Introduction
The assessment of obstructive sleep apnea (OSA) in infants shares many similarities with older children. Polysomnography and clinical assessment are the cornerstones in both. Infants, however, have considerable changes in sleep and breathing across the first year of life, so establishing thresholds for testing results is a challenge. Given this challenge, establishing a firm diagnosis of OSA in an infant can be difficult. Polysomnography (PSG) is the accepted standard for the diagnosis of OSA across childhood, though is not available in many jurisdictions. Understanding how developmental changes in breathing may alter PSG interpretation as well as an understanding of alternative tools for evaluation is important for the identification of OSA in infants as well as considering treatment decisions.
Developmental changes in sleep and breathing that may impact OSA evaluation
Sleep and breathing maturation is a crucial aspect of development and undergoes important changes in early infancy which may impact identification of abnormal sleep or breathing in the first years of life. Both sleep and breathing are evident as physiological processes as early as 25- and 10-weeks’ gestational age, respectively. Neither sleep nor breathing control systems are fully developed at birth and both are not likely to fully mature until late in the first decade of life or into adolescence. With sleep occupying 16 to 18 h/day after birth and decreasing across the first year of life, it clearly plays an important function in early development. Longitudinal studies of breathing response show that healthy infants have greater variability in change in ventilation and arousal from sleep in response to hypoxia at lower postnatal age, with this variability decreasing in the first months of life. These changes in sleep and breathing in early life mean that what can be considered appropriate sleep and breathing changes across infancy.
While sleep and breathing are independent processes, the relationship between sleep and breathing is important to understand when assessing either separately or the interaction between the two. Sleep is divided into different stages, with infants having a predominance of rapid eye movement (REM, also called active in infants) sleep at birth; REM occupies 50% of total sleep time at birth, decreasing such that slow wave sleep (SWS, stage N3 sleep, or deep sleep) is the predominate stage by 1 year of age. , During REM sleep, there is a reduction of the tonic activity of the diaphragm and intercostal muscles that, along with a complaint chest wall, results in paradoxical inward rib cage movement associated with lower oxygen saturation. Respiratory frequency and stability is more variable during REM sleep which, alongside changes in chest wall activity, means that respiratory events, including central, obstructive, and oxygen desaturation events, are more common during REM sleep than other sleep stages. Changes in chest wall mechanics likely also explain lower and more variable oxygen saturations in REM compared to other sleep stages. Like REM sleep, the frequency of respiratory events in healthy infants decreases across the first year of life. Arousals from sleep, an important protective mechanism, are also more common in REM sleep and decrease over the first year of life. Sleep is cyclical across the night with shorter cycles in infants. As nocturnal sleep is established after 3 months of age, there is also a shift such that SWS predominates during daytime sleep as well as in the first part of the night and REM in the later part of the night. These features mean that age-related changes predispose even healthy infants to higher numbers of respiratory events in early life which complicates the identification of OSA.
Primary tools for assessing obstructive sleep apnea in infants
The early interest and investigations of OSA in infants stemmed from a potential association between adult OSA or familial OSA and sudden infant death syndrome (SIDS) and subsequent interest in a potential association between OSA in infants and risk of SIDS. Tools used to assess breathing during sleep in at-risk infants included different types of movement monitors connected to alarms to detect cessation of breathing movement, 12- to 24-hour polygraphy recording a combination of electroencephalograms (EEG), electrocardiograms (ECG), and/or respiratory activity, and transcutaneous O 2 and CO 2 monitoring with less information about clinical or airway assessments. Today, attended in-laboratory polysomnography (PSG) and clinical assessment are the cornerstone to assessing OSA in infants.
