Chapter 61. Special Intensive Care

Respiratory failure is one of the most common problems requiring admission to the neonatal intensive care unit. Respiratory failure is often the final result of restrictive lung disease with regional or global alveolar collapse or consolidation. In order to achieve optimal alveolar gas exchange, the lung needs to be inflated at end expiration (at functional residual capacity) and have sufficient tidal and minute ventilation to eliminate carbon dioxide.1 Under optimal conditions, several factors help prevent the collapse of alveoli at the end of expiration and thereby maintain an adequate functional residual capacity. Surfactant, which is produced in the type II cells of the lung, dramatically reduces surface tension and opposes the tendency of alveoli to collapse when lung volume is at its lowest. In addition, the rigidity of the chest wall opposes lung collapse. In comparison to the adult, several factors present disadvantages for the neonate’s capacity to maintain optimal lung volumes. The neonatal chest is highly compliant, which limits its ability to oppose elastic recoil and collapse during expiration and increases the potential for the development of collapse. This problem is greatly exacerbated in the setting of surfactant deficiency or inactivation, in which unopposed surface tension dramatically increases the chance of alveolar collapse during expiration. Lastly, although the small diameter of the neonatal tracheobronchial tree generally is sufficient to provide unimpeded airflow, further small reductions in its diameter can dramatically increase resistance and adversely affect gas entry and egress.

A hallmark of treatment for most causes of neonatal respiratory failure is the provision of positive airway pressure.2 (See also Chapter 102.) Application of pressure to the proximal airway may result in several beneficial effects. Most importantly, positive pressure opposes the tendency of end-expiratory alveolar collapse caused by elastic recoil and high surface tension. Moreover, positive proximal airway pressure increases the pressure differential between the upper airway and the distal airspace during spontaneous respiratory effort, facilitating increased bulk flow of gas down the tracheobronchial tree. In addition, positive pressure may help to maintain adequate patency of the airway in disease states characterized by inflammation, plugging, or anatomic (either fixed or dynamic) narrowing of the airway.

Continuous Positive Airway Pressure (CPAP)

The least invasive means for consistently delivering proximal distending airway pressure is via CPAP.3 CPAP may be delivered via prongs, which are placed in the nose, or by a mask affixed over the mouth and nose. Generally, CPAP is provided at a pressure range of 4 to 8 cm H2O, with the specific level based on the type and severity of the underlying lung disease, the degree of inflation achieved, and the baby’s tolerance of the therapy. Common indications for use of CPAP in the neonatal intensive care unit include mild hyaline membrane disease, disease or narrowing of the airways, and as a bridge to extubation in preterm infants recovering from hyaline membrane disease.4 CPAP may also decrease the incidence or severity of apnea and bradycardia in some preterm infants, likely by improving chest wall stability and helping to maintain airway patency. In recent years, some investigators have advocated early use of CPAP, either in place of or after surfactant replacement, as a means for providing respiratory support for infants with hyaline membrane disease while avoiding the potentially injurious use of more invasive ventilatory support.5 Although most babies treated with CPAP can be concurrently fed successfully via a nasogastric tube, some may develop intolerable gastric and intestinal distension. Vigilant observation and care of the nasal septum and columnella is necessary because both CPAP prongs and ventilation masks can cause pressure necrosis of these structures.

