Noninvasive respiratory support refers to support provided to the nasal airway opening of spontaneously breathing infants in the absence of an endotracheal tube. This support consists of continuous positive airway pressure (CPAP), noninvasive intermittent mandatory ventilation (NIMV), also known as noninvasive intermittent positive-pressure ventilation (NIPPV), noninvasive high-frequency ventilation, and noninvasive neurally adjusted ventilatory assist (NIV-NAVA). Although not typically classified as noninvasive respiratory support, humidified high-flow nasal cannula (HHFNC) may provide positive pressure and respiratory assistance as well and is discussed in this chapter. All noninvasive respiratory support devices provide some distending pressure to the lungs, thus increasing transpulmonary pressure during the expiratory phase of breathing. The basic goal of using these devices is to recruit collapsed alveoli and terminal airways, maintain end-expiratory lung volume, preserve gas exchange, and minimize work of breathing. Noninvasive respiratory support also may mitigate lung injury and inflammation by avoiding shear injury (atelectrauma) or repetitive cycling of the lungs at low end-expiratory lung volume. Also, distending pressure helps maintain upper airway patency and reduce obstructive apnea. By avoiding endotracheal intubation, the risks of airway trauma, infection, and airway emergencies may be reduced.
Nasal CPAP (NCPAP) is the most prevalent noninvasive form of respiratory assistance and also the most widely debated. It has been used to support spontaneously breathing infants with lung disease for more than 40 years. Following reports that mechanical ventilation contributes to pulmonary growth arrest and the development of chronic lung disease, there is a renewed interest in using NCPAP as the prevailing method for supporting newborn infants. Animal research has shown that NCPAP is less injurious to the lungs than is mechanical ventilation. A report in the presurfactant era suggested that early NCPAP, as an alternative to invasive ventilation, might mitigate lung injury and development of chronic lung disease (CLD) in premature infants, and animal studies supported this view. Three clinical trials suggested that CPAP use was no worse and perhaps better than immediate mechanical ventilation in most infants A meta-analysis found decreased death and CLD in infants stabilized and maintained on NCPAP after birth. Despite the successes, little is known about how best to manage patients using NCPAP or other noninvasive strategies. It is also unclear whether some devices used to maintain NCPAP may be better than others in improving outcomes. Additionally, neonates have a high prevalence of NCPAP failure (∼35% to 50%). Clinicians have devised new intermediary approaches to support neonates that would otherwise fail NCPAP and receive invasive ventilation. These intermediary forms of respiratory support demonstrated increased transpulmonary pressure during inspiration and expiration and will also be reviewed.
Background and Historical Aspects
Many neonatal clinicians believe that noninvasive respiratory support is a relatively recent innovation, but CPAP, though not known by that name at the time, was described for use in newborn infants over a century ago. In his 1914 textbook on diseases of the newborn infant, Professor August Ritter von Reuss describes an apparatus ( Fig. 17-1, A ) that is virtually equivalent to the “bubble CPAP” that is used today ( Fig. 17-1, B ). CPAP helps maintain the functional residual capacity (FRC), thus helping to mitigate the natural physiologic reflex, “grunting,” that is frequently exhibited in infants with low lung compliance and low end-expiratory lung volumes (EELVs). Grunting is the dynamic expiratory braking phenomenon resulting from vocal cord adduction and diaphragmatic contraction, which limits airflow during exhalation and maintains transpulmonary pressure and EELV above the critical closing volume of the lungs. Early attempts to replicate the beneficial effects of grunting resulted in the first widely used CPAP systems developed by Gregory et al. In 1971, Gregory et al. were the first to report the successful application of CPAP provided through an endotracheal tube or a head box in a series of spontaneously breathing premature infants with respiratory distress syndrome (RDS) during an era when mortality rates of 60% were common in premature infants receiving invasive mechanical ventilation.
Use of CPAP in neonates during the 1970s was welcomed with enthusiasm as the “missing link” between supplemental oxygen and mechanical ventilation to treat RDS. During this decade a simple, noninvasive approach to providing CPAP was widely used, by application via binasal prongs or oronasal mask. Alternative methods of providing CPAP were occasionally described (see subsequent section on delivery of CPAP). During the 1970s, it was commonly believed that air leaks (such as pneumothoraces) were more common with CPAP than with mechanical ventilation. Gastric distension during CPAP was also frequently observed. Hard nasal prongs, nasopharyngeal orotracheal tubes, and oronasal masks often were not tolerated well by neonates. The head chamber (head box) and face chamber, although noninvasive, never gained wide acceptance because of technical difficulties and mechanical disadvantages. The head chamber seals around the infant’s neck, thus limiting access to the child’s face. It is also difficult to administer in infants weighing less than 1500 g. The devices are very noisy and have been associated with complications such as hydrocephalus, nerve palsies, and local neck ulceration from mechanical compression by the neck seal. The face chamber was originally described by Alhstrom et al. and consists of the application of CPAP via a mask covering the entire face. The mask is held in place by negative pressure. This system is simple and effective, and there are no reported patient complications or mechanical problems such as loss of pressure during administration. There is reported success in using the face chamber for treating RDS and in weaning infants. The major limitations to both methods are lack of access to the infant’s face and the cumbersome method of administration. Other approaches included use of continuous negative pressure applied to the infant’s chest wall using a chest wall chamber.
