Journal of Applied Microbiology 2003, 95, 891–903

doi:10.1046/j.1365-2672.2003.02042.x

Salmonella on pig carcasses: positive pigs and cross contamination in the slaughterhouse N. Botteldoorn, M. Heyndrickx, N. Rijpens, K. Grijspeerdt and L. Herman Department for Animal Product Quality and Transformation Technology, Centre for Agricultural Research-Ghent, Brusselsesteenweg, Melle, Belgium 2003/0012: received 8 January 2003, revised 2 May 2003 and accepted 4 May 2003

ABSTRACT N . B O T T E L D O O R N , M . H E Y N D R I C K X , N . R I J P E N S , K . G R I J S P E E R D T A N D L . H E R M A N . 2003.

Aims: The purpose of this study was to investigate the prevalence of Salmonella in pigs at the moment of slaughter and in the slaughterhouse environment. Methods and Results: In total, five different commercial slaughterhouses were sampled during eight slaughterhouse visits. Carcass swabs, colon content and mesenteric lymph nodes were taken to reflect the animal status and from the slaughterhouse environmental samples were taken. Salmonella was isolated from 37% of the carcass samples as a mean value. High variations were noticed between different slaughterhouses (between 0 and 70%) and sampling days in the same abattoir (between 3 and 52%). A correlation was found between the carcass contamination and the status of the delivered animals (P ¼ 0Æ01675). Cross contamination was estimated to account for 29% of the positive carcasses. The slaughterhouse environment was highly contaminated; before starting the slaughtering activities 25% of the samples were positive on average. The most prevalent serotypes isolated at the slaughterhouse environment and from the colon content were S. Typhimurium, S. Livingstone and S. Derby. On carcasses S. Typhimurium was predominately isolated (71%). The biggest variability of serotypes was found in the mesenteric lymph nodes. Serologically 56Æ3% of the pigs were found positive for Salmonella using a cut-off level of the optical density percentage higher than 10 (O.D.% ‡ 10). While on individual pig level the correlation between the bacteriological and serological data was poor, because of recent Salmonella infections, a better correlation was found at the herd level on the moment of slaughtering. Conclusion: A high degree of carcass contamination is noticed after slaughtering. This contamination resulted from the delivery of Salmonella-positive pigs and cross-contamination from the slaughterhouse environment. Significance and Impact of the Study: In pigs, Salmonella carriage is high, but it is obvious that slaughterhouse hygiene is a determinative factor for managing carcass contamination. Keywords: contamination source, pig carcass, Salmonella, serotyping, slaughterhouse.

INTRODUCTION Salmonella is a well-known human pathogen causing gastroenteritis. In the last decade there has been a dramatic Correspondence to: N. Botteldoorn, Department for Animal Product Quality and Transformation Technology, Centre for Agricultural Research-Ghent, Brusselsesteenweg 370, 9090 Melle, Belgium (e-mail: [email protected]).

ª 2003 The Society for Applied Microbiology

increase in the number of human cases of salmonellosis in most western countries. In 2001 Salmonella enterica subsp. enterica serovar Enteritidis and serovar Typhimurium caused, respectively, 64Æ2 and 21Æ5% of the Belgian human salmonellosis cases (Belgian National Reference Centre for Salmonella and Shigella, Brussels, Belgium). In different epidemiological studies pork has been described as a source for human salmonellosis (Molbak and Hald 1997; Delpech

892 N . B O T T E L D O O R N ET AL.

et al. 1998; Pontello et al. 1998; Murase et al. 2000). In Denmark, the Netherlands and Germany it was estimated that 15–20% of all human cases of salmonellosis were associated with the consumption of pork (Borch et al. 1996; Berends et al. 1998; Steinbach and Kroell 1999; van Pelt et al. 2000). Within the serovar S. Typhimurium, the phage type DT104 is of particular concern because of its acquisition of multiple antibiotic resistance (Threlfall et al. 1994; Wall et al. 1994; Baggesen and Aarestrup 1998). Salmonella has the potential to colonize the gut of the pigs. Most often pigs are healthy carriers of Salmonella. With the exception of infections with S. Choleraesuis and some types of S. Typhimurium, Salmonella infections usually produce no severe disease in pigs. Salmonella Typhimurium is frequently associated with Salmonella infections of pigs (Daube et al. 1998; van der Wolf et al. 1999). The major contamination sources of pig carcasses are pig (faeces, pharynx and stomach) and environment related (contact surfaces and handling by workers) (Borch et al. 1996). The aim of different surveillance programmes is to improve our understanding of the sources and the extent of the contamination. Reduction of Salmonella in pork and pork products should include monitoring and intervention at farm level (Mousing et al. 1997). The widespread occurrence of Salmonella at the herd level suggests that control at this level is difficult (Fedorka-Cray et al. 1994). The monitoring systems applied in the Danish and German Salmonella control programmes are based on the serological examination of blood samples or meat juice samples. In these programmes farms are classified into three different contamination levels on the basis of the antibody levels measured by a specific Salmonella ELISA (Mousing et al. 1997). This study describes the contribution of different Salmonella contamination sources on pig carcasses. Environmental and carcass samples were taken to reflect the handling and sanitary practices in different slaughterhouses. In addition, colon content and mesenteric lymph node samples were taken to determine the prevalence of Salmonella in pigs at slaughter and the animal-related transmission.

M A T E R I A LS A N D M E T H O D S Sampling methods Experimental design. Samples were collected in five commercial slaughterhouses denoted as A, B, C, D and E (Table 1) processing ca 300, 500, 500, 600 and 400 pigs per hour, respectively. Figure 1 is a schematic ground plan of a slaughter line, valid for all five slaughterhouses. They were located in the western part of Belgium, where a highdensity pig production area is situated. Slaughterhouses A, C and D were visited twice; visits were indicated as A1, A2, C1, C2, D1 and D2. At A2 and D1 samples were

