Pediatric obstructive sleep apnea affects a large number of children and has multiple end-organ sequelae. Although many of these have been demonstrated to be reversible, the effects on some of the organ systems, including the brain, have not shown easy reversibility. Progress in this area has been hampered by lack of a preclinical model to study the disease. Therefore, perioperative and sleep physicians are tasked with making a number of difficult decisions, including optimal surgical timing to prevent disease evolution, but also to keep the perioperative morbidity in a safe range for these patients.
Key points
- •
Pediatric obstructive sleep apnea affects up to 7.5% of children and has significant, long-lasting effects on memory, learning, and other executive functions.
- •
Adenotonsillectomy, which is first-line therapy, has equivocal results on neurocognitive function.
- •
There are certain subgroups, including low socioeconomic status and disadvantaged minorities, who seem to carry a higher morbidity from the illness and surgery.
Introduction
Pediatric obstructive sleep-disordered breathing (oSDB) affects up to 7.5% of the pediatric population. Pediatric obstructive sleep apnea (OSA) is characterized by intermittent hypoxia and hypercapnia, sleep fragmentations, frequent arousals, and circadian rhythm disturbances. Pediatric OSA can have a variety of end-organ manifestations with effects on the heart, lungs, brain, gut microbiome, and genitourinary systems ( Fig. 1 ). Clinical manifestations of pediatric oSDB are on a continuum with symptoms ranging from nasal turbulence, to snoring, to obstructive apnea. The population typically affected are children between 2 and 14 years of age. Adolescent OSA is considered a separate phenotype and has variable progression into adult OSA. A main feature of pediatric OSA is the difference in both risk factors and disease expression and response to treatment when compared with adult OSA, often resulting in missed or delayed diagnosis ( Tables 1–3 ).
Adult | Pediatric | |
---|---|---|
Gender predilection | Male >> Female | Male = Female |
History |
|
|
Physical examination |
|
|
Treatment |
|
|
Impact on growth |
|
|
Adult | Pediatric | |
---|---|---|
OSA Severity by AHI | ||
No OSA | 0–5 | 0 |
Mild | 6–20 | 1–5 |
Moderate | 21–39 | 6–9 |
Severe | >40 | >10 |
Adult | Pediatric | |
---|---|---|
Apnea definition | Cessation of airflow for at least 2 respiratory cycles | Cessation of airflow for at least 10 s |
Cortical arousal | Low frequency of cortical arousals (arousal following obstructive event) | Generally, apneic events are followed by a cortical arousal |
Sleep state of OSA | Apnea and hypopnea events generally occur in REM phase | Apnea and hypopnea events Occur in both REM and non-REM phases |
Sleep architecture | Normal sleep architecture | Fragmented sleep |
Transcutaneous CO 2 | Pa co 2 often elevated | Often normal |
Type of obstruction | Persistent and partial obstructions | Cyclical partial or complete obstruction of the upper airway |
Severity scale of OSA | Vastly different between adult and pediatric |
Although obesity has clearly been identified as a causative factor in pediatric OSA, , 2 parts of the clinical phenotype exist, with a significant number of children being normal or even underweight. Emerging data suggest that weight may be used to stratify pediatric OSA severity, as patient weight has been found to correlate with the number of sites of airway obstruction. This finding may, in part, explain the lack of resolution of OSA symptomatology following adenotonsillectomy surgery in a number of symptomatic children. More critically, 2 recent large multicenter studies demonstrated the persistence of neurocognitive deficits following surgery. , Thus, early identification and treatment of oSDB is critical, as neurocognitive deficits at least in the short term may be irreversible once developed. The objective of this article was to summarize the neurocognitive expression of pediatric OSA, enumerate comorbid risk factors, discuss surgical timing/risk factors, and describe perioperative complications, all of which contribute to decision making with this common disorder.
