Effects of Sleeping Position and Time After Feeding on the Organization of Sleep/Wake States in Prematurely Born Infants Michael M. Myers, William P. Fifer, Liza Schaeffer, Rakesh Sahni, Kiyoko Ohira-Kist, Raymond I. Stark, and Karl F. Schulze Departments of Psychiatry and Pediatrics, Columbia College of Physicians and Surgeons, and Division of Developmental Psychobiology, New York State Psychiatric Institute, New York, NY Summary: Epidemiologic studies provide strong evidence for the conclusion that sleeping in the prone position places infants at greater risk for sudden infant death syndrome (SIDS). Prior studies in newborn infants found that in the prone sleeping position there is less time awake and more quiet sleep, but little change in the amount of active sleep. To determine whether the effects of sleeping position on state distribution vary with time after feeding, we studied prematurely born infants in both the prone and supine sleeping positions. Sleep states were recorded each minute during interfeed intervals. Results demonstrate expected effects of sleep position on state distribution: prone sleeping is associated with a 79% increase in quiet sleep and a 71% decrease in time awake. While the decreases in time awake are seen throughout the interfeed interval, increases in quiet sleep in the prone position are found only within the first hour and again near the end of the interfeed interval. These results are consistent with the hypothesis that prone sleeping could increase risk for SIDS by altering the organization of sleep, and that time after feeding may play an important role in the expression of these effects. Key words: Sudden infant death syndrome; sleep positions; sleep states; time after feeding SUDDEN INFANT DEATH SYNDROME is the third leading cause of infant mortality in the US, and accounts for nearly 30% of deaths during the postneonatal period.1 However, the physiologic mechanisms that underlie this tragic syndrome remain unknown. At this time, nearly all clinical information about SIDS is based on retrospective data derived from the general population. Although accurate prospective identification of individual babies that are likely to succumb is not yet possible, a number of factors have been identified that are reliably associated with SIDS. These associations continue to provide researchers with important clues for directing their studies. After accounting for other major risk factors, low birth-weight associated with premature delivery is the most significant and consis- tently reported risk factor for SIDS.2-4 This suggests that systematic studies of prematurely born infants might provide key information concerning the physiologic disturbances that underlie SIDS. During the 1970s, the view evolved that newborn babies and infants should, in general, be placed in a prone sleeping position. The data to support this recommendation ranged from fewer occurrences of diaper rash to reports of lower behavioral and physiological arousal, more sleep time, and less crying time in the prone position.5 However, in the late 1980s, retrospective examination of data from various countries indicated that SIDS infants were much more likely to be found in the prone sleeping position.6 This prompted medical communities to advocate a non-prone sleep position for healthy infants. By 1991, the Netherlands, Norway, Australia, New Zealand, and the United Kingdom had joined the worldwide trend of discouraging prone sleeping in babies under one year of age. In each country, these recommendations were followed by Accepted for publication February, 1998 Address correspondence and reprint requests to Michael M. Myers, PhD, Unit 40, New York State Psychiatric Institute, 722 West 168th Street, New York, NY 10032 SLEEP, Vol. 21, No. 4, 1998 343 Sleeping position and sleep states in premature infantsMyers et al diet on the rate and composition of weight gain, and on cardiorespiratory and neuroelectric activity. All infants had normal neurological examinations as well as normal cerebral ultrasound studies. The gestational ages of these infants ranged from 28 to 33 weeks, and birth weights from 1000 to 1600 g. At the time of study, their post-conceptional ages ranged from 31 to 36 weeks and postnatal ages from 2 to 6 weeks. Each infant had at least 150 minutes of sleep/wake state data after each of two consecutive feedings. Recordings of their states began within 10 minutes after these feedings. At the