Volume 2: Science
7. Conclusions drawn from the scientific response to BSE
Assessment of research in response to BSE
The origin of BSE - transmission of scrapie to cattle
Maternal transmission
Risk of BSE becoming endemic in sheep
Minimum oral infective dose - attack rate study
Sensitivity of the mouse bioassay
Post-mortem test for BSE in cattle
Test for ruminant protein in feed - ELISA test
Uses of bovine material

7.5 A substantial body of research was undertaken in response to the emergence and identification of BSE. We discuss in vol. 3: The Early Years, 1986-88 our concerns about the delay in involving the NPU. We were also concerned that there were other specific areas in which research should have made a more valuable contribution than it did - whether because relevant research was not carried out at all before March 1996, or because it was delayed, or because there were problems with the design of individual experiments. However, when we examined those specific areas, we were satisfied that it was not appropriate to criticise individuals. We discuss the reasons why we reached this view in relation to a number of projects in this section. We discuss a number of other specific items of research in other relevant volumes of the Report, and we identify them below.

7.6 We emphasise that we have not undertaken a critical examination of the whole programme of research carried out by MAFF and the Research Councils. Some aspects of the research programme might have been approached or prioritised differently and, with hindsight, other approaches might have produced better or quicker answers - which in turn might have assisted in risk assessment and policy-making. However, once again, we do not suggest that this was a matter for individual criticism.

7.7 Much of the research commissioned during the period covered by our terms of reference was timely and of a high quality, and in this regard the work of the NPU, of the CJD Surveillance Unit, and of the Prion Diseases Group at St Mary's Hospital in London deserves specific commendation.

7.8 We also think it important to note that we do not suggest that a different approach to the research programme would have led to the earlier recognition of vCJD.

The origin of BSE - transmission of scrapie to cattle

7.9 The early epidemiology gave rise to the belief that the cause of BSE was the transmission of conventional scrapie agent(s) from sheep to cattle. It was thought that each new case of BSE was an index case, and that changes in the rendering process had enabled this to occur. The belief that BSE was caused by conventional scrapie agent(s) meant that it was thought likely to present the same risk to humans as scrapie. Reassurance was derived from the fact that scrapie-infected sheep had long been eaten without apparently causing CJD. In recommending in June 1988 that MBM should be tested on cattle and laboratory animals, the Southwood Working Party recognised the importance of validating the hypothesis upon which the assessment that BSE posed only a remote risk to humans was based.

7.10 Research, and development of knowledge (eg, the wide host range of BSE, the occurrence of a TSE in domestic cats and transmission of BSE to a pig), progressively cast doubt on the theory that BSE was caused by conventional scrapie agent(s), or that it would behave in the same way as conventional scrapie, but the uncertainty was not immediately appreciated.

7.11 One experiment that would have offered some insight into the origin of BSE was attempted transmission of conventional scrapie agent(s) to cattle. This was recognised by Mr Bradley in his 'logical approach to research' paper, written at the end of 1987, which included as item 6: 'Demonstrate transmissibility of sheep scrapie to cattle.' Yet such research was not initiated until 1997 and 1998 (projects SE1942 and SE1941), outside our terms of reference. Why was it not done earlier?

7.12 Consideration was given to such an experiment at various stages during the development of the R&D programme - Dr MacOwan and Mr Wells both described this consideration in detail in written statements to us. ¹ We note that the Tyrrell Committee did not recommend such an experiment in June 1989; that the minutes of the March 1991 SEAC meeting stated that the transmission of natural scrapie to cattle by the oral route would not be pursued; ² and that at a meeting between MAFF's Animal Health Policy Division and the Chief Scientist's Group on 9 February 1993, the policy viewpoint was that the study was of academic interest only, had not been recommended by the Southwood Report or the Tyrrell Committee, and would not lead to new policy initiatives. ³ The proposal was rejected and it was stated that it would not be considered further.

7.13 We describe in Chapter 3 some of the reasons why the experiment was not pursued. Dr Tyrrell provided information explaining why the Tyrrell Committee did not recommend such an experiment, ⁴ and Mr Bradley also outlined some of the reasons in a written statement to us:

24. The control measures which were put in place operated to cut off the source of infection. In practical terms it did not matter that policy advisers did not know where the disease had come from because the control measures would still be effective in eradicating the disease in cattle. Further, the results of such an experiment would not be available for a number of years. By the time they would have been available, the results would not have altered policy in a significant way because the control measures put in place by 1990 took account of the need to protect all species from the BSE agent in cattle, and all ruminants from any TSE agent (including scrapie agent) via feed.

