- Email: [email protected]
The mesencephalic nucleus of the trigeminal nerve and the SIDS
Medical Hypotheses 84 (2015) 8–10
Contents lists available at ScienceDirect
Medical Hypotheses journal homepage: www.elsevier.com/locate/mehy
The mesencephalic nucleus of the trigeminal nerve and the SIDS Giovanni Andrisani ⇑, Giorgia Andrisani
a r t i c l e
i n f o
Article history: Received 22 July 2014 Accepted 9 November 2014
a b s t r a c t Sudden infant death syndrome (SIDS) is a major cause of infant mortality throughout the world, yet its cause and mechanism of action remain poorly understood. Here, we discuss a novel model of the etiology of SIDS which ties together what is known about the brain regions thought to be affected in SIDS infants with a defined neuroanatomical circuit and a documented preventative factor in young children. We propose that SIDS occurs due to a lack of sufficient development and plasticity of glutamatergic synapses in the mesencephalic nucleus of the trigeminal nerve (Me5) and reticular formation (RF) of the brainstem. This model is supported by evidence of brainstem dysfunction in SIDS as well as evidence of signaling through the Me5 and RF in another means of regulating cortical arousal. Furthermore, long-term plasticity of glutamatergic synapses is well known to play a critical role in learning and memory in other regions of the brain, implying that those mechanisms may also be relevant in the development of brainstem circuitry. This model clearly explains why SIDS deaths appear so suddenly with little pathological explanation and suggests a potentially novel way to prevent these deaths from occurring. Ó 2014 Elsevier Ltd. All rights reserved.
Introduction Sudden infant death syndrome (SIDS) is the leading cause of death in children below one year of age, with an incidence ranging from 0.5 to 1 per thousand births. It manifests as the sudden and unexpected death of a seemingly healthy infant, usually after the child has gone to sleep. Children who die of SIDS generally show no signs of suffering, and definitive diagnosis only occurs if the post-mortem examination of the child shows no clear explanation for their death. The Lino Rossi Research Center in Milan examines post-mortem brains of SIDS infants in comparison to control cases for whom cause of death could be definitively determined. The most significant alterations observed in the brains of SIDS or sudden intrauterine undefined death syndrome (SIUDS) cases are insufficient development of the brainstem and cerebellum, two brain regions critical for modulating critical survival functions such as respiration, heartbeat, digestion, and arousal [1,2]. This is characterized by delayed maturation at the cerebellar ventricular (ependymal) and bulbar level and apoptosis in the hippocampus and dorsal brainstem nuclei [3]. Even today, pathological anatomical studies on the young patients who died from SIDS fail to find evidence of life-threatening injury, instead merely seeing seemingly minor alterations, most notably at the brainstem level [1,4]. The hypothesis described herein is particularly suited to these findings, as a ⇑ Corresponding author at: via della croce, 47, Matera 75100, Italy. Cell: +39 3392062102, +39 0835387878 (Clinic); fax: +39 0835387878. E-mail address: [email protected] (G. Andrisani). http://dx.doi.org/10.1016/j.mehy.2014.11.005 0306-9877/Ó 2014 Elsevier Ltd. All rights reserved.
