Apnea of Prematurity and Sudden Infant Death Syndrome

Abstract

Apnea is a frequent occurrence in prematurity and its prevalence in the most severely preterm population is indicative of an immature respiratory neural control system. Preterm infants are also at increased risk of Sudden Infant Death Syndrome (SIDS), which has been associated with similar respiratory neural control dysfunction seen in prematurity. Generally, abnormalities in both central and peripheral mechanisms of respiratory control are thought to be key underlying features of abnormal respiratory system development. Numerous factors contribute to the etiology of apnea and respiratory control dysfunction including the environment (e.g. substance use/misuse), sex, genetics, a vulnerable neonate and various underlying comorbidities. However, there are major gaps in our understanding of both normal and abnormal respiratory control system development, which highlights the need for continued research using novel and innovative methods.

Introduction

Apnea of prematurity is almost universal in preterm infants of less than around 28 weeks’ gestation, and a clinically significant problem at the lowest gestation which now approaches 23 weeks. In this population it is associated with pulmonary immaturity and, in combination with immature respiratory control, results in the need for interventions that may establish a vicious cycle of lung injury. The consequences may be life-long (Martin, 2018). Therefore, apnea resulting from immature respiratory control is probably of greater consequence than apnea occurring at any time later in life (Mitchell et al., 2021).

Fetal breathing activity has been described in many species and is present very early in gestation. Breathing activity in the human fetus can be detected using ultrasound by 11 weeks’ gestation. Although the placenta is the site of gas exchange in utero, fetal breathing movement is important for lung growth and development. Moreover, decreased diaphragmatic activity has been associated with pulmonary hypoplasia.

Fetal breathing movements occur only during rapid eye movement [REM] in the third trimester with cessation of breathing during non-REM sleep, possibly secondary to descending inhibitory pontine input to the medullary rhythm generating center.

Fetal hypercapnia increases the incidence and depth of fetal breathing movements only during REM sleep, without affecting breathing frequency. This response is present from 24 weeks’ gestation and CO₂ sensitivity increases with advancing gestational age, as discussed later. Although the fetus lives in a relatively hypoxemic state [PaO₂ 23–27 mm Hg], oxygen delivery in utero is adequate because it matches oxygen consumption and allows for fetal activity and growth. Unlike adults, the fetus responds to hypoxemia with a decrease in breathing activity. The cause of this hypoxic ventilatory depression in utero appears to be central in origin. Brainstem transection or lesions in the lateral upper pons allow acute hypoxemia to stimulate fetal breathing movements even in the absence of the carotid bodies. This is consistent with the concept that hypoxic depression is the result of descending pontine or suprapontine inhibition. Unlike the fetal breathing response to hypercapnia, the hypoxic response may be adaptive in the sense that an increase in fetal respiratory activity in response to hypoxia would be counterproductive. The rapid rise in oxygen saturation at birth and resultant loss of hypoxic respiratory depression is probably an important stimulus for respiration to become continuous. Unfortunately, this transition is not always successful after preterm birth.

Natural History of Apnea of Prematurity

Apnea is an almost universal manifestation of immature respiratory control in premature infants. Such infants may experience respiratory pauses of varying duration, with decreasing gestational age increasing vulnerability to such events. Short respiratory pauses may be self-limiting, while longer episodes may necessitate intervention, especially in the most immature infants. Apnea has traditionally been defined as a respiratory pause lasting at least 20 seconds, or a pause accompanied by bradycardia, cyanosis, or pallor. However, shorter apnea events [e.g., 15 seconds in duration], if recurrent, may also require clinical intervention, particularly if they result in hypoxemia.

Apnea should be distinguished from periodic breathing in which the infant exhibits regular cycles of rapid respiration for approximately 10 seconds interspersed with pauses of similar duration, in a recurring pattern [≥3 cycles]. Periodic breathing is a benign respiratory pattern in the premature or young term infant, although there may be accompanying hypoxemia and bradycardia.

