Estimating the prevalence of BSE in dairy birth cohorts and predicting the incidence of BSE cases in Japan

Following the detection of the first case of BSE in Japan in September 2001, four million cattle were subjected to a rapid test for BSE up to the end of 2004. A further 10 cases were detected in the dairy cattle population and two cases in Holstein steers. We focused on the dairy population and estimated the prevalence of BSE infected animals within each birth cohort for the years 1992–2001 using Bayesian inference. From this we were able to predict historic and future trends in the number of infected animals culled from each cohort and whether or not they could be detected using a rapid test. Assuming that BSE infectivity entered Japan in 1995, 225 (95%CI: 111–418) infected animals were predicted to have been culled from 1995 to 2001, of which 116 (56–219) would have been slaughtered for human consumption, and 33 (12–65) cases would have been detected during this period if a BSE surveillance program as comprehensive as the one in place as of April 2004 was applied. Assuming that BSE infectivity entered Japan in 1992, 905 (366–4633) infected animals were predicted to have been culled from 1992 to 2001, of which 694 (190–2473) would have been slaughtered for human consumption, and 201 (53–693) cases would have been detected during this period. Assuming the April 2004 level of surveillance continues and that the feed ban introduced in 2001 is completely effective, 18 (3–111) BSE cases are likely to be detected in the future. The BSE epidemic in Japan most likely reached a peak between 1998 and 2001 and should be eradicated around 2012.     

Adjusted incidence risks of BSE in risk subpopulations of cattle in Japan

Immediately after the detection of the first case of BSE in 2001, the Japanese government introduced active surveillance targeting fallen-stock and all cattle slaughtered for human consumption. By the end of 2004, four million animals were tested with rapid tests under the passive and active surveillance. As a result 13 additional cases were detected. I focused on the 1996 birth cohort, in which nine cases of BSE were detected during 2001–2004, and estimated the prevalence of BSE infection of that birth cohort using maximum-likelihood methods. Using the estimated prevalence of infection as an input variable, I calculated the adjusted incidence risk of BSE by different ages and risk subpopulations (clinical-suspects, fallen-stock, sick-slaughter and healthy-slaughter animals). The adjusted incidence risk of BSE in sick-slaughter animals (animals showing clinical signs not compatible with BSE when slaughtered for human consumption) was 18.7 and 4.5–78.4 times higher than the incidence risk in fallen-stock and healthy-slaughter animals, respectively.     

Simulation-based estimation of BSE infection in Japan

The first case of bovine spongiform encephalopathy (BSE) in Japan was found in September 2001. As a result, national BSE surveillance systems in slaughterhouses and farms were introduced between October 2001 and April 2004. All cattle, with the exception of those under 24 months of age that die at farms, now undergo compulsory testing when they die or are slaughtered. The removal of specified risk material (SRM) from all slaughtered cattle and a ban on the feeding of meat-and-bone meal to all farm animals were implemented in October 2001. However, infected cattle that died or were slaughtered before these measures were put into practice could have been a source of infection to other cattle through the rendering process. The slaughtered cattle could also have been a source of infection to humans via SRM that entered the food chain. The purpose of this study was to estimate the number of BSE-infected cattle that could have been a source of infection to cattle and humans before October 2001. Since all typical cases were dairy cattle, this study focused on the dairy cattle population. We developed a simulation model to obtain the year of death and the final disposition of infected cows born in each year from 1996 to 2001. In this model, the dairy cattle population was divided into birth cohorts, and parameters regarding its population dynamics were assumed to be constant. Using this model, the total number of infected cattle in each birth year was estimated by maximum likelihood estimation using data on the number of detected cases from 2002 to 2006. Finally, the number of infected cattle that died or were slaughtered each year was estimated by Monte-Carlo simulation using the same model with the total number of infected cattle estimated by maximum likelihood estimation. It was estimated that the majority of infected cattle that could have been sources of infection before 2001 were born in 1996. The total number born in 1996 was estimated to be 155 (95% confidence interval: 90-275). Of these 155 cattle, 56 died or were slaughtered before October 2001, after the accumulation of infectious agent in their bodies. Only 5 of these 56 cattle were estimated to have been slaughtered. Therefore, the number of infected cattle that could have served as a source of human infection would appear to have been a very limited subset of the BSE-infected cattle in Japan.     

Modelling the expected numbers of preclinical and clinical cases of bovine spongiform encephalopathy in Switzerland.