gradual decrease in glutamate signaling in the RF would not necessarily lead to any serious pathological injury, but rather to the slow and steady cessation of vital functions. Hypothesis In accordance with this data, we have developed a theory for the etiology of SIDS that explains both the lack of gross neuropathological findings and why these deaths occur in this way, as well as suggests a rationale behind certain preventive factors. We hypothesize that SIDS occurs as a result of insufficient glutamate signaling in the reticular formation (RF) in the brainstem, which can be ameliorated by stimulation of that region through mechanoreceptors in the oral cavity via the mesencephalic nucleus of the trigeminal nerve (Me5). Evaluation of the hypothesis Involvement of the brain stem and reticular formation in SIDS The RF is a brainstem structure which serves to control vital functions such as breathing, temperature regulation, cardiac function, and blood pH, amongst other things [5]. It is highly evolutionarily conserved and is critically involved in controlling the sleep–wake cycle in humans, as well as alertness and arousal via the ascending reticular activating system (ARAS) [6]. Loss of function in the reticular formation often results in changes in consciousness and can lead to coma or death. Importantly, many
G. Andrisani, G. Andrisani / Medical Hypotheses 84 (2015) 8–10
researchers have long believed that SIDS originates in the brainstem and, as the RF is a key part of this brain region and is the seat of control of many vital functions, we can logically conclude that the RF plays a critical role in the pathogenesis of SIDS. Indeed, this conclusion is supported by a number of pathological studies of brain anatomy in SIDS deaths which reveal damage to RF structures [1,7]. These findings together strongly support the idea that SIDS occurs as a result of aberrant signaling in the brainstem, and suggest that understanding potential mechanisms by which this develops and can be treated will be key to decreasing the incidence of SIDS. Potential role for Me5 in strengthening synapses in reticular formation Intriguingly, one particular brain region, the Me5, is known to both connect to and stimulate the RF, and may play a critical role in facilitating activation of the cortical arousal response via the ARAS in response to oral stimulation. Many researchers have noted that gum chewing is positively correlated with attentiveness and reaction time, and functional magnetic resonance imaging (fMRI) studies showed that the brain regions most active during mastication were related to both movement and attention. This suggests that chewing induced an increase in the level of arousal and attention, as well as had an impact on motor control, which, when combined, could lead to improved cognitive performance. Previous studies show that the Me5 has a close relationship with the RF and implicate the Me5 in maintaining sufficient RF function during the early stages of sleep, when other means of RF stimulation are absent [8]. The Me5 is the only formation of the central nervous system in which there are cell bodies of primary afferent sensory neurons, and it functions primarily as an intraneuraxial ganglion [9]. Me5 cells are primarily activated due to stimulation of mechanoreceptors in the periodontal ligament. The degree of excitation and the resulting response of mechanoreceptors depend not only on the characteristics of the receptor, but also the amount of force required to induce the movement of the tooth. Receptors that project to the Me5 are concentrated around the root tip of the tooth, and recordings made from the Me5 have demonstrated the presence of neurons that respond to stresses of the teeth, the hard palate, and the opening of the mouth [10]. The caudal region of the Me5 is continuous with the reticular formation and projects to many nuclei therein, indicating that activation of the Me5 by chewing has the ability to alter RF activity. The RF, in turn, signals through the ARAS to stimulate cortical arousal and alertness [11]. Importantly, all of these structures are present and functional before the presence of teeth, which is of particular relevance when one takes into account the fact that pacifier use is associated with an up to 90% reduction in the risk of SIDS [12]. Importance and mechanism of development of glutamatergic synapses in the reticular formation
9
positive feedback mechanism called long-term potentiation (LTP), synapses that are highly used become stronger and increase in number (Hebb’s rule), and therefore the neuron is more likely to be activated due to stimulation at that synaptic site [14]. Though it remains to be understood precisely why or how presynaptic release of glutamate leads to the maturation of the brainstem synapses, it stands to reason that the same mechanisms that are at play in the cerebral cortex and hippocampus exist in neurons in the Me5 and RF if the mechanism demonstrated for the cortex and hippocampus (LTP) was also demonstrated in the FR under the stimulation of glutamatergic Me5, that would clarify the reasons for the lack of specific neurotransmitters in the FR (especially ACH and 5-HT) that are object of the latest research. If this is indeed the case, one could suppose that activating the Me5 and RF through oral stimulation, for example by the use of a pacifier, could induce LTP in those