The Effect of Extended Continuous Positive Airway Pressure on Changes in Lung Volumes in Stable Premature Infants: A Randomized Controlled Trial

Abstract

Objective

To compare changes in lung volumes, as measured by functional residual capacity (FRC), through to discharge in stable infants randomized to 2 weeks of extended continuous positive airway pressure CPAP (eCPAP) vs CPAP discontinuation (dCPAP).

Study design

Infants born at ≤32 weeks of gestation requiring ≥24 hours of CPAP were randomized to 2 weeks of eCPAP vs dCPAP when meeting CPAP stability criteria. FRC was measured with the nitrogen washout technique. Infants were stratified by gestational age (<28 and ≥ 28 weeks) and twin gestation. A linear mixed-effects model was used to evaluate the change in FRC between the 2 groups. Data were analyzed blinded to treatment group allocation.

Results

Fifty infants were randomized with 6 excluded, for a total of 44 infants. Baseline characteristics were similar in the 2 groups. The infants randomized to eCPAP vs dCPAP had a greater increase in FRC from randomization through 2 weeks (12.6 mL vs 6.4 mL; adjusted 95% CI, 0.78–13.47; P = .03) and from randomization through discharge (27.2 mL vs 17.1 mL; adjusted 95% CI, 2.61–17.59; P = .01).

Conclusions

Premature infants randomized to eCPAP had a significantly greater increase in FRC through discharge compared with those randomized to dCPAP. An increased change in FRC may lead to improved respiratory health.

Trial registration

ClinicalTrials.gov: NCT02249143.

Nasal continuous positive airway pressure (CPAP) is the standard of care for the initial support of premature infants with respiratory distress syndrome to establish and maintain their functional residual capacity (FRC) while minimizing potential volutrauma from mechanical ventilation.^1–3 Infants who require intubation are usually extubated to CPAP to maintain alveolar recruitment and prevent atelectasis. A meta-analysis demonstrated that premature infants randomized to CPAP vs intubation at delivery have decreased bronchopulmonary dysplasia at 36 weeks of postmenstrual age (PMA).⁴ In addition to acute physiological benefits, studies in animal models have shown an improvement in lung growth after CPAP,⁵ as well as decreased inflammatory cytokines,⁶ decreased airway reactivity,⁶ and increased angiogenic growth factors.⁷ The beneficial effects of CPAP in preterm infants may stem from avoidance of mechanical ventilation and promotion of lung growth through mechanical stretch.

Although early application of CPAP after delivery is currently standard management, there is no consensus as to when or how to wean stable premature infants off CPAP.⁸ Prolonged CPAP also has potential adverse side effects, including nasal breakdown, delayed oral feeds, prolonged hospitalization, decreased parental bonding, and developmental therapies.⁸ In addition, parents often urge neonatal intensive care unit (NICU) providers to discontinue CPAP as soon as possible to keep the infant’s face un-obstructed or to facilitate oral feeding. Trials to evaluate CPAP weaning have done so when specific CPAP respiratory stability criteria were met.^9,10 The primary aim of the present study was to quantify and compare changes in lung volumes (as measured by FRC) in stable, convalescing preterm infants receiving CPAP and randomized to an additional 2 weeks of CPAP (extended CPAP [eCPAP]) vs CPAP discontinuation (dCPAP; usual care). We hypothesized that compared with infants randomized to dCPAP, infants randomized to eCPAP would demonstrate a greater increase in FRC at end of the 2-week study period and at hospital discharge.

Secondary aims of the study were to determine whether eCPAP was associated with improvements in other pulmonary function test (PFT) measures, including FRC normalized per kg of body weight, passive respiratory compliance (Crs), Crs normalized per kg, passive respiratory resistance, and tidal volumes. The PMA at which the infants achieved full nipple feeds and were discharged to home was also assessed.

Methods

This study was conducted in the level IV NICU at Oregon Health & Science University (OHSU). The protocol was approved by the OHSU Institutional Review Board, and informed consent was obtained for all enrolled patients. The study was registered at ClinicalTrials.gov (NCT02249143), and recruitment occurred from September 2014 through May 2016. No changes to the protocol or outcome measures were made after trial registration. Infants who were born at ≤32 weeks of gestation, treated with CPAP for ≥24 hours for respiratory distress (as initial treatment or following extubation), and met predesignated CPAP stability criteria were eligible for inclusion. Exclusion criteria included significant congenital heart disease, major congenital malformations, chromosomal abnormalities, culture-proven sepsis at consent, complex maternal conditions, clinical instability, and greater than twin gestation.

