Sleep Architecture in Term and Preterm Infants Beyond the Neonatal Period: The
Influence of Gestational Age, Steroids, and Ventilatory Support
Toke Hoppenbrouwers, PhD1; Joan E. Hodgman, MD1; Denis Rybine MS4; Galina Fabrikant, BA2; Michael Corwin, MD3; David Crowell, PhD5; Debra E. Weese-Mayer,
MD6; CHIME Study Group
Department of Pediatrics, Division of Neonatology and Neonatal Medicine, University of Southern California, Keck School of Medicine, Los Angeles,
CA; 2Formerly at LAC+USC Medical Center, Division of Neonatology, Los Angeles, now at Providian Financial Corporation; 3Department of Pediatrics and Epidemiology and Biostatistics, Boston University Schools of Medicine and Public Health, Boston, MA; 4Data Coordinating Center, Boston
University School of Public Health, Boston, MA; 5Department of Pediatrics, John A. Burns School of Medicine, University of Hawaii at Manoa, Kapiolani Medical Center for Women and Children, Honolulu, HI; 6Department of Pediatrics, Rush Medical College of Rush University, Rush Children’s
Hospital, Chicago, IL
1
37 weeks postmenstrual age) exhibited immature architecture compared
with infants of term postmenstrual age. The latter exhibited similar sleep
except that they had a lower percentage of quiet sleep and longer mean
indeterminate and longest indeterminate episodes. Preterm infants with
the youngest gestational age lagged behind older preterm infants. Neither
sex nor use of steroids affected sleep. Assisted ventilation was associated
with a delay in maturation, small-for-gestational age status with increased
active sleep, and smoking with increased awake time.
Conclusion: With few exceptions, asymptomatic premature infants do
not exhibit significant delays in sleep architecture compared with term
infants at comparable postmenstrual age. The preterm infant with an early
gestational age and morbidity exhibited delayed sleep architecture.
Keywords: Premature and term-born infant, sleep-state development,
steroids, gestational age, postmenstrual age, ventilatory support
Citation: Hoppenbrouwers T; Hodgman JE; Rybine D et al. Sleep architecture in term and preterm infants beyond the neonatal period: the
influence of gestational age, steroids, and ventilatory support. SLEEP
2005;28(11): 1428-1436.
Study Objective: To examine (1) sleep architecture of infants at varied
risk for sudden infant death syndrome, (2) delays or advances in preterm
infants at term postmenstrual age, (3) whether ventilatory support and
gestational age alter sleep, (4) whether steroids alter sleep, (5) confounding influences of sex, small for gestational age, and maternal smoking.
Design: Overnight polysomnography. Dependent variables: Percentage
of active sleep, quiet sleep, indeterminate, and awake time per total recording time; mean and longest duration of state epochs; number of episodes ≥ 10 minutes; and sleep efficiency.
Setting: Collaborative Home Infant Monitoring Evaluation (CHIME).
Participants: Two hundred one preterm and 198 term infants between 33
and 58 weeks postmenstrual age during polysomnography. Fifty-one term
infants with an apparent life-threatening event without known etiology (apnea of infancy), 59 subsequent siblings of babies who died of sudden
infant death syndrome, and 88 healthy term infants.
Results: Tracings of infants with apnea of infancy and healthy term infants were similar. Subsequent siblings of babies who died of sudden
infant death syndrome spent less time in quiet sleep. Preterm infants (≤
result in advances, delays in development, or both. In that case,
it would be more appropriate to establish what is normal physiologic development for the preterm infant rather than to focus on
the comparison with term-born infants.
Maturation of sleep architecture follows a well-described
course throughout infancy, with a gradual decrease in rapid eye
movement (REM) or active sleep and a simultaneous increase in
quiet sleep.1,2 Scher and his investigative team3 reported on a sizable sample of asymptomatic preterm infants in terms of electroencephalogram (EEG)-sleep analysis. They demonstrated how,
compared with term-born infants, the preterm cohort showed
advances and delays in a host of variables that were present into
early infancy, such as alpha- and beta-range spectral values of
the EEG. Morbidity issues have typically not been evaluated. We
previously reported that preterm infants with well-documented
prolonged apnea and bradycardia in the neonatal intensive care
unit (NICU), just prior to discharge, exhibit more sustained quiet
sleep without interruptions by awakenings at 44 weeks PMA than
their age-matched controls, suggesting an acceleration in their development or, alternatively, a blunting of self-protective mechanisms.4 Subsequent to that study, the administration of maternal
steroids to enhance lung maturation has become routine.5 A number of reports have shown that antenatal steroids accelerate gastrointestinal and renal systems, although findings about the brain
vary.6-9 Maturation of the auditory brainstem evoked potential is
not affected,9 but the evoked somatosensory response is, at least
INTRODUCTION
THE DEVELOPMENT OF THE PRETERM INFANT IS FREQUENTLY ASSESSED USING THE TERM INFANT AS A
REFERENCE. FOR EXAMPLE, WHEN PRETERM INFANTS
are 38 weeks postmenstrual age (PMA), they are compared with
term infants of similar age. One problem with this comparison is
that the stress of preterm birth with its accompanying morbidity
is confounded with the effect of prolonged extrauterine existence.
An alternative approach would be to assume that the stress of
preterm birth, treatments with supplemental oxygen, mechanical
ventilation, or corticosteroids, together with prolonged extrauterine existence confer an advantage or a disadvantage that could
Disclosure Statement
This was not an industry supported study. Dr. Corwin has received research
support from CareStat, Inc. Drs. Hoppenbrouwers, Hodgman, Fabrikant,
Crowell, Weese Mayer, and Rybin have indicated no financial conflicts of
interest.
Submitted for publication December 2004
Accepted for publication July 2005
Address correspondence to: Toke Hoppenbrouwers, PhD, Clinical Professor of Pediatrics, Keck School of Medicine, University of Southern California,
Women’s and Children’s Hospital, 1240 Mission Rd, Los Angeles, CA 90033;
Tel: (323) 226-3406; Fax: (323) 226-3440; E-mail: [email protected]
SLEEP, Vol. 28, No. 11, 2005
1428
Sleep in Preterm Infants with Ventilatory Support—Hoppenbrouwers et al
on postnatal day 1, with normalization from 1 week onward.10 In
young rats, eyes open a day sooner than in the unexposed group,
but locomotion is adversely affected.11 Whether antenatal or postnatal steroids affect sleep-state development in infants has not
been established. Shinwell et al have expressed a concern about
the long-term sequelae of steroid treatment.12
As part of the protocol of the Collaborative Home Infant Monitoring Evaluation (CHIME), a national study with 5 clinical sites,
overnight infant polysomnograms (PSG) have been collected.
