Pathological correlates of dementia

Abstract

Authors' reply Sir—David Kay and Martin Roth overstate the intention of our report, which is not to replace the current paradigm of dementia diagnosis in the elderly, to which they made seminal contributions,1Roth M Tomlinson BE Blessed G The relationship between quantitative measures of dementia and of degenerative changes in the cerebral grey matter of elderly subjects.Proc R Soc Med. 1967; 60: 254-260PubMed Google Scholar, 2Britton PG Bergmann K Kay DW Savage RD Mental state, cognitive functioning, physical health and social class in the community aged.J Gerontol. 1967; 22: 517-521PubMed Google Scholar but to question its validity in a community setting. They seek to defend their original observations that NNTs rather than amyloid plaques are the major pathological substrate of cognitive decline in the elderly. They are essentially saying that, if the study had been done better, we would have found evidence to support current orthodoxy. We readily concede that the methods used in CFAS include compromises: dementia status was based on the geriatric mental state equivalent of clinical diagnosis, and there may be some underdiagnosis. Current work in CFAS includes more sensitive instruments for early cognitive impairment. There are no validated protocols to assess the severity of smallvessel pathology, although this is the most prevalent vascular lesion type, and we were unable to incorporate a severity calibration in the pathology protocol. Kay and Roth suggest that more sensitive instruments would detect cognitive changes in respondents with mild neocortical tangle formation who were not assigned a dementia diagnosis, and presumably also in the third of respondents who have a moderate tangle score who are apparently nondemented. However, two-fifths of the demented respondents lack neocortical tangles with no notable excess of vascular lesions over their nondemented counterparts. We do not assert that the lesions of neurodegenerative diseases and vascular pathology in the ageing brain are unimportant. We confirm them as common pathologies in an elderly community-based cohort, showing statistical association with dementia status. We entirely agree that the lesions of Alzheimer-type pathology, Lewy bodies, and vascular pathology are bad things for the human brain, and that pharmacological interventions, taking proper account of multiple pathologies, will bring benefit. Undoubtedly we could select a cohort of cases from the study who fulfill diagnostic criteria for Alzheimer's disease. Our main point is that, even if we used this approach for all accepted dementia diagnoses (eg, multi-infarct dementia and dementia with Lewy bodies), we would be left with a sizeable segment of the cohort in whom cognitive status, including inappropriate intellectual decline and preserved cognition, was not readily explicable from the pathological burden. This finding is exposed in a community-based sample, whereas such cases are normally excluded in the standard secondary referral cohorts from whom much understanding of dementia pathology is based. A categorical, threshold-based, clinicopathological model of dementia substrates would need to perform better than this to impress us of its validity. This is the basis on which we suggest that the impact of mixed pathologies, and of individual variation in the functional impact of mild to moderate pathological burdens in the ageing brain, need to be synthesised into our understanding of the relation between pathological findings and cognitive decline. Pathological correlates of dementiaThe Neuropathology Group of the Medical Research Council Cognitive Function and Ageing Study (MRC CFAS), in their report on pathological correlates of late-onset dementia, derived from a community-based population study (Jan 20, 2001, p 169)1 reach conclusions that seem to downgrade the importance of specific neuropathological criteria in the dementias of the elderly. Full-Text PDF