Introduction
The transition from fetal to neonatal life requires a rapid conversion from intermittent fetal respiratory activity not associated with gas exchange to continuous breathing upon which gas exchange is dependent once the baby is born. This encompasses the development of neural circuitry that regulates respiratory control and serves as a unique link between the maturing lung and the brain. The frequent apneic events exhibited by preterm infants may be akin to the episodic pauses in respiratory movements that characterize fetal breathing. However, after birth, frequent apnea—often associated with bradycardia and oxygen desaturation events—may be one of the most troublesome problems in neonatal intensive care. The problem of vulnerable neonatal respiratory control is typically enhanced by the mechanical disadvantages of a compliant chest wall and unfavorable lung mechanics. This is compounded by the clinical observation that neonatal respiratory control is vulnerable to a diversity of pathophysiologic conditions ( Fig. 3-1 ). Understanding the maturation of neonatal respiratory control is essential to providing a rational approach to ventilatory support for neonates.
Pathogenesis of Apnea of Prematurity
Our understanding of the pathogenesis of apnea of prematurity is hampered by our limited understanding of the integration of chemo- and mechanosensitive inputs to the autonomic control circuitry of the developing human brainstem. Neonatal animal models, such as rodents, are immature at birth compared to the human trajectory but do not typically exhibit apnea. Nonetheless, we are clearly dependent on animal models to characterize the maturation of neuroanatomic architecture and neurochemical transmitter changes in the brain-stem. Undoubtedly there are significant changes in adenosine, γ-aminobutyric acid (GABA), and serotonin content and corresponding receptor subtypes in respiratory-related brain-stem regions.
Central (CO 2 ) Chemosensitivity
CO 2 is sensed primarily at or near the ventral medullary surface, but also by the carotid bodies, and is the major chemical driver of respiration at all ages. It has been recognized for several decades that preterm infants exhibit a diminished ventilatory response to CO 2 compared to more mature infants. The response to CO 2 in preterm neonates results in an increase in tidal volume with little, if any, increase in frequency. Furthermore, apneic preterm infants have a diminished CO 2 response compared to nonapneic preterm controls. In preterm and term infants, the baseline PaCO 2 has been shown to be only up to 1.5 mm Hg above the apneic threshold; this narrow margin might predispose these children to apnea in the face of only minor oscillations in PaCO 2 . Breathing patterns tend to be more irregular in rapid eye movement (REM) than in quiet sleep, and it is possible that the closeness of the eupneic and apneic CO 2 thresholds may contribute to greater breath-to-breath respiratory irregularity in REM sleep. Unfortunately, REM sleep, quiet sleep, transitional sleep, and even wakefulness are often difficult to distinguish in preterm infants. This complicates the ability to draw conclusions about sleep state and respiratory control in the preterm population.
Peripheral (Hypoxic) Chemosensitivity
The peripheral chemoreceptors are located primarily in the carotid body and are responsible for stimulating breathing in response to hypoxia. Both enhanced and reduced peripheral chemoreceptor functions have been proposed as contributors to apnea of prematurity. In utero, carotid chemoreceptor oxygen sensitivity is adapted to the normally low PaO 2 of the mammalian fetal environment (∼23 to 27 mm Hg). After birth, in response to the increase in PaO 2 with the establishment of breathing, the peripheral chemoreceptors are silenced, followed by a gradual increase in hypoxic chemosensitivity. Once peripheral chemosensitivity is established, hyperoxic resuscitation rapidly elicits apnea, as clearly shown in rat pups. It follows that inappropriate hyperoxic ventilatory support of an apneic infant may hinder recovery of the respiratory drive. Interestingly, infants with bronchopulmonary dysplasia seem to exhibit blunted peripheral chemoreceptor responses compared to controls, which may increase their vulnerability to apnea.
Excessive peripheral chemoreceptor sensitivity in response to repeated hypoxia may also destabilize breathing patterns in the face of significantly fluctuating levels of oxygenation. This is consistent with an earlier finding in preterm infants that a greater hypoxia-induced increase in ventilation correlates with a higher number of apneic episodes. Data from rat pups indicate that conditioning with intermittent hypoxic exposures results in facilitation of carotid body sensory discharge in response to subsequent hypoxic exposure. This effect appears to persist into adult life, raising questions about a longer lasting effect of early apnea in human respiratory control.
