Inhaled Nitric Oxide via High-Flow Nasal Cannula in Postrepair Congenital Heart Disease Patients With Pulmonary Arterial Hypertension Following Extubation: A Cohort Study With Propensity Score Matching

Introduction

In low- and middle-income countries and regions, patients with congenital heart disease (CHD) of the systemic-to-pulmonary artery shunt type often undergo surgical closure at a relatively advanced age. By this time, pulmonary arterial hypertension (PAH) and elevated pulmonary vascular resistance have typically already developed. ¹ For pediatric patients with this type of CHD, a pulmonary vascular resistance index (PVRi) below 6 WU × m² is generally regarded as the optimal surgical indication. ² If the PVRi exceeds this threshold, postoperative mortality increases significantly. In such cases, the administration of pulmonary vascular-targeted medications becomes necessary to prevent postoperative pulmonary hypertensive crisis in the intensive care unit (ICU) and reduce postoperative mortality.^{1, 3}

Inhaled nitric oxide (iNO) is the most commonly used agent in the early postoperative period, typically initiated at 5 to 20 ppm. Its efficacy can be enhanced by adjunctive medications (eg, sildenafil), and gradual withdrawal is required to avoid rebound PAH. ⁴ However, this conventional iNO-based regimen is suboptimal: the combination of iNO, oral sildenafil, and prolonged tapering often prolongs postoperative recovery. ⁵

Recent studies have confirmed the feasibility of noninvasive nitric oxide inhalation via high-flow nasal cannula (HFNC) in patients with acute respiratory failure, showing benefits including improved hypoxemia, reduced mortality, and decreased need for mechanical ventilation.^6,7

In CHD postoperative care, Tominaga et al investigated postextubation iNO therapy via HFNC following the Fontan procedure. Their single-center retrospective study demonstrated that a combined rapid iNO weaning strategy (intubated iNO followed by HFNC-delivered iNO postextubation) significantly shortened the duration of postoperative mechanical ventilation, pleural drainage, and hospital stay compared with conventional strategies. ⁸

This pilot study, however, did not address the issue of why the slow process of conventional weaning protocols for iNO therapy prolong postoperative recovery. Notably, Fontan (single ventricle)-related pulmonary vascular lesions are usually classified into the fifth category, ⁹ which differs from the first category of PAH caused by systemic-to-pulmonary artery shunt. Therefore, conclusions obtained in Fontan patients may not be generalizable to other categories of CHD-PAH. ¹⁰ Against this background, we specifically assess the safety and efficacy of the combined fast iNO weaning strategy in the cohort of postoperative patients with systemic-to-pulmonary artery shunt type CHD-PAH.

Methods

Results

During the study period, 126 patients were screened. After excluding 26 patients who underwent only right heart catheterization without receiving repair (high PVRi, no repair indication), 12 cases of fenestrated repair, seven cases of interventional repair, and six cases with other complex malformations (such as correction for anomalous pulmonary venous drainage, mitral or aortic valvuloplasty, etc), 75 eligible cases were finally included. Among them, Group 1 included 22 cases, and Group 2 included 53 cases. After propensity-score matching, 22 matched pairs were generated. The demographics and patient characteristics before and after propensity-score matching were shown in Table 1 and Table 2.

OPEN IN VIEWER

Table 1.

The Demographics and Patient Characteristics Before Propensity-Score Matching.

Variables	Group 1 (n = 22)	Group 2 (n = 53)	P Value	SMD
Age (years)	5.7 (2.8, 9.3)	3.4 (2.0, 5.6)	.021	–0.586
PVRi (WU × m2)	9.8 (7.4, 11.4)	7.8 (6.8, 9.5)	.066	–0.364
PVRi > 8 WU × m2 (n, %)	13 (59)	24 (45)	.276	NA
Cardiopulmonary bypass (minutes)	94 (74, 119)	84 (60, 114)	.243	–0.115

Abbreviations: NA, not applicable; PVRi: pulmonary vascular resistance index; SMD: standardized mean difference.