collected during a whole slaughter day; pigs originating from 28 and 38 farms were sampled, respectively. At A1, B, C1, C2 and E, pigs originating from one farm, slaughtered during the morning were sampled. Only in slaughterhouse E were the pigs slaughtered at first. In D2, the slaughterhouse environment was sampled intensively, including different carcass swab samples at different points in the slaughter line. Sample collection. Environmental samples were taken at different places in the slaughterhouse (Fig. 1, Table 1) to determine the Salmonella prevalence before and during slaughter activities.Totally 190 environmental samples were taken during eight slaughterhouse visits. Overshoes samples or swabs, with a sterile wad of cotton wool, from the lairage (C1 and D1) and from the dirty zone (A1, A2, D1, D2 and E) were taken. In abattoir A, C and D a sample of the scalding water was taken with a sterile collection tube and the temperature was measured. From the equipments swabs with a sterile wad of cotton wool was taken. The knife blade was swabbed from its tip to its base twice. The blade from the splitting machine was swabbed on both sites. In the clean zone the environment was sampled by taking overshoe samples at the beginning of the slaughter day (A2 and D1) and/or during slaughter (every slaughterhouse visit). Samples were collected in the chilling room by taking swabs of the floor and the wall. Colon content and mesenteric lymph node samples were taken to determine the supply of Salmonella-positive pigs to the slaughter line. For every slaughterhouse visit the slaughter pigs that were followed were marked after polishing at the beginning of the slaughter line. In A2 and D1 where samples were taken during a whole slaughter day, each 25th and 50th pig was marked, respectively. From these marked pigs different samples were taken along the slaughter line. At the point of evisceration a piece of the colon was taken and put in a plastic bag with the corresponding number of the pig. Also the mesenteric lymph nodes were cut out of the intestine packet and put in a plastic bag with the pig number on it. In total 345 faecal samples and 346 mesenteric lymph nodes were taken. A piece of the diaphragm was cut out at the point of the meat inspection and at the end of the slaughter line carcass swabs before and after chilling in the chilling room were taken, which reflected the contamination of the end product. Half of the carcasses were swabbed, according to the procedure described by Korsak et al. (1998). The inside and outside of the ham were swabbed together with the sternum along the incision line. Sterile gloves were used and changed after each sampling. In total 370 carcass swabs were taken at the end of the slaughter line. In slaughterhouse C the same carcasses before and after chilling were swabbed, at slaughterhouse D1 carcass swabs were taken at random.

ª 2003 The Society for Applied Microbiology, Journal of Applied Microbiology, 95, 891–903, doi:10.1046/j.1365-2672.2003.02042.x

PIG CARCASS CONTAMINATION WITH SALMONELLA DURING SLAUGHTERING

893

Table 1 Isolation of Salmonella in five commercial slaughterhouses on the different types of samples. All samples were taken during slaughter activities, unless otherwise indicated Slaughterhouse (+/total samples) Sample type Environment Lairage Dirty zone Scalding tank Rectal pistol Knives Splitting machine Overshoes before slaughtering Overshoes slaughter line Overshoes chilling room Carcass Carcasses Carcasses chilling room Pig Faeces MLN Faeces and/or MLN

A1*

A2

B*

C1*

C2*

D1

D2

E*$

Total x/y (%)§

ND 1/3 0/1 0/1 0/5 ND ND 3/4 1/1

ND 2/2 ND ND 1/5 ND 1/3 10/12 ND

ND ND ND ND 1/2 0/2 ND 7/7 ND

4/4 ND 0/1 0/1 3/5 1/1 ND 4/5 1/2

ND ND ND 0/1 1/3 0/1 ND 8/8 ND

7/7 11/12 ND ND 0/13 1/1 1/5 14/14 2/3

ND 19/19 0/1 1/2 5/8 1/6 ND 2/4 0/4

ND 1/1 ND ND ND ND ND 7/10 ND

11/11 (100) 34/37 (92) 0/3 (0) 1/5 (20) 11/41 (27) 3/11 (27) 2/8 (25) 55/64 (86) 4/10 (40)

0/14 ND

32/120 ND

5/26 ND

13/25 9/25

1/30 0/30

76/108 ND

2/30 3/20

9/17 ND

138/370 (37) 12/75 (16)

0/15 0/15 0/15

25/120 28/120 38/120

0/28 0/29 0/29

17/25 22/25 23/25

2/30 6/30 6/30

14/110 14/110 24/110

ND ND ND

7/17 3/17 7/17

65/345 (19) 73/346 (21) 98/346 (28)

*Sampling of one herd. Sampling during a whole day. At random sampling during the morning. $Herd sampled at the start of slaughtering. §x/y (%): positive/total samples examined (percentage of positive samples). ND, not determined; MLN, mesenteric lymph nodes.

All samples were put in sterile plastic bags and transported to the laboratory in a cooled container for immediate analysis. Salmonella isolation method. Carcass swabs and 5 g of faeces, obtained aseptically from each piece of colon, were incubated in 50 ml of buffered peptone water (BPW; Oxoid Ltd, Basingstoke, UK), environmental swabs were incubated in 100 ml of BPW, and overshoes in 225 ml of BPW. The lymph nodes were trimmed of any attached fat, meat or other tissue to keep the lymph node intact. Ten grams of an intact lymph node was homogenized in a stomacher bag with 50 ml BPW for 2 min. All the samples were incubated at 37C overnight. For the isolation of Salmonella, different methods were compared; two methods with a liquid enrichment step in Rappaport Vassiliadis (RV; 10 and 2 ml) and a method based on motility enrichment on the semisolid medium Diassalm (LabM, Bury, UK) (Table 2). In total, 1337 different samples were tested. Following enrichment 0Æ1 ml was transferred to 9Æ9 ml RV broth (Oxoid) and on diagnostic semisolid Salmonella agar (Diassalm). Additionally, for the samples collected in slaughterhouse A1, B, C1, C2 and E, 0Æ2 ml of BPW was

added to 1Æ8 ml of RV. After incubation at 42C for 24 h a 10-ll loop of the RV culture was streaked on xylose lysine desoxycholate agar (XLD; Oxoid) and incubated for 24 h at 37C. When a purple migration zone was observed on Diassalm after 24 h of incubation at 42C, a 1-ll loop from the suspected zone was streaked on XLD. Presumptive Salmonella colonies (black colonies) on XLD were confirmed by PCR (see below). All the Salmonella isolates were stored at )80C in brain–heart infusion (BHI; Oxoid) supplemented with 15% glycerol (w/v). Molecular confirmation and characterization of isolates Preparation of crude cell lysate. The bacterial colonies were suspended in 100 ll of H2O and centrifuged for 2 min at 13 000 g. The pellet was resuspended in 100 ll of 0Æ05 mol l)1 NaOH, 0Æ125% sodium dodecyl sulphate (SDS) and heated for 17 min at 90C. Molecular confirmation. For PCR confirmation of presumptive Salmonella colonies, Salmonella-specific primers ST11 5¢-AGCCAACCATTGCTAAATTGGCGCA-3¢

