Clinical validity of Braak neuropathological staging in the oldest-old

Several studies have demonstrated a good correlation between clinical sever ity and Braak's neuropathological staging in Alzheimer's disease (AD). Howe ver, nonagenarians and centenarians display a different pattern of cortical vulnerability to the neurodegenerative process compared to younger elderly , and it is not known whether correlations between clinical severity and ne uropathological stages remain valid in this age group. To address this issu e we compared Clinical Dementia Rating scale (CDR) scores and Braak stages in 116 patients over 90 years of age with either no cognitive impairment or very mild to severe AD. There is a strong positive correlation between CDR scores and Braak staging (Spearman coefficient = 0.66; P < 0.01), However, neuropathological staging does not distinguish cases with normal cognition (CDR 0) from those with mild cognitive changes (CDR 0.5). Unlike younger c ohorts, Braak stages I and II are frequently associated with questionable d ementia in this age group. Braak stage III overlaps with all CDR levels and correlates poorly with cognitive function. Braak stages IV or greater are consistently associated with at least mild dementia. Consistent with our pr evious neuropathological analyses of nonagenarians and centenarians, the pr esent data suggest that the substantial involvement of the hippocampus whic h characterizes Braak stage IV is a key step in the development of overt cl inical signs of dementia in the oldest-old. Moreover, they indicate that Br aak staging represents a broad concept of the evolution of neurofibrillary tangles rather than a precise hierarchical model associated with a stepwise deterioration of cognitive abilities near the upper limit of life.