circuits, thereby maintaining a higher level of activation in the RF and ARAS even during sleep and decreasing the likelihood of SIDS. Conclusion In conclusion, the hypothesis that we have presented here describes a plausible model of how SIDS occurs, while also explaining the lack of major neuropathological findings in post-mortem brains of SIDS patients. It is in agreement with previous studies [1,15,16] that identify the brainstem as the primary site of dysfunction in SIDS and provides a means to understand why these infant deaths almost exclusively occur during sleep in children who show no signs of cardiac or respiratory distress during wakefulness. Furthermore, this model implicates a clear mechanism by which pacifier use could be preventative for SIDS. Future studies should address whether or not pacifier use is sufficient to induce Me5 and RF activation, as well as the role of Me5 in the CNS as a producer of glutamate and the consequences that derive from its position and its connections with main brainstem nuclei. Most importantly, thorough exploration of this hypothesis can result in the development of better preventative care in infants to avoid SIDS. With better understanding of the mechanism by which SIDS occurs, combined with heart rate monitoring, particularly in high risk infants, it may be possible to both recognize the onset of a sudden cardiac arrhythmia in a sleeping infant prior to death and prevent death from occurring by immediately stimulating the brainstem, potentially through something as simple as using a pacifier. This has the potential to be a low cost means of significantly reducing infant death worldwide and warrants further scientific attention. Conflict of interest statement None. References
The Me5 is in the vicinity of nuclei that produce the major neurotransmitters in the CNS, and neurons within the Me5 express receptors for serotonin, acetylcholine, dopamine, norepinephrine, histamine, and glutamate, suggesting that Me5 activity can be modulated by many different neurotransmitters. However, studies suggest that most cells in the Me5 respond primarily to glutamate [13]. Glutamate is a major excitatory neurotransmitter throughout the brain, and the development of glutamatergic synapses has been well characterized in many brain regions. Conventionally, developing neurons are thought to release glutamate as a means of attracting synaptic partners and consolidating the connections between pairs of neurons. Upon release of glutamate by the presynaptic neuron, specific receptors, most notably AMPA, kainate and NMDA receptors, are activated in the postsynaptic neuron. Through a
[1] Matturri L, Ottaviani G, Alfonsi G, Crippa M, Rossi L, Lavezzi AM. Study of the brainstem, particularly the arcuate nucleus, in sudden infant death syndrome (SIDS) and sudden intrauterine unexplained death (SIUD). Am J Forensic Med Pathol 2004;25:44–8. [2] Mecchia D, Casale V, Oneda R, Matturri L, Lavezzi AM. Sudden death of an infant with cardiac, nervous system and genetic involvement – a case report. Diagn Pathol 2013;8:159. [3] Lavezzi AM, Ottaviani G, Mauri M, Matturri L. Alterations of biological features of the cerebellum in sudden perinatal and infant death. Curr Mol Med 2006;6:429–35. [4] Lavezzi AMM. Hypoplasia of the parafacial/facial complex: a very frequent finding in sudden unexplained fetal death. Open Neurosci J 2008;2:1–5. [5] Kandel ER. Principles of neural science. New York: McGraw-Hill; 2013. [6] Torterolo P, Vanini G. New concepts in relation to generating and maintaining arousal. Rev Neurol 2010;50:747–58. [7] Lavezzi AM, Ottaviani G, Ballabio G, Rossi L, Matturri L. Preliminary study on the cytoarchitecture of the human parabrachial/Kolliker-fuse complex, with
10
[8]
[9]
[10]
[11]
G. Andrisani, G. Andrisani / Medical Hypotheses 84 (2015) 8–10 reference to sudden infant death syndrome and sudden intrauterine unexplained death. Pediatr Dev Pathol 2004;7:171–9. Capra NF, Dessem D. Central connections of trigeminal primary afferent neurons: topographical and functional considerations. Crit Rev Oral Biol Med 1992;4:1–52. Deriu S, La Stimolazione elettrica trigeminale modifica l’espressione genica di c-Fos e diminuisce le crisi epilettiche nel ratto [vol. Ph.D.], Università di Sassari, Università degli studi di Sassari; 2012. Linden RWA. Periodontal mechanoreceptors and their functions. In: Taylor A, editor. Neurophysiology of the jaws and teeth. New York: Macmillan; 1990. p. 52–88. Silva S, Alacoque X, Fourcade O, et al. Wakefulness and loss of awareness: brain and brainstem interaction in the vegetative state. Neurology 2010;74:313–20.
[12] Moon RY, Fu L. Sudden infant death syndrome: an update. Pediatr Rev 2012;33:314–20. [13] Lazarov NE. Comparative analysis of the chemical neuroanatomy of the mammalian trigeminal ganglion and mesencephalic trigeminal nucleus. Prog Neurobiol 2002;66:19–59. [14] Smolen P. A model of late long-term potentiation simulates aspects of memory maintenance. PLoS One 2007;2:e445. [15] Lavezzi AM, Mauri M, Mecchia D, Matturri L. Developmental alterations of the prefrontal cerebral cortex in sudden unexplained perinatal and infant deaths. J Perinat Med 2009;37:297–303. [16] Matturri L, Minoli I, Lavezzi AM, Cappellini A, Ramos S, Rossi L. Hypoplasia of medullary arcuate nucleus in unexpected late fetal death (stillborn infants): a pathologic study. Pediatrics 2002;109:E43.
