Journal of Otology & RhinologyISSN: 2324-8785

Reach Us +1 850 754 6199

Research Article, J Sleep Disor Treat Care Vol: 3 Issue: 1

Positive Airway Pressure Levels for Children with Sleep Disordered Breathing

Mohsen Alajmi1, Leah Schmalz1, Alberto Nettel-Aguirre2,3 and Valerie Kirk1*
1Department of Pediatrics, Section of Respirology, Division of Sleep Medicine, University of Calgary, Canada
2Departments of Pediatrics and Community Health Services, University of Calgary, Canada
3Alberta Children’s Hospital Research Institute for Maternal and Child Health, Canada
Corresponding author : Dr. Valerie Kirk
2888 Shaganappi Trail NW, Calgary AB, T3B 6A8, Canada
Tel: (403) 955-2923; Fax: (403) 955-3065
E-mail: val. [email protected]
Received: July 17, 2013 Accepted: December 09, 2013 Published: December 11, 2013
Citation: Alajmi M, Schmalz L, Nettel-Aguirre A, Kirk V (2014) Positive Airway Pressure Levels for Children with Sleep Disordered Breathing. J Sleep Disor: Treat Care 3:1. doi:10.4172/2325-9639.1000129


Positive Airway Pressure Levels for Children with Sleep Disordered Breathing

Study Objectives: Positive airway pressure (PAP) is the mainstay therapy for children presenting with sleep disordered breathing/ obstructive sleep apnea (SDB/OSA) who are either not a candidate for, or have failed other treatments. There are few published reports on PAP therapy in the pediatric population. We describe PAP settings used to treat clinical groups of children with and without Down syndrome. Methods: A retrospective cohort study was performed. Records from children, aged 1-18 years with prescribed PAP therapy to treat sleep disordered breathing (OSA/SDB) between 1997 and 2011 were reviewed. Demographic and polysomnographic data were extracted for six sub-groups of children with OSA/SDB: nonobese, overweight and obese (children with Down syndrome and otherwise typical children).



CPAP; Bi-level PAP, pediatrics; OSA; sleep apnea; Children


PAP: Positive Airway Pressure; SDB/OSA: Sleep disordered breathing and/or Obstructive sleep apnea; BMI: Body Mass Index; PSG: Polysomnography; RPSGT: Registered Polysomnography Technologist; AHI: Apnea Hypopnea Index


Obstructive sleep apnea affects two to three percent of schoolaged children; however, the prevalence is much higher in obese children ranging from 13% to 66% [1]. Untreated, this common respiratory condition leads to serious consequences including metabolic derangements, cardiovascular disease, neurobehavioral problems and poor daytime performance, all of which contribute to a reduction in quality of life [2-7]. Obesity is now a common cause of morbidity in children and adolescents. The prevalence of overweight children has doubled among children 6 to 11 years of age and tripled among children 12 to 17 years in the United States since 1980 [8,9]. Obstructive sleep apnea and/or sleep disordered breathing (OSA/ SDB) has been identified as one of the most evident complications of childhood obesity. A second group of children known to be at higher risk of OSA/SDB is that of children with Down syndrome. Prevalence rates have been well described in this group and range from 50-100% [10-12]. It has been shown that treatment of OSA/SDB in children results in improvement in blood glucose, blood pressure, daytime performance and improved quality of life [13-16]. Adenotonsillectomy is generally the first line in treating children with OSA/SDB [17], but the cure rate of OSA/SDB by adenotonsillectomy is much lower among both children with obesity and those with Down syndrome in contrast to non-obese children; thus, Positive Airway Pressure (PAP) is used often as treatment for these children [18]. Our lab and others have shown that with adequate parental support, PAP adherence rates in children are equivalent to adult rates [19,20]. However, although it has been shown to be tolerated and effective, there is no published literature describing PAP settings specific for the pediatric patients with OSA/SDB. We aimed to examine the relationship between different anthropometric and polysomnographic variables known either to correlate with OSA/SDB severity or contribute to upper airway obstruction and PAP settings. We also wished to identify the most common PAP levels used in our laboratory to treat OSA/SDB in children.