25. Notwithstanding the points made above, there remains the question of whether the experiment should have been carried out and if so when. This is a complicated question to answer. I have had the benefit of reading Dr MacOwan's letter dated 22nd July, 1998 ⁵ to the Inquiry on this matter and am broadly in agreement with the points which he makes. In particular, it should be borne in mind that similar research had been and was being carried out in the United States and there was therefore a question of whether the resources could be better directed into other experiments. Other questions surrounded the objective and design of any experiment.

26. The hypothesis that BSE was derived from scrapie would only be supported by an experiment which demonstrated that when scrapie was transmitted to cattle the disease which developed in cattle was the same as BSE. Any other result would be inconclusive. However, there are a large number of different strains of scrapie and it was possible that BSE was caused by one particular strain. It is therefore arguable that it would have been necessary to attempt to transmit every strain of scrapie to cattle to properly test the hypothesis. This would have been extremely difficult as well as expensive, especially as the agents would have to be cloned in sufficient quantity for a cattle challenge. Alternatively, pooled sheep brains could have been used but this caused a difficulty because it was not known if BSE existed in sheep. If pooled sheep brains were used (thus aiming to include all the strains that occurred in the UK including possibly the BSE precursor strain or BSE strain itself), it would not be known whether BSE was present or not unless additional strain typing studies were also carried out. Further, it would not establish whether the BSE precursor strain had been naturally present in sheep historically (ie, before BSE) or whether it had been introduced into sheep via MBM and feed during the BSE era. ⁶

7.14 It seemed to us that the principal reasons were:

• a scrapie to cattle experiment had been done in the US;
• the exercise was regarded as academic - although the origin of disease in cattle was of great scientific interest, it was not of fundamental importance to the animal and public heath consequences of BSE;
• other information about the disease, such as the possible influence of genetic makeup on the susceptibility of cattle, was not known and needed to be ascertained before such an experiment was attempted;
• a large and expensive experiment was needed because there were 20 different known strains of scrapie that would have required inoculation;
• other experiments were being conducted on cattle at the CVL and would either have to be abandoned or new accommodation would be required; and
• the results would be hard to interpret and might well be inconclusive.

7.15 Some of these reasons were less persuasive than others. The US research showed that it was possible to produce a scrapie-like disease in cattle inoculated with scrapie, but the resulting TSE differed in many respects from BSE. This was not a satisfactory substitute for the scrapie-into-cattle experiment proposed in the UK.

7.16 On the other hand, there were clearly difficulties with the design of the experiment: inoculation of each of the 20 known different scrapie strains would have been difficult and expensive, and, for the reasons Mr Bradley explained, using pooled scrapie would not have provided a conclusive answer.

7.17 It is clear that other matters were regarded as being of more direct relevance and thus of immediate value. This study, by contrast, was regarded as being of academic interest - policy had to be determined before the results would be available in any event - and we can understand why it was not immediately initiated. The general public, however, did not consider the question to be academic. Nor, indeed, did scientists such as Sir Donald Acheson.

7.18 Discussion of whether attempts should be made to transmit scrapie to cattle began again in 1993. Dr MacOwan described in a written statement the progress of those discussions, leading ultimately to the initiation of the work. ⁷ We note that in 1995, in his report of the Review of MAFF-funded research into TSEs, Dr Kimberlin identified this as an 'underemphasised' area of research. ⁸ He said that the origin of BSE was the one major question about the disease that had not been addressed scientifically, and that the scientific community was becoming increasingly critical of the gap in knowledge, particularly in the light of two recent findings: (a) the BSE agent was different from scrapie; and (b) the experimental transmissions of scrapie to cattle in the USA produced a disease that did not resemble BSE. Dr Kimberlin argued that the fact that a bovine origin of BSE was quite plausible undermined considerably the risk assessment for other species of food animals. In his view it was essential to attempt an oral transmission of UK scrapie to cattle.

