Recognition of Sleep-disordered Breathing in Children

(1)

Recognition

of Sleep-disordered

Breathing

in Children

PEDIATRICS Vol. 98 No. 5 November 1996 871

Christian Guilleminault, MD*; Rafael Pelayo, MD*; Damien Leger, MD*; Alex Clerk, MD*; and

Robert C. Z. Bocian, MD, PhD

ABSTRACT. Objective. To determine whether upper

airway resistance syndrome (UARS) can be recognized

and distinguished from obstructive sleep apnea

syn-drome (OSAS) in prepubertal children based on clinical

evaluations, and, in a subgroup of the population, to

compare the efficacy of esophageal pressure (Pes) moni-toring to that of transcutaneous carbon dioxide pressure

(tcPco2) and expired carbon dioxide (CO2) measurements

in identifying UARS in children.

Study Design. A retrospective study was performed

on children, 12 years and younger, seen at our clinic since

1985. Children with diagnoses of sleep-disordered

breathing were drawn from our database and sorted by

age and initial symptoms. Clinical findings, based on

interviews and questionnaires, an orocraniofacial scale,

and nocturnal polygraphic recordings were tabulated

and compared. If the results of the first polygraphic

recording were inconclusive, a second night’s recording

was performed with the addition of Pes monitoring. In

addition, simultaneous measurements of tcPco2 and

end-tidal CO2 with sampling through a catheter were

per-formed on this second night in 76 children. These 76

recordings were used as our gold standard, because they

were the most comprehensive. For this group, 1848

ap-neic events and 7040 abnormal respiratory events were

identified based on airflow, thoracoabdominal effort,

and Pes recordings. We then analyzed the

simulta-neously measured tcPCo2 and expired CO2 levels to

as-certain their ability to identify these same events. Results. The first night of polygraphic recording was

inconclusive enough to warrant a second recording in 316

of 411 children. Children were identified as having either

UARS (n = 259), OSAS (n 83), or other sleep disorders

(n 69). Children with small triangular chins,

retropo-sition of the mandible, steep mandibular plane, high

hard palate, long oval-shaped face, or long soft palate

were highly likely to have sleep-disordered breathing of

some type. If large tonsils were associated with these

features, OSAS was much more frequently noted than

UARS. In the 76 gold standard children, Pes, tcPCo and

expired CO2 measurements were in agreement for 1512 of

the 1848 apneas and hypopneas that were analyzed. Of

the 7040 upper airway resistance events, only 2314 events

were consonant in all three measures. tcPco2 identified

only 33% of the increased respiratory events identified

by Pes; expired CO2 identified only 53% of the same

events.

From the ‘Stanford University Sleep Disorders Clinic, Stanford, California;

Unite de Sommeil, Service de Physiologie, Hotel Dieu de Paris, Paris, France; and §Department of Pediatrics, Lucile Salter Packard Children’s

Hospital, Stanford, California.

Received for publication Jul 28, 1995; accepted Jan 19, 1996.

Reprint requests to (C.G.) Stanford University, Sleep Disorders Center, 701

Welch Rd. Suite 2226, Palo Alto, CA 94304.

Conclusions. UARS is a subtle form of

sleep-disor-dered breathing that leads to significant clinical

symp-toms and day and nighttime disturbances. When clinical

symptoms suggest abnormal breathing during sleep but

obstructive sleep apneas are not found, physicians may,

mistakenly, assume an absence of breathing-related

sleep problems. Symptoms and orocraniofacial

informa-tion were not useful in distinguishing UARS from OSAS

but were useful in distinguishing sleep-disordered

breathing (UARS and OSAS) from other sleep disorders. The analysis of esophageal pressure patterns during sleep was the most revealing of the three techniques used for recognizing abnormal breathing patterns during

sleep. Pediatrics 199698:871-882; sleep-disordered

breath-ing, upper airway resistance, obstructive sleep apnea, som-nambulism, night terrors.

ABBREVIATIONS. OSAS, obstructive sleep apnea syndrome;

UARS, upper airway resistance syndrome; NREM, non-rapid eye

movement; EEG, electroencephalographic/electroencephalogram;

Pes, esophageal pressure; tcPco, transcutaneous carbon dioxide

pressure; COP, carbon dioxide; EMG, electromyogram; MSLT,

multiple sleep latency test; Sao, oxygen saturation.

Although obstructive sleep apnea syndrome

(OSAS) has been well described in the literature,

little has been written about the upper airway

resis-tance syndrome (UARS). In 1982,’ we found that

some children have abnormally increased upper

air-way resistance during sleep, leading to increased

respiratory effort and sleep fragmentation, without

classically defined apneas and hypopneas2 or signif-icant drops in blood oxygen levels. In these children,

increased airway resistance may be compensated for

by increased respiratory effort during non-rapid eye

movement (NREM) sleep, which allows for the

main-tenance of appropriate minute ventilation. In rapid

eye movement sleep, the physiologic muscle atonia

impairs the ability to mobilize the respiratory

acces-sory muscles, but a moderate increase in breathing

frequency helps to maintain minute ventilation.

However, when upper airway resistance reaches a

certain point, it leads to an a

electroencephalo-graphic (EEC) arousal, which may be preceded by a

drop in tidal volume over a period of one to three

respiratory cycles.2 This pattern of increased effort

followed by arousal is repeated throughout the night

and leads to clinical symptoms similar to those seen

in OSAS. However, when apneas and hypopneas

alone are tabulated, this abnormal respiratory

pat-tern is missed, and children with UARS test negative

on nocturnal polygraphic recording. This may

mis-lead the pediatrician, and, as a result, leave daytime

at Viet Nam:AAP Sponsored on September 1, 2020

www.aappublications.org/news

(2)

symptoms and nighttime disturbances untreated or

only symptomatically treated.

The following report is a retrospective study of

children, drawn from our clinic database, who had

similar initial symptoms but who ultimately had

di-agnoses of either UARS, OSAS, or other sleep

disor-ders. In a subgroup of the studied population, we

compared measurements of esophageal pressure

(Pes), transcutaneous carbon dioxide pressure

(tcPco2), and expired carbon dioxide (CO2) to

deter-mine which of these methods is most appropriate for

distinguishing UARS from OSAS in children.

Population

METHODS

We identified all children in our database, since 1985, who had

a final diagnosis of sleep-disordered breathing and were younger

than 12 years (n = 469). We then identified the initial symptoms

that led these patients to consult the sleep disorders clinic and

sorted the group by symptom and age for comparison.