ª 2003 The Society for Applied Microbiology, Journal of Applied Microbiology, 95, 891–903, doi:10.1046/j.1365-2672.2003.02042.x

894 N . B O T T E L D O O R N ET AL.

Dirty zone

Clean zone

Lairage Overshoe samples Rectal pistol

Killing Evisceration

Colon and MLN

Polishing Scalding tank

Splitting

Veterinary inspection

Scalding water

Diaphragm

Dehairing Flaming

SKG-II and classification on carcass quality

Cleaning up the carcasses Swabs of the post-dehairing table Carcass swabs

Sampling points

Rapid cooler (-25˚C)

Car. swabs

Chilling room (4˚C)

Carcass swabs

Fig. 1 Schematic floor diagram of the different stages of habitual processing lines in the Belgian pig slaughterhouses

and ST15 5¢-GGTAGAAATTCCCAGCGGGTACTG-3¢ described by Aabo et al. (1993) were used. The PCR was performed in a final volume of 25 ll containing 50 mmol l)1

KCl, 10 mmol l)1 Tris–HCl (pH 8Æ3), 1Æ5 mmol l)1 MgCl2, 0Æ5% Tween 20, 0Æ01% gelatine, 200 lmol l)1 of each dNTP, 0Æ75 U of AmpliTaq DNA polymerase (Applera,

ª 2003 The Society for Applied Microbiology, Journal of Applied Microbiology, 95, 891–903, doi:10.1046/j.1365-2672.2003.02042.x

PIG CARCASS CONTAMINATION WITH SALMONELLA DURING SLAUGHTERING

Table 2 Sensitivity of the different methods used for the isolation of Salmonella

Samples

Medium

Total

895

Number of Number of Number of samples Sensitivity samples positive samples* positive with the medium (%)

Diassalm 1337 RV 10 ml 1337 RV 2 ml 240 Faecal samples Diassalm 345 RV 10 ml 345 RV 2 ml 57

408 408 109 65 65 24

389 344 75 58 48 11

95Æ3 84Æ3 68Æ8 89Æ2 73Æ8 45Æ8

*Number of positive samples with at least one method.

Norwalk, CT, USA), 0Æ25 lg of each primer and 1 ll of crude cell lysate. The mixture was subjected to 30 cycles of amplification; the first cycle was preceded by denaturation for 1 min at 95C. Each cycle consisted of a denaturation for 15 s at 95C, annealing for 15 s at 57C and elongation for 30 s at 72C. The last cycle was followed by a final extension for 8 min at 72C. The PCR products (x ll) were subjected to electrophoresis on a 1Æ5% Seakem ME agarose gel (FMC Bioproducts, Rockland, ME, USA) under 20 V cm)1 for 25 min and then analysed (Sambrook et al. 1989). REP (Repetitive extragenic palindromic elements)PCR. REP-PCR was used as a rapid fingerprinting method to classify Salmonella isolates to the serotype level. For REPPCR, the primers REP1R-I 5¢-IIIICGICGICATCIGGC-3¢ and REP2-I 5¢-ICGICTTATCIGGCCTAC-3¢ described by Versalovic et al. (1991) were used. The PCR reaction was performed in a total volume of 25 ll containing 50 mmol l)1 KCl, 10 mmol l)1 Tris–HCl (pH 8Æ3), 1Æ5 mmol l)1 MgCl2, 0Æ5% Tween 20, 0Æ01% gelatine, 200 lmol l)1 of each dNTP, 1Æ4 U of Goldstar RED DNA polymerase (Eurogentec, Seraing, Belgium), 0Æ25 lg of each primer and 1 ll of crude cell lysate. The mixture was subjected to 30 cycles of amplification; the first cycle was preceded by denaturation for 7 min at 95C. Each cycle consisted of a denaturation for 1 min at 94C, annealing for 1 min at 40C and elongation for 8 min at 65C. The last cycle was followed by a final extension for 16 min at 65C. The PCR products were subjected to electrophoresis on a 1Æ5% Seakem ME agarose gel (FMC Bioproducts) in a 1· Tris–borate–EDTA (TBE) buffer under 5Æ6 V cm)1 for 4 h and analysed (Sambrook et al. 1989). The REP-PCR patterns were clustered with GelCompar 4Æ1 (Applied Maths, Kortrijk, Belgium) in comparison with the known Salmonella serotypes. Serotyping One representative strain of each REP type was sent to the Veterinary and Agrochemical Research Center (CODA, Ukkel, Belgium) for conventional serotyping according to the Kauffmann–White scheme.

ELISA-detection of Salmonella-specific antibodies A piece of each diaphragm was frozen and the meat juice harvested after thawing. The meat juice was examined serologically for specific antibodies against Salmonella. The test kit Salmotype-ELISA (Labor Diagnostik, Leipzig, Germany) was used following the suppliers instructions. The Salmotype-ELISA detects the O-antigens 1, 4, 5, 6, 7 and 12, representing more than 90% of the Salmonella serovars commonly isolated from pigs. The test result is presented as the optical density of the sample, relative to the optical density of positive reference samples (O.D.%). As cut-off values for a positive result, O.D.% ‡ 40 as recommended by the kit, and O.D.% ‡ 10 as described by van der Wolf (2000) were used here. A sero-negative herd was defined as a herd with at least 80% of the samples negative (O.D.% ‡ 10) in the Salmonella ELISA, whereas a herd was classified as a sero-positive (O.D.% ‡ 40) when at least one sample from that herd had an O.D.% ‡ 40. Statistical analysis Animal-related contamination was determined as the total number of pigs positive in their faeces and/or mesenteric lymph nodes relative to the total of carcasses positive. The carcass samples from the same pigs as the faeces and lymph nodes were compared and the value was expressed as percentages. Association between carcass contamination and the contamination of faeces and mesenteric lymph nodes was assessed using a Pearson chi-square test. For the ELISA results, the correlation with the bacteriological results on herd level (faecal sample only) and on individual pig level (faeces only or faeces and mesenteric lymph nodes together) was determined by the Pearson chisquare test and the tetrachoric correlation. The tetrachoric correlation assumes that the ordered, categorical variables of a 2 · 2 frequency table have an underlying bivariate normal distribution. It is calculated as the maximum likelihood estimate of the product–moment correlation between the

ª 2003 The Society for Applied Microbiology, Journal of Applied Microbiology, 95, 891–903, doi:10.1046/j.1365-2672.2003.02042.x

896 N . B O T T E L D O O R N ET AL.

normal variables, estimating thresholds from the observed table frequencies. The range of the tetrachoric correlation is from )1 to 1.

