Lunesta vs Dayvigo: Side-Effect Profile Head-to-Head

Two Different Mechanisms, Two Different Side-Effect Fingerprints

Eszopiclone enhances GABA-A receptor activity, producing sedation by broadly dampening neuronal excitability. Lemborexant blocks orexin-1 and orexin-2 receptors, silencing the wake-promoting signal rather than amplifying the sleep signal. This mechanistic split explains why their adverse-event profiles barely overlap.

GABA-A agonists as a class share predictable risks: next-day psychomotor impairment, complex sleep behaviors, amnesia, and tolerance development 1. DORAs, by contrast, were designed to avoid many of these liabilities. The FDA's 2014 advisory committee review of suvorexant (the first approved DORA) noted that the orexin mechanism "does not produce the amnesia, ataxia, or respiratory depression associated with benzodiazepine receptor agonists" 2. Lemborexant, the second DORA to reach the market, extended this profile with additional data on next-morning function from the SUNRISE clinical program 3.

No randomized trial has ever pitted eszopiclone directly against lemborexant. Every comparison below is a cross-trial synthesis. Effect sizes come from each drug's own placebo-controlled data, and differences in trial populations, endpoints, and run-in periods make precise numerical comparisons imperfect.

Dysgeusia: Lunesta's Most Distinctive Burden

The metallic or bitter taste after taking eszopiclone is the single side effect that most often drives patients to discontinue the drug. In Krystal et al.'s 6-month efficacy trial (N=788), unpleasant taste occurred in 33.9% of patients on eszopiclone 3 mg versus 3.1% on placebo 1. Even at the lower 2 mg dose, dysgeusia affected roughly 17% of patients in pooled Phase III data submitted to the FDA 4.

The taste is not trivial. Patients describe it as persisting into the next morning, altering the flavor of breakfast foods and coffee. Discontinuation rates due to taste alone reached 5.3% in the 3 mg arm of the Krystal trial, compared to 2.0% for placebo 1.

Lemborexant does not cause dysgeusia. Zero SUNRISE trial participants reported taste disturbance as an adverse event 3. For patients who have already tried and rejected Lunesta because of the taste, this absence alone may justify a switch.

Next-Morning Residual Effects

Both drugs can cause daytime somnolence, but the pattern differs. In 2014, the FDA lowered Lunesta's recommended starting dose from 2 mg to 1 mg after post-marketing driving simulation studies showed that 3 mg eszopiclone impaired driving performance up to 11 hours after dosing 5. Blood levels of eszopiclone at 7.5 hours post-dose remained above the threshold for psychomotor impairment in a significant proportion of women.

Dayvigo's key SUNRISE-1 trial (N=1,006) specifically measured next-morning residual effects using the Digit Symbol Substitution Test (DSST) and a postural stability assessment. At the 5 mg dose, lemborexant showed no statistically significant difference from placebo on either measure 3. The 10 mg dose produced a small, transient DSST decrement on the first night only, which resolved by the second night of dosing.

Somnolence as a reported adverse event occurred in 10% of lemborexant 10 mg patients versus 1% on placebo in SUNRISE-1, and in 5% of lemborexant 5 mg patients 3. For eszopiclone, somnolence rates were 8.5% at 3 mg and 4.5% at 2 mg in pooled registration data 4. The raw percentages look comparable, but the Lunesta data did not include objective driving or psychomotor testing in the key trial era, and the subsequent FDA safety communication confirmed real-world impairment at doses that were initially considered safe 5.

Dr. Andrew Krystal, who led multiple eszopiclone trials and later served as a consultant on DORA development, noted in a 2020 review: "The orexin antagonists represent a pharmacologically distinct approach that does not appear to carry the same risk of next-morning psychomotor impairment seen with GABA-A modulators, particularly at recommended doses" 6.

Dependency, Tolerance, and Rebound Insomnia

Tolerance development is a concern with any controlled substance prescribed nightly. Eszopiclone's 6-month trial was, at the time of its publication, the longest placebo-controlled insomnia study conducted. It showed sustained efficacy without significant tolerance over 6 months on the primary sleep endpoints 1. That finding was genuinely notable for a GABA-A agonist.

