6.0.1.13.2c Subject deaths (coat)
`The individual patient deaths are listed in the table below. No death narratives for the individuals are available.
`
`NDA 20-845 Nitric Oxide
`
`
`
`
`
`Table 6. 0. 1. 13.2c.2 from table 8 1.1.2) DeathsIn the NINOS study“.
`
`Trial
`Received
`Time of Death
`Description"
`
`ECMO?
`da 5
`Control group
` Muiti-organ failure,
`
`withdrawal of support
`
`Severe hypoxia
`Severe intracranial hemorrhage
`
`Refractory pulmonary hypertension
`
`Severe intracranial hemorrhage,
`
`withdrawal of support
`-
`
`Suspected sepsis/infection
`
`Alveolar-capillary dysplasia
`
`Alveolar-capillary dysplasia
`
`Suspected sepsis/infection
`
`Lefi ventricular failure
`
`RDS
`
`Suspected sepsis/infection
`
`
`Severe pulmonary hypertension
` .h
`
`Polycystic kidneys
`Proven sepsis/infection
`
`Pulmonary hypoplasia
`Proven sepsis/infection
`Suspected sepsis/infection
`
`Severe intracranial hemorrhage
`
`Broneho-pulmonary dysplasia
`
`
` l-NO group
`
`Respiratory failure
`
`Severe CNS ischemia
`Suspected sepsis/infection
`RDS
`
`'Thrombi',
`
`Withdrawal of support
`
`Alveolar-capillary dysplasia
`
`Proven sepsis/infection
`
`Pulmonary lymphangiectasia
`Meconium aspiration
`
`
`
`
`
` Withdrawal of support
`
`Proven sepsis/infection
`
`
`Suspected sepsis/infection
`
`Proven sepsis/infection
`Surgical death
`
`RDS
`
`
`a. Any death prior to no days:5 includedin the NINOS data.
`b. Study subjects are identified by center it and patient # (e.g. 05--A04).
`c. Cause of death from electronic datasets of summary clinical data
`
`6.0.1.13.3 Long-term safety results of the NINOS trial
`Data on the neurodevelopmental outcomes of the survivors is to be collected at 18 to 24 months corrected age.
`No interim results are available.
`
`
`
`46
`
`

`

`NDA 20-845 Nitric Oxide
`
`6.0.1.14 NINOS Efficacy Summary
`Inamen
`This was a multi-center, multi-national, double-blind, placebo-controlled trial to evaluate the efficacy of I-NO in
`thetreatment of term and near-term infants with hypoxic respiratory failure.
`-
`Subjects with hypoxic respiratory failure (see inclusion and exclusion criteria) were randomized to receive either
`01 (no flow of l-NO) or l-NO, 20 ppm for up to 336 hours (14 days). A total of 121 control and 114 I-NO subjects were
`enrolled.
`
`Subjects who responded fully to treatment gas (either control or l-NO) were continued on the 'low-flow‘ study
`gas. For the subjects who received control gas, 17/117 (14.5%) had a full response. In the 20 ppm l-NO group, 57/113
`(50.4%) had a full response (p value vs. control <0.001).
`Subjects who had no response, or responded partially, were entered into the 'high-flow gas' protocol. These
`subjects were administered either placebo gas (02) or to l-NO, 80 ppm), depending on their initial randomization, and
`their response measured after another 30 minutes. For the subjects who received control gas, none had a full response (0%)
`to high-flow control gas (02).
`In the 80 ppm I-NO group, 1/17 (6%) had a full response (no statistical comparison
`possible).
`.
`Non-responders to the high-flow gas were weaned off of the study gas. They were eligible for a repeat trial of the
`same study gas (either low- or high-flow) after 6 hours, so long'as the infant was still otherwise eligible. This process
`could be repeated 3 times. lfno positive response was observed afier 3 repeat trials (a total of 4 trials), the subject was
`labeled a non-responder. Despite this detailed repeat trial protocol, only 3 subjects in the control group (3%) and 2 in the '
`l-NO group (2%) underwent re-initiation of study gas.
`
`WM
`Primary endpoint
`The incidence of'death before discharge or 120 days (whichever comes first), and/or the initiation of ECMO
`between placebo- and l-NO-treated subjects.
`'
`The primary endpoint included one part looking at an unquestioned clinical benefit (reduction in mortality) and a
`component with a less-clear clinical benefit (initiation of ECMO). The results (see below) were completely driven by the
`reduction in the percentage of infants who received ECMO.
`
`Secondary endpoints
`1. Change in PaO; levels meaured 30 minutes after initial administration of the study gas.
`. Change in mean 01 levels measured 30 minutes after initial administration of the study gas.
`. Change in Aa-DO; levels before and 30 minutes after initial administration of the study gas.
`. Neurodevelopmental outcomes assessed at 18-24 months corrected age.
`. The average length of hospitalization among surviving infants.
`. The number of days of assisted ventilation.
`. The incidence of air leak.
`
`. The incidence of chronic lung disease.
`. The proportion of infants transferred for potential ECMO.
`
`\OOOQONLth-UJM
`
`The secondary endpoints can be broken into three groups: 1) measures of acute effects of l-NO on oxygenation; 2)
`measures of clinical outcomes measured at time of discharge; and 3) long-term neurodevelopmental outcomes.
`
`He 1
`C
`l'
`I
`l
`.
`.
`A total of 250 subjects were planned for enrollment. A total of235 enrolled: 121 subjects in the control group
`and 114 in the l—NO group. While the trial was multi-center, three centers accounted for 37% of the enrolled infants
`(Wayne State University, Stanford University/Packard Children’s Hospital, and Baylor Hospital/Texas Children's
`Hospital).
`-
`
`Two teams were used to accomplish the blinding in the trial. The first team consisted of the patient caregivers,
`who were blinded to the treatment gas being administered. The second team consisted of a least one unblinded
`investigator, who was responsible for all activities that revealed the treatment gas. These activities included maintenance of
`the bedside stock of treatment gas, daily calibration of the gas blender, and recording the methemoglobin, l-NO, and NO;
`levels.
`
`
`
`47
`
`

