Sleep disorders in Down syndrome: a systematic review

ABSTRACT Background: Sleep disorders are commonly observed in children with Down syndrome (DS) and can lead to significant behavioral and cognitive morbidities in these individuals. Objective: To perform a systematic review evaluating sleep disorders in individuals with DS. Methods: Search strategies were based on combinations of keywords: “Down syndrome”; “trisomy 21”; “sleep disorders”; “dyssomnias”; “sleep apnea”; “obstructive”; “sleeplessness”; “insomnia”; “parasomnias”; and “excessive daytime sleepiness”. PubMed and Science Direct were used. Only original studies and retrospective reviews in English published between January 2011 and March 2021 were included. Results: 52 articles were included, most of them involving children and adolescents under 18 years of age. The main sleep disorder associated with DS was obstructive sleep apnea (OSA). Some studies reported the presence of cognitive dysfunction in patients with DS and sleep-disordered breathing, and few have been found about parasomnia, insomnia, and daytime sleepiness in these patients. Movement disorders and unusual postures during sleep may be related to disordered sleep breathing in DS. The main treatment options for OSA are continuous positive airway pressure therapy (CPAP), surgery, and weight control. Computational modeling associated with MRI has been used to plan surgical interventions in these patients. Conclusions: Individuals with DS are at high risk of developing sleep-related breathing disorders. The main sleep disorder associated with DS was OSA. The presence of sleep-disordered breathing contributes to a worsening of cognitive function in patients with DS.


INTRODUCTION
Down syndrome (DS) was first characterized in 1866 by John Langdon Down, who described "individuals with peculiar clinical manifestations". Furthermore, in 1958, Jérôme Lejeune and Pat Jacobs stated that DS is a genetic syndrome related to a trisomy of chromosome 21. The DS prevalence in USA is around 13.56 for every 10,000 live births [1][2][3] .
Clinical manifestations vary widely from person to person, but cognitive impairment is commonly noted in this syndrome 4,5 . Also, there are some common phenotypic features in individuals with DS, such as muscle hypotonia, macroglossia, brachycephaly, epicanthal folds, flat nasal bridge, micrognathia, low-set ears, excessive skin on the nape, single transverse palmar crease, clinodactyly of the fifth finger, and a larger gap between the first and second toes 6,7 .
Sleep plays a critical role in good health and well-being. For this reason, sleep disorders in children and adolescents are associated with problems in physical, behavioral, and physiological development and pose an additional risk for obesity, endocrine disorders, depression, immunological, and heart diseases [8][9][10] . These disorders are commonly observed in children with DS and can lead to significant behavioral and cognitive morbidities in individuals with DS [11][12][13] .
The aim of this study was to provide a systematic review to evaluate sleep disorders in people with Down syndrome, focusing on clinical presentation, pathophysiology, and treatment strategies.

METHODS
A systematic review of the literature, based on the PRISMA statement and the recommendations for systematic review and meta-analysis, was conducted to investigate the main sleep disorders in patients with Down syndrome and their treatment 13 . Search strategies were based on combinations of keywords "Down syndrome", "trisomy 21", "sleep disorders", "dyssomnias", "sleep apnea", "obstructive", "sleeplessness", "insomnia", "parasomnias", and "excessive daytime sleepiness", which were defined based on previous research in the Medical Subject Headings (MeSH) system. PubMed and Science Direct were used as databases, with a publication period of January 2011 to March 2021.
Researchers 1 and 2 (R.A.S and L.H.C) considered the topics covered in each article searched, in addition to the inclusion and exclusion criteria. Treatment-only studies were excluded; the focus was on studies that addressed sleep disorders in patients with DS.
Inclusion criteria were: original studies and retrospective chart reviews written in in English and with no restriction on health, age, or gender of subjects. Exclusion criteria were: papers not related to sleep disorders in DS patients after reading the full text and editorials, letters to the editor, review articles, case reports, and meeting abstracts. The collected data were compiled into a spreadsheet containing all relevant information from the studies, including authors, year of publication, journal name, sample characteristics (size, gender, age, and geographic area), data collection methods, clinical diagnosis, and assessed sleep disorder.

