Abstract
Hypertension in midlife contributes to cognitive decline and is a modifiable risk factor for dementia. The relationship between late-life hypertension and dementia is less clear. We have investigated the relationship of blood pressure and hypertensive status during late life (after 65 years) to post-mortem markers of Alzheimer's disease (amyloid-β and tau loads); arteriolosclerosis and cerebral amyloid angiopathy; and to biochemical measures of ante-mortem cerebral oxygenation (the myelin-associated glycoprotein:proteolipid protein-1 ratio, which is reduced in chronically hypoperfused brain tissue, and the level of vascular endothelial growth factor-A, which is upregulated by tissue hypoxia); blood-brain barrier damage (indicated by an increase in parenchymal fibrinogen); and pericyte content (platelet-derived growth factor receptor β, which declines with pericyte loss), in Alzheimer's disease (n = 75), vascular (n = 20) and mixed dementia (n = 31) cohorts. Systolic and diastolic blood pressure measurements were obtained retrospectively from clinical records. Non-amyloid small vessel disease and cerebral amyloid angiopathy were scored semiquantitatively. Amyloid-β and tau loads were assessed by field fraction measurement in immunolabelled sections of frontal and parietal lobes. Homogenates of frozen tissue from the contralateral frontal and parietal lobes (cortex and white matter) were used to measure markers of vascular function by enzyme-linked immunosorbent assay. Diastolic (but not systolic) blood pressure was associated with the preservation of cerebral oxygenation, correlating positively with the ratio of myelin-associated glycoprotein to proteolipid protein-1 and negatively with vascular endothelial growth factor-A in both the frontal and parietal cortices. Diastolic blood pressure correlated negatively with parenchymal amyloid-β in the parietal cortex. In dementia cases, elevated late-life diastolic blood pressure was associated with more severe arteriolosclerosis and cerebral amyloid angiopathy, and diastolic blood pressure correlated positively with parenchymal fibrinogen, indicating blood-brain barrier breakdown in both regions of the cortex. Systolic blood pressure was related to lower platelet-derived growth factor receptor β in controls in the frontal cortex and in dementia cases in the superficial white matter. We found no association between blood pressure and tau. Our findings demonstrate a complex relationship between late-life blood pressure, disease pathology and vascular function in dementia. We suggest that hypertension helps to reduce cerebral ischaemia (and may slow amyloid-β accumulation) in the face of increasing cerebral vascular resistance, but exacerbates vascular pathology.