Empirically defining preclinical Alzheimer’s disease in the era of biomarkers

Background The National Institute on Aging and Alzheimer’s Association released a research framework defining an Alzheimer’s Continuum (AC) based on the presence of at least one b‐amyloid biomarker, independent of clinical symptoms. Staging of the AC will occur along 6 levels of severity. Stages 3‐6 map onto current conceptualizations of mild cognitive impairment and mild, moderate, and severe dementia. However, stages 1 and 2 are preclinical stages, which are less well characterized. These early stages are defined by three criteria: subjective cognitive complaints, transitional declines on cognitive testing, and neurobehavioral symptoms. The goals of the current study were to 1) examine base rates of these criteria among groups of individuals defined as amyloid positive (A+) and amyloid negative (A‐); 2) evaluate inter‐rater agreement for subjective cognitive complaints and neurobehavioral symptoms; and 3) assess the impact of different diagnostic standards on rates of classification into AC stages 1 and 2. Method Analyses were completed on a subsample of cognitively normal participants from the Uniform Data Set, categorized as A+ (n = 101) or A‐ (n = 447). Result Overall rates of subjective cognitive complaints (A+: 34.90%, A‐: 29.90%; χ 2 = 3.33, p = .068) and neurobehavioral symptoms (A+: 22.40%, A‐: 22.40%; χ 2 = 0.00, p = .990) did not significantly differ across amyloid groups. Inter‐observer agreement was fair for subjective cognitive complaints (k = .24) and fair‐to‐moderate for neurobehavioral symptoms (k ranged from .34 to .53). Multivariate base rate analysis revealed that subtly low cognitive scores were more common in amyloid positive than amyloid negative individuals (e.g., the base rate of two scores £9th percentile was 38.00% for A+ individuals but only 24.90% for A‐ patients, χ 2 = 6.98, p = .008). The rates of classification into AC stage 1 versus 2 depended on the number of criteria required for stage 2: 57.40% versus 42.60% (one criterion), 28.70% versus 71.30% (two criteria), and 6.90% versus 93.10% (all three criteria). Conclusion Implementing a two criteria standard may be preferred to provide a balance between false negatives and false positives. Directions for future research are discussed.