- Email: [email protected]
Risk factors for the infection with Brachyspira hyodysenteriae in pig herds
Journal Pre-proof Risk factors for the infection with Brachyspira hyodysenteriae in pig herds Friederike Zeeh, Beatriz Vidondo, Heiko Nathues
PII:
S0167-5877(19)30479-9
DOI:
https://doi.org/10.1016/j.prevetmed.2019.104819
Reference:
PREVET 104819
To appear in:
Preventive Veterinary Medicine
Received Date:
5 August 2019
Revised Date:
28 October 2019
Accepted Date:
29 October 2019
Please cite this article as: Zeeh F, Vidondo B, Nathues H, Risk factors for the infection with Brachyspira hyodysenteriae in pig herds, Preventive Veterinary Medicine (2019), doi: https://doi.org/10.1016/j.prevetmed.2019.104819
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier.
Risk factors for the infection with Brachyspira hyodysenteriae in pig herds
Friederike Zeeha, 1*, Beatriz Vidondob, Heiko Nathuesa
Affiliations: a
Clinic for Swine, Vetsuisse Faculty, University of Bern, Bremgartenstrasse 109a, PB 3350, 3001 Bern,
Switzerland b
ro of
Veterinary Public Health Institute, Vetsuisse Faculty, University of Bern, Schwarzenburgstrasse 155,
3097 Liebefeld, Switzerland
[email protected] (F. Zeeh)
re
[email protected] (B. Vidondo)
-p
E-mail addresses:
Corresponding author:
na
*Friederike Zeeh
lP
[email protected] (H. Nathues)
Clinic for Swine, Vetsuisse Faculty, University of Bern, Bremgartenstrasse 109a, PB 3350, 3001 Bern,
ur
Switzerland
[email protected] +41 (0)31 631 23 44
Fax:
+41 (0)31 631 26 31
Jo
Tel.:
Present address 1
Jockey Club College of Veterinary Medicine and Life Sciences, Department of Infectious Diseases
and Public Health, City University of Hong Kong, Hong Kong SAR. [email protected] 1
Abstract Swine dysentery (SD), caused by infection with Brachyspira hyodysenteriae, is a serious disease in pig production worldwide. Quantitative risk factors triggering the occurrence of infection are unknown. The present case-control study aimed at identifying major risk factors related to presence of B. hyodysenteriae in pig herds. Twenty case herds and 60 randomly selected control herds with a minimum herd size of ‘10 sows/ 80 fattening pigs’ were examined by means of a questionnaire-based interview and a herd examination. Herds with previous eradication of SD were excluded. Logistic
ro of
regression models revealed that the ‘positive/suspicious SD status of source herds’, the regular application of treatment, purchasing more than 4 batches/ year, contact to foxes, diagnostics performed during last 12 months, liquid feeding systems, rats on farm, and >250 fatting places were
-p
associated with higher chances of a herd to be infected. On the contrary, having different sources of grower pigs within one batch, the presence of raptor birds and the presence of martens in the region
re
were associated with fewer chances of being infected. The final multivariable logistic regression model identified purchasing more than 4 batches/ year (OR = 7.5, 95% CI 1.8-54.3) and contact to foxes (OR
lP
= 5.9; 97.5% CI 1.2-34.6) as the two main risk factors in our study. ‘More than 4 batches/ year’ implies continuous herd management supporting persistence of B. hyodysenteriae in an infected herd, but
na
also increased number of purchases each increasing the risk of B. hyodysenteriae introduction by carrier pigs or transport vehicles. Foxes might be infected with B. hyodysenteriae by feeding on positive
ur
piglets and rodents. Besides, ‘contact to foxes’ might represent a lack in biosecurity. In conclusion, the
Jo
risk factors detected underline the importance of biosecurity in SD prevention and control.
Keywords
Swine Dysentery; Epidemiology; Protective factor; Biosecurity; Fox
2
1. Introduction
Swine dysentery (SD), caused by Brachyspira hyodysenteriae, is of high clinical and economic importance worldwide due to severe intestinal lesions but also subclinical infections resulting in decreased growth rate and feed conversion (Hampson et al., 2015). Transmission of B. hyodysenteriae takes place via the faecal-oral route and treatment requires application of effective antimicrobials over a prolonged period and sometime repeatedly. Control and eradication measures of B. hyodysenteriae
ro of
have gained considerable importance during the last years, especially because of an increase in frequency and amount of antimicrobial resistance (Hidalgo et al., 2009; Mirajkar et al., 2016; Massacci et al., 2018) and the absence of a commercially available vaccine. Basis for a sustainable control or
factors for transmission and spread of B. hyodysenteriae.
-p
eradication are reliable detection methods, but also broad knowledge on epidemiology including risk
re
In respect to risk factors for B. hyodysenteriae infection, only a limited number of studies have been published. In a report from Australia, data from a postal survey performed in herds with serologic
lP
results indicative of B. hyodysenteriae and in serologic negative herds were analysed (Robertson et al., 1992). Factors linked to uncontrolled pig purchase, visitors and rodents (amongst others) had a
na
significantly higher odds ratio >1 for being positive and protective factors were for instance protective clothes or footbaths. Beside this one study using statistical analysis for description of potential risk
ur
factors, a number of studies with empiric data have been published. Transmission of B. hyodysenteriae by carrier pigs is considered to be the main route. Hence, pig trade and transport may pose a high risk.
Jo
This was also supported by a recent study describing B. hyodysenteriae in transport vehicles (Giacomini et al., 2018). Besides pig movement also other factors like management, positive neighbour herds, pig transport, feed lorry, and birds were involved in regional spread of SD (Windsor and Simmons, 1981). Also feral pigs and wild boars (Phillips et al., 2009; Reiner et al., 2011), rodents (Backhans et al., 2010), dogs (Songer et al., 1978) and birds like mallards (Jansson and Råsbäck, 2009), geese (Rubin et al., 2013), rheas (Jensen et al., 1996) and a crow (Zeeh et al., 2018) have been tested positive for 3
B. hyodysenteriae. Furthermore, B. hyodysenteriae can survive in slurry (Boye et al., 2001) and thus rapidly spread within herds and potentially between herds. Additionally, in an experimental challenge study, Musca flies and cockroaches were B. hyodysenteriae positive in their gastrointestinal tract and also excreted viable bacteria (Blunt et al., 2010). Factors enhancing spread within a herd are e.g. continuous management of pens or barns and poor husbandry (Alvarez-Ordóñez et al., 2013; Burrough, 2016). All these findings indicate that there might be a number of potential risk factors for B. hyodysenteriae infection warranting further and more detailed (statistical) analysis.
ro of
The aim of the study was to systematically examine potential risk factors for the infection with B. hyodysenteriae. Identification of such factors would support prevention and eradication of
-p
B. hyodysenteriae infections and by this improvement of pig health.
lP
2.1. Study design and herd selection
re
2. Methods
The case-control study was performed in Swiss pig herds. Twenty case herds were selected from
na
herds with previous detection of B. hyodysenteriae. Per case herd, three control herds were randomly selected from a database containing information of about 90% of all Swiss pig herds. Randomisation
ur
was done by ‘random number’ in Excel. Case and control herds were selected according to the production type (piglet producing versus grower-finisher herd). As it became obvious during the first
Jo
months that the category ‘piglet producer without grower-finisher pigs’ would be underrepresented in the study population, this herd type was not considered in the project anymore. Only herds with more than 10 sows and/or more than 80 grower-finisher pigs were included because this herd size is more likely to exist in the future. Herds that had undergone an SD eradication were excluded as biosecurity measures might have been changed after the eradication. The sample size of 20 case and 60 control herds was chosen based on the limited number of B. hyodysenteriae positive herds meeting 4
all inclusion criteria and the pig owners’ availability to perform the interviews (see below) during the study period. This sample size rendered a power of 80% at 0.05 significance level to detect odds ratios larger than 4 (Demidenko, 2006).
2.2. Questionnaire and interview
Pig farmers were contacted by phone and asked to participate in the study. The participation
ro of
was voluntary. A questionnaire was created and validated in three herds not included in the present study. The validated questionnaire was completed by the principal investigator on farm together with the farmer.
-p
A total of 497 questions covered general information about the herd, housing, management, hygiene, internal and external biosecurity (with emphasis on slurry management and fly and rodent control),
re
pig trade and feed and feeding, during the last 12 months to reduce recall bias. To ensure a standard across farms, subjective parameters, such as ‘hygiene’ or ‘rodent burden’,
lP
etc., were judged by the principal investigator during a herd examination immediately after the
na
interview.
ur
2.3. Data processing and statistical methods for data analysis
The data from the questionnaire were transferred in a data management programme (Microsoft
Jo
InfoPath 2010), subsequently transformed in an Excel table (Microsoft Excel, 2010) and imported to the statistical software R (R version 3.3.1) for further analysis. Data were visually checked for completeness and, if biological plausible, combined or imputed
when necessary. Continuous data were tested for normality using histograms and by Shapiro-Wilk test. Using logistic regression models, each of the measured farm-level properties was tested for association with the probability of being a case. Correlations between explanatory variables were assessed by 5
means of Spearman correlation coefficients. Only those that did not correlate to each other were allowed in the same multivariable model. Multivariable logistic regression models were built using both forward and backward stepwise variable selection. This ensured robustness of the final model. For forward selection, variables with the highest odds ratios (OR) were selected first. Only models with biological significance were retained. Control for confounding was carried out during the variable selection processes because variables were retained if they caused a change of more than 25% in the odds ratios of the other variables in the model. The geographic location of herds was allocated to one
ro of
of the three regions East (cantons Aargau, Appenzell Inner Rhoden, St Gallen, Thurgau, Zurich), Middle (Basel-Land, Lucerne, Schwyz) and West (Berne, Friborg, Solothurn, Vaud). Region was not associated with the chances of having SD and did not affect the results of the final model. For all models, the level
-p
of significance was set at 0.05.
re
Geographical distribution of the examined herds was visualized using QGIS (www.qgis.org).
lP
3. Results
Of 120 contacted pig farmers, 32 refused to participate because of ‘no interest’ (n = 15), ‘no pigs
na
anymore’ (n = 8) or ‘personal reasons incl. not reached’ (n = 9), resulting in a response rate of 73.3%. Additional four herds were excluded after the interview because of not having met the initial inclusion
ur
criterion of herd size. Further four herds (piglet producers without fattening pigs) were retroactively excluded because of the revised inclusion criteria. Finally, 20 case herds and 60 control herds,
Jo
examined between August 2014 and November 2017, were included in the study. The 80 herds were located in 12 Swiss cantons. The majority of the herds was situated in the
cantons of Berne (30.0%), Lucerne (23.8%), and Aargau (15.0%) (Fig. 1).
3.1. Univariable Analysis 3.1.1. Herd characteristics 6
The majority of herds were grower-finisher herds (85% (17/20) and 90% (54/60) in the cases and controls, respectively. Case herds also comprised two wean-to-finish herds (10%), and one farrow-tofinish herd (5%). Within the control herds, the remaining six were farrow-to-finish herds (10%). Three of these farrow-to-finish herds were closed herds with no pig purchase in the last 12 months. All other 77 pig owners reported pig purchases. Due to the low number of herds with sows, piglets and/or weaned pigs, parameters describing these age categories are not analysed further.
ro of
All case herds and 45 (75%) control herds were monitored by a pig health service (SUISAG). Production under ‘animal-friendly’ labels (e.g. requiring more space and/or access to outdoor areas) was present in 45 herds (50% of the case herds, 59% of the control herds).
