Estimating the power of a Mycobacterium bovis vaccine trial in Irish badgers

The aim of this study was to estimate the power, using simulation techniques, of a group randomized vaccine field trial designed to assess the effect of vaccination on Mycobacterium bovis transmission in badgers. The effects of sample size (recapture percentage), initial prevalence, sensitivity and specificity of the diagnostic test, transmission rate between unvaccinated badgers, Vaccine Efficacy for Susceptibility (VES) and Vaccine Efficacy for Infectiousness (VEI), on study power were determined. Sample size had a small effect on power. Study power increased with increasing transmission rate between non-vaccinated badgers. Changes in VES had a higher impact on power than changes in VEI. However, the largest effect on study power was associated with changes in the specificity of the diagnostic test, within the range of input values that were used for all other modelled parameters. Specificity values below 99.4% yielded a study power below 50% even when sensitivity was 100% and, VEI and VES were both equal to 80%. The effect of changes in sensitivity on study power was much lower. The results from our study are in line with previous studies, as study power was dependent not only on sample size but on many other variables. In this study, additional variables were studied, i.e. test sensitivity and specificity. In the current vaccine trial, power was highly dependent on the specificity of the diagnostic test. Therefore, it is critical that the diagnostic test used in the badger vaccine trial is optimized to maximize test specificity.

Prevalence and risk factor analysis of livestock associated MRSA-positive pig herds in The Netherlands

In 2005, methicillin-resistant Staphylococcus aureus was found in pig herds and in humans in contact with pigs. To determine the prevalence of, this now-called livestock-associated (LA) MRSA among pig herds in The Netherlands and to identify and quantify risk factors, an observational study of 202 pig herds was performed between 2007 and 2008. Five environmental wipes and 60 nasal swabs from each herd were collected, and microbiological analysis was performed on single environmental samples and pooled nasal samples. A herd was considered MRSA-positive if ≥1 sample tested positive. The prevalence of MRSA-positive herds was 67% in breeding herds and 71% in finishing herds. Multivariable logistic regression analysis was then performed on data from 171 breeding herds. The number of MRSA-positive herds increased from ∼30% at the start to ∼75% at the end of the study, most likely due to transmission between herds. The prevalence of MRSA increased with herd size, as ∼40% of smaller herds (<250 sows) were MRSA-positive compared to >80% of larger herds (>500 sows). Other risk factors (e.g. antimicrobial use, purchase of gilts and hygiene measures) were not significantly associated with MRSA, though associated with herd size. Herd size appeared to be a compilation of several factors, which made larger herds more often MRSA positive.

Design and analysis of small-scale transmission experiments with animals

Interactions between pathogens and hosts at the population level should be considered when studying the effectiveness of control measures for infectious diseases. The advantage of doing transmission experiments compared to field studies is that they offer a controlled environment in which the effect of a single factor can be investigated, while variation due to other factors is minimized. This paper gives an overview of the biological and mathematical aspects, bottlenecks and solutions of developing and executing transmission experiments with animals. Different methods of analysis and different experimental designs are discussed. Final size methods are often used for analysing transmission data, but have never been published in a refereed journal; therefore, they will be described in detail in this paper. We hope that this information is helpful for scientists who are considering performing transmission experiments.

Transmission of bovine leukaemia virus within dairy herds by simulation modelling

In Argentina, bovine leukaemia virus (BLV) infection is common in dairy herds. The country currently has a National Voluntary Control Programme but relatively few farms have enrolled. However, there is increased interest from authorities and farmers to implement regional compulsory programmes but there is scarce quantitative information of the transmission of BLV in cattle herds. This information is a prerequisite to develop effective BLV control strategies. Mathematical modelling offers ways of integrating population-level knowledge and epidemiological data to predict the outcomes of intervention scenarios. The purpose of the current paper is to gain understanding about the dynamics of the transmission of BLV in dairy herds from Argentina by simulation and to compare various BLV transmission models and select the one that is most appropriate. The hypothetical herd is conceptually described in terms of BLV status as a population of individuals that are protected by maternal antibodies (M), that are susceptible (S), that are in the latent period (E) or that are infectious (I). BLV is spread by horizontal and vertical transmission. We used an age-structured population model and within-herd transmission was simulated by Monte Carlo techniques. The next-generation approach has been used for the systematic computation of the basic reproduction ratio (R0). Parameter values for disease transmission were derived from previously published data; rates of entry, exit or transition between age groups were calculated based on our previous study, observational data, expert opinions and literature. With these parameter values the probability of a minor outbreak was estimated to be 10%, the probability of extinction was estimated as <0.001% and the expected time to extinction as more than 80 years. The probability of a minor outbreak and changes in prevalence were different when the index case was an adult cow compared to introduction by a heifer. Prediction of prevalences from MSI models fit the data satisfactorily. R0 was estimated as 9.5. The sensitivity analysis on R0 showed that all measures directed to reduce the transmission rate are potentially effective given operational control measures. An important prediction of these models is that, even in a relatively small, closed dairy herd, the time-scale for a BLV outbreak may be as long as several years and within-herd control of BLV requires intensive efforts.