Intermittent mandatory ventilation using an endotracheal tube was first described in the early 1970s. It quickly became the standard of care for supporting the lungs of sick newborn infants and remained so for nearly 3 decades. For these reasons the use of CPAP fell out of favor during this period. Though Kattiwinkel et al. and Caliumi-Pellegrini et al. described the initial experiences using soft short binasal prongs to deliver CPAP in the early 1970s, renewed interest in noninvasive support did not occur until the late 1980s following the report by Avery that infants treated at a hospital using NCPAP seemed to have better outcomes. This ushered in a proliferation of new strategies and more widespread acceptance of NCPAP as a gentle way to maintain airway patency and allow sufficient gas exchange. CPAP and other noninvasive strategies are at the center of some of the mostly intensely focused research in neonatal medicine. Premature infants with RDS represent the most widely studied patient population for all forms of noninvasive support. However, NCPAP is used for other respiratory disorders including transient tachypnea of the newborn, meconium aspiration syndrome, primary pulmonary hypertension, pulmonary hemorrhage, patent ductus arteriosus, and consequent pulmonary edema. NCPAP improves lung function following surgical repair of congenital cardiac anomalies and paralysis of a hemidiaphragm and is also an effective option for infants following surgical repair of diaphragmatic hernias. NCPAP is also effective in managing infants with respiratory insfections, such as congenital pneumonias and respiratory syncytial virus bronchiolitis. NCPAP is useful for treating obstructive and central apnea of prematurity and congenital and acquired airway lesions. NCPAP is often used for the management of laryngo-, broncho-, and/or tracheomalacia. The positive pressure will distend these large airways and mitigate their tendency to collapse, particularly during expiration. Conceptually, NCPAP should help neonatal pulmonary disorders in which there is excessive lung fluid, including not only transient tachypnea of the newborn and patent ductus arteriosus, but also congestive heart failure, hydrops fetalis, and other causes of pulmonary edema. Data, however, are lacking. NCPAP is contraindicated in patients with upper-airway abnormalities (i.e., cleft palate, choanal atresia, tracheoesophageal fistula), unrepaired diaphragmatic hernia, severe cardiovascular instability, severe apneic episodes, and severe ventilatory impairment (pH <7.25, and PaCO 2 >60 mm Hg).
Physiologic Effects of CPAP
Several short-term studies have evaluated outcomes in infants supported with NCPAP. NCPAP has been shown to reduce tachypnea, increase FRC and PaO 2 , decrease intrapulmonary shunting, improve lung compliance, and aid in the stabilization of the highly compliant infant chest wall. NCPAP also decreases thoracoabdominal asynchrony and labored breathing index.
Apnea of prematurity (AOP) is a common disorder in premature infants born before 34 weeks’ gestation. These infants exhibit various combinations of apnea, bradycardia, and oxygen desaturation. Apnea is classified as obstructive, central, or mixed. Methylxanthines are effective in treating AOP. The sole trial comparing CPAP with methylxanthine therapy was performed more than 25 years ago. In that trial, face mask CPAP at the very low levels of 2 to 3 cm H 2 O was compared with theophylline in 32 infants of 25 to 32 weeks’ gestation. The investigation found theophylline to be more effective than face mask CPAP with the low setting in reducing (1) prolonged apnea episodes, (2) the need for intubation and ventilation because of worsening AOP, and (3) the number of bradycardia spells. The Cochrane review regarding CPAP use for AOP concludes that this topic needs additional evaluation. There is widespread use of NCPAP for management of AOP despite the dearth of supportive evidence. Kurtz et al. evaluated the effect of discontinuing CPAP and found that when infants were supported by CPAP they had significantly lower respiratory rates, fewer obstructive apneas, shorter central apneas, and less severe apnea-associated desaturation and spent more time in a state of normal quiet breathing than infants breathing without NCPAP. NCPAP is effective for obstructive apneas because it splints the upper airway open, thereby reducing the risk for pharyngeal or laryngeal obstruction.
Clinical Management of Patients on Nasal CPAP
Clinical management of infants supported by CPAP is based on decades of experience and is often regarded as more of an art than a science. As such, management strategies vary greatly from one institution to another. However, following the publication of several large clinical studies described below, there is more definitive evidence to support clinical management of patients receiving noninvasive support than ever before. Premature infants should be stabilized on NCPAP in the delivery room. Additionally, NCPAP is generally indicated in infants with increased work of breathing (WOB), substernal and suprasternal retractions, grunting, and nasal flaring. The chest radiograph may show poorly expanded and/or increased lung opacification.
Ongoing management for optimal NCPAP levels is based on adequacy of lung inflation without overdistending the lung parenchyma. Determining the optimal level of NCPAP is a technically challenging process, because few objective measurements exist to determine adequacy of lung volume recruitment. Blood gases and chest X-rays can be helpful in determining patient response to NCPAP; however, frequent X-rays and blood gases can also be detrimental to neonatal patients because of repositioning needs, exposure to ionizing radiation, and blood loss. Transcutaneous monitoring of CO 2 and pulse oximetry offer reliable correlates for determining gas exchange in patients supported by NCPAP. In practice, oxygen requirement, which reflects ventilation perfusion matching, is a good proxy of adequacy of lung aeration. The goal is to keep FiO 2 below 0.30 to 0.40 by increasing the NCPAP level stepwise up to 8 cm H 2 O, if necessary. Because of the lack of physiologic monitoring, many institutions have embraced respiratory scoring tools, such as the Silverman–Anderson respiratory severity score, to guide clinical management. These scores have been shown to have good reliability between clinicians and can be useful for determining when the patient requires support, ongoing settings adjustments, or weaning during noninvasive support.