In the OHSU NICU, all CPAP is administered as bubble CPAP through appropriately sized Hudson nasal prongs¹¹ (Teleflex, Morrisville, North Carolina) with an 8 Fr orogastric tube to vent the stomach. The nares are assessed every 4 hours for redness or inadvertent pressure, and the prongs are repositioned as needed. The OHSU NICU follows established clinical consensus guidelines for CPAP discontinuation. In these guidelines, preterm infants are maintained on CPAP until CPAP stability criteria⁹ are met for a minimum of 12 hours, after which CPAP is discontinued and the infant is placed on room air. The published CPAP stability criteria⁹ are CPAP of 4–5 cm H₂O, no oxygen requirement, respiratory rate <70 breaths/minute, retraction score of ≤1, <3 self-resolving apneic episodes (<20 seconds) and/or bradycardia (<100 bpm) and/or desaturations (pulse oximeter oxygen saturation, SpO₂ <86%) per hour for the last 6 hours, average SpO₂ >86% for at least 90% of the time over the past 24 hours, not being treated for PDA or sepsis, and tolerating time off CPAP during routine care (up to 15 minutes).

The clinical team determined the patient’s respiratory management. When an infant met the CPAP stability criteria, study consent was obtained, baseline noninvasive PFTs without sedation were performed, and the infant was randomized to eCPAP or dCPAP. PFTs were also performed at the end of the 2-week treatment period and before discharge, at 35–36 weeks PMA. A standard operating protocol with specific acceptance criteria was followed for the PFTs, and a single, highly trained respiratory therapist performed all the tests. One investigator masked to treatment allocation reviewed all tests for quality.

Permuted block randomization with opaque sealed envelopes was used to stratify the infants by gestational age at delivery (<28 vs ≥28 weeks) and twins. The envelopes were prepared by a statistician not involved in the study to optimize the masking of randomization. After the randomization PFT, the envelope was opened and the patient placed on the allocated therapy. All infants were monitored daily for adverse events. If both twins were recruited into the study, the second twin was placed in the same treatment group as the first randomized twin.¹²

Infants randomized to the dCPAP group failed the room air trial if they exhibited at least 2 of the following⁹: respiratory rate >80 breaths/minute for >12 hours, increased apnea and/or bradycardia and/or desaturations (>2 episodes requiring moderate to vigorous stimulation per hour for the previous 6-hour period), any oxygen requirement to maintain SpO₂ >86%, a blood gas pH of <7.2, PaCO₂ >65 mm Hg, and a major apnea or bradycardia event requiring resuscitation. An infant who failed dCPAP resumed CPAP at the previous level, and dCPAP was reattempted within the next 7 days. If this second attempt again was unsuccessful, the infant was dropped from the study. Infants randomized to eCPAP were taken off CPAP for up to 7 days if abdominal distention with feeding intolerance or nasal septal breakdown occurred. If symptoms resolved within 7 days, eCPAP was reattempted. The infant was dropped from the study if signs of CPAP intolerance reoccurred a second time.

PFTs were measured with an infant pulmonary function cart (SensorMedics 2600; SensorMedics, Yorba Linda, California) at the infant’s bedside in the NICU. All testing was done with the infant supine and quiet at least 30 minutes following a feeding. The measurements were done with the infant breathing through a face mask connected to a 3-way valve.¹³

For the nitrogen washout technique, calibration was done with 2 known volumes, and a calibration line was constructed at the specific flow rate and used to correlate the nitrogen washed out to the infant’s FRC. The system corrected for dead space and corrected the FRC to body temperature, pressure, and water-saturated conditions. Acceptance criteria included infant supine and quietly asleep, test initiated at end expiration, no leak on washout tracing, consistent tracings, and ≥3 measurements with a coefficient of variation <10%.^13,14

Tidal volume was calculated by digital integration of the flow signal.^13,15 For the single-breath occlusion technique, the airway was briefly occluded at end inspiration until an airway pressure plateau was observed and the Hering–Breuer reflex was invoked. The linear portion of the passive flow–volume curve was identified, a regression line was drawn to obtain the best fit, and respiratory system compliance and resistance were calculated. Acceptance criteria as outlined by the American Thoracic Society/European Respiratory Society were followed.^15,16