Similar to periodic breathing, clinically significant apneic episodes often only develop after several days or even weeks of postnatal life (Patel et al., 2016; Di Fiore et al., 2010). Consistent with this observation, Di Fiore and associates have shown that intermittent hypoxemic episodes [IH] in very preterm infants are infrequent in the first week, followed by a progressive increase over weeks 2 to 4 before decreasing in weeks 6 to 8 (Di Fiore et al., 2010). Thereafter, both the frequency and duration of apnea and IH events decrease with advancing postnatal age. Given that IH events are largely precipitated by apnea, these observations serve to emphasize the developmental immaturity of respiratory control that underlies neonatal apnea.

Apnea can increase in severity at any time during hospitalization in response to a variety of adverse events, most notably sepsis. Apnea can also recur in premature infants after the neonatal period in response to specific clinical situations in which respiratory drive is altered. Respiratory syncytial viral infection is well known to elicit apnea. These spells may be severe, necessitating endotracheal intubation. The cause of respiratory depression in respiratory syncytial viral infection is uncertain; however, inhibitory reflexes from upper airway afferents have been implicated.

Former premature infants also may experience apnea during recovery from anesthesia and, in preterm infants, cardiorespiratory monitoring during the acute postoperative period is an important part of their care. Routine care of the premature infant before discharge includes several practices that may cause temporary recurrence of apnea. For example, eye examination for retinopathy of prematurity may be associated with apnea. Immunization also has been found to be associated with apnea, bradycardia, and desaturation. With such practices, therefore, infants require continued monitoring until they are stable, but evidence of a subsequent detrimental effect is lacking.

Characterization of Neonatal Cardiorespiratory Events

Respiratory Muscle Imbalance

The expression cardiorespiratory events refer to the triad of apnea, bradycardia, and intermittent hypoxemia [IH] or desaturation in preterm infants, although apnea is the major precipitant in such events. Apneic events are distinguished not only by their duration, but also by the presence or absence of airway obstruction during the episode of apnea. Thach and Stark initially described an increase in the frequency of apnea when the premature infant’s neck was flexed (Thach and Stark, 1979). Subsequently, upper airway obstruction was found to accompany apnea in preterm babies, even though neck flexion was not present (Milner et al., 1980). The location within the upper airway at which obstruction occurs is usually within the pharynx, but may vary between pharyngeal and laryngeal structures.

The presence or absence of upper airway obstruction forms the basis of the classification of apnea into three types. Mixed apnea is the most commonly observed clinically significant, even in small premature infants, and consists of obstructed inspiratory efforts as well as a central pause. In obstructive apnea, obstructed breaths are characterized by chest wall motion without nasal airflow which continue throughout the entire apnea. In central apnea, inspiratory efforts cease entirely, and obstructed breaths are not observed. Mixed apnea accounts for approximately 50% to 75% of all instances of apnea in premature infants; obstructive apnea, 10% to 20%; and central apnea, 10% to 25% (Miller et al., 1985). These data differ somewhat from adults with so-called sleep apnea where upper airway obstruction plays the major role. Whereas many upper airway muscles, including the alae nasi, laryngeal abductor, and adductor muscles, modulate patency of the extrathoracic airway, failure of genioglossus activation has been most widely implicated in mixed and obstructive apnea in both adults and infants. Carlo and co-workers compared activity of the genioglossus muscle with that of the diaphragm in response to hypercapnic stimulation. Consistent with data obtained in animal models, genioglossus activation in preterm infants was delayed for approximately one minute after initiation of CO₂ rebreathing and occurred only after an end-tidal PCO₂ threshold of approximately 45 mm Hg had been reached (Carlo et al., 1988). By contrast, diaphragmatic EMG activity increased linearly with progressive hypercapnia. Thus, it is possible that an absent, small, or delayed upper airway response to hypercapnia may result in upper airway instability when accompanied by a linear increase in chest wall activity. This instability may predispose affected infants to obstruction of inspiratory efforts after a period of central apnea. Consistent with this, short apneic episodes are more likely to be central, and longer episodes [lasting longer than 15 seconds] are more likely to be accompanied by obstructed breaths. Furthermore, airway obstruction often occurs toward the end of the longer episodes of mixed apnea, when diaphragmatic activity may be enhanced before that of the upper airway muscles. Subsequently, Gauda and co-workers evaluated the activity of the genioglossus and diaphragm during spontaneously occurring mixed and obstructive apneic episodes. During mixed apnea, the amplitude of the diaphragmatic EMG activity decreased on the initial obstructed inspiratory effort and did not exceed that for the breath preceding apnea until flow was re-established (Gauda et al., 1989). Thus, decreased diaphragmatic activity is a major component of spontaneous apnea associated with airway obstruction, and neither diaphragm nor genioglossus activity is increased until resolution of apnea. These findings suggest that central, mixed, and obstructive apneas are caused by a common mechanism - a reduction in central drive affecting the diaphragm and dilating muscles of the upper airway. However, the finding that only 40% of spontaneous apneic episodes were terminated with genioglossus activation indicates that this is not the sole mechanism by which upper airway obstruction is relieved in premature infants.