In the neonatal period, it is well known that the ventilatory response to hypoxia, an initial increase in minute ventilation, is followed by a posthypoxia decline in frequency of breathing. This so-called hypoxic ventilatory depression is seen in less dramatic form in later life, and may be an appropriate response to sustained hypoxia when coupled with a decrease in metabolic rate. Descending inhibition from the midbrain and other structures appears to cause this hypoxic depression rather than a decline in peripheral chemoreceptor firing, although a contribution from the latter cannot be excluded. The role of hypoxic depression in contributing to apnea of prematurity is unclear; however, low baseline oxygenation is associated with more episodic desaturation in preterm infants.
Role of Mechanoreceptor (Laryngeal) Afferents
Activation of the laryngeal mucosa elicits a potent airway-protective reflex, which, in preterm and term neonates and immature animals of various species, results in a host of autonomic perturbations including apnea, bradycardia, hypotension, closure of upper airways, and swallowing movements. While this strong inhibitory reflex, termed the laryngeal chemoreflex , is thought to be an important contributor to apnea and bradycardia associated with excessive suctioning or aspiration, its relationship to apnea of prematurity is less clear. The pronounced inhibitory effect on ventilation in early life may be the result of enhanced central inhibitory pathways, and GABA has been proposed to mediate this effect. Despite the physiologic rationale for a relationship between stimulation of laryngeal afferents and apnea of prematurity, a temporal relationship between apnea of prematurity and gastroesophageal reflux is rare in preterm infants, as discussed later. Of interest are the data from piglets, confirmed in infants, showing that respiratory inhibition may precede a loss of lower esophageal sphincter tone and theoretically predispose to reflux.
Genesis of Central, Mixed, and Obstructive Apnea
Apnea is classified into three categories traditionally, each based upon the absence or presence of upper airway obstruction: (1) central, (2) obstructive, and (3) mixed. Central apnea is characterized by total cessation of inspiratory efforts with no evidence of obstruction. In obstructive apnea, the infant tries to breathe against an obstructed upper airway, resulting in chest wall motion without airflow throughout the entire apneic episode. Mixed apnea consists of obstructed respiratory efforts, usually following central pauses. The site of obstruction in the upper airways is primarily in the pharynx, although it also may occur at the larynx and possibly at both sites. Interestingly, upper airway closure may also occur during central apnea.
Unlike adult sleep apnea, which is primarily obstructive, apnea of prematurity has a predominantly central etiology with loss of respiratory drive initiated in the brainstem. During mixed apnea it has been assumed that there is an initial loss of central respiratory drive and the resumption of inspiration is accompanied by a delay in activation of the upper airway muscles superimposed upon a closed upper airway. This may be due to a lower CO 2 threshold for activation of chest wall vs upper airway muscles. Mixed apnea typically accounts for more than 50% of long apneic episodes, followed in decreasing frequency by central and obstructive apnea. Purely obstructive spontaneous apnea in the absence of a positional problem is probably uncommon. As standard impedance monitoring of respiratory efforts via chest wall motion cannot recognize obstructed respiratory efforts, mixed (or obstructive) apnea is frequently identified by the accompanying bradycardia or desaturation, although these are not the initiating events.
Relationship between Apnea, Bradycardia, and Desaturation
Cessation of respiration or hypoventilation is almost invariably the event that initiates various patterns of apnea, bradycardia, and desaturation in preterm infants. There is no clear consensus as to when a respiratory pause, which is universal in preterm infants, can be defined as an apneic episode. It has been proposed that apnea may be defined by its duration (e.g., >15 seconds) or by accompanying bradycardia and/or desaturations. However, even the 5- to 10-second pauses that occur in periodic breathing may be associated with bradycardia or desaturation. Periodic breathing—ventilatory cycles of 10- to 15-second duration with pauses of 5- to10-second duration—is considered a “normal” breathing pattern in infants who should not require therapeutic intervention, as discussed earlier. Periodic breathing is speculated to be the result of dominant peripheral chemoreceptor activity responding to fluctuations in arterial oxygen tension. The rapidity of the fall in oxygen saturation after a respiratory pause is directly proportional to baseline oxygenation, and this, in turn, is related to lung volume and severity of lung disease.
