More than 10% of the US population have obstructive sleep apnea (OSA). Standard of care is therapy with CPAP (continuous positive airway pressure) which virtually eliminates OSA. However, most patients use CPAP only for part of the night (4-5hours) and about 50% patients discontinue CPAP long-term. Alternative therapies are limited, thus many OSA patients remain at risk of OSA sequelae (e.g. sleepiness, memory issues, high blood pressure, etc.). Importantly, different patients get OSA for different reasons, and recent data show that some of the underlying causes of OSA ("endotypes") such as having a low arousal threshold (i.e. waking up easily) are associated with lower CPAP adherence. Using a randomized controlled trial design, this will be the first study using a targeted intervention to manipulate the underlying OSA causes (i.e., giving a safe hypnotic to patients with OSA to increase the arousal threshold) to test the hypothesis that endotype-targeted therapy increases CPAP-adherence in patients who have low but continued CPAP usage. Ultimately, this strategy may improve the care and outcomes of millions of undertreated OSA patients.

This double-blind, randomized, placebo-controlled, single-center, phase II trial, will test if endotype-targeted drug therapy increases CPAP adherence. Patients with low CPAP adherence (i.e. CPAP use 0.5-4h/night) will be randomized in parallel to CPAP+Eszopiclone vs CPAP+Placebo for 2 weeks. The primary outcome is mean TimeOnCPAP/night on days 2-14 after initiation of study drugs.

Aim 1: To test the effect of eszopiclone on CPAP usage in unselected OSA patients with low CPAP adherence.

Hypothesis 1: Eszopiclone improves CPAP usage more than placebo.

Aim 2: To test the effect of eszopiclone on CPAP usage in OSA patients with a low vs a high arousal threshold (ArTH) endotype.

Hypothesis 2: Eszopiclone improves CPAP usage substantially more in OSA patients with a low vs a high ArTH.

Further, this study will explore the mechanisms through which eszopiclone may affect CPAP usage, the effect of eszopiclone on underlying sleep apnea severity, and if the intervention and its potential effects on adherence translate into changes in clinically important outcomes (i.e., vigilance and sleepiness).

Inclusion Criteria:

• Ages 21-65 years old
• Body Mass Index <32 kg/m^2
• Physician diagnosis of OSA (AHI ≥ 10) based on a clinical sleep study within the past 2 years (subjects who report a history of sleep apnea but do not have a sleep study report from the past 2 years available will be offered an overnight home sleep apnea test to verify OSA diagnosis)
• AHI>5/h on the overnight research sleep study #1
• Subject had the opportunity to use CPAP for at least 1 month
• Continuous positive airway pressure (CPAP) data can be queried remotely and shows usage 0.5-4h/night (based on most recent 30 day period)
• Interest to continue trying CPAP

Exclusion Criteria:

• "SAVE CPAP Side Effect Score" >3 (1 point for each: dry mouth, nasal symptoms, CPAP pressure intolerance, claustrophobia, noise problems, soreness/skin irritation, mask fit/leak problems)
• Any high-risk features: Epworth sleepiness score ≥18, safety-critical profession (e.g., commercial driver), prior sleep-related car accident, substantial hypoxemia during sleep [SpO2<70% for >5min] or awake [SpO2<92%]
• Any uncontrolled medical/psychiatric condition, or safety concern based on MD judgment
• Pregnancy/Breastfeeding (current or planned during the next month)
• Inability to complete study procedures, such as questionnaires that are only available/validated in English
• Other known untreated sleep fragmenting disorder, such as periodic limb movement disorder, or narcolepsy
• Unwilling or unable to withhold CPAP during polysomnography
• Presence of tracheostomy
• Hospitalization within the past 90 days
• Allergy to the study drug
• Regular use of opioids, or benzodiazepines
• Chronically using study drug or other hypnotic
• Significant circadian rhythm disorder or sleepwalking as an adult
• Active illicit substance use or >3 oz nightly alcohol use
• Prisoners
• Cognitive impairment, unable to provide consent, or unable to carry out research procedures
• Safety concern based on MD judgment