Objective To describe short-term and long-term outcomes of preterm neonates with severe acute pulmonary hypertension (aPHT) in relation to response to rescue inhaled nitric oxide (iNO) therapy.
Design Retrospective cohort studyover a 6 year period.
Setting Tertiary neonatal intensive care unit.
Patients 89 neonates <35 weeks gestational age (GA) who received rescue iNO for aPHT, including 62 treated at ≤3 days of age (early aPHT).
Interventions iNO ≥ 1 hour.
Main outcome measures Positive responders (reduction in fraction of inspired oxygen (FiO2) ≥0.20 within 1 hour of iNO) were compared with non-responders. Primary outcome was survival without moderate-to-severe disability at 18 months of age.
Results Mean (SD) GA and birth weight was 27.7 (3.0) weeks and 1077 (473) gm, respectively. Median (IQR) pre-iNO FiO2 was 1.0 (1.0, 1.0). Positive response rate to iNO was 46%. Responders showed improved survival without disability (51% vs 15%; p<0.01), lower mortality (34% vs 71%; p<0.01) and disability among survivors (17% vs 50%; p=0.06). Higher GA (adjusted OR: 1.44 (95% CI 1.10 to 1.89)), aPHT in context of preterm prolonged rupture of membranes (6.26 (95% CI 1.44 to 27.20)) and positive response to rescue iNO (5.81 (95% CI 1.29 to, 26.18)) were independently associated with the primary outcome. Compared with late cases (>3 days of age), early aPHT had a higher response rate to iNO (61% vs 11%; p<0.01) and lower mortality (43% vs 78%; p<0.01).
Conclusion A positive response to rescue iNO in preterm infants with aPHT is associated with survival benefit, which is not offset by long-term disability.
- neurodevelopmental outcomes
- hypoxic respiratory failure
- preterm prolonged rupture of membranes
Statistics from Altmetric.com
What is already known on this topic?
A proportion of severe hypoxaemia in preterm neonates presenting during the immediate postnatal period is in association with acute pulmonary hypertension.
In this setting, treatment with inhaled nitric oxide can result in acute improvement in oxygenation status.
Compared with matched historic untreated cases, inhaled nitric oxide treatment may be associated with lower mortality.
What this study adds?
Response patterns to rescue inhaled nitric oxide treatment for preterm neonates with acute pulmonary hypertension is similar to those previously reported for term neonates.
An acute positive response to therapy is associated with lower mortality as well as lower long term disability among survivors.
Neonates presenting after 3 days of age rarely respond to nitric oxide therapy and have poor prognosis, particularly when associated with culture positive sepsis.
In spite of equivocal evidence and published statements,1–4 5%–7% of preterm infants in tertiary neonatal intensive care units (NICUs) receive inhaled nitric oxide (iNO) treatment.5–7 It is postulated that contemporary practice is aimed at acute stabilisation of physiological parameters when refractory hypoxaemia is perceived to be from acute pulmonary hypertension (aPHT),8 a catastrophic disorder conspicuously absent from iNO trials of preterm neonates. Despite recent observations of improved oxygenation following iNO in preterm neonates with aPHT secondary to pulmonary hypoplasia,9–11 in some centres iNO may be completely withheld from this population irrespective of the underlying aetiology.12 This may be due to concerns regarding subsequent clinical outcomes of these extremely sick infants. It is not known whether successful use of iNO to acutely improve physiological parameters translates into improved survival and if that occurs without long-term disability.
The primary objective of this study was to characterise iNO response patterns in preterm infants with aPHT and investigate their relationship to short-term and long-term neurodevelopmental outcomes. Our secondary objective was to identify factors associated with improved outcomes and favourable response to iNO therapy. We hypothesised a priori that in preterm neonates with aPHT, a positive response to rescue iNO therapy will be associated with improved clinical outcomes.
