Prospective Polysomnographic Analysis of Obstructive Sleep Apnea in Down Syndrome

Objectives  To investigate obstructive sleep apnea (OSA) in a consecutively encountered, nonselected population of young patients with Down syndrome using standard overnight polysomnography and to determine the effects of therapy.

Methods  In a population of patients seen for routine developmental evaluations, 9 boys and 10 girls were studied using standard overnight polysomnography.

Results  Using pediatric standards, OSA was found in 79% of the subjects (95% confidence interval, 54%-94%), with a median apnea index of 3 events per hour (interquartile range, 2-5), a median apnea-hypopnea index of 6 events per hour (interquartile range, 3-8), and a median arterial oxygen saturation (SaO₂) low point of 88% (interquartile range, 84%-90%). Higher body mass index was significantly associated with a higher apnea index and a lower SaO₂ level, and there was a significant inverse relationship between age and the lowest SaO₂ value as well as a possible association between sleep-related symptoms at the time of diagnosis and the lowest SaO₂ value. In addition, patients with OSA had a significantly higher movement arousal index than those without OSA.

Conclusions  Using rigid polysomnographic standards, this pilot study revealed OSA in a high percentage of young subjects with Down syndrome and an association between OSA and obesity, age, and poor sleep quality. These findings justify larger and more detailed population studies to further define clinical factors that are concomitant with OSA in Down syndrome and to improve therapy.

POLYSOMNOGRAPHY (PSG) using the criteria of Rechtschaffen and Kales¹ is the gold standard to confirm the presence and severity of OSA. To our knowledge, this is the first study using complete and detailed overnight PSG to address the prevalence of OSA in an unselected population of young patients with Down syndrome who received follow-up for therapeutic interventions.

Our study was approved by an institutional review board of human studies, and written consent was obtained from the patient or guardian. Individuals were recruited from a consecutively encountered patient population as 2 of us (D.C.L-D. and S.P.) saw them for routine developmental evaluations in the Down syndrome clinic at the Center for Disabilities and Development of the University of Iowa (Iowa City). The criteria for participation included a diagnosis of Down syndrome on the basis of physical examination findings and karyotype as well as an agreement to undergo overnight PSG.

Four subjects declined participation in this study. A total of 2l subjects underwent complete overnight PSG. Two of these patients were not used in the final analysis because one had technically inadequate PSG and another was considered an adult (age, 22 years). Difficulties encountered in scheduling PSG for a rural research population frequently traveling long distances, using a single-bed pediatric laboratory concurrently evaluating clinical patients, resulted in a relatively extended period for enrollment and completion of the baseline PSG (September 29, 1993-November 9, 1995).

Since 1947, the Center for Disabilities and Development has served individuals from rural communities in a large multistate area where there is a relative paucity of comprehensive developmental services for children with special needs. At the time of entry, all patients in this study were receiving follow-up for developmental concerns and overall health care. No patient was initially referred to the center with a primary diagnosis of OSA.

A standard polygraph (Nicolet 1A97; Nicolet Instrument Corp, Madison, Wis) was used to record the electroencephalogram, electro-oculogram, electromyogram of the chin and bilateral tibialis anterior muscle, electrocardiogram, airflow measurement, and chest wall movement. Airflow was measured using nasal and oral thermistors. Respiratory effort was monitored with a combination of intercostal electrodes, an abdominal strain gauge, and visual assessment of chest and abdominal wall movement using video monitoring. Oximetry was measured using a disposable finger probe placed on the index finger. Snoring was recorded using overhead microphones. Audiovisual information was obtained using a black-and-white video camera.

Using the criteria of Rechtschaffen and Kales,¹ the studies were analyzed in an American Academy of Sleep Medicine–accredited sleep center by technicians and physicians who were board certified in sleep disorders medicine. Repeated PSG was performed for patients with OSA who experienced a recurrence of sleep-related symptoms (defined as snoring, insomnia [difficulty initiating and/or maintaining sleep], excessive sleepiness, and behavioral problems).

Children should not be assessed using adult criteria; they generally have faster respiratory rates and smaller functional residual capacity, so significant oxygen desaturation can occur during short periods of apnea. There is no universally accepted normative data that clearly define hypopnea in children, so we used both an apnea index (AI) and an apnea-hypopnea index. Most experts consider an apnea with a duration of 2 or more normal respiratory cycles (for the individual patient) as significant in children. Some physicians consider an AI of more than 1 event per hour and an arterial oxygen saturation (SaO₂) low point of less than 92% (when the resting, waking baseline value is >92%), in association with an obstructive event, as abnormal.^2-4 These are the standards we used to diagnose OSA.