Neurocognitive expression of obstructive sleep apnea in children
There is a summative, temporal net effect of these physiologic perturbations on nearly every end-organ system, including the brain. Multiple areas of the brain are affected by persistent pediatric OSA. Because sleep is an essential piece of mammalian memory consolidation, , learning, memory, and attention deficits are seen in the clinical phenotype of idiopathic sleep behavior disorders. Learning, memory, and attention are largely hippocampally mediated in early life, pointing to hippocampal damage as being seminal to the symptomatology seen from OSA in the developing brain. One study demonstrated gray matter volume reductions in the left hippocampus/entorhinal cortex, left posterior parietal cortex, and left superior frontal gyrus in children with OSA. These findings were similarly observed in adult patients with OSA, albeit with greater changes in the temporal gyri, insula, and posterior cingulate cortex. All of these changes affect learning, memory, and other areas that govern critical thinking and outcomes processing. In addition, gray matter loss was noted within sites involving motor regulation of the upper airway as well as those responsible for cognitive function. , When taken together, effects on these brain structures result in a number of clinically observed changes.
Clinical manifestations of obstructive sleep apnea
The neurocognitive effects from pediatric OSA can be broadly classified into those affecting executive function, learning and memory, or behavior. Extensive discussion on the effects of pediatric OSA on the brain can be found in our recent review article.
Changes in executive functioning
Executive functioning encompasses a set of cognitive processes that includes a number of basic and higher-order functions. , Executive functioning is primarily measured by testing working memory, cognitive flexibility, and inhibitory control. Because arousal from sleep is the primary mechanism by which cognitive function is preserved in response to an obstructive event, it seems logical that arousals from sleep would predict neurobehavioral morbidity. This has been demonstrated not to be the case. Furthermore, adenotonsillectomy has not been shown to facilitate executive function recovery in children with mild-moderate sleep apnea.
Behavioral changes
Attention deficit/hyperactivity disorder (ADHD) has been reported at far higher rates in children with OSA. Even very low apnea hypopnea index (AHI) rates have been correlated with ADHD. There is also a significant increase in ADHD symptomatology with increased degree of hypoxia. However, to date there have been no large-scale trials to elucidate the link between ADHD and OSA, and therefore much remains unknown. However, this has led to multiple clinicians proposing to screen children with ADHD for sleep-disordered breathing phenotypes before initiating pharmacologic intervention. Furthermore, early sleep deprivation is associated with a higher risk of developing ADHD later on. MRI has demonstrated deficits in areas of the brain involved with emotional processing. There have also been significant clinical functional deficits demonstrated within multiple domains, including compulsion, aggression, and somatization. Adenotonsillectomy results in significant improvement in behavioral outcomes, , suggesting there is a strong link between pediatric OSA and ADHD. However, many of the issues related to pediatric OSA translate into academic difficulties for a number of children compared with controls. However, because of the behavioral and academic difficulties, a number of children with OSA are theorized to be incorrectly diagnosed with ADHD.
Learning and memory deficits
Learning and memory deficits have been well characterized in pediatric OSA. These deficits are diverse and involve spatial memory, working memory, verbal comprehension, and reasoning. These deficits affect both acquisition and retention of newly learned material , and are worse in children with Trisomy 21. The question that has eluded researchers for some time is the reversibility of these changes after therapy. Some measures of neuropsychological functioning seem to improve, although specific improvements are difficult to demonstrate even with large cohorts.
Preclinical models of pediatric obstructive sleep apnea
There are 2 preclinical rodent models of OSA that have been used. One model is the intermittent hypoxia model, which uses 14 days of rapid cycling oxygen concentrations of between 10% and 21%. The neurobehavioral outcome was deficits in the Water Maze Task, which involves hippocampally based decision making. The second model is the intermittent tracheal occlusion model which uses surgery for implantation of a silicone tube, which is intermittently inflated to occlude the trachea. This model has not been neurobehaviorally validated. To date, however, there is no preclinical model for pediatric OSA.
Comorbid risk factors for pediatric obstructive sleep apnea
Obesity has been demonstrated to be a significant risk factor for the development of pediatric OSA. Obesity also has a significant effect on worsening the neurocognitive profile of OSA, which has been attributed to sleep fragmentation and poorer sleep quality. Obesity also correlates with worse outcomes following adenotonsillectomy and a higher risk of respiratory complications on the first postoperative night. Although most patients achieve a reduction in oSDB symptoms following adenotonsillectomy, obese patients are noted to have a significantly greater incidence of residual disease. In addition, children with very high body mass index ( z -scores >3) appear to benefit the least from adenotonsillectomy. This subgroup may benefit from preoperative drug-induced sleep endoscopy to assess regions of obstruction, assess response to adenotonsillectomy, and identify additional areas of airway obstruction. ,
Rhinosinusitis
A number of pediatric patients present to the ear, nose, and throat (ENT) surgeon with chronic rhinosinusitis concomitant with sleep-disordered breathing. The argument can be made in these children, that with a low impact on quality of life, an initial course of intranasal steroids can be used to assess symptomatology. Intranasal steroids can also alleviate comorbid allergies and reduce mouth breathing. African American children with rhinosinusitis seem to have an increased risk of OSA compared with white children.