Contents lists available at ScienceDirect
Medical Hypotheses journal homepage: www.elsevier.com/locate/mehy
The mesencephalic nucleus of the trigeminal nerve and the SIDS Giovanni Andrisani ⇑, Giorgia Andrisani
a r t i c l e
i n f o
Article history: Received 22 July 2014 Accepted 9 November 2014
a b s t r a c t Sudden infant death syndrome (SIDS) is a major cause of infant mortality throughout the world, yet its cause and mechanism of action remain poorly understood. Here, we discuss a novel model of the etiology of SIDS which ties together what is known about the brain regions thought to be affected in SIDS infants with a defined neuroanatomical circuit and a documented preventative factor in young children. We propose that SIDS occurs due to a lack of sufficient development and plasticity of glutamatergic synapses in the mesencephalic nucleus of the trigeminal nerve (Me5) and reticular formation (RF) of the brainstem. This model is supported by evidence of brainstem dysfunction in SIDS as well as evidence of signaling through the Me5 and RF in another means of regulating cortical arousal. Furthermore, long-term plasticity of glutamatergic synapses is well known to play a critical role in learning and memory in other regions of the brain, implying that those mechanisms may also be relevant in the development of brainstem circuitry. This model clearly explains why SIDS deaths appear so suddenly with little pathological explanation and suggests a potentially novel way to prevent these deaths from occurring. Ó 2014 Elsevier Ltd. All rights reserved.
Introduction Sudden infant death syndrome (SIDS) is the leading cause of death in children below one year of age, with an incidence ranging from 0.5 to 1 per thousand births. It manifests as the sudden and unexpected death of a seemingly healthy infant, usually after the child has gone to sleep. Children who die of SIDS generally show no signs of suffering, and definitive diagnosis only occurs if the post-mortem examination of the child shows no clear explanation for their death. The Lino Rossi Research Center in Milan examines post-mortem brains of SIDS infants in comparison to control cases for whom cause of death could be definitively determined. The most significant alterations observed in the brains of SIDS or sudden intrauterine undefined death syndrome (SIUDS) cases are insufficient development of the brainstem and cerebellum, two brain regions critical for modulating critical survival functions such as respiration, heartbeat, digestion, and arousal [1,2]. This is characterized by delayed maturation at the cerebellar ventricular (ependymal) and bulbar level and apoptosis in the hippocampus and dorsal brainstem nuclei [3]. Even today, pathological anatomical studies on the young patients who died from SIDS fail to find evidence of life-threatening injury, instead merely seeing seemingly minor alterations, most notably at the brainstem level [1,4]. The hypothesis described herein is particularly suited to these findings, as a ⇑ Corresponding author at: via della croce, 47, Matera 75100, Italy. Cell: +39 3392062102, +39 0835387878 (Clinic); fax: +39 0835387878. E-mail address: [email protected] (G. Andrisani). http://dx.doi.org/10.1016/j.mehy.2014.11.005 0306-9877/Ó 2014 Elsevier Ltd. All rights reserved.