Polysomnography: PSG is the accepted standard for the diagnosis of OSA in children including infants though whether different standards are needed for the definition of OSA in infants is controversial. The American Academy of Sleep Medicine (AASM) published infant sleep scoring criteria in 2015 completing the challenge of standardized rules from 0 to 17 years of age. The International Classification of Sleep Disorders (ICSD) 3rd edition defines pediatric OSA based on the presence of at least one symptom of OSA (i.e., snoring, labored/paradoxical/obstructive breathing during sleep, sleepiness/hyperactivity/behavioral/learning problems) and PSG results demonstrating at least one of: (1) an obstructive-mixed apnea-hypopnea index (OMAHI) >1 event/second; or (2) a pattern of obstructive hypoventilation (i.e., P a CO 2 >50 mmHg for >25% of total sleep time) with snoring, flattening of the inspiratory nasal pressure waveform and/or paradoxical thoracoabdominal motion. While the definition acknowledges that adult diagnostic criteria may be used for ages 13 to 18 years of age, infants are not considered separately. Other authors have recommended infant OSA definitions that include apnea-hypopnea index (AHI) >1 event/hour, AHI ≥2 events/hour, obstructive apnea-hypopnea index (OAHI) ≥2 events/hour, mixed obstructive apnea index (MOAI) >3 events/hour, obstructive mixed apnea-hypopnea index (OMAHI) >3 events/hour, apnea index (AI) >5 events/hour, obstructive respiratory distress index (ORDI) >5 events/hour, and AHI >10 events/hour. The range of variables used to define OSA and cut-offs show variability in what is considered OSA in infants with values that overlap with data from otherwise healthy infants.
Normative PSG data in infants comes both from older studies, where definitions of respiratory events differ from the currently recommended standards, as well as studies that have defined respiratory events based on AASM criteria. In a review of normative PSG data by Ng and Chan the most common definition of apnea across studies was based on time (events ≥3 seconds) rather than the two missed breaths used in the AASM rules. , The authors note that this is in contrast to events considered significant in neonatology being ≥20 seconds. , Three of the included studies provided age-related trends for the apnea index and showed apnea indices were highest in the neonatal period with decreasing trends for obstructive, central, and mixed apneas indices with increasing age in the first year of life. , , Median values for obstructive apnea index (OAI) were 0 events/hour across all studies with two studies reporting 90th percentiles of 0.6 to 0.7 events/hour from 2 weeks of age and 0.2 to 0.4 events/hour by 2 to 3 months of age. Of note, hypopneas are not included in these events. A more recent summary of PSG parameters is provided by Daftary and colleagues in their study of PSG values in healthy infants. The results include ranges for typical PSG parameters that highlight the high degree of variability for healthy infants <30 days of age. This includes AHI (1.0–37.7 events/hour), OAI (0.2–12.5 events/hour), central apnea index (0–27.2 events/hour), mixed apnea index (0–8.3 events/hour), and hypopnea index (0.7–12.9 events/hour). The authors provide a summary table comparing their results to those from three other studies and demonstrate a trend for decreasing respiratory events from <30 days of age to 1 to 2 years of age. Not all studies used the AASM criteria for collecting and scoring PSG data. The three studies that included sleep staging showed a mean AHI of 14.9 events/hour <30 days, 21.4 events/hour <45 days of age, and 2.8 events/hour at 1 to 2 years of age. There was less variation by age of the median duration of respiratory events. This decrease in respiratory events with age mirrors that described in the Childhood Home Monitoring Evaluation, where extreme events decreased from birth to 43 weeks gestational age in both term and preterm infants, and what is describe in a study of infants <2 years of age undergoing PSG. While further studies with the same measurement methodology are needed to more accurately define the upper limits of normal for different respiratory events, the results highlight variability in the number of respiratory events in otherwise healthy infants in early life as well as decreasing events from birth through the first years of life.