Quality criteria have been fulfilled in 399 PSGs. The first objective of the present study is to examine the similarity in sleep architecture in infants at perceived increased risk for sudden infant
death syndrome (SIDS) compared with healthy term infants. The
second objective is to confirm previous findings of sleep and waking in preterm infants, in particular to uncover delays or advances
in sleep architecture of preterm compared with term infants. The
third objective is to determine the effects of ventilatory support,
an estimate of morbidity, and gestational age on the rate of sleep
maturation. The fourth objective is to examine whether steroid
administration affects sleep maturation. In addition, the potential
confounding influences of sex, being small for gestational age
(SGA), and maternal smoking during pregnancy were assessed.
PMA). All of the sleep variables outlined below were examined
as a function of PMA interval at the time of PSG. Subsequently,
the cross-sectional PSGs of 112 term infants were compared with
those of 119 preterm infants at the following PMA intervals (3843, 44-49, 50-55 weeks). Maternal and infant characteristics of
the preterm cohort are listed in Table 1 as well.
The Effect of Ventilatory Support, Steroids and Gestational Age on
Sleep Architecture
A subset of 42 preterm infants who had experienced mechanical
ventilation and/or received supplemental oxygen was matched
with an equal size subset that had experienced none or a minimal
number of days of such treatments. The number of treatment days
in the exposure group ranged between 5 and 74. The mean number
of days with mechanical ventilation was 20.5, and the mean number
of days on supplemental oxygen was 28.0. Each infant in the
exposure group was matched with a control of similar gestational
age (± 2 weeks) and similar PMA at the time of the PSG (± 2
weeks). The first infant on the list after the index case, who fit these
criteria, was selected. The mean for the matched, low mechanical
ventilation subset was 0.3 days, and that for supplemental oxygen
was 2.2 days duration. In the matched control group, 32 infants
had no mechanical ventilation, 7 infants had 24 hours or less, 2
had 48 hours, and 1 had 72 hours. Of these same infants, 20 had
24 hours or less of supplemental oxygen, 15 had 60 or more hours,
and 2 had 8 and 12 days, respectively. The latter 2 were paired
with index infants who had received oxygen for 41 and 72 days.
The mean gestational ages of the exposure and the control groups
were, respectively, 29.9 and 30.5 weeks. It was impossible to find
an untreated match for the additional 36 youngest infants with a
PSG whose mean gestational age at birth was 25.9 weeks. These
were not included in the above matched groups.
Steroid administration to the mother occurred independently
from that given to the infant, although a partial overlap was present.
Infants whose mothers had received steroids before delivery (n = 32)
were matched with infants of similar gestational age and PMA with
no steroid exposure either prenatally or in the nursery. Similarly,
28 infants who received steroids in the NICU were matched by
gestational age and PMA at the time of the PSG with those who had
no steroids in the NICU, regardless of maternal steroid exposure.
Seventeen additional infants with a PSG whose gestational age was
≤ 26 weeks could not be matched.
For an analysis that targeted the effect of gestational age on
sleep architecture, 2 subsets were identified consisting of infants
with a gestational age ≤ 27 weeks who underwent a PSG between,
respectively, 36 and 38 weeks (n = 12) and 39 and 41 weeks (n =
13). These subsets were compared with 2 gestational age groups
of 28-31 weeks who underwent PSGs during the same 2 PMA
intervals as the first subsets (n = 28 and n = 24, respectively).
METHOD
Participants
The Institutional Review Board at each of the 5 clinical sites
approved the study. An informed consent was obtained from the
mothers of 201 preterm infants, defined as a gestational age ≤
37 weeks, and 198 term infants for participation in CHIME and
the collection of 1 overnight, in-hospital PSG. Of the preterm
infants, 193 underwent their PSG at or after a PMA of 34 but
before 54 weeks. An additional 8 preterm infants had their PSG
either earlier or later. Preterm infants of 34 weeks or less gestation
at birth were excluded from the CHIME study if they had any
of the following: pneumonia confirmed by chest radiograph;
home treatment with (1) continuous oxygen, (2) bronchodilators,
(3) diuretics, (4) steroids, (5) medications for gastroesophageal
reflux or seizure; congenital heart disease except asymptomatic
patent ductus arteriosus, atrial septal defect, or small muscular
ventricular septal defect; ventricular peritoneal shunt; congenital
brain anomaly that would result in non-SIDS diagnosis in the event
of sudden death; chromosomal abnormality; midfacial hypoplasia
or cleft palate; inborn error of metabolism; caregiver currently
using illicit drugs; or parental inability to communicate (language
barrier or no telephone). Among both the term and the preterm
groups were subsequent siblings of SIDS (SSS) and infants with
an apparent life-threatening event with unknown etiology, thus
designated apnea of infancy (AOI).
SIDS Risk Group Comparisons
The Effect of SGA, Sex and Maternal Smoking During Pregnancy on
Sleep Architecture
The cross-sectional PSGs of 88 healthy term (HT) infants, 51
term infants with AOIs, and 59 term SSS were compared. Maternal and infant characteristics are presented in Table 1.
Thirty SGA babies were identified and each matched with a
control of similar gestational age and PMA at the time of the PSG,
following the same strategy as described above. The full range of
dependent variables was examined in this cohort as in the study
of the effect of sex upon sleep architecture (Table 1). Forty-four
mothers smoked 1 or more cigarettes a day during pregnancy.
Maturational Comparisons
The cross-sectional PSGs of 193 preterm infants were divided
into 5 age intervals (35-38, 39-42, 43-46, 47-50, and 51-54 weeks
SLEEP, Vol. 28, No. 11, 2005
1429
Sleep in Preterm Infants with Ventilatory Support—Hoppenbrouwers et al
Physiological Recording Method
Table 1—Infant and Maternal Characteristics
Preterm
n=195
30.5 ± 3.3*
Term
HT
AOI
n=88
n=51
39.4 ± 1.0 39.4 ± 1.0
Sleep variables included 2 EEG derivations (C4-A1 and C3A2 according to the International 10-20 system), a chin electromyogram, 2 electrooculograms, and a movement sensor attached
to the diaper. Inductance plethysmography was used to measure
thoracic and abdominal breathing excursions. In addition, a CO2
monitor sampled expired CO2 through a miniature cannula taped
under the infant’s nostril. This cannula also contained an oronasal
thermistor. The electrocardiogram was recorded with 2 disposable electrodes placed under the clavicle and equidistance from
the midline. Oxygen saturation values (Nellcor, Pleasanton, Calif)
were obtained with a pulse oximeter probe taped to the infant’s
foot.13 The data were recorded on an ALICE 3 System (Healthdyne, Marietta, GA) and stored on a Sony optical disk.