Conventional Positive Pressure Ventilation

In many cases of respiratory compromise in the neonate, CPAP alone does not provide adequate support, and more invasive therapy, including intubation and formal mechanical ventilation, is required.6Conventional ventilation of the newborn generally refers to the use of intermittent positive pressure ventilation, or IPPV, whereby a ventilator provides alternating cycles of positive end-expiratory pressure (PEEP) and discrete inflating “breaths.” A common feature of these ventilators is the use of a continuous flow of oxygen-enriched gas through the ventilator circuit, providing an ongoing source of fresh gas during spontaneously breathing. Partial closure of a valve in the expiratory portion of the circuit during the inspiratory cycle diverts gas down the endotracheal tube, delivering a breath to the patient. The size of each full inflating breath is determined by the specific ventilator mode employed. Typically, such ventilators are time-cycled, meaning that they provide a breath at discrete intervals, based on the set rate, and for a defined period (the inspiratory time). Should the patient be breathing spontaneously, most modern ventilators now have the capacity to detect the patient’s spontaneous effort and to synchronize each delivered breath to a spontaneous inspiratory effort during each respiratory cycle (a technology called SIMV, or synchronized intermittent mandatory ventilation). In addition, some neonatal ventilators now offer a pressure support option whereby all intervening spontaneous breaths that do not qualify for a full ventilator breath are provided with a modest increase in inspiratory pressure (generally 4–8 mm Hg) intended to help overcome the resistance to airflow in the endotracheal tube and decrease the work of breathing. An alternative mode now available on some ventilators is “assist-control” in which each spontaneously initiated respiratory effort detected by the ventilator is supported with a full tidal volume breath as shown in Figure 61-1.

In pressure-limited ventilation, gas is diverted into the endotracheal tube at a preset pressure, the peak inspiratory pressure (PIP). The volume of gas delivered in this mode depends on the underlying lung compliance; highly compliant lungs receive a greater volume of gas than those that are poorly compliant. Optimal inflating pressures cannot be generalized because of the wide variability in the type and severity of underlying lung disease among infants. With some newer generation neonatal ventilators, volume-targeted ventilation is now possible.7 In this mode, a specified volume of gas is diverted down the endotracheal tube. The pressure required to deliver this volume depends on the compliance and resistance in the lung. Volume-targeted ventilation offers the theoretic advantage of providing consistent gas volumes in the setting of rapidly changing lung compliance, a situation common to many forms of neonatal lung disease, such as hyaline membrane disease. It should be noted that most ventilators, when in a volume mode, require that the operator set a pressure limit that will not be exceeded during the inspiratory phase even if the desired volume is not delivered. This can lead to inconsistent tidal volumes and inadequate ventilation. Target settings for tidal volume (VT) generally are 4 to 8 ml/kg. Regardless of the inspiratory mode used with conventional ventilation, positive end-expiratory pressure should be provided to minimize end-expiratory alveolar collapse. Generally, a positive end-expiratory pressure of 4 to 8 mm Hg is used for most neonates, though higher levels may be optimal in the setting of significant disease of the airways. Most practitioners recommend assessing multiple parameters to guide the pressure or volume used to ventilate a newborn, including gas exchange, clinical assessment of air entry (both visual and auscultatory), and radiographic assessment of lung inflation.

Currently, there is no consensus on the most efficacious and least injurious conventional mode of ventilation for neonates. Many of the newer neonatal ventilators can employ either a pressure-limited or volume-targeted strategy and allow the practitioner to monitor a number of sensitive and sophisticated ventilatory measurements regardless of the mode employed. In addition, many of these newer ventilators offer modes that are beyond the scope of this discussion, each of which may theoretically improve gas exchange or reduce ventilator-induced lung injury. None of these newer modes have been widely embraced, but ongoing studies may ultimately provide compelling evidence of benefit over the more basic strategies just discussed.

High-Frequency Ventilation (HFV)

HFV offers an alternative to the conventional ventilation approaches described previously. As the name implies, HFV employs supraphysiologic rates, generally in the range of 300 to 700 breaths per minute, and very small tidal volumes, sometimes less than the anatomic dead space. The physiologic mechanisms by which HFV achieves ventilation are very complex, and this mode of ventilation can be extremely efficient, including in babies who are not effectively ventilated using a conventional strategy. In the United States, 3 different high-frequency modes are employed clinically. In high-frequency oscillation, gas is delivered by the forward and aft movement of a diaphragm, which “oscillates” about a set mean airway pressure. This mode of HFV is distinguished from others because of its active expiratory phase. In high-frequency jet ventilation, brief pulses of gas are delivered directly into a port on a specially designed endotracheal tube, while a conventional ventilator simultaneously provides a set positive end-expiratory pressure or mean airway pressure. In high-frequency flow interruption, a central stream of gas is interrupted at a rapid rate. Unlike high-frequency oscillation, both high-frequency jet ventilation and high-frequency flow interruption allow for the use of regular “background” breaths, which are conventional tidal breaths typically provided several times per minute.

