SALMONELLA IN PORK PRE-HARVEST AND HARVEST CONTROL OPTIONS BASED ON EPIDEMIOLOGIC, DIAGNOSTIC AND ECONOMIC RESEARCH

Deliverables

The purpose of the present work was to investigate the application of ready-to-go Salmonella PCR tests, based on dry chemistry, for final identification of rough presumptive Salmonella isolates. The results were compared with two different bio-typing methods performed at two different laboratories. The sensitivity of the BAX Salmonella PCR test was assessed by testing a total of 80 Salmonella isolates, covering most serogroups, which correctly identified all the Salmonella strains by resulting in one 800-bp band in the sample tubes. The specificity of the PCR was assessed using 20 non-Salmonella strains, which did not result in any DNA band. A total of 32 out of the 36 rough presumptive isolates were positive in the PCR. All but one isolate were also identified as Salmonella by the two biochemical methods. All 80 Salmonella strains were also tested in the two multiplex serogroup tests based on PCR beads. All strains belonging to the serogroups B, C1, C2-C3, and D were grouped correctly. Among the 32 rough presumptive isolates identified, 19 isolates resulted in a band of 882 bp (sergroup B), 11 isolates resulted in a band of 471 bp (serogroup C1), and 2 isolates showed a band of 720 bp (serogroup D). In conclusion, rough presumptive Salmonella isolates can be conveniently confirmed to the serogroup-level, using the pre-mixed PCR tests. The system can be easily implemented in accredited laboratories with limited experience in molecular biology. The sensitivity of the PCR was assessed by testing a total of 80 Salmonella isolates, covering most serogroups, which correctly identified all the Salmonella strains by resulting in one 800-bp band in the sample tubes. The specificity of the PCR was assessed using 20 non-Salmonella strains, which did not result in any DNA band. A total of 32 out of the 36 rough presumptive isolates were identified as Salmonella in the PCR and the biochemical methods. However, one isolate was positive only by the API-method, suggesting a false-positive identification. All the 80 Salmonella strains were also tested in the two multiplex serogroup PCR tests. All the strains belonging to the serogroup B, C1, C2-C3, and D were grouped correctly. All the other Salmonella strains, not from these serogroups, were negative in the serogroup tests. The 20 non-Salmonella strains were not tested in the serogroup tests, since the tests were only optimised to distinguish within the Salmonella enterica species.

The costs to society imposed by the presence of Salmonella in the pig meat production chain possess a number of components-the costs of the disease as manifested in pig herds, the costs incurred by illness in humans, outbreak costs to the food production, processing and distribution industry, averting behaviour costs to consumers, and consumer demand effects. Intervening in an attempt to control Salmonella involves the costs of operating control measures, as well as those of surveillance, regulation and enforcement of compliance where control measures are mandatory. These costs can be compared with the benefits to be gained from successful intervention. These comprise largely avoidance of the economic losses to society of Salmonella contamination or infection. Intervention targeted at controlling Salmonella in pigs and pig meat may also coincidentally reduce the prevalence of other diseases in pigs, the presence of other pathogens or bacteria in pig meat, or the incidence of other food borne illnesses in humans. The measures for the control of Salmonella in pigs and pig meat that are likely to be most cost-effective are additional consumer education, additional training of food service sector employees, improved hygiene practices at pig slaughterhouses and processing plants and treatment of pig carcasses with a chlorine dioxide wash. Unfortunately, the economic aspects of feed as a herd control measure could not be assessed here. In the short run, Salmonella control tends to be more cost-effective closer to the point of food consumption. In the longer run measures for the control of food borne illness (or livestock disease) which eliminate sources of infection, and may subsequently be discontinued, may ultimately be more cost-effective than measures of decontamination at a later stage in the food production chain, which may have to be retained indefinitely. Ranked in ascending order of annual costs these estimates suggest that, for the EU, construction of additional pen separations for finishing herds, improved hygiene practices at pig slaughterhouses and processing plants, and additional education of consumers as to food safety risks and hygiene practices are the least costly options. The annual benefits exceed the annual costs for all but three control measures, these being additional cleaning and disinfection on farms with breeding herds or finishing herds and gamma irradiation of pig meat using on-site facilities. The most cost-effective control options (as measured by the ratio of benefits to costs) are indicated to be additional education of consumers, additional on-site training of food service sector employees, improved hygiene practices in pig slaughterhouses and processing plants, and construction of additional pen separations for finishing herds. Whilst these control measures do not secure the largest reductions in human illness costs in absolute terms, they are indicated to be the most efficient means of reducing the human illness costs of Salmonella infection from pig meat in terms of the reduction in human illness costs secured per ECU of control expenditure. Differences in the extent to which intervention to control Salmonella in pigs and pig meat is cost-effective in different countries will be influenced, for most control measures, by national differences in the importance of foreign travel and consumption of pig meat as sources of human infection. A key indicator identified in this study is the quantity of pig meat produced relative to the size of the human population. Similar patterns of results are derived for the EU and the UK, for which the majority of control measures are found to be cost-effective, whilst the results for The Netherlands are similar to those for Denmark, where the majority of control measures are indicated to incur costs in excess of the benefits that they would secure. Denmark and The Netherlands produce much greater quantities of pig meat relative to the population size than the UK or the EU in total.