RESULTS
An initial search identified 3559 studies from the past 10 years. Subsequently, editorials, letters to the editor, review articles, case reports, meeting abstracts, and laboratorybased studies, including animal studies, were excluded, remaining 163 articles. After reading full-text articles, that met all predefined criteria, and excluding duplicates, 52 articles were included in this systematic review ( Figure 1).
Among the selected studies were papers from 14 countries, most of them from the USA and Belgium. Regarding the population studied, most studies included children and adolescents under 18 years of age, and only 9 included the adult population. The main results are summarized in Tables 1  to 4. Almost all studies were case series, and about 50% of the manuscripts used PSG to define OSA.

Prevalence, etiology, and correlating factors for sleep disorders in individuals with Down syndrome
The main sleep disorder associated with DS in the selected articles was obstructive sleep apnea (OSA), with a prevalence ranging from 60 to 95%, depending on the criteria used for diagnosis and the age of the patients. However, the heterogeneity between the studies in terms of the method used for the diagnosis of the respiratory disorder is noteworthy: polysomnography (PSG), home polysomnography (HPSG), home night sleep records, cardiorespiratory polygraphy, housekeeping, McGill oximetry score, and actigraphy. In some cases, only questionnaires or scales were used, such as: Pittsburgh Sleep Quality Index (PSQI); Epworth Sleepiness Scale (ESS); Berlin Questionnaire (BQ); Child Sleep Habits Questionnaire (CSHQ), which may have compromised the assessment of prevalence [14][15][16][17][18][19][20][21][22] .
We found few studies of parasomnia, insomnia, and daytime sleepiness in individuals with DS. Two studies found that some sleep problems were significantly more common in the population with DS, such as: resistance to bedtime, sleep duration, sleep anxiety, night watch, parasomnias, and daytime sleepiness 20,23,24 . However, none of the studies addressed the presence of parasomnias and their most frequent types isolated.
Maris et al. studied the occurrence of parasomnias, insomnia, and daytime sleepiness by comparing two groups    There was no significant change in oAHI, oxyhemoglobin saturation nadir, ETCO 2 , or percent TST in REM after treatment for any treatment group. There was no association between reported symptoms and AHI severity or change in AHI.
In this cohort, the resolution of mild AOS was low for all treatment groups. These findings are consistent with the current understanding that OSA in children with DS is probably the result of multiple overlapping abnormalities contributing to the obstructive pathology      The high prevalence of disease found in these previously undiagnosed 8-year-old children underlines the importance of performing OSA diagnostics in children with DS throughout childhood. These findings suggest that the prevalence of OSA remains high up to early school years. of DS patients, the first with younger individuals (4 to 6.9 years) and the second with older children (over 11 years). Parasomnia was reported significantly less frequently with increasing age, which is also seen in normally developed children. In children with DS, in contrast to children with normal development, a decrease in the prevalence of sleep anxiety with increasing age was observed. Delay in falling asleep occurred more frequently in children with DS than children with normal development. Sleep onset delay in DS was significantly more common with increasing age and in children with sleep anxiety. Daytime sleepiness occurred more frequently among boys, regardless of age 20      Tonsillectomy with adenoidectomy was the most common procedure in both groups, but the proportion of tonsillectomy with adenoidectomy decreased over time. -The proportion of palatal surgery and tracheostomy also decreased significantly, whereas there was an increase in the proportion of lingual tonsillectomies, tonguebase reduction procedures, and supraglossoplasty performed in both groups over time. The relative rates of change in these procedures were higher in the DS population.
Tonsillectomy with adenoidectomy remains the most commonly performed procedure, although there was a significant increase in other sleep surgeries performed (LT, tongue-base reduction, and supraglossoplasty) between the two study periods, especially in children with DS. Children with asthma, autism and a history of enlarged adenoids and tonsils had more frequent sleep problems than children without these comorbidities.
Sleep disorders are important but also underrecognized problems in children with DS. It appears to be correlated with some prevalent comorbidities, which may provide guidance to augment current practice guidelines to evaluate sleep problems in this population. Sensor-recorded position (supine, prone, lateral) was expressed as the percentage of total sleep time. The apnea-hypopnea index (AHI) was calculated in each sleep state (NREM, REM), position, and position-sleep state combination. AHI was higher in REM than NREM; nonetheless, the NREM AHI was higher in DS than NREM AHI that controls.
In DS and non-DS children alike, respiratory events are predominantly REM-related. However, when matched for OSA severity, children with DS have a higher NREM AHI, which is worse in the supine position, perhaps indicating a positional effect compounded by underlying hypotonia inherent to DS.
These activities are atypical in these patients, indicating that DS patients are at greater risk for overweight/obesity, cardiovascular/metabolic diseases, OSA severity, and a salivary flow reduction 25 .