For inclusion in the study, the chart of each child contained all of the following: a validated sleep/wake questionnaire,3 results of a clinical interview, results of a pediatric evaluation (including a

neuropediatric examination), results of crarnofadal and upper

airway examinations, information on familial sleep/wake habits

and craniofacial features, and results of at least one night of

polygraphic recording performed in the laboratory. The charts of

58 children had to be discarded because one or more of the above

items were missing. The charts of 411 children (69 girls) 12 years

or younger at time of their initial consult were eligible for this

study. Of these children, 259 (12 girls) had UARS, 83(39 girls) had OSAS, and 69 (18 girls) had other sleep disorders. Children were

referred by general or specialized pediatricians (n = 294) and

other health practitioners indudmg orthodontists (n = 68) or were

brought by their parents (n = 49). Over the years, the number one

cause for referral to our sleep clinic from pediatric practices has

been suspicion of sleep-disordered breathing (about 80%). This

was also the case during the surveyed period.

To ensure that appropriate comparisons could be made, the

population was divided into three age groups: the infant and

toddler group (up to 24 months old; n = 54); the preschool

children group (2 through 6 years old; n = 253); and the

school-aged children group (7 through 12 years old; n = 104).

Clinical Evaluation

Parents filled out a questionnaire on the children’s sleep/wake

habits and sleep complaints before consultation. Responses to this

questionnaire were reviewed with the parents and the child

dur-ing the clinical interview. Each child had a general pediatric

examination, a neuropediatric evaluation, an evaluation of cranio-facial features, and a clinical evaluation of upper airway patency. Evaluation of family sleep habits (parents, siblings, and

grandpar-ents) and sleep-related problems was performed, as was a clinical

evaluation of craniofacial features and upper airway patency of

the parents present at the time of the interview (n = 705).

Nocturnal Polygraphic Monitoring

After this initial evaluation, all children underwent a nocturnal

polygraphic recording. The following variables were monitored:

EEG (C3/A, C4/A,, 0,-A2 of the 10-20 International electrode

placement system), electro-oculogram, chin electromyogram

(EMG), and electrocardiogram (modified V2 lead). Respiration

was monitored with an oronasal thermistor and thoracic and

abdominal bands to measure respiratory efforts. Blood oxygen

levels were measured with pulse oximetry (Neilcor, Oakland,

CA). For infants, tcPco measurement were added. Snoring

sounds were recorded with a microphone, and the intercostal

diaphragmatic EMG was recorded with superficial EMG elec-trodes.

Depending on the findings of the first recording, a second

nocturnal polygraphic monitoring may have been performed

mea-suring the same variables but with the addition of esophageal

pressure recording.4 To obtain baseline values, recording of Pes

was performed on awake children resting quietly in the supine

position for at least 30 minutes, either before sleep onset and/or after morning awakening.

A tcPco2 electrode was placed on the chest of 224 of 411

children, and the percentage of expired CO2 was measured during the night in 98 of 411 children using asampling catheter placed in

front of the nose and mouth. Simultaneous recording of Pes,

tcPco2, and expired CO2 was obtained in 77 of 411 children, 76 of

whom had sleep-disordered breathing of some type (OSAS or

UARS). The results obtained in these 76 children were used to

compare the ability of these three methods to best identify not

only OSAS but also UARS.

Multiple Sleep Latency Test

After the first night of polygraphic monitoring, 88 children

between the ages of 5 years, 8 months and 12 years, 2 months had

a Multiple Sleep Latency Test (MSLT). This test comprised four

20-minute naps in 31 children and five 20-minute naps in 48

children. (The change from 4 to 5 naps was the result of a decision

by the American Sleep Disorders Association to standardize the

MSLT for dinical studies.) In the school-aged group (7 through 12

years old), 76 children complained of daytime sleepiness, and,

with the exception of 2 children who had health insurance

reim-bursement problems, all had the MSLT. The lack of good

norma-five data limited the prescription of the test in younger children.54 Although there were 17 children complaining of daytime

sleepi-ness between the ages of 5 years, 8 months and 7 years, only 5

children were given the MSLT, and 4 of these were within 8 weeks

of their seventh birthdays.

Clinical Evaluation of Craniofacial Features and Upper

Airway Patency

To evaluate craniofacial features, we used a simple dinical scale

applied by the same individuals on all children seen. The scale is

based on visual evaluation and covers the following items: tonsil

size (from 0 = not visible, to 4 = tonsils touching at the middle);

length of the soft palate (from 0 = short, to 2 long); width of the

tongue (from 0 = normal, to 2 = enlarged); facial shape (from 0 =

square, to 3 = long face); position of the maxilla compared with

the mandible (from 0 prognathic, to 4 retrognathia); height of

the hard palate (from 0 = low-placed, to 2 = high-arched);

inter-molar width (from 0 = wide, to 2 = narrow); steepness of the

mandibular plane (from 0 = horizontal, to 3 = steep); and chin

size (from 0 wide, to 3 = small and triangular).7 A photograph

of the child’s face was also obtained during the first examination.

Analysis of Polygraphic Recordings

Sleep-Wake and Arousals

Sleep and wakefulness was scored using the international

Re-chtschaffen and Kales for children 12 months and older.

The Guilleminault and Souquet Infant Sleep AtIaS was used in

younger infants. The American Sleep Disorders Association

At-las’0 was used for the scoring of microarousals in children older

than 1 year. A systematic tabulation of transient (3- to 15-second)

or Rechtschaffen and Kales (more than 15 seconds) arousals

al-lowed for the calculation of an arousal index, the number of short

arousals (up to 60 seconds) per hour of sleep, and a wake after

sleep onset value that induded all types of arousals.

Apnea and Hypopnea

Apneas were divided into central, obstructive, and mixed after

the international definitions.8 A hypopnea was defined as a

de-crease in nasobuccal airflow by at least half compared with the

previous tracing. Recognition of hypopneas was aided by the

simultaneous analysis of oxygen saturation (Sao2) using pulse

oximetry.” To be considered valid, an Sao drop had to be of 3%

or more compared with the previous normal recording page. This

value of 3% or more is used internationally with oximeters that

average up to 12 pulsations, which was the case for this study.

Apneas and hypopneas were scored if events lasted 5 seconds

or more in children up to 9 months old and 9 seconds or more in

older children. Scoring of pathologic respiration during sleep was

performed following the recommendations of Marcus et al,’2

re-quiring the presence of at least one obstructive sleep apnea per

(3)

Chin EMG

2O

ARTICLES 873

hour of sleep and also the presence of an apnea-hypopnea index of

five or more events per hour of sleep.

