Customize your JAMA Network experience by selecting one or more topics from the list below.
Taylor RW, Zimmerman JL, Dellinger RP, et al. Low-Dose Inhaled Nitric Oxide in Patients With Acute Lung Injury: A Randomized Controlled Trial. JAMA. 2004;291(13):1603–1609. doi:10.1001/jama.291.13.1603
Author Affiliations: Critical Care Medicine, St Louis University/St John's Mercy Medical Center, St Louis, Mo (Dr Taylor); Department of Medicine, Baylor College of Medicine and Ben Taub General Hospital, Houston, Tex (Dr Zimmerman); Critical Care Medicine, Robert Wood Johnson Medical School–UMDNJ, Cooper University Hospital, Camden, NJ (Dr Dellinger); INO Therapeutics Inc, Clinton, NJ (Dr Straube and Mr Small); Pulmonary and Critical Care Medicine, Temple Lung Center, Temple University Hospital, Philadelphia, Pa (Dr Criner); Department of Surgery, University of Cincinnati Medical Center, Cincinnati, Ohio (Dr Davis); Department of Clinical Research, Ortho Biotech, Bridgewater, NJ (Dr Kelly); and Department of Pulmonary Medicine, Albany Medical College, Albany, NY (Dr Smith).
Caring for the Critically Ill Patient Section Editor: Deborah J. Cook, MD, Consulting Editor, JAMA.
Context Inhaled nitric oxide has been shown to improve oxygenation in acute
Objective To evaluate the clinical efficacy of low-dose (5-ppm) inhaled nitric
oxide in patients with acute lung injury.
Design and Setting Multicenter, randomized, placebo-controlled study, with blinding of
patients, caregivers, data collectors, assessors of outcomes, and data analysts
(triple blind), conducted in the intensive care units of 46 hospitals in the
United States. Patients were enrolled between March 1996 and September 1999.
Patients Patients (n = 385) with moderately severe acute lung injury, a modification
of the American-European Consensus Conference definition of acute respiratory
distress syndrome (ARDS) using a ratio of PaO2 to FiO2 of
≤250, were enrolled if the onset was within 72 hours of randomization,
sepsis was not the cause of the lung injury, and the patient had no significant
nonpulmonary organ system dysfunction at randomization.
Interventions Patients were randomly assigned to placebo (nitrogen gas) or inhaled
nitric oxide at 5 ppm until 28 days, discontinuation of assisted breathing,
Main Outcome Measures The primary end point was days alive and off assisted breathing. Secondary
outcomes included mortality, days alive and meeting oxygenation criteria for
extubation, and days patients were alive following a successful unassisted
Results An intent-to-treat analysis revealed that inhaled nitric oxide at 5
ppm did not increase the number of days patients were alive and off assisted
breathing (mean [SD], 10.6 [9.8] days in the placebo group and 10.7 [9.7]
days in the inhaled nitric oxide group; P = .97;
difference, –0.1 day [95% confidence interval, –2.0 to 1.9 days]).
This lack of effect on clinical outcomes was seen despite a statistically
significant increase in PaO2 that resolved by 48 hours. Mortality
was similar between groups (20% placebo vs 23% nitric oxide; P = .54). Days patients were alive following a successful 2-hour unassisted
ventilation trial were a mean (SD) of 11.9 (9.9) for placebo and 11.4 (9.8)
for nitric oxide patients (P = .54). Days alive and
meeting criteria for extubation were also similar: 17.0 placebo vs 16.7 nitric
oxide (P = .89).
Conclusion Inhaled nitric oxide at a dose of 5 ppm in patients with acute lung
injury not due to sepsis and without evidence of nonpulmonary organ system
dysfunction results in short-term oxygenation improvements but has no substantial
impact on the duration of ventilatory support or mortality.
Inhaled nitric oxide has been shown to be a selective pulmonary vasodilator
with minimal systemic effects.1,2 Nitric
oxide has been shown to improve outcome, as measured by the need for extracorporeal
membrane oxygenation, in persistent pulmonary hypertension of the newborn.3,4 Inhaled nitric oxide has also been
shown to improve gas exchange both in animal models of acute respiratory distress
syndrome (ARDS)5-8 and
in humans.9-16 Two
single-center studies17,18 demonstrated
the ability of inhaled nitric oxide to improve oxygenation in ARDS patients
with no difference in clinical outcome. Three multicenter, randomized, placebo-controlled
failed to demonstrate an impact on mortality. In one of these studies,19 fixed doses of nitric oxide at 0, 1.25, 5, 20, and
40 ppm were given to patients with ARDS from causes other than severe sepsis.
