Introduction

Exogenous pulmonary surfactant (SRT) is most commonly administered to extremely preterm and very preterm infants. According to European data, approximately 50% of infants born at or before 32 weeks of gestation receive SRT.1 As a result, most scientific evidence focuses on this gestational age group. However, in clinical practice, SRT is also used in more mature newborns, including those born after 34 weeks of gestation. Yet, prospective studies focusing on SRT in this population remain limited, with only few studies available to date.2,3,4,5

In clinical practice—particularly among late preterm infants – respiratory distress syndrome (RDS) frequently coexists with other respiratory conditions, including e.g., air leaks, pulmonary hemorrhage, or congenital pneumonia.6,7,8 In term neonates, where RDS is uncommon, these conditions often occur independently and may be the primary cause of respiratory failure. Pulmonary comorbidities can influence the effectiveness of surfactant therapy; therefore, it is important to consider their potential impact on short-term outcomes, such as post-treatment improvements in systemic oxygenation. The epidemiology and impact of such conditions in late preterm infants are less well described than those in very preterm infants, and there are even fewer data on the effectiveness of SRT in decreasing oxygen requirements depending on respiratory comorbidities.

The primary aim of this study was to evaluate the response to SRT therapy in late preterm infants by assessing changes in systemic oxygenation, with a specific focus on the influence of pulmonary comorbidities. As a secondary objective, we investigated differences in the timing of SRT between infants who responded to treatment and those who did not.

Materials and methods

Study design and patient enrollment

This multicenter, prospective cohort study was conducted from February 2023 to August 2024 in 35 tertiary-level neonatal intensive care units (NICUs). Eligible patients were infants who were born at or after 34 weeks of gestation, had established RDS or were considered at risk of developing RDS, and required SRT administration according to the criteria outlined in the European Guidelines for RDS Management.1 In accordance with these guidelines, the diagnosis of RDS was made based on a comprehensive clinical assessment rather than relying on particular lung imaging or blood gas criteria. This encompassed signs of increased work of breathing shortly after birth and a need for respiratory support, including supplemental oxygen. In all enrolled infants, homogeneous lung recruitment was confirmed by lung imaging before surfactant administration. Infants at risk for RDS were defined as those with early signs of respiratory distress and the presence of recognized perinatal risk factors, such as lower gestational age within the late preterm range, absence of antenatal corticosteroid exposure, male sex, cesarean delivery, and maternal diabetes.1,9,10,11 Surfactant criteria were met when an oxygen requirement exceeded 30% to maintain adequate systemic oxygenation in infants receiving noninvasive ventilation with a positive airway pressure of at least 6 cm H2O.

Babies with major congenital abnormalities of the respiratory system or hemodynamically significant congenital cardiovascular disorders were excluded.

Study procedures and definitions

Surfactant administration was carried out in the NICU, and all infants received poractant alfa (Curosurf®, Chiesi Farmaceutici, Parma, Italy). Unless intubation was required for stabilization, respiratory support in the NICU consisted of noninvasive ventilation and supplemental oxygen, with the goal of maintaining target percutaneous oxygen saturation (SpO2) values within the range of 90–94%. Oxygen saturation was monitored using a pulse oximeter with the probe placed on the right hand (pre-ductal site). The mode of noninvasive ventilatory support, as well as the dose and technique of surfactant administration, was at the discretion of the attending physician. The choice of surfactant delivery method reflected routine practice at each study site in accordance with its established protocols, clinical experience, and physician preference, consistent with the noninterventional design of the study. Oxygenation parameters, including the fraction of inspired oxygen (FiO2) and corresponding SpO2, were recorded before SRT administration, as well as at 1 h, 6 h, 24 h, and 48 h after administration. The response to SRT was measured by changes in systemic oxygenation, specifically through variations in the SpO2/FiO2 (S/F ratio) over time.