`

`NBA 20-845 Nitric Oxide
`
`6.0.1.14 NINOS Efficacy Summary (cont)
`1
`l
`.
`l E
`l
`.
`: .
`.
`‘
`NINOS has several important differences from the INOSG and INO-Ol/ -02 trials with regard to the subjects
`included in the trial.
`'
`‘
`
`First, subjects did not have to have echocardiographic proof of pulmonary hypertension. Indeed, l9% of control
`and 26% of l-NO infants did not have the clinical diagnosis of PPHN (see table 6.0.1.1213). Additionally, 37% of the
`control subjects and 41% of the I-NO subjects had lefi-to-right shunting of blood across the patent foramen ovale.
`Second, infants who had previously received surfactant and/or high-frequency ventilation were not excluded from
`the trial (see table 6.0.1.1216). Over 70% of the infants in both groups had received surfactant, and over 30% had
`received high-frequency ventilation.
`Congenital diaphragmatic hernia (CDH) was nOt an exclusion criteria in the NINOS trial, although those subjects
`were not included in the subjects for the primary analysis. This contrasts with the INOSG'trial and INO-Oll ~02 trial,
`where CDH was an exclusion criteria. No data on the effects of l-NQ in the CDH population were submitted with this
`NDA.
`'
`
`The impact of these differences in the NINOS trial was that even critically ill neonates, who were already
`receiving maximal standard therapy, were eligiblg for enrollment. ln distinction, the INO-Ol/ -02 trial excluded infants
`"who had recently received surfactant or high-frequency ventilation, and required that the infant have PPHN prior to*
`enrollment. For this reason, they had to be stable enough that the investigator was not forced to start these interventions
`while the enrollment process went on (including the determination of PPHN by ECHO). Additionally, the investigators in
`the [NO-011' -02 trial may have selected only 'less-ill'
`infants for consideration for that trial, preferring to start other,
`established, therapies for the critically ill infants. These differences are reflected in the significantly higher 01 in the NlNOS
`trial, relative to the IND-0U -02 trial (averaging 22-25 in the INO-Olf -02 trial versus 42-44 in the NINOS trial (see tables
`6.0.3.1213 and 6.0.1.1214).
`'
`
`Of the infants enrolled in the trial, 117 infants received control gas and 113 received l-NO, 20 ppm. In the I-NO
`group, 57/113 of the infants responded initially to 20 ppm I-NO, while 55/113 infants did not have a full response to 1-
`NO 20 ppm, and so were administered I-NO 80 ppm.
`in the
`Individuals in both groups received treatment gas promptly after randomization, save for one individual
`placebo group who started treatment gas 43 hours after randomization: 26.3 minutes was the mean time to start cf
`treatment gas in the control group and 29.3 minutes in the I-NO group. Over 50% of the infants in both groups started
`treatment gas in <15 minutes.
`Five individuals did not receive study gas after enrollment in the trial. Another seven individuals received l-NO
`after being randomized to receive control gas. These individuals are listed in 'section 6.0.1.12.3a above, along with an
`analysis of the results according the actual gas received.
`Finally,
`isolated individuals received non-standard amounts of I-_NO. The table below lists the subjects in the
`NlNOS trial by the concentration of l-NO actually received.
`
`Trial
`
`Table 601.14.] (from table 8.0.3.1) Enumeration of sub'ects from NINOS accordin- to' stud
`Control
`l-NO
`l-NO
`I-NO
`5mm
`80 mm
`100 um
`
`. as received”.
`Combined
`I-NO
`
`
`
`
`
`a. All subjects in the l-NO group in NlNOS were first exposed to 20 ppm. A subset of the subjects who did not respond were then given 1-
`NO, 80 ppm. Small numbers of subjects also received either more, or less, then the intended 20 or 80 ppm (protocol violations).
`b. Does not include the 4 control and l l-NO infants who were randomized but did not receive study gas (see section 6.0.1.12.3a).
`
`.
`Dumtignl Adjgflmgm of ”many
`In the NINOS trial, the median duration of exposure to control gas was 21 hours, compared with 71 hours for the
`I-NO group. This reflects the higher fraction of control
`infants who were discontinued from study gas after failing to
`increase their PaOz.
`
`
`
`
`
`48'
`
`

