OBJECTIVE:

To examine whether the association between clinical Alzheimer disease (AD) diagnosis and neuropathology and the precision by which neuropathology differentiates people with clinical AD from those with normal cognition varies by age.

METHODS:

We conducted a cross-sectional analysis of 2,014 older adults (≥70 years at death) from the National Alzheimer's Coordinating Center database with clinical diagnosis of normal cognition (made ≤1 year before death, n = 419) or AD (at ≥65 years, n = 1,595) and a postmortem neuropathologic examination evaluating AD pathology (neurofibrillary tangles, neuritic plaques) and non-AD pathology (diffuse plaques, amyloid angiopathy, Lewy bodies, macrovascular disease, microvascular disease). We used adjusted logistic regression to analyze the relationship between clinical AD diagnosis and neuropathologic features, area under the receiver operating characteristic curve (c statistic) to evaluate how precisely neuropathology differentiates between cognitive diagnoses, and an interaction to identify effect modification by age group.

RESULTS:

In a model controlling for coexisting neuropathologic features, the relationship between clinical AD diagnosis and neurofibrillary tangles was significantly weaker with increasing age (p < 0.001 for interaction). The aggregate of all neuropathologic features more strongly differentiated people with clinical AD from those without in younger age groups (70-74 years: c statistic, 95% confidence interval: 0.93, 0.89-0.96; 75-84 years: 0.95, 0.87-0.95; ≥85 years: 0.83, 0.80-0.87). Non-AD pathology significantly improved precision of differentiation across all age groups (p < 0.004).

CONCLUSION:

Clinical AD diagnosis was more weakly associated with neurofibrillary tangles among the oldest old compared to younger age groups, possibly due to less accurate clinical diagnosis, better neurocompensation, or unaccounted pathology among the oldest old.