Clinical assessment: Clinical history alone is insufficient to distinguish OSA from non-OAS in snoring children. Combining this clinical information with PSG results is important to add to the assessment of OSA risk and guide decisions around treatment, monitoring, and follow-up. Symptoms of OSA in infants differ in some respects from those of older children. Parents or caregivers may not identify noises made by an infant as “snoring” so probing about noisy or audible breathing during sleep may be needed. The absence of snoring or noisy breathing does not rule out OSA as complete airway obstruction is silent. Apnea, while a primary feature of OSA, may be more difficult to recognize in infants because of more subtle respiratory effort. Difficulties with feeding or tiring with feeding may be a sign of airway obstruction, given the challenge of coordinating suck and swallow with a compromised airway. With infants sleeping for a large proportion of the 24 hour day, it may be difficult to appreciate a disruption in sleep duration or daytime sleepiness; inquiring about periods where an infant is awake and alert may be helpful to assess whether sleep is restorative even when awake periods are relatively short. Sweating during sleep is not uncommon in otherwise healthy infants though may be important in an infant with other features of OSA. Similarly, restlessness can be difficult to assess in infants who have not yet developed suppression of muscle activity in REM sleep though may point to struggles to open an obstructed airway in an infant with other symptoms of OSA. Asking specifically about tolerance to different sleep positions is important as parents or caregivers may not admit to trying a position other than supine sleep given the recommendations for “back to sleep” because of SIDS risk. They may, however, admit if asked that they have noted differences in snoring, noisy breathing, or other OSA-related symptoms when the infant sleeps prone on their chest or otherwise supervised during sleep. Recurrent respiratory tract infections, which are also common in infants, can be related to OSA because narrower airways become congested more easily, and disrupted sleep may alter immune function. In a retrospective chart review of infants with OSA, the most common presenting symptoms were pauses in breathing (27%), followed by a combination of characteristics (22%) and snoring (18%). In another cohort of infants with OSA, snoring (53%) and nocturnal desaturations (24%) were the most common indications for PSG. While there is a broad range of potential symptoms of OSA, focusing on the concerns of an individual infant can highlight important features of OSA in that infant.
Physical examination adds additional important information about OSA risk in infants. Growth measurements, ideally serial ones, are useful to assess growth as additional energy expended to support breathing during sleep or decreased oral intake related to difficulty in feeding can impact growth. In an early study of 14 infants with OSA, failure to thrive was the main complaint for 20%. Mid-face hypoplasia or retro/micrognathia may indicate a smaller airway that is more likely to be compromised during sleep. This can be best assessed by looking at the infant from the side where the size of the lower jaw (mandible) can be more easily assessed in relationship to the upper jaw (maxilla). Tonsillar hypertrophy may be seen in infants beyond 3 months of age though is a less common feature in infants versus children with OSA. Developmental milestones are important as impaired sleep may impact development, and impaired development may be a flag to other causes of developmental delay, including musculoskeletal and neurological conditions, that may heighten OSA risk. Poor weight gain or developmental progress may also influence decisions about follow-up when a diagnosis of OSA is unclear; early reassessment may be indicated for an infant with these concerns compared to an infant with the same symptoms and PSG results who has appropriate growth and development. The clinical assessment provides important information about the impact of OSA and the potential benefits of treatment to important measures of overall health.
Risk factors are important to consider, though not all infants with any particular risk factor will develop OSA. Common risk factors for OSA in infants include craniofacial anomalies, prematurity, gastroesophageal reflux, and adenotonsillar hypertrophy. Hypotonia and central nervous system immaturity can each contribute to OSA risk because of their potential effect of compromising airway function and response to respiratory challenges, such as sleep and acute respiratory tract infection. In a study of 139 infants 0 to 17 months of age with OSA, the most common comorbidity was gastroesophageal reflux disease (GERD, 68%), followed by craniofacial disease (37%), neuromuscular disease (34%), prematurity (30%), genetic syndrome (30%), laryngo/tracheomalacia (27%), and epilepsy (17%). In another study, the most common risk factors for OSA were hypotonia (53%), GERD (30%), and laryngomalacia (24%), with 34% of infants having a genetic abnormality. Understanding the risk factors for OSA may help to identify OSA earlier in its course, especially in infants with complex medical illnesses.