An emergency protocol was established for split-night diagnostic and therapeutic interventions. This protocol included the initiation of continuous positive airway pressure (CPAP) for any desaturation level lower than 70% and any obstructive event that was clearly associated with cardiac dysrhythmia.

Evidence of possible medical factors concomitant with OSA such as a history of pulmonary hypertension and cardiac disease, abnormal electrocardiogram or echocardiogram results, cardiac catheterization, obesity, snoring, insomnia, excessive daytime sleepiness, and behavioral problems was retrospectively collected from all patients. Patients were determined to have a history of pulmonary hypertension after a thorough review of their medical records, which provided documentation of any previous diagnosis of pulmonary hypertension with clinical and procedural supporting evidence.

The Wilcoxon rank sum test was used to compare the AI and lowest SaO₂ value of subjects with and without histories of snoring, insomnia and/or excessive daytime sleepiness, behavioral problems, pulmonary hypertension, and tonsillectomy and/or adenoidectomy. The Pearson correlation coefficient was computed to examine the association of AI and lowest SaO₂ value with body mass index and age. The Wilcoxon rank sum test was also used to compare the movement arousal index (MAI) between patients with and without OSA.

A secondary goal was to address the effects of therapeutic intervention vs no therapy in patients diagnosed as having OSA. Treatment was recommended and offered to all patients who received this diagnosis.

An isolated tonsillectomy or tonsillectomy with adenoidectomy (as deemed most appropriate by the otolaryngologist) was the treatment of choice. If a tonsillectomy and/or adenoidectomy had already been performed (or was refused), CPAP was recommended. These therapies were often refused by the family or caregiver because of the potential risks of surgery and/or concerns about the patient's compliancy with CPAP. Positional therapy was recommended when previous tonsillectomy and/or adenoidectomy had been performed, CPAP was refused, and the obstructions were documented by PSG to occur only in the supine position. In addition, complementary recommendations were made including good dietary habits, low-level exercise when possible, the avoidance of all unnecessary central nervous system depressants (especially prior to sleep), and the institution of good sleep hygiene practices.

Initially, CPAP was considered successful when there was a complete resolution of all major obstructive events, including during rapid eye movement sleep while the subject was lying in the supine position. At follow-up, therapy was considered beneficial when the family or caregiver reported an improvement in sleep-related symptoms. Effects of therapy were documented during telephone interviews, medical record review, and follow-up clinic visits at 4-year and 7- to 8-year intervals after enrollment.

Obstructive sleep apnea was found in 15 (79%) of the 19 subjects included in the analysis (95% confidence interval [CI], 54%-94%) (Table 1 and Table 2). No patient had central sleep apnea; in the patient population with OSA, we recorded 1297 obstructive events.

The median AI in patients with OSA was 3 events per hour with an interquartile range of 2 to 5 events per hour. The median apnea-hypopnea index was 6 events per hour with an interquartile range of 3 to 8 events per hour. The median lowest SaO₂ value was 88% with an interquartile range of 84% to 90%.

In the total population studied, a higher body mass index was significantly associated with higher AI (r = 0.62; 95% CI, 0.23-0.84; P = .005) and lower SaO₂ level (r = −0.55; 95% CI, −0.13 to −0.80; P = .02). A significant inverse correlation was noted between age and lowest SaO₂ level (r = −0.48; 96% CI, −0.04 to −0.77; P = .04) (Figure 1). This relationship held true both when assessing the entire study group as a whole (patients with OSA and without OSA) and when specifically addressing the subgroup with OSA.

Overall, there was a suggested possible association between the presence of insomnia and/or sleepiness at the time of diagnosis and the lowest SaO₂ value. The median lowest SaO₂ level was 85% in subjects with insomnia and/or sleepiness compared with 90% in those without insomnia or sleepiness (P = .05).

In addition, the MAI was higher in the OSA group compared with the patients without OSA. The median MAI in the OSA group was 1.9 with an interquartile range of 0 to 9.2. This was significantly greater than the non-OSA group, which had a median MAI of 0 (P = .03) (Table 3 and Table 4).

By 8 years, 7 patients were lost to follow-up. Seven individuals with OSA were treated: 1 with tonsillectomy followed by CPAP, 3 with tonsillectomy and adenoidectomy, 2 with CPAP (of which 1 was noncompliant), and 1 with positional therapy (see Table 1 and Table 2).

Five patients with OSA had repeated PSG to address recurrent sleep-related symptoms (Table 5). Female patient 9 underwent 2 subsequent CPAP studies. Residual obstructions noted on the first CPAP trial necessitated a second trial, which resulted in a complete resolution of all major obstructions, including during prolonged rapid eye movement sleep in the supine position. Nevertheless, this patient was noncompliant with CPAP and was lost to follow-up at 8 years.