In that study, non–statistically significant decreases were noted in
both the intensity of mechanical ventilation (oxygenation index) and the duration
of mechanical ventilation in the 5-ppm dose group. On the basis of that subgroup
analysis, the current multicenter, randomized, blinded, placebo-controlled
trial was initiated.
Patients in intensive care units were enrolled from 46 academic, teaching,
and community hospitals in the United States. The study was approved by the
institutional review board at each participating hospital. Written informed
consent was obtained from each patient or his or her legal representative
Eligible patients had moderately severe acute lung injury due to causes
other than severe sepsis, using a modification of the American-European Consensus
Conference definition of ARDS (a ratio of PaO2 to fraction of inspired
oxygen [FiO2] of ≤250 instead of ≤200).22 Because
inhaled nitric oxide was expected to affect only the lung, study entry criteria
were established to exclude patients in whom poor outcome and duration of
mechanical ventilation were unlikely to be altered by improvements in oxygenation.
Therefore, patients with evidence of nonpulmonary system failure at the time
of randomization and sepsis-induced ARDS were excluded. Patients with sustained
hypotension, vasopressor support with evidence of high-output failure, severe
head injury, severe burns, or evidence of other organ system dysfunction (renal,
hepatic, thrombocytopenia, and disseminated intravascular coagulopathy) were
excluded. Entry criteria are presented in Box 1.
1. Nonpregnant adults (≥18 years)
2. Developed ALI within the preceding 72 hours as defined as:
PaO2/FiO2 ≤250, regardless of the amount of PEEP
infiltrates on frontal chest radiograph
artery occlusion pressure ≤18 mm Hg when measured or no clinical evidence
3. ALI resulting from at least 1 of the following:
Toxic gas inhalation
blood transfusion (including transfusion reactions)
Elective or emergency major surgery
4. FiO2 of0.50-0.95 or a set PEEP ≥8 cm H2O
1. History of immunocompromise,
Received chemotherapy or radiation therapy
within the last 30 days
≥20 mg of prednisone or
equivalent for ≥30 days
≥50 mg of prednisone
or equivalent continually for >10 days within the last 30 days
AIDS (human immunodeficiency virus–positive
patients could be entered into the study provided they
lavage results negative for Pneumocystis carinii)
2. Persistent systemic hypotension, defined as systolic blood pressure
<90 mm Hg, or a nonpurposeful reduction of systolic pressure by ≥40
mm Hg; patients who had severe sepsis or a systolic blood pressure >90 mm
Hg but were receiving >5 µg/kg per minute of dopamine (or equivalent)
and who met any of the following conditions within 4 hours before the initiation
of treatment gas:
Systemic vascular resistance <800
dynes·sec·m3 and an elevated cardiac index >4 L/m2/min
White blood cell count >20 000/µL
Urine output <0.5 mL/kg/h
for 1 hour
3. Evidence of nonpulmonary organ dysfunction, defined as 1 or more
of the following:
Creatinine ≥1.5 mg/dL (132.60
Total bilirubin ≥4.0 mg/dL (68.40
µmol/L) and aspartate aminotransferase or alanine aminotransferase
>2 times the upper limit of normal
≤50 × 103/µL
time ≥1.5 times the upper limit of normal
Abbreviations: ALI, acute lung injury; FiO2, fraction of
inspired oxygen; PEEP, positive end-expiratory pressure.
