Inhaled Nitric Oxide for Acute Lung Injury: Title and subTitle BreakNo Place for NO?

It is estimated that up to 107 000 cases of acute lung injury (ALI) and acute respiratory distress syndrome (ARDS) occur yearly in the United States.¹ Although the mortality rate from ARDS has been decreasing, between 17 000 and 43 000 patients die from ALI and ARDS annually.² Despite substantial progress in understanding the pathophysiology of this syndrome and numerous randomized controlled trials (RCTs) of both pharmacological agents and mechanical ventilatory strategies, the only interventions shown to reduce mortality have been lung-protective strategies of mechanical ventilation that decreased tidal volume.^{3 - 5} In the ARDS Network trial, tidal volume limitation was associated with a clinically and statistically significant 22% relative reduction in mortality for patients with ALI and ARDS.⁴ Central to the success of these trials is the concept that mechanical ventilation itself perpetuates lung injury. Although this concept has led to positive studies of interventions to decrease iatrogenic lung damage complicating ARDS, no multicenter trial has convincingly demonstrated benefit from interventions such as surfactants, corticosteroids, ketoconazole, or other therapies that treat the underlying pathogenic mechanisms or physiological perturbations that characterize ALI and ARDS.³

In the context of an empty therapeutic armamentarium for lung injury, considerable research efforts and clinical enthusiasm greeted the discovery of nitric oxide (NO).^{6 - 10} From a clinical perspective, inhaled nitric oxide has many of the properties of an ideal pulmonary vasodilator. It acts on the endothelial surface of the lung to produce regional vasodilation in ventilated lung units. As a result, nitric oxide was expected to produce relatively selective pulmonary vasodilation that in theory would improve the ventilation-perfusion mismatch, hypoxemia, and pulmonary hypertension characteristic of ARDS. The short half-life of nitric oxide suggested that it was free of significant systemic hemodynamic effects. Given the potential role of nitric oxide as a pulmonary vasodilator, its use as a therapeutic agent increased dramatically and rapidly. A survey of members of the European Society of Intensive Care Medicine in 1997 found that 63.2% of respondents (196/310) prescribed inhaled nitric oxide , primarily for treatment of ALI and ARDS.⁹

Consistent with the role of nitric oxide as a pulmonary vasodilator, earlier case series^{10 - 11} and a single-center RCT¹² demonstrated that it induces short-term improvements in oxygenation and pulmonary vascular resistance in patients with lung injury. Unfortunately, these physiologic gains have not translated into improved patient outcomes when nitric oxide was evaluated in larger controlled studies. A recent systematic review¹³ of 5 RCTs^{14 - 18} of inhaled nitric oxide vs placebo or no therapy for acute hypoxemic respiratory failure (including ALI, ARDS, and other diagnoses) in adults and children concluded that inhaled nitric oxide produced modest improvements in oxygenation for up to 72 hours but had no effect on mortality (pooled relative risk using fixed effects model, 0.98; 95% confidence interval, 0.66-1.44) or on the duration of mechanical ventilation.

In addition to direct effects on gas exchange, nitric oxide has potentially beneficial immunological properties; it inhibits platelet aggregation and neutrophil adhesion.^{6 - 8} The anti-inflammatory properties of nitric oxide have been postulated to explain decreased ischemia-reperfusion lung injury in animal models and in small observational studies of lung transplantation.^{19 - 21} However, a recent RCT of nitric oxide administered immediately after lung transplantation failed to demonstrate an improvement in either oxygenation variables or clinical outcomes.²²

In this issue of THE JOURNAL, Taylor and colleagues²³ make an important contribution to this literature. Between 1996 and 1999, the investigators enrolled 385 patients with moderately severe ALI and ARDS in a randomized placebo-controlled multicenter trial of inhaled nitric oxide at 5 ppm. The rationale for this dose was based on an earlier placebo-controlled RCT that evaluated several doses of nitric oxide (1.25, 5, 20, 40, or 80 ppm) in 177 patients with ARDS.¹⁵ Although there was no clinical benefit in the nitric oxide group overall in that study, a post-hoc analysis suggested that patients receiving nitric oxide at 5 ppm had significantly improved ventilation-free survival 28 days after randomization.

Patients in the current study by Taylor et al²³ met modified American-European Consensus Conference criteria for ALI (PaO₂/FiO₂ ratio ≤250 rather than ≤200) and could be enrolled up to 72 hours after meeting ALI criteria.²⁴ Patients with nonpulmonary organ dysfunction and sepsis, for whom improvements in pulmonary function might not improve survival and duration of mechanical ventilation, were excluded. The study design incorporated methods to maximize internal validity: centralized randomization to conceal the allocation schedule; convincing procedures to blind clinicians, data collectors, and outcomes assessors; complete follow-up; and inclusion of all randomized patients in the analysis. Although the treatment groups were balanced at baseline, mean pulmonary artery pressure was significantly lower in the placebo group; however, this difference was small and is unlikely to have confounded the results.

The major goals for mechanical ventilation settings were established a priori, but actual ventilator settings were not recorded. The goals of mechanical ventilation included limitation of inspiratory plateau pressure to 35 cm H₂O. However, there was no limitation placed on the tidal volumes used during the study. Other cointerventions were applied similarly between the groups (prone positioning) or not used at all (high-frequency oscillatory ventilation). Although the use of "rescue" therapies and other aspects of critical care were not standardized, differential application of cointerventions is unlikely to have confounded the results given the presence of effective caregiver blinding.

