Introduction

Persistent pulmonary hypertension of the newborn (PPHN) is a complication during neonatal transition characterized by a sustained elevation in pulmonary vascular resistance (PVR). The incidence of PPHN is approximately 2 per 1000 live births in late preterm and term neonates [1]. Despite appropriate therapy, PPHN is associated with a mortality rate of 10% and significant neurodevelopmental, cognitive, and hearing disabilities among survivors [1,2,3,4,5].

PPHN is secondary to a number of underlying causes associated with impaired relaxation of the pulmonary vasculature, such as perinatal asphyxia, infection, and meconium aspiration syndrome (MAS) [6,7,8]. Elevated PVR leads to decreased pulmonary blood flow (PBF), resulting in hypoxemia and acidosis. In turn, hypoxemia and acidosis cause the pulmonary vessels to constrict, thereby worsening the PPHN [8].

Postnatal increase in oxygen tension in the lungs is one of the most crucial factors to promote pulmonary vasodilatation [9]. Therefore, supplemental oxygen therapy is the mainstay in the treatment of PPHN to reduce PVR and increase oxygenation [6,7,8]. Inhaled nitric oxide (iNO) is a selective pulmonary vasodilator and is considered the first-line PPHN-specific therapy when oxygen therapy alone is insufficient to reduce PVR [6, 8].

High concentrations of oxygen are often needed to prevent hypoxemia-induced deterioration of PPHN, but this can also increase the risk of hyperoxemia, especially when the infant stabilizes, and PVR reduces. However, titrating oxygen therapy within the narrow therapeutic range is difficult in infants with PPHN. Infants with PPHN often experience respiratory and hemodynamic instability, which warrants caution regarding weaning off oxygen therapy.

As a result, infants with PPHN may unintentionally face an increased risk of hyperoxemia, which can cause organ injury [10]. Avoiding hyperoxemia is potentially especially relevant in the context of PPHN, as it can directly promote pulmonary vasoconstriction, further worsening the PPHN [11]. Previous translational studies have shown that there is a certain “oxygenation threshold” above which pulmonary vasodilatation is no longer enhanced and subsequent response to iNO therapy is impaired [12,13,14,15]. Moreover, reoxygenation after a period of hypoxemia may amplify or even directly cause organ injury, rendering infants with PPHN, especially those associated with perinatal asphyxia, particularly susceptible to hyperoxemia [16]. In line with this, Kilmartin et al. showed that PPHN increased the risk of death and brain injury in infants with perinatal asphyxia [17].

There is no data available regarding how often hyperoxemia occurs during PPHN treatment in infants. Therefore, we performed a retrospective cohort study to determine the occurrence of hyperoxemia in late preterm and term infants treated for PPHN.

Methods

A retrospective study was performed including late preterm and term infants (≥ 34 + 0 weeks) admitted to the Neonatal Intensive Care Unit (NICU) at Leiden University Medical Center (LUMC) who received iNO therapy between November 2011 and June 2023. As this study aims to investigate the association between PPHN therapy and hyperoxemia, echocardiographic confirmation of PPHN was not considered an inclusion criterion. Exclusion criteria were as follows: iNO therapy initiated > 48 h after birth, < 2 arterial blood gasses during iNO therapy, death < 12 h after birth, and congenital malformations that affected the ability to adequately oxygenate (e.g., cyanotic congenital heart defects and lung hypoplasia).

According to the protocol of the respective NICU, iNO therapy (PrinterNOx, CareFusion UK 232 Ltd., Kent, UK) is indicated for infants with an arterial oxygen tension (PaO2) < 13.3 kPa (100 mmHg) at a fraction of inspired oxygen (FiO2) 1.0. iNO is started at a dose of 20 parts per million (ppm). If oxygenation improves, iNO is gradually reduced at the discretion of the respective neonatologist in decrements of 50% until 5 ppm; then, further weaning is done at a rate of 1 ppm per hour. iNO is administered through conventional or high-frequency oscillation (HFO) ventilation, with peripheral oxygen saturation (SpO2) limits set at 92–98% for the cardiorespiratory monitor. PaO2 and arterial carbon dioxide tension (PaCO2) targets are 10–13 kPa (75–98 mmHg) and 5–6 kPa (38–45 mmHg), respectively. Automated oxygen control is not used in PPHN. In case of inadequate improvement, sildenafil can be considered. Extracorporeal membrane oxygenation (ECMO) is used as a rescue treatment.

