Department of Public Health and Food Safety (UU) Dutch Food and Consumer Product Safety Authority (Ministry of LNV) Inspectorate for Health Protection and Veterinary Public Health National Inspection Service for Livestock and Meat (Ministry of) Public Health, Welfare and Sport This report is available on the website of the Dutch Food and Consumer Product Safety Authority: The extended dataset on antimicrobial resistance and trends in The Netherlands, has been published recently as a report: Maran 2002 and Maran These reports will also be placed on the VWA-website: 3

4 Introduction In the Netherlands, various laboratories are involved in the investigation of zoonoses in humans and animals and determining the presence of zoonotic agents in humans, food, feed and animals. VWA/KvW has several laboratories in which investigations are carried out into food and feed samples and into animal waste products. Materials of human origin are examined for pathogens by the regional public health laboratories. On behalf of VWA/KvW and IGZ, the National Institute for Public Health and the Environment (RIVM) conducts several monitoring and surveillance programmes regarding zoonoses and zoonotic agents (bacteria, viruses and parasites) in humans and animals or animal material. RIVM also houses the National and Community Reference Laboratory for Salmonella. General Zoonoses are diseases that are transmittable between animals and humans. Both the Ministry of Public Health, Welfare and Sport (VWS) and the Ministry of Agriculture, Nature and Food Quality (LNV) in the Netherlands are responsible for the monitoring and the control of these zoonotic diseases in the food production chain. Since 2003 the Dutch Food and Consumer Product Safety Authority (VWA) of the Ministry of LNV has been responsible for inspection and supervision of the food production chain, including food, feed and animals. The VWA consists of a central co-ordinating unit and two subsidiary units: the Inspectorate for Health Protection and Veterinary Public Health (VWA/KvW) and the National Inspection Service for Livestock and Meat (VWA/RVV). The main tasks of both subsidiary units, that in 2002 belonged to the Ministry of VWS, are still the same as before 2003: the VWA/KvW has a public health responsibility with regard to food-borne infections and zoonoses and the VWA/RVV is involved in meat inspections and in the registration and control of diseases, including zoonoses, in animals. As a result of the activities of the VWA/KvW and the VWA/RVV, the protection of food-safety during all stages of the production chain and the health protection of animals are now the responsibility of a single authority, VWA. Other institutions are also involved in the protection of animal health. At the request of the Ministry of LNV, the Animal Health Service (GD) is responsible for the sampling in some animal disease surveillance programmes. In addition, the Product Boards for Livestock, Meat and Eggs (PVE) conduct the Salmonella and Campylobacter monitoring and control programme prescribed by Directive 92/117 EC under the responsibility of the Ministry of LNV. Relevant research is also carried out at the Central Institute for Animal Disease Control (CIDC). The investigation of animals for the presence of rabies is a major activity of that institute. In collaboration with RIVM, this institute is also involved in a study on the risk factors concerning the development of campylobacteriosis in humans. The department of virology of the Erasmus Medical Centre (EMC) in Rotterdam plays an important role in studies on viral zoonoses, such as influenza. The National Influenza Centre, consisting of both EMC and RIVM, co-ordinates investigations on influenza in the Netherlands. A zoonotic agent can be transmitted from animals to humans in various ways. Foodstuffs of animal origin are the most important source of zoonoses. Salmonella and Campylobacter are the major bacterial agents of food-borne zoonoses in the Netherlands. This report on zoonoses and zoonotic agents includes recent findings on antibiotic resistance from both humans and farm animals. An important purpose of that research is the detection of potential public health risks related to the use of antibiotics in animal husbandry. In the Netherlands notifiable human zoonotic diseases must be reported to the Municipal Health Services (GGD), whereas the registration of these diseases is a responsibility of the Inspectorate for Health Care (IGZ) of the Ministry of VWS. When a zoonotic disease is reported, the local GGD is responsible for the control of the disease. If more than one GGD is involved, the National Co-ordinating centre for Infectious Diseases (LCI) is responsible for co-ordination of the control activities. IGZ can, often in close collaboration with VWA, initiate monitoring and surveillance programmes for zoonotic agents and zoonotic diseases in humans.

5 Demographic data (source: CBS) Dutch population in January 2002 Age distribution Total 0-19 years 3,940, years 4,685, years 5,280, years 1,667,107 > 80 years 513,607 Total 16,105,285 Dutch population Man 7,971,967 7,909,855 7,846,317 Woman 8,133,318 8,077,220 8,017,633 Total 16,105,285 15,987,075 15,863,950 Growth +/ , ,125 === Number of animals and farms registered in 2002 Animals (x1000) Farms Cattle 3,858 41,266 Pigs 10,962 11,851 Sheep 1,186 15,254 Goats 255 4,853 Horses and ponies ,774 Ducks and turkeys 2, Poultry 101,052 3,358 - broilers 54,660 1,096 - layers 38,889 1,880 Notifiable diseases Zoonosis IZW GWWD Anthrax X X 1) Botulism X - Brucellosis (B. abortus) X X 1) Brucellosis (B. suis) X X 2) BSE - X 3) Campylobacteriosis - X 3) Echinococcosis - X 3) EHEC/STEC X - Leptospirosis (L. hardjo) X X 3) Listeriosis - X 3) Psittacosis X X 4) Rabies X X 1) Salmonellosis - X 3) Toxoplasmosis - X 3) Trichinellosis X X 2) Tuberculosis X X 1) Yersiniosis - X 3) IZW: GWWD: 1) 2) 3) 4) mammals cattle all animals birds (except poultry) Infectious Diseases Act (human) Animal Health and Welfare Act (Animals) Number of animals slaughtered in 2002 (source: VWA/RVV) Animals (x1000) Total Cattle, incl. veal calves 1,874 Pigs 15,413 Sheep 440 Goats 11 Horses and ponies 3 Ducks and turkeys 10,159 Poultry 478,656 - broilers 463,322 - layers 15,333 5

6 Chapter 1 General features of the Laboratory Surveillance of Infectious Diseases Since 1989 a sentinel-based surveillance programme on bacterial pathogens has been operative, called Laboratory Surveillance Infectious diseases (LSI). Sixteen regional public health laboratories (PHLs) participate in this programme, covering >60% of the Dutch population. All first isolates of Salmonella are sent to the National Reference Laboratory (NRL) at RIVM for serotyping, phagetyping and sensitivity tests to antibiotics. Basic information from the patient is collected, such as age, gender, residence, country of infection and the possible source of infection. From April 1995 onwards, on a weekly basis, these laboratories also report the total number of detected Campylobacter spp. and the total number of stool samples examined. transmission routes causing outbreaks of gastroenteritis is unknown. In a national project in 2002, the RIVM studied the role of specific pathogens, transmission routes and sources in outbreaks of gastroenteritis with at least five cases involved. The project also aimed to serve as a surveillance tool for Norwalk-like virus and to improve the process of investigation and management of gastroenteritis outbreaks in the Netherlands. GGDs, VWA/KvW and medical microbiological laboratories collaborated in the project. In the national 2002 study, all outbreaks reported to a GGD or to VWA/KvW, were recorded and examined. In total, 281 outbreaks were reported affecting over 8,700 patients, including 62 hospitalisations and 16 deaths. Outbreaks mainly occurred in nursing homes (161), restaurants (32), hospitals (25), day-care centres (20) and other locations (43). Most of the outbreaks were caused by person-toperson transmission (78%), (cross-) contaminated food (13%) or a combination of both (8%). One large water-borne outbreak was thoroughly examined. In 86% of outbreaks, faecal samples were examined (1,474 faecal samples for cases; 202 for controls) and in almost two-thirds of all outbreaks the causative agent was found. Norovirus was found most frequently (and explained 54% of the examined outbreaks), followed by Salmonella spp. (4%) and Campylobacter spp. (1%). Giardia lamblia and Cryptosporidium parvum were systematically included in the diagnostics, but were found only once as the cause of an outbreak (in a day-care centre). Since April 1996, Escherichia coli serotype O157 (sorbitol-negative isolates that agglutinate with E. coli O157 antiserum) that may produce shiga toxin, has been added to the programme. First isolates of E. coli O157 are also sent to the NRL for confirmation and further typing. Except for O-typing and H-typing, isolates are typed for the presence of shiga toxin genes (stx1 and stx2), the E. coli attaching and effacing gene (eae-gene), the enterohemolysin gene and they are characterised by pulsed-field gel electrophoresis. From April 1999 onwards, all Dutch medical microbiological laboratories have contributed to the surveillance of shiga toxin-producing E. coli (STEC). In this enhanced STEC-surveillance programme, the municipal health services interview all diagnosed cases in order to obtain detailed information on risk factors and clinical aspects, using a standardized questionnaire. The Salmonella-data from this surveillance are presented in section 3.11 and the Campylobacter and STEC data are reported in sections 3.2 and 3.5, respectively. Frame 1 Explosion project: a study of outbreaks of gastroenteritis in the Netherlands The knowledge on outbreaks of gastroenteritis in the Netherlands is limited. In the Dutch surveillance system, outbreak surveillance is limited to food-borne outbreaks, based on mandatory notification to IGZ and voluntary reports to VWA/KvW. In the majority of these reported food-borne outbreaks, no causative agent has been found (40-90%). In addition, the relative importance of other possible

7 Chapter 2 General features of surveillance, monitoring and control programmes in feed, animals and food 2.1 Monitoring programme for Salmonella in feed Within the framework of EU Directive 92/117/EU, samples of feed ingredients and compound feed have been taken regularly for several years in the Netherlands. In line with this, the feed sector has implemented a national monitoring programme for Salmonella in compound feed and feed materials. Standards and the necessary control measures have been laid down in a GMP regulation for the feed chain sector for the purpose of controlling Salmonella in (poultry) feed. The aim of the programme is to minimise the introduction of Salmonella into the poultry chain via animal feed. The programme started a few years ago, but has been intensified during the past few years. Especially the monitoring of feed ingredients has been intensified since The current GMP standards (maximum Salmonella incidence and process standards for enterobacteriaceae) are presented in table 2.1. In 2002 the product standards for poultry feed were made more stringent for laying hens and broilers. The control measures are mostly aimed at processing during the production of poultry feed and at the supply of Salmonella-critical feed materials that are used in feed. In addition there are general control measures in the GMP feed sector standard for other animal feed than for poultry in order to minimise the introduction of Salmonella by way of feed. For these kinds of feed so far only standards for enterobacteriaceae after heat treatment have been laid down: the target value is < 100 cfu enterobacteriaceae per gram, the intervention limit is 1,000 enterobacteriaceae per gram. Monitoring is performed to verify the effectiveness of the control measures for both the Salmonella-critical feed materials and poultry feed. The following feed ingredients have been assessed as Salmonella-critical for 2002: Brazilian extracted soy beans and expeller, South American fish meal, extracted rapeseed and expeller, toasted soy beans and eggshells. In addition, the programme Table 2.1 Maximum Salmonella incidence and process standards for enterobacteriaceae (PDV) Product norms 2002 Maximum Salmonella spp. Maximum % in batches to be contamination percentage in delivered with S. Enteritidis or batches to be delivered Typhimurium Poultry compound feed and feed for single delivery to poultry companies, for: Top breeding 0+ % 0+ % Raising parent stock 0+ % 0+ % Parent stock 0+ % 0+ % Rearing hens laying sector 1 % 1 0+ % Laying hens 1 % 1 0+ % Consumption turkeys 0+ % 2 0+ % Broilers 0+ % 0+ % Process norms Maximum cfu enterobacteriaceae per gram Target value Action limit Poultry compound feed for: Top breeding 100 Raising parent stock 100 Parent stock 100 Other poultry compound feed if given heat treatment, for: Breeding hens laying sector <100 1,000 Laying hens <100 1,000 Meat turkeys <100 1,000 Broilers <100 1,000 cfu colony forming units 1 Action limit value, in 2001 the product norm for laying hens was 2%. 7 2 In 2001, the product norm for broilers was 0.4%.

8 monitors feed for other animal species apart from poultry and of non-salmonella-critical feed materials, to avoid unexpected contamination from a formerly unsuspected source. The monitoring is primarily done by the companies involved as part of their GMP programmes. A national, independent monitoring programme is carried out by the Animal Feed Sector Inspection Services (KDD) under the authority of the Product Board Animal Feed (PDV). The results of the KDD-monitoring programme for Salmonella in feed ingredients and compound feed are presented in section Surveillance of zoonotic bacteria in farm animals To enable adequate control of zoonotic bacteria, such as Salmonella spp., Campylobacter spp. and Escherichia coli O157, reliable data on the prevalence and trends of these agents in their natural reservoir(s) are necessary. Furthermore, associations between the types of these pathogens occurring in their reservoirs and those causing disease in humans should be determined. For the three mentioned zoonotic bacteria, farm animals serve as the main natural reservoirs, whereas food products act as important vectors for transmission to humans. Therefore, VWA/KvW commissioned RIVM to run a monitoring programme in farm animals in the Netherlands for the mentioned bacteria, which started in Faecal samples of laying hens, broilers, veal calves, finishing pigs and dairy cattle have been and still are collected weekly on farms throughout the Netherlands. GD performs the random selection of farms for all animal species, except for veal farms. Selection for the latter species is supplied by the Foundation for Quality Guarantee of Veal (SKV). The selected farm managers are requested to participate in the monitoring programme on a voluntary basis. The number of farms in the selection is based on a statistical approach, with a flock or herd as the subject of analysis. Approximately 1,000 flocks/herds are sampled each year. The sample sizes and input parameters for the calculations for 2002 are presented in table 2.2. The unit of measurement is a flock or herd of animals, indicating for layers, broilers and finishing pigs all animals of similar age kept in one building. For dairy cattle, a herd indicates all lactating animals; for calves all animals housed within one building. In the sample selection, stratification is applied for the region in which a farm is located. Stratification is also applied according to farm size for dairy cattle, finishing pigs and veal farms, and an additional stratification factor for veal calves is age. Sampling is performed on one flock or herd per farm. If several flocks/herds are present, one is randomly selected. Related to the size of the flock or herd, a number of faecal samples, with a maximum of 60 per flock/herd, is taken from the floor or the manure conveyor. The number of faecal samples is sufficient to detect, by approximation, 5% shedding animals with a 95% confidence. These samples are subsequently pooled to a maximum of five pooled samples per flock, with a maximum of 13 faecal samples per pooled sample. These pooled samples are sent to RIVM within 48 hours after sampling for bacteriological investigation for Salmonella spp., Campylobacter spp., and/or E. coli O157. A flock or herd is considered positive for a specific target bacterium if it is cultured from at least one of the pooled samples. Subsequently, the samples are released for investigation for zoonotic parasites and viruses as part of other projects. Just after or prior to sampling, a questionnaire regarding farm and flock characteristics is completed in co-operation with the farm manager. These data can be used to identify factors that are associated with the presence of the bacteria on farms. This information can then be used in the developmental stage of effective intervention strategies to reduce the prevalence at farm level. The results of the investigations for Salmonella, Campylobacter and E. coli during are described in sections 3.11, 3.2 and 3.5, respectively. 2.3 Control programmes for Salmonella and Campylobacter in poultry In 1997, the poultry sector started an eradication programme for Salmonella and Campylobacter, called the "Plan of Approach for Salmonella (and Campylobacter)". The rules of this programme are based on five principles: hygiene requirements, cleaning and Table 2.2 Calculated sample size per animal species for 2002, based on the a priori estimate of the prevalence for 2002 and the desired accuracy level of the annual estimate at a 90% confidence level (Surveillance programme VWA/KvW - RIVM) Layers Broilers Finishing pigs Dairy cows Veal calves Total number of flocks per year 1 2,843 17,367 70,914 25,824 13,050 Estimated prevalence for % s 25% c 30% s 10% e 20% e Accuracy 5% 5% 5% 4% 5% Sample size Based on data from GD (layers, broilers, finishing pigs and dairy cows) or SKV (veal calves) Based on data from the surveillance programme in previous years ( ). The prevalence of the bacterium with the highest estimate is used in the calculation s, c, e Based on the prevalence of Salmonella, Campylobacter or Escherichia coli, respectively

9 disinfection, incoming and outgoing investigations, reporting results and measures taken after infection. The objectives of the plan of approach were to achieve a reduction in the infection level of the broiler flocks with Salmonella at the end of the slaughter process to less than 10% infected and a reduction in the infection level of flocks of layer hens to less than 5% with Salmonella Enteritidis and S. Typhimurium. in poultry meat and the control programme in the egg sector New objectives were set along with these new programmes. The main objective for the poultry meat sector was set at a maximum Salmonella contamination level of produced meat of 10% at the end of The objective in the egg sector was tightened to 0+ % S. Enteritidis/S. Typhimurium infected eggs in the consumption channel. For both sectors the end target is a contamination degree with Salmonella of 0+ %. In 2000, the objectives were not attained and additional measurements were considered necessary. These adjustments resulted in the control programme Salmonella and Campylobacter Table Salmonella control programme in the broiler sector (PVE) Link Incoming Process control + outgoing Elite-broiler-breed rearing leaflets (40 pieces) at 4 weeks faeces samples (2x30 cloaca swabs by GD) max. 21 days before transfer faeces samples per flock (6x25 cloaca swabs by GD) Elite-broiler-breed production at 20 weeks faeces samples, (6x25 cloaca swabs by GD) from 24 weeks, every 4 weeks faeces samples per flock (6x25 cloaca swabs by GD) Hatchery logistics consider the outgoing results from earlier batches fluff samples (25g) or meconium (250) or layers (60) per hatching entity Broiler-breed rearing leaflets (40 pieces) at 4 weeks faeces samples (2x30 cloaca swabs by GD) max. 21 days before transfer faeces samples per flock (6x25 cloaca swabs by GD) Broiler-breed production at weeks, faeces samples per flock (6x25 swabs by GD) every 9 weeks blood samples per flock by GD (1%: min 30 max 60 animals) or from 26 weeks fluff samples (25g) or meconium (250) or layers (60) per hatching entity Hatchery logistics consider the outgoing results from earlier batches fluff samples (25g) or meconium (250) or layers (60) per hatching entity Broiler production leaflets per batch (40 pieces) 2 pair of overshoe samples per flock Slaughterhouse 1x30 swabs from appendix per flock of every batch 1 sample of end product (25g breast skin) Table Campylobacter control programme in the broiler sector (PVE) Link Incoming Process control + Outgoing Broiler production 1 x 5 swabs from fresh droppings 2 times/year for each farm Slaughterhouse 1x30 swabs from appendix per flock negative on farm invest. (NB. Only the flocks that were already examined at farm level) irrespectively of investigation at farm, 1 breast skin (25g) per day/slaughterhouse is sampled and analysed. 9