Serum samples from Denmark, Germany, Sweden and Greece within SALINPORK were analyzed by the Salmonella ELISA's at the Dept. of Biochemistry and Immunology, Danish Veterinary Laboratory, Denmark, and at the Animal Health Service, The Netherlands. Samples were normally received marked with a barcode. Sera and data are stored in a serum bank and a database, respectively, for use in future intercalibration studies for serological Salmonella assays or serology for other infections. The sera are derived from controlled inoculation studies in pigs with Salmonella serovars of various serogroups (B, C1, C2, D and E) performed in Denmark and The Netherlands as well as from cross-sectional and longitudinal studies of pig herds in the 5 countries, participating in herd studies within SALINPORK. The controlled inoculation (cohabitation) studies provided evidence for a low sensitivity of detection of S. Infantis and S. Livingstone infections (both of serogroup C1), at least under the experimental conditions chosen. More scattered information from various herd studies in Denmark and within SALINPORK also has pointed to this possibility. Inoculation studies with S. Goldcoast (serogroup C2) and studies of a Dutch herd furthermore demonstrated an almost complete lack of detection of infection with C2-serotypes by the Mix-ELISA. The data available for evaluation of the sensitivity towards serotypes of the D1-group appear too limited to draw a conclusion. Infection with S. Typhimurium at the herd level seems to be almost optimally detected in the Mix-ELISA. This was shown in both the inoculation studies and by the better sensitivity estimates for detection of infection with S. Typhimurium rather than any serotype in the cross-sectional studies. Thus, the Mix-ELISA appears as a strong tool for demonstration of S.Typhimurium infection in swine herds, with a co-determination of infections with other serotypes to a variable, and in many cases limited extent. Detailed results on the cross-sectional studies can be found in Chapter IV of the comprehensive report. The studies showed maximal levels of relative sensitivity and specificity between 80% and 90% for classification of herds as infected with S. Typhimurium, but significantly lower sensitivities towards detection of Salmonella infections of any serotype. Analysis of the longitudinal herd studies demonstrates that seropositivity occurred in 13/15 (87%) at least once during the study period in bacteriological positive herds, while 6/17 (35%) of herds without Salmonella isolations were classified as seropositive at least once during the study period. The bacteriological positive herds from DE, DK and NL showed a clear dominance of S. Typhimurium isolations, and the high level of sensitivity obtained is therefore in good concordance with results of the cross-sectional studies.

The objective of this study was to estimate the degree of cross-contamination of carcasses brought about by manual handling and processing during slaughter. The degree of cross-contamination was estimated by first slaughtering pigs from one or more Salmonella positive herds followed by pigs from one or more Salmonella negative herds. By sampling the carcasses at several point during the slaughter process, the measured contamination of the carcasses from the negative herds provided information on the degree of cross-contamination brought about by manual handling and processing. The results from this study showed that pigs from the Salmonella negative herds do not necessarily remain negative during and after slaughter. The source of contamination might have been the lairage, since it was possible for faecal matter to pass between the pens with the positive and the negative pigs. Another source of contamination of the carcasses was presumably the slaughter equipment, especially the carcass splitter. Carcasses of pigs may be cross-contaminated from either Salmonella positive pigs slaughtered previously on the same day or from contaminated slaughter equipment. The latter can of course also be contaminated from Salmonella positive pigs slaughtered on the same day, but the results strongly suggest, that remaining and/or persistent infections of the equipment also is an important source. Salmonella was not isolated from the hands of personnel. Though not thoroughly investigated in this study, stress imposed by transport and lairage is likely to increase the Salmonella contamination of slaughter pigs. Especially, the length of time spent in lairage seems to influence the level of contamination. This study showed that carcasses of pigs may be cross-contaminated from either Salmonella positive pigs slaughtered previously on the same day or from contaminated slaughter equipment. The latter can of course also be contaminated from Salmonella positive pigs slaughtered on the same day, but the results of this study strongly suggest, that remaining and/or consistent infections of the equipment also is an important source. Finally, though it was not that thoroughly investigated in this study, stress imposed by transport and lairage is likely to increase the Salmonella contamination of slaughter pigs. Especially, the length of time spent in lairage seems to influence the level of contamination. Based on the results of this study, it is recommended to: -Separate pigs from infected and non-infected herds during transport and lairage, and reduce the time in lairage to a minimum. -Proper cleaning of the lairage between batches of pigs and at the end of a slaughter day is also highly recommended. -Separate pigs from infected and non-infected herds during slaughter. This will reduce cross-contamination of otherwise negative carcasses brought about by manual or mechanical handling during slaughter. -Application of special hygiene practices during slaughter of pigs from high risk herds. -Clean and disinfect the slaughter equipment at least once a day, but preferably several times during a slaughter day e.g. during the breaks. This will reduce the level of carcass contamination from both consistently infected equipment and from equipment infected from Salmonella positive pigs slaughtered at the same day. -Bacteriological control of the cleaning and disinfecting procedures. This study indicates that these procedures are not always adequate. Slaughterhouses experiencing problems with persistently contaminated equipment should reconsider their current cleaning and disinfecting procedures.

