From the Adult Intensive Care Unit and
`Intensive Care Services, Royal Brompton
`Hospital, and Imperial College London —
`both in London. Address reprint requests
`to Dr. Evans at the Unit of Critical Care,
`Imperial College London, Royal Bromp-
`ton Hospital, Sydney St., London SW3
`6NP, United Kingdom, or at t.evans@
`rbh.nthames.nhs.uk.
`
`N Engl J Med 2005;353:2683-95.
`Copyright © 2005 Massachusetts Medical Society.
`
`T h e ne w e ngl a nd jou r na l o f m e dicine
`
`review article
`
`drug therapy
`
`Inhaled Nitric Oxide Therapy in Adults
`
`Mark J.D. Griffiths, M.R.C.P., Ph.D., and Timothy W. Evans, M.D., Ph.D.
`
`background and historical perspective
`
`Nitric oxide was largely regarded as a toxic pollutant until
`
`1987, when its biologic similarities to endothelium-derived relaxing factor
`were demonstrated.1 Subsequently, nitric oxide and endothelium-derived
`relaxing factor were considered a single entity, modulating vascular tone through
`the stimulated formation of cyclic guanosine 3',5'-monophosphate (Fig. 1).2 Endog-
`enous nitric oxide is formed from the semiessential amino acid L-arginine by one
`of three (neural, inducible, and endothelial) isoforms of nitric oxide synthase. The
`physiologic role of endogenous nitric oxide was first shown when an infusion of an
`inhibitor of all forms of nitric oxide synthase in healthy volunteers led to systemic
`and pulmonary pressor responses.3 However, the role of nitric oxide in maintaining
`low pulmonary vascular resistance in healthy persons has since been challenged.4
`Inhaled nitric oxide had a negligible effect on pulmonary blood flow in healthy
`humans,5 but when healthy persons were breathing 12 percent oxygen, it reversed
`the pulmonary hypertension that was induced without affecting systemic hemody-
`namics.6 In 1991, inhaled nitric oxide was shown to be a selective pulmonary vaso-
`dilator in patients with pulmonary hypertension,7 as well as in animals with pul-
`monary hypertension induced by drugs or hypoxia.8 Two years later, inhaled nitric
`oxide emerged as a potential therapy for the acute respiratory distress syndrome
`(ARDS), because it decreased pulmonary vascular resistance without affecting sys-
`temic blood pressure and improved oxygenation by redistributing pulmonary blood
`flow toward ventilated lung units in patients with this condition.9
`Despite such promise, the potential therapeutic role of inhaled nitric oxide in
`adults remains uncertain; licensed indications are restricted to pediatric practice.
`Furthermore, recent changes in the marketing of inhaled nitric oxide have dra-
`matically increased its cost, which has inevitably led to a need to justify continuing
`its administration to adults. This review will consider the biologic actions of in-
`haled nitric oxide, discuss clinical indications for its administration in adults, and
`assess possible future developments.
`
`chemical reactions of inhaled nitric oxide
`Nitric oxide is a gas that is colorless and odorless at room temperature and is rela-
`tively insoluble in water. It is poorly reactive with most biologic molecules, but be-
`cause it has an unpaired electron, it can react very rapidly with other free radicals,
`certain amino acids, and transition metal ions.10 In biologic solutions, nitric oxide
`is stabilized by forming complexes with — for example — thiols, nitrite, and pro-
`teins that contain transition metals.11
`Atmospheric concentrations of nitric oxide typically range between 10 and 500
`parts per billion but may reach 1.5 parts per million (ppm) in heavy traffic12 and
`1000 ppm in tobacco smoke.13 When inhaled with high concentrations of oxygen,
`
`n engl j med 353;25 www.nejm.org december 22, 2005
`
`2683
`
`Downloaded from www.nejm.org at HELLENIC SOCIETY INTENSIVE CARE on January 24, 2006 .
`Copyright © 2005 Massachusetts Medical Society. All rights reserved.
`
`2683
`
`

