CPAP can be generated by either constant-flow or variable-flow devices. The most common constant-flow (also referred to as continuous-flow) methods of CPAP delivery are generated by bubble CPAP or by conventional ventilators. In the case of bubble CPAP, heated, humidified, blended gas flows are delivered to the infant, typically by nasal prongs or a mask, with the distal end of the expiratory tubing submerged in sterile water (or in some centers, 0.25 % acetic acid) to the depth required to generate the desired airway pressure (e.g., 6 cm below the surface to generate 6 cmH₂O pressure). Most, if not all, modern conventional neonatal ventilators are able to provide CPAP via a continuous gas flow source directed against a controlled resistance in the expiratory limb of the circuit. Although some conventional ventilators are also able to provide variable-flow CPAP, modulating the expiratory resistance valve and the circuit flow to maintain the pressure, this technology is different from what is generally considered true variable-flow CPAP, as described below. In addition to bubble and ventilator CPAP, the Benveniste gas-jet valve (Dameca, Copenhagen, Denmark) has been used (predominantly in Scandinavia) as a constant-flow CPAP device that works via the Venturi principle using two coaxially positioned tubes connected by a ring.

Variable-flow CPAP became increasingly popular in the 1990s and beyond. A number of devices and name changes have occurred over the past two decades, with the Infant Flow® CPAP/SiPAP system (CareFusion, Yorba Linda, CA, USA) being the dominant current device. Applying several fluidic principles of operation (including the Bernoulli, Coanda, and “fluidic flip” effects), gas delivered via dual injector jets at high velocity generates CPAP at the airway by the gas flow into the device and the leak around the nasal prongs.

Bubble CPAP produces measurable pressure oscillations around the baseline CPAP level. In preterm lambs, bubble CPAP may be associated with a higher pH, better oxygenation and ventilation, and less ventilation inhomogeneity, suggesting that the stochastic recruitment effect system may result in the need for a lower mean airway pressure to achieve the same level of volume recruitment, with potential reduced risk of adverse effects of higher CPAP pressures [1]. However, studies in human neonates are conflicting as to possible beneficial effects on gas exchange [2, 3]. Comparing the various CPAP devices in terms of efficacy, studies have shown improved lung compliance and decreased inspiratory work of breathing (WOB_I) and its component resistive work of breathing (RWOB) with variable-flow versus constant-flow CPAP via a ventilator [4]. Variable flow CPAP was also shown to reduce RWOB (but not WOB_I) and respiratory asynchrony when compared with bubble CPAP [5]. Despite these observed differences, studies to date comparing the different devices have not reported consistent improvements in clinically important outcomes. Stefanescu et al. [6] showed fewer days on supplemental oxygen and shorter hospital stay post-extubation using either Infant Flow CPAP or ventilator CPAP, but no differences in extubation failure. Comparing similar devices, Mazzella [7] demonstrated a lower oxygen requirement and decreased respiratory rate on Infant Flow CPAP, but no differences in successful weaning, need for mechanical ventilation, or duration of treatment. In a study comparing variable-flow and bubble CPAP in preterm infants, extubation failure was lower and the duration of support was shorter in infants ventilated less than 14 days when supported with bubble CPAP following extubation [8].

In the face of incomplete and often conflicting evidence, it seems prudent to use bubble or variable-flow CPAP (rather than ventilator CPAP) for their potential benefits on pulmonary mechanical parameters and some clinical outcomes. When deciding between these two options, the stochastic recruitment benefits of bubble CPAP may offer advantage in acute atelectasis-prone patients with respiratory distress syndrome. In contrast, the lower RWOB associated with variable-flow CPAP may benefit infants with chronic respiratory disease, impaired respiratory muscle contractility, and susceptibility to fatigue [9]. However, selection of an optimal patient interface and attention to detail in CPAP delivery (such as optimal prong size, minimizing leaks around the prongs and via the mouth, and appropriate airway care and positioning) may be at least as important as the CPAP device chosen. In all cases, adequately heated and humidified gas should be delivered at appropriate flows to avoid airway mucosal injury and condensation.

Phasic NIV consists of continuous positive pressure augmented with additional breaths delivered by periods of increased airway pressure, potentially facilitating ventilation and oxygenation. These breaths may be regular and nonsynchronized (nasal intermittent positive pressure ventilation – NIPPV) or synchronized with the infant’s respiratory efforts (synchronized nasal intermittent positive pressure ventilation – SNIPPV). NIPPV/SNIPPV is typically delivered by conventional mechanical ventilators, in some cases with adaptations for these specific modalities, although the ventilator used in studies of SNIPPV is no longer commercially available. SNIPPV has been shown to reduce respiratory asynchrony, improve tidal and minute volumes, decrease respiratory rate, and augment respiratory drive when compared with CPAP [10]. No strong evidence exists for benefits of SNIPPV over NIPPV. Something of a hybrid form of phasic ventilation is that of bi-level CPAP (commonly using the Infant Flow® SiPAP system), in which intermittent modest increases in airway pressure are superimposed on a baseline CPAP. Evidence is lacking to support either safety or efficacy of bi-level CPAP versus NIPPV/SNIPPV.

When compared with CPAP in neonates, NIPPV has been shown to further reduce the frequency of apnea [11], the need for intubation when used as primary respiratory support [12], and extubation failure (need for reintubation after extubation) [13]. However, NIPPV has not been shown to be superior to CPAP in preventing bronchopulmonary dysplasia (BPD) [14] or preventing death or BPD. [15] These data suggest that a significant subpopulation of neonates may enjoy short-term benefits from NIPPV as opposed to CPAP when avoiding or weaning from invasive ventilatory support. Because many infants can be successfully managed on CPAP alone for apnea, primary respiratory support, and following extubation, the difficulty is identifying those infants who may require the additional benefits of NIPPV and instituting this support modality in a timely manner. Evidence-based guidelines to aid in this important task are lacking.

Increasing interest has been paid in recent years to the potential role of nHFV, in particular noninvasive high-frequency oscillatory ventilation and nasal high-frequency percussive ventilation via nasal prongs. A limited number of laboratory and small clinical trials have demonstrated the feasibility of providing nHFV and addressed some of the technical aspects of this modality [16, 17], as well as suggested that nHFV may be associated with superior CO₂ elimination [18–20], shorter duration of supplemental oxygen support and respiratory distress [21], and sustained extubation success [22, 23]. However, additional large trials of safety and efficacy are needed before nHFV can be recommended for routine use in neonates.