Bradycardia is a prominent feature in preterm infants with apnea. The mechanism underlying bradycardia associated with apnea in preterm infants is not entirely clear. A significant correlation between decrease in oxygen saturation and heart rate has been noted, and the bradycardia during apnea might be related to hypoxic stimulation of the carotid body chemoreceptors, especially in the absence of lung inflation. On the other hand, bradycardia may occur simultaneously with apnea during stimulation of laryngeal receptors, suggesting a vagally mediated central mechanism for the production of both. Data in preterm infants indicate that isolated bradycardic events (<70/min) in the absence of accompanying hypoxemia are unlikely to significantly affect tissue oxygenation measured by near-infrared spectroscopy.
Cardiorespiratory Events in Intubated Infants
If apnea is the precipitant of episodic bradycardia and desaturation in spontaneously breathing events, what is the etiology of such events in intubated and ventilated infants? The likely answer is that mechanical ventilation does not always result in effective ventilation in intubated infants. Loss of lung volume and excessive abdominal expiratory muscle activity have been shown to accompany desaturation events in such infants. Loss of spontaneous ventilatory effects may also occur under these conditions. The presence of episodic desaturation and bradycardia in ventilated infants reinforces the need to synchronize spontaneous and ventilator-delivered breaths in infants. Extubation, if feasible, may effectively decrease the frequency of such events.
Therapeutic Approaches
Impaired respiratory control is clearly a major contributor to the need for neonatal assisted ventilation. Our ability to enhance neonatal respiratory control will, therefore, probably diminish potential morbidity induced by ventilation. Aggressive therapy for apnea may be beneficial in avoiding intubation, enhancing extubation, and decreasing any adverse effects of apnea or gas exchange. To achieve these goals we have both nonpharmacologic and pharmacologic and ventilatory strategies. Many of the latter are addressed elsewhere in this book.
Optimization of Mechanosensory Inputs
The respiratory rhythm-generating circuitry within the central nervous system depends on intrinsic rhythmic activity and sensory afferent inputs to generate breathing movement. Bloch-Salisbury et al. have demonstrated that their novel technique of stochastic mechanosensory stimulation, using a mattress with embedded actuators, is able to stabilize respiratory patterns in preterm infants as manifest by a decrease in apnea and an almost threefold decrease in percentage of time with oxygen saturations <85%. Interestingly, the level of stimulation employed was below the minimum threshold for behavioral arousal to wakefulness, thus inducing no apparent state change in the infants, and the effect could probably not be attributed to the minimal increase in sound level associated with stimulation. Such an approach is clearly still a research tool but worthy of future study. It points to the important consideration that environmental stimulation of the infant must be optimized. Similarly, skin-to-skin care is a highly desirable practice in the neonatal intensive care unit (NICU) to encourage parental attachment and potentially influence respiratory control. Data have shown that this practice is not only safe but also associated with decreased electrical diaphragm activity, potentially benefiting energy expended on respiratory efforts.
Optimization of Blood Gas Status
Intermittent hypoxic episodes are almost always the result of respiratory pauses, apnea, or ineffective ventilation, aggravated by poor respiratory function ( Fig. 3-2 ). Targeting lower baseline oxygen saturation has been associated with persistence of intermittent hypoxic episodes. It is unclear whether this lower baseline oxygen saturation increases the incidence of apnea with resultant hypoxemia (via hypoxic depression of breathing) or whether the incidence of apnea is comparable between oxygen targets, but the lower oxygen saturation baseline predisposes to more frequent or profound intermittent hypoxemia. Regardless of mechanism, a low baseline oxygen saturation (e.g., <90%) should be avoided in the face of immature respiratory control superimposed upon poor respiratory function. It is also unclear whether the beneficial effect of packed red cell transfusion is secondary to improved respiratory control or decreased vulnerability to hypoxia in the face of apnea. These studies demonstrate that improvement in intermittent hypoxic episodes after red cell transfusion is manifest only after the first weeks of postnatal life and that nursing reports significantly underestimate the frequency of such cardiorespiratory events. Obviously any benefit of packed cell transfusion on apnea, bradycardia, and desaturation must be balanced against potential hazards of transfusion.
