Inhaled Nitric Oxide for the Early Treatment of Persistent Pulmonary Hypertension of the Term Newborn: A Randomized, Double-Masked, Placebo-Controlled, Dose-Response, Multicenter Study
Objectives. To assess the dose-related effects of inhaled nitric oxide (I-NO) as a specific adjunct to early conventional therapy for term infants with persistent pulmonary hypertension (PPHN), with regard to neonatal outcome, oxygenation, and safety.
Methods. Randomized, placebo-controlled, double-masked, dose-response, clinical trial at 25 tertiary centers from April 1994 to June 1996. The primary endpoint was the PPHN Major Sequelae Index ([MSI], including the incidence of death, extracorporeal membrane oxygenation (ECMO), neurologic injury, or bronchopulmonary dysplasia [BPD]). Patients required a fraction of inspired oxygen [Fio 2] of 1.0, a mean airway pressure ≥10 cm H2O on a conventional ventilator, and echocardiographic evidence of PPHN. Exogenous surfactant, concomitant high-frequency ventilation, and lung hypoplasia were exclusion factors. Control (0 ppm) or nitric oxide (NO) (5, 20, or 80 ppm) treatments were administered until success or failure criteria were met. Due to slowing recruitment, the trial was stopped at N = 155 (320 planned).
Results. The baseline oxygenation index (OI) was 24 ± 9 at 25 ± 17 hours old (mean ± SD). Efficacy results were similar among NO doses. By 30 minutes (no ventilator changes) the Pao 2 for only the NO groups increased significantly from 64 ± 39 to 109 ± 78 Torr (pooled) and systemic arterial pressure remained unchanged. The baseline adjusted time-weighted OI was also significantly reduced in the NO groups (-5 ± 8) for the first 24 hours of treatment. The MSI rate was 59% for the control and 50% for the NO doses (P = .36). The ECMO rate was 34% for control and 22% for the NO doses (P = .12). Elevated methemoglobin (>7%) and nitrogen dioxide (NO2) (>3 ppm) were observed only in the 80 ppm NO group, otherwise no adverse events could be attributed to I-NO, including BPD.
Conclusion. For term infants with PPHN, early I-NO as the sole adjunct to conventional management produced an acute and sustained improvement in oxygenation for 24 hours without short-term side effects (5 and 20 ppm doses), and the suggestion that ECMO use may be reduced.
- extracorporeal membrane oxygenation
- bronchopulmonary dysplasia
- neonatal outcome
- nitrogen dioxide
- PPHN =
- persistent pulmonary hypertension of the newborn •
- ECMO =
- extracorporeal membrane oxygenation •
- I-NO =
- inhaled nitric oxide •
- BPD =
- bronchopulmonary dysplasia •
- NO2 =
- nitrogen dioxide •
- ppm =
- parts per million •
- MSI =
- Major Sequelae Index •
- RAD =
- reactive airway disease •
- Fio2 =
- fraction of inspired oxygen •
- NO =
- nitric oxide •
- TWOI =
- time-weighted oxygenation index •
- OI =
- oxygenation index
Persistent pulmonary hypertension of the newborn (PPHN) is a syndrome of acute respiratory failure, characterized by systemic hypoxemia associated with extrapulmonary shunting of venous blood and evidence of elevated levels of pulmonary artery pressure in the absence of congenital heart disease. This syndrome is seen more commonly in term infants who have underlying diseases such as meconium aspiration, respiratory distress syndrome, sepsis, or lung hypoplasia, or it may be idiopathic PPHN.1 2 In the United States, approximately 10 000 newborns per year suffer from PPHN.3 4 The diagnosis of PPHN is usually made by 24 hours after birth and most patients are born at hospitals without extracorporeal membrane oxygenation (ECMO).4
The standard therapy for PPHN typically includes conventional mechanical ventilation, oxygen, sedation, paralysis, alkalosis, inotropic support, intravenous vasodilators, and antibiotics.4 5 In 1994, the efficacy and safety of surfactant and high-frequency ventilation for PPHN were unproven, and the use of these therapies was becoming more widespread, before resorting to ECMO.4-7 However, previous case series indicated that inhaled nitric oxide (I-NO) improved oxygenation acutely by selective pulmonary vasodilation.8 9 Therefore, the overall hypothesis for the present study was that early use of I-NO could reduce inspired oxygen and conventional ventilator requirements. This, in turn, might lead to less secondary lung injury due to the inflammatory effects of hyperoxia and barotrauma, a process that begins within 24 hours10 11 and may be deleterious to outcome. Accordingly, the first objective of this trial was to determine if early I-NO therapy would reduce the overall incidence of death, need for ECMO, neurologic sequelae, and bronchopulmonary dysplasia (BPD).
At the start of this study, data from animal and human studies had not indicated what was the most safe and effective dose of prolonged I-NO for PPHN.12-15 The secondary objective of this trial was to determine if there was dose-related efficacy, methemoglobinemia, inspired nitrogen dioxide (NO2) levels, intracranial hemorrhage, or unsuspected adverse effects of I-NO in the neonatal period.16 Therefore, eligibility requirements were defined narrowly and potentially confounding, and investigational rescue treatments, such as high-frequency ventilation and surfactant, were prohibited.
This clinical trial was a randomized, double-masked, placebo-controlled, dose-response study. Twenty-five neonatal intensive care units enrolled patients; fifteen of the sites were ECMO centers. The protocol, protocol amendments and each institution's Informed Consent forms were approved by the local Institutional Review Board before patient enrollment. Written informed consent was obtained for each patient before enrollment. Equipment, treatment gases and funding based on patient recruitment at each site was provided by Ohmeda, PPD (Liberty Corner, NJ).