The objective of this study was to model the expected numbers of cattle incubating bovine spongiform encephalopathy (BSE) and the numbers of clinical cases of BSE in the Swiss cattle population between 1984 and 2005. The results were compared with the observed number of clinical BSE cases and with the results of a culling and testing scheme on herdmates of cattle with BSE. The age distribution of the Swiss cattle population, the age-at-death distribution of the first 235 BSE cases and exposure information were used to calculate the expected number of infected cattle in each birth cohort and the resulting numbers of clinical cases and survivors incubating the disease for each year. The model which did not assume any under-reporting of cases fitted the observed epidemic curve of clinical cases reasonably well, and predicted that the Swiss BSE epidemic would come to an end between 2003 and 2005. The age of survivors incubating BSE is increasing. The higher than expected incidence of subclinical cases observed in animals from the culling scheme is most probably the result of the heterogeneous distribution of infected animals and affected herds in the population. The results of the model need to be taken into account when designing surveillance and testing schemes for BSE.     

Imported and indigenous BSE cases in Germany

Bovine spongiform encephalopathy (BSE) is a transmissible fatal neurodegenerative disease in cattle with an average incubation time of five years. The first BSE case in an indigenous cow was detected in Germany in November 2000. This was almost eight years after the huge BSE epidemic in the United Kingdom had peaked, and several years after many EU member states had seen their first BSE case. In the 1990s, BSE had been diagnosed in six imported animals in Germany. However, after the implementation of an active surveillance programme using BSE rapid testing systems, 399 indigenous German BSE cases have been found up to the end of July 2006. The birth cohorts of 1995-1997 contribute to the vast majority of the first 250 German cases that were diagnosed between 2000 and 2003. However, the most recent German BSE cases belong primarily to the birth cohorts 1998-2000 which is indicative of a recycling of BSE infectivity at that time. Moreover, there were two BSE cases in cattle born in spring 2001, i.e. after the meat and bone meal feed ban had come into effect on 2nd December 2000. In this article, we describe the dynamics of the German BSE epidemic and compare these data with those of other countries that observed larger numbers of cases.     

A spatio-temporal analysis of BSE cases born before and after the reinforced feed ban in France

A spatio-temporal analysis was carried out to see how the risk distribution of bovine spongiform encephalopathy (BSE) in France changed depending on the period of birth. The data concerned the 539 BSE cases born in France after the ban (BAB) of meat and bone meal (MBM) in 1990 and detected between July 1, 2001 and December 31, 2003, when the surveillance of BSE was comprehensive. Seventy-two of these cases were born after the reinforced (second) ban (BASB) in 1996, which involved the removal of BSE-risk materials and cadavers from the processing of MBM. The Ederer-Myers-Mantel (EMM) time and space cluster test was applied, after classifying the cases by trimester and region of birth, BAB or BASB status, and dairy or beef status. Then disease mapping was performed for four successive birth periods, three for the BAB cases (January 1991 through June 1994, July 1994 through June 1995, July 1995 through June 1996), and one for the BASB (July 1996 through October 1998). It was elaborated with the Bayesian graphical modelling methods and based on a Poisson distribution with spatial smoothing. The parameters were estimated by a Markov Chain Monte Carlo (MCMC) simulation method. The main finding was that the areas with the highest risk of BSE changed largely from one birth period to another; from the west, it reached the east of France for birth cohort 1994-1995 and the southwest for birth cohort 1995-1996. The EMM test identified a peak risk in this region both for dairy and beef cattle in the fall 1995. The spatial distribution of the risk for the BASB cases matched the spatial pattern of risk for the preceding BAB birth cohort quite well; this was in favour of a common origin of the infection of the BAB and BASB cases, despite the complementary control measures.     

Estimating the trend of the French BSE epidemic over six birth cohorts through the analysis of abattoir screening in 2001 and 2002

