key: cord-0038530-nb49gj3f
authors: Madec, F; Kobisch, M; Leforban, Y
title: An attempt at measuring health in nucleus and multiplier pig farms
date: 2003-10-03
journal: Livest
DOI: 10.1016/0301-6226(93)90113-v
sha: 55d5ea25045f73ffdd9bb81993030e5bec9a3449
doc_id: 38530
cord_uid: nb49gj3f

A pilot epidemiological inquiry was undertaken in France in a group of 205 nucleus and multiplier pig farms. The aim was to find out a method for a quantitative evaluation of the health level in farms selling young breeders. An exhaustive protocol was prepared for data collection. The sources of information were clinical inspections on the farms, meat inspection data at slaughter and laboratory investigations. Data processing issued in the selection of a profile made of 14 prevailing health indicators. These were then associated so as to set up a health index with an overall score. In a second phase, the relevance of the method with respect to disease transmission was assessed. The principle was a contact challenge within totally controlled facilities between SPF pigs hysterectomy-derived and gilts taken from farms with different health scores as previously checked. The contact lasted for 28 days. Eleven farms were chosen and in every one of them 7 gilts were sampled and 10 SPF pigs were assigned to each of these farms. All the pigs were submitted to a detailed observation. At the end of it the pigs were euthanized at the laboratory and checked to find out any lesions and infectious agents. A wide range of symptoms appeared among the SPF pigs. Mortality rate was 14.6%. Pneumonia affected 23.7% of them. A relationship was found between the germ transfer and the severity of the troubles. The degree of illness in SPF pigs was clearly related to the score obtained previously in the corresponding farms. Consequently, the method was considered as valid with respect to health evaluation.

To measure health standards is a delicate task for an epidemiologist whatever the species concerned. With humans, the problem of selecting accurate and reliable criteria has been underlined many times (Feinstein, 1977 , Holland et at. 1979 , Wagner et al. 1983 . For this purpose, the concept of health indicators has been widely used (Hetzel, 1972, Jenicek and Cleroux, 1982) . The approach is also available for farm animals and, therefore, health requirements have to be defined and listed (Drummond, 1989) . The question of health evaluation is of particular importance in farms selling new replacement stock. As regards pigs, the onset of breeding pyramids with a wide spread of gilts and boars from a limited number of specialized nucleus herds emphasize the need for measuring health in these herds. Descriptions of health statuses of farms have been proposed (Goodwin and Whittlestone, 1983, Muirhead, 1989) . The evaluation must not be confined to the microbial agents that can be detected. Nevertheless, these are usually granted great interest in health monitoring (Alexander, 1986 , Skovgaard, 1987 . Health evaluation methods consist in binary responses for specific health indicators taken one by one rather than in a figure giving an overall quantitative evaluation for the health profile. Furthermore, such measurements can hardly ever be judged as reliable through experimentation. This paper presents a method of scoring health in nucleus and multiplier pig farms obtained from a field study. Then, it describes an experimental trial undertaken as a validating challenge.

The method used to assess the health level A pilot study was undertaken to describe the health situation of nucleus and multiplier pig farms: 205 farms were considered. They were scattered all over France respecting the territorial pig farming distribution. Farmers and veterinarian inquirers were volunteers.

Two types of diseases were considered in the protocol: • those determined by specific microbial agents: • those resulting from carelessness in herd management and husbandry.

Two sorts of measurements were therefore undertaken:

Indicators of risk of actual disease transmission. In order to obtain information for this, wide laboratory investigations were carried out on different samples of material, with the aim of detecting infectious agents. They have been described in detail elsewhere (Madec et al. 1990 ). The material used was 61ood from sows and fatteners, feces and urine from sows, and lungs from slaughtered pigs. Furthermore, 3 ten-week-old-piglets, randomly selected on the farms, were necropsied at the laboratory to be fully examined. The notifiable diseases (swine fever, Aujeszky's disease) were kept aside at the time of data processing because all the farms were non infected. Furthermore, since there is a specific regulation for these, the present work was mainly focused on the other health disorders.