The primary hypothesis was that a fixed dose of I-NO at either 5, 20, or 80 parts per million (ppm), delivered to term infants with PPHN, would reduce the PPHN Major Sequelae Index (MSI) by 30%. This composite endpoint included the incidence of death, ECMO, neurologic sequelae in the neonatal period, or bronchopulmonary dysplasia/reactive airway disease (BPD/RAD). Neurologic sequelae in the neonatal period were defined as clinical or electroencephalogram-proven seizures or an abnormal cranial imaging study (demonstrating either hemorrhage, infarct, or other diagnoses) during the neonatal hospitalization. BPD was defined as the need for supplemental oxygen at 28 days after birth with a concurrent abnormal chest x-ray. RAD was defined as the need for bronchodilator therapy at discharge from the nursery.
Secondary hypotheses were that I-NO would: 1) produce a sustained improvement in oxygenation during the first 24 hours of treatment, 2) reduce the incidence of treatment failures due to hypoxemia and/or hypotension leading to institution of other forms of rescue therapy, and 3) reduce days on the ventilator, on supplemental oxygen, and length of hospital stay.
It was also hypothesized that there would be no increase in adverse events experienced by the neonates receiving I-NO as measured by methemoglobin levels, inspired NO2 levels, worsening of primary or secondary efficacy outcome measures, delay in initiating rescue therapy, adverse events described by the investigators, or abnormalities in clinical hematologic and biochemical blood studies. A separate comprehensive 1-year follow-up study was designed to examine the general pediatric, neurodevelopmental, and audiologic outcomes.
Patient Entry Criteria
Term infants (≥37 weeks gestation) with birth weights of ≥2500 g requiring mechanical ventilation having a fraction of inspired oxygen (Fio 2) of 1.0 were eligible within 72 hours of birth. Small for gestational age infants with birth weights of ≥2000 g were included if the gestational age was assessed to be ≥39 weeks. On study entry, an Infant Star conventional ventilator (Infrasonics, Inc, San Diego, CA) was used, with a continuous flow rate between 10 and 15 L/min. Intermittent mandatory ventilator rates >100 breaths/minute and inverse inspiratory to expiratory ratios were not permitted. For study entry, patients required an arterial blood gas with a Pao 2 of ≥40 and ≤100 Torr drawn from an indwelling postductal arterial catheter when the mean airway pressure was ≥10 cm H2O and the Fio 2 was 1.0. They also required a color Doppler echocardiogram with evidence of PPHN or a preductal versus postductal transcutaneous O2 saturation gradient of ≥10%. The following were considered echocardiographic evidence of PPHN: 1) a right-to-left or bidirectional ductal shunt, or 2) if the ductus was closed, a right-to-left or bidirectional foramen ovale shunt with either a tricuspid insufficiency jet with an estimated systolic pulmonary artery pressure ≥75% of systolic aortic pressure or posterior systolic bowing of the interventricular septum. Before starting the treatment gas, the patient had to have a chest x-ray within 12 hours and a head ultrasound within 24 hours. A standardized history and physical examination were required. Preductal and postductal transcutaneous O2 saturations were obtained with all arterial blood gas samples.
Exclusion criteria were lung hypoplasia syndromes, congenital heart disease (other than a small, hemodynamically insignificant ventricular septal defect) as determined by echocardiography, intracranial hemorrhage ≥grade 2, uncorrected polycythemia (hematocrit ≥70%), mean systemic arterial pressure <35 Torr, a lethal syndrome, a suspected or confirmed chromosomal abnormality, use of intravenous vasodilators after entry criteria were met at the study site, uncontrollable coagulopathy or serious bleeding, and enrollment in any other investigational drug or interventional study. Patients were excluded if they had received previous or concomitant surfactant therapy. Patients who received a trial of high-frequency ventilation within 6 hours before starting the treatment gas were also ineligible.
Masking Procedures and Randomization
Strict masking procedures and personnel designations were approved by the steering committee before a site enrollment. Clinical investigators remained masked to the group assignment through the 1-year follow-up for all patients in the study. The clinical investigator managed patient care, assured compliance to the protocol, and assigned adverse events.
The site's unmasked laboratory investigator randomized patients to a placebo or nitric oxide (NO) dose group from a scratch off card. The randomization was blocked for each site in a block size of four patients allocated to one of the four treatment gases. The laboratory investigator set up, calibrated, and operated the I-NO delivery device and measured methemoglobin. All sites used Ciba-Corning 270 Co-oximeters (Ciba Corning Diagnostics Corporation, Medford, MA) for methemoglobin levels.
As soon as the patient was randomized, baseline hemodynamic, ventilator, and blood gas analyses were obtained at three time points 15 to 30 minutes apart. A baseline methemoglobin level was also obtained. If the patient met entry criteria at the first two time points, the Fio 2 was reduced to 0.95, and the third baseline measurement was obtained. If the Pao 2 remained ≥40 and ≤100 Torr, treatment gas was begun immediately. A patient who failed baseline criteria was allowed one additional opportunity to meet baseline oxygen criteria if deemed stable by the clinical investigator.
The following laboratory studies were also obtained before starting the treatment gas: a complete blood count, serum creatinine, blood urea nitrogen, total protein, albumin, total bilirubin, alkaline phosphatase, lactic acid dehydrogenase, serum glutamic-oxaloacetic transaminase, total calcium, inorganic phosphorous, uric acid, and glucose.
