No Relationship between Neurocognitive Functioning and Mild Sleep Disordered Breathing in a Community Sample of Children

Abstract

Study Objectives:

Our objective was to examine the relationship between sleep disordered breathing (SDB) and neurocognitive functioning in a large general-population sample of children who underwent a full-night polysomnogram and comprehensive neuropsychological testing.

Methods:

A population-based study of 571 school-aged children (6-12 years) underwent a 9-hour polysomnogram and a comprehensive neuropsychological battery.

Results:

No significant relationship was found between children with a mild apnea-hypopnea index (1 ≤ apnea-hypopnea index < 5) and any measure of neuropsychological functioning (intelligence, verbal and nonverbal reasoning ability, attention, executive functioning, memory, processing speed, and visual-motor skill). Partial correlations between apnea-hypopnea index and neuropsychological test scores and polynominal trend analysis were all nonsignificant.

Conclusions:

Children with mild SDB showed no significant neuropsychological impairment compared to children without SDB. This study suggests that children with mild SDB are not at risk for neuropsychological impairment and that the coexistence of mild SDB with any neuropsychological impairment should be considered comorbid and not causal. However, the association between neurobehavioral issues and children with mild SDB remains uncertain.

Citation:

Calhoun SL; Mayes SD; Vgontzas AN; Tsaoussoglou M; Shifflett LJ; Bixler EO. No relationship between neurocognitive functioning and mild sleep disordered breathing in a community sample of children. J Clin Sleep Med 2009;5(3):228-234.

Many children are affected by some form of sleep disordered breathing (SDB), a spectrum of sleep abnormalities ranging from primary snoring to upper airway resistance to obstructive sleep apnea. Reported prevalences range from 1% to 4% for obstructive sleep apnea^1–3 to 3% to 12% for habitual snoring.^1,4 Recently, Bixler et al.⁵ reported a prevalence of an apnea-hypopnea index (AHI) of at least 5 was 1.2% in The Penn State Children's Cohort. The literature has been mixed on the definition and significance of SDB in children and potential associations with neuropsychological functioning, behavior disturbances (i.e., inattention), and psychiatric status (i.e., attention-deficit/hyperactivity disorder [ADHD]). Our initial paper focused on the relationship between academic achievement and sleep problems.⁶ This paper will focus on the potential relationship between neuropsychological functioning and SDB.

Significant associations between subjective parent report of primary snoring and impaired neuropsychological test scores in children have been reported in some studies^7–9 but not in others.^7,10 A few studies that have compared clinically referred children with SDB to controls using a full polysomnogram and neurocognitive test scores^10–12 have reported inconsistent and contradictory results. Correlations between neuropsychological functioning and polysomnography indexes have provided little insight. For example, Kaemingk et al.¹³ reported a modest correlation between AHI and neuropsychological outcomes, but others have not.^12,14 In fact, some children with mild SDB demonstrate poorer neuropsychological outcomes than do children with severe SDB.^10,12 Three treatment studies using an untreated control group and polysomnography data have reported improvement in attention test scores for children with SDB following adenotonsillectomy.^15–17 A single study conducted by Friedman et al.¹⁵ demonstrated significant improvement in cognitive test scores after adenotonsillectomy in children with mild to moderate SDB. Furthermore, 2 general-population studies^13,18 using polysomnography data and neuropsychological testing have reported nonsignificant findings overall, although a pattern of scores in these studies suggests that children with severe SDB performed less well than do children with mild SDB.

Interpretation and generalizability from previous studies is limited by many factors. Generally, small sample sizes that are not necessarily representative of the general population, lack of control groups, heterogeneity in terms of severity of the disorder, heavy reliance on parent report of sleep problems, referral bias, inconsistent measurement tools, and limited statistical treatment are some of the factors that play a role in the inconsistent and contradictory results that have been reported in previous studies. The present study attempts to overcome most of these limitations by determining the relationship between SDB and neuropsychological functioning in a large general-population sample of children (The Penn State Children's Cohort) who underwent a 9-hour polysomnogram and a comprehensive neuropsychological test battery measuring intelligence, verbal and nonverbal reasoning, processing speed, executive functioning, attention, memory, and visual-motor skill as part of a study for the assessment of the prevalence of SDB in children. To date, this is the largest population-based study using full polysomnography and a comprehensive neuropsychological battery that has been conducted in school-aged children.

