CAN WE BALANCE EARLY EXOGENOUS SURFACTANT THERAPY AND NON-INVASIVE RESPIRATORY SUPPORT TO OPTIMIZE OUTCOMES IN EXTREMELY PRETERM INFANTS? A NUANCED REVIEW OF THE CURRENT LITERATURE

Abstract

Therapeutic advances have significantly improved the survival of premature infants. However, the high burden of bronchopulmonary dysplasia (BPD) persists. Aiming at the prevention of neonatal lung injury, continuous positive airway pressure (CPAP) and non-invasive ventilation (NIV) strategies have replaced mechanical ventilation for early respiratory support and treatment of respiratory distress syndrome. Multiple RCTs have demonstrated that broad application of CPAP/NIV decreases exposure to mechanical ventilation and reduces rates of BPD. Here, we explore why this treatment effect is not larger. We discuss that today’s NICU population evolving from the premature to the extremely premature infant demands better targeted therapy and indicate how early and accurate identification of preterm infants likely to fail CPAP/NIV could increase the treatment effect and minimize the potential harm of delaying exogenous surfactant therapy in these infants. Finally, we argue that less invasive modes of surfactant administration may represent both a pragmatic and beneficial approach in combining CPAP/NIV and early surfactant. Beneficial treatment effects might be higher than reported in the literature when targeting this approach to those preterm infants suffering from respiratory failure primarily due to surfactant deficiency. Considering ongoing limitations of current approaches and focusing both on prospects and potential harm of modified strategies, this commentary ultimately addresses the need and the challenge to prove that pushing early CPAP/NIV and strategies of early and less invasive surfactant application prevents lung injury in the long term.

INTRODUCTION

Non-invasive respiratory support is widely recommended to stabilize preterm infants at high-risk of developing long-term pulmonary morbidity.^1–3 This approach, based on high-quality evidence, has been widely adopted in practice.^{2 4–10} However, increased use of non-invasive support has not decreased rates of bronchopulmonary dysplasia (BPD) in the highest-risk neonates.^{4 5 8–11} It has been speculated that the increased survival of these infants has contributed to this finding.^{4 11} Alternatively, it has been argued that a single intervention is not sufficient to significantly decrease the risk of BPD – given its multifactorial pathogenesis.¹²

As we approach 15 years since the landmark randomized controlled trials (RCTs) evaluating CPAP versus intubation and prophylactic surfactant were published,^13–16 where do we stand? In practice, we have not seen the decrease in BPD collectively demonstrated in these trials – why? For this targeted review, we focus on infants at highest risk of long-term pulmonary morbidity, specifically those born <28 weeks’ gestational age (GA), as studied in these trials. The impact of non-invasive support should be evaluated in this population, as these infants have the highest pulmonary morbidity associated with mechanical ventilation (MV) exposure.^{4 11 17} We will discuss the limitations of a broad application of continuous positive airway pressure (CPAP)/non-invasive ventilation (NIV) in this population, and how modifying the risk profile associated with exogenous surfactant administration may allow us to improve outcomes even when we are unable to accurately diagnose surfactant deficiency as the primary driver of respiratory failure.

THE PROMISE OF NON-INVASIVE SUPPORT: A SIGNAL BLUNTED BY DEFINING THE TARGET SIMPLY BY GESTATIONAL AGE?

Several studies performed in the mid- to late noughties sought to answer the following question: in high-risk preterm infants, does broad application of non-invasive respiratory support, compared to intubation for prophylactic or early surfactant treatment, decrease pulmonary morbidity? A number of meta-analyses have been performed using data from these trials.^18–20 Four of these trials enrolled appropriately high-risk neonates (<30 weeks).^13–16 In each of these, initial respiratory support with CPAP was compared to interventions that necessitated at least some exposure to MV. The Vermont Oxford DRM Study Group trial compared CPAP to two other treatment arms: “INSURE” (INtubate, SURfactant, Extubate) and prophylactic surfactant followed by MV.¹⁶ SUPPORT directly compared routine CPAP to routine prophylactic surfactant followed by MV.¹⁴ The COIN trial compared CPAP to intubation and ventilation, but did not stipulate that those randomized to intubation receive surfactant.¹³ Finally, the CURPAP study randomized infants to CPAP alone versus INSURE. Together, these trials demonstrate a protective effect of avoiding MV through preferential CPAP on the combined outcome of death or BPD, with a number needed to treat (NNT) ranging from 17.7 to 35, depending on which trials are included in the meta-analyses.^{21 22} Given that ventilator-induced lung injury is a strong risk factor for BPD, the treatment effect achieved by avoiding MV in subjects randomized to CPAP is somewhat disappointing. However, it is critical to realize that across these trials, ~50% of the neonates randomized to CPAP nonetheless required intubation and MV and exogenous surfactant therapy in the first 5–7 days (Table 1).^13–16 This need for early intubation should be considered separately from rates of intubation over the duration of hospitalization, as these rates are both higher and have different underlying causes.