A retrospective, cohort study design was employed. Records of all children who underwent titration polysomnography in the sleep laboratory at the Alberta Children’s Hospital over a fourteenyear period were reviewed. Data were collected from both the sleep laboratory database and the supplementary sleep records stored in the sleep laboratory. No hospital charts were accessed for the study. Data were entered into a password protected computerized study database (Excel®). The study protocol was approved by the Conjoint Health Research Ethics Board, University of Calgary. In addition, all primary care givers signed a generic consent form agreeing to anonymous use of data collected during polysomnography for research purposes.
Patient selection
Records of children who were diagnosed with OSA/SDB and underwent a titration polysomnography to adjust PAP settings between October 1999 and December 31, 2011 were reviewed. All subjects aged 1 to 18 years were included in the study providing they had no major known clinical factors that could contribute to OSA/SDB severity other than obesity and/or Down syndrome (eg. craniofacial abnormalities, cerebral palsy, tracheotomy and other genetic syndromes). Otherwise typical subjects meeting inclusion criteria were classified as non-obese (BMI < 85th percentile), overweight (BMI 85th – 95th percentile) or obese (BMI > 95th percentile). Due to the standardized use of specialized Down syndrome growth charts at our institution, recent weight and heights were frequently not recorded in the lab supplementary charts of children with Down syndrome; their weight classifications were based on the sleep technologists’ report recorded on the night of PSG. As part of the general assessment on the night of study, the night technologists complete a checklist form describing observations about the child and their breathing during the night (available upon request). They are asked to classify patients into non-obese, mildly, moderately or severely obese based on their appearance. For the purposes of our analyses, we classified children with Down syndrome as non-obese if described as normal; overweight if reported as mildly obese, and obese if reported as moderately or severely obese by the night technologist.
Laboratory polysomnography
Computerized laboratory PSG (Sandman® NT, Nellcor Puritan Bennett, Ottawa, ON) was performed according to American Thoracic Society guidelines at the Alberta Children’s Hospital sleep laboratory [21]. Monitoring included electroencephalogram (F4-M1, C4-M1, O2-M1; with the backup electrodes F3-M2, C3-M2, O1- M2). Prior to the 2007 AASM scoring manual [22], we referenced derivations to the auricular instead of mastoid electrodes. In addition, bilateral electrooculograms, submental, intercostal, diaphragmatic and bilateral tibial electromyograms, electrocardiogram, oxygen saturation monitoring, chest, abdominal wall, and sum channel movements using respiratory inductance plethysmography, nasal/oral airflow using a thermistor device, nasal pressure, end-tidal carbon dioxide monitoring, and transcutaneous carbon dioxide monitoring were recorded. A registered sleep technologist (RPSGT) scored sleep architecture and respiratory events using standard scoring criteria [23]. Prior to release of the American Academy of Sleep Medicine (AASM) Manual for the Scoring of Sleep and Associated Events in 2007, our lab used the standard sleep scoring system published by Rechtschaffen and the following definitions of events [24]. Apnea was defined as an 80% or greater decrease in amplitude on the respiratory inductive plethysmographic sum channel. Secondary ancillary surrogate measures of airflow included nasal pressure and thermistor signals. Mixed and obstructive events were counted as significant if greater than two breaths in duration, whereas central apneas were counted only if accompanied by a decrease in SaO2 ≥ 4% from baseline or if ≥ 20 seconds in duration. Hypopnea was defined as at least a 50% decrease in amplitude associated with ≥ 4% drop in SaO2 from baseline on the respiratory inductive plethysmographic sum channel and/or evidence of cortical arousal. Central, mixed/obstructive, and total apnea/hypopnea (AHI) indices were calculated as number of events per hour of sleep. Obstructive sleep apnea was defined as an overall apnea hypopnea index of greater than 1.5 events per hour of total sleep time [24,25].
Positive airway pressure titration
PAP is titrated in our laboratory using a standard protocol (available upon request). Briefly, technologists are instructed to start PAP at a low pressure setting and increase as needed to eliminate respiratory events and hypoventilation. If there is evidence of flow limitation present on the nasal pressure channel, the technologists may increase PAP pressure further in an attempt to normalize that flow signal. The interpreting physicians (three in total over the time period of this study) review the raw data of the study and titration and choose a PAP pressure that is most efficient at eliminating events, normalizing gas exchange. If the pressure at that point is low and there is persistent evidence of flow limitation with arousals on the nasal pressure signal, a higher PAP pressure will be prescribed. For the purpose of this study, we identified a clinically significant difference in PAP pressure of two cmH2O.
Given the subjectivity involved in titrating PAP in the lab using written protocols, pressure settings within two cmH2O were considered more or less equal.
Data analysis
Demographic variables were summarized via means and standard deviation for numerical variables. Categorical variables were summarized via proportions. Two way ANOVA (allowing for interaction between the two factors) was used to compare mean PAP received between groups. Tukey Honest Significant Differences were used to illustrate pair-wise comparisons. Multiple linear regression models were used to assess association between both demographic and polysomnographic variables and PAP settings such as Down syndrome status, oxygen saturation and apnea hypopnea index [26].