Patients were randomly assigned to receive either inhaled placebo gas
(nitrogen) or 5 ppm of nitric oxide (INO Therapeutics Inc, Port Allen, La)
(Figure 1). Patients received gas
labeled only with a study code and without designation of contents. The trial
used concealed allocation, with randomization occurring centrally at the manufacturing
plant. Patient numbers were preassigned sequentially by site (ie, site 01,
patient 001, 002, etc). Drug cylinders were labeled to identify the patient
number without revealing the contents. All drug cylinders were prepared and
labeled before shipment to an investigative site by a research pharmacist
at the sponsor's manufacturing facility. No individual at any clinical site
had a copy of the randomization code before analysis. All patients, clinicians
(physicians, nurses, and respiratory care practitioners), and investigators
were blinded to treatment assignment. The monitors on the inhaled nitric oxide
delivery system were covered with a locked metal device that was opened only
if the high-dose nitric oxide or nitrogen dioxide alarm sounded. Each site
had a separate laboratory investigator team not involved in patient care that
was responsible for the monitoring and recording of methemoglobin levels and
nitric oxide and nitrogen dioxide alarms. Alarm episodes were infrequent (3
episodes reported) during the trial.
The inhaled nitric oxide was delivered through a commercially available
delivery system (INOvent; Datex-Ohmeda, Madison, Wis) that blended the treatment
gas (nitrogen or nitric oxide at 100-ppm balance nitrogen) 1:20 with the ventilator
gases to deliver a target parts per million value into the inspiratory limb
of the ventilator. An analysis in the inspiratory circuit immediately before
patient treatment ensured that the delivered concentration of gas was accurate.
Nitric dioxide was not removed. Continuous monitoring of nitric oxide, nitrogen
dioxide, and FiO2 concentrations occurred at the distal inspiratory
limb. All patients received ventilatory support while using the inhaled nitric
oxide delivery system.
All patients continued treatment with active or placebo gas until the
end of the trial (28 days), death, or adequate oxygenation was achieved. Adequate
oxygenation was defined as pulse oximetry oxygen saturation of 92% or more
or PaO2 of 63 mm Hg or more (PaO2 took precedence when
both values were known), without treatment gas at ventilator settings of an
FiO2 of 0.4 or less, and a positive end-expiratory pressure (PEEP)
of 5 cm H2O or less. As long as these oxygenation criteria were
met, decreases in treatment gas continued in 20% decrements (titrated down
by 1 ppm if inhaled nitric oxide was being administered) every 30 minutes
until either the treatment gas concentration was decreased to 0% or the oxygenation
criteria were not satisfied. If the latter occurred, the treatment gas was
titrated up until oxygenation criteria were reachieved. Clinicians determined
increments of upward titration based on degree of desaturation. If 0% treatment
gas was tolerated for 24 hours, treatment gas was permanently discontinued.
If procedures were required outside the intensive care unit and treatment
gas could be reinstituted within 24 hours, patients were continued in the
study. If not, they were classified as premature discontinuations.
The investigators participating in the trial agreed to guidelines for
prioritizing the mechanical ventilation settings as detailed in Box 2. No other management guidelines were provided to investigators.
Initially institute positive end-expiratory pressure (PEEP) to optimize
compliance (usually 8-12 cm H2O) and to prevent shear force injury
Decrease inspiratory plateau pressure to ≤35 cm H2O
(this level achieves total lung capacity in healthy patients)
Decrease FiO2 to ≤0.60 (to minimize theoretical concern
for oxygen toxicity)
Decrease FiO2 to ≤0.40 and decrease PEEP to 5 cm H2O (allowing extubation from an oxygenation criteria standpoint, one
of the study's secondary end points)
*These recommendations are expert opinion as proposed by the clinical
advisory committee and refined at the investigator meeting prior to study
Oxygenation and ventilation parameters were recorded at baseline, 4
hours, and 12 hours after initiation of study gas and then every 12 hours
thereafter for the 28-day study period. Methemoglobin levels were measured
at baseline, 30 minutes, 4 hours after initiation of the study gas, and then
every other day while patients received the treatment gas. Chest radiographs
were obtained at baseline and then at days 7, 14, 21, and 28 while the patient
was hospitalized. Complete blood cell counts and serum biochemistry values
were collected at baseline and then on days 1, 3, 5, 14, 21, and 28.
The prospectively defined primary efficacy end point for this trial
was the duration of mechanical ventilation measured by number of days patients
were alive and not receiving assisted breathing, defined as the time of extubation
(≥72 hours) or the reduction of both pressure support and continuous positive
airway pressure to 5 cm H2O or less in patients with tracheostomies.