Respiratory Rate and Noisy Breathing

Respiratory rate was calculated in different sleep stages. In

younger children, the normative data of Silberstein et al’3 were

used. Noisy breathing was scored based on behavioral

observa-tion, and snoring was scored as present or absent based on

con-tinuous monitoring of breathing sounds through a microphone

ta_ to the front of the neck (MESAM, Munich, Germany).

Intercostal Diaphragmatic EMG

Intercostal diaphragmatic EMG activity was monitored with

electrodes placed at 10-mm intervals on the right thoracic side

between the mamillary and axillary lines (usually at the seventh

intercostal space, but possibly above or below depending on the

baseline, awake EMG recording obtained), but the signal was not

integrated, and the envelope of the overall signal was analyzed

using artificial units determined for each child, considering the

envelope of the signal during quiet, supine breathing. This signal

was considered at a level of 100% activity during this baseline

monitoring. The percentage of increase or decrease of the envelope

of the EMG signal was determined. An increase of EMG activity

was considered present when the envelope of the EMG signal

reached an amplitude of 150% or more (see Figs I and 2).

Esophageal Pressure

Pes recordings were analyzed for determination of the peak

negative end inspiratory pressure and the peak end expiratory

pressure. A determination of Pes nadir (ie, peak negative end

inspiratory pressure) was performed in awake children resting

quietly in the supine position and in sleeping children during

periods of unobstructed breathing. A mean peak end inspiratory

Pes value was determined from nonobstructed wake and sleep

breathing periods for each child. At least 100 breaths were used to

calculate a mean Pes nadir value and the 2 SD values from this

mean. In children 2 years and older, a cutoff Pes nadir of -10

cm H20 was used (see full description of cutoff values below). This

cutoff point was always less negative in younger infants. Mean +

2 SD peak Pes values were identified for each infant, and the time

spent with a Pes nadir more negative than mean + 2 SD or more

negative than -10 cm H20 was calculated (the time calculated

induded the total respiratory cycle, not just inspiratory time).

Esophageal pressure recordings during wakefulness were not

in-terpretable in 15 infants, from whom waking Pes measurements

were difficult to obtain because of crying or rapid sleep onset.

In addition, specific breathing patterns were identified and

considered indicative of abnormal breathing during sleep:

1. A pattern called crescendo (Fig 1), indicated by a progressive increase of the A value related to an increasingly more negative

peak end inspiratory pressure. This crescendo ended, most

commonly, with an easily recognizable EEC arousal of variable

446 ., 447

c?. .- . .:-. #{149}#{149}

. .

k#{248}kr

LOC/A

.-.. -. - /

EKG #{149}#{149} &;A*

i:_#{149} J..

14 f

.

f

.

Mic. . . tI I

- -- - .-s-

-

-.-- -=ib - -*---.

--

,---

-5s.c

50%

0

Fig 1. Example of a crescendo in a 2-year-old child in NREM sleep. Note at the left the low (ie, difference from peak inspiration to peak

expiration) of the esophageal pressure (Pes) tracing, with progressively more negative peak end inspiratory pressure over time. In the

middle an arousal occurs with movement artifacts and large swings of the bands (channel 10 from the top), possibly associated with a

drop in Sao2 of 2% are shown. After the arousal, the Pes is smaller for the successive breaths and increases again thereafter. Snoring

is noted on the microphone (mic) channel. One can see that the simple recording of intercostal diaphragm activity (channel 7 from the

top: RIC) is not very useful to demonstrate increase of respiratory efforts as EMG bursts are present even after arousal. Use of an

integrated EMG may have been more indicative. This is the printout of a recording from a Nicolet computer with the following channels,

from the top: EEC (C4 /A1); EEC (02 IA1); chin EMC; left electrooculogram (LOC/A2); right electrooculogram (ROC/A1); EKC (modified

V2 lead); right intercostal diaphragmatic EMC (RIC); microphone (Mic); naso-oral airflow (Airflow); Thoracic band (Th Effort); esophageal

pressure (Pes); oxygen saturation (Sao2).

(4)

C/A2

O/A1

A4-R/A

I

Air Flow

Effort

I

4ft

*, -:- - -S .- -4i..- ,.

50%

0

766

. ,

,,-.

Fig 2. Example of abnormal respiratory effort during stage 3 NREM sleep (right) compared with normal breathing (left) during stage 2

NREM sleep, monitored near sleep onset in a 4-year-old child. Channels are the same as in Fig 1. Effort (channel 10 from the top) is from

thoracic and abdominal band recordings. Note the continuous increase in effort indicated by esophageal pressure recording, with a A Pes

equal to 20 to 25 cm H20.

duration (most frequently less than 15 seconds). The arousal

was associated with an abrupt return to less negative Pes

values. This crescendo was frequently associated with snoring.

An abrupt transient (3- to 15- second) a EEG arousal typically

terminated the crescendo, bringing about a decrease in Pes

nadir and period of normal quiet breathing.

2. A pattern of a very fast increase (three to four respiratory

cycles) in the Pes A (defined as the positive difference between peak end expiration and peak end inspiration for a respiratory cycle), with very little difference in the Pes nadir during these

respiratory cycles, and an EEC arousal with areturn to a much smaller Pes &

3. A pattern of tachypnea with a modest decrease in Pes &

fol-lowed by an abrupt termination of tachypnea (usually about 21

to 23 respirations per minute compared with 16 to 17 for

age-matched preschool and school-aged controls) with a

tran-sient EEC arousal.

4. In infants, a longer period of tachypnea ending with an EEG

arousal and a decrease in respiratory rate (see Fig 3).

These polygraphic patterns were commonly associated with

audible breathing (from loud breathing to loud snoring) that

became more quiet, even inaudible, at the time of the arousal.

Often a brief period of tachypnea, with hyperventilation involving

three to five breaths, was observed with the arousal. Our

tabula-tion identified the type of pattern and the duration of the pattern

before its interruption by an arousal and/or a return to normal

breathing. The repetitive presence of one of these patterns

associ-ated with fragmented sleep was considered as abnormal breathing

during sleep.

Scoring and Cutoff Values for Polygraphic Recording With

Pes Monitoring

To be considered pathologic, infants had to spend more than

10% of total sleep time with a peak end inspiratory Pes in the

pathologic range (eg, in children older than 2 years, 10% of the

sleep time had to be spent with Pes nadir more negative than 10

cm H20, as indicated above). This criterion alone was sufficient to

indicate the presence of abnormal breathing during sleep.