`

`
`
`6.0.1.14 NINOS Efficacy Summary (cont)
`Sta .
`.
`I 2
`.
`l
`r‘
`
`NDA 20-045 Nitric Oxide
`
`There are several statistical issues which need to be addressed with regards to the NINOS trial. The fast is the
`pivotal efficacy analysis, comparing the incidence of death and/or initiation of ECMO in the control and I-NO groups.
`Based on the sponsor's primary analysis, shown in Table 6.0.1 .12.2d.2,
`there was a highly significant advantage for the
`group who received I-NO, such that the trial was stopped early and a clinical alert issued. This analysis, which was intent-
`to-treat, included Several subjects who were randomized but never received study gas, as well as subjects who were
`randomized to control, but received I-NO. This latter group, listed in section 6.0.1.12.3a above, included a large number
`of subjects who later went on to die and/or receive‘ECMO. When an analysis is performed according to the gas actually
`received, the p value for the difference between the two groups is significantly diminished.
`Second, the use of an arbitrary p value <0.05 as the threshold for significance needs to be re-thought. The trial
`had three interim analyses. Under such circumstances,-the p value for significance at the end of the trial must be adjusted
`downwards to a p value of 0.044.
`,
`Third, there was evidence of a center effect, as detected by variability in the rate of the primary endpoint among
`the centers. Correction of this variability can be performed by analyzing the data using the Cochran-Mantel—Hacnszel test,
`rather that the unadjusted chi-square analysis.
`.
`Finally, one can analyze the data for the primary endpoint fiom the NINOS trial using the Cochran-Mantel-
`I-laenszel test, separating the subjects according to the study gas they actually received, and excluding the subjects who did
`not receive any study gas. The results of such an analysis are presented in the section below for the primary endpoint and
`for the incidence of ECMO.
`
`to 6.0.1.1216. Overall, the two
`The baseline data‘for the NINOS trial are summarized in tables 6.0.1.12.1.l
`groups were well-balanced with regards to their demographics and baseline characteristics.
`
`Dismsitignl 91 Subjects
`A larger percentage of the subjecm screened for the NINOS trial were enrolled when compared with the INO-Ol/ -
`02 trial. From a personal conversation with the principle investigator of the NINOS trail, Dr. Ehrenkranz, a large fraction
`(>50%) of subjects who were evaluated were ultimately randomized. This contrasts with the IND-011 ~02 study, where
`only 12% of the'screened infants Were randomized.
`Table 6.0.1.12.2c shows the treatments received in addition to study gas afier randomization. The two groups
`were well-matched with regard to the other therapies they received in addition to study gas,
`including HFOWHFJV,
`surfactant and alkalinization.
`
`"'
`Ergjggg! Violations é’ Iggviatigns
`The protocol violations and deviations are listed in Table 6.0.1.12.2b.l above. The two most significant
`violations were the infants who were randomized but did not receive study gas, and the infants who received I-NO after
`being randomized to control gas. These infants are discussed above in section 6.0.1.1233. The fact that all of the infants
`who received the wrong study gas were randomized to receive placebo, and instead received I-NO, raises the possibility
`that the treatment was unblinded somehow for these infants. There is no other evidence of unblinding for these subjects,
`who all received ECMO .and/or died. The 8 subjects came from ‘7 different centers, all of whom administered I-NO and
`control gas to other infants without reported protocol violation.
`'
`ed
`er
`'a
`'i
`
`As summarized in table 6.0.-1.12.2c.1,
`concomitant therapies received.
`
`the two treatment groups were well-balanced with regard to the
`
`a
`'
`'
`e
`flies
`u c
`.
`'
`1. Incidence of death and/or initiation of ECMO .
`The table below summarizes the results of the NINOS trial from the primary and secondary endpoints, based on
`either the Intent-to-Treat study population or on population according to the actual gas received. In the latter population,
`those infants who were randomized but did not receive study gas are eliminated from analysis.
`
`
`
`
`
`ITT population
`'Gas received' populationh
`'Gas received' population”
`
`'
`
`Table 6.0.1.142 Incidence of rimarv end-oint (death and/or ECMO) in NINOS trial.
`% of control sub'ects
`% of I-NO sub'ects
`
`77/12] (63.6%)
`52/ll4 (45.6%)
`71/112 (63%)
`56/118 (47.4%)
`71/112 (63')
`56/118 (47.4%)
`
`a. p value calculated using unadjusted chi-Square
`b. Subjects who did not receive any study gas were excluded from the analysis, while the remaining subjects were classified according to
`the actual gas received.
`e. p value calculated using Cochran-Mantel-Haenszel adjusted chi~square test.
`
`
`
`
`