241 subjects met inclusion criteria for one of the six subgroups of interest: Group 1a (typically developing, non-obese, n=62), Group 1b (typically developing, overweight, n=17), Group 1c (typically developing, obese, n=59), Group 2a (Down syndrome, non=-obese n=75), Group 2b (Down syndrome, overweight, n=15) and Group 2c (Down syndrome and obese, n=13). Subject characteristics are summarized in Table 1. Overall, the groups of otherwise typical children with OSA/SDB had a median age of 9 years while the group with Down Syndrome had a median age of 7 and their median carbon dioxide levels measured during sleep were 45.4 and 46.4, respectively.
Table 1: Demographic and polysomnographic characteristics of children treated with PAP (median, IQR shown).
Two way ANOVA was used to determine if differences in pressure used to treat OSA/SDB existed by both weight status and whether or not the subject had Down syndrome. A significant interaction between weight status and Down syndrome status (P=0.008) was identified, meaning an effect modification of weight status by Down syndrome status. Briefly, the presence of Down syndrome made a difference in PAP levels in each of the non-obese and overweight groups but did not make a difference in PAP levels for the most obese group when comparing to otherwise healthy subjects. Table 2 shows the mean PAP for the six groups. Subjects having Down syndrome had a higher PAP by 1.84 (95% CI (0.23 to 3.45) cm H2O when compared to those without for those children with normal weight. This difference decreased as weight increased. The pressure used to treat OSA/SDB was statistically significantly higher among nonobese and overweight Down syndrome groups in comparison with otherwise healthy children. Obese Down syndrome subjects and obese otherwise healthy subjects were prescribed similar PAP levels to treat their OSA/SDB (Table 2).
Table 2: Comparison of PAP levels among children of various weight categories with and without Down. syndrome.
We analysed the effect of weight and Down syndrome status on AHI. The assumption of constant variance was violated in the ANOVA analysis for AHI. Thus, logarithm-transformed AHI was analyzed in relation to both Down syndrome status and weight. Interaction between weight and Down syndrome status was significant, and this is reflected in both non-obese and overweight children with Down syndrome having statistically significant higher AHI compared to otherwise typical children in the same category; meanwhile, such difference was not seen between the two obese groups.
Looking into the factors that may have affected how PAP settings were set up among patients with OSA/SDB, we studied the relationship between PAP and AHI, age, weight, Down syndrome status, minimum and mean saturation, sex and proportion of time with saturation under 90%. We used these variables as we a) had them and b) seemed to be clinically relevant. The informed regression modelling showed that Down syndrome, AHI and age at polysomnography were factors significantly associated with prescribed PAP pressures. In the presence of AHI, weight status did not add significantly to the prediction or difference of PAP. The linear relationship of these predictors with PAP only explained 23% of the total variability in PAP. The effects of the regression are shown in Table 3 and we include Figure 1 to illustrate what the linear relationship in this set of children was estimated to be between PAP and each of age and AHI (separately) with Downs.
Table 3: Regression coefficients.
Figure 1: These plots illustrate the estimated effect of each AHI (a) and Age (b) separately on initial PAP pressure settings for children with and without Down syndrome.