TABLE 1.

Rates of CPAP failure (first 5–7 days), surfactant administration and in-hospital intubation in the groups randomized to CPAP/non-invasive respiratory support in large RCTs of adequately high-risk neonates

Trial	Year	GA	ACS (any)	CPAP Failure (%) (when assessed)		Surfactant administration in CPAP group (%)	Intubation rates in CPAP group (%; entire hospitalization)
				Day 5	Day 7
Sandri	2010	25 0/7–28 6/7	>95%	48.5%*		48.5%	NO INFO
COIN	2008	25 0/7–28 6/7	94%	46%		38%	NO INFO
SUPPORT	2010	24 0/7–27 6/7	>95%		45%#	67%	83.1% (Text, next to Table 1)
Dunn	2011	26 0/7 – 29 6/7	>98%		45.1%	45%	52.3%

^*trial reports 33% were intubated, but 48.5% randomized to CPAP needed early surfactant administered via endotracheal tube

^#trial reports “survival without need for high-frequency or conventional ventilation at 7 days,” thus rate of intubation may be higher if subject was extubated by day 7

Knowing that the treatment effect is likely blunted by this high failure rate, interventions aimed at minimizing CPAP failure have been identified and implementation of these approaches have demonstrated the protective effect of non-invasive support.^{21 23–25} However, given the relatively consistent signal for high rates of failure in this population, it is appropriate to ask whether broad application of non-invasive support based on GA and/or birth weight (BW) should be re-examined, and whether a more nuanced approach is justified. Consider the cohort of babies at high risk of long-term pulmonary morbidity (<28 weeks’ GA) admitted to our NICUs today: With the goal of avoiding MV exposure, it is reasonable to uniformly apply CPAP as an initial form of respiratory support. However, by one week of age, published data demonstrate that rather than a homogenous cohort solely defined by GA and risk of lung injury, we have distinct cohorts: one cohort stable on CPAP/NIV, and one cohort that required invasive support. Is it possible, shortly after birth, to use more than GA to risk-assess and appropriately target respiratory support? Furthermore, is it possible that the true protective effect in the cohort destined to succeed on non-invasive support (~50%) is blunted by a lack of effect (or even harm) imposed on those infants destined to fail (the remaining ~50%)? If that is true, we should aggressively seek new ways to understand this treatment heterogeneity to more accurately risk-assess individual infants and more appropriately target early respiratory support.

First, it is important to ask whether there is potential harm in exposing babies to NIV – if they are, in fact, destined to fail and ultimately need invasive support. As discussed above, broad populations of preterm infants randomized to receive CPAP/NIV have better average outcomes than those randomized to exogenous surfactant followed by invasive support. However, this perspective does not answer the question of whether the protective signal of non-invasive support is enhanced in – or even fully limited to –those that don’t fail. Furthermore, it does not answer whether poor outcomes in those randomized to receive CPAP/NIV and ultimately fail contribute to blunting the observed treatment effect. While outcomes of those randomized to CPAP/NIV are modified by early failure have not been reported, observational data are available to help answer this question.