Proper airway management is perhaps the single most important aspect for improving outcomes and reducing complications in infants receiving NCPAP. This becomes particularly important because infants are now being supported on NCPAP for long periods of time. Clinicians caring for infants receiving NCPAP must be mindful of selecting the proper prong size. Prongs should fill the entire nares without blanching the tissue. Prongs that are too small will not provide NCPAP pressure because of leakage. They can also increase airway resistance and the imposed WOB and be more easily dislodged. It is essential to secure the prongs well to avoid dislodgment while keeping pressure off the nasal septum.
As previously mentioned, the 1987 publication of Avery et al. surveyed eight neonatal intensive care units (NICUs) to assess the incidence of CLD. The frequency of CLD in that report was lowest at Babies and Children’s Hospital, Columbia University, New York, USA. That center reportedly used NCPAP considerably more often than the other seven NICUs. Many clinicians have been influenced by the “Columbia” approach in which bubble CPAP is used early in the course of respiratory distress of both premature and term-gestation infants.
As part of this strategy, clinicians often accept hypercarbia with PaCO 2 levels up to 65 mm Hg (8.7 kPa) or even higher, PaO 2 levels as low as or lower than 50 mm Hg (6.7 kPa), and pH values as low as 7.20. This general approach has been used in that institution for more than 30 years. Despite the promulgation and widespread acceptance of this approach, to date there are no published randomized, controlled trials (RCTs) that validate its superiority over any other management strategy or technology. There are no long-term outcome studies comparing neurologic, pulmonary, and other findings among infants treated in this manner with others who are managed differently. Additionally, clinicians should be concerned about the potentially deleterious effects of such high PaCO 2 levels, which are beyond the usual limits of “permissive hypercarbia” and can affect cerebral autoregulation and the developing brain (see Chapter 42 ).
Van Marter and colleagues assessed the differences in outcomes between the Columbia NICU and two NICUs in Boston. Although CLD was less common at Columbia, this review has been criticized because of differences in patient populations, indications for mechanical ventilation, and other treatment strategies, as well as the definition of CLD that was used. Much of the apparent success of the Columbia approach has been attributed to the diligent management of sick neonates by a single senior clinician. A rigorously designed RCT is sorely needed to assess whether bubble CPAP, as opposed to other forms of noninvasive respiratory support, will truly prevent or mitigate CLD. Nevertheless, knowledge of the Columbia experience has contributed to the flurry of research concerning NCPAP since 1990.
Methods of Generating Continuous Distending Pressure
Following Gregory et al.’s initial publication demonstrating success using CPAP in premature infants, efforts were made to simplify the manner in which continuous distending pressure was generated, as well as the mode of delivery. In the mid-1970s Kattwinkel et al., as well as Caliumi-Pellegrini and colleagues, described devices in which binasal prongs were used for delivery. Binasal prongs connected to a ventilator for flow and thus pressure delivery were standard for a number of years. In the subsequent section, these and other methods of NCPAP delivery are described.
The pressure delivered via NCPAP can be derived from either a continuous flow or a variable flow source. From the 1970s through the 1980s, only continuous flow was used. Continuous-flow NCPAP consisted of gas flow generated at a source (usually with an infant ventilator) and directed against the resistance of the expiratory limb of the NCPAP circuit. In ventilator-derived NCPAP, variable resistance in a valve is adjusted to provide this resistance to flow.
A second method of continuous-flow NCPAP is bubble or water-seal CPAP (see Fig. 17-1, B ), the method advocated at the Columbia University NICU. With bubble NCPAP, blended gas flows to the infant after being heated and humidified. Typically, binasal prongs, such as the Hudson prongs (Hudson Respiratory Care, Inc., Arlington Heights, Illinois, USA) ( Fig. 17-2 ) or Inca prongs (Ackrad Laboratories, Inc., Cranford, New Jersey, USA), are secured in the infant’s nares. The distal end of the expiratory tubing is immersed under either 0.25% acetic acid or sterile water to a specific depth to provide the approximate level of CPAP desired. Clinicians must be cautious when using this method, however, because the level of CPAP may be higher than the submerged depth of the expiratory tubing and is flow dependent with some systems but not all. Bench studies of bubble NCPAP have shown that different systems produce inherently different high-frequency pressure profiles. In a neonatal lung/anatomic nasal airway model affixed with a leaky nasal airway interface, detectable high-frequency pressure oscillations were observed in the lung. A new device has been developed that increases the amplitude of pressure oscillations by altering the configuration by which gas exits into a water column. These pressure oscillations have been shown to deliver volumes that are similar to those generated by high-frequency ventilation. Compared with standard bubble NCPAP, this form of bubble NCPAP has also been shown to substantially reduce WOB in a surfactant-deficient, lung-injured animal model.