METHODS

Sample

This study was designed in 2 phases, with the first phase designed for collecting general information from the parents about their child's sleep and behavior patterns. In the first phase, elementary schools (kindergarten through grade 5) were selected each year so that approximately 1500 students were enrolled. A screening questionnaire based on the survey published by Ali et al.,¹⁹ validated to identify children at high risk for having SDB, was sent home to parents of every student in these school districts (N = 7312) and had a 78.5% response rate. In the second phase of this study, 200 children were selected each year from the questionnaires that were returned that year. Using a stratification of grade, sex, and risk for having SDB, we randomly selected children from each stratum to maintain representativeness of the original sample. Seven hundred children completed phase 2, for a response rate of 70%. We contrasted the subjects who completed the polysomnography recordings with those who were selected and did not complete the phase 1 questionnaire. There were no significant differences between the 2 groups on grade, sex, and risk for having SDB. Children who completed the majority of the neuropsychological tests and had a full-scale IQ of at least 80 were included in the study. Children diagnosed with medical problems (26.1% allergies, 9.6% asthma, 1.2% juvenile diabetes, < 1% epilepsy), mental health disorders (7.2% ADHD, 1.7% depression/anxiety, 0.8% autism), or a learning disability (9.1%) were not excluded from the study so that the sample is representative of the general population. Thus, our final sample for this study consisted of 571 children. This study was approved by the Institutional Review Board of Penn State College of Medicine. Informed consent was obtained from parents of all participants, and assent was obtained from all children prior to participation.

Neuropsychological Assessment

All children underwent a 2.5-hour neuropsychological evaluation prior to their overnight stay in the sleep laboratory at approximately the same time each afternoon. The standardized tests were administered individually to each child by a trained psychometrist over 1 session. Test scores were converted to age-referenced standard scores with a mean of 100 and an SD of 15, based upon published normative data. Tests in the neuropsychological battery were chosen because they measure intelligence, and key neuropsychological functions including attention, executive functioning, memory, processing speed, mental flexibility, verbal fluency, and visual-motor skill.

Wechsler Abbreviated Scales of Intelligence

The Wechsler Abbreviated Scales of Intelligence consists of 4 subtests corresponding to the same subtests on the WISC-IV (Block Design, Matrix Reasoning, Vocabulary, and Similarities). The WASI correlates 0.87 with WISC-III Full Scale IQ in the general population²⁰ and 0.89 with the Kaufman Brief Intelligence Test in a psychiatric sample.²¹

Gordon Diagnostic System

The Gordon Diagnostic System Vigilance and Distractibility subtests²² were administered. The Gordon Diagnostic System is a continuous performance test that is well established as a measure of attention. Numerous studies have demonstrated significant agreement between Gordon Diagnostic System subtest scores and other measures of attention, including behavior rating scales, standardized observations, performance tests, and diagnoses of ADHD.²³ The Gordon Diagnostic System identified referred children with and without ADHD with 88% accuracy in a study by Mayes and Calhoun.²⁴

Digit Span

The WISC-III Digit Span requires the child to repeat numbers after the evaluator in forward and reverse order. Digit Span has been used as a measure of attention and working memory in several neuropsychological studies.²⁵ In studies of children with ADHD, Digit Span is the lowest of the WISC-III verbal subtest mean scores,^24,26,27 and children with ADHD score significantly lower than comparison children on Digit Span. ^24,28

Developmental Test of Visual-Motor Integration

The Developmental Test of Visual-Motor Integration²⁹ is an untimed test on which the child copies geometric forms with a pencil. The Developmental Test of Visual-Motor Integration has validity as a measure of dysgraphia or difficulty with handwriting.³⁰ Children with neurologic disorders, such as ADHD and autism, earn low scores on the Developmental Test of Visual-Motor Integration relative to other abilities. ³¹

Coding

The WISC-III Coding subtest requires the child to copy geometric forms with a pencil. The Coding subtest is timed and gives bonus points for speedy and accurate performance. Coding has demonstrated sensitivity to neuropsychological impairment and is low in children with neurologic disorders such as ADHD, autism, and learning disability.^{24,27,28,31–33}

Animal Naming Test

The Animal Naming Test requires the child to name as many animals as possible in 1 minute. The test was originally developed for adult aphasia evaluations³⁴ and was later normed on children.³⁵ The test is widely used as a measure of verbal fluency (specifically, retrieval from semantic memory) in child neuropsychological assessment batteries³⁶ and in research studies.³⁷

California Verbal Learning Test

The California Verbal Learning Test³⁸ measures verbal learning and memory and is sensitive to neurologic impairment³⁹ and severity of brain injury.⁴⁰ Children are asked to recall as many words as possible from a shopping list of 15 fruits, clothes, and toys read by the evaluator 5 times.

