Safety and Efficacy of Inhaled Nitric Oxide Treatment for Premature Infants with Respiratory Distress Syndrome: Follow-Up Evaluation at Early School-Age

Abstract

Aims

To assess long-term safety and compare neurodevelopmental outcomes in school age children born prematurely who received inhaled nitric oxide or placebo during the first week of life in a randomized, double blind study. Children treated with inhaled nitric oxide had previously been shown to have decreased intraventricular hemorrhage and periventricular leukomalacia as newborns, and decreased cognitive impairment at 2 years^1–2.

Methods

Follow-up study of medical outcomes, neurodevelopmental assessment and school readiness in 135 of 167 (81%) surviving premature infants seen at 5.7 ± 1.0 years.

Results

Compared to placebo (n=65), iNO-treated children (n=70) demonstrated no difference in growth parameters, school readiness, or need for subsequent hospitalization. However, iNO-treated children were less likely to have multiple chronic morbidities or technology-dependence (p=0.05). iNO-treated children also had less functional disability (p=0.05).

Conclusions

These results demonstrate the long-term safety of iNO in premature infants. Furthermore, iNO treatment may improve health status by decreasing the incidence of severe ongoing morbidities and technology-dependence, and may also decrease the incidence of educational and community functional disability of premature infants at early school age.

Trial Registration

clinicaltrials.gov Identifier: NCT00152542

Introduction

With advances in neonatal intensive care medicine, critically ill premature infants born with respiratory distress syndrome (RDS) increasingly survive to school-age and beyond. With this improved survival, there has been an increase in neurodevelopmental impairment and ongoing morbidities.¹

In a randomized, double blind, placebo-controlled trial, we previously reported that inhaled nitric oxide (iNO) treatment of premature infants having moderate RDS decreased the incidence of death and chronic lung disease (CLD), as well as decreasing severe (Grades III–IV) intraventricular hemorrhage and periventricular leukomalacia (IVH/PVL).² At two years corrected age, these iNO-treated infants had improved neurodevelopmental outcomes, primarily due to improved cognitive ability³. After adjusting for neonatal morbidities known to affect neurodevelopmental outcome such as CLD and severe IVH/PVL, improved neurodevelopment in the iNO-treated group persisted. Thus, iNO appears to be neuro-protective in preterm infants with RDS.

We evaluated the long-term safety at school-age in children who have received iNO as premature infants. We also assessed whether the beneficial effects of iNO therapy persisted at school age, using our previously reported multi-dimensional measure of school readiness.

Patients and Methods

Initial Study

The initial study was a randomized, placebo-controlled, single center study of 207 premature infants with RDS requiring mechanical ventilation and surfactant therapy. Treatment with iNO was initiated at a dose of 10 ppm on the first day, followed by 5 ppm for the next six days or until extubation.

Present Study

The study population consisted of children from the original cohort who survived to early school-age. The present protocol was approved by the institutional review board of the University of Chicago and informed parental/legal guardian consent was obtained.

Clinical data were obtained from the medical record. Children were examined in a single visit at early school-age by a general pediatrician and/or a developmental and behavioral specialist unaware of the child’s perinatal course and treatment assignment evaluated. Z scores for height and weight were calculated using the Centers for Disease Control and Prevention (CDC) growth charts ⁴. Z scores for head circumference were calculated using criteria by Roche⁵.

Children with abnormalities of posture or tone that impacted on motor control were classified as having one of the cerebral palsy (CP) syndromes. The Gross Motor Function Classification System (GMFCS) ⁶ was used to classify motor function.

Caretakers completed questionnaires regarding their child’s race, need for rehabilitation/special education services, ongoing morbidities, number of hospitalizations, and socioeconomic status. Socioeconomic status was assigned using Hollingshead Index of Social Position ⁷. Independent functioning of the child was determined by interview of the caretaker utilizing the Pediatric Functional Independence Measure (also known as the WeeFIM), ⁸. The NICHQ Vanderbilt Parent Assessment Scale ⁹ was used as a screen for the presence of pediatric behavioral disorders. Children underwent the Bracken School Readiness Assessment¹⁰, the Peabody Picture Vocabulary Test 3^rd Ed., ¹¹ and the Beery Test of Visual-Motor Integration ¹².