`

`T h e ne w e ngl a nd jou r na l o f m e dicine
`
`Blood vessel
`
`Endothelial cells
`
`Vascular smooth-muscle cells
`
`Nitric oxide
`
`Soluble guanylyl
`cyclase
`
`Phosphodiesterase
`type 5
`
`Inhibition by sildenafil
`and zaprinast
`
`Inositol 1,4,5,
`-triphosphate
`
`Ca2+
`
`Vascular
`smooth-muscle
`cell
`
`Activation
`
`GTP
`
`cGMP
`
`Activation
`
`cGMP-dependent
`protein kinase
`
`Inhibition
`of calcium release
`
`Decreased
`sensitivity of myosin
`
`Inhibition
`
`Phosphory-
`lated myosin
`(contraction)
`
`Myosin light-chain
`phosphatases
`
`Myosin
`(relaxation)
`
`Activation
`
`Sarcoplasmic
`reticulum
`
`L-type calcium
`channel
`
`Ca2+
`
`K+
`
`Calcium-sensitive
`potassium channel
`
`Figure 1. Regulation of the Relaxation of Vascular Smooth Muscle by Nitric Oxide.
`Nitric oxide activates soluble guanylyl cyclase, leading to the activation of cyclic guanosine 3´, 5´-monophosphate
`(cGMP)–dependent protein kinase (cGKI). In turn, cGKI decreases the sensitivity of myosin to calcium-induced con-
`traction and lowers the intracellular calcium concentration by activating calcium-sensitive potassium channels and
`inhibiting the release of calcium from the sarcoplasmic reticulum. cGMP is degraded by phosphodiesterase type 5,
`which is inhibited by sildenafil and zaprinast. GTP denotes guanosine triphosphate.
`
`gaseous nitric oxide slowly forms nitrogen diox-
`ide.14 Once dissolved in airway-lining fluid, nitric
`oxide may react with reactive oxygen species such
`as superoxide to form reactive nitrogen species
`such as peroxynitrite, a powerful oxidant that can
`decompose further to yield nitrogen dioxide and
`hydroxyl radicals (Fig. 2).15 Therefore, nitric ox-
`ide is potentially cytotoxic, and covalent nitra-
`tion of tyrosine in proteins by reactive nitrogen
`species has been used as a marker of oxidative
`stress.16
`Nitric oxide is rapidly inactivated by hemoglo-
`bin in blood, by haptoglobin–hemoglobin com-
`plexes in plasma, and by a reaction with heme
`
`ferrous iron and ferric iron that forms nitrosyl-
`hemoglobin.17 Nitric oxide forms methemoglo-
`bin and nitrate on reaction with oxyhemoglobin,
`which predominates in the pulmonary circulation.
`Most of the methemoglobin is reduced to ferrous
`hemoglobin by NADH–cytochrome b5
` reductase
`in erythrocytes. In healthy subjects who have in-
`haled nitric oxide (80 ppm) for one hour, plasma
`nitrate concentrations may be four times as high
`as baseline levels.18 Almost 70 percent of inhaled
`nitric oxide is excreted as nitrate in the urine
`within 48 hours.19
`More than 100 proteins, including hemoglo-
`bin20 and albumin,21 contain reduced sulfur (thiol)
`
`2684
`
`n engl j med 353;25 www.nejm.org december 22, 2005
`
`Downloaded from www.nejm.org at HELLENIC SOCIETY INTENSIVE CARE on January 24, 2006 .
`Copyright © 2005 Massachusetts Medical Society. All rights reserved.
`
`2684
`
`

