Approach to mechanical ventilation in very preterm neonates

INTRODUCTION — Mechanical ventilation (MV) is a lifesaving intervention, but it also risks injury to the lungs, brain, and other organ systems. Supporting gas exchange while minimizing harm is the key therapeutic goal and challenge of MV in neonates.

The approach to MV in very preterm (VPT) neonates (ie, gestational age ≤32 weeks) will be reviewed here. A discussion of the general principles of MV in neonates, a broad overview of MV modes, and detailed discussion of MV in other specific neonatal populations at high risk of respiratory failure, including neonates with persistent pulmonary hypertension of the newborn, congenital diaphragmatic hernia, and older preterm-born infants with evolving or established bronchopulmonary dysplasia (BPD) are provided in separate topic reviews:

●(See "Overview of mechanical ventilation in neonates".)

●(See "Persistent pulmonary hypertension of the newborn (PPHN): Management and outcome", section on 'Respiratory support'.)

●(See "Congenital diaphragmatic hernia in the neonate", section on 'Ventilation'.)

●(See "Bronchopulmonary dysplasia (BPD): Management and outcome", section on 'Respiratory support'.)

TERMINOLOGY — The following terms are used throughout this topic:

●Preterm neonates – Different degrees of prematurity are defined by gestational age (GA) or birth weight, as detailed in the table (table 1).

●Very preterm (VPT) neonates – VPT neonates are those born at GA <32 weeks.

●Terms related to MV – Terms used to define different modes and settings for MV in neonates are summarized in the table (table 2).

●Ventilator-induced lung injury (VILI) – VILI is lung injury caused by MV. It can result from exposure to excessive pressure (barotrauma), excessive stretching of the lung tissue (volutrauma), cyclic collapsing of the alveolar spaces (atelectrauma), and exposure to high fraction of inspired oxygen (FiO₂). (See "Ventilator-induced lung injury".)

RESPIRATORY FAILURE IN VPT NEONATES

Contributing factors — Respiratory failure is the primary cause of morbidity and mortality in very preterm (VPT) neonates [1]. Respiratory failure results from multiple contributing factors, including respiratory distress syndrome (RDS) from surfactant deficiency. The pathophysiology of RDS is discussed in detail separately. (See "Respiratory distress syndrome (RDS) in the newborn: Clinical features and diagnosis", section on 'Pathophysiology'.)

Many factors that contribute to respiratory failure in this population have important implications for MV management. Examples include:

●Immature respiratory drive – Preterm neonates are prone to central apnea and hypopnea due to an inconsistent respiratory drive from an immature nervous system. Because of this, VPT neonates typically require a MV mode with a set mandatory or back-up breath rate. (See "Pathogenesis, clinical manifestations, and diagnosis of apnea of prematurity".)

●Narrow and collapsible airways – The supraglottic and subglottic airways of VPT neonates are narrow and lack a mature supporting tissue architecture, making them unstable and prone to collapse. This is one of several reasons why VPT neonates benefit from continuous distending airway pressure (noninvasive continuous positive airway pressure [CPAP] and/or invasive positive end-expiratory pressure [PEEP]) during MV. (See 'Initial conventional MV settings' below.)

●Reduced alveolar air space – The alveolar air space is reduced due to incomplete alveolarization, with less surface area available for gas exchange; inadequate surfactant quantity and quality; and edema from a secretory alveolar epithelium, excessive pulmonary blood flow, underdeveloped pulmonary lymphatics, and inflammation. As a result, gas exchange in VPT neonates is compromised by low pulmonary compliance, ventilation/perfusion mismatch, and poor diffusion. Compliance is not static and marked dynamic changes can occur (eg, after surfactant administration). (See "Respiratory distress syndrome (RDS) in the newborn: Clinical features and diagnosis".)

●Compliant chest wall – VPT neonates have biomechanical disadvantages, including a compliant chest wall and the suboptimal shape and position of the immature ribs and diaphragm. These disadvantages contribute to a low functional residual capacity and risk atelectasis, while increasing the work of breathing.

Need for respiratory support — Because the immature respiratory system fails to maintain adequate gas exchange, most VPT neonates require some form of respiratory support (either noninvasive support or invasive MV) [2]. The likelihood of requiring invasive MV increases with decreasing gestational age (GA). Despite the trend towards increasing use of noninvasive respiratory support in managing VPT neonates since the 1990s and early 2000s, use of invasive MV remains common in this population, especially in extremely preterm (EPT) infants (GA <28 weeks).

In a study from the Australian and New Zealand Neonatal Network (2007 to 2013) of nearly 12,000 VPT neonates who were initially managed with nasal continuous positive airway pressure (nCPAP), one-quarter required endotracheal intubation within 72 hours of birth [2]. CPAP failure was more common among neonates born at 25 to 28 weeks GA compared with 29 to 32 weeks GA (43 versus 21 percent, respectively).

In a prospective registry study that included nearly 11,000 neonates <29 weeks GA in the United States Neonatal Research Network (2013 to 2016), 84 percent required MV at some time in the postnatal course [3]. MV use increased with decreasing GA, with >99 percent of neonates born at 23 weeks GA requiring MV compared with 88 percent of infants born at 26 weeks GA and 66 percent of neonates born at 28 weeks GA [4].

CLINICAL APPROACH — The following sections detail the approach to MV in very preterm (VPT) neonates. Management is tailored to meet the needs of the individual neonate, which will differ between patients and within the same patient over time. Individualized assessment and frequent reassessment of the adequacy of ventilator settings is critical, especially with changes in cardiopulmonary physiology during the early postnatal period.

General principles — While lifesaving, MV can also cause lung injury and impact hemodynamics, with secondary consequences on the brain and other organs. MV can cause ventilator-induced lung injury (VILI) at any age and the general measures used to minimize VILI are fairly consistent regardless of age; however, the impact of VILI and the imperative for utilizing lung-protective strategies are particularly important for VPT neonates with immature, injury-prone lungs. Bronchopulmonary dysplasia (BPD) is a common and consequential morbidity of preterm birth caused by concurrent injury and maldevelopment of the immature lungs. (See "Bronchopulmonary dysplasia (BPD): Clinical features and diagnosis", section on 'Risk factors'.)