`

`6.0.1.14 NINOS Efficacy Summary (cont)
`
`
`
`2. Incidence of death.
`
`No difference in the rate of death was detected between the two groups. This was true both tor the ITT population
`* -
`as well as the population analyzed according to the gas actually received.
`
`N'DA 20-845 Nitric Oxide
`
`Table 6.0.1.143 Incidence ofdeath in NINOS trial.
`% of control sub'ects
`20/121 (16.5%)
`17/112 (15%)
`
`in population
`'Gas received‘ population”
`
`% of I-NO suh'ects
`16/114 (14%)
`17/119 (14.2%)
`
`0.596
`0.869
`
`
`
`
`
`
`
`
`
`a p value calculated using unadjusted chi-square
`b. Subjects who did not receive an).r study gas were excluded from the analysis, while the remaining subjects were classified according to
`the actual gas received.
`.
`-
`
`3. Initiation of ECMO
`
`_ Significantly more subjects in the control group received ECMO using the FIT population. Ifthe populations
`were corrected to reflect the actual gas received, however, the p value for the difierence became less significant. Correcting
`for center effect, using the rare-specified Cochran-Mantel-Haenszel adjusted chi-square test, the reduction in ECMO is not
`significant. Given the three interim looks used in the trial, Dr. Nuri recommends using 0.044 as the cut-oil" for nominal
`significance.
`-
`
`Table 6.0.1.14.4 Incidence of ECMO in NINOS trial.
`Initiation of ECMO
`0/. or control sub'ects
`% of l-NO sub'eets
`ITT population
`66/121 (54.5)
`44/114 (38.5%)
`
`
`'Gas received' populationb
`62/112 (55%)
`48/118 (41%)
`'Gas received' population”
`62/112 (55%)
`48/118 (41%)
`
`
`
`
`a. p value calculated using unadjusted chi-square.
`b. Subjects who did not receive any study gas were excluded from the analysis, while the remaining subjects were classified according to
`the actual gas received.
`c. p value calculated using Cochran-ManteI-Haensael adjusted chi-square test.
`
`4. Meeting criteria for ECMO
`Importantly, there was no significant difference in the-number of subjects who met the criteria for ECMO between
`the two groups.
`‘
`'
`
`‘
`
`Table 6.0.1.145 Incidence of meeting criteria for ECMO in NINOS trial.
`
`Met criteria for ECMO
`% of control sub'ects
`% of l-NO sub'ects m_
`ITT population
`83/121 (69%)
`67/114 (59%)
`0.12
`'Gas received' population”
`76/111 (68%)
`72/119 (60.5%)
`0.208
`
`a, p value calculated using unadjusted chi-square
`6. Subjects who did not receive any study gas were excluded from the analysis, while the remaining subjects were classified according to
`the actual gas received.
`
`. The reasons for subjects meeting the HOMO criteria but not receiving it are listed in table 6.0.1.12.2d.5. The
`most common reason was 'Improved', which occurred more frequently in the infants receiving I-NO. Unfortunately, no
`further details are-available for the crucial time period between when an infant was evaluated and 'met'
`the criteria fir
`ECMO, and when the decision wasmade either to go to ECMO or not. In discussions with Dr. Ehrenkranz, the NINOS
`principle investigator, there was no set time when the infants were evaluated for ECMO criteria. Thus, some infants were
`evaluated pjjg: to initiating study gas, while other infants were evaluated after significant time on study gas had elapsed.
`The time the evaluation took place was also not recorded. The effect of this missing data is to make interpretation of this
`discrepancy between meeting criteria and receiving ECMO impossible to resolve.
`
`5. Survivor endpoints.
`There was no significant difference between the two groups in either the length of hospital stay or assisted
`ventilation (see Table 6.0.1.12.2d.2). The length of hospital stay was numerically longer in the I-NO group (36.41945 in
`the I-NO group versus 29.5:23 in the control group).
`
`50
`
`
`
`