Dargaville and colleagues reported the outcome of ~12,000 preterm infants (25–32 weeks) initially managed on CPAP.²⁶ Of the ~2,000 neonates born at 25–28 weeks’ GA, 43% failed CPAP and required intubation within 72 hours of life. This rate of failure is consistent with numbers reported in RCTs. Rates of pneumothorax, BPD, Grade III/IV IVH and mortality were higher in the group failing non-invasive support compared to the group that did not. Adjusted odds ratios remained significant for BPD, death or BPD, and major morbidity. Observational data similarly demonstrate worse outcomes in those failing CPAP/NIV,^{23 27 28} with outcomes following failure similar to those initially receiving invasive support.^{23 27}

How do we interpret these data? It could be argued that the higher prevalence of known risk factors for various neonatal morbidities (lower GA/BW, less ACS exposures) in the cohort of high-risk neonates destined to fail non-invasive support contributes to their worse outcomes – independent of failing. But in addition, those infants destined to fail and yet inadequately supported on CPAP/NIV might experience additional harm from the failed effort. In consequence, it is reasonable to ask whether we could improve outcomes by preventing exposure to “non-invasive support failure” – and whether we could ultimately identify and separate these cohorts shortly after birth to better target therapy and maximize the beneficial effect of CPAP/NIV. In turn, if the primary cause of failure is treatable, early identification could avoid non-invasive support failure and the potentially associated harm.

The various reasons underlying non-invasive support failure, among others, include the inability to effectively deliver non-invasive pressure and to establish functional residual capacity (FRC) – due to small anatomical structures, limited lung volume and chest wall instability, interface constraints, apnea or impaired respiratory drive, and a variable degree of lung immaturity necessitating more support than provided by non-invasive delivery. However, availabe data support the hypothesis that the primary driver of CPAP/NIV failure is surfactant deficiency. Although immature surfactant quality and quantity is typical of extremely preterm infants, some have sufficient function to altogether forego exogenous replacement, while others have a surfactant deficiency that leads to CPAP/NIV failure. Both observational^{27 29} and RCT data^{13 15} demonstrate that CPAP failure occurs early (within first ~8 hours in life) and is often accompanied by an increasing oxygen requirement, consistent with surfactant deficiency. In the Dargaville study, 97% of those 25–28 wks’ GA failing early NIV received surfactant shortly following intubation.²⁶

Given this, the ability to diagnose clinically relevant thresholds of surfactant deficiency early and accurately would allow the clinician to better define these cohorts of interest. However, in 2022, as discussed by Bancalari and Jobe a decade ago, “there is no clinically acceptable way to securely diagnose surfactant deficiency“³⁰ – as diagnostic tests, neither chest radiograph nor oxygen requirement, peform well enough to guide therapy.^{31 32} For example, FiO₂ requirement has been suggested as an objective marker to which treatment with exogenous surfacant therapy has been anchored.^{2 3} It is undeniable that increasing FiO₂ requirement is consistent with the diagnosis of RDS. However, in babies whose respiratory insufficiency is multifactorial, FiO₂ requirement performs poorly as a diagnostic marker as evaluated by two separate studies in high-risk neonates (<28 weeks).^{27 29} Fuchs et al. investigated performance of FiO₂ requirement at NICU admission, while Dargaville evaluated it within the first two hours of life (Figure). Similar to RCT data, overall rates of CPAP/NIV support failure were ~50% in these high-risk cohorts. In consequence, it might be instructive to consider how FiO₂ requirement performs in the two designated cohorts of preterm infants: those that succeed, and those that fail non-invasive support. Broad application of CPAP/NIV leaves 50% of all high-risk neonates on respiratory support that will subsequetly fail and need treatment optimization. Data from Fuchs et al demonstrate that using an admission FiO₂ of 0.3 leaves 25% needing optimization, primarily in the NIV failure cohort. Finally, making the decision at 2h of life increases the number treated with exogenous surfactant, but potentially leads to overtreatment, with 35% now needing optimization. These data demonstrate that exogenous surfactant treatment cutoffs that rely primarily on FiO₂ only marginally improve the clinician’s ability to cohort high-risk preterm infants into appropriate treatment groups (Figure). Other objective measures of surfactant deficiency and pulmonary dysfunction are under active investigation, but not yet ready for routine clinical implementation.³²

Figure: How the inability to diagnose surfactant deficiency limits our ability to target CPAP/NIV and exogenous surfactant therapy in high-risk preterm infants.