Therapeutic strategies to support gas exchange while minimizing lung injury are key to neonatal care. These strategies include (table 3):

●Avoidance of MV through preferential use of noninvasive respiratory support (eg, nasal continuous positive airway pressure [nCPAP]) when possible (see "Respiratory distress syndrome (RDS) in preterm infants: Management", section on 'Noninvasive positive airway pressure')

●Use of lung protective strategies for invasive MV when noninvasive support fails, including:

•Preferential use of volume-targeted ventilation (VTV) with tidal volumes (Tv) of 4 to 6 mL/kg to minimize volutrauma (see "Overview of mechanical ventilation in neonates", section on 'Volume-targeted ventilation')

•Use of positive end-expiratory pressure (PEEP) to maintain lung recruitment and avoid atelectasis (see 'Initial conventional MV settings' below)

•Avoidance of high inspired oxygen levels (see 'Gas exchange targets' below and "Neonatal target oxygen levels for preterm infants")

•Setting targets for gas exchange that do not aim for normal levels (ie, modest permissive hypercapnia) (see 'Gas exchange targets' below)

•Use of high-frequency oscillatory or jet ventilation (HFOV or HFJV) as a rescue therapy for neonates with refractory respiratory failure while on conventional mechanical ventilation (CMV) or as an initial ventilation strategy in neonates at high risk of developing VILI (see 'Refractory respiratory failure' below and 'Role of high-frequency ventilation' below)

Indications for invasive MV — Our general approach to determining the need for invasive MV in VPT neonates is based on clinical judgement, determining when efforts of noninvasive respiratory support fail to provide adequate gas exchange (table 3). Although the parameters below provide a useful framework, consensus is lacking, and there is no standardized definition to determine when invasive MV should be initiated. Our approach is generally consistent with definitions used in major clinical trials evaluating respiratory support in VPT infants [5,6]:

●We attempt to avoid MV through the preferential use of noninvasive respiratory support (eg, nasal continuous positive airway pressure [nCPAP]). In our center, nCPAP is the preferred initial modality. The evidence supporting preferential use of different noninvasive respiratory support modes such as nCPAP and nasal intermittent positive pressure ventilation (nIPPV) preterm infants to prevent VILI and BPD is discussed in detail separately. (See "Respiratory distress syndrome (RDS) in preterm infants: Management", section on 'Noninvasive positive airway pressure'.)

●We allow a trial of noninvasive support (eg, nCPAP) in all preterm neonates, irrespective of gestational age (GA). We titrate nCPAP to levels as high as 8 cm H₂O when appropriate in an effort to optimize gas exchange in the first few weeks of life. (See "Respiratory distress syndrome (RDS) in preterm infants: Management", section on 'Initial management'.)

●If, despite efforts to optimize noninvasive support, the neonate develops any of the following signs of inadequate gas exchange, we typically intubate and initiate invasive MV:

•pH <7.20, with a PaCO₂ >65 mmHg.

•Requiring FiO₂ >0.4 to 0.5 to achieve target peripheral SpO₂ (oxygen saturation measured by pulse oximetry). (See "Neonatal target oxygen levels for preterm infants", section on 'Oxygen target levels'.)

•Multiple apneic episodes per hour associated with desaturations and bradycardia or more than one episode requiring positive pressure ventilation within a few hours.

●We rely mostly on these objective measures of gas exchange to define failure of noninvasive respiratory support. However, the thresholds are not hard set and these parameters should be interpreted in conjunction with other clinical findings. For neonates with signs of labored breathing, hemodynamic instability, or persistent metabolic acidosis, we generally use a lower threshold for transitioning to invasive MV. The threshold to transition to invasive MV is lower in the most immature infants who are at highest risk of failing noninvasive support.

Our approach to determining when efforts to optimize noninvasive respiratory support have failed is generally consistent with the definitions used in clinical trials evaluating nCPAP in preterm neonates. For example, in the COIN trial, infants randomized to nCPAP were deemed to have failed if they had pH <7.25 with a PaCO₂ >60 mmHg, persistent metabolic acidosis, FiO₂ >0.60, or excessive apneic episodes (ie, ≥6 episodes requiring stimulation or >1 episode requiring positive-pressure ventilation within a six-hour period) [5]. Similarly, in the SUPPORT trial, failure criteria for infants randomized to nCPAP were a PaCO₂ >65 mmHg, FiO₂ >0.50 to maintain SpO₂ above 88 percent for over one hour, or hemodynamic instability [6].

Choice of mode

Our approach — The approach to choosing a mode of ventilation and initial ventilator settings is not standardized and practice varies from center to center. Our general approach is as follows (table 3):

●We use CMV as the initial approach and reserve high-frequency ventilation (HFV) for cases of refractory respiratory failure. Other centers may use HFV more commonly as the initial ventilation approach, particularly in neonates at high risk of developing VILI. Data comparing CMV and elective HFV are discussed below. (See 'Role of high-frequency ventilation' below.)

●We use a synchronized mode with both mandatory and spontaneous breaths (ie, synchronized intermittent mandatory ventilation plus pressure support [SIMV + PS] or assist control ventilation [ACV]). Data comparing these modes are discussed separately. (See "Overview of mechanical ventilation in neonates", section on 'Synchronized modes'.)

●We preferentially use VTV as our initial approach in all VPT neonates. Pressure-limited ventilation (PLV) may be used if VTV is not available or when there is a technical challenge that limits the reliable delivery of measured Tvs (eg, large endotracheal tube [ETT] leak). Data supporting preferential use of VTV are described below. (See 'Volume-targeted versus pressure-limited ventilation' below.)

Volume-targeted versus pressure-limited ventilation — Use of VTV in neonates has increased over time as technical advances have allowed for more accurate measurement of small Tvs and better compensation for endotracheal tube (ETT) leaks. In a 2018 report of respiratory management of extremely preterm infants among various international neonatal networks, approximately one-third of NICUs reported using VTV as the most common initial mode of MV [7]. VTV was the predominant mode in Australia and New Zealand, Canada, Finland, and Spain, but remained less commonly used in other regions.