time of study, the 23 infants weighed between 1328 and 2311 g. They were all growing appropriately and exhibited no active medical problems. All babies were on full enteral feeds of 180 cal/day/kg. They were fed one of five formulas that contained a fixed level of protein but differed in fat-to-carbohydrate ratios. Infants were enrolled in this study after we obtained written consent from their parents. The studies were performed in the Human Infant Physiology Laboratory at Columbia Presbyterian Medical Center, NY. Infants were brought to the laboratory, which is adjacent to the nursery, at approximately 07:30 and electrodes for recording impedance respiration, electrocardiogram, electroencephalogram, and body temperature were attached. They were then placed in a radiantly warmed, clear-plastic incubator in which ambient temperature was monitored, and were maintained under thermoneutral conditions. Subjects were observed for approximately 6 consecutive hours by one of three investigators, who coded behavioral state once each minute. Studies began following an 08:00 feeding and continued until the 14:00 feeding. The studies were interrupted for 10 to 20 minutes for the 11:00 feeding, during which time no behavioral state assignments were made. Otherwise the infants were left undisturbed during the intrafeed intervals. The starting sleeping position was randomly assigned (10 supine, 13 prone). After the 11:00 feeding, position was reversed. reports demonstrating sharp decreases in the incidence of SIDS.7-11 In 1992, after review by the National Institute of Child Health and Human Development of the international evidence, the American Academy of Pediatrics (AAP) recommended that healthy term infants be put to sleep on their sides or backs.12 Within 6 months of publication of the AAP sleep position guideline, there was a 12% drop in incidence of SIDS in the US, even though the percentage of infants sleeping on their sides or supine had not reached the 75% to 90% levels reported in other countries.13 Although statistics from many countries confirmed a correlation between prone sleep positioning and increased incidence of SIDS,6,14 it remains unclear why prone sleeping has such a profound effect. One possible factor is the influence of sleeping position on sleep-state organization. Babies who succumb to SIDS do so in their sleep, and most deaths occur in the morning.15 Although no information is available regarding the actual sleep state in which SIDS deaths occur, abnormal state regulation has long been suspected to be a possible risk factor for SIDS.16-18 Accordingly, the mechanisms that control sleep and factors that influence sleep organization remain areas of key interest to SIDS researchers. In prior studies of prematurely born infants,19 full-term neonates,5,20,21 and older infants,22 it was found that in the prone sleeping position, babies spend less time awake and more time in quiet sleep. This suggests the hypothesis that the risk of prone sleeping may be mediated by changes in state distribution. However, sleep states can be affected by many other factors. Relevant to this current study, Harper and colleagues noted that in normal infants born at term and studied repeatedly from 1 week to 6 months of age, the sequence and cyclicity of active or rapid-eye-movement (REM) sleep followed by quiet or non-rapid eye movement (NREM) sleep was more predictable following feedings than following non-feeding periods of wakefulness. Based on these findings, it was suggested that feeding serves to reset the mechanisms that regulate the timing of sleep-state cycles.23 Taken together, these studies suggest that investigating the effects of prone sleeping in combination with the effects of feeding could provide new insights into the understanding of SIDS. Our working hypothesis was that sleep-state distribution would be altered by sleeping position as in prior studies, but that these effects would differ with time after feeding. State Coding Sleep-state coding began within 10 minutes of the completion of the 08:00 feeding, continued until the 11:00 feeding, resumed after the completion of this feeding, and terminated prior to the 14:00 feeding. Behavioral codes were assigned each minute according to a previously validated coding system.24 Briefly, active sleep (AS) was coded whenever at least one rapid eye movement was observed during the minute, and quiet sleep (QS) was coded when the infant was asleep without