`

`drug therapy
`
`groups that react reversibly with nitric oxide to
`form S-nitrosothiols; these compounds are vaso-
`dilators that inhibit platelet aggregation.22 S-nitro-
`sothiols may also “store” nitric oxide within the
`circulation. For example, S-nitrosohemoglobin in
`red cells has been postulated to regulate micro-
`vascular flow and oxygen delivery.23
`
`physiologic effects of inhaled nitric oxide
`on the cardiovascular system
`
`Inhaled nitric oxide relaxes pulmonary vessels,
`thereby decreasing pulmonary vascular resistance,
`pulmonary arterial pressure, and right ventricu-
`lar afterload (Table 1).6-8 The selectivity of nitric
`oxide for the pulmonary circulation is the result
`of rapid hemoglobin-mediated inactivation of ni-
`tric oxide.29 In the presence of biventricular car-
`diac failure, inhaled nitric oxide may sufficiently
`increase pulmonary blood flow and, hence, left
`atrial end-diastolic pressure to precipitate pulmo-
`nary edema.30
`Early studies in patients with ARDS com-
`pared the effect of inhaled nitric oxide with an-
`other vasodilator (epoprostenol, or prostacyclin
`or prostaglandin I2 ) administered intravenously.9
`The intravenously administered vasodilator wors-
`ened oxygenation owing to antagonism of hypoxic
`pulmonary vasoconstriction. In contrast, the ad-
`vantage of inhaled nitric oxide was that only the
`vasculature associated with ventilated lung units
`was within reach of an inhaled gas diffusing
`across the alveolar-capillary membrane. Selective
`dilatation of these vessels would improve venti-
`lation–perfusion matching (Fig. 3).
`Circulating modulators of vascular tone, such
`as the potent vasoconstrictor endothelin-1 and
`endogenous nitric oxide, influence the effect of
`inhaled nitric oxide. Decreased responsiveness is
`associated with the induction of nitric oxide syn-
`thase by endotoxin both in patients with ARDS
`associated with septic shock31 and in animal mod-
`els (Fig. 3E).32 Conversely, the positive effect of
`inhaled nitric oxide on gas exchange depends on
`the extent to which pulmonary vasoconstriction
`and ventilation–perfusion mismatching are con-
`tributing to impaired oxygenation. For example,
`in a study of mountaineers who were either sus-
`ceptible or not susceptible to high-altitude pul-
`monary edema, inhaled nitric oxide decreased the
`pulmonary arterial pressure of susceptible sub-
`jects, but improved oxygenation only in the sub-
`jects with the greatest degree of hypoxemia (those
`
`Air space
`
`Type I
`alveolar cell
`
`Type II
`alveolar cell
`
`O2
`
`NO2
`
`Nitric oxide
`
`Release of
`reactive
`oxygen species
`
`Formation of
`reactive nitrogen
`species
`
`Inactivation by
`hemoglobin
`
`Red cell
`
`Formation of
`S-nitrosothiols
`
`Plasma proteins
`
`Leukocyte
`
`Vascular space
`
`Endothelial cell
`
`Figure 2. Biochemical Fates of Inhaled Nitric Oxide at the Alveolar-Capillary
`Membrane.
`Small amounts of nitrogen dioxide (NO2) may be formed if inhaled nitric
`oxide mixes with high concentrations of oxygen (O2) in the air space.
`Depending on the milieu of the lung parenchyma, nitric oxide may react
`with reactive oxygen species (derived from activated leukocytes or is che-
`mia–reperfusion injury) to form reactive nitrogen species such as peroxyni-
`trite. In the vascular space, dissolved nitric oxide is scavenged by oxyhemo-
`globin (forming methemoglobin and nitrate) and to a lesser extent, plasma
`proteins (e.g., forming nitrosothiols, which are stable intravascular sources
`of nitric oxide activity).
`
`who had pulmonary edema) by increasing the
`blood flow to the areas of lung that were rela-
`tively unaffected.33
`The effects of inhaled nitric oxide also depend
`on vascular selectivity. For example, dispropor-
`tionate arterial, as opposed to venous, dilatation
`would increase the pulmonary-capillary pressure
`and exacerbate pulmonary edema. Although many
`studies have not shown evidence of selectivity,
`others have demonstrated that 40 ppm of nitric
`oxide induced venodilatation with decreased pul-
`monary-capillary pressure34 and reduced the risk
`of pulmonary edema in patients with acute lung
`injury.35 Apart from changing the pulmonary-
`capillary pressure, nitric oxide may influence the
`development of edema through pulmonary vas-
`cular recruitment or by decreasing inflammation
`and helping maintain the integrity of the alveo-
`lar-capillary membrane. Such specific effects are
`difficult to identify with certainty in vivo. Be-
`cause the effects of nitric oxide probably vary in
`different settings, apparently contradictory clin-
`ical and experimental observations have been
`produced.
`
`n engl j med 353;25 www.nejm.org december 22, 2005
`
`2685
`
`Downloaded from www.nejm.org at HELLENIC SOCIETY INTENSIVE CARE on January 24, 2006 .
`Copyright © 2005 Massachusetts Medical Society. All rights reserved.
`
`2685
`
`