Traditionally, bubble NCPAP systems have incorporated pressure generators from ventilator water-seal positive end-expiratory pressure (PEEP) valves or by using homemade systems devised from sterile water bottles. Today, there are three commercially available bubble NCPAP systems. Because of the low cost of maintenance, the simplicity, and no requirement for an electrical power source, these devices are also frequently used to support patients in resource-limited settings (see Chapter 38 ).
Lee and colleagues observed vibrations of infants’ chests during bubble CPAP at frequencies similar to those used with high-frequency ventilation. Compared to ventilator-derived CPAP, Lee’s group found that bubble CPAP resulted in decreased minute ventilation and respiratory rate. These authors speculated that the observed vibrations enhanced gas exchange. Pillow et al. described similar findings in the lamb model. However, in both of these studies, bubble CPAP was delivered via a nasopharyngeal tube, not nasal prongs. Data obtained using an NCPAP model suggest that these oscillations are quite minimal and unlikely to contribute in a significant way to ventilation. Morley et al. assessed bubble CPAP in a randomized, crossover trial. The bubbles were generated at various rates, from “slow” to “vigorous.” These investigators found that bubbling rates had no effect on carbon dioxide, oxygenation, or respiratory rate. The gas-exchange mechanisms of the bubble NCPAP setup must be further explored to elucidate whether there is a to-and-fro oscillatory waveform that truly augments ventilation. One study supporting bubble NCPAP is that of Gupta et al., who evaluated the efficacy and safety of bubble NCPAP compared with the Infant Flow NCPAP system (described below) for the postextubation management of preterm infants with RDS. Extubation failure rate was lower and the duration of support was shorter in infants ventilated <14 days when supported with bubble CPAP following extubation.
A major concern with bubble NCPAP is that these systems may not reliably monitor pressure or provide pressure-relief valves or alarms. This may place infants at greater risk for excessive pressure delivery. As mentioned previously, many noninvasive devices do not provide clinical monitoring of pressure or have alarms. Thus, it is important to provide continuous physiologic monitoring. In a bench model, condensation forming in the expiratory limb of a commercially available bubble NCPAP system resulted in substantially higher NCPAP levels than desired. Whenever possible, stand-alone pressure manometers, alarms, and pressure-relief devices should be used. Also, bedside clinicians must frequently empty the exhalation limb of condensate, provide water traps, or use circuits that incorporate heated wires or are constructed from material that wicks moisture to the environment.
Since 1995, variable-flow NCPAP has come into widespread use. The technique was developed by Moa et al. to reduce the patient’s WOB. NCPAP is generated by varying the flow delivered to the infant’s nares and a specially constructed nosepiece is employed. These devices use the Bernoulli effect and gas entrainment via dual injector jets directed toward each nasal prong to maintain a constant pressure ( Figs. 17-3 to 17-7 ). With the variable-flow system, when the infant makes a spontaneous expiratory breathing effort, there is a so-called “fluidic flip,” which causes the flow of gas going toward the nares to flip around and to leave the generator chamber via the expiratory limb ( Fig. 17-7, A and B ), thus assisting exhalation. This phenomenon is due to the Coandă effect, which describes the tendency of a fluid or gas to follow a curved surface. A residual gas pressure is provided by the constant gas flow, enabling stable NCPAP delivery at a particular pressure during the entire respiratory cycle.
An extensive description of the physiology of variable-flow CPAP can be found elsewhere. Klausner et al. used a simulated breathing apparatus and found the WOB via nasal prongs to be one-fourth that of continuous-flow NCPAP. Pandit et al. assessed WOB in premature infants treated with either continuous-flow or variable-flow NCPAP. They found the WOB to be significantly less with variable-flow NCPAP. Additionally, the variable-flow devices appear to be able to maintain a more uniform pressure level compared to continuous-flow NCPAP. This may be the reason for the improved lung recruitment seen with variable-flow NCPAP of this type.
As of this writing two variable-flow NCPAP systems are commercially available. The Infant Flow has been the most extensively evaluated and is marketed by Cardinal Health (Dublin, Ohio, USA). The Arabella system (Hamilton Medical, Reno, Nevada, USA) has a flow-generating chamber that varies slightly from the Infant Flow system (IFS), although the same principles (Venturi, Bernoulli, and Coandă) apply. These two systems appear to function similarly.
Several investigators have assessed whether differences exist among the various methods of delivering NCPAP. Liptsen et al. compared WOB in bubble vs variable-flow NCPAP in 18 premature infants. These investigators found more labored and asynchronous breathing with bubble NCPAP compared to variable-flow NCPAP. Boumecid and colleagues compared variable-flow NCPAP with ventilator-driven, continuous-flow NCPAP. They described increased tidal volume and improved breathing synchrony with the variable-flow device compared to the ventilator-driven NCPAP. On the other hand, Stefanescu and colleagues found identical rates of extubation failure in preterm infants who had been weaned from mechanical ventilation and randomized to continuous-flow or variable-flow NCPAP.
