Ambulatory Sleep Monitoring





Polysomnography is a diagnostic sleep medicine tool in which multiple different physiological parameters are simultaneously recorded across a sleep period to characterize sleep and identify sleep disorders. The resulting study is called a polysomnogram (PSG) or sleep study. PSG becomes a much more powerful diagnostic tool than could be provided by recording only one or two simultaneous measurements. Simultaneous recording of multiple physiological measurements permits correlation of how a particular change in sleep/wake state or specific abnormalities in one biological signal affect other signals (eg, a prolonged sinus pause and severe oxyhemoglobin desaturation occurring coincidentally with a long-lasting obstructive apnea during rapid eye movement (REM) sleep or brief periodic limb movements of sleep (PLMS), some causing arousals from nonrapid eye movement (NREM) sleep).

In-laboratory overnight PSG remains the gold standard for confirming a clinical diagnosis of sleep disordered breathing, most often obstructive sleep apnea (OSA). Prevalent and consequential, OSA affects 10% of men and 3% of women ages 30 to 49 years, 17% of men and 9% of women ages 50 to 70 years (1). OSA often underlies pharmacoresistant hypertension, recurrent atrial fibrillation, wake-up ischemic stroke, and obesity (2). Many cases of OSA go undiagnosed, in part because of lack of access (3). In-laboratory PSG is also indicated to diagnose narcolepsy type 1, REM sleep behavior disorder, and distinguish atypical parasomnias from sleep-related epilepsy.

The Centers for Medicaid and Medicare Services (CMS) (4) and the American Academy of Sleep Medicine (AASM) (5) classify sleep studies based on how many different sensors are recorded and whether a sleep technologist is present throughout the recording (attended or unattended). The gold standard PSG is classified as a level 1 study performed attended in a sleep laboratory recording a minimum of seven channels (usually 14–16) including electroencephalography (EEG), electro-oculography (EOG), submentalis electromyography (EMG) also known as chin EMG, EKG, airflow, thoracic and abdominal respiratory effort, and pulse oximetry (SpO2), video, and sometimes carbon dioxide (CO2) monitoring (6,7). A level 2 PSG is simply a level 1 PSG recorded unattended, in or out of the sleep laboratory. A level 3 study records a minimum of four channels including ventilation, oximetry, EKG or heart rate, and done at home or out of the sleep laboratory unattended. A level 4 study is typically done at home unattended and records 2 to 3 cardiorespiratory signals (usually pulse oximetry, sometimes airflow).

Level 1 PSG as a gold standard has been tarnished by high costs, limited access, long wait times, night-to-night variability, and first night effect (6,7). Simpler and less costly methods for recording sleep and sleep disorders have become increasingly requested and available. How long and well a person sleeps are measures of health and fitness, and a plethora of smartphone health applications are marketed that claim to provide information about sleep quality and disruption. We review what is known of these here.


Driven by long wait times, limited access, and patient preference, home sleep apnea testing (HSAT) is now accepted as a reasonable alternative to level 1 PSG for diagnosing and treating OSA. In less than a decade, HSAT, most often using level 3 devices, is rapidly replacing level 1 PSG as the first line diagnostic test for identifying OSA in adults in the United States. Driven in no small part by cost containment, CMS advised in 2008 it would accept prescriptions for positive airway pressure (PAP) therapy to treat OSA in adults diagnosed using either level 1, 2, or 3 devices (and a level 4 device if at least 3 variables were recorded) (8). Within less than a decade, this has led to a seismic change in how OSA is diagnosed and treated in adults in the United States. Before availability and third party payer approval of HSAT, the only acceptable strategy for diagnosing and treating OSA in adults in the United States was to perform a level 1 PSG followed by in-lab continuous positive airway pressure (CPAP) titration. This often requires two nights of PSG, but sometimes CPAP can be introduced after 2 hours of recording time identifying sufficient OSA to titrate CPAP (called a “split night PSG”).

Driven by increasing third-party payer demand and patient preference, HSAT has reduced the annual volume of level 1 PSG done in AASM-accredited sleep centers by 40% or more and has led to recent closure (or even bankruptcy) of many sleep centers (9). A 2013 survey of AASM-accredited sleep centers found that 64% of centers were offering adult HSAT and 48% reducing their plans for expansion of laboratory beds as a result of home testing (10).