Reflex Bradycardia

The reflex effects of apnea include characteristic changes in heart rate. Bradycardia may begin within seconds of the onset of the apneic episode. Henderson-Smart and co-workers noted a significant correlation between the decrease in oxygen saturation and heart rate, and postulated that bradycardia during apnea could result from hypoxic stimulation of the carotid body chemoreceptors (Henderson-Smart et al., 1986). The relationship between reflex control of heart rate and breathing is complex. When ventilation is allowed to increase in response to hypoxia, tachycardia occurs. When this reflex increase in ventilation is prevented, bradycardia results. At the onset of apnea, at which time cessation of ventilation and onset of hypoxemia occur almost simultaneously, hypoxemia would be expected to produce bradycardia.

Other reflex input may accentuate the bradycardia during hypoxemia. For example, the reflex effects of apnea in infants also have been likened to the physiologic responses in diving mammals. During reflex apnea in these animals, upper airway afferent input from superior laryngeal and trigeminal nerve stimulation may produce greatly enhanced bradycardia. The contribution of upper airway reflexes to the bradycardia that occurs during apnea is difficult to study in human infants. In summary, the rapid onset of bradycardia during apnea may be a complex reflex deriving from multiple sources, including trigeminal receptors and carotid chemoreceptors.

Intermittent Hypoxemia

Continuous recording of oxygen saturations via pulse oximetry has revealed a much higher incidence of intermittent hypoxemic [IH] events associated with apnea than was previously thought, and permitted analysis of high-risk IH patterns that correlate with undesirable clinical outcomes. Previous studies examining the long-term effects of apnea may, in fact, be proxy for the effects of IH. IH events are associated with multiple negative outcomes, including retinopathy of prematurity [ROP], bronchopulmonary dysplasia [BPD], sleep-disordered breathing, wheezing, unfavorable neurodevelopmental outcomes, and death, but it can be difficult to separate the effects of IH itself from the therapies, such as oxygen, used to treat it. Furthermore, the association between IH events and adverse outcomes may not be causal. Although the mechanisms underlying the pathologic correlates of IH are still under investigation, there is evidence from rodent models that IH causes changes in inflammatory signaling and generation of reactive oxygen species (Del Rio et al., 2011; Yang et al., 2016). Normally only a small amount of oxygen is incompletely reduced to form reactive oxygen species which, in most cases, are neutralized by endogenous antioxidants and free radical scavengers. However, preterm infants do not have well-developed antioxidant defenses, and under metabolically stressful situations the free radicals generated could cause direct tissue damage and trigger proinflammatory and proapoptotic pathways. It, therefore, appears that any adverse effects of apnea on neurorespiratory outcomes may be mediated by the resultant IH events that, in turn, may be precipitants of an oxidant stress (Dennery et al., 2019; Shah et al., 2020).