RESULTS Sensitivities of the different isolation methods For the isolation of Salmonella, different methods were compared; two methods with a liquid enrichment step in RV (10 and 2 ml) and a method based on motility enrichment on the semisolid medium Diassalm (Table 2). A sample was determined positive when Salmonella was isolated with at least one of the isolation methods. Overall, the highest sensitivity (95Æ3%) was obtained with Diassalm. RV 10 ml gave a sensitivity of 84Æ3% whereas RV 2 ml gave a much lower sensitivity of 68Æ8%. In faecal samples a general lower sensitivity was obtained. Diassalm and RV 10 ml gave a sensitivity of 89Æ2 and 73Æ8%, respectively. RV 2 ml offered no additional sensitivity (45Æ8%), no samples were found positive exclusively with RV 2 ml. Prevalence of Salmonella Table 1 shows the prevalence of Salmonella in the environmental samples, the carcasses and the pig-related samples taken at the different slaughterhouses. Salmonella was present in the environmental samples of all the slaughterhouses, from 33% for slaughterhouse A1 to 75% in slaughterhouse C2. In the lairage and the dirty zone of different slaughterhouses the Salmonella prevalence was 100% to 92% for slaughterhouse C and D respectively, where the overshoe samples were obtained during one visit. However, the scalding water samples taken during slaughter activity were never positive. The temperature of the scalding water during activity was between 60 and 62C. An average of 25% of the overshoe samples taken at two different slaughter lines before the start of the slaughter activity were found positive. Overshoe samples taken during activity, at different points along the slaughter line, were positive in 86% of the cases. In four slaughterhouses, the knife samples were found positive (in overall 27% of the samples) and for one of the slaughterhouses (C ) the knives were found positive on two occasions. The carcass splitter was positive during three slaughterhouse visits (C1, D1 and D2). A total of 370 carcass swabs were taken and an average of 37% of them turned out to be positive before chilling. However, high variations were observed between the different abattoirs (ranging from 19% at B to 56Æ5% at D) and between the different sampling days at the same slaughterhouse (e.g. ranging from 3% at C2 and 52% at C1). After chilling, the carcass contamination of the same pigs originating from one herd decreased in C1 (52% before and

36% after chilling) and C2 (3% before and 0% after chilling). However, in slaughterhouse D2 an increase from 6 to 15% was noticed after chilling. In this case swabs were taken at random and not from the same carcasses before and after chilling. Overall, Salmonella was isolated from 19% of the faecal samples, taken from the colon after evisceration, and from 21% of the mesenteric lymph nodes. However, high variations were observed between the herds slaughtered in different slaughterhouses (ranging from 0% in A1 and B to 64% in C1 for faecal samples and from 0% in A1 and B to 88% in C1 for lymph nodes) and between different visits of one slaughterhouse (ranging from 6% in C2 to 64% in C1 for faecal samples and from 20% in C2 to 88% in C1 for lymph nodes). Overall 28% of the animals carried Salmonella in the faeces and/or the mesenteric lymph nodes. Distribution of the Salmonella serotypes The three most prevalent serotypes (Table 3) isolated from the slaughterhouse environment, by taking overshoe samples at the slaughter line before and during slaughtering, were S. Typhimurium (38Æ6% of the isolates, six of eight slaughterhouse visits), S. Livingstone (21%, three of eight slaughterhouse visits) and S. Derby (21%, seven of eight slaughterhouse visits). Also Infantis, Ohio and Brandenburg were isolated from overshoe samples. The overshoe analysis allowed the isolation of different serotypes from the environment, for example at A2 and B it was possible to isolate four different serotypes. Salmonella Typhimurium and S. Derby were the most prevalent serotypes in the dirty zone of slaughterhouses D1 (37Æ7% of the isolates) and D2 (31Æ1% of the isolates), respectively. The largest variation in serotypes was found in pig-related samples (Table 3). Nine different serotypes (Typhimurium, Brandenburg, Derby, Infantis, Virchow, Livingstone, Ohio, Anatum and London) were isolated from the carcasses. Salmonella Typhimurium was isolated from 71% of the contaminated carcasses and during five of eight slaughterhouse visits. At D1, where the carcass contamination was 70%, S. Typhimurium was almost exclusively found (97% of the isolates). At C1 additional serotypes S. Derby, S. Virchow and S. Mbandaka were found on the carcasses after chilling. At slaughterhouse D2 S. Typhimurium and S. Derby were found after chilling but for this specific case carcasses from different pigs were swabbed as before chilling. Salmonella Typhimurium was isolated from the colon content of 40% of the Salmonella-positive pigs, while this figure was 30Æ7% for S. Livingstone and 15Æ3% for S. Derby. Salmonella Livingstone, S. Typhimurium and S. Derby were isolated from 35Æ6, 26 and 10Æ9% of the mesenteric lymph nodes, respectively. In slaughterhouse C1,

ª 2003 The Society for Applied Microbiology, Journal of Applied Microbiology, 95, 891–903, doi:10.1046/j.1365-2672.2003.02042.x

Infantis (1)

A1

Carcasses chilling room

Overshoes chilling room Carcass Carcasses

Bovismorbificans (1)

Splitting machine Overshoes before slaughtering Overshoes slaughter Infantis (2) line during Derby (1) slaughter

Rectal pistol Knives

Dirty zone

Environment Lairage

Samples

Slaughterhouse C2

London(1) Typhimurium(5) Anatum(1)

D1

D2

Typhimurium (11) Derby (13) 4 :– : –(1) Derby (1) Typhimurium (1) Typhimurium (1) Derby (2) Brandenburg (1) NT (3) Livingstone (1) Typhimurium (1) Typhimurium (1) London (1) Derby (1)

Typhimurium(1) Livingstone(1) Virchow(1)

Brandenburg(1)

C1

Typhimurium (1)

Typhimurium (1) Derby (2)

Typhimurium (74) Derby (2) Infantis (1) Anatum (1)

Typhimurium (2)

Typhimurium (3) Typhimurium (3) Typhimurium (3) Typhimurium (11) Derby (2) Livingstone (1) Derby (1) Livingstone (4) Derby (3) Derby (1) Infantis (1) Infantis (2)

Bovismorbificans (1)