Study design and patient selection
We conducted a retrospective cohort study in the tertiary NICU at Mount Sinai Hospital, Toronto, over a 6-year period. The study was approved by the Institutional Research Ethics Board. All patients who received iNO during the study period were identified from an electronic database maintained by the Respiratory Therapy Department, and health charts were screened for eligibility. The inclusion criteria were gestational age (GA) <35 weeks at birth and exposure to iNO ≥60 min. Our exclusion criteria were: (1) iNO used for chronic pulmonary hypertension related to chronic lung disease (CLD), (2) infants transferred in or out of our unit while receiving iNO, (3) congenital or genetic anomalies and (4) structural cardiac defects except patent ductus arteriosus, patent foramen ovale or ventricular septal defects.
Mount Sinai Hospital is a high-risk feto-maternal centre, whose NICU comprises predominantly inborn preterm patients. Practices related to iNO use are governed by a standard operating policy which did not change during the study period. The only stated use of iNO is as rescue therapy for refractory hypoxaemia secondary to aPHT, as diagnosed by the attending physician, irrespective of the underlying aetiology. Routine use of iNO for prevention of CLD is not practised at our centre. For this study, aPHT was categorised as ‘definite’ if diagnosis was documented on echocardiography reports performed before or within 24 hours of iNO initiation or if there was pre-iNO documentation of a pre–post ductal peripheral oxygen saturation difference ≥5% on at least two occasions 1 hour apart. Otherwise, aPHT was classified as ‘clinically presumed’. Echocardiograms, whenever performed, were reported either by a paediatric cardiologist or by the neonatologist heading our unit’s functional echocardiography programme. Management strategies commonly employed in our unit before iNO include invasive mechanical ventilation—conventional and high frequency, surfactant therapy and sedation or muscle relaxant as deemed medically appropriate. The use of these measures, as well as the decision to initiate iNO, its timing and dose, was at the discretion of the attending physician. The starting dose was 20 parts per million (ppm) for the majority of infants, the minimum being 10 ppm. In latter cases, dose was escalated to 20 ppm if no clinical improvement was observed. For this study, patients were categorised after 1 hour of iNO therapy as ‘positive responders’ if FiO2 reduced by ≥0.20 or as ‘non-responders’ if FiO2 increased, remained unchanged or reduced by <0.20. Our policy stated that iNO could be discontinued if the reduction in FiO2 after an hour was <0.10, else weaned according to a predefined protocol. The final decision, however, to maintain, wean or discontinue iNO was at the discretion of the attending physician.
Data collected from health records included baseline demographics, antenatal and perinatal factors such as maternal illness, fetal anomalies, timing of rupture of membranes, documented oligohydramnios, antenatal corticosteroid use, mode of delivery, Apgar scores and surfactant use. Preterm prolonged rupture of membrane (PPROM) was defined as >18 hours before birth while premature PPROM (PPPROM) was defined as PPROM occurring before onset of labour. Severity of illness during the first 12 hours of age was assessed using the score for neonatal acute physiology (SNAP) II, which is prospectively recorded in our unit; SNAP II >20 was considered a marker of significant illness severity. The primary diagnosis attributable to aPHT was ascertained and categorised as: PPPROM >2 weeks, culture positive sepsis or others (included intrauterine growth restriction, twin-to-twin transfusion syndrome, pulmonary haemorrhage, congenital pleural effusion, hydrops fetalis and idiopathic/possible respiratory distress syndrome (RDS)). Data related to iNO use included age at treatment, pre-iNO variables such as FiO2, ventilation mode, mean airway pressure, mean blood pressure and use of sedation or muscle relaxant. To distinguish aPHT occurring during postnatal transition versus later in postnatal course and based on our observation of significant change in iNO responsiveness after 3 days of age, we classified aPHT as early (iNO initiation ≤3 days of age) or late (>3 days of age). In addition, we recorded FiO21-hour post-iNO, total duration of iNO exposure and if PHT was documented on echocardiography reports whenever available. Detailed echocardiography data were not collected as part of this study.