SSS
n=59
39.8 ± 0.9
Gestational age
at birth, wk
Birth weight, gm 1478 ± 691 3318 ± 301 3316 ± 421 3548 ± 382
Sex, %
Male
52.6
50.0
52.9
54.2
Female
47.4
50.0
47.1
45.8
Maternal age, y
28 ± 6
31 ± 6 a
26 ± 7
30 ± 6
Parity, no. live
2.3 ± 1
2.0 ± 1.2 2.4 ± 1.4
3.3 ± 1.2
births
Married, %
61.5
80.7
48.0
79.7
Race, %
Asian
12.0
11.4
9.8
1.7
Black
13.7
9.1
7.8
10.2
Caucasian
32.6
56.8
41.2
72.9
Hispanic
30.3
4.6
15.7
11.9
Other
11.4
18.2
25.5
3.4
Education, %
< High school
61.5
26.1
70.6
48.3
graduate
Some college
23.0
21.6
13.7
27.6
College graduate
15.5
52.3
15.7
24.1
Hours of
8.1 ± 0.4
8.0 ± 0.5 8.0 ± 0.4
8.1 ± 0.4
polysomnography
Maternal
18.3
10.2
39.2.
17.0
smoking, %
SGA, %
10.3
4.5
9.8
1.7
Data Analysis
After local review, optical disks were shipped to the central
data analysis center in Boston, where trained personnel analyzed
all PSG tracings. These technicians visually scanned the recordings. On a 30-second basis, the EEG (mixed and low-voltage irregular activity, high-voltage slow activity, and trace alternant14),
electrooculogram, the movement, and the breathing channels
were each examined separately and assigned a code. These codes
were subsequently entered into computer files. Reliability studies have been published elsewhere.15 Smoothing of the data was
accomplished in 2 steps, first, at the level of the individual variable, where if an eye movement was present in one 30-second
epoch, it was considered present in the adjacent epochs as well.
This decision is predicated on the knowledge that eye movements
are phasic, and demanding that they be present in each 30-second epoch of active sleep would unrealistically fragment sleep
states. A CHIME sleep-state algorithm was developed based on
EEG, respiratory variability, eye movements, and body movements conforming to either active sleep or quiet sleep.14,16 The
data-collection personnel derived wakefulness epochs from the
movement sensor and annotations made during the actual recordings. The second step of state smoothing was then accomplished
with a 5-epoch moving window based on the state representing
the majority within the window. If there was no majority, then
the five 30-second epochs were assigned indeterminate status. A
previously published report by our group evaluated the typical
smoothing strategies encountered in the infant sleep literature, including the one used here.16
Sleep states as a percentage of total recording time, longest
state episode, number of 10-minute or longer episodes, and mean
duration of each state were calculated as a function of (1) SIDS
risk status (2) gestational age status of preterm infants (3) SGA
status and (4) preterm or term infant group. The percentages of active sleep and quiet sleep of the total recording time were derived
for each matched group. Sleep efficiency (percentage of active
sleep plus percentage of quiet sleep) was selected as an additional
dependent variable in SIDS risk groups, as a function of term and
preterm status, and to examine the influence of maternal smoking, SGA, and sex. Statistical comparisons were performed with
the help of a general linear model type III sum of squares, assuming an unequal number of entries in each cell, a 1-way analysis of
variance and paired Student t tests. In some cases, when matched
*Data are presented as mean ± SD.
AOI refers to apnea of infancy; SSS, subsequent sibling of sudden
infant death syndrome. HT refers to healthy terms.
Monitoring Procedure
Personnel at the sleep or respiratory physiology laboratories in
Los Angeles, Chicago, Cleveland, Honolulu, and Toledo underwent training in a standardized protocol for obtaining PSG data.
Extensive control and feedback procedures were implemented to
ensure comparability of the data. Each infant was admitted to the
laboratory in time for a 9:00 PM onset of calibration and recording. Often the infants were fed during preparation of monitoring
and application of electrodes. Throughout the remainder of the
night, a demand-feeding schedule was followed with opportunity
for breastfeeding. In some laboratories, mothers or caregivers
were provided with a recliner and remained in the recording room
with the instruction to be as quiet as possible; in others, a separate
room was provided for the mother or caregiver where she could
rest. Newborns were typically wrapped in a blanket, while cloth
bands around the arms of older infants were secured with safety
pins.
Monitoring was carried out in a darkened room with the recording equipment in an adjacent room. The infants were placed
in a supine or side-lying position, and, in most sites, they were
observed with the help of a low-illumination camera and monitor.
Activities such as closing and opening of eyes, startles, crying,
movements, and staff interventions were charted. PSG personnel
made annotations on the record to establish episodes of behavioral wakefulness.
SLEEP, Vol. 28, No. 11, 2005
1430
Sleep in Preterm Infants with Ventilatory Support—Hoppenbrouwers et al
Sleep States as Percentage of Total
Recording Time
Table 2—Comparison of Term Infants
AOI
34.43 (10.96) 39.17 (13.28)
33.11 (11.16) 28.33 (15.19)
5.20 (5.58)
6.38 (7.01)
27.26 (10.73) 26.11 (11.73)
4.88 (2.20)
5.31 (2.60)
4.69 (2.11)
3.96 (3.19)
0.10 (0.66)
0.22 (0.64)
2.95 (1.42)
3.10 (1.64)
7.11 (2.61)
8.17 (3.20)
5.98 (1.51)
5.74 (2.85)
1.23 (0.54)
1.41 (0.71)
15.68 (8.73)
18.14 (14.16)
P value
Preterm Infant
.04
50
QS
40
30
AS
20
10
.02
IN
0
35 - 38
39 - 42
43 - 46
47 - 50
51 - 54
Postmenstrual Age (weeks)
Longest Episode of State
40
27.03 (9.92)
30.61 (10.83)
24.55 (7.19)
22.12 (12.62)
3.47 (2.67)
4.22 (3.71)
.02
35
QS
30
56.25 (27.35) 57.17 (36.10)
25
20
AS
15
10
Preterm Infant
5
67.54 (10.86) 67.50 (13.28)
Term infant
0
35 - 38
39 - 42
43 - 46
47 - 50
51 - 54
Postmenstrual Age (weeks)
*Data are presented as mean (SD).
SSS refers to subsequent sibling of sudden infant death syndrome;
AOI, apnea of infancy; #, mean number of episodes ≥ 10 minutes;
Mn Dur, mean state duration; Long, mean longest state episode. Significant P values refer to the difference between SSS and healthy term
only.
Figure 1—Maturational changes in sleep and waking architecture in both preterm and term infants.
(P < .0001). No maturational effect was observed in either the
percentage of indeterminate or awake time. Percentages of quiet
sleep continued to rise (47%) in preterm infants up to 54 weeks
PMA, accompanied by a decline in active sleep to 21% (Figure
top and Table 3).
The number of episodes of quiet sleep of at least 10 minutes
in duration increased (P < .0001); they doubled between 34 to 37
and 42 to 45 from 2 to 4 and continued to rise significantly to 6 after 46 to 49 weeks PMA. A decrease in the number of active sleep
episodes of at least 10 minutes in duration was observed during
this age span, from 5 between 34 and 37 weeks to 2 between 50
and 53 weeks PMA. Only the mean duration of all quiet sleep
episodes showed a significant trend (P < .0001) with a rise from 6
minutes in the youngest age group to 11 minutes in the oldest.