The I-NO delivery system (Ohmeda, PPD, Madison, WI) was designed expressly to deliver NO mixed with nitrogen, or nitrogen alone (BOC Specialty Gases, Port Allen, LA) into the inspiratory limb of the ventilator circuit using a mass flow controller. A sample gas catheter was attached to the inspiratory limb of the ventilator immediately before the patient connection. Electrochemical detectors attached to the delivery device provided a continuous measurement of NO and NO2 (model EC90 NO monitor and model EC40 NO2 monitor, Bedfont Scientific Ltd, Kent, England). The accuracy of NO measurement for values between 0 to 5 ppm was ± 0.5 ppm and for 5 to 80 ppm, was ± 2 ppm, regardless of the Fio 2.17 Validation was performed by using a known standard, blended by mass flow controllers and verified by chemiluminescence (error of ± 1%). Electrochemical cell analyzers for NO2 have been shown to over estimate NO2 levels, due to the formation of NO2 in the sampling circuit.18 In the presence of oxygen, the NO2 monitor overestimates by 1.2 ppm at 80 ppm of NO, 0.3 ppm at 40 ppm of NO, and 0.1 ppm at 20 ppm of NO. Verification was performed using a selective NO2 ultraviolet absorbance analyzer. The linearity of the NO2 analyzers was within ± 3%.
On randomization, the laboratory investigator calibrated the delivery device with standard concentrations of NO (112 ppm) in nitrogen and NO2 (7.2 ppm) in nitrogen. The I-NO delivery device mixed the NO source gas 1:20 with ventilator gas. To deliver NO at doses of 0, 5, 20, and 80 ppm, source tanks of 0, 100, 400, and 1600 ppm of NO, balance N2, were used. Because delivery of any treatment gas diluted the ventilator gas by 5%, the maximal Fio 2 delivered to the patient was 0.95. Therefore, patients were placed on an Fio 2setting of 0.95 before starting treatment gas. The ventilator Fio 2 setting was increased to 1.0 when the treatment gas was started. As a result, an Fio 2of 0.95 was delivered on initiation of the treatment gas for all groups.
Protocol for Management During Treatment Gas Administration
No Fio 2 or ventilator changes were to be made over the first half hour of treatment gas. High-frequency jet and oscillatory ventilation were not permitted. Ventilator settings, heart rate, blood pressure, and postductal arterial blood gases, preductal and postductal transcutaneous oxygen saturations, inspired gas levels, and methemoglobin levels were obtained at 0.5, 1, 2, 3, 4, and then every 4 hours or as needed while on 100% treatment gas. Patients were not permitted to receive treatment gas for >14 days.
Study patients received a fixed dose of either 0, 5, 20, or 80 ppm of NO until one of four events occurred: 1) a treatment success was achieved, based on improved oxygenation (Pao 2≥60 Torr, Fio 2 <0.6, and mean airway pressure <10 cm H2O); 2) a treatment failure resulted, based on a decrease of Pao 2 <40 Torr for 30 minutes in the absence of a reversible mechanical problem, a mean systemic arterial pressure <35 Torr, the patient reached the site's ECMO criteria, 14 days of treatment gas had elapsed, or if remaining in the study was not in the best interest due to cardiopulmonary instability or local ECMO criteria that was not covered by the study's failure criteria; 3) inspired NO2 levels were >3 ppm for 30 minutes19; or 4) a methemoglobin level that exceeded 7%. All patients, whether treatment successes or failures, were included in the data analyses.
For treatment successes and failures, the protocol permitted sequential 20% decrements in treatment gas at a minimum of 30 minutes and maximum of 4 hours. Ventilator settings, an arterial blood gas, preductal and postductal transcutaneous saturations, inspiratory gas levels, and vital signs were required immediately before and 30 minutes after a 20% reduction. During this half hour period, it was requested that no ventilator change be made unless the Pao 2became <40 Torr. If this level of hypoxemia occurred on a reduction, the treatment gas could be increased 20%. The weaning process would begin again when the criteria for success were met or in the case of a treatment failure, the treatment gas would have to be discontinued before other rescue therapy was begun. Treatment gas could not be re-instituted and no other investigational drug or intervention was permitted. I-NO for treatment failures was not permitted.
For patients with elevated methemoglobin and inspired NO2levels based on the protocol definitions, treatment gas could be continued at a lower level (one of the 20% decrements) if the methemoglobin or NO2 levels dropped below threshold levels.
Posttreatment Gas Data
A methemoglobin level was obtained 2 hours after the treatment gas was discontinued. The baseline complete blood count and blood chemistries were repeated within 12 hours of discontinuing the treatment gas. A repeat head ultrasound, computerized axial tomography, or magnetic resonance imaging was required before discharge. A bilateral evoked response hearing screen was obtained before discharge. A chest x-ray was performed on day 28 if the patient still required supplemental oxygen.
An independent data safety and monitoring board was composed of statisticians and pediatric specialists. An interim, blinded safety analysis was performed after data from 100 patients were obtained.
Sample Size and Statistics
We estimated the incidence of PPHN major sequelae before the start of this trial by a retrospective survey of seven sites. Data were obtained on 107 patients who would have been eligible for the present study. The incidences of the major sequelae were: death, 8%; ECMO, 36%; neonatal neurologic sequelae 20%; and BPD, 5%. To determine if NO could reduce PPHN major sequelae by 30% (α level = 0.05, β level = 0.2), a total of 320 patients (80 in each of 4 groups) were required. The Cochran-Mantel-Haenszel χ2 test was used for discrete or categorical data, such as PPHN major sequelae (death, ECMO, neurologic sequelae, BPD, or composite), treatment failures due to cardiopulmonary instability, and adverse events by organ systems. Fisher's exact test was used if the frequencies were small.
