Incidence and Remission of Sleep Disordered Breathing and Related Symptoms in 6-17 Year Old Children-the Tucson Children's Assessment of Sleep Apnea Study (TuCASA)

Abstract

Objective

To determine the incidence and remission of sleep disordered breathing in adolescent children.

Study design

319 children completed two home polysomnograms approximately 5 years apart. Sleep disordered breathing (SDB) was determined to be present if a child had a respiratory disturbance index ≥ 1 event per hour associated with a ≥3% oxygen desaturation. Subjective symptoms such as witnessed apnea, excessive daytime sleepiness, difficulty initiating and maintaining sleep, and habitual loud snoring were considered present if they occurred frequently or almost always. BMI percentiles were calculated using CDC childhood growth charts adjusted for sex and age.

Results

The mean age at assessment was 8.5 years at Baseline and 13.7 years at Follow-up respectively. Incident SDB was more common in boys (OR=3.93, p=.008, CI= 1.41-10.90). Children with Prevalent SDB were more likely to be boys (OR=2.48, p=.006) and had a greater increase in BMI percentile change (OR 1.01, p=.034). Children with Prevalent SDB also had 3.41 greater odds of developing obesity from Baseline to Follow-up in comparison with children with Prevalent NoSDB.

Conclusions

Adolescent boys are more likely to have persistent and incident SDB than girls. Children with prevalent SDB are more likely to have developed obesity. These risks are similar to those observed in adults.

Sleep disordered breathing (SDB) in children has been associated with a number of physiological, neurocognitive, and behavioral problems.^1-4 Moreover, there is increasing evidence that sleep disordered breathing has an important adverse impact on a child's physical and cognitive health and social development.^{1, 5-7}

Although there are several studies specifying the prevalence of snoring and other sleep related problems in children,^{8-10 11, 12} there is little evidence documenting the incidence and remission of symptoms associated with sleep disorders in children, such as witnessed apnea, excessive daytime sleepiness, habitual snoring, and insomnia.^13-15 Furthermore, there are no data to date indicating the incidence and remission rates of SDB in a large sample of children using polysomnography as well as identifying factors associated with incidence or remission. This gap in the research literature was highlighted in an extensive recent review of the epidemiology of pediatric obstructive sleep apnea (OSA).¹⁶ The Tucson Children's Assessment of Sleep Apnea (TuCASA) study is a longitudinal cohort study designed to investigate the physiologic, anatomic and behavioral correlates of SDB and to document their changes with development. In this analysis, we present data on the incidence, remission, and persistence of SDB and related symptoms.

Methods

Details of the TuCASA study design have been published previously.^{10, 17} Briefly, Hispanic and Caucasian children ages 6 to 11 years were recruited to undergo unattended home polysomnography and perform a neurocognitive assessment. Subjects were recruited through the Tucson Unified School District (TUSD), a very large district with a substantial elementary school population. Parents were asked to complete a short screening questionnaire and to provide their contact information if they were willing to allow study personnel to contact them to determine if their child was eligible for the study. A total of 7,055 screening questionnaires were sent home with children in a “notes home” folder. Of these, 2,327 (33%) were returned. Recruitment information was supplied on 52% of the returned questionnaires from which we selected children, based on pre-established inclusion and exclusion criteria, to undergo polysomnography. An unattended home polysomnogram was scheduled as soon as possible after recruitment. From 1999-2003, 503 children aged 6-11 years completed home polysomnograms (Baseline). Approximately five years later (Follow-up, mean 4.7 years), 348 children participated in the second phase of the study; 319 children had home visits where acceptable in-home polysomnography was completed. On both occasions, all of the families completed sleep screening, sleep habits, and morning questionnaires. The TuCASA study was approved by the University of Arizona Institutional Review Board (IRB) as well as the Tucson Unified School District (TUSD) Research Committee.

The methods for obtaining data have been previously described¹⁷. In brief, a two-person, mixed sex team arrived at the home approximately one hour prior to the child's normal bedtime. Prior to performing any study procedures, parents gave informed consent and the child gave assent to the study using language appropriate IRB approved forms. Each child's height, weight, neck circumference, and blood pressure were measured. A parent was asked to complete a comprehensive Sleep Habits Questionnaire (SHQ) that inquired about their child's sleep history and sleep characteristics.

A single, unattended overnight polysomnogram was obtained using the Compumedics PS-2 system (Abbotsford, Victoria, Australia). The following signals were acquired as part of the TuCASA montage: C3/A2 and C4/A1 electroencephalogram, right and left electrooculogram, a bipolar submental electromyogram, thoracic and abdominal displacement (inductive plethysmography band), airflow (nasal/oral thermister), nasal pressure cannula, finger pulse oximetry, ECG (single bipolar lead), snoring microphone, body position (Hg gauge sensor), and ambient light.