10 Developments in the control programme Salmonella and Campylobacter poultry meat As the new control programme is focused on the Salmonella contamination of the end product, alterations were introduced in the monitoring system at slaughterhouses, and a monitoring commitment was added for the cutting plants. In slaughterhouses samples are taken from every batch of end product and these are then analysed for Salmonella and occasionally for Campylobacter. Based on these monitoring results, contamination percentages can be calculated. Slaughterhouses that deliver more than 10% Salmonella-contaminated poultry meat, are obliged to formulate a working plan to improve the situation. The slaughterhouses receive certificates from the Commodity Board, which they can apply for publication. On these certificates the results of contamination percentages for a period of three months (quarterly) are presented. An essential change in the control programme concerns the start of tracing experiments in the poultry chain. When a Salmonella infection is present in a flock, the Salmonella serotype that caused the infection must always be identified. Last year, it was made mandatory in the control programme to carry out further serotyping of samples of end product contaminated with Salmonella. In this way the kind of Salmonella serotypes in the poultry meat chain can be known at any time. In addition, since 1 January 2002, mandatory eradication of S.Enteritidis/S.Typhimurium positive breeding and production flocks and hatching eggs has been put into force. The monitoring activities for Salmonella and Campylobacter are summarized in tables and 2.3.2; monitoring results for Salmonella are described in section 3.11 and results of the investigations for Campylobacter are described in section 3.2. Table Salmonella control programme in the layer sector (PVE) Link Incoming Process control + Outgoing Elite-layer-breed rearing leaflets (40 pieces) at 4 weeks faeces samples (2x30 cloaca swabs by GD) max. 21 days before transfer faeces samples (6x25 cloaca swabs by GD) Elite-layer-breed production at 20 weeks faeces samples (6x25 cloaca swabs by GD) from 24 weeks, every 4 weeks faeces samples (6x25 cloaca swabs by GD) Hatchery logistics consider the outgoing results from earlier batches fluff samples (25g ) per hatching entity Layer-breed rearing leaflets (40 pieces) at 4 weeks faeces samples (2x30 cloaca swabs by GD) max. 21 days before transfer faeces samples (6x25 cloaca swabs by GD) decision for vaccination against S. Enteritidis Layer-breed production Not S.E.-vaccinated*: weeks, blood samples by GD (1% of flock: min 30 max 60 animals) Not S.E.-vaccinated*: every 9 weeks blood samples by GD (1%: min 30 max 60 animals) or from 26 weeks fluff samples (25 g) or meconium (250) or layers (60) per hatching entity S.E.-vaccinated*: weeks, faeces samples (6x25 cloaca swabs by GD) S.E.-vaccinated*: every 9 weeks faeces samples, (6x25 cloaca swabs by GD) from 26 weeks fluff samples (25 g) or meconium (250) or layers (60) per hatching entity Hatchery logistics consider the outgoing results from earlier batches fluff samples (25 g) or meconium (250) or layers (60) per hatching entity Layer rearing Decision for vacc. S. gallinarum max. 21 days before delivery, blood samples per flock by GD (0.5 % of flock: min 24, max 60) Layer production max. 9 weeks before slaughter blood samples per flock by GD (0.5% of flock: min 24, max 60) *S.E. = S. Enteritidis

11 Developments in the control programme egg sector In 2000, the target of the original control programme in the egg sector of less than 5% of the flocks infected with S. Enteritidis/ S. Typhimurium was not reached by far. Therefore the control programme was adjusted in The most important alterations concerned an obligation to trace-back a S. Enteritidis/ S. Typhimurium infection, as in the poultry meat sector. Moreover, only flocks free of S. Enteritidis/S. Typhimurium are allowed to enter a laying hen farm. If an infection with S. Enteritidis/S. Typhimurium occurs at the laying hen farm, the flocks must be vaccinated for S. Enteritidis. Eggs from Salmonella-positive flocks must be delivered to the egg products industry. The monitoring and vaccination activities regarding Salmonella are summarized in table and results of investigations concerning Salmonella are described in section Salmonella and Campylobacter monitoring of poultry meat at retail During the period poultry products such as whole carcasses, drumsticks, wings, breasts and fillets were sampled randomly at retail by VWA/KvW inspectors. Samples were only taken from cooled products without added spices. At this moment products of Dutch origin and imported products are not differentiated in the surveillance programme. The sampling and laboratory tests were performed according to the "Guidelines for the determination of the prevalence of Salmonella contamination in consumer poultry at retail level". The samples were taken at the main retail points for poultry meat: real butchers, regular poulterers (poultry sellers), market poulterers and supermarkets, which account for 95% of the poultry sold to Dutch consumers. The number of these retail points to be sampled depended on their market share, and based on this a total of 65 real butchers, 44 regular poulterers, 37 market poulterers and 243 supermarkets should be sampled each year. The sample size was 750 gram meat per product group per shop and is based on the quantity of chicken the average consumer buys at a time. For the sampling procedure the chicken meat was divided into four product groups: whole carcasses, parts of leg, parts of breast and "other parts". Each of these groups should be sampled at the four types of retail points, in theory resulting in a total of 1,556 samples taken each year for the investigation for the presence of Salmonella and Campylobacter. The results of the monitoring programme regarding Salmonella are presented in section 3.11; those for Campylobacter are presented in section

12 Chapter 3 Zoonotic pathogens 3.1 Brucella spp. Introduction Brucellosis is a bacterial zoonosis with a worldwide distribution and is caused by Brucella spp. The species of concern are B. abortus in cattle, B. melitensis in goat and sheep and B. suis in pigs. Dogs and cats can be infected by the above-mentioned Brucella spp. and also with B. canis. Brucella spp. can cause abortion in the animals concerned and a variety of clinical symptoms in man of which undulating fever is the most typical. B. melitensis infections in man are known as Malta fever. Bovine brucellosis used to have a worldwide distribution. However, in a number of countries this infection has been eradicated with intensive control programmes. B. melitensis and B. suis are irregularly distributed. Especially B. melitensis is a problem in European and African countries around the Mediterranean Sea. Man is infected by ingestion of infected raw milk and fresh cheese or by contact with infected animals. Farm animals General All infections with Brucella spp. must be notified in the Netherlands. Bovine brucellosis caused by B. abortus used to be endemic in the Netherlands. Since the Dutch B. abortus control programme for cattle started in 1959 its prevalence has dropped from about 30% to 1% in Sporadic cases have occurred annually up to 1995 and the last infected herd was eradicated in In August 1999 by Decision 99/466/EC amended by Decision 00/69/EC, the Netherlands was declared officially free of bovine brucellosis by the European Commission. B. melitensis has never been reported in animals in the Netherlands and by order 93/52/EC, the Netherlands was also declared officially free of B. melitensis in The Netherlands is also free of B. suis with only two cases being reported in the late 1960s and another two in Three of these cases were imported, infected pigs and one was caused by feeding infected offal from imported hares. Monitoring and control B. abortus From October 1999, VWA/RVV has been responsible for controlling eventual outbreaks of brucellosis and GD, under the responsibility of the Ministry of LNV, carries out monitoring. According to Directive 98/46 all animals older than 24 months in 20% of the present cattle herds are examined once a year; the selection of herds is based on risk analysis. On a statistical basis, a selection is made of herds according to the following criteria: 1. Herds producing milk for cheese production at the farm without pasteurization. 2. Herds with imported cattle. 3. Herds with a high percentage of animals moved from and to the herd. 4. Herds in which on a yearly basis no abortions were reported. Dairy farms are examined by analysing bulk milk tank samples using an ELISA test; positive samples are re-examined using the milk ring test. Individual blood samples are taken from herds on non-dairy farms and examined by a micro-agglutination test; positive samples are re-examined using an ELISA test. Blood from cows that have aborted is first examined using a micro-agglutination test; positive samples are re-examined using the complement fixation test. Bulls are serologically examined for brucellosis according to Directive 88/407/EC before entering cattle sperm centres and on a yearly base thereafter. In addition, before export to some non-member states, cattle are serologically checked for brucellosis. B. melitensis To maintain the official free status, an annual screening programme under Directive 94/953/EEC has been set up to prove with 95% confidence that less than 0.2% of sheep and goat herds in the Netherlands are infected with B. melitensis. B. suis The free status is maintained by the notification system. According to Directive 90/429/EEC, all boars are examined before entering and when leaving the sperm centre. Breeding animals exported to non-member states are only examined upon request of the importing country. Results B. abortus For several years no cases of bovine brucellosis have been diagnosed in the Netherlands. In 2002 a total of 10,155 cows were examined as the result of an abortion. From those cows 42 were declared positive in the first test. After repeated sampling and investigation, 16 of those animals seemed serologically positive. These animals were killed and material from the animals was investigated by culture. All of those bacteriological tests appeared to be negative. The results of the monitoring and control programme for bovine brucellosis are listed in table 3.1.

13 Table 3.1 B. abortus in 2002 (Monitoring programme GD) Type of sample Pos./total % Tank milk from herds 0/7, Blood samples from herds * 0/38, Blood samples for export * 0/3, Blood samples in case of abortions * 0/10, Aborted material * 0/ Blood samples related to sperm centres * 0/6, * Samples derived from individual animals Farm animals Results of the surveillance of zoonotic bacteria in farm animals The results of the surveillance in farm animals showed a decrease of the prevalence of Campylobacter spp. in broilers between 1997 (first quarter not sampled) and 1999, an increase in 2000 and a decrease again in 2001 to a similar level as that observed in 1999 (table 3.2.1). No samples were taken from 6 March to 9 July in 2001 due to the Foot and Mouth disease (FMD) epidemic, which biased the annual prevalence estimate. FMD-adjusted percentages are given to make trend observations possible. After adjustment, a continuing decreasing trend from 1998 until 2000 can be observed, followed by a plateau situation in 2001 and an increase in B. melitensis In 2002 a total of 16,617 blood samples from sheep and 1,712 from goats were taken for examination. All samples in this screening programme, conducted since 1994, tested negative. B. suis In 2002 a total of 7,508 pigs were sampled. All specimens appeared to be negative as has been the case in the preceding years. Humans In the Netherlands brucellosis in humans is notifiable after positive laboratory diagnosis. Only a very few cases of human brucellosis are reported each year. In 1999 and 2000, 1 and 3 cases were reported, respectively, and in 2001 and 2002, 1 and 5 cases of brucellosis were reported, respectively. These notified cases could not be traced back to a Dutch origin. Conclusions All cattle, sheep, goats and pigs in the Netherlands are free of brucellosis. All reported cases of human brucellosis were acquired by travelling abroad or by consumption of contaminated, imported products. Based on these findings it can be concluded that the relevance of Brucella spp. as a zoonotic agent in the Netherlands is virtually zero. In the future special attention must be paid to eventual human cases possibly related to imported products from an expanding EU market. 3.2 Campylobacter spp. Introduction Thermophilic Campylobacter spp. (predominantly Campylobacter jejuni and C. coli) are the most frequently identified bacterial agents of gastroenteritis in the Netherlands and they pose a serious public health problem. Recent epidemiological studies performed in the Netherlands show that in 1999 approximately 100,000 cases of campylobacteriosis occurred in the general population, of whom about 25,000 consulted a general practitioner. In addition, there are about 60 cases annually of Guillain-Barre syndrome, several thousands of reactive arthritis and several dozen deaths related to this disease. The economic damage of this has not formally been estimated but it is expected to be more than 100 million Euro per year. Campylobacter occurrence in poultry meat in retail stores showed a similar pattern until 2000, but increased from this point onward. Results from PVE monitoring of broilers corresponded with the surveillance observations. Up to 2001 the majority of isolates from broilers were determined to be C. jejuni by using a mixed-pcr technique. Isolates cultured in 2002 were not further discriminated into species. Samples from dairy cattle, veal calves and finishing pigs were also frequently contaminated with Campylobacter. Samples from dairy cattle were mainly contaminated with C. jejuni, isolates from pigs were mainly C. coli and veal calves showed a similar occurrence of both C. coli and a mixture of C. jejuni/c. coli. Contamination of these food product categories with Campylobacter spp. is rarely found in retail stores, for which reason the surveillance for Campylobacter spp. in these animal species has been discontinued. However, if future indications point to an increased need for the Campylobacter surveillance, it may be re-instated. Table Campylobacter spp. in farm animals in the Netherlands (Surveillance programme VWA/KvW - RIVM) Animal species Year Pos./total % 1 Broilers (random) / /189 31; [30] /154 18; [20] /127 24; [16] / /166 27; [25] Dairy cattle (random) / /169 7 Veal calves (random) / /60 58 Finishing pigs (random) / / [..] = Foot and Mouth disease adjusted prevalence for comparison with the estimate of only samples taken in the fourth quarter 13

14 Results of the monitoring in the poultry meat production chain Within the control programme of the poultry meat production chain, additional screening has been performed since the end of Targets were set for reduction of the of Salmonella contamination in the end product for 2000 and again for Amongst others this was thought to be achieved through additional hygiene measures at different levels in the poultry meat production chain. The monitoring results show that the control programme in poultry has hardly been successful with respect to the reduction of contamination with Campylobacter (table 3.2.2). The same picture appears for findings in the surveillance on healthy chickens at farms (table 3.2.1) and in the monitoring of poultry products at retail by VWA/KvW (table 3.2.3). Table Percentage of Campylobacter-positive broiler flock/batches from different links in the chain (Control programme PVE) Matrix Link Faeces broiler farm Caecum slaughterhouse Neck skin slaughterhouse n.i.* n.i.* Breast skin slaughterhouse n.i.* n.i.* In the programme, breast skins are tested as a more realistic proxy of the end product. However, compared to the results of the monitoring of poultry at retail the levels of contamination are much lower in breast skins. Retail contamination levels are close to that found in caeca and neck skins at slaughter; however, such a discrepancy was not seen with respect to Salmonella (section 3.11). Figure shows the quarterly change of contamination at the farm and at the slaughterhouse; both trends show a strong seasonality. Food In the years a total of more than 9,000 samples were taken by VWA/KvW. The numbers examined yearly varied from 859 in 1999 to 1,600 in The isolation rates for Campylobacter show a decrease from 36.2% in 1996 to 23.5% in 1999, but since then the rate has gone up to a level above 30% (table 3.2.3). Humans In 2002, 3,421 cases of campylobacteriosis were found. The recorded cases are first isolates of 15 regional public health laboratories (coverage 62% of the Dutch population) from relevant patient material. This corresponds to an incidence of 34 per 100,000 inhabitants in Dutch epidemiology studies show that to arrive at the incidence of patients with campylobacteriosis visiting a GP or in the general population this figure should be multiplied by about 4 and 19, respectively. *n.i. = no investigation Figure Percentage of Campylobacter-positive flocks from hatchery to the end of the slaughter line Positive flocks (%) th st nd 3rd 4th 1st nd 3th 4th 1st nd 3th 4th 1st nd 3th 4th 1st nd 3th 4th Quarter Faeces (boiler farm) Caecum (slaughterhouse) Neck skin (slaughterhouse) Breast skin (slaughterhouse)

15 Human cases of campylobacteriosis decreased between 1996 and 1999 but increased in 2000 and increased once again in 2001 (table 3.2.4). The number of reported isolates in 2002 decreased and had returned to the level found in Exactly the same trend has been found for the contamination levels of poultry meat at retail (table 3.2.3). In the Netherlands routine surveillance of Campylobacter is restricted to laboratory-specific weekly frequencies. Information on age, gender, residence, travel, species and antibiotic resistance has been registered for one year in the so-called CaSa project. This project covered most of the public health laboratories between April 2002 and April Table shows that 22% of the cases were travel-related. The routine information accompanying a diagnostic request shows that 11% was travel-related, indicating that travel is strongly under-reported, as was also found for Salmonella. 90% of the Campylobacter infections, presumably domestically derived, concerned C. jejuni, 7% C. coli and involved 3% other C. spp. or was unknown. The contribution of C. coli to travel-related infections was twice as high as to those domestically derived. Antimicrobial resistance differed as well, depending on species and travel (see chapter 4). An increase was found in the number of reported food outbreaks related to Campylobacter contamination. Of 81 food-related outbreaks that were reported by the IGZ in 2002, 15 (20%) were related to Campylobacter contamination. In 2001, 9 out of 81 outbreaks were Campylobacter-related (11%) and in 2000, 4 out of 78 outbreaks (5%) were caused by Campylobacter. In 2002 the number of explosions (and the number of cases involved) of contamination with Salmonella was twice as high as that for Campylobacter and in 2000 this proportion was still 4 times as high. Part of the apparent increase in importance of proved Campylobacter-related explosions might be the result of recent improvements in the reporting and investigation of food-borne outbreaks. Seasonal evolution of human campylobacteriosis Similar to the data of the occurrence of salmonellosis (section 3.11), the incidence of campylobacteriosis in humans shows a seasonal variation. Isolates of Campylobacter from humans strongly increase in May and peak in early September (figure 3.2.2). In contrast with Salmonella however, the numbers of Campylobacter-positive poultry flocks at slaughter (caeca) show an explicit seasonal fluctuation. The testing of flocks is compulsory for the control programme in poultry meat. The secular trend in humans (down until 1999, up until 2001 and down again in 2002) is completely different from the gradually decreasing secular trend in poultry meat over the whole period (table 3.2.2). Table Campylobacter in poultry meat, fresh and chilled (Monitoring programme VWA/KvW) Poultry meat sample size at retail 1,325 1,314 1, ,578 1,600 Campylobacter spp. positive (%) Table Confirmed human cases of Campylobacter spp. and faeces tested for various reasons (Laboratory surveillance RIVM) Laboratory confirmed cases* 3,741 3,641 3,427 3,175 3,474 3,682 3,421 Campylobacter spp. cases / 100,000 inhabitants Faeces tested / 100,000 inhabitants 1,076 1,07 1,043 1, * First isolates from the 15 PHLs with an estimated coverage of 62% of the Dutch population Table Campylobacter-species distribution and travel, isolated from humans between 4/2000 and 3/2003 in public health laboratories, covering > 60% of the Dutch population (Laboratory surveillance RIVM) Casa- project Number of Travel- Species Travel- Domestic/ 4/2002-3/2003 patients indicated distribution related Unknown Laboratory registrations 3,176 11% C. jejuni 84% 90% Questionnaires received 1,434 22% C. coli 14% 7% C. spp., other/unknown 2% 3% 15

16 Conclusions Data from the public health laboratories indicate that the incidence of campylobacteriosis has changed little between 1996 and This is contrary to data seen in other industrialised countries in which the incidence has steadily increased (with the exception of the USA). Although data from other countries are difficult to compare, the incidence of campylobacteriosis in the Netherlands appears to be lower than in the surrounding countries. Data on campylobacteriosis obtained from two public health laboratories indicate that the incidence varies with place, time and age, the highest incidence being observed during the summer, in urbanised areas, in children below five and in young adults (particularly women). From 2004 onwards this information will also be available from the other public health laboratories. The reasons for the variability in incidence are not understood. As shown, food and probably the environment, including the domestic environment, undergo continuous contamination from the reservoirs of Campylobacter. These, including farm animals, wild animals and pets, create many pathways by which humans can come into contact with Campylobacter. The laboratory-driven case-control study (CaSa project) between April 2002 and April 2003 has been performed to indicate the relative importance of the transmission routes involved and to quantify the role of possible risk factors. This will be the input for the CARMA (Campylobacter Risk Management and Assessment) project, started in 2001 (frame 3).