The primary study endpoint, defined as a good response to surfactant, required that 6 h after SRT administration, infants achieve either a minimum 50% increase in the S/F ratio or an S/F ratio of 428 or higher (corresponding to SpO2 ≥ 90% in room air), and did not require escalation to mechanical ventilation (MV). In practical terms, a 50% increase in the S/F ratio typically results in a reduction in the oxygen requirement, e.g., from an FiO2 of 0.40 to approximately 0.27, while maintaining a blood oxygen saturation of 90%. For the good response to be considered sustained, improvements in systemic oxygenation had to be maintained at 24 and 48 h after SRT administration; otherwise, the response was classified as transient. In accordance with the product label, any additional SRT doses were administered no earlier than 12 h after the initial dose and therefore did not influence the primary endpoint assessment. The secondary endpoint was the change in oxygen requirements, evaluated by the reduction in oxygen exposure, measured as the area under the FiO₂ curve over time. Other study outcomes included the need for MV within 72 h in infants who received SRT without prior intubation, the duration of respiratory support, and the length of hospital stay. Additionally, we analyzed the relationship between the initial surfactant dose—typically targeted at 200 mg/kg per RDS guidelines but subject to minor differences due to vial-based rounding—and the oxygen requirements following surfactant administration.

Pulmonary comorbidities were documented and included as explanatory variables in the analysis of the response to SRT. Air leaks were diagnosed using lung ultrasound (LUS) or chest X-ray, following standard diagnostic protocols at each participating site. Meconium Aspiration Syndrome (MAS) was diagnosed in infants with a history of meconium-stained amniotic fluid, early-onset respiratory distress, and characteristic imaging findings. Congenital pneumonia was suspected in the absence of meconium exposure, but with respiratory symptoms, imaging suggestive of infection such as persistent infiltrates on chest X-rays or subpleural consolidations on LUS, and supporting evidence such as maternal risk factors (e.g., chorioamnionitis), laboratory findings (e.g., inflammatory markers, blood cultures), and clinical response to antibiotics, as well as the overall clinical course; not all criteria needed to be present in each case. Pulmonary hypertension was diagnosed with bedside functional echocardiography. Diagnostic criteria included the presence of a right-to-left shunt through the ductus arteriosus, estimation of pulmonary artery pressure via tricuspid regurgitation gradient, and assessment of pulmonary artery flow acceleration time.

Statistical analysis and sample size calculation

To evaluate the impact of pulmonary comorbidities on the SRT response, a linear mixed-effects model was employed, with baseline oxygenation (S/F ratio), gestational age, time after SRT administration, and the presence of comorbid pulmonary conditions included as fixed effects. Patient effects were treated as random variables, and p values were computed using the Satterthwaite approximation for degrees of freedom.

Oxygen exposure was quantified as the area under the FiO2 curve (AUC) over time, which was calculated with the trapezoidal rule. Group differences in the AUC were analyzed using t tests, with the Bonferroni correction for multiple comparisons of Box‒Cox-transformed data.

Spearman correlation analysis was performed to assess the relationship between the initial SRT dose and FiO2 level at various time points. Confidence intervals (95% CIs) were generated through bootstrapping with 1000 iterations.

A sample size calculation was performed to estimate the proportion of patients showing a good response to surfactant (SRT) with a desired precision. Assuming a 50% response rate (maximizing sample size for a binomial distribution), a total of 340 patients was required to achieve a two-sided 95% confidence interval with a margin of error of ±5%. To account for potential dropouts or missing data (estimated at up to 3%), the target enrollment was increased to 350 infants.

All analyses were conducted in R software (version 4.4.1; R Foundation for Statistical Computing, Vienna, Austria).

Results

Participants

A total of 350 infants were enrolled, with a median gestational age of 35.6 (IQR: 34.6–36.5) weeks. Among these infants, 64% were male, and the median birth weight was 2645 g (IQR: 2336–3000 g). The detailed characteristics of the study patients are presented in Table 1.