To assess the prevalence and location of Salmonella contamination in European slaughterhouses, samples from products and the environment were collected at various slaughter processes throughout slaughter days in several sampling rounds. The initial screening of the slaughter process variables in the basic model indicated that the isolation of Salmonella from several slaughter processes was associated with the proportion of positive carcasses. In particular, isolation of Salmonella from the polishing, trimming, scalding and pluck removal operations seemed to be associated with an increased risk of carcass contamination. In the final analysis, the probability of recovering Salmonella from a carcass was found to be positively associated with the isolation of Salmonella from the polishing equipment. Further, the finding of Salmonella during the pluck removal procedure was found to increase the probability of finding Salmonella on the carcass, but only if the also the scalding water was found positive for Salmonella. Sufficiently high temperatures of the scalding water (62 degrees Celsius) and appropriate cleaning and disinfection of the polishing equipment at least once a day is recommended in order to reduce the level of carcass contamination. Slaughterhouses without special preventive measures during bung loosening and evisceration will most likely benefit from establishing such procedures. The probability of detecting Salmonella positive carcass was found to be positively associated with Salmonella contamination on the polishing equipment. When samples from the scalding process were positive, the probability of finding a contaminated carcass was almost 8 times higher when the pluck removal procedure yielded a positive sample than when it did not. In contrast, the finding of Salmonella during the pluck removal procedure had no effect on the level of carcass contamination, when Salmonella could not be isolated from the scalding water. Scalding usually reduces the number of Salmonella spp. on the carcass surface. However, if the water temperature drops below the recommended 62 degrees Celsius and/or the amount of organic material is sufficient to protect the bacteria against the heat, the probability of bacteria surviving this process is increased. These results indicate that samples taken from the polishing equipment may be used as an indication of contamination of the slaughter line. In addition, measurements of scalding water temperature can serve as a critical control point for the risk of carcass contamination during pluck removal.

To investigate whether purchasing gilts constitutes a risk for introduction of Salmonella, a source analysis study was performed on a total of 84 breeding or multiplying herds from 5 European countries, being Germany, Denmark, Greece, The Netherlands and Sweden. In these herds, blood samples were taken from 20 sows and 20 ready-to-ship breeding animals. There appeared to be a good correlation between the proportion seropositive sows and gilts in Germany and The Netherlands. A weak and non-significant correlation was found in Denmark and Sweden. In Greece, there were not enough observations to calculate a meaningful correlation coefficient. High proportions of seropositive gilts were found only in herds with high proportions of seropositive sows. This may indicate the occurrence of vertical transmission of infection (from dam to offspring) or merely reflect a generally established Salmonella contamination of these farms. In herds with a low proportion of seropositive sows, low proportions of seropositive gilts were found. This means that the status of sows in a herd with a low Salmonella prevalence among sows can be used as an indication of a low prevalence among gilts. However, this is not the case in herds with a high prevalence of seropositive sows, since both high and low proportions of seropositive gilts were found in these herds. With the current study protocol, it was not possible to further investigate why in some herds with infected sows, gilts remained seronegative and not in others. However, seropositive gilts were found in herds with seropositive sows, suggesting exposure to Salmonella and thus potentially constituting a risk of introducing infection to a new herd. Like other pig herds, breeding and multiplying herds can become infected with Salmonella. Infected sows may pass the infection on to their offspring, which then constitute a risk for the herds in which they eventually are introduced. In the SALINPORK study, evidence for the presence of seropositive ready-to-ship breeding animals was found in herds with seropositive sows, suggesting exposure to Salmonella and thus potentially constituting a risk when introduced to a new herd. However, the presence of seropositive sows in a herd was not always found to result in positive offspring. Still, these results point out that the transmission of infection between sow and offspring is an epidemiological phenomenon that deserves further attention, especially with regard to the potential function as a pre-harvest control point. An additional, often overlooked aspect is that since high infection levels can be found in sow herds, pork products derived from culled breeding stock represent a potential source of human infection.

A risk factor analysis was performed on the data collected from 94 pig finishing herds and their Salmonella-status based on serum samples collected at two slaughterhouses. Feeding of wet fermented by-products was found to have a lowering effect on the prevalence of Salmonella in finishing herds. Possible explanations for this observation are the low pH and the presence of organic acids, which are known to have a bacteriostatic action. Possibly, organic acids added to the drinking water of the pelleted compound feed of finishers has the same effect. Another protective factor is the presence of closed pen separations in a section. This prevents the exchange of faeces and urine between the pens and might prevent the circulation of any possible Salmonella infection. The use of bacitracin as an antibiotic growth promoter increases the chance for a herd to have a positive Salmonella status. A possible explanation is the working mechanism of bacitracin. Apparently, bacitracin is effective against Gram-positive bacteria and Salmonella being a Gram-negative bacteria is possibly not as much effected. Other antibiotic growth promoters might have more effect against Salmonella. All antimicrobial growth promoters have been banned from finishing feed for pigs. The main risk factor, which is the result of this study, is the strong protective effect of the feeding of fermented liquid feed to pigs against Salmonella infections in these pigs. This is of major importance for the pork production chain. Consumers want products that are safe to consume and governments are increasing the pressure on producers to take their responsibility with regard to the (microbiological) quality of their own products. To be able to guarantee the (microbiological) quality of pork, e.g. free from Salmonella, the entire production chain starting with feed ingredients through to the retailer will have to be involved in the quality assurance chain. The Major source of Salmonella contamination of pork is the healthy carrier of Salmonella: the slaughter pig. Risk factors with an association to the Salmonella infection status of finishing pigs were studied in this project. It was found that liquid feed containing fermented co-products from the human food industry had a strong protective effect for Salmonella infection. This is important for three reasons: - Salmonellosis is still one of the main causes of gastroenteritis in humans and pork is an important source after broilers and eggs; - Use of these co-products in the pig industry as feed ingredients improves pig health and production; - Recycling of co-products from the human food industry avoids dumping or burning and therefore loss of valuable nutrients and prevents an environmental burden.

