Biomarker affliction classes contribute additively to observed dementia severity and prospective conversion risk

Background Dementia is likely to be overdetermined by the independent contributions of its biomarkers, of which Alzheimer's disease (AD)-specific biomarkers are a subset. Objective To assess the impact of affliction by multiple biomarkers on dementia severity. Methods Using previously validated algorithms, N = 988 PET (+) subjects of the Alzheimer's Disease Neuroimaging Initiative (ADNI) were assigned to groups “afflicted by” or “resilient against” the effects of central nervous system amyloid-β (Aβ), plasma adipokines, and/or neurodegeneration and tested against Clinical Dementia Rating Scale-Sum of Boxes (CDR-SB) and time to dementia conversion. Results CDR-SB rose as a function of the number of afflicting biomarkers in both demented and non-demented cases. 327/988 (33.1%) were afflicted by a single biomarker. That biomarker was Aβ in only 19.88% of these PET (+) subjects. 221/988 (22.4%) were afflicted by all three. Only 124/988 (12.6%) were resilient to all three. Affliction by adipokines had the strongest effect (r = 0.33, p < 0.001). Affliction by Aβ was weakest (r = 0.18, p < 0.001). The number of afflicting biomarkers explained 25.2% of CDR variance and significantly impacted conversion risk over 48 months (by χ2 (df = 3) = 56.69, p < 0.001) independently of baseline CDR-SB (by Cox Proportional Hazards Wald χ2 (df = 3) = 26.24, p < 0.001). Conclusions The biomarkers that determine dementia severity in PET (+) subjects may ultimately comprise ad hoc combinations and do not necessarily include Aβ. These findings have implications for A/T/N diagnoses.