The world’s population is ageing, as improving living standards have increased life expectancy across the globe. The most affluent countries have the longest average life expectancy, and the greatest proportions of older adults. However, population ageing is occurring faster in the low-and middle-income (‘developing’) countries than in the wealthier (‘developed’) countries. It is determined by both birth rates and death rates. Populations with high fertility tend to have low proportions of older people, and vice versa. The ‘demographic transition’ is the process in which, over time, a high-fertility-high-mortality population becomes a low-fertility-low-mortality population, and its average age increases. As people live longer, the public health concern becomes whether their years of extended life are characterized by relative health and independence or by ill-health, disability and dependence on others. Thus, we measure not only overall life expectancy but healthy life expectancy and disability rates.

In affluent countries, disability rates appear to be decreasing as the population ages. In low-and middle-income countries, it seems likely that disability rates will climb as people live longer with chronic disease. The least affluent countries are the least prepared to deal with the challenges of population ageing, given the rapidity with which the demographic transition is occurring in their societies. However, one factor in their favour is the elderly support ratio, which is the number of elderly people for every 100 younger adults in a given population, with younger individuals assumed to be directly or indirectly supporting older ones. This ratio is increasing rapidly in the affluent countries, with greater numbers of older individuals needing support from a shrinking number of younger persons. In the less affluent countries, this ratio is expected to change quite slowly³.

Dementia is a major cause of functional disability, dependence and mortality in the elderly. It therefore reduces both overall life expectancy and active (disability-free) life expectancy, although the extent to which it does so might vary across populations. Dementia is also responsible for increasing use and cost of health services, particularly long-term-institutional care, in countries where such institutions are available and culturally acceptable. The cost to a given society of caring for its members with dementia depends on its age structure, the numbers and proportions of affected individuals, their length of survival and degree of dependency, the availability of appropriate goods and services, and who bears the costs of these items. These direct costs, as well as indirect costs (e.g. loss of income by the patient and family caregiver) will vary according to the affluence and age structure of a given society.

At the time of this writing, the most widely employed diagnostic criteria for dementia at the syndrome level are those of the Tenth Revision of the International Classification of Diseases (ICD-10)¹ and the Fourth Edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV-TR)². Over the years, studies in different countries have used different versions of these criteria. Variations in definitions and criteria can potentially raise and lower estimates of prevalence and incidence across studies and populations. One study demonstrated that the proportion of individuals classified as having dementia varied from 3% using the ICD-10 criteria to 29% using the DSM-III criteria, within the same population⁴.

According to current definitions, dementia is a syndrome of acquired, chronic cognitive impairment, occurring in clear consciousness, sufficient to interfere with social and occupational functioning, and characterized by impairment in at least two cognitive domains. For the past few decades, the definition has further required that memory be one of the impaired domains², but this requirement may not persist in future revisions of diagnostic criteria. For example, it may be possible that an individual with significant impairments in executive and language functioning can be diagnosed as having dementia, even without memory impairment.

Further, there is currently no standard diagnostic entity reflecting a minor level of cognitive decline that is not accompanied by functional impairment, thus falling short of the threshold for being called dementia. This situation too may change as knowledge evolves and preclinical detection of dementing illnesses becomes a possibility. Besides the functional dimension of the diagnosis, the standard neuropsychological criteria for dementia typically require test performance in two cognitive domains to be at least two standard deviations below the mean for the individual’s peer group, as defined by age, education, linguistic/ethnic group and sometimes gender. However, appropriate norms are not always available. Thus, the issue of ‘caseness’, or what makes an individual a true case, has yet to be universally defined, and will become harder as the demand increases for earlier detection of the dementia syndrome.

Alzheimer’s disease (AD) is a primary neurodegenerative disease which appears to be the single most common cause of the dementia syndrome in older adults, in most populations which have been studied worldwide. The next single most frequent cause of dementia is cerebrovascular disease; either cortical infarcts or white matter disease, or both, can lead to significant cognitive impairment and decline. Diagnostic criteria for vascular dementia and cognitive impairment have varied as to whether the mere presence of cerebrovascular disease with dementia is sufficient, or whether evidence from neuroimaging is also required, for the diagnosis²‘⁵. However, mixed Alzheimer’s and vascular pathology is more common than pure vascular disease or pure degenerative disease, as borne out by the few community studies that have been able to conduct autopsies⁶. Although prominent in specialty clinical settings, dementias related to other primary brain disorders such as frontotemporal lobar degeneration, Lewy body disease and Parkinson’s disease appear relatively infrequent in the community at large; population-based estimates are scarce. Opinion is moving towards accepting co-morbidity and multiple causes in a given individual with dementia, rather than selecting one subtype over another in the presence of mixed aetiology.

The overall prevalence of dementia, based on many studies in affluent countries worldwide, is in the range of 5-10% of individuals aged 65 years and older. One meta-analysis showed a doubling of dementia prevalence with every 5-year increase in age; for AD the doubling occurred with every 4.5 years and for vascular dementia every 5.3 years⁷. Population studies typically do not employ neuroimaging or neuropathology for diagnosis, but rather use operationalized standard clinical diagnostic criteria. In these studies, AD appears to account for two-thirds to three-quarters of observed cases. Previous reports of vascular dementia being more prevalent than degenerative dementia in some Asian populations may have been the result of varying diagnostic criteria. However, prevalence has not been established for the newer, broader entity called vascular cognitive impairment (VCI) which encompasses all levels of cognitive decline from mild deficits to dementia. The prevalence of dementia of Parkinson’s disease has been estimated at about 0.5% of the population aged 65 years and older⁸.

Although incidence rates vary, one meta-analysis showed rates of AD ranging from 0.33% per year for those aged 65-74 to 8.68% for those aged 95+ years⁹. Evidence is inconclusive as to whether incidence continues to increase after age 90. In a study including MRI scans for the diagnosis, the age-adjusted incidence rate for AD was 19.2, and for vascular dementia 14.6, per 1000 person-years¹⁰. While many studies have shown higher dementia prevalence among women, many incidence studies have not shown a gender difference. If incidence is indeed the same in men and women, gender difference in prevalence may be due to longer survival in women. There are few population-based data on early-onset dementias.

Based on a handful of studies from Asia, Africa, Central and South America, and Eastern Europe, prevalence in low-and middle-income countries is much lower than in the affluent countries. It is, however, increasing as their populations age, and seems to be divided between Alzheimer’s and vascular subtypes as in the rest of the world¹¹. The prevalence difference is not solely a function of life expectancy. Opinion is divided as to how much of the discrepancy is due to mis-classification, given the difficulties in measurement discussed earlier, and how much reflects true differences in disease occurrence. However, all studies report an exponential increase in both prevalence and incidence with age. With the rapid population ageing of the less affluent countries, among which India and China have very large populations, even a lower proportion affected by dementia will reflect a larger number of individual cases in these countries than in the wealthier countries¹².