Volume 4: The Southwood Working Party, 1988-89
10. Discussion
Assessment of the risk of transmission of BSE to humans
The scrapie experience
The approach to risk

10.20 We have referred above to passages in the Report dealing with the risk posed by BSE to humans, ie:

5.3.5. In these, as in other circumstances, the risk of transmission of BSE to humans appears remote.

8.2. . . . Although the risks appear remote.

9.2. It is . . . most unlikely that BSE will have any implications for human health.

10.21 The Report does not expressly explain the basis upon which the terms 'remote' and 'most unlikely' are used to describe the risks to human health.

10.22 We asked the Working Party for clarification of their assessment of risk. Their starting point was to explain the significance of the word 'remote' when used to qualify a risk in a medical context. In that context a remote risk is one that is highly unlikely to prove significant, but which it is unreasonable to ignore. Reasonable precautions should be taken to try to prevent a remote risk. ¹ The Working Party considered that the risk that BSE would prove transmissible to humans by eating beef or bovine products was remote. The slaughter and compensation policy and the advice in relation to baby food were precautions that they considered it reasonable to take to guard against this remote risk.

10.23 The risk from eating bovine tissue was seen as lower than that from other potential methods of transmission. An injection of a medicinal product derived from infective bovine tissue, or infection of an open wound as a result of occupational hazard, posed a more serious risk than that from oral ingestion. Before writing their Report, the Working Party had taken steps to see that the responsible authorities had addressed the medicinal and occupational risks. It seems that it was on the assumption that appropriate precautions were being taken that the Working Party had described those risks as 'remote' in their Report (see paragraphs 10.83-10.109 below).

10.24 We explored with the Working Party why it was that they concluded that the risk of transmission from eating bovine tissues was remote. They confirmed our understanding that this conclusion was based on the premise that BSE was likely to behave like scrapie.

The scrapie experience

10.25 Sections 3.5 to 3.7 of the Report consider whether there is any evidence of an association between scrapie and CJD. The Report concludes that there is not. Chapter 4 of the Report was written by Mr Wilesmith. It sets out in detail the epidemiological basis for Mr Wilesmith's conclusion that BSE had resulted from infection by the scrapie agent. That conclusion is set out in the Summary at 10.2 as follows:

The epidemiological evidence suggests this new disease has appeared as a result of contamination of meat and bone meal derived partly from sheep offal and fed to British cattle from the early 1980s. Contamination had arisen because modern rendering practices failed to destroy the agent of scrapie, the endemic spongiform encephalopathy of sheep.

10.26 In paragraph 6.3 of the Report the Working Party state:

It cannot automatically be assumed that animals and man will react to BSE agent exposure as they have done to scrapie, which in the human case has not led to any clear association with disease. The BSE agent may, for example, be an adapted or particularly virulent form of scrapie agent although the results of the epidemiological study indicate otherwise.

10.27 We think that the inference from the Report is that, because BSE is probably scrapie in cattle, it is likely to behave in the same way as scrapie in sheep and thus not to transmit to humans or to animals that have not proved susceptible to scrapie. The Working Party have confirmed to us that this was indeed the basis of their approach to risk:

There was no, or virtually no scientific knowledge concerning BSE available at this time. We worked on the basis that scrapie was the most likely cause of the BSE epidemic; we had to base our advice on the science relating to scrapie at that time . . . Our assessment of the likelihood of the risk posed by BSE followed from this and recognised that BSE might be more hazardous to man than scrapie had been; our recommendations were made on this basis. ²

