This month’s Archives of Neurology provides a gold mine of studies on the neuropathological correlates of mild cognitive impairment, and the predictive value of this syndrome as a precursor of Alzheimer disease. One paper from Ron Petersen, Gregory Jicha, and colleagues at the Mayo Clinic in Rochester, Minnesota, describes an autopsy study of amnestic MCI (aMCI) patients who died while still in this stage of limited cognitive impairment. The results confirm that aMCI lies at a midway point between normal aging and AD as judged by the extent and localization of plaque and tangle pathology. But is aMCI truly an AD prodrome? That question was addressed by a second study from the same group, which examined subjects who progressed from aMCI to dementia before their death. Those results showed that while most (70 percent) of the elderly with amnestic MCI developed Alzheimer disease, the rest developed dementia from other causes.

These studies should reassure researchers that they are on the right track in their pursuit of MCI, and particularly aMCI as a clinical predictor and early intervention point for AD. However, the results also point out an important problem with this approach, which is that in the general population, many people with MCI do not go on to develop AD, or even dementia. Trying to find better predictors of progression is the topic of a third report, where investigators Liana Apostolova, Paul Thompson, and their colleagues at the University of California, Los Angeles, used imaging to measure hippocampal size and shape in aMCI patients, and then followed the cohort over 3 years. They found that a smaller hippocampus with evidence of atrophy was associated with progression from MCI to dementia.

The Mayo studies are part of the clinic’s decades-long longitudinal community-based study of aging and dementia, which pairs years of clinical follow-up of subjects with postmortem pathological studies. Petersen and Jicha worked with Joseph Parisi and colleagues in Minnesota, plus Dennis Dickson at the Mayo Clinic in Jacksonville, Florida, and Heiko Braak at the J. W. Goethe University, Frankfurt, Germany, on an extensive neuropathological workup, including ,quantification of plaques, tangles and other disease markers throughout the brain.

In the first paper, lead author Petersen and colleagues describe the neuropathology of 15 elderly subjects who died between 1986 and 2004, while they were classified with MCI of the amnestic type (aMCI). Compared to age-matched healthy controls and people with AD, the aMCI group generally fell right in the middle—their brains were not normal, but the pathology did not rise to the level of AD. The quantity and distribution of tau-containing neurofibrillary tangles (rated by Braak score) were higher in the aMCI group than in normal aging, but not as high as in AD. Consistent with previous studies, amyloid plaque scores were very similar to normal aged brains, and much lower than in AD brains. Other abnormalities were commonly seen in a subset of the cases, including argyrophilic grain disease, hippocampal sclerosis, and vascular lesions.

Despite significant heterogeneity in the amount and type of pathology, all 15 subjects showed some involvement in the medial temporal lobe, which was consistent with their diagnosis of memory loss. In particular, the researchers observed NFTs in the entorhinal cortex and hippocampus, and diffuse amyloid in the neocortex—all previously reported signs of prodromal AD. From these results, the authors conclude that, “The regional involvement by NFTs correlated best with the degree of clinical impairment across the spectrum of healthy to aMCI to AD. In contrast, the amyloid plaque burden was less discriminating.” This observation fits with previous studies (see ARF related news story).

The companion study takes a further step down the road to dementia by asking about the neuropathology of patients with MCI who progress to dementia. First author Jicha describes the characterization of a group of 34 subjects who died after progression from aMCI to dementia. Most (71 percent) fit the pathological criteria for AD. But the remaining 29 percent showed non-AD primary pathologies including vascular disease, argyrophilic disease, and Lewy body disease.

No matter what the pathology, though, all had involvement of the medial temporal lobe sufficient to cause memory loss. In most cases, the problem was NFT degeneration, but in some instances the researchers observed hippocampal sclerosis, leading them to write that the “anatomy of neurodegeneration is more predictive of the clinical syndrome than the specific pathologic process. That which unites these various processes is their predilection for the medial temporal lobe.”