Scoring of sleep was performed by a single registered polysomnographic technologist using Rechtschaffen and Kales criteria.¹⁸ Arousals were identified using criteria published by the American Academy of Sleep Medicine.¹⁹ Apneas were scored if the amplitude (peak to trough) of the airflow signal using the thermistor decreased below at least 25% of the amplitude of baseline breathing (identified during a period of regular breathing with stable oxygen levels), if this change lasted for more than 6 seconds or 2 breath cycles. Hypopneas were designated if the amplitude of any respiratory signal decreased below (approximately) 70% of the amplitude of baseline and if the thermister signal did not meet the criterion for apnea. Central events were marked if no displacement was noted on both the chest and abdominal inductance channels. However, central events that occurred after movement were not included. Otherwise, events were scored as obstructive. After full scoring, analysis software was used to link each event to data from the oxygen saturation and EEG channels. The Respiratory Disturbance Index (RDI) was defined as the number of respiratory events (apneas and hypopneas) per hour of total sleep time. For this analysis, a 3% oxygen desaturation was required for an event to be counted in the total RDI. In our primary analyses, we considered a child to have SDB if their RDI was greater than or equal to 1 event per hour of total sleep time. Use of this definition is supported by previous evidence that a RDI of one, based on events with a 3% oxygen desaturation, is clinically significant.^{12, 20, 21} We also explored the impact of using definitions of RDI with cutpoints up to 10 events per hour.

Witnessed Apnea (WITAP) was defined as stopping breathing, struggling to breathe, or having to shake the child awake while sleeping. Excessive Daytime Sleepiness (EDS) was defined as being sleepy in the daytime, falling asleep at school or while watching television. Snoring (SN) was present if the parent reported their child snored loudly. Insomnia (DIMS) was present if the child currently experienced trouble falling asleep, staying asleep, or waking too early. The symptom was defined as present if the parent reported it occurred “frequently” or “almost always”.

Body mass index (BMI) percentiles were calculated using U.S. Centers for Disease Control childhood growth charts adjusted for sex and age.²² Children were categorized as obese if they exceeded the 95th percentile of BMI for their age, sex, and ethnicity. Variables pertaining to incidence and remission of SDB were defined as follows: IncidentSDB denotes no SDB at Baseline and SDB at Follow-up; RemissionSDB denotes SDB at Baseline and no SDB at Follow-up; PersistentSDB denotes SDB at Baseline and at Follow-up; PersistentNoSDB denotes no SDB at Baseline and no SDB at Follow-up; PrevalentSDB denotes Persistent SDB group + Incident SDBgroup; PrevalentNoSDB denotes RemissionSDB group + PersistentNoSDB group.

Data Analysis

The Student t-test and the 2 sample test of proportion were used to compare differences in characteristics at baseline and follow-up. Potential predictors of incident and remissive SDB were also investigated. These variables included change in neck circumference, BMI percentile, sleep efficiency, and total sleep time. Additional variables included parent reported witnessed apnea, snoring, excessive daytime sleepiness, insomnia, sex, and ethnicity. Logistic regression was used to determine predictors of incidence, remission, and persistence or absence of SDB as defined above.

Results

The mean age at first assessment was 8.5 years (range: 6-12), and mean age at second assessment was 13.7 years (range: 10-18). The mean time between assessments was 4.6 years (range: 2.9-7.3). There were 50.4% males and 49.6% females at the time of initial enrollment; this sex ratio remained approximately the same at the second assessment. Ethnicity was 58.5% Caucasian and 41.5% Hispanic.

Table I shows the characteristics of the cohort at Baseline and Follow-up. Of note was that the prevalence of both SN and WITAP, as well as polysomnographically determined SDB, decreased from Baseline to Follow-up. SN declined from 15.0% to 9.7% and WITAP decreased from 5.2% to 2.2%. Moreover, the percentage of children with a RDI ≥1 fell from 23.9% to 15.3%. In contrast, the rate of obesity increased from 12.9% to 19.2%.

Table 1.

Characteristics of the Tucson Children's Assessment of Sleep Apnea Study Participants at Baseline and Follow-Up^*