Table 1 Characteristics of the study participants, stratified by response to surfactant. A good response to SRT was defined as either a ≥ 50% increase in the S/F ratio or an S/F ratio ≥428 at 6 h postadministration, along with no need for escalation from non-invasive ventilation to mechanical ventilation.
Full size table

Among the 350 infants, 72 (21%) required primary intubation and subsequent MV. The remaining patients were initially managed with noninvasive respiratory support: 29% with continuous positive airway pressure (CPAP), 25% with bilevel CPAP, 22% with noninvasive nasal intermittent positive pressure ventilation, 2.3% with noninvasive oscillatory ventilation, and 0.3% with high-flow nasal cannula.

All infants underwent lung imaging prior to surfactant administration: 60% had only LUS, 21% had only chest X-ray, and 19% had both LUS and chest X-ray. Functional echocardiography was performed when clinically indicated, most often in the context of an unsatisfactory response to surfactant therapy, typically on the second day of life.

Surfactant was administered at a median of 8.2 h after birth, with 13% of the infants treated within 2 h, 28% treated between 2 and 6 h, and 28% treated more than 24 h after birth. The dosages and methods of administration are detailed in Table 1.

Primary study outcome and S/F changes

A good response to SRT was observed in 222 infants (64%). Among these SRT responders, 179 (81%) were classified as having a sustained response, whereas 43 (19%) were classified as having a transient response. Figure 1 illustrates the patient flow paths leading to the primary study outcome.

Fig. 1: Flowchart illustrating the distribution of ventilation modalities, surfactant administration methods, and treatment responses.
figure 1

One patient was discontinued from the study due to the investigator’s relocation and incomplete data entry. DR Delivery Room, MV Mechanical Ventilation, SRT Surfactant, LISA Less Invasive Surfactant Administration, INSURE Intubation, Surfactant, Extubation, SALSA Surfactant Administration through Laryngeal or Supraglottic Airways.

Full size image

In the logistic regression model evaluating the relationship between surfactant administration method and baseline FiO₂ on treatment response, neither LISA (OR 1.05; 95% CI: 0.55–2.00) nor INSURE (OR 1.40; 95% CI: 0.71–2.78) was associated with significantly different odds of a good response compared to the standard method (SRT plus mechanical ventilation). A higher baseline FiO₂, expressed as a z-score (per one standard deviation increase; SD = 0.18), was significantly associated with greater odds of a good response (OR 3.28; 95% CI: 2.25–5.01).

The baseline S/F ratio, prior to SRT administration, was 220 (IQR: 164–263), which corresponds to an SpO2 of 88% while receiving 40% oxygen. A significant increase in the S/F ratio was observed after SRT administration, increasing to 323 (IQR: 240–388) after 1 h, 369 (IQR: 280–430) after 6 h, 392 (IQR: 313–457) after 24 h, and 413 (IQR: 320–462) after 48 h (all p < 0.001 vs. baseline). Figure 2 shows the trajectory of the S/F ratio in surfactant responders and nonresponders over the first 48 h.

Fig. 2: Trajectories of S/F ratio in responders and nonresponders to surfactant therapy.
figure 2

Box-and-whisker plots show median, interquartile range, and range of S/F ratio over time. The horizontal lines mark the S/F threshold of 428 (SpO₂ = 90% at FiO₂ = 0.21). Changes over time were assessed by repeated measures ANOVA with Bonferroni post-test.

Full size image

According to the linear mixed-effects model, systemic oxygenation improvement was significantly influenced by the time since SRT administration, baseline oxygenation status, and the presence of certain comorbidities (Table 2).

Table 2 Linear mixed-effects model of S/F ratio changes after surfactant administration: impact of pulmonary comorbidities.
Full size table

Air leaks and persistent pulmonary hypertension were associated with a significant negative impact on oxygenation improvements. Oxygenation outcomes appeared less favorable in infants with MAS or pulmonary hemorrhage; however, these associations did not demonstrate a statistically significant effect. Congenital pneumonia had no measurable effect on S/F ratio changes following SRT administration.

Oxygen requirement

The oxygen requirement showed a significant decreasing trend over time, starting from a baseline FiO2 of 0.40 (IQR: 0.35–0.50) before surfactant administration and decreasing to 0.30 (IQR: 0.25–0.40) at 1 h, 0.26 (IQR: 0.22–0.35) at 6 h, 0.25 (IQR: 0.21–0.30) at 24 h, and 0.23 (IQR: 0.21–0.30) at 48 h postadministration (all p < 0.001 compared with baseline).