How-ever, in each case, we also required that the abnormal breathing

pattern had consequences that were visible on the polygraphic

recording, ie, an EEC arousal directly associated with one of the

abnormal breathing patterns described above. Fragmented sleep

was determined from the polygraphic recordings. Currently, there

is no agreement on what is considered a fragmented polygraphic

recording (normative data published in the literature vary signif-icantly in range and are affected by the differences in sleep

labo-ratory settings-noise level, technical staff, type of equipment

used, and many other idiosyncratic variables), so we established clear criteria for our sleep laboratory. These criteria, along with

those for determining abnormal breathing patterns, were

deter-mined and validated at our laboratory between 1980 and 1985,

during which time we monitored all of our controls, for specific

research grants, using Pes measurements to create a substantial

normative database. We reevaluated these criteria at

approxi-mately 2-year intervals on a sample of children’s recordings

per-formed in the Sleep Disorders Center to ascertain whether they

were still valid. During the 9-year survey period, the same three

experienced technicians recorded the children and scored their

polygraphic recordings. An EEC arousal was scored on a

mini-mum of 3-second durations. If the preceding EEC background

presented too much mixed frequency, the duration used to

(5)

ARTICLES 875

,‘#{149}v , r-- . T

LEG .. - - --

-AIRFLOW

-,wJIUJAvAMAjw1AJ’\iV\JvVcu*

4nn

50%

0

5sec

Fig 3. Tachypnea monitored in a 4-month-old infant with noisy breathing during sleep. Note the arousal response on the right. With

increased breathing frequency, despite the observed laborious breathing, Sa02 is maintained at a normal level. Note some noisy breathing

picked up by the microphone (channel 7 from the top) indicated on the right before the arousal.

firm the arousal may have been longer. EEC arousals were iden-tified using the central and occipital EEC leads, whereas a burst of

chin EMC and/or acceleration of heart rate were considered as

adjunctive information. From the data collected, fragmented sleep

was considered present: (1) if there were more than 11 arousals of

3 seconds or longer per hour of sleep; or (2) if parents at home

reported more than three long (15 minutes or longer) awakenings

at least four nights per week in children older than 2 years.

tcPco2 and Expired CO2

A change of 4 mm Hg was required on the tcPco curve to

affirm the presence of a nonartifactual change in tcPco2. This

change must not have been abrupt and immediately associated

with a large movement artifact. Expired CO2 was calculated as a

percentage. It must have been at least I % to be considered as not

artifact related, and it must not have been concomitant with a

large movement artifact.

Statistical Analysis

To determine whether clinical symptoms, signs, or polygraphic

findings were significantly different between OSAS and UARS, or

between children positive or negative for sleep-disordered breath-ing, parametric or nonparametric statistics were used, depending on the normalcy of distribution. Two-tailed analysis of variance

and Wilcoxon’s signed rank test were performed on the studied

variables. Pearson product moment correlations were also

calcu-lated. The Statview computerized statistical package was used.

RESULTS

Total Group

Of the 411 children in the total group, 259 had a

final diagnosis of UARS, 83 had a final diagnosis of

OSAS, and 69 had a final diagnosis of other sleep

TABLE 1. Birth Information

Total Studied Croup n = 411 (100%)

UARS Croup

n = 259 (63%)

OSAS Croup

n = 83 (20%)

Other Croup

n = 69 (17%)

Cestational age, (wks) Birth weight (g)

Duration of hospital stay after delivery

Up to 48 hours (Number of children)

2 to 6 days (Number of children)

6 to 8 days (Number of children) Age group at time of sleep consult Number of infant-toddlers

Number of 2- to 6-year-olds Number of 7- to 12-year-olds Number of girls

39 ± 0.9 2797 ± 194

385

16

10

54 253 104 69 (16.7%)

39 ± 0.7 2703 ± 162

248

8

3

37 161 61 12

40 ± 1.0

2702 ± 210

71

6

7 52 24 39

39 ± 1.0

2805 ± 200

66

2

1

10 40 19 18

Abbreviations: UARS, upper airway resistance syndrome; OSAS, obstructive sleep apnea syndrome.

(6)

Lt) OLt)

(“l OOO

“i t ©LrLr

a’ C©’Lr)

,-I- \O m’s’

,-c1cxmcflN.

c..1 a-mo’o’m

N. LrLrC1

ci\oo\cfmLn

O\ OOLflLIC’4 C) 4.-’

C’? \o\oq I% Lt4Lr)OQr1 0’

mmmm

m

c9)cf)\qt-:

? #{176}‘

.

i: ,c

u

_aJ)

0

L) o

0

bO’

.o

,

8

II

(I)

4)

(_1

II

(11 S

S

< II

.- S

r)R

HI

Os S

o

.ll

(4

(4 S

(a

(4

‘0 S (4 S 0

U) S

I-0 S

(4

N

(4

CA’S

U

CA

0

. 2

C-

0-#{149}CA5

Ii!

(44)0

,(4 .‘..

‘S

2

.

(t .U)U

U

.

0

:s (4

. 2

<z

disorders. Twenty-two children were Tanner stage 2;

all others were Tanner stage 1#{149}14Table 1 presents data

on birth history of these children and group

subdi-vision. The most striking finding of the table is the

low percentage of girls in the UARS group compared

with the total percentage of girls in the studied

group. This low number may be a result of the fact

that, for several years, we did not systematically look

for UARS in girls (see “Discussion”), and from

refer-ral bias.

Infant and Toddler Group

This group comprised 54 children. The clinical

symptoms that led to consultation were: (1)

repeti-tive disruption of nocturnal sleep, sometimes

associ-ated with night terrors (n = 33); (2) low growth with

snoring or noisy breathing and frequent airway

in-fection or otitis (n = 11); (3) disrupted nocturnal

sleep with esophageal reflux and airway infections

(n = 4); (4) abnormal daytime lethargy with snoring

and frequent airway infection (n = 2); and (5)

repet-itive airway infection, snoring, and enlarged tonsils

(n = 4). Referrals came from pediatricians for 24

children and from other health professionals for 8

children. For 22 of the children, parents requested

the consult. Referrals from health professionals were

for suspicion of OSAS in 19 children, for

investiga-tion of esophageal reflux in 4 children, and for

noc-turnal sleep disruption with or without night terrors

in 3 children. Six children were referred by social

workers, nurse practitioners, or clinical psychologists

because of sleep disruption of the family.

The symptoms expressed at clinical interviews are

indicated in Table 2. Some of these symptoms were

not causes of referral, and the same infant may have

had several symptoms. Suspicion of OSAS was the

conclusion of the initial sleep clinic evaluation in 35

of the 54 infants in this group.