Although the zoonotic Salmonella types can occur in almost all food-producing animals, there are often rather strong associations between certain types and a particular animal reservoir. Using Danish estimation principles to quantify sources of human salmonellosis, an attempt was made to assess the role of pork in Denmark, The Netherlands, Germany, Sweden and England and Wales. A comparison of Salmonella types isolated from animals and food with isolates from humans makes it possible to produce estimates of the number of human cases attributable to certain animal sources. It is prerequisite that some of the predominating Salmonella types are found almost exclusively in a single animal reservoir. Further, it is assumed that all human infections with these so-called "typical" types originate only from that particular source. Human infections caused by Salmonella types that are found in several reservoirs, may then be distributed in proportion to the occurrence of the typical types, assuming that human infections caused by different Salmonella types have equal chance of appearing in the national Salmonella statistics. Pork was estimated to be associated with 10 to 23 % of human salmonellosis in Denmark, The Netherlands and Germany, with some variation between countries. Pork was presumably of minor importance in England and Wales and negligible in Sweden. No estimates could be made for the Greek situation. It was estimated that in Denmark, 10 to 15% of human infections could be attributed to pork and pork products. In The Netherlands, pork was assumed to be responsible for 14 to 19% of human cases of salmonellosis. This corresponds well with other estimates of the proportion of pork related cases in The Netherlands. In Germany, 18 to 23% of human cases were assumed to be attributable to pork. For England and Wales it was concluded that the majority of the S. Enteritidis cases in England and Wales in 1997 could be attributed to eggs and poultry, whereas the role of pork presumably was of minor importance. However, a proper quantification of the total number of pork-related cases was not attempted. Data obtained from Greece on the occurrence of Salmonella and the distribution of sero- and phage types in animals, food and humans in 1997 was very sparse and unrepresentative. Therefore, no attempt to quantify sources of human salmonellosis was made. The prevalence of Salmonella in Swedish pig herds and pork in 1997 was extremely low. In 1997, no cases of human salmonellosis could be traced to Swedish pork. It was therefore concluded, that the number of cases attributable to Swedish produced pork in 1997 was negligible.

To estimate herd incidence and prevalence of Salmonella infections, as well as their seasonal fluctuations, seropositive and seronegative herds were followed over a two-year period, by serological and bacteriological testing. This study was performed in Germany, Denmark, The Netherlands and Sweden. The stability of an initially allocated Salmonella status, was found to vary noticeably, apparently irrespective of a seropositive or seronegative classification at onset of the study. Sixty-two per cent of herds in this study shifted from their initial status at least once during the observation period. Steady-status periods varied from a month to more than two years. Pig herds are not a closed production systems and contact with the outside world is unavoidable. This contact includes many factors that can lead to a loss of Salmonella status, such as introduction of feed and animals, changes in feed or management strategy or contact with visitors, pets, rodents, insects or wildlife. However, herds that are consistently negative over longer period of time than in this study are not uncommon in most pig producing countries. It is difficult to find a general practice for these herds which could explain why they remain negative, especially since management strategies, commonly accepted as risky with respect to Salmonella infections have no consequences in the absence of Salmonella. Given the dynamic nature of pig production it could be stated that one cannot base a Salmonella herd-status on results of a single sampling round. Regular testing is necessary to enable producers, advisors and authorities to react to a sudden increase in the Salmonella prevalence in a herd or on a national level. Overall, 12 out of 32 herds (38%) continued to have the same Salmonella status as appointed at the start of the study during the observation period. Among the 6 herds, which remained negative, 3 herds were sampled over the entire 2-year period, while sampling was stopped prematurely for the remaining 3 herds (after 12-18 months). Of the 17 seropositive herds, 6 herds (35%) remained positive. However, 2 of the positive herds were not sampled over the entire 2-year period (respectively 6 and 15 months), and 3 herds were missing 3 to 5 sampling rounds. Six out of 9 seronegative herds, which became seropositive, returned to their negative status after 3 to 7 months. Three negative herds changed status already in the first round, 2 of them where back to their negative status within 2 sampling rounds. In 3 herds, the change to a positive status was accompanied by Salmonella isolation from pen faecal samples. However, in 1 negative herd a S.Derby isolation in the grower unit in one sample round was not reflected by a serological response, but a S.Typhimurium isolation in the finishing unit 5 months later was. The period of time it took for 8 seropositive herds to become seronegative during the observation period varied from 2 to 22 months. LA Salmonella herd-status should not be based on a single sample round. Regular testing is necessary to enable producers, advisors and authorities to react to a sudden increase in the Salmonella prevalence in a herd or on a national level, as well to give producers with a Salmonella positive status the opportunity to obtain a negative status after control measures have been successfully applied.

