Is It Time to Retire the Term “Dementia”?

The term dementia (from Latin demens, meaning “without mind”), was incorporated into the European vernaculars in the 17th and 18th centuries.¹ By the early 18th century, it had acquired a medical connotation, and it was included in Blancard's Physical Dictionary² in 1726. Since then, the term has had an interesting history with many vicissitudes in its meanings and usage.¹ By the late 19th century, the primacy of cognitive impairment as the defining feature of dementia was generally accepted. The major debate of this time was between the vascular and parenchymal mechanisms of etiology, on which opinion has also shown many fluctuations over the last century. By early 20th century, the association with old age had become firmly established and the descriptor “senile dementia” was widely applied.³

When Alois Alzheimer described the neuropathology of dementia in a 51-year-old woman in 1906,⁴ the presence of neurofibrils⁵ and plaques⁶ in the brains of senile dementia patients had already been described. Alzheimer did not regard his description as that of a new disease,⁷ but rather of an atypical early presentation of senile dementia. Kraepelin probably did consider it to be a new disease when he called it “Alzheimer's disease” in the 8th edition of his Handbook.⁸ Senile and presenile dementia continued to be regarded as distinct disorders until the 1960s and '70s, when comprehensive neuropathological studies demonstrated that the brain changes described by Alzheimer were a common cause of cognitive decline in the elderly.^9–11 Although the heterogeneity of the concept of dementia was not in doubt, other developments in the later 19th and early 20th century highlighted this fact further. Pick's¹² description of lobar atrophy, the reports of a spongiform “pseudosclerosis” by Creutzfeldt¹³ and Jakob¹⁴ (later demonstrated to be transmissible¹⁵), and other similar developments broadened the specific categories encompassed by the term dementia.

The description of specific diseases under the rubric of dementia has continued into late 20th century.¹⁶ The other major development has been an increase in prevalence of Alzheimer's disease (AD) because of an aging population, with the result that AD has become the prototypical dementia syndrome. Indeed, the most commonly used definitions of dementia, the DSM-IV¹⁷ and ICD-10¹⁸ definitions, reflect this. These criteria emphasize memory impairment in the diagnosis of dementia, along with impairment in other cognitive domains. The main thrust of these definitions is to distinguish dementia from a state of altered consciousness (delirium) and restricted cognitive impairments such as aphasia or amnestic syndrome.

The emphasis on memory impairment is understandable in the criteria for AD, but it is restrictive when applied to dementia from other causes, such as vascular dementia (VaD) and frontotemporal dementia (FTD). Memory impairment is not of primary importance in early VaD, and cerebrovascular disease can produce impairment in several cognitive domains while affecting memory only mildly or not at all.¹⁹ This is also true for FTD, in which frontal-executive and language disturbance may be quite prominent early in the course. The application of the criteria for dementia results in the situation that when patients are diagnosed with VaD they are at a relatively advanced stage of their illness, negating attempts to intervene early.²⁰ Despite recognition of this limitation, only one set of criteria for VaD has had the necessary requirement of memory impairment removed from its definition.²¹ The memory impairment criterion also delays the diagnosis of FTD or other degenerative brain conditions in which deficits are more localized.

The definition of dementia does some disservice to AD itself. Memory loss is an early feature of AD and may be present for some time before other cognitive deficits become apparent.²² Going strictly by published criteria, a diagnosis of amnestic syndrome rather than dementia is warranted at this stage of the disease. Even the nature of memory impairment required for a diagnosis has had its vicissitudes: DSM-III required disturbance in short-term memory, whereas DSM-III-R and DSM-IV require both short-term and long-term memory dysfunction. As drugs for the treatment and prevention of AD become available, it will become important to recognize the disease early, prior to the development of the dementia syndrome. With recent developments in technology, and findings such as early loss of medial temporal lobe volume and the significance of apolipoprotein genes, it may even be possible to diagnosis AD in the preclinical stage. It can therefore be argued that the term dementia does not further the descriptive power of a diagnosis of AD, which is already established as a disease with a recognized pattern of cognitive and noncognitive symptoms, etiology, and course.