The infants with RDS and no comorbid conditions had the lowest oxygen requirement over time (AUC), as presented in Fig. 3.

Fig. 3: Changes in oxygen requirements according to pulmonary conditions.
figure 3

Areas under the FiO₂ curves over time (AUCs) are displayed for patients with various pulmonary comorbidities, with patients with RDS serving as the reference group. Dots represent medians, and error bars indicate interquartile ranges (IQRs). MAS Meconium Aspiration Syndrome, PPHN Persistent Pulmonary Hypertension of the Newborn.

Full size image

Impact of surfactant timing and dose on improvements in oxygenation

Compared with the nonresponder group, the responder group received SRT significantly earlier (median 7.2 vs. 15.6 h; p = 0.003) and had a significantly greater proportion of infants treated 8, 12, and 24 h after birth (all p < 0.01) (Fig. 4).

Fig. 4: Timing of surfactant administration in responders and non-responders.
figure 4

Cumulative frequency curves show the proportion of infants receiving surfactant over time after birth, with responders represented by a solid line and non-responders by a dashed line. The accompanying table summarizes cumulative percentages treated within each time interval.

Full size image

The FiO2 level after SRT administration demonstrated a statistically significant, modest correlation with the initial SRT dose across all studied time points (all p < 0.001), with the strongest correlation observed at 6 h postadministration (Spearman’s R = −0.25; 95% CI: −0.35 to −0.15).

Other study outcomes

Among the 249 newborns who continued to receive noninvasive respiratory support during SRT therapy, 35 required intubation and MV within the first 72 h of life. The proportion of these newborns was significantly greater among nonresponders (22 infants; 22.9%) than among responders (13 infants; 8.5%; p = 0.003).

The median duration of noninvasive support was 72 h (IQR: 48–120), with no significant difference between those with and without a good response to SRT (p = 0.547). Similarly, the median hospital stay (15 days, IQR: 11–23) did not differ significantly (p = 0.275). The median duration of MV was 66 h (IQR: 24–96) for infants with a good response to SRT and 72 h (IQR: 48–120) for those without a good response (p = 0.088). Additional study outcomes, stratified by surfactant response and ventilation status before SRT, are presented in Supplementary Tables S2a and S2b.

Discussion

In this multicenter prospective study, we evaluated the short-term efficacy of SRT therapy in late preterm infants, with a focus on the impact of pulmonary comorbidities. Our findings revealed significant improvements in systemic oxygenation after SRT therapy, although comorbid pulmonary conditions affected the treatment response.

Given that RDS is a key indication for SRT therapy, understanding its prevalence in this population is essential. Although the incidence of RDS in late preterm and term neonates is relatively low at 8–10%, the absolute number of affected infants in this population is considerable.12,13,14 A meta-analysis by Ramaswamy et al.15 revealed that 46% of late preterm and term infants with RDS receive SRT therapy, despite the limited supporting evidence, the absence of comprehensive risk‒benefit assessments, and the fact that SRT treatment in term infants remains off-label.16,17,18 However, other reports indicate lower treatment rates. For instance, a recent multicenter study from Turkey found that only about 8% of NICU-admitted late preterm infants (288 of 3327), or 15% of those presenting with respiratory distress (288 of 1866), received surfactant.5 Nevertheless, these figures still represent a considerable absolute number of patients, given the much larger population of late preterm infants compared to very and extremely preterm neonates.

Current indications for SRT therapy in late preterm neonates are generally extrapolated from guidelines for managing RDS in infants born at <32 weeks gestation. These guidelines recommend administering SRT at an FiO2 threshold of 0.30 for infants receiving noninvasive respiratory support,1 aiming to reduce both mortality and the need for MV. However, SRT therapy has not been shown to reduce the risk of requiring MV in late preterm infants.6,15

In our study, the primary outcome—a good response to SRT—was defined as an immediate improvement in systemic oxygenation, which was observed in 64% of the infants. However, the efficacy of SRT may vary on the basis of how a “good response” is defined. Since our definition was specifically developed for this study, comparisons are only valid with cohorts using the same criteria. In a previously published study of 964 preterm infants born at <32 weeks gestation,19 74% of the sample met this definition (Supplementary Table S1), suggesting that the treatment response differs depending on gestational age, likely due to differences in underlying pathophysiology.