Availability of HSAT has led to at least three different clinical pathways by which adults with OSA can be diagnosed and prescribed PAP therapy (Flow Chart 8.1) (11). Patients who have an HSAT positive for OSA (apnea-hypopnea index [AHI] ≥ 5/h) can (a) undergo an in-laboratory level 1 PSG titrating PAP, and when scored and read be prescribed particular PAP therapies; (b) titration of CPAP using an auto-adjusting CPAP device (APAP); or (c) APAP treatment without titration. At least four well-designed studies have shown adults with OSA diagnosed by HSAT then prescribed either APAP therapy or APAP titration at home had similar PAP compliance, nightly use, discontinuation rates, and improvement in sleepiness as did subjects in whom CPAP was titrated by level 1 PSG and then prescribed CPAP (11–14). HSAT followed by APAP titration and treatment permits a more accessible and cost-effective alternative for diagnosing and treating OSA in adults without compromising PAP compliance.

HSAT is most likely to confirm a diagnosis of OSA in patients who have a moderate-to-high pretest probability of OSA (Table 8.1) (5). In 2007, the AASM published clinical practice guidelines for performing HSAT (Table 8.2) (5). These recommend HSAT on patients with (a) a high pretest probability of moderate-to-severe OSA and without significant comorbid medical conditions that require level 1 PSG; (b) level 1 PSG is not possible by virtue of immobility, safety, or critical illness; (c) to retest a patient after surgery for OSA, after initial use and adjustment of an oral appliance to treat OSA, or to reassess a patient with OSA after weight loss or weight gain; and/or (d) after two negative or technically inadequate HSAT tests in patients at high risk for OSA (5,15). Table 8.3 shows those who are usually not suitable for HSAT.


FLOW CHART 8.1 Alternative Clinical Pathways for Starting PAP Following a Diagnosis of Obstructive Sleep Apnea (OSA) by Home Sleep Apnea Testing (HSAT)

TABLE 8.1  Adults With a High Pretest Probability That Obstructive Sleep Apnea Will Be Found on Home Sleep Apnea Testing (HSAT)

  A history of habitual loud snoring, moderate-to-severe daytime sleepiness, and witnessed apneas when sleeping

  Obese (BMI >35 kg/m2), increased neck circumference (>43 cm/17 in. in men, >40 cm/16 in. in women), tonsillar hypertrophy, craniofacial anomaly, and/or retrognathia

  Snoring, weight gain, BMI >30 kg/m2, and mild daytime sleepiness

  Comorbid congestive heart failure, atrial fibrillation, treatment-resistant systemic hypertension, diabetes mellitus type 2, ischemic stroke, TIA, nocturnal cardiac dysrhythmias, pulmonary hypertension

  History of an accident or near miss at work or when driving that could relate to sleepiness

  Commercial drivers with large neck size or obesity

  Considering surgery for snoring, mandibular advancement splint, and bariatric surgery

Source: Adapted from Ref. (5). Collop NA, Anderson WM, Boehlecke B, et al. Clinical guidelines for the use of unattended portable monitors in the diagnosis of obstructive sleep apnea in adult patients. Portable Monitoring Task Force of the American Academy of Sleep Medicine. J Clin Sleep Med. 2007;3(7):737–747.

TABLE 8.2  AASM Clinical Guidelines for HSAT

  HSAT should be performed on patients with a high pretest probability of moderate-to-severe obstructive sleep apnea (OSA) and without significant comorbid medical conditions which require more extensive polysomnographic monitoring.

      May be indicated to diagnose OSA in patients for whom in-laboratory PSG is not possible by virtue of immobility, safety, or critical illness.

      Could be used to monitor response to non-CPAP treatments for OSA (eg, oral appliance or positional devices).

  The HSAT device should record a minimum of three signals (airflow, respiratory effort, and blood oxygenation).

      The sensors used to record HSAT should be the same as those used in level 1 PSG.

      Allow display of raw data capable of manual scoring or editing of automated scoring by a qualified sleep technologist.

  HSAT should be done in conjunction with a comprehensive sleep evaluation, be recorded in an AASM-accredited sleep center, and read by an AASM board-certified (or eligible) sleep specialist.

      An experienced sleep technologist must apply the sensors and educate the patient in sensor application.

      The HSAT test raw data should be available for display to assess their quality. The study should be reviewed and interpreted by a board-certified or eligible sleep specialist using scoring criteria consistent with published AASM standards.

      Patients undergoing HSAT should have a follow-up visit with a sleep specialist to discuss the test results.

      Two negative or technically inadequate HSAT tests in patients with a high pretest probability of moderate to severe OSA should prompt performing a level 1 PSG.

Source: Adapted from Ref. (5). Collop NA, Anderson WM, Boehlecke B, et al. Clinical guidelines for the use of unattended portable monitors in the diagnosis of obstructive sleep apnea in adult patients. Portable Monitoring Task Force of the American Academy of Sleep Medicine. J Clin Sleep Med. 2007;3(7):737–747.