What do the results of this trial add to the literature? Oxygenation improved during the first day in patients receiving inhaled nitric oxide, but this effect did not persist. This finding is consistent with results of previous trials, although the absolute degree of improvement cannot be directly compared across studies because the current trial reports only the effects of nitric oxide on the partial pressure of arterial oxygen. More commonly, trials have reported the PaO₂/FiO₂ ratio or the oxygenation index (defined as mean airway pressure divided by PaO₂/FiO₂ × 100), since these measures incorporate both an absolute measure of oxygenation and the intensity of therapy required to obtain that level of oxygenation. Except for this temporary physiological effect, in this RCT nitric oxide did not improve any clinically important outcome such as mortality or ventilator-free days (the number of days alive and off ventilation to day 28 after randomization). Overall, the number of adverse events was similar between groups, but more nosocomial infections occurred in the nitric oxide group.

The results of this trial consolidate earlier findings and support the notion that nitric oxide is not useful in the treatment of the majority of patients with ALI or ARDS. There are likely several explanations for this result. First, oxygenation may have little bearing on survival, a point supported by the observation that most patients with ALI and ARDS do not die of refractory hypoxemia but rather due to multiple organ system failure.²⁵ Consequently, therapy that only serves to improve oxygenation is unlikely to have a significant impact on mortality.¹³ Although ventilator-induced lung injury likely perpetuates multiple organ system failure, there is little evidence from RCTs that nitric oxide significantly alters the intensity of mechanical ventilation. Dobyns et al¹⁴ and Dellinger et al¹⁵ both reported a significantly improved oxygenation index with nitric oxide; however, this effect appeared to be related to an improved PaO₂/FiO₂ ratio rather than a change in mean airway pressure. No trial has examined whether nitric oxide use might allow for lower tidal volumes as a component of a lung-protective strategy. Thus, the notion that a small and temporary improvement in an intermediate physiologic outcome (oxygenation) would lead to improved clinical outcomes is likely flawed. This concept is supported by the ARDS Network trial, in which tidal volume limitation worsened short-term oxygenation but improved survival.⁴

A second possible explanation for the failure of nitric oxide to improve clinical outcome is that its benefit in clinical trials was masked by the overwhelming negative effect of ventilator-induced lung injury. Although most trials limited plateau pressure, none adopted a lung-protective strategy incorporating a low tidal volume as the standard of care. This distinction is important because there is no safe level of plateau pressure below which tidal volume limitation is not beneficial.²⁶

Third, it is possible that the long-term application of nitric oxide caused harm, overshadowing the acute physiological benefit, especially considering that nitric oxide might worsen oxygenation as therapy continues in the absence of dose titration. In a nonblinded trial, Gerlach et al²⁷ randomly assigned 40 patients with ARDS to receive either inhaled nitric oxide at 10 ppm or no therapy. Dose-response curves at baseline and at 2 and 4 days after starting therapy revealed that 10 ppm of nitric oxide produced maximal oxygenation benefit at baseline and day 2, but that peak oxygenation at day 4 was associated with a much smaller dose of 1 ppm. Thus, prolonged fixed dosing regimens may increase sensitivity to nitric oxide, and the lack of clinical benefit observed in the current study by Taylor et al²³ and other RCTs may relate to a left shift of the oxygenation-dose response curve after several days of continuous therapy. Of course, whether this potentially deleterious effect on longer-term oxygenation produces clinical adverse events is unknown. Nevertheless, there is some evidence from large RCTs to suggest that nitric oxide may cause clinically significant harm. In the study by Lundin et al,¹⁷ patients in the nitric oxide group had a higher incidence of renal failure and in the current study patients in the nitric oxide group had a higher incidence of nosocomial infection.

Finally, patients in the current trial²³ may have been randomized too late after the onset of lung injury. Indeed, if the potential clinical benefits of nitric oxide in lung injury are related to an anti-inflammatory mechanism, earlier initiation of therapy may be required, similar to therapy for sepsis. None of the previous trials explicitly reported patient enrollment within 24 hours of meeting criteria for lung injury.^{14 - 18}

Although nitric oxide is not effective for established ALI and ARDS, it is an important therapy in other clinical situations in which the objective is to reduce pulmonary arterial pressure and improve right ventricular function rather than improve ventilation-perfusion mismatching. Nitric oxide has been advocated as a pulmonary vasodilator in the treatment of patients after heart transplantation; with idiopathic pulmonary hypertension, pulmonary hypertension related to chronic obstructive pulmonary disease, pulmonary hypertension related to left ventricular failure during transplant assessment; and postoperatively after congenital cardiac repairs.^{28 - 32} However, evidence for meaningful patient benefit beyond demonstrable physiological improvements for these applications of nitric oxide is lacking. At present the only application of nitric oxide shown to be clinically effective is in the treatment of respiratory failure in newborns. Randomized trials have shown nitric oxide to reduce the need for extracorporeal membrane oxygenation in persistent pulmonary hypertension of the newborn and reduce mortality and chronic lung disease in neonatal respiratory distress syndrome.^{33 - 34}

In conclusion, current cumulative evidence from clinical trials suggests that nitric oxide has no place in the routine therapy of patients with ALI or ARDS. Consequently, this expensive intervention cannot be recommended as standard care for these patients. However, in severe cases of ARDS for which refractory hypoxemia or pulmonary hypertension are major clinical problems, short-term physiological improvements may be crucial for patient survival. In these limited situations, nitric oxide may have a role as "salvage" therapy, as a component of a multimodal approach that includes other strategies such as high-frequency oscillation or prone positioning.³⁵