Introduction

A polysomnogram (PSG), also known to patients as a sleep study, is an integral component of many sleep evaluations.[1][2] Attended (Type I) PSGs are the gold standard for diagnosing sleep-related breathing disorders (SRBD), including obstructive sleep apnea (OSA). Many other sleep disorders can also be diagnosed and managed based on PSG results. The study consists of a simultaneous recording of several physiologic parameters during sleep and wakefulness, including the electroencephalogram (EEG) to identify wake versus sleep and its stages.[1][3] 

This article focuses on sleep center-based, attended PSG, where the patient must stay overnight at a specialized sleep laboratory.[1] This article will provide an overview of the processes of obtaining and interpreting the attended PSG as a part of the diagnosis and management of a wide range of sleep complaints.

Specimen Collection

Several physiologic parameters are recorded during a diagnostic polysomnogram (PSG), each with one or more channels displayed electronically when interpreting a PSG. These data form the basis of the utility of a PSG to diagnose a range of sleep disorders. A PSG includes:[1]

• Bilateral frontal, central and occipital EEG
• Surface chin and leg electromyogram (EMG)
• Left and right eye electrooculogram (EOGs)
• Electrocardiogram (ECG) lead II
• Audio and video recording for snoring, body position, and other abnormalities
• Nasal pressure transducer
• Oronasal thermal flow sensor
• Thoracic and abdominal respiratory effort monitored most commonly with respiratory inductive plethysmography (RIP) belts
• Pulse oximeter

The EEG electrodes are set up according to the international 10-20 system.[4] The recommended electrode placements for an EEG during a PSG are F4/M1, C4/M1, and O2/M1. Backup electrodes should also be placed to allow for the display of F3/M2, C3/M2, and O1/M2 if there is a malfunction with the electrodes used for the recommended derivations. EMG measures the submentalis muscle activity of the chin, essential for the identification of rapid eye movement (REM) sleep. Another set EMG electrodes record anterior tibialis activity to evaluate for periodic limb movements of sleep (PLMS).[5] Left and right eye EOGs measure the change in the electrical field potential between the cornea (positive) and retina. The recommended derivations are E1-M2 and E2-M2. EOGs are used to identify the waking eye movements, slow rolling eye movements of drowsiness into stage N1, and the fast eye movements of stage R (called stage R rather than REM when scoring PSGs). 

Snoring and body position are monitored with the use of audio and video recording. Nasal airflow is measured by changes in pressure from a nasal pressure transducer. Oronasal thermal flow is measured based on the differences in temperature from the nose and mouth during inhalation and exhalation. RIP belts measure respiratory effort.[6][7][8] A continuous pulse oximeter is likewise of prime importance in the evaluation of SRBDs.

Procedures

Polysomnogram (PSG) should be performed during the patient's habitual sleep period. Patients present to the sleep center in the evening and are given questionnaires to quantify their subjective report of perceived sleepiness such as the Epworth Sleepiness Scale (ESS) or the Stanford Sleepiness Scale (SSS). A careful medication review is also performed. Upon completion of these forms, the patient should be allowed to use the restroom before attaching the various methods of data gathering as above. The technician then instructs the patient to await further instructions for calibrations and leaves the room.

The calibrations include equipment and physiologic calibration. The initial calibration of the equipment is done by sending a known signal through all recording channels. The signal is then sent through individual channels and frequency filters so the signals can be adjusted based on the center's standard operating procedures. Physiologic calibration occurs after the monitoring mechanisms are connected to the patient, which includes a series of eye, jaw, and foot movements and breathing changes. The technologist observes the output for correlation with the patient's actions.

After the calibrations are complete, the patient should be left in a cool, dark, and quiet room to promote sleep. At the same time, the technician must remain vigilant while monitoring the patient's PSG recordings and any physiologic abnormalities, as well as the video feed. Emergency procedures to call for medical help should be implemented and practiced regularly. The procedure is ended the following morning, and patients must be rested enough to drive if driving themselves. The patient must sleep for at least 2 hours during the PSG to be considered a valid study.

Indications

Indications for a polysomnogram (PSG) in adults include the following:[1]

• Diagnosis of sleep-related breathing disorders
• Patients with neuromuscular disorders and sleep complaints
• Continuous positive airway pressure (CPAP) or bilevel positive airway pressure (BiPAP) titration
• Evaluation of response to positive airway pressure (PAP), oral appliance, or positional therapy
• Evaluation before upper airway surgery for snoring or OSA
• Evaluating response after surgery for OSA
• Follow-up PSG if the response to treatment for OSA is not adequate or symptoms are not adequately controlled despite compliance with appropriate therapy
• 10% change in weight in patients treated with CPAP to verify the continued need for CPAP in those who have lost weight and optimize therapy settings in those who have gained weight
• Patients with congestive heart failure and symptoms of SRBD or persistent symptoms after optimal medical management of congestive heart failure
• Diagnosis of narcolepsy (in this case, PSG is followed by multiple sleep latency test the next day)
• Diagnosis of nocturnal epilepsy
• Diagnosis of REM sleep behavior disorder (RBD)
• Diagnosis of periodic limb movement disorder (PLMD) in which the patient has significant daytime fatigue or bed partner complaints
• Insomnia only if there is a suspected comorbid sleep disorder as above or after a failure of adequate treatment.[9]