The longest mean episode of quiet sleep was 17 minutes in
preterm infants between 34 and 37 weeks PMA, compared with
26 minutes in the 42- to 45-week group (P < .0001). This value
continued to rise until 54 weeks PMA. The decrease in the duration of the longest episode of active sleep from 31 minutes in
the youngest preterm infants to 17 minutes in the oldest preterm
infants (50-53 weeks PMA, P < .0002) was slower. There was,
however, an abrupt drop of 11 minutes after the equivalent term
age (Figure bottom and Table 3).
pair results were not significant, a power analysis was performed
to obtain least significant numbers and values to ascertain the effect of the sample size.
RESULTS
Comparison of Healthy Term With Term SIDS Risk Groups
There were no significant differences in any of the sleep
variables of healthy term and AOIs (Table 2). After adjusting
for age, 3 significant differences characterized the sleep of SSS,
compared with healthy term infants: a smaller percentage of
quiet sleep (34.43% ± 10.96% vs 43.07% ± 13.29%), the mean
longest quiet sleep episode was reduced compared with healthy
term (27.03 minutes ± 9.92 vs 34.07 minutes ± 13.17), and the
mean number of sustained periods of wakefulness of 10 minutes
or longer was slightly higher in SSS than in healthy term (Table
2).
Maturation of Sleep States in Preterm Infants
The preterm infants of 34 to 37 weeks PMA exhibited immature sleep architecture compared with preterm infants who had
reached a PMA of 42 to 45 weeks. Quiet sleep percentages of total
recording time increased from 18% to 30% (P < .0001) accompanied by a decrease in active sleep percentages from 41% to 26%
SLEEP, Vol. 28, No. 11, 2005
Term infant
60
Percent
% quiet sleep 43.67 (13.29*
% active sleep 26.08 (11.46)
% indeterminate 4.74 (5.18)
% awake time 25.51 (12.42)
# quiet sleep
6.18 (2.28)
# active sleep 3.34 (2.39)
# indeterminate 0.07(0.45)
# awake time
2.74 (1.63)
Mn Dur
9.19 (3.44)
quiet sleep
Mn Dur
5.36 (2.17)
active sleep
Mn Dur
1.23 (0.53)
indeterminate
Mn Dur
20.87 (16.38)
awake time
Long
34.07 (13.17)
quiet sleep
Long
20.73 (9.45)
active sleep
Long
3.23 (2.73)
indeterminate
Long
60.12 (36.10)
awake time
Sleep
69.75 (13.28)
Efficiency
SSS
Minutes
Healthy term
Comparison of Preterm and Term Infants
With 3 exceptions, none of the comparisons between preterm
1431
Sleep in Preterm Infants with Ventilatory Support—Hoppenbrouwers et al
infants ≥ 38 weeks PMA and term infants (healthy term and AOI)
at comparable PMA yielded significant differences (Table 4). The
exceptions were the number of quiet 10-minute or longer sleep
episodes (P < .016); preterm infants had an average of 5.9 (±2.4) of
these quiet sleep episodes per overnight recording, compared with
3.2 (± 2.2) in term infants. The mean duration of indeterminate
epochs (1.4 minutes ± 0.5 vs 1.3 minutes ± 0.6) as well as the
longest indeterminate epoch (4.1 minutes ± 2.9 vs 3.6 minutes ±
3.2) was elevated in preterm infants as well.
None of the independent variables studied had any influence on
sleep efficiency. However, sleep efficiency increased significantly
with increasing PMA intervals (P < .03). No significant interaction
between preterm/term groups and PMA interval was observed,
indicating that this enhancement of sleep efficiency applied to
both term and preterm infants.
The Effects of Ventilatory Support, Steroid Administration and Gestational Age on Sleep Architecture
A review of the infant medical histories revealed that the group
who had undergone appreciable ventilatory support exhibited
more morbidity. Thus, besides the need for ventilatory support,
48% had asymptomatic patent ductus arteriosus versus 10%
in the group with minimal support (infants with symptomatic
patent ductus arteriosus were excluded from the CHIME study.
Nineteen percent had intraventricular hemorrhage grade I
and II (in addition, 1 infant had grade III). In the minimalventilatory-support group, the number was 12%. The numbers
for periventricular leukomalacia were 7% and 2%, respectively;
necrotizing enterocolitis (10%) was only found in the ventilatory-
Table 3—Preterm State as a Function of Postmenstrual Age Interval
at the Time of Polysomnography
34-37*
n=57
% quiet sleep 18 (9)*
% active sleep 41 (14)
% indeterminate 8 (5)
% awake time 34 (13)
# quiet sleep
2 (2)
# active sleep
5 (3)
# indeterminate 0.1(0.4)
#awake time
4 (1)
Mn Dur quiet
6 (2)
sleep
Mn Dur active 4 (2)
sleep
Mn Dur
1 (0.5)
indeterminate
Mn Dur awake 13 (7)
time
Long quiet
17 (8)
sleep
Long active
31 (16)
sleep
Long
5 (3)
indeterminate
Long awake
61 (62)
time
38-41 42-45
n=74 n=34
23 (9) 30 (9)
39 (13) 26 (9)
6 (5) 5 (4)
32 (11) 38 (14)
3 (2) 4 (2)
5 (3) 3 (2)
0.1(0.4) 0
4 (1) 4 (2)
5 (2) 7 (2)
7 (3)
5 (2)
46-49
n=20
36 (8)
26 (11)
6 (5)
32 (11)
6 (2)
4 (3)
0.1(0.2)
4 (1)
8 (3)
50-53 P value
n=8
47 (15) < .0001
21 (7) < .0001
6 (6)
.09
27 (15) .07
(6 (1) < .0001
2 (1) < .0004
0.1(0.4) .7
3 (2)
.06
11(3) < .0001
6 (2)
5 (1) < .02†
1 (0.6) 1 (0.3) 1 (0.5)
1 (0.7)
15 (8) 17 (8)
16 (7)
19 (13) < .04†
20 (8) 26 (7)
31 (9)
37 (8) < .0001
31 (17) 20 (9)
22 (6)
17 (6) < .0002
3 (2)
4 (4) < .02†
4 (3)
3 (1)
63 (42) 65 (28) 61 (23) 47 (25)
Table 5—Percentage of Active Sleep and Quiet Sleep as a Function
of Ventilatory Support, Steroids Given in the Neonatal Intensive Care
Unit, Gestational Age and Maternal Smoking During Pregnancy in
Ventilated Infants
Quiet sleep, %
Ventilatory support
vs
Minimal ventilatory
support
Steroids in the NICU
vs
No steroids in the
NICU
Gestational age
≤ 27 weeks
PSG @ 36-38 weeks
vs
28 ≤ gestational age
≤ 31 weeks
PSG @ 36-38 weeks
Maternal smoking
in ventilated infants
vs
No maternal smoking
in ventilated infants
.2
.09
Data are presented as mean (SD).