The problem emerged at discontinuation. In the Krystal trial, patients who stopped eszopiclone after 6 months experienced a statistically significant rebound in sleep latency during the first two nights off the drug, with values worse than their original baseline 1. Rebound resolved within a week for most patients, but this pattern raised the question of physiological dependence.

Lemborexant's 12-month SUNRISE-2 extension study (N=949) assessed discontinuation effects by switching active-treatment patients to placebo for a 2-week run-out period. No significant rebound insomnia was detected on any polysomnographic measure during that period 7. The American Academy of Sleep Medicine's 2023 clinical practice guideline noted that DORAs "appear to have a lower propensity for tolerance, withdrawal, and rebound insomnia than benzodiazepine receptor agonists," though it cautioned that longer-term data beyond 12 months remain limited 8.

This distinction matters most for patients who need chronic treatment. Short-term use (2 to 4 weeks) may not expose the dependency differential. Long-term nightly use tips the risk-benefit analysis toward lemborexant for patients concerned about discontinuation difficulty.

Complex Sleep Behaviors and Parasomnias

All FDA-approved insomnia drugs now carry a boxed warning for complex sleep behaviors (sleepwalking, sleep-driving, engaging in activities while not fully awake). The FDA issued this class-wide warning in 2019 after reviewing 66 serious injury and 20 death reports spanning all sedative-hypnotics 9.

In clinical trials, the incidence of complex sleep behaviors was low for both drugs. Eszopiclone's prescribing label reports somnambulism in <1% of patients across all trials 4. Lemborexant's label reports "sleep paralysis" and "hypnagogic/hypnopompic hallucinations" as adverse reactions, each in approximately 1% to 2% of patients at the 10 mg dose 10. These phenomena reflect the orexin system's role in regulating REM sleep boundaries. They are uncomfortable but generally not dangerous. Sleep paralysis and hallucinations are also cardinal symptoms of narcolepsy, a condition defined by orexin deficiency, which explains their mechanistic link to DORA therapy.

For patients with a history of sleepwalking or parasomnias, neither drug class is clearly safer. Clinicians should screen for pre-existing parasomnias before prescribing either medication.

Safety in Older Adults

Insomnia prevalence increases with age, and so does vulnerability to sedative-hypnotic side effects. Falls are the critical outcome.

SUNRISE-2 enrolled 192 adults aged 65 and older. Fall-related adverse events occurred in 2.1% of lemborexant 5 mg patients, 3.2% of lemborexant 10 mg patients, and 2.5% of placebo patients in this subgroup, differences that were not statistically significant 7. The American Geriatrics Society Beers Criteria list benzodiazepine receptor agonists (including eszopiclone) as "potentially inappropriate" in adults 65 and older due to fall risk and cognitive impairment, while DORAs are not included on the avoid list 11.

As the Beers Criteria panel wrote in their 2023 update: "Nonbenzodiazepine benzodiazepine receptor agonists (Z-drugs) have adverse events similar to those of benzodiazepines in older adults (e.g., delirium, falls, fractures); increased emergency room visits/hospitalizations; motor vehicle crashes" 11.

This does not mean lemborexant is risk-free in elderly patients. The 10 mg dose still produced somnolence rates above placebo. The 5 mg dose is the recommended starting and maintenance dose for patients 65 and older.

Respiratory Safety

Patients with comorbid obstructive sleep apnea (OSA) represent a common clinical challenge. GABA-A agonists can reduce upper-airway muscle tone and worsen apnea-hypopnea index (AHI) scores. Eszopiclone's prescribing information does not include OSA-specific safety data from controlled trials.

Lemborexant was specifically studied in patients with mild-to-moderate OSA. A Phase I crossover study (N=53) found that lemborexant 10 mg did not worsen AHI compared to placebo, and oxygen saturation nadir did not differ significantly between groups 12. This finding aligns with the DORA mechanism: orexin blockade does not directly affect respiratory drive or airway muscle tone.