17 Frame 2 Study into the similarity of Campylobacter isolates from humans and their pets Introduction From epidemiological studies, it is clear that direct contact with pet animals is a risk factor for human campylobacteriosis. There are some reports with strong indications that direct contact with young dogs or cats have led to human Campylobacter infections, and a few casuistic reports in which the correlation between infected pets and human illness is supported or proven by typing of the Campylobacter isolates. Between April 2002 and April 2003, a collaborative project between ASG Lelystad, RIVM and VMDC was performed to study the role of pets in human campylobacteriosis. Aim The aim of the study was to identify the similarity on the molecular level between Campylobacter isolates from diseased humans and Campylobacter isolates from their pets. Materials and methods The study was part of a gastroenteritis case-control study for the identification of risk factors carried out by RIVM. Gastroenteritis patients (proven infection by isolation of Campylobacter) received a questionnaire. If the patient had stated any contact with pet animals, and if they were willing to co-operate in further research, they were supplied with a short questionnaire about the health status of their animal(s), and with material to collect animal faeces. Faecal samples were sent by post to the diagnostic laboratory of VMDC. Campylobacter isolation was performed on CCDA plates incubated micro-aerobically for 2-4 days at 37 C and 42 C. Five separate colonies per positive pet sample and the connecting human isolate were speciated by a combination of PCRs and genotyped by Amplified Fragment Length Polymorphism (AFLP). Results and discussion Campylobacter was isolated from 19% of a total of 686 animal samples, resulting in 466 pet strains for comparison with 91 human strains. In one case (1%), the human and pet isolates showed an identical genotype and in 90 cases there was no similarity. From the 466 pet strains, 72 (15%) were typed as C. jejuni, 318 (68%) as C. helveticus/c. upsaliensis, 33 (7%) as C. hyointestinalis/c. fetus and 18 strains could not be typed. The percentage of positive pet samples (19%) is low compared to that reported in other studies. This may be due to the fact that the samples were sent by post; it took at least 24 hours before the samples were processed. The percentage rate of C. upsaliensis/c. jejuni is high compared to that reported in other studies. Conclusions In this study a similarity between human and pet animal isolates was found to be rare, suggesting that contact with pets does not often lead to direct transmission of Campylobacter. However, recent case studies concerning specific groups (young children and people with intestinal disorders) have described a severe clinical outcome of Campylobacter infection once this infection has occurred in these specific groups. Figure Seasonal and secular trends of the weekly occurrence of human cases of campylobacteriosis and the percentage of (caeca) positive flocks in the slaughterhouses (RIVM - PVE) Infected flocks, caeca slaughterhouse (%) First isolates, humans (15 PHLs) dec- mrt- jun sep- dec- mrt- jun sep- dec- mrt- jun sep- dec- mrt- jun- sep- dec- mrt- jun sep- dec- 0 % Positive flocks Humans 17

18 General approach of the CARMA project Autonomous developmensts Data - observational - experimental Risk assessment measures Intervention scenarios Epidemiological studies - human - animal number of infections measures Disease burden Cost of illness Supply chain costs Cost - effectiveness analysis Acceptability Risk management decision making process Political/culture limitations Frame 3 Campylobacter risk management and assessment General In 2001, a Campylobacter Risk Management and Assessment (CARMA) project was started. In this project, RIVM collaborates with CIDC, Agricultural Economics Institute (LEI), VWA/KvW and State Institute for Quality Control of Agricultural Products (RIKILT). The goal of the project is to be able to give advice on the effectiveness and efficiency of measures aimed at reducing campylobacteriosis in the Dutch population. This goal will be attained by providing scientific support to a risk management process as defined by the Codex Committee on Food Hygiene. Risk managers are the Ministries of VWS and LNV. Results up to and including 2002 The figure on this page shows the general approach of the CARMA project. In the first (completed) phase, all available information was collected and summarised in an extensive risk profile. This information is used to assess the relative contribution of different sources of contamination for the incidence of human infection. For all major routes of infection, a risk model will be built, starting with a model for the poultry meat chain. The model describes the transmission of thermophilic Campylobacter spp. on broiler farms, in slaughterhouses, in distribution and retail and in kitchens. The resulting exposure estimate will be combined with dose-response information to provide an estimate of the incidence of infection. In the disease-burden model, various outcomes related to infection are quantified: gastroenteritis, Guillain-Barré syndrome, reactive arthritis and mortality. The combined disease burden of these end-points is expressed in Disability Adjusted Life Years (DALYs). In the economic model, the costs of illness are estimated. Developments in 2003 and 2004 Current work focuses on developing transmission models for each stage in the food chain, and combining these with dose-response models. This work will continue in 2003, when it is also planned to estimate the disease burden and cost of illness. In consultation with risk managers and stakeholders, a series of possible interventions has been selected, including interventions at the farm, during slaughtering, and consumer education. According to current plans, these interventions will be implemented in the model in the year

19 2004. Autonomous developments (e.g., new legislation and trends in consumer preferences) will also be implemented in the model. Predicted changes in the incidence of illness lead to changes in disease burden, and in costs of illness. On the other hand, all interventions will lead to changes in the costs of producing poultry meat, and these will also be estimated in In the last phase of the project (2 nd half of 2004) all collected information will be combined in a cost-effectiveness analysis. Various interventions are compared on the basis of their (net) cost per DALY gained (or alternatively per life year gained, or per averted case of illness). Absolute effects, costs and cost-effectiveness ratios are important tools for supporting risk management decisions. However, other social and political factors also have a major influence on these decisions. Therefore, the project also aims at describing these factors and presenting them to decision-makers. A first series of analyses has already been completed and further work will focus on the support for selected interventions. There is a relationship with costs: measures with a low level of stakeholder acceptance will require larger efforts and thus costs for effective implementation. 3.3 Chlamydophila psittaci Introduction Psittacosis (parrot fever, ornithosis) is caused by Chlamydophila psittaci and this bacterium can be transmitted from (pet) birds to humans. Infections in humans usually occur when the organism is inhaled. The organism can be transmitted by means of aerosoles from dried faeces or respiratory secretions of infected birds. The infection can also occasionally be transmitted from mammals (sheep, goat, cattle) to humans by exposure to the birth fluids and placentas of infected animals. C. psittaci can cause influenza-like symptoms and other serious health problems among humans. Animals There is no systemic control programme to eradicate C. psittaci in the Netherlands. Therefore, it is not known what percentage of birds in the Netherlands is affected. At present infections are combatted on an ad-hoc basis. Humans When a positive human case of C. psittaci is confirmed in a laboratory, the result must, in accordance with legal commitments, be reported to GGD. Each GGD reports positive cases to IGZ where a record of these reports is kept. If GGD suspects that a bird or group of birds is the source of a human infection, they contact VWA. A veterinary inspector of the VWA visits this bird(s)location and evaluates the situation. If necessary, the owner of the animal(s) is advised to improve the hygiene situation. In all cases samples are also taken from the bird(s) for investigation. If a bird needs to be treated, the owner is advised to contact his own veterinarian. After consulting GGD, the veterinary inspector can also contact the human patient in order to inform that person about the infection risks (for both humans and animals) and preventive hygiene measures that can be taken. Conclusions Psittacosis is a serious contagious illness in the Netherlands. Especially for susceptible persons such as YOPI' (Young, Old, Pregnant and Immunocompromised) the course of the disease can be very serious or even fatal. There is no systemic control programme in the Netherlands. It seems to be useful to implement a monitoring programme to gain a clear insight into the number of infected birds. Investigation of imported birds from third countries that are kept in quarantine might be seen as an important component of such an investigation. A few years ago the taxonomic classification of Chlamydophila was changed. The family Chlamydiaceae (order Chlamydiales) was divided into two genera, Chlamydia and Chlamydophila. The genus Chlamydophila comprises the genera C. psittaci, C. pecorum, C. pneumoniae and three new species, which are derived from C. psittaci; C. caviae, C. felis and C. abortus. Bird infections appear to be common and latent infections in birds are normal. Signs of infections in birds include lethargy, ruffled feathers and anorexia. Infected birds excrete the organism in the faeces and respiratory secretions and they shed the organism intermittently. This shedding can be activated by stress factors. Humans with a psitaccosis infection usually have acute respiratory symptoms associated with a febrile illness. A headache, a non-productive cough and a pulse-temperature dissociation are also common symptoms. In addition to the respiratory tract, C. psittaci can also affect other organ systems and result in, for example, endocarditis, myocarditis, hepatitis and arthritis. In pregnant women, severe illness with respiratory failure, thrombocytopenia, hepatitis and fatal death has been reported. 19

20 3.4 Echinococcus spp. Introduction Echinococcus granulosus is the tapeworm of carnivores (dogs, wolves, jackals, foxes). The larval stage occurs in the form of smaller or larger cysts in sheep, cattle, horses, pigs, camels, goats and buffalos after ingestion of infectious eggs from the tapeworm. In humans a similar hydatid cyst may develop causing cystic echinococcosis. Alveolar echinococcosis characterized by tumour-like processes with multiple small cysts is caused by E. multilocularis, one of the most serious parasitic zoonoses. The parasitic cycle is mainly sylvatic (wildlife). In Europe, the red fox is the major definite canine host, but adult parasites can also occur in dogs and cats. Humans may get infected by oral intake of eggs shed by infected definite hosts. The disease is endemic in central Europe (Germany, Austria, Switzerland, France) but the parasite has now also been reported in other parts of Europe. Animals E. granulosus and E. multilocularis are both notifiable in animals. Prevalence Determination of the prevalence of E. multilocularis in end hosts is an important parameter for estimating the potential infection risk of humans in endemic areas. In the Netherlands, E. multilocularis was not found in foxes until In previous studies carried out between 1996 and 1998, E. multilocularis was detected in 2 distinct areas in Groningen and Limburg, close to the border with Germany. Based on the results of a study carried out in Groningen between , a baseline prevalence of 9.4% (95% CI %) was found. A similar study in Limburg is ongoing. In 2002, 25 foxes were tested, of which 3 were positive for E. multilocularis. These data indicate that E. multilocularis is present in the Netherlands and that in both Groningen and Limburg the prevalence in foxes is of such concern that periodical monitoring is appropriate. Surveillance Surveillance for E. granulosus is performed by meat inspection. Up to 1985, the percentage of slaughtered cattle with hydatid cysts varied between 0.65 and 0.95%. Since 1985 no official data have been available of the prevalence of hydatid cysts in slaughtered animals in the Netherlands. Meat inspectors report that hydatid cysts are very seldom observed. In 2002 a pilot study to detect and to identify E. granulosus genotypes in suspected cysts observed in the livers of cattle and sheep, using molecular identification methods, was conducted in 5 slaughterhouses. No sterile or fertile E. granulosus cysts were detected although the number of samples was limited. It is our aim to continue this study. Humans Prevalence Echinococcosis is not a notifiable disease in the Netherlands. Patients with hydatid echinococcosis, mostly originating from endemic countries (in the Mediterranean, South America, or Africa) occur annually. In 2002, of the 32 serologically positive patients, 29 patients originated from countries of those endemic for E. granulosus. The other 3 patients were Dutch. One of these patients most likely became infected in South America (Suriname); of the other 2 patients no more data were available. Only one patient with alveolar echinococcosis was diagnosed in the Netherlands. This patient was born and raised in Switzerland. Conclusions In all documented cases human hydatid disease in the Netherlands can be traced back as an imported infection. The prevalence of hydatid cysts in slaughter animals remains very low but further investigation is needed to determine the true prevalence of fertile and non-fertile cysts. It is concluded that the zoonotic importance of E. granulosus in the Netherlands is very low. Although autochtonous alveolar echinococcosis has not been diagnosed in the Netherlands, the prevalence of E. multilocularis in foxes remains of great concern. Due to the long incubation time of alveolar echinococcosis, human cases might only be seen after a delay of 5-10 years. Therefore, risk-based periodical monitoring will be performed in the border regions of the Netherlands. The risk related to the consumption of inadequately heated mushrooms, strawberries and comparable products (forest fruit) must be communicated to the public who visit areas with a relatively high prevalence of this organism. 3.5 Shiga toxin-producing Escherichia coli (STEC) Introduction Shiga toxin-producing Escherichia coli O157 (STEC O157), also called enterohaemorrhagic E. coli O157, was recognized as an important human pathogen in In the first half of the 1990s, attention for this pathogen increased due to several large outbreaks of STEC-related gastroenteritis and complications such as the haemolytic uraemic syndrome (HUS) in Japan, Scotland and the USA. In the Netherlands, so far only two small epidemiologically linked clusters of STEC have been observed: the first in June 1993, where four young children developed HUS and eight household members were infected. The second cluster in April 1998 involved a family living on a farm. Farm animals Results of the surveillance of zoonotic bacteria in farm animals The results of the surveillance in farm animals showed a significantly increasing trend towards E. coli O157 positive herds of veal calves in both periods, characterised by different examination procedures ( and ) (table 3.5.1). Remarkably, pink veal herds were more often found positive for E. coli O157 than white veal herds. Although the crude prevalence estimates for E. coli O157 in herds of dairy cattle also increased in subsequent years, a significantly increasing trend was not present. For both cattle species, a seasonal fluctuation was observed, in which the highest prevalence of positive herds was observed in July-November. This seasonality coincides with that of the incidence of STEC O157 infections in humans (figure 3.5).

21 Table E. coli O157 in farm animals at herd level in the Netherlands (Surveillance programme VWA/KvW - RIVM) Animal species Year Pos./total % 1 Broilers (random) / / /100 4 Layers (random) / / /113 1 Finishing pigs (random) 1997 n.i / /189 0 Dairy cattle (random) / /267 5; [5] /163 9; [11] /158 9; [11] / /148 14; [20] Veal calves (random) / /152 5; [7] /60 13; [12] /133 17; [17] / /159 24; [24] 1 [..] = Foot and Mouth disease adjusted prevalence for comparison with the estimate of n.i.= no investigation 3 different examination procedure of samples from this year onward Fifty-five of the 58 isolates examined from dairy cattle between 1998 and 2002 contained one or both of the stx genes; for veal calves this number was 80 of 84 isolates. Thus, the vast majority of isolates were STEC O157 and thereby potentially pathogenic to humans. STEC O157 isolation from faecal samples collected from layers, broilers and finishing pigs rarely occurred in the period 1997 to Dairy cattle and veal calves were the only farm animals surveyed systematically from 1997 up to and including Food In surveys carried out by VWA/KvW in 1999 and 2002 (table 3.5.2) no STEC O157 was detected in pork and beef sampled at retailers. Humans Results of the enhanced surveillance In the enhanced surveillance of STEC O157 in 2002, 49 cases (51% male) were identified, corresponding with 0.30 per 100,000 inhabitants (figure 3.5). In 1999, 2000, 2001 and 2002, there were 36, 43, 41 and 49 cases of STEC infections, respectively. The contribution of children in the age group 0-4 years increased from 24% in 2001 to 33% in 2002 and that from patients aged from 7% in 2001 to 16% in The distribution of O-, H- and stx-typing was similar to that in previous years: O157:H7, stx2 positive was prominent (50%) followed by O157:H, stx1 and stx2 positive. In 2002, one sorbitol-positive strain was isolated from a 2-year-old boy and fortunately no deaths were reported. Questionnaires on risk factors and clinical course were available for 45 (92%) of the cases in The development of HUS increased from 21% in 2001 to 27% in % of the patients were hospitalised. A known risk factor, such as contact with farm animals or manure, consumption of raw or undercooked beef, raw milk, raw-milk cheese or contact with a symptomatic individual was reported by 56% of the patients in Especially, contact with farm animals or manure seems to play an important role in transmission of STEC O157. In 2002, animal contact via manure as source was confirmed in one case by molecular typing (frame 4). Pulsed-field gel electrophoresis demonstrated that 30 of the 46 strains could be attributed to 12 PFGE clusters, of which 6 were known and 6 were new clusters. For 2 of the 6 new clusters the relationship between patients was known: one family cluster and 2 patients who consumed sausages from the same stall. For another 3 clusters dates of onset between the patients suggested a common source. However, this source could not be identified in additional interviews. Conclusions It is concluded that STEC O157 is a limited public health problem in the Netherlands. However, the incidence is underestimated due to the selective testing policy in Dutch laboratories and the use of less sensitive detection methods that also are unable to detect sorbitol-fermenting isolates. Furthermore, the recent shift among HUS patients in several European countries from infections with STEC O157 to infections with other STEC serogroups demands implementation of testing methods that are able to detect all STEC. Due to the epidemic potential and the seriousness of the illness, the occurrence of STEC should be carefully monitored. Table STEC O157 in pork and beef (Monitoring programme VWA/KvW) Type of meat Year Pos./total % Beef / /456 0 Pork / /

22 Frame 4 Escherichia coli O157 infections - direct and environment-mediated transmission Introduction In January 1999 an enhanced laboratory-based surveillance system for Shiga toxin-producing Escherichia coli O157 (STEC O157) infection was set up in the Netherlands (chapter 1). The results of that surveillance are reported in section 3.5. Of the known risk factors reported in this surveillance system, one of the most prominent was contact with farm animals or their faeces. Since the summer of 2000, field investigations are now done upon notification of recent animal contact of patients. Cases possibly linked to animals (or manure) In the period July 1999 up to the summer of 2003 a total of 27 cases of sporadic STEC O157 infections was linked to recent contact with animals or animal faeces. Seventeen of these cases were examined in more detail (table 4). Ten of these patients developed the haemolytic-uraemic syndrome (HUS), 6 had symptoms of bloody diarrhoea and the remaining patient had only watery diarrhoea. Eight field investigations resulted in the isolation of STEC O157. The bacteria were isolated from sheep, goats, a calf, cattle, fallow deer, a pony and a donkey. In 5 cases, the human and the respective animal isolates were indistinguishable by phagetyping and molecular subtyping, and therefore direct animal-to-human contact or contact with animal manure was the most likely cause of infection for these 5 patients. Although initial information pointed to animal contact, later it became clear that contact with an infected person during a visit abroad was most likely to be the cause of the illness in case IV. Similarly, cases X and XI appeared to be probably food borne and caused by the same type of sausage. Children's farms Visits to children's farm are popular leisure activities and also have become an important feature of education for young children. In the Netherlands, there are about 500 children's farms with a total number of about 15 to 20 million visits annually, mainly in family groups but also in prearranged school parties. Such visits are highly instructive to children in helping them to learn about aspects of animal husbandry and farm produce. Close contact with the animals such as petting and feeding is often encouraged. However, the above cases highlight the risk, especially to the main group of visitors, young children, of acquiring severe zoonotic infections during visits to children's farms. Cattle and other ruminants, such as sheep and goats, are important natural reservoirs of STEC O157. Animals carrying STEC O157 usually do not show clinical symptoms and shedding appears to be intermittent and transient. Although it is not possible to take specific structural preventive actions at present, the creation of a safe farm environment by the owner and the observance of simple hygiene procedures by visitors themselves are probably the most important preventive measures. Owners should provide good standards of hygiene, adequate toilet and hand-washing facilities, separate eating areas from animal contact areas, instructions in simple hygiene measures to visitors, and close supervision of visitors, particularly children. Communication In consultation with the Dutch Foundation for Children's farms a Code for owners and employees of children's farms has been written by VWA/KvW containing general advice on pathogens likely to be present in the farm environment and steps to be taken to minimize the risk of human infection. However, visitors are often not conscious of or disparage the possible risk of acquiring zoonotic infections through contact with farm animals. Therefore, more attention has been paid to informing the public about farm visits and zoonoses and about the risks and their responsibilities. Articles have been published in both newspapers and different popular magazines. Furthermore, a standard sign with instructions for visitors has been provided for children's farms. A similar strategy has been adopted towards dairy and beef farms by publication of articles in different agricultural journals. Figure 3.5 Reported cases of STEC O157, the monthly fraction of positive dairy cattle and veal calf flocks on farms (Laboratory surveillance RIVM) Infected Reported human STEC O157 cases Jan Mrt Dairy cattle (faeces) May July Sep Nov Jan Mrt Veal calves (faeces) May July Sep Nov Jan Mrt May July Sep Nov Jan Mrt May July Sep Nov 40% 35% 30% 25% 20% 15% 10% 5% 0% week of onset of disease Positive flocks

