Mortality from dementias and neurodegenerative disorders in people aged 15-64 in England and Wales in 1979-96

BMJ 1998; 317 doi: (Published 01 August 1998) Cite this as: BMJ 1998;317:320
  1. Azeem Majeed, medical epidemiologist (azeem.majeed{at},
  2. Petra Lehmann, research officer,
  3. Liz Kirby, research officer,
  4. Michel P Coleman, deputy chief medical statistician
  1. Office for National Statistics, London SW1V 2QQ
  1. Correspondence to: Dr Majeed
  • Accepted 6 March 1998

There is now good evidence that the new variant of Creutzfeldt-Jakob disease first identified in 1996 is caused by exposure to bovine spongiform encephalopathy (BSE).13 Because Creutzfeldt-Jakob disease can be difficult to diagnose, it is possible that cases of sporadic and new variant Creutzfeldt-Jakob disease have occurred that were not correctly diagnosed because postmortem or neuropathological examinations were not performed. In such cases patients could have been certified as dying from another dementia or a neurodegenerative disease. If this had occurred, mortality from Creutzfeldt-Jakob disease would have been underestimated, with a corresponding increase in mortality from other dementias and neurodegenerative disorders. We investigated whether there had been any increase in mortality from dementias and neurodegenerative disorders since 1979.

Methods and results

We obtained the number of deaths from selected dementias and neurodegenerative disorders (see appendix on BMJ website) in 5 year age groups in people aged 15-64 years living in England and Wales during 1979-96. We excluded deaths in older people because the large number of deaths from senile dementia and Alzheimer's disease would have made it difficult to detect a relatively small increase in deaths from other dementias and neurodegenerative disorders and because death certification is less accurate in elderly people.4

Using the 5 year age bands, we then calculated mortality directly standardised for age for each year, using the population of England and Wales in 1979 as the standard population. Because changes in the coding of death certificates in 1984 and 1993 led to artefactual changes in mortality for neurological disorders, we present results separately for the periods 1979-83, 1984-92, and 1993-6.

We found 7796 deaths from dementias and neurodegenerative disorders for study. The commonest diagnoses were senile and presenile dementias (2074 deaths), cerebral degenerations (1924), and extra-pyramidal disorders (1797). During 1979-83 (1628 deaths) directly standardised mortality changed little, being 10.4 and 10.0 per million in 1979 and 1983 respectively (figure). Between 1984 and 1992 (4574 deaths), when the population was first known to have been exposed to BSE infected material, mortality decreased from 17.0 to 14.7 per million. During 1993-6 (1594 deaths), mortality increased from 11.1 per million in 1993 to 13.3 in 1995 before falling to 12.6 in 1996.


For statistical analysis, we split the study period into four groups (1979-83, 1984-7, 1988-92, 1993-6). Analysis of covariance was done with mortality as the dependent variable and group, year, and group=D7year interaction as factors. This enabled us to estimate separate slopes for each group. The interaction was significant (P=3D0.03), suggesting that the slope of rate against time had changed over the study period. The only positive slope was between 1993 and 1996 (slope=3D0.53, 95% confidence interval 0.03 to 1.04).


There was no evidence for any increase in mortality from dementias and neurodegenerative disorders in 15-64 year olds during 1979-92. There was, however, a small but statistically significant increase during 1993-6. Because of the limitations of routine mortality statistics, it is not possible to say with any certainty what the cause of the increase was. Our findings suggest, though, that greater surveillance of deaths from dementias and neurodegenerative disorders may be needed to avoid missing possible BSE associated deaths.

Now that the link between BSE and Creutzfeldt-Jakob disease is fairly well established it is likely that further cases of new variant Creutzfeldt-Jakob disease will be identified in the future. However, because the length and variability of the incubation period for development of new variant Creutzfeldt-Jakob disease remain unknown, only limited conclusions can be made as to the total number of BSE related deaths that will occur.5 The Office for National Statistics will therefore continue to monitor death rates from dementias. Research is also under way to identify any cases of Creutzfeldt-Jakob disease that may have been missed in people aged under 45 dying from a dementing illness during 1979-96.


We thank Dr James Ironside and Dr Richard Knight from the National CJD Surveillance Unit for their help in defining the group of diseases for study, and Professor Martin Bland for help with the statistical analysis.

Contributors: AM and MPC planned the study. PL and LK carried out the analysis. AM wrote the paper, with comments from MPC, PL, and LK. AM and MPC are the guarantors for the paper.

Funding: Department of Health and Medical Research Council.

Conflict of interest: None.