Stroop Color and Word Test Children's Version

The Stroop Color and Word Test-Children's Version is a measure of executive functioning.⁴¹ It is most often described as measuring ability to shift cognitive inhibition and ability to inhibit an overlearned dominant response in favor of an unusual one.⁴² Several hundred studies have been published on the Stroop because of its sensitivity to individual differences in response inhibition.⁴³

Wisconsin Card Sorting Test-64 Card Version

The Wisconsin Card Sorting Test-64 Card Version (WCST-64)⁴⁴ is a shortened version of the well-regarded 128-card version that measures executive functioning. The task requirements for the WCST-64 are the same as those for the longer version, and the normative and technical features of the WCST-64 are derived from the same sample as the WCST. The WCST has been shown to be sensitive to frontal-lobe dysfunction in children.⁴⁵

Polysomnography

All children underwent a 9-hour polysomnogram with a parent present in a sound-attenuated, light- and temperature-controlled room in our General Clinical Research Center. Children's bedtime and wake time approximated their typical sleep times (average 2100-2200 to 0600-0700). Each child was monitored with an infrared video and a computerized system (24 analog channel and 10 dc channel TS amplifier using Gamma software, Grass Telefactor, Inc.), including 4 channels of electroencephalogram, 2 channels of bilateral electroculogram, and chin and anterior tibial electromyogram. Respiration was assessed throughout the night by use of a thermocouple at nose and mouth (model TCT R, Grass Telefactor, Inc., West Warwick RI), nasal pressure (MP 45-871 ± 2 cm H₂O, Validyne Engineering Corp, Northridge CA), and piezoelectric thoracic and abdominal respiratory-effort belts (model 1312, Sleepmate, Midlothian VA). We obtained an objective estimate of snoring during the polysomnogram by monitoring breathing sounds with a microphone attached to the throat (model 1250, Sleepmate Technologies) as well as a separate room microphone. All-night hemoglobin oxygen saturation (SaO₂) was obtained by pulse oximeter (model 8800, Nonin Medical, Plymouth MN) attached to the finger. A single-channel electrocardiogram was also recorded. All of the polysomnography records were double scored in accordance with the American Thoracic Society standards for cardiopulmonary sleep studies in children.⁴⁶ Results from the overnight polysomnography evaluation were used to classify the children as having or not having SDB. Obstructive apnea was defined as a cessation of airflow with a minimum duration of 5 seconds and an out-of-phase strain-gauge movement. A hypopneic event was defined as a reduction of airflow of approximately 50% with an associated decrease in SaO₂ of at least 3% or an associated breathing-related arousal. Based on these data, an apnea-hypopnea index (AHI) was calculated ([apnea + hypopnea]/hours of sleep). Central apneas were not included in the AHI calculation. We further defined individuals “without SDB” if AHI was less than 1, as “mild SDB” if AHI was at least 1 but less than 5, and as “moderate SDB” if AHI was 5 or greater. The scoring of snoring was based on the number of snoring related arousals counted by the technician.

Data Analyses

For analysis purposes, subjects were separated into the following categories: 3 groups using AHI (a conventional measure of SDB): AHI less than 1, AHI of at least 1 but less than 5, and AHI of 5 or greater and 4 groups using no snoring and SDB less than 1, primary snoring, snoring with no SDB, and snoring and SDB greater than 1. To assess the differences between the groups, analysis of covariance was performed using race as a covariate. Group differences were also compared by effect size, as measured by partial eta squared (η²_p) and reported as f ⁴⁷, and power analyses was performed. Polynomial trend analysis was conducted to determine if any significant trends in neuropsychological test scores existed between the 3 AHI groups. Partial Pearson correlations were used to calculate the linear relationship between actual AHI, apnea index (only obstructive apnea), minimum degree of SaO₂ (the lowest of either non-rapid eye movement [REM] or REM oxygen desaturation), breathing-related arousals, and neuropsychological scores. Two-tailed tests were used. A probability level of 0.05 was chosen for statistical significance.