B

Typhimurium (11) Typhimurium (5) Typhimurium (5) Infantis (1) Brandenburg (2) Livingstone (8) Derby (7) Infantis (4) Virchow (1) Livingstone (1) Ohio (6) Typhimurium (2) Livingstone (4) Derby (1) Virchow (1) Mbandaka (1)

Livingstone (6) Typhimurium (1) Ohio (1) Derby (2)

Livingstone (1)

Brandenburg (1)

Livingstone (2)

A2

Typhimurium (4) London(1) Livingstone (2) Brandenburg (1) Derby (1)

London (2) Brandenburg (2) Derby (1) Typhimurium (1) Infantis (1)

Derby (1)

E

Table 3 Detail of the Salmonella serotypes isolated from the environmental and the pig-related samples in the slaughterhouse before and during the slaughtering process and number of isolates (between brackets)

PIG CARCASS CONTAMINATION WITH SALMONELLA DURING SLAUGHTERING

ª 2003 The Society for Applied Microbiology, Journal of Applied Microbiology, 95, 891–903, doi:10.1046/j.1365-2672.2003.02042.x

897

A1

Typhimurium (13) Brandenburg (4) Derby (2) (3)* Infantis (1) Rissen (1) Livingstone (2) Ohio (1) London (1) Anatum (1)* Typhimurium (9) Brandenburg (6) Derby (2) (1)* Infantis(1) Virchow(1) Livingstone (3) Ohio (3) London (1) Goldcoast (1) Poona (1)* Rissen (1)

A2

Slaughterhouse B

Livingstone (22)

Livingstone (17)

C1

Derby (4) Typhimurium (2)

Derby (2)

C2

NT: not possible to serotype with the conventional method. *The serotype of the strains isolated by the different isolation methods (RV and Diassalm) was different from each other.

MLN

Pig Colon content

Samples

Table 3 (Contd.)

Typhimurium (5) Anatum (5) Derby (1) Livingstone (1) Havana (1) Agona (1)

Typhimurium (6) Anatum (2) Derby (3) Livingstone (1) Panama (1) 47 : Z4Z23 : – (1)

D1

D2

Typhimurium (3)

Typhimurium (7)

E

898 N . B O T T E L D O O R N ET AL.

ª 2003 The Society for Applied Microbiology, Journal of Applied Microbiology, 95, 891–903, doi:10.1046/j.1365-2672.2003.02042.x

PIG CARCASS CONTAMINATION WITH SALMONELLA DURING SLAUGHTERING

Table 4 Comparison of the bacteriological status and the ELISA result at the individual pig level and on herd level after slaughter. Both cut-off values O.D.% ‡ 40 and O.D.% ‡ 10 are presented

Level

Bacteriological status (number of samples)

Cut-off ‡ 10

Cut-off < 40

Cut-off ‡ 40

131 52

217 29

43 36

142

183

246

79

Cut-off < 10

The Salmonella status defined on the faeces Pig )(260) 129 +(65) 13 Total

325

Herd

)(40)* +(27)

25 7

15 20

36 17

4 10

Total

67

32

35

53

14

The Salmonella status defined on the faeces and MLN Pig )(226) 117 109 +(97) 25 72

198 46

28 51

244

79

Total

323

899

142

181

MLN, mesenteric lymph nodes. *A herd was defined as bacteriologically positive from the moment that one pig of the herd was Salmonella positive in the faeces after slaughter.

(P ¼ 0Æ11463). The animal-related contamination was defined to be 70Æ5% on a total of 324 animals. Ninety-six animals were positive in their faeces or lymph nodes an a total of 136 positive carcasses. In slaughterhouses A and C a correlation between the number of living animals that shed Salmonella spp. in their faeces and/or positive in their mesenteric lymph nodes and the number of contaminated carcasses at the end of the slaughter line could be noticed (P ¼ 0Æ00611 for A2). In contrast, in slaughterhouses B and D1 a high carcass contamination was noticed whereas the delivery of Salmonella-positive pigs was nil or low. In the latter cases the impact of the cross-contamination of the slaughterhouse environment on the carcasses seems to be the most important factor (for B 100%, because all the pigs from that herd were Salmonella negative and for D1 in 68Æ5% of 100

Contamination frequencies

S. Livingstone was found in both the mesenteric lymph nodes, and the colon content. In other slaughterhouses additional serotypes were isolated; S. Goldcoast, S. Havana, S. Agona and S. Virchow were isolated from the mesenteric lymph nodes but not from faecal samples. More so, the first three serotypes were exclusively isolated from mesenteric lymph node samples and not from other animal-related samples or environmental samples of the slaughterhouse. At A2, C1, D1 and D2 the same serotypes were isolated from the carcasses as from knifes or the splitting machine. At the herd level, 27 of 67 (40Æ2%) herds sampled in the different abattoirs were found positive for Salmonella (Table 4). A herd was considered positive when a faecal sample from at least one pig of the herd was found positive. For each herd, an average of 5Æ149 pigs were sampled and in a positive herd an average of 2Æ44 animals were shedding Salmonella at the moment of sampling (result not presented). Salmonella Typhimurium, S. Derby, S. Brandenburg, S. Livingstone and S. Anatum were isolated from the faeces, in respectively 34, 22Æ8, 14Æ2, 8Æ6 and 8Æ6% of the positive herds. The serotypes Infantis, Ohio, London and Rissen were each found in 2Æ8% of the positive herds (result not presented). In six herds, the pigs were contaminated with more than one Salmonella serotype, in the other 21 herds all pigs were shedding the same serotype. Figure 2 shows the Salmonella contamination frequencies of carcass and faeces and mesenteric lymph nodes in the different slaughterhouse visits. A significant association was found between the numbers of animals that shed Salmonella in their faeces and/or were carrier of Salmonella in their mesenteric lymph nodes and the number of contaminated carcasses at the end of the slaughter line (P ¼ 0Æ01675). With the faecal sample alone no correlation could be found

Carcass Colon content Mesenteric lymph nodes

90 80 70 60 50 40 30 20 10 0 A1

A2

B

C1

C2

D1

E

Slaughter house Fig. 2 Salmonella contamination frequencies of carcasses, faeces and mesenteric lymph nodes in the different slaughterhouses

ª 2003 The Society for Applied Microbiology, Journal of Applied Microbiology, 95, 891–903, doi:10.1046/j.1365-2672.2003.02042.x