The Wilcoxon rank sum test was used to analyze continuous variables (eg, the time-weighted oxygenation index [TWOI], duration of supplemental oxygen, acute change in Pao 2, or methemoglobin levels). For ventilatory and hemodynamic data, baseline was considered as the third qualifying time point (Fio 2 = 0.95) immediately before starting the treatment gas. A two-tailed t test was used only for the change from baseline for the clinical hematologic and biochemical variables. The significance level for all tests was set at 0.05. There was no α level adjustment for pairwise comparisons performed in the study. The incidence of adverse events was tabulated using the COSTART body system classification.20
One of the major secondary endpoints of this study was to determine whether I-NO produced a sustained improvement in oxygenation. Therefore, we prospectively defined the TWOI as the change in oxygenation index (OI) from the individual's baseline OI over time, divided by the duration on treatment gas up to 24 hours. This method adjusts for attrition from treatment failure or success. If the patient worsened from his/her own baseline, the TWOI would be a positive number. If the patient improved, the index would be a negative number as shown in Fig 1.
Enrollment began in April 1994. One hundred fifty-five patients were enrolled. The trial was halted in June 1996 because of slow recruitment. The accrual goal was 320 patients.
Patients Screened and Patients Enrolled
A total of 1282 patients were screened. The most common conditions preventing enrollment were: oxygenation outside the eligible range (26%) and lack of echocardiographic evidence of PPHN (19%). Surfactant therapy (12%), high-frequency ventilation (9%), prematurity (8%), lung hypoplasia syndromes (8%), and age >72 hours (5%) were the other conditions preventing enrollment.
A total of 8 patients, 2 in the control group and 6 in the NO groups were erroneously enrolled. All patients were included in the efficacy and safety analyses. Of the 155 randomized patients, the number of patients who received treatment in each group (0, 5, 20, and 80 ppm) were 41, 41, 36, and 37 respectively.
Baseline Patient Profile
Baseline variables were similar among treatment groups. Baseline data are presented (Tables 1 and2) for the control group (N = 41) and the pooled I-NO group (N = 114). The majority of patients were delivered by cesarean section (62%) and at hospitals other than the study site (over 90%). Most infants (77%) received conventional mechanical ventilation and 5% received high-frequency jet or oscillatory ventilation before admission to the study site. Underlying conditions associated with PPHN were also well-balanced among treatment groups with a majority of patients diagnosed with meconium aspiration. Although there were more patients with idiopathic PPHN in the NO groups compared with placebo, this was not statistically significant. Seizures were documented in 17% of the control patients and 20% of the patients who went on to receive NO. Abnormal head ultrasounds, almost all due to low-grade intracranial hemorrhages, were demonstrated in 10% of the control patients and 5% of the pooled NO patients.
The baseline ventilatory (Table 2) and hemodynamic conditions were also very similar between control and NO groups. The patients required high levels of conventional ventilatory support. The baseline inspiratory Fio 2 for all patients was 0.95 by protocol. Systolic, mean, and diastolic, systemic arterial pressures were 67 ± 13, 53 ± 10, and 44 ± 10 Torr, respectively. Dopamine and/or dobutamine (to a lesser extent) were used in 76% and 74% of the control and the pooled NO group, respectively, at the start of the treatment gas. Most patients had echocardiographic evidence for PPHN; only 9% were diagnosed by preductal versus postductal oxygen saturation difference of >10%. The mean preductal and postductal transcutaneous oxygen saturations were similar for all groups; for the control group the saturations were 93.6 ± 3.0% and 93.2% ± 4.1%, respectively.
Acute Changes in Blood Gases, pH, and Hemodynamics
The acute changes in Pao 2 after the first half hour of treatment gas, on stable ventilator settings are shown in Fig 2. There was a statistically significant increase in Pao 2 from baseline for each NO group compared with control. Although a higher mean Pao 2 was observed for the 80 ppm dose, this was not statistically different from the other NO doses. The corresponding OIs at 30 minutes for each group from lowest to highest NO dose were 24 ± 14, 20 ± 11, 21 ± 13, and 15 ± 10; the pooled NO value was 19 ± 11 (cm H2O/Torr). There were no appreciable differences between and within groups for pH, Paco 2, or ventilator settings at 30 minutes compared with baseline. At 30 minutes for control and pooled NO groups respectively, the pH was 7.49 ± 0.11 and 7.52 ± 0.12, the Paco 2 was 32 ± 71 and 29 ± 10 Torr, the peak inspiratory pressure for both groups was 33 ± 6 cm H2O, the ventilator rate was 59 ± 15 and 58 ± 14 breaths/minute, and the positive end expiratory pressure was 5.1 ± 1.6 and 5.0 ± 1.2 cm H2O.
Systolic, mean, and diastolic systemic arterial pressures, as well as heart rate, remained steady during the first half hour of treatment gas. The mean arterial blood pressure for the 0, 5, 20, and 80 ppm groups at 30 minutes of treatment gas were 54 ± 9, 52 ± 9, 50 ± 9, and 50 ± 10 Torr; the value for the pooled NO group was 51 ± 9 Torr. The corresponding heart rates at 30 minutes, from the lowest (control) to highest NO dose were 151 ± 20, 151 ± 24, 155 ± 27, and 155 ± 24 beats/minute; the value for the pooled NO group was 154 ± 25 beats/minute. Blood pressure and heart rate data were similar for all groups during administration of the treatment gas.