Statistical Analyses

The primary outcome was the change in FRC from randomization through discharge in the 2 groups of infants. We hypothesized that a 15% improvement in FRC would be demonstrated at the end of the 2 weeks of treatment and at discharge in stable infants randomized to eCPAP vs dCPAP. Based on our previous studies,¹³ with an α value of 0.05 and a power of 0.80, sample size calculation determined the need for a minimum of 22 subjects in each group. We chose a 15% difference as meaningful, because changes in PFTs of this magnitude are used by the National Heart, Lung, and Blood Institute to define a positive response to bronchodilators.¹⁷

All outcomes were analyzed by a statistician blinded to the infants’ treatment assignments. Maternal and infant characteristics were compared using the independent-samples t-test for continuous measures (or the Wilcoxon-Mann-Whitney test for nonnormal distributions) and the χ² test for categorical variables. FRC values over time were modeled using linear mixed-effects regression with treatment group-time interaction and repeated measures for the randomization, 2 week, and discharge time points. Compound symmetric covariance structures were used to account for the correlation of FRC values within each patient. We compared outcomes between the 2 groups and included adjustments for sex, twin gestation, and weight at randomization. Sensitivity analyses randomly removing 1 of each pair of twins were done to determine the effect of including both twins on the difference in FRC between the 2 treatment groups. Group differences in the secondary outcomes were tested at randomization, 2 weeks, and discharge using the independent-samples t-test. Analyses were run in R version 3.3.1 (R Project for Statistical Computing, Vienna, Austria)¹⁸ using the nlme package lme function¹⁹ for the mixed-effects modeling.

Results

We screened a total of 116 patients born at ≤32 weeks of gestation for eligibility, and 50 patients met the inclusion and exclusion criteria and were randomized. Twenty-four infants were assigned to eCPAP, and 26 were assigned to dCPAP (Figure 1; available at www.jpeds.com). Six infants were excluded because they failed CPAP discontinuation, their parents withdrew consent, or their PFTs were technically unacceptable. The serial PFTs of the remaining 44 infants (22 per group) were analyzed for the primary outcome. Six infants (3 in each group) were transported back to their referring hospital at the end of the 2-week treatment period, and their discharge values were recorded as missing.

Figure 1.

CONSORT diagram of randomized premature infants. Enrollment, randomization, and the study of randomized infants through discharge.

The maternal and infant characteristics of the randomized infants who completed the PFTs are shown in Table I. There were no significant between-group differences in demographic characteristics, although the infants randomized to eCPAP tended to be more mature and thus to have a higher birth weight and tended to have more twins. The lowest gestational age in each group was 25.3 weeks. There were comparable numbers of patients born at <28 weeks of gestation in the 2 groups.

Table I.

Maternal and infant characteristics

Characteristics	eCPAP (N = 22)	dCPAP (N = 22)
Maternal age, y, mean (SD)	29.7 (7.3)	31.2 (7.2)
Gravida, median (IQR)	1.5 (1,6)	2 (1,3.8)
Antenatal steroids, n (%)	19 (91)	17 (77)
Cesarean delivery, n (%)	11 (50)	16 (73)
Birth weight, g, mean (SD)	1405 (417)	1222 (389)
Gestational age, wk, mean (SD)	29.5 (1.9)	28.7 (2.2)
White race, n (%)	17 (77)	19 (86)
Male sex, n (%)	9 (41)	14 (64)
Twin gestation, n (%)	7 (32)	4 (18)
1-minute Apgar score, median (IQR)	6 (5.3–7.0)	4 (2.0–7.0)
5-minute Apgar score, median (IQR)	8 (6.0–8.8)	7 (6.0–8.0)
<28 weeks gestation, n (%)	8 (36)	9 (41)
Surfactant, n (%)	4 (18)	10 (46)
Days of mechanical ventilation, median (IQR)*	0 (0,0)	0 (0,1)
Days of CPAP before randomization into study, median (IQR)	12.0 (5.8–28.8)	16.5 (10.3–32.3)
Days of oxygen, median (IQR)	0 (0–2.8)	0 (0–4.5)

^*P < .05; P values are from the independent-samples t test for means, Wilcoxon–Mann–Whitney test for medians, and χ² test for counts.