Neural mechanisms of Neonatal Apnea

Role of Brainstem

Immaturity or depression of central inspiratory drive to the respiratory muscles is accepted as a key factor in the pathogenesis of apnea of prematurity. Vulnerability of the bulbopontine respiratory centers in the brainstem to inhibitory mechanisms could explain why apneic episodes are precipitated in preterm infants by such a wide diversity of specific clinicopathologic events. In other words, apnea may represent the final common response of incompletely organized and interconnected respiratory neurons to a multitude of afferent stimuli [Fig. 1]. Immature circuits within neuronal networks may be highly susceptible to inhibitory neurotransmitters and neuroregulators such as adenosine, γ-aminobutyric acid [GABA], and endogenous opiates. Unfortunately, the maturation of central respiratory integrative mechanisms and of their biochemical neurotransmitters is inaccessible to study in human infants, and no ideal animal model of spontaneous apnea has been identified for study in the non-anesthetized state. As discussed later, abnormalities in the serotoninergic system have also been proposed for sudden infant death syndrome [SIDS].

Figure 1.

Apnea of prematurity is the result of a combination of integrated factors. These include maturational changes in brainstem neurochemistry, diminished CO₂ responses, up- or downregulated peripheral [hypoxic] chemosensitivity, descending hypoxic depression and enhanced inhibition from upper airway mechanoreceptors. Predisposition to hypoxemic episodes is then enhanced by parenchymal immaturity and possible pulmonary hypertension.

Earlier studies revealed that brainstem conduction times of auditory evoked responses are longer in infants with apnea than in matched premature infants without apnea (Henderson-Smart et al., 1983). This observation provides indirect evidence that infants with apnea exhibit greater-than-expected immaturity of brainstem function on the basis of postmenstrual age and supports the concept that stability of central respiratory drive improves as dendritic and other synaptic interconnections multiply in the maturing brain.

The absence of respiratory muscle activity during central apnea unequivocally points to the depression of respiratory center output. As indicated above, Gauda and co-workers documented a decrease in diaphragm activity using electromyography [EMG] during spontaneously obstructed inspiratory efforts that characterize mixed apnea (Gauda et al., 1989). Thus, both central and mixed apneic episodes share an element of decreased respiratory center output to the respiratory muscles. The role played by the balance of neurotransmitters in modulating this inhibition is not yet known. GABA is a ubiquitous major inhibitory neurotransmitter within the brain. Physiologic studies in neonatal animal models have implicated GABA in inhibition of respiratory frequency and decreased ventilatory responses during hypercapnia, hypoxia, and superior laryngeal nerve stimulation (Abu-Shaweesh et al., 2001). GABA, thus, has the potential to play a key role in the vulnerability of preterm infants to apnea.

Infants with sepsis are prone to respiratory compromise, including apnea. The proinflammatory cytokine IL-1β is commonly released in response to infection and can depress respiration by way of a prostaglandin E₂-related mechanism (Olsson et al., 2003). Thus, activation of this pathway may trigger apnea in the infant with infection (Olsson et al., 2003). Chorioamnionitis is a major precipitant of preterm birth, and is associated with neonatal brain injury in the form of periventricular leukomalacia and chronic neonatal lung injury disease (Mitchell et al., 2021). Antenatal or postnatal exposure of the lung to a proinflammatory stimulus may activate brain circuits via vagally-mediated processes. Lipopolysaccharides [LPSs] instilled into the trachea of 10- to 12-day-old rat pups increased inflammatory cytokine gene expression in the medulla oblongata and attenuated both the immediate and late hypoxic ventilatory response when animals were tested within three hours of treatment (Balan et al., 2011). This brainstem response to intrapulmonary LPS was diminished after vagotomy, suggesting a lung-to-brainstem communication via vagal afferents.

Role of Chemoreceptors

The ventilatory and respiratory muscle responses to increases in inspired carbon dioxide [CO₂] reflect predominantly central chemoreceptor activity and are less well developed in the immature infant before 33 weeks of postmenstrual age (Rigatto et al., 1975).