Clear indications for the use of HFV in place of controlled mechanical ventilation are not well established. Most routinely, HFV is used as a rescue therapy for neonates who respond inadequately to a conventional strategy.8 Additionally, HFV offers an alternative mode of ventilation in the setting of air-leak syndromes, including pneumothorax, pulmonary interstitial emphysema, and bronchopleural fistula.9 Some practitioners also advocate use of HFV as the initial and primary mode of ventilation. Studies comparing a conventional strategy to HFV for the initial management of preterm infants with respiratory distress syndrome are inconclusive.10-12 In term infants with hypoxemic respiratory failure, HFV is frequently employed to recruit atelectatic lung and improve pulmonary hemodynamics.13

As with any approach to mechanical ventilation, HFV devices should be used with caution. They have the potential to rapidly and dramatically lower PaCO2 to unacceptable levels, including in patients who have refractory hypercarbia on a conventional ventilator. In addition, marked global hyperinflation can occur, resulting in adverse hemodynamic effects and air leak. In the setting of inhomogenous lung disease, application of HFV may exacerbate regional hyperinflation and increase the likelihood of air leak. Many neonatologists advocate the routine use of transcutaneous monitoring of carbon dioxide and frequent radiographic evaluation of lung inflation of infants treated with HFV.

Nitric oxide, a highly volatile, inert gas, was discovered in the 1990s by Furchgott. It is produced during the conversion of arginine to citrulline by the nitric oxide synthase (NOS) family of enzymes.14 Three major isoforms of NOS have been described, including endothelial nitric oxide synthase (eNOS), the major isoform in the pulmonary vasculature.

Over the past decade, the importance of nitric oxide in regulating vascular tone has been greatly clarified. Produced primarily by eNOS in the vascular endothelium, nitric oxide diffuses into the adjacent vascular smooth muscle cell. There, nitric oxide increases the activity of soluble guanylate cyclase (sGC), which results in a subsequent increase in cyclic guanosine monophosphate (cGMP) activity. Through a complex series of steps, cGMP promotes vascular smooth muscle cell relaxation and results in vasodilation. eFigure 61.1 provides a schematic view of the generation of nitric oxide in the pulmonary circulation.

The importance of endogenous nitric oxide in the regulation of pulmonary vascular tone in the perinatal lung has recently been established.15 eNOS is expressed in the developing pulmonary circulation throughout the last half of gestation, and animal studies demonstrate that endogenous nitric oxide modulates pulmonary vascular tone during this time. At birth, exposure to numerous physiologic changes, including rhythmic lung distension, shear stress, and exposure to increased oxygen, causes a marked increase in the production of nitric oxide by the lung. Nitric oxide causes a rapid fall in pulmonary vascular resistance and, coupled with the increase in systemic vascular resistance when the umbilical circulation is interrupted, a dramatic and sustained increase in pulmonary blood flow. This highly regulated series of events remains among the most important physiologic adaptations in the transition from fetal to newborn life.

Nitric Oxide Treatment of Persistent Pulmonary Hypertension of the Newborn (PPHN)

PPHN is discussed in detail in Chapter 50. In this disorder the pulmonary circulation fails to undergo the normal fall in pulmonary vascular resistance at birth. When severe, the persistence (or redevelopment) of elevated pulmonary vascular resistance can cause critical hypoxemia as a result of either extrapulmonary right-to-left shunt of deoxygenated blood through persistent fetal channels, including the ductus arteriosus and/or the foramen ovale, or intrapulmonary shunt caused by a ventilation-perfusion mismatch. Whether in isolation or in association with one of a number of other specific disease entities, this pathophysiologic circumstance has been termed persistent pulmonary hypertension of the newborn (PPHN). Until recently, options for the treatment of PPHN were limited because of the lack of selective pulmonary vasodilators.

