Chapter 4

Contents - Previous - Next

4. Epidemiology

4. Epidemiology

4.1 Early findings
4.2 The vehicle of infection
4.3 The start of the BSE epidemic
4.4 The north-south gradient
4.5 The recycling of infection in cattle
4.6 The development of the epidemic
4.7 Spongiform encephalopathy in other animal species

BSE was first recognized in November 1986. Retrospective analysis of case histories indicated that a small number of BSE cases had occurred as far back as April 1985. As the result of extensive surveillance, about 130 cases of BSE had been confirmed in the United Kingdom by the end of 1987 (Matthews, 1990). Over 2 000 cases were confirmed in 1988, but it should be remembered that BSE became a notifiable disease in June of that year (HMSO, 1988a) and the number of reported cases increased dramatically, from around 60 cases a month to 50 to 60 cases a week, after notification had been introduced. During 1989, the first full year of BSE notification, over 7 000 confirmed cases were recorded. By the end of 1990 the total for the United Kingdom was over 20 000.

4.1 Early findings

A major epidemiological study was started in June 1987 (Wilesmith et al., 1988). Although infection was undoubtedly the cause of BSE, it was important to eliminate other possibilities, particularly as the transmissibility of BSE had not yet been demonstrated.

There was no association of the time of onset of BSE with the stage of pregnancy or with calendar month, as might occur following the seasonal use of various pharmaceutical products or agricultural chemicals. Many products were specifically excluded as causes of BSE: for example, vaccines, hormones, organophosphorus fly sprays, synthetic pyrethroid sprays, anthelmintics, herbicides, pesticides, etc. (Wilesmith, 1992).

BSE is clearly not a disease of genetic origin. It has occurred in the majority of United Kingdom dairy breeds and their crosses, in the proportion expected from their representation in the national herd (see Table 3). Analysis of available pedigrees excludes a simple Mendelian pattern of inheritance as the sole cause of the disease.

TABLE 3. Distribution of confirmed cases of BSE in dairy cattle of different breeds and distribution of dairy breeds in the United Kingdom

Breed Percentage of confirmed
BSE cases
of dairy breeds
Holstein-Friesian 91.5 89.7
Ayrshire 3.9 2.2
Channel Islands 3 7 3.4
Others 0 7 4 7

Source: Wilesmith et al, 1992a.

However, the epidemiological data neither exclude nor support the possibility of bovine genetic factors controlling the susceptibility to an infectious disease, such as occurs with scrapie (Wilesmith et al., 1988; Wijeratne and Curnow, 1990).

But two other pieces of evidence suggest that host genetic variation may not be of great importance in BSE. The first is the remarkably (compared to scrapie) uniform pattern of severity and distribution of vacuolar lesions in BSE (see Histopathology, p. 33). The second is the 100 percent susceptibility and high uniformity of incubation periods seen in a total of 16 Jersey and Holstein-Friesian cattle that had been injected with BSE (Dawson, Wells and Parker, 1990, Dawson et al., 1991). A comparable experiment with scrapie in sheep would have given variability of both incidence and incubation period.

The epidemiological studies further showed that the occurrence of BSE was not associated with the importation of cattle, the use of semen, or the movement of breeding animals between herds. In view of the subsequent evidence that infection with scrapie was the cause of BSE, it was particularly important to find that BSE was not associated with the presence of sheep on the same farms (Wilesmith et al., 1988).

FIGURE 3 Number of cases of BSE by the month and year of onset clinical signs (April 1985 to April 1991)

4.2 The vehicle of infection

The form of the epidemic curve (showing the occurrence of BSE by month and year of clinical onset) is typical of that of an extended common-source epidemic (see Fig.3). By a process of elimination, the only common factor to be identified was the feeding of proprietary feedstuffs. Commercial calf pellets, cow cake or protein supplements to home mixed rations have been fed to all cases for which accurate records are available (Wilesmith et al., 1988). This was still true at the end of 1990, when over 20 000 cases of BSE had occurred. Every case is a primary case and there is no evidence of cattle-to-cattle transmission of infection (apart from a report in March 1991 [Anon., 1991b] of a possible instance of maternal transmission to an animal that was born in November 1988, four months after the ruminant protein ban).

Two animal-derived products may be incorporated into proprietary feedstuffs: tallow, and meat and bone meal. The balance of evidence shows meat and bone meal to be the vehicle of infection.

First, the physicochemical properties of the scrapie agent make it more likely to partition with the protein fraction than with the lipids of tallow.

