Effects of Oxygen on Perioperative Vascular Function: A Randomized Clinical Trial

Abstract

Background: Vascular dysfunction contributes to postoperative organ injury. Exposure to high concentrations of oxygen during surgery is common and may impair vascular function. We tested the hypothesis that hyperoxia during cardiac surgery impairs vascular function compared with normoxia.

Methods: We recruited and randomly assigned patients having elective cardiac surgery to hyperoxia or normoxia during surgery, measured endothelium-mediated vasodilation via brachial artery flow-mediated dilation and fingertip pulse amplitude tonometry (reactive hyperemia index), assessed endothelium-dependent, endothelium-independent, and heme-independent soluble guanylyl cyclase activator-induced vasodilation ex vivo in mediastinal fat arterioles using wire myography, and quantified plasma markers of vascular function and oxidative stress.

Results: Two hundred participants completed the study. Oxygen treatment did not affect flow-mediated dilation (primary outcome, P=0.377) or reactive hyperemia index (P=0.898). In isolated mediastinal fat arterioles, however, hyperoxia impaired endothelium-independent relaxation (P<0.001) but not endothelium-dependent relaxation (P=0.759) or heme-independent soluble guanylyl cyclase activation (P=0.650). Hyperoxia also increased plasma plasminogen activator inhibitor-1 postoperatively but not e-selectin or syndecan-1. Hyperoxia increased intraoperative concentrations of F₂-isoprostanes and isofurans, which were associated with plasminogen activator inhibitor-1 but not other measurements of vascular function.

Conclusions: Among adults receiving cardiac surgery, intraoperative hyperoxia did not affect endothelium-dependent vasodilation but impaired endothelium-independent vasodilation, likely via soluble guanylyl cyclase heme oxidation. Soluble guanylyl cyclase is a potential therapeutic target to enhance vascular function.

Registration: URL: https://www.clinicaltrials.gov; Unique identifier: NCT02361944.

Keywords: cardiac surgery; endothelium; nitric oxide; oxygen; vascular function.

Figure 1. Consolidated Standards of Reporting Trials diagram.

Figure 2. Perioperative brachial artery flow‐mediated dilation and reactive hyperemia index responses in participants treated with hyperoxia and normoxia.

Boxes represent interquartile range, and horizontal line indicates median value. Whiskers represent the 10th to 90th percentile concentrations and values outside 10th and 90th percentile are indicated with dots. P values indicate statistical significance for effect of treatment across both assessment time points and adjusted for shear rate. ICU indicates intensive care unit.

Figure 3. Ex vivo mediastinal fat arteriole vasodilation in hyperoxia and normoxia treated participants.

Endothelium‐dependent vasodilation was assessed with increasing doses of acetylcholine (A). The endothelium‐independent response was assessed with increasing doses of sodium nitroprusside (B). Finally, the heme redox state independent response to soluble guanylyl cyclase activation was assessed with increasing doses of cinaciguat (C). Dots indicate the predicted response at each dose of the indicated drug, and the shaded area indicates the 95% CI of the response.

Figure 4. Plasma plasminogen activator inhibitor‐1 (A), e‐selectin (B), and syndecan‐1 (C) concentrations in participants treated with hyperoxia and normoxia.

Boxes represent interquartile range and horizontal lines indicates median value. Whiskers represent the 10th to 90th percentile concentrations, and values outside the 10th and 90th percentile are indicated with dots. ICU indicates intensive care unit; and PAI‐1, plasminogen activator inhibitor 1.