Tendency for a farm to show out signs of enzootic diseases. Visits were planned so to obtain information about the disease history, current clinical signs and production inefficiencies. The sources of data were the following:

• on-farm performance records and clinical inspections of the pigs during standardized visits to the farms • laboratory examinations • slaughter checks

A general analysis of the data was undertaken in order to rank the farms into clusters (families) according to their similarities regarding health standards. Concurrently, the most relevant parameters for this purpose were determined. The statistical process was performed with the use of descriptive multivariate methods, namely correspondence analysis and a cluster analysis (Lebart et al. 1985 , Jambu et al. 1983 . At the end of the statistical process, a combination of fourteen variables (health indicators) was retained to classify the farms according to their health status and a typology could be concluded (Madec et al. 1990 ). Some of the health indicators derive from the association of 2 or more criteria. They are listed in Table 1 . Furthermore, each of the health indicators was divided into 3 levels respecting the sense of the relationship with health. Then, they were used to establish an ultimate health score. Health indicators with a low, medium or high level were given respectively an elementary score of zero, one or two. Afterwards the elementary scores were added up so as to obtain a global score for health, ranging theorically from zero (the 14 indicators at low level) to 28 (all at the high level ).

The scoring index obtained in phase A was tested in a second phase in order to find out whether the detected difference between farms based on the index is expressed in the level of disease in SPF pigs. A contact trial was suggested between gilts taken from breeding or multiplier pig farms previously checked for health using the method described above and hysterectomy-derived SPF piglets. The design of the experimental contact trial is demonstrated in Fig. 1 . The experimentation was conducted in totally isolated rooms. Each room was equipped with 2 flat decks with perforated floors and sides partially open. • Feeding: the same pelleted food was provided ad lib. throughout the trial. • The farms: gilts from 11 farms were tested. These farms did not participate to phase A. The managers of breeding pyramids that accepted to take part in the trial were first asked to give us an initial idea of the profile of the farms candidates. All the pigs from these farms had to be individually and systematically identified at birth on routine basis (Tattoo). Provisional groups of farms were made with respect to health. Then a random selection was undertaken within these groups. The selected farms were widely scattered all over the North-West of France. Their health was checked up just before the contact trial started. • The gilts: On the farms, batches of pigs weighing around 75-85 kg live weight were considered for the choice of the individuals that were to be transferred to the research station. Pigs showing obvious signs of illness or those with a low liveweight were eliminated. Among the others, gilts estimated to be fit for sale were submitted to random selection. They were weighed, and transferred to the experimental facilities in a perfectly clean and disinfected lorry. On average, they were 141 days old (SD: 19 days) at a liveweight of 79.4 kg (SD: 4.5 kg). • The piglets: The piglets used in this trial came from the SPF herd at the "Station de Pathologie Porcine" in Ploufragan. At the beginning of the challenge, they weighed 30.3 kg, on average, at 60 days of age SD: 6 days). • Duration of the challenge: For each farm, the contact period lasted 28 days. • Statistical procedures: The data, concerning the prevalences obtained in the (Mengeling, 1992) Serology ELISA (Nicolet and Paroz, 1980) Culture (Avril, 1989) ELISA (Foged et al. 1988) Culture (Michel-Briand, 1989) Culture (Dabernet and Sanson-Le-Pors, 1989 Culture (Avril, 1989) Culture (Hommez et al. 1986) Culture (Ellis etal. 1976) Coproscopy (Coles, 1974) different groups of piglets, were compared by chi-square. The weight gains were compared by the t test. • Clinical observations and laboratory examinations: rectal temperature, fecal consistency, respiratory signs (sneezing, coughing, etc. ), locomotor disorders and any other particular anomalies were recorded day by day. Weekly food intake was recorded for each flat-deck. Blood samples were taken from every pig on the first day, and then, every 7 days. The sera were tested against pseudorabies, swine Influenza, Porcine respiratory coronavirus, Transmissible gastroenteritis, Porcine Epidemic Diarrohea, porcine parvorirus and Mycoplasma hyopneumoniae. At the beginning of the trial, nasal swabs were taken for bacteriology and faeces were examined for parasites and salmonellae. At the end of the period (28 days ), all the pigs were euthanasied and then, according to a standard protocol, necropsied and sampled for further laboratory testing (bacteriology and parasitology, Table 2 ). Respiratory tract lesions were scored at necropsy, on scales according to Madec et al. 1988 . Pigs dying in the course of the trial were weighed and a similar procedure was carded out in a thorough examination for lesions plus appropriate microbiology.