23 Microbiological investigations and inspections In the period June to November 2002, 132 children's farms were visited. By the use of a standard inspection list an impression of the general hygiene and the degree of implementation of the Code was obtained. Furthermore, fresh faecal droppings were collected and examined for STEC O157, Salmonella spp. and Campylobacter spp. It was concluded that both the hygiene standards at the farms visited and the facilities for hygiene practices by the employees as well as the visitors was good. However, more attention needs to be paid to facilities for hand-washing, cleaning footwear, an eating place for visitors and quarantine. Conclusions Eventually all of these investigations will contribute to a better understanding of the epidemiology of STEC O157, which will help in setting up an intervention strategy and a system to certify children's farms. Controlling the risk of STEC O157 infection will also minimize the risks from most other pathogens commonly present in animals and transmissible to humans by hand to mouth. Fresh faecal droppings collected at 13 (10.2%) of the children's farms visited were positive for STEC O157, 19 (14.5%) for Salmonella spp. and 74 (56.5%) for Campylobacter spp. The percentage of positive samples ranged from 4 to 60% (mean 26% and median 20%), 3 to 50% (mean 12% and median 7%) and 3 to 69% (mean 18% and median 13%), respectively. These numbers show that there is a real risk of acquiring severe zoonotic infections when visiting a children's farm and they underline the importance of the creation of a safe farm environment by the owner and the observance of simple hygiene procedures by visitors. In addition to this prevalence study, a longitudinal study on one children's farm previously found to be (repeatedly) positive for STEC O157 was carried out by VWA/KvW in Every week starting at week 8 up to week 47, faecal droppings and environmental samples were collected and examined for the presence of STEC O157. The goal of the study was to gain more insight into the epidemiology (source, transmission, survival, number of STEC O157 bacteria) of STEC O157 on children's farms. However, STEC O157 strains were not isolated from any of the 851 samples examined. 23

24 Table 4 Investigations of sporadic cases of STEC O157 infection linked with exposure to the farming environment (VWA/KvW - RIVM) Case Setting of animal contact STEC O157 isolated from Subtyping (faeces collected from) I Jul 00 Children's farm (sheep, goats, birds, pig) Sheep, goats + a II Oct 00 A swine as a pet at home (swine) - III Dec 00 Children's farm (sheep, goats, calf, pig, pony) Calf ND b IV Jan 01 Farm (handicapped persons) (dairy cattle, sheep, pony, horse, pigs, poultry, rabbit) - V May 01 Poultry farm (chickens, goats, rabbits, dog) - VI Jul 01 Dairy farm (family) (cattle) Dairy cattle + VII Jul 01 Children's farm (sheep, goats, fallow deer, pony/donkey) Fallow deer, pony/donkey + VIII Aug 01 Children's farm (goats/fallow deer, pony) Goats/fallow deer + IX Feb 02 Pet shop & at home (goats and swine & guinea pig and turtle) - X Mar 02 Riding school (handicapped persons) (horses) - XI Apr 02 Hobby farm (family) (cattle and horses) - XII Jun 02 Garden at home & paddock (goat manure & cattle) - XIII Jul 02 Farmyard campsite (cattle and sheep) Cattle - c XIV Aug 02 Day-care centre & hobby farm (family) (goats and swine & goats and ponies) - & Pony - XV Aug 02 Visit a friend at home (pony) - XVI Sep 02 Dairy farm (children s party visit) (cattle, sheep, pigs) Cattle, sheep + XVII Jun 03 Children's farm (goats, sheep, horses, calf, geese) - a b c +, no difference between phagetype, stx pattern and pulsed-field gel electrophoresis pattern of human and animal isolates ND, not done because the human isolate was not available for subtyping -, the human and animal isolates could be distinguished by subtyping Frame 5 Zoonotic pathogens on children's farms Introduction In 2000 and 2001 V&V conducted a few pilot studies aimed at a broad characterization of farms in the Netherlands that are open to the public (children's farms) and the identification and prioritization of veterinary public health risks that can be associated with these farms. Because of the special risk groups that are involved and the fact that some severe cases of infection of children with STEC O157 could be linked to visits to children s farms, special attention was being paid to this particular category. The findings of these studies formed partly the basis of the Code of good hygiene children's farms practices that was drawn up by VWA/KvW in November Results In the Netherlands there are 400 to 500 farms that officially present themselves as children's farm. The main purpose of these farms is to educate and entertain the public, mostly children, and not to produce animal products. The farms are usually owned by a foundation or a city council. In these cases the farms are primarily part of a therapeutic setting and they can play an important role as a meaningful workplace for the residents of these institutions. The highest densities of children's farms are found in the urban areas in the western part of the country. This part of the Netherlands comprises the major cities Amsterdam, Utrecht, Rotterdam, the Hague and their respective suburbs. Most children's farms in the Netherlands are relatively small, i.e., have a total area of less than one hectare (1 ha = 2.5 acre) and have high animal densities. Sometimes these densities approach those to be found at commercial intensive farms such as pig or poultry fattening farms. A collection of animals is by no means an exception. The most popular species that are kept are sheep, goats, poultry, ducks, rabbits and guinea pigs. Doves, swans, goose, tropical birds, donkeys, ponies, mice and rats are also very popular. Because of their size, horses and cattle are less popular animals. After the large classical swine fever outbreaks in the late 1990s most children's farms got rid of their pigs. The staff of the children's farms usually consists of one or two professional keepers who are assisted by several volunteers from the local community. The level of education, knowledge and motivation of the staff is therefore highly variable. Only a small minority of even the professional staff has had thorough training with regard to the management of children's farms and animal husbandry. It is therefore not surprising that, especially with regard to smaller or more fragile animals, such as small rodents and tropical birds, morbidity and mortality rates are high: often up to 50 and 25%, respectively.

25 Many of the prevalent diseases are a direct or indirect consequence of mistakes with regard to husbandry practices such as the choice of the feed and feeding practices or de-worming strategies. In addition, hygiene and biosecurity policies are almost always inconsistent and inadequate. Coupled with the high animal densities this leads to many opportunities for animal diseases and zoonoses to colonize these farms. Moreover, there is no consistent or permanent veterinary care at these farms that can be compared with the kind of care at commercial farms in the Netherlands. Altogether children's farms in the Netherlands are visited between 9 and 13 million times yearly with some hundreds of thousands of people visiting at least once a month. Several thousands even visit the farm at least once a day. The total number of people that actually generate these 9 to 13 million visits is estimated at about million. About 70% of these visitors belong to the classical risk groups of the YOPIs (Young, Old, Pregnant, Immunocompromised). A visit usually lasts between one and two hours. The young visitors engage themselves primarily with direct contact with the animals (cuddling, playing, grooming, feeding) and playing with other children. About 60% of the farms also sell agricultural products. The products involved are mostly eggs but sometimes also freshly harvested herbs, fruits, vegetables, garden plants, and cheese and in one case even meat. Thirty percent of the children s farms sell snacks and beverages such as candy, ice cream, pastry, soft drinks, coffee and tea. In many cases there are virtually no separate facilities for the sale and/or consumption of these products. In addition, hand-washing and toilet facilities are in general rather inadequate. Factual knowledge with regard to actual numbers of human or animal infections with zoonoses is still largely lacking. On the basis of the characteristics of the zoonotic agents potentially involved and the characterization of the visitors and their activities it was roughly estimated that the human infections with the highest incidence ( cases per million visits) would be salmonellosis, Q fever, cryptosporidiosis and fungal skin infections (e.g., trychophyton and microsporum). Also of importance (30-50 per million visits) would be psittacosis/ornithosis, echtyma (orf), campylobacteriosis, pasteurellosis and staphylococcosis. Only a minute fraction of these infections will be noted because only a fraction of those involved will seek medical attention. Furthermore, of all the relevant zoonotic agents the enteropathogens such as STEC O157, Salmonella and Campylobacter must be considered as the highest veterinary public health risk. That is to say, if a veterinary public health risk is defined as the product of the probability of an infection and the severity of the consequences of such an infection (Risk = P(inf) x Severity of disease ). Conclusions Data suitable for an exact quantification of veterinary public health risks at farms that are open to the public are virtually non-existent and make further research necessary. From an animal and public health perspective as well as an animal welfare perspective it is necessary to draw up uniform Codes of veterinary care in which particular attention is paid to the specific unique situation at children's farms. Furthermore, it is necessary that the professional staff of these farms receive uniform training that deals with these unique aspects. Based on the risk-assessment carried out, in which STEC O157 has been used as frame of reference (figure 5), it may be assumed that the control or elimination of the enteropathogens is the first priority with regard to public health at farms open to the public. Figure 5 STEC O157 related zoonotic risks on children's farms in the Netherlands (V&V) ,00 90 To STEC 0157 (100) related score ,00 18,75 15,00 6,67 4,38 STEC O157 Salmonellosis Campylobact. Psittacosis Q fever Cryptosp. Coli (ETEC) Toxoplsm. Paratuberc. Staphyloc Yersiniosis Larva migrans Pasteurellosis Fungal infect. Ecthyma Listeriosis Cat scratch dis. Caseous lymf. Leptospirosis Avian tuberc. Capnocytoph. 2,33 1,50 0,97 0,63 0,75 0,53 0,38 0,15 0,10 0,10 0,07 0,06 0,04 0,03 0,00 0,00 Bovine tuberc. 25

26 3.6 Hepatitis E virus Introduction Hepatitis E virus (HEV) is a non-enveloped, positive-sense, single-stranded RNA virus with icosahedral symmetry. Although it is related to the alpha-virus superfamily, HEV is classified as a separate hepatitis E-like virus genus. Infection in humans occurs in sporadic and epidemic forms and can cause an acute, self-limited, icteric hepatitis. HEV is endemic in much of the developing world. The most important route of transmission is faecal-orally, and disease outbreaks are often associated with contaminated drinking water or bad hygiene conditions. Recent studies suggest the existence of a virus reservoir in animals. Animals Prevalence Recent studies found IgG antibodies to HEV (anti-hev) in several wild and domestic animal species native to developing and developed countries. Molecular evidence for natural HEV infection in pigs has been reported for HEV endemic and non-endemic countries all over the world. Investigations and results To examine whether HEVs are commonly present in pigs in the Netherlands, pooled stool samples, collected from 115 pig farms were assayed by a reverse transcription-polymerase chain reaction (RT-PCR) amplification. HEV-RNA was detected by RT-PCR and hybridization, in 25 of the specimens (22%). RT-PCR amplification products of open reading frames 1 and 2 were sequenced and the obtained sequences were compared with published sequences of HEV genotypes from humans and pigs. HEVs from pigs in the Netherlands were clustered in at least two groups, together with European and American isolates from pigs as well as humans. These data show that HEVs in pigs in the Netherlands are genetically closely related with HEVs isolated from humans, although so far viruses identical to these pig HEVs have not been found in humans. While zoonotic transmission has not been proven, these findings suggest that pigs may be reservoir hosts of HEVs. Humans Prevalence In the industrialised countries of Europe, the prevalence of HEV in sera is rather low (1-3%) but in recent years there has been an increasing number of diagnoses of this infection in people who have not been abroad. In the Netherlands also, HEV infections have been determined in hepatitis patients who had not travelled abroad. The source of these infections is unclear and it is also not clear whether there may be an under-diagnosis of HEV infections in the Netherlands. In pregnant women the infection may cause serious complications, and can be fatal for mother and child. Investigations and results In 2001 a cluster of three cases of hepatitis associated with HEV was found in the northern part of the Netherlands. The typed strain from one case was related to HEV strains found in North America and Europe, and also related to a cluster of pig strains found in the Netherlands. These findings indicate that locally acquired HEV infections in industrialized countries may be overlooked. Routine testing for HEV infection in cases of acute hepatitis in the Netherlands should be considered before diagnosis of autoimmune hepatitis is made and steroid therapy initiated. Conclusions Further research will be needed to elucidate whether animals can be a source of infections in humans and if HEV can be transmitted by food and water in industrialized countries. More comprehensive testing of serum samples and faecal samples of patients with symptoms of acute viral hepatitis will be needed to gain more insight into the incidence, diagnosis, and transmission of HEV. 3.7 Influenza Introduction Influenza virus types A, B and C belong to the Orthomyxoviridae, a family of negative-stranded RNA viruses with a segmented genome. A key difference between them is their in vivo host range: whereas influenza viruses of types B and C are predominantly human pathogens that have also been isolated from seals and pigs respectively, influenza A viruses have been isolated from many species including humans, pigs, horses, marine mammals and a wide range of domestic and wild birds. It is generally accepted that in human influenza pandemics from recent centuries and numerous outbreaks in domestic and wild animals, interspecies transmission of avian influenza A viruses has played a crucial role. Migratory birds and waterfowl are thought to serve as the natural reservoir for influenza A viruses. Influenza A viruses representing 15 hemagglutinin (HA) and 9 neuraminidase (NA) subtypes have been described in wild birds and poultry throughout the world. Viruses belonging to antigenic subtypes H5 and H7 may become highly pathogenic upon transmission from wild birds to poultry and thus cause fowl plague, in contrast to viruses possessing other HA subtypes. Animals Wild birds After the influenza A virus zoonoses in Hong Kong in 1997, caused by highly pathogenic avian influenza (HPAI) viruses of subtype H5N1, the department of Virology at the Erasmus Medical Centre (EMC) in Rotterdam initiated a surveillance study in wild birds. This study was performed to gain information on the prevalence of influenza A viruses in birds in Northern Europe and to generate a panel of reference reagents (influenza A virus isolates and antisera). Cloacal swabs or fresh dropping samples were collected primarily from bird species known to harbour influenza A virus such as ducks, geese and gulls, and a wide variety of other birds species were also tested. Samples are collected by a large number of ornithologists and sent to our laboratory where they were tested using a highly specific and sensitive RT-PCR procedure. From RT-PCR positive samples, viruses were subsequently isolated in

27 embryonated chicken eggs and characterised using serological tests and genomic sequencing. The results of the surveillance in the year 2002 are summarised in the table below. Table 3.7 Results of the influenza A virus surveillance in wild birds in 2002 (EMC) Type of bird Species Pos./total % Ducks Mallards 200/2, Wigeons 1/ Others (7 species) 0/ Geese Barnacle geese 0/ Others (4 species) 0/6 0.0 Others Gulls, waders and songbirds 0/1, In 2002, virus isolates containing HA subtypes 1, 2, 4, 5 and 10 and NA subtypes 2, 5 and 6 were detected in mallards. A number of virus isolates have not yet been fully characterised. Sequence analysis of the HA gene of the H5N2 isolate revealed the absence of a basic cleavage site, indicating that it was not an HPAI virus. Our accumulated collection of influenza A viruses from Northern Europe now includes viruses containing all known HA subtypes, with the exception of subtypes 8, 9, 12, 14 and 15 and all 9 known NA subtypes. In addition, we have now provided conclusive evidence that viruses isolated from Black-headed gulls in 1999 represent a novel HA subtype (H16) and a novel NA subtype (N10). This collection of virus isolates was obtained from more than 15,000 samples collected from more than 250 bird species. The influenza A virus isolates collected within this surveillance network and the antisera raised against them are used as reference reagents in a number of laboratories. In addition, studies on the zoonotic potential of these viruses, pathogenesis, and virus evolution are currently in progress. Pigs Investigations for influenza infections in pigs are requested in the case of clinical respiratory problems. Mostly blood samples for serological investigation are sent to the GD laboratory. In a small number of occasions material is forwarded for virus isolation. For serological investigation the Haemagglutination Inhibition (HI) tests for H1N1, H3N2 and, since the last months of 2002, also for H1N2 were used. Poultry All breeding flocks of Gallus gallus which are in production are clinically inspected each month for List A diseases in accordance with Directive 90/539/EC, by the authorised practitioner. For the export of birds or hatching eggs some third countries demand serological investigation of the flocks of origin for Avian Influenza. Mostly only the agar gel precipitation test (AGPT) is requested, but sometimes also the HI test for H5 and/or H7 must be included. When the results of the AGPT were positive the HI test for H5 and H7 were used for confirmation. All 4,380 samples received in 2002 from 152 flocks were negative. Humans Influenza B virus and influenza A viruses of subtype H3N2 and H1N1, descendants of the 1968 Hong Kong pandemic and 1918 Spanish influenza pandemic respectively, are the main causes of influenza-like illnesses in humans. The World Health Organization coordinates a global influenza surveillance network, currently consisting of 112 national influenza centres (NIC) and 4 collaborating centres for reference and research. This network routinely characterizes the prevalence, antigenic properties and genetic properties of circulating influenza viruses. The combined antigenic, epidemiological, and genetic data are used to select strains for use in influenza vaccines. The NIC of the Netherlands is operated by the department of Virology at EMC and RIVM. Samples collected within a network of general practices (NIVEL) or sent in by hospitals, nursing homes, and individual general practitioners are analysed. In 2002, no influenza viruses from avian or pig origin were detected in humans in the Netherlands. Conclusions Influenza A viruses continue to circulate in humans, pigs and wild birds in the Netherlands. Viral zoonoses have not been detected recently. Only low pathogenic avian influenza A viruses were detected in wild birds, with a prevalence of about 10 % in mallard ducks. In poultry there was no evidence for infection with avian influenza A virus of subtype H5 and H7. By serology, influenza A viruses of subtypes H1N1, H3N2 and H1N2 were detected in pigs, but these viruses were not isolated in Influenza virus surveillance studies in humans, pigs, poultry and wild birds will be continued in Material was sent in for virus isolation from 5 pig herds. All samples were negative for influenza A virus. A total of 9,619 blood samples were received for serology from 724 herds; of these 3,046 samples were positive for H1N1 and 3,501 for H3N2. For H1N2 only 643 samples were tested from 43 herds and 64 were positive. 27