The shift towards increasing use of VTV in neonates was spurred by preclinical observations that tissue stretch from lung overdistension (volutrauma) caused more lung inflammation and injury than exposure to high pressure without excessive stretch (barotrauma) [8-10]. The superiority of VTV over PLV has been confirmed in meta-analyses of clinical trials, which demonstrate improved short-term outcomes [11]. However, there are few data on long-term neurodevelopmental outcomes.

A 2017 systematic review and meta-analysis identified 20 randomized controlled trials comparing VTV and PLV in 977 neonates (predominantly preterm neonates) [11]. VTV resulted in:

●Shorter duration of MV (mean difference 1.35 days shorter; 95% CI 0.86-1.83 days)

●Lower incidence of pneumothorax (5 versus 9 percent; rate ratio [RR] 0.52, 95% CI 0.31-0.87)

●Lower incidence of BPD at 36 weeks (23 versus 35 percent; RR 0.68, 95% CI 0.53-0.87)

●Lower incidence of periventricular leukomalacia or grade 3 and 4 intraventricular hemorrhage (IVH) (8 versus 16 percent; RR 0.47, 95% CI 0.27-0.80)

●Nonsignificant trend towards lower mortality (12 versus 16 percent; RR 0.75, 95% CI 0.53-1.07)

All of the trials included in the meta-analysis compared volume-targeted with pressure-limited strategies, but there was considerable heterogeneity with respect to the specific mode used in each arm. In addition, the trials were conducted at centers with relative expertise in the use of VTV. The generalizability of these results to centers implementing the novel use of VTV remains uncertain.

Historically, PLV was the standard approach used for neonatal MV. In fact, PLV is the only mode available on many older generations of neonatal ventilators. A key disadvantage of PLV is breath-to-breath variability in delivered Tv. This can be pronounced when the mechanics of the lung and respiratory circuit change dynamically, as may occur after surfactant administration, with changes in lung volume, or when the ETT is partially occluded (eg, from secretions) or has a positional leak. Both unacceptably large and small Tvs can contribute to VILI from volutrauma and atelectrauma, respectively. In addition, rapid fluctuations in carbon dioxide may contribute to unstable cerebral perfusion and brain injury. In a study analyzing data on nearly 12,000 ventilator breaths in 36 preterm neonates managed with PLV or VTV, there was considerably more breath-to-breath Tv variability with PLV, with 61 percent of breaths in the PLV group falling outside the target range of 4 to 6 mL/kg compared with 37 percent of VTV breaths [12].

However, there are some advantages to PLV:

●Cost and availability – Unlike VTV, which generally requires more modern ventilators to function effectively in neonates, PLV can be used with older generations of neonatal ventilators, many of which lack a VTV option and only provide PLV. Thus, in resource-limited settings, PLV may be more widely available, less costly, and easier to use.

●Less prone to error from ETT leaks – In VTV, a large ETT leak can limit the reliable delivery of the desired Tv. This is less of a problem in PLV since the delivered pressure does not depend upon the accurate measurement of Tv.

Role of high-frequency ventilation — HFV can be used either as the primary mode of MV support following endotracheal intubation ("elective" HFV therapy), or as "rescue" therapy for neonates with refractory respiratory failure despite efforts to optimize CMV. In our practice, we do not use elective HFV as a primary mode of ventilation in VPT neonates because CMV is more readily available, easier to use, and conclusive evidence supporting elective HFV is lacking. However, other centers may use HFV more commonly as the initial ventilation approach, particularly in neonates at high risk of developing VILI.

The approach to using HFV as a "rescue" therapy for neonates with refractory respiratory failure is detailed below. (See 'Refractory respiratory failure' below.)

Based on the available evidence, elective HFOV does not appear to be superior to CMV as the initial MV approach in preterm neonates [13-15]. However, the data suggest that elective HFOV is a safe alternative to CMV in the hands of experienced providers.

Most of the studies on elective HFV have evaluated HFOV; there are fewer available data on HFJV in this setting:

●"Elective" HFOV – In a 2015 meta-analysis of 19 trials comparing elective HFOV with CMV in 4096 preterm infants, the incidence of pulmonary air leak was higher in the HFOV group (28 versus 23 percent; RR 1.19 [95% CI 1.05-1.34]) but the rate of BPD in survivors at 36 weeks or discharge was lower in the HFOV group (30 versus 35 percent; RR 0.86 [95% CI 0.78-0.96]) [15]. Mortality at 36 weeks or discharge was similar in both groups (15 versus 16 percent; RR 0.95 [95 % CI 0.81-1.10]). Rates of grade III or IV IVH or periventricular leukomalacia were similar in both groups. An earlier patient-level meta-analysis reported similar findings [13].

Long-term neurodevelopmental outcomes were not reported in the meta-analyses as pooled estimates were not possible due to heterogeneity in measurement. In the largest trial reporting neurodevelopmental outcomes (n = 386 infants), the rate of moderate to severe disability at 16 to 24 months was higher in the HFOV group (46 versus 35 percent; RR 1.28, 95% CI 1.02-1.60) [16]. However, this early trial did not use an open-lung strategy and the finding of worse neurodevelopmental outcomes with HFOV has not been consistently reported in other trials [17-20]. An "open-lung" strategy minimizes atelectasis and increases lung volume, typically through application of higher continuous distending airway pressure settings such as mean airway pressure (during HFOV) or PEEP (during CMV). These distending pressures are sometimes applied transiently, with a hope that increased lung volumes can be maintained despite subsequent pressure decrements.

Based on limited follow-up data, long-term pulmonary outcomes appear to be similar for patients managed with HFOV or CMV [16,21-23]. In a follow-up study of 161 patients enrolled in one large trial, pulmonary function among surviving 16- to 19-year-old participants did not differ between those who received HFOV versus CMV [23].