900 N . B O T T E L D O O R N ET AL.

the carcass samples the corresponding faeces or lymph nodes were negative). Serological analysis A positive ELISA result for anti-Salmonella antibodies (cutoff O.D.% ‡ 10) was measured in 183 of 325 meat juice samples (56Æ3%). Of these meat juice samples, 79 samples showed an O.D.% ‡40 (24Æ3%) (Table 4). At the herd level, sampled at the slaughterhouse, 47Æ7% were serologically negative when O.D.% > 10 was used. A herd can be bacteriologically positive, but all the pigs sampled within the herd at the moment of slaughter are serologically negative (25Æ9% of the positive herds). If a cutoff O.D.% ‡ 40 was used in 62Æ9% of the herds that where bacteriologically positive, all the pigs sampled from these herds at the moment of slaughter were serologically negative (Table 4). At the individual pig level serological ELISA results were compared with bacteriological results (faeces positive or negative) (Table 4). Serologically 43Æ6% of the animals were found negative while 80% of them were bacteriologically negative (faeces negative). Thirteen of the 65 bacteriologically positive pigs (20%) were serologically negative. Seventy per cent of the serologically positive pigs (cut-off O.D.% ‡ 10) were not shedding Salmonella at the moment of sampling. When both faeces and mesenteric lymph nodes were used to identify the bacteriological status of the pig, the best correlation was found with the ELISA results when using a cut-off O.D.% ‡ 40 (P ¼ 0Æ00000) instead of a cutoff O.D.% ‡ 10 (P ¼ 0Æ00002). DISCUSSION For the isolation of Salmonella from pork-related samples the highest sensitivity was reached with the semisolid medium Diassalm (95Æ3%). Other authors recommended, especially for faecal samples, the use of a semisolid medium for the isolation of Salmonella (Aspinall et al. 1992; Voogt et al. 2001) and obtained, in particular for faecal samples, a lower sensitivity by using only one isolation method (Davies et al. 2000). The use of Diassalm together with RV (10 ml) (see Materials and Methods) as isolation method of Salmonella offers the highest sensitivity. Thirty seven per cent of the carcasses sampled at the end of the slaughter line before chilling were positive for Salmonella. Korsak et al. (1998) found about 27% of the pork carcasses positive for Salmonella in four different Belgian slaughterhouses. In comparison with similar studies in some other western countries the Belgian contamination frequency is high. A study from the Netherlands showed only 1Æ4% of the carcasses positive at the end of the slaughter line (Swanenburg 2000). A German study reported that about

10% of the carcass swabs collected in seven different abattoirs were positive (Kasbohrer et al. 2000). Only in the UK, with 7–26% of the carcasses contaminated, the same high frequencies were reported (Davies et al. 1999). Nineteen per cent of the pigs were shedding Salmonella in their faeces and 21% carried Salmonella in their mesenteric lymph nodes. Together 28% of the animals carried Salmonella in their faeces and/or mesenteric lymph nodes. Heuvelink et al. (1998) found also a higher contamination level of the mesenteric lymph nodes (26Æ5%) in comparison with the feacal samples (23Æ8%). Letellier et al. (1999) isolated Salmonella from 5Æ2% of the caecal samples in Canada. In Germany and the Netherlands, respectively, 3Æ7 and 25Æ6% of the faecal samples were positive (Kasbohrer et al. 2000; Swanenburg 2000). Highly contaminated intestines or lymph nodes could be a primary source of carcass contamination during evisceration. In the slaughterhouse not only carcasses of infected animals were contaminated but also cross-contamination of other infected animals and from the environment occurred. Figure 2 shows that in several slaughterhouses carcass contamination is a result of two parameters: the Salmonellastatus of the supplied animals and the slaughterhouse hygiene. Cross-contamination was estimated to account for 29% of the entire carcass contamination. This corresponds well with the results of Berends et al. (1997) and Borch et al. (1996) who estimated that the initial source of pork carcass contamination was the carrier pig itself, calculated to be 70 and 30% of the entire carcass contamination was due to cross-contamination from the slaughterhouse. In our study large variations were noticed between the cross-contamination values of different slaughterhouses sampled. When the carcass contamination was followed as a function of the supply of positive animals, a good correlation was found between both parameters in slaughterhouse A2 (86Æ4%); whereas in slaughterhouse D1 and B the cross-contamination was calculated to be 68Æ5 and 100%, respectively. Certain serotypes present in the faeces at the beginning of the slaughter day were indeed recovered on carcasses slaughtered at the end of the day. In slaughterhouse D1 the high carcass contamination already present at the start of the slaughter day was probably due to a continuous environmental infection source. The overall presence of Salmonella in the environmental samples of the slaughterhouses was excessive from 92 to 100%. Fluctuation from 33 to 75% occurred between the different slaughterhouses. Even before starting the slaughtering process Salmonella could be isolated from the environment, indicative of inefficient cleaning. By taking overshoe samples different Salmonella serotypes could be isolated, which indicates the efficiency of the sampling technique. An overshoe sample could be used for monitoring of the cleaning efficiency at the slaughterhouse. The

ª 2003 The Society for Applied Microbiology, Journal of Applied Microbiology, 95, 891–903, doi:10.1046/j.1365-2672.2003.02042.x