It is important to note that despite all of the published research showing differences in short-term physiologic outcomes in infants supported with the array of NCPAP devices, there are no definitive data to support or refute using one NCPAP system over another. Moreover, the studies cited above were performed in infants of different gestational ages and weights, which may have had a significant effect on the outcomes reported. Rather than be influenced by these relatively limited studies, it is more important that clinicians familiarize themselves with the equipment they are using and understand the limitations of each device.
Nasal Airway Interfaces
Multiple nasal devices are available through which continuous-flow NCPAP may be delivered. The devices may be either short (6 to 15 mm) or long (40 to 90 mm). However, long nasal prongs are not recommended because of the high imposed WOB. It is probably more accurate to refer to the former as nasal prongs and to the latter as nasopharyngeal prongs. A single nasopharyngeal prong typically consists of an endotracheal tube that has been cut and shortened and then inserted through one of the nares into the nasopharynx. However, this practice is less common following the advent of binasal short prongs and a Cochrane review suggesting that binasal prongs are more effective.
Nasal prongs commonly used with bubble NCPAP are depicted in Figure 17-8 . The nasal prongs used with the Infant Flow driver are depicted in Figs. 17-4 and 17-5 . Unfortunately, few comparative data are available to guide clinicians in choosing one type of prong over another. Some prongs, such as those used with the IFS, are specific to the device ( Fig. 17-9 ). Prongs may vary in the type of material, length, configuration, and diameters (both inner and outer). These aspects will affect the resistance to flow in a particular device and, as a result, the pressure entering the device may differ considerably from that entering the child’s nares or nasopharynx. DePaoli et al. compared the pressure drop for five different CPAP devices at various rates of gas flow. These authors found great variation among devices in the pressure drop. Although the least amount of drop-off occurred with the IFS, these authors cautioned that their findings do not establish clinical superiority of one mode of NCPAP or nasopharyngeal CPAP (NPCPAP) over any other. DePaoli and colleagues have published a more in-depth appraisal in their Cochrane review characterizing NCPAP devices and pressure sources. As stated above, they concluded that binasal short prongs are more effective than the nasopharyngeal prong for avoiding reintubation. Binasal short prongs remain the most common method of administering NCPAP in neonates. Because infants are generally obligate nose breathers, NCPAP may be facilitated when delivered directly into the nose. The most common complications are obstruction by secretions and skin breakdown at the nasal septum. These complications can be prevented by careful attention to skin care, nasal suctioning, and use of a protective barrier.
Nasal masks are a relatively recent innovation available with the variable-flow systems. A small, soft mask is attached to the pressure generator ( Fig. 17-10, A and B ). Such masks are markedly smaller than face masks; hence there is little additional dead space. Nasal masks may be useful when the infant’s nares are too small to accept the nasal prongs. Some units also use them in conjunction with nasal prongs, alternating several hours on and off each device to minimize the pressure effects of the prongs on the nares. However, a good seal must be present to prevent pressure loss with the nasal mask. There are no published data concerning the safety and efficacy of nasal masks.
Nasal cannulae (NC) are typically used to provide supplemental oxygen ( Fig. 17-11 ). However, depending on the flow rate, size of the NC, degree of leak, and size of the nares, these devices also provide some distending pressure. As no pop-off valve is present on currently available NC, pressure generated is uncontrolled and may be substantial. Some high-flow NC systems do have a pop-off valve. These will be discussed later in this chapter. Cannulae can also be easily dislodged; it is not unusual to pass by a child being treated with NC and to note that the cannulae are not in the nares but on the cheek or in the mouth or elsewhere.
The RAM Nasal Cannula (Neotech, Valencia, California, USA) was originally marketed for use with oxygen therapy but has been shown to be a useful interface for NCPAP and other forms of noninvasive ventilation. It is a short binasal prong designed with a larger bore tubing than standard oxygen or high-flow NC. The resistance and dead space of these prongs are reported to be similar to those of an endotracheal tube, but the dead space consideration is probably less critical, because infants are likely to exhale around the prong, not through the long narrow tubing. One bench study has shown that CO 2 removal is less efficient during simulated NIMV with a RAM cannula compared to binasal short prongs but there were no differences during NCPAP. Other investigators have shown that even small leaks with the RAM cannula result in large reductions in pressure. However, effective pressures can usually be maintained with small leaks, despite the added tubing length and resistance. This device has gained widespread acceptance to provide NCPAP and NIMV because it is relatively easy to maintain and is fixated similar to an NC. Anecdotal reports suggest that this nasal airway interface is less injurious to the nasal airway than those that require more complex fixation techniques. However, the long segment of narrow tubing leading from the circuit connector to the prongs creates substantial resistance, such that there is a noticeable drop in pressure from the circuit to the patient interface, especially with the smallest size cannula. It has been observed anecdotally that when used for NIPPV/NIMV, the ventilator will be cycling and registering substantial peak inflation pressure and therefore not triggering an alarm even when the prongs are completely out of the patient’s nose. Consequently bedside staff may be unaware that the infant is not receiving adequate support until oxygen saturation or bradycardia alarms sound. Studies in humans need to assess gas exchange, minute ventilation, and WOB among different nasal airway interfaces.
Infants may lose pressure through their open mouths while undergoing NCPAP or other forms of support. Thus many clinicians actively try to prevent pressure loss by means such as placing a pacifier in the child’s mouth or using a strap under the infant’s chin to close the mouth ( Fig. 17-12 ). Fortunately, when NCPAP and NPCPAP are applied, there is often enough downward pressure on the palate that it is frequently contiguous to the tongue, providing a natural seal with minimal to no pressure loss through the mouth.