Results of Polygraphic Recording

After the first night of polygraphic recording, 7

children had diagnoses of OSAS. Four children had

esophageal reflux at 24-hour monitoring. The initial

polygraphic recording indicated disturbed nocturnal

sleep in 46 infants, based on the number of long

nocturnal awakenings and/or short EEC arousals,

but no clear cause was present after this first

record-ing in 43 children. Based on clinical symptoms, there

was still a suspicion of a sleep-disordered breathing

syndrome in 39 of these 43 children. Parasomnia was

suspected in 5 children, because night terrors were

strongly emphasized by parents as a major

distur-bance but was confirmed in only 2 children, because

the event was monitored by video during the

labo-ratory recording.

For 41 children in this age group, results were

inconclusive enough to warrant a second nocturnal

polygraphic recording with the addition of Pes

mea-surements to determine respiratory effort. From this

second recording, 37 infants had diagnoses of

abnor-mal respiratory patterns during sleep, which is

in-dicative of upper airway resistance syndrome, based

on the percentage of time spent during sleep with

I

‘0

9

S

0

S

‘0 S

.E

.

S

lb

z

(7)

4) ‘S (I) 0 4) bc (4 4)5 U 4) bC 0 4)

IL

S S S S

L(t eiriri \ca\.

mcomn

,- ‘- e’4 c’ ‘ ‘r ‘

‘-LrC m Q\ ‘

o Lf) \ I’. \0 ri © ‘

,-0\ (10\

‘- 0 C 00 ‘- ‘- L) Cf) ,- 0’ ri Lt

oc LflOOLOLt) LflLfl

Lt Lt 0 C1 0 Lt Lt) t.. © 0 t

N.. Lft#{149}%0\

ri C © to t ‘ m

e’ Nr1rmm ri

Q\ (‘1 0’ r’4 N m .o m tt 1) m

LrC .o t% 0 N. 0 0 Lt) .Q Q C4

, a o .o 0 ‘Cr ‘ ‘- N. in c’ 0 0 0 N. \Q 0 0 ci ‘ . rl - ,-‘

0” 0 &) L) ‘ m ri r4 ‘ Ci’) r

-ri C ‘ ‘- t% C ri Lt -‘ ‘ t ‘‘ ,- r’)

a’ N. e1 c’1 0’ O LtC ‘ 0\ \O C (4 Cl’)q - - - ‘

‘-0’ t% 0 Q .4’ \O Lt) 0 (4 N. 0\ Q\ ri c’ 0 Lt) C’) Q\ 0 ‘.0 Q\ F 0 LtLt) ‘ m ri r ri ri ri

,-m r.. .= C4 0’ N.. m O P

2 r Lr ‘ e’ ‘ ‘ ‘ ,- ei .0

N. N C “1’ Lt) ri (%4to N. ca ‘

1’ ri ri n ‘- m c4 e4 0 Lfl C’) 0’ N O’ C’) ‘- ri - 00 C’)‘ -

-a C’) C’) 0’ %% = N C’) N 1’ N. ri

. * C’) ‘ C4 ‘- ‘- ‘ - r1 LI’) t 0\ 0\ ‘ N LI’)0 N. c’4 .0 LI)©

‘.0 LI’) LI’) ‘ C’) C’)(1) C’) I’1 r’4

rI r1 , . o a’ o ‘ .= r

1 N ‘ rl ‘ c’ 0 O’ 0\ .0 t ‘ C’)

Pes nadir more negative than 10 cm H20 and the

frequency of associated EEC arousals.

Preschool Children

This group included 253 children; 201 were

re-ferred by physicians, 42 were referred by other

health professionals, and 10 children were

self-re-ferred. The cause of referral was suspicion of OSAS

in 205 children, presence of a parasomnia

(sleep-walking or night terror) in 45 children, investigation

of a depressive reaction in 1 child, and esophageal

reflux in 2 children. The symptoms found at clinical

interviews and from the questionnaire are presented

in Table 3. At the conclusion of the clinical

evalua-tion, there was a suspicion of OSAS in 222 children in

this group.

Results of Polygraphic Recording

After one night of polygraphic recording, 52

chil-dren were identified as having OSAS. All of them

had at least one obstructive sleep apnea and also

more than five apneas and hypopneas per hour of

sleep. There was a clinical suspicion of a parasomnia

sleep disorder in 30 children, and an unresolved

suspicion of sleep-disordered breathing in 171

chil-dren.

In 201 children in this age group, results were

inconclusive enough to warrant a second nocturnal

mea-surements. Based on this second recording, 161

chil-dren in this group had diagnoses of UARS.

School-aged Children

This group included 104 children: 69 were referred

by physicians, 10 by orthodontists, and 8 by other

health professionals, and 17 were self-referred. The

cause for referral was suspicion of OSAS in 77

chil-dren, bruxism in 7 children, suspicion of nocturnal

epilepsy in 2 children, daytime sleepiness and

suspi-cion of narcolepsy in 8 children, abnormal nocturnal

sleep and diminished school performance in 4

chil-dren, and sleepwalking and night terrors in 6

chil-dren. The symptoms identified at clinical interviews

and by questionnaires are summarized in Table 4.

Results of Polygraphic Recording

After the first night of polysomnography, the

di-agnosis of obstructive sleep apnea was confirmed in

24 children. In 6 children other

non-breathing-re-lated sleep problems were strongly suspected. In the

other 74 children, there was an unresolved suspicion

of a sleep-related breathing problem.

In 78 children in this age group, polygraphic

re-sults were inconclusive enough to warrant a second

nocturnal polygraphic recording with the addition of

Pes measurements. Based on this second recording,

61 children in this group had diagnoses of UARS.

Comparison of Children With OSAS and UARS

The possibility of sleep-disordered breathing was

mentioned at clinical evaluation in 367 of our charts,

not necessarily as the first considered diagnosis. The

outside referrals to the center considered

sleep-dis-ordered breathing as a possibility in 300 children.

4) 0 .5 ‘ (I) L) (flCA5 4)0 0 4)0 bO’ ‘ ,S E’o 0 0 0.11 J) < II IIS C’) -(8 5r.1 . II 1=05 (I) S bO (4 (4 S ‘0 S (4 U) 4) bO S (4 4) I-,0

I

S

I

. 4) 0 I.. ‘O

.1

S 4) ‘0 U ‘0

2

> 0 8 .5 4) ‘S S 4) 4) S (4 .

a.