PSG is not indicated for the following:[1]

• Patients adequately treated with CPAP
• Parasomnias (excluding when nocturnal seizures and RBD are considerations in the differential)
• Seizure disorders that are prominent during the day
• Restless legs syndrome
• Insomnia when there is a low concern of secondary cause for insomnia such as OSA or periodic limb movements of sleep (PLMS) significant enough to cause PLMD [9]
• Circadian rhythm disorders
• Other common events of sleep such as bruxism, nightmares, and sleep talking

Potential Diagnosis

The utility of the polysomnogram (PSG) stems from the array of physiologic data gathered during sleep. As such an extensive evaluation, there are also many potential diagnoses, the nuanced scoring, and diagnostic criteria beyond the scope of this article. Several examples of possible diagnoses are evident from the indications above.

OSA, however, deserves special attention because of its prevalence.[10][11][12][13] The pre-test probability of identifying OSA increases with age, male sex, and body mass index (BMI).[14][15][16] The diagnosis of OSA is based on measuring reductions in airflow during sleep. The diagnostic PSG would ideally include supine sleep during REM. Obstructive apneas (OAs) are scored when there is at least a 90% reduction in airflow for ten or more seconds (typically scored based on the oronasal thermal flow channel) with continued respiratory muscle activity. Respiratory muscle activity is evaluated with RIP belts. 

Obstructive hypopneas (OHs) are a partial reduction in airflow, also with continued respiratory activity. Scoring obstructive hypopneas requires a drop in the signal by at least 30% for ten or more seconds (typically scored based on the nasal pressure transducer channel) associated with a decrease in peripheral oxygen saturation of at least 3% from baseline or an EEG arousal.[17] Although the 3% drop in oxygen saturation is the American Academy of Sleep Medicine's (AASM) recommended criteria, medicare requires a 4% drop to be considered an OH, which is regarded as an acceptable alternative by the AASM for non-medicare patients as well. These alternate criteria do not take arousals into account when scoring OHs. 

EEG, EOG, EMG, and body position sensor or visual confirmation of body position are essential to identify supine stage R sleep. For most patients, OAs and OHs are most likely to occur during supine REM. Once identified, OAs and OHs are combined and averaged throughout sleep to give the apnea-hypopnea index (AHI). 

An AHI of 5 events/hour or less is considered normal for adults. Mild OSA is 5 to 14 events/hour, moderate 15 to 30 events/hour, and over 30 events/hour qualifies as a severe disease.[18] The first-line treatment for OSA is PAP during sleep.[19] PAP most commonly includes CPAP and auto-titrating continuous positive airway pressure (APAP).

Normal and Critical Findings

The essential normal findings for a polysomnogram (PSG) are the sleep stages, which are crucial to interpretation. Sleep stage scoring is based on EEG, EOG, and submental EMG criteria. Sleep stage scoring was initially being detailed by Rechtschaffen and Kales (R&K) but has since been replaced by the AASM scoring manual.[20] The AASM manual is mostly consistent with the original sleep scoring manual by Rechtschaffen and Kales and is regularly updated.

The AASM scoring manual uses new nomenclature for the wake and sleep referred to as stage W for wake, stage N1, N2, and N3 for the non-rapid eye movement (NREM) stages of sleep, and stage R for REM sleep. Stage N3 replaces stages 3 and 4 in the new classification. Sleep is staged in 30-second sections, known as epochs. Traditionally, PSGs were recorded on paper, and an epoch of sleep was a standardized length of paper given standard recording time. Today, digital PSG has virtually replaced recording on paper, and sleep is scored in sequential epochs that can be manipulated on the screen. If more than one sleep stage occurs in an epoch, the epoch is generally scored based on the sleep stage occupying most of the epoch.[21]

To identify the sleep stage in an epoch, the EEG plays a central role. The EEG demonstrates several recognizable wave patterns from wake through the stages of sleep. These wave patterns are described in terms of frequency, amplitude, and morphology. Frequencies include:

• Alpha 8 to 13 Hz
• Beta > 13 Hz
• Theta 4 to 7.99 Hz
• Delta < 4 Hz

Amplitude is a measure of the voltage of the EEG waves and varies across different sleep stages. The overall morphology of the waves can also provide insight into the sleep stage. For example, K-complexes and sleep spindles occur in N2 sleep. K-complexes are waves with a sharp initial negative deflection, followed by a positive deflection typically with maximal amplitude in the frontal derivations. Sleep spindles are sinusoidal 11-16 Hz waves lasting at least 0.5 seconds with their maximal amplitude generally in the central derivations. Sawtooth waves, which are 2-6 Hz waves with a serrated appearance with maximal amplitude in the central leads, may be seen during or before stage R.