Bayesian estimation of hepatitis E virus seroprevalence for populations with different exposure levels to swine in The Netherlands

Hepatitis E virus (HEV) is ubiquitous in pigs worldwide and may be zoonotic. Previous HEV seroprevalence estimates for groups of people working with swine were higher than for control groups. However, discordance among results of anti-HEV assays means that true seroprevalence estimates, i.e. seroprevalence due to previous exposure to HEV, depends on choice of seroassay. We tested blood samples from three subpopulations (49 swine veterinarians, 153 non-swine veterinarians and 644 randomly selected individuals from the general population) with one IgM and two IgG ELISAs, and subsets with IgG and/or IgM Western blots. A Bayesian stochastical model was used to combine results of all assays. The model accounted for imperfection of each assay by estimating sensitivity and specificity, and accounted for dependence between serological assays. As expected, discordance among assay results occurred. Applying the model yielded seroprevalence estimates of approximately 11% for swine veterinarians,approximately 6% for non-swine veterinarians and approximately 2% for the general population. By combining the results of five serological assays in a Bayesian stochastical model we confirmed that exposure to swine or their environment was associated with elevated HEV seroprevalence.

Comparing methods to quantify experimental transmission of infectious agents

Transmission of an infectious agent can be quantified from experimental data using the transient-state (TS) algorithm. The TS algorithm is based on the stochastic SIR model and provides a time-dependent probability distribution over the number of infected individuals during an epidemic, with no need for the experiment to end in final-size (e.g., where no more infections can occur). Because of numerical limitations, the application of the TS algorithm is limited to populations with only a few individuals. We investigated the error of using the easily applicable, time-independent final-size (FS) algorithm knowing that the FS situation was not reached. We conclude that the methods based on the FS algorithm: (i) underestimate R(0), (ii) are liberal when testing H(0):R(0)1 against H(1):R(0)<1, (iii) are conservative when testing H(0):R(0)1 against H(1):R(0)>1, and (iv) are conservative when testing H(0):R(control)=R(treatment) against H(1):R(control)>R(treatment). Furthermore, a new method is presented to find a difference in transmission between two treatment groups (MaxDiff test). The MaxDiff test is compared to tests based on FS and TS algorithms. The TS test and the MaxDiff test were most powerful (approximately equally powerful) in finding a difference, whereas the FS test was less powerful (especially, when both R(control) and R(treatment) are >1).

Classification of temporal profiles of F4+ E. coli shedding and faecal dry matter in experimental post-weaning diarrhoea of pigs

Enterotoxigenic F4+ Escherichia coli can colonize the intestine of pigs and cause diarrhoea. Our primary goal was to find a discriminant rule to discriminate between F4+ E. coli shedding profiles as this may reflect differences in the infectiousness of pigs. Our secondary goal was to find a discriminant rule to discriminate between diarrhoeic and non-diarrhoeic pigs. Repeated measurements (bacterial shedding and percentage dry matter of faeces) were taken of 74 weaned pigs that were infected experimentally with F4+ E. coli. These measurements were summarized into two new variables by means of a principal components analysis. Discriminant rules were derived based on these summary variables by fitting a mixture of normal distributions. Finally, the association between the classifications (as derived from the discriminant rules) and the occurrence in the pigs of the F4 receptor, an adhesion site for F4+ E. coli, was studied. We found that only the classification based on bacterial shedding allowed us to distinguish two significantly different groups of pigs (high and low shedders). Presence of the F4 receptor was associated strongly with pigs being high shedders.

Evaluation of natural transmission of bovine leukaemia virus within dairy herds of Argentina

The purpose of this study was to describe patterns of seroconversion to bovine leukaemia virus and to estimate the main parameters needed for future model building. A longitudinal study was carried out between February 1999 and November 2001 in seven commercial dairy farms in Argentina using 1535 lactating cows. Time-interval parameters were analysed using a parametric survival model with shared frailty, time until infection was analysed using a Bayesian interval-censoring survival model and the infection transmission parameter (beta) was estimated by a generalized linear model. The reproduction ratio (R0) was calculated. In total, 1000 cows tested positive and 494 tested negative. The predicted median age at infection was 4.6 years for seroconverted cows. For infected herds, the proportion of positive calves was as high as for infected cows and showed a large proportion of infected breeding heifers. Peaks in the overall average incidence per season-year were observed during autumn and spring. Results reveal that the period around parturition is a high-risk period. Moreover, heavily infected herds seem to have an increased proportion of young stock infected. The overall beta was estimated as 2.9/year (95% CI 1.9-3.7) and combined with a relatively long infectious period it resulted in a high reproductive ratio (R0=8.9). Therefore, a high effectiveness of control measures needs to be achieved to eradicate the disease.