#, mean number of episodes ≥ 10 minutes; Mn Dur, mean state duration; Long, mean longest state episode.
*37 weeks inclusive but not yet 38 weeks.
†A Scheffe test did not reveal significant differences between specific
age intervals.
Active sleep, %
42.0 (13.0)*
34.4 (12.5)
22.7 (13.3)
40.3 (12.7)
27.6 (10.5)
36.2 (13.8)
P value
< .001
NS
14.6 (7.3)
< .001
23.1 (7.8)
35.2 (12.8)
50.6 (7.96)
< .01
Data are presented as mean and (SD). NICU refers to neonatal intensive care unit; PSG, polysomnogram.
Table 4—Comparisons of Term and Preterm Infants
Preterm
n=86
% quiet sleep 24.68 (8.78)^
% active sleep 39.12 (11.66)
% indeterminate 6.02 (4.79)
% awake time 30.17 (8.75)
Sleep
efficiency
63.81 (10.72)
38-43
Term
n=15
27.70 (7.45)
34.51 (8.40)
5.40 (3.97)
32.38 (8.14)
Preterm
n=23
34.27 (7.97)
28.69 (9.98)
5.18 (5.09)
31.86 (8.14)
62.21 (8.39)
62.96 (9.33)
44-49
Term
n=60
37.21 (10.29)
30.17 (11.49)
4.07 (4.53)
28.55 (10.04)
Preterm
n=10
43.81 (12.77)
23.95 (7.86)
5.94 (6.12)
26.30 (11.53)
67.38 (10.89)
67.76 (14.93)
50-55
Term
n=37
46.96 (11.48)
24.88 (13.50)
5.49 (5.60)
22.66 (13.74)
71.84 (13.65)
P value
.0001*
.0001*
.058
.03*
Data are presented as mean and (SD);
*Only age effect is significant; no difference between term and preterm group in this table.
SLEEP, Vol. 28, No. 11, 2005
1432
Sleep in Preterm Infants with Ventilatory Support—Hoppenbrouwers et al
roids on sleep architecture, whether given to the mother or the
newborn. Early gestational age, ventilatory support, and neonatal steroid administration were strongly correlated. With current
standards of treatment, these influences cannot be teased apart.
Significant ventilatory support was associated with a delay in the
maturation of sleep architecture. Interestingly, maternal smoking
in ventilated babies undid the delay mentioned above. Maternal
smoking prolonged waking variables in the overall cohort of term
and preterm infants.
The potential confounding effect of postneonatal side smoke or
feeding on sleep architecture is worth studying, but it requires a
very detailed follow up of the infants. The CHIME database probably contains adequate information that requires careful scrutiny
and several analyses, better performed in a separate paper. The
use of caffeine, on the other hand, its dose, timing, and adherence,
were insufficiently documented to be of value for an examination
as a potential confounder.
This study provides preliminary answers to the original question raised in the Introduction as to whether it would be most
appropriate to establish what is normal physiologic development
for the preterm infant, rather than to focus on the comparison with
term-born infants. Three developmental courses can be observed:
that of the term-born infant, that of the asymptomatic preterm
infant, and that of the preterm infant with significant ventilatory
support. The brain of the asymptomatic preterm infant seems to
follow its own developmental schedule, at least in terms of sleep
development; in-utero or ex-utero existence seems to make little
difference. The baby matures normally, and, using the term baby,
as the reference seems appropriate because sleep architecture is
virtually identical at the equivalent PMA of 38 to 40 weeks and
remains so. With larger magnification, however, even these infants who have seemingly caught up exhibit more subtle delays
and advances that can be demonstrated up to a PMA of 43 to 53
weeks.3,19 In our study, very young preterm infants and those who
received significant ventilatory support indicative of morbidity
exhibited a measurable delay in the maturation of sleep architecture. For these infants, the behavior of their fellow asymptomatic
preterm infants may be the most appropriate reference.
Past findings of alterations in sleep architecture in infants
at increased risk for SIDS have been equivocal. Navelet et al20
combined what were then called near-miss infants with subsequent siblings and reported a decrease in awakening and a higher
amount of active sleep, but these findings seemed to apply mostly
to near-miss infants. Challamel et al21 found fewer awakenings in
the latter, but, once awake, these infants remained so longer. We
published a systematic study of 25 SSS whose PSGs were virtually the same as those of a carefully matched control group. In
that study, none of the variables examined here were different.22
In the present study, differences centered especially on a reduction in quiet sleep, suggesting a delay in maturation. A delay in
maturation is also suggested by a relative decrease in quiet sleep
that others have observed in infants with apparent life-threatening events.23 In the present study, the sleep architecture of infants
with AOI could not be differentiated from those of healthy term
infants. The best explanation for these discrepancies is that our
cohort was not only the largest ever studied under the same protocol, but that it comprised infants with AOI in whom explanations
for the apparent life-threatening event had been ruled out. Most
previous studies were performed before this diagnosis was generally accepted.
support group.
Infants with ventilatory support in the NICU exhibited immature
sleep architecture compared with matched controls that had
experienced no or minimal ventilatory support. The percentage
of active sleep was 42% (± 13.0%) in the former compared with
34.4% (± 12.5%) in the latter (F1,82 = 7.5, P < .001). There was no
difference in percentage of quiet sleep (Table 5).
Steroids in the NICU had no influence on sleep architecture,
as measured by the PSG (Table 5). Maternal steroids similarly
had no effect. Since the least significant number for the NICU
steroid matched groups for percentage of active sleep was 93
rather than the 42 that were actually available, the maternal and
NICU exposure were combined. The means were 36.7% for
exposed infants and 37.8% for those not exposed. In that case, the
power analysis revealed that the least significant number of cases
to possibly yield a significant finding was 19,190 cases.
When they reached a PMA of 36 to 38 weeks, infants of 27
weeks of less gestational age still showed a delay in their sleep
architecture, with 14.6% of quiet sleep, compared with 23.1% in
infants born with a gestational age of 28 to 31 weeks (F1,32 = 8.1, P
< 0.001, Table 5). Two other groups of the same gestational ages
who had their PSGs at 39 to 41 weeks PMA exhibited 21.0% and
26.6% of quiet sleep, respectively (F1,33 = 4.2, P <.09), a similar
but nonsignificant trend. There were no significant differences in
active sleep percentages. The sleep architecture of infants born at
a later gestational age—between 28 and 31 weeks, who underwent
their PSG recordings also between 36 to 38 weeks PMA and 39 to
41 weeks PMA—could not be differentiated statistically either.
The Influence of SGA, Sex, and Maternal Smoking
The 30 SGA infants in this study exhibited a significant
increase in the percentage of active sleep (P < .03) compared
with age-matched appropriate-for-gestational-age infants (38.28
± 2.40 vs 31.52 ± 13.61). Sex had no effect on sleep architecture.