Additional tools as substitutes or additional assessment in infants with suspect OSA
While PSG is the accepted standard for the diagnosis of OSA in infants and children, access to PSG is limited, unavailable, or subject to long wait times in many jurisdictions. As a result, PSG may not be an option and alternative tools may be used to add objective or additional information to the clinical assessment of infants with suspected OSA.
Overnight Oximetry: Continuous overnight oximetry is a common tool used for the assessment of OSA in children with some challenges and fewer studies about its use in infants. Similar to other measures related to breathing, oxygen pulse saturations (S p O 2 ) in healthy infants are more variable at younger age with the range between minimum and maximum S p O 2 narrowing and shifting toward higher values from 2 weeks to 24 months of age ( Fig. 6.1 ). Based on S p O 2 desaturation of ≥3%, desaturations are more frequent at 2 weeks of age than 24 months, decreasing from a mean of 27.2 events/hour to 3.3 events/hour. Desaturations are typically 3% to 4% from baseline and 5 to 6 seconds in duration. Compared to term infants at the same postnatal age, infants born preterm spend longer at S p O 2 <96% and have more desaturation events which are longer and deeper. Of note, the cumulative frequency plots displayed in Fig. 6.1 are not the typical way that S p O 2 is summarized in PSGs or other clinical testing. Cumulative frequency plots have the benefit of including measurement of the variability of S p O 2 and can be used to look at S p O 2 distribution as well as desaturation incidence, depth, and duration. Regardless of how S p O 2 data are analyzed and displayed, understanding how the data will change across the first years of life is vital to what the data can tell you.
A commonly used S p O 2 score, the McGill Oximetry score (MOS), was developed to identify OSA in children from overnight S p O 2 recordings. The dataset included children 6 months to 18 years of age and excluded children with comorbidities including central nervous system disease, airway and lung disease, neuromuscular disease, GERD, and other syndromes or genetic abnormalities. The MOS uses patterns of S p O 2 decreases ≥4% to classify overnight oximetry recordings as positive (3 or more desaturation clusters and at least 3 desaturations to <90%), negative (no desaturation to <90%), or inconclusive (did not meet criteria for positive or negative). Desaturation clusters are defined as 5 or more desaturations occurring in a 10- to 30-minute period. This scoring typically requires manual review of the S p O 2 data across the night as desaturation clusters are not the same as the oxygen desaturation index (ODI) or desaturation index (DI; i.e., number of desaturation events/hour of recording time) which are commonly provided by S p O 2 machine software. The scoring criteria were further modified to enable prioritization of adenotonsillectomy based on severity ( Table 6.1 ) though it is unclear if infants were included in the validation portion of the study. The MOS was compared to PSG results in a study of 53 infants under 1 year of age with laryngomalacia and clinical suspicion of OSA, 46 of whom had OSA, and included 14 infants with additional comorbidities. Using a PSG definition of OSA of OAHI ≥2 events/hour, the MOS showed a sensitivity of 91% and specificity of 25% with positive and negative predictive values of 91% and 0%, respectively. Using a revised MOS definition of two desaturation clusters and one desaturation to <90% worsened OSA prediction (sensitivity 89%, specificity 0%). The MOS provides a systematic way to evaluate overnight S p O 2 recordings and its relatively high sensitivity and low specificity may be useful in ruling out OSA in infants with a negative test result though it should not be used to confirm OSA.
| Score | Description | Clusters | S a O 2 | ||
|---|---|---|---|---|---|
| # drop <90% | # drop <85% | # drop <80% | |||
| 1 | Normal/inconclusive | <3, baseline S p O 2 >95% | <3 | 0 | 0 |
| 2 | Mild OSA | ≥3 | ≥3 | ≤3 | 0 |
| 3 | Moderate OSA | ≥3 | ≥3 | >3 | ≤3 |
| 4 | Severe OSA | ≥3 | ≥3 | >3 | >3 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