The TWOI corrected from baseline was used as a measure of a sustained improvement in oxygenation. Figure3 demonstrates that the TWOI improved (negative number) for all NO groups. For the control group, this index was not statistically improved from baseline. There was a highly significant reduction in the baseline-adjusted, TWOI for the pooled NO group compared with placebo.
The incidence of any treatment failure for the control group was 37% and for the 5, 20, and 80 ppm NO doses, 29%, 28%, and 59%, respectively, (the pooled value for NO subjects was 39%). The higher incidence of treatment gas failures in the 80 ppm group was related primarily to methemoglobinemia (13 of 37, 35%) and elevated NO2 levels (7 of 37, 19%). The incidence of treatment failures due to cardiopulmonary instability (hypoxemia or hypotension criteria) was reduced by NO (control 34%, the pooled value for NO subjects was 25%), but this result did not reach statistical significance (P = .29); the incidences of treatment failures for each group were 34%, 27%, 25%, and 24% for 0, 5, 20, and 80 ppm NO groups, respectively. Hypoxemia was the principal cause of treatment failure due to cardiopulmonary instability. Only 3 patients were classified as treatment failures due to hypotension, 2 in the 5 ppm and 1 in the 80 ppm NO groups. The frequency of premature discontinuation of treatment gas when the investigator felt it was in the best interest of the patient (eg, immediate rescue therapy was needed) was 20% for the control group and 18% for the pooled NO group. One patient in the placebo group had the treatment gas discontinued because of a malfunction in the inspiratory gas monitor. For patients classified as treatment failures based on hypoxemia and/or hypotension, the incidence of rescue therapy for placebo and the pooled I-NO group, respectively, were: ECMO (71% and 48%), high-frequency oscillatory ventilation (64% and 72%), surfactant (36% and 45%), and vasodilator therapy (36% and 21%).
The duration of treatment gas, mechanical ventilation, supplemental oxygen, ECMO, and hospitalization are shown in Table3. These results give a descriptive view of the hospital course but are difficult to analyze because of the issue of dropouts in both the placebo and NO groups. Patients generally received the treatment gas for about 2½ days. The mean durations on treatment gas for patients with treatment success were 107.4, 95.4, 72.2, and 65.1 hours for the control, 5, 20, and 80 ppm groups, respectively. Thus, among treatment successes, it appeared that there was a dose-response trend in duration of treatment gas. Patients who became treatment failures due to cardiopulmonary instability had gas discontinued at similar times in the control group (9.9 ± 11.0 h) and the NO groups (10.0 ± 12.7 h). The mean TWOIs for the treatment failures due to cardiopulmonary instability were +1.25 in the control group and +1.49 in the NO group; this was not statistically different. The mean OI at the time a patient was declared a treatment failure due to cardiopulmonary instability was 52.7 ± 16.3 and 43.2 ± 12.4 for the control and treatment gas, respectively, (P = .15). The duration of supplemental oxygen for the control and the 5, 20, and 80 ppm groups was 155, 122, 124, and 141 hours, respectively, were not statistically different from each other.
The PPHN MSI and its four components for the control and pooled NO groups are shown in Table 4. The number of patients with at least one major sequela was 56% (23/41) in the placebo group and 50% in the pooled NO group; this difference was not statistically significant. There were no statistically significant reductions in the components of the index for any of the NO doses compared with placebo. Although there were numerically more deaths before 28 days in the NO group, this result was not statistically significant and there was no dose-related trend. One patient in the placebo group died 92 days of age from sudden infant death syndrome but was not included in Table 4 (which includes data on death <28 days). There was no dose-related trend for the incidence of neurologic sequelae in the neonatal period or BPD. The incidence of BPD/RAD for the 0, 5, 20, and 80 ppm doses were 13%, 24%, 10%, and 9%, respectively.
To be consistent with another large clinical trial,20deaths through 120 days were included in the analysis of the incidence of death and/or ECMO. The incidence of death and/or ECMO was lower for the NO doses (29%) compared with control group (39%), but this difference was not statistically significant (P= .25). The results incidence of this combined outcome was due to the incidence of ECMO. The use of ECMO for the 0, 5, 20, and 80 ppm doses was 34%, 24%, 25%, and 16%, respectively; the pooled NO value of 22% was not significantly different from control (P = .12).
Adverse Events and Safety Data
With the exception of methemoglobinemia and elevated inspired NO2 levels, the incidence of adverse events was similar between the control and NO groups. The time course of change in methemoglobin levels for each dose of NO over the first 2 days is shown in Fig 4. Methemoglobinemia >7% occurred in 13 patients, all from the 80 ppm NO group (n = 37). By protocol mandate, this level of methemoglobin required either discontinuation of the treatment gas or reduction of the treatment gas by 20% decrements until the methemoglobin level was below 7%. The average time to the peak value was 19.6 ± 27.5 hours (median = 8 h). The highest level of methemoglobin was 11.9% at 8 hours in 1 patient.
The mean NO2 levels over the first 12 hours are shown in Fig 5. NO2 levels for the 80 ppm NO group rose rapidly. Seven patients (only from the 80 ppm group) had NO2 levels briefly >3 ppm. NO2 levels then remained around 1.5 ppm for the 80 ppm NO group over the next 2 days. All other groups remained below 0.5 ppm over the next 2 days.