`

`T h e ne w e ngl a nd jou r na l o f m e dicine
`
`Table 1. Comparison of Ideal Treatment Goals with Those Achieved by Inhaled Nitric Oxide in Adults with the Acute
`Respiratory Distress Syndrome (ARDS).
`
`Ideal Treatment Goals
`
`Improved oxygenation
`
`Decreased pulmonary vascular resistance
`
`Decreased pulmonary edema
`
`Physiological Effects of Inhaled Nitric Oxide
`
`20% Improvement in approximately 60% of patients for only 1 to 2 days in
`clinical trials, with no associated survival benefit24,25; may significantly
`improve oxygenation in very severe cases and buy time for the institu-
`tion of other means of support
`
`Selective pulmonary vasodilator of uncertain benefit in acute lung injury or
`ARDS characterized by mild pulmonary hypertension26; may have a
`supportive role in patients with acute right-sided heart failure, particu-
`larly in association with increased pulmonary vascular resistance and
`hypoxemia
`
`May be influenced by effects on hemodynamics, inflammation, infection,
`and the alveolar-capillary membrane
`
`Reduction or prevention of inflammation
`
`Conflicting evidence of its antiinflammatory efficacy at multiple molecular
`and clinical levels
`
`Cytoprotection
`
`Protection against infection
`
`
`
`May contribute to the formation of cytotoxic reactive nitrogen species and
`reactive oxygen species, especially when administered with high con-
`centrations of oxygen; conversely, may prevent the generation of reac-
`tive oxygen species by free iron and scavenge hydroxyl radicals27
`Direct antimicrobial effects,28 but associated with an increased incidence of
`ventilator-associated pneumonia in one study25
`
`Most clinical studies have provided support for
`the view that inhaled nitric oxide has no effect
`on the systemic circulation. In contrast, experi-
`mental studies have demonstrated a reduction in
`systemic vascular resistance36 and restoration of
`mesenteric perfusion after the inhibition of ni-
`tric oxide synthase.37 Similarly, the inhalation of
`nitric oxide (80 ppm) by healthy volunteers abol-
`ished the vasopressor effect of the inhibition of
`nitric oxide synthase in the circulation of the
`forearm, an effect associated with increased ar-
`terial concentrations of nitrite and S-nitrosylhe-
`moglobin, but not of S-nitrosothiols or S-nitroso-
`hemoglobin.18 The concept of a plasma-based
`repository for nitric oxide activity that may be
`supplemented by inhaled nitric oxide has become
`widely accepted; probable contributors include ni-
`trites,38 iron nitrosyl and N-nitrosamine complex-
`es,39 and nitrated lipids.40
`When inhaled nitric oxide is used therapeuti-
`cally, its rapid withdrawal may induce rebound
`pulmonary hypertension and hypoxemia.9,41 The
`inhalation of nitric oxide by healthy animals de-
`creases endothelial nitric oxide synthase activity
`and increases plasma concentrations of endothe-
`lin-1,42 which inactivates endothelial nitric oxide
`synthase by nitration.43 In practice, rebound phe-
`nomena may be avoided by withdrawing inhaled
`nitric oxide gradually. Despite these concerns, in
`
`large clinical studies of patients with ARDS, the
`abrupt discontinuation of inhaled nitric oxide has
`not caused a deterioration in oxygenation.24,25
`
`direct cytotoxicity and effects
`on inflammation
`
`Inhaled nitric oxide may modulate the acute neu-
`trophilic inflammation of the lung parenchyma
`and dysfunction of the alveolar-capillary mem-
`brane that characterizes ARDS at several levels.
`The protective effects of nitric oxide may derive
`from specific effects on neutrophil function —
`for example, by attenuation of the respiratory
`burst and neutrophil-derived oxidative stress.44
`Inhaled nitric oxide has decreased the accumula-
`tion of neutrophils in the pulmonary vasculature
`and air space in animal models of acute lung in-
`jury,45 consistent with its known effects on the
`adhesion and deformability of neutrophils in vi-
`tro.46 Furthermore, similar effects of inhaled nitric
`oxide outside the lung have been observed in ro-
`dent models of severe sepsis.47 In a model in which
`cecal ligation and puncture were used to induce
`sepsis, mice lacking inducible nitric oxide syn-
`thase had fewer neutrophils sequestered in the
`pulmonary vasculature than normal mice, but
`they had greater neutrophil migration into the air
`spaces.48 Subsequent experiments have confirmed
`that nitric oxide derived from neutrophils acts as
`
`2686
`
`n engl j med 353;25 www.nejm.org december 22, 2005
`
`Downloaded from www.nejm.org at HELLENIC SOCIETY INTENSIVE CARE on January 24, 2006 .
`Copyright © 2005 Massachusetts Medical Society. All rights reserved.
`
`2686
`
`