-p
The median number of fattening places (‘herd size’) of all 80 herds was 255 (range: 80-1000), of the case herds 383 (range: 104-1000), and 230 (range: 80-1000) in the control herds. Seventy of the
re
farmers (85.0% in case herds, 88.3% in control herds) had other lines of business besides the pigs with mainly tillage (n = 53) and dairy cows and/ or raising calves (n = 49; multiple answers possible). Seven
lP
out of the 15 persons working also outside the own farm had contact to foreign pigs. In 85.0%, the interview partner was the pig owner. Tenants (n = 6), contractors or employees (3 each) were the
na
minority.
ur
3.1.2. Trade and transport
Jo
Nineteen different trading companies were involved in the pig trade of incoming pigs and 18 of
them also in trade of outgoing pigs. Only 11.3% (3 out of 20 cases, 6 out of 60 controls) and 6.4% (1 out of 20 cases, 4 out of 58 controls) of the herds had no trader for incoming or outgoing pigs, respectively. In the 12 months prior the interview, the median number of source herds was 1.0 (range 1-36) in the case herds and 2.5 (range: 1-27) in the control herds. Average number of batches per year, 7
according to the 75 pig farmers who bought weaner or grower pigs, was in median 17 (range: 1-52) in the case herds and 6 (range: 1-52) in control herds. When herds were classified into herds with few (04) and many (>4) batches/year, then case herds were significantly more often in the group with many batches (88.9% versus 54.5%; OR = 6.66; P = 0.017, 95% confidence interval (CI): 1.68-44.76). Batches of grower pigs comprised in 25.0% of the case herds and in 56.7% of the control herds pigs of more than one source (OR = 0.3; P = 0.018; CI: 0.1-0.8). The maximum number of sources per batch, according to the data from the interview partners, was in median 1 (1-4) in case herds and 1 (1-9) in
ro of
control herds. The status of the source herds in respect to B. hyodysenteriae infection (according to the pig health service, the herd attending vet or the pig owners; last 12 months) differed significantly between the two herd types. Case herds had received in 44.4% pigs from B. hyodysenteriae positive or
-p
suspicious herds, control herds in 6.7% (OR = 11.2; P < 0.01; CI: 3.0-49.1). The classification as positive herd is most reliable for breeding and piglet producing herds, because most of them belong to the pig
re
health services (100% and over 85%, respectively). SD is regulated in the pig health services programmes and trained veterinarians provide on-farm investigations and order laboratory
lP
confirmation tests either on a mandatory and regular basis (breeding herds) or in case of SD suspicion in the other affiliated herds. For herds of farmers who do not access the pig health services, the
na
detection of B. hyodysenteriae was asked from the herd-attending veterinarian. Suspicious herds were for instance farrowing herds of a B. hyodysenteriae positive sow pool system without own detection
ur
of the pathogen.
Transport of grower pigs included transport by the trader (9 out of 18 of case herds, 38 out of
Jo
57 of control herds), sub-contractors of the trader (5 out of 18 case herds; 3 out of 57 control herds), the piglet producer (4 out of 18 case herds; 3 out of 57 control herds) or the farmer of the fattening herd himself (2 out of 18; 9 out of 57 control herds; multiple answers possible). Slaughter pigs were transported mainly by traders (50.0% of case herds; 50.8% of control herds) and sub-contractors (40.0% of case herds; 25.4% of control herds).
8
Sequence of farms in respect to transport of incoming pigs was in 45.5% of the herds (55.0% of the case herds; 42.1% of the control herds) ‘always the first farm being approached during the delivery tour’. Thirty-nine percent of the interview partners (30% of case herds and 42.1% of control herds) reported the presence of pigs dedicated to other farms on the transport vehicles. Regarding collection of outgoing (slaughter) pigs, none of the case herds (n = 20) was approached first. Of the control herds, 20.7% (12 out of 58) were approached first. When asked, if the transport vehicles for slaughter pigs were always empty at arrival, then 82.3% (95.0% case herds; 78.0% control herds) of the participants
ro of
answered ‘No’.
-p
3.1.3. External biosecurity
A loading bay was present in 45.0% of the case herds whereof four (44.4%) were rated as
re
‘suitable’ (preventing re-entrance of slaughter pigs into the barn), and only one (11.1%) was clean as assessed by the first author. In the 22 out of the 56 control herds with a loading bay, 13 (59.1%) of
lP
them were ‘suitable’ and 10 (45.5%) ‘clean’.
In all 80 herds, at least one external person have had accessed the pig barn in the 12 months
na
prior the interview. The veterinarian had been present in the stable in the last 12 months in 70.0% of the case herds (n = 14) and in 63.3% of the control herds (n = 38). Apart from the veterinarian, between
ur
1 to 8 different person groups had entered the barn of case herds and between 0-9 of control herds. Protective clothes were provided to visitors in 52.2% (11 case herds, 31 control herds) and, as assessed
Jo
by the first author, used in 42.5% of the herds (7 case herds, 27 control herds). Distance to the next pig barn was, according to the farmers, for case herds in median 430 m (50-
8’000) and for controls 600 m (20-5’000). The median number of neighbour pig herds in a radius of 500 m was 1 for cases and 0 for controls, in a radius between 0.6 -1 km and in a radius between 1.1 and 2 km one neighbour herd each.
9
Access to outdoor areas was available for the pigs in 71.3% of the herds (12 case herds, 45 control herds). In 56 herds, this area had concrete floor, and one herd of the control group used a pasture. Dogs and/or cats had access to the pig barns of 51.9% herds (11 case herds, 30 control herds). Presence of wild animals in the region was stated by 97.5% of all farmers (19 case herds, 59 control herds). Mainly sparrows (87.5%), foxes (83.8%) and crows (71.3%) were reported. Wild boars were reported by 36.3% of the interview partners (5 case herds, 24 control herds). Presence of wild raptor birds in the region was significantly related to a negative B. hyodysenteriae status (15.0% of case herds,
ro of
56.7% of control herds; OR = 0.1; P = 0.003; CI: 0.03-0.5). Contact to wild animals was reported in 85.0% of the case herds and 78.3% of the control herds. Especially contact to wild birds, including sparrows and crows, were reported in both herd types (16 case herds, 42 control herds). Contact to foxes
-p
occurred in 35.0% of the case herds (7 out of 20) and 10.0% of the control herds (6 out of 60) (OR = 4.9; P = 0.013; CI: 1.4-17.6). Crows came in contact to the domestic pigs of 71.3% of the herds (14 case
re
herds, 43 control herds) and wild boars to 36.3% (5 case herds, 24 control herds). Slurry is considered a reservoir for B. hyodysenteriae. In our study, foreign slurry was deposed
lP
in the vicinity of 38.8% of the herds (8 out of 20 case herds, 23 out of 60 control herds). The own slurry was stored mainly in pits (80.0% of case herds, 76.7% of control herds). Material used for applying the
na
manure was diverse (barrels in 58 herds, tubes in 43 herds, trailing tubes in 51 herds) and shared with other pig farmers in 43 herds (10 out of 20 case herds, 33 out of 59 control herds). Treatment of slurry
ur
was performed in five case herds and 19 control herds and included i.e. regular agitation or application of products improving the flow-ability. Forty-five herd farmers (14 case herds, 31 control herds) gave
Jo
pig slurry away and only eight received foreign slurry (2 case herds, 6 control herds). Of these 45 herds with slurry contracts, five herds had both contract types (2 case herds, 3 control herds). Vaccination status of incoming pigs was known to 83.5% of the farmers (15 out of 20 cases, 51
out of 59 controls).
3.1.4. Internal biosecurity 10
All-in/All-out management was applied to the fattening units in 23 control herds but in none of the case herds. The floor in the fattening barns was in most cases partially slatted (75.0% case herds, 84.2% control herds) and mainly moderately clean as rated by the investigator. Regular disinfection was applied in 15.0% of the case herds (3 out of 20) and 15.3% of the control herds (9 out of 59). Restocking after clearance of pens was conducted in median after 1 day (range: 0-28) in case herds and after 3 days in control herds (range: 0- 30). Stocking density in pens for grower pigs was in median
ro of
0.75 pigs/ m2 (0.57-2.15) in case herds and 0.86 pigs/ m2 (0.23-2.1) in control herds and in pens for finisher pigs 0.65 pigs/ m2 (0.2-1.79) and 0.63 (0.23-1.6), respectively. Bedding material was used in 75.0% of the case herds (15 out of 20) and in 81.7% of the control herds (49 out of 60) and consisted
-p
in most herds (57 out of 64) of straw. Type and source of activity material, which was provided to all
re
pigs, did not differ between the two herd types. In most cases, organic material (e.g. straw) was used.
lP
3.1.5. Feed and feeding system
Feed supply involved 16 different companies. A home-made mixture of feed was used in 50.0%
na
of the case herds (10 out of 20) and 26.7% of the control herds (16 out of 60). Soy is considered to have an influence on B. hyodysenteriae and occurrence of SD and high soy contents in are used in the diet
ur
of experimental B. hyodysenteriae infection studies to facilitate SD development (Jacobson et al., 2004). In 90.0% of the case herds (18 out of 20) and 86.7% of the control herds (52 out of 60), soy was
Jo
part of the feed. Content of soy in the feed for growers until about 50kg bodyweight, according to declaration of the feed mills or the farmers, was in median 9.5% (range: 0-17; case herds) and 10.5% (0-17.1; control herds), and in the feed for finisher pigs 8.8% (0-17; case herds) and 10.3% (0-21.9; control herds), respectively. By-products were used in 28.8% of the herds (9 case herds, 14 control herds) and consisted in herds of whey (7 case herds, 11 control herds).
11
The feeding system for grower-finisher pigs differed significantly between the case and the control group. A system for liquid feeding was more often present in case (16 out of 20) than in control (32 out of 60) herds (OR = 3.5; P = 0.042; CI: 1.1-13.3). The other herds had feeding systems for dry feed. Feed storage and feeding hygiene did not differ between the two herd types.
3.1.6 Fly and rodent control in fattening units
ro of
Fly control was applied in 58.8% of all herds (10 case herds, 37 control herds). More case herds (9 out of 10) had a more intensive fly control regime (‘every second month to permanent’) compared to the 37 control herds (46.2% with the same frequency). In both herd types, the actual fly burden was
-p
assessed as being in median ‘low’ at the time of the examination.
Presence of rodents was reported by farmers of 19 case herds and 57 control herds. In the case
re
herds, mainly rats and mice (52.6%), followed by mice (47.4%) were observed by the farmers. In the control herds, in 70.2% mice, in 22.8% rats and mice, and in 3.5% each rats and other rodents were
lP
indicated. Rats were significantly more often observed in case herds (10 out of 19, 52.6%) than in control herds (OR = 3.1; P = 0.042; CI: 1.1-13.3). Rodent control in the barn was performed in 72.5% of
na
the herds (15 cases herds, 42 control herds) and in the surrounding of the pig barn in 50.7% (10 out of 16 case herds, 25 out of 53 control herds). Methods used for the fattening barn were mainly
ur
‘rodenticides applied by the farmer’ (12 out of 16 case herds, 37 out of 45 control herds), and ‘cats’ (8 out of 16 case herds, 16 out of 45 control herds). Frequency of rodent control was in case herds most
Jo
often ‘permanent‘, in control herds mainly ‘3-4/year‘. If rodent control in the surrounding was present, then it was mainly ‘permanent‘ in both herd types and comprised rodenticides (6 out of 10 case herds; 15 out of 28 control herds) and cats (2 out of 10 case herds; 18 out of 28 control herds) in case and control herds, respectively. A professional rodent control company was used in only two case herds and one control herd. Signs of rodents, as examined by the first author, were present in 55.6% of the pig barns (10 out of 20 case herds, 32 out of 56 control herds) and in 50.0% of the surroundings (5 out 12
of 16 case herds, 27 out of 48 control herds). The rodent control was assessed by the first author as ‘effective’ (i.e. enough bait boxes at correct place, regular checking) only in two case herds and two control herds.
3.3 Performance and drug usage
Daily weight gain of the grower-finisher pigs in case herds was, according to the farmers, in
ro of
median 780g/d (range: 702-960) and in the control herds 825g/d (range: 493-995). Losses of pigs in the fattening unit were in median 2.0% (range: 0.2-3.9) in case herds and 1.6% (range: 0-7.0) in control herds.
-p
A regular antibiotic therapy was significantly more often used in the 20 case herds (73.7%) compared to the 60 control herds (25.0%) (OR = 8.4; P < 0.01, CI: 2.7-29.8). Within the 20 case herds,
re
14 were regularly treated during the 12 months prior the interview. Seven of the herds were regularly treated against B. hyodysenteriae and additional two against diarrhoea (not further specified). Further
lP
nine case herds have sporadically been treated against B. hyodysenteriae. Tiamulin was used in 12 herds, tylosin in five herds, lincomycin in four herds, and a drug not licensed (and effective) against
na
B. hyodysenteriae was used in one herd. In five of the herds, more than one drug component was used.
ur
A summary of all variables which had a p-value <0.05 in the univariable logistic regression
Jo
models is shown in table 1.