Both groups of patients were randomized at a mean of ~32 weeks PMA, and there was no between-group difference in weight or FRC at the time of randomization (Table II). The infants randomized to eCPAP had a significantly greater increase in FRC from randomization through 2 weeks of treatment (12.6 mL vs 6.4 mL; adjusted mean difference, 7.13; 95% CI, 0.78–13.47; P = .03) and from randomization through discharge (27.2 mL vs 17.1 mL; adjusted difference, 10.10; 95% CI, 2.61–17.59; P = .01) (Table II and Figure 2). At discharge, the FRC was 68.1 mL (28.4 mL/kg) in the eCPAP group vs 57.1 mL (25.4 mL/kg) in the dCPAP group (adjusted difference, 6.62; 95% CI, 0.02–13.22; P = .049). Both groups of infants were discharged at comparable PMAs and had comparable weights at the PFTs performed just before discharge (Table II). Oral feeds were initiated in both groups according to feeding cues, and patients achieved full feeds at comparable PMAs, 35.8 weeks in the eCPAP group and 36.0 weeks in the dCPAP group.

Table II.

Infant characteristics and FRC changes from randomization through discharge

Characteristics	eCPAP (N = 22)	dCPAP (N = 22)	P value
PMA at randomization, wk	32.2 (0.73)	32.0 (0.80)	.58
Weight at randomization, g	1567 (293)	1482 (323)	.36
FRC at randomization, mL	40.5 (11.8)	41.5 (10.6)	.15*
Average change in FRC, randomization to 2 wk, mL	12.6 (11.4)	6.4 (10.1)	.03*
Average change in FRC, randomization to discharge, mL	27.2 (12.5)	17.1 (11.7)	.01*
PMA at discharge PFT, wk	35.8 (0.63)	36.0 (1.02)	.57
Weight at discharge PFT, g	2381 (228)	2271 (440)	.35
FRC at discharge, mL	68.1 (9.2)	57.1 (12.0)	.049*

Values are mean (SD); P values are from an independent-samples t test unless noted otherwise. FRC values are obtained in 44 patients (22 in each group) at randomization, in 22 patients with eCPAP and 21 patients with dCPAP (1 technically unacceptable) at the 2-week time point, and in 17 patients with eCPAP (3 infants back-transported to the referring hospital and 2 with technically unacceptable PFT) and 18 patients with dCPAP at discharge (3 infants back-transported to the referring hospital and 1 with technically unacceptable PFT). PMA and weight at discharge are based on 19 patients in each group.

^*Linear mixed-effects regression models with repeated measures for time, treatment vs time interaction, and adjusting for sex, twins, and weight at randomization.

Figure 2.

FRC in randomized patients. Changes in FRC (mean ± SEM) in the eCPAP group (red) and the dCPAP group (blue) in 44 patients (22 in each group) at randomization, in 22 patients with eCPAP and 21 patients with dCPAP at the 2-week time point, and in 17 patients with eCPAP and 18 patients with dCPAP at discharge. The solid lines denote the treatment period from randomization through the 2 weeks of treatment, and the dashed lines denote the interval from the end of treatment to discharge. P = .03 for the change in FRC from randomization to the end of 2 weeks of treatment between the randomized groups; P = .01 for the change in FRC from randomization to discharge between the randomized groups. FRC values over time were modeled using linear mixed-effects regression models with repeated measures for time, treatment–time interaction, and adjusting for sex, twins, and weight at randomization.

Secondary PFT outcomes are shown in Table III (available at www.jpeds.com). FRC/kg was significantly different between the groups at the end of 2 weeks of treatment and at discharge. There was no difference in passive respiratory compliance (Crs) or in Crs/kg at baseline or any of the subsequent time points. The respiratory resistance was comparable in the 2 groups at baseline, significantly lower in the eCPAP group at 2 weeks, but comparable in the 2 groups again at discharge. The infants randomized to eCPAP had a statistically lower tidal volume at randomization, but these measurements were comparable at the end of the 2-week treatment period and at discharge.

Table III.