To investigate whether purchasing pigs constitutes a risk for introduction of Salmonella, a source analysis study was performed on a total of 97 farrow-to-finish herds from 5 European countries, being Germany, Denmark, Greece, The Netherlands and Sweden. In these herds, blood samples were taken from 20 sows, and pooled faecal samples in 20 pens from ready-to-ship feeders. Only in Denmark, a significant correlation was found between the proportion of seropositive sows and the proportion of positive pen faecal samples in feeder pens. The interesting results from the SALINPORK project made the Federation of Danish Pig Producers and Slaughterhouses expand their part of the project to include an additional 49 herds, and to look into risk factors for isolation of Salmonella in pen faecal samples from feeder pigs. It was shown, that use of pelleted feed was a risk factor for isolation of Salmonella in pen faecal samples. Furthermore it was shown that finisher herds, that received pigs from Salmonella positive herds, had a significantly higher seroprevalence, than finisher herds that received pigs from Salmonella negative herds. Based on these results, the Federation of Danish Pig Producers and Slaughterhouses are implementing a Salmonella certification system for sow to feeder herds, in order to make it possible for finisher herds to demand Salmonella-free pigs from farrow to feeder herds. In the Danish part of the study, where an additional 49 herds were included, the following results were obtained: -Salmonella was isolated from pen samples in 13 (19%) of 69 herds. With the exception of feeders in the age group of four weeks, Salmonella was isolated to the same extent in all age groups. -Fifty-three of 1371 (3.9%) sows were seropositive using the mixed-ELISA with a cut-off of 40 OD%. Seropositive sows were evenly distributed according to parity. Comparing microbiological and serological results, a strong association between a seroprevalence >10% in sows and occurrence of Salmonella typhimurium among feeders was found. A similar association between other Salmonella serotypes among weaners and serology could not be demonstrated. -Among those potential risk factors provided by the questionnaire, feeding purchased, pelleted feed to sows seems to be associated with a higher risk of detecting the presence of Salmonella among weaners. We found, that isolating Salmonella in feeders is a risk factor for high seroprevalence in finishers, and that the feed factors (use of pelleted, purchased feed) found in previous studies on finishers, were important also on the sow-level.

In this study, Salmonella isolates were mainly found in highly seroprevalent herds (>50% seropositive samples). Still, some discrepancy between test results was observed. In general, an apparent false positive serological result can either be caused by cross-reactivity towards other infections, employing a low cut-off value in a test with relatively high background densities (noise) or a false negative reaction in the gold standard test. It is commonly acknowledged that the sensitivity of culture methods in a practical setting is often quite low. Therefore, a serologically positive result, which in this comparison is classified as false-positive, may very well represent a real infection which was not detected by bacteriological sampling. In contrast to culture methods, antibodies against Salmonella can be detected from non-shedding carrier pigs, provided the O-antigens of the Salmonella serotype are included in the test. These antibody titres may or may not reflect a current infection of the individual animal, but they do indicate exposure to Salmonella during production. Therefore, as a measure of the presence of Salmonella on a herd-level, serological testing can be regarded as both a sensitive and practical method. Based on our findings, we concluded that there is a positive correlation between isolation of Salmonella in a herd and seropositivity for all countries in this study. Using serology to determine the Salmonella status, the test and herd cut-off can be adapted to serve the purpose of sampling best. A low herd-level cut off may be chosen in a situation where it is important to detect most infected herds (e.g. from a food safety perspective) hereby keeping the number of false negative herds low and accepting a relatively high proportion of false positive herds. On the other hand, in a situation where a false positive result has large financial consequences, a higher herd-cut off value may be more appropriate (e.g. from a production perspective). It is recommended that each country quantifies the nature of the relationship between bacteriological and serological results for their own situation. Modern serological techniques have proven to be a convenient and cost efficient method for testing Salmonella infection in pigs. Though serological results can not distinguish between current and past infections, they indicate exposure to Salmonella at one stage of production prior to sampling. In comparing serological results to bacteriological results, it is important to remember that, though both are measures of infection, they measure different stages of infection. Serology is a measure of historical exposure which may not correlate closely to the microbiological burden at the time of sampling. In contrast to culture methods, antibodies against Salmonella can be detected from non-shedding carrier pigs, provided the O-antigens of the Salmonella serotype are included in the test. Therefore, as a measure of the presence of Salmonella on a herd-level, serological testing can be regarded as both a sensitive and practical method.