In this group, neither demographic nor diagnostic variables predicted which patients would develop AD, and which would develop dementia from other causes. Complicating the picture, most showed two or more pathological processes that might contribute to dementia.

Two editorials, one accompanying each paper, stress the importance of these kinds of painstaking, long-term human studies to understanding the pathogenesis and potential for treating AD. In one, Harry Vinters of the University of California, Los Angeles, raises the question of whether the early “Alzheimerization” of brains seen in MCI means that irreversible cellular and molecular events have already been set in motion, a question that only pharmacological intervention will answer.

Commenting on the progression study, Lawrence Hansen of the University of California, San Diego, calls the results “a sobering reminder that even with a thoroughly documented clinical course progression from MCI to dementia, a final clinical diagnosis of AD may be wrong in a significant minority of patients. As the authors note, these data raise concern about potential pathologic heterogeneity in subjects recruited for therapeutic trials in MCI.”

The heterogeneity issue has dogged the testing of early treatments for AD. While a diagnosis of MCI raises the risk of conversion to dementia, the majority of people with MCI either remain stable or improve when followed for 1 or 2 years. In prospective studies in a general population, conversion rates run 15 percent per year (and, as Jicha and colleagues have now shown, not all of these will have bona fide AD).

This is where the third paper fits in. To try to separate out the progressors from non-progressors, Apostolova, Thompson, and colleagues are investigating the use of magnetic resonance imaging to measure hippocampal morphology as a predictor of who will progress from MCI to dementia. They applied their technique for comparing hippocampal volume and shape (Thompson et al., 2004) to a cohort of 20 amnestic MCI patients, who were then followed up for 3 years.

During the study, six subjects developed clinical AD, seven remained cognitively stable, and seven improved. The group that progressed started with 9 and 13 percent lower hippocampal volume on the left and right side, respectively, relative to the stable group, and 24 and 27 percent lower volume compared to the improved group. The changes were statistically significant except for the right side changes between the progressors and the non-progressors. Regional analysis showed atrophy in the CA1 and subicular regions, with significant increases in patients that progressed to AD compared to the group that improved. The results agree with work done using similar techniques by John Csernansky and colleagues, who detected hippocampal shape changes in the CA1 region that predicted onset of dementia in elderly subjects (see Csernansky et al., 2005).—Pat McCaffrey


  1. The findings with regard to CA1 are consistent with other data indicating selected neuronal loss in CA1 in AD. The involvment of subiculum is less consistent with data showing neuron loss in subiculum to be approximately the same in AD and control brains (e.g. West et al., 1994).


    . Differences in the pattern of hippocampal neuronal loss in normal ageing and Alzheimer's disease. Lancet. 1994 Sep 17;344(8925):769-72. PubMed.

    View all comments by Paul Coleman
  2. The number of patients with mild cognitive impairment (MCI) is increasing and many of them are referred to memory clinics. The cognitive decline in MCI is greater than in normal aging, but the clinical characteristics are heterogeneous and some people may return to normal. It is, therefore, a challenge to detect subgroups that represent a prodromal stage of Alzheimer disease (AD). In two recent papers, Petersen et al. present the histopathology of some MCI cases that remain as MCI subjects or progress to clinical dementia. The advantage of this study is that the interval between the last clinical investigation (including cognitive tests) and death is fairly short (0.72 years), but the drawback is that the subjects are very old (88.9 years). The neuropathological investigations showed that, with respect to various types of plaques, the amnestic MCI subjects (aMCI) more resembled healthy individuals than AD patients. This contrasts recent in vivo amyloid imaging findings using Pittsburgh compound B (PIB) in MCI patients, where high PIB retentions, similar to AD patients, were measured in MCI patients (1,2,3). These observations might represent differences in plaque properties in different courses of the disease as well as age and/or ApoE genotype properties.