10.28 In the period up to March 1996, the theory that BSE was scrapie in cattle underpinned the belief of many that BSE did not involve a risk to human health. With the benefit of hindsight it is now generally accepted - and acknowledged by Mr Wilesmith - that some of his provisional conclusions, which were adopted by the Working Party, were probably not correct. The affected animals recorded at the time of the Southwood Report were not all index cases that had succumbed to a common source of infection, namely scrapie. It is more probable that they were infected as a consequence of the recycling of earlier cases of BSE which had gone undetected. The initial source of infection is unknown and may never be known, but there are difficulties in the way of the theory that it was scrapie. These matters are dealt with in detail in vol. 2: Science in an analysis which would not have been possible with the limited knowledge at the time of the Southwood Report. We feel, nonetheless, that had the Working Party subjected Mr Wilesmith's theory to a rigorous assessment they might have concluded that it was not easy to accept that scrapie had always been capable of infecting cattle but had, until 1982, been inactivated by rendering. We do not, however, consider that the Working Party are to be criticised for accepting Mr Wilesmith's conclusions as to the probable source of BSE. They had been provided with Mr Wilesmith's assistance as an epidemiology expert adviser and could not be expected to attempt a peer review of his work. They emphasised to us: '. . . we were not a peer group review body. It was not our role to re-examine the specialised epidemiological data presented by Mr Wilesmith.' ³ Mr Wilesmith's conclusions were almost universally accepted at the time and are still considered by some to be a plausible explanation of the emergence of BSE. Nevertheless, the Report did not make it plain that Chapter 4, which dealt with the cause of BSE, and Chapter 6, which dealt with the future course of the disease, were written by and expressed the opinions of Mr Wilesmith, opinions which the Working Party had not felt able to peer review. We believe they should have done.

10.29 The view that BSE was unlikely to be transmissible to humans was shared by both Dr Hope and Dr Kimberlin. It was later endorsed by the Scientific Veterinary Committee of the European Commission and by the Spongiform Encephalopathy Advisory Committee (SEAC). The latter, in its Summary of Present Knowledge and Research into TSEs, published in September 1994, observed:

Our conclusion therefore is that, as the Southwood Working Party determined, taking all the available evidence together, the risk to man from BSE is remote. ⁴

10.30 SEAC went on to refer to the advice that had nonetheless been given aimed at reducing exposure to BSE agents. It ended this report:

Our scientific assessment is that the risk to man and other species from BSE is remote because the control measures now in place are adequate to eliminate or reduce any risk to a negligible level. We do, however, point out that any species exposed already and before any bans were effective could be incubating disease, and therefore continuous monitoring is very important until any possible incubation period has been exceeded.

10.31 The conclusion that BSE would probably prove no more transmissible to humans than scrapie was not founded on a scientific understanding of the factors governing transmission of TSEs in general or scrapie in particular. That some factor prevented the oral transmission of scrapie to man whereas it could be transmitted orally to other animals seemed clear from epidemiological studies. That kuru had probably been transmitted orally from one human to another was known. What was not known was the mechanism by which the TSE agent in one animal infected another, or even the biological nature of that agent. What created a 'species barrier' to the transmission of scrapie from one animal to another was not known. The conclusion that BSE would behave in the same way as scrapie was essentially a matter of judgement in the face of uncertainties, or an 'educated guess'. This was not something that the Working Party sought to conceal, as is clear from private correspondence which Sir Richard had with a number of scientists about BSE. In a letter to a Mr J Granger (a veterinary surgeon) on 25 October 1988, Sir Richard wrote:

I have to say that it is my view that the risk of transmission of BSE to man is relatively small, but this view is based entirely on drawing a parallel with scrapie . . . there's clearly a formidable barrier to its transmission to man. On the other hand as the agent has crossed apparently from sheep to cattle, we must be concerned that it would pass to another species. ⁵

10.32 The other members of the Working Party confirmed that the view expressed by Sir Richard was one which they shared. ⁶

10.33 To Dr E Poole (of the Radcliffe Infirmary in Oxford) Sir Richard wrote on 9 August 1988:

. . . my colleagues and I have made various recommendations based, I have to admit, largely on guesswork and drawing parallels from the existing knowledge of scrapie and CJ disease. ⁷

10.34 To Dr David Doyle, a consultant neuropathologist, he wrote on 8 September 1988, again in a private letter:

Against this rather comforting scenario one could argue that the agent causing BSE has already crossed one species boundary and may be a more 'virulent' strain - whatever this means in the context of scrapie agents. ⁸

10.35 We consider that it was reasonable for the Working Party to conclude by way of judgement, or educated guess, on the basis of the scrapie experience, that BSE was unlikely to be transmissible to humans and that the risk of transmission by oral ingestion was remote, giving that word the rather technical meaning it has in the medical context (see paragraph 10.22 above). Similar conclusions were reached by some of those most experienced in the field of TSEs. The Report did not, however, spell out the crude basis upon which this assessment of risk had been made.

