Sleep-Disordered Breathing and Neurobehavioral Outcomes

Commentary

In children, obstructive sleep apnea (OSA) verified by polysomnography represents a serious disorder characterized by intermittent cessation of breathing accompanied by gas exchange abnormalities, snoring, gasping, restlessness, and frequent awakenings due to repeated closure or narrowing of the upper airway. If OSA is unrecognized or left untreated, significant morbidity can result, including failure to thrive and severe cardiovascular problems, along with daytime behavioral issues, including disruption in attention and learning, hyperactivity, aggression, and antisocial behavior.¹ Obstructive sleep apnea is relatively common among children, with a prevalence of approximately 2%.² Adenotonsillar hypertrophy accounts for most cases of OSA in children, although children with craniofacial abnormalities, morbid obesity, and neurological disorders affecting upper airway patency also may develop OSA. Adenotonsillectomy results in 75% to 100% polysomnographic-defined resolution of OSA,² but success with this procedure may become more limited as children become increasingly overweight.³

Frank obstructive sleep apnea is not the only condition that can affect the upper airway during sleep, and sleep-disordered breathing (SDB) also can manifest in more subtle forms. During the past several years, increasing evidence suggests that milder SDB disturbances (eg, snoring, present in 10%-14% of children), even when accompanied by polysomnographic findings well below conventional thresholds established for OSA, can be detrimental to the day time cognitive, emotional, and behavioral well-being of children.⁴ Although deficits in a range of cognitive and emotional areas have been linked with SDB, by far the clearest associations occur with behaviors that characterize attention-deficit/hyperactivity disorder (ADHD; eg, decrements in sustained and selective attention, conduct problems, and hyperactivity).^{5 - 6} Effective treatment of SDB through adenotonsillectomy appears to alleviate many of these symptoms.^{7 - 8}

This improvement in neurocognitive and neurobehavioral functioning in children following adenotonsillectomy provides important evidence that SDB is linked to adverse behavioral and cognitive outcomes that are amenable to treatment. On the other hand, in studies that included polysomnographic measurement before and after adenotonsillectomy, no clear relationship was found for behavioral or cognitive improvements with polysomnographic measures of SDB.^{7 - 8} In addition, sleep studies in children with ADHD suggest that milder forms of SDB, rather than more severe forms, may be more clearly linked with ADHD-type behaviors.⁹

One possibility for the lack of expected linkage between polysomnographic findings and objective cognitive test performance and parental ratings of behavior may relate to limitations inherent in commonly used polysomnographic measures that are potentially less sensitive to subtle underlying SDB characteristics subjectively recognized by health professionals. If this is so, there appears to be no clear, objective polysomnographic-related marker on the continuum between normal breathing and OSA for which neurobehavioral pathology is expected and treatment is clearly warranted. Another possibility contributing to this lack of a clear cause-and-effect relationship is that the association of SDB with neurobehavioral outcomes is far more complex and multifactorial. Predictive models identifying children at risk for SDB-associated behavioral sequelae can be developed only with a clear understanding of these potential moderating risk factors.

An article in the April issue of the Archives of Pediatrics & Adolescent Medicine, a theme issue on sleep research, effectively underscores this issue. Owens and colleagues¹⁰ report findings from a retrospective study of 235 children undergoing overnight polysomnography for symptoms of SDB. These children were found to demonstrate higher than expected rates of weight problems, behavioral difficulties, diagnoses of ADHD, and comorbid sleep problems. The researchers compared parental report of behavioral difficulties on the Child Behavior Checklist with polysomnographic measures, along with special attention to other possible risk factors for poor neurobehavioral outcomes in the children. Parental ratings on the Child Behavior Checklist did not relate to SDB severity, but significant relationships were found for increased Child Behavior Checklist concerns and other potential risk factors, including increased weight, shortness of sleep duration, and the presence of comorbid sleep disorders. These results emphasize the need to attend to other possible contributing factors that, if ameliorated, could contribute to a positive posttreatment course for children with SDB.

The findings by Owens and colleagues¹⁰ are consistent with related research findings. For example, overweight children have been found to have higher rates of behavioral disruption, in particular internalizing problems, although historically this finding has been found predominantly in overweight girls.¹¹ Sleep disruptions besides SDB, including abbreviated sleep times and other sleep problems, also have been associated with externalizing symptoms and ADHD in children without clearly evident SDB.¹²

Two other articles in the April issue of Archives of Pediatrics & Adolescent Medicine also are relevant to the findings of Owens and colleagues¹⁰ and further point out the complexity in the relationship of SDB with behavioral disturbance. In the study by Taveras and colleagues,¹³ infants who slept less than 12 hours per day up to age 2 years were found to be at risk for being overweight and having higher BMI z scores at age 3 years. This study¹³ suggests that shortened sleep durations early in life may actually predict the later appearance of weight problems similar to those seen in the study by Owens and colleagues.¹⁰ In the study by Gregory and colleagues,¹⁴ children whose parents reported that they slept a shorter time than other children on any of a series of longitudinal ratings across preschool and school years had an increased risk of self-reported internalizing or aggressive behaviors measured in adulthood. Consistent with Owens and colleagues,¹⁰ these studies^{13 - 14} again highlight the importance of considering other sleep disruption–related factors besides SDB.

If the findings of Owens et al¹⁰ emphasize potential contributors other than SDB to neurobehavioral outcomes, the seeming insensitivity of standard polysomnographic assessments, at least with reference to these same outcomes, remains a concern. Newer, objective SDB measures that may better clarify relationships to outcomes have been proposed.¹⁵ In addition, potential contributory risk factors such as those noted by Owens and colleagues¹⁰ must be carefully evaluated and examined along with other potential, newer markers linking neurobehavioral deficits to specific risk factors in potentially vulnerable child populations. For example, serotonin levels may relate both to sleep disruption as well as aggressive and hyperactive behaviors¹⁶ often seen in children with SDB. Genetic factors may contribute to individual susceptibility to certain types of sleep disruption. Vulnerability that may be associated with age and sex,⁶ as well as history of prematurity or black heritage,¹⁷ also must be considered in building relative risk models that can more effectively target interventions and provide better understanding of underlying causal mechanisms.

In conclusion, more complex evidence that links sleep disorders to adverse neurobehavioral outcomes is rapidly increasing. The articles published in the sleep research theme issue of the Archives of Pediatrics & Adolescent Medicine highlight the serious cognitive, behavioral, and family-related morbidity accompanying sleep disruption at the earliest stages of life and support the critical need for continued research in this area. The goal in characterizing these potential risk factors should be to identify early markers and at-risk profiles that can be used to make effective treatment decisions.