●"Elective" HFJV – In a 2002 meta-analysis of three randomized trials comparing HFJV with CMV in 204 preterm neonates, there was a nonsignificant trend towards a reduced rate of BPD in the HFJV group at 28 days (63 versus 71 percent; RR 0.90 [95% CI 0.74-1.09]) and a significant difference at 36 weeks postmenstrual age (20 versus 33 percent; RR 0.59 [95% CI 0.35-0.99]), though the latter outcome was not reported in one trial [24]. Mortality rates were similar in both groups (15 versus 18 percent; RR 0.86 [95% CI 0.49-1.50]). Rates of pulmonary air leak and grade III to IV IVH were similar in both groups. Rates of PVL were reported in the two trials with inconsistent findings. One trial reported a lower rate of PVL in the HFJV group [25], while the other trial found markedly higher rates of PVL in the HFJV group [26]. It is possible that differences in how HFJV was used may account for these differences. Notably, an open-lung strategy was used in the first but not the second trial.

Data on long-term respiratory and neurodevelopmental outcomes in infants managed with HFJV are lacking.

Initial conventional MV settings — We initiate CMV using a minimal volume-targeted approach to prevent VILI while maintaining target hemoglobin oxygen saturation (SpO₂) and allowing permissive hypercapnia. (See 'Gas exchange targets' below.)

Typical initial settings are (table 3):

●Tv 4 to 6 mL/kg

●PEEP 5 to 6 cm H₂O

●Inspiratory time (Ti) 0.35 to 0.4 seconds

The neonate should be assessed immediately upon initiating MV and reassessed frequently thereafter to determine the adequacy of the settings and potential need for further titration of the ventilator. (See 'Monitoring' below and 'Titrating conventional MV' below.)

Our approach is based on clinical trials, observational data, and clinical experience:

●Tv setting – An initial setting of 4 to 6 mL/kg is consistent with the recommendations of the European consensus guidelines on the management of RDS and with the settings used in clinical trials evaluating VTV [27-30]. It is also consistent with usual practice, as demonstrated in a cross-sectional survey of neonatal MV practice in Europe from 2010, in which the mean Tv was 5.7 mL/kg among preterm neonates managed with CMV [31]. Clinical trial data comparing different Tv settings are limited. In a small trial involving 30 preterm neonates randomized to a Tv of 3 or 5 mL/kg within the first hour after birth, neonates in the 5 mL/kg group had shorter duration of MV and lower levels of pro-inflammatory cytokines detected in tracheal aspirates [32]. However, it is reasonable to use slightly larger initial Tv in smaller preterm neonates to account for the relatively larger contribution of fixed instrumental dead space to the overall Tv. (See "Overview of mechanical ventilation in neonates", section on 'Volume-targeted ventilation'.)

●Initial PEEP setting – As PEEP is delivered directly to the subglottic airway during MV, a lower pressure level than used during noninvasive respiratory support is often sufficient. While the optimal initial PEEP level is uncertain, it is common practice to use a level of 5 to 6 cm H₂O and then adjust further, if needed, to achieve target SpO₂ levels. This range is consistent with PEEP levels reported in the previously described cross-sectional European survey study and a secondary analysis of an international multicenter trial involving 278 extremely low birth weight (ELBW, birth weight <1000 g) infants (mean PEEP levels in these studies were 4.5 and 5.7 cm H₂O, respectively) [31,33]. However, considerable between-center variation was noted in the latter study [33].

A 2019 systematic review on PEEP selection in preterm neonates identified little data to inform practice [34]. Two small crossover trials contributed physiologic gas exchange data from 28 subjects, with no statistically significant differences for oxygenation or ventilation noted between low (<5 cm H₂O) versus high (≥5 cm H₂O) PEEP [34]. Data on clinical outcomes were not reported.

●Ti setting – An initial setting of 0.35 to 0.4 seconds is consistent with usual practice as demonstrated in the previously described cross-sectional European survey study, in which the mean Ti was 0.38 seconds [31]. In a 2003 meta-analysis of five trials (694 neonates) comparing "shorter" Ti (ranging from 0.33 to 1.0 seconds) with "longer" Ti (ranging from 0.7 to 1.0 seconds), longer Ti was associated with more air leak (RR 1.56, 95% CI 1.25-1.94) and higher mortality (RR 1.26, 95% CI 1.00-1.59) [35]. However, the trials included in the meta-analysis were conducted between 1978 and 1989, prior to the routine use of antenatal corticosteroids and surfactant and well before the era of preferential noninvasive support. Thus their applicability to modern day practice is limited.

Gas exchange targets — Optimal gas exchange targets for VPT neonates remain uncertain. The goal is not to achieve "normal" oxygen and carbon dioxide levels, but rather levels that balance the harms from hypoxia, acidosis, and respiratory energy expenditure with iatrogenic harms, including oxygen toxicity and VILI.

●Oxygen targets – Based on the available evidence, we recommend an SpO₂ target range of 90 to 95 percent for most VPT neonates. The evidence supporting this recommendation and other issues related to oxygen targets in preterm neonates are discussed in detail separately. (See "Neonatal target oxygen levels for preterm infants".)

●Carbon dioxide targets – For most VPT neonates, we suggest a strategy of modest permissive hypercapnia. We aim for pCO₂ levels between 40 and 65 mmHg in the first few weeks of life. For older preterm infants with evolving BPD, it is reasonable to use more liberally permissive pCO₂ targets as long as the pH remains >7.25. The rationale for permissive hypercarbia is that it may limit the extent and duration of MV support, thereby reducing the risk of VILI.

The evidence supporting modest permissive hypercarbia in preterm neonates comes from randomized clinical trials and meta-analyses [36-38]. While a survival benefit or reduction in respiratory morbidity has not been clearly demonstrated, the available data suggest that this approach is safe and well tolerated in this population.

