Bovine Creutzfeldt-Jakob disease?

Failures of epidemiology must be remedied

Britain's Creutzfeldt-Jakob Disease Surveillance Unit was set up in 1990 to alert the Department of Health and the government's Spongiform Encephalopathy Advisory Committee (SEAC) to any changes that might suggest that humans were affected by exposure to the agent responsible for bovine spongiform encephalopathy (BSE). These would include changes in age specific incidence, occupational distribution, or dietary correlates of cases of Creutzfeldt-Jakob disease. Presentation, if it happened at all, was considered more likely to be atypical, but this could not be described in advance. The surveillance unit has fulfilled its remit spectacularly and speedily: a previously unrecognised and consistent disease pattern in young adults has been found, the most likely explanation for which is exposure to the bovine spongiform encephalopathy agent, most probably (but not necessarily) before specified bovine offals were banned in 1989.

Since 1 May 1990, 10 cases of Creutzfeldt-Jakob disease have been reported in people under the age of 42. It is important to put these 10 cases in a broad epidemiological context. We have three estimates of the incidence of sporadic Creutzfeldt-Jakob disease for people aged 15-39: 0.05 per million person years,1 (the same as for people aged 40-44 years2), giving an expected number of 6.3 cases in the six years since 1 May 1990; 0.0286 per million person years, based on three sporadic cases of Creutzfeldt-Jakob disease in this age group reported from 1985 to 1989 (RG Will, personal communication), giving an expected number of 3.6 cases; or a previous guesstimate of 0.01 per million person years,2 (one fifth the incidence at age 40-44), giving an expected number of 1.26 cases. Against each of these expectations, the probability of 10 or more sporadic cases of Creutzfeldt-Jakob disease occurring in people aged 15-39 in the six years from 1 May 1990 is 11 in a hundred, 4 in a thousand, or 0.9 in a million respectively. The 10 cases were actually referred to the surveillance unit over less than a six year period—how much less is epidemiologically important for estimating the epidemic doubling time. At the moment we don't know the doubling time, from say five to 10 cases; it may be three, six, or 12 months, or more. If in the past six months the number of cases increased from five to 10, a further increase from 10 to 20 cases in the next six months would be consistent with an epidemic; but because of chance, so too would any number of cases from four to 16.

Now that public health rather than agriculture is the top priority, the signal failure of bovine spongiform encephalopathy epidemiology must be remedied (box 1). In sharp contrast to the study of HIV and AIDS in humans, projections about the prevalence of bovine spongiform encephalopathy in cattle have been subjected to neither deliberation by an independent working party nor to statistical peer review. Nor have the projections been regularly published or publicly monitored. This applies critically to projections of numbers of affected cattle born after the July 1988 ruminant feed ban. These cattle now make up 56% (8370/15001) of all cases reported in 1995 (to the end of last October.3

Britain's paramount contributions to the epidemiology of HIV/AIDS included key data acquisition,4 monitored projections,5 peer reviewed statistical methodology,6 7 and quantification of maternal HIV transmission.8 But mistakes were made. We were told, for example, that there was no evidence of HIV transmission via breast milk, but before long this was found to be a false reassurance. Likewise, guidelines for HIV infected healthcare workers were inadequate.9 Such mistakes seem likely to be repeated in dealing with bovine spongiform encephalopathy. Occupational precautions to safeguard abattoir workers against exposure to infected tissues have not been persuasively advocated and, if press and television pictures are typical, have been lax. If careless of themselves, such workers will have been hapless guardians of the public health.

Milk from cows suspected of harbouring the disease is used only for feeding to their calves; such cows are isolated when calving; and the carcases of affected cattle are incinerated. But the British government has abrogated to farmers and their veterinary advisers decisions about breeding from the offspring of affected cows.10 Now that there is risk to humans, but uncertainty about its extent, two things should happen. Every regulation, especially those in relation to calves under 6 months, should be reviewed. What is the evidence, for example, for SEAC's exclusion of cattle under 30 months from the new deboning and specified cattle offals provisions announced last week? Some bovines under 30 months are certainly infected. Let us have done with misleading the profession, the public, and the press with unqualified “no evidence of” statements. All evidence must be quantified.

Short of culling the entire British herd, options range from age specific random selection and slaughter of cattle, using their brain pathology to determine the fate of other cattle of the same age or in the same herd, to the more radical solutions of culling all cattle born before 1990 (allowing a safety margin for transgression of the July 1988 feeding ban), all progeny of affected cattle because of uncertainty about transmission from dam to calf, and all herds with any affected member born after the ban because of evidence of lateral transmission.11

Data on transmission from dam to calf are crucial for determining plausible scenarios for the infection curve in cattle born after the July 1988 feeding ban, as are data on feeding transgressions. The ongoing study of possible vertical transmission in 315 paired calves, now in its seventh year,10 should be unblinded. Its results and other available data on transmission from dam to calf are crucial for projections, but also because of their implications for human health (see box 2). Work on projections has been further complicated by changes in instructions to veterinary officers and in European Union regulations on live exports, not to mention reductions in compensation to farmers for slaughtered animals. This compensation was reduced in April 1994 from the average market price of a young bovine to the value of older bovines,12 which may ironically have jeopardised the selectivity and speed of the very case reporting on which we rely for projections.

Although it is too early to make confident projections about the new Creutzfeldt-Jakob disease variant, it is not too early to identify our data requirements. These include an estimate of the disease's incubation period and plausible scenarios for the infection curve in humans from 1981. Such scenarios should take account of British consumption of bovine tissue at three potential levels of infectivity, the highest being the specified bovine offals banned in 1989; the next being those additionally proscribed in Britain or (which are not the same) elsewhere in the European Union since then.

Wisely, even before the recent moves to ban British beef, the European Union was not content to rely on standard qualitative assays of infectivity13 for reassurance on the safety of beef. It restricted exports to bovines born after the ban and from farms that had not had a recent case of bovine spongiform encephalopathy.3 12 No such restrictions limited British consumption. Practically, we do not have rapid in vitro or in vivo tests for infected cattle and so have continued to play Russian roulette with no information on the odds. Age specific prevalence of infected cattle has not even been monitored by random pathology after slaughter, for which there is now the strongest case.