Inhaled Nitric Oxide for Severe Hypoxemia After CABG: A Propensity Score‒Matched Analysis

To evaluate whether a stepped treatment strategy incorporating inhaled nitric oxide (iNO) is associated with more rapid improvement in oxygenation and a lower likelihood of treatment escalation compared to conventional high-flow nasal cannula (HFNC) therapy, in extubated patients with severe hypoxemia (PaO₂/FiO₂ ≤ 100 mmHg) following coronary artery bypass grafting (CABG).

A single-center, retrospective, propensity score‒matched cohort study.

A cardiac surgery intensive care unit (ICU) at a tertiary care hospital.

Adult patients (≥18 years) who developed severe hypoxemia after extubation following elective off-pump CABG via median sternotomy between January and December 2025. After 1:1 propensity score matching, 122 patients (61 per group) were included.

The iNO‑first strategy group (n = 61) received a stepwise regimen: iNO alone (20 ppm), then iNO with HFNC, then iNO with non-invasive positive pressure ventilation (NPPV) based on predefined oxygenation criteria. The HFNC‑first strategy group (n = 61) received a stepwise regimen: HFNC alone, then NPPV.

The primary outcome was the PaO₂/FiO₂ ratio at 6, 12, and 24 hours. The iNO‑first strategy group had significantly higher median PaO₂/FiO₂ at 6 hours (158 v 132 mmHg, p < .001) and 12 hours (161 v 144 mmHg, p < .001). The difference at 24 hours was not significant (199 v 187 mmHg, p = .056). In the iNO‑first strategy group, 24 patients (39.34%) required escalation to HFNC added to iNO, whereas all patients in the HFNC‑first strategy group received HFNC as initial therapy. NPPV use did not differ (16.39% v 19.67%, p = .639). Postoperative hospital stay was shorter in the iNO‑first strategy group (13 v 14 days, p = .043), though this finding should be interpreted with caution. No reintubation or mortality occurred in either group.

In this retrospective analysis, an iNO-first strategy was associated with superior early oxygenation at 6 and 12 hours compared to an HFNC-first strategy. This study compares 2 non‑randomized treatment strategies rather than the isolated pharmacological effect of iNO. The strategy demonstrated a favorable safety profile. However, the benefits were not sustained at 24 hours, and no differences were found in NPPV utilization or ICU length of stay. Given the non-randomized design, these results should be interpreted as associations, not causal effects.