The relationship between blood pressure and cognitive function

Abstract

The relationship between blood pressure (BP) and cognitive outcomes in elderly adults has implications for global health care. Both hypertension and hypotension affect brain perfusion and worsen cognitive outcomes. The presence of hypertension and other vascular risk factors has been associated with decreased performance in executive function and attention tests. Cerebrovascular reserve has emerged as a potential biomarker for monitoring pressure-perfusion-cognition relationships. A decline in vascular reserve capacity can lead to impaired neurovascular coupling and decreased cognitive ability. Endothelial dysfunction, microvascular disease, and mascrovascular disease in midlife could also have an important role in the manifestations and severity of multiple medical conditions underlying cognitive decline late in life. However, questions remain about the role of antihypertensive therapies for long-term prevention of cognitive decline. In this Review, we address the underlying pathophysiology and the existing evidence supporting the role of vascular factors in late-life cognitive decline.

Figure 1

Normal cerebral autoregulation curve with its lower (50 mmHg) and upper (150 mmHg) limits of mean arterial pressure (green line), and a narrowed range with a steeper curve (red dashed line).

Figure 2

Relationship between arterial BP and BFV in the middle cerebral artery. a | Dominant spontaneous oscillations of BP and BFV in a 72-year-old healthy control woman. b | Dominant spontaneous oscillations of BP and BFV in a 68-year-old man with type 2 diabetes mellitus, in the supine position. BP, BFVL and BFVR were decomposed into different modes, each mode corresponding to fluctuations at a different timescale. BP and BFV fluctuations exhibit continuous and dominant oscillations at frequencies 0.07–0.4 Hz. Instantaneous phases of BP and BFV oscillations (solid lines, bottom graphs) and the mean BP–BFV phase shift (dashed lines) were obtained. c | Phase shift between arterial BP and BFV. Results were obtained from 12 healthy controls, 10 patients with hypertension, and 10 patients with history of stroke, by calculating the instantaneous phase shift during the Valsalva maneuver. Dynamic autoregulation in controls was characterized by specific BP–BFV phase shifts. The reduction of the phase shifts observed in participants with hypertension who had never had a stroke and patients with history of stroke indicates impaired autoregulation. Abbreviations: BFV, blood flow velocity; BFVL, left BFV; BFVR, right BFVR; BP, blood pressure. Parts a and b modified with permission from Hu et al. Part c reprinted with permission from Hu et al.

Figure 3

Anatomical and perfusion images from a patient with hypertension and diabetes mellitus and an age-matched healthy control. a | Brain volume loss. b | Extensive periventricular white matter hyperintensities. c | Reduced perfusion in the frontal and temporal regions. d | Normal brain volume. e | Absence of white matter hyperintensities. f | Normal perfusion throughout the brain. Abbreviations: 3D CASL, three-dimensional continuous arterial spin labeling; FLAIR, fluid attenuated inversion recovery; MPRAGE, T1-weighted magnetization-prepared rapid acquisition with gradient echography.

Figure 4

High resolution 8-T gradient gadolinium-enhanced echography slices (in-plane pixel size 195 μm). a–f | Vascular supply to the lacunar infarctions and vascular patterns in the infarcted region. Black arrows, infarct sites. Red arrows, small vessel ending within low signal intensity foci in the lacunar infarction. Blue arrows, branches of the cerebral artery supplying infarcted areas. White arrow, larger area of low signal intensity suggestive of iron deposits that extends beyond infarction into the basal ganglia. Iron deposition in brain parenchyma may represent blood–brain-barrier breakdown associated with microvascular disease and microinfarcts. g | Clinical T2-weighted image of the infarcted area (indicated by black arrows), obtained with 1.5-T MRI. Modified with permission from Novak et al.