Second, although BSE has shown a wide geographical pattern of occurrence throughout the epidemic, there has always been a striking north-south gradient, with the greater number of cases occurring in the south and east. The basis of this pattern (see The north-south gradient, p.24) is likely to be due to meat and bone meal because both its distribution and incorporation into animal feeds occur within a relatively short distance from its production. Just the opposite is true of tallow (Wilesmith et al., 1988).

The food-borne hypothesis is strongly supported by several other important features of the epidemic (Wilesmith et al., 1988; 1992a; Wilesmith, Ryan and Hueston,1992c). First, BSE occurs much more often in dairy herds than in beef suckler herds. The difference is not related to any variation in breed susceptibility but to different feeding practices. In dairy herds, it is common to feed concentrates containing meat and bone meal during the first six months of life. Such feeds are rarely used for beef suckler calves, whose food consists of milk from the dam, mostly supplemented by conserved forage and cereals. About 85 percent of all BSE cases in beef suckler herds occur in purchased animals and a high proportion of these are cross-bred animals that were born in dairy herds and probably infected there before being sold. A comparison of BSE incidence in herds with only home-bred cases shows that the proportion of dairy herds affected is about 50 times greater than the proportion of beef suckler herds affected (see Table 4).

TABLE 4. Cumulative proportion of herds with BSE in home-bred cattle according to type and size of herd

Herd type Herd size Total
< 50 50-99 100-199 >199  
Dairy and mixed
No. of BSE herds 779 2 384 2 615 424 6 263
No. of herds at risk 21 072 15 728 6 935 1 032 44 767
Percentage affected 3.70 15.16 37.71 41.09 13.99
Beef suckler
No. of BSE herds 77 40 20 4 142
No. of herds at risk 48 182 4 454 1 347 183 54 166
Percentage affected 0.16 0.90 1.48 2.19 0.26

Note: Total includes 61 dairy, herds and one beef herd of unknown size.
Source: Wilesmith et al., 1 992a.

Second, the incidence of BSE-affected herds of either type increases progressively with herd size; the bigger the herd, the more feed is required and the greater the chances of buying an infected batch (see Table 4). Third, over 50 percent of all affected herds have had only one case and another 20 percent have had only two. Both observations strongly indicate a low average exposure to a source of infection outside the cattle population.

Finally, a case-control study shows that the inclusion of meat and bone meal in proprietary calf feeds is a statistically significant risk factor for the occurrence of BSE (Wilesmith, Ryan and Hueston, 1992c). A computer based simulation model has been constructed to analyse some of the time scales of the epidemic (Wilesmith et al., 1988). The model shows that the exposure of the cattle population must have started abruptly, around the winter of 1981/82. Both adults and calves have been exposed but the majority of affected animals were exposed as calves. This means that the age-specific incidences will reflect incubation period.

Since the incubation period was originally estimated to range from 2.5 to at least eight years, the model predicts that, during the earlier years of the epidemic, more and more cases will occur in older animals as the full effect of the incubation period distribution becomes manifest. This prediction is confirmed by the observation of a higher incidence of BSE in animals aged five years and above in 19X8 than in 1987 (Wilesmith, Ryan and Atkinson, 1991; Wilesmith et al., 1992a).

As of July 1991, the incidence of BSE was highest in four- and five-year-olds. The oldest recorded case was 15 years and the youngest was 22 months (but born before the ruminant protein ban was introduced).

4.3 The start of the BSE epidemic

A key epidemiological finding is that exposure to a scrapie-like agent, sufficient to cause clinical disease, started in the winter of 1981/82.

It seems likely that the epidemic began by the infection of cattle with scrapie agent from sheep. However, the present epidemiological evidence does not distinguish between this and an alternative possibility, namely that an endemic infection of cattle existed long before the current epidemic starned, but was undetected because the incidence of clinical disease was so low. To be realistic, the incidence would have had to he less than one case per 100 000 adult cattle per annum, which was the incidence et the very start of the epidemic (Wilesmith et al., 1988). Such a possibility has been suggested as the cause of an outbreak of TME on a ranch in the United States where mink were regularly fed dead cattle (Marsh et u/.,1991). Undetected BSE could exist in other countries but, until there is direct evidence for such a possibility, it must be assumed that scrapie was the original cause of BSE.

