For improving oxygenation in a recruitable lung, iRDS being the optimal example for a highly recruitable lung pathology, at least in the early stage of iRDS before widespread formation of hyaline membranes, cellular infiltrations and fibrinous depositions occur, and if applying an “open-lung” strategy, mean airway pressures (P _AW) are gradually increased by steps of 1–2 cm H₂O with the goal of reducing the FiO₂ to ≤ 0.4 while achieving adequate SaO₂ values (≥90 %, and ≥85 % in the premature infant). During this phase, CO₂ and haemodynamics should be carefully monitored, and at the occurrence of a minimal increase of CO₂ (indicating beginning overdistention) or a decrease of blood pressures (i.e. more than 2 mmHg in the preterm infant), airway pressures have to be reduced despite eventual still poor oxygenation and high FiO₂ needs, before eventually further increasing airway pressures again. Once appropriate saturation is achieved and FiO₂ could be lowered to ≤0.4 (e.g. in experienced hands ≤0.3), airway pressure should be reduced stepwise in increments of 1–2 cm H₂O (moving downward on the deflation limb of the PV-curve that exhibits hysteresis (Bryan and Cox 1999; De Jaegere et al. 2006; Pellicano et al. 2009)) as long as SaO₂ can be maintained, the ultimate goal being to find the lowest pressure needed to maintain “good” saturation (Fig. 45.4). Chest radiography is usually recommended after finding the “optimal” mean airway pressures (P _AW) to ensure adequacy of lung recruitment and to exclude signs of overinflation, although only gross overinflation can be seen on chest X-ray by the loss of the dome shape of the diaphragm. If signs of overinflation are present, airway pressures have to be reduced further. The common definition of a good lung volume by the position of both hemidiaphragms at the level of the 8th to 9th rib lacks any scientific evidence.

Meconium aspiration syndrome is characterised by a variable combination of atelectasis, airflow obstruction and lung inflammation, and is often accompanied by pulmonary hypertension. Maldistribution of ventilation and aeration is therefore common (Yeh 2010), rendering usually efficient but lung-protective ventilation difficult with conventional ventilation. Interestingly, HFOV (see Sect. 47.​2.​3 for indications) allows often to achieve relative good pulmonary ventilation and gas exchange (Clark et al. 1994) reducing the need for extracorporeal membrane oxygenation (ECMO) treatment and decreasing the incidence of air leaks.

With poor oxygenation due to an important intrapulmonary shunt (predominant atelectasis), stepwise lung recruitment attempts as described above are justified and relatively high mean airway pressures (P _AW) might be required, whereas with poor oxygenation because of an extrapulmonary right to left shunt (with suprasystemic pulmonary artery pressures due to acidosis – acidosis being an important pulmonary vasoconstrictor) and/or lung overdistention, high mean airway pressures or formal recruitment attempts are often inefficient and may be in fact dangerous. This is because of the risk of further overdistending the lungs and often poor haemodynamic tolerance (drop in systemic blood pressure and often oxygenation, as well as the risk of exacerbating pulmonary hypertension). In any case, after any recruitment attempt, whether beneficial or not in terms of oxygenation and ventilation efficiency, airway pressures have to be immediately lowered to the lowest pressure needed to maintain adequate lung volumes. Because of the high probability of having to deal with high airway resistance, low frequencies below 10 Hz can in general be advocated (Hachey et al. 1998). For more detailed informations on MAS, see Sect. 47.​2.​3.

Lung hypoplasia syndromes, of which congenital diaphragmatic hernia (CDH) is part of (see Sect. 47.​2.​4), are characterised by a reduced total lung capacity in relation to body weight. This means that using “only” 4–5 ml/kg of tidal volume (Vt) during conventional ventilation might be inappropriately high and is likely to cause secondary lung injury. However, using lower Vt than 3–4 ml/kg will render ventilation in terms of CO₂ washout difficult and not be clinically acceptable in a situation when important respiratory acidosis (i.e. permissive hypercapnia) might haemodynamically not well be tolerated because of already coexisting pulmonary hypertension. To switch to HFOV as an early intervention or even first intention approach as soon as PIP exceeds 25 cm H₂O during conventional ventilation has therefore to be proposed. However, to use a high lung volume approach as commonly recommended during HFOV would be inappropriate since lung hypoplasia syndrome does, like CDH, not represent per se a recruitable lung. The peak-to-peak pressures (amplitude) are adjusted to achieve a PaCO₂ in the range of 35–45 mmHg with a pH in the range of 7.35–7.45 as long as the preductal SaO₂ is >85 %. Mean airway pressure is limited in absence of any additional restrictive lung disease (e.g. iRDS) to 14–16 cm H₂O.

HFOV was established years ago as the choice of a ventilator mode (i.e. rescue treatment) for pre-existing pulmonary air-leak syndromes (PAL) resulting from injurious mechanical ventilation. This was mainly based on a single uncontrolled rescue study (Clark et al. 1986). Furthermore, the use of HFOV was shown to reduce the risk [RR 0.73 (0.55, 0.96), RD −0.174 (−0.321, −0.027), number of infants that need to be treated (NNT) 6 (95 % CI 3, 37)] of developing any new pulmonary air leak (interstitial pulmonary emphysema and/or pneumothorax) in infants ventilated for severe RDS (Bhuta et al. 2001). However, this latter analysis is based on the results of a single randomised controlled trial (HiFO Study Group 1993), and no benefit of HFOV was shown for those infants with established moderate to severe PIE at entry into the trial.

While HFOV might be accepted as a suitable rescue ventilation mode for PAL, the selection of oscillatory frequency is an unresolved issue: selecting the wrong frequency can potentially exacerbate the underlying problem of air leak. In an optimal setting, frequency would be determined by measuring the impedance of the lung and determination of the corner frequency, which defines the optimal frequency in the overdamped newborn lung (Pillow 2005; Venegas and Fredberg 1994). However, as this approach is currently only used in a research setting, clinicians have little to guide them on optimal frequency selection. In general, increased resistance will decrease the corner frequency, whereas decreased compliance will increase the corner frequency. Corner frequency is also dependent on patient size and is higher in smaller infants. Whereas some advocate using a low HFOV frequency for air leak, there are two major concerns with this approach. Firstly, by using a fixed % setting of inspiratory time, the inspiratory time will increase as frequency decreases – so the duration of exposure to the high pressure is longer and higher tidal volumes will be delivered. Secondly, lower resistance and inertance at lower frequency results in a higher percentage of the airway-opening pressure amplitude being delivered to the proximal and distal airspaces (see Sect. 5.2.1.1). Although the higher respiratory resistance in PAL potentially justifies a lower frequency compared with a poorly compliant lung (i.e. classical iRDS), decreasing frequency below 8–10 Hz is probably not justified based on respiratory mechanics (except in very big infants). Excessively low frequencies potentially expose the central airways to large-pressure oscillations as well as those airspaces proximal to areas of high peripheral resistance. Decreasing the % inspiratory time to 30 % or below to limit delivered tidal volume would reduce volutrauma associated with the use of lower frequencies.