All study data were retrieved from two digital medical record databases (Healthcare Information X-change (HIX), Chipsoft B.V., Amsterdam, The Netherlands, and PDMS; MetaVision iMDsoft, Leiden, The Netherlands). Baseline characteristics were determined, including maternal and patient demographics, pathology underlying PPHN (categorized according to the criteria summarized in Online Resource 1), mode of ventilation, hospital course including therapeutic interventions other than oxygen and iNO, and mortality. During iNO therapy, PaO2 in arterial blood gasses (RapidPoint500, Siemens Healthcare, UK) were collected, as well as one-per-minute data of pre- and postductal SpO2 in pulse oximetry and FiO2. The study period varies among infants due to differences in iNO therapy duration. If an infant was transferred for ECMO, the study period was ended.

The primary study outcomes were the incidence and duration of hyperoxemia during iNO therapy, defined as a PaO2 > 13 kPa (98 mmHg) or SpO2 > 98%. Secondary outcomes included the incidence and duration of hypoxemia, defined as a PaO2 < 10 kPa (75 mmHg) or SpO2 < 92%, and the concentration and duration of oxygen therapy. The definitions for hyperoxemia and hypoxemia are based on the oxygenation target ranges for PPHN used in our NICU. Severe hyperoxemia was defined as PaO2 > 30 kPa (225 mmHg) as it has been associated with an increased risk of death or adverse neurodevelopmental outcome in asphyxiated infants [18]. Severe hypoxemia was defined as a PaO2 < 6.7 kPa (50 mmHg) or SpO2 < 80% as per previous studies [19, 20]. The incidences of hyperoxemia and hypoxemia were derived from collected PaO2 measurements, while the durations of hyperoxemia and hypoxemia were derived from one-per-minute SpO2 measurements.

Due to the retrospective design of the study, a convenience sample was used. Data analysis was performed using SPSS 29 (IBM SPSS Statistics). Parameters were checked for normality through visual inspection of histograms. As this observational study intends to describe oxygenation parameters during PPHN therapy, outcomes have not been tested for statistical significance. Results are presented as n (% of N) for categorical variables and mean ± SD or median (IQR) for continuous parameters.

Results

Between 2011 and 2023, 280 late preterm and term infants received iNO therapy in the NICU, of which 99 were excluded according to predefined criteria: 18 received iNO therapy > 48 h after birth, 9 had < 2 arterial oxygen tension measurements, 4 deceased < 12 h after birth, and 68 suffered from a congenital malformation that impaired adequate oxygenation. Thus, a total of 181 infants were included in the study (Fig. 1), with GA 40 (38–40) weeks and birth weight of 3366 ± 677 g (Table 1). The median duration of iNO therapy (study period) was 2.4 (1.5–3.6) days, with a range of 0.3 to 13.8 days. The median number of PaO2 samples per infant was 13 (8–19).

Fig. 1
figure 1

Flow chart of inclusion and exclusion criteria. The numbers in bold represent the number of infants. iNO, inhaled nitric oxide; GA, gestational age; PaO2, arterial oxygen tension

Full size image
Table 1 Baseline characteristics of infants treated for PPHN
Full size table

Oxygen therapy

The median duration of oxygen therapy (FiO2 > 0.21) was 49 (25–83) h, constituting 97.6% (77.6–100%) of the iNO time. FiO2 was 1.0 in 156/181 (86%) infants for a duration of 2.2 (0.3–8.0) h per infant, representing 3.7% (0.6–13.5%) of the iNO time. The median FiO2 supply per infant was 0.43 (0.34–0.56).