`

`
`6.0.1.14 NINOS Efficacy Summary (cont)
`
`
`..
`a .r =
`‘
`o
`' u a
`l I.
`
`
`6. Acute changes in oxygenation parameters.
`The next table summarizes the oxygenation endpoints specified in the NINOS trial. The results clearly show that
`l-NO has a significant efi'ect to improve oxygenation acutely.
`
`NDA 20-345 Nitric Oxide
`
`
`
`Table 6.0.].146 Oxygenation end-oints from NINOS trial'
`
`Acute Changes in Oxygenation .
`Control
`n=121
`
`
`Change in P30; (mmHg)
`
`Change in 01
`
`
`Change in A-a DO; (mmHg)
`
`
`
`3. Taken from the ITT population. measured as change from baseline to 30 minutes after start of study gas.
`
`p value
`
`
`
`
`
`
`I-NO Therapy
`n=114
`
`58.2:852
`-l4.li2l.0
`-60.0d:85.l
`
`Not all subjects, however, responded to I-NO. A larger percentage of the control subjects who had an acute
`improvement ultimately went on to ECMO and/5r death (53 vs 24 %), although the numbers of full responders in the
`control group were small. Subjects who had no acute reSponse to study gas in both groups were at higher risk for the
`primary endpoint (death and/or ECMO).
`
`Table 6.0.1.14] Sub'ects who met arimarv end-oint -rou-ed b res-onse to low -flow study gas'.
`
`
`
`
`
`NINOS low-flow
`9/17 (53%)
`14/57 (25%)
`
`
`
`full-responders
`
`NINOS low-flow
`36/55 (65%)
`
`
`67/100 (67%)
`
`
`
`
`a. Data from NDA volume 2.14, page 029508. Data shown only for evaluable subjects.
`b. NlNOS defines a full response to study gas as an increase in PaOz >20 mml-Ig, and a partial reSponse as an increase of 10-20 mml—lg.
`This row includes those subjects who did not have an increase in Pa01>20 mmHg (full response).
`
`partial or non-responders”
`
`borour
`-
`v
`elN '
`e
`The sponsor performed several post-hoe analyses, looking at the efi‘ect of LNG in several subgroups of the NINOS
`population. The endpoint used was the number of subjects who met the primary endpoint of the NINOS trial. These
`results are summarized in table 6.0.1.12.3b.1. The significant results of the analysis include:
`1) the use of l-NO in subjects with idiopathic pulmonary hypertension and pulmonary hypertension associated
`with pneumonia/sepsis is associated with a greater decrease in the rate of death and/or ECMO than in subjects with either
`meconium aspiration or respiratory distress syndrome (RDS). In the graph below, the % of the control and l-NO subjects
`who met the primary endpoint is grouped according to the underlying disease reSponsible for the pulmonary hypertension.
`
`APPEARS THIS WAY
`0N ORlGINAL
`
`
`
`5]
`
`