Four of the 5 individuals with repeated PSG initially chose no therapy, CPAP, positional therapy, or tonsillectomy. With time, all of these patients experienced a progression of sleep-related symptoms, and persistent or worsened OSA was documented with PSG in every subject. No further therapy was performed for male patient 8, who was subsequently diagnosed as having a severe oppositional behavior disorder. Female patient 3 is presently considering a tonsillectomy.

Male patients 1 and 9 were successfully treated with CPAP, determined on PSG as a resolution of all major obstructive events, including during prolonged rapid eye movement sleep in the supine position, and of sleep-related symptoms at follow-up. Currently, male patient 1 continues to tolerate CPAP well (with reports of significant improvements in behavior and functioning at school), and male patient 9 has reported a "100% improvement" in previously noted insomnia and daytime sleepiness.

Prior to this study, tonsillectomy and adenoidectomy had been performed in 40% of our patients with OSA, whereas 75% of the children without OSA had undergone previous upper airway surgery. Four subjects with OSA had a previous diagnosis of pulmonary hypertension, whereas no individuals without OSA had this condition. In 2 of these patients, previous upper airway procedures were associated with marked improvements in cardiopulmonary health. Because tonsillectomy with adenoidectomy is the main treatment for OSA in children, it is likely that the prevalence of OSA is even higher in this population.

The prevalence of OSA in the general pediatric patient population without Down syndrome has been estimated at 0.7%.⁵ This condition has been associated with snoring, obesity, insomnia, sleepiness, and behavioral problems. It has also been speculated that OSA may contribute to pulmonary hypertension in individuals with or without Down syndrome.^6-13

Down syndrome has been associated with a variety of respiratory abnormalities during sleep including Cheyne-Stokes respiration,¹⁴ central sleep apnea,¹⁵ rib cage to abdominal asynchrony,¹⁶ and OSA.¹⁷ Hypoxia and hypoperfusion may predispose patients with Down syndrome to pulmonary hypertension and cardiac disease.^9,10

The literature on OSA and Down syndrome is composed of case reports, small selected and retrospective treatment studies of patients known to have OSA, a few reviews, and studies that did not include standard PSG.^17-28 To our knowledge, this is the first prospective study of OSA in a nonselected population of patients with Down syndrome using full, detailed, formal PSG with follow-up of treatment during a long period.

Southall et al¹² studied 6 children with overnight recordings monitoring only SaO₂ level, expired carbon dioxide level, and respiratory movement. Stebbens et al²⁸ studied 34 children with Down syndrome primarily in home studies using only oximetry, respiratory inductance plethysmography, and end-tidal carbon dioxide measurements (electroencephalography, electromyography, and electro-oculography were not used). Marcus et al²⁹ monitored 54 children in a population that included patients referred specifically for OSA. This group underwent 1- to 2-hour daytime nap studies (only 16 selected patients had a subsequent night study) without electroencephalography (sleep was determined by observation). Children who did not sleep spontaneously were sedated. Levanon et al³⁰ studied 23 children with Down syndrome and found a mean respiratory disturbance index of 2.3 events per hour; these individuals had been referred because of difficulties in breathing during sleep.

In children, treatment for OSA usually consists of tonsillectomy and adenoidectomy. In patients with Down syndrome, isolated tonsillectomy is often recommended because their upper airway configuration may predispose them to velopharyngeal insufficiency with combined tonsillectomy and adenoidectomy.³¹ In cases in which tonsillectomy and adenoidectomy have not been successful or are not an option, CPAP has been effective and well tolerated.³² Treatment of obesity is recommended but is generally difficult.³³ When required, uvulopalatopharyngoplasty and tracheostomy have been used as the least preferred options.^17-34 In patients with positionally related OSA (OSA only in the supine position), positional therapy can be considered if standard therapy has failed or is refused.

This study demonstrates a high prevalence of OSA in a small population of young patients with Down syndrome. More severe OSA was associated with a higher body mass index, older age, higher MAI, and possibly more sleep-related complaints. These findings justify larger, more focused, and more detailed population studies to improve therapeutic efficacy for OSA.

Article

Corresponding author and reprints: Mark Eric Dyken, MD, Department of Neurology Sleep Disorders Center, University of Iowa, Roy J. and Lucille A. Carver School of Medicine, Iowa City, IA 52242 (e-mail: mark-dyken@uiowa.edu).

Accepted for publication February 25, 2003.

Obstructive sleep apnea has been associated with pulmonary hypertension, cardiovascular disease, and Down syndrome. To our knowledge, this is the first prospective study of OSA in a consecutively encountered, nonselected population of patients with Down syndrome using complete, detailed, overnight PSG with long-term follow-up. Using these standards, we provide stronger evidence that OSA is common in young patients with Down syndrome and is related to obesity, age, and poor sleep quality.