Drugs That Distort Polysomnography (Sleep Study) Results

What Polysomnography Measures and Why Drug Effects Matter

Polysomnography records at least 12 simultaneous physiological signals during sleep: electroencephalography (EEG) for sleep staging, electrooculography (EOG) for eye movements, submental and tibial electromyography (EMG), nasal pressure and thermistor airflow, thoracic and abdominal respiratory effort belts, pulse oximetry, electrocardiography (ECG), body position, and snoring microphone. The test produces a scored hypnogram that maps sleep stages across the night and calculates the Apnea-Hypopnea Index (AHI), the primary metric for OSA diagnosis [1].

How Sleep Architecture Shapes the Diagnosis

Any substance that shifts the proportion of N1, N2, N3, or REM sleep changes the denominator against which respiratory events are counted. A drug that eliminates REM sleep, for instance, removes the stage where upper-airway collapsibility is greatest. The result: an artificially low AHI that underestimates true disease severity [2].

Why Clinicians Need an Unmedicated Baseline

The American Academy of Sleep Medicine (AASM) recommends that, whenever clinically safe, medications known to affect sleep architecture or respiratory drive be held before diagnostic PSG [3]. The goal is to capture the patient's intrinsic sleep-breathing pattern. When a drug cannot be safely discontinued (chronic opioid therapy, anti-seizure regimens), the interpreting physician must note this in the report and adjust clinical decision-making accordingly.

Benzodiazepines and Non-Benzodiazepine Receptor Agonists

Benzodiazepines (diazepam, lorazepam, clonazepam) and "Z-drugs" (zolpidem, zaleplon, eszopiclone) bind GABA-A receptors, increase N2 spindle activity, suppress N3 slow-wave sleep, and reduce REM proportion. These changes compress the sleep stages where apneas concentrate, producing a falsely reassuring AHI [4].

Impact on Respiratory Parameters

A 2003 study in the journal Sleep found that 10 mg zolpidem lowered the arousal index by 18% without changing actual airflow obstruction, meaning the patient obstructed just as often but the brain failed to mount its normal protective arousal response [5]. In clinical terms, the oxygen desaturation index (ODI) may rise even as the scored AHI appears unchanged or lower, because hypopneas that no longer trigger arousal get reclassified.

Impact on Muscle Tone Recordings

Clonazepam is prescribed for REM behavior disorder (RBD). If a patient takes clonazepam the night of a diagnostic PSG intended to confirm RBD, the drug will suppress the very phasic EMG bursts the study is designed to detect. The AASM recommends tapering clonazepam before RBD-diagnostic PSG when safe to do so [3].

Hold Guidance

Zolpidem (half-life ~2.5 hours) clears in roughly 12 hours. Diazepam (half-life 20-100 hours) may require 5 or more days. Discuss the taper plan with the prescribing clinician.

Opioids and Respiratory Drive

Opioids are among the most clinically significant PSG distorters. Mu-receptor agonists (morphine, oxycodone, hydrocodone, methadone, fentanyl) depress medullary chemoreceptor sensitivity, producing central apneas and ataxic (Biot) breathing patterns that do not reflect the patient's native airway anatomy [6].

Central vs. Obstructive Patterns

A landmark 2007 study by Walker et al. In Chest (N=60 chronic opioid users) reported that 75% had sleep-disordered breathing: 24% had primarily central sleep apnea (CSA), 39% had ataxic breathing, and 12% had obstructive events [7]. Morphine-equivalent daily doses above 200 mg were associated with AHI values exceeding 30 events per hour. Methadone, with its long and variable half-life (8-59 hours), is especially problematic. Holding it is rarely safe, so the sleep physician must interpret the study knowing that observed central events may resolve if opioid therapy changes.

Why This Matters for Pain Patients

A PSG performed on a patient taking 90 morphine milligram equivalents (MME) daily may show a mixed apnea pattern. Prescribing CPAP alone for what appears to be OSA will not treat the central component. Adaptive servo-ventilation (ASV) may be needed, but ASV carries its own contraindications in patients with heart failure. The opioid context changes the entire treatment pathway.

SSRIs, SNRIs, and REM-Sleep Suppression

Selective serotonin reuptake inhibitors (fluoxetine, sertraline, paroxetine, escitalopram) and serotonin-norepinephrine reuptake inhibitors (venlafaxine, duloxetine) increase serotonergic tone, which suppresses REM sleep onset and reduces total REM percentage by 20-50% depending on dose and agent [8].