Clinical Use of CPAP: Randomized, Controlled Trials
Several large RCTs have been performed assessing NCPAP for resuscitation or early management in the delivery room and for early management of RDS.
Early CPAP with Rescue Surfactant
Mechanical ventilator-induced lung injury is created by excessive tidal volumes (volutrauma) and/or repetitive cycling of the lungs using insufficient volumes and end-expiratory pressure (atelectrauma), thus propagating the release of inflammatory mediators (biotrauma) in the lungs. Short-term exposure to excessive delivered tidal volume during ventilation can exacerbate lung injury and compromise the therapeutic effect of surfactant replacement therapy. Additionally, oxidative stress can occur from excessive FiO 2 administration in the newborn lungs and impair or arrest lung development. These forms of injury have been implicated as major causes of CLD and other morbidities associated with ventilation of premature lungs (see Chapter 30 ).
To investigate the use of NCPAP immediately after birth, Morley and colleagues [CR] randomized 610 premature infants (25 to 28 weeks’ gestation) in the delivery room at 5 minutes of age to NCPAP or to intubation and ventilation in the CPAP or Intubation (COIN) trial. The NCPAP infants were initially treated with either a short single nasal prong or binasal prongs at a pressure of 8 cm H 2 O. The primary outcome was the combined endpoint of death or bronchopulmonary dysplasia (BPD), defined as the need for oxygen at 36 weeks’ postmenstrual age. Although there was a trend in CPAP babies toward less death/BPD (34% vs 39%), this difference was not statistically significant. There was a 50% decrease in the use of surfactant in CPAP-treated neonates ( p < 0.001). Although there was a significant decrease in the number of ventilator days in the CPAP group, this difference averaged only 1 day. Of note, the CPAP-treated infants were significantly more likely to develop pneumothoraces (9% vs 3%, p < 0.001). Although this is often attributed to the higher initial NCPAP pressure, the more likely explanation is the high threshold for intubation and rescue surfactant (FiO 2 0.60). Many of the pneumothoraces occurred on days 2 to 3 in infants with a high and increasing oxygen requirement, indicative of extensive atelectasis that would result in maldistribution of tidal volume and thus predispose to overdistension and air leak.
In 2010, the National Institute of Child Health and Human Development Neonatal Research Network published the largest trial to date, the SUPPORT trial. In this study 1316 infants were randomized in a factorial design to NCPAP in the delivery room vs intubation and surfactant; participants were also assigned to one of two ranges of oxygen saturation. The primary outcome variable was death or CLD. The early NCPAP and surfactant groups did not differ in the primary outcome, and no increase in pneumothoraces was found. In this trial NCPAP was usually started at 5 cm H 2 O, but rescue intubation and surfactant administration occurred at a lower threshold than in the COIN trial. Though two-thirds of the infants in the NCPAP group did ultimately receive surfactant, this group had a shorter duration of mechanical ventilation and less use of postnatal steroids. In post hoc analysis, the infants at 24 and 25 weeks who were randomized to NCPAP had a lower death rate. Importantly, there was no difference in death or neurodevelopmental impairment at 18 to 22 months’ corrected age and less respiratory morbidity in the NCPAP group.
The Vermont–Oxford Network Study Group studied 648 infants at 27 centers, comparing in a three-arm protocol prophylactic surfactant/mechanical ventilation, prophylactic surfactant/extubation, and bubble NCPAP followed by intubation/surfactant if necessary (the Delivery Room Management or DRM trial). There were no differences in mortality, CLD, or other complications for any of the three groups. Both intubation and surfactant were needed less in the NCPAP group. This study was stopped early before the planned sample size of 876 infants, because of declining enrollment.
These large studies indicate that, at the very least, early NCPAP is equivalent to early intubation in most cases of newborns of at least 24 weeks’ gestation and may be preferable. Additionally, a meta-analysis of prophylactic vs selective surfactant demonstrated less CLD/death when early NCPAP was used, followed by selective surfactant, if needed.
If surfactant is needed, a common approach is the INSURE technique, an acronym whereby infants are in tubated, given sur factant, and then e xtubated to NCPAP. This strategy results in improved short-term outcomes, though long-term outcomes have not been assessed. Importantly, clinicians should be aware that very immature infants or critically ill infants may need continued intubation and ventilator support.
Until now, surfactant administration has required intubation for drug instillation. However, many other techniques for surfactant administration are now being investigated. Most of these techniques involve the simultaneous application of NCPAP. These alternatives include the use of a small tube passed through the cords, a laryngeal mask, pharyngeal administration, and aerosolized surfactant. None of these techniques has yet been shown to be both equivalent in effect and safer than intubation; however, as of this writing these studies are in their early stages.