TABLE 8.3  Comorbid Conditions That Warrant In-Laboratory Polysomnography

  BMI >50

  Obesity-hypoventilation syndrome (paCO2 awake >45 and BMI >30)

  Chronic obstructive pulmonary disease (COPD) with a FEV1 <60%

  Moderate-to-severe cardiac disease (NYHA class III or IV)

  Moderate-to-severe neuromuscular disease (FEV <60–80% of predicted)

  Documented pulmonary hypertension >40 mmHg

  Hypoxia and/or hypercapnia at rest, or oxygen dependent for any reason

  Chronic nightly narcotic use

  Violent behaviors during sleep, REM sleep behavior disorder, atypical parasomnias, sleep-related epilepsy

  Ability to understand instructions

  History of central sleep apnea

  Inadequate HSAT results after ≥2 attempts

  Unsuitable home environment (noise level, partner/family interactions, distance from sleep laboratory)

  Severe intellectual, physical, and/or psychiatric disability with inadequate caregiver attendance

  Second opinion where symptoms/results of previous testing do not equate with the clinical impression, the original diagnosis is uncertain, and/or serious medical–legal consequences may be relevant

  Sleep duration typically less than 4 hours per night

Source: Adapted from Ref. (5). Collop NA, Anderson WM, Boehlecke B, et al. Clinical guidelines for the use of unattended portable monitors in the diagnosis of obstructive sleep apnea in adult patients. Portable Monitoring Task Force of the American Academy of Sleep Medicine. J Clin Sleep Med. 2007;3(7):737–747.

Level 3 Cardiorespiratory Devices Most Often Used for HSAT

Most often, level 3 cardiorespiratory devices are used to record HSAT in AASM-certified sleep centers. Level 3 devices vary widely in the number and types of biological signals recorded, technical specifications, and limitations (16). These typically record at least 4 to 7 cardiorespiratory variables, most often airflow, respiratory effort, heart rate or EKG, and SpO2. Airflow is usually monitored by nasal pressure (more reliable for detecting hypopneas) and/or thermal sensors (more reliable for detecting apneas), and respiratory effort by thoracic and abdominal belts using respiratory inductance plethysmography (RIP) or a chest wall impedance monitor. The heart rate is often derived from the pulse oximetry signal. Some devices record snoring by microphone or vibration detection, and many record body or head position. Most level 3 HSAT devices do not record EEG so they cannot identify sleep/wake stages or arousals.

In 2011, the AASM published an evidence-based technology evaluation to help clinicians decide how to select devices appropriate for HSAT (16). These classify HSAT devices using the so-called SCOPER system (Sleep, Cardiac measures, Oximetry, Position, Effort, and Respiration). The evaluation found that all HSAT devices need to record pulse oximetry. If airflow is recorded using either nasal pressure or a thermal sensor, then thoracic and abdominal respiratory effort should also be recorded (in order to identify central from obstructive and mixed apneas (16).

The AASM recommends selecting an HSAT device that has been validated against level 1 PSG, records pulse oximetry, and measures respiratory events with at least one airflow sensor and two respiratory effort belts or peripheral arterial tonometry (PAT) and another validated measure of respiratory disturbance (15). Of note, nasal pressure airflow sensors more reliably identify hypopneas and respiratory event related arousals (RERAs) than thermal airflow sensors. However, apneas are more reliably identified using thermal sensors.

One novel level 4 HSAT device available warrants particular attention. Worn on the nondominant wrist, the Watch-PAT (Itamar Medical Ltd., Israel) records PAT, actigraphy, pulse oximetry, body position, and snoring. Using a fully automated proprietary algorithm, it scores respiratory events based on PAT, oxygen desaturation, and heart rate changes. It provides surrogate measures of sleep, wake, NREM, and REM sleep using actigraphy and PAT as surrogate measures for sleep. Because Watch-PAT does not record airflow, it can be used to test the efficacy of CPAP therapy. Many sleep specialists are bothered that data analysis by the Watch-PAT is unmodifiable (you read what you get). However, studies comparing the use of it with level 1 PSG show it reliable (17–20). Strong correlation for AHI scores (often 0.82–0.94), oxygen desaturation indexes (ODI, 0.9 or higher), but also for mean recording sleep times and recognition of REM sleep periods using surrogate measures for sleep when Watch-PAT was compared with level 1 PSG (17–20).