For a high-risk (<28 wks GA) neonate on non-invasive support, a clinician must decide when to the neonate has “failed” and would benefit from exogenous surfactant. Broad application of CPAP/NIV at birth ensures ~50% of the cohort will maximally benefit from CPAP/NIV, but exposes ~50% of this cohort to the potential harms of CPAP failure. An FiO₂ requirement of 0.3 at NICU admission provides 60% sensitivity and 90% specificity in predicting CPAP failure, leaving 25% of the cohort still needing treatment optimization (20% still exposed to CPAP failure, 5% exposed to exogenous surfactant with questionable benefit). An FiO₂ requirement of 0.3 at 2 hours of life provides 80% sensitivity and 50% specificity in predicting CPAP failure and leaves 35% of the cohort still needing treatment optimization (10% still exposed to CPAP failure, 25% exposed to exogenous surfactant with questionable benefit). In the absence of objective criteria to diagnose surfactant deficiency, using risk factors (GA, BW) and proxies (FiO₂ requirement, CXR) we can expect that the true benefit of appropriate treatment strategies will be attenuated.

Has the inability to target exogenous surfactant by diagnosing RDS swung the pendulum towards broad implementation of less invasive administration of surfactant without exposure to MV?

Broad application of CPAP/NIV can result in delaying targeted surfactant therapy. This risk is acceptable if exogenous surfactant treatment brings with it subsequent and necessary exposure to MV. However, it is important to consider whether we can improve outcomes by either early and targeted exogenous surfactant to those with surfactant deficiency, or, alternatively, delivering exogenous surfactant more broadly but with a lower risk profile than that associated with exposure to MV. Given the difficulty with early diagnosis of surfactant deficiency in this population, more work has been done investigating methods to decrease the risks associated with exogenous surfactant therapy.^{33 34} By lowering the risk profile of the intervention required to deliver exogenous surfactant, it stands to reason that these approaches could be applied earlier and more broadly without increasing exposure to harm. It is important to consider whether these approaches improve outcomes when compared to a broad application of NI support and delayed surfactant therapy in those that ultimately fail.

Less Invasive Surfactant Administration (LISA)

Among different modes, less invasive surfactant administration (LISA) constitutes the most promising and most intensively studied approach today.^34–38 Described by Verder et al in Denmark in the early 1990s and rediscovered in Germany ~10 years later,^{39 40} the technique has gained popularity across neonatal units in Germany, Europe and worldwide:^{37 41–43} A thin catheter or feeding tube is inserted into the trachea of a spontaneously breathing preterm infant on CPAP and surfactant is administered.^{37 40 43} A very similar method developed in Australia (minimally invasive surfactant therapy, MIST) uses a more rigid catheter positioned without Magill forceps.⁴⁴ It is important to note that while these interventions are “less” or “minimally” invasive, they are still complex procedures performed in a high-risk population. These interventions still involve direct visualization of the airway via laryngoscopy and demand practitioner and team experience. Placement of catheters may involve premedication with analgesics or paralytics. Alternatively, the use of comforting measures (positioning, facilitated tucking, sucrose) are commonly used instead of sedation/paralysis in GNN centers.³⁷

Several RCTs have proven feasibility, safety and efficacy of the maneuver. A Cochrane review and a recent network meta-analysis including 16 RCTs and 20 observational studies (in total, 13,234 preterm neonates with RDS) provide evidence that LISA/MIST as compared to surfactant replacement via ETT and INSURE, respectively, lowers rates of BPD, death or BPD, the risk of MV in the first 72h, and the incidence of major complications other than BPD (severe intraventricular hemorrhage, periventricular leukomalacia and necrotizing enterocolitis).^{34 42} While these are promising findings, it is important to note that many of these trials have not exclusively studied infants at highest-risk of CPAP/NIV failure and lung injury. It is in this high-risk population that broad application of early surfactant without associated MV exposure would be expected to be most beneficial, and it is important to consider this subgroup of infants separately.³³ It is also reasonable to consider that targeting this intervention to those with surfactant deficiency would maximize the treatment effect, and minimize exposure to the known short-term risks associated with the procedure including laryngoscopy, bradycardia and hypoxemia. The strategy of CPAP-assisted spontaneous breathing and early surfactant therapy via LISA – often within the first hour of life – is widespread in German Neonatal Network (GNN) centers and includes the tiniest babies.^{37 45 46} In ~60% of infants, this early treatment remains the only surfactant dose. A second dose is required in ~27%, a third in 8.5% and a fourth dose or more in ~4% of infants.⁴⁶ The first RCT evaluating benefits and harms of prophylactic LISA treatment in very immature preterm infants as well as FEV₁ values at 5 years of age is underway (PRO-LISA trial, GNN).

