Neurovascular decoupling is associated with severity of cerebral amyloid angiopathy

Abstract

Objectives: We used functional MRI (fMRI), transcranial Doppler ultrasound, and visual evoked potentials (VEPs) to determine the nature of blood flow responses to functional brain activity and carbon dioxide (CO2) inhalation in patients with cerebral amyloid angiopathy (CAA), and their association with markers of CAA severity.

Methods: In a cross-sectional prospective cohort study, fMRI, transcranial Doppler ultrasound CO2 reactivity, and VEP data were compared between 18 patients with probable CAA (by Boston criteria) and 18 healthy controls, matched by sex and age. Functional MRI consisted of a visual task (viewing an alternating checkerboard pattern) and a motor task (tapping the fingers of the dominant hand).

Results: Patients with CAA had lower amplitude of the fMRI response in visual cortex compared with controls (p = 0.01), but not in motor cortex (p = 0.22). In patients with CAA, lower visual cortex fMRI amplitude correlated with higher white matter lesion volume (r = -0.66, p = 0.003) and more microbleeds (r = -0.78, p < 0.001). VEP P100 amplitudes, however, did not differ between CAA and controls (p = 0.45). There were trends toward reduced CO2 reactivity in the middle cerebral artery (p = 0.10) and posterior cerebral artery (p = 0.08).

Conclusions: Impaired blood flow responses in CAA are more evident using a task to activate the occipital lobe than the frontal lobe, consistent with the gradient of increasing vascular amyloid severity from frontal to occipital lobe seen in pathologic studies. Reduced fMRI responses in CAA are caused, at least partly, by impaired vascular reactivity, and are strongly correlated with other neuroimaging markers of CAA severity.

Figure 1. Functional MRI responses in patients with CAA and controls

Group blood oxygen level–dependent functional MRI responses for the motor (left) and visual (right) tasks. Reduced functional MRI activation is observed for patients with CAA. Red-to-yellow color scale indicates the significance of activity, expressed as a z score ranging from 3.1 (red) to 6.0 (yellow). CAA = cerebral amyloid angiopathy.

Figure 2. Functional MRI BOLD responses in patients with CAA and controls

Group average time curves of BOLD fMRI signal in patients with CAA and controls, expressed as the percent change from baseline, in the 50 most active voxels in the primary visual cortex during the visual stimulation task. Percentage increase in BOLD fMRI signal is reduced in patients with CAA compared with controls. Error bars indicate the SEM. BOLD = blood oxygen level–dependent; CAA = cerebral amyloid angiopathy; fMRI = functional MRI.

Figure 3. BOLD response amplitude is lower in patients with CAA than in controls

BOLD fMRI amplitude responses are lower in patients with CAA than in controls in the visual task (A) (p = 0.01) but not the motor task (B) (p = 0.22). The horizontal bar indicates the mean response of the group. Significance testing is by paired t test. BOLD = blood oxygen level–dependent; CAA = cerebral amyloid angiopathy; fMRI = functional MRI.

Figure 4. Lower fMRI responses are correlated with higher WMH volume and microbleed count

In patients with CAA, lower BOLD fMRI activity in visual cortex is correlated with (A) higher WMH volume (normalized to average subject head size, see the Methods section for details) (p = 0.002) and (B) higher microbleed count (p = 0.001). The “r” indicates the Pearson correlation coefficient. Neither age nor hypertension was related to either WMH or microbleed count (p > 0.05 for all comparisons). BOLD = blood oxygen level–dependent; CAA = cerebral amyloid angiopathy; CMB = cerebral microbleed; fMRI = functional MRI; ICH = intracerebral hemorrhage; WMH = white matter hyperintensity.