Asthma
A number of children with OSA also have coexisting asthma; , however, the presence of asthma seems to decrease the likelihood of concomitant OSA. In children with both disorders, there are a number of other abnormalities in pulmonary and chemoreceptor function. Given the overlap, it is important for the clinician to identify the presence of both conditions and focus therapies to address both. In concomitant oSDB and asthma, there appears to be some improvement in asthma symptomatology with adenoidectomy and adenotonsillectomy. However, comorbid asthma increases the likelihood of postoperative respiratory complications , as well as predicts a higher likelihood of postsurgical residual OSA disease.
Impact of race and socioeconomic status on pediatric obstructive sleep apnea
Several studies have characterized the relationship between demographics and pediatric OSA with regard to preoperative, intraoperative, and postoperative outcomes. Parental education has been demonstrated to be a mitigating factor for the development of pediatric OSA, whereas lack of education represents a risk factor for the disease. Furthermore, children with OSA are more likely to come from disadvantaged neighborhoods. It is hypothesized that neighborhood distress may play a role in disease evolution. A larger study looking at disease demographics demonstrated a higher incidence in African American children; when adjusting for poverty rate and single-female head of household factors, race was less of a factor.
One smaller study suggests that access to care is a significant determinant of health care outcomes, as distance from an urban center predicted loss to follow-up. However, even in an urban cohort, most of whom were on public insurance, almost 50% were lost to follow-up after initial ENT evaluation. This finding was confirmed in a cohort of publicly insured patients in whom there was a significant delay in obtaining polysomnography (PSG) and definitive surgical intervention. As far as school performance, only socioeconomic status (SES) is the common independent causative variable, suggesting that other factors may be less important when studying the effects of OSA on academic indices alone.
Several studies have looked at disadvantaged minority children and the impact of pediatric OSA. African American race and environmental tobacco smoke exposure have both been demonstrated to increase disease severity. These factors have similarly been demonstrated in Australia. Despite these observations and the higher incidence of OSA in this population, these patients are less likely to present for ENT services and are at higher risk of becoming lost to follow-up. Unfortunately, many patients with OSA in the at-risk populations (minorities, Medicaid recipients) often fail to seek care. A recent meta-analysis identified low SES, uninsured or underinsured status, and nonwhite race as potential disparities for access to care. Attempts to strictly define race as a standalone factor in pediatric studies has been obfuscated by the fact that race is largely self-supplied in most studies, although African ancestry carries a genetic signature in an adult genome-wide association study analysis of OSA.
Surgical Factors
There are multiple surgical techniques described for pediatric adenotonsillectomy in patients with oSDB. However, bleeding rates may differ by surgical technique used. A large Cochrane analysis found that existing data are insufficiently powered to demonstrate a difference in bleeding rates. Over time, the pediatric ENT community is slowly shifting to monopolar electrocautery and coblation techniques and away from cold techniques. Intracapsular tonsillectomy has been suggested to be superior in select patients with OSA with regard to complication rates. However, one international study reviewed 1087 patients and found that monopolar techniques were associated with lower hemorrhage rates as compared with electrocautery. Therefore, significant controversy remains in the literature surrounding the “optimal” surgical technique for pediatric adenotonsillectomy.
Surgical Timing
Despite adenotonsillectomy being widely accepted as first-line treatment for pediatric OSA, there are limited data as to the optimal timing of surgery. Approximately 67% of all pediatric adenotonsillectomies are performed due to airway obstruction, which represents a shift away from the historical indication for this surgery, chronic tonsillitis.