Sleep disordered breathing in patients with Down syndrome and its negatives effects on cognitive function
Several studies have been done associating sleep parameters and cognitive functions. In one interesting study, neurophysiological parameters obtained in the PSG and multiple sleep latency test (MSLT) were correlated with the answers in cognitive tests, and found that shorter total sleep duration and greater sleepiness were associated with poorer cognitive function in patients with DS. Furthermore, the lowest percentage of slow-wave sleep was found to be a predictor of better adaptive behavior and academic performance in individuals with DS. Another important finding was that appropriate treatment of sleep-disordered breathing in DS patients resulted in better cognitive performance, especially in the area of attention 26 .
Lee et al. compared the results of PSG studies and cognitive scales assessing language, behavior, and intellectual performance in patients with DS. They found that reduction in the percentage of REM sleep and the presence of OSA were associated with impaired language function in patients with DS 18 . Other studies with similar designs have correlated a reduction in slow-wave sleep with poorer performance in verbal learning and executive functions in patients with DS 26-30 .
In addition, children with DS are at higher risk for sleep disordered breathing (SDB), which can negatively affect the cardiovascular system. Besides, the risk of future cardiovascular events is increased in these patients due to decreased nocturnal reduction in heart rate (HR) and blood pressure (BP) 31 .
Another study discussed the unstable ventilatory control that is more common in children with DS. This finding indicates that these children are at greater risk for sleep disordered breathing than patients without DS 32 .

Sleep related movement disorders and unusual sleep postures in Down syndrome patients
Sleep problems in children with DS go beyond OSA and other sleep-disordered breathing. Sleep-related movement disorders are also more common in individuals with DS 20,23 . Hoffmire et al. observed that 21% of children with DS were positive for sleep-related movement disorders measured with the CSHQ. Also, this risk was associated with asthma, autism, and a history of enlarged adenoids and tonsils 23 .
Other previous studies applied questionnaires and found that atypical positions such as leaning forward with legs back, leaning forward with legs forward, leaning forward with legs crossed, and sitting were common and were often related to the presence of OSA diagnosis 33 . Additionally, patients with DS commonly present the unique position of sitting with a flopped-forward body in which the head rests on the bed while asleep, which contributes to optimized airflow 34 . The reason for this position is unclear, but authors conjectured that this may be a protective mechanism for airway patency 35 .
Another study used PSG and recording of body positions during sleep using sensors. Subjects with DS spent a significantly longer duration of sleep in the prone position and less in the right lateral decubitus position compared to subjects without the paired syndrome by age and sex 36 .

OBSTRUCTIVE SLEEP APNEA IS THE MOST PREVALENT SLEEP DISORDER IN PATIENTS WITH DS
As previously mentioned, OSA is the most prevalent sleep disorder in these patients, and there are a few reasons for this. Maris et al. found that children with DS have anatomical narrowing of the upper airway at different levels and are more prone to collapse and thus at higher risk for OSA. Other factors contribute to explain the association between OSA and DS such as muscle hypotonia, higher incidence of congenital heart disease, hypothyroidism, lung disease, immunodeficiency, relative macroglossia (due to smaller bone framework of mandible and maxilla) 37 .
Some studies have included a control group of children and adolescents without DS and found a prevalence of less than 20% of OSA, highlighting the important association between DS and OSA 16,21,22,38,39 . Some authors claim that individuals with DS have more severe OSA and greater refractoriness to treatment 16,19,39,40 . According to studies by Konstantinopoulou et al., left ventricle diastolic function correlates with the severity of OSA, which improves with the use of continuous positive airway pressure (CPAP). In addition, they noted a tendency for increased sleepiness in individuals with OSA, which was correlated with the awakening index. Further studies are needed to confirm the findings described 41 .
Coverstone et al. evaluated the probability of developing OSA in DS patients with pulse oximetry and classified them according to the McGill score. Patients with McGill score 3 or 4 (more than 3 desaturations below 80-85% in one night of sleep) or McGill score 2 with increased body mass index (BMI>25 kg/m 2 ) were referred by an otorhinolaryngologist due to their increased risk of adenotonsillar hypertrophy. The authors suggest that patients with low McGill scores should be monitored regularly by a specialist to obtain continuous assessment 15 .
Nicolas et al. conducted the first study to compare nocturnal gas exchange in children with DS with a control group of children with similar OSA. They concluded that patients with DS have respiratory muscle hypotonia and/or an alteration in ventilatory control 42 .
Nisbet's study showed that children with DS and OSA had a similar dominance of rapid eye movements (REM) in breathing events compared to children with OSA and without DS, but the children with DS had a higher NREM apneahypopnea index (AHI), even though they were similar in terms of total AHI and had a similar percentage of sleep time in NREM. Notably, children with DS in supine position had a higher NREM AHI than in the non-supine position 36 .