Salmonella can enter the food chain at any point throughout its length, from livestock feed, via the on-farm production site, at the slaughterhouse or packing plant, in manufacturing, processing and retailing of food, through catering and food preparation in the home. In order to obtain, reduce and control the occurrence of Salmonella contamination at various levels of pork production in relation to human salmonellosis, an EU-wide surveillance and control programme is needed. A minimum requirement of an EU-wide approach of monitoring the time trends of Salmonella infections throughout production and in humans should be to design a system which collects standardised data for comparison between EU member states, and monitors the effectiveness of control measures taken. The process of planning and executing a Salmonella surveillance and control programme can be divided in 5 phases: 1) Orientation - assessing the Salmonella situation by a bacteriological and/or serological screening to estimate the Salmonella prevalence at the pre-harvest and harvest level of pork production. 2) Preparation - identifying or establishing the prerequisite elements needed to implement a surveillance and control programme. 3) Implementation - commencing the mobilisation of the relevant authorities and institutes, and routine data collection, data management and data communication. 4) Evaluation - National and EU authorities evaluate gathered data to determine the present Salmonella status in relation to previous assessments as well as the achieved goals and consider if any adjustments of definitions, assumptions or sampling schemes are necessary. 5) Modification - adjustments are made towards optimising and updating the programme. The revised programme is subsequently implemented. This continuous process of evaluation, modification and implementation is necessary to adapt to changes in both local and global Salmonella epidemiology. The general approach proposed here follows the stable-to-table principle, where every link in this chain is responsible for reducing, controlling and preventing the Salmonella contamination of its product, which serves as the basis for the production in the next link. Elements of such a monitoring and intervention system for the control of Salmonella in pig production is discussed and proposed. Phase I: Orientation - The orientation phase should ideally result in the following estimates: -Serological prevalence of Salmonella infected pig herds. -Bacteriological prevalence of Salmonella infected pig herds, and distribution of Salmonella sero- and phage types and resistance patterns. -Bacteriological prevalence of Salmonella contaminated carcasses at the slaughterhouse level, including distribution of Salmonella sero- and phage types and resistance patterns. -Bacteriological prevalence of Salmonella contaminated pork and pork products at retail level. -Incidence of reported Salmonella infections in humans, including distribution of Salmonella sero- and phage types and resistance patterns. Phase II: Preparation - The aim of this phase is to identify or establish the prerequisite elements needed to implement a surveillance and control programme. These elements are: -Well defined objectives, and long- and short-term goals. -A unique herd identification system, laboratory facilities, and central data base facilities. -Exploring the possibility of incorporation in existing disease control programmes. -A sampling scheme containing a description of sample material, sample size, sample frequency, and sample location. -Definition of threshold values at which predetermined control measures are activated. -A task and responsibility structure encompassing relevant authorities and institutions for sampling, data-management, analysis, interpretation and reporting, and implementation of intervention measures. Phase III.: Implementation - In this phase the relevant authorities and institutes are mobilised, and routine data collection, data management and data communication commences. Producers exceeding the threshold value, as defined in phase II, will be required to implement control and intervention measures in order to reduce the Salmonella occurrence to an acceptable level. Phase IV: Evaluation - National and EU authorities evaluate gathered data to determine the present Salmonella status in relation to previous assessments and consider if any adjustments of definitions, assumptions or sampling schemes are necessary. This phase should also be used to evaluate the achieved goals, and if appropriate new goals and recommendations are formulated. Phase V: Modification - Based on the results and recommendations of the evaluation performed in phase IV (i.e. trends of the Salmonella status, assessment of efficiency and efficacy of the system and new technological developments), adjustments are made towards optimising and updating the programme. The revised programme is subsequently implemented. This continuous process of evaluation, modification and implementation is necessary to adapt to changes in both local and global Salmonella epidemiology e.g. shifts in predominant serotypes or antimicrobial resistance patterns.

A new serological test the AQ mix-ELISA detecting Salmonella Typhimurium and Salmonella Choleraesuis antibodies was developed by the Danish Veterinary Institute in collaboration with Exiqon A/S using a technology owned by Exiqon A/S (http://www.exiqon.com). LPS derived antigens were chemically coupled to a photo-reactive anthraquinone-derivative and coupled covalently to microtiter plates by irradiation of UV-light, resulting in a stable and durable diagnostic surface. Comparisons of the new test with the mix-ELISA showed a high correlation between the two tests. Furthermore the pre-coated AQ mix-ELISA plates had a high reproducibility and durability. The high reproducibility of the AQ mix-ELISA, the robustness and the easy handling of the plates makes this test highly suitable for screening programs, where a uniformity of results are required over a long-term period. The fact that the plates can be manufactured centrally, stored for lengthy periods and distributed by ordinary mail also makes it useful for standardization and comparison of assays between different laboratories. As a further development of the present test for detecting Salmonella antibodies (the mix-ELISA) a new serological test the AQ mix-ELISA, was developed. By using a new technology to attach modified Salmonella antigens to microtiter plates, a stable and durable diagnostic surface was produced. For further details see Jauho et al 1999 and Wiuff et al. 1999b. The Danish Agency for Trade and Industry additionally supported the project. Sensitivity and specificity of the AQ mix-ELISA. A panel of 40 sera collected from Danish multiplying pig herds was tested repeatedly in the AQ mix-ELISA and compared to similar testing in the mix-ELISA. The samples were classified as sero-negative and sero-positive, using the experimental cut-off at 10 OD% and the monitoring cut-off at 40%, respectively. Based on the classifications in the mix-ELISA the sensitivity of the AQ mix-ELISA was 100% and the specificity 93% at the 10 OD% cut-off for both tests. Using the monitoring cut-off at 40 OD%, the sensitivity was 100% and the specificity 86%. Correlation of the AQ mix-ELISA with the established mix-ELISA. The panel of 40 serum samples was tested 10 times during one month in the AQ mix-ELISA and the mix-ELISA, respectively. The association between the two sets of OD-values obtained in this trial was strong with a rank correlation coefficient of 0.92 (p<0.001). The sample standard deviations between test-runs were at similar levels in the two ELISAs, not exceeding 15%. Durability of the ready-to-use AQ mix-ELISA plates. To test for the stability of the coated AQ mix-ELISA plates, these plates were repetitively tested with the 7 reference sera over a period of 6 months. It was found that the non-calibrated OD-values of these sera were stable throughout this period with a maximum coefficient of variation of 10% for the positive sera and 29% for the negative serum. The variation coefficients were equal to or lower than those obtained by tests of the same sera during periods of 1 or 2 weeks. The reproducibility of the diagnostic surface in the AQ mix-ELISA. Using automated production machinery made a bulk production of 10.000 AQ mix-ELISA plates. The performance of the test was examined during 3 test periods of each 3 weeks, by applying a panel of 47 sera (15 sero-negative, 32 sero-positive) +1 blank to every plate. During each period 28 plates were tested. Estimates for the average OD-level of a single sub-batch was calculated for each of the three test periods on a transformed scale with a "true value" at 1.00. The results strongly indicated a high degree of uniformity, a constant reproducibility and stability over time of the diagnostic surface of the AQ mix-ELISA plates.

