Introduction Supplemental oxygen has long been used as a therapeutic agent in the management of ischaemic chest pain, but without a sound scientific basis. We proposed to investigate the effect of high-flow oxygen on coronary physiology, invasively and non-invasively, in those with severe coronary artery disease (CAD).

Methods 30 volunteers and 30 patients with severe CAD were allocated to receive high-flow air and oxygen in a random blinded order. They underwent cardiovascular magnetic resonance (CMR) T2* blood-oxygen level dependent (BOLD) and dynamic contrast-enhanced perfusion imaging. Myocardial blood flow (MBF), perfusion reserve (MPR), first-pass extraction fraction (E), blood volume fraction (v_b) and capillary permeability-surface area product (PS) were quantified using bookend T₁ based non-linearity correction and distributed parameter model constrained deconvolution.

Invasive coronary physiology measurements were performed in severe CAD patients (Fractional Flow Reserve <0.75), also randomised to high-flow air and oxygen, using a pressure wire (Radi, St Jude Medical, Uppsala, Sweden). The mean transit time of a room temperature 3−4 ml bolus of saline was recorded at rest and hyperaemic (adenosine) stress for both inspired gases. Resistance Index (RI), Resistance Reserve Index (RRI), Index of Microvascular Resistance (IMR) and Coronary Flow Reserve (CFR) were derived.

Results In volunteers the change in BOLD Air-T2* was 6.1 ± 7.5 ms vs. O₂-T2* 7.8 ± 8.1, P = 0.575. Rest Air-MBF was 0.88 ± 0.40 ml/min/g vs. O₂-MBF 0.95 ± 0.45, P = 0.168 and hyperaemic Air-MBF 3.46 ± 1.67 vs. O₂-MBF 3.35 ± 2.05, P = 0.821.

In patients, ischaemic region change in BOLD Air-T2* was 1.1 ± 6.6 ms vs. O₂-T2* 2.7 ± 8.2ms, P = 0.583 compared to Air-T2* 11.2 ± 8.7 vs. O₂-T2* 2.9 ± 2.5, P = 0.004 in remote myocardium. Remote Air-MBF was 0.81 ± 0.19 vs. O₂-MBF 0.76 ± 0.15, P = 0.087 and hyperaemic Air-MBF 1.31 ± 0.047 vs. O₂-MBF 1.50 ± 0.53, P = 0.249; whilst ischaemic region Air-MBF was 0.83 ± 0.20 vs. O₂-MBF 0.85 ± 0.28, P = 0.725 and hyperaemic Air-MBF 1.26 ± 0.055 vs. O₂-MBF 1.35 ± 0.59, P = 0.407. Rest ischaemic Air-E was 0.69 ± 0.14 vs. O₂-E 0.66 ± 0.17, P = 0.408 and remote Air-E was 0.73 ± 0.13 vs. O₂-E 0.64 ± 0.15, P = 0.044. Ischaemic region Air-PS was 0.67 ± 0.38 ml/min/g vs. O₂-PS 0.58 ± 0.32, P = 0.227; whilst remote region Air-PS was 0.72 ± 0.38 vs. O₂-PS 0.53 ± 0.32, P = 0.024.

Invasively there was an increase in rest transit time 0.69 ± 0.35 vs. 0.97 ± 0.50, P = 0.001, resistance index 51.0 ± 26.6 vs. 75.0 ± 38.0, P < 0.001, coronary flow reserve 2.21 ± 1.30 vs. 2.90 ± 1.51, P = 0.019 and resistance reserve index 2.62 ± 1.37 vs. 3.54 ± 1.74, P = 0.006 in the air vs oxygen groups respectively.

Conclusion Supplemental high-flow oxygen (and hence hyperoxaemia) results in increased microvascular resistance in patients with severe CAD, demonstrated during invasive coronary physiology studies. We have also shown with novel CMR techniques that, in patients with severe CAD, whilst there is no difference in absolute myocardial blood flow, high-flow supplemental oxygen results in a reduction in first-pass extraction fraction and capillary permeability-surface area product. This results in a reduction in myocardial oxygenation in the remote, non-ischaemic, myocardium. Supplemental high-flow oxygen should therefore be avoided in patients who are not hypoxic.