rapid eye movements. In addition, QS coding required that the infant appear rag-doll floppy and with few or no movements apart from brief generalized startles. There was also an indeterminate state (IND), in which small body movements Methods Patient Population and Testing Procedures Twenty-three preterm infants (9 males, 14 females) were subjects for this experiment. These infants were enrolled in an ongoing prospective study of the effects of SLEEP, Vol. 21, No. 4, 1998 344 Sleeping position and sleep states in premature infantsMyers et al were more common, but in which there were no rapid-eye movements. Codes were also assigned for minutes in which the infant appeared to be awake (AW) and when they were crying or fussing (CRY). The behavior of the infant was observed continuously while the study was in progress, and minute-by-minute codes were logged to the patient record. The number of minutes of each state was summed over the entire study period, within each of the two interfeed intervals, and for the five half-hour periods within each interfeed interval. In addition, the minute-by-minute codes were subjected to a second analysis after applying a smoothing algorithm. The rules for this 3-minute window approach to behavioral state coding were similar, though were not identical, to those used by Prechtl.25 They were as follows: (1) AW and CRY were collapsed into a single AW state. (2) Starting at the beginning of the record (ie, at minute 10 after feeding), each minute was recoded as unscorable until 3 consecutive minutes of QS, AS, or AW were encountered. (3) QS, AS, or AW codes were maintained as long the ongoing state did not change for more than 2 minutes. When changes in original state codes of 2 minutes or less were encountered, these were recoded to reflect the ongoing state. (4) When there were changes from one state to another (QS to AS, AS to AW, etc.) and the transition occurred within 3 minutes or less, the intervening minutes (no matter what their original code) were recoded as transitions. If changes between states took longer than 3 minutes, the intervening minutes were recoded as unscorable. (5) After applying rules 1 through 4, periods between dissimilar states longer than 3 minutes (ie, too long to be a transition), and periods between epochs of like-state longer than 2 minutes (ie, interruptions too long to be smoothed) were recoded as unscorable. Several measures of state organization were extracted from this recoding procedure. In particular, the latency after feeding to the first epoch of QS and the duration of this QS period were noted. In addition, for each baby and in each sleeping position, the original state codes were compared with the recoded states. The percent agreement between these two codes was taken as a measure of state stability. In other words, if a minute of QS in the original code remained as QS in the new code, that minute was embedded in a period of QS that lasted at least 3 minutes. If, on the other hand, that minute changed codes, it must have occurred either in isolation or in a period of QS that lasted less than 3 minutes and was not part of a stable state. Some babies (4 while in the supine position and 13 while in the prone position) had more than one epoch of QS within the interfeed interval. For these infants, estimate of the sleep cycle duration was taken as the number of minutes SLEEP, Vol. 21, No. 4, 1998 from the beginning of one QS epoch to the beginning of the next. Data Analysis For each baby, a total of 300 minutes of data were analyzed, 150 minutes after the completion of feeding in each sleeping position. Each interfeed interval was divided into five half-hour periods. The first minute in each data set was minute 10 after feeding, and the last was minute 160 after the feeding. Thus, what is labeled as the first halfhour period after feeding is that from minute 10 to minute 40. Statistical analyses were performed using the Systat statistical-analysis package. All analyses included initial sleep position as a categorical variable. To determine effects of sleeping position on the total incidence of each sleep/wake state, repeated-measures ANOVAs were performed. Analyses with time after feeding as an independent variable were conducted in two steps. First, for each behavioral state, a repeated-measures ANOVA was run. In these analyses there were two factors over which repeated measures of state were recordedsleeping position and half-hour