Out of the eleven farms selected and having their health standards checked, 5 were breeding farms and 6 were multipliers. Ten breeding pyramids were involved. Table 3 shows the score which these farms obtained. The average score was 15.6 (min= 10, max=22).

The overall results concerning the SPF pigs are listed in Table 4 . Mortality rate was 14.8%. Glasser's disease and streptococcal septicemia due to were the main causes of death. Clinical signs of coughing and sneezing were recorded in respectively 25.2 and 47.7% of the piglets. Fever (rectal temperature >_,40.5°C) was detected in 53.3%, and diarrhoea affected 22.4% of the piglets. Daily liveweight gain varied a lot: 13 piglets lost weight while alive; 10 gained at least 900 g daily.

Overall results from the SPF piglets: clinical signs, lesions and microbiology at necropsy after contact with gilts from nucleus and multiplier farms (n = 107 piglets) Necropsy showed pneumonia in 2 3.7% of the piglets and turbinate atrophy in 12.2%, the latter being mild. Many specific pathogens were transferred from the gilts to the corresponding piglets. These were mainly pneumotropic agents, but, neither swine influenza nor porcine coronavirus were concerned. But surprisingly a seroconcersion against porcine parvovirus was found out in a group of piglets. All the fecal samples proved negative as regards salmonellae.

When the gilts and the piglets were put together on the same flat-deck, the challenge was more severe ( Table 5 ). Signs of illness attested by rectal temperature (Fig. 2) were also detected earlier in this case. Furthermone the piglets had a significantly lower weight gain. The mortality rate was higher and fever more largely prevalent. The general trend was also for more severe lesions together with a higher frequency of isolation of P. multocida. On the other hand, the figures obtained for the seroconversion against M. hyopneumoniae were similar. In both cases rectal temperature became normal after 20 days.

The farms fell into 3 groups according to the figures obtained for the score on health. The response of the piglets is listed in Table 6 . There is a clear difference between the extreme groups ( ~< 12 vs. >1 18 ) for all the criteria. When the score reached 18 or more, prevalence of illness indicated in the table by rectal temperature was low and none of the piglets died. The pathogen transfer rate was lower. Nevertheless, P. multocida was isolated from 4 4 5 6 7 8 9 10 11 12 15 14 15 16 17 18 19 20 21 22 23 24 25 26 27 (2) days X pigs, % piglets ( 10% in group >1 18 ). However, in the piglets matched to the 3 farms getting a poor score, the microbial transfer rate was high and some agents like H. parasuis and M. hyopneumoniae were found in at least half of the piglets.

The response of the piglets concerned by the medium scores was more irregular but considering the whole sample of farms, as regards many criteria, a gradual trend could be detected between the severity of the response of the piglets and the score on health.

Transmission of disease to a recipient farm through the replacement breeding stock accounts for the transfer of specific pathogens. However, diseases cannot develop unless these pathogens find certain specific conditions in the recipent farm. The immunity status of the herd is believed to play an important role in determining this. At first sight the general process looks simple in the case of highly infectious viral diseases like TGE or swine influenza. In a non-immune herd, viral shedding from newly-purchased pigs will undoubtedly lead to disease. These diseases usually show a marked wave pattern according to the level of protection of the swine population.

Unfortunately, the situation is much more complex with respect to other health disorders like enzootic respiratory diseases or streptococcosis. In such cases, microbial transfer is also necessary for a disease to break out but in most situations there is the need for many other additional conditions before disease does actually occur. This kind of multifactorial disease has a major economic impact on the pig industry. The problem is that most of the pathogens involved in such diseases are widely spread in the breeding pyramids and it is assumed that the diseases can be transmitted by pigs from herds harbouring the pathogens even though subclinical troubles only are present in the selling farms (BackstrSm and Hoefling, 1986, De Jong and Nielsen, 1990 ). On the other hand, it has been demonstrated earlier that, at least for the respiratory tract, the frequency of germ carriage is related to the severity of the lesions (Morrison et al. 1985 , Cowart et al. 1989 .

