Use of Sleep Endoscopy to Predict Outcomes of Pediatric Adenotonsillectomy

The purpose of this study is to determine whether sleep endoscopy performed in high-risk pediatric patients with obstructive sleep apnea (OSA) at the time of adenotonsillectomy (AT) can predict whether the AT will be successful as an initial treatment for OSA. We hypothesize that patients with multiple sites of obstruction in addition to adenotonsillar hypertrophy (e.g. the nasal airway, velum, base of tongue, supraglottis) will be more likely to have residual sleep apnea on postoperative sleep testing.

Obstructive sleep apnea syndrome OSAS is defined as the symptomatic repetitive obstruction of the upper airway during sleep and has been estimated to affect 1-6% of the general pediatric population. Untreated OSAS in children has been associated with childhood hypertension, autonomic dysfunction, attention deficit/hyperactivity disorder, neurobehavioral impairment, and poor quality of life. These sequelae contribute to a 226% increase in health care utilization among children with OSAS compared to controls, primarily in the form of increased hospitalizations, emergency department visits, and medication use. Adenotonsillar hypertrophy is considered the most common risk factor for OSAS in children, therefore unlike in adult OSAS, adenotonsillectomy (AT) is the recommended first line treatment. In large part due to the increasing awareness and diagnosis of pediatric OSAS, the incidence of AT increased dramatically from 1980 to 2005. With more than 500,000 procedures performed per year, AT is now the second-most common procedure performed in children in the US, and 77% of these have OSAS as the primary indication.

Current guidelines recommend AT as a first line treatment for pediatric OSAS even for those patients who may have significant risk of post-AT OSAS. Estimates of the prevalence of persistent OSAS after AT vary widely due to use of different polysomnographic criteria for diagnosis. Studies that assessed the risk of post-AT OSAS using a conservative adult threshold for diagnosis demonstrated that even with this high threshold at least 13-29% of children undergoing AT for pediatric OSAS will have significant residual disease and approximately 75% of children will fail to achieve normalization on polysomnography. Specific populations of patients that have been recognized to be particularly at risk for post-AT OSAS include those with severe baseline OSAS, Down syndrome, obesity, and age > 7 years. In obese patients, the prevalence of post-AT OSAS has been reported as high as 73-88%. Since obesity has tripled over the last three decades and now affects approximately 8% of children aged 2-5 years, 18% of children aged 6-11 and 21% of adolescents aged 12-19 years, the problem of persistent OSAS after AT is likely to continue to grow.

Even within populations at risk for AT failure, there is a wide variation in treatment response. One study of morbidly obese children undergoing AT demonstrated only a 37% cure rate while 53% had sufficient residual OSAS to require further treatment with continuous positive airway pressure (CPAP). However, no significant baseline differences were identified between surgical responders and non-responders. The mechanism for failure in this population is unclear, but it is presumed that increased generalized adiposity leads to multilevel obstruction similar to obese adults, thus decreasing the likelihood of success with AT. Similarly a poor but still variable response to AT was observed in children with Down syndrome with post-AT success varying between 18% and 55% depending on the specific criteria used. There are no studies that have clearly identified predictors of AT outcome within the Down syndrome population, however, some studies of Down syndrome patients who failed AT have suggested that multilevel obstruction is common. Thus, although specific populations of patients are known to have greater risk of post-AT OSAS on average, the individual characteristics causing persistent disease remain unclear. Accurate prognostication of the risk of residual OSA after AT for any individual patient remains a challenge. Studies of patients with persistent post-AT OSAS have suggested that multilevel obstruction at locations besides the tonsils or adenoids are likely contributors, but this has not been clearly demonstrated. In this study, we present a novel concept for building a composite model to predict the outcome of AT in children with OSAS. This model will include not just baseline features of history and physical exam but also the findings of dynamic sleep-related collapse at specific anatomic sites in the pharynx observed during sleep endoscopy. This model will give further insight into the mechanisms of airway obstruction as well as the possible reasons for persistent OSAS after AT.

A rating scale for DISE has previously been described in an attempt to standardize the reporting of endoscopic findings in adults with OSAS. This rating scale evaluates the degree and pattern of obstruction at four levels of the pharynx: the Velum (soft palate), Oropharynx (including the tonsils), Tongue base, and Epiglottis (VOTE). The VOTE rating scale has been demonstrated to have moderate to substantial inter-rater reliability with kappa values ranging from 0.4-0.8 depending on the specific structures being compared. Other investigators have utilized modified versions of the VOTE rating scale in children, including other levels of the airway such as the nasal airway, nasopharynx, and supraglottis. One recent study demonstrated that sleep endoscopy findings in children were more reliable than during awake endoscopy and noted a strong correlation between polysomnography results and the overall impression of OSA severity during endoscopy.

Dexmedetomidine, which will be used in this study, is a highly selective α2-adrenoceptor agonist that has been demonstrated to result in a sedated sleep similar to natural sleep without causing respiratory depression. Though there have been some reports of transient bradycardia and blood pressure changes in response to dexmedetomidine infusion (usually transient hypotension of 10% with slow infusion) these cardiovascular effects are mitigated by co-administration of a bolus of ketamine. These patients would already have been using dexmedetomidine or a different anesthesia for their tonsil surgery. The dexmedetomidine is not an intervention or part of this study.

In a preliminary retrospective review of our patient population and surgical volume, we examined the electronic medical record of all patients who underwent AT over a 12 month period. 498 patients were identified, operated on by the four pediatric otolaryngologists in the group. Approximately 200 (40%) of these were performed for OSAS in patients that could be considered high risk for residual post-AT OSAS and would meet inclusion/exclusion criteria described below.

Untreated OSA in children has been associated with childhood hypertension, autonomic dysfunction, attention-deficit/hyperactivity disorder, poor school performance, and poor quality of life. These sequelae contribute to a 226% increase in health care utilization among children with OSA compared to controls. Residual OSA after AT in the pediatric population remains a serious concern; as the patient grows and changes, their airway physiology also changes. Although there is a growing body of research suggesting demographic and comorbidity risk factors for post-AT residual OSA, the ability to accurately predict the likelihood or severity of residual OSA for any given individual remains elusive. Possible tools for the evaluation of post-AT OSA in pediatric populations include radiologic examinations, cine MRI scanning, and endoscopic evaluation, along with polysomnography, validated questionnaires, and physical examinations. Determining what instruments best predict residual OSA after surgical intervention is an important step towards more effective OSA management.