	Baseline (N=480)		Follow-up (N=319)
Characteristic	Mean ± SD	Range	Mean ± SD	Range	p-value
Age	8.5 ± 1.6	6-12	13.7 ± 1.8	10.2-18.3	<0.001
Respiratory Disturbance Index
≥ 1	2.6 ± 3.4	1-29.1	2.6 ± 4.4	1-30.7	0.89
< 1	0.4 ± 0.3	0-0.9	0.2 ± 0.2	0-0.9
	No.	%	No.	%
Respiratory Disturbance Index
≥ 1	115	23.9%	49	15.3%	0.003
< 1	365	76.1%	270	84.6%
Ethnicity
Hispanic	203	42.3%	114	35.7%	0.06
Non-Hispanic	277	57.7%	205	64.3%
Sex
Male	240	50.0%	163	51.1%	0.76
Female	240	50.0%	156	48.9%
Obesity
Yes	62	12.9%	61	19.2%	0.02
No	418	87.1%	256	80.%
Snoring
Yes	72	15.0%	31	9.7%	0.03
No	408	85.0%	288	90.3%
Excessive Daytime Sleepiness
Yes	78	16.35%	48	15.1%	0.67
No	402	83.8%	269	84.9%
Witnessed Apnea
Yes	25	5.2%	7	2.2%	0.03
No	455	94.8%	312	97.8%
Insomnia
Yes	141	29.9%	80	25.3%	0.21
No	339	70.6%	236	74.7%

^*Representing children with acceptable PSG

The incidence and remission rates for SDB and symptoms of sleep disturbances are displayed in Table II. Incidence rates for key symptoms of sleep disturbance were: SN 4.1%, EDS 11.8%, WITAP 0.9%, and DIMS 16.6%. Remission rates were SN 60.3%, EDS 63.8%, WITAP 84.0%, and DIMS 47.8%. For polysomnographically determined SDB, the incidence rate was 10.0% and the remission rate was 70.8%.

Table 2.

Incidence and Remission of Sleep Disturbances

	Incidence of Symptoms		Remission of Symptoms		Reporting “No” at Baseline and Follow-up		Reporting “Yes” at Baseline and Follow-up
	%	No./No. at risk	%	No./No. at risk	%	No./No. at risk	%	No./No. at risk
Snore	4.1%	12/290	60.3%	35/58	95.9%	278/290	39.7%	23/58
Excessive Daytime Sleepiness	11.8%	35/297	63.8%	30/47	88.2%	262/297	36.2%	17/47
Witnessed Apnea	0.9%	3/323	84.0%	21/25	99.1%	320/323	16.0%	4/25
Insomnia	16.6%	42/253	47.8%	43/90	83.4%	211/253	52.2%	47/90
SDB	10.0%	23/230	70.8%	63/89	90.0%	207/230	29.2%	26/89

Using a logistic regression model and empirically testing each predictor variable, children with IncidentSDB were compared with children with Prevalent NoSDB. We also compared children with RemissionSDB to those children who had PrevalentSDB. Boys were more likely to develop SDB (OR=3.93, p=.008, CI = 1.41-10.90).

In addition to exploring incident and remissive cases alone, we explored predictors for the combination of incident and persistent cases (Prevalent SDB compared with Prevalent NoSDB). Children with PrevalentSDB were more likely to be boys (OR=2.48, p=.006, CI (1.29-4.76)) and had a greater increase in BMI percentile change (OR 1.01, p=.034, 1.01-1.03) (Figure). Children with PrevalentSDB also had a 3.41 greater odds (p<.001, CI: 1.76-6.61) of developing obesity from Baseline to Follow-up in comparison with children with Prevalent NoSDB. Parent reported symptoms or other measured physiological variables on the night of the PSG were not predictive of future incident SDB or remission of SDB.

Figure.

Change in BMI Percentile from Baseline to Follow-up. Children with PrevalentSDB had greater increase in BMI Percentile than children with Prevalent NoSDB.

We examined the number of incident, remission and persistent cases from Baseline to Follow-up for higher cutpoints of RDI, such as 2, 5, and 10 events per hour. Unfortunately, the number of cases were too few to do any meaningful analysis with only 13, 3, and 1 case respectively for each of the above cutpoints.

Seventeen children had tonsillectomies between the time of their first and second polysomnographic evaluation. Eight children had PersistentNoSDB and two children had PersistentSDB, resulting in no change to SDB classification. Following tonsillectomy, 3 of the 17 children had incident SDB, and 4 had remission of SDB. Subsequent analyses showed that having a tonsillectomy had no influence on our conclusions.

Discussion

In this analysis, we found that in children between the ages of 6 and 11 years, the incidence of parent reported sleep symptoms over the ensuing 5 years is relatively low. Moreover, except for symptoms consistent with insomnia, most symptoms of sleep disorders remit. In addition, we found that not all children with polysomnographically determined SDB remit and that some children develop SDB during this interval. Male sex and obesity appear to be the most important risk factors for prevalent SDB in this age group. These data indicate that SDB does not necessarily resolve with the onset of adolescence in children and risks for SDB in adolescence are similar to those observed in adults.