The purpose of this work was to test the value of a method used to estimate quantitatively the health standards on farms in order to measure the level of "microbe pressure" as evidenced by disease transmission. The idea put forward being that disease is all the more likely to be transmitted as the amount of microbes transferred is high. According to the authors, the health level can be assessed through the measurement of direct health indicators (mortality rate, clinical signs, etc. ) but also indirectly through the evaluation of the degree of exposure to particular circumstances that have been proved to be linked to a given disease (Jenicek and Cleroux, 1982 ) . Most of the parameters here included in the index are direct health indicators. To elaborate a parameter in order to find out how vulnerable the farms are to viral epidemics is also -though indirectly -a health indicator in itself and this is in accordance with the above definition. Collectively, the different criteria selected in the index have a descriptive value for the whole farm health situation but they do not have an individual direct causative value with respect to a specific disease (Evans, 1978) or to disease transmission per se by the concerned pigs since for example none of the criteria gives a measure of viral shedding. Thus the health indicators are meant to describe the trend and/or to find out the factors which make that a farm is more or less liable to show such or such diseases. The relative economic impact of the different diseases of swine was not directly taken into account when selecting the health indicators. Since the target farms were farms selling young replacement stock, the objective was not to assess the relation between health and financial income. Such attempts have been undertaken elsewhere (Ellis and James, 1979 , Dijkhuisen, 1989 , Madec et al. 1992 . Nevertheless, it can be noticed in the list of the health indicators that respiratory and reproductive problems, which are very detrimental in financial terms, hold a place of choice in the profile.

The experiment reported in this paper which consists of using young gilts for a standard challenge was carried out in order to assess infection pressure more objectively on their original farm. The results obtained showed great discrepancies between the groups of piglets according to the farms the gilts came from. When the farms offered top health standards, the response of the piglets was either nil or mild regarding both clinical signs and lesions or microbial transfer. This might indicate a low infection pressure. Whereas, when most of the health indicators were at a bad level, the response of the piglets was severe. Nevertheless a new issue might be raised: were the gilts selected truly representative of the real health status of the farms. It has not been possible to solve this problem completely. However, the retrospective comparative observation of the respiratory tract of the gilts at the end of the trial and the viscera of the slaughtered pigs from the farms concerned gave substantial results (Madec et al. 1991 ) .

Besides, one might question a 28-day challenge, why 28? This might not be long enough for certain diseases to have time to develop properly. However, the curves indicating the daily prevalences of the clinical signs showed a slight decrease from day 12 post-contact onwards, and in most of the groups of pigs, the signs of disease vanished during the last days of the trial. Since the signs of illness (and therefore contamination) occurred within the first 12 days, at least 16 days remained for antibody detection.

The rapidity and the severity of the response were higher in the case of close contact between the piglets and the gilts than when the pigs had limited contact with one another. This is not without practical consequences in commercial farms where young high health-status breeders are bought-in. A progressive contact with the local microflora is recommended whenever the two herds show great discrepancies regarding their own health standards.

Furthermore the introduction of high health-status animals into a given commercial farm with a poor health level could also argued: will this not create a booster effect on disorders in the whole herd through an acute response in the replacement pigs? Our experimental model could not confirm this hypothesis since no significant recrudescence in the clinical signs could be observed in the gilts even when their own piglets were severely affected.

The results of the present study show that it is possible to estimate the health level of breeding farms through a rational epidemiological approach. To do so, it is necessary to combine as many different health indicators as possible and these have to provide data as hard as possible (Feinstein, 1977 ) to avoid subjective deviation by the inquirer (Heard, 1981 ) . The method described above requires certain investigations which have practical constraints, for example laboratory investigations and the sacrifice of 3 ten-week-old-piglets per herd. New diagnosing methods might also appear for current diseases and new syndromes might occur. Therefore, actualization and further research are made necessary in order to find out new, refined or more accurate criteria just as relevant but more practical to use than some of the health indicators mentioned in this study.