In very preterm neonates, respiratory distress is driven by primary SRT deficiency, making SRT therapy highly effective as a direct replacement. However, in late preterm infants, primary SRT deficiency is less common or less severe, whereas additional pulmonary conditions are more common. Many neonatal respiratory disorders, such as pneumonia or MAS, can cause SRT inactivation and secondary SRT deficiency.20 This results in less favorable responses to SRT, as ongoing destructive processes, such as inflammation, further impair its function. Consistent with this, we observed the most significant improvements in infants with isolated RDS, without additional comorbidities.

An unexpected finding was the significantly longer median time from birth to SRT administration in late preterm infants than in our previous cohorts of very preterm infants. In the present study, the median time to SRT administration was 8.2 h—considerably longer than in our previous studies: 1.5 h (IQR: 0.75–3.68) in 236 infants born at less than 30 weeks of gestation,21 2.1 h (0.8–6.7) in 500 infants with a median gestational age of 30 weeks,22 and 2.6 h (1.6–6.7) in 986 infants born at less than 32 weeks of gestation.19 This suggests that a more conservative, wait-and-see approach was taken for late preterm infants in our study, which may partly explain their less pronounced improvements compared with very preterm infants. In contrast, another study on late preterm infants reported a shorter time to surfactant administration (median 4 h, IQR: 2–10).5 This difference may reflect variations in clinical practice, guideline adherence, or local thresholds for intervention.

Our findings further emphasize the importance of early treatment, as infants who responded positively to SRT received it significantly earlier than nonresponders. Earlier administration may also have an indirect impact on other outcomes, including lower rates of intubation and mechanical ventilation. The higher rate of mechanical ventilation among nonresponders likely reflects the clinical course following inadequate improvement after surfactant therapy. In our cohort, delayed administration and concomitant respiratory complications appeared to be the main contributors to poor response, making escalation to invasive support a predictable outcome in these cases.

While one could argue that nonresponders had more severe RDS, no significant differences were observed in perinatal characteristics, maternal history, or pulmonary morbidity between the groups.

In our cohort, persistent pulmonary hypertension of the newborn (PPHN) and air leaks were associated with reduced SRT effectiveness in improving systemic oxygenation—an outcome that could reasonably be expected. When the cause of the pulmonary problem lies outside surfactant’s target site (i.e., the alveoli) and involves vascular dysfunction in PPHN, SRT administration may not yield the desired effect. Lack of immediate response to SRT should prompt urgent hemodynamic assessment (echocardiography). Additionally, lung imaging, preferably ultrasound, should be mandatory before SRT administration to rule out pneumothorax. In our study cohort, lung imaging prior to SRT was performed in 100% of patients, though it is not standard practice everywhere.

While PPHN and air leaks were the most significant comorbidities affecting SRT effectiveness, other conditions, such as MAS and pulmonary hemorrhage, exerted a similar but somewhat lesser effect. Although the treatment response was not as favorable as in isolated RDS, a notable decrease in oxygen demand was still observed post-SRT in both comorbidities. Available evidence suggests that SRT may reduce the severity of MAS and the need for ECMO.23,24 However, to date, the effects of surfactant in pulmonary hemorrhage have not been evaluated in randomized controlled trials.25

