The purpose of this study is to compare the effects of two different techniques of non-invasive ventilation (nCPAP and nHFOV) on gas exchange in preterm infants recovering from respiratory distress syndrome.

Background - Extremely low birth weight (ELBW) infants usually develop respiratory distress syndrome (RDS), due to lung immaturity, surfactant deficiency and immature respiratory control mechanisms (1). Even though mechanical ventilation is frequently lifesaving, complications are common (2). Tracheal intubation and mechanical ventilation are associated with ventilator-induced lung injury (VILI) and airway inflammation, leading to bronchopulmonary dysplasia (BPD) (3). The mechanisms of this injury involve alveolar over distension, the presence of shear forces and the release of pro-inflammatory cytokines (6), moreover prolonged duration of intubation and mechanical ventilation is associated with an increased risk of death or survival with neurologic impairment (3). In an effort to reduce VILI and so BPD in premature infants, there has been a trend toward increased use of non-invasive forms of respiratory support: nasal continuous positive airway pressure (nCPAP), nasal intermittent positive-pressure ventilation (NIPPV), high-flow nasal cannula (HFNC), nasal high-frequency oscillatory ventilation (nasal-HFOV)(2, 1, 7).

NCPAP is an alternative to intubation and a meta-analysis trials of early nasal CPAP versus intubation and ventilation showed that nasal CPAP reduces the risk of BPD. Nonetheless use of NCPAP in the delivery room may fail in ELBW, with 34 to 83% of such infants requiring subsequent intubation. Furthermore, post extubation support with nCPAP in these infants is associated with a 16-40% failure rate at 1 week (3, 4, 5, 9). NCPAP stabilized the surfactant deficient alveoli and improves oxygenation, but does not necessarily improve alveolar ventilation or partial pressure of carbon dioxide (pCO2) elimination (2, 8).

Some Authors reported the use of nasal high-frequency ventilation (nHFOV) in 14 very low birth weight (VLBW) infants with respiratory failure, using nasopharyngeal tube (3) and they have shown that this technique can lower pCO2 (1).

Other Authors investigated the efficacy of nasal HFOV applied on a single nasopharyngeal tube in an heterogeneous group of 21 infant with moderate respiratory insufficiency and they shown that was effective in reducing pCO2 (10).

No randomized controlled trials have directly evaluated the efficacy of nasal HFOV versus nCPAP with use of nasal prongs/mask in ELBW. There is rationale and support for the idea that high-frequency oscillation using a nasal prongs may improve carbon dioxide elimination in infants and minimizing the need for intubation and mechanical ventilation. In premature infants, HFOV is believed to cause less lung injury than conventional ventilation. So the question can be whether the benefits of HFOV and non invasive (nasal) ventilation are synergistic.

End-point - Our first end-point is that short-term application of nasal HFOV compared with nasal continuous airway pressure (nCPAP) in infants with persistent oxygen (O2) need recovering from RDS would improve CO2 removal. We also hypothesized that use of nHFOV can reduce the inspired fraction of O2 (FiO2) levels requiring to maintain normal percutaneous saturation of O2 (SpO2) levels .

Design - Multi-center non-blinded, randomized, observational four period crossover study Setting/population Level III Neonatal Intensive Care Unit (NICU): very-low birth weight infants (< 1500 g) and/or gestational age < 32 weeks requiring nCPAP and oxygen while recovering from RDS.

Methods - Infants requiring nasal CPAP (4-8 cm H2O) for > 24 hours prior to study enrolment, and FiO2 more than 0.21, will be randomized to either nCPAP or nHFOV delivered by Medin-CNO device. A crossover design with four 1 h treatment periods will be used such that each infant will receive both treatments twice. Oxygen saturation (SaO2), transcutaneous CO2 (tcCO2) and O2 (tcO2) and vital signs will be monitored continuously. Cardio-respiratory monitor recordings will be analyzed for apnoea, bradycardia and oxygen desaturations.

Study - After written informed consent will be obtained, the patient will be placed in the supine position. A transcutaneous CO2 and O2 monitor as well as pneumocardiography sensors will be placed on the infant and monitored. The patient will be randomized by sealed envelope shuffle to a starting treatment mode of either nCPAP (4-8 cm H2O, with the same CPAP level used prior to entering into the study) or nHFOV with the following starting parameters: CPAP level: 4-8 cm H2O with the same CPAP level used prior to entering into the study; Flow: 7-10 l/min (providing the desired CPAP level); Frequency: 10 Hz, Amplitude: 10, (eventually modified to obtain tcPCO2 values in the normal range (45-65 mmHg; 5.9-8.6 kPa). Short binasal prongs of right size will be always used. Prior to entering into the study all the patients will receive caffeine.