4) 0 C) -J I-0. U) 40 .. . 0 ‘0 S (4 ARTICLES 877

(8)

0.2 #0

E

L)’

n,0 ,

I

.c

0. II

II S

1=

o ‘r Lf C’) 00 Lt 00 t.

qq ce Lt t-.OLt)N C’

LI)LI) 0’ O 0 0’ 00 0’

I-. I’1 0 ‘- C’) rl ‘ C

t) C’) L( C”) C’) C’) LI)

C’) C’) LI) LI) 0 0 t . ‘0 LI) O 0

t%(’ I- 1= I’1I’C’1 r

: ‘- 0 1’) 0 (‘4 0 0’ ‘ LI) C’)U) 0 CI 0\

. ‘C C’)t. LI) C C’) C’)e- i- \O 00 %0 C’) 0’ Q r N. O C) (‘40 LI) 0 C) ‘

F LI) ‘0 C’) C’)C’) ‘-‘ . r’i CN .-

-. 0’ LI)0 0’ C’1 t’4 0 N. 0’ 0’

NC4N(4,

0 e .- ‘i 0’ ‘0 t% C 0 LI) .

t’. o N. C’) C’) C’) I- (- I’I c1 i#{149}- ,-

‘-N. LI) LI)C$ F C) ‘-‘ LI) \0 00 0’ 1’)

‘- C’) 0’ 0’ \O 0’ ‘-4 C’) . C 0’ C’)

0’ N N. rI rl ri ,- C4 .- .

C’) 00 O LI) C’) C’) LI) LI) I1 O C’) C’)

LI’)0 N t \O C4I’) t 0’ 0 (‘1 LI) t

rLt1

0 O I”4 4 0’ C’) t”. LI) i- ( O

0’mC’)I”4 C%1_

0

S

0o.

H

S

‘0 S

4)

I

1;

I

J

I

.

i

0

‘U -J

1=

After polygraphic monitoring, sleep-disordered

breathing was the final diagnosis in 342 children. A

total of 83 children had been identified with OSAS

based on one night of polygraphic recording. A

group of 320 children underwent a second night of

mea-surements, and 259 of them were shown to have

upper airway resistance syndrome.

The arousal index was more than 12 per hour of

sleep in 229 of the 259 children with UARS, with a

mean value of 16 ± 4 arousals per hour of sleep in

those children older than 12 months. The mean

arousal index in OSAS patients of similar age was

15 ± 7 arousals per hour of sleep (not statistically

different, two-way analysis of variance). The mean

lowest Sao2 was 95 ± 1 % for the children with UARS;

it was 88 ± 5% for the children with OSAS (P < .001,

Wilcoxon rank signed test).

Results of the Orocraniofacial Clinical Scale

The distribution of scores on the various scales

were divided into thirds: lower, middle, and upper.

We paid particular attention to children with

ex-treme scores on these scales, in the lower or upper

third. The evaluation of tongue width proved to have

little clinical value and was eliminated from our

analysis. Tongue width was noted to be large only in

several trisomy 21 cases. Otherwise, 80% of the

chil-dren had very similar scores, largely because of the

difficulty of visually scoring this characteristic. Table

5 presents the mean score for each of these groups.

We identified the individuals who had a combined

score in the top third (>13.8) and those in the bottom

third (with a score 6.5). A group of 110 children

scored more than 13.8 and 49 scored 6.5 or less. Of

the 110 children scoring in the top third, 106 had

sleep-disordered breathing (OSAS or UARS). The

orocraniofacial scale thus identified a subgroup of

sleep-disordered breathing children with common

craniofacial features: a small chin, a steep

mandibu-lar plane, a retroposition of the mandible, a long face,

a high hard palate, and an elongated soft palate. Of

the population with sleep-disordered breathing, 32%

had these anatomical features, including 35 of 83

OSAS children (42%) and 71 of 251 UARS children

(28%). However, these features were not helpful in

distinguishing between OSAS and UARS.

Clinical evaluation of the tonsils was treated

sep-arately. Of the children, 146 scored in the bottom

third (1.3), and 51 scored in the top third (2.7) for

tonsil size. Forty-seven (92%) of the 51 children in the

top third had sleep-disordered breathing. All

chil-dren who had high scores on both the orocraniofacial

scale and the tonsil size scale were identified as

having OSAS on the first polygraphic recording

night (n = 33). Of the 18 other children in the top

third for tonsil size, 8 scored in the lower third on the

orocraniofacial scale; these children were all in the

UARS group.

if

a child had high scores on both scales, OSAS was

the most common diagnosis. If a child had a high

score on only one of the two scales, UARS was the

most common finding. In children who had low

scores on both scales, sleep-disordered breathing

(9)

S 400 S . +1 0 VS U) (4 0.5 S S . +1 0 VS U) (4 4) S C) .5 ‘-I Q) 4) (4 S 5 0 (4 0 1= a) ‘5 S 0 S 0 ..s (4 40 ..s . (4 .a (4 0 ARTICLES 879 LI)

t#{149}%I’1\O 0’ c’4 ‘- C’LI)

gi ci 0 C’) 0 I’l 0 C’) 0 . 0 C’4 0

+1 +1 +1 +1 +1 +1 +1

C . ( \0 LI)

0 0 0 0 0 0

C’) 0 C’) 0 0 N 0 0 C’4 0 LI) 0

+1 +1 +1 +1 +1 +1 +1

C’) c r ,-c’i C ,-. LI) ri LI) .-C’) rI (4 ;E (

t

_S ₂

. .

L.. ‘0

‘0 :‘

S (4U(4

4t}:u;.

o’

.e E o .

CA’S CA 0.ll

II II II II II II II II II II II II II II

0 C’) 0 C’) 0 0 C”l 0 C’) 0 C1 0 Cl

4) ‘0 4) 5 S (4 (4 E C 5 (4 ‘ll .

‘0 (4 .

S (4 .5

E’0

(4 . r’

_ ‘S U)

4) ll #{176}

. 4) 5 0 0

CA

was very uncommon. Of 49 children who scored in

the bottom third of the orocraniofacial scale and had

a score of 2 or less for tonsil size, only 1 child was

identified as having sleep-disordered breathing; this

particular child had a diagnosis of UARS.

The clinical scales were systematically obtained at

the initial interview. As used (ie, considering the top

third of the scores), they identified only 124 (36%) of

342 children with sleep-disordered breathing. There

was a correlation between high scores on both of our

clinical scales and sleep-disordered breathing

(Pear-son product moment correlation, P < .0001), but this

involved only 10% of the children with

sleep-disor-dered breathing and 40% of the children with OSAS.

