The new work is part of the U.K. Medical Research Council’s prospective Cognitive Function and Ageing Study (CFAS). Led by senior author Carol Brayne, University of Cambridge, it appeared in the May 28 New England Journal of Medicine. The CFAS has followed 18,000 people since the 1980s, using cognitive and diagnostic evaluations to capture incident AD. Its advantage over similar studies is its size and that patients were picked at random from truly population-based health registries, suggested Ince. “If you ask for volunteers or if you use a secondary referral population then you are going to introduce bias,” he said. “It is a rare and outstanding approach to have a population-based study of cognition that extends into very late life and shows neuropathological features at the time of death,” write Douglas Ewbank and Steven Arnold, University of Pennsylvania, Philadelphia, in an accompanying editorial.

An autopsy collection program allowed joint first authors George Savva and Stephen Wharton, at Cambridge and Sheffield, respectively, and their colleagues to examine 456 brains for neuropathological signs of AD, including diffuse and neuritic plaques and neurofibrillary tangles. They found that the association between dementia (diagnosed based on an algorithm equivalent to the DSM-III-R criteria) and both neuritic plaques and tangles was strong at age 75, but three- to fourfold weaker at age 95. Between these two ages, the prevalence of neuritic plaques in the neocortex and the hippocampus actually decreased in people with dementia and increased in people without dementia. “We find that AD pathology is a much better predictor in younger old people, those toward the 70-75 end of the age spectrum, whereas by 85 and older the severity of Alzheimer pathology in demented people becomes virtually indistinguishable from severity of pathology in non-demented people,” said Ince.

This is not the first study to report that the oldest old buck the trend seen in younger cohorts. The Honolulu Asia Aging Study, The Adult Changes in Thought Study, the Nun Study of the School Sisters of Notre Dame, the Religious Order Study, the 90+ Study, and the Baltimore Longitudinal Aging Study all have similar findings, said Ince. He noted, though, that while these are community based, the “community is never the whole community. Those studies are not saying anything widely different [from ours], but because they do not have absolutely kosher population-based selections, we can’t be certain there is no bias.” The smaller Einstein Aging Study also reported that dementia of unknown etiology increases dramatically in nonagenerians (see Crystal et al., 2000). “The [CFAS] data are the same kind of data that we have, and that most people have,” agreed Kawas. She added that these findings deserve more attention. “Within a few decades, the oldest old alone will account for more dementia than we currently have in the United States at all ages,” said Kawas.

The findings raise important issues for clinical practice, suggested Ince. “One implication is that biomarkers may not be as directly related to illness as we’d be happiest with,” he said. “We are entering an era where there are multiple potential disease-modifying therapies around. Everyone knows that, ideally, those therapies will be used as early as possible. We are saying that there needs to be careful thinking and caution about applying biomarkers for those pathologies in the oldest old.”

The study also raises questions about etiology. If typical plaque and tangle pathology does not explain dementia in some of the oldest old, then what does? One possibility is brain atrophy. “We found that cerebral atrophy, both moderate and severe, distinguish the demented from the non-demented group at all ages. Our work, to a certain extent, is validating the potential usefulness of atrophy as a biomarker for cognitive failure,” said Ince (see ARF related news story). Synapse loss may also distinguish people with and without dementia among the oldest old (see Head et al., 2007).

The other side of the etiology coin is why some long-lived people who have plaques and tangles apparently remain free of dementia? This might relate to the concept of cognitive reserve, which posits that some people are better able to maintain function in the face of moderate to severe pathology. “There is a major gap between things that you can see in brain pathologically and how the brain is actually working. We need to stop thinking that pathology is such a direct correlate of function,” said Ince. On what contributes to cognitive reserve, Ince suggested a combination of innate and acquired factors. Recent findings from Kawas’s group speak to the former. She reported this past March that the ApoE2 isoform seems to protect the oldest old against cognitive sequelae of AD pathology (see Berlau et al., 2009). In contrast, some people with the ApoE4 isoform show subtle differences in brain imaging early on (e.g., Filippini et al., 2009 and ARF related news story).

One shortcoming of autopsy studies is that they cannot inform on the trajectory of disease progression. It’s unknown, for example, for how long some of the oldest old in this study lived with pathology but no dementia. Plaque imaging with PIB or other amyloid ligands may shed some light on that. The question then becomes “when you visualize plaques during life in someone who is cognitively intact, what does it mean?” asked Kawas. Longitudinal studies may help address that question (see ARF related HAI conference story).

“As we learn more about the common course of Alzheimer’s disease, it will become even more important to understand the atypical cases of older persons who are cognitively intact despite having neurodegenerative disease lesions,” write Ewbank and Arnold.—Tom Fagan.

References:
Savva G, Wharton SB, Ince PG, Forster G, Matthews FE, Brayne C. Age, neuropathology, and dementia. New England Journal of Medicine. 2009 May 28; 360:2302-2309. Abstract

Ewbank DC, Arnold SE. Cool with plaques and tangles. New England Journal of Medicine. 2009 May 28; 360:2357-2359. Abstract