28 3.8 Listeria monocytogenes Introduction In the genus Listeria, human illness is predominantly caused by Listeria monocytogenes, especially serotypes 1/2a, 1/2b and 4b. This bacterium is found in humans and a variety of animal species, but is also widespread in the environment. Most human infections are caused by food (of animal origin). Food VWA/KvW monitors the occurrence of L. monocytogenes in foods at the retail level. According to the Dutch Commodity Act investigations must be performed on food prepared for direct consumption. The maximum level of contamination of food with L. monocytogenes is set at 100 bacteria per gram. The results of this monitoring as carried out in 2001 and 2002 are reported in table 3.8. Humans Fifteen regional public health laboratories weekly report first isolates of Listeria spp. Between the incidence of laboratory-confirmed infections with Listeria spp. was consistently low, on average 1 case per 0,5 million inhabitants per year, i.e., 31 cases per year for the whole of the Netherlands (figure 3.8.1). Cases were found more often amongst very young children, i.e., 1 case per 300,000 children 0-4 years old, especially babies, and older people, i.e., 1 case per 150,000 inhabitants 60 years and older. Up to 4 times as many cases were found amongst young adult women up to 35 years of age as amongst males of comparable age, but this concerns low numbers (figure 3.8.2). Between 1995 and 1998 findings were about 70% higher than in the following and preceding years. In 93% of the cases of listeriosis, L. monocytogenes was involved. Serotyping was reported in 16% of the isolates and showed no evidence of a shift in serotype distribution between , consisting predominantly of 4b (44%) followed by 1/2a (36%), 1/2c (7%) and 1/2b (13%). Between 1998 and 2002 the national reference laboratory for Listeria spp. at RIVM typed 147 L. monocytogenes isolates from patients all over the country (figure 3.8.1). In this selection serotype 4b also dominates (40%) followed by 1/2a (30%), 1/2b (24%) and 6% other types such as 1/2c, 3a, 3b and 4e % was isolated from blood and about 10% from liquor which might be considered as proxys of cases with septicaemia and meningitis, respectively. Conclusions It is concluded that listeriosis is a limited public health problem in the Netherlands with a low relatively stable incidence. However, due to the epidemic and transnational potential and the seriousness of the illness, the occurrence of Listeria should be more carefully monitored. No clear epidemiological information on related cases and their epidemiology exists in the Netherlands. From 2004 onwards, the surveillance of Listeria spp. will be intensified according to the model followed since 1999 for STEC, involving all Dutch microbiological laboratories, more extensive case finding, followed by a questionnaire on risk factors and course of the disease and PFGE typing of the isolates. Table 3.8 Foods at retail level containing >100 cfu/g L. monocytogenes (Monitoring programme VWA/KvW) Sample type Pos./total (%) Pos./total (%) Ready-to-eat meat (& products) Beef and veal 4/ /1, Poultry 1/ n.i.* Fermented sausages 1/ n.i.* Sliced meat products 12/1, / Other 11/1, Ready-to-eat dairy products Cheese (soft) 0/ / Cheese (hard) 0/ n.i.* Ice cream 0/ / Custard 0/ n.i.* Milk products 0/ Milk, raw 2/ Fish and fish products 1/ Smoked eel 0/ n.i.* Smoked salmon 0/ / Smoked mackerel 3/ / Herring 23/ n.i.* Shrimps 1/ / *n.i. = no investigation

29 Figure L. monocytogenes infections in humans reported by 15 PHLs and typed by the RIVM for all Dutch laboratories (Laboratory surveillance RIVM) Figure Age distribution of cases with a L. monocytogenes infection, , all data of projects pooled (Laboratory surveillance RIVM) Number per year PHLs Typed RIVM Number Males Females 3.9 Mycobacterium spp. Introduction Mycobacterium bovis is the causative agent of tuberculosis in cattle and a wide range of other animals. In a large number of countries reservoirs of M. bovis have been found in wildlife species such as badgers, brush-tailed possum, elk and deer. Fortunately M. bovis infections have never been diagnosed in wildlife in the Netherlands. The Dutch cattle population is virtually free of M. bovis. Occasionally M. bovis and M. tuberculosis infections are diagnosed in zoo or companion animals. Human tuberculosis is mainly caused by M. tuberculosis. With the aid of DNA fingerprinting it has recently been proven that M. microti can also cause severe forms of tuberculosis in man. M. tuberculosis, M. bovis and M. microti belong to the so-called M. tuberculosis complex. Tuberculosis in birds is mainly caused by M. avium, which is also related to caseous lymphadenitis in pigs and children. In immunocompromised individuals, such as patients with AIDS, M. avium can be the cause of very severe forms of mycobacteriosis. Paratuberculosis in cattle and goats is caused by M. paratuberculosis. The possible role of M. paratuberculosis in the aetiology of Crohne s disease remains a subject for ongoing debate and research. Animals General The Dutch cattle population has been officially declared free of tuberculosis since 1992, a situation that was ratified by the decision of the European Commission 95/138/EC. However, M. bovis still occasionally occurs in humans, cattle and other animals. To protect visitors and personnel at zoos, susceptible animals are intensively screened for tuberculosis. Most commercial poultry is kept indoors and also because of the all-in/all-out regime and good hygiene procedures, avian tuberculosis has disappeared from commercial poultry. As far as birds are concerned, M. avium is nowadays only seen in wildlife and companion animals. The incidence of caseous lymphadenitis in slaughter pigs is rather high in comparison with countries such as Denmark. In a surveillance programme performed in 1998 M. avium was isolated from more than 50% of mandibular and mesenteric lymphnodes with caseous lesions. Slaughterhouse companies together with the CIDC and the VWA/RVV are developing plans to reduce the incidence. With the help of voluntary control programmes for paratuberculosis, cattle farmers have the opportunity to free their infected herds of the disease. 29

30 Surveillance The surveillance of tuberculosis in farm animals depends mainly on official post-mortem examination of all animals slaughtered and on post-mortem investigation of animals with suspected lesions. Due to the high speed of the slaughter process and the fact that no palpation is carried out, the sensitivity of the detection method is relatively low. The sensitivity of the monitoring is separately discussed in frame 6. In addition to post-mortem examination, measures concerning tuberculosis consist of tracing and screening of contact herds (when bovine tuberculosis is confirmed by culture), tuberculin testing of cattle as part of the accreditation procedure for cattle sperm centres (Directive 88/407/EC) and examination of cattle before export to non-member states. Zoos in the Netherlands have a specific monitoring system to avoid tuberculosis in their animals. Zoo animals that have died are sent to various institutes for post-mortem or the veterinarian from the Zoo does the post-mortem examination. Animals suspected of tuberculosis that have died are sent to the Dutch reference laboratory CIDC at Lelystad. If a positive case is diagnosed all contact animals are screened. Slaughter pigs are checked for caseous lymphadenitis by slicing both mandibular lymph nodes at post-mortem inspection. The head, stomach and intestines of animals with suspect lesions are destroyed. Results of the investigations in 2002 M. bovis Two imported cows showed suspect lesions at post-mortem inspection at the slaughterhouse (table 3.9). One cow was imported at the age of one and a half years old and fattened for another year at a farm in the Netherlands. The other case was a veal calf imported as a newborn. In both cases the diagnosis was confirmed by CIDC by culture. The tracing resulted in 36 contact farms. All PPD investigations carried out at these contact farms appeared to be negative. From 3 positive herds in other member states, 16 contact herds in the Netherlands were traced. The PPD investigations carried out at these contact herds appeared also negative. Three animals that were exported to other member states appeared to be PPD-positive. The herds of origin were declared suspect for tuberculosis. However, none of them tested positive. In conclusion in 2002 a total of 57 herds were declared suspect and their official tuberculosis free (OFT) status was suspended. At the end of 2002, 34 herds regained their OFT status. M. tuberculosis Tuberculosis due to M. tuberculosis was diagnosed in a group of 10 bonobos in a zoo. In a short period 1 animal died and 2 animals were euthanised, being clinically ill. At post-mortem examination tuberculosis was suspected and this pre-emptive diagnosis was confirmed by culture. PPD investigation of the remaining 7 animals resulted in 2 positive results. One of the animals was also found to be positive in plain thorax X-rays. This animal was euthanised and was positive at post-mortem by culture as well. The remaining 6 bonobos were treated following the 'human protocol' with two antibiotics for six months. Intake of medication was observed carefully and registered. In the meantime 2 bonobos gave birth. After treatment the 8 monkeys will be held under close observation for a long period. The health-control programme for the other primates in this zoo has been intensified. As a result of that 234 PPD tests were carried out; each of them was found negative. M. avium The prevalence of lymphadenitis registered in slaughter pigs at 3 big slaughterhouses in the south-eastern part of the Netherlands is still fluctuating around 0.5%. Humans In recent decades, less than 1% of the tuberculosis cases diagnosed in the Netherlands were caused by M. bovis. In 1999, 19 human M. bovis cases were reported. In 2000 and 2001 the numbers of registered cases were even less, namely 13 and 5, respectively. In 2002 tuberculosis caused by M. bovis was registered 8 times. The age distribution of the patients with M. bovis infections is not even. Most cases are found among Dutch patients 65 years and older, and fewer in the age group 45-64, and are clearly caused by endogenous reactivation from latent infections. M. bovis infections among Dutch patients in the age group < 45 years are highly infrequent. Most cases of M. bovis infections among foreign-born patients are found in the age category from 15-64, and it is conceivable that these infections are contracted outside the Netherlands. Table 3.9 Tuberculosis (M. bovis) in farm animals in 2002 (VWA/RVV - CIDC) Activity Type of animal Indiv. animals Pos./total % M. bovis/ % or herds positive Meat Inspection Cattle Animals 2/1,874,066 < 0.1 2/2 100 Export Cattle Animals 0/3, /0 0 Sperm centres Cattle Animals 0/6, /0 0 Tracing Cattle Herds 2/57 0/0 0 Meat inspection Pigs Animals 0/15,413, /0 0 Meat inspection Sheep Animals 0/439, /0 0 Meat inspection Goats Animals 0/10, /0 0

31 As a result of the earlier mentioned M. tuberculosis infection of several zoo primates, all employees that had contact with the bonobos (8 persons) have been checked for tuberculosis. Seven persons were PPD positive and X-ray examination showed small abnormalities at one of them. All positive human contacts were treated and will be clinically followed by GGD for a longer period. Conclusions Although tuberculosis in humans derived from animal sources now plays a minor role compared with tuberculosis caused by transmission among humans, it is still highly important to monitor these infections in animals. M. bovis infections in man are old latent infections in elderly people or infections contracted outside the country by young people. Therefore M. bovis infections occasionally detected in animals in the Netherlands are of relatively low zoonotic importance. As M. tuberculosis infections occasionally pass from monkeys, elephants or parrots to man, an intensive surveillance system for these animals must be maintained. Especially the animals that are in close contact to man must be tested periodically for the presence of tuberculosis. M. avium infections in pigs can be a source for infection in man. More research in this field is needed and will be carried out in due time. Frame 6 An epidemiological and economic evaluation of detection methods for bovine tuberculosis Introduction For some zoonotic agents, monitoring or surveillance data are only used for an estimate of the prevalence in feed, animals and food (i.e., Campylobacter spp.). For other agents, e.g., mycobacteriaceae, positive results must be followed immediately by control measures (risk management). The performance of these monitoring systems in terms of possibilities for risk reduction largely depends on the sensitivity of the system. One hundred percent sensitivity may result in a virtually zero risk, assuming that appropriate action is undertaken. However the costs will be considerable. In contrast, a relatively low sensitivity of the monitoring system, accompanied by lower initial costs, may lead to severe and/or costly consequences when infections are diagnosed rather late. A large financial burden can be related to a greater spread of an infection and greater severity of longer lasting infections that could have been detected and controlled at an earlier stage with a more sensitive monitoring system. Animal health and also public health is at stake when decisions are made on the type and sensitivity of the monitoring system. Moreover, decisions must be based on dependable data that, at least partially, can be derived from a model. For many zoonotic agents, such a model will have epidemiological and economic components and, whenever relevant, an estimate of the human disease burden. In 2001/2002, researchers of CIDC, in association with Wageningen University, evaluated several possible monitoring systems for the detection of bovine tuberculosis. In this study single and combined detection methods used in EU Member States as well as newly developed methods were taken into account: 1. Inspection at slaughter (the Netherlands). 2. Combination of simple and compared tuberculination (the current surveillance method in many EU Member States). 3. Compared tuberculination (the current method in Ireland). 4. Gamma-interferon test (80% sensitivity and 99.8% specificity). 5. ELISA on blood samples at slaughter (50% sensitivity for infectious animals and 99.9% specificity). 6. ELISA on tank milk. The definitive diagnosis of bovine tuberculosis is always based on isolation and cultivation of the agent. Other methods however can be used in a more or less sensitive monitoring system as early warning. These methods were evaluated using an epidemiological model and an economic model. In the epidemiological model intra-farm and inter-farm transmission were calculated and the sensitivity and specificity of each detection method are set up in the report. The different stages of an infected and infectious animal are taken into account when statements are made on the performance of the method. With regard to the Inspection at slaughter it is assumed that only animals in an infectious stage will be detected by visual inspection. Based on literature and a small amount of data collected in the Netherlands, a sensitivity of this detection method of 10% is used in the model. The economic model was fed with data from the epidemiological model and pays special attention to the estimated total costs of both single and combined methods, in relation to the risk and extent of a possible outbreak. Based on data of the surveillance of M. bovis in humans it was decided that a special module of the disease burden in man would not have additional value. Conclusions The following conclusions can be drawn from the results: The estimated, quantified costs of inspection at slaughter as a single method are relatively low. However the risk of spreading the infection is substantially greater compared to a combination with other methods. This is due to the fact that an existing infection will be recognised at a later stage, resulting in an increased intra-farm and inter-farm transmission. Inspection at slaughter can best be combined with one of the tuberculination procedures or with the gamma-interferon test. Such a combination results in a substantial increase of the estimated costs and also a substantial decrease in the number of animals in which infections are not noticed. Nevertheless this combination of methods is described in the report as optimal. In the case of an outbreak and depending on the extent, supplemental diagnostic procedures such as the gamma-interferon test may be considered. The relationship between risk reduction and costs is not linear. A reduction of risk by 50% results in a more than two-fold increase in the accompanying costs. Future data derived from an eventual outbreak have to be incorporated in the model to retest its applicability. 31

32 Risk managers must decide, considering the quantified risks and costs given by the model, which system should be implemented. It is concluded that under the Dutch circumstances the inspection at slaughter as a single method with a relatively low sensitivity has nevertheless sufficient performance resulting in an acceptable risk compared to the associated costs. A prerequisite is that a positive diagnosis is followed immediately by the correct control measures. Continuous education of meat inspectors so that they can recognise suspect lesions remains essential. The cornerstone of the monitoring system for bovine tuberculosis in Dutch cattle by the single detection method of inspection at slaughter will therefore remain in place for the foreseeable future. Whenever circumstances make it necessary, the monitoring system will be reconsidered. Frame 7 Mycobacterium avium subsp. paratuberculosis Introduction As in many other countries Mycobacterium avium subsp. paratuberculosis (Map) is present in the Dutch cattle population. The agent is of economic importance because the resulting infection (Johne s disease) causes moderate to severe gastroenteritis with weight loss and profuse diarrhoea as prominent clinical symptoms. Impairment of production causes considerable financial losses for Dutch dairy farmers. In 2000 the Paratuberculosis Program Netherlands (PPN) was started in which the Ministry of LNV and industry participate in a co-ordinating platform focussing on reduction of the prevalence of Map in Dutch cattle. GD is responsible for the implementation of the different components of the programme. Until now there has been no conclusive evidence that Map is transmissible to man. Its possible role in the aetiology of Crohn s disease is the subject of continuous debate. Research on the eventual zoonotic properties of Map is therefore continuing in many institutes, including RIVM. In the meantime it cannot be excluded that monitoring and control programmes, mainly focussed on the improvement of animal health and reduction of economic losses, could also play a role in the reduction of the risk for public health. It is therefore worthwhile to discuss the programme in the framework of this report. Despite the activities set up by PPN, meat, milk and dairy products coming also from Map-positive farms are sold at retail. It is assumed that the prescribed pasteurisation of milk is sufficient to kill Map. However cross-contamination and insufficient heating of raw meat as well as products made from raw milk could be related to a certain risk for public health. As stated before the choice for a special monitoring and control system in animals and food must be based upon thorough cost-benefit analysis, also taking into account the risk factors for animal and human health. In 2004 a desk study will be presented that will provide us with state-of-the-art information on the possible public health risk in relation to Map in dairy cattle and products thereof. For now the control of Map in Dutch cattle is based on the two voluntary programmes that have been instituted since 2000 by PNN, the so-called Paraplanner and the Intensive programme. The key issue in the new approach of these programmes is prevention, based on the results of an epidemiology model and a cost-benefit study by the Wageningen University. The programmes consist of a step-by-step approach and follow each other according to the prevalence of Map in a herd. The Paraplanner focusses on measures to prevent and reduce transmission from dam to calf and young stock in all national herds. The spin-off of the general hygiene measures of the programme will also lead to a reduction of other infectious agents transmittable by faeces. The goal of the Paraplanner is to reduce the number of infected animals in a herd to a certain, preset level. When this level of reduction has been achieved, the Intensive programme is started to reduce the infection level of these herds to zero. Additional testing and culling of infected individual animals are important elements of the Intensive programme. All animals 3 years and older from clinically negative herds are serologically examined using an ELISA test. If all tests are negative the herd obtains the lowest suspected status. Each year faeces samples of all animals 2 years and older from these herds are examined by culture of 5 pooled samples. If these cultures stay negative each year the status will improve. The free status is gained when all cultures have been negative on four occasions. This intensive programme is expensive due to the high costs of the test methods involved and the necessary culling of positive animals. Results At the end of 2002 a total of 1,140 herds were participating in the voluntary Paraplanner. Of the 1,428 dairy herds that participated in the Intensive programme, 76% obtained the status unsuspected at different levels Rabies virus Introduction The rabies virus belongs to the Lyssavirus of the Rhabdoviridae family. In Europe, two types of this virus are important: the classic virus (street virus) with the fox as the most important reservoir and the European Bat Lyssa (EBL) viruses. The EBL virus can be subdivided into several subtypes. The EBL-1 and EBL-2 are classified genotypes 5 and 6 of the genus Lyssavirus. Animals with symptoms of classic rabies have not been reported in the Netherlands for many years. The EBL-1 virus on the other hand is endemic in the Serotine bat (Eptesicus serotinus), one of the most common free living insectivorous bats in the Netherlands. European Bat Lyssavirus infections in humans have been reported and have resulted in fatal rabies. Therefore, bats involved in direct-contact incidents including biting incidents (bats->humans and bats<->companion animals) are examined in the Netherlands for the presence of EBL virus.