Outcome data included predischarge mortality and common neonatal morbidities. Based on days after iNO initiation, mortality was subcategorised as ‘early’ when it occurred ≤7 days or ‘late’ if after 7 days. Among morbidities, intraventricular haemorrhage ≥ grade 3 was determined as per Papile’s classification, while CLD was defined as need for supplemental oxygen or respiratory support at 36 weeks corrected GA. Neurodevelopmental data were abstracted from our follow-up clinic. We defined ‘disability’ as composite of the following at 18 months of age: (1) score of <70 in any domain (motor, cognitive or language) on Bayley Scale of Infant and Toddler Development, Third Edition; (2) cerebral palsy classified by Gross Motor Function Classification System (GMFCS) score ≥3; (3) severe hearing loss needing bilateral hearing implants; and (4) visual impairment defined as uncorrectable vision loss in at least one eye.
The primary outcome was survival without disability. Mortality, common neonatal short-term morbidities, disability and cerebral palsy with GMFCS score ≥3 were considered as secondary outcomes.
For the primary analysis, ‘positive responders’ were compared with ‘non-responders’. In addition, early aPHT cohort was also evaluated separately due to notable differences seen on comparison to late aPHT. Intergroup comparison was performed using the chi square test or Fischer’s exact test for categorical variables and Student’s t-test or Wilcoxon rank-sum test for continuous variables as appropriate. Multiple logistic regression analysis was performed to identify factors associated with the primary outcome in the whole cohort using the following covariates: gestational age, early versus late aPHT, positive versus non-response to iNO, gender and PPROM >18 hours. These covariates were chosen based on the result of univariate analysis or known clinical relevance. Primary diagnosis was not used as a factor in regression analysis due to its collinearity with early and late aPHT. Neonates with PPPROM >2 weeks and sepsis, however, are described separately. Regression analysis was also performed to identify factors associated with a positive iNO response in the whole cohort and early aPHT.
Of 101 neonates with GA <35 weeks who received iNO during the study period, 89 with a mean (SD) GA and birth weight of 27.7 (3.0) weeks and 1077 (473) gm, respectively, were included. Four infants were excluded as iNO was given for chronic PHT and 8 because of being transferred while receiving iNO. Median age at treatment was 1 (1, 53) days. Sixty-two (70%) cases were of early aPHT. In total, 69 (78%) cases were classified as definite aPHT; 44 based on diagnosis documented on echocardiography reports and 25 from documented pre-iNO pre–post ductal saturation difference. Pre-iNO, all neonates were invasively ventilated (93% on high frequency ventilation) and 66 (74%) were receiving 100% oxygen. Among 32 cases of PPPROM >2 weeks duration, mean latency period of ROM was 6.4 (5.2) weeks. In the ‘other’ diagnostic category, only eight cases were classified as idiopathic/possible RDS. Long-term follow-up data were missing for four survivors; all had early aPHT with a positive iNO response.
Response to iNO and clinical outcomes
Overall, positive response rate was 46%. iNO responsiveness declined significantly after 3 days of age; with 31 of 54 (57%), 5 of 6 (83%), 1 of 2 (50%) and 0 of 5 (0%) neonates responding to iNO on days 1, 2, 3 and 4–7, respectively. Positive responders were more likely to be female, had received surfactant at birth, were of lower postnatal age at treatment and were less likely to have aPHT in context of sepsis (table 1). Restricting analysis to early aPHT cohort revealed lower GA and higher surfactant treatment among responders versus non-responders, but no differences were seen in distribution of diagnosis. In all but five cases, iNO was continued >1 hour despite a lack of positive response. Positive response to rescue iNO therapy was associated with higher rate of survival without disability, lower mortality as well as less disability among survivors (table 2).