To avoid the misclassification of patients with short duration of mechanical
ventilation due to death, the end point was measured as the number of days
alive from the time the patient was both alive and not receiving assisted
breathing to the end of the 28-day study. Secondary end points were mortality,
days patients were alive and meeting oxygenation criteria for extubation,
and days patients were alive following a successful unassisted ventilation
Esteban et al23 suggested that testing
patients who received mechanical ventilation for their ability to maintain
spontaneous breathing off the ventilator (2-hour unassisted ventilation test)
was associated with a high likelihood of successful extubation. Each patient
in this trial who met oxygenation criteria for extubation was assessed daily
with an unassisted ventilation test to determine if he or she could breathe
without mechanical support. Oxygenation criteria, determined by a panel of
critical care clinicians participating in this study, were prospectively established
as representing criteria that would make a patient a candidate for extubation.
These criteria included an FiO2 of 0.40 or less, a PaO2 of
60 mm Hg or more, and a PEEP of 5 cm H2O or less in a patient no
longer receiving treatment gas. Because inhaled nitric oxide would not be
expected to influence ventilatory capability or airway protection, these criteria
would be more relative to inhaled nitric oxide effect.
For the purposes of this trial, the following criteria were used to
establish a diagnosis of the etiologies of acute lung injury: (1) pneumonia:
pulmonary infiltrates thought to be due to primary lung infection, fever,
and/or leukocytosis and a sputum Gram stain with more than 25 white blood
cells and less than 10 epithelial cells per low-power field; (2) aspiration:
witnessed or clinical history compatible with aspiration of gastric contents;
(3) pulmonary contusion: pulmonary infiltrates that appear within 24 hours
of blunt trauma to the chest; (4) acute pancreatitis: clinical syndrome consistent
with pancreatitis associated with increased serum amylase and lipase concentrations;
(5) massive blood transfusion of 10 U or more; (6) postpartum acute lung injury
occurring within 72 hours of delivery without evidence of sepsis or cardiac
dysfunction; and (7) acute lung injury associated with surgical procedure:
patients fitting trial definition of acute lung injury who had undergone a
surgical procedure with no other cause of acute lung injury identified.
Nonpulmonary organ system dysfunction was defined by 1 or more of the
following: creatinine, 1.5 mg/dL or more (≥132.60 µmol/L); total
bilirubin, 4.0 mg/dL or more (≥68.40 µmol/L) with an aspartate aminotransferase
or alanine aminotransferase level more than 2 times the upper limit of normal;
platelets, 50 × 103/µL or less; or prothrombin time,
at least 1.5 times the upper limit of normal.
An intention-to-treat analysis was performed. Continuous variables were
compared using either the t test or, if the distribution
of the variable was not normal, the Wilcoxon rank sum test. Categorical variables
were compared using the Fisher exact test. Variables are reported as mean
(SD). No interim analyses were planned or performed. The level of statistical
significance was prospectively set at P≤.05. The
statistical software used was SAS version 6.12 (SAS Institute Inc, Cary, NC).
The sample size determination was based on the following assumptions
derived from data generated in previous clinical trials: (1) the desired type
I error of .05 was the threshold for statistical significance (2-tailed);
(2) the difference in the number of days alive without assisted breathing
was at least 3.5 days; (3) the standard deviation of the mean number of days
alive without assisted breathing was 9.54 days and this was the same in the
placebo and treatment arms; and (4) the desired power (1–β) for
the trial was 80%. Using these assumptions, the calculated minimum sample
size for each of 2 identical but separate trials was determined to be approximately
258 patients. Before completion of the 2 trials, the decision was made by
the clinical advisory committee to recommend prospective merger of the databases
and shorten both trials at a combined sample size of 385. If this trial had
produced positive results and the US Food and Drug Administration had required
a second trial for approval, a second trial would have been performed.
Between March 1996 and September 1999, 385 patients (193 placebo, 192
inhaled nitric oxide) were enrolled at 46 sites. No patients were lost to
follow-up or withdrawn from the study. All patients had complete data collected
until death, discharge, or end of study. Total protocol violations were similar
between treatment groups. Major protocol violations occurred in 31 patients
and more frequently in placebo patients (22 placebo, 9 inhaled nitric oxide).
Most frequent violations were a PaO2/FiO2 ratio greater
than 250, intubation for more than 72 hours, and unilateral pulmonary infiltrates.