`

`drug therapy
`
`A
`
`Ventilation
`
`Pulmonary
`arterial
`blood
`flow
`
`Maintenance of Oxygenation
`
`B
`
`C
`
`Improved oxygenation
`
`Normal
`ventilation–perfusion
`
`Minimization of
`ventilation–perfusion
`mismatching owing to
`hypoxic pulmonary
`vasoconstriction
`
`Pulmonary blood
`flow increased by
`inhaled short-acting
`vasodilator
`
`Nitric
`oxide
`
`Pulmonary
`venous blood
`flow
`
`D
`
`E
`
`F
`
`Decreased Oxygenation
`
`Hypoxic pulmonary
`vasoconstriction
`counteracted by
`intravenous vasodilator
`
`Dysregulation of
`pulmonary vascular
`tone by disease
`
`Nitric
`oxide
`
`Nitric
`oxide
`
`Accumulation or
`leakage of nitric oxide
`owing to long-term
`administration of
` inhaled
` nitric
` oxide
`
`Figure 3. Mechanism of Action and Inaction of Inhaled Nitric Oxide.
`Panel A shows normal ventilation–perfusion. Hypoxic pulmonary vasoconstriction (Panel B) minimizes ventilation–
`perfusion mismatching in the presence of abnormal ventilation. Inhaled vasodilators with a short half-life improve
`oxygenation by increasing blood flow to ventilated lung units (Panel C). If a vasodilator is administered intravenous-
`ly (Panel D) or if diseases are associated with dysregulated pulmonary vascular tone, such as sepsis and acute lung
`injury (Panel E), hypoxic pulmonary vasoconstriction is counteracted, leading to worsening oxygenation. Long-term
`administration of inhaled nitric oxide, with the accumulation of nitric oxide or leakage between lung units associat-
`ed with collateral ventilation, as may occur in chronic obstructive pulmonary disease (Panel F), may negate the ben-
`eficial effects of inhaled nitric oxide on oxygenation.
`
`an autocrine modulating factor in infiltration of
`neutrophils into the lungs during sepsis.
`The toxic potential of nitric oxide is well
`known; endogenously produced nitric oxide con-
`tributes to the control and killing of multiple
`pathogens28 and malignant cells.49 Studies in-
`volving inhibitors of nitric oxide synthase50 and
`mice lacking inducible nitric oxide synthase51
`have suggested that nitric oxide–derived reactive
`nitrogen species contribute to epithelial damage
`after a variety of insults. The results of interac-
`tions between nitric oxide and reactive oxygen
`species are unpredictable and probably depend
`on the relative local concentrations of the par-
`ticipants in these reactions.52 Increased concen-
`trations of oxidative products of nitric oxide were
`found in the airway-lining fluid of patients with
`ARDS,53 and these may be further increased by
`
`inhalation of nitric oxide.54 In rodents, inhalation
`of nitric oxide (20 ppm) did not increase protein
`nitration unless hyperoxia was superimposed.55
`Taken together, these observations suggest an
`important role for oxidative damage and reactive
`nitrogen species in these pulmonary diseases, but
`the role of exogenous nitric oxide in modulating
`these processes is uncertain.
`
`other effects
`Endogenous nitric oxide inhibits the adhesion of
`platelets to endothelial cells and subsequent ag-
`gregation.2 In experimental microsphere-induced
`pulmonary embolism, inhaled nitric oxide attenu-
`ated increases in pulmonary arterial pressure and
`platelet aggregation.56 However, in animals, healthy
`volunteers, and patients with pulmonary diseas-
`es, the effects of inhalation of nitric oxide on the
`
`n engl j med 353;25 www.nejm.org december 22, 2005
`
`2687
`
`Downloaded from www.nejm.org at HELLENIC SOCIETY INTENSIVE CARE on January 24, 2006 .
`Copyright © 2005 Massachusetts Medical Society. All rights reserved.
`
`2687
`
`

