Data from 11,926 older twins (aged 65+) has found measurable cognitive impairment in 25% of them and subjective cognitive impairment in a further 39%, meaning that 64% of these older adults were experiencing some sort of cognitive impairment.
Although subjective impairment is not of sufficient magnitude to register on our measurement tools, that doesn’t mean that people’s memory complaints should be dismissed. It is likely, given the relative crudity of standard tests, that people are going to be aware of cognitive problems before they grow large enough to be measurable. Moreover, when individuals are of high intelligence or well-educated, standard tests can be insufficiently demanding. [Basically, subjective impairment can be thought of as a step before objective impairment, which itself is a step before mild cognitive impairment (MCI is a formal diagnosis, not simply a descriptive title), the precursor to Alzheimer’s. Note that I am calling these “steps” as a way of describing a continuum, not an inevitable process. None of these steps means that you will inevitably pass to the next step, but each later step will be preceded by the earlier steps.]
Those with subjective complaints were younger, more educated, more likely to be married, and to have higher socio-economic status, compared to those with objective impairment — supporting the idea that these factors provide some protection against cognitive decline.
The use of twins reveals that environment is more important than genes in determining whether you develop cognitive impairment in old age. For objective cognitive impairment, identical twins had a concordance rate of 52% compared to 50% in non-identical same-sex twins and 29% in non-identical different-gender twins. For subjective impairment, the rates were 63%, 63%, and 42%, respectively.
National variation in MCI prevalence
Another very large study, involving 15,376 older adults (65+), has explored the prevalence of amnestic MCI in low- and middle-income countries: Cuba, Dominican Republic, Peru, Mexico, Venezuela, Puerto Rico, China, and India. Differences between countries were marked, with only 0.6% of older adults in China having MCI compared to 4.6% in India (Cuba 1.5%, Dominican Republic 1.3%, Peru 2.6%, Mexico 2.8%, Venezuela 1%, Puerto Rico 3% — note that I have selected the numbers after they were standardized for age, gender, and education, but the raw numbers are not greatly different).
Studies to date have focused mainly on European and North American populations, and have provided prevalence estimates ranging from 2.1%-11.5%, generally hovering around 3-5% (for example, Finland 5.3%, Italy 4.9%, Japan 4.9%, the US 6% — but note South Korea 9.7% and Malaysia 15.4%).
What is clear is that there is considerable regional variation.
Interestingly, considering their importance in Western countries, the effects of both age and education on prevalence of aMCI were negligible. Granted that age and education norms were used in the diagnosis, this is still curious. It may be that there was less variance in educational level in these populations. Socioeconomic status was, however, a factor.
Participants were also tested on the 12-item WHO disability assessment schedule (WHODAS-12), which assesses five activity-limitation domains (communication, physical mobility, self-care, interpersonal interaction, life activities and social participation). MCI was found to be significantly associated with disability in Peru, India, and the Dominican Republic (but negatively associated in China). Depression (informant-rated) was also only associated with MCI in some countries.
All of this, I feel, emphasizes the situational variables that determine whether an individual will develop cognitive impairment.
Caracciolo B, Gatz M, Xu W, Pedersen NL, Fratiglioni L. 2012. Differential Distribution of Subjective and Objective Cognitive Impairment in the Population: A Nation-Wide Twin-Study. Journal of Alzheimer's Disease, 29(2), 393-403.
 Prevalence, Distribution, and Impact of Mild Cognitive Impairment in Latin America, China, and India: A 10/66 Population-Based Study. PLoS Med [Internet]. 2012 ;9(2):e1001170 - e1001170. Available from: http://dx.doi.org/10.1371/journal.pmed.1001170
Full text available at http://www.plosmedicine.org/article/info%3Adoi%2F10.1371%2Fjournal.pmed....