Secondary pulmonary function measurements in randomized patients

Measures	eCPAP (N = 22)	dCPAP (N = 22)	P value
FRC/kg at randomization, mL/kg	26.0 (5.5)	27.9 (4.6)	.20
FRC/kg at 2 weeks of treatment, mL/kg*	28.1 (4.9)	25.2 (3.7)	.03
FRC/kg at discharge, mL/kg†	28.4 (2.5)	25.4 (4.9)	.03
Crs at randomization, mL/cm H2O	2.20 (0.56)	1.93 (0.49)	.10
Crs/kg at randomization, mL/cm H2O/kg	1.37 (0.28)	1.32 (0.28)	.63
Crs at 2 weeks of treatment, mL/cm H2O‡	2.50 (0.54)	2.22 (0.36)	.06
Crs/kg at 2 weeks of treatment, mL/cm H2O/kg‡	1.28 (0.28)	1.17 (0.18)	.16
Crs at discharge, mL/cm H2O§	2.57 (0.51)	2.48 (0.55)	.61
Crs/kg at discharge, mL/cm H2O/kg§	1.08 (0.18)	1.12 (0.21)	.55
Rrs at randomization, cm H2O/mL/s	0.054 (0.014)	0.053 (0.009)	.86
Rrs at 2 weeks of treatment, cm H2O/mL/s‡	0.061 (0.014)	0.074 (0.021)	.03
Rrs at discharge, cm H2O/mL/s§	0.073 (0.019)	0.082 (0.022)	.19
Vt at randomization, mL/kg	7.1 (1.01)	7.8 (0.97)	.02
Vt at 2 weeks of treatment, mL/kg‡	7.6 (1.06)	7.4 (0.90)	.40
Vt at discharge, mL/kg§	7.3 (1.04)	7.3 (0.64)	.99

Rrs, passive respiratory resistance; Vt/kg, tidal volume per kilogram.

Values are mean (SD); P values are from an independent-samples t test unless noted otherwise.

^*Acceptable tests for FRC/kg in 22 patients with eCPAP and 21 patients with dCPAP.

^†Acceptable tests for FRC/kg in 17 patients with eCPAP and 18 patients with dCPAP.

^‡Acceptable tests obtained in 20 patients with eCPAP and 21 patients with dCPAP.

^§Acceptable tests obtained in 18 patients with eCPAP and 19 patients with dCPAP.

The infants randomized to eCPAP tolerated it well with no significant adverse events, including no events of nasal breakdown or feeding intolerance. One patient in the dCPAP arm failed the trial off CPAP twice and was dropped from the study. The parents of 1 patient in the eCPAP group withdrew consent.

Safety Monitoring

No significant adverse events were attributed to the study.

Discussion

This single-center, randomized pilot trial demonstrated that eCPAP administered to convalescing preterm infants who were born at ≤32 weeks of gestation, meeting predefined stability criteria for CPAP discontinuation, had a significantly greater increase in FRC from randomization through the 2-week treatment period and, importantly, also through discharge.

Despite advances in neonatal care, prematurity still results in impaired lung development²⁰ with decreased airway²¹ and parenchymal function²² and increased respiratory morbidity into adulthood. Currently, there are no known therapies to stimulate lung growth and decrease respiratory morbidity after preterm birth. Our present results suggest that in stable preterm infants, eCPAP may be a nonpharmacologic and safe therapeutic strategy to stimulate lung growth. Infants follow along a pulmonary function percentile established shortly after birth,^23,24 so a greater increase in FRC at NICU discharge could translate into longer-term respiratory benefits over the infant’s lifetime.

We are aware of no randomized trials defining the optimal PMA to discontinue CPAP in stable preterm infants. Early CPAP may benefit the preterm lung by minimizing injury, but clinical and animal data indicate that the mechanical stretch of the lung with CPAP may stimulate lung growth. Congenital diaphragmatic hernia prevents in utero lung expansion and results in hypoplastic lungs,^25,26 and complete congenital airway obstruction prevents lung fluid from leaving the trachea and produces a large, hyperplastic lung.²⁷ Animal models of in utero congenital diaphragmatic hernia have shown that fetal trachea occlusion increases the lungs’ distending pressure and results in increased fetal lung growth,^28,29 capillary growth, and epithelial branching.^30–33

The mechanical strain from CPAP applied to the lungs of young animals can stimulate lung growth and have other beneficial molecular effects. Tracheostomized ferrets exposed to 6 cmH₂O CPAP vs atmospheric pressure for 2 weeks demonstrated a 40% increase in total lung capacity with unchanged pressure volume curves indicating accelerated lung growth vs simple lung distention.⁵ High levels vs low levels of CPAP in ferrets significantly increased lung volume and airway size by volumetric computed tomography scan.⁶ In addition, cyclic stretch of cultured airway epithelial cells increases the release of nitric oxide and placental growth factor,⁷ which are important determinants of angiogenesis. In animal models, inhibition of angiogenic factors inhibits alveolarization, and replacement restores alveolarization. These studies support a physiological basis for eCPAP-related lung growth in the premature lung.