3.2. Multivariable analysis
Based on their high OR, absence of collinearity, statistical significance and biological plausibility, the following variables were considered in the final multivariable model: ‘number of pig purchases (batches) per year’, ‘contact to foxes’, ‘feeding system’, ‘rats on farm’, and ‘herd size’. The variables 13
‘BH-status of source herds’, ‘Diagnostics performed during last 12 months’, and ‘Regular treatment’ were considered to be (also) an consequence (rather than a risk or protective factor) of SD and were therefore not included in the multivariable models. The two variable selection processes gave consistent results. The final model thus consisted of two variables ‘Number of batches/year’ (OR = 7.5; P = 0.02; CI: 1.8-54.3) and ‘Contact to foxes’ (OR = 5.9; P = 0.03; CI: 1.2-34.6) as risk factors.
ro of
4. Discussion
Brachyspira hyodysenteriae infection is one of the re-emerging pig diseases with high economic
-p
impact and it is affecting pig herds regardless of whether the herds are thought to be free from it or having effectively combated the disease. Particular attention should be paid to this bacterium due to
re
its increasing resistance to antimicrobials.
The results of this first systematic study on risk factors for being infected with B. hyodysenteriae
lP
indicate that purchasing more than four batches per year (i.e. continuous management) and contact to foxes are the most important risk factors.
na
‘More than four batches per year purchased’ is linked to an increased number of purchases each increasing the risk of B. hyodysenteriae introduction by carrier pigs or transport vehicles (Thakur et al.,
ur
2016; Giacomini et al., 2018). Secondly, the underlying continuous management of the farms enables the persistence of the pathogen within the herd (Alvarez-Ordóñez et al., 2013).
Jo
‘Contact to foxes’ was a risk factor in our study. Presence of foxes has been linked to presence
of and problems with rodents (Fleming et al., 2016). According to that study, foxes feed of piglets and rodents. Piglets might already harbour B. hyodysenteriae and rodents are known vectors of that pathogen, thus potentially infecting foxes. However, further research is needed to elucidate if foxes are actually vectors. In addition to foxes being potentially a vector, reporting of ‘contact to foxes takes
14
place’ represents gaps in biosecurity. Although ‘access to outdoor areas’ was not a significant risk factor in our study, contact to foxes occurred only in farms with outdoor areas. Herds with liquid feeding were 3.5 times more likely to be in the case group. Although feed has been described as an important factor to enhance or reduce SD (Durmic et al., 2002; Pluske et al., 2002; Jacobson et al., 2004), the feeding system per se has not been discussed in the context of SD until now. Alterations in intestinal microbiome favouring B. hyodysenteriae or other digestive features might play a role, but further research is needed in order to confirm this hypothesis. Also potential
ro of
bias should be considered. Larger herd size (i.e. > 250 places for fattening pigs) was related with an increased likelihood to be in the case group (OR = 3.1). In our study population, larger herds had purchased more batches per
-p
year (rho = 0.65, P < 0.01, Spearman’s rank correlation test), thus increasing the risk to introduce B. hyodysenteriae by pigs, contaminated transport vehicles or other routes as discussed above. Larger
re
herds could also have more source herds, but this was not the case in the present study (Spearman rank: rho = - 0.11, p>0.05). However, large herds were not necessarily managed in a more professional
lP
way and also did exhibit deficits in biosecurity.
Presence of rats was significantly related to case herds. Rats are known carriers and vectors of
na
B. hyodysenteriae (Backhans et al., 2011). Our finding confirms earlier statements that rodents are a risk factor for SD ( Robertson et al., 1992; Pearson et al., 2016;).
ur
Although performance of diagnostics and regular treatment had increased odds ratios in our study, we interpret it as a result of the B. hyodysenteriae infection. In eleven of the case herds,
Jo
diagnostics had been performed whereof seven were in the context of B. hyodysenteriae and treatment in case herds was often directed against B. hyodysenteriae.
Three variables were significantly related to a negative B. hyodysenteriae status (protective factors) in our study. Raptor birds prey on rodents and smaller birds and thus might prevent or reduce B. hyodysenteriae transmission. The low pH in their stomach possibly inhibits or kills Brachyspira 15
bacteria potentially ingested by infected rodents. Also martens (Mustelidae) feed of rodents, but their actual role in SD like that of raptor birds remains unclear too. The biological background of the variable ‘different sources within one batch’ being protective was unexpected and is difficult to explain. The hypothesis that such mixed batches might be related to regular (prophylactic) treatment at time of placement could not be confirmed in our data (rho = 0.02, P > 0.05, Spearman’s rank correlation).
Waterfowl have been described in several studies as carrier and potential vector of
ro of
B. hyodysenteriae ( Jansson et al., 2004; Medhanie et al., 2013; Elmberg et al., 2017;). Our data did not confirm these findings, as waterfowl seems to have an inferior role in the transmission of B. hyodysenteriae in the examined Swiss population. Only 26 out of 80 farmers reported of ducks or
-p
geese in the region and none of contact of these birds to their pigs. Crows can harbour B. hyodysenteriae as recently described in Switzerland (Zeeh et al., 2018). Presence of crows in the
re
neighbourhood of their herds was reported by 71.3% of the farmers. Twenty of them explicitly described contact of these birds with outdoor areas and hence pigs, but there was no difference
lP
between case and control herds. However, this study did not analyse the frequency of these contacts which might be important in successful B. hyodysenteriae transmission. Swallows (Hirundinidae) feed
na
of flies which are vectors of B. hyodysenteriae. In our study, swallows were not related to the B. hyodysenteriae status, although 50% of the herds had contact to these birds (data not shown).
ur
Sparrows (Passer sp.) were the most common bird species (87.5% positive answers) and contact to them was reported by about 60% of the farmers. Although percentage of herds with contact was
Jo
higher in the case group, this difference was not statistically significant. Wild boars can harbour B. hyodysenteriae (Reiner et al., 2011). In our study population, they do not seem to be involved in SD epidemiology. Only 36.3% of the farmers had noted wild boars near their pig herds and only six had observed contact between the two pig groups. Also none of the variables like slurry and slurry management, soy content of feed, access to outdoor areas or a considerable number of different person groups entering the pig barns was a risk factor in our study. However, we cannot exclude that 16
some potential risk factors remained unnoticed because of the limited number of study herds only allowing detecting the factors with the highest impact. The fact that over 50% of the farmers shared the slurry material with other pig farmers, that access to outdoor areas was provided in 71.3% and that up to ten different person groups (the actual number of entries is higher but has not been examined in detail) had access to the pigs represent a potential risk of introduction of B. hyodysenteriae. This is of special concern because the overall biosecurity measures like fences, nets warding birds or protective clothes and boots for the visitors require improvement in many of the examined herds.
ro of
Another concern is the pig transport. Suitable ramps were lacking in most herds and combination of different batches from different source herds of unknown B. hyodysenteriae status and for different fattening herds increase the risk for transmission of B. hyodysenteriae in a large number of herds. All
-p
interview partners confirmed that at least once per year other pigs were already present on the vehicles when arriving at their farm. Although the source and number of the foreign batches was not
re
examined in the present study, data from other examinations in Switzerland underline the presence and importance of contact between different batches during transportation (Kümmerlen et al., 2019).
lP
Further studies are needed to examine the importance for B. hyodysenteriae epidemiology in detail.
na
Participation in the study was voluntary. This might have caused bias towards farmers with more interest in SD and its prevention. A disadvantage (but inevitable feature) of the study design was the
ur
retrospective aspect of herd classification. The classification into control herds was based on data from the farmers (‘no detection of B. hyodysenteriae, no signs of SD’). No clinical signs of SD were present
Jo
in the control herds during the examination. However, we cannot fully exclude that subclinically infected herds were among the group of control herds. Testing of these herds would have required large sample sizes resulting in high costs. Also, our sample did not contain breeding and piglet producing herds although they can be B. hyodysenteriae positive too (Speiser, 2008; Löbert et al., 2016) so future studies should specifically address this production type. Due to rather small number of cases
17
(20) in our study, some of the factors not found significant should not be disregarded as non-influential but rather be specifically adressed in future studies. Purchase of pigs (from uncontrolled sources) was a risk factor in the one study using statistical analysis to examine risk factors for SD (Robertson et al., 1992). Serologic results had been used to classify positive and negative herds thus making results are not fully comparable to our study. Here, herds receiving pigs from B. hyodysenteriae positive or suspicious source herds were 11.2 times more likely to be in the case group. However, detection of B. hyodysenteriae in the case herds might have
ro of
influenced some factors. For instance, several farmers of case herds reported that pig trade and transported had changed after the diagnosis. Now, they systematically received pigs from positive source herds. Secondly, none of the case herds had been approached as first herd during transport of
-p
slaughter pigs. On the contrary: many farmers of case herds reported being the last farm approached to load slaughter pigs because of their positive B. hyodysenteriae status and consecutive adaption of
re
transport mode by the pig traders. Therefore, we consider these factors more a result than a source of SD infection. Also other factors might have changed after diagnosis which is inevitable and hard to
lP
control in retrospective studies. Therefore, some factors might have been underestimated or missed in the present study. Furthermore, random control of actual pig herds in the neighbourhood showed
na
that the farmers did not know or mention every other pig herd in their region resulting in an underestimation of neighboured pig herds in the present study. Additionally, no information was
ur
available on the SD status of the neighbour pig herds. Therefore, the potential risk originating from pig herds in short distance, e.g. 430 m, the median distance to case herds, can be crossed especially by
Jo
rats (Backhans and Fellström, 2012), could not be evaluated. A national database registering and monitoring the status of every herd would help to control
SD. Even though the reporting of pig transports is mandatory in Switzerland, many aspects are not considered (mixing of batches in trucks) which makes it very difficult to reconstruct the chains of transmission.
18
In recent years, two new pathogenic Brachyspira spp. causing clinical signs of SD have been identified and designated B. suanatina (Råsbäck et al., 2007; Oren and Garrity, 2016) and B. hampsonii (Chander et al., 2012; Oren and Garrity, 2017). Both pathogens have not been described in Swiss pig herds until now (Scherrer et al., 2016) and have therefore not been subject of the present study. However, as especially B. hampsonii causes increasingly problems in some areas (Mirajkar et al., 2015; Johnson et al., 2018), future risk factor studies should include also this pathogen and potentially
ro of
B. suanatina as well.
5. Conclusions
In conclusion, our study highlights the importance of biosecurity and All-In-All-Out management.
-p
Although the importance of pig trade could not be fully determined in this retrospective study, we would like to emphasize also the knowledge in the infection status of source herds to prevent infection.
re
Biosecurity and management are (under the current policy) in the responsibility of the particular farmer. With respect to pig trade, combined efforts of the supplier farmers, the pig traders and the
lP
receiving farmers are necessary to prevent spread of B. hyodysenteriae.
Jo
ur
na
.
19
Acknowledgment We would like to thank all pig farmers for their valuable participation in the study, the feed supplier for data of soy contents and Giulia Paternoster for creating the map.
Conflict of interest None.
ro of
Funding: This research did not receive any specific grant from funding agencies in the public, commercial, or
Jo
ur
na
lP
re
-p
not-for-profit sectors.
20
References
Alvarez-Ordóñez, A., Martínez-Lobo, F.J., Arguello, H., Carvajal, A., Rubio, P., 2013. Swine dysentery: Aetiology, pathogenicity, determinants of transmission and the fight against the disease. Int. J. Environ. Res. Public Health 10, 1927–1947. Backhans, A., Fellström, C., 2012. Rodents on pig and chicken farms – a potential threat to human and animal health. Infect. Ecol. Epidemiol. 2.
chickens on Swedish farms. Vet. Microbiol. 153, 156–132.
ro of
Backhans, A., Jansson, D., Aspán, A., Fellström, C., 2011. Typing of Brachyspira spp. from rodents, pigs and
Backhans, A., Johansson, K.-E., Fellström, C., 2010. Phenotypic and molecular characterization of Brachyspira spp. isolated from wild rodents. Environ. Microbiol. Rep. 2, 720–7.
