Published ahead of print on August 13, 2003, doi:10.1164/rccm.200301-134OC
© 2003 American Thoracic Society Incomplete Arousal Processes in Infants Who Were Victims of Sudden DeathDepartment of Pediatrics, Nagoya City University Medical School, Nagoya, Japan; Pediatric Sleep Unit, Neuropediatrics, Erasmus Hospital; Pediatric Sleep Unit, University Children's Hospital, Free University of Brussels, Brussels, Belgium; and St. Petersburg State Pediatric Medical Academy, St. Petersburg, Russia Correspondence and requests for reprints should be addressed to Andre Kahn, Hôpital Universitaire des Enfants Reine Fabiola, Avenue J.J. Crocq 15, B-1020 Brussels, Belgium. E-mail: akahn{at}ulb.ac.be
Infants who became victims of sudden infant death syndrome (SIDS) aroused less from sleep than control infants. This study was conducted to determine the characteristics of arousal from sleep of infants who eventually died of SIDS. Sixteen infants were monitored some days or weeks before they died of SIDS. Their polygraphic sleep recordings were compared with those of matched control infants. Arousals were scored as subcortical activation (incomplete arousals) or cortical arousal (complete arousals). Cortical arousals were significantly less frequent in the victims who would succumb to SIDS in the future than in the control infants during both REM and non-REM sleep (p = 0.039). The frequency (p = 0.017) and duration (p = 0.005) of subcortical activation were significantly greater in the infants who died of SIDS than in the control infants during REM sleep. Compared with the control infants, the infants who later died of SIDS had more frequent subcortical activation in the first part of the night, between 9:00 P.M. and 12:00 A.M. (p = 0.038), and fewer cortical arousals during the latter part of the night, between 3:00 and 6:00 A.M. (p = 0.011). The present data are suggestive of incomplete arousal processes in infants who eventually died at a time they were presumed to have been asleep.
Key Words: sleep • arousal • death • sudden infant death syndrome • apnea Failure to arouse from sleep has been suggested to contribute to sudden infant death syndrome (SIDS) (1, 2). An insufficient propensity to arouse could lower the chance of infants to survive when exposed to noxious conditions during sleep (3). Infants with an apparent life-threatening event had fewer arousals than control infants (4). A report on the analysis of heart rate and breathing variability of 16 infants who died of SIDS led to the conclusion that these infants were more frequently awake in the early part of the night and slept more during the end of the night than did control infants (5). Polysomnographic recordings confirmed that the victims who would succumb to SIDS in the future had fewer body movements and spent less time awake than control subjects (6, 7). Arousals reflect a progressive activation of various structures, from subcortical to cortical areas (8–12). Evidence has been collected showing that autonomic and brainstem arousals can occur without changes in cortical activity. Apneic events during sleep may end without electroencephalographic changes (13, 14). Transient arousal associated with rises in blood pressure occur without apparent EEG changes (13, 15). In healthy infants, sequential arousal behaviors such as sighs, startles, and thrashing limb movement precede full arousal (16). Complete arousals include both autonomic and cortical activation. Alternatively, subcortical or autonomic activation with no cortical involvement represents incomplete arousal reactions. The present study was undertaken to evaluate whether infants who later died of SIDS were characterized by an incomplete arousal process.
Infants Nighttime sleep studies were conducted in 10 Belgian sleep laboratories to assess infants' sleep characteristics or to alleviate parental anxiety about sleep apnea. All studies were conducted after 1977 according to the same standard protocol. Sixteen recordings were made on infants who eventually died within days or weeks of the sleep studies. The group of 16 victims of SIDS was composed of 2 infants studied because they were siblings of infants who died suddenly during the first year of life, 4 infants with loud snoring or heavy sweating during sleep, 2 with breath-holding spells, and 8 because of parental anxiety over sleep apnea. The group consisted of 6 girls and 10 boys. Their median age at recording was 11 weeks (range, 7–19 weeks). One child was born premature at 36 weeks of gestation. Nine children usually slept supine. The median time between sleep recording and death was 21 days (range, 4 days–7 weeks). Their deaths were sudden and unexpected, remained unexplained despite a complete history and postmortem examination, and were attributed to SIDS. No infant was undergoing sleep monitoring at the time of death. The sleep and breathing characteristics of some of the victims of SIDS had been reported previously (6, 7, 17). Control infants from the same prospective studies were matched with the victims of SIDS for sex, gestational age, parental smoking, usual sleep position, and age at the time of recording. The control infants had no family or personal history of apnea or SIDS and survived the first year of life. At the time of study, all infants were healthy, not sleep deprived, and receiving no medication.
Polygraphic Recordings
Data Analysis
Sleep Characteristics
Apnea
Cortical Arousal and Subcortical Activation
Statistical assessments were made with the Mann–Whitney U test, Kruskal–Wallis rank test, Spearman's rank correlation, and  2 test with Yates's transforms. Values were expressed as median and range. The level of significance adopted was p value less than 0.05. The aim and methodology of the study were approved by the Ethics Committee and were explained at the time of recording to the parents, who gave their informed consent.