RESULTS

In this sample, we identified 6 individuals (1%) as having moderate SDB, 152 as mild SDB (26.6%), and 413 without SDB (72.3%). The major characteristics of the study sample are presented in Table 1 according to SDB status, defined by AHI less than 1, AHI 1 to less than 5, and AHI of 5 or greater. Significant differences in all groups were found between children in AHI and race, but not based on age, sex, or parent occupation (socioeconomic status). No significant differences or linear trends were found between the 3 AHI groups on any of the neuropsychological measures (intelligence, verbal and nonverbal reasoning ability, attention, executive functioning, memory, processing speed, and visual-motor skill), and effect sizes were small (f ≤ 0.10). The observed power ranged from limited (0.11) to adequate (0.80). All correlations between neuropsychological test scores and continuous AHI were nonsignificant at 0.05, and explained variance was small (r² ≤ 0.01) (Table 2). Additional analyses were conducted to investigate the association between breathing-related arousals and neuropsychological test scores, minimum degree of SaO₂ and test scores, as well as apnea index and test scores. Similar to the AHI results, all correlations between additional variables and neuropsychological test scores were nonsignificant at 0.05, and explained variance was small (r² < 0.03). No significant differences at 0.05 were found between the 4 snoring or SDB groups on any neuropsychological measure except on nonverbal IQ (Table 3). Although the p value for this measure is significant at 0.01, the effect size is small (f ≤ 0.14), and the power is adequate (0.80).

Table 1.

Demographic Data of Study Population

	AHI<1	≥ 1 AHI<5	AHI ≥ 5	p Value
Age, y	8.7 ± 1.6	8.7 ± 1.7	9.4 ± 0.98	0.44
Male	52	58	29	0.18
Non-Caucasian	20	31	0	0.01
Parents are not professionals	48	44	0	0.08

Data are presented as percentage, except age, which is mean ± SD. AHI refers to apnea-hypopnea index.

Table 2.

Mean Neuropsychological Standard Scores for Children in the 3 AHI Groups Covarying for Race, Correlations Between Test Scores and Continuous AHI, and Trend Analysis

	AHI<1 n = 413	≥ 1 AHI<5 n = 152	AHI ≥ 5 n = 6	ANCOVA		Trend p value	Pearson ra
				p value	f
Wechsler Abbreviated Scales of Intelligence
Verbal	106 ± 12.5	106 ± 12.9	113 ± 8.3	0.36	0.06	0.47	0.11
Nonverbal	107 ± 15.1	104 ± 14.3	116 ± 13.8	0.04b	0.10	0.91	0.01
Full-scale	107 ± 12.7	106 ± 13.2	116 ± 9.5	0.21	0.08	0.63	0.08
Attention and executive functioning
Digit span	103 ± 14.2	102 ± 15.5	101 ± 13.9	0.47	0.05	0.75	0.01
Gordon Diagnostic System
Vigilance
Correct	97 ± 17.5	95 ± 20.7	106 ± 10.6	0.30	0.06	0.55	0.01
Errors	96 ± 22.6	93 ± 27.6	102 ± 10.9	0.53	0.04	0.45	0.03
Distractibility
Correct	99 ± 18.8	98 ± 20.2	109 ± 14.3	0.45	0.05	0.51	0.03
Errors	92 ± 27.8	91 ± 29.1	101 ± 9.3	0.70	0.03	0.42	0.02
WCST total errors	100 ± 19.1	100 ± 19.4	113 ± 1.2	0.68	0.05	0.40	0.07
Stroop
Color word	86 ± 11.0	89 ± 10.6	77 ± 15.9	0.30	0.09	0.06	−0.12
Interference	107 ± 11.3	109 ± 13.5	106 ± 8.5	0.38	0.08	0.78	0.07
Processing speed
Coding	102 ± 16.4	98 ± 14.1	106 ± 17.5	0.06	0.10	0.46	0.02
Symbol search	107 ± 15.9	110 ± 17.1	117 ± 10.6	0.18	0.10	0.55	0.01
Memory
California Verbal Learning Test recall
Immediate	105 ± 14.5	105 ± 16.2	114 ± 14.5	0.36	0.06	0.21	−0.04
Delayed	108 ± 13.9	106 ± 16.3	104 ± 5.3	0.49	0.06	0.84	0.03
Animal naming	112 ± 21.1	109 ± 22.1	116 ± 20.7	0.58	0.04	0.91	−0.08
Visual-motor skill
VMI	96 ± 13.5	93 ± 13.0	102 ± 20.4	0.07	0.10	0.45	0.00