    In the second paper from Petersen’s group, Jicha et al. studied the pathological outcome of progression of MCI to dementia in 34 subjects. The majority of aMCI patients progressed to both clinical and pathological AD (19 out of 24 subjects). If in vivo imaging will reveal high amyloid load in aMCI patients, amyloid imaging might be a promising diagnostic technique. The papers by Petersen et al. stress the importance of postmortem follow-up studies of patients undergoing amyloid imaging. By combining data from in vivo and autopsy studies, as well as CSF analysis, we will obtain a deeper insight into the pathophysiological mechanism of AD.

    See also:

    DeKosky ST, Mathis CA, Price JC et al. Human amyloid imaging studies with Pittsburgh compound-B in mild cognitive impairment (MC): Is MCI the critical period of amyloid plaque deposition? AAN San Diego 2006: S01.004


    . Simplified quantification of Pittsburgh Compound B amyloid imaging PET studies: a comparative analysis. J Nucl Med. 2005 Dec;46(12):1959-72. PubMed.

    . Amyloid imaging in MCI patients. Neurobiol Aging 2006; 27(Suppl 1):S6.

  3. These studies confirm the confusion that developed over the years that AD and AD with dementia are the same issue. Progression to AD with dementia is a separate issue from Braak stage. A growing number of autopsy studies confirm the findings of the Nun Study that brain atrophy, i.e. brain cell death, is the key issue in AD with dementia. Forty-four percent of the nondemented group of the 33 religious sisters studied by Grosch et al. met the AD neuropathological criteria upon autopsy at ages ranging from 88 to 93. Hippocampal volume was the primary variable in determining whether a person had just AD with near normal cognition or AD with dementia. Much is known about the environmental, nutritional, social and medical factors that effectively kill brain cells: e.g. folic acid deficiency, aluminum overload, strokes, etc. The research community needs to throw a much wider net beyond Braak stage, to integrate epidemiology, nutrition, toxicology, etc., so that prevention and therapy programs can be developed to assist the patients. We have recently proposed software for this purpose.


    . Hippocampal volume as an index of Alzheimer neuropathology: findings from the Nun Study. Neurology. 2002 May 28;58(10):1476-82. PubMed.

    . Alzheimer disease is substantially preventable in the United States -- review of risk factors, therapy, and the prospects for an expert software system. Med Hypotheses. 2005;64(5):960-7. PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Buildup of Neurofibrillary Pathology Parallels Cognitive Decline

Paper Citations

  1. . Mapping hippocampal and ventricular change in Alzheimer disease. Neuroimage. 2004 Aug;22(4):1754-66. PubMed.
  2. . Preclinical detection of Alzheimer's disease: hippocampal shape and volume predict dementia onset in the elderly. Neuroimage. 2005 Apr 15;25(3):783-92. PubMed.

Further Reading


  1. . Mapping cortical change in Alzheimer's disease, brain development, and schizophrenia. Neuroimage. 2004;23 Suppl 1:S2-18. PubMed.
  2. . Mild cognitive impairment. Lancet. 2006 Apr 15;367(9518):1262-70. PubMed.
  3. . Neuropathologic substrate of mild cognitive impairment. Arch Neurol. 2006 Jan;63(1):38-46. PubMed.

Primary Papers

  1. . Neuropathologic features of amnestic mild cognitive impairment. Arch Neurol. 2006 May;63(5):665-72. PubMed.
  2. . Neuropathologic outcome of mild cognitive impairment following progression to clinical dementia. Arch Neurol. 2006 May;63(5):674-81. PubMed.
  3. . Neuropathology of amnestic mild cognitive impairment. Arch Neurol. 2006 May;63(5):645-6. PubMed.
  4. . The neuropathology of aging, mild cognitive impairment, and Alzheimer disease. Arch Neurol. 2006 May;63(5):647-8. PubMed.
  5. . Conversion of mild cognitive impairment to Alzheimer disease predicted by hippocampal atrophy maps. Arch Neurol. 2006 May;63(5):693-9. PubMed.