The highest incidence rates we observed were for insomnia-related symptoms (DIMS) and excessive daytime sleepiness. These findings are consistent with the physiologic sleep phase delay occurring in adolescents and the potentially hectic social environment in which they live. In this age group, factors such as enhanced autonomy, and familial, educational and psychosocial stresses may contribute to poor sleep hygiene.^{23, 24} Interestingly, remission rates for EDS were higher than remission rates for DIMS. The explanation for these differences is not clear. However, the factors contributing to EDS and DIMS in 6-11 year old children are not the same as in 12-18 year old adolescents. Thus, whereas SDB may be an important contributing cause of EDS in younger children, insufficient sleep may be more relevant in the older age group. Similarly, the aforementioned social environment not only contributes to the higher incidence of DIMS, but also to a relatively low remission rate.

In contrast to DIMS and EDS, the incidence rates for snoring and witnessed apnea were much lower. It is difficult to contrast our findings with respect to snoring with prior studies because of differences in the age of the cohorts and length of follow-up. Nevertheless, our snoring incidence rate of 4.1% over an approximate 5 year interval is comparable with the 4.5% rate over a 3 year interval reported by Anuntaseree et al and the 4.3% rate observed by Chervin et al over a 4 year interval.^{7, 25} Incidence and remission rates for the symptom of witnessed apnea have not been previously reported. However, the prevalence rate observed at follow-up in the current study (2.0%) is similar to the findings of Sánchez-Armengol et al in a cohort of 12-16 year old children (3%),²⁶ but higher than observed by Joo et al in a group of 15-18 year old Korean high school students. ²⁷ It is possible that the older age of the cohort in the latter study accounts for this difference.

We observed a 10% incidence rate and a 70.8% remission rate for polysomnographically confirmed SDB in our study. There have been few prior studies documenting the incidence and remission rates of SDB in children, with most occurring in small cohorts recruited from clinical sources. Marcus et al initially studied a small group of children and found that 2 of 20 with primary snoring had evidence of SDB on polysomnography 1-3 years later.¹⁴ Similarly, Topol and Brooks found 1 of 13 children had evidence of SDB 3 years later.¹⁵ In contrast, Anuntaseree et al found 5 new cases of SDB in 61 children (9.1%) with habitual snoring 3 years previously.²⁵ Our study extends these limited observations to a much larger cohort studied after a long duration spanning the time period between preadolescence to adolescence. Our results indicate that there is a small but significant number of preadolescent children who will develop SDB as they become adolescents. Moreover, although it is generally expected that both snoring, and consequently SDB, should resolve over time in preadolescents in parallel with regression of tonsillar and adenoidal tissue, our finding that the remission rate is only 70.8% indicates that a considerable number of children with SDB will not spontaneously resolve. Given the increasing data that SDB in children adversely impacts neurocognition and behavior, and may increase the risk of hypertension and obesity, our results provide additional evidence that active intervention to treat SDB in preadolescents is indicated.

We observed that prevalence of SDB or no SDB was associated with sex and increasing weight. Boys as well as those children who had excessive weight gain were more likely to have SDB at the time of the 2^nd TuCASA examination. In adults, SDB also is more common in males and those who are obese. Thus, our findings support the hypothesis that factors important in the development of SDB in adolescents are similar to those operative in adults. Given the increasing prevalence of obesity in both children and adults in the general population, it is likely that SDB will be a more common occurrence in children, especially adolescents.

One limitation of our study is that we used only one RDI threshold to classify SDB. Unfortunately, there were too few cases for a meaningful analysis using other definitions. Although we acknowledge that other definitions of SDB have been used (e.g., 5 events/hour), an RDI ≥ 1 event per hour has been associated with significant physiologic and behavioral outcomes in a number of studies. ^{12, 20, 21, 28, 29} Therefore, we believe our analyses have important clinical implications.

In summary, our findings in a general population indicate that although symptoms of sleep disorders in preadolescents generally resolve over time, they will not remit in some children, and a significant number will develop them. Moreover, SDB in preadolescents may not spontaneously remit and a significant number will develop SDB when they are adolescents. Risk factors for SDB in adolescents are similar to that observed in adults. Given the potential for adverse consequences on health and the likelihood of an increasing prevalence of SDB, clinicians should be both vigilant in diagnosis and aggressive in treatment.

List of Abbreviations

BMI

Body Mass Index

CDC

United States Center for Disease Control

CI

Confidence Interval

DIMS

Difficulty Initiating and Maintaining Sleep

ECG

Electrocardiogram

EDS

Excessive Daytime Sleepiness

EEG

Electroencephalogram

IRB

Institutional Review Board

OR

Odds Ratio

OSA

Obstructive Sleep Apnea

RDI

Respiratory Disturbance Index

SDB

Sleep Disordered Breathing

SHQ

Sleep Habits Questionnaire

SN

Snoring

TuCASA

Tucson Children's Assessment of Sleep Apnea

TUSD

Tucson Unified School District

WITAP

Witnessed Apnea