Given the randomized trials and Cochrane meta-analyses showing lower rates of CLD/death compared to prophylactic or early surfactant, the American Academy of Pediatrics (AAP) now recommends using NCPAP immediately after birth, with selective surfactant as needed. Intubation solely for the purpose of prophylactic surfactant administration is no longer recommended in the era of early NCPAP; the practice was based on older studies in which the late surfactant group did not receive distending airway pressure, a practice that predictably led to atelectasis and more severe RDS. Additionally, the AAP concludes that there is no evidence for increased adverse outcomes and that, in fact, early NCPAP may lead to a reduction in the duration of mechanical ventilation and use of postnatal steroids.
Noninvasive ventilation (NIV) or NIPPV is a form of support that is typically used for patients failing NCPAP or following extubation from invasive (i.e., intubated) ventilation. Synchronized and nonsynchronized NIMV with a conventional ventilator is the most common form, but some institutions have begun using high-frequency ventilators as well. The concept is attractive: NIV would avoid potential complications of prolonged ventilator support via an endotracheal tube (volutrauma, subglottic stenosis, infections). NIV is not a new concept. It was initially described in the early 1970s when clinicians applied time-cycled pressure-controlled inflations using a ventilator via an oronasal mask. In the mid-1980s, more than half of the level III NICUs in Canada were using this technique. However, this practice was associated with increased risk of neurological injury from the mask fixation. In 1985 Garland and colleagues reported an increased risk of gastrointestinal perforation among infants ventilated noninvasively with either nasal prongs or a face mask. Of note, however, subsequent publications concerning NIV have not confirmed higher rates of this complication. Moreover, NIV may have advantages over NCPAP in stabilizing a borderline FRC, reducing dead space, preventing atelectasis, and improving lung mechanics. Compared with NCPAP, NIV has been shown in some short-term studies to be associated with larger tidal and minute volumes, reductions in thoracoabdominal asynchrony (chest wall stabilization) and respiratory rates, as well as improved gas exchange and reduction in WOB. In general, NIV has been studied to determine its potential usefulness (1) in preventing extubation failure, (2) in treating AOP, and (3) as a primary mode of treating respiratory disorders.
The most widely used and studied form of NIV is synchronized and nonsynchronized NIMV. The majority of clinical trials in humans have compared outcomes in premature infants between NIMV and NCPAP. Friedlich et al. randomized 41 premature infants, after extubation, to either NPCPAP or nasopharyngeal synchronized mandatory ventilation (NPSIMV). These authors used the Infant Star ventilator (Infrasonics, Inc., San Diego, California, USA) with the StarSync abdominal capsule-triggering device (Graseby capsule, Infrasonics, Inc.) for synchronization. Binasal nasopharyngeal prongs were used in both groups. Treatment failure was defined as one of multiple parameters: (1) pH of 7.25 or less, (2) increased PaCO 2 , (3) increased FiO 2 requirement, (4) need for an NPSIMV rate greater than 20/min, (5) need for a peak inflation pressure on NPSIMV of 20 cm H 2 O or more, (6) need for PEEP on NPSIMV of 8 cm H 2 O or more, or (7) severe apnea. They reported significantly fewer extubation “failures” with NPSIMV (1/22, 5%) compared to NPCPAP (7/19, 37%) ( p = 0.016). Barrington et al. randomized 54 very low birth-weight infants to NCPAP or NPSIMV after extubation. They used binasal Hudson prongs with the Infant Star ventilator as the generating source for both groups, as well as the StarSync triggering device. Extubation failure criteria were similar to those of Friedlich et al. Barrington and colleagues found the NPSIMV group to have a lower incidence of failed extubation (4/27, 15%) compared with the NCPAP group (12/27, 44%) ( p < 0.05). Khalaf et al. randomized 64 premature infants to either NPSIMV or NCPAP applied after extubation using either the Bear Cub Model BP 2001 (Bear Medical Systems, Inc., Riverside, California, USA) or the Infant Star ventilator with the StarSync triggering device and Argyle nasal prongs (Covidien, Minneapolis, Minnesota, USA). Failure criteria were similar to those of the two previous trials. Treatment failure occurred in 2 of 34 (6%) NPSIMV infants compared to 12 of 30 (40%) NCPAP infants ( p < 0.01).
Bhandari and colleagues performed an RCT comparing NPSIMV after an initial dose of surfactant followed by mechanical ventilation and then extubation. They found significantly less CLD in the NPSIMV group. This was a small trial with 41 total babies enrolled, so it was not powered to find a difference in CLD. In another small study, Kugelman and colleagues randomized 84 premature infants to NIPPV or to NCPAP. These authors reported a decreased need for mechanical ventilation in the NIPPV group, as well as significantly less BPD. Ramanathan et al. randomized 110 infants to either NIPPV or NCPAP postextubation and found a reduced need for invasive ventilation and a reduction in BPD.
Lampland et al. compared differences in pathophysiologic and pathologic conditions in surfactant-deficient, lung-lavaged piglets supported by invasive SIMV or NPSIMV. Animals supported by NPSIMV had higher arterial blood gas pH, lower PaCO 2 , and lower respiratory rates. Also, piglets in the invasive SIMV group had higher PaO 2 /PA o 2 ratios and more pulmonary interstitial inflammation than did the NPSIMV-treated piglets. The results from this short-term study demonstrate that NPSIMV may be less injurious to the lung and provide better ventilation with less need for support than invasive SIMV.