-p
Blunt, R., Hancox, L., Mellits, K., McOrist, S., 2010. Likely carriage of Brachyspira hyodysenteriae in cockroaches and flies on pig farms, in: Proceedings of the 21th International Pig Veterinary Society Congress in
re
Vancouver. Vancouver, p. 93.
Boye, M., Baloda, S.B., Leser, T.D., Møller, K., 2001. Survival of Brachyspira hyodysenteriae and B. pilosicoli in
lP
terrestrial microcosms. Vet. Microbiol. 81, 33–40.
Burrough, E.R., 2016. Swine Dysentery: Etiopathogenesis and diagnosis of a reemerging disease. Vet. Pathol. 0300985816653795.
na
Chander, Y., Primus, A. E., Olivera, S.R., Gebhart, C.J., 2012. Phenotypic and molecular characterization of a novel strongly hemolytic Brachyspira species, provisionally designated “Brachyspira hampsonii.” J. Vet.
ur
Diagnostic Investig. 24, 903–10.
Demidenko, E., 2006. Sample size determination for logistic regression revisited. Stat. Med. 26, 3385–3397.
Jo
Durmic, Z., Pethick, D.W., Mullan, B.P., Accioly, J.M., Schulze, H., Hampson, D.J., 2002. Evaluation of largeintestinal parameters associated with dietary treatments designed to reduce the occurrence of swine dysentery. Br. J. Nutr. 88, 159–169.
Elmberg, J., Berg, C., Lerner, H., Waldenström, J., Hessel, R., 2017. Potential disease transmission from wild geese and swans to livestock, poultry and humans: a review of the scientific literature from a One Health perspective. Infect. Ecol. Epidemiol. 7, 1300450. Fleming, P., Dundas, S., Lau, Y., Pluske, J., Fleming, P.A., Dundas, S.J., Lau, Y.Y.W., Pluske, J.R., 2016. Predation
21
by red foxes (Vulpes vulpes) at an outdoor piggery. Animals 6, 60. Giacomini, E., Gasparrini, S., Lazzaro, M., Scali, F., Boniotti, M.B., Corradi, A., Pasquali, P., Alborali, G.L., 2018. The role of transportation in the spread of Brachyspira hyodysenteriae in fattening farms. BMC Vet. Res. 14, 10. Hampson, D.J., La, T., Phillips, N.D., 2015. Emergence of Brachyspira species and strains: reinforcing the need for surveillance. Porc. Heal. Manag. 1, 8. Hidalgo, Á., Carvajal, A., García-Feliz, C., Osorio, J., Rubio, P., 2009. Antimicrobial susceptibility testing of Spanish field isolates of Brachyspira hyodysenteriae. Res. Vet. Sci. 87, 7–12.
ro of
Jacobson, M., Fellstrom, C., Lindberg R., Wallgren, P., Jensen-Waern, M., 2004. Experimental swine dysentery: comparison between infection models. J. Med. Microbiol. 53, 273–280.
Jansson, D.S., Johansson, K.E., Olofsson, T., Råsbäck, T., Vågsholm, I., Pettersson, B., Gunnarsson, A., Fellström, C., 2004. Brachyspira hyodysenteriae and other strongly -haemolytic and indole-positive spirochaetes
-p
isolated from mallards (Anas platyrhynchos). J. Med. Microbiol. 53, 293–300.
re
Jansson, D.S., Råsbäck, T., 2009. Experimental Challenge of Mallards (Anas platyrhynchos) with Brachyspira hyodysenteriae and “ Brachyspira suanatina” isolated from pigs and mallards. J. Comp. Pathol. 141, 211–
lP
222.
Jensen, N.S., Stanton, T.B., Swayne, D.E., 1996. Identification of the swine pathogen Serpulina hyodysenteriae in rheas (Rhea americana). Vet. Microbiol. 52, 259–69.
na
Johnson, L.A., Fernando, C., Harding, J.C.S., Hill, J.E., 2018. Characterization of Brachyspira communities from clinical cases of swine mucohaemorrhagic diarrhea through deep sequencing of the NADH oxidase (nox)
ur
gene. Vet. Microbiol. 214, 81–88.
Löbert, S., Zimmermann, W., Bürki, S., Frey, J., Nathues, H., Scheer, P., Doherr, M.G., Stalder, U., Zeeh, F., 2016.
Jo
Occurrence of Brachyspira hyodysenteriae in multiplier pig herds in Switzerland. Tierarztl. Prax. Ausgabe G Grosstiere - Nutztiere 44, 13–18.
Kümmerlen, D., Hartmann, S., Riklin, A., Figi, R., Sidler, X., 2019. Aspects of animal health, animal welfare and biosecurity during 101 transports of piglets in Switzerland. Schweiz Arch Tierheilkd 161, 153–163. Massacci, F.R., De Luca, S., Cucco, L., Tentellini, M., Perreten, V., Pezzotti, G., Magistrali, C.F., 2018. Multiresistant Brachyspira hyodysenteriae shedding by pigs during the fattening period. Vet. Rec. vetrec2018-104886.
22
Medhanie, G.A., McEwen, S.A., Slavic, D., Guerin, M.T., 2013. Brachyspira spp. and avian intestinal spirochaetosis: an epidemiological review. Worlds. Poult. Sci. J. 69, 541–552. Mirajkar, N.S., Bekele, A.Z., Chander, Y.Y., Gebhart, C.J., 2015. Molecular epidemiology of novel pathogen “Brachyspira hampsonii” reveals relationships between diverse genetic groups, regions, host species, and other pathogenic and commensal Brachyspira species. J. Clin. Microbiol. 53, 2908–18. Mirajkar, N.S., Davies, P.R., Gebhart, C.J., 2016. Antimicrobial susceptibility patterns of Brachyspira species isolated from swine herds in the United States. J. Clin. Microbiol. 54, 2109–2119. Oren, A., Garrity, G.M., 2017. List of new names and new combinations previously effectively, but not validly,
ro of
published. Int. J. Syst. Evol. Microbiol. 67, 529–531. Oren, A., Garrity, G.M., 2016. List of new names and new combinations previously effectively, but not validly, published. Int. J. Syst. Evol. Microbiol. 66, 4299–4305.
Pearson, H.E., Toribio, J.-A.L.M.L.M.L., Lapidge, S.J., Hernández-Jover, M., 2016. Evaluating the risk of pathogen
-p
transmission from wild animals to domestic pigs in Australia. Prev. Vet. Med. 123, 39–51.
re
Phillips, N.D., La, T., Adams, P.J., Harland, B.L., Fenwick, S.G., Hampson, D.J., 2009. Detection of Brachyspira hyodysenteriae, Lawsonia intracellularis and Brachyspira pilosicoli in feral pigs. Vet. Microbiol. 134, 294–
lP
9.
Pluske, J.R., Pethick, D.W., Hopwood, D.E., Hampson, D.J., 2002. Nutritional influences on some major enteric bacterial diseases of pig. Nutr. Res. Rev. 15, 333–371.
na
Råsbäck, T., Jansson, D.S., Johansson, K.-E., Fellström, C., 2007. A novel enteropathogenic, strongly haemolytic spirochaete isolated from pig and mallard, provisionally designated “Brachyspira suanatina” sp. nov.
ur
Environ. Microbiol. 9, 983–991.
Reiner, G., Winkelmann, M., Willems, H., 2011. Prevalence of Lawsonia intracellularis, Brachyspira
Jo
hyodysenteriae, and Brachyspira pilosicoli infection in hunted wild boars (Sus scrofa) in Germany. Eur. J. Wildl. Res. 57, 443–448.
Robertson, I., Mhoma, J., Hampson, D., 1992. Risk factors associated with the occurrence of swine dysentery in Western Australia: results of a postal survey. Aust. Vet. J. 69, 92–93. Rubin, J.E., Harms, N.J., Fernando, C., Soos, C., Detmer, S.E., Harding, J.C.S., Hill, J.E., 2013. Isolation and characterization of Brachyspira spp. Including “Brachyspira hampsonii” from lesser snow geese (Chen caerulescens caerulescens) in the Canadian arctic. Microb. Ecol. 66, 813–822.
23
Scherrer, S., Borgström, A., Frei, D., Wittenbrink, M.M., 2016. First screening for Brachyspira hampsonii in Swiss pigs applying a new high resolution melting assay. Vet. Microbiol. 193, 17–21. Songer, J.G., Glock, R.D., Schwartz, K.J., Harris, D.L., 1978. Isolation of Treponema hyodysenteriae from sources other than swine. J. Am. Vet. Med. Assoc. 172, 464–6. Speiser, S.A., 2008. Untersuchungen zu “Lawsonia intracellularis-, Brachyspira hyodysenteriae”- und “Brachyspira pilosicoli”-Infektionen bei Absetz- und Mastschweinen in der Schweiz. Diss. Vetsuisse Fac. Berne. Bern. Thakur, K.K., Revie, C.W., Hurnik, D., Poljak, Z., Sanchez, J., 2016. Analysis of Swine Movement in Four Canadian
ro of
Regions: Network Structure and Implications for Disease Spread. Transbound. Emerg. Dis. 63, e14–e26. Windsor, R.S., Simmons, J.R., 1981. Investigation into the spread of swine dysentery in 25 herds of East Anglia and assessment of its economic significance in five herds. Vet. Rec. 109, 482–4.
Zeeh, F., Klausmann, S., Masserey, Y., Nathues, H., Perreten, V., Rohde, J., 2018. Isolation of Brachyspira
Jo
ur
na
lP
re
-p
hyodysenteriae from a crow (Corvus corone) in close proximity to commercial pigs. Vet. J. 236, 111–112.
24
Table 1 Results of univariable logistic regression models of risk factors and protective factors for Swine Dysentery in Swiss pig herds. The first category presented is the reference category. Variable
Level
OR
CI 95%
P
-
-
-
11.2
3.0-49.1
0.001
No
-
-
-
Yes
8.4
Risk factors BH-status of source herds
Negative
Regular treatment
ro of
Positive/suspicious
2.7-29.8
< 0.001
-
-
1.7-44.8
0.017
-
-
-
4.9
1.4-17.6
0.013
-
-
-
Yes
4.4
1.5-13.3
0.007
Dry
-
-
-
Wet
3.5
1.1-13.3
0.042
No
-
-
-
Yes
3.1
1.1-9.3
0.034
-
-
-
3.1
1.1-9.3
0.044
Few (0-4)
-
(batches) per yeara
Many (>4)
6.7
Contact to foxesa
No
last 12 months
unita
ur
Rats on farma
na
Feeding system in fattener
No
lP
Diagnostics performed during
re
Yes
-p
Number of pig purchases
Jo
Herd size (fattening places) a
Small (80-250) b Big (>250)
Protective factors Different sources of grower
Noc
-
-
-
pigs within one batch
Yes
0.3
0.08-0.8
0.018
25
Raptor birds present in region
Martens present in region
No
-
-
-
Yes
0.1
0.03-0.5
0.003
No
-
-
-
Yes
0.1
0.01-0.5
0.024
OR: odds ratio CI: confidence interval
ro of
BH: Brachyspira hyodysenteriae Variable was included in the final multivariable model.
b
The minimum number of fattening places was 80 (see methods, inclusion criteria).
c
’No’ includes ‘only own pigs’.