Following the discovery of nitric oxide and the recognition that PPHN is associated with a lack of endogenous nitric oxide, animal studies were undertaken to determine the effects of exogenous nitric oxide, supplied as an inhaled gas through a ventilator circuit, on animals with experimental PPHN. In these animals, inhaled nitric oxide (iNO) selectively lowers pulmonary vascular resistance and improves gas exchange, raising the possibility of its application to newborns with PPHN. Use of nitric oxide as a therapeutic agent for PPHN offered numerous theoretic benefits. First, as with nitric oxide that is produced endogenously, iNO would be expected to cause smooth muscle cell relaxation and vasodilation resulting in a sustained reduction in pulmonary vascular resistance, decreased extrapulmonary right-to-left shunt, and an improvement in oxygenation. In addition, iNO would be expected to selectively improve blood flow to well-ventilated regions of the lung, thereby improving the matching of ventilation and perfusion (V:Q). Lastly, because nitric oxide becomes tightly bound by hemoglobin once it reaches the bloodstream, it should not have vasoactive effects remote from the lung, providing the first truly selective pulmonary vasodilator.

Several large, prospective, randomized trials have now conclusively demonstrated that iNO improves oxygenation and reduces the combined endpoint of death or need for extracorporeal membrane oxygenation in term and near-term neonates with moderate to severe hypoxemic respiratory failure and PPHN.13,16,17 These findings have been demonstrated in studies both with and without inclusion of discrete evidence (either clinical or echocardiographic) of pulmonary hypertension as an entry criteria. Provision of iNO has therefore become an important adjunct in the therapy of hypoxemic respiratory failure in term and near-term infants. Many experts continue to advocate targeting the use of iNO to babies with defined pulmonary hypertension. In addition, strategies to achieve adequate lung inflation, including consideration of high-frequency strategies, should be employed prior to initiation of iNO. Optimizing lung inflation can itself dramatically lower pulmonary vascular resistance and ameliorate the need for iNO. Moreover, because inhaled gas reaches only inflated lung, the response to nitric oxide in the setting of inadequate lung inflation may be significantly reduced.

Consensus on the optimal dose of iNO has not been established. In general, low-dose (< 20 ppm) nitric oxide is effective, and most controlled studies have demonstrated little or no benefit to routine use of higher doses. Strategies for weaning and discontinuation of nitric oxide are also poorly defined. As with initiation of therapy, use of echocardiography to document resolution of pulmonary hypertension at the time of weaning and discontinuation can be a valuable guide. When nitric oxide is discontinued, caution should be exercised because some babies develop rebound pulmonary hypertension, which may be clinically evident by a substantial increase in oxygen requirement. The cause of this phenomenon is poorly understood, but most patients respond to reinitiation of therapy. In some cases, significant pulmonary hypertension persists beyond the need for mechanical ventilation. In this circumstance, nitric oxide can be delivered by nasal cannula or head hood, though adjustments in the dose must be made to ensure adequate delivery to the distal lung.21

Nitric Oxide Use in the Preterm Infant

A clear role for iNO in the treatment of preterm infants has been more difficult to define. Based on its anti-inflammatory and antioxidant effects and its promotion of lung vascular and alveolar growth in laboratory settings, its potential to reduce the incidence of bronchopulmonary dysplasia in preterm infants has been investigated in several studies. However, interpretation of these studies has been complicated by the wide variability in study design, including differences in initial disease severity, age at enrollment, and duration of treatment. Taken together, these studies suggest that early, prolonged (7–24 days) iNO reduces bronchopulmonary dysplasia and death without worsening intraventricular hemorrhage or periventricular leukomalacia.18-20 However, assessment of long-term neurodevelopmental outcome of these patients will be critical, and routine use of iNO in this population will likely remain experimental until those results are available.