PIG CARCASS CONTAMINATION WITH SALMONELLA DURING SLAUGHTERING

knives and the splitting machine could be a source of contamination; at slaughterhouse D1 the same serotype (S. Typhimurium) was found on the splitting machine and on the carcasses and this was possibly the reason for the continuous contamination of the carcasses during the whole day. During a second visit 1 month later (D2), this S. Typhimurium strain was not isolated from the cutting equipments, but other Salmonella serotypes were isolated such as S. London, S. Derby and a nonserotypable strain on the SKG-II apparatus and knives. From one carcass hanging in the refrigerator and killed 1 week earlier S. Typhimurium could be recovered by swabbing. In most cases the Salmonella flora in the slaughterhouse environment reflected the microbial flora of the pigs that were delivered that day. The most frequently isolated serotypes from the animal-related samples were S. Typhimurium, S. Derby and S. Livingstone, and these were also the most prevalent serotypes in the environment of the slaughterhouse. Differences in serotypes between sampling days and between slaughterhouses were observed, which demonstrated that sampling results depend on slaughterhouse (hygienic parameters and qualification of personnel), the moment of sampling and the origin and number of infected pigs that were delivered during the sampling period. After tracing the Salmonella data from the colon content isolated in the slaughterhouse back to the herd level, it was estimated that 40% of the herds were Salmonella positive at the moment of slaughter. A high level of herd contamination was also found in the Netherlands with 23% of the herds Salmonella positive sampled on the farm (van der Wolf et al. 1999) and in the UK with 63% positive farms (Davies et al. 1999). For interpretation of our data, it has to be kept in mind that the pigs with positive colon content and/or mesenteric lymph nodes in the slaughterhouse could have been infected on the farm and during transport or during the waiting period in the lairage before slaughtering. There are indeed indications that the contamination could already be detected in the faeces and the mesenteric lymph nodes as early as 3 h after infection (Fedorka-Cray et al. 1994). Especially the lairage and the high contamination level of the slaughterhouse environment are probably the major source for Salmonella infections prior to slaughter (Hurd et al. 2001; Swanenburg et al. 2001). Hurd et al. (2002) demonstrated that rapid infection during transport, and particularly during holding, is a major reason for increased Salmonella prevalence in swine: a sevenfold higher Salmonella isolation rate and twice as many different serovars were observed from pigs necropsied at the abattoir than from those necropsied on the farm. On the serological level Nielsen and Wegener (1997) found 4–7% positive meat drip samples of finishing pigs in Denmark when a cut-off O.D.% ‡ 40 was used. In a

901

German study, from 2947 serologically investigated fattening pigs 7Æ3% were found positive at a cut-off O.D.% ‡ 40 (von Altrock et al. 2000). In our study a much higher seroprevalence of 24Æ3% was noticed when cut-off O.D.% ‡ 40 was used. When we compared the bacteriological results with the serological results of the individual pigs at the moment of slaughter, the correlation between both was poor. It is to be expected that recent Salmonella infections cannot be detected by serological examination. It is indeed possible that the pigs get infected during transport or during the waiting period in the lairage before slaughter (Fedorka-Cray et al. 1994). In the abattoir, Hurd et al. (2002) found a bacteriological Salmonella prevalence rate of 39Æ9% in the pigs compared with only 17% seroprevalence with O.D.% ‡ 40. In our study in 17Æ7% of the cases where the ELISA was negative (cut-off value O.D.% < 10) Salmonella could be isolated from the mesenteric lymph nodes (12 of the 141 samples) or from the faeces (13 of the 142 samples). These infected pigs can probably be important in the transmission of Salmonella in the slaughterhouse and to the carcasses. However, it is also possible that an animal with a positive serological status does not shed Salmonella in the faeces anymore. The Danish control programme for reducing Salmonella in the pork production chain is based on serological monitoring of the pig herds. On herd prevalence they found a good correlation between bacteriological and serological examination (Stege et al. 2000). However, it is not so clear which cut-off value has to be used: in the Danish monitoring programme a cut-off value O.D.% ‡ 40 is used, but other authors suggested the use of the lower cut-off value (O.D.% ‡ 10) (van der Wolf 2000). From the results of our study it seems that by taking meat juice samples in the slaughterhouse on individual pig level a cut-off O.D.% ‡ 40 is preferential, while on herd level, measured at the slaughterhouse, O.D.% ‡ 10 gives the best correlation with the bacteriological analysis. Our results indicate clearly that a Salmonella reducing control programme has to include measures to decrease the Salmonella contamination at the pig level and in the slaughterhouse. The use of an adequate method to determine the Salmonella status of the pigs together with an improvement of the slaughterhouse hygiene seems mandatory. ACKNOWLEDGEMENTS This work was financially supported by the Flemish government (IWT). We thank Petra Vanmol for the excellent technical assistance and Ann Vanhee, Ann Van de Bossche, Elly Engels, Leen Duboccage, Jurgen Depuydt and Piere Delputte whose help was crucial for sampling at the slaughterhouses. Further on we would like to thank the different slaughterhouses for their participation in the study.

ª 2003 The Society for Applied Microbiology, Journal of Applied Microbiology, 95, 891–903, doi:10.1046/j.1365-2672.2003.02042.x

902 N . B O T T E L D O O R N ET AL.

REFERENCES Aabo, S., Rasmussen, O., Rossen, L., Sørensen, P. and Olsen, J. (1993) Salmonella identification by the polymerase chain reaction. Molecular and Cellular Probes 7, 171–178. von Altrock, A., Schutte, A. and Hildebrandt, G. (2000) Results of the German investigation on the EU project Salmonella in pork (Salinpork) 1. Investigations in the farms. Berliner und Mu¨nchener Tiera¨rztliche Wochenschrift 113, 191–201. Aspinall, S.T., Hindle, M.A. and Hutchinson, D.N. (1992) Improved isolation of salmonellae from faeces using a semisolid RappaportVassiliadis medium. European Journal of Clinical Microbiological Infection Diseases 11, 936–939. Baggesen, D.L. and Aarestrup, F.M. (1998) Characterisation of recently emerged multiple antibiotic-resistant Salmonella enterica serovar typhimurium DT104 and other multiresistant phage types from Danish pig herds. Veterinary Record 25, 95–97. Berends, B.R., Van Knapen, F., Snijders, J.M. and Mossel, D.A. (1997) Identification and quantification of risk factors regarding Salmonella spp. on pork carcasses. International Journal of Food Microbiology 36, 199–206. Berends, B.R., Van Knapen, F., Mossel, D.A.A., Burt, S.A. and Snijders, J.M.A. (1998) Impact on human health of Salmonella spp. on pork in The Netherlands and the anticipated effects of some currently proposed control strategies. International Journal of Food Microbiology 44, 219–229. Borch, E., Nesbakken, T. and Christensen, H. (1996) Hazard identification in swine slaughter with respect to foodborne bacteria. International Journal of Food Microbiology 30, 9–25. Daube, G., Dumont, J.-M., Libotte-Chasseur, M.-L., Ghafir, Y. and De Zutter, L. (1998) Serovars of Salmonella strains isolated from foods of animal origin in Belgium. In Proceedings of the Third Conference in Food Microbiology, Lie`ge, Belgium, 10 September, pp. 70–71. Davies, R.H., McLaren, I.M. and Bedford, S. (1999) Distribution of Salmonella contamination in two pig abattoirs. In Proceedings of the Third International Symposium on the Epidemiology and Control of Salmonella in Pork, Washington, DC, 5–7 August. pp. 286–288. Davies, P.R., Turkson, P.K., Funk, J.A., Nichols, M.A., Ladely, S.R. and Fedorka-Cray, P.J. (2000) Comparison of methods for isolating Salmonella bacteria from faeces of naturally infected pigs. Journal of Applied Microbiology 89, 169–177. Delpech, V., McAnulty, J. and Morgan, K. 1998. A salmonellosis outbreak linked to internally contaminated pork meat. Australian and New Zealand Journal of Public Health 22, 243–246. Fedorka-Cray, P.J., Whipp, S.C., Isaacson, R.E., Nord, N. and Langer, K. (1994) Transmission of Salmonella Tyhimurium to swine. Veterinary Microbiology 41, 333–344. Heuvelink, A., Tilburg, J., Ritmeester, W., Sturm, J. and van de Giessen, A. (1998) The occurrence of Salmonella in feces and mesenteric lymph nodes of normal slaughter pigs in the Netherlands. Rijksinstituut voor volksgezondheid en milieu (RIVM). The Netherlands rapport 285859006. Hurd, H.S., McKean, J.D., Wesley, I.V. and Karriker, L.A. (2001) The effect of lairage on Salmonella isolation from market swine. Journal of Food Protection 64, 939–944.