gradual decrease in glutamate signaling in the RF would not necessarily lead to any serious pathological injury, but rather to the slow and steady cessation of vital functions. Hypothesis In accordance with this data, we have developed a theory for the etiology of SIDS that explains both the lack of gross neuropathological findings and why these deaths occur in this way, as well as suggests a rationale behind certain preventive factors. We hypothesize that SIDS occurs as a result of insufficient glutamate signaling in the reticular formation (RF) in the brainstem, which can be ameliorated by stimulation of that region through mechanoreceptors in the oral cavity via the mesencephalic nucleus of the trigeminal nerve (Me5). Evaluation of the hypothesis Involvement of the brain stem and reticular formation in SIDS The RF is a brainstem structure which serves to control vital functions such as breathing, temperature regulation, cardiac function, and blood pH, amongst other things [5]. It is highly evolutionarily conserved and is critically involved in controlling the sleep–wake cycle in humans, as well as alertness and arousal via the ascending reticular activating system (ARAS) [6]. Loss of function in the reticular formation often results in changes in consciousness and can lead to coma or death. Importantly, many
G. Andrisani, G. Andrisani / Medical Hypotheses 84 (2015) 8–10
researchers have long believed that SIDS originates in the brainstem and, as the RF is a key part of this brain region and is the seat of control of many vital functions, we can logically conclude that the RF plays a critical role in the pathogenesis of SIDS. Indeed, this conclusion is supported by a number of pathological studies of brain anatomy in SIDS deaths which reveal damage to RF structures [1,7]. These findings together strongly support the idea that SIDS occurs as a result of aberrant signaling in the brainstem, and suggest that understanding potential mechanisms by which this develops and can be treated will be key to decreasing the incidence of SIDS. Potential role for Me5 in strengthening synapses in reticular formation Intriguingly, one particular brain region, the Me5, is known to both connect to and stimulate the RF, and may play a critical role in facilitating activation of the cortical arousal response via the ARAS in response to oral stimulation. Many researchers have noted that gum chewing is positively correlated with attentiveness and reaction time, and functional magnetic resonance imaging (fMRI) studies showed that the brain regions most active during mastication were related to both movement and attention. This suggests that chewing induced an increase in the level of arousal and attention, as well as had an impact on motor control, which, when combined, could lead to improved cognitive performance. Previous studies show that the Me5 has a close relationship with the RF and implicate the Me5 in maintaining sufficient RF function during the early stages of sleep, when other means of RF stimulation are absent [8]. The Me5 is the only formation of the central nervous system in which there are cell bodies of primary afferent sensory neurons, and it functions primarily as an intraneuraxial ganglion [9]. Me5 cells are primarily activated due to stimulation of mechanoreceptors in the periodontal ligament. The degree of excitation and the resulting response of mechanoreceptors depend not only on the characteristics of the receptor, but also the amount of force required to induce the movement of the tooth. Receptors that project to the Me5 are concentrated around the root tip of the tooth, and recordings made from the Me5 have demonstrated the presence of neurons that respond to stresses of the teeth, the hard palate, and the opening of the mouth [10]. The caudal region of the Me5 is continuous with the reticular formation and projects to many nuclei therein, indicating that activation of the Me5 by chewing has the ability to alter RF activity. The RF, in turn, signals through the ARAS to stimulate cortical arousal and alertness [11]. Importantly, all of these structures are present and functional before the presence of teeth, which is of particular relevance when one takes into account the fact that pacifier use is associated with an up to 90% reduction in the risk of SIDS [12]. Importance and mechanism of development of glutamatergic synapses in the reticular formation
9
positive feedback mechanism called long-term potentiation (LTP), synapses that are highly used become stronger and increase in number (Hebb’s rule), and therefore the neuron is more likely to be activated due to stimulation at that synaptic site [14]. Though it remains to be understood precisely why or how presynaptic release of glutamate leads to the maturation of the brainstem synapses, it stands to reason that the same mechanisms that are at play in the cerebral cortex and hippocampus exist in neurons in the Me5 and RF if the mechanism demonstrated for the cortex and hippocampus (LTP) was also demonstrated in the FR under the stimulation of glutamatergic Me5, that would clarify the reasons for the lack of specific neurotransmitters in the FR (especially ACH and 5-HT) that are object of the latest research. If this is indeed the case, one could suppose that activating the Me5 and RF through oral stimulation, for example by the use of a pacifier, could induce LTP in those