The definition of dementia has created uncertainty for some other neuropsychiatric disorders. A useful example is alcohol-related brain damage. Most investigators agree that alcohol is an important cause of the amnestic syndrome and frontal-executive dysfunction, but there has been considerable debate as to whether alcohol causes dementia. Whereas some authors have suggested that alcohol-induced dementia is relatively common,²³ others have argued against its acceptance as an entity until a definitive neuropathological basis is established.²⁴ The debate distracts attention from the fact that alcohol is responsible for much cognitive dysfunction in society, and it has hindered the conduct of treatment studies in this field. Traumatic brain damage is another instance of the difficulties in applying the dementia concept. Trauma is a common cause of cognitive impairment in young people, but most physicians are uncomfortable in using the term dementia to describe their disability even when this is appropriate by strict definition, and this diagnosis is usually restricted to those with severe impairment.

The definition of dementia is limited by its other features as well. It lacks operationalized definitions for the criteria, introducing considerable subjectivity to the definition. There is no consensus on what constitutes memory “impairment,” “disturbance in executive functioning,” or “significant impairment in social or occupational functioning.” A 70-year-old man with cognitive impairment who is still employed will receive a diagnosis of dementia, whereas his retired counterpart who is similarly impaired may escape the diagnosis. Admittedly, the consequences of the diagnosis are different in the two men, but it does serve to highlight the subjective nature of this diagnosis, an inappropriate situation for a disorder rooted in biological dysfunction. The “impaired functioning” criterion also creates a threshold effect on the diagnosis. This may be one reason why a premorbidly high intellectual level is suggested to be a protective factor against dementia: it is likely to widen the distance between the premorbid level of functioning and the level of functional impairment that sets the threshold for a dementia diagnosis. The lack of operational criteria has compromised the reliability of the diagnosis of dementia between raters and across sets of criteria,^25,26 without which the validity of the concept can become elusive. Two recent studies of the reliability and validity of diagnostic criteria for dementia highlight this point. The first study examined 1,879 subjects age 65 years and over from the Canadian Study of Health and Aging²⁵ and found that the proportion of subjects receiving a diagnosis of dementia varied from 3.1% using ICD-10 criteria to 29.1% using DSM-III criteria, a 10-fold difference! An investigation of the reliability and validity of the diagnosis of VaD²⁶ reported that the concordance between different pairs of criteria varied from 18% to 48%, and much of the variation was due to differences in definitions of dementia between criteria sets—even when these sets were being used by the same group of investigators.

The acceptance of an acute onset and reversibility within definitions of dementia such as in DSM-III was an attempt to keep the definition broader than what AD would entail. This change has been successful among neuropsychiatrists and other concerned specialists, but the general medical community, much like the lay public, continues to use the term dementia to imply chronicity and irreversibility. Evidence of progressive deterioration is required by at least one commonly used criteria set for dementia, the CAMDEX.²⁷ This requirement is arguably the result of AD having become synonymous with dementia, an unacceptable situation for both concepts. Even the implicit understanding that the diagnosis of dementia is underpinned by a disturbance in brain function evident on neurological investigations and/or neuropathological examination has become a liability. It has been cogently argued that depression results in a syndrome of dementia, and there is neuropsychological, neuroimaging, and neurochemical support for the argument.²⁸ Yet depressive dementia remains excluded from most definitions of dementia, wherein the cognitive syndrome of depression, although rooted in neurobiological dysfunction, is labeled “pseudodementia.” Such a distinction is a disincentive to the search for overlapping pathophysiological mechanisms in the cognitive impairment of depression and dementia, and it further underscores the importance of clarity in the terminology.

Not only is the term dementia limiting in its scope, it has a pejorative connotation in its general usage. The Concise Oxford Dictionary defines demented as “driven mad, crazy,”²⁹ and the term is commonly used as a form of insult. Its use by the medical profession arouses great anxiety in the patients, evoking images of extreme disability and dependence. The diagnosis has serious consequences in terms of a person's competence in employment, legal transactions, driving, and even the conduct of ordinary activities of living, far beyond what might be immediately suggested. Third-party disbursements depend on the diagnosis, and a dementia diagnosis may make it difficult for the patient to obtain insurance or make a valid will. The diagnosis of dementia may, therefore, become a cross for the patient and the family to bear.