interval after feeding. When there was evidence of a possible interaction between time and sleeping position (ie, p<.10), post-hoc paired t-tests were run to elucidate the nature of the these interactions. Latency, duration, and stability data were analyzed using paired t-tests. For records in which there were no epochs of QS within the 150 minute analysis period, latency was set to 150 minutes and duration to zero. Results Effects of Sleeping Position on State In the combined data set across both sleeping positions, the distribution of minutes across sleep/wake states was that expected for a population of prematurely born infants. The vast majority of minutes were coded as AS (73%), followed by QS (13%), AW (6%), IND (5%), and CRY (3%) (see Table 1). When the data were divided with regard to sleeping position, three differences in state distribution emerged. The two non-sleep states (AW and CRY) were both significantly increased when babies were placed in the supine position. In contrast, the incidence of QS was almost doubled in the prone position. Table 2 presents results from the analysis of states after applying the 3-minute smoothing window described in Methods. There were three significant effects of sleeping position on sleep structure. First, the latency to the onset of QS after feeding was decreased from 84.2 minutes in the supine position to 55.5 minutes when the babies were prone (F(1,21)=6.64, p<.02). Second, the duration of the first QS epoch was longer when in the prone position (16.3 vs 8.3 345 Sleeping position and sleep states in premature infantsMyers et al Table 1.Number of minutes (mean±sd) coded as one of the five sleep/wake states. Total minutes coded was 300. Also shown are state distributions (mean number of minutes±sd) within each sleeping position and, the mean(±sem) difference between sleeping positions along with results (F) from repeated measures ANOVA (df=21) for each state. State Total Minutes Quiet Sleep Indeterminate Active Sleep Cry/Fuss 39±16 15±8 219±27 9±8 Minutes in Prone Supine 25±12 8±5 111±17 2±2 Table 2.Characteristics of sleep architecture by sleeping position using state codes derived from the application of a 3-minute smoothing window to the original state codes. All values are means±SEM. See text for details of how each characteristic was derived. Results (F) from repeated measures ANOVA (df=21) for each characteristic are also given. Characteristic latencya to onset of first epoch of QS durationa of first epoch of QS stabilityb of QS stabilityb of AS Prone Supine ANOVA p 55.5+8.8 84.2+9.9 4.72 <.05 16.3+1.8 82.0+5.7 97.5+0.5 8.3+1.6 58.8+8.3 96.7+0.7 12.02 6.75 0.62 <.01 <.02 ns ANOVA p 11.1±2.6 0.9±1.3 3.0±4.0 -5.2±1.7 22.35 0.49 0.38 10.50 <.001 ns ns <.01 of AW and CRY were increased when babies were in the supine position (AW: F(1,21)=13.84, p<.01; CRY: F(1,21)=10.50, p<.01). In addition, the number of awakenings (AW or CRY), regardless of duration, were counted in each sleeping position. There were more than twice as many awakenings in the supine position as in the prone position (7.74±1.30 vs 3.26±0.70; F(1,21)=13.46, p<.01). A different pattern of results was seen for the two sleep states over the five half-hour periods after feeding (see Fig. 2). For AS, there was no significant effect of sleep position; however, there was a significant effect of time after feeding. Post hoc polynomial tests indicated that this effect was due to a cubic trend in the data (p<.05). There was no significant interaction between sleep position and time after feeding on the distribution of AS. For QS, there was a significant main effect of sleeping position with a higher incidence of QS in the prone position (F(1,21)=22.35, p<.001). There was also a significant interaction between sleeping position and time after feeding (F(4,84)=3.50, p<.02). Post-hoc tests demonstrated that this interaction was attributed to the effects of sleeping position being significant (p<.05) only in the first, second and fifth half-hour intervals after feeding. Further analyses of these data indicated that the differences in QS between sleeping positions are due not only to increases in the average duration of QS epochs (see Table 2), but also to a greater percentage of babies with extensive amounts of QS. Specifically, in the prone position there were many more babies with 10 or more minutes of QS in the first, second and fifth half-hour intervals (see Fig. 3). (a) minutes (b) % of minutes