In one trial, 362 ELBW infants were randomized to a MV protocol targeting modestly permissive pCO₂ levels (40 to 50 mmHg on postnatal days 1 to 3, 45 to 55 mmHg on days 4 to 6, and 50 to 60 mmHg on days 7 to 14) or a protocol targeting more liberally permissive pCO₂ levels (55 to 65 mmHg on postnatal days 1 to 3, 60 to 70 mmHg on days 4 to 6, and 65 to 75 mmHg on days 7 to 14) [36]. Mortality at 28 days was lower in the modest permissive hypercapnia group, but the finding was not statistically significant (9 versus 12 percent; RR 0.72 [95% CI 0.39-1.33]). Rates of moderate to severe BPD at 36 weeks postmenstrual age (PMA) were similar in both groups (19 versus 22 percent; RR 0.87 [95% CI 0.58-1.3]). In a follow-up report describing outcomes at two years corrected age, mortality was similar in both groups (12 versus 15 percent; RR 0.77 [95% CI 0.45-1.32]), as were rates of neurodevelopmental impairment (48 versus 50 percent; RR 0.96 [95% CI 0.73-1.24]) [37].

In an earlier meta-analysis from 2001 that included two small trials with a total of 269 neonates randomized to permissive hypercapnia (defined in one trial as target pCO₂ 45 to 55 mmHg and as >52 mmHg in the second trial) or normocapnia (defined as pCO₂ 35 to 45 mmHg and <48 mmHg, respectively), the composite outcome of death or BPD at 36 weeks PMA was similar in both groups (56 versus 59 percent; RR 0.94 [95% CI 0.78-1.15]) [38].

Ongoing MV management

Monitoring — The adequacy of MV support is primarily determined by the success of maintaining adequate gas exchange. (See 'Gas exchange targets' above.)

Appropriate monitoring includes (table 3):

●Continuous SpO₂ monitoring – Peripheral oxygen saturation (SpO₂) monitoring with pulse oximetry has some limitations, but it provides a pragmatic approach to continuously monitor the adequacy of oxygenation without a need for blood sampling. It is the routine standard for oxygen monitoring during MV in most neonatal units. (See "Respiratory support, oxygen delivery, and oxygen monitoring in the newborn".)

●Serial physical examinations – The neonate's work of breathing should be assessed frequently. Marked tachypnea or labored breathing indicate that the neonate is not receiving adequate ventilatory support.

●Serial blood gases – Arterial blood gases (ABGs) measure pH, pCO₂, and arterial oxygen tension (PaO₂). The latter is the most accurate direct measure of oxygenation readily available in the clinical setting. However, ABGs require an arterial puncture (painful procedure) or an indwelling arterial catheter (risk of infection and vascular injury). Capillary and venous blood gases (CBGs and VBGs) are useful alternatives to ABGs to monitor pH and pCO₂, though both tend to modestly underestimate pH and overestimate CO₂. In our practice we typically use CBGs and VBGs for monitoring adequacy of CO₂ removal and we rely primarily on continuous pulse oximetry for monitoring of oxygenation. We do not routinely place indwelling arterial catheters to monitor PaO₂, unless the infant's clinical status necessitates blood sampling at least every 4 to 6 hours or there is another indication for placement of an arterial catheter (eg, hemodynamic instability requiring active titration of vasoactive medications).

●Noninvasive carbon dioxide monitoring – Transcutaneous CO₂monitoring (TCOM) and end-tidal CO₂ (EtCO₂) devices monitor CO₂ continuously and can provide a real-time indication of dynamic changes in pCO₂ levels. Dynamic changes in minute ventilation can result from changes in the neonate's clinical status and/or acute changes in ventilator support (eg, displaced ETT). These changes are particularly important in VPT infants, since both hypocapnia and hypercapnia can be associated with intraventricular hemorrhage. (See "Germinal matrix and intraventricular hemorrhage (GMH-IVH) in the newborn: Risk factors, clinical features, screening, and diagnosis".)

In our practice, we do not routinely use TCOMs or EtCO₂ devices, though we may use them in select circumstances such as neonates with severely compromised ventilation or when dynamic changes in pCO₂ levels are anticipated, particularly when transitioning to, or titrating, HFV. Of note, EtCO₂ monitoring is not feasible in infants receiving HFV; TCOM is the preferred noninvasive method for CO₂ monitoring in this setting.

TCOM and EtCO₂ monitors have become attractive options for noninvasive monitoring in critically ill neonates. Historically, use of TCOM in preterm infants was limited by technical limitations and concerns that use of heat on the TCOM sensor risks injury to the immature skin. Similarly, use of EtCO₂ monitoring in VPT neonates was limited due to concerns about the additional weight (and therefor risk of ETT displacement) and added dead space to the respiratory circuit. Technical advances have largely mitigated these concerns [39,40]. However, additional data on the accuracy and utility of these devices in VPT infants are needed before adopting them as a routine part of neonatal practice.

●Chest radiographs – Chest radiographs can provide a crude evaluation of lung volumes and can be used to inform decisions about ventilator settings (eg, PEEP if using CMV or mean arterial pressure [MAP] if using HFV). Chest radiographs also help identify adverse occurrences or complications of MV, such as air leaks and malpositioned ETTs. However, they should be obtained judiciously since their use can lead to handling and agitation of the newborn and potentially harmful radiation exposure.

●Ventilator data – For neonates managed with VTV, it is important to monitor the ventilator's peak inspiratory pressure (PIP) measurements to ensure that the neonate isn't exposed to excessively high pressures (ie, barotrauma). Some ventilators allow the clinician to set an upper PIP limit in VTV modes to minimize intermittent barotrauma or alert the provider of consistently high PIP needs through a ventilator alarm. Similarly, when using PLV, it is important to monitor exhaled Tv to ensure that breaths are not excessively small or large.

Titrating conventional MV — Based upon ongoing monitoring, we adjust MV settings to achieve the desired gas exchange targets (see 'Gas exchange targets' above):

●Optimizing oxygenation is accomplished primarily through lung volume recruitment (ie, increasing PEEP) to improve ventilation/perfusion (V/Q) matching. The fraction of inspired oxygen (FiO₂) also can be increased to improve oxygenation, but high FiO₂ levels (ie, >0.50) should generally be avoided.