Escherichia coli O157 prevalence in Dutch poultry, pig finishing and veal herds and risk factors in Dutch veal herds

In the period October 1996 through December 2000, a total of 7163 pooled faecal samples of laying hen and broiler flocks, finishing-pig herds and veal herds were examined for the presence of Salmonella spp., Campylobacter spp. and verocytotoxin-producing Escherichia coli O157 as part of a national monitoring programme in The Netherlands. Isolates were tested for eae and VT genes. Risk factors for Dutch veal herds were quantified. For all herd/flock types, faecal samples were cultured for E. coli O157. Of broiler flocks, laying flocks and finishing pig herds, respectively, 1.7%, 0.5% and 0.4% were E. coli O157 positive. In total, 42 of the 454 veal herds (9.3%) showed at least one positive pooled sample. E. coli O157-positive herds were compared (with logistic regression) to negative herds, regarding variables obtained from the questionnaire taken from the farm manager. To account for season, a sine function was included in the logistic regression as offset variable. In the final model, 'pink-veal production' (compared to white-veal production), 'group housing of the sampled herd' (compared to individual housing), 'more than one stable present' (compared to one stable present), 'hygienic measures regarding visitors' (compared to no hygienic measures), 'interval arrival-sampling of a herd of >20 weeks' (compared to < or =10 weeks), and 'presence of other farms within 1 km' (compared to no presence of farms <1 km) showed associations (P<0.05) with the presence of E. coli O157. These results need careful interpretation; they should be considered as indications for further (experimental or cohort-based) research rather then causal associations.

Transmission of F4+ E. coli in groups of early weaned piglets

The aim of this study was to estimate transmission parameters of enterotoxigenic F4+ Escherichia coli F4 (F4+ E. coli) in groups of early weaned piglets with F4-receptor-positive (F4R+) and F4-receptor-negative piglets (F4R-). Transmission of F4+ E. coli was quantified in four heterogeneous groups of F4R+ and F4R- piglets. Infectiousness was determined by the number of F4+ E. coli/g faeces shed during 8 days. Transmission parameters were estimated using generalized linear models assuming a stochastic SIR model. F4R+ piglets were found to be more susceptible than F4R- piglets, but F4R+ and F4R- piglets were not different in infectiousness. The reproduction ratios for homogeneous F4R+ and F4R- populations were estimated as 6.37 (95% CI 1.89-21.48) and 0.02 (95% CI 0.00-1.13) respectively. The implication of these results is that in order to prevent major outbreaks, the fraction of F4R+ piglets should be small (approximately 10% or less). Therefore, selective breeding programmes could contribute to reducing F4+ E. coli-related diarrhoea and transmission.

Estimating transmission parameters of F4+ E. coli for F4-receptor-positive and -negative piglets: one-to-one transmission experiment

F4+ Escherichia coli is an important agent of post-weaning diarrhoea in piglets. Piglets that express an adhesion site for F4+ E. coli in their small intestine (F4R+) shed higher numbers of F4+ E. coli than piglets lacking this site (F4R-). We hypothesized that F4R+ piglets are more infectious and more susceptible for F4+ E. coli. This implies that in populations with F4R+ and F4R- piglets, the transmission would be dependent on the frequency of both types of animals. To quantify the difference in infectiousness and susceptibility, a one-to-one transmission experiment was performed with 20 pairs consisting of one inoculated and one contact piglet. Based on the contact infections observed, transmission parameters were estimated with generalized linear models. F4R+ piglets were infectious for other piglets and the reproduction ratio (R0) for homogeneous F4R+ populations, that is the average number of secondary infections that one F4R+ pig will cause during its entire infectious period in a population of susceptible F4R+ individuals only, was estimated as 7.1. F4R+ piglets were more susceptible than F4R- piglets and reducing the fraction of F4R+ piglets of a population will reduce transmission. It was calculated that in order to prevent major outbreaks of F4+ E. coli (R0 < 1), the fraction of F4R+ piglets must be lower than 0.14.