33 As most of the human rabies cases worldwide are caused by the street virus, all other animals with abnormal behaviour in the Netherlands are, as a precaution, also examined for the presence of rabies (street virus). Animals In the Netherlands wild animals, farm animals and pets displaying abnormal behaviour (inexplicable aggression) are considered to be suspected cases of rabies. Investigation of these animals takes place at CIDC. The numbers of animals examined in 2001 and 2002 are summarized in table Table Animals examined for rabies in 2001 and 2002 (VWA/KvW - CIDC) Animal species Pos./total % Pos./total % Bats (wildlife) 9/ / Foxes 0/ / Dogs 0/ /3 0.0 Cats 0/ /4 0.0 Deer 0/ /0 0.0 Muskrats 0/ /0 0.0 Ferrets 0/ /1 0.0 Bats (held in captivity) 0/ /1 0.0 Total 9/ / All of the animals diagnosed as rabies-positive during that period were bats and all those Lyssavirus-positive bats were identified as Serotine bats. The incidences of Lyssavirus in the examined Serotine bats in 2001 and 2002 were 6.9% (9/131) and 2.5% (3/121), respectively. PCR characterization of EBL strains Laboratory research at RIVM for the sequences of the RT-PCR amplified products of nucleoprotein regions of EBL-1 strains detected in the Netherlands in 2001 resulted in % homologies with EBL-1 strains isolated in Europe over the last ten years. EBL-2 sequences have not been detected in bats in the Netherlands since Humans People who have been in contact with animals diagnosed as rabid receive post-exposure vaccinations, if possible combined with the application of anti-rabies immunoglobulin (MARIG). This treatment is also applied when a suspected animal cannot be investigated. The treatment can be carried out by general practitioners, who can obtain information about the treatment and its necessity from the National Poisons Control Centre (NVIC) of RIVM. No human rabies cases were reported in 2001 and In 2001 information about 25 cases of post-exposure treatment was provided by NVIC and the majority of these cases (16) were related to contacts with bats. The other cases were related to contact with other animals (table ). In 22 cases post-exposure treatments were carried out. Vaccines were applied for 3 persons; 19 persons were treated with vaccine and MARIG. In 2002 treatment information was provided 110 times. Vaccines were given to 16 persons and 21 persons were treated with both vaccine and MARIG. As in 2001 most of the information provided was related to contacts with bats (90). Conclusions Rabies caused by the classic (street) virus has not been reported for many years in the Netherlands, whereas EBL virus is endemic in bats. Table Information supplied about post-exposure treatments, number of exposures and animals involved (NVIC, RIVM) Animal species Exposure Vaccines only Vaccines Exposure Vaccines only Vaccines and MARIG and MARIG Bats Dogs Cats Rats Polecats Deer Mice Foxes Squirrels Ferrets Rabbits Llamas Total

34 The results of the investigation of animals in 2001 and 2002 reflect that status and these findings correspond with those from previous years (report 2001), in which only bats were also found to be rabies-positive. Furthermore the results of the identification of the bats show that the Serotine bat is still the main reservoir of Lyssavirus in the Netherlands. The investigation of the virus strains detected in those animals in 2001 has made clear that only the EBL-1 virus is present in those bats. As the risk of introduction of the street virus is still present and both the street virus and the EBL virus may cause fatal rabies in humans, it is important that investigation of rabies-suspected animals is continued in the Netherlands Salmonella spp. Introduction Salmonella has been recognised as an important pathogen for many years. In the Netherlands between 1984 and 2002 about 1,000 serotypes and phagetypes have been found. The pathogen has drawn a lot of attention and much progress has been made in the control of the pathogen in feed, animal husbandry and at slaughter and has also led to improvements to the HACCP standards in the production of foods and to ongoing education of the public for improvement of food handling and kitchen hygiene. Recent epidemiological studies performed in the Netherlands estimate that in 1999 about 55,000 cases of salmonellosis occurred in the general population, of whom about 9,000 consulted a general practitioner leading to an estimated economic cost of million Euro (not counting the chronic sequelae of the disease). Feed Results of the monitoring programme in feed ingredients and end products (compound feed) Table shows that the percentage of positive Salmonella samples from imported by-products of rape seed was high (12%) in both 2001 and Imported by-products of sunflower-seed, toasted soya beans and fish meal were also regularly contaminated with Salmonella (4.4%, 2.4% and 3.5% in 2002, respectively). The rape seed was mainly of German origin (12.9% positive in 2002) and the fish meal was mainly of South American origin (4.2% positive in 2002). In 2002, Brazilian extracted soybeans and expeller had a Salmonella contamination of 3.2%, which is higher than the total contamination of extracted soybeans (2.1%). Through additional treatment the level of contamination of the end products is much lower than found in the feed ingredients. Progress has been made between 1999 and 2002 in reduction of the contamination of poultry feed (table ). In 2002, only 0.3% of the samples of the end products (i.e., the compound feed) for poultry feed were positive for Salmonella. The percentage of positive Salmonella samples of the end products for pig feed and cattle feed was less than 1% (0.6% and 0.9%, respectively). When Salmonella is isolated from feed materials and compound feed, in most cases the isolates are further classified by serotyping. The purpose of this classification is to establish more accurately any relationship among Salmonella serotypes in feed materials, the compound feed produced from them, animals and animal products. It is an aid in determining the possible cause of Salmonella contamination in subsequent links in the production chain. The classification of Salmonella types for 2002 is shown in the the tables on the next page. Table Percentage of Salmonella positive samples of feed ingredients (Monitoring programme KDD) Feed ingredients Pos./total % Pos./total % Barley and by-products 0/ / Wheat and by-products 5/ /1, Maize and by-products 0/ / Ground-nut by-products 0/ / Extracted rape seed and expeller 1 27/1, /3, Palm kernel and by-products 0/ / Toasted soybeans n.i.* n.i.* 14/ Extracted soybeans and expeller 22/1, /1, Extracted sunflower seed 9/ / Linseed flakes 3/ / Fishmeal n.i.* n.i.* 28/ *n.i. = no investigation

35 Table Percentage of Salmonella positive samples of compound feed, taken under the national programme (Monitoring programme KDD) Compound feed Pos./total % Pos./total % Pos./total % Pos./total % Poultry feed (total) 67/6, /6, /9, /7, Top breeding 4/ / /1, / Raising parent stock 6/ / / / Parent stock 12/1, /1, /1, /1, Laying hens 33/2, /2, /2, /2, Consumption turkeys 0/ / / / Broilers 15/2, /1, /2, /2, Pig feed 17/2, /2, /3, /3, Cattle feed 17/2, /2, /3, /1, Table Classification of Salmonella serotypes in feed ingredients in 2002 (Monitoring programme KDD) Feed ingredient No. of positives Salmonella serotypes Extracted rape seed and expeller x Agona; 29x Mbandaka; 41x Lexington; 16x Senftenberg; 8x Kentucky; 8x Havanna; 4x Oranienberg; 3x Rissen; 3x Anatum; 3x Altona; 2x Montevideo; 2x Indiana; 1x Llandoff; 1x Livingstone; 1x Cubana and 257x unknown. Extracted sunflower seed 17 11x Lexington; 4x Mbandaka; 1x Cubana and 1x unknown. Fishmeal 28 7x Cerro; 3x Ohio; 2x Havanna; 2x Montevideo; 1x Livingstone; 1x Tennessee; 2x Senftenberg; 2x Infantis; 1x Mbandaka; 1x Anatum; 1x Oranienburg; 1x Agona; 1x virchow; 1x Orion and 2x unknown. The main causes of contamination of the end products are inadequate processing of the (positive) feed components and/or post-process contamination (cross- and recontamination). Corrective actions undertaken by the feed industry include: repelletizing of the feed (if positive) at higher process temperatures, acidification of the feed with (for example) formic acid, intensified cleaning and disinfection of the plant and processing equipment, feedback of results of the monitoring regarding Salmonella to their suppliers of the feed materials and more intensive monitoring of the Salmonella status of the (critical) feed materials delivered at the compound feed plant. Considering that in the percentage of Salmonella positive compound feed for pigs and poultry was (much) higher than in 2002, it may be concluded that the industry has again steadily improved its performance. 35

36 Farm animals Results of the surveillance of zoonotic bacteria in farm animals The results of the surveillance in farm animals showed a decrease in Salmonella occurrence over the period 1998 to 2002 (table ) in broilers. The increase in 2001 was not statistically significant. An increase in S. Paratyphi B var. Java was measured in this period for broilers, contrasted by a decrease in S. Enteritidis. For layers, a decrease in the percentage of positive flocks was observed during the period A similar fraction of S. Enteritidis of all Salmonella isolates found was observed over the years, with Pt-4 dominating other S. Enteritidis phagetypes. Dairy cattle and veal calves were not massively colonised by Salmonella spp.. Faecal samples from finishing pigs were frequently contaminated with salmonellae, but a possible decreasing trend was observed between 2000 and The main isolated serotype was S. Typhimurium, with an increasing role for the Dutch phagetypes ft-401 and ft-506 (corresponding to the English phagetype DT-104) in the past few years. Results of the control programmes in the poultry meat production chain and the egg sector In 2000, the target of the original control programme in the egg sector of less than 5% infection with S. Enteritidis/S. Typhimurium of the layer production flocks was not reached. In 2000, the percentage S. Enteritidis/S. Typhimurium contamination of layer production flocks was, based on blood sampling, estimated at 11.8% (table ). In the layer rearing chain, however, the percentage S. Enteritidis/S. Typhimurium contaminated flocks was very low, 0.4%. After these findings, the control programme was adjusted in Table Classification of Salmonella serotypes in compound feed in 2002 (Monitoring programme KDD) Compound feed No. of positives Salmonella serotypes Poultry feed 18 2x Agona; 1x Lexington; 1x Senftenberg; 3x Mbandaka; 1x Hadar; 1x Ibadan, 1x Indiana; 1x Virchow; 1x Infantis;1x Montevideo; 1x Stanley, 1x Anatum, 1x Anatum and 3x unknown. Pig feed 18 2x Worthington; 1x Mbandaka; 1x Bareilly; 1x Carrau and 11x unknown. Cattle feed 14 3x Panama; 1x Livingstone; 1x Kentucky; 1x Montevideo and 8x unknown.

37 Table Salmonella spp. in farm animals in the Netherlands (Surveillance programme VWA/KvW - RIVM) RV medium RV + DIA/MSRV medium Animal species Year Pos./total % 1 Pos./total % 1 Broilers /63 25; (4) n.i. 2 (random) / /192 28; [29]; (8) / /154 21; [23]; (13) / /128 17; [20]; (38) / /123 23; (18) / /161 11; [13]; (33) Layers / (18) n.i. 2 (random) / /207 24; [24]; (62) / /168 23; [23]; (43) / /166 20; [20]; (46) / /84 14; (33) / /134 13; [14]; (44) Dairy cattle /82 0 n.i. 2 (random) /263 3; [2] n.i /169 2; [2] n.i /156 1; [1] n.i / / /148 3; [2] 8/148 5; [5] Veal calves /114 3 n.i. 2 (random) /148 1; [0] n.i /60 5; [4] n.i /132 1; [1] n.i /89 6 9/ /155 4; [6] 9/155 6; [7] Finishing pigs 1997 n.i. 2 n.i. 2 (random) / ; (7) n.i /189 13; [12]; (8) n.i /194 15; [15] 68/194 35; [34]; (10) / /154 29; (20) /157 20; [19] 47/157 30; [26]; (19) [..] = Foot and Mouth disease adjusted prevalence for comparison with the estimate of 2001, (..) = % of positive flocks/herds for S. Paratyphi B var. Java (broilers), S. Enteritidis (layers) or S. Typhimurium DT-104 (finishing pigs) n.i.= no investigation samples taken in the fourth quarter only 37

38 Table Number of broiler breed and layer flocks examined and found positive for S. Enteritidis and S. Typhimurium (Control programme PVE) Poultry meat Egg sector Year Type of flocks Flocks S. E S.T Year Type of flocks Flocks S. E S.T examined (%) (%) examined (%) (%) 1999 Elite broiler breed: 1999 Elite layer breed: rearing and production rearing and production * Broiler breed: rearing Layer: rearing 1, , , , Broiler breed: production Layer: production 1, , , , * One flock positive for S. Enteritidis as well as for S. Typhimurium Table Salmonella serotypes other than S. Enteritidis and S. Typhimurium found in positive broiler breed, (elite) layer breed flocks in 2002, both during rearing and production (Control programme PVE) Control programme Poultry meat Egg sector Period Broiler breed (faeces of >700 flocks) Elite layer breed (fluff, 38 flocks) Layer breed (cloaca, >2700) January-March x Braenderup; 3x Heidelberg; 2x Virchow 0 1x Infantis; 2x Kottbus April-June x Braenderup; 1x Infantis; 1x Mbandaka 2x Virchow 0 July-September x Heidelberg; 1x Indiana 12x Virchow 0 October-December x Heidelberg; 2x Paratyphi B var. Java 0 1x P. B var Java

39 The most important alterations concern an obligation to trace-back a S. Enteritidis/S. Typhimurium infection, as in the poultry meat sector. The objective in the control programme in the egg sector was sharpened to 0+ % S. Enteritidis/S. Typhimurium infected eggs in the consumption channel. The table shows that in 2002 this target was not yet reached by far. An essential change in the control programme poultry meat and egg sector concerns the start of tracing experiments in the poultry chain. In case of a Salmonella infection in a flock, the Salmonella serotype that caused the infection must always be identified (table ). Table Percentage of Salmonella positive broiler flocks/batches, from different links in the chain (Control programme PVE) Matrix link Fluff hatchery Leaflets transport Faeces farm (overshoes) Caeca slaughterhouse Neck skin slaughterhouse n.i. * n.i. * Breast skin slaughterhouse n.i. * n.i. * n.i. * The 2000 objective of the plan of approach in the control programme for poultry meat was to achieve a level of infection with Salmonella spp. of less than 10% in broiler flocks. Although contamination in broiler breeding flocks was consistently low between 1999 and 2002 (table ) contamination strongly increased just before and after arrival of the chickens at the farm. Table shows that much progress has been made between 1998 and 2002 in the reduction of Salmonella contamination in all the links of the poultry meat production chain. However, in 2000 the target of 10% was not attained for the tested caeca or neck skins at the slaughterhouse nor for breast skins tested in In the control programme the new objective was set at a maximum Salmonella contamination level for produced meat of 10% at the end of Since 2001, breast skins have been tested as a more realistic proxy of the end product. The large difference between the results for caeca and neck skins suggests a strong influence of cross-contamination at slaughter which is not representative for the real contamination of the meat. Over the whole of 2002 breast skin tested at an average contamination of 12%, but in the 4th quarter of 2002 fell below the target of 10%. The 12% contamination of breast skins in 2002 is comparable to the 13.4% contamination found in poultrymeat at retail in the same year (table ). * n.i. = no investigation Figure shows that the change of contamination in the different links of the poultry meat production chain observed over the years (table ) is generally in good agreement with the changes in the different quarters of a year. Serotype distribution of Salmonella in animals Isolates received and typed at the Salmonella reference laboratory at RIVM in general confirm the predominance of serotypes in pigs, cattle and poultry as found in the more specific surveillances of healthy animals at farms (VWA/KvW - RIVM), slaughterhouses (PVE), at retail (VWA/KvW) and in animals for clinical diagnostic purposes (GD). In 2002, S. Typhimurium was still predominant in pigs but to a much lower extent than in the period , related to the (re-)emerging of other serotypes as indicated in the table below. No shift of serotypes has occurred in cattle examined for diagnostic purposes, where S. Dublin still predominates, followed by S.Typhimurium. However, it is striking that S. Dublin was also found in healthy cattle at farms (surveillance, data not shown) in both 2001 and 2002 and that this serotype also occurred incidentally in pigs. 39

40 Since 1998 the RIVM database for Salmonella allows the distinction within poultry between broilers (including products) and layers (including animals from reproduction and eggs). It clearly shows the continuing increase of S. Paratyphi B var. Java in broilers as well as in layers (table ). The (relative) importance of S. Enteritidis continues to decrease in broilers but it is still the predominant serotype among layers. Several other more minor changes in poultry are also indicated. Phagetype distribution of S. Enteritidis in animals In 2002 the serotype S. Enteritidis was rarely found in pigs and cattle, was strongly reduced in poultry and rarely found among broilers (table ). In layers S. Enteritidis is still the predominant serotype although the Salmonella contamination level itself is low. As in humans the predominant Pt-4 continued to decrease in poultry and was replaced by Pt-1, Pt-6, Pt-8 and Pt-21. In 2002 Pt-21 was the strongest emerging phagetype in humans. Phagetype distribution of S.Typhimurium in animals In pigs S. Typhimurium as a whole was less important in 2002 (table ). However, ft-507 increased and that increase might be related to an increasing number of small outbreaks in humans in 2002 and In cattle DT-104 became relatively more important. In poultry the fraction of S. Typhimurium isolates was very low in Phagetype ft-150, causing problems in humans and poultry in the late 90s and in 2001, was not found at all in Clear differences exist in types circulating among broilers and layers and between the examined years. From these detailed typing data it is estimated that >35% of all human Salmonella infections in 2001 and 2002 were still related to the consumption of contaminated eggs. Food In 2002, the monitoring programme of poultry meat at retail from VWA/KvW showed that the gradual decrease of contamination of Figure Percentage of Salmonella positive flocks from hatchery to the end of the slaughterline Positive flocks (%) th st nd 3rd 4th 1st nd 3th 4th 1st nd 3th 4th 1st nd 3th 4th 1st nd 3th 4th Quarter Fluff (hatchery) Inlay leaflets (hatchery) Faeces (boiler farm) Neck skin (slaughterhouse) Breast skin (slaughterhouse) Caecum (slaughterhouse)