How REM Suppression Masks Apnea Severity

OSA is typically worst during REM sleep, when postural muscle atonia allows maximal pharyngeal collapse. A patient on sertraline 200 mg may spend only 8% of total sleep time (TST) in REM instead of the normal 20-25%. If the physician does not request a REM-specific AHI calculation, the overall AHI will undercount the patient's true disease burden. The Endocrine Society's 2018 guideline on testosterone therapy notes that PSG should capture adequate REM sleep for reliable OSA assessment, especially in men starting TRT [9].

SSRI-Induced Bruxism and PLMs

SSRIs also increase periodic limb movements (PLMs) and bruxism during sleep. A 2018 systematic review in the Journal of Clinical Sleep Medicine found a 2.7-fold increase in PLM index among SSRI users compared to controls [10]. If the clinical question is "does this patient have restless legs syndrome or PLM disorder," the SSRI must be factored into interpretation or ideally held.

Withdrawal Caution

Abruptly stopping an SSRI causes REM rebound (vivid dreams, fragmented sleep, increased REM density) that equally distorts PSG. A supervised taper over 2-4 weeks before the study is preferred when clinically appropriate.

Stimulants: Amphetamines, Methylphenidate, Modafinil

CNS stimulants prescribed for ADHD or narcolepsy suppress total sleep time, delay sleep onset, reduce sleep efficiency, and decrease REM and N3 percentages [11].

Effect on Multiple Sleep Latency Test (MSLT)

For narcolepsy evaluation, PSG is followed the next day by an MSLT, which measures how quickly a patient falls asleep and whether sleep-onset REM periods (SOREMPs) occur. Stimulants taken within 2 weeks of the MSLT can prevent SOREMPs entirely, producing a false-negative result. The AASM clinical practice guideline for narcolepsy diagnosis requires a 2-week stimulant-free washout before MSLT [12].

Modafinil and Armodafinil

Modafinil's half-life is approximately 15 hours. It primarily reduces N3 sleep and may not suppress REM as aggressively as amphetamines, but it still delays sleep onset and lowers sleep efficiency. A single dose on the morning of the PSG night may not fully clear by lights-out.

Antihistamines and Over-the-Counter Sleep Aids

First-generation antihistamines (diphenhydramine, doxylamine, hydroxyzine) cross the blood-brain barrier, block H1 receptors, and produce sedation. They increase total sleep time and reduce sleep latency, but they also suppress REM sleep and increase N2 proportion [13].

The "Good Sleep" Illusion

Patients frequently take diphenhydramine 25-50 mg the night of a sleep study because they worry about not falling asleep in the lab. This well-intentioned choice lowers arousal threshold and may reduce the scored AHI while worsening true hypoxemia burden, as the brain mounts fewer protective arousals. Sleep labs should explicitly instruct patients to avoid OTC sleep aids for at least 48 hours before PSG.

Second-Generation Antihistamines

Cetirizine and loratadine have minimal CNS penetration and negligible effects on sleep architecture at standard doses. These generally do not need to be held.

Cannabis (THC) and Cannabinoids

Delta-9-tetrahydrocannabinol (THC) reduces REM sleep, shortens sleep latency, and may increase N3 slow-wave sleep at low doses while fragmenting sleep at high doses [14]. Synthetic cannabinoids (dronabinol, nabilone) produce similar EEG changes.

Emerging Data on THC and Apnea

A small randomized trial by Carley et al. Published in Sleep (2018, N=73) explored dronabinol for OSA and found a modest AHI reduction of 10.7 events per hour versus placebo at the 10 mg dose [15]. While that trial was investigating therapeutic potential, the finding also means that a patient who uses THC regularly will present to the sleep lab with an artificially suppressed AHI. Cessation of THC, conversely, triggers REM rebound. The optimal approach: document use, maintain the patient's usual pattern if cessation is impractical, and note the confound in the report.

Alpha-2 Agonists and Muscle Relaxants

Clonidine and tizanidine, both alpha-2 agonists, reduce sympathetic outflow, lower arousal threshold, and decrease REM sleep. Clonidine specifically can produce central apneas in predisposed individuals by depressing ventilatory drive [16]. Baclofen and cyclobenzaprine deepen sedation and reduce upper-airway tone, potentially worsening obstructive events while simultaneously reducing cortical arousals that would otherwise flag them.