`

`6.0.1.14 NINOS Efficacy Summary (cont)
`'
`'
`a
`a
`e
`
`c
`
`NDA 20-345 Nitric Oxide
`
`Figure 6.0.1.14.8 Primary Endpoint According to
`76
`Underlying Disease
` 67
`
`80 "g
`
`701'
`
`so —'
`so
`
`40
`
`
`
`a Control
`a I-NO
`
`30 J1
`10 ll
`
`
`20 —f
`
`0 ;
`
`%ofSubjectsWhoMet
`
`
`
`PrimaryEndpomt
`
`PPHN
`
`Sepsis
`
`MAS
`
`RDS
`
`Underlying Disease
`
`2) the response to l-NO may also depend on the severity of the overall clinical condition of the infant and/or
`duration of disease when I-NO was initiated. In the graph below, the rate at which the subjects met the primary endpoint is
`grouped according to thebaseline 01 of the subjects. Subjects with the lowest OI who received I-NO had the lowest rate cf
`meeting the primary endpoint.
`
`Figure 6.0.1.143 Primary Endpoint
`by Baseline OI
`
`m D
`
`D Control
`
`I l-NO
`
`U103CO
`
`(A) O
`
` NI 0
`EvaluablePts.(‘70) M48
`
`
`
`.4
`
`00
`
`0125—30
`
`01 30-40
`
`'01 >40
`
`52
`
`
`
`

`

`
`
`6.0.1.14 NINOS Efficacy Summary (cont)
` ; -V I ' "2| 1|! '|I!-_.l .'
`
`
`
`
`
`
`(f
`abSence
`are grouped according to the presence or
`the' subjects
`in the second graph below,
`3)
`echocardiographicaliy demonstrated PPHN. There was a 27% reduction in the risk for meeting the primary endpoint when
`pulmonary HTN was present (65 to 47%), compared with 22% in the group without baseline evidence of pulmonary
`hypertension (50 to 39%).
`
`NDA 20-845 Nitric Oxide
`
`Figure 6.0.1.14;10 Primary Endpoint
`by Baseline ECHO'
`
`70
`
`60
`
`I
`
`%ofEvaluablePts
`
`55
`
`.
`
`-
`-
`DPulmonary HTN Present
`
`-
`
`Pulmonary HTN Absent
`
`Control
`
`I—NO
`
`the primary endpoint,
`4) the use of surfactant appears to have lowered the percentage of subjects who met
`independent of an efiect of I-NO. 1n the presence of surfactant-,- 47/87 control subjects met the primary endpoint (54%)
`compared with 30134 subjects who did not receive surfactant (88%).
`5) No differences in the response rate to I-NO related to either sex or race Were detected (see table 6.0.1.123b. l).
`
`-
`
`APPEARS nus war
`0N URlGlNAL
`
`
`
`53‘
`
`