Our data clearly shows that Down syndrome status was an important factor related to prescribed PAP levels in our population. Although weight was a significant factor, we found that the presence of Down syndrome changed the way weight status affected PAP when the two way anova with interaction was carried out. Children with Down syndrome have hypotonia which predisposes them to upper airway collapse during sleep. In addition, they have a relatively small hard palate, reduced mandibular volume, and a general reduction in airway volume caused by soft tissue crowding [27]. These differences likely explain our finding that children with Down syndrome required higher PAP settings to treat their OSA/SDB.
Although the presence of Down syndrome had the most significant effect on PAP pressure in our subjects (when only looking at Down’s syndrome and weight), this effect only held true for those subjects who were either non-obese or overweight. When exploring other factors that could impact PAP, we identified that age of subject at the time of the titration PSG also had an important role in PAP setting. A previous analysis in 2005 of PAP pressure settings in our lab identified age of child as an important predictor of pressure (unpublished data) with children less than 12 yr of age typically requiring a lower settings (mean 7cm H2O) compared to older children (mean 8cm H2O). This could be related to changes in muscle and fat mass as well as hormonal changes that children undergo at this age. Having said that, there is little data in the published literature to support this theory.
Lastly, we showed that AHI correlated with PAP pressure setting in our subjects. The AHI is generally used to grade severity of OSA/ SDB. There is conflicting data with respect to the relationship between severity of OSA/SDB and the AHI in children with obesity. Canapari et al. reported that visceral fat distribution is strongly predictive of AHI, whereas age, gender and BMI Z score were not predictive values [28]. However, others have shown that actually BMI Z score was significantly correlating with log-transformed AHI score [10,29]. Our data showed that children with Down syndrome in our sample had a higher mean AHI (19.0/hr vs 16.2/hr non-Down syndrome) and that both AHI and Down syndrome status were independent predictors of PAP setting. It has been well documented in the literature and further confirmed in our study population that children with Down syndrome who suffer from OSA/SDB often have hypoventilation during sleep documented by increase in end tidal carbon dioxide [10].
Our data was collected over a long period of time during which scoring criteria was changed based on AASM standards. At times, data collection was incomplete and BMI percentiles were not available for many of the children with Down syndrome. These issues all result in limitations in the interpretation of our findings.
Predicting the approximate PAP required to treat OSA/SDB should be ultimately based on clinical outcomes after specific PAP settings. Starting a patient on PAP therapy based on information from the patient, and based on a model that has a better fit and good prediction, followed by obvious clinical improvement, might eliminate the need for routine laboratory titration polysomnography studies which are both time consuming and financially expensive particularly in the pediatric population [30]. Our analysis only showed us what seems to have been playing a role in how children were prescribed in our dataset, however, its low percentage of variability explained does not allow us to think of it as a prediction tool, but it sets the stage for looking into more variables that could influence setting PAP levels and for searching into other prediction methods and assess their usefulness.
In conclusion, we have shown that the PAP level needed to treat OSA/SDB in our pediatic population is mainly dependent on the presence or absence of Down syndrome with further influence of age and disease severity.


The authors have no conflicts of interest to declare. No off-label products were used in this study. Financial support in part, was provided by the Alberta Children’s Hospital Foundation.


Track Your Manuscript

Share This Page

Media Partners