Jo
ur
na
lP
re
-p
a
26
ro of -p re
lP
Figure 1: Geographic distribution of 20 case herds (Brachyspira hyodysenteriae positive, red asterix) and 60 control herds (green square) included in the risk-factor study. Actual number of symbols is
na
lower than 80 because of overlapping of neighbouring herds. The inserted map illustrates the distribution of pigs in Switzerland. The size of the pink circles
ur
correlates to the number of pigs. (Source: Federal Statistical Office, Switzerland. Data: 2016. Accessed: 28 September 2018 with Link:
Jo
https://www.atlas.bfs.admin.ch/maps/13/map/mapIdOnly/0_de.html. Current link: https://www.atlas.bfs.admin.ch/maps/13/de/12451_5892_5872_4801/20564.html (09 July 2019))
27
PII:
S0167-5877(19)30479-9
DOI:
https://doi.org/10.1016/j.prevetmed.2019.104819
Reference:
PREVET 104819
To appear in:
Preventive Veterinary Medicine
Received Date:
5 August 2019
Revised Date:
28 October 2019
Accepted Date:
29 October 2019
Please cite this article as: Zeeh F, Vidondo B, Nathues H, Risk factors for the infection with Brachyspira hyodysenteriae in pig herds, Preventive Veterinary Medicine (2019), doi: https://doi.org/10.1016/j.prevetmed.2019.104819
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier.
Risk factors for the infection with Brachyspira hyodysenteriae in pig herds
Friederike Zeeha, 1*, Beatriz Vidondob, Heiko Nathuesa
Affiliations: a
Clinic for Swine, Vetsuisse Faculty, University of Bern, Bremgartenstrasse 109a, PB 3350, 3001 Bern,
Switzerland b
ro of
Veterinary Public Health Institute, Vetsuisse Faculty, University of Bern, Schwarzenburgstrasse 155,
3097 Liebefeld, Switzerland
[email protected] (F. Zeeh)
re
[email protected] (B. Vidondo)
-p
E-mail addresses:
Corresponding author:
na
*Friederike Zeeh
lP
[email protected] (H. Nathues)
Clinic for Swine, Vetsuisse Faculty, University of Bern, Bremgartenstrasse 109a, PB 3350, 3001 Bern,
ur
Switzerland
[email protected] +41 (0)31 631 23 44
Fax:
+41 (0)31 631 26 31
Jo
Tel.:
Present address 1
Jockey Club College of Veterinary Medicine and Life Sciences, Department of Infectious Diseases
and Public Health, City University of Hong Kong, Hong Kong SAR. [email protected] 1
Abstract Swine dysentery (SD), caused by infection with Brachyspira hyodysenteriae, is a serious disease in pig production worldwide. Quantitative risk factors triggering the occurrence of infection are unknown. The present case-control study aimed at identifying major risk factors related to presence of B. hyodysenteriae in pig herds. Twenty case herds and 60 randomly selected control herds with a minimum herd size of ‘10 sows/ 80 fattening pigs’ were examined by means of a questionnaire-based interview and a herd examination. Herds with previous eradication of SD were excluded. Logistic
ro of
regression models revealed that the ‘positive/suspicious SD status of source herds’, the regular application of treatment, purchasing more than 4 batches/ year, contact to foxes, diagnostics performed during last 12 months, liquid feeding systems, rats on farm, and >250 fatting places were
-p
associated with higher chances of a herd to be infected. On the contrary, having different sources of grower pigs within one batch, the presence of raptor birds and the presence of martens in the region
re
were associated with fewer chances of being infected. The final multivariable logistic regression model identified purchasing more than 4 batches/ year (OR = 7.5, 95% CI 1.8-54.3) and contact to foxes (OR
lP
= 5.9; 97.5% CI 1.2-34.6) as the two main risk factors in our study. ‘More than 4 batches/ year’ implies continuous herd management supporting persistence of B. hyodysenteriae in an infected herd, but
na
also increased number of purchases each increasing the risk of B. hyodysenteriae introduction by carrier pigs or transport vehicles. Foxes might be infected with B. hyodysenteriae by feeding on positive
ur
piglets and rodents. Besides, ‘contact to foxes’ might represent a lack in biosecurity. In conclusion, the
Jo
risk factors detected underline the importance of biosecurity in SD prevention and control.
Keywords
Swine Dysentery; Epidemiology; Protective factor; Biosecurity; Fox
2
1. Introduction
Swine dysentery (SD), caused by Brachyspira hyodysenteriae, is of high clinical and economic importance worldwide due to severe intestinal lesions but also subclinical infections resulting in decreased growth rate and feed conversion (Hampson et al., 2015). Transmission of B. hyodysenteriae takes place via the faecal-oral route and treatment requires application of effective antimicrobials over a prolonged period and sometime repeatedly. Control and eradication measures of B. hyodysenteriae
ro of
have gained considerable importance during the last years, especially because of an increase in frequency and amount of antimicrobial resistance (Hidalgo et al., 2009; Mirajkar et al., 2016; Massacci et al., 2018) and the absence of a commercially available vaccine. Basis for a sustainable control or
factors for transmission and spread of B. hyodysenteriae.
-p
eradication are reliable detection methods, but also broad knowledge on epidemiology including risk
re
In respect to risk factors for B. hyodysenteriae infection, only a limited number of studies have been published. In a report from Australia, data from a postal survey performed in herds with serologic
lP
results indicative of B. hyodysenteriae and in serologic negative herds were analysed (Robertson et al., 1992). Factors linked to uncontrolled pig purchase, visitors and rodents (amongst others) had a
na
significantly higher odds ratio >1 for being positive and protective factors were for instance protective clothes or footbaths. Beside this one study using statistical analysis for description of potential risk
ur
factors, a number of studies with empiric data have been published. Transmission of B. hyodysenteriae by carrier pigs is considered to be the main route. Hence, pig trade and transport may pose a high risk.
Jo
This was also supported by a recent study describing B. hyodysenteriae in transport vehicles (Giacomini et al., 2018). Besides pig movement also other factors like management, positive neighbour herds, pig transport, feed lorry, and birds were involved in regional spread of SD (Windsor and Simmons, 1981). Also feral pigs and wild boars (Phillips et al., 2009; Reiner et al., 2011), rodents (Backhans et al., 2010), dogs (Songer et al., 1978) and birds like mallards (Jansson and Råsbäck, 2009), geese (Rubin et al., 2013), rheas (Jensen et al., 1996) and a crow (Zeeh et al., 2018) have been tested positive for 3
B. hyodysenteriae. Furthermore, B. hyodysenteriae can survive in slurry (Boye et al., 2001) and thus rapidly spread within herds and potentially between herds. Additionally, in an experimental challenge study, Musca flies and cockroaches were B. hyodysenteriae positive in their gastrointestinal tract and also excreted viable bacteria (Blunt et al., 2010). Factors enhancing spread within a herd are e.g. continuous management of pens or barns and poor husbandry (Alvarez-Ordóñez et al., 2013; Burrough, 2016). All these findings indicate that there might be a number of potential risk factors for B. hyodysenteriae infection warranting further and more detailed (statistical) analysis.
ro of
The aim of the study was to systematically examine potential risk factors for the infection with B. hyodysenteriae. Identification of such factors would support prevention and eradication of
-p
B. hyodysenteriae infections and by this improvement of pig health.
lP
2.1. Study design and herd selection
re
2. Methods
The case-control study was performed in Swiss pig herds. Twenty case herds were selected from
na
herds with previous detection of B. hyodysenteriae. Per case herd, three control herds were randomly selected from a database containing information of about 90% of all Swiss pig herds. Randomisation
ur
was done by ‘random number’ in Excel. Case and control herds were selected according to the production type (piglet producing versus grower-finisher herd). As it became obvious during the first
Jo
months that the category ‘piglet producer without grower-finisher pigs’ would be underrepresented in the study population, this herd type was not considered in the project anymore. Only herds with more than 10 sows and/or more than 80 grower-finisher pigs were included because this herd size is more likely to exist in the future. Herds that had undergone an SD eradication were excluded as biosecurity measures might have been changed after the eradication. The sample size of 20 case and 60 control herds was chosen based on the limited number of B. hyodysenteriae positive herds meeting 4
all inclusion criteria and the pig owners’ availability to perform the interviews (see below) during the study period. This sample size rendered a power of 80% at 0.05 significance level to detect odds ratios larger than 4 (Demidenko, 2006).
2.2. Questionnaire and interview
Pig farmers were contacted by phone and asked to participate in the study. The participation
ro of
was voluntary. A questionnaire was created and validated in three herds not included in the present study. The validated questionnaire was completed by the principal investigator on farm together with the farmer.
-p
A total of 497 questions covered general information about the herd, housing, management, hygiene, internal and external biosecurity (with emphasis on slurry management and fly and rodent control),
re
pig trade and feed and feeding, during the last 12 months to reduce recall bias. To ensure a standard across farms, subjective parameters, such as ‘hygiene’ or ‘rodent burden’,
lP
etc., were judged by the principal investigator during a herd examination immediately after the
na
interview.
ur
2.3. Data processing and statistical methods for data analysis
The data from the questionnaire were transferred in a data management programme (Microsoft
Jo
InfoPath 2010), subsequently transformed in an Excel table (Microsoft Excel, 2010) and imported to the statistical software R (R version 3.3.1) for further analysis. Data were visually checked for completeness and, if biological plausible, combined or imputed
when necessary. Continuous data were tested for normality using histograms and by Shapiro-Wilk test. Using logistic regression models, each of the measured farm-level properties was tested for association with the probability of being a case. Correlations between explanatory variables were assessed by 5
means of Spearman correlation coefficients. Only those that did not correlate to each other were allowed in the same multivariable model. Multivariable logistic regression models were built using both forward and backward stepwise variable selection. This ensured robustness of the final model. For forward selection, variables with the highest odds ratios (OR) were selected first. Only models with biological significance were retained. Control for confounding was carried out during the variable selection processes because variables were retained if they caused a change of more than 25% in the odds ratios of the other variables in the model. The geographic location of herds was allocated to one
ro of
of the three regions East (cantons Aargau, Appenzell Inner Rhoden, St Gallen, Thurgau, Zurich), Middle (Basel-Land, Lucerne, Schwyz) and West (Berne, Friborg, Solothurn, Vaud). Region was not associated with the chances of having SD and did not affect the results of the final model. For all models, the level
-p
of significance was set at 0.05.
re
Geographical distribution of the examined herds was visualized using QGIS (www.qgis.org).
lP
3. Results
Of 120 contacted pig farmers, 32 refused to participate because of ‘no interest’ (n = 15), ‘no pigs
na
anymore’ (n = 8) or ‘personal reasons incl. not reached’ (n = 9), resulting in a response rate of 73.3%. Additional four herds were excluded after the interview because of not having met the initial inclusion
ur
criterion of herd size. Further four herds (piglet producers without fattening pigs) were retroactively excluded because of the revised inclusion criteria. Finally, 20 case herds and 60 control herds,
Jo
examined between August 2014 and November 2017, were included in the study. The 80 herds were located in 12 Swiss cantons. The majority of the herds was situated in the
cantons of Berne (30.0%), Lucerne (23.8%), and Aargau (15.0%) (Fig. 1).
3.1. Univariable Analysis 3.1.1. Herd characteristics 6
The majority of herds were grower-finisher herds (85% (17/20) and 90% (54/60) in the cases and controls, respectively. Case herds also comprised two wean-to-finish herds (10%), and one farrow-tofinish herd (5%). Within the control herds, the remaining six were farrow-to-finish herds (10%). Three of these farrow-to-finish herds were closed herds with no pig purchase in the last 12 months. All other 77 pig owners reported pig purchases. Due to the low number of herds with sows, piglets and/or weaned pigs, parameters describing these age categories are not analysed further.
ro of
All case herds and 45 (75%) control herds were monitored by a pig health service (SUISAG). Production under ‘animal-friendly’ labels (e.g. requiring more space and/or access to outdoor areas) was present in 45 herds (50% of the case herds, 59% of the control herds).