The reduced ventilatory response to CO₂ in small preterm infants may be the result of decreased central chemosensitivity, or mechanical factors preventing an appropriate increase in ventilation (Gerhardt and Bancalari, 1984). Unlike adults, preterm infants do not tend to increase frequency of ventilation during hypercapnia. In these infants, hypercapnia may be accompanied by prolongation of expiratory duration (Noble et al., 1987). The CO₂ response curve has a decreased slope [indicating a less steep ventilatory response to increasing CO₂ concentrations] in preterm infants who exhibit apnea (Gerhardt and Bancalari, 1984). However, a cause-and-effect relationship between decreased CO₂ responsiveness and apnea of prematurity has not been clearly established. Both entities may simply reflect decreased respiratory drive. Administration of CO₂ at low inspiratory concentrations would be expected to relieve apnea, as it does periodic breathing. This approach has been investigated, but may not be therapeutically feasible in human infants. Interestingly, the hypercapneic response of infants born to smoking and substance-abusing mothers is reduced, and as discussed earlier, may contribute to vulnerable respiratory control in this population (Ali et al., 2014).

Infants respond to a fall in inspired oxygen concentration with a transient increase in ventilation over approximately one minute, followed by a return to baseline or even depression of ventilation. The characteristic response to low oxygen in infants appears to result from initial peripheral chemoreceptor stimulation, followed by overriding depression of the respiratory center as a result of hypoxemia. This biphasic response in preterm infants may persist for several weeks. In convalescing premature infants, those with the most markedly increased ventilatory response to hypoxia also had some more apneic events recorded at baseline, suggesting that enhanced peripheral chemoreceptor activity in these infants causes respiratory control instability (Nock et al., 2004). Thus both decreased and increased peripheral chemosensitivity, may increase vulnerability of neonatal respiratory control. In very preterm infants [<1500 gm], the biphasic response may be absent, and such infants show only a sustained decrease in ventilation in response to hypoxia (Alvaro et al., 1992). Such hypoxic respiratory depression may be useful to conserve energy in the hypoxic intrauterine environment where respiratory activity is only intermittent and not contributing to gas exchange. In the postnatal environment, a progressive decrease in inspired oxygen concentration causes a significant flattening of CO₂ responsiveness in preterm infants, consistent with decreased excitability of the respiratory control system. This unstable response to low inspired oxygen concentration may play an important role in the origin of neonatal apnea. It explains why the incidence of apnea is reduced when a slightly increased concentration of inspired oxygen is administered to apneic infants. However, this approach to treat apnea should not be recommended because of potential adverse effects of hyperoxemia. Prolonged vulnerability of respiratory control to hypoxic stress in preterm infants is consistent with persistence of the characteristic biphasic ventilatory response to hypoxia into the second month of postnatal life.

Role of Mechanoreceptors

Afferent neural input from pulmonary stretch receptors modulates respiratory timing in human neonates. This vagally-mediated response, called the Hering-Breuer reflex, inhibits inspiration, prolongs expiration, or both, in response to increasing lung volume, thereby limiting lung overinflation. The apnea induced by sudden lung inflation, and resultant passive exhalation, may be useful to assess respiratory system mechanics during pulmonary function testing. Another manifestation of this reflex response to lung inflation is that inspiratory duration of an obstructed inspiratory effort appears to be an appropriate compensatory mechanism during airway occlusion, and may be limited in preterm infants.

Another manifestation of the Hering-Breuer reflex is the immediate slowing of respiratory rate when continuous positive airway pressure [CPAP] is administered to stabilize lung volume and improve oxygenation. This therapy actually decreases apnea by enhancing upper airway patency and eliminating the obstructive component of mixed apneas described earlier.

The so-called laryngeal chemoreflex refers to the apnea, bradycardia and glottic closure elicited by stimulation of upper airway afferents as occurs with tactile or fluid stimuli. It clearly serves a protective function to prevent lung aspiration. In neonates the response is frequently enhanced and clinically significant apnea may result, at times requiring intervention.