Values represent mean ± SD. Test scores are standard scores with a mean of 100 and an SD of 15. ANCOVA refers to analysis of covariance; WCST, Wisconsin Card Sorting Test; VMI, Developmental Test of Visual-Motor Integration.

^aAll Pearson correlation coefficients are nonsignificant at 0.05.

^bApnea-hypopnea index (AHI) ≥ 5 highest test score.

Table 3.

Mean Neuropsychological Standard Scores for the Snoring Groups

	AHI<1		AHI ≥ 1		ANCOVA
	No snoring n = 329	Snoring n = 84	No snoring n = 96	Snoring n = 62	p value	f
Wechsler Abbreviated Scales of Intelligence
Verbal	106 ± 13.0	104 ± 10.2	108 ± 13.2	105 ± 12.1	0.32	0.08
Nonverbal	108 ± 15.2	106 ± 14.5	107 ± 14.5	100 ± 13.5	0.01a	0.14
Full-scale	108 ± 12.9	106 ± 11.6	108 ± 13.6	103 ± 12.0	0.08	0.11
Attention and executive functioning
Digit span	103 ± 14.4	103 ± 13.7	101 ± 15.3	102 ± 15.6	0.60	0.05
Gordon Diagnostic System
Vigilance
Correct	97 ± 17.3	96 ± 18.2	97 ± 18.3	94 ± 23.6	0.72	0.06
Errors	96 ± 22.1	96 ± 24.2	93 ± 29.0	95 ± 24.4	0.88	0.03
Distractibility
Correct	100 ± 18.3	97 ± 20.6	101 ± 18.1	96 ± 22.6	0.41	0.07
Errors	92 ± 28.2	93 ± 26.2	94 ± 26.0	87 ± 31.9	0.42	0.07
WCST errors	101 ± 19.2	97 ± 18.7	101 ± 21.5	99 ± 15.8	0.73	0.07
Stroop
Color word	88 ± 10.5	91 ± 12.7	88 ± 10.9	90 ± 10.9	0.52	0.09
Interference	107 ± 11.6	106 ± 10.2	111 ± 14.3	107 ± 11.9	0.40	0.10
Processing speed
Coding	102 ± 16.5	101 ± 16.1	99 ± 14.4	97 ± 14.1	0.16	0.09
Symbol search	107 ± 16.1	108 ± 15.1	110 ± 17.2	111 ± 16.8	0.16	0.13
Memory
California Verbal Learning Test, recall
Immediate	105 ± 14.3	105 ± 15.4	105 ± 17.1	106 ± 14.7	0.85	0.03
Delayed	108 ± 13.7	110 ± 14.7	105 ± 15.1	107 ± 17.7	0.52	0.08
Animal naming	112 ± 21.2	111 ± 20.7	109 ± 22.3	111 ± 21.8	0.76	0.04
Visual-motor skill
VMI	97 ± 13.5	95 ± 13.6	94 ± 13.3	93 ± 13.6	0.33	0.08

Values represent mean ± SD. Test scores are standard scores with a mean of 100 and an SD of 15. ANCOVA refers to analysis of covariance; WCST, Wisconsin Card Sorting Test; VMI, Developmental Test of Visual-Motor Integration.

^ap ≤ 0.05; no snoring apnea-hypopnea index (AHI)<1 versus snoring AHI ≥ 1.

DISCUSSION

Our general-population study of 571 young children with full overnight polysomnography showed no relationship between mild sleep apnea and any measure of neuropsychological functioning, including IQ, verbal and nonverbal reasoning, attention, executive functioning, memory, processing speed, and visual-motor skill (Table 2).