`

`T h e ne w e ngl a nd jou r na l o f m e dicine
`
`duration of bleeding and other indexes of platelet
`function are variable.52
`Surfactant dysfunction contributes substan-
`tially to the pathophysiological characteristics of
`lung injury. Reactive nitrogen species react with
`and impair the functions of the surfactant pro-
`teins; it has been shown that the surfactant from
`animals receiving inhaled high-dose nitric oxide
`(80 to 100 ppm) had a reduced capacity to lower
`surface tension.57 Conversely, inhaled nitric ox-
`ide increased the production of surfactant pro-
`teins in four-week-old lambs.58 The relevance of
`these observations to adult humans treated with
`inhaled nitric oxide is uncertain.
`Inhaled nitric oxide has a dose-dependent bron-
`chodilator effect on drug-induced bronchocon-
`striction in animal models59 and causes mild
`bronchodilation in patients with asthma.60 An in-
`teresting finding is that the nitric oxide–derived
`S-nitrosothiols, which act as bronchodilators, were
`present at lower concentrations in the fluid lining
`the airways of patients with severe asthma than
`of healthy subjects, suggesting that this mecha-
`nism may contribute to bronchospasm.61
`
`administration of inhaled nitric oxide
`to adults
`Route, Monitoring, and Safety
`Nitric oxide is most commonly administered to
`patients receiving mechanical ventilation, although
`it may also be given through a face mask or nasal
`cannulae. Limiting the mixing of nitric oxide and
`high concentrations of inspired oxygen reduces
`the risk of adverse effects resulting from the for-
`mation of nitrogen dioxide (Fig. 2). This is mini-
`mized further by introducing the mixture of ni-
`tric oxide and nitrogen into the inspiratory limb
`of the ventilator tubing as near to the patient as
`possible62 and synchronizing injection of the mix-
`ture with inspiration.63
`Although a massive overdose of inhaled nitric
`oxide (500 to 1000 ppm) is rapidly fatal,64 stud-
`ies in animals have provided reassuring data in-
`dicating that nitric oxide has minimal pulmonary
`toxicity when it is inhaled at a concentration of
`less than 40 ppm for up to six months.65 Electro-
`chemical analyzers can be used to monitor the
`concentrations of nitric oxide and nitrogen diox-
`ide in the inspired gas mixture to an accuracy of
`1 ppm. More sensitive (chemiluminescence) mon-
`itors can detect nitric oxide and its oxidative de-
`rivatives in parts per billion.
`
`Up to 40 ppm of inhaled nitric oxide admin-
`istered clinically should not cause methemoglo-
`binemia in adults in the absence of methemo-
`globin reductase deficiency.66 However, guidelines
`in the United Kingdom recommend measurement
`of methemoglobin concentrations within six hours
`after the initiation of nitric oxide therapy and
`after each increase in the dose.62 The Control of
`Substances Hazardous to Health Regulations sug-
`gest that environmental concentrations of nitric
`oxide and nitrogen dioxide should not exceed a
`time-weighted average of 25 ppm and 2 ppm,
`respectively, over an eight-hour period.67 Clearly,
`it is unlikely that such levels would accumulate
`from therapeutic administration of nitric oxide in
`a well-ventilated room (10 to 12 air changes per
`hour). Consequently, the use of environmental
`monitoring and equipment to adsorb nitric oxide
`(nitric oxide scavenging) in the clinical setting is
`rarely necessary.62
`
`Dose–Response Relationship
`Early clinical experience with the use of inhaled
`nitric oxide to treat patients with respiratory fail-
`ure indicated that higher doses were required to
`treat pulmonary hypertension than to improve
`oxygenation. When nitric oxide is administered,
`only a minority of patients have no response when
`a response is defined as a 20 percent increase in
`oxygenation.68 Although this threshold is widely
`accepted, its biologic relevance has not been vali-
`dated across a range of respiratory failure; for ex-
`ample, a 10 percentage point improvement in he-
`moglobin saturation in a patient with hypoxemia
`who is breathing 100 percent oxygen may be clini-
`cally very important. No radiologic or physiological
`variables predict a response to inhaled nitric oxide
`in patients with acute lung injury or ARDS, and the
`response varies over the clinical course.69,70
`In the treatment of pulmonary hypertension,
`a 30 percent decrease in pulmonary vascular re-
`sistance during the inhalation of nitric oxide (10
`ppm for 10 minutes) has been used to identify an
`association with vascular responsiveness to agents
`that can be helpful in the long term71; indeed, a
`positive response to nitric oxide was associated
`with a favorable response to calcium-channel
`blockers in a small cohort of patients with pri-
`mary pulmonary hypertension.72
`Numerous small studies involving patients
`with acute respiratory failure have examined the
`dose–response relationship of inhaled nitric oxide
`
`2688
`
`n engl j med 353;25 www.nejm.org december 22, 2005
`
`Downloaded from www.nejm.org at HELLENIC SOCIETY INTENSIVE CARE on January 24, 2006 .
`Copyright © 2005 Massachusetts Medical Society. All rights reserved.
`
`2688
`
`