Children with hearing aids were classified as having hearing loss. Children with hearing aids who were unable to communicate were classified deaf. Visual acuity was assessed using Lea Symbols¹⁴. Children with corrected visual acuity between 20/60–20/200 were classified visually impaired, and children with acuity worse than 20/200 were classified blind.

Ongoing morbidities were classified as none/mild (e.g. eczema, obesity, seasonal allergies, or mild asthma not requiring controller medication), chronic (e.g. asthma requiring controller medications, failure to thrive requiring nutritional supplements, and epilepsy), or multiple chronic/technology dependence (e.g. ventriculoperitoneal shunt, gastrostomy tube, or tracheostomy).

Based on standardized scores, children were assigned one of four school-readiness levels¹⁵. The criterion for school readiness was a level of greater than 2. Children who could not be evaluated in person were assessed by phone questionnaire of the caretaker.

Statistical Analyses

Continuous and normally distributed baseline characteristics such as birth weight and gestational age were compared between the iNO and the placebo group using the two sample t test. Ordinal baseline characteristics such as Apgar score and skewed distributed variables such as oxygenation index were compared between the two groups using the Wilcoxon Rank-Sum test. Categorical baseline variables were compared using Fisher’s exact test. Similarly, we compared the outcome variables at age school-age between the two groups using Wilcoxon Rank-Sum tests for ordinal outcomes and Fisher’s exact tests for categorical outcomes.

Results

Early school age (mean age 5.7 ± 1.0 year) outcome data were obtained for 135 of the 167 (81%) surviving children from the original cohort (Fig. 1). The 32 children lost to follow-up had similar birth weights (1136 ± 380 grams, p=0.12) and gestational ages (27.9 ± 2.7 weeks, p=0.44) as those assessed. As previously reported, phone interviews allowed unequivocal assessment of developmental status of the five children who could not be assessed in person¹⁶.

Figure 1.

Flow diagram showing the number of patients involved at various stages of the study.

Of children assessed, 70 (51.8%) had received iNO. Neither the mean age at follow-up nor the rates of follow-up significantly differed between children who received iNO (79.6%) and those who received placebo (82.3%). The clinical characteristics of the children and maternal demographics at the time of enrollment in the initial study did not differ significantly between the two follow-up groups (Table 1). The Hollingshead Indices of Social Position were not significantly different between the two groups (p=0.89). Consistent with our original report, the group receiving iNO had a decreased incidence of severe IVH/PVL (p=0.04).

Table 1.

Clinical Characteristics and Demographics of the Follow-up Cohort

	Placebo (n=65)	iNO (n=70)	P
Patient characteristic
Birth weight, g	1002±403	1028±381	0.70
Gestational age, weeks	27.4 ± 2.7	27.6±2.6	0.71
Age at follow-up, years	5.6±1.1	5.7±0.9	0.43
Male sex, n (%)	32 (49)	35 (50)	1.0
Initial oxygenation index
Median	7.1	6.1	0.24
Interquartile range	4.5–15.0	3.7–12.3
1-minute Apgar score
Median	5	5	0.53
Interquartile range	4–6	3.5–6
5-minute Apgar score
Median	7	7	0.56
Interquartile range	6–8	7–8
Antenatal corticosteroids, n (%)	37/65 (57)	41/68 (60)	0.73
Postnatal corticosteroids>7days, n (%)	7/64 (11)	8/69 (11)	1.0
Surfactant, no. of doses	2.3±0.9	2.2±0.9	0.44
Neonatal Morbidities, n (%)
Chronic lung disease	35 (54)	28 (40)	0.12
Severe intraventricular hemorrhage or periventricular leukomalacia	16 (25)	7 (10)	0.04
Hollingshead Index of Social Position, n (%)
Level 1–2	7 (11)	9 (13)	0.89
Level 3	22 (34)	19 (27)
Level 4	16 (24)	21 (30)
Level 5	20 (31)	21 (30)
Maternal Race (self-reported)
Black	47 (72)	48 (69)	0.66
White	9 (14)	14 (20)
Other	9 (14)	8 (11)

^*Plus-minus values are means ± standard deviation

Long-term Safety of iNO

To understand whether iNO treatment during the initial hospitalization for prematurity increases the risk of adverse outcomes during the first five years of life, we assessed children along three axes: somatic growth, need for re-hospitalization and neurodevelopment, as reflected by school readiness scores.

Although z scores for median weight (−0.45) and head circumference (−1.47) in all children were significantly lower than term infants, scores did not differ between study groups. These data indicate that iNO-treated premature infants (median weight 19.3 kg) did not have increased rates of growth failure compared with placebo-treated premature infants (median weight 18.2 kg).

Fewer than half of all infants, regardless of treatment status, were hospitalized during the 5–6 years following initial discharge. Of those children, half were hospitalized once. The distribution of the number of hospitalizations in the iNO-treated group was not different from that of the placebo-treated group (p=0.94).

To assess whether iNO treatment adversely impacted neurodevelopmental outcomes, we asked whether the number of children developmentally ready for school differed between treatment groups. To assess school readiness, we employed our multidimensional assessment of neurodevelopment¹⁵. As previously reported, two-thirds of children were scored at Levels 3 and 4¹⁵. Children treated with iNO had similar distributions of school-readiness to the children who received placebo (p=0.87). In addition, there were no significant differences between groups in developmental assessment components or the incidences of cerebral palsy, hearing impairment, or visual impairment (Table 2). Taken together, these data indicate that iNO treatment during prematurity adversely affects neither physical development nor neurodevelopment at early school age, supporting its safety for use in premature infants.

Table 2.

School-Readiness Levels of the Follow-Up Cohort at Early School-Age

	Placebo N=65	iNO n=70	P
Bracken*
>85	32	36	0.60
70–84	14	15
<70	18	16
PPVT-III
>85	26	34	0.47
70–84	23	16
<70	14	16
VMI
>85	37	44	0.33
70–84	15	12
<70	13	11
WeeFIM
>85	49	56	0.36
70–84	8	5
<70	8	7
GMFCS
No Cerebral Palsy	59	63	0.77
Level 1–2	2	1
Level 3	1	0
Level 4	1	2
Level 5	0	2
Hearing
No Problems	60	67	0.44
Requires Hearing Aids (can hear and speak)	5	1
Requires Hearing Aids (no language)	0	2
Vision
No Problems	55	54	0.35
Glasses or strabismus	4	11
Visually impaired (can see objects)	4	4
Blindness (no functional vision)	2	1

^*Bracken: Bracken School Readiness Assessment; PPVT-III: Peabody Picture Vocabulary Test 3^rd Ed; VMI: Beery Test of Visual-Motor Integration; GMFCS: Gross Motor Function Classification System

Efficacy of iNO

To assess whether beneficial effects of iNO treatment during prematurity persisted to early school age, we first compared the incidences of chronic morbidities between groups. There were no significant differences between groups in the percentage of children having one or no chronic morbidities. Importantly, however, significantly fewer iNO-treated infants had multiple chronic morbidities or were dependent on technology compared with placebo-treated infants (Fig. 2). Because the presence of multiple chronic morbidities or dependence on technology strongly reduces the likelihood of school readiness¹⁵, these data suggest that, by decreasing the total morbidity burden, iNO may decrease the number of premature children requiring additional societal resources.

Figure 2. Ongoing morbidities of study patients at school-age.

*P=0.05.