-p
The median number of fattening places (‘herd size’) of all 80 herds was 255 (range: 80-1000), of the case herds 383 (range: 104-1000), and 230 (range: 80-1000) in the control herds. Seventy of the
re
farmers (85.0% in case herds, 88.3% in control herds) had other lines of business besides the pigs with mainly tillage (n = 53) and dairy cows and/ or raising calves (n = 49; multiple answers possible). Seven
lP
out of the 15 persons working also outside the own farm had contact to foreign pigs. In 85.0%, the interview partner was the pig owner. Tenants (n = 6), contractors or employees (3 each) were the
na
minority.
ur
3.1.2. Trade and transport
Jo
Nineteen different trading companies were involved in the pig trade of incoming pigs and 18 of
them also in trade of outgoing pigs. Only 11.3% (3 out of 20 cases, 6 out of 60 controls) and 6.4% (1 out of 20 cases, 4 out of 58 controls) of the herds had no trader for incoming or outgoing pigs, respectively. In the 12 months prior the interview, the median number of source herds was 1.0 (range 1-36) in the case herds and 2.5 (range: 1-27) in the control herds. Average number of batches per year, 7
according to the 75 pig farmers who bought weaner or grower pigs, was in median 17 (range: 1-52) in the case herds and 6 (range: 1-52) in control herds. When herds were classified into herds with few (04) and many (>4) batches/year, then case herds were significantly more often in the group with many batches (88.9% versus 54.5%; OR = 6.66; P = 0.017, 95% confidence interval (CI): 1.68-44.76). Batches of grower pigs comprised in 25.0% of the case herds and in 56.7% of the control herds pigs of more than one source (OR = 0.3; P = 0.018; CI: 0.1-0.8). The maximum number of sources per batch, according to the data from the interview partners, was in median 1 (1-4) in case herds and 1 (1-9) in
ro of
control herds. The status of the source herds in respect to B. hyodysenteriae infection (according to the pig health service, the herd attending vet or the pig owners; last 12 months) differed significantly between the two herd types. Case herds had received in 44.4% pigs from B. hyodysenteriae positive or
-p
suspicious herds, control herds in 6.7% (OR = 11.2; P < 0.01; CI: 3.0-49.1). The classification as positive herd is most reliable for breeding and piglet producing herds, because most of them belong to the pig
re
health services (100% and over 85%, respectively). SD is regulated in the pig health services programmes and trained veterinarians provide on-farm investigations and order laboratory
lP
confirmation tests either on a mandatory and regular basis (breeding herds) or in case of SD suspicion in the other affiliated herds. For herds of farmers who do not access the pig health services, the
na
detection of B. hyodysenteriae was asked from the herd-attending veterinarian. Suspicious herds were for instance farrowing herds of a B. hyodysenteriae positive sow pool system without own detection
ur
of the pathogen.
Transport of grower pigs included transport by the trader (9 out of 18 of case herds, 38 out of
Jo
57 of control herds), sub-contractors of the trader (5 out of 18 case herds; 3 out of 57 control herds), the piglet producer (4 out of 18 case herds; 3 out of 57 control herds) or the farmer of the fattening herd himself (2 out of 18; 9 out of 57 control herds; multiple answers possible). Slaughter pigs were transported mainly by traders (50.0% of case herds; 50.8% of control herds) and sub-contractors (40.0% of case herds; 25.4% of control herds).
8
Sequence of farms in respect to transport of incoming pigs was in 45.5% of the herds (55.0% of the case herds; 42.1% of the control herds) ‘always the first farm being approached during the delivery tour’. Thirty-nine percent of the interview partners (30% of case herds and 42.1% of control herds) reported the presence of pigs dedicated to other farms on the transport vehicles. Regarding collection of outgoing (slaughter) pigs, none of the case herds (n = 20) was approached first. Of the control herds, 20.7% (12 out of 58) were approached first. When asked, if the transport vehicles for slaughter pigs were always empty at arrival, then 82.3% (95.0% case herds; 78.0% control herds) of the participants
ro of
answered ‘No’.
-p
3.1.3. External biosecurity
A loading bay was present in 45.0% of the case herds whereof four (44.4%) were rated as
re
‘suitable’ (preventing re-entrance of slaughter pigs into the barn), and only one (11.1%) was clean as assessed by the first author. In the 22 out of the 56 control herds with a loading bay, 13 (59.1%) of
lP
them were ‘suitable’ and 10 (45.5%) ‘clean’.
In all 80 herds, at least one external person have had accessed the pig barn in the 12 months
na
prior the interview. The veterinarian had been present in the stable in the last 12 months in 70.0% of the case herds (n = 14) and in 63.3% of the control herds (n = 38). Apart from the veterinarian, between
ur
1 to 8 different person groups had entered the barn of case herds and between 0-9 of control herds. Protective clothes were provided to visitors in 52.2% (11 case herds, 31 control herds) and, as assessed
Jo
by the first author, used in 42.5% of the herds (7 case herds, 27 control herds). Distance to the next pig barn was, according to the farmers, for case herds in median 430 m (50-
8’000) and for controls 600 m (20-5’000). The median number of neighbour pig herds in a radius of 500 m was 1 for cases and 0 for controls, in a radius between 0.6 -1 km and in a radius between 1.1 and 2 km one neighbour herd each.
9
Access to outdoor areas was available for the pigs in 71.3% of the herds (12 case herds, 45 control herds). In 56 herds, this area had concrete floor, and one herd of the control group used a pasture. Dogs and/or cats had access to the pig barns of 51.9% herds (11 case herds, 30 control herds). Presence of wild animals in the region was stated by 97.5% of all farmers (19 case herds, 59 control herds). Mainly sparrows (87.5%), foxes (83.8%) and crows (71.3%) were reported. Wild boars were reported by 36.3% of the interview partners (5 case herds, 24 control herds). Presence of wild raptor birds in the region was significantly related to a negative B. hyodysenteriae status (15.0% of case herds,
ro of
56.7% of control herds; OR = 0.1; P = 0.003; CI: 0.03-0.5). Contact to wild animals was reported in 85.0% of the case herds and 78.3% of the control herds. Especially contact to wild birds, including sparrows and crows, were reported in both herd types (16 case herds, 42 control herds). Contact to foxes
-p
occurred in 35.0% of the case herds (7 out of 20) and 10.0% of the control herds (6 out of 60) (OR = 4.9; P = 0.013; CI: 1.4-17.6). Crows came in contact to the domestic pigs of 71.3% of the herds (14 case
re
herds, 43 control herds) and wild boars to 36.3% (5 case herds, 24 control herds). Slurry is considered a reservoir for B. hyodysenteriae. In our study, foreign slurry was deposed
lP
in the vicinity of 38.8% of the herds (8 out of 20 case herds, 23 out of 60 control herds). The own slurry was stored mainly in pits (80.0% of case herds, 76.7% of control herds). Material used for applying the
na
manure was diverse (barrels in 58 herds, tubes in 43 herds, trailing tubes in 51 herds) and shared with other pig farmers in 43 herds (10 out of 20 case herds, 33 out of 59 control herds). Treatment of slurry
ur
was performed in five case herds and 19 control herds and included i.e. regular agitation or application of products improving the flow-ability. Forty-five herd farmers (14 case herds, 31 control herds) gave
Jo
pig slurry away and only eight received foreign slurry (2 case herds, 6 control herds). Of these 45 herds with slurry contracts, five herds had both contract types (2 case herds, 3 control herds). Vaccination status of incoming pigs was known to 83.5% of the farmers (15 out of 20 cases, 51
out of 59 controls).
3.1.4. Internal biosecurity 10
All-in/All-out management was applied to the fattening units in 23 control herds but in none of the case herds. The floor in the fattening barns was in most cases partially slatted (75.0% case herds, 84.2% control herds) and mainly moderately clean as rated by the investigator. Regular disinfection was applied in 15.0% of the case herds (3 out of 20) and 15.3% of the control herds (9 out of 59). Restocking after clearance of pens was conducted in median after 1 day (range: 0-28) in case herds and after 3 days in control herds (range: 0- 30). Stocking density in pens for grower pigs was in median
ro of
0.75 pigs/ m2 (0.57-2.15) in case herds and 0.86 pigs/ m2 (0.23-2.1) in control herds and in pens for finisher pigs 0.65 pigs/ m2 (0.2-1.79) and 0.63 (0.23-1.6), respectively. Bedding material was used in 75.0% of the case herds (15 out of 20) and in 81.7% of the control herds (49 out of 60) and consisted
-p
in most herds (57 out of 64) of straw. Type and source of activity material, which was provided to all
re
pigs, did not differ between the two herd types. In most cases, organic material (e.g. straw) was used.
lP
3.1.5. Feed and feeding system
Feed supply involved 16 different companies. A home-made mixture of feed was used in 50.0%
na
of the case herds (10 out of 20) and 26.7% of the control herds (16 out of 60). Soy is considered to have an influence on B. hyodysenteriae and occurrence of SD and high soy contents in are used in the diet
ur
of experimental B. hyodysenteriae infection studies to facilitate SD development (Jacobson et al., 2004). In 90.0% of the case herds (18 out of 20) and 86.7% of the control herds (52 out of 60), soy was
Jo
part of the feed. Content of soy in the feed for growers until about 50kg bodyweight, according to declaration of the feed mills or the farmers, was in median 9.5% (range: 0-17; case herds) and 10.5% (0-17.1; control herds), and in the feed for finisher pigs 8.8% (0-17; case herds) and 10.3% (0-21.9; control herds), respectively. By-products were used in 28.8% of the herds (9 case herds, 14 control herds) and consisted in herds of whey (7 case herds, 11 control herds).
11
The feeding system for grower-finisher pigs differed significantly between the case and the control group. A system for liquid feeding was more often present in case (16 out of 20) than in control (32 out of 60) herds (OR = 3.5; P = 0.042; CI: 1.1-13.3). The other herds had feeding systems for dry feed. Feed storage and feeding hygiene did not differ between the two herd types.
3.1.6 Fly and rodent control in fattening units
ro of
Fly control was applied in 58.8% of all herds (10 case herds, 37 control herds). More case herds (9 out of 10) had a more intensive fly control regime (‘every second month to permanent’) compared to the 37 control herds (46.2% with the same frequency). In both herd types, the actual fly burden was
-p
assessed as being in median ‘low’ at the time of the examination.
Presence of rodents was reported by farmers of 19 case herds and 57 control herds. In the case
re
herds, mainly rats and mice (52.6%), followed by mice (47.4%) were observed by the farmers. In the control herds, in 70.2% mice, in 22.8% rats and mice, and in 3.5% each rats and other rodents were
lP
indicated. Rats were significantly more often observed in case herds (10 out of 19, 52.6%) than in control herds (OR = 3.1; P = 0.042; CI: 1.1-13.3). Rodent control in the barn was performed in 72.5% of
na
the herds (15 cases herds, 42 control herds) and in the surrounding of the pig barn in 50.7% (10 out of 16 case herds, 25 out of 53 control herds). Methods used for the fattening barn were mainly
ur
‘rodenticides applied by the farmer’ (12 out of 16 case herds, 37 out of 45 control herds), and ‘cats’ (8 out of 16 case herds, 16 out of 45 control herds). Frequency of rodent control was in case herds most
Jo
often ‘permanent‘, in control herds mainly ‘3-4/year‘. If rodent control in the surrounding was present, then it was mainly ‘permanent‘ in both herd types and comprised rodenticides (6 out of 10 case herds; 15 out of 28 control herds) and cats (2 out of 10 case herds; 18 out of 28 control herds) in case and control herds, respectively. A professional rodent control company was used in only two case herds and one control herd. Signs of rodents, as examined by the first author, were present in 55.6% of the pig barns (10 out of 20 case herds, 32 out of 56 control herds) and in 50.0% of the surroundings (5 out 12
of 16 case herds, 27 out of 48 control herds). The rodent control was assessed by the first author as ‘effective’ (i.e. enough bait boxes at correct place, regular checking) only in two case herds and two control herds.
3.3 Performance and drug usage
Daily weight gain of the grower-finisher pigs in case herds was, according to the farmers, in
ro of
median 780g/d (range: 702-960) and in the control herds 825g/d (range: 493-995). Losses of pigs in the fattening unit were in median 2.0% (range: 0.2-3.9) in case herds and 1.6% (range: 0-7.0) in control herds.
-p
A regular antibiotic therapy was significantly more often used in the 20 case herds (73.7%) compared to the 60 control herds (25.0%) (OR = 8.4; P < 0.01, CI: 2.7-29.8). Within the 20 case herds,
re
14 were regularly treated during the 12 months prior the interview. Seven of the herds were regularly treated against B. hyodysenteriae and additional two against diarrhoea (not further specified). Further
lP
nine case herds have sporadically been treated against B. hyodysenteriae. Tiamulin was used in 12 herds, tylosin in five herds, lincomycin in four herds, and a drug not licensed (and effective) against
na
B. hyodysenteriae was used in one herd. In five of the herds, more than one drug component was used.
ur
A summary of all variables which had a p-value <0.05 in the univariable logistic regression
Jo
models is shown in table 1.