●Optimizing ventilation (ie, carbon dioxide clearance) is accomplished primarily by adjusting minute ventilation, which is primarily determined by Tv and respiratory rate.

If the neonate has persistently inadequate or worsening gas exchange despite efforts to optimize CMV, or when the required CMV settings provoke high concern of VILI, we transition to HFV. (See 'Refractory respiratory failure' below.)

Weaning and discontinuation of MV — After the neonate has been stabilized and lung function improves, subsequent titration of MV support consists of a gradual reduction or "weaning” of the MV settings while maintaining gas exchange targets to eventual discontinuation of MV and extubation.

●Weaning – The aim of MV weaning is both to minimize VILI and to progress towards discontinuation of MV. In our practice, reduction of ventilatory setting is individualized and is based on clinical judgement. Although the practice of using of standardized MV weaning protocols is common in pediatric and adult populations, their role in preterm neonates is uncertain. In an international survey including 321 neonatal units embedded within 10 neonatal networks, fewer than 25 percent reported using a protocol to guide MV weaning [7]. Randomized trials comparing protocolized versus non-protocolized weaning in neonates are lacking [41]. In a single-center observational study of 301 preterm neonates receiving MV before and in the two years after implementing a respiratory therapist-driven MV weaning protocol, implementation of the protocol was associated with a shorter time to first extubation attempt, shorter overall duration of MV, and lower rate of extubation failure [42]. Rates of mortality and BPD were similar in both eras.

●Discontinuing MV – MV should be discontinued once it is anticipated that acceptable gas exchange can be achieved with noninvasive support. However, identifying precisely when this has occurred is challenging and there is no standard approach. If extubation is performed too early, the infant may experience poor gas exchange, clinical instability, and additional harm, including potential reintubation.

In our practice, MV is discontinued as soon as both of the following criteria are met:

•The neonate has no concrete contraindications to extubation (eg, severe neurologic impairment or airway anomalies requiring ongoing maintenance of an artificial airway)

•Adequate gas exchange is maintained with a low MV demand rate (typically 10 to 20 breaths per minute)

As discussed separately, discontinuation of MV in VPT neonates usually consists of extubation to noninvasive positive airway pressure (eg, nCPAP, NIPPV). (See "Respiratory distress syndrome (RDS) in preterm infants: Management", section on 'Noninvasive positive airway pressure'.)

Our approach to assessing extubation readiness is based largely on our clinical experience. We do not use spontaneous breathing trials (SBTs) or other readiness testing beyond our clinical judgement to inform the decision to proceed with a trial of extubation. The reluctance to extubate infants too early resulting in a high risk of extubation failure stems in part from concern that exposure to multiple tracheal intubations and courses of MV may place the neonate at greater risk of morbidity (eg, upper airway injury, risk of BPD, and organ injury from associated hemodynamic instability and hypoxemia). However, the number of intubations and extubations does not appear to increase the risk of BPD. An observational study in ELBW infants found that the risk of BPD was progressively higher for infants exposed to greater numbers of MV courses [43]. But the association was no longer significant after adjusting for the total duration of MV, except for infants exposed to ≥4 MV courses. This suggests that duration of MV impacts the risk of BPD more than the number of intubations and extubations.

An objective assessment tool that could accurately identify extubation readiness would be desirable since it could potentially shorten MV duration for some neonates, while maintaining adequate support in those at high risk of failure. However, based on the available data, SBTs and other readiness tests lack sufficient sensitivity and specificity to be clinically useful as a routine practice. [44,45].

A 2019 systematic review identified 35 studies reporting on 31 different clinical and physiologic assessments of extubation readiness in preterm neonates; the sensitivity ranged from 75 to 100 percent and the specificity ranged from 18 to 95 percent [44]. One of the most commonly used and studied assessment tools was the SBT, which consists of monitoring for apnea, desaturations, bradycardia, or increased oxygen requirement while the infant remains intubated and maintained on CPAP alone via the ETT. In the systematic review, the pooled sensitivity and specificity of the SBT were 95 and 62 percent, respectively [44]. In a subsequent prospective multicenter study of 259 preterm neonates <28 weeks GA who underwent SBTs, passing the SBT (defined as no apnea or desaturation requiring stimulation and no or minimal increase in oxygen requirement) predicted extubation success with a sensitivity and specificity of 93 and 39 percent, respectively [45]. In this cohort, the SBT misclassified 59 neonates (23 percent). Thus, the overall accuracy of the SBT (ie, the ability to correctly predict extubation success or failure) was 77 percent. The study authors concluded that the SBT adds little to other clinical assessments of extubation readiness.

Refractory respiratory failure

Transition to HFV — When efforts to optimize CMV settings fail to provide adequate gas exchange, or when the required CMV settings provoke high level of concern of VILI, we typically transition the neonate to a trial of high-frequency ventilation (HFV). The decision to use HFV as rescue therapy in this setting should be individualized based upon the clinical status of the neonate, comorbidities, and the experience of the providers in managing HFV. Observations that may prompt consideration of transitioning to HFV include:

●Requirement of PIP levels >25 cm H₂O to achieve target tidal volumes

●Requirement of PEEP levels >8 cm H₂O or FiO₂ >0.40 to >0.50 to meet SpO₂ targets

●Development of air leaks, such as pneumothorax or pulmonary interstitial emphysema

When the decision is made to use HFV for rescue therapy, we primarily use HFOV, but HFJV may be considered in neonates with pulmonary air leak. (See "Pulmonary air leak in the newborn".)

The general properties of HFOV and HFJV are discussed separately. (See "Overview of mechanical ventilation in neonates", section on 'High-frequency ventilation (HFV)'.)

Initiating and titrating HFV — In our practice, we typically use the following approach for initial settings and subsequent titration when using high-frequency ventilation (HFV) such as HFOV or HFJV in VPT neonates:

●HFOV:

•Initial settings:

-MAP is set 2 cm H₂O higher than the level provided with CMV and then adjusted to achieve adequate lung expansion on chest radiograph.

-The amplitude is started in the mid to high teens and increased steadily until chest wall vibrations are observed.

