Is it Alzheimer disease or just old age? What senior hasn’t thought as much while searching for misplaced car keys or trying to remember a favorite cousin’s daughter’s name? For clinicians, telling the difference is critically important, in light of increasing evidence that irreversible brain damage occurs early in AD, long before the clear-cut diagnosis of dementia is made. The good news, from a paper in the August issue of Neuropsychology, is that there is a detectable difference between the normal loss of brain power with aging and the early signs of Alzheimer disease between groups of elderly subjects. But the bad news is that it is still difficult to say for a given person exactly where the boundary sits.
That’s the upshot of a meta-analysis of nearly 20 years of clinical studies looking at cognitive impairment in the years before AD is diagnosed. The work, from Lars Backman and colleagues at the Karolinska Institute, and Brent Small at the University of South Florida, shows that relatively large and widespread cognitive impairment can be measured in preclinical AD compared to normal aging. At the same time, the results show that there is considerable overlap in test scores between individuals with normal aging and nascent AD, meaning that distinguishing the two prospectively by cognitive measures alone will continue to be a challenge.
Clinicians agree that measurable mental decline occurs in the years before an AD diagnosis, but no one knows whether memory deficits, cognitive speed, or even a composite measure of brain function best separates normal aging individuals from those who will progress to the frank dementia of AD. To get at this problem, first author Backman and coworkers combed the scientific literature for studies that looked at baseline cognitive performance in subjects who were followed up to see who later developed AD. The 47 studies they identified covered 1,207 preclinical AD cases and 9,097 controls. Measures of cognitive function that were assessed at baseline included global cognitive ability (the Mini Mental State Examination, or MMSE), as well as seven more specialized tests of episodic memory, executive functioning, verbal ability, visuospatial skill, attention, perceptual speed, and primary memory.
Consistent with current thinking, Backman and colleagues found that episodic memory deficits were strongly associated with a later AD diagnosis, but so were other measures, including the MMSE, executive function, and perceptual speed. In fact, they detected statistically significant deficits for seven of eight tests in the AD incident group, with only primary memory showing no preclinical deficits. Their observation of generalized shortfalls in brain function is consistent with histopathological findings in AD brains, showing early and widespread changes in brain structure and function.
The authors also asked to what extent study design influences the strength of association between global cognitive impairment or defects in episodic memory (the two measures with the largest sample sizes) and preclinical AD. Studies that featured younger age and shorter follow-up interval resulted in stronger associations, and the use of more difficult tasks to measure memory revealed greater differences between the normal and AD groups. In general, however, the differences due to study design were small compared to the very large differences observed between the groups.
This convincing evidence of widespread cognitive deficits appearing well in advance of AD diagnosis, along with biochemical studies showing synapse loss and brain toxicity occurring early on in AD models (see ARF related news story) only strengthens the argument that accurate identification of nascent disease may well hold the key to prevention and treatment. But despite the strong associations they found between early cognitive performance and later AD diagnosis, Backman and colleagues also saw extensive overlap in the performance of normal aging individuals and those with preclinical AD on all the tests. There are many reasons for this, including the multiplicity of causes of cognitive decline with age that have nothing to do with AD, including psychiatric, hormonal, or circulatory problems, to name a few. And the variable rates of cognitive decline among individuals with AD may also make some cases hard to spot.
So what is the best strategy to predict if an individual’s future holds a diagnosis of AD? Data show that the prognostic value of cognitive tests is improved when several specific tests are combined, as in the MCI model (see ARF related news story). In addition, Backman and colleagues propose that other, independent measures such as brain volume, glucose metabolism, plaque load, as well as genetic risk and life experience could also be added to cognitive tests to provide a more predictive diagnosis of incident AD while there is still time to halt or reverse the progression to dementia.—Pat McCaffrey
- Borroni B, Perani D, Broli M, Colciaghi F, Garibotto V, Paghera B, Agosti C, Giubbini R, Di Luca M, Padovani A. Pre-clinical diagnosis of Alzheimer disease combining platelet amyloid precursor protein ratio and rCBF spect analysis. J Neurol. 2005 Nov;252(11):1359-62. PubMed.
- Bäckman L, Jones S, Berger AK, Laukka EJ, Small BJ. Cognitive impairment in preclinical Alzheimer's disease: a meta-analysis. Neuropsychology. 2005 Jul;19(4):520-31. PubMed.