that were unchanged after smoothing minutes, F(1,21)=12.02, p<.01). Finally, the stability of QS, as measured by the percentage of minutes unaffected by state smoothing, was greater in the prone position (82.0%) than while supine (58.8%, F(1,21)=6.71, p<.02). In contrast, there was no significant effect of position on the stability of AS. In the prone sleeping position, 13 infants had two epochs of QS separated by a segment of AS within the 150 minutes of the postfeed interval. In these records, the average (±sd) time from the beginning of the first QS epoch to the beginning of the next was 66±32 minutes. In the supine position, only four infants had two QS epochs within the 150 minute postfeed interval. For these four records, the sleep cycle duration was similar to that in the prone position (73±16 minutes). Discussion The working hypothesis for these studies was that in the prone sleeping position, babies would spend less time in awake states and more time in QS, but these effects of prone sleeping would vary with time after feeding. The overall effects of sleeping in the prone position were as predicted, and there were effects of time after feeding, but not for all states. In the prone position there were decreased amounts of AW and CRY, and this was apparent Effects of Time After Feeding Figure 1 shows the incidence of AW and CRY in each of the five half-hour intervals after feeding with respect to the two sleeping positions. Repeated-measures ANOVA indicate there are no significant effects of time after feeding and no significant interaction with time and position. Consistent with the results shown in Table 1, the incidence SLEEP, Vol. 21, No. 4, 1998 14±8 7±5 108±16 7±8 Difference Prone-Supine 346 Sleeping position and sleep states in premature infantsMyers et al Figure 1.These graphs display the means of the number of minutes of Awake (top) and Cry (bottom) that were coded in each of the first 5 half-hour intervals after feeding, for each of the two sleeping positions (open=supine: black=prone). ANOVAs indicate there are significant (p<.05) main effects of position on these states, both being increased in the supine position, but no significant interaction with time after feeding. Figure 2.These graphs display the means of the number of minutes of Quiet sleep (top) and Active sleep (bottom) that were coded in each of the first 5 half-hour intervals after feeding, for each of the two sleeping positions (open=supine: black=prone). There is no significant effect of sleep position on AS butt here is a significant increase in QS in the prone position over all intervals (p<.001). In addition, for QS, there is a significant interaction with time after feeding and position, with significant (p<.05) differences during the first, second, and fifth half-hour intervals. It is not clear why sleeping position leads to differences in sleep architecture. It has been proposed that in the prone position there is an increase in arousal threshold to both internal and external stimuli.5,26 This is supported by a recent study in which it was found that heart-rate responses of infants to auditory stimulation are decreased in the prone position.27 Thus, it may be that in the prone sleeping position, sleep is deeper and the threshold to move from sleep to states of wakefulness is increased. Increases in arousal threshold in the prone sleeping position may not only decrease the likelihood that infants awaken, but also might promote longer periods of QS. On the other hand, awakenings are more likely to be followed by AS than QS.23 Thus, increases in awakenings in the supine position could indirectly lead to decreases in the amount of QS. Consistent with this hypothesis, we found an increase in the number of arousals in the supine position. This might be due to a decreased arousal threshold linked to factors such as a lower body temperature in the supine position. Alternatively, increases in the number of arousals might be caused by an increase in physiological events that promote arousal, such as gastroesophageal reflux.28 However, differences in QS were seen only toward the beginning and end of the interfeed interval, yet changes in the amount of time awake due to sleeping position were seen throughout the interfeed interval. Moreover, there was no significant change in the amount of AS as a function of sleeping position. This pattern of results suggests Figure 3.This figure shows the percentages of the 23 infants