10.36 Is this a matter for criticism? We have concluded that it is. The evaluation of risk was the primary task that the Working Party had undertaken to perform. Each member was a scientist of high distinction. Their Report was to be published, and their reasoning would need to be understood at the time and reassessed in the future as new information came to light. Their position required them to use the greatest care in the manner in which they communicated their assessment of risk. It does not seem to us that they did so. In our view the Working Party should have set out the reasoning which led them to use the term 'remote', and explained what they meant by that term. It was certainly possible to do so. We can illustrate this by quoting from an article on 'Transmissible Encephalopathies in Animals' which Dr Kimberlin submitted for publication in the Canadian Journal of Veterinary Research in June 1989: ⁹

A final question concerns the public health risks posed by BSE. On the one hand, there is the evidence that scrapie can infect other species to produce diseases such as TME and BSE; on the other, the absence of an epidemiological link between scrapie and CJD suggests that even if scrapie could infect humans, usually it does not. For this reason alone, BSE is unlikely to be a major threat to humans.

The article went on to refer to the possibility of a change in the agent on crossing the species barrier. Dr Kimberlin concluded:

It is therefore possible that the transmission of scrapie from sheep to cattle may alter the population of scrapie strains to which humans are exposed. If the strains selected by cattle have a greater pathogenicity for humans than sheep strains, the risks to humans would be increased accordingly.

From a practical standpoint, it is virtually impossible to quantify these risks. Strain typing is a protracted exercise and even if a selection of scrapie strains in cattle was demonstrated, there is no way of evaluating directly their increased or decreased pathogenicity for humans. Neither are there diagnostic tests of infection (other than bioassay) to measure the level of human exposure. And the long incubation periods of these diseases mean that it could be at least a decade before an actual risk revealed itself by an increased incidence of CJD. Because of these difficulties there is no alternative but to assume that BSE poses a real risk, however small, and take precautionary steps to reduce it to an absolute minimum.

10.37 We do not understand Dr Kimberlin's conclusions to differ from those that had been reached by the Working Party. They are, however, expressed with a clarity lacking in the Southwood Report. They are not alarmist but at the same time leave no room for complacency.

The approach to risk

10.38 The Working Party were asked not merely to advise whether BSE posed a risk to man but to advise upon the appropriate precautionary measures should they conclude that a risk existed. So far as the latter is concerned, the Working Party explained to us:

Our approach to risk was in accord with the then developing application of analysis to public risk which involved the balancing of the perceived magnitude of the risk against the practicability or achievability of successive steps for its reduction. The magnitude of a risk comprises both its likelihood and the scale of the danger. ¹⁰

10.39 This approach to risk is sometimes described as ALARP (As Low As Reasonably Practicable). It requires an exercise in proportionality. When deciding whether a precaution is 'reasonably practicable' it is necessary to weigh the cost and consequences of introducing the precaution against the risk which the precaution is intended to obviate. As the Working Party observed, in the case of BSE this was a difficult exercise:

If the homology with scrapie was complete, then there was no risk. But we judged that there was a possibility that BSE might be more dangerous to man than scrapie. This was a particularly difficult situation to evaluate, for though the likelihood of the risk was very low, the magnitude of the danger should it prove to be real was very great. ¹¹

10.40 The Working Party did not proceed upon the basis that it was certain that BSE was caused by the scrapie agent nor on the basis that it was certain that BSE would behave as scrapie behaved. They proceeded on the basis that while it was likely that BSE would prove no more transmissible than scrapie, precautions needed to be taken to guard against the possibility that this would prove not to be the case. As they pointed out, the implications of that possibility were extremely serious. We believe that this was a proper approach to be adopted by the Working Party.

10.41 We now propose to consider the recommendations made by the Working Party in relation to risk management, dealing first with the risk from eating BSE-infected tissues, then the risk from medicinal products, then occupational risk and finally the risk to animal health.