Hyperoxemia

In 149/181 (82%) infants, at least one PaO2 measurement > 13 kPa was observed and severe hyperoxemia (PaO2 > 30 kPa) occurred in 46/181 (25%) infants (Fig. 2). The maximum PaO2 during iNO therapy was 21.6 (15.4–30.3) kPa, with a range of 5.6–59.7 kPa. In 179/181 (99%) infants, SpO2 > 98% was observed for 17.7% (8.2–35.6%) of the iNO time (Fig. 2).

Fig. 2
figure 2

Oxygenation parameters outside the target range. a Proportion of infants (N = 181) with at least one PaO2 outside the target range during iNO therapy. b Proportion of iNO time with a SpO2 outside the target range. Values are presented as median (IQR). PaO2, arterial oxygen tension; SpO2, oxygen saturation (pulse oximetry); iNO, inhaled nitric oxide

Full size image

Hypoxemia

Of the 181 infants, 160 (88%) infants had at least one PaO2 measurement < 10 kPa, and severe hypoxemia (PaO2 < 6.7 kPa) was observed in 81/181 (45%) infants (Fig. 2). The minimum PaO2 during iNO therapy was 6.9 (5.6–8.5) kPa, with a range of 1.4 and 22.5 kPa. In 169/181 (93%) infants, SpO2 < 92% was observed for 1.6% (0.5–4.3%) of the iNO time, and severe hypoxemia (SpO2 < 80%) occurred in 93/181 (51%) infants for 0.02% (0.00–0.22%) of the iNO time (Fig. 2).

Discussion

In this study, we demonstrated for the first time that hyperoxemia occurs often in late preterm and term infants treated for PPHN.

A previous retrospective study by Solberg et al. in mechanically ventilated infants, irrespective of PPHN, showed that 36% of term infants with measurements of arterial blood gasses had at least one hyperoxemic PaO2 within 48 h, compared to 82% of the infants during treatment for PPHN in the present study [21]. Infants with PPHN constitute a subgroup of ventilated infants with an increased risk of hyperoxemia due to their high oxygen requirement, as observed in our current study in which most infants received maximum FiO2 for a considerable period of time. In line with this, the median FiO2 was higher than in the study of Solberg et al. (0.43 vs. 0.21) [21].

Remarkedly however, despite the high concentrations of oxygen therapy, the occurrence of hypoxemia was similar to that of hyperoxemia, and severe hyperoxemia (PaO2 < 6.7 kPa) was observed in almost half of the infants. In the acute phase, infants with PPHN typically present with severe hypoxemia, and may remain cyanotic even when exposed to a high oxygen concentration [7]. After stabilization, the infants remain labile, with even small changes causing acute pulmonary vasoconstriction and rapid clinical decompensation [22]. As a result, neonatologists are cautious in weaning FiO2, which increases the risk of hyperoxemia. In line with this, we found longer time periods of SpO2 > 98% compared to SpO2 < 92% within this study, indicating that the neonatologists were attentive in preventing hypoxemia but were permissive of hyperoxemia. These findings are consistent with a previous survey among 492 neonatologists in the USA that evaluated oxygen management in infants with PPHN [23]. The survey demonstrated that a significant number of neonatologists preferred to target higher SpO2 and/or PaO2 to avoid hypoxemia but did not use an upper limit of SpO2 and PaO2 to prevent hyperoxemia. Furthermore, a small but notable portion of neonatologists (6%) opted to administer 100% oxygen, irrespective of the oxygenation parameters, until they were confident that the infant had clinically stabilized [23]. The permissive attitude towards hyperoxemia is also reflected by the current significant variations in practice with respect to oxygen titration strategies and the lack of evidence-based guidelines for oxygen weaning in the management of PPHN [23, 24]. Of the 492 neonatologists from the survey, 72% did not use specific oxygen titration guidelines [23]. In line with this, in our NICU, oxygen titration is left to the discretion of the respective neonatologist. Based on the clinical perceptions of the neonatologists in this study, despite the protocol’s recommendation to avoid hyperoxemia in PPHN, it is hypothesized that titration of oxygen is initiated once PaO2 values are > 13 kPa to reduce the risk of rebound PPHN. However, the clinical tolerance towards hyperoxemia is challengeable in infants with PPHN as hyperoxemia can promote pulmonary vasoconstriction and add to organ injury due to reoxygenation after an hypoxemic episode. Considering the high incidence of hyperoxemia in infants with PPHN, evaluation of the risks of hyperoxemia in infants with PPHN is warranted.