-The frequency is started at 12 Hz.

•Titration – HFOV settings are adjusted as follows to achieve gas exchange targets:

-CO₂ clearance – CO₂ clearance is primarily achieved by titrating the pressure wave amplitude which can be increased or decreased to achieve more or less CO₂ removal, respectively. Increased CO₂ elimination can also be achieved by decreasing the oscillatory frequency, although experts discourage this as a primary approach to avoid rapid fluctuations in CO₂ [46].

-Oxygenation – Oxygenation is primarily achieved by titrating the MAP.

●HFJV:

•Initial settings:

-Jet frequency – 360 to 420 breaths per minute (bpm), with lower rates applied to smaller neonates (eg, <750 g).

-Jet PIP – Initial setting is usually in the low 20s or a few cm H₂O higher than had been required during CMV.

-Jet pinch valve "on-time" – Typically 0.02 seconds.

-Sigh breath PIP – Usually set at or below typical CMV levels to minimize volutrauma.

-Sigh breath rate – 4 to 6 bpm.

-PEEP (conventional ventilator setting) – Initially set similar to PEEP level that was used during CMV, with an expectation that up-titration to a higher level than what is typical during CMV may be needed to facilitate oxygenation.

•Titration – HFJV settings are adjusted as follows to achieve gas exchange targets:

-CO₂ clearance is achieved by titrating the pressure difference (delta) between the jet PIP and PEEP, with larger delta pressures facilitating greater CO₂ elimination and lower pCO₂ values.

-Oxygenation – Oxygenation is primarily achieved by titrating the PEEP (conventional ventilator setting).

-Sigh breaths can be up-titrated for poor gas exchange, particularly if there is atelectasis, and down-titrated or discontinued if the neonate has adequate gas exchange and good lung recruitment. If the neonate has air leak, sigh breaths should be minimized.

If a trial of HFV is unsuccessful, we revert to CMV.

Efficacy of HFV as rescue therapy — Limited data are available on the use of high-frequency ventilation (HFV) such as HFOV and HFJV as a rescue therapy for neonates with refractory respiratory failure or air leak on CMV [24,47-49]:

●Efficacy of "rescue" HFOV therapy – In a multicenter trial conducted between 1988 and 1990, 176 preterm infants who were on high PIP or had developed pulmonary air leak on CMV were randomized to HFOV or continued CMV [47]. Fewer infants assigned to HFOV developed new or worsening pulmonary air leak (47 versus 64 percent; RR 0.73 [95% CI 0.55-0.96]). However, more patients in the HFOV group developed IVH (36 versus 20 percent; RR 1.77 [ 95% CI 1.06-2.96]), the majority of which were grade I. The trial did not detect differences between groups in 30-day mortality or the need for ongoing MV at 30 days. Of note, this trial was conducted prior to the routine use of antenatal corticosteroids and surfactant, and well before the era of preferential noninvasive support. This limits the generalizability of findings to clinical practice in the modern era.

●Efficacy of "rescue" HFJV therapy – In a multicenter trial conducted between 1987 and 1989, 144 preterm infants who developed pulmonary interstitial emphysema (PIE) on CMV were randomized to HFJV or ongoing CMV [49]. Crossover was permitted if prespecified failure criteria were met. Patients managed with HFJV were more likely to achieve treatment success (defined as resolution of PIE or improvement in chest radiograph with ability to wean MAP by ≥60 percent) compared with patients managed with CMV (61 versus 37 percent, respectively). Rates of new-onset pulmonary air leak, BPD, and grade III or IV IVH were not statistically different between both groups. Overall mortality in the trial was 33 percent and did not differ significantly between groups. However, when patients who crossed over to the alternative treatment were excluded from the analysis (17 patients in the HFJV group; 21 in the CMV group), mortality was lower in the HFJV group (35 versus 53 percent; RR 0.66 [95% CI 0.45-0.97]) [50]. Similar to the trial of rescue HFOV, this trial was conducted over 30 years ago, and the applicability of these findings to modern-day practice is uncertain.

Data on elective use of HFOV and HFJV are discussed above. (See 'Role of high-frequency ventilation' above.)

Disadvantages of HFV — Disadvantages of high-frequency ventilation (HFV) include the following:

●Risk of hemodynamic instability – HFV applies continuous high pressure to the lungs to achieve an "open-lung strategy." This high intrathoracic pressure can decrease systemic venous return (ie, reduce preload), which can negatively impact cardiac output and hemodynamic stability. These effects tend to be more pronounced in the setting of significant acidosis, which is sometimes present in neonates receiving rescue HFV therapy.

●Risk of brain injury – There is concern that use of HFV in preterm infants may be associated with increased risk of brain injury (eg, intraventricular hemorrhage [IVH] and periventricular leukomalacia [PVL]). Proposed mechanisms that may contribute to the risk of brain injury in this population include:

•Rapidly changing pCO₂ – HFV is very efficient in clearing CO₂ and neonates may experience rapid changes in pCO₂, particularly in the period after initiating HFV. High, low, or rapidly changing pCO₂ levels can negatively impact cerebral blood flow.

•Hemodynamic instability – As discussed above, HFV often negatively impacts cardiac output and hemodynamic stability.

Both processes can alter cerebral blood flow and cause brain injury from ischemia and/or excessive perfusion. However, many other factors contribute to brain injury in preterm infants and the relative contribution of HFV to the overall risk in this population remains unclear. The clinical trials described above had inconsistent findings, with some reporting similar rates of IVH and PVL in the HFV and CMV groups [13,15,25,49], while others reported increased risk of adverse neurologic outcome in the HFV group [16,26,47]. However, the applicability of these findings to modern-day practice is uncertain since these trials were conducted in an earlier era when antenatal corticosteroid therapy and preferential noninvasive support were not part of routine clinical practice. (See 'Efficacy of HFV as rescue therapy' above and 'Role of high-frequency ventilation' above.)

Additional details on IVH and PVL, including pathogenesis and risk factors, are provided in a separate topic review. (See "Germinal matrix and intraventricular hemorrhage (GMH-IVH) in the newborn: Risk factors, clinical features, screening, and diagnosis".)