However, in cases of mild OSA (AHI 1–5), nonsurgical management has been proposed as an option. Kohn and colleagues retrospectively reported on 201 pediatric patients with mild OSA of whom 101 (52%) opted for nonsurgical management. Of the 91 patients completing a follow-up sleep study, 46% had a greater than 20% decrease in AHI and 41% had a greater than 20% increase in AHI. However, 11 patients had an increase in AHI greater than 5, and 6 of these patients were upgraded from mild to moderate OSA and 5 were upgraded to severe OSA. Twenty-four patients (26%) had resolution of OSA with AHI less than 1. Older children between the ages of 12 and 18 tended to have greater AHI means on repeat PSG testing. Kohn and colleagues concluded that mild pediatric OSA had approximately equal chances of worsening or improving over time without surgical intervention.
Children with mild OSA often report poor quality of life (QOL) and can be very symptomatic. These children generally improve with surgical management. However, medical management is a possible option for children with mild OSA, particularly if the child is noted to have allergic rhinitis symptoms and a mild impact on QOL.
Therefore, in the absence of neurocognitive dysfunction, there are clearly subgroups with OSA for whom medical therapy may be of benefit. In the presence of neurocognitive sequelae, the optimal therapeutic path remains unclear. The CHAT study assessed children with mild-moderate OSA assigned to early intervention with adenotonsillectomy or symptomatic management. They found early adenotonsillectomy improves behavioral symptoms and QOL; however, these patients did not benefit from improvement in executive function. This logically led to the hypothesis that the neurocognitive dysfunction could be mediated by time with disease burden, and therefore surgery earlier in life would allow for executive function recovery. This finding was the rationale behind the recently published POSTA study. This study demonstrated that early adenotonsillectomy in young children was not beneficial in reversing changes in executive function. At the minimum, these data suggest that the mechanism mediating neurocognitive recovery maybe different than purely the obstructive component of pediatric OSA. What remains unknown is the progression of neurocognitive dysfunction without therapy. The behavioral and QOL benefits, however, have to be weighed against the complications associated with adenotonsillectomy.
Postsurgical complications of adenotonsillectomy
The complications of adenotonsillectomy can be put into 2 major categories: respiratory and nonrespiratory.
Respiratory Complications
A major consideration for adenotonsillectomy is the risk of perioperative respiratory complications. One study retrospectively analyzed more than 3000 children younger than 6 and identified that children younger than 3 were at higher risk for respiratory complications. Similarly, age-based risk stratification also demonstrated that children at or younger than 2 years of age were at higher risk than children between 2 and 3 years of age. Further, children weighing less than 14 kg also had a higher rate of post tonsillectomy respiratory complications and generally, children younger than 2 years of age weigh less than 14 kg. In addition, obese children, as well as those with congenital/craniofacial syndromes, are also at higher risk of respiratory complications from OSA. Therefore, despite adenotonsillectomy on the whole being a safe procedure, there are certain subgroups for whom increased caution is warranted. ,
Nonrespiratory Complications
The major nonrespiratory complications of this surgery include dehydration, hemorrhage, and postoperative fever. , Dehydration has typically been coupled with oropharyngeal pain, preventing adequate oral rehydration. Many patients are not compliant with postoperative pain regimens. Furthermore, there is also an association with low SES and underprivileged minority status having a higher rate of urgent revisits not related to bleeding. Younger children also have a higher risk of non–bleeding-related revisits. This has led a push to focus on symptom control to help transition these patients in the postdischarge period. In addition, societal guidelines have been projected to stratify patients by risk of complications. There are some data to suggest that these guidelines have reduced the revisit rates for non–life-threatening complications.
Post tonsillectomy hemorrhage (PTH) remains a major, and potentially life-threatening concern. , This complication is more common in young children, specifically those younger than 3, although a recent meta-analysis also found children older than 8 to be at increased risk. Obese children have also been found to be at greater risk. Children with OSA are also known to have lower PTH rates than children with chronic tonsillitis. Secondary PTH is more likely to be life threatening. Further, there is a fair amount of debate within the ENT community about the contribution of surgical technique to postoperative PTH rates. Last, the severity of OSA on presentation does not seem to influence the incidence of postoperative complications.
Summary
In summation, pediatric OSA is a common, heterogeneous disorder affecting many children. There are certain subgroups, including younger children, underprivileged minorities, and children with a low SES background, that have a higher morbidity of disease at presentation and a higher rate of complications with surgical therapy. These risks have to be balanced against the end-organ consequences of untreated disease, including neurocognitive morbidity in a vulnerable population. Although much is known about the disease, much remains to be elucidated, especially the reversibility of neurocognitive dysfunction after therapy.