The occurrence of death for the control, 5, 20, and 80 ppm groups was 2, 2, 5, and 3, respectively. Deaths in the first 28 days generally occurred late (10 ± 9 days), well after stopping the treatment gas. The duration of treatment gas, for the patients who died was 68 ± 51 hours. The age at treatment failure was 24 ± 14 hours. Four patients were successfully weaned off treatment gas, but died later. Of the two control patients who died, 1 had an XXXXY karyotype (and severe respiratory failure) and 1 had sudden infant death syndrome listed as the cause of death at 92 days. Of the patients who received NO, 2 patients were on ECMO at or near death, 1 died of alveolar capillary dysplasia and the other died ofPseudomonas sepsis. Five patients expired with pneumothorax, 1 with group B streptococcal sepsis, 1 with perinatal asphyxia, and 1 was listed as a cardiopulmonary arrest (probably related to sepsis) at 33 days. The deaths occurred mostly in ECMO centers. The investigators listed two deaths in the NO group as possibly, but remotely, related in time and the other 10 deaths were listed as no relationship to treatment gas.
A total of 24 adverse events in 18 patients were judged to be serious by the local principal investigators, yielding a 7% rate for the control group and an 11% rate for the NO groups. One surviving infant in the 80 ppm group had persistent conjugated hyperbilirubinemia, which was considered by the investigator as a serious adverse event possibly related to study gas. Otherwise, the serious adverse events were considered unrelated or as a remote possibility to the treatment gas.
The incidence of all adverse events when tabulated by the COSTART body system classification, was not statistically significant between placebo and the pooled NO doses except for hematologic adverse events due to methemoglobinemia in the 80 ppm group. The mean values for the baseline clinical laboratory tests were normal except for a mean absolute band count of approximately 2500 for the entire study population. Blood chemistries and complete blood counts taken at baseline and within 12 hours of discontinuing the treatment gas showed no evidence of dose-dependent trends.
This trial, involving term infants with PPHN, succeeded in studying patients who were seriously ill but early in the course of this labile disorder. Control patients had a mean OI of 25, and were started on the treatment gas at 26 hours of age. Of these patients, 37% either died or received ECMO. This patient profile includes infants in whom I-NO therapy may improve oxygenation and prevent the progression of cardiopulmonary failure to ECMO criteria.5 6 In general, and not unexpectedly, the patients in this study were mildly hyperventilated at baseline using conventional ventilators.4 5 The results of the trial were not confounded by efficacy and safety issues that could have arisen from previous or concomitant surfactant therapy22 23 or concomitant high-frequency ventilation.24-26 The trial demonstrated that I-NO produced an acute and sustained improvement in oxygenation for up to 24 hours using 5, 20, or 80 ppm in a majority of patients. The major safety concerns were methemoglobinemia and elevated NO2 levels, which were problematic only in the 80 ppm NO group.
This study was performed during a rapidly closing window of opportunity, in which I-NO could be studied independently of the widespread use of other investigational rescue therapies for PPHN.27 Rigorous enrollment criteria to determine the efficacy and safety of I-NO led to a drop in enrollment from 10 to 15 patients per month to <5 per month. In addition, investigators felt that economic disincentives to transfer patients to study sites, and the large number of sites using I-NO in and outside of the large clinical trials posed problems.27 The trial was halted prematurely after only half the planned number of patients had been enrolled. Therefore, the trial was underpowered to demonstrate statistically significant improvements in the primary outcome endpoint, the PPHN MSI. In addition, the trial suggested, but did not prove, that early I-NO reduced the need for ECMO because there was a 35% reduction (P = .12) halfway to the patient accrual goal.
One of the goals in the management of PPHN is to improve oxygenation by selective pulmonary vasodilation. In several pioneering case series published before this trial, it was shown that I-NO produced acute improvements in oxygenation within 15 to 30 minutes.8 9 In another case series, it appeared that patients with PPHN responded appreciably better with acute improvements in oxygenation than those with hypoxemia due to parenchymal lung disease only; however, no dose-response could be found.14 Similarly, in the present study, acute improvements in oxygenation were observed for 5, 20, and 80 ppm NO groups, although there appeared to be a trend toward better oxygenation in the 80 ppm group. The lack of change in systemic arterial pressure and heart rate for any of the NO doses supports previous claims of the selective pulmonary vasodilating properties of I-NO. In a recently completed clinical trial21 (the Neonatal Inhaled Nitric Oxide Study—NINOS) that enrolled patients ≥34 weeks gestational age with PPHN as well as patients with hypoxemia but no PPHN, NO at 80 ppm only rarely improved oxygenation when 20 ppm did not. In the NINOS and present studies, approximately 50% to 60% of the patients had ≥20 Torr increase in Pao 2within 30 minutes of administering I-NO.
One of the unique contributions of the present study is the demonstration that I-NO caused a sustained improvement in oxygenation for the duration of the treatment gas or 24 hours (whichever came first) without surfactant or high-frequency ventilation. The TWOI was used to prospectively measure oxygenation over time in a patient population where discontinuation of the various treatment gases could occur at differing rates if the patient reached treatment failure or success criteria. Furthermore, this study did not indicate that there were any dose-dependent differences between the 5, 20, and 80 ppm doses' ability to produce a sustained improvement in oxygenation.