Smoking during pregnancy consistently influenced wakefulness.
The mean duration of wakefulness (17.6 minutes ± 13.7 vs 16.6
minutes ± 12.3), as well as the longest episode of wakefulness
(57.9 minutes ± 31.6 vs 55.4 minutes ± 28.3), was prolonged in
infants born to smoking mothers (P < .034 and .004, respectively).
A comparison of sleep-state percentages in ventilated babies
whose mothers smoked during pregnancy, versus the absence of
smoking, revealed a normalizing of the delay discussed in above
(Table 5).
DISCUSSION
The sleep architecture of SSS was significantly different from
that of healthy term and AOI, the latter 2 being comparable.
The findings suggest a delay in maturation in these SSS. This
prompted an elimination of this risk group for the comparison
between preterm infants and term infants at comparable PMA.
Preterm infants between 34 and 38 weeks PMA exhibited immature sleep-state patterns when they were compared with preterm
infants that had reached term PMA. The latter exhibited similar
sleep architecture as term-born infants, except for more episodes
of long awakenings between 50 and 54 weeks PMA, a finding that
has been reported by others, quite sleep episodes > 10 min and
more indeterminate sleep.17-18 Preterm infants with the youngest
gestational age lagged behind the older ones even if they were
measured at the same PMA. There was no apparent effect of steSLEEP, Vol. 28, No. 11, 2005
1433
Sleep in Preterm Infants with Ventilatory Support—Hoppenbrouwers et al
that has only risen.28 Another manifestation of variance is the inability to replicate findings, even within sleep laboratories.22 Preterm infants prior to nursery discharge experience, in their own
way, a highly stimulating environment, and their sleep behavior
shows variance similar to that of term infants.
The developmental trend in sleep and waking behavior, however, is robust despite these variances. Sleep resembles the neural
organization of language, a preprogrammed capability that cannot but manifest itself within a certain time window. Case studies in conjoined twins, who are always identical, reveal similarities in these maturational sleep trends, yet dissimilarities in the
temporal distribution of sleep and individual sleep states.29 Thus,
even though the genetic and environmental input in such infants
is virtually identical, each exhibits aspects of his or her own sleep
architecture; in other words, infant temperament and other idiosyncrasies can have profound effects. In the 1980s, Monod and
Guidasci30 examined the sleep-waking behavior of severely damaged neonates and revealed surprisingly normal values in some,
but by no means all, of these infants. Against the background of
such variance, is it reasonable to expect that we can unequivocally attribute outcome to these small transient adaptations, confounded with numerous social and environmental influences?
Yet, it seems unwise to jettison all attempts to find potential central nervous system predictors of infant outcome in light of the
high stakes.
Our study had ended for most infants around 6 months of age.
Maturation in sleep architecture will, of course, continue beyond
that age. During the last decade, a significant body of information
has accumulated that suggests that sleep, in particular, although
not exclusively, REM sleep, is crucially involved in memory consolidation, as well as in neural development. The delays and advances in the maturation of sleep architecture can be considered
manifestations of plasticity. The phasic activity of REM sleep in
infancy peaks at the time of the most rapid neural development.31
It is, however, not just the temporal association that lends credence to the contemporary variant of what Roffwarg et al. termed
the ontogenetic hypothesis,32 the idea that phasic activity during
REM sleep provides an important source of endogenous neural
activity that, together with exogenous input, sculpts and reinforces neural networks in the developing brain. In fact, the contemporary variant of that hypothesis posits that the endogenous
neural activity might even afford protection against excessive
experience-driven plasticity. Findings in maturation of the visual
system in kittens, with manipulation of monocular deprivation,
REM-sleep deprivation, and pharmacologic elimination of pontogeniculooccipital spikes, has suggested a likely causal role for
REM sleep in neural development.33 Synaptic mechanisms are extremely malleable and intrinsically unstable. The neural networks
that underlie early cognitive and motor behavior in status nascendi could be considered the epitome of instability. In that light, the
minor adaptations in sleep architecture that can be observed, for
instance a decrease in the percentage of time spent in active sleep
in SGA infants, may diminish endogenous stimulation at an inopportune time. If, indeed, REM sleep or active sleep controls neural activity, then a lack may have more deleterious consequences
than previously appreciated. In this context, the role of cytokines
other than the anti-inflammatory variety in synaptic plasticity is
intriguing.34 Preterm infants have been found with elevated levels
of anti-inflammatory cytokines that, upon further search, may be
correlated with cytokines involved in synaptic plasticity.35
Most studies in the literature have documented the physiologic
adaptations such as delays and advances in asymptomatic preterm
infants. By contrast, Karch and coworkers24 explicitly examined
sleep architecture in 3 preterm cohorts, 2 of which received assisted ventilation, with sample sizes of 24, 14, and 23, respectively.
All were monitored between 30 and 37 weeks PMA. Unlike our
results, they found a larger proportion of quiet sleep and longer
durations of quiet sleep in symptomatic infants when studies at
30 and 36 weeks were combined. Scher et al25 compared preterm
infants with chronic lung disease at term age with 2 groups: similarly aged preterm infants without lung disease and term-born
infants. Their finding of a reduction in percentage of active sleep
was different from either the findings of Karch and co-workers
and ours. In our study, controls were matched for gestational age
and PMA, and the latter varied between 33 and 57 weeks PMA.
Even when only matches of infants with PSG at term PMA were
examined, the increase of active sleep percentage in the infants
with ventilatory support persisted. Some of these differences may
be accounted for by the changes in NICU support in the more than
20 years since publication of the paper by Karch and coworkers.
The morbidity of surviving infants in the CHIME study during
their nursery stay was almost certainly higher. Maternal smoking
during pregnancy, rather than exacerbating the delay in maturation, seemed to compensate for it.
An effect on sleep of steroids, which are often used in clinical
practice, was not measurable in our cohort. Neither prenatal nor
postnatal steroid exposure had a significant effect. Since a power
analysis suggested that a sample twice the size of the 42 exposed
infants might result in significant findings, we collapsed maternal
and infant steroid-exposure groups. This strategy added measurably to the variance. The least significant number of subjects to
obtain a potentially significant result was extremely large. This
study does not establish unequivocally that steroids per current
clinical protocol have no or little effect on subsequent sleep behavior, but this conclusion is nonetheless likely. Latencies of the
auditory brainstem averaged evoked potential are similarly unaffected.9 Although, on the first day after birth, the somatosensory evoked response showed a slight change in infants prenatally exposed to steroids, this effect was extremely transient and
had disappeared after 1 week.10 When the PSG in our study was
performed, the postnatal age of every preterm infant, exceeded 1
week by far.