`

`drug therapy
`
`and oxygenation, demonstrating marked variation
`in any one patient and between patients, as well
`as some evidence of a plateau in effect when the
`dose was between 1 and 10 ppm. The time-depen-
`dent variation in the dose–response relationship
`of inhaled nitric oxide in patients with severe
`ARDS has been explored with the use of a pro-
`spective, randomized protocol in which patients
`received either inhaled nitric oxide (10 ppm) or
`a placebo.73 Dose–response relationships (nitric
`oxide, 0 to 100 ppm) were constructed in the two
`groups on days 0, 2, and 4 of the study. Two im-
`portant observations emerged: first, the dose–
`response curves for changes in oxygenation and
`mean pulmonary pressure were shifted to the left
`only in patients who inhaled nitric oxide (10 ppm)
`continuously. Second, “supramaximal” doses of
`nitric oxide were associated with worsening oxy-
`genation. These observations imply that the op-
`timal dose of inhaled nitric oxide must be deter-
`mined by titration against the therapeutic target
`in each patient at least every two days, and prob-
`ably more frequently.
`
`clinical indications for administering
`inhaled nitric oxide to adults
`
`Acute Lung Injury and the Acute Respiratory Distress
`Syndrome
`In adults with acute lung injury, inhaled nitric
`oxide is used more often to improve oxygenation
`than to decrease pulmonary vascular resistance.
`Two small (a total of 70 patients), single-center
`studies74,75 and four multicenter, randomized, pla-
`cebo-controlled trials24,25,76,77 have failed to deter-
`mine the therapeutic role of inhaled nitric oxide in
`patients with acute respiratory failure.78 A French
`multicenter study that recruited 203 patients re-
`ported no decrease in the duration of mechanical
`ventilation or the mortality rate among patients
`treated with inhaled nitric oxide as compared with
`those taking a nitrogen placebo, but that study has
`been published only in abstract form.76 A phase
`2 American study that was not statistically pow-
`ered to demonstrate a benefit in mortality rate re-
`ported that doses of 1.25 to 40 ppm of inhaled
`nitric oxide were well tolerated (Table 2).24 The
`percentage of patients having a response (defined
`by a 20 percent increase in the arterial partial
`pressure of oxygen) to the various doses was sim-
`ilar: approximately 60 percent of patients in both
`studies.
`
`A European multicenter study that planned to
`enroll 600 subjects enrolled 268 patients with
`early acute lung injury and then changed the pro-
`tocol after 140 patients had been recruited.77 Ul-
`timately, three groups of patients were analyzed:
`those who had less than a 20 percent increase in
`arterial partial pressure of oxygen in response to
`inhaled nitric oxide, patients with a response who
`were treated conventionally, and patients with a
`response who were treated with the lowest effec-
`tive dose of inhaled nitric oxide. The mortality
`rates in the three groups were similar at 30 days.
`Another American multicenter study performed
`between 1996 and 1999 compared the effects of
`continuously inhaled nitric oxide (5 ppm) with those
`of a placebo in patients with ARDS that was not
`associated with severe sepsis or multiorgan fail-
`ure.25 Despite the lower dose, the increase in oxy-
`genation (specifically in the partial pressure of
`arterial oxygen) lasted only for the first day of
`therapy, a finding similar to that in the first
`American study. Nitric oxide had no significant
`effect on any outcome measure (Table 2).
`Two important questions are raised by these
`studies. First, why are the effects of inhaled nitric
`oxide so short-lived? Increasing sensitivity to nitric
`oxide during its inhalation may diminish its ben-
`eficial effects and increase toxicity.73 Alternatively,
`constant inhalation may lead to equilibration of
`the vasodilator effect between ventilated and non-
`ventilated areas (Fig. 3E). Such effects might be
`mitigated by performing daily dose–response as-
`sessments or by including regular nitric oxide–free
`periods in the regimen, depending on whether re-
`bound phenomena occur. Clearly, any continued
`benefit may depend on the use of other therapeutic
`approaches such as maintaining alveolar recruit-
`ment. Second, if the clinical benefits are real, why
`do they not translate into improved outcome?
`Because ARDS is a heterogeneous condition with
`multiple causes requiring different interventions
`that independently affect the outcome, very large
`numbers of patients would be required for a study
`to demonstrate benefit. Furthermore, many large
`studies evaluating modes of ventilation80,81 and
`prone positioning82 in patients with ARDS have
`shown no correlation between improved oxygen-
`ation and the outcome. This result is partly ex-
`plained by the observation that only a minority of
`patients with ARDS die from respiratory failure;
`the majority die from multiorgan failure.83
`
`n engl j med 353;25 www.nejm.org december 22, 2005
`
`2689
`
`Downloaded from www.nejm.org at HELLENIC SOCIETY INTENSIVE CARE on January 24, 2006 .
`Copyright © 2005 Massachusetts Medical Society. All rights reserved.
`
`2689
`
`

