The joint association of arteriosclerosis and plasma amyloid with cognitive function: the population‐based Rotterdam Study

Abstract

BackgroundAlzheimer's disease and vascular brain injury are the most common causes of dementia, and are present jointly in 40% of individuals at time of death. Yet, the interaction between Alzheimer's and vascular pathology on cognition is largely undetermined. We studied the interaction between arteriosclerosis and plasma biomarkers of Alzheimer's disease on cognition in a general population without dementia.MethodOn stored plasma samples, obtained between 2002‐2005, we measured plasma levels of amyloid‐β40 and amyloid‐β42 by Simoa assay in 2252 participants of the population‐based Rotterdam Study (mean age: 68.9 years, 52% women). Arteriosclerosis was assessed using arterial calcification volumes that were quantified (semi‐)automatically on unenhanced computed tomography (CT). We captured the systemic burden of calcification, using a principal component analysis of calcification volumes in 4 different arterial vessel beds (intracranial and extracranial carotid arteries, coronary arteries and aortic arch). Cognitive assessment included the verbal fluency test, the letter‐digit substitution task, a 15‐word learning test, the Stroop test, and Purdue pegboard task. Global cognition (g‐factor) was calculated using a principal component analysis of these tests. We then determined the interaction between plasma amyloid‐β and calcification on cognitive performance, using multivariable linear regression models.ResultPlasma levels of amyloid‐β and systemic arteriosclerosis were both associated with poorer cognitive performance. After mutual adjustment in multivariable models only calcification remained significantly associated (beta [95%CI] for g‐factor per 1‐SD increase in amyloid‐β40: ‐0.02 [‐0.07; 0.03]; amyloid‐β42: 0.03 [‐0.01; 0.08]; and calcification burden: ‐0.06 [‐0.09; ‐0.04]). The association between amyloid‐β and cognition was stronger in individuals with higher burden of calcification (Figure 1), particularly for amyloid‐β40 (interaction amyloid‐β40×calcification: p = 0.04; amyloid‐β42×calcification: p = 0.001). Results were broadly similar across cognitive tests. The observed interaction between amyloid‐β40 and calcification volume was most profound for calcification burden in the intracranial carotid artery and aortic arch.ConclusionArteriosclerosis and amyloid‐β display synergistic effects on cognition in the general population. This interaction was more profound for amyloid‐β40 than for amyloid‐β42, suggesting our observations may relate particularly to amyloid angiopathy.