3.2. Multivariable analysis
Based on their high OR, absence of collinearity, statistical significance and biological plausibility, the following variables were considered in the final multivariable model: ‘number of pig purchases (batches) per year’, ‘contact to foxes’, ‘feeding system’, ‘rats on farm’, and ‘herd size’. The variables 13
‘BH-status of source herds’, ‘Diagnostics performed during last 12 months’, and ‘Regular treatment’ were considered to be (also) an consequence (rather than a risk or protective factor) of SD and were therefore not included in the multivariable models. The two variable selection processes gave consistent results. The final model thus consisted of two variables ‘Number of batches/year’ (OR = 7.5; P = 0.02; CI: 1.8-54.3) and ‘Contact to foxes’ (OR = 5.9; P = 0.03; CI: 1.2-34.6) as risk factors.
ro of
4. Discussion
Brachyspira hyodysenteriae infection is one of the re-emerging pig diseases with high economic
-p
impact and it is affecting pig herds regardless of whether the herds are thought to be free from it or having effectively combated the disease. Particular attention should be paid to this bacterium due to
re
its increasing resistance to antimicrobials.
The results of this first systematic study on risk factors for being infected with B. hyodysenteriae
lP
indicate that purchasing more than four batches per year (i.e. continuous management) and contact to foxes are the most important risk factors.
na
‘More than four batches per year purchased’ is linked to an increased number of purchases each increasing the risk of B. hyodysenteriae introduction by carrier pigs or transport vehicles (Thakur et al.,
ur
2016; Giacomini et al., 2018). Secondly, the underlying continuous management of the farms enables the persistence of the pathogen within the herd (Alvarez-Ordóñez et al., 2013).
Jo
‘Contact to foxes’ was a risk factor in our study. Presence of foxes has been linked to presence
of and problems with rodents (Fleming et al., 2016). According to that study, foxes feed of piglets and rodents. Piglets might already harbour B. hyodysenteriae and rodents are known vectors of that pathogen, thus potentially infecting foxes. However, further research is needed to elucidate if foxes are actually vectors. In addition to foxes being potentially a vector, reporting of ‘contact to foxes takes
14
place’ represents gaps in biosecurity. Although ‘access to outdoor areas’ was not a significant risk factor in our study, contact to foxes occurred only in farms with outdoor areas. Herds with liquid feeding were 3.5 times more likely to be in the case group. Although feed has been described as an important factor to enhance or reduce SD (Durmic et al., 2002; Pluske et al., 2002; Jacobson et al., 2004), the feeding system per se has not been discussed in the context of SD until now. Alterations in intestinal microbiome favouring B. hyodysenteriae or other digestive features might play a role, but further research is needed in order to confirm this hypothesis. Also potential
ro of
bias should be considered. Larger herd size (i.e. > 250 places for fattening pigs) was related with an increased likelihood to be in the case group (OR = 3.1). In our study population, larger herds had purchased more batches per
-p
year (rho = 0.65, P < 0.01, Spearman’s rank correlation test), thus increasing the risk to introduce B. hyodysenteriae by pigs, contaminated transport vehicles or other routes as discussed above. Larger
re
herds could also have more source herds, but this was not the case in the present study (Spearman rank: rho = - 0.11, p>0.05). However, large herds were not necessarily managed in a more professional
lP
way and also did exhibit deficits in biosecurity.
Presence of rats was significantly related to case herds. Rats are known carriers and vectors of
na
B. hyodysenteriae (Backhans et al., 2011). Our finding confirms earlier statements that rodents are a risk factor for SD ( Robertson et al., 1992; Pearson et al., 2016;).
ur
Although performance of diagnostics and regular treatment had increased odds ratios in our study, we interpret it as a result of the B. hyodysenteriae infection. In eleven of the case herds,
Jo
diagnostics had been performed whereof seven were in the context of B. hyodysenteriae and treatment in case herds was often directed against B. hyodysenteriae.
Three variables were significantly related to a negative B. hyodysenteriae status (protective factors) in our study. Raptor birds prey on rodents and smaller birds and thus might prevent or reduce B. hyodysenteriae transmission. The low pH in their stomach possibly inhibits or kills Brachyspira 15
bacteria potentially ingested by infected rodents. Also martens (Mustelidae) feed of rodents, but their actual role in SD like that of raptor birds remains unclear too. The biological background of the variable ‘different sources within one batch’ being protective was unexpected and is difficult to explain. The hypothesis that such mixed batches might be related to regular (prophylactic) treatment at time of placement could not be confirmed in our data (rho = 0.02, P > 0.05, Spearman’s rank correlation).
Waterfowl have been described in several studies as carrier and potential vector of
ro of
B. hyodysenteriae ( Jansson et al., 2004; Medhanie et al., 2013; Elmberg et al., 2017;). Our data did not confirm these findings, as waterfowl seems to have an inferior role in the transmission of B. hyodysenteriae in the examined Swiss population. Only 26 out of 80 farmers reported of ducks or
-p
geese in the region and none of contact of these birds to their pigs. Crows can harbour B. hyodysenteriae as recently described in Switzerland (Zeeh et al., 2018). Presence of crows in the
re
neighbourhood of their herds was reported by 71.3% of the farmers. Twenty of them explicitly described contact of these birds with outdoor areas and hence pigs, but there was no difference
lP
between case and control herds. However, this study did not analyse the frequency of these contacts which might be important in successful B. hyodysenteriae transmission. Swallows (Hirundinidae) feed
na
of flies which are vectors of B. hyodysenteriae. In our study, swallows were not related to the B. hyodysenteriae status, although 50% of the herds had contact to these birds (data not shown).
ur
Sparrows (Passer sp.) were the most common bird species (87.5% positive answers) and contact to them was reported by about 60% of the farmers. Although percentage of herds with contact was
Jo
higher in the case group, this difference was not statistically significant. Wild boars can harbour B. hyodysenteriae (Reiner et al., 2011). In our study population, they do not seem to be involved in SD epidemiology. Only 36.3% of the farmers had noted wild boars near their pig herds and only six had observed contact between the two pig groups. Also none of the variables like slurry and slurry management, soy content of feed, access to outdoor areas or a considerable number of different person groups entering the pig barns was a risk factor in our study. However, we cannot exclude that 16
some potential risk factors remained unnoticed because of the limited number of study herds only allowing detecting the factors with the highest impact. The fact that over 50% of the farmers shared the slurry material with other pig farmers, that access to outdoor areas was provided in 71.3% and that up to ten different person groups (the actual number of entries is higher but has not been examined in detail) had access to the pigs represent a potential risk of introduction of B. hyodysenteriae. This is of special concern because the overall biosecurity measures like fences, nets warding birds or protective clothes and boots for the visitors require improvement in many of the examined herds.
ro of
Another concern is the pig transport. Suitable ramps were lacking in most herds and combination of different batches from different source herds of unknown B. hyodysenteriae status and for different fattening herds increase the risk for transmission of B. hyodysenteriae in a large number of herds. All
-p
interview partners confirmed that at least once per year other pigs were already present on the vehicles when arriving at their farm. Although the source and number of the foreign batches was not
re
examined in the present study, data from other examinations in Switzerland underline the presence and importance of contact between different batches during transportation (Kümmerlen et al., 2019).
lP
Further studies are needed to examine the importance for B. hyodysenteriae epidemiology in detail.
na
Participation in the study was voluntary. This might have caused bias towards farmers with more interest in SD and its prevention. A disadvantage (but inevitable feature) of the study design was the
ur
retrospective aspect of herd classification. The classification into control herds was based on data from the farmers (‘no detection of B. hyodysenteriae, no signs of SD’). No clinical signs of SD were present
Jo
in the control herds during the examination. However, we cannot fully exclude that subclinically infected herds were among the group of control herds. Testing of these herds would have required large sample sizes resulting in high costs. Also, our sample did not contain breeding and piglet producing herds although they can be B. hyodysenteriae positive too (Speiser, 2008; Löbert et al., 2016) so future studies should specifically address this production type. Due to rather small number of cases
17
(20) in our study, some of the factors not found significant should not be disregarded as non-influential but rather be specifically adressed in future studies. Purchase of pigs (from uncontrolled sources) was a risk factor in the one study using statistical analysis to examine risk factors for SD (Robertson et al., 1992). Serologic results had been used to classify positive and negative herds thus making results are not fully comparable to our study. Here, herds receiving pigs from B. hyodysenteriae positive or suspicious source herds were 11.2 times more likely to be in the case group. However, detection of B. hyodysenteriae in the case herds might have
ro of
influenced some factors. For instance, several farmers of case herds reported that pig trade and transported had changed after the diagnosis. Now, they systematically received pigs from positive source herds. Secondly, none of the case herds had been approached as first herd during transport of
-p
slaughter pigs. On the contrary: many farmers of case herds reported being the last farm approached to load slaughter pigs because of their positive B. hyodysenteriae status and consecutive adaption of
re
transport mode by the pig traders. Therefore, we consider these factors more a result than a source of SD infection. Also other factors might have changed after diagnosis which is inevitable and hard to
lP
control in retrospective studies. Therefore, some factors might have been underestimated or missed in the present study. Furthermore, random control of actual pig herds in the neighbourhood showed
na
that the farmers did not know or mention every other pig herd in their region resulting in an underestimation of neighboured pig herds in the present study. Additionally, no information was
ur
available on the SD status of the neighbour pig herds. Therefore, the potential risk originating from pig herds in short distance, e.g. 430 m, the median distance to case herds, can be crossed especially by
Jo
rats (Backhans and Fellström, 2012), could not be evaluated. A national database registering and monitoring the status of every herd would help to control
SD. Even though the reporting of pig transports is mandatory in Switzerland, many aspects are not considered (mixing of batches in trucks) which makes it very difficult to reconstruct the chains of transmission.
18
In recent years, two new pathogenic Brachyspira spp. causing clinical signs of SD have been identified and designated B. suanatina (Råsbäck et al., 2007; Oren and Garrity, 2016) and B. hampsonii (Chander et al., 2012; Oren and Garrity, 2017). Both pathogens have not been described in Swiss pig herds until now (Scherrer et al., 2016) and have therefore not been subject of the present study. However, as especially B. hampsonii causes increasingly problems in some areas (Mirajkar et al., 2015; Johnson et al., 2018), future risk factor studies should include also this pathogen and potentially
ro of
B. suanatina as well.
5. Conclusions
In conclusion, our study highlights the importance of biosecurity and All-In-All-Out management.
-p
Although the importance of pig trade could not be fully determined in this retrospective study, we would like to emphasize also the knowledge in the infection status of source herds to prevent infection.
re
Biosecurity and management are (under the current policy) in the responsibility of the particular farmer. With respect to pig trade, combined efforts of the supplier farmers, the pig traders and the
lP
receiving farmers are necessary to prevent spread of B. hyodysenteriae.
Jo
ur
na
.
19
Acknowledgment We would like to thank all pig farmers for their valuable participation in the study, the feed supplier for data of soy contents and Giulia Paternoster for creating the map.
Conflict of interest None.
ro of
Funding: This research did not receive any specific grant from funding agencies in the public, commercial, or
Jo
ur
na
lP
re
-p
not-for-profit sectors.
20
References
Alvarez-Ordóñez, A., Martínez-Lobo, F.J., Arguello, H., Carvajal, A., Rubio, P., 2013. Swine dysentery: Aetiology, pathogenicity, determinants of transmission and the fight against the disease. Int. J. Environ. Res. Public Health 10, 1927–1947. Backhans, A., Fellström, C., 2012. Rodents on pig and chicken farms – a potential threat to human and animal health. Infect. Ecol. Epidemiol. 2.
chickens on Swedish farms. Vet. Microbiol. 153, 156–132.
ro of
Backhans, A., Jansson, D., Aspán, A., Fellström, C., 2011. Typing of Brachyspira spp. from rodents, pigs and
Backhans, A., Johansson, K.-E., Fellström, C., 2010. Phenotypic and molecular characterization of Brachyspira spp. isolated from wild rodents. Environ. Microbiol. Rep. 2, 720–7.