Although many previous studies have reported positive findings between children with moderate to severe SDB and neuropsychological functioning,^9–12,16 our study suggests that children between the ages of 6 and 12 who have mild SDB are not susceptible to neuropsychological deficits. The strongest support for the position that SDB causes neurocognitive and behavioral deficits is found in a study¹⁵ demonstrating improvement in cognitive test scores following adenotonsillectomy in children with mild to moderate SDB and in 3 additional studies^1,16,17 demonstrating improvement in behavior and attention after adenotonsillectomy. It is not possible, however, to conduct blind, placebo-controlled studies, and, therefore, it is not known if the reported improvement in attention and cognition in these studies before and after adenotonsillectomy is the result of placebo effects and/or regression to the mean, instead of treatment effects. It should be noted that some children who had adenotonsillectomy prior to our study could have been exposed to SDB and have permanent cognitive consequences but no longer have SDB. The most parsimonious explanation for the differences between our study results and that of others is the lack of children with moderate SDB in our sample (n = 6) that limited our power. It is certainly reasonable to assume that the more severe the SDB, the greater the impact on neuropsychological functioning. Further considerations for inconsistencies across studies include methodologic issues (i.e., small sample size, referral bias, lack of control for number of comparisons, and no control group) that can limit generalizability and make interpretation and comparison difficult.

Although our analyses suggested that neurocognitive impairment is not associated with mild childhood SDB, we did not investigate the potential association between neurobehavioral issues and SDB. Several studies have reported positive findings between parent-reported behavior and emotional problems (i.e., attention deficit, “externalizing” behaviors, and anxiety/depression) and SDB in children.^12,14,48 Because we collected parent rating scales on all children in our cohort, we plan to explore this potential association in our next study.

In addition to investigating the association of SDB with neuropsychological functioning, some studies^7–9 have assessed the association between subjectively reported snoring and neuropsychological functioning. No studies to date have reported on the association of objectively measured snoring and neuropsychological functioning. In our study, using the same cohort, snoring was based on the objective evaluation obtained during the polysomnography (microphone applied to the child's throat during sleep as well as room microphone). We found that mild SDB and objectively measured snoring were not associated with performance on any measure of neuropsychological functioning (IQ, verbal reasoning ability, attention, executive functioning, memory, processing speed, or visual-motor skill) except nonverbal reasoning ability (Table 3). Although this finding is statistically significant at 0.01, the effect size was small.

The exact mechanisms by which SDB and neuropsychological functioning may be associated remain unknown. Recently, however, several researchers have suggested that the mechanisms are most likely multifactorial.^49–52 For example, 1 study found that children with severe SDB and low IQ levels (≤ 85) had significantly higher levels of C-reactive protein (a serum marker of inflammation) than did children with severe SDB and normal IQ levels.⁴⁹ This suggests that low IQ level may be associated more strongly with systemic inflammation than with SDB per se. Perhaps children present with different types of severe SDB: those with inflammation and associated neuropsychological impairment (subtype 1) and those without inflammation and neuropsychological impairment (subtype 2). Differences between subtypes in underlying medical conditions (e.g., visceral adiposity, metabolic syndrome) or protective mechanisms (e.g., genetics) may make some children more or less resilient to the effects of SDB and, in turn, neuropsychological impairment. Robust associations between SDB and neuropsychological functioning may result from the additive effect of genetic, metabolic, and environmental risk factors.

From a clinical standpoint, professionals who evaluate and treat children with SDB should be cognizant of comorbid risk factors associated with this disorder. Our study suggests that children aged 6 to 12 years who have mild SDB are likely not to be at risk for neuropsychological impairment, and our study results suggest that the coexistence of mild SDB with any neuropsychological impairment should be considered comorbid and not causal. Certainly, there may be individual children for whom mild SDB affects neuropsychological functioning, but, for a large group of community children, mild SDB was not significantly related to neuropsychological functioning in our study. Future research should focus primarily on those children with moderate to severe SDB and the mechanisms associated with neuropsychological impairment. Ultimately, we believe it will be the convergence of neuropsychological and neurobiologic data that clarifies the morbidity associated with SDB.

DISCLOSURE STATEMENT

This was not an industry supported study. The authors have indicated no financial conflicts of interest.