7.19 We share Dr Kimberlin's view, but accept that in doing so we are applying hindsight.

Maternal transmission

7.20 We describe in Chapter 3 research carried out to ascertain whether maternal and lateral transmission played a part in the spread of BSE. It was believed that scrapie was maintained in the UK sheep flock by a combination of vertical (maternal) and lateral transmission (although the emergence of BSE raised the question of whether MBM might play a role in transmitting scrapie). We discuss below the difficulties in obtaining funding for the maternal transmission study in cattle. There were also problems with the design of the experiment. In particular, it was recognised that both case and control animals might have been exposed to infected feed. MAFF officials and the Tyrrell Committee gave consideration to this problem. However, it was thought preferable to proceed with the experiment, which was designed to show whether calves then available from affected cows had a significantly greater incidence of disease than contemporary controls, rather than to delay it by a further two years in order to use only animals that had definitely not been exposed to infected feed. Dr Tyrrell explained this balancing process in his letter to Mr Andrews on 21 March 1989. We discussed it with Mr Bradley when he gave oral evidence:

Mr Bradley: We were under the clear scientific knowledge that results about maternal transmission were highly important in the control of BSE. The stimulation from the Southwood Committee and subsequently by the Tyrrell Committee reinforced that, so there was a time - the quicker we could get the study underway, the quicker we would know the result. But the earlier we started it, the greater the risk we ran with the feed ban problem . . .

And so, obviously, an ideal situation as I suggested previously would be to start with offspring from cattle which were removed essentially at birth and reared independently, if that was what we meant by maternal transmission. Actually John Wilesmith defined it as in utero or in the immediate post natal period. How long is this? Let us say until the calf is taken from the cow which, in the case of the dairy cow, would have been 24 or 48 hours. In that case it would almost eliminate a source from feed, because they would not be eating a concentrate at that age. That would be ideal. Then it would mean every calf you purchased at that particular point was only a few days old. Then you had to get 600 of them that satisfied the conditions. This would have taken longer, so you would have delayed the end result.

Lord Phillips: It would have been very difficult, would it not, to get your 300 calves of BSE infected mothers? Unless you could find a cow which was infected with BSE about to [calve], you would not know?

Mr Bradley: Exactly, so all the practical difficulties are all weighed in the balance. You asked the question, was it ideal; scientifically it was not, and I explained why, practically it was as ideal as we could get it. ⁹

7.21 This pragmatic decision seemed to us reasonable, given that all were agreed on the urgency of starting the study. It was better to obtain some, albeit imperfect, information than to wait a further two years.

7.22 As discussed in Chapter 3, the results of the experiment were finally obtained in 1996, and showed that BSE occurred in 14 per cent of the offspring of affected dams and in 4.3 per cent of the controls. With hindsight, as we note earlier, it appears that the increased risk for offspring of BSE-affected dams would have been apparent several years before the completion of the study, had emerging results been monitored and analysed by a Data Monitoring Committee. We consider in Chapter 3 the results of the 1993 case control study, which investigated maternal and lateral transmission.

7.23 When the results were obtained, there was a difficulty in analysing them. Although the experiment had been designed to be large enough to give an answer despite the possible exposure of the animals to infected feed, it was not possible to establish with certainty whether the increased incidence in offspring of affected dams was the result of maternal transmission of BSE, or maternal transmission of genetic susceptibility to BSE.

7.24 We explain in Chapter 3 that a complex genetic association study using multiple polymorphic genetic markers would have been needed to determine if there were significant differences in genotype between subjects and controls. Such a study would not have been possible in 1989. However, the consequence is that uncertainty remains about the interpretation of the results of the study.

Risk of BSE becoming endemic in sheep

7.25 We discuss in vol. 11: Scientists after Southwood the consideration of research into the risk of BSE transmitting to, and becoming endemic in, sheep. The need for such research was discussed on a number of occasions. We identify in volume 11 some of the experimental and practical difficulties in carrying it out, and indicate the reasons why it was not pursued until relatively late in the day. The question whether BSE is endemic in sheep is perhaps now the most important unanswered question about the epidemic. With hindsight it is plain that it would have been desirable to set about answering this question when it was first asked.

Minimum oral infective dose - attack rate study

7.26 We discuss in detail in Chapter 3 the widespread misconception about the amount of infective material necessary to transmit infection by the oral route. The implications of this misconception feature in volumes 5 and 6, on Animal and Human Health (1989-96) respectively.