A bovine spongiform encephalopathy (BSE) testing programme at the abattoir started in 2001 in France. A total of 5 281 293 bovines were tested in 2001 and 2002; 87 were found positive in 2001--37 per million (95% CI 30-46)--, whereas only 71 in 2002--24 per million (95% CI 19-30). Logistic regression models were run to compare the prevalence of BSE on successive birth cohorts, using a pair-wise method of controlling for age at testing; the prevalence on the first one, determined on animals slaughtered in 2001, was compared to the prevalence on the following one determined on animals slaughtered in 2002. Five models were performed in order to compare the birth cohorts preceding and following the months of June 1993 (i.e. July 92-June 93 birth cohort compared to July 93-June 94 birth cohort) (8.5 years old cattle), June 1994 (7.5 years old cattle), June 1995 (6.5 years old cattle), June 1996 (5.5 years old cattle) and June 1997 (4.5 years old cattle). The models were adjusted for the production type of cattle and the test used. The results showed a significant increase (OR = 2.31, 95% CI 1.08-4.9) of the BSE prevalence between the July 93-June 94 and July 94-June 95 cohorts, and then a significant decrease over the next two birth cohorts; the July 95-June 96 birth cohort was significantly less affected than the July 94-June 95 one (OR = 0.46, 95% CI 0.27-0.78), and the July 96-June 97 birth cohort was significantly less affected than the July 95-June 96 one (OR = 0.17, 95% CI 0.07-0.37). The increase in BSE prevalence between the July 93-June 94 and July 94-June 95 cohorts was in agreement with modelling studies, but needs to be confronted to the data on fallen stock at the national level. The decrease in BSE prevalence on the birth cohorts born after June 1995 was in agreement with the findings on the fallen stock in the western part of France and matches the implementation of the removal of specified risk materials (SRM) and dead animals from the processing of meat and bone meal (MBM) since June 1996.     

Prevalence of BSE in cattle found dead, euthanased or emergency slaughtered on farms in western France in 2000, 2001 and 2002

The overall trend and the trend within birth cohorts of the prevalence of bovine spongiform encephalopathy (BSE) in cattle found dead, euthanased or emergency slaughtered on farms in the Bretagne, Basse Normandie and Pays de la Loire regions of France, during the periods from August 7 to December 22 in 2000, 2001 and 2002, were analysed by non-conditional logistic regression, adjusted for the region and for the type of animals. The overall prevalence of BSE during these three periods decreased from 2.71 per 1000 in 2000 to 1.41 per 1000 in 2001 and 0.42 per 1000 in 2002. The prevalence within birth cohorts started to decrease for the cohort born between July 1, 1995 and June 30, 1996 (cohort 95/96) and the trend was reinforced for cohort 96/97, suggesting that the exposure of animals to the BSE agent had started to decrease for animals born after July 1995, that is, one year before the ban on specified risk materials in meat and bone meal was implemented in France. However, considering that most of the animals would have been infected at between six and 18 months of age, the decrease may have been due, at least partly, to this control measure.     

The unrecognised French BSE epidemic

In France, implementation of systematic screening programs in 2000, as a complement to the mandatory reporting of animals with clinical signs of BSE (passive surveillance), revealed certain limitations of the mandatory system. Indeed, systematic screening showed that some BSE cases were not detected by the clinical surveillance system, implying considerable BSE case under-reporting throughout the epidemic. As the most likely explanation for variant Creutzfeldt-Jakob disease (vCJD) is exposure to the aetiologic agent of BSE, it is essential to reconstruct the French BSE epidemic pattern accounting for this under-reporting. We estimated age- and year-specific incidence rates of BSE by using a back-calculation method. This approach relies on the principle that the number of clinical BSE cases is the consequence of the number of BSE-infected animals after a known incubation time, defined as the time between infection and clinical onset. We generalized this model to take into account epidemiological characteristics of BSE, such as French cattle mortality, BSE case reporting probability, and age-dependent susceptibility and/or exposure to the BSE agent. We confirmed that the average BSE incubation period is five years and that the peak risk of bovine infection occurs between 6 and 12 months of age. The results also showed that the proportion of underreporting is the most influential parameter in the model, and that BSE was substantially underreported until rapid tests were introduced. Indeed, only 103 BSE cases were detected by passive surveillance up to June 2000, while we estimated that there was 301 200 (95% confidence interval (CI) [27 600-837 600]) cattle infected by the BSE agent. Despite uncertainty over the beginning of the epidemic, we showed that the French BSE epidemic in the late 1980s was completely undetected, and only the second wave, after 1990, was observed.     

Direct Costs of Bovine Spongiform Encephalopathy Control Measures in Germany

On 26 November 2000, the first autochthonous case of bovine spongiform encephalopathy (BSE) was detected in Germany. Since then, a total of 413 BSE cases have been confirmed, resulting in the culling and destruction of 17 313 heads of cattle. In view of the possible risks for human and animal health, Germany has adopted EU regulations along with some additional requirements concerning active surveillance and response measures after detecting a BSE‐positive animal. In this study, we used a stochastic model to estimate the costs incurred by the ensuing legislative amendments responding to BSE between November 2000 and December 2010. The total costs were estimated to range between 1847 and 2094 million Euros. They peaked in 2001 (about 394 million Euros) and declined since. About 54% of the costs (approximately 1000 million Euros) were incurred by the extension of the feed ban for animal protein to all farmed livestock. Active surveillance accounted for 21% (405 million Euros), the incineration of animal protein for 13% (249 million Euros) and the removal of specified risk material for 11% (225 million Euros). Only 1% of the costs was related to response measures after detecting a BSE‐positive animal, including indemnity payments for culled cattle and confiscated carcasses at the slaughterhouse.     