Physiologic Basis for Therapies

Non-pharmacologic approaches rely primarily on non-invasive modes of ventilatory support and optimization of blood gas status. Bloch-Salisbury and colleagues have demonstrated that their novel technique of stochastic mechanosensory stimulation, using a mattress with imbedded actuators, is able to stabilize respiratory patterns in preterm infants as manifested by a decrease in apnea and an almost three-fold decrease in percentage of time with oxygen saturations below 85% (Bloch-Salisbury et al., 2009). 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 worthy of future study.

The mainstay of pharmacologic therapy for apnea of prematurity is the xanthines, especially caffeine. Their primary mechanism of action in the perinatal period is thought to be blockade of inhibitory adenosine A₁ receptors with resultant excitation of respiratory neural output (Herlenius et al., 2002). An alternative mechanism of caffeine action is blockade of excitatory adenosine A_2A receptors at GABAergic neurons and resultant decrease in GABA output, resulting in excitation of respiratory neural output (Mayer et al., 2006). The xanthines also inhibit phosphodiesterase, which normally breaks down cyclic adenosine monophosphate [cAMP], although the relationship of cAMP accumulation to relief of apnea in infants is questionable.

The complex neurotransmitter interactions elicited by caffeine led to concerns regarding its safety. A large multicenter trial, however, demonstrated that caffeine treatment [used to treat apnea or enhance extubation] is effective in decreasing the rate of BPD and improving neurodevelopmental outcome at 18 to 21 months, especially in those receiving respiratory support (Davis et al., 2010; Schmidt et al., 2007). There is also evidence for reduction in developmental coordination disorder in the caffeine-treated cohort at 5 years of age (Doyle et al., 2014), and reduced risk of motor impairment at 11 years of age, without clear academic benefit (Schmidt et al., 2017). This benefit may be secondary to decrease in apnea and resultant intermittent hypoxemic episodes; however, that is speculative.

Xanthines increase central respiratory drive in the neonatal period without a detectable increase in arousal or change in sleep patterns, which is surprising, given the relatively high caffeine levels to which these infants are exposed. Data in neonatal rodents demonstrate an anti-inflammatory effect of caffeine in proinflammatory states elicited by postnatal hyperoxia or antenatal endotoxin exposure (Dumpa et al., 2019; Endesfelder et al., 2017; Köroğlu et al., 2014; Weichelt et al., 2013). In these studies, improved lung pathology and respiratory system mechanics were observed after caffeine treatment. In contrast, other data raise concerns about potential adverse effects of neonatal caffeine exposure in various animal models (Dayanim et al., 2014; Desfrere et al., 2007). Data on the effects of caffeine on the developing brain are controversial and include no effect in an ovine model (Atik et al., 2014), a protective effect on hypoxia-induced perinatal white matter injury (Back et al., 2006), and improved white matter structural development in preterm infants (Doyle et al., 2010). These conflicting data in the face of clinical benefit suggest that changes in dosing and indications for caffeine that deviate from proven beneficial protocols should proceed with caution.

Sudden Infant Death Syndrome [SIDS]

