|Home | About | Journals | Submit | Contact Us | Français|
Theoretical promise of benefit does not translate to improvements in morbidity and mortality
In this week's BMJ, Adhikari and colleagues report a systematic review of the impact of inhaled nitric oxide on physiological outcomes, morbidity, and mortality in people with acute respiratory distress syndrome.1 They found that nitric oxide resulted in a limited improvement in oxygenation but did not reduce mortality (risk ratio 1.10; 95% confidence interval 0.94 to 1.30), the duration of ventilation, or the number of days free of ventilation.
Acute respiratory distress syndrome is an important public health problem. It is a catastrophic form of acute respiratory failure that arises after pulmonary (for example, pneumonia or aspiration of gastric contents) or extrapulmonary (for example, sepsis or polytrauma) insults. Not only does it have an incidence as high as 64 cases per 100000 people per year,2 but it has a high mortality rate (30-60% in unselected populations) and risk of subsequent morbidity in survivors.3
The many cellular and molecular actions of nitric oxide (the 1992 “molecule of the year”4), also known as endothelial derived relaxing factor, are incompletely understood.5 However, intensive care physicians were quick to use inhaled nitric oxide to treat acute respiratory distress syndrome because of its immediately observable beneficial effects on oxygenation and pulmonary vascular pressures in these patients, and because of the lack of other effective treatments.
Inhaled nitric oxide acts as a selective vasodilator. It is inactivated by haemoglobin and so acts only on the pulmonary circulation, lowering pulmonary vascular resistance and improving cardiac output. As it is only delivered to lung units that are ventilated, it preferentially increases blood flow to these units, improving ventilation-perfusion matching and oxygenation. Theoretically, these actions could improve oxygenation and oxygen delivery, allow for less injurious mechanical ventilation, and ultimately reduce the prevalence or severity of the multiorgan dysfunction that is the cause of death in most patients with acute respiratory distress syndrome. Unfortunately, randomised controlled trials of inhaled nitric oxide have not borne out this theory.
Adhikari and colleagues' methodologically rigorous and clinically sound systematic review and meta-analysis of 12 randomised controlled trials provides evidence that nitric oxide can provide modest short term improvements in oxygenation. However, all but one of the nine relative risk point estimates for mortality favoured patients in control groups. This lack of concordance between physiological improvements and outcome for patients is a recurring theme in critical care,6 7 and it probably plays a part in our collective reluctance to abandon these interventions.
As Adhikari and colleagues' study is a meta-analysis that includes several small underpowered randomised controlled trials, the results are best treated as hypothesis generating.8 The review suggests that nitric oxide has no significant effect on mortality. Alternatively, and in keeping with the post hoc finding that people receiving nitric oxide had an increased risk of developing renal dysfunction (1.50, 1.11 to 2.02), we could speculate that patients with acute respiratory distress syndrome receiving nitric oxide have slightly higher mortality. What this meta-analysis does not provide is any suggestion of benefit with inhaled nitric oxide.
Given these findings, routine use of inhaled nitric oxide in such patients cannot be recommended. Indeed given the costs and potential for harm, its routine use in acute respiratory distress syndrome should be actively discouraged. I suggest that it should be used in this syndrome only for rare cases of refractory hypoxaemia, after considering modalities such as high frequency oscillation or prone positioning.9
In the light of this new evidence should we persist with trying to show that nitric oxide improves outcomes in acute respiratory distress syndrome? Not finding a benefit is not proof of no benefit. However, the trend towards harm seen across these many trials makes it unlikely that we will eventually prove a benefit through persistence alone. This does not mean that nitric oxide is without merit; it is still useful in other diseases such as persistent pulmonary hypertension of the newborn.10
Even in acute respiratory distress syndrome, nitric oxide may yet have a role outside salvage therapy for severe hypoxaemia, but this work by Adhikari and colleagues indicates that we need to change radically the designs of randomised controlled trials of nitric oxide in these patients. For example, future trials could explore alternative dosing schedules for nitric oxide (such as titrated decremental dosing11), the application of nitric oxide to specific subgroups of patients, or combining nitric oxide with other treatments (such as mechanical ventilation strategies to reduce pressure, or early aggressive fluid resuscitation12).
Future trials should be grounded in sound observations from physiological animal and early phase human studies, along with observations from completed randomised trials. Making the successful leap from a solid mechanistic theory to a large randomised controlled trial is challenging,13 and trying to do so without this grounding seems destined for failure, regardless of persistence.
Competing interests: NDF is a friend of, and collaborator with, the authors of the paper that is related to this editorial; he accepted this editorial commission only after disclosing such to the BMJ editorial staff. He did not peer review the paper.
Provenance and peer review: Commissioned; not externally peer reviewed.