studied that had 10 or more minutes of quiet sleep within each of the half-hour intervals after feeding. Solid squares are results when the infants were sleeping in the prone position, solid circles are data obtained when the infants were in the supine position. The state data were those from the 3-minute smoothing of the original minute-by-minute coding (see Methods for details). throughout the interfeed interval. These findings are in agreement with several prior studies.5,19-22 Also consistent with this prior work, we found there were increases in QS in the prone position; however, this effect was significant only at the beginning and end of the interfeed interval. The increases in QS in the prone position were attributable to a greater percentage of babies spending longer times in QS. SLEEP, Vol. 21, No. 4, 1998 347 Sleeping position and sleep states in premature infantsMyers et al that although arousals are more frequent in the supine position, the decreases in QS in the supine sleeping position are not necessarily a secondary effect of changes in arousal. Harper and colleagues have proposed that feeding serves as a stimulus for entraining sleep cycles of infants.23 Their data, obtained from infants during the first 6 months of life, show a peak probability of QS at about 45 minutes post-feeding with a second peak about 1 hour later. In our study, the mean latency to the onset of the first epoch of QS after feeding was about 55 minutes (see Table 2). In most of our recordings there was only one epoch of QS during the 150 minutes postfeed. However, in 17 of these records (4 in the supine position, and 13 in the prone position), two epochs of QS occurred. Similar to the findings of Harper et al,23 in these 17 records the average time from the beginning of the first epoch of QS to the beginning of the second was 68 minutes. From these results, we would expect to find two peaks in the incidence of QS during the interfeed interval, one within the second half-hour after feeding, and one near the end of the interfeed interval. This pattern of QS distribution was observed when infants were in the prone position, but not when supine (see Figs. 2 and 3). Thus, another possible contribution of sleeping position to changes in the distribution of QS is that the effects of feeding on resetting the sleep cycle are much less effective when infants are supine. Prone sleeping changes not only state distribution, but fundamental aspects of physiologic function as well. In the prone position, babies have lower metabolic activity,19 yet they have higher heart rates.20,21,29,30 Premature infants also have decreased heart-rate variability, increased respiratory rates, and lower respiratory rate variability in the prone position.30 Many of these same physiologic variables exhibit systematic changes with time after feeding. Abrupt increases in circulating levels of nutrients, metabolites, and humoral mediators of anabolic activity occur during the post-absorptive period. The physiologic correlates of these anabolic processes include increases in energy, body temperature, heart rate, and respiratory rate.31 There is also evidence that neuroelectric activity and states of sleep are systematically altered by diet.32 Given the pervasive changes related to the prandial cycle, it is not surprising that time after feeding interacts with body position in regulating sleep architecture. However, how these interactions influence vulnerability to SIDS remains largely conjectural. Perhaps periods of decreased metabolic activity and heart rate that occur later in the feeding cycle represent a time of increased risk. Alternatively, it may be that the high metabolic and autonomic activity following feeding create a mismatch between the physiologic states associated with sleep and nutrient processing. What do these data contribute to our understanding of the possible mechanisms that underlie the heightened risk SLEEP, Vol. 21, No. 4, 1998 of SIDS in the prone position? As noted by Gould et al,17 the sleep state in which the cascade of events that leads to death is initiated is not known. It may be that REM presents the more vulnerable state because of phasic autonomic activity, alterations in thermoregulation, or irregularity of respiratory control. Alternatively, QS may be the more vulnerable state because of higher arousal thresholds. Our results lend indirect support for the hypothesis that QS represents a greater risk, since