Prevalence Estimates for Dementia Fall Within Common Range
People aged 85 and over, often referred to as the oldest old (see ARF related news story on Savva et al., 2009), represent the fastest growing segment of the U.S. population, according to the U.S. Census, and are expected to grow in number from 5.8 million in 2010 to 8.7 million in 2030 (see The Next Four Decades). Although it is known that dementia and cognitive problems increase with more advanced age (see ARF related news story on Rodriguez et al., 2008), few studies have so far focused on the oldest old. “We don’t have a lot of studies in this group, so it is very important to try to define the nature of cognitive impairment,” said Francine Grodstein at Harvard Medical School, who was not involved in the current study.

Yaffe and colleagues from UCSF and five other institutions examined data from 1,299 women 85 years of age and older enrolled in the Women Cognitive Impairment Study of Exceptional Aging (WISE) to determine the prevalence of different subtypes of cognitive impairment. During a clinic visit, participants performed a battery of tests designed to determine cognitive function. Based on the results, Yaffe and colleagues calculated that 17.8 percent of participants had dementia and 23.2 percent mild cognitive impairment (MCI)—a set of problems with memory, language, and other mental functions severe enough to be noticed by others and picked up by tests, but not so severe that they interfere with daily tasks.

The prevalence of dementia calculated for the WISE cohort was lower than that identified by a number of earlier studies, including the widely cited annual Alzheimer’s Disease Facts and Figures report issued by the Alzheimer’s Association. Yaffe said part of the explanation might be that the figure for dementia reported in those studies may also have included cases of MCI. “When you look at our figure for MCI and dementia combined—41 percent—our estimate is not wildly different from other studies,” she said.

Yaffe and coauthor Laura Middleton, at Sunnybrook Health Sciences Center in Toronto, Canada, wrote in the paper that another factor that may have resulted in a lower estimate is that the mean ages of women in WISE (88 years in the 85-90 age group and 92 for those 90 years and older) is younger than those of participants in other studies. For example, in the 90+ Study—which includes men and women aged 90 and above who were residents of Leisure World, a large retirement community in Orange County, California—the mean age of all participants was 94 years. In that study, the overall prevalence of dementia among women, which was diagnosed using in-person examinations as well as telephone and informant questionnaires, was reported to be 45 percent (Corrada et al. 2008).

In addition, participants in WISE were recruited from volunteers already enrolled in the multicenter Study of Osteoporotic Fractures. It is possible that women with dementia were less likely to choose to also participate in WISE. “Dr. Yaffe’s participants were not clearly representative of the population, and it is not clear to what extent they represent the world of 88 year olds. Her prevalence of 41 percent cognitive impairment, if anything, is likely to be an underestimate, given that these individuals were research volunteers,” wrote Claudia Kawas in an e-mail to ARF. “But overall, her numbers are within the range of other studies.” Kawas, who is at the University of California at Irvine, heads the 90+ Study.

Consistent with other studies, including that by Kawas and colleagues, Yaffe and colleagues found that the prevalence of dementia in women 90 years and older was approximately double that of women aged 85 to 89 years (28.2 versus 13.9 percent). The prevalence of MCI was only slightly higher in women 90 years or older compared to those aged 85 to 89 years (24.5 vs. 22.7 percent). These trends suggest that “by the middle of the century there will be more people with cognitive impairment and dementia in [the oldest old] age group alone than we currently have at all ages combined,” Kawas wrote.

Defining Dementia Types and Subtypes
Although dementia prevalence increases with age among the oldest old, the relative proportion of the different subtypes of dementia remains the same regardless of age. In the WISE cohort, Alzheimer's disease and mixed dementia accounted for nearly 80 percent of dementia cases combined, and vascular dementia accounted for 12.1 percent.

When Yaffe and colleagues looked at MCI subtypes, they found that the most common one, in 33.9 percent of study participants, is amnestic multiple domain MCI, which affects multiple cognitive functions, including memory. It was followed by non-amnestic single domain MCI (28.9 percent) and amnestic single domain MCI (21.9 percent). Again, this distribution was similar across all age groups of oldest old women.

“Few studies have looked at MCI separately from dementia in the oldest old, and they have not determined different subtypes,” Yaffe said. This distinction is important, she adds, because different types of MCI might have different prognoses and might require different forms of treatments. There is some debate in the field about MCI, and the terminology to describe the pre-dementia stages of AD is flux (see ARF related news story). Overall, however, researchers agree that amnestic MCI represents early AD and needs to be treated accordingly, whereas non-amnestic MCI tends to have different causes and different prognoses.

“As the population ages, we are going to need accurate methodologies to diagnose people at risk for cognitive decline, as treatment may be most effective at these mild stages,” said Jennifer Manly at Columbia University Medical Center, New York. The value of Yaffe’s study, said Manly, lies in showing that standard methods used to measure MCI in younger populations are also effective in the oldest old. An important follow-up question to address, she adds, is whether oldest old individuals with different subtypes of MCI go on to develop dementia and how quickly that progression occurs.

Establishing the effectiveness of methodologies in different groups of people is an important component of epidemiological research, says Manly. In the same issue of Archives of Neurology, Manly published results of a study to determine whether telephone-based interviewing could be applied to a racially diverse group of 377 individuals residing in Northern Manhattan neighborhoods, who are enrolled in a large longitudinal study of aging and cognitive function. For the oldest old, in particular, repeated visits to study sites can become burdensome, and alternative means of assessing them would be valuable. This new study follows a previous one showing that telephone-based interviews can be effective in diagnosing dementia (Knopman et al., 2010).

Working with Richard Mayeux and colleagues at Columbia, Manly reported that telephone-based interviews, when used in combination with information gathered during in-person visits, were useful for distinguishing demented individuals from non-demented ones (including both people with normal cognitive function and MCI). They did not, however, distinguish dementia from MCI. “We are not saying you should use only telephone-based interviews to look at the prevalence of cognitive impairment, but it can be a useful tool for getting information about individuals who are unable or unwilling to be seen in person,” said Manly. Normally, these elders would be “lost” from a study, so telephone interviews can help provide an outcome measure to use in the study’s analysis.

Unlike WISE, which included primarily white women, Manly and Mayeux’s study, known as the Washington Heights-Inwood Community Aging Project, is community-based and includes a large number of African Americans and Latinos.—Laura Bonetta.

Laura Bonetta is a freelance writer in Garrett Park, Maryland.

References:
Yaffe K, Middleton LE, Lui LY, Spira AP, Stone K, Racine C, Ensrud KE, Kramer JH. Mild cognitive impairment, dementia, and their subtypes in oldest old women. Arch. Neurol. 2011 May;68(5):631-6. Abstract

Manly JJ, Schupf N, Stern Y, Brickman AM, Tang MX, Mayeux R. Telephone-based identification of mild cognitive impairment and dementia in a multicultural cohort. Arch Neurol. 2011 May;68(5):607-14. Abstract