7.27 The question of the minimum oral infective dose of the BSE agent was not addressed experimentally until January 1992, and then only indirectly as part of the attack rate study. As we indicate, the implications of experiments carried out at the NPU, in which BSE had by November 1990 been successfully transmitted to a sheep by oral administration of 0.5 g of brain material, had apparently not been appreciated by that stage. Nor had the epidemiological pointers been appreciated: cattle feed contained generally no more than 5 per cent MBM, not all of which would have been made from infective material; and compound feed, with the exception of that for young calves from dairy herds, comprised only part of cattle rations.

7.28 Once more, Mr Bradley had identified the need to demonstrate 'minimal infective doses' for cattle, primates and laboratory animals in his 'logical approach' paper in December 1987. Why were experiments not set in hand to do so? The Tyrrell Committee did not recommend either epidemiological investigations into the minimum oral infective dose, or an attack rate experiment to establish this. They explained to us why they had not done so:

The members of the Committee were aware of the importance of ascertaining the size of the dose that could transmit BSE by oral ingestion. The members of the Committee had thought about epidemiological calculations of the amount of infective material likely to have been included in a cow's rations.

The Committee could have recommended epidemiological calculations of the amount of infected material likely to have been included in a cow's rations. This approach, however, fails to identify an important part of the question, which is that you are concerned with the concentration of infectivity and not just the amount of infected material. The practical or scientific value of such a calculation would have been severely limited by the large number of unknown factors to which a range of arbitrary values would have had to be assigned. For example, calculations of the exposure of cattle to sheep scrapie required estimates of the prevalence of infection in sheep (which was unknown), the effects of different rendering processes on infectivity (which was unknown) and the extent of the species barrier between sheep and cattle (which was unknown). Furthermore, the extent of the species barrier would depend on the individual strains of agent and variations in the PrP genotypes in sheep and cattle. Similarly, calculations of the exposure of cattle to recycled BSE, that is BSE through meat and bone meal in a cow's rations would require estimates of the prevalence of BSE infection in cattle, the infectivity titres of BSE in different tissues at different times after infection and the effects of rendering processes. These are the difficulties, which led the members of the Committee not to recommend specifically the first of the two suggested possible approaches. However, some of these matters were the subject of recommendations for research made by the Report, albeit primarily for other purposes which were judged to be of greater potential value. For example, paragraph C1(e) of the Report dealt with bio-assays of BSE agent in tissues of infected cattle and paragraph C3 of the Report recommended, and gave two stars to, research into different rendering processes so that knowledge could be acquired about inactivating infectious agents.

The Committee could also have recommended that an attack rate experiment be undertaken. The members of the Committee believed that such an experiment belonged to a later generation of studies in cattle which would only become possible when information had been obtained on the variation of bovine genetic factors and their effect on susceptibility to BSE infection. To do otherwise might lead to an experiment being performed using animals which were found to be genetically not susceptible to the agent. ¹⁰

7.29 As regards the last point, Dr Watson told us that he learnt on 26 July 1989 that BSE had recently been successfully transmitted from cattle to cattle by experimental means at the CVL for the first time. ¹¹ He continued:

This meant that the species in which the disease was occurring could be infected experimentally and these findings opened the way to major experiments on dose rate and pathogenesis studies in this species.

7.30 We agree that this study did not belong in the first generation of experiments but had to await the outcome of other work. Perhaps with prompting from policy-makers to whom this knowledge would have been useful, work might have begun earlier. With hindsight it is regrettable that other indications about the small quantity of physical material capable of transmitting infection were not appreciated.

7.31 The attack rate study provided experimental evidence in 1994 that as little as 1 gram of infective BSE brain material given orally could cause disease. It is now known that this quantity is sufficient to cause disease in 70 per cent of animals. This fact has implications that are serious and wide-ranging. They feature in many different aspects of the story, for example, in relation to the role of cross-contamination of cattle feed in infecting cattle born after the ruminant feed ban (BABs) (see vol. 5: Animal Health, 1989-96); the related question of the importance of developing an ELISA test to detect contamination of feed; and the need for strict enforcement of the SBO Regulations (see volumes 5 and 6, on Animal and Human Health respectively).