Results are presented for the intent-to-treat group, but findings were similar
for evaluable patients.
The primary causes of ARDS and baseline characteristics of patients
in the treatment groups are shown in Table
1. The groups were well balanced with respect to the primary causes
of ARDS and baseline respiratory dysfunction. A statistically significant
higher mean pulmonary artery pressure was noted in the inhaled nitric oxide
group at baseline (P = .02). Glucocorticoids were
administered in 15% of placebo and 16% of nitric oxide patients at randomization.
A pulmonary artery catheter was used in 54% of placebo patients and 57% of
nitric oxide patients. No patients received extracorporeal membrane oxygenation
or high-frequency oscillation. During the study, prone positioning was performed
in 7.3% of placebo and 5.7% of nitric oxide patients.
The primary outcome variable, days patients were alive and not receiving
assisted breathing to day 28, was not different in the placebo and intervention
groups (mean [SD], 10.6 [9.8] vs 10.7 [9.7] days; P =
.97; difference, –0.1 day [95% confidence interval, –2.0 to 1.9]).
The results for all of the efficacy variables are presented in Table 2. There was no statistically significant difference in mortality
between treatment groups: the 28-day mortality rate was 20% (39/193 patients)
in the placebo group and 23% (44/192 patients) in the inhaled nitric oxide
group (P = .54). There were no significant differences
in any other secondary outcome between groups. Changes in PaO2 and
PEEP over time for the 2 groups are shown in Figure 2. There was a statistically significant increase in the
group means during the initial 24 hours that resolved by 48 hours.
A total of 1296 adverse events were reported to have occurred in these
critically ill patients (630 in the inhaled nitric oxide group and 666 in
the placebo group). There was 41 infections reported in the placebo group
and 66 in the inhaled nitric oxide group. None of the infections was judged
by blinded investigators to have been related to treatment gas administration.
The total number of cardiovascular, gastrointestinal, endocrine, hematologic,
metabolic and nutritional, and nervous system adverse events were similar
in the treatment groups. Respiratory system adverse events were more frequent
in the placebo group (61% vs 51% in the nitric oxide group). This difference
resulted from an increased number of placebo patients with pneumonia (20%
vs 16%), pneumothorax (16% vs 13%), and apnea (7% vs 4%). There was no difference
in the percentage of patients developing any elevations of creatinine (≥3.0
mg/dL [265.2 µmol/L]; 4% placebo vs 6% inhaled nitric oxide) or in patients
with markedly elevated creatinine (≥3.5 mg/dL [309.4 µmol/L]; 3%
placebo vs 5% inhaled nitric oxide).
As expected, none of the inhaled nitric oxide patients had clinically
relevant levels of methemoglobin. One patient in the placebo group had a methemoglobin
level higher than 5%. No nitrogen dioxide levels above 2 ppm were reported.
The incidence of air leak syndrome (pneumothorax, pneumomediastinum, pneumopericardium)
was 21% in both treatment groups. The hematologic and clinical chemistry values
and changes from baseline values were similar between groups.
This trial assessed the effects of 5 ppm of inhaled nitric oxide in
patients with ARDS and severe acute lung injury, defined as a PaO2/FiO2 ratio of less than 250. As seen in previous trials, the addition of
inhaled nitric oxide induced a rapid improvement in the oxygenation of these
patients, which was maintained for 24 hours. It was not associated with any
clinically relevant change in patient outcomes measured by days alive without
assisted breathing, the number of patients alive and not using assisted breathing
at day 28, the days alive after a successful 2-hour unassisted ventilation
trial, days alive after reaching oxygenation criteria, or mortality. The difference
in mean pulmonary artery pressure at baseline, although statistically significant,
is not clinically relevant and unlikely to have influenced response. Although
PEEP-induced lung recruitment will influence the effect of inhaled nitric
oxide on oxygenation,24 this protocol did not
address PEEP interaction with inhaled nitric oxide to optimize oxygenation
The lack of correlation between oxygenation changes and long-term outcome
in our study was also seen in the ARDS Network study of low and traditional
tidal volumes. In that trial of 861 patients with ARDS, those receiving the
6 mL/kg of predicted body weight tidal volume during mechanical ventilation
had lower mean PaO2 than those randomized to receive the higher
tidal volume but had a statistically lower mortality rate.25
The lack of clinical outcome benefit from the use of inhaled nitric
oxide in this general population of patients with non–sepsis-induced
ARDS is consistent with the results reported in smaller studies and large,
randomized trials.19-21 We
previously19 described 176 patients with non–sepsis-induced
ARDS who received inhaled nitric oxide at doses ranging from 1.25 to 40 ppm.