`

`
`
`NDA 20-345 Nita-i: Oxide
`
`'
`
`6.0.1.14 NINOS Efficacy Summary (cont)
`v
`arv
`
`Several patient benefit parameters were prospectively defined and followed. There were no statistically significant
`differences between the two groups with regards to any of the following endpoints: incidence of den-it, meeting criteria for
`need for ECMO; length of hospital stay; duration of assisted ventilation; duration of continuous positive airway pressure
`(CPAP); duration of either low or high-dose 01 therapy; the incidence of Air Leak Syndrome; the incidence of chronic lung
`disease; or the number of infants discharged home on 02 therapy.
`The NlNOS trial met its primary endpoint with a p value of 0.022 according to the analysis favored by this
`reviewer, which corrected for the protocol violations and for the center effect. The group that received l-NO also had a IOWer
`incidence of initiation of ECMO, although this decrease was also not of overwhelming significance (p=0.026 using the
`unadjusted chi square test, p=0.067 using Cocluan-Mantel-Haenszel adjusted chi-square test). The use of l-NO in subjects
`with idiopathic pulmonary hypertension and pulmonary hypertension associated with pneumoniafsepsis is associated with
`a greater decrease in the rate of death and/or ECMO than in subjects with either meconium aspiration or: respiratory distress
`syndrome (RDS).
`'
`There was no significant difference between the control and I-NO groups in terms of meeting the pre-specifred
`criteria for needing ECMO. Given the available information, the most likely reason infants who met the criteria for ECMO
`improved is an improvement in 0!, secondary to‘the effect of l-NO on oxygenation. In fact, failure to improve from the .
`average baseline 01, in either group, met criteria for transfer to ECMO (see section 6.0.1.9). The concern is that if the
`primary reason the subjects did not receive ECMO was that they 'improved‘ clinically, and 'improvement' was a reflection
`of the decreased 01 caused by I-NO, then the reduced need for ECMO was nothing more than a reflection of improved
`oxygenation. It could be argued that by delaying the transfer of a given subject for ECMO, by improving oxygenation and
`reassuring the clinician ('piuk‘ babies are better than 'blue‘ babies), I-NO allows other, effective therapies time to work. In
`this scenario, I-NO has no clinically beneficial effect, beyond reassurance of the clinician.
`-
`Another piece of evidence suggesting that the acute efi'ect of I-NO on oxygenation was responsible for the
`decreased rate of use of ECMO is seen in table 6.0.1.12.3d, where the rate of ECMO and/or death is grouped according to
`the initial respOnse to study gas. Infants in both control and l-NO groups who did not have an improvement in their
`oxygenation after 30 minutes ultimately went on to receive ECMO andr‘or die at almost exactly the same rate.
`Finally, if there is another beneficial efl‘ect of I-NO beyond oxygenation, an improvement in other markers of
`clinical outcome (i.e., duration of hospitalization or ventilation) might be expected to be detected (none were).
`
`_
`
`1. l-NO acutely improves oxygenation in a substantial fraction of neonates exposed. In the NINOS study, 50% of
`I-NO subjects and 17% of control subjects had an increase of >20 mm Hg in their PaO; after 30 minutes.
`2.
`l-NO use is associated with a significant decrease in the incidence of both death before 120 days and/or the
`initiation of ECMO. In the NINOS trial, there was a reductio'rt‘in 29% in the rate of initiation of ECMO in the I-NO group
`for the intent to treat population, and 25% in the 'gas received' population.
`3. l-NO appears to have a more substantial impact on subjects with idiopathic PPHN and pneumonia/sepsis with
`regards to the rates of death and/or ECMO.
`4. l-NO appears to have a similar impact on subjects with and without echocardiographically-proven PPH‘N with
`regards to the rates of death and/or ECMO.
`5. No effect of I-NO on several efficacy parameters was detected: incidence of death, meeting criteria for need fir
`ECMO; length of hospital stay; duration of assisted ventilation; duration of continuous positive airway pressure (CPAP);
`or the duration of either low or high-dose O; therapy.
`
`6.0.1.1‘5 NINOS Safety Summary
`'
`a
`a
`c
`a
`d
`
`'
`
`' v v'
`
`af
`
`In addition to the incidence of death and ECMO, which were included in the efficacy summary above, several (f
`the secondary endpoints measured safety endpoints, including:
`1. Methemoglobin levels.
`2. Inhaled NO; concentrations.
`3. Incidence of chronic air leak (pneumothorax, pneumopericardium, pneumoperitoneum, pneumomediastinum,
`interstitial emphysema).
`4. Incidence of bronchopulmonary dysplasia (defined as 02 >21% required at 28 days of age with abnormal chest
`x-ray) or reactive airway disease (defined as requiring bronchodilators at discharge).
`5.
`Incidence of intracranial abnormalities and risk factors for abnormal neurological
`intraventricular hemorrhage and brain infarct).
`6. Incidence of pulmonary or gastrointestinal hemorrhage.
`7. The average length of hOSpitalization among surviving infants.
`8. The number of days of assisted ventilation.
`9. The proportion of infants transferred for potential ECMO.
`10. Neurodevelopmental outcomes assessed at 18-24 months corrected age.
`
`sequelae (seizures,
`
`54
`
`
`
`