circuits, thereby maintaining a higher level of activation in the RF and ARAS even during sleep and decreasing the likelihood of SIDS. Conclusion In conclusion, the hypothesis that we have presented here describes a plausible model of how SIDS occurs, while also explaining the lack of major neuropathological findings in post-mortem brains of SIDS patients. It is in agreement with previous studies [1,15,16] that identify the brainstem as the primary site of dysfunction in SIDS and provides a means to understand why these infant deaths almost exclusively occur during sleep in children who show no signs of cardiac or respiratory distress during wakefulness. Furthermore, this model implicates a clear mechanism by which pacifier use could be preventative for SIDS. Future studies should address whether or not pacifier use is sufficient to induce Me5 and RF activation, as well as the role of Me5 in the CNS as a producer of glutamate and the consequences that derive from its position and its connections with main brainstem nuclei. Most importantly, thorough exploration of this hypothesis can result in the development of better preventative care in infants to avoid SIDS. With better understanding of the mechanism by which SIDS occurs, combined with heart rate monitoring, particularly in high risk infants, it may be possible to both recognize the onset of a sudden cardiac arrhythmia in a sleeping infant prior to death and prevent death from occurring by immediately stimulating the brainstem, potentially through something as simple as using a pacifier. This has the potential to be a low cost means of significantly reducing infant death worldwide and warrants further scientific attention. Conflict of interest statement None. References
The Me5 is in the vicinity of nuclei that produce the major neurotransmitters in the CNS, and neurons within the Me5 express receptors for serotonin, acetylcholine, dopamine, norepinephrine, histamine, and glutamate, suggesting that Me5 activity can be modulated by many different neurotransmitters. However, studies suggest that most cells in the Me5 respond primarily to glutamate [13]. Glutamate is a major excitatory neurotransmitter throughout the brain, and the development of glutamatergic synapses has been well characterized in many brain regions. Conventionally, developing neurons are thought to release glutamate as a means of attracting synaptic partners and consolidating the connections between pairs of neurons. Upon release of glutamate by the presynaptic neuron, specific receptors, most notably AMPA, kainate and NMDA receptors, are activated in the postsynaptic neuron. Through a
[1] Matturri L, Ottaviani G, Alfonsi G, Crippa M, Rossi L, Lavezzi AM. Study of the brainstem, particularly the arcuate nucleus, in sudden infant death syndrome (SIDS) and sudden intrauterine unexplained death (SIUD). Am J Forensic Med Pathol 2004;25:44–8. [2] Mecchia D, Casale V, Oneda R, Matturri L, Lavezzi AM. Sudden death of an infant with cardiac, nervous system and genetic involvement – a case report. Diagn Pathol 2013;8:159. [3] Lavezzi AM, Ottaviani G, Mauri M, Matturri L. Alterations of biological features of the cerebellum in sudden perinatal and infant death. Curr Mol Med 2006;6:429–35. [4] Lavezzi AMM. Hypoplasia of the parafacial/facial complex: a very frequent finding in sudden unexplained fetal death. Open Neurosci J 2008;2:1–5. [5] Kandel ER. Principles of neural science. New York: McGraw-Hill; 2013. [6] Torterolo P, Vanini G. New concepts in relation to generating and maintaining arousal. Rev Neurol 2010;50:747–58. [7] Lavezzi AM, Ottaviani G, Ballabio G, Rossi L, Matturri L. Preliminary study on the cytoarchitecture of the human parabrachial/Kolliker-fuse complex, with
10
[8]
[9]
[10]
[11]
G. Andrisani, G. Andrisani / Medical Hypotheses 84 (2015) 8–10 reference to sudden infant death syndrome and sudden intrauterine unexplained death. Pediatr Dev Pathol 2004;7:171–9. Capra NF, Dessem D. Central connections of trigeminal primary afferent neurons: topographical and functional considerations. Crit Rev Oral Biol Med 1992;4:1–52. Deriu S, La Stimolazione elettrica trigeminale modifica l’espressione genica di c-Fos e diminuisce le crisi epilettiche nel ratto [vol. Ph.D.], Università di Sassari, Università degli studi di Sassari; 2012. Linden RWA. Periodontal mechanoreceptors and their functions. In: Taylor A, editor. Neurophysiology of the jaws and teeth. New York: Macmillan; 1990. p. 52–88. Silva S, Alacoque X, Fourcade O, et al. Wakefulness and loss of awareness: brain and brainstem interaction in the vegetative state. Neurology 2010;74:313–20.
[12] Moon RY, Fu L. Sudden infant death syndrome: an update. Pediatr Rev 2012;33:314–20. [13] Lazarov NE. Comparative analysis of the chemical neuroanatomy of the mammalian trigeminal ganglion and mesencephalic trigeminal nucleus. Prog Neurobiol 2002;66:19–59. [14] Smolen P. A model of late long-term potentiation simulates aspects of memory maintenance. PLoS One 2007;2:e445. [15] Lavezzi AM, Mauri M, Mecchia D, Matturri L. Developmental alterations of the prefrontal cerebral cortex in sudden unexplained perinatal and infant deaths. J Perinat Med 2009;37:297–303. [16] Matturri L, Minoli I, Lavezzi AM, Cappellini A, Ramos S, Rossi L. Hypoplasia of medullary arcuate nucleus in unexpected late fetal death (stillborn infants): a pathologic study. Pediatrics 2002;109:E43.