Sensitivity of the mouse bioassay

7.32 The need to establish the infectivity of different organs, tissues and body fluids was identified at an early stage. It, too, featured in Mr Bradley's 'logical approach to research' paper. We describe in Chapter 3 the results of the NPU tissue infectivity study and of the CVL pathogenesis study. Following the successful transmission of BSE to mice, it was decided that a mouse bioassay should be used. The use of calves was regarded, in our view fairly, as prohibitively expensive. Corresponding studies into scrapie infectivity carried out by Hadlow had successfully used mice.

7.33 However, as we describe in Chapter 3, the preliminary results of the NPU studies, published in 1992, showed that tissues which were known to be infective in scrapie-infected sheep, particularly spleen, failed to produce disease in the mouse strain most susceptible to BSE infection. By the time of the eighth BSE R&D meeting between the CVL and NPU in October 1991, spleen had failed to produce disease after 810 days. This and other hitherto negative results raised the question of whether the mouse assay was sufficiently sensitive to detect BSE infectivity, and it was suggested that spleen should be inoculated into calves. Others, for example Dr Dealler and Professor Lacey, raised concerns about the sensitivity of the mouse assay as early as April 1991. ¹²

7.34 The spleen proposal was considered at the beginning of 1992. At a meeting on 10 January between staff of the CSG, MAFF's Animal Health Division (AHD) and the CVL to decide on the allocation of some available research funds (£280,000), it was agreed that inoculation of spleen from a clinical case into calves was a priority. An experiment was to be designed and costed, and Mr Bradley was to produce a paper for MAFF colleagues on whether or not it was scientifically desirable to validate the mouse assay for BSE by comparing the sensitivity of calves and mice. ¹³ Mr Robert Lowson of AHD and Mr Bradley later spoke to the CVO and it was agreed that BSE should be inoculated into calves as a priority. ¹⁴

Mr Wells told us:

At the seventh CVL/NPU R&D meeting on 23rd April, 1991 (YB91/4.23/1.1) it was noted that the absence of detected infectivity in spleens to date from field cases of BSE was an unexpected result. One of the issues this raised was whether the mouse assay system was sufficiently sensitive. A study was therefore initiated in April 1992, to determine the extent of the under-estimation of infectivity titre in BSE tissues when titrated across a species barrier in mice. The study is incomplete, but has demonstrated so far that there is at least a thousand fold difference in the sensitivity of the bioassay in mice and cattle. That is, a thousand fold greater sensitivity of detection of the BSE agent by bioassay in cattle. ¹⁵

7.35 A progress report on the experiment produced by Mr Wells in April 1993 stated that the mice had actually been inoculated in January 1993. ¹⁶

7.36 With hindsight it is unfortunate that the use of mice to assay BSE infectivity in different tissues did not prove as successful as their use to assay scrapie infectivity. This has necessitated further work in calves in order to provide more definitive answers - these are needed for a full range of tissues, including milk (to test for maternal transmission), faeces and urine (to test for lateral transmission).

Post-mortem test for BSE in cattle

7.37 We note in vol. 11: Scientists after Southwood that there was a practical need for a simple post-mortem test for BSE, but that this did not feature in the report of the Tyrrell Committee. Such a test would have been invaluable in screening to establish the extent of subclinical cases of BSE, and even for screening meat. We discuss below the possibility that a research supremo or coordinating body might have assisted in this area. The work of Dr Narang and the appraisal of his touch test to detect BSE infectivity in cattle is considered in volume 11.

Test for ruminant protein in feed - ELISA test

7.38 We also note in volume 11 that the need for a test in order to identify the presence of ruminant protein in feed was not drawn to the attention of the Tyrrell Committee and did not feature in their report. Accordingly it did not receive priority treatment. One approach to this - the development of an ELISA test for detection of ruminant protein in feed - is discussed fully in vol. 5: Animal Health, 1989-96.

Uses of bovine material

7.39 The only three-star recommendation of the Tyrrell Committee that was not set in hand immediately was A1d, namely that consideration should be given to commissioning a study into the fate of bovine (and ovine) tissues and products that could lead to infection being spread by as-yet-unrecognised routes. It seemed to us that one reason why it fell through the cracks was that it was not scientific research in the way that the other projects were. We consider the need for such an audit, and the fate of the Tyrrell recommendation, in vol. 7: Medicines and Cosmetics.