OPEN IN VIEWER

Table 2.

The Demographics and Patient Characteristics After Propensity-Score Matching.

Variables	Group 1 (n = 22)	Group 2 (n = 22)	P Value	SMD
Age (years)	5.7 (2.8, 9.3)	4.6 (3.4, 7.0)	.518	–0.196
PVRi (WU × m2)	9.8 (7.4, 11.4)	8.6 (7.5, 9.4)	.481	–0.062
PVRi > 8 WU × m2 (n, %)	13 (59)	13 (59)	>.99	NA
Cardiopulmonary bypass (minutes)	94 (74,119)	100 (66,153)	.778	0.042

Abbreviations: NA, not applicable; PVRi: pulmonary vascular resistance index; SMD: standardized mean difference.

After propensity-score matching, the postoperative recovery conditions of the two groups of patients were shown in Table 3. In addition, no major complications, such as re-intubation, emergency sternal reopening, extracorporeal membrane oxygenation support, renal replacement therapy, or death were observed in either group.

OPEN IN VIEWER

Table 3.

The Postoperative Recovery Conditions of the Two Groups of Patients.

Variables	Group 1 (n = 22)	Group 2 (n = 22)	P Value
Pulmonary hypertensive crisis (n, %)	2 (9.1)	3 (13.6)	>.99
Mechanical ventilation (hours)	18 (8.5, 22.3)	21.5 (19.8, 27.3)	.014
Postoperative ICU stay (days)	2 (1, 4)	3 (3, 5)	.023
Postoperative hospital stay (days)	9 (7, 14.3)	8 (7, 15.5)	.82

Abbreviation: ICU, intensive care unit.

Discussion

This study demonstrates that administration of iNO via HFNC following extubation is noninferior to conventional iNO delivery via endotracheal intubation in pediatric patients with pulmonary hypertension following corrective surgery for systemic-to-pulmonary artery shunt congenital heart disease. Transition to HFNC-based iNO following extubation is safe and is not associated with an increased risk of postoperative complications. Compared with the standard weaning strategy, the fast-weaning strategy was associated with reduction in median mechanical ventilation duration and postoperative ICU stay. However, it should be noted that although postoperative recovery was faster, the magnitude of improvement was modest, and the clinical significance may be less pronounced than the statistical significance. As this study was a retrospective analysis, its conclusions require further validation by studies with a higher level of evidence.

Our results are consistent with those reported by Tominaga et al, ⁸ who observed that postextubation iNO delivery via HFNC after Fontan surgery significantly shortened the duration of mechanical ventilation. Although Fontan physiology differs from that of systemic-to-pulmonary artery shunt CHD, the underlying principle of maintaining pulmonary vasodilation during the vulnerable postextubation period remains relevant. This strengthens the inference that HFNC-iNO therapy contributes directly to shorter ventilator duration. This finding is similar to the classification in the 2022 ESC/ERS Guidelines for the diagnosis and treatment of pulmonary hypertension, which categorizes both conditions as Group 1 PAH. ¹⁶

In our cohort, the incidence of pulmonary hypertensive crisis was comparable between the two groups, suggesting that HFNC-iNO does not compromise safety. The absence of major complications such as reintubation, extracorporeal membrane oxygenation support, or death further reinforces the safety profile of this approach. Although Group 1 patients were weaned off iNO more rapidly, HFNC oxygen therapy delivers warmed, humidified oxygen at a slight positive end-expiratory pressure, enhancing alveolar recruitment, oxygenation, CO₂ clearance, and lung compliance. This reduces both the work of breathing and the need for intubation in patients with respiratory distress. ¹⁷ Moreover, HFNC markedly potentiates the response to iNO, ¹⁸ which is likely the principal safeguard of its safety profile.

