Patterns of Cortical Neurodegeneration in Alzheimer's Disease: Subgroups, Subtypes, and Implications for Staging Strategies

Normal brain aging into the ninth decade is associated with cognitive impairments that are quantitative in nature rather than forming subgroups of changes and it is not associated with appreciable neuron degeneration. In contrast, although age is a risk factor for Alzheimer's disease (AD), cognitive deficits form statistically unique subgroups and the patterns of neocortical neurodegeneration occur according to early deficits. There is substantial evidence for clincopathological subgroups and some studies suggest the possibility of subtypes. A subtype is comprised of cases with unique clinical deficits, pattern of brain lesions, and etiology. One possible AD subtype are patients with early spastic paraparesis associated with the deletion of exon 9 from the presenilin 1 gene, a unique type of senile plaque, and degeneration of the corticospinal tract. Also suggestive of subtypes are different laminar patterns of neurodegeneration in posterior cingulate cortex. These patterns of neuron death are not compatible with a linear regression model of large neuron death, but rather a multivariate model based on principal components analysis. Because AD is not composed of a single pattern of clinical, cognitive, and pathological lesions, staging at postmortem examination can only be used as a general guide to the level of neurofibrillary and amyloid deposition in the cerebral cortex. The presence of clinicopathological subgroups, the lack of a strong relationship between neuron death and neurofibrillary tangles, and the possibility of subtypes raise the specter that no current staging strategy precisely defines disease status. Thus, linear regression of population variables and simple staging need to be replaced by multivariate models and assessments of disease progression that accommodate many statistical and topographical subgroups and/or subtypes of AD.