Results from a large multicenter, international RCT comparing outcomes in premature infants of <1000 g birth weight randomized to either NCPAP or NIV were not able to support any of the above findings. There were no differences in mortality or CLD between the two groups. However, in this trial any form of NCPAP and any form of NIPPV could be utilized. Additionally, the NIV could be used either initially or after extubation (within 28 days). It is thus possible that specific differences due to device or timing of application could have been missed. A Cochrane meta-analysis concluded that NIMV reduces extubation failure and need for reintubation within 48 hours to 1 week more effectively than does NCPAP; however, NIMV has no effect on CLD or mortality.
A major controversy as of this writing surrounding the use of NIMV is whether it would be better to synchronize NIV breaths with the infant’s intrinsic efforts. While this is preferred by many clinicians, there are limited technologies to allow this in the face of a large interface leak, especially because the Infant Star ventilator StarSync abdominal capsule-triggering system is no longer available in the United States. Much of the preceding NIMV data have come from trials assessing the efficacy of synchronized NIMV using the StarSync triggering device and Graseby capsule. One study has shown reduced respiratory efforts between synchronized and nonsynchronized NIMV with no differences in tidal volumes, minute ventilation, gas exchange, chest wall distortion, apnea, hypoxemia spells, and abdominal girth.
Bilevel NCPAP or sigh intermittent positive airway pressure (SiPAP) has been marketed as an alternative to constant NCPAP (Viasys, Inc., Conshohocken, Pennsylvania, USA). Using the technology of the Infant Flow driver, these devices can alternate between a lower and a higher CPAP pressure. Synchronization using the Graseby capsule is available in Europe and Canada. In a prospective RCT, Lista et al. compared outcomes in preterm neonates between CPAP and SiPAP as an initial form of support in the acute phase of RDS. Infants supported by SiPAP underwent shorter duration of mechanical ventilation, showed less O 2 dependency, and were discharged sooner. The major flaw in all the comparisons between CPAP and NIPPV/NIV/SiPAP is that they are comparing two different distending pressures; for example, CPAP of 5 cm H 2 O is compared with NIPPV of 20/5 at a rate of 25/min. The latter translates to a mean distending pressure of 7 to 8 cm H 2 O and it is not clear whether the apparent benefits of these NIV techniques are due to a generation of cyclic tidal volume or simply higher mean airway pressure, i.e., more effective CPAP. A clinical trial is needed that would compare these techniques at an equal mean airway pressure.
Neurally Adjusted Ventilatory Assist
NAVA is a relatively new and unique form of assisted ventilation. It can be used in both intubated and nonintubated patients; here we will concentrate on NIV-NAVA. NAVA controls the ventilator by using the electrical activity of the diaphragm (EAdi). The EAdi signal is obtained by nine miniaturized electrodes embedded on a conventional naso/orogastric tube. When properly positioned in the lower esophagus the EAdi signal represents the spontaneous central respiratory drive (see foocus.com/power-point/Chatburn-NAVA-for-Neonates.pdf ).
Conveniently, the tube can also be used for feeding. The baseline signal (at end exhalation) represents the tonic activity of the diaphragm, and the peak level represents the inspiratory effort. Signals are recorded in microvolts. The EAdi signal can be used without NAVA support, and as such is a useful tool to assess neural breathing pattern (e.g., neural inspiratory effort, neural respiratory rate, central apnea).
In the NAVA mode, the EAdi signal is used to trigger and cycle-off the ventilator and also determines the amount of pressure delivered to the patient. The level of assistance is proportional to the EAdi signal and NAVA level. The assistance is adjustable and is dependent on a gain factor between EAdi and pressure delivered (so-called “NAVA level,” expressed as cm H 2 O/μV). For example, if the NAVA level is 1 cm H 2 O and EAdi is 10 μV, the pressure delivered is 10 cm H 2 O. Breaths are triggered at 0.5 μV above the baseline EAdi and terminate when the EAdi signal is 70% of the highest value. Thus, the pressure change is equal to the EAdi change times the NAVA level: Δ P = ΔEAdi × NAVA level.
NAVA provides excellent ventilator synchrony as it is based on a neural signal and not affected by circuit leak. Breathing out of synchrony with the ventilator is uncomfortable for the patient and may cause serious side effects, especially in neonates, including increased intracranial pressure and pneumothorax.
An EAdi signal will not be present in some situations such as during apnea, overassist, oversedation, or severe brain injury. Thus, use of NIV-NAVA is possible only when the infant has a reasonably stable respiratory drive. NIV-NAVA may not be an option in the extremely low birth-weight infant who is prone to frequent and prolonged apnea or who may not be able to generate an effective EAdi signal. Frequent use of the backup pressure control mode in these cases may not be optimal, as volume-targeted ventilation has been shown to reduce death/CLD, intraventricular hemorrhage, periventricular leukomalacia, pneumothoraces, and hypocarbia compared to pressure-control ventilation.
Nonetheless, NAVA is an intriguing technique, which uses feedback control to optimize synchronization and ventilator support. As of this writing it is available through only one company (Maquet Critical Care, Rastatt, Germany) with their Servo ventilators. Additionally, large randomized trials assessing NAVA have not been done. A randomized crossover trial of 15 infants of <32 weeks compared NIV-NAVA with NIV-pressure support. The authors found better patient–ventilator synchrony with NAVA.