Two RCTs comparing LISA/MIST to continuation of CPAP in neonates ≤28 weeks’ GA, the Avoidance of mechanical ventilation (AMV) trial and the OPTIMIST trial, provide proof that LISA/MIST reduces the need for MV within 72h of birth (AMV: n=220 infants, 26–28 weeks’ GA; RR 0.61 [95% CI 0.42–0.88]; OPTIMIST: n=485 infants, 25–28 weeks’ GA; RR 0.50 [95% CI 0.40–0.64]).^{35 38} In addition, the OPTIMIST trial found a reduced risk of pneumothorax requiring drainage, incidence of BPD at 36 weeks’ GA, and duration of respiratory support in infants allocated to MIST.³⁸ It must be noted that OPTIMIST was closed prior to full enrollment due to the COVID-19 pandemic. The largest cohort studies comparing LISA with standard surfactant replacement therapy, published by the GNN, cover > 2,500 preterm infants ≤ 28 weeks’ GA managed with LISA.^{41 46} LISA was associated with reduced need for MV within the first 72 hours, reduced risk of mortality and BPD as well as reduced risk of pneumothorax and most secondary outcomes.^{41 46} Data from these trials may indicate that these “lower-risk” interventions can be broadly applied to improve outcomes. Yet, potential risks associated with the intervention and/or indiscriminate use need to be taken into consideration. While mortality in the OPTIMST trial was not different among study groups in the cohort of 25–28 week GA neonates, in the subgroup of 25–26 week GA infants, mortality was 8% in the control versus 17.8% in the MIST group.⁴⁷ Although these differences did not reach statistical significance, they demand caution. Of note, in contrast, the recently published investigation of the GNN in 6,542 preterm infants born at 22 0/6 to 26 6/7 weeks GA documented lower risks of all-cause death, and BPD or death in LISA treated infants in experienced centers.⁴⁶ However, a previous GNN study found increased rates of focal intestinal perforation (FIP) in the LISA group – primarily in the most immature babies born at 22–24 weeks.⁴¹ The Nonintubated Surfactant Application (NINSAPP) trial and the OPTIMIST trial did not confirm a potential association with FIP.^{36 38} In summary, more trials and comparative studies including LISA/MIST and INSURE are needed to evaluate benefits and potential harms, especially in the tiniest babies.

Laryngeal Mask Airway and Pharyngeal Instillation

Alternative modes of surfactant administration include laryngeal mask airway (LMA), pharyngeal instillation and nebulization.^{33 34} LMA and pharyngeal instillation constitute attempts to apply exogenous surfactant less invasively than laryngoscopy and/or endotracheal intubation via deposition to the pharynx or administration using supraglottic airway devices.⁴⁸ LMA, though, still requires pharyngeal positioning – with attendant risks. In the early 1970s, Enhorning and Robertson demonstrated that tracheal and pharyngeal deposition prior to the first breath improved lung expansion in preterm rabbits.⁴⁹ 15 years later, pharyngeal instillation of surfactant was revisited in preterm infants,⁵⁰ followed by a multicenter pilot study comparing prophylactic pharyngeal deposition with routine assistance⁵¹ and a feasibility study of intrapartum pharyngeal instillation⁵² – each study enrolling a small cohort of preterm infants. Data from RCTs on pharyngeal instillation are lacking. A randomized trial in preterm infants <29 weeks’ GA is ongoing.⁵³ Feasibility and safety of LMA was first described in a case report in 2004,⁵⁴ followed by small observational studies and five RCTs including 307 neonates in total.³³ A meta-analysis covering these RCTs found a reduced need for MV in LMA cohorts.⁴⁸ Moreover, LMA seemed superior to CPAP but inferior to INSURE with respect to oxygen requirement in the first 6h.⁴⁸ However, study cohorts were mainly comprised of moderate and late preterm infants, and study designs were heterogenous.⁴⁸ Efficacy and safety in extremely preterm infants remains to be determined.