Obesity and other possible predictive variables for obstructive sleep apnea in patients with Down syndrome
The association between obesity and the occurrence or severity of OSA in patients with DS is controversial. Most studies included that no correlation exists between higher BMI and OSA in this population 14,17,19,33,40,43,44 , but it should be noted that most of these studies included children only.
On the other hand, Chamseddin et al. correlated obesity not only with a higher occurrence of OSA in DS patients, but also with a high severity of OSA 45 . Similarly, two other studies reported that patients with DS, who had high BMI and/ or hypothyroidism, had greater upper airways narrowing and consequently a higher severity of OSA. They also highlight the importance of preventing obesity in adolescence to reduce the incidence of OSA in adults with the syndrome 15,16 . Therefore, there is no consensus among researchers on the relationship between OSA and overweight/obesity.
There are some predictive variables for the occurrence of OSA in patients with DS, such as presence of parasomnias, longer total sleep time, daytime sleepiness, snoring, witnessed apnea and nocturia 18,23,27,33,43,46 . Hoffmire et al. described that the presence of asthma or allergic rhinitis is not related to an increased risk of OSA in patients with DS 23 . In addition, there is no consensus among researchers on the association between gastroesophageal reflux disease (GERD) and OSA in this population 23,43 . Nehme et al. pointed out that the symptoms of GERD may be similar to those of OSA, leading to a better performance of the PSG exam, which could contribute to a greater identification of OSA in these patients 43 .

Screening methods and biomarkers for obstructive sleep apnea in patients with Down syndrome
Although PSG is considered the gold standard examination to define OSA, screening methods have been investigated to evaluate sleep disorders in this population. Considering the technical difficulties in performing PSG, the lack of availability of the exam, and its high cost, alternatives must be sought. In this manuscript, it was shown that only about 50% of selected studies used PSG to define OAS in DS patients. Although the presence of restlessness and snoring are important indicators of OSA in patients with DS, no significant association between these indicators and low oxygen saturation was found in the Stores et al. study. Therefore, the authors suggest that the presence of restlessness may be an important clinical feature to assess the need for a PSG 47,48 . Questionnaires and clinical and laboratory data are used to identify moderate to severe OSA in this population 49 .
Another alternative is screening by home pulse oximetry (HPO), which could halve the number of children with DS who need multichannel sleep studies 50 . Although these tests are useful, they cannot be used in isolation to diagnose breathing-related sleep disorders 47,50 .
Two studies have used questionnaires as a tool for diagnosing OSA. In the first study, conducted by Hoffmire et al., the CSHQ and the Pediatric Sleep Questionnaire (PSQ) were applied 23 . In the second study, conducted by Maris et al., the CSHQ and the PSG were used as auxiliary tools for diagnosis 20 . Both studies concluded that a large number of children with DS had sleep behavior disorders (insomnia, parasomnias) and sleep-related breathing problems, but curiously, their caregivers did not complain of such conditions. No relationship was found between the scores obtained in the CSHQ and the OSA index 20,23 . Therefore, the isolated use of questionnaires as a screening tool for OSA does not seem to be an effective method.
In an interesting study conducted at Boston Children's Hospital, a predictive model was created to help screen for OSA in patients with DS. The variables used were age, sex, race, height, weight, BMI, sedentary behavior, blood pressure, peripheral O 2 saturation, neck circumference, macroglossia assessment, Mallampatti classification, Friedman/ Brodksy scores, classification of scores, and current treatment for asthma, GERD, or thyroid disease. Results of the following scales and questionnaires were also used: PSQ, CSHQ, and Sleep Disorders Scale (SRBD), which were applied to parents and/or guardians. Using a logic learning machine, the best model had a validated negative predictive value of 73% for mild OSA and 90% for moderate or severe OSA. The final model revealed that the most relevant variables (out of 101) were certain CSHQ questions, SRBD questions, and the hypertension percentile. The study shows promising results with models using clinical data and questionnaires and may be an interesting tool for screening OSA in patients with DS 51 .
Similarly, Beppler et al. have developed a prototype called PediBand to help diagnose OSA in patients with DS. PediBand assesses the following physiological parameters: heart rate and its variability, respiratory rate, and O 2 saturation. This model is a promising tool to investigate sleep disordered breathing in DS. However, as it is still a prototype, further clinical studies are needed to strengthen the evidence for its use 52 .
OSA biomarkers have also been studied in individuals with DS. Elsharkawi et al. measured biomarkers such as epinephrine, norepinephrine, dopamine, serotonin, glycine, taurine, γ-aminobutyric acid (GABA), glutamate, phenylethylamine (PEA), aspartic acid, histamine, 3,4-dihydroxyphenylacetic acid (DOPAC), 5 hydroxy acid (5-HIAA), tyramine, and tryptamine in DS patients with OSA, DS patients without OSA, and in healthy controls, which were equal in age and gender. The results showed that epinephrine, norepinephrine, dopamine, and taurine were good predictors of the presence or absence of DS, but these results were not statistically significant in distinguishing the presence or absence of OSA in these patients. Thus, these urine biomarkers were ineffective tools for screening OSA in individuals with DS 53 . It should also be noted that the low availability of the tests and the technical difficulties in performing them are major obstacles to its use in clinical practice.
Jayaratne et al. conducted a 3D comparison of patients with and without OSA. An anthropometric analysis scheme was developed to quantify facial norms with well-defined reference points focusing on the soft tissues of the external morphology. Most anthropometric measures were lower in individuals with DS, indicating maxillomandibular hypoplasia and reduced measures of the nose, ears, and eyes. However, the authors compared patients with DS and OSA versus patients with DS and without OSA and found no significant differences in these measures. A limiting factor was the restriction to ethnicity (Caucasians only), which requires a more in-depth analysis of different ethnicities and a wider age range 54 .