This study also provides supporting data related to the improvement (decrease) in the TWOI, indicating that patients who received NO become more stable in terms of oxygenation. Although the following variables did not reach statistical significance individually, together they indicate a trend toward stability. For the pooled NO group, there was a reduction in the time to reach success criteria, the median time to treatment failure due to cardiopulmonary instability was longer, and the time on treatment gas for those patients who required ECMO was also longer. However, it could be argued that any potential stability in oxygenation for the treatment failure group could delay the use of ECMO, which is an accepted and proven therapy.28
The primary outcome endpoints (death, need for ECMO, neonatal neurologic sequelae, and bronchopulmonary dysplasia) were not statistically different between the pooled or separate NO groups and the control group. These results are not significant, possibly because only about half of the planned enrollment occurred; however, the results are comparable to other trials.21 29 The only trend toward efficacy was the 35% reduction in the need for ECMO for the pooled NO group compared with placebo; the need for ECMO was lower for each of the three NO doses. The NINOS Group21 and Roberts et al30 reported that NO produced a significant reduction in the need for ECMO in their patients who, at study entry, were generally older (approximately 1.3 to 2 days) and more ill (mean OI of approximately 45); a 30% and 44% reduction in the use of ECMO was demonstrated in each study respectively. As in these previous trials,21 29 we found no difference in BPD, duration of supplemental oxygen, or neonatal neurologic sequelae.
Another unique contribution of the present trial is the large data set that was acquired to examine the safety of I-NO, independent of other rescue therapies that could have their own adverse events.26 Other than methemoglobinemia and elevated NO2 levels in the 80 ppm NO group, no adverse events in the neonatal period or abnormalities in clinical laboratory data could be definitely attributed to NO. An increase in bleeding was not observed despite previous reports of a prolongation of bleeding time with I-NO.30 However, the low incidence of grade 1 intraventricular hemorrhage (4%) and the entry exclusion criteria for Grades 2–4 does not provide any conclusive safety information regarding the use of I-NO in patients with intraventricular hemorrhage. No adverse effects reflecting increased lung inflammation by I-NO could be found.31 A separate long-term follow-up study of patients in this trial will be reported in the near future.
Methemoglobinemia, defined in our study as >7%, occurred only in the 80 ppm group (1/3 of the patients). There was a rapid rate of rise in methemoglobin with a peak occurring at a median of 8 hours in contrast to the findings by Roberts et al29 who started NO at 80 ppm and saw a gradual rise in methemoglobin over several days. The difference may be due to study design; their protocol permitted NO to be reduced twice a day if oxygenation was satisfactory. Furthermore, many commonly used co-oximeters are less accurate in distinguishing fetal hemoglobin from methemoglobin than the device used in the present study. A study by Wessel et al32 who used 80 ppm of NO in children with pulmonary hypertension, found an increase in methemoglobin levels closer to what was observed in this study.
NO2 levels are proportional to the second power of the inspired NO concentration and to the first power of the inspired oxygen concentration.33 Accordingly, the NO2 levels in the 80 ppm NO group of this study were significantly higher (NO2 generally 2–3 ppm) compared with the lower NO doses (NO2 generally under 0.5 ppm), with seven patients in the 80 ppm NO group reaching an NO2 level ≥3 ppm. Part of the elevation in NO2 levels are related to the electrochemical technique, which overestimates NO2 levels due to formation of NO2 in the sampling circuit. However, elevated NO2 levels could have been observed more frequently if treatment gas had not been reduced or discontinued because of methemoglobinemia in one third of the 80 ppm group. Comparison of NO2 levels among studies is complicated by different assay methods and sampling circuitry formation with different delivery devices. The present study shows that using NO doses of ≤20 ppm and the delivery system does not result in NO2 levels above safety standards.19 In addition, it is worth noting that NO therapy was not associated with an increase in BPD, an important finding because of the theoretic concern that low levels of NO2 and peroxynitrite could theoretically cause lung inflammation.31 In fact, the days on oxygen were lower in the NO groups, and lowest in the 80 ppm NO group (although not statistically significant). This study does not rule out the possibility that more prolonged use of I-NO at high doses could lead to clinically significant airway inflammation.
I-NO produced an acute and sustained improvement in oxygenation for 24 hours in a majority of term infants with early but moderately severe PPHN. Furthermore, oxygenation appeared to be stabilized to a greater degree by NO whether the patient became a treatment success or failure. The suggestion that ECMO be reduced by 35% was similar to statistically proven reductions in ECMO observed in two other clinical trials. The 80 ppm NO dose did not seem to have any advantages over the 20 and 5 ppm doses and resulted in elevated methemoglobin and NO2 levels. Although this truncated study may be underpowered for definitively answering the primary hypotheses, the rigorous entry criteria render the findings highly reliable and specific to the therapeutic benefit and safety of I-NO for PPHN. Future studies are needed to determine early predictors of success, adverse effects of discontinuing NO34 35 (especially for treatment failures), as well as long-term safety.
This study was sponsored by Ohmeda, PPD.