The Salmonella ELISA is a serological diagnostic test used in pigs, which is capable of detecting antibodies against Salmonella, which are formed after infection with Salmonella in pigs. The ELISA is based on LPS containing the O-anitgens 1,4,5,6,7 and 12 and is capable of detecting infections by Salmonellae belonging to the serogroups B, C1 and D1. About 90% of the Salmonella occurring in pigs in The Netherlands and in Denmark belong to these serogroups. Inoculation and vaccination studies with other (related) pathogens (Yersinia, E. coli, Campylobacter, Pasteurella, Bordetella, Pseudorabies, mange) indicate that this ELISA is highly specific (specificity approaches 100%). Serum, meat drip or colostrum can be used as test material in this ELISA. This ELISA is ideally suited to screen groups of pigs for the presence of Salmonella infection. On an individual animal level, a positive result does not necessarily mean that the animal is still infected. Antibodies can be demonstrated long after the infection can be eliminated by the animal. The screening of herds for the presence of Salmonella infection can for example be part of an intervention program to reduce the number of Salmonella infected slaughter pigs entering the slaughterhouse. By reducing the number of Salmonella infected animals entering the slaughterhouse, a reduction of the contamination of pork and pork products can be reached which can result in a reduced risk of Salmonella infection in people consuming pork and pork products. The ELISA described above has been used in this joint research program to estimate herd and population prevalence for breeding sows, replacement gilts, multiplying sows and finishers. Serological herds statusses of finishing herds have been used to perform risk factor analysis for Salmonella infection in finishing herds. Longitudinal studies in finishing pigs using serological screening of herds have been performed. Inoculation studies with different serotypes / serogroups of Salmonella have been performed to test the sensitivity of the test. Specificity testing with sera from several inoculation and vaccination trails has been performed.

The standard method to detect Salmonella positive pigs is bacteriological examination. In the last years the use of the Salmonella-LPS-Mix-ELISA became available in order to screen pig herds on their Salmonella status. Seeder pigs were used to test the the transmission of Salmonella and to test the feasibility of Salmonella-LPS-Mix-ELISA . The following Salmonella serovars were used in fattening pigs: S. typhimurium, S. brandenburg, S. panama, S. livingstone, and S. goldcoast. The transmission of these strains turned out to be different. The Salmonella-LPS-mix-ELISA proved to be useful in detecting S. typhimurium and S. brandenburg, and doubtful for S. livingstone, S. panama and S. goldcoast. In order to elucidate which phases of the pork production chain contribute to the Salmonella contamination of pork, research was focussed on the last part of the pigs lives (transport from farm to slaughterhouse, waiting period in the lairage) and the slaughter phase. The slaughter line was the most important contamination source for carcasses of pigs from both sero-positive and sero-negative herds. The lairage was the most important contamination source for livers, tongues, rectal contents, mesenterial lymphnodes and tonsils of pigs from sero-negative herds and contributed to about 75% of the Salmonella contamination of pork. The farm was the most important contamination source for all samples, except carcasses, of pigs from sero-positive herds. This knowledge can be used to set up a chain control programme for Salmonella on pork. An important source of Salmonella contamination of pork is the healthy carrier of Salmonella: the slaughter pig. Pigs that were originating from a Salmonella-infected farm have a higher chance to end up as Salmonella contaminated pork and are also a substantial risk for the contamination of the lairages, the slaughter line and pigs from other herds. The high contamination of the drain water in slaughterhouses, from which many Salmonella serotypes were isolated, shows that there is a continuous flow of Salmonella through the slaughter line, most obvious originating from the slaughtered pigs. A large part of the Salmonella contamination of pork could be attributed to the lairage. Herd serology was significantly associated with Salmonella in rectal contents and lymphnodes of the pigs, this parameter can be useful to distinguish between Salmonella risks of herds and farms. To control Salmonella at pig farms it is necessary to detect all the Salmonella serovars antibodies which can be present in the pig. The most common serovar are detected by the Salmonella LPS Mix ELISA. However, not all serovars can be detected and more research needs to be done to detect these serovars.

Information on exposure to Salmonella through feed and the differences between feed types with respect to Salmonella occurrence is of great importance for Salmonella control strategies at the herd level of pig production. Sampling of feed was performed in pens with fattening pigs as close to slaughter weight as possible. The samples were collected from newly supplied feed at the top of the feed container or at the feed outlet in the pen and analysed for the presence of Salmonella. Feed samples taken at trough level or in the herd are not just representing the Salmonella status of feedstuffs leaving the feed mill, but also any possible occurrences of contamination during transport and storage or in the feeding system at the farm. Therefore, these samples can be considered a measure of what pigs actually are exposed to. Salmonella was isolated from both pelleted and non-pelleted feed, as well as both wet (though not fermented) and dry feed. Since Salmonella could be isolated from feedstuff samples, it can be concluded that pigs can be exposed to Salmonella through feed. However, it is not possible to differentiate between a new Salmonella infection introduced into the herd through feed and a re-infection of feed somewhere between storage and the feeding trough with this sampling protocol. The critical control points for contamination of especially pelleted and heat treated feed between production and the feeding trough in pig pens should be investigated. Another question which remains to be answered is the relative importance of feed as a source of Salmonella, given the fact that the serotypes of Salmonella isolated from feed are frequently not those found most commonly in animal populations nor human cases. The results indicated above point out the need for further investigation into the role of feed (type) as a potential source of Salmonella introduction into a pig herd. Efforts to decontaminate feed from Salmonella at the feed mill, by pelleting and heat treatment, are without benefit if recontamination takes place before the feed is served to pigs. Though there is a general consensus among researches involved in Salmonella epidemiology that feed has the potential of introducing Salmonella into a herd, the relative importance of this factor is difficult to assess. Both SALINPORK and other research has shown that non-heat treated feed types have a quality which, to some extend, protects pigs from infection if Salmonella is present in the herd compared to heat treated and pelleted feed. Understanding the mechanics behind this phenomenon can prove vital in an effective strategy of increasing pig resistance through a change in feed management as part of Salmonella control measures at the herd level.