In each group there were 16 infants (10 boys) with a median postnatal age of 11 weeks (range, 7–19 weeks); two of the mothers smoked during gestation, one infant was born premature, and seven infants slept prone. Comparing the infants who died of SIDS and the control infants, no differences were seen in total sleep duration (median value of 381 minutes in the infants who died of SIDS, 395 minutes in the control infants; range, 311–539 minutes), amount of REM sleep (median value of 56.4% in the infants who died of SIDS, 55.4% in the control infants; range, 35.2–69.0%), NREM sleep (median value of 43.7% in the infants who died of SIDS, 44.6% in the control infants; range, 31–64.8%); sleep efficiency (median value of 75.8% in the infants who died of SIDS, 75.2% in the control infants; range, 56–99.8%), or time awake (median value of 24.2% in the infants who died of SIDS, 24.8% in the control infants; range, 0.2–44.4%). There was no sex-related difference in sleep characteristics. In both the infants who died of SIDS and the control infants, cortical arousals were more frequent in REM than in NREM sleep (p < 0.001, in both groups) (Table 1) . Subcortical activation occurred more frequently in REM than in NREM sleep in the infants who would succumb to SIDS in the future (p = 0.038), not in the control subjects.
The infants who would succumb to SIDS in the future tended to have fewer total arousals than the control subjects during sleep, but the difference was not significant. Compared with the control subjects, the victims of SIDS had significantly fewer cortical arousals during REM and NREM sleep (p = 0.039). The infants who died of SIDS had more frequent subcortical activation than the control infants during REM sleep (p = 0.017). The difference in subcortical activation did not reach significance during NREM sleep. The ratio of cortical arousal to subcortical activation was significantly smaller in the infants who died of SIDS than in the control infants during REM sleep (p = 0.001). A similar trend was seen during NREM sleep (p = 0.073) (Figure 3) .
The duration of cortical arousals was similar in the two groups of infants. The duration of subcortical activations was significantly greater in the infants who died of SIDS than in the control infants, during both REM (median duration of 7 seconds in the infants who died of SIDS, 5 seconds in the control infants; range, 3–46 seconds; p = 0.005) and non-REM sleep (median duration of 5 seconds in the infants who died of SIDS, 4 seconds in the control infants; range, 3–26 seconds; p = 0.045). Although the frequency of total arousals increased during the night in both groups of infants, only in the control infants was the increase statistically significant (p = 0.007). Compared with the control infants, the victims who would succumb to SIDS in the future had significantly more frequent subcortical activations in the first part of the night, between 9:00 P.M. and 12:00 A.M. (frequency of subcortical activation in the victims of SIDS, 2.9; range, 0.4–5.3, and in the control subjects, 1.8; range, 0–3.0; p = 0.038). The victims who would succumb to SIDS in the future had fewer cortical arousals in the latter part of the night, between 3:00 and 6:00 A.M. (frequency of subcortical activation in the victims of SIDS, 11.1; range, 3.2–15.7, and in the control subjects, 14.5; range, 7.7–22.9; p = 0.011). Compared with the control infants, the victims who would succumb to SIDS in the future had significantly more frequent central apnea (median value of 9.6 apnea/hour in the infants who died of SIDS, 5 in the control infants; range, 2–16.4 apnea; p = 0.017) and more frequent obstructive apnea (median value of 0.8 apnea/hour in the infants who died of SIDS, 0.2 in the control infants; range, 0–5.1 apnea; p = 0.021) during REM sleep. In the two groups of infants, the frequency of apnea was not associated with the time of the night. The maximum duration of both types of apnea was similar in the two groups of infants (median value of 11 seconds for central apnea and 6 seconds for obstructive apnea). Cortical arousals that occurred spontaneously were seen more frequently in the control infants than in the victims of SIDS (median values of 98 arousals per hour in the control infants, 91 arousals per hour in the subjects with SIDS; range, 76–100 per hour; p = 0.021). The differences were seen in REM sleep only. The frequencies of arousals that followed either central or obstructive apneas were similar in both groups of infants in all sleep stages. There was likewise no correlation between the frequency of either central or obstructive apnea and the frequency of either subcortical activation or cortical arousal. In both groups of infants, body position during sleep was not associated with significant differences in the frequency or duration of subcortical or cortical arousals.