What is the significance of advances and delays from both a research and clinical point of view? The age at which these advances
and delays, collectively coined “dysmaturity” by Scher,3 have occurred coincide with the typical age window of SIDS—between
42 and 52 weeks PMA in preterm infants and 10 and 12 weeks
of postnatal age in the term infant. SIDS is far more prevalent in
preterm infants than in term-born infants, and the rate increases
with decreasing gestational age.26 It is also the age that a number
of nonmonotonic physiologic changes occur in both the central
and autonomic nervous systems, suggesting a profound reorganization of neuronal circuits and their interaction during a discrete
age period that represents a risk time for SIDS.27 Change and variability are a hallmark of infancy; in both infancy and adulthood,
they imply flexibility, or even plasticity, but, in excess, such a
reorganization may imply risk.
Variability in sleep architecture remains the norm, as has been
amply demonstrated previously. In 1982, we published a tabulation of influences on sleep that amounted to 2 dozen, a number
SLEEP, Vol. 28, No. 11, 2005
1434
Sleep in Preterm Infants with Ventilatory Support—Hoppenbrouwers et al
Department of Pediatrics, John A. Burns School of Medicine,
University of Hawaii at Manoa, Honolulu: Kapiolani Medical
Center for Women and Children, David H. Crowell*, Larry E.
Tinsley (currently at Pediatric Critical Care Medical Group/Pediatric Management Group, Children’s Hospital of Los Angeles),
Linda E. Kapuniai**
Department of Pediatrics, Divisions of Neonatology and Neonatal Medicine, USC Keck School of Medicine, Los Angeles, CA;
Los Angeles County & USC Medical Center; Women’s and Children’s Hospital, Los Angeles; Good Samaritan Medical Center,
Children’s Hospital of Los Angeles, Toke T. Hoppenbrouwers*,
Rangasamy Ramanathan, Paula Palmer**, Thomas G. Keens,
Sally L. Davidson Ward, Daisy B. Bolduc, Technical Coordinator
There is evidence that some advances and delays predict developmental outcome.3,36 During the last decades, enormous progress
has been made to ensure the survival of micro-preemies with a
gestational age of 23 to 26 weeks. It has become clear that even
though their central nervous systems are relatively undamaged
and severe encephalopathies are not the rule, between 25% and
70% of these infants wind up with neurologic and developmental handicaps, depending on their gestational age.37-39 Why this
is so is not well understood, and there is a scarcity in treatment
strategies. Might this cohort be particularly suited to examine the
validity and sensitivity of EEG/sleep variables in the prediction
of neurologic and developmental outcome? The term catch-up
seems to take on an added significance because it defines a rate
of maturation. These low-birth-weight infants could be compared
with asymptomatic preterm infants at a PMA of 30 to 32 and 34 to
36 weeks and at several later intervals to establish a timetable for
the nature and appearance of these adaptations and their respective catch-up times. It might answer such questions as whether
and how the timetable differs in the youngest infants who have
overcome medical illnesses in their early life.
In conclusion, the asymptomatic premature infant is not obviously delayed in sleep architecture compared with the term infant.
Increased waking behavior throughout the first months at home
constitutes an exception. By and large, these infants mature appropriately to become equivalent to the term infant at 40 weeks
PMA. Upon more detailed examination, however, others have uncovered subtle delays and advances.3,19,25 The clinical significance
of the findings has yet to be firmly established for both short- and
long-term outcomes. The preterm infant with an early gestational
age and morbidity truly differs when compared with the appropriate, asymptomatic preterm control. Catch-up details may provide
clues to long-term outcome. Sex and exposure to steroids do not
appear to affect sleep and waking in the asymptomatic preterm
infant, but maternal smoking and SGA status have subtle impacts
on sleep architecture.
CLINICAL TRIALS OPERATIONS CENTER:
Department of Obstetrics and Gynecology, Case Western University School of Medicine and MetroHealth Medical Center, Cleveland, OH, Michael R. Neuman* (currently at Michigan Technological University, Houghton), Rebecca S. Mendenhall**
DATA COORDINATING AND ANALYSIS CENTER
Departments of Pediatrics and Epidemiology and Biostatistics, Boston University Schools of Medicine and Public Health,
Boston, MA,Michael J. Corwin*, Theodore Colton, Sharon M.
Bak**, Mark Peucker, Technical Coordinator, Howard Golub,
Physiologic Data Biostatistician, Susan C. Schafer, Clinical Trials Coordinator
STEERING COMMITTEE CHAIRMAN
Formerly Department of Pediatrics, Yale University School of
Medicine, New Haven, CT: George Lister currently at Southwestern Medical School, Dallas, TX
NATIONAL INSTITUTES OF HEALTH
Pregnancy and Perinatology Branch, Center for Research for
Mothers and Children, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, MD,
Marian Willinger
*Principal Investigator
**Study Coordinator
ACKNOWLEDGEMENTS
Supported by NICHD Grants HD 29067, HD 29071, HD 28971,
HD 29073, HD 29056, HD 34625 and a generous donation from
the Orange County Guild for Infant Survival.
REFERENCES
The participants in the CHIME Study Group include:
1.
CLINICAL SITES:
Department of Pediatrics, Case Western Reserve University
School of Medicine, Cleveland, OH MetroHealth Medical Center, Terry M. Baird* (presently at Rainbow Babies and Children’s
Hospital), Rainbow Babies and Children’s Hospital, Richard J.
Martin, Lee J. Brooks (currently at The Children’s Hospital of
Philadelphia) Roberta O’Bell**
Department of Pediatrics, Medical College of Ohio, Mercy
Children’s Hospital, and Children’s Medical Center, Toledo, OH,
Carl E. Hunt* (currently at National Center on Sleep Disorders
Research, National Heart, Lung and Blood Institute, National Institutes of Health, Bethesda, MD) David R. Hufford, Mary Ann
Oess**
Department of Pediatrics, Division of Respiratory Medicine,
Rush Medical College of Rush University, Chicago, IL: Rush
Children’s Hospital at Rush University Medical Center: Debra E.
Weese-Mayer*, Jean M. Silvestri, Sheilah M. Smok-Pearsall**
SLEEP, Vol. 28, No. 11, 2005
2.
3.
4.
5.
6.
7.
8.
1435
Parmelee A, Wenner W, Akiyama Y, Schulz M, Stern E. Sleep states
in premature infants. Dev Med Child Neuro 1967;9:70-7.
Hoppenbrouwers T, Hodgman J, Arakawa K, Geidel S, Sterman
M. Sleep and waking states in infancy: normative studies. Sleep
1988;11:387-401.
Scher M. Neurophysiologic assessment of brain function and maturation ii. a measure of brain dysmaturity in healthy preterm neonates. Pediatr Neurol 1997;16:287-95.
Hoppenbrouwers T, Hodgman J, Arakawa K, Durand M, Cabal L.
Polygraphy after discharge in preterm infants with and without apnea in the nursery. Neuropaediatrie 1992;23:75-81.
Ward RM. Pharmacologic enhancement of fetal lung maturation.
Clin Perinatal 1994;21:523-42.