`

`T h e ne w e ngl a nd jou r na l o f m e dicine
`
`∥ The group receiving inhaled nitric oxide had an increased incidence of acute renal failure (as defined by a serum creatinine concentration of more than 3.5 mg per deciliter or the need
`¶ Of 268 patients with a response to nitric oxide, 180 underwent randomization.
`§ There were significant differences in this outcome between the control group and the group receiving inhaled nitric oxide.
`‡ The 80-ppm dose was stopped, owing to consensus that the dose was likely to be higher than the peak of the dose–response curve.
`† The definition of the American–European Consensus Conference on the Acute Respiratory Distress Syndrome (ARDS) was used.79
`* PaO2:FiO2 denotes the ratio of the partial pressure of arterial oxygen to the fraction of inspired oxygen.
`
`for renal replacement therapy) (P<0.03).
`
`met for extubation
`urvival after oxygenation criteria
`trial of unassisted ventilation;
`survival after successful two-hour
`tory pressure§; 28-day survival;
`
` s
`
`Oxygenation and positive end-expira-
`
`Survival without need
`
`the first 28 days
`ventilation during
`for mechanical
`
`5 ppm in 192 patients
`
`Nitrogen in 193
`
`patients
`
`injury; organ failure∥
`of hospitalization and acute lung
`dency on intensive care; duration
`
`30-day and 90-day survival; depen-
`
`Reversal of acute lung
`
`injury
`
`days) in 87 patients
`±SD, 9±8 ppm for 9±6
`effective dose; mean
`2, 10, or 40 ppm (lowest
`
`93 patients
`no placebo in
`therapy with
`
`Conventional
`
`Oxygenation§; pulmonary arterial
`
`urvival
`pressure§; response; 28-day
`
` s
`
`Duration of mechani-
`
`cal ventilation
`
`80 ppm in 8 patients‡
`40 ppm in 27 patients
`20 ppm in 29 patients
`5 ppm in 34 patients
`1.25 ppm in 22 patients
`
`Nitrogen in 57
`
`patients
`
`Inhaled Nitric Oxide
`
`Control
`
`cluded
`organ failure, or both ex-
`vere sepsis, non-pulmonary
`mm Hg†; patients with se-
`ARDS and a PaO2:FiO2 <250
`
`Patients with acute lung injury
`
`tion 18–96 hr¶
`ceiving mechanical ventila-
`mm Hg who had been re-
`with a PaO2:FiO2 <165
`
`or both, excluded
`nonpulmonary organ failure,
`patients with severe sepsis,
`within 72 hr after diagnosis†;
`
`Patients with ARDS, enrolled
`
`28
`
`2004
`
`Taylor et al.25
`
`30
`
`1999
`
`Lundin et al.77
`
`28
`
`days
`
`1998
`
`et al.24
`Dellinger
`
`Secondary Outcomes
`
`Primary Outcome
`
`Intervention
`
`Patients*
`
`Intervention
`Duration of
`
`Year
`
`Study
`
`Table 2. Results of Multicenter Clinical Studies of the Use of Inhaled Nitric Oxide in Patients with Acute Respiratory Failure.
`
`2690
`
`n engl j med 353;25 www.nejm.org december 22, 2005
`
`Downloaded from www.nejm.org at HELLENIC SOCIETY INTENSIVE CARE on January 24, 2006 .
`Copyright © 2005 Massachusetts Medical Society. All rights reserved.
`
`2690
`
`

`

`drug therapy
`
`Targeting Pulmonary Vascular Resistance
`The inhalation of nitric oxide by patients with
`acute lung injury, which is characterized by mild
`pulmo