Hurd, H.S., McKean, J.D., Griffith, R.W., Wesley, I.V. and Rostagno, M.H. (2002) Salmonella enterica infections in market swine with and without transport and holding. Applied and Environmental Microbiology 68, 2376–2381. Kasbohrer, A., Protz, D., Helmuth, R., Nockler, K., Blaha, T., Conraths, F. and Geue, L. (2000) Salmonella in slaughter pigs of German origin: an epidemiological study. European Journal of Epidemiology 16, 141–146. Korsak, N., Daube, G., Ghafir, Y., Chahed, A., Jolly, S. and Vindevogel, H. (1998) An efficient sampling technique used to detect four foodborne pathogens on pork and beef carcasses in nine Belgian abattoirs. Journal of Food Protection 61, 535–541. Letellier, A., Messier, S., Pare, J., Menard, J. and Quessy, S. (1999) Distribution of Salmonella in swine herds in Quebec. Veterinary Microbiology 67, 299–306. Molbak, K. and Hald DT. (1997) An outbreak of Salmonella Typhimurium in the country of Funen during late summer. A case-controlled study. Ugeskrant Laeger 159, 5373–5377. Mousing, J., Jensen, P.T., Halgaard, C., Bager, F., Feld, N., Nielsen, B., Nielsen J.P. and Bech-Nielsen, S. (1997) Nation-wide Salmonella enterica surveillance and control in Danish slaughter swineherds. Preventive Veterinary Medicine 29, 247–261. Murase, T., Yamada, M. Muto, T., Matsushima, A. and Yamai, S. (2000) Fecal excretion of Salmonella enterica serovar Typhimurium following a food-borne outbreak. Journal of Clinical Microbiology 38, 3495–3497. Nielsen, B. and Wegener, H.C. (1997) Public health and pork and pork products: regional perspectives of Denmark. Revue Scientifique et Technique 16, 513–524. van Pelt, W., van Giessen, A., van Leeuwen, W., Wannet, W., Henken, A., Evers, E., De Wit, M. and van Duynhoven, Y. (2000) Oorsprong, omvang en kosten van humane salmonellose. Infectieziekten Bulletin 11, 4–8. Pontello, M., Sodano, L., Nastasi, A., Mammina, C., Astuti, M., Domenichini, M., Belluzzi, G., Soccini, E., et al. (1998) A community-based outbreak of Salmonella enterica serotype Typhimurium associated with salami consumption in Northern Italy. Epidemiological Infections 120, 209–214. Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989) Molecular Cloning: a Laboratory Manual, 2nd Edn. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press. Stege, H., Christensen, J., Nielsen, J.P., Baggesen, D.L., Enoe, C. and Willeberg, P. (2000) Prevalence of subclinical Salmonella enterica infection in Danish finishing pig herds. Preventive Veterinary Medicine 44, 175–188. Steinbach, G. and Kroell, U. (1999) Salmonella infections in swine herds-epidemiology and importance for human diseases. DTW Dtsch. Tiera¨rztliche Wochenschrift 106, 282–288. Swanenburg, M. (2000) Salmonella in pork production chain: sources of Salmonella on porc. PhD thesis, Utrecht University, Utrecht, The Netherlands. Swanenburg, M., Urlings, H.A.P., Keuzenkamp, D.A. and Snijders, J.M.A. (2001) Salmonella in the lairage of pig slaughterhouses. Journal of Food Protection 64, 12–16. Threlfall, E.J., Frost, J.A., Ward, L.R. and Rowe, B. (1994) Epidemic in cattle and humans of Salmonella Typhimurium DT 104 with

ª 2003 The Society for Applied Microbiology, Journal of Applied Microbiology, 95, 891–903, doi:10.1046/j.1365-2672.2003.02042.x

PIG CARCASS CONTAMINATION WITH SALMONELLA DURING SLAUGHTERING

chromosomally integrated multiple drug resistance. Veterinary Record 134, 577. Versalovic, J., Koeuth, T. and Lupski, J.R. (1991) Distribution of repetitive DNA sequences in eubacteria and application to fingerprinting of bacterial genomes. Nucleic Acids Research 19, 6823–6831. Voogt, N., Raes, M., Wannet, W.J., Henken, A.M. and van de Giessen, A.W. (2001) Comparison of selective enrichment media for the detection of Salmonella in poultry faeces. Letters in Applied Microbiology 32, 89–92. Wall, P.G., Morgan, D., Lamden, K., Ryan, M., Griffin, M., Threlfall, E.J., Ward, L.R. and Rowe, B. (1994) A case control study of

903

infection with an epidemic strain of multiresistant Salmonella Typhimurium DT104 in England and Wales. Communicable Diseases Report. CDR Review 14, R130–R135. van der Wolf, P.J. (2000) Salmonella in the porc production chain: feasibility of Salmonella-free pig production. PhD thesis, University of Utrecht, Utrecht, The Netherlands. van der Wolf, P.J., Bongers, J.H., Elbers, A.R., Franssen, F.M., Hunneman, W.A., van Exsel, A.C. and Tielen, M.J. (1999) Salmonella infections in finishing pigs in The Netherlands: bacteriological herd prevalence, serogroup and antibiotic resistance of isolates and risk factors for infection. Veterinary Microbiology 67, 263–275.

ª 2003 The Society for Applied Microbiology, Journal of Applied Microbiology, 95, 891–903, doi:10.1046/j.1365-2672.2003.02042.x