sleeping in the prone position increases both the amount of QS and the incidence of SIDS. Prone sleeping may confer an immediate risk for SIDS through acute changes in cardiorespiratory physiology, sleep state organization, and/or effects of rebreathing. However, a history of having slept in the prone position may also be of importance. Chronic prone sleeping may place demands on key physiological systems that lead to alterations in the developmental course of these systems. These alterations may be of little consequence to the vast majority of infants. However, in some these, changes in maturation may amplify weaknesses initiated by other factorssuch as adverse conditions during fetal lifeand thereby confer an increased risk for SIDS. Vital statistics from several countries strongly suggest that changing from the prone to the supine sleeping position has had an enormous impact on the incidence of SIDS.9,14 Thus, investigation of the effects of prone sleeping may afford new clues about the mechanisms that underlie these deaths. Studying these effects in groups that differ in vulnerability to SIDS is also key. This current investigation focused on the responses of prematurely born infants to changes in sleep position. Comparing these results with those obtained from term infants who are at less risk, as well as with infants at increased risk for other reasons (eg, exposure to effects of maternal smoking), would be informative. It is also important to note that our studies were conducted during the daytime. Since most SIDS deaths occur at night, in particular during the early morning hours, it should be determined whether the effects of sleeping position are dependent upon time of day. It would also be of interest to determine whether the sidesleeping position has an intermediate effect on sleep organization, since side-sleeping also confers risk for SIDS, although to a lesser degree than that for the prone position.33 Finally, in order to better link the results of such studies to SIDS, they should be conducted at a later stage of development. The physiological responses of prematurely born infants during their stay in the hospital may not provide the most relevant set of data. Repeating this study in infants born premature and those born at term when they reach 2 to 4 months term-adjusted age would be informative, because diurnal rhythms are more mature and this is the age of greatest vulnerability. 348 Sleeping position and sleep states in premature infantsMyers et al 16. Harper RM, Leake B, Hoffman H, Walter DO, Hoppenbadewers T, Hodgman J, Storman T.. Periodicity of sleep states is altered in infants at risk for the sudden infant death syndrome. Science 1981;213:10301032. 17. Gould JB, Lee AF, Morelock S. The relationship between sleep and sudden infant death. Annals of the New York Academy of Sciences 1988;533:62-77. 18. Thoman EB, Denenberg VH, Sievel J, Zeidner LP, Becker P. State organization in neonates: developmental inconsistency indicates risk for developmental dysfunction. Neuropediatrics 1981;12:45-54. 19. Masterson J, Zucker C, Schulze K. Prone and supine positioning effects on energy expenditure and behavior of low birth weight neonates. Pediatrics 1987;80:689-692. 20. Hashimoto T, Hiura K, Endo S, Fukada K, Mori A, Tayama M, Miyao M.. Postural effects on behavioral states of newborn infantsa sleep polygrahic study. Brain and Development 1983;5:286-291. 21. Amemiya F, Vos JE, Prechtl HF. Effects of prone and supine position on heart rate, respiratory rate and motor activity in fullterm infants. Brain Development 1991;13:148-154. 22. Kahn A, Groswasser J, Sottiaux M, Rebuffat E, Franco P, Dramaix M. Prone and supine body position and sleep characteristics in infants. Pediatrics 1993;91:1112-1115. 23. Harper RM, Hoppenbrouwers T, Bannett D, Hodgeman J, Sterman MBX, McGinty DJ. Effects of feeding on state and cardiac regulation in the infant. Developmental Psychobiology 1977;10:507-517. 24. Stefanski M, Schulze K, Bateman D, Kairam R, Pedley TA, Masterson JX, James LS.. A scoring system for states of sleep and wakefulness in term and preterm infants. Pediatric Research 1984;18:58-62. 25. Prechtl H: The behavioral states of the newborn infant (a review). Brain Research 1974;76:185-212. 26. Newman NM, Trinder JA, Phillips KA, Jordan K, Cruickshank J. Arousal deficit: mechanism of the sudden infant death syndrome? Australian Paediatric Journal 1989;25:196-201. 