We found short-term improvements in oxygenation, but there was no benefit
of pooled inhaled nitric oxide vs placebo on duration of mechanical ventilation,
hospital stay, or mortality.
Lundin et al20 assessed the effects of
inhaled nitric oxide in an open-label study of 180 patients with acute lung
injury who responded to a test dose of inhaled nitric oxide of 2, 10, or 40
ppm with a 20% increase in PaO2. Inhaled nitric oxide was administered
at the lowest effective dose observed for each patient during the testing
phase of the study. Although the development of severe respiratory failure
was lower in the inhaled nitric oxide group, there was no difference between
groups with respect to the course of reversal of acute lung injury, the number
of patients alive and not receiving mechanical ventilation, or mortality.
Finally, the Groupe d'Etude sur le NO inhalé au cours de l'ARDS (GENOA)
trial, published in abstract form, reported physiologic improvements with
inhaled nitric oxide but no clinical outcome benefits as measured by mortality
or the duration of mechanical ventilation.21 A
meta-analysis26 also concluded that a composite
of clinical trials showed no benefit of inhaled nitric oxide in ARDS. Inhaled
nitric oxide as a potential adjunct therapy following lung transplantation
has produced mixed results.27,28
There are many potential reasons for this lack of long-term benefit
despite initial improvements in oxygenation. In the trials described herein
and in smaller studies,17,18 inhaled
nitric oxide induced improvements in oxygenation that were maintained only
during the first 24 to 48 hours. The reason for this is unclear, because withdrawal
or reinstitution of inhaled nitric oxide in earlier studies of individual
patients suggested a continued effect for more than 7 days.16 These
initial observations may have reflected rebound deterioration with the withdrawal
of inhaled nitric oxide rather than continued effectiveness. It is also possible
that the 5-ppm inhaled nitric oxide dose will, over time, diffuse into poorly
ventilated areas and negate the selective pulmonary vasodilation achieved
with initial inhaled nitric oxide administration. Other explanations include
the possibility that the trial was not optimally designed to allow the acute
physiologic effects of inhaled nitric oxide to translate into long-term benefit.
Different trial conditions that may have yielded better results could include
combination with other interventions, such as lung recruitment, a different
selection of patients, or a different dosing regimen. Recently it has been
demonstrated that when the initial dose of inhaled nitric oxide in ARDS is
chosen by dose response curve, oxygenation benefit is subsequently lost in
many patients, but maintained at a lower dose.29 Daily
dose response curves may optimize inhaled nitric oxide effect. Finally, it
is possible that the oxygenation benefit of inhaled nitric oxide was offset
by toxicity. Considerable controversy exists about the cytotoxic vs cytoprotective
effects of nitric oxide.30 The higher incidence
of infection in the inhaled nitric oxide group is interesting in light of
a recent report31 of antibacterial properties
of inhaled nitric oxide in an animal model of pneumonia.
It is clear that using the current dosing regimen, inhaled nitric oxide
does not improve clinical outcomes in patients with moderately severe acute
lung injury. There were no subgroups that benefited from inhaled nitric oxide,
either primary vs secondary, degree of hypoxemia, or degree of pulmonary pressures.
The results of this study do not support the routine use of inhaled nitric
oxide in hospitalized patients with acute lung injury.
In patients with documented moderately severe acute lung injury but
without sepsis or other organ system failure, inhaled nitric oxide at 5 ppm
did not improve any of the measured patient benefit outcomes. This lack of
effect on patient benefit outcomes was seen despite a statistically significant
improvement in the acute physiology that resolved between 24 and 48 hours.
These data do not support the routine use of inhaled nitric oxide in the treatment
of acute lung injury or ARDS. Inhaled nitric oxide may be considered as a
salvage therapy in acute lung injury or ARDS patients who continue to have
life-threatening hypoxemia despite optimization of conventional mechanical
Create a personal account or sign in to: