Response to Inhaled Nitric Oxide and Mortality Among Very Preterm Neonates With Pulmonary Hypertension

Key Points

Question

Is there an association between inhaled nitric oxide (iNO) responsiveness and mortality among very preterm neonates with early (age, ≤72 hours) or late (age, >72 hours) acute pulmonary hypertension?

Findings

In this cohort study of 362 very preterm neonates who received iNO for 371 acute pulmonary hypertension episodes across 12 Canadian centers, neonates with early acute pulmonary hypertension demonstrated a higher iNO response rate (71% vs 49%) and lower mortality (34% vs 49%) than those with late acute pulmonary hypertension. Accounting for pretreatment illness severity, iNO responsiveness and echocardiography use were associated with greater survival in the early acute pulmonary hypertension cohort.

Meaning

This study suggests that iNO responsiveness was associated with improved survival among very preterm neonates presenting with acute pulmonary hypertension during the early postnatal transition period; its role in late acute pulmonary hypertension remains unclear.

Abstract

Importance

Clinical observations of immediate improvement in fraction of inspired oxygen (FiO₂) in a proportion of cases is often cited as the rationale for using inhaled nitric oxide (iNO) in the management of acute pulmonary hypertension among very preterm neonates (gestational age, <32 weeks). However, the clinical effectiveness of such a response pattern remains underinvestigated.

Objective

To identify factors associated with predischarge mortality among very preterm neonates receiving iNO for acute pulmonary hypertension, with specific a priori emphasis on iNO responsiveness.

Design, Setting, and Participants

This prospective observational cohort study was conducted from January 1, 2018, to December 31, 2022, at 12 Canadian tertiary neonatal intensive care units. Consecutive very preterm neonates who received iNO for a diagnosis of acute pulmonary hypertension and pretreatment FiO₂ of 0.50 or more were included. Neonates with congenital anomalies or those who were treated for chronic pulmonary hypertension with iNO were excluded. Early acute pulmonary hypertension (≤72 hours of age) and late acute pulmonary hypertension (>72 hours of age) cohorts were analyzed separately. Statistical analysis was performed from January 2023 to January 2024.

Exposure

Treatment with iNO for acute pulmonary hypertension.

Main Outcomes and Measures

The study cohorts were divided for comparison based on the primary outcome of predischarge mortality. Logistic regression analyses were used with predefined variables, including iNO responsiveness, to identify factors associated with mortality. A positive response to iNO was defined as a pre-iNO minus 4-hour post-iNO FiO₂ of 0.20 or more.

Results

The early acute pulmonary hypertension group (mean [SD] birth gestational age, 26.3 [2.4] weeks; median treatment age, 1 day [IQR, 1-2 days]; 147 boys [56%]) included 262 neonates; 179 (68%) had a pre-iNO FiO₂ of 1.0. The late acute pulmonary hypertension group (mean [SD] birth gestational age, 24.9 [1.7] weeks; median treatment age, 13 days [IQR, 9-20 days]; 72 boys [66%]) included 109 neonates; 51 (47%) had a pre-iNO FiO₂ of 1.0. Neonates with early acute pulmonary hypertension more frequently had a positive iNO response (71% [186 of 262] vs 41% [45 of 109]) and lower mortality (34% [90 of 262] vs 49% [53 of 109]) than those with late acute pulmonary hypertension. Accounting for pretreatment illness factors, greater reduction in FiO₂ with iNO remained associated with lower mortality for neonates with early acute pulmonary hypertension (adjusted odds ratio per FiO₂ reduction of 0.10, 0.74 [95% CI, 0.65-0.84]). For those with late acute pulmonary hypertension, however, only pretreatment illness severity (lower pre-iNO FiO₂ and higher pre-iNO pH), and not positive response to iNO (adjusted odds ratio, 0.47 [95% CI, 0.17-1.30]), was associated with mortality.

Conclusions and Relevance

In this cohort study of very preterm neonates with acute pulmonary hypertension treated with iNO, responsiveness to iNO was associated with improved outcomes during the first 72 hours of age. The prognostic role of iNO response in acute pulmonary hypertension presenting after 72 hours of age remains unclear. Future studies should investigate the distinct pathophysiological mechanisms associated with late acute pulmonary hypertension in this population.

This cohort study examines factors associated with predischarge mortality among very preterm neonates receiving inhaled nitric oxide for acute pulmonary hypertension, with an emphasis on responsiveness to inhaled nitric oxide.

Introduction

Inhaled nitric oxide (iNO), a selective pulmonary vasodilator with an established safety profile, is a standard treatment for acute pulmonary hypertension in full-term and near-term neonates.^1,2 Its role among preterm neonates is debated due to inconsistent evidence from randomized clinical trials.^3,4 However, in contemporary practice, iNO is predominantly used among preterm neonates as a rescue therapy for acute pulmonary hypertension physiology; this patient population has not been adequately studied in randomized clinical trials. Approximately 5% of neonates with a gestational age (GA) younger than 34 weeks receive iNO in neonatal intensive care units (NICUs), with exposure rates inversely related to GA, and mortality observed in approximately 40% of neonates.^5,6

With challenges in conducting adequately powered randomized clinical trials for this critically ill population, observational studies using prospective data registries have emerged as an alternative. Recent registries have found similar oxygenation improvement patterns among preterm and term-born neonates with iNO after acute pulmonary hypertension–associated hypoxic respiratory failure (HRF) during the first week of age.^7,8 Whether this positive response to iNO is associated with improved clinical outcomes, and its role beyond the early postnatal period, remains underinvestigated. Few single-center retrospective studies report greater survival in association with iNO responsiveness among preterm neonates with acute pulmonary hypertension; however, data were limited by small sample sizes and unaccounted confounders.⁹ Therefore, our objective was to investigate the association between iNO responsiveness and other clinical factors with mortality among preterm neonates receiving iNO for acute pulmonary hypertension–associated HRF. We hypothesized that if iNO use in this context is clinically ineffective, patient outcomes may be associated with baseline pretreatment factors rather than iNO responsiveness.

Methods

Design

A multicenter prospective observational cohort study was conducted between January 1, 2018, and December 31, 2022, at 12 tertiary NICUs across Canada. Two centers contributed data for 4 years, 6 centers contributed for 3 years, 3 centers contributed for 2 years, and 1 center contributed for 1 year. Centers entered data for all eligible patients who received iNO during the study periods. All centers received approval and a waiver of informed consent from local research ethics boards (Mount Sinai Hospital Research Ethics Board; Sunnybrook Health Sciences Centre Research Ethics Office; University of Calgary Conjoint Health Research Ethics Board; Hamilton Integrated Research Ethics Board; University of Manitoba Bannatyne Campus Research Ethics Boards [for 2 sites in Winnipeg: Sainte Boniface Hospital and Health Sciences Centre]; NAGANO Comité d'éthique de la recherche du CHU de Québec-Université Laval [for 3 sites in Quebec: St. Justine, CHU Québec, and McGill]; University of Alberta Health Research Ethics Board [for Royal Alexandra/Stollery Children’s Hospital in Edmonton]; UBC C&W Research Ethics Board [for Vancouver site]; and Western University Health Science Research Ethics Board [for London HSC]) because deidentified data were collected within the existing framework of the Canadian Neonatal Network and enacted a data transfer agreement with the coordinating center at Mount Sinai Hospital, Toronto. This study was conducted and reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for observational research.¹⁰

Participants

The registry collected data for all very preterm neonates (birth GA, <32 weeks) who received iNO in participating NICUs during the study periods. This study included patients with a given diagnosis of acute pulmonary hypertension, as deemed by the attending clinicians, and HRF, defined as a pre-iNO fraction of inspired oxygen (FiO₂) of 0.50 or more. Exclusion criteria were congenital cardiac anomalies (except for patent ductus arteriosus, patent foramen ovale or atrial septal defect, or ventricular septal defect), known genetic anomalies, congenital diaphragmatic hernia, and given diagnosis by clinicians of chronic pulmonary hypertension.

Data Sources

This study expanded on the Canadian Neonatal Network database, which collects daily clinical data on very premature neonates in Canadian tertiary NICUs to improve quality benchmarks.¹¹ Trained personnel at each participating center abstracted maternal and neonatal baseline demographics, perinatal characteristics, neonatal morbidities, and mortality outcomes from health records for entry into the Canadian Neonatal Network database. Additional data were collected for each iNO exposure encounter, including treatment indication (acute pulmonary hypertension, chronic pulmonary hypertension), age at treatment, primary diagnosis, and pre-iNO circulatory and ventilatory status (blood pressure, vasoactive drug use, blood gases, ventilation mode, FiO₂, mean airway pressure [MAP], preductal and postductal peripheral oxygen saturation, and other pulmonary vasodilator use). Furthermore, we measured FiO₂ at the 4-hour mark after iNO initiation (to assess responsiveness) and collected information on other pulmonary vasodilator use and diagnosis confirmation by echocardiography (within 24 hours before or after iNO initiation). If multiple echocardiography results were available, preference was given to the pre-iNO scan or the scan closest to iNO initiation.

Study Definitions

We defined episodes of acute pulmonary hypertension as early acute pulmonary hypertension if iNO was initiated at 72 hours of age or less, and as late acute pulmonary hypertension if iNO was initiated at more than 72 hours of age, to distinguish between episodes treated during postnatal transition or later in the NICU course. These groups were a priori considered likely to be distinct disorders stemming from different underlying pathologic conditions. Neonates with both early and late acute pulmonary hypertension episodes were included in the respective cohorts, while those with multiple late acute pulmonary hypertension episodes (repeated exposure after ≥7 consecutive iNO-free days) had only the initial episode included. We defined echocardiography-confirmed acute pulmonary hypertension as the presence of any of these features, in hierarchical order: (1) right-to-left or bidirectional (right-to-left component >30% of cardiac cycle) patent ductus arteriosus shunt; (2) right ventricular systolic pressure greater than or equal to systolic blood pressure, where right ventricular systolic pressure = 4V² + 5 (where V is the peak velocity of tricuspid regurgitant jet in meters per second squared; 5 mm Hg was presumed right atrial pressure); (3) paradoxical interventricular septal motion; or (4) exclusively right-to-left patent foramen ovale shunt. Episodes were classified as confirmed acute pulmonary hypertension if neonates met these echocardiography criteria or had a preductal and postductal peripheral oxygen saturation gradient of 5% or more. Those not meeting these criteria were classified as having presumed acute pulmonary hypertension. Respiratory severity score (FiO₂ × MAP) was calculated to quantify respiratory illness severity before treatment.^12,13 A positive response was defined as an FiO₂ reduction of 0.20 or more at 4 hours after iNO compared with the pretreatment FiO₂. Conversely, a reduction of less than 0.20, no change, or FiO₂ increase was categorized as nonresponse. If iNO was discontinued before 4 hours, FiO₂ at discontinuation was used.