Whitelaw A and Thoresen M. Antenatal steroids and the developing
brain. Arch Dis Child Fetal Neonatal Ed 2000;83:F154-7.
Catarelli D, Chirico G, Simeoni U. Renal effects of antenatally or
postnatally administered steroids. Pediatr Med Chir 2002;24:15762.
Modi N, Lewis H, Al-Naqeeb, Ajayi-Obe M, Doré CJ, Rutherford
M. The effects of repeated antenatal glucocortocoid therapy on the
Sleep in Preterm Infants with Ventilatory Support—Hoppenbrouwers et al
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
developing brain. Pediatr Res 2001;50:581-5.
Amin SB, Orlando MS, Dalzell LE, Merle KS, Guillet R. Brainstem
maturation after antenatal steroid exposure in premature infants
as evaluated by auditory brainstem-evoked response. J Perinatol
2003;23:307-11.
de Zegher F, de Vries V, Daniels H et al. Effects of prenatal betamethasone/thyrotropin releasing hormone treatment on somatosensory
evoked potentials in preterm newborns. Pediatr Res 1992;32:2124.
Gramsbergen A, Mulder EJ. The influence of betamethesone and
dexamethasone on motor development in young rats. Pediatr Res
1998;44:105-10.
Shinwell ES, Karplus M, Reich D et al. Early postnatal dexamethasone treatment and increased incidence of cerebral palsy. Arch Dis
Child Fetal Neonatal Ed 2000;83:F177-81.
Crowell D, CHIME study Group. An atlas of infant polysomnography. Boca Raton: A CRC Press Co. The Parthenon Publishing
Group; 2003.
Anders T, Emde R, Parmelee A, eds. A Manual of Standardized Terminology, Techniques and Criteria for Scoring of States of Sleep
and Wakefulness in Newborn Infants. Los Angeles: UCLA Brain
Information Service, BRI Publications Office; 1971.
Crowell DH, Brooks LJ, Colton T. Infant polysomnography: reliability. Sleep 1997;20:553-60.
Kulp T, Corwin M, Brooks LE, et al. The effect of epoch length and
smoothing on infant sleep and waking architecture for term infants
at 42-46 weeks. Sleep 2000;23:893-9.
Yokochi K, Shiroiwa Y, Inukai K, Kito H, Ogawa J. Behavioral state
distribution throughout 24-h video recordings in preterm infants at
term with good prognosis. Early Human Dev. 1989;19:183-90.
Giganti F, Fagioli I, Ficca G, Cioni G, Salzarulo P. Preterm infants
prefer to be awake at night. Neurosci Lett 2001;312:55-7.
Booth C, Leonard H, Thoman E. Sleep states and behavior patterns
in preterm and fullterm infants. Neuropediatrics 1980;11:354-64.
Navelet Y, Payan C, Guilhaume A, Benoit O. Nocturnal sleep organization in infants “at risk” for sudden infant death syndrome.
Pediatr Res 1984;18:654-7.
Challamel M, Revol M, Leszczynski M, Debilly G. Organization
nycthemerale des etats de vigilance chez le nourrison normal et le
nourrison dit “rechappe d’un syndrome de mort subite” Rev Electroencephalogr Neurophysiol Clin 1981;11:28-36.
Hoppenbrouwers T, Hodgman J, Arakawa K, Sterman M. Polysomnographic sleep and waking states are similar in subsequent siblings
of SIDS and control infants during the first six months of life. Sleep
1989;12:265-76.
Haddad G, Walsh E, Leistner H, Grodin W, Mellins R. Abnormal
maturation of sleep states in infants with aborted sudden infant
death syndrome. Pediatr Res 1981;15:1055-7.
Karch D, Rothe R, Jurisch R, Heldt-Hildebrandt R, Lubbesmeier A,
Lemburg P. Behavioral changes and bioelectric brain maturation of
preterm and fullterm newborn infants: a polygraphy study. Develop
Med Child Neurol 1982;24:30-47.
Scher MS, Richardson gestational age, Salerno DG, Day NL, Guthrie RD. Sleep architecture and continuity measures of neonates with
chronic lung disease. Sleep 1992;15:195-201.
Malloy MH, Hoffman HJ. Prematurity, sudden infant death syndrome, age of death. Pediatrics 1995;96:464-71.
Hoppenbrouwers T, Hodgman JE. Sudden infant death syndrome
(SIDS): an integration of ontogenetic, pathologic, physiologic and
epidemiologic factors. Neuropaediatrie 1982;13:36-51.
Hoppenbrouwers T. Electronic monitoring in the newborn and
young infant: theoretical considerations. In: Sleeping and Waking
Disorders: Indications and Techniques. Guilleminault C, ed. Menlo
Park: Addison Wesley Publishing Co; 1982:17-59.
Sackett G, Korner A. Organization of sleep-waking states in conjoined twin neonates. Sleep 1993;16:414-27.
Monod N, Guidasci S. Sleep and brain malformation in the neonatal
SLEEP, Vol. 28, No. 11, 2005
period. Neuropaediatrie 1976;7:229-49.
31. Mirmiran M, Ariagno R. Role of REM sleep in brain development
and plasticity. In: Maquet P, Smith C and Stickgold R, eds. Sleep
and Brain Plasticity. Oxford: Oxford University Press; 2003:181-8.
32. Roffwarg H, Muzio J, Dement W. Ontogenetic development of the
human sleep-dream cycle Science 1966;152:604-19.
33. Frank M, Stryker M. The role of sleep in the development of central
visual pathways. In: Maquet P, Smith C and Stickgold R, eds. Sleep
and Brain Plasticity. Oxford: Oxford University Press; 2003:189206.
34. Vitkovic L, Bockaert J, Jacque C. “Inflammatory” cytokines: neuromodulators in normal brain? J Neurochem 2000;74:457-71.
35. Dembinski J, Behrendt D, Martini R, Heep A, Bartmann P. Modulation of pro- and anti-inflammatory cytokine production in very
preterm infants. Cytokine 2003;21:200-6.
36. Whitney M, Thoman E. Early sleep patterns of premature infants
are differentially related to later developmental disabilities. J. Dev
Behav. Pediatr. 1993;14:71-80.
37. Rijken M, Stoelhorst G, Martens SE, et al. Mortality and neurologic, mental, and psychomotor development at 2 years in infants
born less than 27 weeks gestation: the Leiden follow-up project on
prematurity. Pediatrics 2003;112:351-8.
38. Hack M, Friedman H, Fanaroff A. Outcomes of extremely low birth
weight infants. Pediatrics 1996;98:931-7.
39. Sweet MS, Hodgman JE, Pena I, et al. Two-year outcome of infants
weighing 600 grams or less at birth and born 1994 through 1998.
Obstet Gynecol 2003;101:18-23.
T, Hunt
1436
Sleep in Preterm Infants with Ventilatory Support—Hoppenbrouwers et al