Level 4 HSAT Using Only Nocturnal Home Pulse Oximetry

Sleep specialists sometimes record nocturnal home pulse oximetry in adults or children screening for either extended periods of sleep-related hypoxemia or brief repetitive oxygen desaturations. Nocturnal home pulse oximetry alone in adults as a diagnostic test for OSA depends on whether the criteria used to define a positive are quantitative or qualitative (15). A study that scored oxygen desaturations in overnight oximetry in 200 adults with suspected OSA found using quantitative criteria (≥4% from baseline SpO2 and ≤90%) had a sensitivity of 41% and a specificity of 97% (15). Another study that used qualitative criteria (a pattern of brief repetitive arterial oxyhemoglobin desaturations and ≥3% fluctuations per hour of recording time as abnormal illustrated in Figure 8.1) on an overnight pulse oximetry in 240 adults with suspected OSA had a high sensitivity of 98%, a low specificity of 48%, a positive predictive value of 61% (likely to identify those who have OSA), but a negative predictive value of 97% (15).

Scoring Apneas and Hypopneas Using Home Sleep Apnea Devices

When using HSAT devices, which measure airflow and respiratory effort, apneas and hypopneas can be scored using the same rules for level 1 PSG (15). An apnea is scored if there is greater than or equal to 90% reduction in peak signal excursion of the airflow sensor for greater than or equal to 10 seconds, a hypopnea when there is a greater than or equal to 30% reduction in airflow sensor lasting for greater than or equal to 10 seconds, and associated with a 3% oxygen desaturation (CMS rule for hypopnea requires a 4% oxygen desaturation) (Figure 8.2). However, most HSAT are recorded using level 3 devices, which cannot identify sleep/wake states or arousals. HSAT, which do not record sleep (or a surrogate for it), calculate the mean number of apneas and hypopneas per hour of recording time (not sleep time). An AHI based on recording time can be falsely low AHI if much of it is spent lying in bed awake (21). Worse yet, hypopneas or RERAs that cause arousal without significant oxygen desaturation cannot be scored in most HSAT and contribute to spuriously low AHI (Figure 8.3) (21).


FIGURE 8.1 Overnight oximetry showing recurring desaturations that occur during periods of night suggestive of either REM sleep and/or sleeping supine.


FIGURE 8.2 A 2 minute epoch of home sleep apnea test (HSAT) showing three obstructive hypopneas, one obstructive apnea, and one central apnea from a HSAT.


FIGURE 8.3 Tracing from an HSAT showing obstructive hypopneas some of which cannot be scored because they do not cause ≥4% arterial oxygen desaturations.

Limitations, Advantages, and Disadvantages of HSAT

The limitations of HSAT vary with the level of device and sensors recorded. Different HSAT device manufacturers use different sensors and different scoring algorithms to identify respiratory events. The scoring algorithm for some is proprietary, which vendors usually refuse to share. Many devices do allow the raw data to be reviewed to judge the quality of the data. Some score the studies automatically and do not allow correction of the scoring. Automated scoring tends to underestimate the AHI compared with manual scoring (22). Automated scoring is especially likely to result in false-negative scores when OSA is mild in severity.

How HSAT are recorded and scored impacts on the validity of the results. It is best to either score the study manually or use automated scoring with manual correction (5). HSAT can be scored with strong agreement when using only the nasal pressure airflow signal with an inter-scorer agreement of 0.96 for AHI, 0.93 for AI, but only 0.75 for HI; less so when studies were scored solely using uncalibrated respiratory inductive plethysmography (RIP) with AHI 0.98, AI 0.66, and HI 0.78 (23).

False-negative rates for HSAT can be as high as 20%, often due to how respiratory events are tallied: AHI are calculated on the number of events per time device is on (15). If the patient is awake much of the recording time, the AHI can be falsely low. A recent study compared HSAT with level 1 PSG in 28 adults with Parkinson disease and moderate-to-severe OSA. They found HSAT underestimated AHI (by an average of 12 events per hour) and had a high technical failure rate (27%) (24).

Technically inadequate HSAT rates range from 3% to 18% depending on who sets up the device (technologist versus patient) and type of device used (10). False negative rates up to 17% have been reported (5). Night-to-night variability is greater in HSAT recorded in patients with mild OSA (AHI 5–15/h) (25). A recent retrospective study of 238 consecutive patients with either a normal or technically inadequate HSAT found (a) a normal HSAT was more likely to be found in patients younger than 50 years of age, and technically inadequate studies in those older than 50; (b) younger subjects with a normal HSAT were more likely to have a normal level 1 PSG (76%); and (c) the majority (71%) of patients with technically inadequate HSAT had OSA (and were often >50 years of age) (26). HSAT technical failures can be reduced by choosing an HSAT device that is simple to use, instructing patients on how to attach the sensors, giving the patient a handout with written instructions and a picture of the setup, and a phone number for the patient to call if problems arise.