Phenotypic selection for residual feed intake and its effect on humoral immune responses in growing layer hens

According to the resource allocation theory, animals have to make a trade-off between resource-demanding life traits to obtain maximal fitness. Artificial selection toward efficient producing farm animals, however, may have created animals that have an impaired ability to divert resources to maintenance processes, such as responding to immune challenges. Residual feed intake (RFI), defined as the difference between observed feed intake (FI) and expected feed intake based on metabolic BW and growth, was used as a measure for feed efficiency. Individual BW and FI of 352 pullets were recorded weekly from 4 until 14 wk of age to estimate RFI. The top 50 efficient R- and the top 50 nonefficient R+ birds were selected. BW and BW gain in both groups were similar. FI and RFI, however, were significantly higher in R+ birds. Thirty animals out of every group were randomly allocated to 1 of 3 treatments: immunization with keyhole limpet hemocyanin (KLH), Mycobacterium butyricum, or heat-inactivated Salmonella enteritidis bacteria. Antibody titers against KLH, M. butyricum, or Salmonella lipopolysaccharide did not differ between R+ and R- birds. The antibody titer against Salmonella protein was higher in R+ birds. We concluded that a population of chickens from a commercial breed shows considerable variation in RFI. Specific antibody production against KLH, M. butyricum, and S. enteritidis lipopolysaccharide, however, is not influenced by efficiency in terms of RFI. R+ animals may have a higher level of nonantigen specific antibodies, as indicated by the higher antibody response to Salmonella protein.

Prevalence estimation and risk factors for Escherichia coli O157 on Dutch dairy farms

To estimate the prevalence of Escherichia coli O157 on Dutch dairy herds, faecal samples were collected once from 678 randomly selected dairy farms in the period October 1996-December 2000. Samples were cultured for E. coli O157. Thirty-eight isolates were tested for virulence genes (eae, VT1 and VT2). A questionnaire about farm characteristics was taken from the farm manager, resulting in variables that could be analysed to identify and quantify factors associated with presence of E. coli O157. In total, 49 of the 678 herds (7.2%) showed at least one positive pooled sample. E. coli O157 was not isolated from herds sampled in December-April in consecutive years (except for one isolate found in March, 2000). VT- and eae-genes were found in 37 and 38 isolates, respectively. Logistic regression was performed on variables obtained from the questionnaire, comparing E. coli O157-positive herds to negative herds. To account for season, a sine function was included in the logistic regression as an offset variable. In the final model, the presence of at least one pig at the farm (OR = 3.4), purchase of animals within the last 2 years before sampling (OR = 1.9), supply of maize (OR = 0.29) to the cows, and sampling a herd in the year 1999 or 2000 (compared to sampling in 1998; OR = 2.1 and 2.9, respectively) had associations with the presence of E. coli O157.

Design and analysis of an Actinobacillus pleuropneumoniae transmission experiment

This paper describes a methodology to quantify the transmission of Actinobacillus (A.) pleuropneumoniae from subclinically infected carrier pigs to susceptible contact pigs, and to test the effect of possible interventions on the transmission. The methodology includes the design of a transmission experiment, and a method with which A. pleuropneumoniae transmission can be quantified and with which the effect of an intervention on the transmission can be tested. The experimental design consists of two parts. First, subclinically infected carrier pigs are created by contact exposure of specific-pathogen-free pigs to endobronchially inoculated pigs. Second, transmission is observed from the group of carrier pigs to a second group of susceptible contact pigs after replacing the inoculated pigs by new contact pigs. The presented analytical method is a generalised linear model (GLM) with which the effect of an intervention on the susceptibility and infectivity can be tested separately, if the transmission is observed in heterogeneous populations. The concept of the experimental transmission model is illustrated by describing an A. pleuropneumoniae transmission experiment in which the effect of vaccination on the susceptibility is quantified. Although it could not be demonstrated that vaccination has an effect on the susceptibility of pigs, it was demonstrated that nasal excretion of A. pleuropneumoniae is related to the infectivity of pigs.

A review of porcine reproductive and respiratory syndrome virus in Dutch breeding herds: population dynamics and clinical relevance

Understanding the spread of porcine reproductive and respiratory syndrome virus (PRRSV) in pig populations is essential to the development of effective PRRS prevention and control strategies. Moreover, knowledge of the field dynamics of PRRSV in pigs will provide insights into the clinical relevance of PRRS, and will enable the targeting of interventions. This review of PRRSV includes discussion on the occurrence of outbreaks, the persistence of infection and the fade-out of infection in Dutch breeding herds. The dynamic character of PRRSV infections in endemically infected herds and the relevance of the disease under Dutch field conditions are also highlighted. Furthermore, several strategies aimed at controlling the spread of PRRSV are discussed.