It is well established that scrapie can be experimentally transmitted both within and between species, by the feeding or intragastric administration of infected material (Kimberlin, 1990a). Scrapie has been endemic in the United Kingdom for nearly three centuries, and the country also has a very large sheep population. In terms of animal waste, about 15 percent of all rendered material is ovine compared with about 45 percent of bovine origin (HMSO, 1985). But if a large sheep population (relative to cattle) with endemic scrapie, and the use of meat and bone meal as a feed supplement, are necessary preconditions for BSE why did the disease not occur before the 1980s?

During the production of meat and bone meal the temperatures achieved by most rendering processes operating at normal atmospheric pressure would not be high enough to guarantee the total elimination of large amounts of scrapie infectivity (Taylor 1989a). Yet they may have been adequate to disinfect lower levels of contamination until recent changes in rendering processes allowed sufficient infection to survive in meat and bone meal (Wilesmith, Ryan and Atkinson, 1991).

A detailed investigation has been carried out on the 46 rendering plants in operation in the United Kingdom in 198X (Wilesmith. Ryan and Atkinson, 1991).

During the period from 1972 to 19887 the proportion of meat and bone meal produced by continuous processes as opposed to batch processes, increased from 0 percent to about 75 percent. However, this is unlikely to have been a major factor in causing BSE, for two reasons. First the change was too gradual to account for the sudden onset of exposure of cattle in 1981/82. Second, the survey did not reveal a difference in the mean maximum temperatures between continuous and batch processes. In addition the particle size of the raw material is smaller in the continuous processes: this would favour a more, not less, efficient inactivation of scrapie.

The same period (1972-1988) also witnessed a decline in the use of solvent extraction which was employed to increase the yield of tallow. But this change was quite abrupt. The proportion of meat and bone meal produced by the use of solvents decreased by nearly 50 percent between 1980 and 1983. Not only does this fit the predicted onset of exposure but the move away from solvent extraction would have meant the loss of two partial scrapie-inactivation steps.

It is likely that the usual conditions of solvent extraction for about eight hours at 70" C, would have reduced infectivity and/or made the residual infectivity more heat sensitive. The second step was the direct application of superheated steam to meat and bone meal for 15 to 30 minutes to remove the last traces of solvent. Wet heating is much more effective against scrapie than dry heating Wilesmith, Ryan and Atkinson, 1991). It can be concluded that the cessation of solvent extraction was a major factor causing BSE.

4.4 The north-south gradient

Only two rendering plants in the United Kingdom still use solvent extraction, and both are in Scotland. This helps to explain the much lower incidence of BSE in Scotland (see Fig. 2).

In addition, half of the remaining plants produce greaves as an intermediate product which is then sold to other plants for further processing to produce meat and bone meal. Some of the greaves is mixed with raw material and subjected to a complete processing cycle. About 15 percent of all meat and bone meal receives this second heat treatment.

However, there are major regional differences in the amount of reprocessing of greaves. Very little meat and bone meal is produced in this way in the south of England, but increasing proportions are produced in the Midlands, the north of England and Scotland. This variation would contribute to the north-south gradient in the incidence of BSE (Wilesmith, Ryan and Atkinson, 1991 ).

Other factors could also be involved. Although both sheep and scrapie are widely distributed in the United Kingdom, some regional variation would be expected in the amount of infected sheep material entering different rendering plants. There would also be variation in the use of meat and bone meal by different commercial compounders of cattle feedstuffs, as well as differences in their geographical market share of sales.

Present epidemiological studies seek to evaluate these factors. However, the situation in much of the country is complex (Wilesmith et al., 1992a) and a more fruitful approach has been to focus on BSE in Northern Ireland (Denny et al., 1991) and the Channel Islands. It is interesting that the dramatically different occurrences of BSE in Guernsey and Jersey can be associated with differences in the manufacturers supplying feedstuffs from the mainland (Wilesmith, 1992).

4.5 The recycling of infection in cattle

There is now sufficient information to reconstruct the salient events in the BSE epidemic in the United Kingdom.

As explained above, it is now assumed that scrapie was the original cause of the BSE epidemic. It is theoretically possible that BSE originated with a mutant scrapie strain that arose spontaneously in sheep and simply happened to be more pathogenic for cattle than other scrapie strains. This possibility is discounted by the form of the epidemic, which would require the simultaneous emergence of this mutant strain in many flocks throughout the country (see Fig. 1). This is improbable (Wilesmith et al., 1988). It is far more likely that the epidemic was started by one scrapie strain that is common in different breeds of sheep, or possibly, a few strains that behaved in a similar manner when crossing the sheep-to-cattle species barrier.