Results of the MSLT in a Subgroup of Children Older

Than 5.8 Years

Among the children younger than 7 years who had

been given the MSLT, there were 3 children with

UARS, 1 child with OSAS, and 1 child with another

sleep disorder. Among the children 7 years and

older, 46 had UARS, 23 had OSAS, and 14 had other

sleep disorders. The mean sleep latency was 15 ± 3

minutes in the other sleep disorders category, and

13 ± 4 minutes in the sleep-disordered breathing

category (barely not significant at P < .06). The

dif-ference between the MSLTs of children in the OSAS

and UARS groups was not significant. MSLT was not

useful in the subdivision of children according to the

cause of their daytime sleepiness.

Comparison of tcPco, Expired CO and Pes

Measurements in the Recognition of UARS

Seventy-six children with sleep-disordered

breath-ing (21 with OSAS and 55 with UARS) had

simulta-neous recording of Pes, tcPco2, and expired CO2. The

results of Pes measurements were used as the gold

standard for the diagnosis of UARS. All of the

ana-lyzed events were identified during NREM sleep. A

total of 1848 apneas and hypopneas were analyzed,

as were 7040 respiratory segments with an

abnor-mally increased end inspiratory pressure nadir

asso-ciated with an arousal.

In 1512 of the 1848 apneas and hypopneas that were

analyzed, changes in tcPco, expired COP, and Pes were

noted simultaneously. The percent of expired CO2 was

not interpretable in 336 events, primarily because of

dogged catheters that led to erroneously low expired

CO2 values. tcPco2 measurements failed to identify 73

of the 1848 events, 57 of which were hypopneas and 16

of which were not identified because of artifacts.

Of 7040 upper airway resistance events, the

corre-lation between the three measures was much lower.

There were only 2314 events for which there was an

agreement between the three measures. Only 2346

(33%) of the 7040 Pes-identified events showed an

increase of 4 mm Hg or more for tcPco2. Artifacts

limited the interpretation of tcPco2 in only 369 events

(about 5%). The remaining 4325 events (61%) would

not have been recognized based on tcPco2

measure-ments alone.

When a comparison between percent of expired

CO2 and Pes was performed, there were 1435 events

identified with Pes recording that were not

(10)

able with expired CO2 because of artifacts resulting

from inappropriate air sampling. When sampling

was valid, however, expired CO2 still failed to

iden-tify 1816 respiratory segments with increased airway

resistance. Expired CO2 was, therefore, only able to

recognize 3749 (53%) of 7040 increased airway

resis-tance events identified by Pes recordings.

DISCUSSION

Since 1982,1 it has been documented that abnormal

upper airway obstruction during sleep may exist

without apnea; in fact, complete obstructive sleep

apnea may be rare in children.12”5 However, in the

absence of apneas, hypopneas, and oxygen

desatu-rations, the problem of increased upper airway

re-sistance162#{176} is still overlooked, despite the fact that it

may lead to nocturnal sleep disruption, daytime

sleepiness, and an assortment of symptoms related to

sleepiness.

Depending on the age of the child, symptoms may

vary, and sleep-disordered breathing may not be

suspected because many of the symptoms can be

associated with several causes. Symptoms such as

sleepwalking, night terrors, disturbed nocturnal

sleep, abnormal shyness, and rebellious and

aggres-sive behavior may not be immediately related to

abnormal breathing during sleep, but analysis of the

reported symptoms presented in Tables 3 and 4

in-dicate that such symptoms are not uncommon with

sleep-disordered breathing in children. These events

appear secondary to the arousals induced by the

abnormal breathing during sleep. However, the

anal-yses also indicate that such symptoms are not

dis-criminative in our population. Sleep-disordered

breathing can be strongly suspected based on reports

of well-known symptoms, such as snoring or noisy

breathing during sleep, daytime mouth breathing,

low growth, or repetitive upper airway infection.

Comparisons of symptoms between UARS and

OSAS in children, however, indicates that there is a

very large symptom overlap between the two

syn-dromes, which explains the discrepancy between

clinical suspicion based on symptom analysis and

polygraphic findings when only obstructive events

are sought.

More than one fourth of the children studied were

not referred by their pediatrician or by an

otorhino-laryngology specialist but were self-referred or sent

by uncommon sources of referral, such as school

psychologists. In many of these patients, parents had

mentioned to their health professional the problem

(commonly involving noisy breathing during sleep)

that ultimately led them to consult our clinic;

how-ever, the symptoms were not considered indicative

of disordered breathing during sleep. However,

when symptoms were suggestive enough (eg,

repet-itive upper airway infections, slow growth, or failure

to thrive) pediatricians were often surprised when no

obstructive apnea was noted during nocturnal

poly-graphic recording and often questioned why a

sub-sequent, more complicated test was recommended.

Finally, the absence of snoring was commonly

con-sidered by pediatricians as an indicator that there

was no sleep-disordered breathing, even if there

were other clinical symptoms that might indicate the

presence of a sleep-disordered breathing syndrome.

A total of 294 children were referred to the Sleep

Disorders Center by more or less a specialized

pedi-atric unit. As mentioned previously, the referrals

were for suspicion of sleep-disordered breathing,

which may introduce a bias into our overall

popula-tion but is probably, unfortunately, a fair

represen-tation of the type of disorder referred to a specialized

pediatric sleep disorders clinic. This report indicates

another referral bias in that there were many fewer

girls than boys referred to the clinic. One may only

speculate why such a bias is present. For years, it has

been reported that women were protected from

OSAS, and it is only recently that it has been shown

that women may have different symptoms and more

often have UARS than full-blown OSAS.21 This, by

now well-demonstrated, erroneous statement,

asso-dated with polygraphic results negative for OSAS,

may have led pediatricians to consider other

diag-nostic pathways and limit the number of referrals to

a specialized sleep clinic and laboratory. Hopefully,

our report will eliminate any doubt considering the

near equal presence of sleep-disordered breathing in

both genders. Because disorders during sleep are

difficult to investigate from reports of parents who

are only aware of the daytime symptoms or

symp-toms that disturb their own sleep, all children seen in

the clinic had at least one nocturnal polygraphic

recording. Because the very new ambulatory

equip-ment with EEC monitoring was not available at the

time of the survey, all recordings were performed in

the laboratory and in the same recording room. This,

at least, limits bias; once a child was seen in the sleep

clinic, the same procedures were followed.

