Analysis Alzheimer’s disease

What do we mean by Alzheimer’s disease?

BMJ 2010; 341 doi: (Published 12 October 2010) Cite this as: BMJ 2010;341:c4670
  1. Marcus Richards, programme leader, Medical Research Council1,
  2. Carol Brayne, professor2
  1. 1MRC Unit for Lifelong Health and Ageing, London
  2. 2Institute of Public Health, University of Cambridge
  1. m.richards{at}
  • Accepted 25 July 2010

In its most common late onset form, the term Alzheimer’s disease is unlikely to refer to a discreet neuropathological entity, but to a diffuse clinical syndrome that represents the gradual accumulation of multiple pathologies, which arise from risk factors over the course of life

In the United Kingdom more than 700 000 people are estimated to have dementia. It costs the UK economy £23bn (€28bn; $35bn) per year1—more per patient than the median UK salary, and it is likely to become more prevalent as life expectancy increases. No effective treatments are available for most people with these disorders.

In 2009 the Department of Health National Dementia Strategy proposed several key objectives for relevant services in the UK, including improving public and professional awareness; achieving good quality early diagnosis and intervention; and developing and improving structured personal and peer support services.2 Alzheimer’s disease is the most common type of dementia. It is therefore important to look at what the diagnosis of Alzheimer’s disease represents, not least because the large volume of research on its pathophysiology has not led to any effective preventive breakthrough in the population, unlike cardiovascular disease and stroke. This article focuses on late onset Alzheimer’s disease that occurs after age 65 years—the most common expression of this syndrome—rather than early-onset Alzheimer’s disease, which is a comparatively rare, mostly autosomal dominant familial disease.

Challenges to the pathology-led view

Since the 19th century, when neuritic plaques and neurofibrillary tangles were identified as features of senile dementia, Alzheimer’s disease has been framed as a neuropathological entity—driven for example by amyloidosis or accumulation of oligomers and leading to neuronal death. Yet the diagnosis of Alzheimer’s disease in the living patients is made on the basis of clinical information that only rarely includes neuropathology. Suggestions have been made for the clinical formulation to take account of biomarkers, including evidence from brain imaging and cerebrospinal fluid assay.3

Such investigations, however, are unlikely to be feasible at community level; in any case the diagnosis is confirmed or ruled out at necropsy, as reflected by the labels “definite,” “probable,” and “possible” specified by standard diagnostic criteria.4

Two challenges have been made to the pathology led model of Alzheimer’s disease. Firstly, dissociation, or lack of a consistent relationship, between the clinical syndrome and the classical pathological features of Alzheimer’s disease is not unusual. Although these brain features are rarely absent in patients with a clinical diagnosis of Alzheimer’s disease, plaques and tangles are not pathognomonic, and some people with postmortem evidence of these were cognitively spared at death.5 6 7 This led to the notion of asymptomatic Alzheimer’s disease (ASYMAD),8 and the possibility of different people with the same degree of hallmark pathologies showing different clinical profiles.

Secondly, although the construct of ASYMAD preserves the supremacy of pathology, the pathology is almost certainly more heterogeneous than traditionally believed. Nearly 46% of brains of people with a clinical diagnosis of probable Alzheimer’s disease who were resident in the community, and therefore were probably relatively mild cases, had mixed pathologies. These were most commonly plaques and tangles plus cerebrovascular disease,9 and evidence is accumulating that the latter plays a causal role in Alzheimer’s disease.10

Ironically, Alois Alzheimer described atherosclerotic changes in the syndrome to which his name was given, although this feature was then underplayed, and eventually became an exclusion criterion.11 This exclusion is still proposed,3 although cerebrovascular disease is recognised as a comorbid condition by the draft fifth edition of the Diagnostic and Statistical Manual of Mental Disorders guidelines.12

The contribution of cerebrovascular disease to Alzheimer’s disease is consistent with evidence that the gradual accumulation of cardiovascular risk factors—such as obesity, hypertension, and hypercholesterolaemia—from midlife or earlier is a modifiable risk for this syndrome (at particular ages although not, perhaps, in the whole lifetime). Also, evidence shows that health related lifestyle risk factors—such as smoking, heavy drinking, low fruit and vegetable consumption, and low physical activity—for all the major chronic disorders closely associated with ageing tend to cluster. This perspective is acknowledged by the Department of Health National Dementia Strategy, and summed up by their device that “what’s good for your heart is also good for your head.” Evidence from a large representative population study indicated that plaques, tangles, small vessel disease, and other vascular pathology all contribute to dementia at death.13


Dementia as a diffuse multiform syndrome

It is timely to interrogate the term Alzheimer’s disease. In older age groups Alzheimer’s disease seems to be a diffuse clinical syndrome representing the gradual accumulation of multiple pathologies, arising from multiple interlocking risk factors over the life course. The term Alzheimer’s syndrome seems more appropriate, and might represent a large range of clinical profiles at different ages, with different functional consequences. What does this mean for diagnosis, for doctors who look after people with this syndrome, for pathologists, for researchers, and for patients?

Implications for practitioners

The National Audit Office estimates that only around one third of people with some form of dementia are clinically diagnosed as such.14 A core aim of the Department of Health National Dementia Strategy is early detection with a view to early intervention. Using a method of service delivery such as the Croydon Memory Service Model could lead to a modest but cost effective increase in quality adjusted years of life through reduced institutionalisation.15 This is an evaluation of a single service, however, and the generalisability of its effectiveness is unknown.