The results of our study regarding the efficacy and safety of HFNC-iNO are also similar to the findings reported by Inoue et al ¹⁹ They investigated iNO therapy delivered via HFNC in adults after open-heart surgery and observed that iNO given during mechanical ventilation and continued via HFNC after extubation both effectively lowered postoperative mean pulmonary artery pressure, with no significant difference in mean pulmonary artery pressure between mechanical ventilation-iNO and HFNC-iNO. No cases of reintubation, perioperative mortality, or iNO-related adverse events were noted. The authors concluded that HFNC-iNO is a useful method for maintaining reduced mean pulmonary artery pressure and improved oxygenation following extubation in adult patients after open-heart surgery.

Several aspects of HFNC-iNO therapy require special attention. First, the use of HFNC-iNO still raises environmental safety concerns. Because NO is delivered continuously and the HFNC-iNO circuit lacks any built-in gas-scavenging system, the potential exposure of the patient, nearby medical staff, and other patients to leaked NO must be carefully managed. The iNO is rapidly oxidized by oxygen in the breathing circuit and airway to form NO₂, a toxic gas whose production increases with higher NO concentration, FiO₂, gas flow, and longer contact time. Monitoring NO₂ is critical to prevent airway and alveolar epithelial injury, ensure compliance with international safety thresholds, and verify normal function of the delivery system. In our study, we used clinically approved iNO devices with integrated monitoring to keep NO₂ within safe ranges. While ambient NO₂ was not measured, published evidence confirms that NO₂ levels during HFNC-iNO are well below safety limits. ²⁰ Furthermore, for a novel technique to be feasible, safety of both patient and staff is paramount. Training of nursing and respiratory care staff to safely manage this therapy, as well as immediate availability of medical staff to escalate care, are essential to ensure patient safety. Therefore, the use of HFNC-iNO therapy is recommended in ICUs with appropriate air ventilation equipment and manned by an exclusive staff of doctors and nurses.

In addition to systemic to pulmonary artery shunt CHD-PAH patients and Fontan patients, HFNC-iNO therapy is likely to be indicated even in other complex congenital heart diseases with postoperative pulmonary hypertension, such as tetralogy of Fallot with pulmonary atresia and major aortopulmonary collateral arteries (segmental pulmonary hypertension). ²¹ However, given the significant interindividual variability in the HFNC-iNO beneficiary population, ²² dedicated studies are warranted to further validate these findings.

Limitations of this study include its retrospective design, single-center nature, and relatively small sample size. Although propensity score matching was used, the small sample size restricted its effectiveness. This is evidenced by a persistent age disparity (standardized mean difference = −0.196) between groups. Furthermore, the characteristics of single-center study leads to limited extrapolation of the results, and there may be center-specific bias. Additionally, the lack of blinding and protocol variations over time may introduce bias. Future prospective and multicenter studies with a larger sample size are warranted to validate our present conclusions for HFNC-iNO therapy. Future studies should also evaluate the differences between HFNC-iNO and conventional strategies in terms of medical resource utilization and hospitalization costs, thereby providing an economic basis for clinical promotion. Another limitation is that systematic postoperative follow-up right heart catheterization was not performed to evaluate changes in pulmonary vascular PVRi during mid- to long-term follow-up. Therefore, the long-term impact of the two iNO weaning strategies on pulmonary vascular remodeling and PVRi recovery could not be assessed in this study. Future investigations with standardized long-term catheterization follow-up are warranted to clarify these important clinical outcomes.

Conclusion

In summary, continuing iNO via HFNC after extubation in pediatric patients with PAH following CHD repair is a well-tolerated therapeutic strategy that is not associated with an increased incidence of adverse events. This fast-weaning strategy may be associated with shorter mechanical ventilation duration and postoperative ICU stay, which requires further investigation with higher-level evidence. This approach therefore represents a promising alternative to conventional iNO weaning protocols in clinical practice.