-p
Blunt, R., Hancox, L., Mellits, K., McOrist, S., 2010. Likely carriage of Brachyspira hyodysenteriae in cockroaches and flies on pig farms, in: Proceedings of the 21th International Pig Veterinary Society Congress in
re
Vancouver. Vancouver, p. 93.
Boye, M., Baloda, S.B., Leser, T.D., Møller, K., 2001. Survival of Brachyspira hyodysenteriae and B. pilosicoli in
lP
terrestrial microcosms. Vet. Microbiol. 81, 33–40.
Burrough, E.R., 2016. Swine Dysentery: Etiopathogenesis and diagnosis of a reemerging disease. Vet. Pathol. 0300985816653795.
na
Chander, Y., Primus, A. E., Olivera, S.R., Gebhart, C.J., 2012. Phenotypic and molecular characterization of a novel strongly hemolytic Brachyspira species, provisionally designated “Brachyspira hampsonii.” J. Vet.
ur
Diagnostic Investig. 24, 903–10.
Demidenko, E., 2006. Sample size determination for logistic regression revisited. Stat. Med. 26, 3385–3397.
Jo
Durmic, Z., Pethick, D.W., Mullan, B.P., Accioly, J.M., Schulze, H., Hampson, D.J., 2002. Evaluation of largeintestinal parameters associated with dietary treatments designed to reduce the occurrence of swine dysentery. Br. J. Nutr. 88, 159–169.
Elmberg, J., Berg, C., Lerner, H., Waldenström, J., Hessel, R., 2017. Potential disease transmission from wild geese and swans to livestock, poultry and humans: a review of the scientific literature from a One Health perspective. Infect. Ecol. Epidemiol. 7, 1300450. Fleming, P., Dundas, S., Lau, Y., Pluske, J., Fleming, P.A., Dundas, S.J., Lau, Y.Y.W., Pluske, J.R., 2016. Predation
21
by red foxes (Vulpes vulpes) at an outdoor piggery. Animals 6, 60. Giacomini, E., Gasparrini, S., Lazzaro, M., Scali, F., Boniotti, M.B., Corradi, A., Pasquali, P., Alborali, G.L., 2018. The role of transportation in the spread of Brachyspira hyodysenteriae in fattening farms. BMC Vet. Res. 14, 10. Hampson, D.J., La, T., Phillips, N.D., 2015. Emergence of Brachyspira species and strains: reinforcing the need for surveillance. Porc. Heal. Manag. 1, 8. Hidalgo, Á., Carvajal, A., García-Feliz, C., Osorio, J., Rubio, P., 2009. Antimicrobial susceptibility testing of Spanish field isolates of Brachyspira hyodysenteriae. Res. Vet. Sci. 87, 7–12.
ro of
Jacobson, M., Fellstrom, C., Lindberg R., Wallgren, P., Jensen-Waern, M., 2004. Experimental swine dysentery: comparison between infection models. J. Med. Microbiol. 53, 273–280.
Jansson, D.S., Johansson, K.E., Olofsson, T., Råsbäck, T., Vågsholm, I., Pettersson, B., Gunnarsson, A., Fellström, C., 2004. Brachyspira hyodysenteriae and other strongly -haemolytic and indole-positive spirochaetes
-p
isolated from mallards (Anas platyrhynchos). J. Med. Microbiol. 53, 293–300.
re
Jansson, D.S., Råsbäck, T., 2009. Experimental Challenge of Mallards (Anas platyrhynchos) with Brachyspira hyodysenteriae and “ Brachyspira suanatina” isolated from pigs and mallards. J. Comp. Pathol. 141, 211–
lP
222.
Jensen, N.S., Stanton, T.B., Swayne, D.E., 1996. Identification of the swine pathogen Serpulina hyodysenteriae in rheas (Rhea americana). Vet. Microbiol. 52, 259–69.
na
Johnson, L.A., Fernando, C., Harding, J.C.S., Hill, J.E., 2018. Characterization of Brachyspira communities from clinical cases of swine mucohaemorrhagic diarrhea through deep sequencing of the NADH oxidase (nox)
ur
gene. Vet. Microbiol. 214, 81–88.
Löbert, S., Zimmermann, W., Bürki, S., Frey, J., Nathues, H., Scheer, P., Doherr, M.G., Stalder, U., Zeeh, F., 2016.
Jo
Occurrence of Brachyspira hyodysenteriae in multiplier pig herds in Switzerland. Tierarztl. Prax. Ausgabe G Grosstiere - Nutztiere 44, 13–18.
Kümmerlen, D., Hartmann, S., Riklin, A., Figi, R., Sidler, X., 2019. Aspects of animal health, animal welfare and biosecurity during 101 transports of piglets in Switzerland. Schweiz Arch Tierheilkd 161, 153–163. Massacci, F.R., De Luca, S., Cucco, L., Tentellini, M., Perreten, V., Pezzotti, G., Magistrali, C.F., 2018. Multiresistant Brachyspira hyodysenteriae shedding by pigs during the fattening period. Vet. Rec. vetrec2018-104886.
22
Medhanie, G.A., McEwen, S.A., Slavic, D., Guerin, M.T., 2013. Brachyspira spp. and avian intestinal spirochaetosis: an epidemiological review. Worlds. Poult. Sci. J. 69, 541–552. Mirajkar, N.S., Bekele, A.Z., Chander, Y.Y., Gebhart, C.J., 2015. Molecular epidemiology of novel pathogen “Brachyspira hampsonii” reveals relationships between diverse genetic groups, regions, host species, and other pathogenic and commensal Brachyspira species. J. Clin. Microbiol. 53, 2908–18. Mirajkar, N.S., Davies, P.R., Gebhart, C.J., 2016. Antimicrobial susceptibility patterns of Brachyspira species isolated from swine herds in the United States. J. Clin. Microbiol. 54, 2109–2119. Oren, A., Garrity, G.M., 2017. List of new names and new combinations previously effectively, but not validly,
ro of
published. Int. J. Syst. Evol. Microbiol. 67, 529–531. Oren, A., Garrity, G.M., 2016. List of new names and new combinations previously effectively, but not validly, published. Int. J. Syst. Evol. Microbiol. 66, 4299–4305.
Pearson, H.E., Toribio, J.-A.L.M.L.M.L., Lapidge, S.J., Hernández-Jover, M., 2016. Evaluating the risk of pathogen
-p
transmission from wild animals to domestic pigs in Australia. Prev. Vet. Med. 123, 39–51.
re
Phillips, N.D., La, T., Adams, P.J., Harland, B.L., Fenwick, S.G., Hampson, D.J., 2009. Detection of Brachyspira hyodysenteriae, Lawsonia intracellularis and Brachyspira pilosicoli in feral pigs. Vet. Microbiol. 134, 294–
lP
9.
Pluske, J.R., Pethick, D.W., Hopwood, D.E., Hampson, D.J., 2002. Nutritional influences on some major enteric bacterial diseases of pig. Nutr. Res. Rev. 15, 333–371.
na
Råsbäck, T., Jansson, D.S., Johansson, K.-E., Fellström, C., 2007. A novel enteropathogenic, strongly haemolytic spirochaete isolated from pig and mallard, provisionally designated “Brachyspira suanatina” sp. nov.
ur
Environ. Microbiol. 9, 983–991.
Reiner, G., Winkelmann, M., Willems, H., 2011. Prevalence of Lawsonia intracellularis, Brachyspira
Jo
hyodysenteriae, and Brachyspira pilosicoli infection in hunted wild boars (Sus scrofa) in Germany. Eur. J. Wildl. Res. 57, 443–448.
Robertson, I., Mhoma, J., Hampson, D., 1992. Risk factors associated with the occurrence of swine dysentery in Western Australia: results of a postal survey. Aust. Vet. J. 69, 92–93. Rubin, J.E., Harms, N.J., Fernando, C., Soos, C., Detmer, S.E., Harding, J.C.S., Hill, J.E., 2013. Isolation and characterization of Brachyspira spp. Including “Brachyspira hampsonii” from lesser snow geese (Chen caerulescens caerulescens) in the Canadian arctic. Microb. Ecol. 66, 813–822.
23
Scherrer, S., Borgström, A., Frei, D., Wittenbrink, M.M., 2016. First screening for Brachyspira hampsonii in Swiss pigs applying a new high resolution melting assay. Vet. Microbiol. 193, 17–21. Songer, J.G., Glock, R.D., Schwartz, K.J., Harris, D.L., 1978. Isolation of Treponema hyodysenteriae from sources other than swine. J. Am. Vet. Med. Assoc. 172, 464–6. Speiser, S.A., 2008. Untersuchungen zu “Lawsonia intracellularis-, Brachyspira hyodysenteriae”- und “Brachyspira pilosicoli”-Infektionen bei Absetz- und Mastschweinen in der Schweiz. Diss. Vetsuisse Fac. Berne. Bern. Thakur, K.K., Revie, C.W., Hurnik, D., Poljak, Z., Sanchez, J., 2016. Analysis of Swine Movement in Four Canadian
ro of
Regions: Network Structure and Implications for Disease Spread. Transbound. Emerg. Dis. 63, e14–e26. Windsor, R.S., Simmons, J.R., 1981. Investigation into the spread of swine dysentery in 25 herds of East Anglia and assessment of its economic significance in five herds. Vet. Rec. 109, 482–4.
Zeeh, F., Klausmann, S., Masserey, Y., Nathues, H., Perreten, V., Rohde, J., 2018. Isolation of Brachyspira
Jo
ur
na
lP
re
-p
hyodysenteriae from a crow (Corvus corone) in close proximity to commercial pigs. Vet. J. 236, 111–112.
24
Table 1 Results of univariable logistic regression models of risk factors and protective factors for Swine Dysentery in Swiss pig herds. The first category presented is the reference category. Variable
Level
OR
CI 95%
P
-
-
-
11.2
3.0-49.1
0.001
No
-
-
-
Yes
8.4
Risk factors BH-status of source herds
Negative
Regular treatment
ro of
Positive/suspicious
2.7-29.8
< 0.001
-
-
1.7-44.8
0.017
-
-
-
4.9
1.4-17.6
0.013
-
-
-
Yes
4.4
1.5-13.3
0.007
Dry
-
-
-
Wet
3.5
1.1-13.3
0.042
No
-
-
-
Yes
3.1
1.1-9.3
0.034
-
-
-
3.1
1.1-9.3
0.044
Few (0-4)
-
(batches) per yeara
Many (>4)
6.7
Contact to foxesa
No
last 12 months
unita
ur
Rats on farma
na
Feeding system in fattener
No
lP
Diagnostics performed during
re
Yes
-p
Number of pig purchases
Jo
Herd size (fattening places) a
Small (80-250) b Big (>250)
Protective factors Different sources of grower
Noc
-
-
-
pigs within one batch
Yes
0.3
0.08-0.8
0.018
25
Raptor birds present in region
Martens present in region
No
-
-
-
Yes
0.1
0.03-0.5
0.003
No
-
-
-
Yes
0.1
0.01-0.5
0.024
OR: odds ratio CI: confidence interval
ro of
BH: Brachyspira hyodysenteriae Variable was included in the final multivariable model.
b
The minimum number of fattening places was 80 (see methods, inclusion criteria).
c
’No’ includes ‘only own pigs’.
Jo
ur
na
lP
re
-p
a
26
ro of -p re
lP
Figure 1: Geographic distribution of 20 case herds (Brachyspira hyodysenteriae positive, red asterix) and 60 control herds (green square) included in the risk-factor study. Actual number of symbols is
na
lower than 80 because of overlapping of neighbouring herds. The inserted map illustrates the distribution of pigs in Switzerland. The size of the pink circles
ur
correlates to the number of pigs. (Source: Federal Statistical Office, Switzerland. Data: 2016. Accessed: 28 September 2018 with Link:
Jo
https://www.atlas.bfs.admin.ch/maps/13/map/mapIdOnly/0_de.html. Current link: https://www.atlas.bfs.admin.ch/maps/13/de/12451_5892_5872_4801/20564.html (09 July 2019))
27