SIDS is a tragic event which, by definition, typically occurs without either antecedent warning or a clearly definable etiology in an otherwise healthy term infant. Apnea of prematurity and SIDS are considered separate conditions. This distinction is hampered by the fact that SIDS has a higher incidence in former preterm infants related to the lingering medical, environment, and sociodemographic vulnerability associated with prematurity (Ostfeld et al., 2017). The peak occurrence of SIDS occurs at a slightly earlier age in preterm vs term infants, in whom the incidence peaks at 2–5 postnatal months. For both preterm and term infants this is beyond the postmenstrual [corrected gestational] age of approximately 43 weeks at which apnea of prematurity has largely resolved (Ramanathan et al., 2001). Given that the incidence of SIDS has decreased by more than half with the introduction of the international campaign to avoid all sleep positions except supine, what can we learn about its etiology? A widely held belief is that an infant may asphyxiate in the prone position, implying that increasing hypercapnia and/or hypoxia will lead to apnea in the absence of a mature arousal response. Unfortunately, this is untestable in the case of SIDS. Supportive data have shown that infants of 36 to 42 weeks’ gestation have a greater decline in minute ventilation during a hypoxic challenge in the prone position if their mothers had misused substances and/or smoked during pregnancy (Rossor et al., 2018). Both of these maternal habits are known risk factors for SIDS (Bednarczuk et al., 2020). The most classic autopsy studies were performed by Duncan and colleagues on 41 cases of SIDS and a smaller number of controls to whom non-SIDS deaths were ascribed (Duncan et al., 2010). The focus of this work was on the serotonin system which was hypothesized to be deficient in SIDS cases, although it remains unclear whether underproduction of serotonin has a developmental, genetic, or environmental origin. Their underlying concept is vulnerability during a critical developmental period in addition to an exogenous stressor to align for SIDS to occur. The study found lower levels of medullary serotonin and its key biosynthetic enzyme in the SIDS group (Duncan et al., 2010). It remains unclear whether underproduction of serotonin has a developmental, genetic, or environmental etiology.

Despite important apparent distinctions between apnea of prematurity and SIDS, an influence of prematurity on peripheral chemosensitivity has been proposed as a contributor to SIDS (Gauda et al., 2007). For example, prior extremes of hypoxia or hyperoxia may induce plasticity in peripheral chemoreceptors that has adverse consequences when the infant is then placed in an adverse [e.g., prone] environment. Similarly, while laryngeal chemoreceptor-induced apnea may have a protective effect to prevent inhalation of gastric aspirate, an excessive response may, theoretically, trigger profound and potentially lethal apnea.

Animal models have been employed to try and elucidate SIDS pathobiology. For example, anesthetized piglets exposed to recurrent hypoxia in the first 10 days of life demonstrated increased inhibition of phrenic neural output during successive hypoxic exposures. Central GABAergic inhibition contributed to this effect which resolved beyond two weeks of life (Miller et al., 2000). In rodents, extensive analysis has been conducted to elucidate the postnatal changes in brain neurochemistry with the aim of identifying critical and vulnerable windows of development. Identification of such windows could serve as neurodevelopmental correlates of the vulnerable period of SIDS, which peaks at 2–5 months. In the rat brainstem, for example, various neurotransmitters (e.g. glutamate, GABA, and 5-HT) change abruptly and, in some cases, transiently toward the end of the second postnatal week (Liu and Wong-Riley, 2010b; Liu and Wong-Riley, 2010a). While these appear to represent important stages of normal CNS development, such changes coincide with a functional loss of ventilatory defense responses to acute hypoxia (Liu et al., 2006), which has been a proposed feature of SIDS. For example, SIDS victims and their siblings have reduced ventilatory defense responses to hypoxia and hypercapnia as well as prolonged sleep apnea and excessive periodic breathing (Valdés-Dapena, 1980). In a rodent model of SIDS, neonatal hypoxia exposure during this critical window decreased brainstem serotonin levels, which was associated with loss of the hypoxic ventilatory defense response and an unexpected rate of mortality (MacFarlane et al., 2016; Mayer et al., 2014). In that study, an increased brainstem microglia expression was observed, which is consistent with similar reports of gliosis in SIDS cases (Kinney et al., 1983; Obonai et al., 1997), suggesting that postnatal hypoxia maybe a sufficient insult to disrupt important milestones in neurochemical development (e.g. serotonin) leading to impaired arousal and cardiorespiratory defense responses. Thus, the repeated prolonged bouts of hypoxia that may occur in infants during the prone sleeping position may be a significant “environmental” stressor that contributes to the brainstem serotonergic abnormalities seen in SIDS cases. This is especially significant since prolonged hypoxia exposure has been implicated in SIDS.