Aerosolization of Surfactant

Constituting a truly non-invasive approach of exogenous surfactant administration, nebulization – or aerosolization – altogether avoids airway manipulation.³³ Following a first report of aerosolization of synthetic dipalmitoyl-phosphatidylcholine into the incubators of 11 preterm infants with RDS by Robillard et al. 60 years ago,⁵⁵ implementation of CPAP/NIV in neonatal care has reawakened and driven efforts in animal and in vitro models to advance surfactant nebulization and establish appropriate pulmonary deposition.⁵⁶ The most recently published trials compared nebulized surfactant to CPAP alone,^{57 58} and intubation and standard surfactant administration, respectively,⁵⁹ in neonates with RDS born at 28 0/7 – 32 6/7,⁵⁸ 29 0/7 – 33 6/7⁵⁷ and 23 0/7 – 41 6/7 weeks’ GA (mean GA 33.2±3.2 and 33.1±3.1).⁵⁹ Data confirmed feasibility and suggested safety of the intervention.^{58 59} However, nebulization of surfactant either was not superior to CPAP alone,⁵⁸ or the benefit of nebulization was limited to preterm infants born >31 weeks.^{57 59} In summary, aerosolization remains an attractive route but adequately powered and well-designed trials in high-risk neonates are needed to evaluate efficacy and safety in today’s increasingly premature NICU population.

QUESTIONS FOR 2023 AND BEYOND

As long as we cannot accurately diagnose surfactant deficiency in these high-risk neonates, clinicians will be unable to harness the true benefit of CPAP/NIV and early exogenous surfactant therapy as two powerful therapies. For now, besides further improvements in early NIV application,^{23 25} the focus might remain on broad application of early exogenous surfactant using methods that do not necessitate standard intubation and MV exposure. Yet, we are left comparing these approaches to CPAP/NIV alone. Whether such broad application of early exogenous surfactant would perform better than targeted diagnosis-based surfactant administration remains unknown. More trials and long-term follow-up are needed to evaluate side effects of alternative modes of surfactant administration, especially LISA/MIST, and to exclude potential harm especially in the most immature preterm infants.

Other issues remain unanswered, including the question if biochemical markers of surfactant production could be used to target therapy. This remains an area of active investigation.^60–65 It is unknown whether any of this discussion is applicable to the periviable NICU population and future studies must consider the unique challenges facing this population. It is reasonable to hypothesize that the etiology of NI support failure becomes increasingly more complex and multifactorial as GA decreases. The observation that in RCT the beneficial treatment effect of LISA/MIST falls with decreasing GA. It is likely that adjunctive approaches will be needed to achieve optimal outcomes in these babies. Whether exposing these high-risk neonates to the potential harms associated with these interventions in the absence of clear benefit is detrimental necessitates caution.

In addition, while BPD is routinely assessed as an outcome measure in high-risk infants, it remains unclear whether this diagnosis as assessed at 36 weeks’ GA⁶⁶ accurately predicts long-term pulmonary morbidity. Today’s BPD in those born at very early GA is no longer a disease primarily caused by oxygen, pressure and time as proposed by Philip in 1975.⁶⁷ Increasingly, BPD is a broad term encompassing heterogenous lung disease phenotypes and subsequent chronic morbidities evolving from a complex interplay of physiologically and immunologically immature lungs with prenatal and postnatal inflammation, growth restriction, oxygen toxicity and exposure to shear stress, barotrauma and volutrauma – culminating in alveolar and pulmonary vascular maldevelopment. Given the multifactorial etiology of BPD, assessments of the impact of early respiratory support may need to be more precise. Furthermore, one might expect some degree of pulmonary dysfunction in all periviable infants, and interventions may be best aimed at minimizing rather than eliminating this dysfunction.

These facts highlight the relevant shortcomings of BPD as a surrogate for long-term pulmonary function and overall morbidity. Today’s widely used definition⁶⁶ has not demonstrated appropriate prediction of long-term lung function in childhood and, in particular, adulthood. Furthermore, evading a diagnosis of BPD at 36 weeks’ does not ensure lack of future pulmonary morbidity.⁶⁸ The importance of severity of BPD as a predictive marker is increasingly recognized.⁶⁹ It may be time to critically re-evaluate how we design our trials and judge the impact of our interventions and develop new and more predictive markers of long-term pulmonary health. Without this, we may underestimate the benefit of interventions such as LISA/MIST that clearly decrease MV exposure, and yet have a marginal impact on the incidence of BPD.

In conclusion, well-designed RCTs have proven the beneficial effect of early exogenous surfactant, as well as non-invasive respiratory support. Further advancement of these interventions, and improvement of outcomes of highest-risk neonates, will depend on minimizing the risk of MV exposure associated with surfactant administration through both the development of less invasive delivery measures, and progress in diagnosing surfactant deficiency earlier and more accurately.