All support will be delivered by the Medin-cno device. Research personnel will adjust the FiO2 to attain a targeted oxygen saturation of 87-94%. The patients will be maintained in the usual thermoneutral environment (incubator) throughout the study, and will receive the standard routine care by the primary care team.

At study initiation, the infant will be started on the randomized starting mode of either nCPAP or nHFOV. The study will consist of four 1 h study blocks, alternating from the initial mode to the alternate mode twice. During the study neonates will be monitored continuously with a cardio-respiratory monitor, a pulse oximeter and a transcutaneous gas monitoring. All the data will be recorded continuously at 1-min intervals directly from the monitor and/or observed directly by an experienced neonatologist and manually recorded on a respiratory sheet. During each study block, the following data will be recorded: transcutaneous partial pressure of CO2 (tcPCO2), transcutaneous partial pressure of O2 (tcPO2), heart rate, respiratory rate, SaO2, Silverman score. Apnoeic episodes will be defined as absence of thoracic impedance change for a minimum of 20 s. Bradycardic episodes will be defined as persistent heart rate <80 beats per minute for a minimum of 10s. Significant desaturation episodes will be defined as persistent pulse oximetry values <80% for a minimum of 10s. Manual blood pressure will be taken with appropriate sized neonatal blood pressure cuff 30 minutes after the beginning of each treatment block.

Immediately after entering the study, at the beginning of the first study period, a transcutaneous monitoring of TcPCO2 and TcPO2 will be started and a capillary blood gas analysis (BGA) will be performed in order to test the reliability of the TcPCO2 data. A second capillary BGA will be performed at the end of second study period in both CPAP and nHFOV. Cerebral (cer-rSO2) and renal (ren-rSO2) tissue oxygenation will be measured by near-infrared spectroscopy (NIRS) as additional variable during the study period, based on the instrument's availability in each participating center.

The study will be ended when the patient will complete the 4 h study or will be terminated early if the patient will develop any signs of intolerance during the study, including an increase of >50% in the number of episodes of apnea or bradycardia compared with the prestudy baseline noted 1 h preceding study entry, or increased supplemental FiO2 > 0.3 from pre-study baseline for at least 15 min (i.e. from 0.30 to 0.60). To allow for equilibration, we will group and analyze data points from the last 20 min of each treatment block.

A sample size of 30 has been calculated to detect a mean difference of 3 mmHg tcCO2 based on a two-tailed p value of 0.05, power of 0.9 and a within-patient standard deviation (SD) of 2.5 mmHg (2).

Duration of study - To be defined, depending on the number of participating NICUs.

Compliance to protocol - Compliance will be defined as full adherence to protocol. Compliance with the protocol will be ensured by a number of procedures as described below.

Site set-up - Local principal investigators will participate to preparatory meetings in which details of study protocol, data collection and procedures will be accurately discussed. All centers will receive detailed written instruction on web based recording data, and, to solve possible difficulties, it will be possible to contact the Chief investigators. Moreover, it has been ascertained that the procedure is equally made in all participating centers.

Data processing and monitoring - All study data will be

1. Screened for out-of-range data, with cross-checks for conflicting data within and between data collection forms by a data manager.
2. Referred back to relevant centre for clarification in the event of missing items or uncertainty.

The Chief Investigator and trial statistician will review the results generated for logic and for patterns or problems. Outlier data will be investigated.

Safety - Safety end-point measures will include incidence, severity, and causality of reported serious adverse effects (SAE), namely changes in occurrence of the expected common prematurity complications and clinical laboratory test assessments, and the development of unexpected SAEs in this high risk population. All SAEs will be followed until satisfactory resolution or until the investigator responsible for the care of the participant deems the event to be chronic or the patient to be stable. All expected and unexpected SAEs, whether or not they are attributable to the study intervention, will be reviewed by the local principal investigators to determine if there is reasonable suspected causal relationship to the intervention. If the relationship is reasonable SAEs will be reported to Chief Investigators who will report to Ethics Committee and inform all investigators to guaranty the safety of participants.

Inclusion Criteria:

• Birthweight < 1500g and/or
• Gestational age < 32 weeks
• nCPAP treatment for > 24 h
• Oxygen supply to keep SaO2 87-94% for a minimum of 1 h prior to initiation of the study
• Parents written informed consent

Exclusion Criteria:

• Active medical treatment for patent ductus arteriosus
• culture proven sepsis
• Major congenital malformations
• Genetic syndromes
• Postoperative recovery period of <24 h