41 Table Serotype distribution of Salmonella in animals (NRL, RIVM) Pigs Cattle Poultry Broilers Layers Pigs Cattle Poultry Broilers Layers Total number , ,860 1,377 5,569 1,959 1,475 Serotype%Total % % % % % % % % % % Typhimurium Paratyphi B var. Java Dublin Enteritidis Infantis Bovismorbificans Derby Virchow Mbandaka Brandenburg Livingstone Indiana Manhattan Agona London Goldcoast Senftenberg Heidelberg Panama Bareilly Kentucky Anatum Hadar Cubana Tennessee Montevideo Cerro Lexington Rissen Oranienburg Other serotypes * Poultry: all chicken categories together; Broilers: including chicken products; Layers: including repro and eggs poultry meat with Salmonella spp. since 1996 has continued (table ). In % was contaminated with Salmonella spp. More than 50% of these consisted of S. Paratyphi B var. Java. The S. Enteritidis and S. Typhimurium contamination reached a level of approximately 1% in 2002; S. Enteritidis continued to decrease whereas the level of S. Typhimurium stayed the same as in S. Hadar, regularly found in the years up to 2001, was rarely found in The results of the investigations of meat derived from cattle and pigs indicate that there was an elevation of contamination when compared to the previous sentinel year 1999, especially in pork (table ). 41

42 Table Phagetype distribution of S.Typhimurium in animals (NRL, RIVM) Pigs Cattle Poultry Broilers Layers Pigs Cattle Poultry Broilers Layers Typhimurium%Total 45.4% 32.6% 2.6% 2.7% 3.8% 70.4% 31.7% 4.8% 3.7% 2.4% Total number , ft%typhimurium % % % % % % % % % % ft-506±dt ft-507, some DT ft-655, some DT ft-401±dt-104, DT ft-296±dt-12, some DT ft ft ft-3, some- DT ft-350±dt-193,some DT ft-510 some DT ft-301 some DT ft-20±dt ft ft-2±dt ft-60±dt-12, some DT ft-508 some DT ft-351 some DT ft- 530±DT ft-61±dt ft-150±dt-161, some DT ft-651, some DT Other ft * Poultry: all chicken categories together; Broilers: including chicken products; Layers: including repro and eggs. S. Typhimurium phagetypes are based on the Dutch phagetyping system. If correspondence is known with the Colindale scheme (Definitive Type: DT), this is indicated. Table Phagetype distribution of S. Enteritidis in animals (NRL, RIVM) Pigs Cattle Poultry Broilers Layers Pigs Cattle Poultry Broilers Layers Enteritidis%Total 0.2% 0.0% 6.8% 1.9% 41.8% 0.3% 1.1% 22.6% 8.9% 41.2% Total number , Pt%Enteritidis % % % % % % % % % % Pt Pt Pt Pt Pt Pt Pt Pt Other Pt Poultry: all chicken categories together; Broilers: including chicken products; Layers: including repro and eggs. S. Enteritidis phagetypes correspond to those in the Colindale scheme (Pt) used since 1997

43 Table Salmonella (including the serotype distribution) in poultry meat at retail (Monitoring programme VWA/KvW) Poultry meat sample size at retail 1,325 1,314 1, ,454 1,578 1,600 Salmonella spp. positive (%) Enteritidis positive (%) n.i. * Paratyphi B var. Java positive (%) Main serotypes As a fraction of the positive isolates (%) Paratyphi B var. Java Enteritidis Hadar Indiana Infantis Virchow Typhimurium (DT-104) (1.8) 1.3(0.7) 0.1(0.1) 7.4(7.0) 7.4(2.8) Other types * n.i. = no investigation Humans Serotype distribution of Salmonella in humans In 2002 a considerable reduction was observed by the 16 PHLs (coverage of 64% of the Dutch population) in the numbers of Salmonella isolates from relevant patient material. In total 1,588 isolates were sent in for typing that year, i.e., a 25% reduction with respect to 2001 (table ). This decrease holds for both the predominant serotypes S. Typhimurium and S. Enteritidis that have a completely different food origin. This reduction came on top of the 30% reduction between 1996 and 2001 of laboratory-confirmed salmonellosis. As in 2002 the decrease primarily concerned a reduction of infections among young children. The increase in 2001 thus appeared incidental and can be explained by the temporary epidemic rise of multiresistant S. Typhimurium DT-104 at the end of 2000 and in Hospitalization related to a Salmonella infection confirms the trend in the laboratory data, however, with a more pronounced increase in 2001 and an even stronger reduction afterwards in Table Salmonella in meat from cattle and pigs retail (Monitoring programme VWA/KvW) Product Year Pos./ % Salmonella % total serotype/ positives Meat, Beef / / Meat, Pork / Enteritidis (1/33) 3.0 Typhimurium (14/33) / The 1,588 laboratory-confirmed cases in 2002 amounts to an incidence of 15 cases per 100,000 inhabitants. Dutch epidemiological studies show that to arrive at the incidence of patients with salmonellosis visiting a GP or in the general population, this figure should be multiplied by about 2.5 and 14.5, respectively. The main serotypes in 2002 were still S. Enteritidis and S. Typhimurium comprising >75% of all isolates. The next serotypes in rank concerned about 2% or less. Strong differences exist between serotypes in being travel-related or not. An average of 6.5% of the Salmonella infections between 1998 and 2002 was travel-related but was 8% in 2001 and This figure however, is an underestimation due to underreporting. After interviewing cases in 2001 and 2002 it proved to amount to 12-17% on average. Obvious extremes are the typically non-endemic infections by S. Typhi and S. Paratyphi B that in reality should have been close to 100%. S. Paratyphi B var. Java in humans did not enter the list of the top serotypes found in humans. The fraction in humans (0.4%) is negligible compared to the >50% of all isolates as found in poultry at the end of the production chain. Phagetype distribution of S. Typhimurium in humans Phagetyping shows multiresistant DT-104 (ft-506 and ft-401) to have become predominant (table ). In 2001 the temporary epidemic rise in that year of DT-104 can also be seen in S. Typhimurium overall, Salmonella spp. overall and in hospital discharges for salmonellosis in that year (table ). As in pigs, ft-507 is emerging in humans as well and is related to numerous small explosions in the north-eastern provinces in the Netherlands in 2001 and Ft-507, but to a lesser extent than DT-104, is often multiresistant (see chapter 4). S. Typhimurium infections are rarely reported to be travel-related and are typically domestic. 43

44 Table Evolution of the main Salmonella serotypes in humans, reported by 16 PHLs (NRL, RIVM) Travel All hospitalized cases (1,165) (955) (855) (779) (781) (848) (465) Total number 2,975 2,889 2,556 2,266 2,127 2,059 2,082 1,588 6,5% Serotype%total % % % % % % % % Enteritidis % Typhimurium % Brandenburg % Infantis % Virchow % Hadar % Bovismorbificans % Manhattan % Goldcoast , % Derby % Paratyphi A , % Kentucky % Typhi , % Newport % Agona % Braenderup % SI 9.12:NM % Paratyphi B var. Java % Dublin % Panama % Paratyphi B % Livingstone % Others * Changes in a year as compared to other years of any significance are indicated Phagetype distribution of S. Enteritidis in humans S. Enteritidis Pt-4 is still the predominant phagetype within S. Enteritidis but is considerably reduced as compared to the end of the 1990 s (table ). Since that time Pt-21, Pt-1, Pt-6, Pt-8 and Pt-28 became more important. In 2002 the emerging of Pt-21, Pt-8 and Pt-1 became noteworthy. To some extent this is reflected in the emerging of these phagetypes in poultry. However, travel related cases may differ considerably between the Enteritidis phagetypes and is most noteworthy for Pt-1 and reflected in its higher level of fluoroquinolone resistance compared to more domestic phagetypes (chapter 4). The most important phagetypes Pt-4 and emerging Pt-21 seem more strongly domestic derived. Seasonal evolution of human salmonellosis Similar to the data of the occurrence of campylobacteriosis (section 3.2), the incidence of salmonellosis in humans shows a seasonal variation. Isolates of Salmonella from humans strongly increase at the end of the second quarter of each year and peak in early September. In contrast with Campylobacter however, the number of Salmonella-positive poultry flocks does not show an explicit seasonal fluctuation. On average 500 flocks are tested weekly, compulsory to the monitoring programme in poultry meat. The secular trend in humans is loosely related to the consistent gradual decrease of the level of contamination of poultry flocks. The trend in humans for example slightly increased in 2001 and strongly decreased in 2002 (table ). Estimated contribution of travel, farm animals and their products to human salmonellosis Using typing data of isolates of Salmonella spp. it was estimated which fraction of cases of human salmonellosis can be attributed to which category of farm animal and their products, or which fraction is of unknown origin including travel (figure ). Retrospective data were used of isolates derived from humans and farm animals that were routinely sent in to NRL at RIVM for serotyping and phagetyping. The estimate exploits the relative host-specificity of Salmonella serotypes and phagetypes. In addition to typing data from human isolates, data were used from isolates sent in from broilers (droppings from farms, caeca and meat products), layers (including raw materials for egg products, consumption eggs and materials from farms and hatcheries such as inlay leaflets, fluff, etc.), pigs (both adults, piglets, healthy and sick animals) and cattle

45 Table Phagetype distribution of S. Typhimurium in humans (NRL, RIVM) Travel Typhimurium%total Total number 823 1, % phagetype%typhimurium % % % % % % % % ft-506±dt % ft-507, some DT % ft-510, some DT % ft-296±dt-12, some DT % ft-508, some DT % ft % ft % ft-655, some DT % ft-295, some DT % ft-401±dt % ft-61±dt % ft-560±dt % ft-350±dt-193, some DT % ft-20±dt % ft % ft-60±dt-12, some DT % ft-651, some DT % ft-351, some DT % ft-301, some DT % ft-530±dt % ft-150±dt-161, some DT % ft-10±dt % ft % ft-656±dt % ft-204±dt-204b % Other ft * Typhimurium phagetypes are based on the Dutch phagetyping system. If correspondence is known with the Colindale scheme (Definitive Type: DT), this is indicated. Changes in a year compared to other years of any significance are marked. (mainly dairy cattle and veal calves, healthy and sick animals). Apart from the general decrease in human salmonellosis, within poultry a gradual shift is visible from the contribution of types circulating in broilers to those found in layers, the latter most probably related to the consumption of contaminated eggs. The slight increase in 2001 can be attributed to the epidemic increase of S. Typhimurium DT-104, explaining the increase of the estimated contribution of pigs and cattle in that year. The figure on page 45 shows the estimated contributions to human salmonellosis in The human cases were estimated to be related to types circulating in broilers (15%), eggs (37%), pigs (25%) and cattle (13%). About 10% of the human cases were either travelrelated (probably an underestimation) or from an unknown source. 45

46 Table Phagetype distribution S. Enteritidis in humans (NRL, RIVM) Travel Enteritidis%total Total number 1, % Pt%Enteritidis % % % % % % Pt % Pt % Pt % Pt % Pt % Pt % Pt-14b % Pt-6a % Pt % Pt % Pt % Pt-6b % Pt % Pt % Pt-13a % Pt-4b % Pt % Pt-21b % Pt-1b % Pt % No Pt % Other Pt % Enteritidis phagetypes correspond to those in the Colindale scheme (Pt) used since Changes in a year compared to other years of any significance are indicated. Figure Seasonal and secular trend of the weekly occurrence of human cases of salmonellosis and the percentage of positive flocks in the slaughterhouses (caeca). Compare with tables and (Laboratory surveillance RIVM) Infected flocks, caeca slaughterhouse (%) First isolates, humans (15 PHLs) dec- mrt- jun sep- dec- mrt- jun sep- dec- mrt- jun sep- dec- mrt- jun sep- dec- mrt- jun sep- dec- 0 % Positive flocks Humans

47 Figure Estimated contribution of travel, farm animals and their products to laboratory-confirmed human salmonellosis and estimated salmonella infection cases in the general population (Laboratory surveillance RIVM) Lab-confirmed cases of salmonellosis Travel/other Cattle Pigs Layers/Repro/Eggs Broilers (-products) 0 Figure Estimated contribution of travel, farm animals and their products to laboratory-confirmed human salmonellosis in 2002 (Laboratory surveillance RIVM) Cattle 13% Travel / other 10% Broilers incl. products 15% Conclusions Human salmonellosis has been decreasing steadily during the past 25 years from about an estimated 150,000 cases in the general population at the beginning of the eighties up to about 50,000 at the beginning of the 21st century. The decline in 2002 is especially large if compared to 2001 with 25% less laboratory confirmed cases and 45% less hospital admissions. Part of the reduction in recent years can probably be attributed to the control programmes put into effect since the end of 1997 in the poultry production chain. In all the links of the poultry meat production chain including at retail, a gradual decrease in level of contamination is seen, especially with respect to S. Enteritidis. Pigs 25% Layers / Repro / Eggs 37% 47

48 However, the improvement had already started a few years before the start of the control programme, furthermore only about 15% of human salmonellosis can be attributed to salmonella types predominating in poultry meat. More than 35% of human salmonellosis can be attributed to the consumption of eggs and the number of cases involved has not changed much since 1999, presumably reflecting the fact that among layers S. Enteritidis decreased only slightly between 1998 (11.1%) up to 2002 (8.8%). A point of concern is the fact that although S. Enteritidis Pt-4 strongly decreased, other phagetypes are replacing it. This has been happening in all European countries in the past few years and it is hoped for that the control strategies against S. Enteritidis are not now selecting for the more persistent phagetypes. Another matter of concern is the explosive increase of S. Paratyphi B var. Java in poultry meat although so far this seems less harmful for humans than other salmonellae. Apart from the economic impact on the poultry industry of this highly persistent strain it is developing resistance to fluoroquinolones very quickly and thus poses a threat to public health. Not much progress has been made in the reduction of the levels of contamination of cattle and especially of pigs. This is of concern because pigs and cattle are the main reservoirs for multiresistant S. Typhimurium DT-104, one of the more important causes of human salmonellosis in the Netherlands. In 2001, an estimated 250 extra laboratory-confirmed cases were due to DT-104 infections, many of whom were hospitalised. It is worrying that it is neither understood why suddenly in the beginning of 2001 so many cases of salmonellosis occurred due to DT-104, nor why it decreased so much in The implementation of the control programme for the pig sector during 2003 may give more control of the DT-104 problem, but the first results of the programme will not appear before the end of The results of the monitoring of antimicrobial resistance are reported in chapter 4. Frame 8 Monitoring of feed ingredients and compound feed from outside the European Union Introduction Apart from the monitoring of Salmonella contamination of feed by KDD described in section 3.11, VWA/RVV investigates imported feed ingredients and compound feed from outside the European Union. This is done within the framework of Directive 95/53/EC. On arrival in the Netherlands of feed ingredients and compound feed from so called third countries, VWA/RVV takes random samples of Salmonella-critical imported products from vegetable and animal origin and examines these samples for contamination with Salmonella spp. The minimum sample size of the mentioned products is 500 gram. Results The results of the monitoring programme are listed in table 8. Table 8 Percentage of Salmonella positive samples of imported feed ingredients and compound feed from outside the EU in 2002 (Monitoring programme VWA/RVV) Type of sample Pos./Total % Grain and grain products 1/ Barley and by-products 0/ Wheat and by-products 0/ Maize and by-products 1/ Oil seed and derived products 21/ Rape seed and by-products 2/ Palm kernel and by-products 0/ Soy bean and by-products 17/ Sunflower seed and by-products 2/ Linseed and by-products 0/1 0.0 Coconut expeller 0/ Sesam expeller 0/1 0.0 Miscellaneous 0/ Tuber, roots 0/9 0.0 Other seeds and fruits 0/2 0.0 Forages and Roughage 0/2 0.0 Citrus pulp 0/ Copra flakes and expeller 0/5 0.0 Tapioca 0/1 0.0 Bakery by-products 0/2 0.0 Total ingredients of vegetable origin 22/ Fish Meal 2/ Total ingredients of animal origin 2/ Cattle Feed 0/ Pig Feed 0/ Poultry Feed (miscellaneous) 2/ Poultry Feed (Broilers) 0/ Duck Feed 0/2 0.0 Total compound feed 2/ Total feed ingredients and compound feed 26/ Conclusions The results of the monitoring in 2002 of imported feed ingredients and compound feed from outside the EU by VWA/RVV were in good agreement with the independently derived results from the more extensive monitoring of companies in the feed sector by KDD, carried out within the framework of the GMP-standard. Comparable high levels of Salmonella contamination were found in the feed ingredients rape seed and, to a lesser extent, soy bean and sunflower seed. Low levels of Salmonella contamination were found in the examined compound feed samples.