The victims who would succumb to SIDS in the future tended to arouse less by the end of the night than the matched control infants. In addition, the victims who would later succumb to SIDS had more frequent and longer subcortical activations and fewer cortical arousals than the matched control infants. Infants who died of SIDS appeared to be characterized by incomplete arousal processes. We must acknowledge some potential caveats in this study. First, the limited number of infants studied may prevent reaching significance in some analyses. The data depended on the limited number of polysomnographic studies prospectively collected in infants who eventually died of SIDS. The number of control subjects was limited to that of the victims of SIDS because of the lengthy process associated with the visual analysis of arousal reactions during whole-night sleep recordings. Second, the scoring of cortical arousal and subcortical activation depended on the combination of autonomic and electroencephalographic changes (19). Scoring was based on the evidence that arousal is a continuous process (8, 16) that includes subcortical structure-induced autonomic changes and cortical activation (8, 12, 16, 19). Complete arousal reactions (cortical arousals) included both autonomic and cortical activation. Incomplete arousals (subcortical activation) included autonomic but no cortical activation. As scoring was done visually, it cannot be excluded that spectral or other automatic techniques might lead to a different outcome (15, 20). Third, because of the limited number of subjects with SIDS available for analysis, this report was limited to the description of arousal characteristics. No analysis was done of infant factors that could be associated with arousal types. Last, we cannot explain why the victims who would succumb to SIDS in the future had more frequent subcortical activations and fewer cortical arousals than the control infants. Neither the level nor the mechanisms involved can be assessed within the scope of this study. As recordings of all the infants were performed in similar controlled conditions, the observed differences between the two groups could not be related to variance in experimental factors that modify arousal thresholds in infants, such as prenatal exposure to smoking (21, 22), sedatives (23, 24), previous sleep deprivation (25), prone sleeping (26, 27), sleeping with the face covered (28), or sleeping in high environmental temperatures (29). Arousal thresholds may be modified by respiratory events such as apnea (30). Although arousal may represent a normal response to obstructive apnea in adults (1), both central and obstructive apnea were reported to be poor arousal stimuli in infants (14, 31). Had obstructive apnea contributed to arouse the infants, the higher apnea index seen in the victims of SIDS, as already reported previously (6, 7, 17), would favor an increased—not a reduced—frequency of arousals. Other endogenous factors that favor arousals, such as pain (32), tracheobronchial irritation (33) were not evaluated, but would be expected to occur at random in the two groups of infants. Future studies could evaluate the role of hypoxic episodes, changes in heart rate or blood pressure (34) in the development of subcortical activation and cortical arousals. The reduced frequencies of cortical arousal in the victims who would later die from SIDS could result from structural or functional changes within the infants' arousal systems. Pathologic changes described in victims of SIDS include brainstem gliosis (35, 36), hypoplasia (37), or apoptosis (38, 39). Functional changes could involve specific synaptogenesis or synaptic activities within various arousal systems such as the noradrenergic, serotonergic, dopaminergic, cholinergic, histaminergic, or orexin-binding sites (8, 39–42). Abnormalities of serotonergic neurons were shown in the ventral medulla of victims of SIDS, in brainstem structures associated with respiratory, cardiovascular, and arousal controls (43–45). Abnormal variations of serotonin transporter gene were also reported (46). Prenatal nicotine exposure reduces arousal response to hypoxia or auditory stimulation (22, 47). This impairment of arousal responses was partially attributed to ß2-containing nicotine acethylcholine receptors (48). Such changes could favor maturational or permanent dysfunction of the infant's arousal process. Descending cortical inhibition of the arousal pathways might also interfere with the complete arousal sequence (9, 49). We cannot explain why a progressive increase in arousal frequency occurred through the night in the control infants, but not in the infants who died of SIDS. In healthy subjects, arousability increases as a function of accumulated sleep time (27, 29). In the victims who would succumb to SIDS in the future, the smaller amount of arousals seen by the end of the night could result from the infants' endogenous characteristics. It might also result from sleep deprivation in the early part of the night. Compared with the control infants, the victims who would later die from SIDS had more frequent subcortical activation in the first part of the night. This finding is reminiscent of the report of more frequent fluctuations in breathing and heart rates during the early part of the night reported in the victims of SIDS (5). It is not known whether this increase in early night subcortical activation resulted in sleep fragmentation, with a decreased tendency to arouse from sleep in the latter part of the night. Indirect evidence in adults associates autonomic changes with sleep fragmentation (11, 20). It may thus be hypothesized that subcortical activation induced sleep discontinuity and a decreased tendency to arouse from sleep by the end of the night. A relationship between the present findings and the occurrence of the sudden death of infants remains to be determined, as the sleep studies were recorded some days or weeks before the death of the infants. Compared with the control infants, the victims who would succumb to SIDS in the future exhibited fewer cortical arousals in the latter part of the night, when most SIDS events occur (2, 50). Cortical arousal permits the initiation of a behavioral response that protects from life-threatening stimuli (1) by moving away from bedding obstructing the airways when sleeping facedown (50). Sleeping prone would further increase the infant's arousal threshold (26, 27). It has, however, been suggested that full arousal is not necessary to protect an infant from a hypoxic stimulus (9). It is unclear why the death of the infants could not be prevented by the increased frequency in subcortical activation as measured in the victims of SIDS. A central loss of cerebellar control or other central loss of autoresuscitative mechanisms, as hypothesized to occur in SIDS, could have contributed to the death of the infants (51, 52). In conclusion, although the present findings do not conclusively demonstrate a relationship with the terminal fatal event, they indicate that infants who eventually died of SIDS were characterized by incomplete arousal processes during sleep.
Conflict of Interest Statement: I.K. has no declared conflict of interest; P.F. has no declared conflict of interest; J.G. has no declared conflict of interest; S.S. has no declared conflict of interest; I.K. has no declared conflict of interest; H.T. has no declared conflict of interest; A.K. has no declared conflict of interest. Received in original form January 30, 2003; accepted in final form August 13, 2003
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