Methods of disease Control

Actinobacillus, Pasteurella

Transmission of atrophic rhinitis in swine under different environmental conditions

Enzyme-linked Immunosorbent Assay for the detection of the coronavirus-like agent and its antibodies in pigs with porcine Epidemic diarrhea

Pasteurella rnultocida and Bordetella bronchiseptica in atrophic rhinitis and pneumonia in swine

Haemophilus

Definition of progressive atrophic rhinitis

Economic aspects of common health and fertility problems for the individual pig producer: an overview

Health requirements and problems. Pig

Culture methods for the detection of animal salmonellosis and Arizonosis

The economics of animal health II. Economics in farms practice

Causation and disease: a chronological journey

Hard science, soft data and the challenges of choosing clinical variables in research

Differenciation of toxigenic from non toxigenic isolates of P. multocida by Elisa

Monitoring for atrophic rhinitis: five years experience with a pilot control scheme

Methods of approach to the diagnosis and resolution of pig health problems

The implications of health indicators: a comment

Measurement of level of health. W.H.O. regional publications. Europeans series

Identification and characterization of Streptococcus suis

Data analysis and cluster analysis

Epidemiologie: principes, techniques, applications

Multidimensional descriptive data analysis

Economie et sant6 en production porcine

Etude exp6rimentale de la notion de pression de contamination dans les 61evages porcins commercialisant des reproducteurs

616ments de situation sanitaire des 61evages porcins de la r6gion de Bretagne. 1 -Bilan 16sionnel de l'appareil respiratoire des porcs h l'engrais

Evaluation quantitative du niveau sanitaire des 61evages porcins de s61ection et de multiplication: les bilans sanitaires approfondis

Porcine Parvovirus: Properties and prevalence of a strain isolated in the United States

Bordetella. In: L. Le Minor and M. Veron Medecine Sciences

Microorganisms associated with pneumonia in slaughtered weight swine

The high health status herd

Tween 20 soluble proteins ofM. hyopneumoniae as antigen for an Enzyme Linked Immunosorbent Assay

Manual of Recommended Diagnostic Techniques and Requirements for Biological Products 1989

Manual of Recommended Diagnostic Techniques and Requirements for Biological Products

Advanced laboratory techniques for influenza diagnosis. U.S. Department of Health, Education and Welfare. Immunology series

Prevention of micorbial contamination in the antemortem phase: the SPF (specific Pathogen-free) concept

Statistical validation of a severity of illness measure

Un essai a estim6e le santee dans 6levages de s61ection et de multiplication

Au cours d'une seconde phase, la pertinence de la m6thode en regard du risque de transmission de maladies par les jeunes reproducteurs dans un 61evage d'accueil a 6t6 test6e dans des conditions exp6rimentales standardis6es et totalement contr616es. Le principe est un "contact" entre des porcs SPF et des cochettes pr61ev6es dans des 61evages ayant obtenu des scores diff6rents pour leur 6tat de sant6. Le contact dure 28 jours. Onze 61evages ont 6t6 choisis

La signification de l'outil 6pid6miologique pour estimer la pression d'infection dans les 61ev-ages est discut6e

Ein Versuch zur Messung der Gesundheit in Nukleus-und Vermehrungsbetrieben f'tir Schweine

Als Informationsquellen dienten klinische Untersuchungen der Best~inde, Fleischbeschaudaten am Schlachthof und Labortest. Fiir 14 wichtigere Gesundheitsindikatoren wurden aus den Daten Profile berechnet. Diese wurden dann zu einem Gesundheitsindex als "overall score" zusammengef'tihrt. In der zweiten Phase wurde die Bedeutung der Methode f'tir die Krankheitsiibertragung bewertet. Das Prinzip war eine Kontaktinfektion der Jungsauen yon Earmen mit verschiedenen Gesundheitsindices in einer total kontrollierten Anlage mit SPF-Prim~ir-Schweinen. Der Kontakt wurde f'tir 28 Tage gegeben. Aus 11 ausgew~lalten Betrieben wurden je 7 Jungsauen genommen und l 0 SPF-Schweine wurden jeder dieser Gruppen zugeordnet. Alle Schweine wurden genau beobachtet, am Ende gesch~itet und im Labor auf L~isionen und Infektionsagencien untersucht. Bei den SPF-Schweinen wurde eine breite Palette von Symptomen registriert. Die Mortalit~itsrate war 14.6% und 23.7% waren mit Pneumonie infiziert. Zwischen Keimtransfer und Schwere der Erkrankungen wurde eine Beziehung gefunden. Der Grad der Erkrankung der SPF-Tiere war deutlich mit dem vorher ermittelten

The authors thank Prof. J.R. Buddle from Murdoch University (Australia) for reviewing the manuscript. They also thank R. Cariolet, F. Paboeuf and J.F. Pansart for their excellent technical assistance.