The conclusion to be drawn from the above arguments is that the limitations and disadvantages of the use of dementia outweigh any usefulness that remains in the term, and it is time that we abandoned it. Admittedly, since it has been used for such a long period and has been the subject of a vast literature and mythology, it will leave a major void. How will this be filled? Some inheritors of the legacy are obvious. AD, Huntington's disease, Creutzfeldt-Jakob disease, frontotemporal degeneration, and cortical Lewy body disease are established as diseases with well-characterized clinical profiles and neuropathological findings such that a syndromal approach to them is in fact unnecessary. A diagnosis of AD usefully encapsulates the cognitive and noncognitive features of the disorder and its pathology and prognosis much better than “dementia” does. The term Huntington's disease is inclusive of the cognitive, motor, and behavioral symptoms that characterize the disorder of which dementia is but one aspect. In fact, the descriptions of mood, thought, perceptual, and behavioral disturbances in the syndrome of dementia are anomalous, since the syndrome itself is being defined by cognitive impairment. On the other hand, diagnoses like AD or Huntington's disease do not have such constraints imposed upon them. The noncognitive features are of major importance in many of these disorders, and the move away from the use of dementia is likely to facilitate this recognition.

The proposal to abandon the term dementia is of course quite radical and is likely to encounter considerable inertia and some active resistance. The term has wide acceptance that goes beyond the medical fraternity to public health, legislation, community affairs, and lay usage. There is a large and growing industry dependent on it. The nonmedical groups, less likely to be versant with the limitations discussed above, may feel more disadvantaged by a change in terminology. How does one conceptualize, they would argue, the “coming epidemic of dementia” or “a dementia-specific nursing home,” for instance? For a public health legislator or a lay caregiver, there is much commonality in the dementias to warrant continued retention of the term. This, however, is largely because of their exclusive concern with the severe end of the spectrum at which many of the differences between the various cognitive disorders disappear. Many clinicians share this viewpoint, along with a therapeutic pessimism that makes a refined approach seem like wasted effort. Although these arguments cannot be dismissed, the situation seems to be changing with the promise of specific remedies in various dementing disorders. Moreover, care for the cognitively impaired should be informed by the nature of the cognitive and noncognitive impairments in an individual, and the diagnosis of dementia does not advance this understanding in most cases. Instead of knowing how many individuals in the community have “dementia,” a legislator is probably more interested in the number of individuals who would need hostel or nursing home placement because of cognitive impairment.

Another counterpoint to the theme of this paper is the overlap of different etiologies in the dementia syndrome. It is not unusual for vascular and Alzheimer-type changes to coexist. AD and frontotemporal dementia or dementia with Lewy bodies may be quite indistinguishable clinically, especially among elderly persons. The term dementia acknowledges this “natural” commonality of the presentation. However, the alternative of cognitive disorder does the same and yet presses harder for the contributions made by the different etiologies.

In summary, the concept of vascular dementia makes us reencounter the fact that cognitive impairment is often a continuous variable on which we impose a categorical construct when we diagnose “dementia.” The arbitrariness of this imposition is obvious and is well highlighted by the low reliability and validity of dementia diagnosis described above. Terms such as cognitive impairment or cognitive disorder^29,30 have been suggested to broaden the concept and deemphasize the later stages when severe dysfunction has resulted. Impairment is appropriate as a term for a continuous variable, but since clinicians prefer to deal with categories, cognitive disorder may be the appropriate appellation, with the prefix vascular, HIV-related, posttraumatic, etc., providing the necessary etiological descriptor. For instance, vascular cognitive disorder overcomes many of the limitations of vascular dementia.³⁰ This change of terminology would not obviate the need for operational criteria, but a statistical or other objective approach is more easily applied to a newly developed “cognitive disorder” diagnosis than to the history-laden concept of dementia. It would be a bold move and one that is likely to meet much resistance, but psychiatry has done this before with “neurosis” without having to look back, and we will have to continue to make such decisions in the future.

ACKNOWLEDGMENTS

The author is grateful to Dr. Julian Trollor for his insightful comments on an earlier draft.