However, the continued exposure of cattle to sheep scrapie was not the ultimate driving force of the epidemic. On the contrary, the epidemic would inevitably have been amplified by the subsequent recycling, via meat and bone meal, of infected cattle material within the cattle population. Recycling would have produced the equivalent of a serial passage of the infection, as happened with kuru. Because of the length of BSE incubation periods, recycling would have established the pattern of the epidemic long before BSE was even recognized (see The development of the epidemic, p. 26).

One consequence of recycling is that it would favour the selection of cattle adapted strains of agent, and these strains could differ from those in the sheep population. Present evidence suggests little or no allelic variation in any cattle genes that might affect the incubation period. This means that all cattle would tend to exert a similar selective pressure, favouring strains with the shortest incubation period. Since some scrapie strains are known to be more heat stable than others, the rendering process itself could also have exerted a uniform selective pressure favouring heat-resistant strains.

Several isolates of the BSE agent are being studied by experimental passage in mice (Fraser et al., 1988; Fraser, Bruce and McConnell, 1991). Preliminary evidence shows that BSE isolates from geographically separate sources have strikingly similar incubation periods and other properties. This suggests that BSE was caused by a single common scrapie strain in sheep.

This strain could have been selected and passaged unchanged in cattle (through recycling). Alternatively, it could have given rise to a mutant strain in cattle that was subsequently selected because it had a shorter incubation period in cattle than the parental strain.

The latter possibility is somewhat favoured by the evidence that isolates of BSE, while strikingly similar to one another, are different from past isolates of sheep scrapie in mice and also from one recent scrapie isolate (Fraser, Bruce and McConnell, 1991). Unfortunately, natural scrapie isolates that were contemporary with the start of the exposure of cattle are not available to make the most appropriate comparison.

4.6 The development of the epidemic

With experimental scrapie, serial passage in a new species usually involves a reduction of incubation period, even when there is no selection of strains. In addition, strain selection always favours the strain with the shortest incubation period (Kimberlin, Cole and Walker, 1987; Kimberlin, Walker and Fraser, 1989). Therefore, a likely consequence of the recycling of BSE in cattle is a reduction in incubation period (and this would be independent of any reduction resulting from an increase in infectivity, as discussed below). Evidence of a reduced incubation period is being sought by analysing the age at onset of disease, which reflects incubation period, as a function of year of birth, which is when a majority of cases would have been exposed (Wilesmith and Ryan, personal communication).

The second consequence of recycling is the multiplication of infectivity during each passage. This would increase the total amount of infectivity circulating in the cattle population. The third consequence is that there is no longer a species barrier. In terms of effective dose, this could be the most important consequence because the species barrier is usually the limiting factor in the interspecies transmission of scrapie-like agents. The epidemic would inevitably be driven by cattle BSE, which would then have a selective advantage over sheep scrapie.

A change in effective dose is reflected in the stepwise increase in the incidence of BSE, starting in mid- 1989 (see Fig. 3). This change is unlikely to be due to an increased ascertainment of cases, which would have reached a consistent high level by this stage of the epidemic. Moreover, the same change was observed in the Channel Islands, where the reporting of cases has always been close to 100 percent (Wilesmith et al., 1992a). Therefore, the increase of BSE in mid- 1989 suggests a substantial degree of recycling of infected cattle material around 1984/85 (Wilesmith and Wells, 1991).

The way in which the incidence of BSE increased is particularly interesting. One would expect a higher effective dose to increase both the number of affected herds and the incidence within affected herds. In practice, the latter has changed little, but there has been a large increase in the number of herds with BSE. This means that the average dose of infectivity was extremely low and the main effect of recycling was to increase the number of batches of meat and bone meal with the minimum "threshold" amount of infectivity necessary to infect cattle, rather than the concentration of infectivity within batches. Later on, one might expect the average concentration to have risen sufficiently to cause a more obvious increase in the incidence of BSE within affected herds (Wilesmith, personal communication).

Up to July 1991 about 70 percent of all affected herds had only had one or two cases. This is in marked contrast to some outbreaks of TME in which morbidity approached 100 percent. The tragedy of BSE is that, despite the low exposure, a substantial proportion of a large cattle population (about four million adults) was exposed from 1981/82 until the summer of 1988 when the ruminant protein ban came into effect. This explains why the total number of BSE cases in the United Kingdom is so high.

It is difficult to estimate the attack rate for BSE because the distribution of infectious agent in meat and bone meal would not have been homogeneous and feeding regimes would have varied from herd to herd and from year to year.