49 Frame 9 Salmonella control programme in the pig sector Introduction On the initiative of PVE an investigation was carried out between September 2001 and April 2002 at a large slaughterhouse to examine the Feasibility of controlling Salmonella in pig meat. The aim of the project was to explore the possibility of producing pork with the lowest possible levels of Salmonella contamination in an economically feasible way under the conditions present in the Netherlands. Salmonella-free pigs were slaughtered separately from other pigs with an unknown Salmonella status. At the end of the slaughter line Salmonella was found in less than 2% of the carcasses, measured using the USDA sponge method. The results did not indicate that separate slaughter had any positive effect. It was however possible to observe a positive effect when the influence of slaughter-line contamination, caused by strains of Salmonella in the C group (which were persistently present on the slaughter equipment), was not taken into account. The conclusion was that separate slaughter can only have an effect on the prevalence of Salmonella on carcasses if slaughter-line hygiene is considerably improved. This includes more regular cleaning of parts of equipment that are difficult to reach and effective cleaning of equipment during slaughtering, which will require modification of the slaughter equipment. Moreover, separate slaughter becomes more worthwhile as the number of Salmonella-free farms grows. The pig industry has already achieved good results with regard to reduction in the prevalence of Salmonella in recent years. This reduction can mainly be explained by the introduction of generic hygiene measures in the pig industry (such as IKB, hygiene regulations for all chains working in the pig sector and a Code of Hygiene for pig slaughterhouses). Despite that, further reduction in contamination of pork with Salmonella is regarded as necessary. In order to reach that objective, a Salmonella control programme has been drawn up that will be implemented in The results of the investigations already conducted and the results of similar investigations in other countries have been used in the development of that programme. Main features of the Salmonella control programme for the pig sector in 2003 Objective The main purpose of the control measures is to reduce the contamination of pig carcasses with Salmonella by means of optimizing slaughter hygiene and a reduction of the number of heavily contaminated pig farms. Following a 12-month monitoring period this objective will be quantified and tied to a time schedule. The purpose of the monitoring phase of the control programme is to establish the baseline situation and track trends in contamination, to make it possible to determine the effectiveness of the control measures. Scope If the outlined objective is achieved it is important for all fattening farms and slaughterhouses to participate in the Salmonella programme. The industry has therefore decided to impose various measures on all fattening farms and slaughterhouses by means of a PVV Regulation. The programme will initially concern pig fatteners and pig slaughterhouses. In the EU the Danish programme is regarded as the guiding approach. The Dutch approach is based on the broad guidelines of the Danish programme supplemented with elements that are considered specific for the Dutch situation. The control programme comprises two aspects: monitoring of all pig fattening farms and slaughterhouses and intervention plans for contaminated pig fattening farms and slaughterhouses. Pig fatteners All farms in which fattening pigs are kept will be obliged to participate in the Salmonella monitoring programme. From the animals of each farm 36 blood samples a year must be examined for Salmonella, with a regularity of 12 samples once every four months. After being monitored for a year, pig-fattening units will be divided into three Salmonella-status categories (categories 1, 2 and 3 for slightly. moderately and heavily contaminated farms respectively). Heavily contaminated farms must draw up and implement an intervention plan in order to be then reclassified in a lower category. This intervention plan contains a number of proposed measures, objectives and a time schedule. Categorisation is based on the monitoring data. Slaughterhouses All slaughterhouses will start with a monitoring phase, during which carcasses will be sampled for Salmonella using a method that is now common worldwide (USDA sponge method). Microbiological tests carried out under the Code of Hygiene for pig slaughterhouses also help to create a picture of slaughter hygiene. After a year an evaluation will be carried out and the standard for the maximum permitted percentage of Salmonella contamination will be determined. Slaughterhouses that exceed this standard will then need to draw up and implement an intervention plan with the aim of getting their farm below the set Salmonella standard. Time schedule The monitoring phase will be implemented as soon as consensus is reached with the Ministry of LNV about the preconditions connected with the project (possibly in 2004). After being monitored for 12 months the pig fattening farms and the slaughterhouses will be categorized. After that, the details of the intervention plans will be worked out. Frame 10 Purchase of pig manure as a preventable risk factor for Salmonella Typhimurium DT-104 infections in dairy cattle Introduction Pigs and cattle are considered to be the main sources for human infections with Salmonella Typhimurium in the Netherlands and especially so for the multiresistant DT-104 strain. Of all Salmonella isolates from pigs and cattle received at RIVM between 1998 and 2002, 66% and 32%, respectively, were S. Typhimurium and about 49

50 14% was DT-104 in both species. Phagetype DT-104 has been steadily increasing between 1998 and 2002 in both pigs and cattle. The epidemic increase of human infections with DT-104 in 2001 has drawn heightened attention towards studies in farm animals such as that reported below in 1999 that may provide instruments for prevention of the disease. Results S. Typhimurium is, after S. Dublin, the second most important serotype involved in clinical outbreaks of salmonellosis in dairy herds in the Netherlands. The incidence of clinical outbreaks of S. Typhimurium increased in 1999, especially in the southern part of the Netherlands. Veterinary practitioners often observed clinical signs only among dry cows and in several cases S. Typhimurium DT-104 was isolated. Risk factors for these outbreaks were studied in a matched case-control study with 47 case farms and 47 control farms. All 47 case farms had at least one positive bacteriological culture for S. Typhimurium in one or more samples. S. Typhimurium ft-401 and ft-506 (DT-104) were most frequently isolated (13 isolates). On most case farms (66%) clinical signs were seen among adult cows only. The main clinical signs were diarrhoea (in 92% of the farms) and depression (79%). Each case farm was matched with a control farm in the same region. Control farms had no history of clinical salmonellosis. A questionnaire was held on case and control farms. The relationship between the S. Typhimurium status of the farm and possible risk factors was analysed using multivariate conditional logistic regression. Significant risk factors (P < 0.10) in the final model (table 10) were purchase of (pig) manure, no cats present on the farm, feeding colostrum from more than one dam, a non-seasonal calving pattern, restricted grazing of lactating cows and a high mean mowing percentage of pasture. Conclusions The role of manure in the introduction of an infection with S. Typhimurium has been recognised in earlier studies. However, experiments showed that it was not possible to infect grazing cattle with S. Dublin 10 days after infected manure was spread on the pasture. Recently it was shown that the spread of poultry manure on bordering property is a risk factor for Salmonella spp. infections in cattle herds. In the Netherlands dairy farms purchase pig manure for its low price as fertiliser, especially in the southern part of the Netherlands. In an earlier study S. Typhimurium was isolated from faeces of 24% randomly selected finishing pig herds in the southern part of the Netherlands. With the objective of finding effective ways of reducing the spread of S. Typhimurium by pig manure the GD is now studying transport of pig manure and the ASG (Lelystad) is studying treatment of pig manure. Frame 11 Explosive increase of Salmonella Paratyphi B variation Java in poultry Introduction In the Netherlands Salmonella Paratyphi B var. Java has increased explosively in poultry since It proves to be very persistent once introduced into a poultry farm and the organism is quickly developing resistance to fluoroquinolones. The current status of the epidemic is described below together with its possible public health implications. Results In the Netherlands S. Paratyphi B var. Java increased in poultry from less than 2% of all isolates before 1996 up to 70% in Figure 11 shows that S. Paratyphi B var. Java is also appearing in layers and material from hatcheries and reproduction flocks. Thus, S. Paratyphi B var. Java has overtaken S. Enteritidis, which was the dominant serotype between 1996 and 1999 (table 11.1), S. Hadar in 1994 and 1995 and S. Virchow (data not shown) at the start of the 1990s. This development in poultry runs parallel with that described for Germany up to 2000 (figure 11). The Germans showed that in the late nineties it concerned isolates of only one multiresistant clone of S. Paratyphi B var. Java, with a distinctive PFGE pattern. In the middle 1990s it concerned a few multiresistant clones, whereas isolates in the beginning of the 1990s that were not all derived from poultry were genetically much more heterogeneous and sensitive to antibiotics. This change in the Netherlands from a few multiresistant clones in poultry to one clone is reflected in the antibiotic resistance pattern and suggests that up to 1995 S. Paratyphi B var. Java from poultry was not found in humans. Table 10 Risk factors for S. Typhimurium infection; 47 cases and 47 matched control dairy farms; G = with 6 degrees of freedom. P < (GD) Frequencies Case Control Risk factor in multivariate model Yes No Yes No Odds ratio 90% CI P Purchase of manure Presence of cats on farm Colostrum only from own dam Calving of dams throughout the year Unrestricted grazing of lactating cows Mean mowing percentage for grass-land Most frequently pig manure

51 Figure 11 Increase of S. Paratyphi B var. Java as a proportion of all Salmonella isolates from all chicken material (faeces, neck skin, caeca, inlay sheets, fluff, meat, etc.) investigated in the Netherlands and the absolute number of isolates (right-hand Y axis) received by the CRL Berlin (Laboratory surveillance RIVM) % of all isolates Number of isolates from chicken in Germany august 0 Chicken miscellaneous (NSC) Broilers (-products) Layers (+ repro) Chicken (Germany) Table 11.1 S. Paratyphi B var. Java in chicken products from butchers, supermarkets and poulterers (Monitoring programme VWA/KvW) Samples investigated 1,359 1,325 1,314 1, ,454 1,578 1,600 Salmonella spp. positive (%) S. Hadar positive (%) S. Enteritidis positive (%) nd S. Paratyphi B var. Java positive (%) Table 11.2 Development of resistance of S. Paratyphi B var. Java isolated from various chicken material such as faeces, meat, neck skin, caeca, chicken products, inlay sheets, fluff, etc. (Monitoring programme CIDC) MIC value (µg/ml) Res% Ciprofloxacin Flumequine * Clinical relevant resistance starts to be noted from the first category in the grey shaded areas. The resistance percentage is given in the last column. The results are based on the frequency distribution of the MIC values of isolates for the above-mentioned antibiotics. 51

52 Up to the end of 2002 it appeared that the resistant clone which is spreading now is a local problem and it seemed to be less pathogenic for humans. Although the exposure of humans to contaminated poultry meat is relatively high, human patients with S. Paratyphi B var. Java infection are rare (0.3% of all isolates). Between 1998 and 2002 however, 50% of the human isolates preserved and typed with PFGE were identical to the poultry clone. Moreover, in December 2002 a request for information from Scotland about an emerging S. Paratyphi B var. Java problem in humans could be traced to imported poultry with an identical PFGE pattern to the clone found in Germany and the Netherlands. This suggests that S. Paratyphi B var. Java infections may be a serious public health threat and puts the problem into an international perspective. Presumably through treatment of poultry flocks with quinolones, resistance of S. Paratyphi B var. Java to flumequine increased from 3% between 1996 and 2000 to 19% in 2001 and 37% in 2002 (table 11.2 below) while that of other serotypes in poultry remained at about 7% (not shown). S. Paratyphi B var. Java is also becoming less sensitive to ciprofloxacin (table 11.2) which is the antibiotic of first choice in serious cases of salmonellosis. Conclusions As it has been established that S. Paratyphi B var. Java can produce typhoid-like clinical symptoms or can lead to outbreaks, the increase of S. Paratyphi B var. Java in poultry in combination with the resistance of S. Paratyphi B var. Java to quinolones poses a threat for public health. The same development in antibiotic resistance in Campylobacter jejuni has been a topic of discussion around the world for more than a decade with respect to authorising the use of quinolones in poultry. Even reduced sensitivity can make the standard treatment less effective and may lead to the selection of clinically resistant strains during treatment. The public health risk of quinolone-resistant Salmonella is greater than that for Campylobacter, as fluoroquinolones are the medicines of first choice for the treatment of a serious Salmonella infection. In 1996 the poultry sector in the Netherlands first noted problems with S. Paratyphi B var. Java in 5 chicken farms. Currently about 100 farms are battling this infection. Neither the efforts of several integrations nor a large intervention study by CIDC resulted in effective ways to control the infection. S. Paratyphi B var. Java infection persists despite perfectly adequate scores for hygiene, whereas S. Enteritidis, S. Hadar, S. Infantis, S. Virchow. etc. can be eliminated successfully using the standard cleaning and disinfection procedures. Still tests have made clear that S. Paratyphi B var. Java is no more resistant to disinfectants than other Salmonella types. Developments The Ministries of VWS and LNV and the production boards with their research institutes together with the poultry meat production integrations have decided in 2002 to work together to find measures for effective control of S. Paratyphi B var. Java in the poultry industry Tick-transmittable zoonotic pathogens: Borrelia burgdorferi, Ehrlichia/Anaplasma spp., tick-borne encephalitis virus and Babesia spp. Introduction There are many different tick species, but the most prevalent species in the Netherlands is Ixodes ricinus, also known as the sheep tick. This tick species will bite many different hosts ranging from reptiles to birds and mammals including man and can transmit a variety of pathogens. In Europe I. ricinus is the predominant carrier and transmitter of Borrelia burgdorferi, the causative agent of Lyme disease. There are no reports that describe B. burgdorferi exerting clinical manifestations in vertebrate animal hosts with the exception of dogs. In the last decade several researchers have shown that 10 to 35% of I. ricinus ticks in the Netherlands are infected with various closely related B. burgdorferi species also designated as B. burgdorferi sensu lato. Surveys have shown that there are regional differences in the rate of tick infection. Although it is clear that birds and small rodents are competent reservoir hosts for B. burgdorferi sensu lato it is unknown what is the rate of infection of these animals in the Netherlands. Another tick-transmitted human pathogen is a bacterium that in the past has been designated as the agent causing human granulocytic ehrlichiosis (HGE). This organism is now considered to be the same species as the pathogen causing tick-borne fever in ruminants, Ehrlichia phagocytophila, and the causative agent of tick-borne fever in horses, E. equi, and these species have recently been united under the name Anaplasma phagocytophilum. Ehrlichia and Anaplasma species may also cause disease in canines. In countries such as Germany, Austria and Russia the tick-borne encephalitis virus (TBEV), which is also transmitted by I. ricinus, causes many cases of serious human disease each year. In most Mediterranean countries in Europe infections with Rickettsia species, in particular R. conorii, frequently occur. The protozoan pathogen Babesia divergens and B. microti are pathogens that rarely cause disease in humans and only 25 European cases of disease caused by infection with B. divergens have been described so far. B. divergens is the cause of occasional outbreaks of babesiosis among cattle in the Netherlands. Animals Prevalence and results There is no surveillance system for detecting tick-borne infections in wild or domestic animals in the Netherlands. However, several researchers have determined the prevalence of infection with B. burgdorferi in ticks in the Netherlands. In most surveys ticks were collected by flagging from the vegetation and subsequently investigated for the presence of B. burgdorferi using microscopy (dark-field analysis) or by PCR.

53 In 2001 and 2002 RIVM analyzed Ixodes ticks from the Netherlands and other European countries for the presence of Borrelia, Ehrlichia/Anaplasma and Rickettsia species. These studies confirmed previous findings that B. burgdorferi is present in approximately 10% of ticks in the Netherlands. Furthermore, 15% of the ticks collected from vegetation by colleagues from CIDC were shown to carry Anaplasma or Ehrlichia species. One-third of these positive ticks carried an Ehrlichia-like species identified by RIVM in PCR analysis of blood samples from mice captured in various parts of the Netherlands revealed that DNA from this organism could be detected in 20% of the mice. However, the clinical significance of this organism still remains to be determined. Serological assays using a recombinant antigen from A. phagocytophilum showed that 40% of roe deer sera contained antibodies to this antigen. Such antibodies were also detected in 40% of dog serum samples This strongly suggests that roe deer and dogs are frequently infected with A. phagocytophilum. A limited number of ticks in the Netherlands were also screened for the presence of Rickettsia species and 20% of the ticks were shown to be infected with R. helvetica. Despite extensive testing of a large number of ticks by Van der Poel and coworkers (MGB, RIVM) no TBEV has been found in ticks in the Netherlands. During 2001 and 2002 no studies were carried out for the presence of Babesia in ticks or other animals in the Netherlands. Humans Prevalence and results As in most other European countries, the Netherlands has no surveillance system to monitor the number of cases of Lyme disease or other cases of tick-borne infectious diseases. Therefore, the true incidences of these diseases remain unknown. Using postal questionnaires a survey showed that in 1994 general practitioners saw 6,500 patients with clinically manifest cases of Lyme disease (erythema migrans) accounting for an incidence of 43 per 100,000 population. Furthermore, these GPs also reported that they saw 33,000 patients who had sustained tick bites. In 2002 this study was repeated to determine whether there had been an increase or decline in the number of cases of Lyme disease and in the incidence of tick bites (frame 12). In 1999 the first and so far only case of endogenous human granulocytic ehrlichiosis in the Netherlands was reported. However, the high degree of A. phagocytophilum infection in ticks in the Netherlands would suggest that more cases of human ehrlichiosis may occur. In collaboration with Norwegian researchers, RIVM has discovered that in Norwegian sheep at least two variants of A. phagocytophilum exist with clearly distinct virulence. This might indicate that the variants found in the Netherlands may be less virulent for humans, resulting in a low incidence of ehrlichiosis despite high infection rates in ticks. So far there have been no reports of endogenous rickettsiosis, TBE or Babesia infections in humans in the Netherlands. Conclusions Lyme borreliosis remains the most important tick-transmitted disease in the Netherlands. The number of infected ticks in the Netherlands is large and the number of tick-bite cases of Lyme disease in humans is considerable. This makes Lyme disease the most frequently occurring and clinically most important non-food-related zoonosis in the Netherlands. Based on present knowledge B. burgdorferi does not pose a real threat to animals, with the possible exception of dogs. Prevention of tick bites and prompt removal of attached ticks is the best strategy for reducing the number of cases of Lyme disease. It will be virtually impossible to control the numbers of vectors and reservoir animals. The number of tick bites and cases of erythema migrans has doubled since 1994 (frame 12). However, it is unclear whether this also means that the number of ticks and/or the rate of infection of these ticks have increased. Longitudinal studies will be required to study possible shifts in tick density and infection. Many ticks are infected with Anaplasma and Ehrlichia spp.. However, only a single human case and very few cases of ehrlichiosis in animals have occurred in the Netherlands. The most likely explanation for this observation is that these bacteria species have only limited virulence. However, under-diagnosis cannot yet be ruled out completely. Physicians and veterinarians must remain alert and should be aware of the possible diseases caused by Anaplasma and Ehrlichia. Ticks in the Netherlands do not seem to carry TBE virus. There have been reports that the area in which TBEV is prevalent is expanding to the more northern parts of Germany. However, these reports have not been substantiated with solid evidence. Also, no TBEV containing ticks have been identified so far in the Netherlands, suggesting that TBEV does not yet pose a threat for the Dutch population. It was surprising to find that 20% of ticks in the Netherlands were infected with R. helvetica. The virulence of R. helvetica for humans is unclear and only a single case of classic rickettsial disease caused by R. helvetica has been reported in France. In addition, R. helvetica infection has been associated with sudden cardiac death in two young and otherwise healthy individuals and with a large number of cases of sarcoidosis in Sweden. However, others have not corroborated these Swedish reports. Further research to determine the clinical relevance and possible animal reservoir are required. Frame 12 Doubling of tick bites and erythema migrans between 1994 and 2001 in humans and related changes in risk factors Introduction In April 1995 and 2002 all general practitioners (GPs) in the Netherlands were asked to complete a short questionnaire on the 53

54 Table 12.1 Average number of consultations per general practice, incidence and estimated* total number of tick bites and erythema migrans (RIVM) 95 Number of consultations / GP incidence/100,000 (CI )** Total number of consultations (15,4 million inhabitants) 2001 (16 million inhabitants) Tick bites 4,7 9,0 193 ( ) 381 ( ) 30,000 61,000 EM*** 0,9 1,8 39 (38-41) 73 (71-75) 6,000 12,000 * Estimates after Monte Carlo simulation of the underlying theoretical distributions of the response categories in the questionnaires 95 ** CI = 95% confidence interval *** EM = erythema migrans Figure 12 Geographical distribution of the incidence of EM and the relative increase (relative risk: RR) of incidence of EM in the Netherlands in 1994 and 2001 (RIVM) number of tick bites and erythema migrans (EM), the first symptom of Lyme disease, for which they were consulted in the previous year and the size of their practice. Associations with possible behavioural and ecological risk factors were determined at the municipal level for both 1994 and 2001, if available. The aim of these studies was to obtain detailed information about the actual occurrence of tick bites and erythema migrans in the Netherlands and changes thereof, as well as the geographical distribution and regional differences in risk factors that might be involved in the general population among 9,000 people in 43 cities indicates that only 1 out of 15 tick bites leads to consultation with a doctor. Risk areas were found in the north and east of the country and a strip along the coast. The strongest increase in tick bites and EM was seen in the south and a small area in the northeast (figure 12). Results On the municipal level, tick bites and EM were associated with areas covered with wood, areas with sandy soil, the number of roe deer, urbanisation and tourism (table 12.2). Regional increases in tourism, new wood in urban regions and an increase in horse riding (probably a regional proxy of outdoor recreation) were positively associated with the increase in tick bites and EM. There was a negative association with the increase of the degree of urbanisation. In 2001 all GPs together saw approximately 61,000 patients with tick bites and 12,000 patients with EM (table 12.1). The incidence of EM was estimated at 73 per 100,000 inhabitants. A general practice on average had about 2 patients consult them for EM and 9 related to a tick bite. This is a two-fold increase for all estimates with respect to Answers to a question in a Dutch study in In order to obtain detailed information about the relative risks (RR) associated with the incidence of EM, the data of 1994 and 2001 were combined in multivariate Poisson regression models.