Outcomes and Study Groups

The primary outcome was predischarge mortality. Secondary outcomes examined in survivors included incidence of chronic lung disease (defined as the need for respiratory support and/or supplemental oxygen at 36 weeks’ corrected GA¹⁴); stage 2a or higher necrotizing enterocolitis as per the modified criteria of Bell et al¹⁵; grade 3 or higher intraventricular hemorrhage as per the classification of Papile et al¹⁶; and retinopathy of prematurity requiring treatment. We a priori decided to investigate early acute pulmonary hypertension and late acute pulmonary hypertension as separate study groups. Each study group was divided based on the primary outcome for comparison: predischarge mortality vs survival until discharge.

Statistical Analysis

Statistical analysis was performed from January 2023 to January 2024. Data were summarized and presented as numbers and percentages, mean (SD) values, or median (IQR) values. Intergroup comparisons were performed using the χ² test or the Fisher exact test for categorical variables and the t test or the Mann-Whitney test for continuous variables, as appropriate. For each study group, multivariable logistic regression analysis was conducted to identify factors associated with the primary outcome. The analysis incorporated the following predefined variables: birth GA, small for GA size as per Kramer curves less than 10%, male sex, Score for Neonatal Acute Physiology (SNAP II) of 20 or more,¹⁷ cesarean delivery, primary underlying diagnosis, change in FiO₂ from baseline at 4 hours after iNO initiation, echocardiography performed, confirmed pulmonary hypertension (vs presumed), and variables identified through bivariate analysis (P < .10). Kaplan-Meier curves were generated using a predefined definition of iNO responsiveness (change in FiO₂ of 0.20 or more) to illustrate mortality timing between responders and nonresponders in early and late acute pulmonary hypertension groups. Surviving neonates were censored on the day of discharge. Additional planned analyses were undertaken in the cohort where, and if, change in FiO₂ was independently associated with survival. These analyses included (1) area under the receiver operating characteristic curve (AUROC) analysis to assess the discriminatory ability of change in FiO₂ to differentiate survivors from nonsurvivors (if relevant, change in FiO₂ exhibiting the highest combination of sensitivity and specificity using the Youden index was reported); and (2) multivariable logistic regression analysis to identify factors associated with a positive response, defined as change in FiO₂ of 0.20 or more, including the following predefined variables: birth GA, small for GA size, male sex, SNAP II of 20 or more, primary underlying diagnosis, pre-iNO variables (respiratory severity score, partial pressure of carbon dioxide, blood pressure, vasoactive drug use), confirmed pulmonary hypertension, preductal and postductal peripheral oxygen saturation gradient of 5% or more, and variables identified through bivariate analysis (P < .10). Furthermore, prematurity-related morbidities among survivors were described, comparing iNO responders with nonresponders. Finally, to examine the interaction between lung maturity, type of acute pulmonary hypertension, and iNO responsiveness, multivariable logistic regression analysis was conducted in the whole cohort and adjusted odds ratios (AORs) reported for postmenstrual age at treatment and early acute pulmonary hypertension (vs late acute pulmonary hypertension) for the outcome of positive iNO response, adjusting for sex, premature rupture of membranes at greater than 24 hours, confirmed acute pulmonary hypertension, pre-iNO FiO₂, pre-iNO blood pH, and pre-iNO vasoactive drug use. All P values were from 2-sided tests, and results were deemed statistically significant at P < .05. All analyses were performed using SAS, version 9.4 (SAS Institute Inc).

Results

Of the 426 preterm neonates who received iNO during the study period, 64 met the exclusion criteria, leaving 262 neonates in the early acute pulmonary hypertension group (mean [SD] birth gestational age, 26.3 [2.4] weeks; median treatment age, 1 day [IQR, 1-2 days]; 147 boys [56%] and 115 girls [44%]) and 109 in the late acute pulmonary hypertension group (mean [SD] birth gestational age, 24.9 [1.7] weeks; median treatment age, 13 days [IQR, 9-20 days]; 72 boys [66%] and 37 girls [34%]) (Figure 1). Nine neonates were included in both groups; 12 experienced multiple late acute pulmonary hypertension episodes, with only the first included in the analysis.

Figure 1. Patient Inclusion Flowchart.

Neonates had a gestational age of less than 32 weeks at birth. FiO₂ indicates fraction of inspired oxygen; iNO, inhaled nitric oxide; and NICU, neonatal intensive care unit.

^aNine neonates had both early acute pulmonary hypertension and late acute pulmonary hypertension episodes and were included in both cohorts.

Cohort Characteristics

The early acute pulmonary hypertension group had a mean (SD) birth GA of 26.3 (2.4) weeks, distributed as follows: 66 (25%) at less than 25 weeks, 106 (41%) at 25 weeks and 0 days to 27 weeks and 6 days, and 90 (34%) at 28 weeks and 0 days to 31 weeks and 6 days. A total of 147 (56%) were boys. The median age at iNO treatment was 1 day (IQR, 1-2 days), with 234 (90%) receiving high-frequency ventilation and 179 (68%) having a pre-iNO FiO₂ of 1.0.

In contrast, the late acute pulmonary hypertension group was more immature, with a mean (SD) birth GA of 24.9 (1.7) weeks, and 102 (94%) born at less than 28 weeks. A total of 72 (66%) were boys. The median age at iNO treatment initiation was 13 days (IQR, 9-20 days) and 94 treatments (86%) occurred at 7 days of age or older; the number of episodes on day 4 of age was 2, on day 5 of age was 5, and on day 6 of age was 8. At the time of treatment, 89 (82%) were receiving high-frequency ventilation, and 51 (47%) had a pre-iNO FiO₂ of 1.0.

Echocardiography-confirmed suprasystemic acute pulmonary hypertension was present in 189 (72%) of early and 69 (63%) of late acute pulmonary hypertension cases, and iNO was used as the primary pulmonary vasodilator in 254 neonates (97%) in the early acute pulmonary hypertension cohort and 106 neonates (97%) in the late acute pulmonary hypertension cohort. Pre-iNO vasoactive drug use was higher in late acute pulmonary hypertension cohort than the early acute pulmonary hypertension cohort (43 [40%] vs 76 [29%]; P = .05), whereas a positive response occurred more often in the early acute pulmonary hypertension cohort than the late acute pulmonary hypertension cohort (186 [71%] vs 45 [41%]; P < .001). High predischarge mortality rates were observed in both groups: 90 (34%) for early acute pulmonary hypertension and 53 (49%) for late acute pulmonary hypertension. Response to iNO was categorized based on predeath FiO₂ because of mortality occurring less than 4 hours from treatment among 8 neonates with early acute pulmonary hypertension (3 positive responses) and 11 neonates with late acute pulmonary hypertension (2 positive responses).

Survival vs Death Before Discharge

Comparison of baseline and clinical characteristics between surviving and nonsurviving neonates revealed several differences (Table 1). In the early acute pulmonary hypertension group, mortality was associated with lower GA and birth weight, less use of antenatal corticosteroids, higher incidence of prolonged rupture of membranes, lower 5-minute Apgar score, and increased pre-iNO vasoactive drug use. Greater survival was associated with a larger post-iNO change in FiO₂ and more frequent use of echocardiography. In the late acute pulmonary hypertension group, most demographic factors were similar, but significant differences were observed in pretreatment illness severity. Mortality was associated with younger postnatal age at treatment; higher frequency of late-onset sepsis as the underlying diagnosis; greater pretreatment FiO₂, MAP, and vasoactive drug use; and lower blood pH. There were no intergroup differences in change in FiO₂ and echocardiography use.

Table 1. Comparison of Neonates Who Survived vs Died After iNO Exposure for Early and Late Acute Pulmonary Hypertension.

Characteristic	Early acute pulmonary hypertension (n = 262)			Late acute pulmonary hypertension (n = 109)
	Survived (n = 172)	Died (n = 90)	P value	Survived (n = 56)	Died (n = 53)	P value
Birth GA, mean (SD), wk	26.9 (2.2)	25.3 (2.5)	<.001	24.9 (1.7)	24.9 (1.7)	.81
Birth weight, mean (SD), g	1023 (341)	862 (370)	≤.001	704 (182)	764 (249)	.16
SGA, No. (%)	14 (8)	8 (9)	.84	13 (23)	7 (13)	.18
Male sex, No. (%)	99 (58)	48 (53)	.51	38 (68)	34 (64)	.68
Female sex, No. (%)	73 (42)	42 (47)		18 (32)	19 (36)
Maternal diabetes, No. (%)	20 (12)	13 (15)	.55	3 (6)	1 (2)	.62
Maternal hypertension, No. (%)	15 (9)	6 (7)	.53	15 (27)	5 (9)	.02
Receipt of ANS, No. (%)	155 (91)	72 (80)	.02	54 (96)	49 (92)	.43
Antenatal MgSO4, No. (%)	115 (69)	56 (64)	.47	49 (88)	39 (74)	.07
PROM >24 h, No. (%)	67 (40)	52 (58)	.005	13 (24)	17 (33)	.32
SNAP II ≥20, No. (%)	124 (73)	69 (83)	.06	20 (36)	21 (40)	.67
Cesarean delivery, No. (%)	107 (62)	49 (54)	.22	31 (55)	22 (42)	.15
5 min Apgar score <7, No. (%)	95 (56)	72 (82)	<.001	31 (56)	35 (66)	.30
Received surfactant, No. (%)	169 (98)	85 (94)	.13	53 (95)	49 (92)	.64
Primary underlying diagnosis, No. (%)
RDS	102 (59)	48 (53)	.34	NA	NA	.03
Pulmonary hypoplasia	55 (32)	29 (32)		NA	NA
Chronic lung disease	NA	NA		19 (34)	9 (17)
Late-onset sepsis	NA	NA		13 (23)	24 (45)
Others	15 (9)	13 (14)		24 (43)	20 (38)
Pre-iNO clinical characteristics
FiO2, median (IQR)	1.0 (0.87-1.0)	1.0 (0.95-1.0)	.28	0.90 (0.70-1.0)	1.0 (0.85-1.0)	.002
MAP, cmH2O, median (IQR)	12 (11-14)	12 (10-14)	.29	14 (13-16)	14.5 (13-16)	.18
RSS (FiO2 × MAP), median (IQR)	12 (9.0-14)	11 (9.6-14)	.74	12 (9.7-14)	14 (12-16)	.007
pH, median (IQR)	7.2 (7.1-7.3)	7.0 (7.1-7.3)	.39	7.27 (7.23-7.34)	7.21 (7.12-7.26)	<.001
Pco2, mm Hg, median (IQR)	54 (44-69)	49 (41.5-58)	.02	56.5 (48-68)	59 (52-65)	.68
Vasoactive drug use, No. (%)	34 (20)	42 (47)	<.001	15 (27)	28 (53)	.007
Echocardiography performed, No. (%)	135 (78)	55 (61)	.003	35 (63)	33 (62)	.98
Confirmed pulmonary hypertension, No. (%)a	159 (92)	84 (93)	.79	45 (80)	46 (87)	.37
Age at iNO, median (IQR), d	1 (1-2)	2 (1-2)	.05	15 (11-22)	12 (7-18)	.02
Change in FiO2, median (IQR)	60 (30-73)	20.5 (0-52)	<.001	19 (0-28)	11 (0-31)	.86
Positive iNO response, No. (%)b	144 (85)	42 (48)	<.001	26 (48)	19 (37)	.23
Duration of iNO therapy, median (IQR), h	45 (24-80)	33 (17-70)	.03	82 (44-154.5)	58.5 (13.5-108)	.04

Abbreviations: ANS, antenatal corticosteroids; FiO₂, fraction of inspired oxygen; GA, gestational age; iNO, inhaled nitric oxide; MAP, mean airway pressure; NA, not applicable; Pco₂, partial pressure of carbon dioxide; PROM, premature rupture of membranes; RDS, respiratory distress syndrome; RSS, respiratory severity score; SGA, small for gestational age; SNAP II, Score for Acute Neonatal Physiology.

^a Confirmed pulmonary hypertension defined as, if meeting echocardiography-confirmed acute pulmonary hypertension criteria: presence of any of the following features, in hierarchical order: (1) bidirectional or right-to-left patent ductus arteriosus shunt; (2) right ventricular systolic pressure greater than or equal to systolic blood pressure, where right ventricular systolic pressure = 4V² + 5 (where V is the peak velocity of tricuspid regurgitant jet in meters per second squared and 5 mm Hg was presumed right atrial pressure in all cases); (3) paradoxical interventricular septal motion; or (4) patent foramen ovale shunting right to left or a preductal and postductal peripheral oxygen saturation gradient of 5% or more.

^b Positive iNO response: a reduction in FiO₂ of 0.20 or more at 4 hours after treatment initiation compared with pretreatment FiO₂.

After adjusting for potential confounders, in the early acute pulmonary hypertension group, older birth GA (AOR, 0.74 [95% CI, 0.63-0.87]), change in FiO₂ after iNO (AOR per FiO₂ reduction of 0.10, 0.74 [95% CI, 0.65-0.84]), and more frequent use of echocardiography (AOR, 0.35 [95% CI, 0.16-0.81]) were associated with greater survival probability, while greater pre-iNO vasoactive drug use was associated with lower survival probability (AOR, 3.06 [95% CI, 1.41-6.62]) (Table 2). Change in FiO₂ discriminated between survivors and nonsurvivors, with an AUROC of 0.72 (95% CI, 0.65-0.79). The optimal change in FiO₂ for identifying survivors was 0.16, with a sensitivity of 0.87 (95% CI, 0.82-0.93) and specificity of 0.49 (95% CI, 0.38-0.59). For the late acute pulmonary hypertension group, only lower pre-iNO FiO₂ and higher pre-iNO blood pH were associated with greater survival probability. Change in FiO₂ was not associated with survival (AOR, 0.47 [95% CI, 0.17-1.30]).

Table 2. Factors Associated With Mortality in Preterm Neonates After iNO Exposure for Early and Late Acute Pulmonary Hypertension.

Factor	Odds ratio (95% CI)
	Crude	Adjusted
Early acute pulmonary hypertension cohort (iNO use ≤72 h of age)
Birth GA	0.72 (0.63-0.82)	0.74 (0.63-0.87)
SGA	1.17 (0.47-2.90)	1.27 (0.38-4.25)
Male sex	0.83 (0.49-1.42)	0.78 (0.39-1.55)
Receipt of ANS	0.44 (0.21-0.94)	0.49 (0.18-1.30)
PROM >24 h	0.44 (0.26-0.75)	0.32 (0.15-0.73)
SNAP II ≥20	1.82 (0.93-3.56)	1.03 (0.43-2.46)
Cesarean delivery	0.72 (0.42-1.22)	0.92 (0.44-1.92)
Primary diagnosis (reference: pulmonary hypoplasia)
RDS	1.01 (0.57-1.82)	0.52 (0.22-1.26)
Others	1.51 (0.59-3.85)	0.89 (0.25-3.20)
Change in FiO2, per reduction of 0.10	0.75 (0.67-0.83)	0.74 (0.65-0.84)
Echocardiography use during episode	0.42 (0.24-0.75)	0.35 (0.16-0.81)
Confirmed acute pulmonary hypertension (vs clinically presumed)a	1.08 (0.40-2.96)	0.66 (0.19-2.35)
Pre-iNO vasoactive drug use	3.73 (2.08-6.69)	3.06 (1.41-6.62)
Late acute pulmonary hypertension cohort (iNO use >72 h of age)
Birth GA	0.96 (0.77-1.20)	0.99 (0.74-1. 32)
SGA	0.41 (0.14-1.18)	0.71 (0.18-2.80)
Male sex	0.80 (0.36-1.79)	0.81 (0.28-2.36)
Age at iNO initiation	0.99 (0.97-1.01)	0.99 (0.96-1.02)
Primary diagnosis (reference: chronic lung disease)
Late-onset sepsis	4.75 (1.62-13.96)	2.24 (0.61-8.19)
Others	2.07 (0.74-5.73)	0.94 (0.25-3.49)
Confirmed acute pulmonary hypertension (vs clinically presumed)a	1.65 (0.58-4.63)	1.10 (0.32-3.79)
Positive response to iNOb	0.62 (0.29-1.35)	0.47 (0.17-1.30)
Pre-iNO FiO2, per 0.10	1.04 (1.01-1.07)	1.04 (1.01-1.08)
Pre-iNO vasoactive drug use	3.03 (1.35-6.80)	1.67 (0.59-4.76)
Pre-iNO pH, per 0.05 units	0.68 (0.55-0.85)	0.72 (0.56-0.94)

Abbreviations: ANS, antenatal corticosteroids; FiO₂, fraction of inspired oxygen; GA, gestational age; iNO, inhaled nitric oxide; PROM, premature rupture of membranes; RDS, respiratory distress syndrome; SGA, small for gestational age; SNAP II, Score for Acute Neonatal Physiology.

^a Confirmed pulmonary hypertension defined as, if meeting echocardiography-confirmed acute pulmonary hypertension criteria, presence of any of the following features, in hierarchical order: (1) bidirectional or right-to-left patent ductus arteriosus shunt; (2) right ventricular systolic pressure greater than or equal to systolic blood pressure, where right ventricular systolic pressure = 4V² + 5 (where V is the peak velocity of tricuspid regurgitant jet in meters per second squared and 5 mm Hg was presumed right atrial pressure in all cases); (3) paradoxical interventricular septal motion; or (4) patent foramen ovale shunting right to left or a preductal and postductal peripheral oxygen saturation gradient of 5% or more.

^b Positive iNO response: a reduction in FiO₂ of 0.20 or more at 4 hours after treatment initiation compared with pretreatment FiO₂.

Positive Response to iNO vs Nonresponse

Kaplan-Meier analysis demonstrated lower mortality rates among responders compared with nonresponders in both groups, with statistical significance only in the early acute pulmonary hypertension group (Figure 2). Surviving neonates had a high incidence of prematurity-related morbidities in both groups (Table 3); iNO responders in the early acute pulmonary hypertension cohort showed a relatively lower bronchopulmonary dysplasia rate than nonresponders. Analysis of factors associated with a positive response, restricted to the early acute pulmonary hypertension group, revealed several variables of interest (eTable 1 in Supplement 1); however, on multivariable logistic regression analysis only older birth GA was associated with greater response (AOR 1.27 [95% CI, 1.05-1.54]). In the whole cohort, early acute pulmonary hypertension (AOR 5.40 [95% CI, 2.94-9.93]) was the strongest independent factor associated with positive iNO response, followed by greater postmenstrual age at treatment (AOR, 1.14 [95% CI, 1.04-1.26]). Within subgroups of neonates with birth weight less than 1000 g or less than 750 g, positive iNO response continued to be associated with lower mortality with no difference in rates of severe intraventricular hemorrhage (eTable 2 in Supplement 1).

Figure 2. Kaplan-Meier Survival Curves Based on Inhaled Nitric Oxide (iNO) Responsiveness Among Neonates With Early and Late Acute Pulmonary Hypertension.

Responsiveness was defined by a minimum reduction in fraction of inspired oxygen of 0.20 or more at 4 hours after initiation of iNO. Circles indicate censoring.

Table 3. Rates of Major Morbidities Among Surviving Neonates With Early or Late Acute Pulmonary Hypertension Categorized by iNO Response Pattern^a.

Morbidity among survivors	Early acute pulmonary hypertension (n = 172)			Late acute pulmonary hypertension (n = 56)
	Positive response, No. (%) (n = 144)b	Absence of positive response, No. (%) (n = 26)	P value	Positive response, No. (%) (n = 26)b	Absence of positive response, No. (%) (n = 28)	P value
BPD at 36 wk	98 (68)	23 (88)	.03	25 (96)	26 (93)	>.99
Necrotizing enterocolitis, stage 2a or higher	9 (6)	3 (12)	.40	1 (4)	8 (29)	.03
Interventricular hemorrhage, grade 3 or higher	32 (22)	6 (23)	.94	6 (23)	2 (7)	.10
ROP needing treatment	18 (15)	6 (25)	.24	13 (52)	13 (48)	.78

Abbreviations: BPD, bronchopulmonary dysplasia; FiO₂, fraction of inspired oxygen; iNO, inhaled nitric oxide; ROP, retinopathy of prematurity.

^a Two neonates missing each from early and late acute pulmonary hypertension iNO response were not included in this analysis.

^b Positive iNO response: a reduction in FiO₂ of 0.20 or more at 4 hours after treatment initiation compared with pretreatment FiO₂.

Discussion

The absence of quality data has contributed to the ongoing debate regarding the effectiveness of iNO for preterm neonates with acute pulmonary hypertension–associated HRF, leading to conflicting expert consensus guidelines and significant variability in physician practices and perceptions. A National Institutes of Health Consensus Panel, while not ruling out the use of iNO in preterm neonates with acute pulmonary hypertension, advised prior discussion with families regarding potential risks and benefits, as well as the remaining uncertainty of evidence.¹⁸ The Pediatric Pulmonary Hypertension Network recommend iNO use for preterm neonates with proven acute pulmonary hypertension physiology, citing established safety, lack of alternatives, and observations of rapid resolution of symptoms.¹⁹ A recent survey among 304 members of the American Association of Pediatrics on iNO initiation in preterm neonates with acute pulmonary hypertension–associated HRF reported that 36% supported treatment, 22% opposed it, and 42% involved parents in decision-making.²⁰ All groups cited similar published guidelines to support their decisions. Conversely, an Australasian survey reported that 19 of 20 NICUs administered early rescue iNO (<72 hours of age) to preterm neonates with HRF.²¹ Among these, 70% administered early rescue iNO when acute pulmonary hypertension was confirmed by echocardiography.

Study Design Considerations and Registry Data

An adequately powered, placebo-controlled, randomized clinical trial without open-label treatment would be ideal to examine the effectiveness of iNO for preterm neonates with acute pulmonary hypertension–associated HRF. However, the infrequency of this condition, widespread availability and safety of iNO, lack of alternatives, and absence of equipoise among many clinicians to withhold open-label treatment limits its feasibility. Similarly, identifying an adequate number of untreated controls for comparison, matched for illness severity, is also challenging. Although network-level data may offer valuable insights, so far these have lacked data granularity to account for underlying pathophysiology.²² Ellsworth et al²³ examined data from 92 635 preterm neonates in the Pediatrix Medical Group registry, identified 767 (0.8%) with a diagnosis of pulmonary hypoplasia, and analyzed 151 matched pairs of patients exposed to iNO and unexposed to iNO, finding no difference in clinical outcomes. However, echocardiographic diagnosis of acute pulmonary hypertension and pretreatment oxygenation status was unaccounted for. This is important, as only approximately 60% of patients with pulmonary hypoplasia are reported to develop acute pulmonary hypertension–associated HRF needing rescue iNO.²⁴ In fact, Pediatrix investigators observed a 33% mortality risk reduction with iNO exposure in the subgroup with a concomitant given diagnosis of persistent pulmonary hypertension of the newborn, albeit limited by small sample size (hazard ratio, 0.67 [95% CI, 0.45-1.01]).²³ This finding highlights the importance of and challenges in accounting for baseline illness severity, a key factor governing iNO exposure decision in practice.

The clinical intent with the use of iNO in acute pulmonary hypertension–associated HRF is to improve oxygenation. An alternative investigative approach, given the aforementioned limitations, may be designing prospective registries to quantify effectiveness (acute response) and examine if response translates to better clinical outcomes, as known among term and near-term neonates.¹ Such datasets must account for pretreatment clinical status and illness severity. The PaTTerN registry included 55 very preterm (27 weeks and 0 days to 33 weeks and 6 days) and 85 full-term or near-term (≥34 weeks) neonates who received iNO for 24 hours or more for confirmed acute pulmonary hypertension at 7 days of age or less in the US between 2017 and 2020.⁷ The groups showed a 25% or greater reduction in oxygenation index (OI) in 91% of very preterm neonates and 88% of full-term or near-term neonates, with similar time to improvement across HRF severity categories. This registry, however, excluded critically ill patients deemed at risk of death within 24 hours at treatment onset. Prospective registry data from Japanese NICUs, including 431 neonates younger than 34 weeks and 675 neonates 34 weeks or older who received iNO for acute pulmonary hypertension–associated HRF at 7 days of age or younger between 2010 and 2012 also noted similar oxygenation index reduction at 1 hour and 24 hours after iNO in both groups of patients and across HRF severity categories.²⁵ Providing data on iNO responsiveness in preterm neonates, its interaction with timing in postnatal NICU course, and associated factors, we further supplement previous evidence.

In the absence of evidence, as with previous reports,^26,27,28,29 we used the pragmatic definition of FiO₂ reduction of 0.20 or more sustained for 4 hours to define a positive response. However, we also investigated and found that an FiO₂ reduction of 0.16 or more provided the highest combination of sensitivity and specificity for identifying survival, potentially providing the first evidence of an outcome-based definition of iNO response for preterm neonates with early acute pulmonary hypertension. Even a modest reduction in FiO₂, as small as 0.10, may be clinically relevant in early acute pulmonary hypertension and was associated with reduced odds of mortality.

Early vs Late Acute Pulmonary Hypertension Among Preterm Neonates

By virtue of prolonged hospitalization and associated morbidities, preterm neonates may present with acute pulmonary hypertension at different ages and with different underlying causes. To distinguish perinatal and new-onset postnatal-acquired pathologic conditions, we studied early and late acute pulmonary hypertension separately. Although our findings of different iNO response patterns and factors associated with mortality provide validation of our methods, we acknowledge that the definition based on 72 hours of age is somewhat arbitrary and does not rule out some overlap, where patients treated on days 4 or 5 of age may in fact have early acute pulmonary hypertension physiology that presented or was treated late. However, most patients with late acute pulmonary hypertension patients were older than 7 days at treatment and likely represent a new-onset postnatal illness. Perinatal acute pulmonary hypertension phenotype (early acute pulmonary hypertension) had a high iNO response rate, with greater FiO₂ reduction associated with greater odds of survival, and surviving responders demonstrated a lower bronchopulmonary dysplasia rate than nonresponders. The latter likely reflects quicker HRF resolution, lowering the burden of oxygen therapy and mechanical ventilation. Lower GA and vasoactive drug exposure are established risk factors for mortality among neonates, also identified in this study.^30,31,32 In the absence of mechanistic data, we speculate that the association between echocardiography use and lower mortality in early acute pulmonary hypertension may indicate the impact of accurate diagnostics and delineation of circulatory pathophysiology, facilitating tailored treatment strategies.^33,34 Our data provide further support for iNO use as a rescue therapy for critically ill preterm neonates with acute pulmonary hypertension–associated HRF presenting during the transitional period (ie, persistent pulmonary hypertension of the newborn).

Our late acute pulmonary hypertension cohort demonstrated findings that were distinct from those in the early acute pulmonary hypertension cohort, namely, lower iNO response rate (41%), modest post-iNO change in FiO₂, and higher mortality. Although late acute pulmonary hypertension responders also had higher survival, regression analysis showed it to be associated with pretreatment illness status. Although this observation raises questions on the relevance of iNO and associated short-term oxygenation improvement in postnatal-acquired acute pulmonary hypertension, it may also suggest differences in circulatory pathophysiology and a need to examine the role of iNO differently. Unlike early acute pulmonary hypertension, right-to-left shunting is often not a significant component of late acute pulmonary hypertension pathophysiology due to closure of fetal shunts. Thus, iNO may induce pulmonary vasodilation but not an acute FiO₂ improvement due to nonfeasibility of shunt reversal. Whether pulmonary vascular resistance lowers with iNO in late acute pulmonary hypertension, as well as its clinical relevance, needs further investigation. These data may assist administrators and regulatory bodies in standardizing, rationalizing, and defining the off-label use of iNO among preterm neonates.

Mechanistically, NO and the cyclic guanosine monophosphate pathway regulate pulmonary arterial vasodilation after birth.^35,36 Disrupted endogenous NO production during transition prompts consideration of exogenous iNO to reverse pulmonary arterial vasoconstriction.³⁷ However, pathways governing late acute pulmonary hypertension episodes remain understudied. It is possible that NO pathway disruption is not the predominant pathophysiology. Biological considerations for late acute pulmonary hypertension may include acute respiratory distress syndrome, characterized by sudden and severe gas exchange impairment from lung-related processes or dysregulated pulmonary vascular function.³⁸ Endothelial and microcirculatory dysfunction during systemic inflammation in late-onset sepsis or necrotizing enterocolitis could also play a role.^30,39,40 Additional research is warranted to elucidate mechanisms and identify therapeutic targets for late acute pulmonary hypertension phenotype in preterm neonates.

Limitations

This study has some limitations. Causality cannot be established given the lack of untreated controls and the observational study design. However, iNO responsiveness identifying neonates more likely to survive an early acute pulmonary hypertension episode may help risk stratification and inform family counseling. In the absence of harmonized management protocols, some acute pulmonary hypertension cases may have remained untreated and unidentified. In addition, while iNO was standard in NICUs treating acute pulmonary hypertension–associated HRF cases, other measures to improve FiO₂ could not be ruled out, although no patient received another pulmonary vasodilator within 4 hours of iNO initiation. In addition, only 72% of patients had diagnoses confirmed by echocardiography, with remaining cases diagnosed using preductal and postductal peripheral oxygen saturation gradient and clinician assessments. Last, due to lack of data, we are unable to examine the association between iNO exposure and subsequent need for patent ductus arteriosus treatment or its interaction with clinical outcomes. However, no patient received concomitant patent ductus arteriosus treatment during the included iNO episode.

Conclusions

This cohort study found that most very preterm neonates presenting with acute pulmonary hypertension–associated HRF during postnatal transition responded positively to iNO treatment, and iNO responsiveness identifies survival benefit independent of pretreatment illness severity status. On the other hand, preterm neonates receiving iNO treatment for acute pulmonary hypertension acquired during the postnatal NICU course demonstrated only modest improvement in oxygenation, with mortality associated primarily with pretreatment illness severity and not with change in FiO₂ with iNO. Further research is needed to delineate pathophysiological mechanisms underlying late acute pulmonary hypertension presentations in preterm neonates to guide effective management strategies.

Supplement 1.

eTable 1. Comparison of Very Preterm Neonates With Early aPH Who Had a Positive Response to iNO vs Those Who Did Not

eTable 2. Mortality and Severe Intraventricular Hemorrhage (IVH) Between Positive iNO Responders and Nonresponders Among Neonates With Extremely Low Birth Weight

Supplement 2.

Nonauthor Collaborators. Canadian Neonatal Network Investigators

Supplement 3.

Data Sharing Statement