27. Franco P, Groswasser J, Sottiaux M, Broadfield E, Kahn A. Decreased cardiac responses to auditory stimulation during prone sleep. Pediatrics 1996;97:174-178. 28. Kahn A, Rebuffat E, Sottiaux M, Dufour D, Cadranel S, Reiterer F. Arousals induced by proximal esophageal reflux in infants. Sleep 1991;14:39-42. 29. Fox RE, Viscardi RM, Taciak VL, Niknafs H, Cinoman MI. Effects of position on pulmonary mechanics in healthy preterm infants. Journal of Perinatology 1993;13:205-211. 30. Sahni R, Schulze K, Kashyap S, Towers H, Ohira-Kist K. Postural differences in cardiorespiratory activity in quiet and active sleep in low birth weight infants. Pediatric Res 1997;41:51A. 31. Schulze KF, Stefanski M, Masterson J, Spinnazola R, Ramakrishnan R, Dell RB, Heird WC. Energy expenditure, energy balance, and composition of weight gain in lowbirth weight infants fed diets of different protein and energy content. Journal of Pediatrics 1987;110:753-759. 32. Schulze K, Kashyap S, Sahni R, Fifer W, Myers M. Diet, sleep and the developing autonomic nervous system. Proceedings of the Third European Congress of the European Society for the Study and Prevention of Infant Deaths 1995;36-42. 33. Fleming PJ, Blair PS, Bacon C, Bensley D, Smith I, Taylor E, Berry J, Golding J, Tripp J.. Environment of infants during sleep and risk of the sudden infant death syndrome: results of 1993-5 case-control study for confidential inquiry into stillbirths and deaths in infancy. British Medical Journal 1996;313:191-198. ACKNOWLEDGMENTS This work was supported by Public Health Service grants HD13063, HD32774, HD27564, and a clinical research center grant RR00645. We also wish to acknowledge the contributions of Ms. Vivian Reidy at the Bronx High School of Science. This study was undertaken in partial fulfillment of Liza Schaeffers senior research thesis, and Ms. Reidy was instrumental in promoting this effort. REFERENCES 1. Guyer B, Martin JA, MacDorman MF, Anderson RN, Strobino DM:.Annual Summary of Vital Statistics-1996. Pediatrics 1997;100:905-918. 2. Hoffman HJ, Damus K, Hillman L, Kongrad E. Risk factors for SIDS: Results of the National Institute of Child Health and Human Development SIDS Cooperative Epidemiological Study. Annals of the New York Academy of Sciences 1988;533:13-20. 3. Lipsky CL, Gibson E, Cullen JA, Rankin K, Spitzer AR. The timing of SIDS deaths in premature infants in an urban population. Clinical Pediatrics 1995;34:410-414. 4. Malloy MH, Hoffman HJ. Prematurity, sudden infant death syndrome, and age of death. Pediatrics 1995;96:464-471. 5. Brackbill Y, Douthitt TC, West H. Psychophysiological effects in the neonate of prone versus supine placement. Journal of Pediatrics 1973;82:82-84. 6. Fleming PJ, Blair PS. The role of sleeping position in the aetiology of the sudden infant death syndrome. in McIntosh N (ed): Current Topics in Neonatology. London, Saunders; 1997. 7. de Jonge GA, Burgmeijer RJ, Engelberts AC, Hoogenboezem J, Kostense PJ, Sprij AJ. Sleeping position for infants and cot death in The Netherlands 1985-91. Archives of Disease in Childhood 1993;69:660663. 8. Irgens LM, Markestad T, Baste V, Schreuder P, Skjaerven R, Oyen N. Sleeping position and sudden infant death syndrome in Norway 196791. Archives of Disease in Childhood 1995;72:478-482. 9. Dwyer T, Ponsonby AL, Blizzard L, Newman NM, Cochrane JA. The contribution of changes in the prevalence of prone sleeping position to the decline in sudden infant death syndrome in Tasmania. JAMA 1995;273:783-789. 10. Davidson-Rada J, Caldis S, Tonkin SL. New Zealand's SIDS prevention program and reduction in infant mortality. Health Education Quarterly 1995;22:162-171. 11. Wigfield R, Gilbert R, Fleming PJ. SIDS: risk reduction measures. Early Human Development 1994;38:161-164. 12. American Academy of Pediatrics Task Force on Infant Positioning and SIDS: Positioning and SIDS. Pediatrics 1992;89:1120-1126. 13. Spiers PS, Guntheroth WG. Recommendations to avoid the prone sleeping position and recent statistics for sudden infant death syndrome in the United States. Archives of Pediatrics & Adolescent Medicine 1994;148:141-146. 14. Willinger M. SIDS prevention. Pediatric Annals 1995;24:358-364. 15. Kelmanson IA. Circadian variation of the frequency of sudden infant death syndrome and of sudden death from life-threatening conditions in infants. Chronobiologia 1991;18:181-186. SLEEP, Vol. 21, No. 4, 1998 349 Sleeping position and sleep states in premature infantsMyers et al
© Copyright 2026 Paperzz