Abstract
INTRODUCTION: Cerebrovascular reactivity (CVR) to changes in blood carbon dioxide and oxygen levels is a robust indicator of vascular health. Although CVR is typically assessed with hypercapnia, the interplay between carbon dioxide and oxygen, and their ultimate roles in dictating vascular tone, can vary with pathology. Methods to characterize vasoreactivity to oxygen changes, particularly hypoxia, would provide important complementary information to established hypercapnia techniques. However, existing methods to study hypoxic CVR, typically with arterial spin labeling (ASL) MRI, demonstrate high variability and paradoxical responses. METHODS: To understand whether these responses are real or due to methodological confounds of ASL, we used phase-contrast MRI to quantify whole-brain blood flow in 21 participants during baseline, hypoxic, and hypercapnic respiratory states in three scan sessions. RESULTS: Hypoxic CVRreliability was poor-to-moderate (ICC = 0.42 for CVR relative to P(ET)O(2) changes, ICC = 0.56 relative to SpO(2) changes) and was less reliable than hypercapnic CVR (ICC = 0.67). DISCUSSION: Without the uncertainty from ASL-related confounds, we still observed paradoxical responses at each timepoint. Concurrent changes in blood carbon dioxide levels did not account for paradoxical responses. Hypoxic CVR and hypercapnic CVR shared approximately 40% of variance across the dataset, indicating that the two effects may indeed reflect distinct, complementary elements of vascular regulation. The data included in this article were collected as part of a randomized cross-over clinical trial, but do not assess the outcomes of this trial: Improving Human Cerebrovascular Function Using Acute Intermittent Hypoxia (NCT05164705), https://clinicaltrials.gov/study/NCT05164705.