●Cost and availability – HFV requires specialized ventilators, which can be costly and may not be available in some settings.

●Need for experienced staff – It generally is not feasible to use these ventilators if the respiratory therapists and other clinical staff lack appropriate training and experience.

●Limitations to routine clinical assessments – Certain routine clinical assessments are limited while on HFOV (eg, heart and lung examination).

INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at th^e 5^th t^o 6^th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the ¹0^th to ¹2^th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)

●Basics topic (see "Patient education: What to expect in the NICU (The Basics)")

SUMMARY AND RECOMMENDATIONS

●Importance – Respiratory failure is the primary cause of morbidity and mortality in very preterm (VPT) neonates (ie, ≤32 weeks gestational age). It results from multiple contributing factors, including surfactant deficiency, incomplete alveolarization, and an immature respiratory drive. Despite the trend towards increasing use of noninvasive respiratory support in managing VPT neonates, use of invasive mechanical ventilation (MV) remains common in this population. (See 'Respiratory failure in VPT neonates' above and "Respiratory distress syndrome (RDS) in the newborn: Clinical features and diagnosis".)

●Ventilator-induced lung injury – While lifesaving, MV can also cause lung injury and contribute to hemodynamic instability and brain injury. The impact of ventilator-induced lung injury (VILI) and the imperative for utilizing lung-protective strategies are particularly important for VPT neonates. Bronchopulmonary dysplasia (BPD) is a common and consequential morbidity of preterm birth caused by concurrent injury and maldevelopment of the immature lungs. (See 'General principles' above and "Bronchopulmonary dysplasia (BPD): Clinical features and diagnosis", section on 'Risk factors'.)

●MV management – MV management in VPT neonates is tailored to meet the needs of the individual neonate, which will differ between patients and within the same patient over time. Our general approach is as follows (table 3) (see 'Clinical approach' above):

•Indications – Initiation of invasive MV is generally required for neonates who, despite efforts to optimize noninvasive support, develop any of the following signs of inadequate gas exchange (see 'Indications for invasive MV' above):

-pH <7.20 with a PaCO₂ >65 mmHg

-Requiring FiO₂ >0.4 to 0.5 to achieve target oxygen saturation

-Multiple apneic episodes per hour associated with desaturations and bradycardia, or more than one episode requiring positive pressure ventilation within a few hours

•Choice of mode – Our approach to selecting an initial mode for MV in VPT neonates is as follows (see 'Choice of mode' above):

-We suggest conventional mechanical (CMV) rather than elective high-frequency ventilation (HFV) as the initial approach (Grade 2C). Elective HFV is a reasonable alternative for neonates at high risk of developing VILI. (See 'Role of high-frequency ventilation' above.)

-We suggest a synchronized mode that provides both mandatory and spontaneous breaths (ie, synchronized intermittent mandatory ventilation plus pressure support [SIMV + PS] or assist control ventilation [ACV]) rather than only mandatory breaths (ie, SIMV alone) (Grade 2C). (See "Overview of mechanical ventilation in neonates", section on 'Synchronized modes'.)

-We suggest volume-targeted ventilation (VTV) rather than pressure-limited ventilation (PLV) (Grade 2B). PLV is a reasonable option if VTV is not available or there is a technical challenge that limits the reliable delivery of measured tidal volumes (Tvs) such as a large endotracheal tube leak. (See 'Volume-targeted versus pressure-limited ventilation' above.)

•Initial settings – Typical initial settings are (see 'Initial conventional MV settings' above):

-Tv 4 to 6 mL/kg

-PEEP 5 to 6 cm H₂O

-Inspiratory time (Ti) 0.35 to 0.4 seconds

•Gas exchange targets – For most VPT neonates, we recommend a target oxygen saturation of 90 to 95 percent (Grade 1B) and we suggest a strategy of modest permissive hypercapnia (Grade 2C). In the first few weeks of life, target pCO₂ levels are between 40 and 65 mmHg. For older preterm infants with evolving BPD, it is reasonable to use more liberally permissive pCO₂ targets as long as the pH remains >7.25. (See 'Gas exchange targets' above and "Neonatal target oxygen levels for preterm infants" and "Bronchopulmonary dysplasia (BPD): Management and outcome".)

•Monitoring – Appropriate monitoring for neonates receiving MV includes (see 'Monitoring' above):

-Continuous pulse oximetry

-Serial physical examinations

-Blood gases

-Judicious use of chest radiography

-Monitoring of ventilator data

•Titrating MV – MV settings are adjusted as follows to achieve the desired gas exchange targets (see 'Titrating conventional MV' above):

-Optimizing oxygenation is accomplished primarily through increasing positive end-expiratory pressure (PEEP). The fraction of inspired oxygen (FiO₂) also can be increased to improve oxygenation, but high FiO₂ levels (ie, >0.50) should generally be avoided.

-Optimizing ventilation (ie, CO₂ clearance) is accomplished primarily by adjusting minute ventilation, which is determined by Tv and respiratory rate.

•Refractory respiratory failure – For VPT neonates who fail to achieve adequate gas exchange despite efforts to optimize CMV settings or require CMV settings that provoke a high level of concern for VILI (eg. air leak), we suggest a trial of HFV (Grade 2C). Use of HFV is limited to centers with the appropriate resources and experience with this modality. (See 'Refractory respiratory failure' above.)

•Weaning and discontinuing MV – After the neonate has been stabilized and lung function improves, MV settings can be weaned gradually towards eventual discontinuation of MV and extubation to noninvasive support. In our practice, MV is discontinued as soon as both of the following criteria are met:

-The neonate has no concrete contraindications to extubation (eg, severe neurologic impairment or airway anomalies requiring ongoing maintenance of an artificial airway)

-Adequate gas exchange is achieved with a low MV demand rate (typically 10 to 20 breaths per minute)

ACKNOWLEDGMENT — The editorial staff at UpToDate acknowledges James Adams, Jr., MD, who contributed to an earlier version of this topic review.