Compared with late aPHT, early aPHT had a higher positive response rate (38/62 (61%) vs 3/27 (11%); p<0.01) and lower mortality (27/62 (44%) vs 21/27 (78%); p<0.01). Of the three late aPHT responders, one died and one developed disability; while, all non-responders either died (n=20) or survived with disability (n=4). iNO responsiveness and related outcomes for neonates with sepsis and PPPROM >2 weeks is shown in figure 1.
On regression analyses, advanced GA, PPROM >18 hours and positive response to iNO were predictive of survival without disability (table 3), while lower GA at birth and early aPHT were associated with iNO responsiveness (figure 2A). Early aPHT cohort showed a trend towards association between having definite aPHT and iNO responsiveness; however, this did not reach statistical significance (figure 2B).
Comparison with literature
The indication for iNO therapy in preterm neonates is not well defined. Unlike term neonates, placebo controlled trials enrolling preterm infants have primarily engaged iNO’s anti-inflammatory and pro-lung development properties in mild to severe lung disease, largely secondary to RDS.1 3 13–16 With greater penetration of antenatal steroids and postnatal surfactant therapy, RDS is no longer a frequent cause of oxygenation failure.17 Instead, as seen in our cohort, most preterm cases of severe persistent hypoxaemia are secondary to disorders associated with pulmonary vascular development, sepsis or disruption of postnatal circulatory transition; this likely explains the high response rate to iNO therapy observed in our study.
Our current knowledge regarding acute pulmonary hypertensive disorders in preterms is limited to case series of neonates with PPPROM/oligohydramnios presenting with hypoxaemia during the immediate postnatal period.9 10 18 19 Post-hoc analysis of such patients enrolled in the larger Preemie Inhaled Nitric Oxide trial (12 out 449 neonates) reported improvement in oxygenation and lower mortality with iNO therapy; however, due to sample size (n=6 per group) this difference was not statistically significant.20 Retrospective comparative PPPROM case–control studies have yielded similar findings.10 19 The overall response rate to rescue iNO in preterm neonates, its role in aPHT from disorders other than PPPROM and long-term outcome of these neonates has not been explored before. This is an important knowledge gap contributing to the clinical uncertainty surrounding the rescue use of iNO for these extremely sick preterm neonates.
Our study suggests that patterns of iNO response in preterm neonates with aPHT may mirror that of term neonates. A significant proportion of cases, but not all, improve acutely with iNO and have better outcomes compared with non-responders.21 This was particularly true in early aPHT where we observed a response rate as high as 61%.22 The reduced mortality in iNO responders likely reflects the fact that, unlike term neonates, extracorporeal membrane oxygenation is not available for preterm infants. Importantly, we confirmed that the survival benefit is not offset by long-term disability; the overall disability rate being comparable to previous reports for iNO-treated term neonates.23 Though not specifically quantified in this study, our interesting observation of lower disability among surviving responders versus surviving non-responders may be explained by the expected shorter duration of severe hypoxaemia in responders.
The use of iNO in preterm neonates has been a subject of much recent debate with differing opinions from expert bodies.4 12 Given the lack of a control group, our study cannot provide definitive proof; it is however an important step towards solving this critical knowledge gap. As recommended in the recent guideline by Pediatric Pulmonary Hypertension Network, our results support rescue iNO treatment for refractory hypoxaemia in preterm infants with aPHT.12 In this regard, confirming the diagnosis of aPHT may be an important factor. Echocardiography has been shown to aid early diagnosis and initiation of iNO in PPPROM neonates, resulting in lower mortality compared with historical controls. Using a combination of echocardiography reports and clinical evidence of pre-iNO right-to-left ductal shunt, we were able to confirm aPHT in 78% cases. Indeed in early aPHT cohort, we saw a trend towards association between definitive diagnosis and iNO responsiveness; however, it did not reach statistical significance. Future larger studies will be required to confirm this relationship as well as its utility in disorders other than PPPROM, in particular late aPHT.