Certain orocraniofacial features are highly

sugges-tive of a breathing disorder during sleep when

asso-ciated with specific clinical symptoms and need to be

given some attention by pediatricians. As we

re-ported, children with a small triangular chin, a steep

mandibular plane, a retroposition of the mandible, a

long face, a high hard palate, and/or an elongated

soft palate are very likely to have some type of

sleep-disordered breathing, as are children with

clearly enlarged tonsil linguae.24 These features,

however, identify only a subset of children with

sleep-related breathing disorders, and the absence of

such features should not lead one to dismiss the

possibility of sleep-disordered breathing. These

chin-ical measures were not useful for distinguishing

OSAS from UARS.

Obstructive hypopnea during sleep was

recog-nized early on, and these partial occlusions were

focused on for many years, because they induce an

easily recognizable hypoventilation when they have

a certain duration (at least 10 through 15 seconds)

and are associated with Sao2 drops that are easily

measurable with noninvasive equipment such as a

pulse oximeter. A hypopnea may involve only one

respiratory cycle, however, and it is then difficult to

identify. Upper airway resistance syndrome is

char-acterized by increasing respiratory efforts that can

lead to a hypopnea of one or two respiratory cydes

and subsequently to an arousal.26 The responses to

(11)

ARTICLES 881

these increased respiratory efforts during sleep can

lead to changes in EEC spectra during sleep,2627 and

to increases in interventricular septum shift with

changes in right and left intraventricular dimensions

to the point at which pulsus paradoxus can be

ob-served.28 It has been hoped that noninvasive

moni-toring of CO2 could help recognize these changes in

breathing patterns? If only one respiratory cycle is

involved, however, changes in CO2 may be too small

to detect without more invasive technology. Even if

two or more respiratory cycles are involved,

success-ful monitoring of expired CO2 can be a real challenge

during a single night’s recording. Neither end-tidal

CO2 as measured with a face mask or tcPco2 are a

consistently accurate reflection of Paco2;#{176}these

tech-niques may therefore miss some respiratory changes

that can have an impact on sleep continuity.

It is important to investigate sleep with as little

disturbance to the child’s sleep as possible, although

obtaining as much information as needed. We agree

with Morielli et a129 that the nasal sampling catheter

used to monitor end-tidal CO2 can be uncomfortable

for some children, that it can be easily displaced by

movement, and that the catheter may become

blocked by mucus. Some children are mouth

breath-ers, and two catheters are then needed, which

in-creases the risk of error. In addition, the mechanical

impact of increased inspiratory efforts will not be

indicated by changes in respiratory gases.

During the past 15 years,”162#{176}it has been

demon-strated in adults and in children that there is a

bal-ance between negative intrathoracic pressure and the

dilating forces exercised by the upper airway

mus-des during inspiration. Esophageal pressure

moni-toning indirectly investigates intrathoracic pressure

and, therefore, gives information on respiratory

ef-forts. If monitored in association with a

pneumota-chometer, it gives information on quantification of

airflow, which allows for the calculation of tidal

vol-ume and minute ventilation.26’ We have previously

reported results obtained on a small group of

chil-dren3’ measured for quantification of airflow, using a

face mask and pneumotachometer, and for effort

using esophageal pressure recordings. After having

correlated these measures, we now use only Pes

measurements. Although the measurement of

esoph-ageal pressure is more invasive than techniques

us-ing thoracoabdominal bands or impedance systems,

it brings much more information. Of course, even in

experienced hands and with a very small (1 .6-mm

diameter) pressure catheter, problems do occur. We

had to repeat the nocturnal Pes recording in 7% of

our population to obtain valid results. Despite these

problems, however, this approach best fulfilled the

combined criteria of ease of use and ability to collect

the needed information.

As indicated by one of the referees of this report,

36 of 37 children 6 to 24 months old with diagnoses

of UARS were reported by parents to present

chronic noisy breathing. Chronic noisy breathing

(which includes snoring) was associated with

sleep-disordered breathing in 74% of the affected

children in that age group. Two issues are, thus,

worth emphasizing: (1) children in that age range

commonly sleep during the daytime, and parents

may recognize more easily a sleep-related

disor-der; and (2) considering the seemingly frequent

clinical association with positive laboratory

find-ings, it may be possible to avoid Pes

measure-ments. Many consider Pes measurements invasive

in infants. Our technique is based on the use of a

very supple infant feeding tube as the sensor.32

However, avoidance of Pes monitoring would be a

gain. Based on the reports of Stoohs and

Cuillemi-nault3 and Berthon-Jones, trials are made to

monitor airflow through a nasal cannula placed in

each nare and connected to an external pressure

transducer, if the infant is a nose breather, the

shape of the flow curve of each breath may allow

determination of flow limitation (Dr David

Rap-poport, Division of Pulmonary Medicine,

Colum-bia University, New York, NY, oral presentation,

1995). However, mouth breathing will be a

hand-icap, and most probably a face mask, with all its

inconveniences, will be needed to collect the

ap-propriate information.

One may also question how common certain

para-somnias, such as night terrors and sleepwalking, are

in the general population compared with the

fre-quency reported in our case study. The best

pub-lished longitudinal prospective study on the subject

is that of Klackenberg5 performed in Stockholm,

Sweden. A total of 25% of the children (up to 5 years

old) followed had frightening awakenings on and

off. Pavor nocturnus, however, was much more

un-common, and 3.5% of the children (the oldest was 7

years old) presented the characteristics of this

para-somnia. Similarly, rare episodes of somnambulism

were noted between 6 and 16 years in 40% of the

children, with the greatest prevalence in those I I to

12 years old (16.7%). However, independent of age,

only 2% to 3% had at least one episode per month. A

recent investigation of violence during sleep37

mdi-cated that nonviolent repetitive sleepwalking is often

considered of little concern by family and

pediatri-cians, and it is often only after an accident has

oc-curred that referrals are made. In the population

reported here, night terrors and sleepwalking had

occurred at least once per week during the past 4

months (and commonly much more often). Tables 2

through 4 indicate that these two clinical symptoms,

at the frequency emphasized above, had occurred in

a much higher proportion of the referred population

than would be expected from the general population

data. The difference was marked and reached a

fac-tor of 10. Finally, we acknowledge that in Tables 2

through 4 the other sleep disorder groups are small

(eg, in Table 3 there were 213 children with

sleep-disordered breathing, and 40 children were referred

with other sleep disorders. Undoubtedly, this is a

limitation of the study when comparisons between

the groups are made.

Despite these drawbacks, a fact is clear: although

more than 114 articles have been devoted to OSAS in

children during the past 5 years (based on a

MED-LINE search), the study of UARS in children led to

only one report during the same period. More than

likely, children with this syndrome are currently