Treatment options for obstructive sleep apnea in patients with Down syndrome and new perspectives
The main treatment options for OSA in DS patients are CPAP, surgery, and weight control. Several therapeutic alternatives have been studied, considering that CPAP therapy is not always available or tolerated, that surgical intervention is not always appropriate, and that there is no consensus on whether there is a direct relationship between obesity and OSA in patients with DS.
Several studies indicate that adenotonsillectomy (AT) is still the gold standard for the treatment of OSA in patients with DS 37,39,44,[55][56][57][58] . Other possible interventions include lingual tonsillectomy (LT) and supraglossoplasty (SGP). LT may be considered in the context of residual OSA after AT, despite its lower efficacy 56 . The authors emphasize the importance of surgical planning with the identification of upper airway obstruction sites and the main tool for this purpose is druginduced sleep endoscopy (DISE) 20,59,60 .
Concerning drug treatment, further studies are needed to clarify its role in the OSA in patients with DS. Intranasal corticosteroids may contribute to a local anti-inflammatory effect by reducing apnea, but the effectiveness has not been fully demonstrated. A retrospective study showed that children who underwent AT and used nasal corticosteroids had less residual OSA than children who did not undergo this drug treatment. Considering the small sample size of the study, the role of medication in the treatment of residual OSA in DS remains uncertain 60,61 .
One of the new interventions that have been studied is myofunctional orofacial training (MT). MT is based on the principle of strengthening orofacial and cervical functions for muscular balance, thereby reducing the chances of recurrences due to the maintenance of inadequate functional patterns 58 . Diercks et al., on the other hand, pioneered the investigation of a hypoglossal nerve stimulation implant in the pediatric range as a prospect for treating a patient with DS associated with severe OSA. The study demonstrated that the therapeutic intervention produced a well-tolerated and effective outcome and significantly reduced the patient's respiratory impairment 62,63 .
Three-dimensional reconstruction models from imaging exams -such as computer tomography (CT) and magnetic resonance imaging (MRI) -look promising, but studies with a larger sample of patients are needed to verify their real effectiveness 64,65 .
In conclusion, individuals with DS are at high risk of developing sleep-related breathing disorders, mainly due to anatomical changes in the upper airway. The presence of sleep disorders contributes to the deterioration of cognitive function in patients with DS. PSG is the gold standard exam for determining OSA, but the high cost and difficulty of technical approach are pushing for better options. OSA is the most studied sleep disorder in patients with DS and its main treatment is AT. There are some emerging perspectives on OSA treatment in patients with DS, but high-quality trials of multimodal interventions are needed to provide robust evidence for the treatment of OSA in DS individuals.