This clinical trial was a multicenter collaboration sponsored by Ohmeda, PPD. The steering committee was composed of the co-authors of this article. In addition, the following institutions and investigators participated in this trial. University of Virginia School of Medicine, Charlottesville, VA—M. Pamela Griffin, MD, Martina Compton, RRT; Children's Hospital of Orange County, CA—Robert Langdon Hillyard, MD, Elizabeth Phyllis Drake, RN, PNP, Albert Rocky Stone, RRT; Children's Hospital and Health Center, San Diego, CA—Marva L. Evans, MD, Gail R. Knight, MD;Schneider Children's Hospital, New Hyde Park, NY—Andrew M. Steele, MD, Robert Koppel, MD, Angela Romano, MD; University of Kentucky Children's Hospital, Lexington, KY—Thomas H. Pauly, MD, John R. Walker, DO, Vicki Whitehead, RN; University of Alabama at Birmingham, Birmingham, AL—Virginia A. Karle, MD, Monica V. Collins, RN; St Joseph's Hospital and Medical Center, Phoenix, AZ—Mark L. Shwer, MD, Donna C. Hamburg, MD; Boston Perinatal Center, Floating Hospital for Children at New England Medical Center, Boston, MA—Barbara A. Shephard, MD, John M. Fiascone, MD, Ivan D. Frantz, MD; Legacy Emanuel Children's Hospital, Portland, OR—John V. McDonald, MD; Medical College of Georgia Hospital & Clinics, Augusta, GA—D. Spencer Brudno, MD, William Grayson, RRT; Children's Hospital of Oklahoma, University of Oklahoma, Oklahoma City, OK—K. C. Sekar, MD, Mary Anne McCaffree, MD, Mike McCoy, ARNP; Pennsylvania Hospital, Philadelphia, PA—Vinod K. Bhutani, MD, Emidio M. Sivieri, Mary Kay Grous; Columbia Wesley Medical Center, Wichita, KS—Barry T. Bloom, MD, Pamela S. Keyes, RN, ARNP, Tom L. Rose, RRT; Crouse Irving Memorial Hospital, Syracuse, NY—Ellen M. Bifano, MD; Children's Hospital of the King's Daughters/Eastern Virginia Medical School, Norfolk, VA—Jamil H. Khan, MD, Marilyn M. Reininger, RN; The Bowman Gray School of Medicine, Winston-Salem, NC—Steven M. Block, MD; Children's Hospital of New Jersey, Newark, NJ—Morris Cohen, MD, David Brown, MD;University of California, Davis Medical Center, Davis, CA—T. Allen Merritt, MD, Boyd Goetzman, MD; Ochsner Foundation Hospital, New Orleans, LA—Erick M. Fajardo, MD, Janet E. Larson, MD, Marie C. McGettigan, MD; East Carolina University School of Medicine, Pitt County Memorial Hospital, Greenville, NC—Steven C. Covington, RRT, John E. Wimmer, MD, Arthur E. Kopelman, MD; Columbia Presbyterian/St Luke's Medical Center, Denver, CO—Barbara J. Quissell, MD, Delphine M. Eichorst, MD;Hope Children's Hospital, Christ Hospital and Medical Center, Oak Lawn, IL—Arvind Shukla, MD, Manohar Rathi, MD, Alison Miklos, RNC; Duke University Medical Center, Durham, NC—Richard L. Auten, MD, Kathy J. Auten; University of South Dakota School of Medicine, Sioux Falls, SD—Dennis C. Stevens, MD, David P. Munson, MD, Rachel D. Klinghagen, RN, CNP; Santa Rosa Children's Hospital, San Antonio, TX—Anthony Corbet, MD, Sue Pape, RN, William Papp, RRT.
Data Safety and Monitoring Board: Statistics Collaborative, Inc., Washington, DC—Janet Wittes, PhD (Chair); Medical College of Virginia, Richmond, VA—Stephen Ayres, MD; Boston Floating Hospital, Boston, MA—Jonathan Rhodes, MD; St Vincent's Hospital, New York, NY—Jayne Rivas, MD; Yale School of Medicine, New Haven, CT—Stanley Rosenblum, MD;Children's National Medical Center, Washington, DC—Billie Short, MD.
The scientific and technical support of Anne Berssenbrugge & Greg Beltrand from Ohmeda, PPD, and Michael Karol from Long Island Jewish Medical Center is gratefully acknowledged. The investigators are indebted to the neonatologists, nursing staffs, and respiratory therapists at their respective sites. The assistance of Irene Barling, Barbara Wilkens, and Diane Davidson at the coordinating study center (Schneider Children's Hospital) is greatly appreciated.
- Received October 8, 1997.
- Accepted December 11, 1997.
Reprint requests to (D.D.) Division of Neonatal-Perinatal Medicine, Schneider Children's Hospital, Long Island Jewish Medical Center, New Hyde Park, NY 11040.
Members of the I-NO/PPHN Study Group are listed in the Appendix.
- Gersony WM,
- Duc GV,
- Sinclair JC
- Linde-Zwirble WT,
- Angus DC,
- Clermont G,
- Roberts MS,
- et al.
- Davis JM,
- Whitin J
- Roberts JD,
- Chen TY,
- Kawai N,
- et al.
- ↵Sokol GM, Van Meurs KP, Thorn WJ, et al. Limitations in nitrogen dioxide measurement with commercially available analyzers. Presented at the Twelfth Annual Children's National Medical Center Symposium on ECMO and Advanced Therapies for Respiratory Failure; Keystone, CO. (Abstract #71). 1996
- Centers for Disease Control
- ↵US Department of Health and Human Services. COSTART. Coding symbols for thesaurus of adverse events. FDA Office of Epidemiology and Biostatistics. 3rd ed. Washington, DC: US Department of Health and Human Services; 1989
- Lotze A,
- and The Survanta in Term Infants Study Group
- Findlay RD,
- Taeusch HW,
- Walther FJ
- Wiswell TE,
- Graziani LJ,
- Kornhauser MS,
- et al.
- Stark AR,
- Davidson D
- Copyright © 1998 American Academy of Pediatrics