However, the theoretical approach discussed by Wilesmith (1991) gives some idea of the attack rate. If the average dairy herd has 70 adult cows and the annual replacement rate is 20 percent, then each new birth cohort joining the adult herd would comprise 14 heifers. A single case in the cohort thus represents an attack rate of 7 percent (Wilesmith, 1991).

The most important remaining question is the future course of the epidemic. This is discussed in Chapter 9.

4.7 Spongiform encephalopathy in other animal species

The epidemiology of BSE shows how animals can be exposed to scrapie infection via processed feedstuffs, not just by the feeding of untreated sheep carcasses or offal as was indicated by studies of TME.

Prior to BSE, a transmissible spongiform encephalopathy known as chronic wasting disease (CWD) (see Table 1) was described in captive mule deer (Odocoileus hemionus hemionus) and Rocky Mountain elk (Cervus elaphus nelsoni) held in wildlife facilities in Colorado and Wyoming (Kimberlin, 1990a). About 100 cases have been diagnosed since 1967 (Williams et al., 1990). The majority of affected animals were born in the wild and found as orphans. A source of infection in captivity has not been identified. Scrapie occurs in the United States and an exogenous feed source of infection is a possibility. However, cases of disease have been diagnosed in free-ranging deer and elk (Williams et al., 1990).

Concurrently with the BSE epidemic in the United Kingdom, cases of spongiform encephalopathy have occurred in five exotic species of ruminants kept in wildlife parks and zoos in England. The species are:

• nyala (Tragelaphus angasi) in 1986 (Jeffrey and Wells, 1988);
• gemsbok (Oryx gazella) in 1987 (Jeffrey and Wells, 1988; Wilesmith et al., 1988);
• Arabian oryx (Oryx leucoryx) in 1989 (Kirkwood et al., 1990);
• greater kudu (Tragelaphus strepsiceros) in 1989 (Kirkwood et al.; 1990).
• eland (Taurotragus oryx) in 1989 (Fleetwood and Furley, 1990).

The pathology of these diseases leaves no doubt about their nature and likely aetiology. Most, if not all, of the animals involved were fed diets containing meat and bone meal before the ruminant protein ban was instituted. Brain material from clinical cases in a nyala and a greater kudu has produced spongiform encephalopathy after injection into mice.

The diseases in zoo animals differ from BSE in three respects. First, the clinical duration was considerably shorter than in BSE, sometimes just a few days. Second, the incidence of disease in the zoo animals was disproportionately high in relation to the size of the population exposed. Third, the age of onset of clinical disease was much younger (30 to 38 months) in the zoo animals than in cattle. The youngest case of BSE so far recorded is at 22 months, but the onset of clinical signs is typically at four to five years.

These findings suggest a higher effective exposure of the zoo animals than cattle. But the route of exposure and the inclusion rate of meat and bone meal would have been similar. It seems, therefore, that the species barrier was lower for the exotic ungulates, resulting in an increased susceptibility to infection.

A second case in a greater kudu was reported at the end of 1990 (Anon., 1990b). The animal was born in April 1989 and was never fed ruminant protein. Disease developed 19 months later and the clinical duration was only a few days. The calf was born to the first case of spongiform encephalopathy in a greater kudu and the simplest explanation is that it became infected from its mother (Kirkwood et al., 1992). Maternal transmission of infection is well established in sheep scrapie and the possibility of it occurring in cattle should now be considered more likely (see The possibility of endemic infection, p. 55).

Since January 1990 cases of a new spongiform encephalopathy have occurred in several adult domestic cats in various parts of the United Kingdom (Wyatt et al., 1990; 1991; Leggett, Dukes and Pirie, 1990). Feline spongiform encephalopathy (FSE) is experimentally transmissible to mice by the injection of affected brain. As with the zoo animals, it seems likely that FSE was also the result of food-borne infection, but the fact that it was only recognized in early 1990 suggests that the infection may have been BSE rather than scrapie.

Cats are fed a wide variety of different foods, including offals and prepared pet foods containing meat and bone meal. There have been insufficient cases of FSE to identify the source of infection. The total number of cases in 1990 was 12, and only six had occurred in the United Kingdom by the end of July 1991. Because there would be no recycling of infection from cat to cat via feed, a large-scale epidemic like BSE seems unlikely.

Before the first case of FSE had been recognized, the United Kingdom pet food industry had already instituted a voluntary "specified offals ban" in 1989, as a precautionary measure against the infection of cats and other domestic species. This measure became mandatory in 1990 (HMSO, 1990b) (see Minimizing the exposure of other species, p. 53).

Contents - Previous - Next