A strength of our study was the consistent management of all preterm infants on bubble CPAP, with randomization occurring when conservative CPAP stability criteria were met.⁹ This was adhered to by all of the neonatologists and health care providers at our NICU. Worldwide, there is wide variation in when and how premature infants are weaned off CPAP, even within the same NICU.^34,35 Many neonatologists will attempt to wean patients off CPAP as quickly as possible. In contrast, the longstanding practice at Columbia University Medical Center, where early CPAP was championed, is to continue infants on CPAP until 34.5 weeks PMA.³⁶ Some approaches to weaning premature infants off CPAP include weaning the CPAP pressure with variable frequency to low levels before discontinuation, time cycling between on and off CPAP for variable durations, and transitioning from CPAP to high-flow to low-flow nasal cannula and then to room air. All of our patients met the predefined stability criteria at approximately 32 weeks of age, and only 1 of the 26 (3.8%) assigned to the dCPAP group failed these criteria, necessitating reapplication of CPAP. Many centers wean patients from CPAP to high-flow nasal cannula, but this was not evaluated in our study.

Another strength of the study is the application of reproducible PFTs to document changes in lung volumes. We powered the study to show a 15% increase in FRC in the eCPAP group vs the dCPAP group and demonstrated a significant increase of >15%. To provide perspective, one of the most effective therapies to decrease morbidity in preterm infants is a single course of antenatal corticosteroids, which has been shown to increase FRC and Crs by approximately 50% in preterm animals and humans.³⁷ We have previously reported that neonatal PFT measurements of FRC and Crs correlate with clinical outcomes; a 10%–12% improvement in Crs measured soon after birth correlated with a significant decrease in wheeze through 12 months of age in offspring of pregnant smokers randomized to vitamin C vs placebo during pregnancy.¹⁵

CPAP has a number of acute and potentially long-term beneficial effects for the preterm lung, but it also has potential side effects, including nasal breakdown and gastric distention.⁸ Although our sample size was small for these outcomes, there was no increase in adverse events and no differences in time to full oral feeds, weight, and PMA at discharge between the 2 groups.

Although this was a randomized trial, we acknowledge several limitations. This was a pilot, single-center trial with a relatively small sample size; however, the desired sample size was achieved to address the hypothesis with the application of reproducible PFTs. Four of the 50 randomized infants were excluded from the overall PFT analysis, and even though this is a small number, it may have biased the outcome. It was not ethical or feasible to blind the study to investigators or parents, but PFTs were reviewed and statistical analyses were performed blinded to treatment. In addition, standard American Thoracic Society/European Respiratory Society criteria were applied to the acceptance of the PFT testing. The median gestational age of the randomized infants was 29.6 weeks (IQR, 27.5–30.8 weeks), so these results might not apply to more immature infants. We also did not perform follow-up PFTs after discharge to continue to follow the PFT trajectories.

There are several potential explanations for the significantly greater increase in FRC in the infants randomized to eCPAP vs those randomized to dCPAP. Similar to animal studies of CPAP application,⁵ there was no significant difference in the Crs between the 2 randomized groups, even though the FRC was larger in the eCPAP group. These measurements support increased lung growth vs lung overdistension with eCPAP. Passive respiratory resistance was not different in the 2 groups, although the measurement technique used in this study primarily measures resistance in the large airways. Measurements of forced expiratory flows, considered a more sensitive measure of the peripheral airways,³⁸ were not performed. Although the eCPAP infants may have larger lungs, the present study did not assess whether there was improved gas exchange. Some published evidence suggests that CPAP may also stimulate angiogenesis, which could translate into improved lung diffusion capacity³⁹; however, lung diffusion capacity was not measured in this study, because this technique is not currently available for infants in the NICU.

In conclusion, stable convalescing preterm infants born at ≤32 weeks of gestation and randomized to eCPAP have a significantly greater increase in FRC at the end of the 2-week treatment and through discharge compared with infants randomized to dCPAP. Premature infants have altered lung development with impaired alveolarization and an increased risk of subsequent respiratory disease. eCPAP in stable preterm infants may be a novel, safe, and nonpharmacologic therapeutic strategy to stimulate lung growth, alter pulmonary function trajectories, and improve childhood respiratory outcomes.

Glossary

CPAP

Continuous positive airway pressure

Crs

Passive respiratory compliance

dCPAP

Discontinued continuous positive airway pressure

eCPAP

Extended continuous positive airway pressure

FRC

Functional residual capacity

NICU

Neonatal intensive care unit

OHSU

Oregon Health & Science University

PFT

Pulmonary function test

PMA

Postmenstrual age