Discussion

In this physiologic study of 26 patients with moderate to very severe HPS, supine positioning, Trendelenburg positioning and iNO significantly improved oxygenation without negative impacts on CO. To our knowledge, this is the largest report of any of these interventions.

The PaO₂ was 3.3 mm Hg higher in the supine compared with the seated position, and 0.9 mm Hg higher in the Trendelenburg position compared with the supine position (table 2). These position changes are theorised to work through diversion of pulmonary blood flow away from pathologically dilated pulmonary vessels at lung bases, reducing shunt and V/Q mismatch.2 7 Although improvements in PaO₂ were small, it is more relevant to consider the potential impact on tissue oxygen delivery (DO₂). DO₂ is proportional to oxygen saturation, which is dependent on PaO₂ through the sigmoidal oxyhaemoglobin dissociation curve. Patients with severe post-transplant hypoxaemia have saturations of 80%–85% or lower,5 corresponding to a baseline PaO₂ ≤45 mm Hg. On this steep portion of the oxyhaemoglobin dissociation curve, even small positional PaO₂ improvements could result in clinically significant improvements in saturation and DO₂. These benefits would need to be weighed against the risk of aspiration and ventilator-associated pneumonia. In routine care, Trendelenburg positioning might also be considered in an HPS exercise programme and/or as a sleep positional strategy to reduce overnight oxygen requirements.

Inhaled NO was the most effective intervention, demonstrating a mean PaO₂ improvement of 12.4 mm Hg and responses of ≥10% in over half of patients (table 2). Given that basilar lung vessels may already be maximally dilated,8 vasodilatory effects of iNO are thought to be more effective in mid and upper lung units, whereby flow is redistributed away from basilar lung units, improving V/Q matching and oxygenation. The same mechanism likely explains observed benefits of garlic in HPS, the active compound of which (allicin) is a potent vasodilator.9 Unfortunately, we were not able to formally assess regional changes in lung perfusion in order to prove this theory. Overall, our findings suggest that a trial of either iNO or a comparable inhaled vasodilator for severe post-transplant hypoxaemia2 (and/or for intra-operative hypoxaemia or as a bridge to transplant) would be reasonable. Ambulatory iNO, longer-acting inhaled vasodilators such as inhaled iloprost,10 and systemic vasodilators might also be evaluated as chronic medical therapies for HPS.

MB infusion did not produce changes in PaO₂ (table 2). As a potent vasoconstrictor, MB is thought to act by directly vasoconstricting dilated vessels at lung bases to improve V/Q matching, and by restoring impaired hypoxic vasoconstriction in poorly ventilated areas.2 There was a significant increase in MPAP but no significant increase in PVR with MB (table 1, online supplemental table 3). In a previous study of seven subjects, Schenk et al6 reported improvement in oxygenation from 58 mm Hg to 74 mm Hg with MB.6 Divergence in our results may be attributable to our more severe disease population (baseline PaO₂ 46 mm Hg), which may have been less MB responsive due to end-stage vascular remodelling with resulting vasoplegia of dilated HPS vessels.11 This is supported by Schenk, et al’s reported significant PVR increase with MB,6 compared with a non-significant change in our cohort. Furthermore, this group reported a significant correlation between improvement in oxygenation and increase in PVR.6 However, both Schenk, et al’s report and our own demonstrate variable PaO₂ responses, likely reflective of a heterogeneous population, with some but not all patients having preserved vasoconstrictor responses. Accordingly, a trial of MB may still be merited in severe post-transplant hypoxaemia, as long as impact on CO (and thus DO₂) is monitored. We also note that adding MB to iNO did not significantly improve PaO₂ compared iNO alone. However, responses were again variable and there was a 13% increase in CO compared with iNO alone, suggesting a DO₂ benefit. Accordingly, we believe that a trial of combined inhaled vasodilator and MB may be merited in severe situations.

Our study has several limitations. The iNO was delivered with FiO₂ of 0.4 as opposed to room air. However, this more closely approximates the real-world scenario in which this salvage therapy would be used, and effects of all interventions were determined based on comparison to a baseline on the same FiO₂. Although much larger than any prior reports, our sample size remains small and may have been underpowered to detect a small MB effect. We did not pursue molecular biomarker testing to confirm or refute the proposed mechanisms of action of reported agents. We also have not prospectively evaluated the effectiveness nor sustainability of these strategies in the post-transplant context.

In summary, we demonstrated that iNO significantly improves oxygenation in severe HPS, without deleterious effects on CO. Although more data are needed, it may be considered early in the management of severe hypoxaemia, whether as a bridge to transplant, during transplant or after. Longer-acting vasodilators require study as possible maintenance therapy for HPS. Small but significant effects were seen with Trendelenburg positioning, but not with MB. However, observed individual variability in effects suggests that pre-transplant testing of all agents in high-risk patients may be helpful in guiding post-transplant management in the event of severe hypoxaemia. Future studies are required to prospectively evaluate the effect of iNO in patients with severe post-transplant hypoxaemia and measure correlations between pre-transplant and post-transplant responses.