Advantages of HSAT include: can be recorded at home in the patient’s natural sleep environment for more than one night if needed, easier access, reduced wait times for diagnosis and treatment, lower labor costs, and patient preference for home recordings (15,27). HSAT is far less expensive (often a tenth that of a level 1 PSG), preferred by patients, and associated with equal CPAP adherence outcomes compared with level 1 PSG (14,27,28).

Disadvantages of HSAT are many and include (a) a myriad of devices with many different sensors to choose; (b) most devices do not measure sleep; (c) absence of a trained sleep technologist to identify and fix artifacts, equipment adjustments, or intervene in medically unstable patients; (d) potential to misinterpret the results because of limited data, and (e) higher technical failure rate than PSG and greater potential for data distortion or loss (15).

A recent study performed a multicenter randomized controlled trial comparing home-based versus laboratory-based treatment pathways in 373 subjects at high risk for moderate to severe OSA (29). The laboratory-based pathway cost was $1,840 compared with $1,575 for the home-based pathway ($264 less when recorded at home). Per patient costs for the laboratory arm were $40 higher than the home arm ($1,697 vs. $1,736). The operating margin was $142 in the laboratory arm, compared with a loss of −$161 in the home arm. A home-based diagnostic pathway for OSA for third-party payers incurs fewer costs than a laboratory-based pathway (29). However, costs for the sleep center are comparable if not higher, resulting in a negative operating margin.


Actigraphy identifies sleep and wakefulness based on the tenet that body movements when awake are frequent and large, and absent or small during sleep (30,31). Actigraphs are compact, lightweight, computerized piezoelectric accelerometer-based microelectromechanical systems (MEMS), which use proprietary algorithm-based processing of the MEMS signal (Figure 8.4).

Usually worn on the nondominant wrist, actigraphs are capable of storing digitized data for extended periods, permitting the data to be transferred from the device to computer for scoring, analysis, and reporting. Wrist actigraphy provides information about the day-to-day timing, duration, and continuity of sleep in the natural sleep environment over long time periods (typically 1–2 weeks, sometimes months).

The International Classification of Sleep Disorders version 3 (ICSD-3) recommends actigraphy (a) to identify different sleep/wake patterns suggestive of circadian rhythm disorders, idiopathic hypersomnia, or insufficient sleep; (b) when evaluating patients for suspected narcolepsy for at least one week prior to multiple sleep latency testing to confirm the patient has obtained sufficient sleep over the week prior to ensure the results are valid; and (c) in patients with insomnia or circadian rhythm disorders to first characterize the sleep disorder and then evaluate the effects of treatment strategies. Actigraphy can provide objective evidence of sleep/wake activity in patients who provide poor sleep histories or sleep/wake complaints without clear explanation.

The Society of Behavioral Sleep Medicine recently published a scoring and instructional manual to assist clinicians and inform researchers in the use of actigraphy (32). The most important consideration when selecting an actigraph is whether the design is suitable and has been validated for the patient population(s) who will use it (32). The second most important consideration is cost, not only for the device, but the proprietary software, computer interface, batteries, licenses, warranty, maintenance, and technical support. So-called accessories such as an event marker are mandatory, a light meter very helpful, and a watch perhaps unnecessary. Table 8.4 summarizes technical considerations these recommend when purchasing a commercially available actigraph for use in patients or clinical research.


FIGURE 8.4 Actigraph showing delayed sleep phase type sleep/wake schedule.

How Reliable Is Actigraphy in Identifying Sleep and Wakefulness?

Actigraphy estimates sleep 0.88 to 0.95 compared with level 1 PSG in healthy populations. Actigraphy typically identifies sleep onset earlier than PSG especially among poor sleepers who may lie quietly trying to fall asleep (33). Actigraphy is more useful if accompanied by reports of bed, nap, and nonwear times chart on daily logs (not in the waiting room before follow-up). Brief daytime naps may go unrecognized. User fatigue limits how long an individual (or the caregiver) will continue to use and chart their use of it. Validation of actigraphy in patients with medical, neurological, psychiatric, and/or sleep disorders is limited. Before embarking on a study of such populations, it is best to review what has been done. The reliability of actigraphy is limited by whether the patient (or caregiver) is diligent in using it as directed.

TABLE 8.4  Society for Behavioral Sleep Medicine Technical Considerations When Purchasing an Actigraph for Use in Patients or Clinical Research

Only gold members can continue reading. Log In or Register to continue

Jul 9, 2018 | Posted by in NEUROLOGY | Comments Off on Ambulatory Sleep Monitoring
Premium Wordpress Themes by UFO Themes