Suppression of the Hemodynamic Response Function Demonstrates Altered Cerebral Vasoreactivity in Sickle Cell Disease

Introduction: Children with sickle cell disease (SCD) face progressive cognitive decline compared to their unaffected siblings, even prior to the occurrence of stroke, suggesting damage to the underlying functional brain organization. Hence, functional MRI (fMRI) studies are needed to help characterize the pathophysiological basis of cognitive decline in SCD.

Functional MRI techniques measure the blood oxygenation level dependent (BOLD) signal, which is sensitive to changes in the ratio of oxy- to deoxy-hemoglobin driven by local changes in neural activity. Regional brain activity raises local oxygen consumption that is overly compensated for by increases in local blood flow, bringing an excess of oxygenated blood and elevating the ratio of oxy- to deoxy-hemoglobin in the active brain regions. This stereotyped neurovascular coupling sequence unfolds over 20 seconds and gives rise to the canonical hemodynamic response function (HRF). Given the severe chronic anemia with attendant hypoxia, alterations in oxygen binding, increased cerebral blood flow (CBF), and prevalence of vasculopathy associated with SCD, careful comparison of the HRF between children with SCD and controls must occur prior to pursuing fMRI analyses.

Children with SCD have increased CBF and oxygen extraction fraction at baseline when compared to controls, suggesting that they require these compensatory mechanisms to meet the brain's metabolic demands at rest, which could affect their ability to further increase blood flow in response to local neural activity. Thus, we hypothesized that children with SCD would have a blunted HRF with lower peak amplitudes when compared to sibling controls. To test our hypothesis, we conducted the gold-standard visual task-driven fMRI paradigm for measuring HRF in children with SCD and their siblings.

Methods: Children with HbSS or HbS beta thalassemia null disease and healthy, sibling controls between the ages of 5 and 21 years were recruited from St. Louis Children's Hospital. Participants were excluded for history of overt stroke, vasculopathy, stem cell transplant or chronic exchange transfusion therapy. Participants completed a fMRI task that required them to press a button with their right index fingers at the onset of a 30 Hz counterphase-flickering checkerboard and to press a button with their left index finger at its offset. Nine time points (TR = 2.2 s) were included in the general linear model used to describe the HRF in response to each stimulus. No assumptions were made regarding the shape of the HRF. A peak-finding algorithm extracted regions of interest (ROI) from the group fMRI activation map. Activations were averaged across all voxels within each ROI to generate region-wise time courses for each participant. Results are reported as median [interquartile range], and comparisons are made using a Mann Whitney U test or Pearson chi-square test.

Results: Fourteen children with SCD and 13 healthy sibling controls were included in the analysis. There was not a statistically significant difference in age (10.0 [9.8-15.5] years vs. 13.0 years [11.0-14.5]; p = 0.66) or gender (p = 0.31) between the SCD and control cohorts. Eighty-six percent of SCD participants were treated with hydroxyurea at a dose of 30.3 [28.6-33.1] mg/kg/day. The baseline hemoglobin of the SCD cohort was 8.8 [8.3-10.1] g/dL. Two children with SCD had a history of silent cerebral infarcts.

The peak amplitude of the HRF in the primary visual cortex (V1) was significantly reduced in children with SCD (right: 0.29 [0.23-0.50]% and left: 0.29 [0.20-0.53]%) when compared to sibling controls (right: 0.56 [0.53-0.95]%, p < 0.001 and left: 0.52 [0.38-0.90]%, p = 0.01). While the HRF amplitude was reduced by almost 50%, there was no difference in the timing of the BOLD response (i.e. lag) between groups, both demonstrating a time-to-peak of 8.8 seconds (Figure).

Conclusion: The striking decrease in HRF amplitude suggests that patients with SCD have altered cerebral vasoreactivity to neural stimuli. Further investigation into the pathophysiology behind and clinical implications of the blunted hemodynamic response in SCD is warranted. We are currently investigating the relationship between HRF suppression, cognitive outcomes and stroke risk in SCD, and the effects of HRF suppression on functional connectivity networks using resting-state functional connectivity MRI collected simultaneously in the same cohorts.

Figure.

View largeDownload slide

Figure.

View largeDownload slide

Close modal