A model (BSurvE) for estimating the prevalence of bovine spongiform encephalopathy in a national herd

We developed the BSurvE spreadsheet model to estimate the true prevalence of bovine spongiform encephalopathy (BSE) in a national cattle population, and evaluate national BSE surveillance programs. BSurvE uses BSE surveillance data and demographic information about the national cattle population. The proportion of each cohort infected with BSE is found by equating the observed number of infected animals with the number expected, following a series of probability calculations and assuming a binomial distribution for the number of infected animals detected in each surveillance stream. BSurvE has been used in a series of international workshops, where analysis of national datasets demonstrated patterns of cohort infection that were consistent with infection-control activities within the country. The results also reflected the timing of known events that were high-risk for introduction of the infectious agent.     

BSE/TSE--the present epidemiologic and diagnostic situation

In 1992, 1994 and 1997, first BSE cases were diagnosed among imported cattle. The first domestic BSE case in Germany was confirmed on 26 November 2000. Altogether 192 cases (7 cases in 2000, 125 cases in 2001 and 60 cases in 2002) were reported (at the date: 7. August 2002). Comparing the BSE situation in Germany with other European countries under consideration of the surveillance schemes applied, Germany has a similar BSE incidence based on 100,000 adult cattle as France and Spain but a higher incidence than neighboring countries like The Netherlands, Austria and Denmark. At present, on average 2,3 cases per week are confirmed in Germany. With respect to the age distribution of the BSE cases, about 84% of all domestic cases confirmed originated out of the years 1995 and 1996 with a accumulation in 1996 (about 53% of all cases). Single BSE cases with the years of birth 1990, 1991 and 1993 demonstrate, that the BSE agent is circulating among the German cattle population at least since the beginning of the nineties. Between 1985 und 7 August 2002, scrapie was diagnosed in 26 sheep flocks in Germany, but alone during the year 2002 (until 7 August 2002) 11 scrapie affected sheep flocks were discovered.     

Active surveillance of BSE in cattle in The Netherlands

Since January 2, 2001 a large-scale active surveillance programme for BSE started in the Netherlands in addition to the passive surveillance programme of cattle with clinical symptoms compatible with BSE. Based on decisions of the Council of European Ministers of Agriculture, the European Union launched an active surveillance system for BSE in cattle of 30 months and older. Until April 1, more than 100,000 head of cattle were tested in this scheme, including all cattle slaughtered and a large part of the cattle that died on the farm. Four animals were found positive in the active surveillance system and one cow from the passive surveillance tested positive for BSE during the first three months.     

Ten years of BSE surveillance in Italy: neuropathological findings in clinically suspected cases

Between 2001 and 2010, 244 clinically suspected cases of bovine spongiform encephalopathy (BSE) were reported in Italy. This report summarizes the neuropathological findings in cattle displaying clinical signs consistent with a diagnosis of BSE. All animal specimens were submitted for confirmatory testing; samples testing negative underwent neuropathological examination to establish the differential diagnosis. Immunohistochemistry for scrapie prion protein (PrPSc) at the level of frontal cortex was carried out to exclude atypical BSE. Neuropathological changes were detected in 34.9% of cases; no histological lesions were found in 52.3% of subjects; 12.8% of samples were found unsuitable for analysis. BSE was detected in one case, but no cases of atypical BSE were observed. This study identified the diseases most commonly encountered in the differential diagnosis of BSE; furthermore, it demonstrated that the surveillance system is necessary for monitoring neuropathological disease in cattle and for the detection of BSE cases.     