While studies of the brainstem neurochemical and physiological abnormalities in animal models have provided some anecdotal insight into the pathobiology of SIDS, significant peripheral abnormalities in respiratory neural control have also been implicated (Gauda et al., 2007; Porzionato et al., 2018). As discussed above, changes in carotid body function (and morphology) likely play a key role in genesis and resolution of apnea in preterm infants (see also (MacFarlane et al., 2013) for review) and could contribute to their increased risk of SIDS. Analysis of the carotid bodies of SIDS cases have revealed decreased carotid body size and increased expression of the inhibitory neurotransmitter dopamine (Perrin et al., 1984). Several studies have highlighted the importance of a functionally competent carotid body in respiratory control, particularly during sensitive periods of development. Specifically, denervation of the rat carotid body at 1 week of age resulted in a higher incidence of mortality than in older animals who were denervated (Serra et al., 2001), which was also a pattern observed in carotid body denervated pigs (Côté et al., 1996; Haddad and Donnelly, 1988) and lambs.(Bureau et al., 1985) Further, piglets (who survived carotid body denervation) exhibited pronounced apnea and hypoventilation (Donnelly and Haddad, 1990), and carotid body denervated dogs failed to arouse from hypoxia during sleep and required resuscitation (Bowes et al., 1981). These data suggest ineffective or loss of carotid body inputs during critical stages of development could have lethal consequences, which highlights the significance such an abnormality might have in the susceptibility to SIDS.

An immature respiratory neural control system of the preterm infant implies a functionally compromised carotid body, at least in the immediate postnatal period. However, data points to a progressive development of carotid body hyper-excitability manifesting as sensory facilitation resulting from the IH events associated with apnea of prematurity. Specifically, the worsening of the IH events during the first postnatal weeks likely contributes to the carotid body hyper-excitability. However, it is not clear how carotid body excitability changes beyond the neonatal ICU period and whether it contributes to the risk of SIDS, especially since the IH events begin to resolve during the 3–4 postnatal weeks. In one study, arousal from hypoxia was blunted in neonatal rats pretreated with acute IH, which became more exaggerated with age and also depended on the period of recovery between episodes of hypoxia (Darnall et al., 2010). In contrast, hyperoxia exposure in neonatal rats to mimic supplemental O₂ therapy in preterm infants, resulted in abnormal development of the carotid bodies and degeneration of the associated afferent neurons (Bavis et al., 2013; Erickson et al., 1998). Further, infants with BPD have attenuated ventilatory defense responses to hypoxia suggesting impaired carotid body sensitivity, which might also be accompanied with hypoplasia (Bates et al., 2018; Calder et al., 1994).

At a minimum, it would seem that ineffective carotid bodies are a major risk factor for SIDS, but whether such abnormalities underlie the increased risk for preterm infants is still largely speculative. This is due in part to the complex ways in which preterm infants “experience” variations in postnatal O₂: 1) preemies experience intermittent hypoxemia as a consequence of AOP; 2) “excessive” O₂ exposure can occur either by virtue of being born into a relatively hyperoxic environment prematurely and in combination with delivery of supplemental O₂; and 3) perhaps to a lesser extent, prolonged hypoxemia from sustained periods of ineffective maintenance of O₂ saturation. All of these features can have complex effects on carotid body development and function, making it difficult to determine how such an organ, which is also involved in the control of apnea, could provide the mechanistic link between AOP and vulnerability to SIDS in this population.

Conclusion

Immature respiratory control remains a subject of great clinical and research interest. For clinicians, apnea of prematurity with the resultant intermittent hypoxemia and bradycardia presents a major clinical problem requiring both invasive and non-invasive interventions. Physiologic and basic biologic studies have largely employed animal and cellular models and considerable advances have been made in our understanding of respiratory control developmental. Additionally, there is increasing interest in the potential longer-term adverse effects of recurrent apnea. Beyond the immediate neonatal period, SIDS is an uncommon but devastating event in which immature or abnormal respiratory control is likely an underlying feature. While recognizing the diverse range of risk factors for SIDS has advanced our understanding of its etiology, there are major gaps in our knowledge of the specific pathophysiology. Filling these gaps will require ongoing novel research employing innovative animal and clinical models in order to advance preventive and therapeutic interventions.