Current RDS management guidelines recommend noninvasive ventilation combined with early rescue SRT administered via the less invasive surfactant administration (LISA) technique as the optimal approach.1 The LISA technique was the most frequently employed method for SRT administration in the patients, followed by equal proportions of patients treated using the INSURE method and during conventional MV. Surfactant administration via a laryngeal mask airway (LMA) using the SALSA method was performed in two neonates at a single center. While the primary limitation of SALSA in very preterm infants has been the lack of appropriately sized devices, it may be a more feasible option in late preterm infants. Gestational age and individual tolerance of different surfactant administration techniques (e.g., LISA vs. SALSA) may also influence outcomes, highlighting the need for further research. Interest in SALSA is growing due to its ease of use and potential to reduce the need for mechanical ventilation.26

In descriptive analyses, LISA recipients appeared more frequently among non-responders, compared to other surfactant administration techniques. However, this trend did not reach statistical significance and may reflect underlying differences in patient condition rather than the administration method itself. In the logistic regression model, which accounted for key confounder such as baseline FiO2, LISA was not independently associated with response outcome. These findings underscore the importance of adjusting for clinical context when evaluating the apparent effectiveness of administration techniques in observational data. The positive association between baseline FiO₂ and likelihood of improvement may seem counterintuitive, as higher FiO₂ typically indicates more severe initial respiratory compromise. Nevertheless, this finding aligns with the definition of response, which reflects relative improvement in oxygenation rather than absolute post-treatment status. Infants with higher initial FiO₂ had greater potential for measurable improvement following surfactant therapy, a pattern consistent with a “ceiling effect.”

A key strength of our study is that it provides data on a less-documented patient population, collected in a real-world clinical setting using a prospective design. Notably, it is one of the few prospective studies investigating SRT efficacy in late preterm neonates.

However, certain limitations should be acknowledged. Our study primarily assessed short-term outcomes on the basis of oxygenation measures. While we collected data on in-hospital mortality and typical complications of prematurity, the sample size was insufficient to analyze the impact of pulmonary comorbidities on these outcomes.

Additionally, our definition of “good response” was selected a priori based on clinical relevance and feasibility. However, this was a discretional definition, not grounded in established pathophysiologic thresholds, and should be interpreted accordingly until more definitive criteria are validated in future research. This definition limits direct comparisons with other studies of late preterm infants, which assessed oxygenation changes after SRT therapy but used different metrics. These metrics include arterial oxygen tension (PaO2),27 the arterial-to-alveolar oxygen ratio (Pa/PAO₂),15,28,29,30 the PaO₂/FiO₂ ratio,31 and the oxygenation index.28,30,31

Although noninvasive oxygenation indices such as the oxygen saturation index (OSI) or oxygenation index (OI) may offer greater precision than the S/F ratio used in our study, their calculation requires mean airway pressure (MAP), which was not collected. Furthermore, although the participating centers followed the European RDS guidelines, the observational nature of the study introduced variability in intervention schemes across sites. However, certain aspects—such as noninvasive ventilation settings—were likely quite uniform. While we did not collect detailed data on airway pressures or nasal interfaces, national data suggest that CPAP pressures in Polish NICUs are relatively consistent, typically ranging between 5 and 6 cm H₂O; nasal masks or prongs are the most commonly used interfaces, with RAM cannulas used infrequently (< 8%).22 Nonetheless, the absence of site-specific ventilation data remains a limitation of the study.

Finally, we did not capture details of pneumothorax management (drainage vs. conservative). In our clinical practice, a minor pneumothorax is viewed as a relative contraindication to SRT, though most neonatologists avoid treatment in this setting, whereas a clinically significant pneumothorax is drained and SRT is not given. This should be considered when interpreting the association between air leaks and oxygenation outcomes.

In summary, our results indicate that SRT therapy effectively enhances oxygenation in late preterm infants; however, its efficacy is moderated by the presence of pulmonary comorbidities and the timing of administration. Oxygenation improvements were most significantly hindered by air leaks and persistent pulmonary hypertension, while MAS and pulmonary hemorrhage had a minor impact. In contrast, congenital pneumonia did not show a significant effect. These findings highlight the need for timely intervention and consideration of underlying conditions when managing respiratory distress in late preterm infants. The ongoing SURFON Trial32 aims to address uncertainties by investigating whether early SRT use in late preterm and early-term infants with RDS can reduce disease severity and shorten hospital stays.