Model calculation to explain the BSE-incidence in Germany

The future development of BSE-incidence in Germany is investigated using a simple epidemiological model calculation. Starting point is the development of the incidence of confirmed suspect BSE-cases in Great Britain since 1988, the hitherto known mechanisms of transmission and the measures taken to decrease the risk of transmission as well as the development of the BSE-incidence in Germany obtained from active post mortem laboratory testing of all cattle older then 24 months. The risk of transmission is characterized by the reproduction ratio of the disease. There is a shift in time between the risk of BSE transmission and the BSE incidence caused by the incubation time of more than 4 years. The observed decrease of the incidence in Germany from 2001 to 2003 is not a consequence of the measures taken at the end of 2000 to contain the disease. It can rather be explained by an import of BSE contaminated products from countries with a high BSE incidence in the years 1995/96 being used in calf feeding in Germany. From the future course of the BSE-incidence in Germany after 2003 a quantification of the recycling rate of BSE-infected material within Germany before the end of 2000 will be possible by use of the proposed model if the active surveillance is continued.     

Age-period-cohort analysis of the Bovine Spongiform Encephalopathy (BSE) epidemic in Switzerland

Cattle born after animal-feed control measures were implemented in 1990 have become BSE cases in Switzerland, indicating sub-optimal effectiveness of these measures. To evaluate these measures, the incidence of BSE cases in Switzerland recorded through clinical case reporting from January 1991 to June 2000 (categorized into age groups and birth cohorts of 6-month duration) was analyzed by Poisson log-linear regression using an age–period–cohort model. The incidence was maximum in the cattle cohort born from October 1989 to March 1990, and dropped to zero in the cohort born from April to September 1991. A second peak was observed in a cohort born from April to September 1994. The first drop of incidence is interpreted as a result of initial implementation of the feed ban in 1990. The second peak might be related to exposure of cattle to feed intended for pigs and poultry.     

Simulating the BSE epidemic and multiplication factor in dairy herds in Japan

With the objective of evaluating the effectiveness of an administrative guidance on the use of ruminant meat-and-bone meal in ruminant feed, effective from April 1996 to September 2001, we developed a model to simulate the evolution of the BSE epidemic and to estimate the BSE multiplication factor (K) in the Japanese dairy population. The output that provided the best fit to the number of BSE cases both observed and predicted to date suggest that the probability that bovine MBM was fed back to cattle was 14.2-75.2% and 0.129-0.570% during the periods from 1992 to April 1996 and from April 1996 to October 2001, respectively. Given these estimates, the value of K would have peaked in 1995 at 40-48 and then declined to 0.32-0.67 between 1997 and 2001. These results suggest that the administrative guidance was effective in reducing the amount of MBM fed to cattle by a factor of 104-141 and was perhaps enough to drive the epidemic towards extinction.     

Analysis and prediction of the BSE incidence in Ireland

Our purpose was to report the statistical methodology that was used to describe the nature of bovine spongiform encephalopathy (BSE) propagation in the Irish cattle population, to predict the number of future cases and to assess the risk to humans in terms of the number of infected animals that were processed. We used a nonlinear Poisson-regression model for the available birth-cohort data and an iterative method to compute the parameter estimates. Standard errors for the estimates were computed from the nonlinear model and these were validated using a bootstrap procedure.

We illustrated the use of the model for prediction and risk assessment using the BSE incidence data between 1981 and 2000. The change in case ascertainment or reporting level was a crucial parameter that determined the observed pattern of clinical BSE. Significant propagation risk was detected from 1985 onwards, with peaks in 1986 and 1994. The trough in the propagation risk in 1990 coincided with a ban of the use of meat-and-bone meal for ruminant feed. Excluding the newly adopted active surveillance method in 2001, the predicted and observed data were comparable.     

BSE, feed and cattle in Switzerland: is there a spatial relation?

Cross-contamination of cattle feed with meat and bone meal (MBM) allowed in feed for other species is regarded as the current hypothesis for the infection pathway of Bovine Spongiform Encephalopathy (BSE) cases occurring after the implementation of a ban on feeding MBM to cattle. This study was aimed at establishing a spatial relation between BSE cases in Switzerland and the findings of MBM in cattle feed. A cluster analysis and a cohort study were performed. Two hundred sixteen BSE cases born after December 1990 and detected until August 1st 2005, screening data of 504 feed producers between 1996 and 2001 and population data from the Swiss 2001 cattle census were included. The cluster analysis showed feed producer, positive for MBM contaminations in cattle feed, as possible cluster centres for BSE cases. In the cohort study, farms within a radius of 2 and 10 km around positive feed producers showed significantly higher odds to have a BSE case than the control group. The odds ratio and its 95% confidence interval were 2.23 (1.26-3.93) for the 2 km radius and 1.38 (1-1.9) for the 10 km radius. The results provide evidence for a spatial relation between cross-contamination and BSE occurrence. These findings support the hypothesis of cross-contamination to be an important route for BSE transmission after a feed ban.