Factors associated with iNO responsiveness and clinical outcomes
The disparity in iNO responsiveness between early and late aPHT suggests pathophysiological differences governing its phenotype. Tracheal aspirate analysis over the first 4 days of age in hypoxaemic aPHT preterm neonates with PPROM and/or sepsis has shown a transient defect in endogenous nitric oxide generation.17 This confirms a key role of nitric oxide during postnatal transition, providing a biological rationale for the observed response to exogenous iNO. The mechanisms responsible for late aPHT have not been investigated; however, sepsis being the dominant aetiology suggests inflammation to be an important factor. In spite of echocardiographic diagnosis of aPHT, neonates with late-onset sepsis in our study were non-responsive to iNO and all had adverse outcomes. Hence, we would caution clinicians against the sole use of iNO in such cases as valuable time may be lost in initiating and establishing a response to therapy. Future research should focus on eliciting mechanisms and evaluate alternate therapies for this patient subset.
We also observed an inverse relationship between GA and iNO responsiveness. This may be a reflection of underlying aetiologies, most of which are known to be associated with in-utero alterations in pulmonary vascular development. We speculate that longer exposure to in-utero factors may result in relatively advanced disease, decreasing iNO efficacy. Nevertheless, higher GA was still predictive of survival without disability. Positive iNO response and aPHT in context of PPROM were also associated with better outcome; the former likely by limiting exposure to severe hypoxaemia while the latter may reflect a relatively less severe disorder compared with the remaining cohort.
Our study has some important limitations. First, we were unable to confirm pre-iNO optimisation of ventilatory management and relied on FiO2 to classify response, as measuring pre-iNO PaO2 is not a mandated practice in our unit. The majority of patients, however, were receiving high-frequency ventilation, FiO2 of 1.0 and had definite aPHT. Additionally, in positive responders, FiO2 decreased significantly within an hour, mimicking the anticipated response to iNO in aPHT. Second, without a control group of untreated infants, we cannot confirm iNO’s impact on neonatal outcomes. However, a placebo controlled trial in this patient subset may not be feasible due to factors such as low event rate, severity of illness, accessibility and safety of iNO and absence of alternate therapies.12 At minimum, our data suggest that a positive iNO response identifies infants with better clinical outcomes. Third, we did not collect detailed echocardiographic data as this was not the focus of our study. Investigating sensitivity of various echocardiographic indices of aPHT in predicting iNO response in early and late cases is an important area for future research.23 Additionally, although we were able to confirm aPHT in the majority of cases, we cannot rule out a degree of misclassification with clinically presumed aPHT, as echocardiograms in our study were not always performed pre-iNO. Lastly, our sample size was relatively small; however, it remains the largest report to date of preterm infants with hypoxaemia secondary to aPHT physiology. This is not surprising given that this disorder is reported to occur in only 2% of the population.17
Irrespective of underlying aetiology, a significant proportion of preterm infants with severe hypoxaemia may respond acutely to rescue iNO therapy when aPHT is the underlying physiology, particularly during the first 3 days of age. A positive response to iNO is independently associated with survival without long-term disability, lower predischarge mortality and less disability among survivors. This study provides the basis for future clinical trials of rescue iNO for aPHT in hypoxic preterm neonates, except when aPHT occurs secondary to late onset sepsis, for which alternate therapies need to be evaluated.
Contributors MB, SG, DEW, PJM, EK, PSS and AJ made substantial contribution to the study conception and design. MB, SG and AJ finalised data collection forms and DEW, PJM, EK, PSS provided critical input. MB and SG were responsible for data collection which were directly supervised by AJ. AJ performed the statistical analysis for the study with critical inputs from DEK and PSS. MB was responsible for producing the first draft of the manuscript which was directly supervised by AJ. SG, DEW, PJM, EK and PSS made significant contributions in data interpretation and revising the manuscript critically.
Competing interests None declared.
Ethics approval Mount Sinai Hospital Research Ethics Board.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.