Clinical Guidance

For patients on baclofen for spasticity, discontinuation is dangerous (risk of withdrawal seizures). The sleep physician should annotate the PSG report and consider that true AHI may be higher than scored.

Alcohol: The Overlooked Confounder

Alcohol is not a prescription drug, but it deserves mention because patients often drink on study nights to "help sleep." Ethanol suppresses REM sleep in the first half of the night, triggers REM rebound in the second half, relaxes pharyngeal dilator muscles, and increases upper-airway resistance. A 2018 meta-analysis in Sleep Medicine Reviews (N=625 across 21 studies) found that even moderate alcohol intake (2 standard drinks) increased AHI by an average of 32.5% in non-OSA subjects [17].

Pre-Study Instructions

Sleep labs should explicitly ask about alcohol use. A standard instruction is to avoid alcohol for at least 24 hours before the study.

Testosterone and GLP-1 Receptor Agonists: Context for HealthRX Patients

Exogenous Testosterone

The Endocrine Society's 2018 guideline recommends baseline PSG or home sleep apnea testing (HSAT) before initiating testosterone replacement therapy (TRT) in men with obesity or existing symptoms of OSA, and re-evaluation 3-6 months after starting TRT [9]. Testosterone does not distort PSG per se, but it can worsen OSA severity over time by promoting fat deposition in the parapharyngeal space and altering central ventilatory sensitivity. A 2014 randomized trial (N=67) published in the Journal of Clinical Endocrinology & Metabolism found that testosterone gel increased AHI by an average of 7 events per hour over 18 weeks versus placebo in older men with low testosterone [18].

GLP-1 Receptor Agonists and Weight Loss

Weight loss from semaglutide or tirzepatide can reduce AHI. The SURMOUNT-OSA trial (N=469) demonstrated that tirzepatide 10-15 mg produced a 55.0% reduction in AHI at 52 weeks compared to 23.2% with CPAP-placeholder in the non-PAP arm [19]. Patients on GLP-1 therapy who undergo repeat PSG may show improved AHI from weight reduction, not a drug-induced artifact on the test itself. The distinction matters: this is a real physiological improvement, not a confound.

A Practical Hold-or-Continue Framework

Not every medication can or should be stopped before PSG. The decision depends on the clinical question, the drug's half-life, and the risk of withdrawal.

Drugs Typically Held (With Prescriber Approval)

Benzodiazepines (non-epilepsy indications), Z-drugs, OTC antihistamines, stimulants (especially before MSLT), cannabis, and melatonin at supraphysiologic doses (>3 mg) should be held for a minimum of 5 half-lives. For stimulants before MSLT, the AASM recommends a full 14-day washout [12].

Drugs Usually Continued (With Annotation)

Chronic opioid therapy, anti-seizure medications, SSRIs/SNRIs that cannot be tapered safely, antipsychotics, and baclofen should be continued with clear documentation in the PSG report. The interpreting physician adjusts conclusions accordingly.

The 5-Half-Life Rule

Five elimination half-lives clears approximately 97% of a drug. For zolpidem (half-life 2.5 h), that is about 12.5 hours. For diazepam (half-life up to 100 h with active metabolites), that could exceed 3 weeks. For methadone (half-life 8-59 h), safe discontinuation is rarely an option at all.

"The single biggest source of uninterpretable polysomnograms in our lab is undisclosed medication use on the study night," notes the AASM's 2020 clinical practice parameters update [3].

"Physicians ordering sleep studies should provide patients with an explicit, written medication-hold list at least 2 weeks before the scheduled study," per the American College of Chest Physicians (ACCP) sleep medicine board review guidance [20].

When to Repeat the Study

A PSG performed under heavy pharmacologic confounding may need to be repeated. Specific triggers for repeat testing include: AHI results that are discordant with clinical symptoms (severe daytime sleepiness but AHI of 3), discovery after the study that a patient took an undisclosed sedative, split-night studies where insufficient REM was captured in the diagnostic half, and MSLT results that are negative in a clinically convincing narcolepsy presentation where stimulants were not adequately washed out. Insurance carriers (including Medicare) generally cover repeat PSG when the ordering physician documents the confound in the medical record [21].