To assess whether iNO-treatment improved more subtle neuro-developmental outcomes at school age, we compared performances on the WeeFim, a component of our multi-dimensional neurodevelopmental battery. Scores on the WeeFim of less than 80% signify significant educational and community disability. Because children with school readiness level 1 were so profoundly disabled, we restricted our analysis to children having school readiness levels of 2–4. Five placebo-treated children had WeeFim scores of less than 80%. In contrast, no iNO-treated child fell into this group (p=0.05). These data suggest that some improved neurodevelopment in iNO-treated infants persists at early school age.

Discussion

In this longitudinal follow-up study of iNO- or placebo-treated premature infants, we found that iNO neither impairs somatic growth, increases hospitalizations nor worsens neurodevelopmental outcomes at school age. These data indicate that iNO treatment in premature infants is safe. Furthermore, fewer iNO-treated infants had multiple chronic morbidities. More iNO-treated infants also had improved functional outcomes compared with placebo-treated infants.

Several randomized placebo-controlled studies in addition to ours have reported improved short-term pulmonary and/or neurologic outcomes following iNO therapy ^{2, 17–19}, although a recent multi-center trial of iNO in premature infants having extremely mild respiratory disease failed to find improved short-term outcomes²⁰. Follow-up studies of these infants have yielded conflicting results: in our study population, neurodevelopmental outcomes were improved at two years corrected age³. In contrast, despite decreased CLD following late-onset iNO treatment in the NO-CLD study²¹, investigators failed to find differences in neurodevelopment in iNO-treated infants at two years of age²². Nonetheless, in the Kinsella trial²³, investigators found that, at one year corrected age, iNO-treated infants weighing 750–999g had a lower rate of death or neurodevelopmental impairment compared with placebo-treated infants²⁴. This improved outcome occurred in the group in which decreased IVH/PVL/ventriculomegaly had been previously reported²³.

Because previous therapies in premature infants have provided short-term benefits at the expense of neurodevelopmental impairment, (most notably dexamethasone therapy²⁵), assessment of the five year safety of iNO treatment is essential. While the sub-population of premature infants who benefit from iNO therapy has yet to be precisely determined, our findings that iNO-treated premature infants do not differ from placebo-treated infants in somatic growth, hospitalization rates or overall neurodevelopmental outcomes provide reassurance of the safety of iNO treatment in premature infants with moderate RDS. These findings should help facilitate future studies intended to better elucidate optimal iNO treatment regimens in this population.

At two years of age, iNO-treated premature infants in our study had approximately half the risk of cognitive impairment exhibited by the placebo-treated group³. That this marked difference in cognitive impairment was not apparent at school age indicates that factors outside of the neonatal period, such as the home environment and socioeconomic status, become increasingly important as infants acquire more complex repertoires of skills and behaviors. For example, low socioeconomic status has roughly four times the negative impact on school readiness compared with that of severe IVH/PVL or CLD¹⁵. It is also possible that the battery of neurodevelopmental assessments we used may be insufficiently sensitive to detect cognitive differences between iNO- and placebo-treated children. Studies of school performance at older ages may be more useful to detect subtle differences in neurodevelopment.

The iNO-induced decreases in neurodevelopmental impairment we observed at two years corrected age were statistically independent of the decreased incidences of both severe IVH/PVL or CLD³. This observation raises the possibility that iNO acts directly on the developing brain, possibly through delivery of NO to brain vasculature (reviewed by Shechter and Gladwin²⁸). Providing iNO to cerebrovascular beds supplying vulnerable pre-oligodendrocytes and neurons, therefore, could result in both improved blood flow during hypoxic episodes common during the initial hospitalization, perhaps overcoming hypoxia-induced eNOS down-regulation²⁹. In our study, infants received iNO for one week, irrespective of gestational age and, hence, the stage of brain development. It is possible, therefore, that longer iNO treatment provided throughout critical periods of brain development may further improve neurodevelopmental outcomes in this at-risk population.