Furthermore, a wider range of challenges cannot be underestimated. For the practitioner and the pathologist there is unlikely to be a straightforward correspondence between the accumulation of this pathology and the psychological gradient that runs from casual forgetfulness, through mild cognitive impairment, to frank dementia. For example, 33% of brains of people classified with mild cognitive impairment who were resident in the community had macroscopic infarcts, but 54% had pathologically diagnosed Alzheimer’s disease.9 The formulation of Alzheimer’s disease as a yes-no clinical diagnostic category—a process that has long been challenged16 17—makes this mapping unnecessarily difficult. It is not helped by the proliferation of subdiagnostic categories of mild cognitive impairment such as early, late, amnestic, non-amnestic, single or multiple domain, which can be a further source of anxiety to patients and their families. We cannot even say where the symptom gradient begins because risk of Alzheimer’s disease is also increased by cognitive disadvantage early in life.18 We do not know the odds that casual forgetfulness in any individual will progress to cognitive decline of sufficient severity to impair quality of life.19 Evidence from community dwelling samples is mixed, and does not necessarily help practitioners to identify individuals at risk. These problems are further compounded by the way in which the expression of cognitive problems varies by education, health, cultural norms, and even by the pathways through which patients with different profiles are clinically reviewed. For example, an older person with memory or behavioural problems but otherwise reasonable health may see an old age psychiatrist; whereas if such a person has complex physical morbidity they may see a geriatrician, or, if there is no acute illness and the individual is moderately coping, a primary care physician. Some people in residential care who are frail and cognitively impaired might not be assessed for dementia at all; and it is unclear whether there would be any benefit from so doing. Should we institute routine screening of the population for cognitive impairment? This is not recommended by NICE or its equivalent US institution; it is likely to overwhelm services with false positive or incidental findings20; it could cause unnecessary distress, perhaps to an older couple who cannot see any purpose in a diagnosis when they are able to carry on as an effective unit; and it might have other unintended consequences—for example, on the right to remain in an institution that is not registered for dementia care.

Implications for prevention and quality of life

If Alzheimer’s disease is a diffuse clinical syndrome, there is unlikely to be a therapeutic silver bullet, notwithstanding the dominant research endeavour towards finding effective drug treatments. But if “what is good for our hearts is also good for our heads,” should we be equally focusing on reducing cardiovascular risk?

Yes of course; this is a public health goal in its own right. But an integrative public health strategy should also recognise that risk of Alzheimer’s disease evolves throughout life.18 This approach would target risk of insults to the brain beginning in the uterus, during childhood poverty and its consequences such as overcrowding and inadequate nutrition, through poor schooling, risky lifestyle; and through effects of low attainment such as hazardous occupational exposures and chronic stress. Although the life course process is also associated with cumulative risk of chronic physical diseases that are linked to Alzheimer’s disease, such an outlook goes beyond the medical dimension as an approach to labelling and intervention in Alzheimer’s disease.

Paradoxically, those who escape premature death are likely to be at increased risk of Alzheimer’s disease through survival. In the same population study used to estimate attributable risk of dementia from different neuropathologies,13 dementia at death was present in 6% of those who died in their mid to late 60s compared with 58% who died aged 95 years and older.21 The most important strategy in this context for those in later life is therefore one of compression of morbidity, that is, to maximise quality adjusted life years by delaying the onset of dementia towards age at death. We invoked a life course approach to reducing risk of Alzheimer’s disease, but here make special mention of childhood education and lifelong learning. These are often thought of as providing a buffer that masks the clinical expression of Alzheimer pathology. They may also facilitate compression of morbidity by a combination of increasing survival through reduction of premature mortality, maintaining self efficacy in those who do survive, and preserving the richness of mental life throughout this time of survival—for example, by increasing knowledge.22

The view that Alzheimer’s disease is a clinical and a pathological entity has a long history11; but it was not until the late 1960s that the potential dissociation between these two aspects began to be appreciated; and it was not for another 30 years that the relationship between cognitive function, activities of daily living, and neuropathology would be systematically investigated in older people in the community.7 No straightforward correspondence exists between higher mental function and the burden of lesions in the ageing brain. If this shifts the focus away from detailed diagnostic classification made on the basis of assumed clinical-pathological correlation and towards a global pragmatic approach to the needs of patients and carers, and to modifiable lifetime risk factors, then the apparent loss of scientific precision is a gain to clinical practice.


Cite this as: BMJ 2010;341:c4670


  • Acknowledgments: MR and CB are funded by the UK Medical Research Council. CB is additionally funded by the NHS.

  • Contributors and sources: MR is a programme leader at the Medical Research Council Unit for Lifelong Health and Ageing, and a reader in cognitive epidemiology at the department of epidemiology and public health, University College London. He began his postdoctoral research career in the late 1980s at Columbia University in New York, in the epidemiology and neuropsychology of neurodegenerative diseases of ageing. For the past 15 years he has been developing a life course approach to cognitive ageing, principally on the basis of British birth cohorts. CB is professor of public health medicine at the department of public health and primary care, Cambridge University, director of the Institute of Public Health, and is an epidemiologist and a public health physician. For many years she has led population based research into neurocognitive disorders, healthy ageing, and neuropsychiatric epidemiology. Both authors contributed equally to the development of this paper, using published material as cited.

  • MR will act as guarantor for the paper.

  • Both authors have completed the Unified Competing Interest form at (available on request from the corresponding author) and declare that (1) MR and CB have support from the MRC for the submitted work; (2) MR and CB have no relationships with companies that might have an interest in the submitted work in the previous 3 years; (3) their spouses, partners, or children have no financial relationships that may be relevant to the submitted work; and (4) MR and CB have no non-financial interests that may be relevant to the submitted work.