In addition, re-assessment of current strategies for oxygen therapy is needed to improve the time of having oxygenation parameters within the target range in infants with PPHN. The use of standardized protocols or innovations such as automated oxygen control [25] has the potential to improve oxygen titration, but more data is required to investigate its application in infants with PPHN.

An important consideration in optimizing oxygen titration in infants with PPHN is that SpO2 may not accurately reflect the arterial oxygen saturation due to underlying pathologies and therapies that compromise the peripheral circulation (e.g., asphyxia, inotropic agents) and shift the oxygen-hemoglobin dissociation curve (e.g., hypothermia) [26, 27]. An alternative titration parameter could be transcutaneous oxygen tension (TcPO2) [28], but its accuracy in infants with PPHN requires more research. Furthermore, the optimal SpO2 and PaO2 ranges in the management of PPHN are unknown. Based on translational studies, it is recommended to maintain SpO2 in the low to mid-90s and PaO2 between 55 and 80 mmHg (7.3–10.7 kPa) [8, 11]. However, clinical studies comparing different oxygenation targets are lacking.

This study has several limitations due to its retrospective design. In infants with PPHN, preductal PaO2 and SpO2 are typically higher compared to postductal values through right-to-left shunting over the ductus arteriosus [8]. Therefore, the probability of a hyperoxic PaO2 and SpO2 is higher in preductal than postductal measurements. This study was not able to differentiate between pre- and postductal PaO2 and SpO2, which may have led to an underestimation of the occurrence of hyperoxemia. However, this would only emphasize the need to improve oxygen weaning in PPHN.

Moreover, it was not possible to compare the occurrence and duration of hyperoxemia and hypoxemia based on PaO2 due to sampling bias: the decision to sample an arterial blood gas depends on the clinical condition of the infant. During a period of hypoxemia or hyperoxemia, the number of PaO2 samples will be increased to closely monitor the infant, resulting in an increase in the occurrence of hypoxemia or hyperoxemia. Similarly, a higher number of PaO2 samples result in a more adequate estimation of the duration of hyperoxemia and hypoxemia. However, we were able to estimate the duration of hyperoxemia and hypoxemia in terms of SpO2, which is the most commonly used parameter to titrate oxygen in PPHN [23].

Furthermore, considering the wide variety in the management of PPHN [23], in particular the variation in oxygenation target ranges, the external validity of the study may be limited by its single-center design. However, the majority of neonatologists are tolerant towards hyperoxemia [23], suggesting that the trend of high occurrence of hyperoxemia demonstrated in this study may be generalizable to a broader population. Lastly, although the data collection over 11 years significantly increased the number of infants included in the study, variations in clinical practice may have occurred, resulting in a historical bias.

In conclusion, while there is an increasing awareness of oxygen toxicity, the management of PPHN remains focused on reversing hypoxemia by means of high concentrations of oxygen. As a result, hyperoxemia occurs often during manual oxygen therapy in late preterm and term infants treated for PPHN. These findings warrant clinical awareness regarding the high risk for hyperoxemia in infants with PPHN. Further research is required to improve current strategies of oxygen therapy in these infants.