Stage W scoring on a PSG depends on whether the patient has their eyes open or closed. With eyes open, the EEG demonstrates low amplitude mixed-frequency (LAMF) waves predominantly 4-7 Hz. With eyes closed, the EEG demonstrates an alpha (posterior dominant) rhythm in most patients. Alpha rhythm is generally most prominent in the occipital leads. The EOG shows blinking, reading eye movements or irregular, conjugate, sharply peaked rapid eye movements. EMG shows a high chin tone. Stage W is scored when greater than 50% of an epoch is alpha with patterns consistent with the above.[21]

Stage N1 is the transition from wakefulness into sleep. LAMF theta waves predominate on EEG. The EOG shows slow, rolling eye movements. Chin EMG tone is lower than stage W but higher than stage R. An epoch with over 50% of these characteristics is scored as stage N1 in the absence of evidence of another sleep stage.[21]

K-complexes characterize stage N2 and sleep spindles with a background of low amplitude mixed-frequency EEG activity. There are usually no eye movements on EOG. EMG is generally lower than during wake. N2 is scored when a K-complex without associated arousal or sleep spindle occurs in the first half of the current epoch or the last half of the previous epoch with less than 20% slow-wave activity. N2 continues until a transition into stage N3 or stage R, arousal, a major body movement followed by slow eye movements, or an awakening.[21]

Stage N3 is the deepest sleep stage, has high-amplitude, and slow waves. The delta frequency predominates between 0.5 and 2 Hz with amplitudes greater than 75 microvolts. There are generally no eye movements on EOG, and EMG is variable, but often the tone is lower than N2. This stage is scored when greater than 20% of an epoch (6 seconds) meets these frequency and amplitude criteria.[21]

Stage R represents REM sleep. As the name implies, the EOG demonstrates rapid eye movements, similar to wakefulness. The EEG is LAMF without K-complexes or sleep spindles. The chin EMG falls below previous amplitude levels in the recording. Sawtooth waves, although often associated with stage R, are not always present. Stage R is scored when the combination of the three EOG, EMG, and EEG criteria are satisfied.[21]

Interfering Factors

Despite being the gold standard for diagnosing multiple sleep disorders, polysomnograms (PSGs) have several limitations. For example, the first-night effect, which is due to poor sleep because of a new sleeping environment, may lead to underestimating OSA due to the potential for decreased REM sleep being captured.[22] Medication changes before a PSG may also interfere with the quality and quantity of sleep during a PSG. Additionally, some sleep disorders such as nocturnal seizures and RBD may occur too infrequently to be captured during a single night study reliably. Equipment issues can contribute to PSG inaccuracies, given the many physiologic recordings and potential for electrodes and instrument cable malfunction. 

In addition to diagnostic limitations, there are also systemic interfering factors with PSG. Performing a PSG requires an adequate sleep period, highly trained technicians to administer the study, and sleep providers interpreting the study, resulting in high cost. As a result, insurance companies usually require prior authorizations, placing many barriers between the patient and their ability to get a PSG even when indicated. The PSG data itself is also not portable in many cases, requiring unique software that is not readily exportable.

Complications

The most common complication is skin irritation from electrode attachment sites. Other complications are rare in polysomnogram (PSG). However, because of the length of the study and potential for evaluating individuals with multiple other medical comorbidities, technicians must always remain watchful for potential physiologic derangements requiring immediate medical assistance during a PSG.

Patient Safety and Education

When a polysomnogram (PSG) is determined to be necessary, the patient should be educated on the procedures and potential outcomes of the study. Electrodes, other monitoring devices, and their uses should be clearly explained to the patient and the reasons for the monitoring in their specific case. The patient's regular sleep schedule should be identified to allow for optimum timing of the sleep period in conjunction with laboratory procedures and availability. Based on the indication of the study, patients may need to be aware of the potential for the initiation of PAP initiation, if indicated. PAP may be uncomfortable, and setting expectations can often lead to a more successful titration.

In general, patients should take their medications as prescribed the day and night of the PSG, including sleep aids. However, medication reconciliation must also be performed and recorded before a sleep study to avoid unwanted or unexpected impacts on sleep. This is particularly true if benzodiazepines or opioids are among the medications, since these medications may exacerbate sleep-disordered breathing.[23][24]

Alcohol use before a sleep study can have a similar effect.[24] Patients being evaluated for narcolepsy should also be carefully screened for medications before PSG and MSLT. Antidepressants, in particular, can suppress REM sleep and should be noted to facilitate appropriate PSG interpretation.

Clinical Significance

Sleep complaints significantly impact patients both physically and psychologically and, when left undiagnosed or untreated, correlate with suffering in many patients.[25] Polysomnograms (PSGs) plays a crucial role in thoroughly evaluating many sleep complaints, and identifying treatable sleep disorders can reduce morbidity and mortality.[26][27][28][29]

The PSG itself is technically complex and time-consuming because of its many monitoring systems, which are also the source of its high clinical utility in diagnosing and managing a wide range of sleep disorders. With proper training, execution, interpretation, and patient preparation, a PSG is a powerful tool in diagnosing and subsequent management of sleep disorders.