`

`NBA 2044!- Nlu'ic Oxide
`
`6.0.1.15 NINOS Safety Summary (com)
`I
`V"
`.
`v‘
`.
`
`- .
`
`1. Wrong gas given
`The most critical protocol violations involved the infants who were randomized to receive control gas, and
`instead received I-NO. These seven infants, along with the one infant who received both l-NO and control gas, were
`discussed in section 6.0.].12.3a,
`including an analysis of the data corrected for the gas actually received by each of the
`infants.
`'
`
`2. Overdose
`
`a. Two infants received >80 ppm I-NO.
`Two subjects in the NINOS trial were given -100 ppm I-NO inadvertently. One infant developed an elevated N02
`level, ultimately received ECMO and was discharged home requiring supplemental 02.
`'
`_
`1. Subject #54-A02: this 3.7 kg female received I-NO for PPHN with meconium aspiration syndrome, with a
`baseline 01 of 23. After responding to 20 ppm l-NO, she was continued on study gas for 159 hours (day 5), at which time
`her dose was inadvertently increased to 100 ppm for approximately 36 minutes. While no coincident methemoglobin and
`NOZ'levels were obtained, her higher recorded levels were 2.2% methemoglobin on day 3 and 0.4 ppm NO; on day 1. She
`was weaned with difficulty, did not receive ECMO, and was discharged, with no' chronic lung disease or excess bleeding
`or other major organ dysfunctions.
`2. Subject #3-A02: this 2.9 kg female received I-NO for PPHN with pneumoniafsepsis, with a baseline 01 of 56
`and 26. She had a partial response to l-NO 20 ppm, and no response to l-NO 80 ppm. An improper flow-meter setting led
`to her exposure to I-NO lOlppm for approximately 1 hour. Her methemoglobin level at that time was 6% and her N02
`level was 5.1 ppm. Study gas was weaned down to 20 ppm and the N02 fell to 3.4 ppm. Afier 14 hours more, she
`a
`received ECMO for persistently elevated A-aD02. She survived, but was discharged to home on 02.
`b. 0;... infant received I-NO for more than 240 hours.
`3. Subject #SI-A 12: this female received l-NO for PPHN with meconium aspiration. She had a partial response to
`I-NO 20 ppm. At 240 hours, the infant was on 20 ppm, and was ultimately weaned afier a total of 253 hours and 25
`minutes. Her maximum NO; and methemoglobin levels were 0.] and 3.6 respectively. The subject received high-
`fi'equency ventilation while receiving study gas, and developed both pneumothorax and periventricular leukomalacia.
`Chronic lung disease was also diagnosed, after the infant required ventilation for 16 days.
`
`_
`3. Unblinded subjects
`There was no significant difference in the reported rate of unblinding between the two groups.
`
`SAW
`1. Methem oglobin levels.
`There was a clear association between the dose of LNG and the risk of elevated methemoglobin levels, as was
`seen in table 6.0.].13.2b.l.
`'
`‘
`1n the NINOS trial, the mean peak methemoglobin level was significantly higher in the l-NO group. During the
`first 12 hours of exposure to 80 ppm l-NO, the peak in the control group was 1.21-0.13, versus 2.0il.5 in the l-NO group
`(p<0.001).
`1n the NINOS trial, methemoglobin was defined as >5%, and a total of 11 subjects (4 controls, '7 I-NO) had their
`study gas decreased because their methemoglobin levels exceeded this level. All of these subjects continued on study gas
`at lower flow rate. No subject was discontinued because ofmethemoglobin >10%. No short-term adverse outcomes were
`identified that can be linked to the elevated methemoglobin levels.
`
`2. Inhaled NO; concentrations.
`In the NINOS trial, the average peak N02 level was significantly higher in the l-NO group than in the control
`group (see table 6.0.1.13.2a.l).
`There was a clear association between the dose ofI-NO and the risk of elevated methemoglobin levels. The mean
`peak N02 level occurring in the first 12 hours was 0.0i0.3 in the control group and 0.8112 in the I-NO group, excluding
`the infants who received the wrong study gas.
`In the NTNOS trial, one subject receiving I-NO 80 ppm had a N0; level >7.0 % during the trial. Subject #55-08,
`a Caucasian male,-had a peak level of 9.1, and the subject underwent a successful wean of study gas, without withdrawal
`from the trial. He received ECMO and was discharged home with chronic lung disease.
`
`
`
`‘55
`
`

`

`NDA 20445 Nitric Oxide
`
`6.0.1.15 NINOS Safety Summary (cont)
`_
`~.
`3. Incidence of chronic air leak syndrome.
`Table 6.0.1.1215 shows that a smaller % of infants in the control group had 11.2.2 prior to study gas
`-.
`administration than in the I-NO group (21 vs. 18%). However, more subjects developed ALS during study gas
`administration in the I-NO group (1 1% vs. 6% in the control group, see Table 6.0.1.1312). A higher % of I-NO infants
`- also developed ALS, looking just at those infants who