Neonatal extracorporeal membrane oxygenation (ECMO) is performed by continuously draining a portion of the baby’s blood from the circulation and exposing it to a highly efficient membrane, which removes carbon dioxide and adds oxygen as shown in eFigure 61.2. The blood is then returned to the baby’s circulation in an isovolemic fashion. ECMO was widely adopted in the 1980s as a final treatment option for term neonates judged to have high mortality who had failed more conservative treatment modalities.22 Over the past 2 decades, experience and technical advances have led to improved outcomes.

Prior to initiation of ECMO, a full circuit of tubing must be primed with compatible, heparinized blood. Vascular cannulae are then surgically placed, the first in the right atrium via the right internal jugular vein, for removal of blood. In venoarterial ECMO, a second cannula is placed in the right carotid artery, which is ligated, and advanced to the ascending aorta, to which blood is returned. By returning oxygenated blood to the arterial circulation, venoarterial ECMO provides true cardiopulmonary bypass, resulting in nonpulsatile arterial blood flow. Venoarterial ECMO can thereby be used to support neonates who suffer from both cardiac and respiratory failure. In contrast, in venovenous ECMO, blood is removed from and returned to the venous circulation, sparing ligation of the carotid artery. Venovenous ECMO requires that the baby’s cardiac performance be adequate to maintain systemic perfusion. Neonatal venovenous ECMO is typically performed using a double lumen catheter placed in the right atrium (via the right internal jugular vein). An alternative is to place the second cannula in a large systemic vein, such as a femoral vein.

Regardless of the type of ECMO, the circuitry and the physics are similar. Venous blood is passively drained, by gravity, from the right atrium into a collapsible collecting bladder. Use of such a bladder avoids the potential application of negative pressure directly on the venous drainage cannula. Blood is then pumped from the bladder via an occlusive roller pump into a highly efficient, semipermeable membrane oxygenator, in which exchange of oxygen and carbon dioxide occurs. After leaving the oxygenator, blood is rewarmed and then returned to the patient via the second cannula.

The decision to proceed with ECMO is usually guided by the perception that the baby has a high risk of mortality in spite of maximal therapy with more conventional therapies. Definitions of what constitutes a “high risk of mortality” vary by ECMO site and experience and have not been nationally standardized. For neonates with hypoxemic respiratory failure, many centers have developed defined criteria, often based on calculated indices of gas exchange, such as the alveolar-oxygen gradient or oxygenation index, measured at discrete intervals, and on “maximal support” as defined locally. Additional factors that must be weighed before undertaking ECMO include the size and gestational age of the patient and the underlying diagnosis. ECMO requires systemic heparinization, increasing the risk of catastrophic bleeding for very preterm newborns or babies with an underlying coagulopathy. For this reason, most centers limit the provision of ECMO to babies of more than 34 weeks’ gestation. In addition, technical considerations preclude the routine use of ECMO in babies weighing less than 1800 to 2000 grams. Most centers also exclude babies with significant preexisting intracranial hemorrhage or injury because of the high risk of long-term neurodevelopmental impairment. Moreover, ECMO should be provided only to babies with a reasonable expectation that the underlying disease can be reversible within a reasonable (days to weeks) period of time.

Considering the severity of baseline illness for which ECMO is initiated, the outcome of neonates treated with ECMO has been encouraging.23 Irrespective of the underlying diagnosis, survival rates for infants treated with ECMO in the United States are approximately 80%.22 The most common cause of death among babies treated with ECMO is hemorrhage. Recent follow-up data from the United Kingdom suggest that ECMO does not worsen neurodevelopmental outcome among survivors with severe neonatal respiratory failure compared with those survivors provided conventional therapy.24 Although infants with respiratory failure treated with ECMO carry a high risk of progressive sensorineural hearing loss, studies suggest that ECMO does not increase the risk beyond that conferred by the underlying cause of respiratory failure and its initial treatment.25 In the last several years, as the application of newer therapies such as HFV and iNO to neonates with respiratory failure has increased, the use of ECMO in this population has decreased significantly. Whether results of ECMO will remain as encouraging in the population of babies that do not respond to these newer therapeutic modalities, in terms of both survival and neurodevelopmental outcomes, requires ongoing investigation.