Exploitable results

To estimate herd incidence and prevalence of Salmonella infections, as well as their seasonal fluctuations, seropositive and seronegative herds were followed over a two year period, by serological and bacteriological testing. This study was performed in Germany, Denmark, The Netherlands and Sweden. The selected herds were followed by 7 subsequent serological testing rounds. At each testing round, 25 blood samples from slaughter pigs close to slaughter and 25 blood samples from growers (with a minimum weight of 20 kg) were taken. In addition, 10 faecal pen samples of approximately 25 grams, each representing 5 pigs, were taken per visit. Testing was done approximately 3 months apart, so that each herd was followed for 2 years. To study the development of the seroprevalence within batches of pigs, pens with growers which were sampled in one round were to be sampled as pens with finishers 3 months later. In herds with a high proportion of seroreactors among growers (> 0.3), a high proportion of seroreactors among finishers in the next sampling round was found. Therefore, a high proportion of seropositive growers in a herd can be regarded as indicative for a high proportion of seropositive finishing pigs in the next sampling round. However, in herds with low proportions of seroreactors among growers, both low and high proportions of seropositive finishers were observed. In this study, two distinct infection scenarios were found: the first showing a positive correlation between the proportion seropositive growers and -finishers, the second where the serological status of finishers appears to be independent from a negative or low serological status of growers. In herds with a high proportion of seroreactors among growers, only high proportions of seroreactors among finishers were found. Therefore, a high proportion of seropositive growers in a herd can be regarded as indicative for a high proportion of seropositive finishing pigs. At least in these herds, antibodies from an infection contracted in the grower unit can likely still be measured in the finishing unit. However, one can not be sure that the response measured in the finishing unit are residual antibody levels from an earlier infection or a new infection obtained in the finishing unit. Nonetheless, such a situation reflects an infection that is well spread throughout the herd. In herds with low proportions of seroreactors among growers, both low and high proportions of seropositive finishers have been observed. This could mean that in those cases, infection first occurs in the finishing unit, indicating a source of Salmonella related to the finishing unit (e.g. contamination of the environment, different feed or older pigs).

A risk factor analysis was performed on the data collected from 358 slaughter pig herds and their Salmonella-status based on serum. Feed factors play an important role with respect to Salmonella. Pigs fed non-pelleted feed, wet or dry, were less likely to test seropositive, compared to pigs fed pelleted feed. The protective effect of non-pelleted feed over pelleted feed is believed to be due to poor growing conditions for Salmonella on non-pelleted feed in combination with the increased coarseness of the feed. Also, pigs who got whey were found to be less likely to test seropositive than pigs not getting whey. Pigs produced in batches, in herds with a hygiene-barrier facilities were less likely to test seropositive than pigs in herds where only one or neither factor was present. In herds where the caretaker did not wash hands consistently before taking care of the animals, pigs were more likely to test seropositive than pigs in herds where the caretaker did. Whether this factor is a measure of the risk of introduction and spread of infection via hands or of the general mentality towards hygiene is open for discussion. Pigs, which were able to have snout, contact with pigs from neighbouring pens, because pen separations were open or too low, were more likely to test seropositive compared to pigs for which such contact was prevented. Closed pen separations of sufficient height prevent the exchange of faeces and urine between pens and thus the spread of infection and contamination. Pigs in herds recruiting from more than 3 supplier herds were more likely to test seropositive than pigs in herds breeding their own replacement stock or recruiting from a maximum of 3 supplier herds. More sources of stock increase the probability of introduction of infection through one of these contacts. The strongest risk factors found in this study are related to feed. Other studies have also shown that the probability of detecting (sero) positive animals can be significantly reduced by feeding non-pelleted feed, preferably wet so that fermentation of feed can take place. Other important factors relate to hygienic precautions that can be taken in a herd. These results indicate that control of Salmonella at the herd level is a combination of minimising exposure and maximising resistance of pigs to Salmonella. Batch production in combination with the use of hygiene-barrier facilities (including washing hands and changing clothes and boots) could be made mandatory for all European (indoor) pig production to avoid introduction of Salmonella, as well as other pathogens, to a herd. In addition, the number of supplier herds should be kept to a minimum, with a maximum of 3 suppliers. In the situation where Salmonella is present in the herd, some form of acidification of feed or drinking water, be it through fermentation of the feed or the addition of organic acids or whey, can be used as an intervention strategy. Closed pen separations of sufficient height are useful in the prevention of spread of infection between pens. In combination with cleaning between batch production, the formation of re-infection cycles on a pen level can be avoided.

Searching for OpenAIRE data...

Deliverables

Exploitable results

Share this page

Download