For insomnia patients with untreated or partially treated sleep apnea, lemborexant offers a theoretical safety advantage, though this should not replace adequate OSA treatment with CPAP or oral appliance therapy 8.

Drug Interactions and Metabolism

Both drugs are metabolized by CYP3A4. Strong CYP3A4 inhibitors (ketoconazole, itraconazole, clarithromycin, ritonavir) increase exposure to both drugs. Eszopiclone's Cmax increased 1.4-fold with ketoconazole co-administration 4. Lemborexant's AUC increased approximately 4-fold with itraconazole, prompting a contraindication for concurrent use with strong CYP3A4 inhibitors 10.

This interaction profile is more restrictive for lemborexant. Patients on HIV protease inhibitors, certain antifungals, or strong CYP3A4-inhibiting antibiotics cannot use lemborexant at any dose, while eszopiclone may still be usable at a reduced dose with monitoring.

Moderate CYP3A4 inhibitors (fluconazole, erythromycin, diltiazem, grapefruit juice) require a dose reduction to 5 mg for lemborexant. Eszopiclone's label recommends not exceeding 2 mg with moderate inhibitors.

Comparative Adverse Event Summary

The following profiles emerge from cross-trial synthesis of registration data:

Eszopiclone (Lunesta) at the 3 mg dose: Dysgeusia 33.9%, somnolence 8.5%, dizziness 5.1%, dry mouth 5.0%, headache 5.2%, infection 5.8%. Discontinuation due to adverse events: 12.4% vs. 7.5% placebo 1.

Lemborexant (Dayvigo) at the 10 mg dose: Somnolence 10%, headache 6%, sleep paralysis 2%, abnormal dreams 2%, hypnagogic hallucinations 1%. Discontinuation due to adverse events: 3.0% vs. 1.4% placebo 3.

The discontinuation rate gap (12.4% vs. 3.0%) is notable, though cross-trial comparisons should be interpreted cautiously. Trial durations differed (6 months vs. 1 month for the acute phase of SUNRISE-1), and dysgeusia alone accounts for a large share of eszopiclone discontinuations.

When Each Drug May Be Preferred

Lemborexant may be the better starting choice for older adults (Beers Criteria favors DORAs over Z-drugs), patients with comorbid mild-to-moderate OSA, patients who previously discontinued a Z-drug due to taste or next-morning impairment, and patients requiring long-term nightly therapy where rebound insomnia at discontinuation is a concern.

Eszopiclone may still be appropriate when a patient is already stable on the drug without problematic side effects, when a patient takes a strong CYP3A4 inhibitor that contraindicates lemborexant, when cost is a deciding factor (generic eszopiclone averages $15 to $30 per month; brand-name Dayvigo averages $400+ without insurance), and when sleep-onset latency is the dominant complaint (eszopiclone has strong data for reducing time to fall asleep within 10 to 15 minutes of dosing).

Generic lemborexant does not yet exist. Eisai's patent protection extends through 2032 based on current filings with the FDA Orange Book 10.

The Cost-Access Reality

Even when clinical data favors one drug, insurance formulary placement dictates real-world prescribing. Most commercial plans and Medicare Part D formularies place generic eszopiclone on Tier 1 or Tier 2 (preferred generic), while Dayvigo sits on Tier 3 or requires prior authorization 8. Patients prescribed Dayvigo should be directed to Eisai's patient assistance program, which covers up to 12 months of copay support for commercially insured patients with out-of-pocket costs above $30 per fill.

The cost differential narrows the clinical advantage. A drug with fewer side effects is only useful if the patient can actually fill the prescription. Clinicians should verify formulary status before writing for lemborexant and have a step-therapy plan documented if the insurer requires a Z-drug trial first.

Lemborexant 5 mg taken 30 minutes before bed, with at least 7 hours of intended sleep time remaining, is the starting regimen for most adults. For patients who cannot access or tolerate it, eszopiclone 1 mg (the current FDA-recommended starting dose) remains a well-characterized alternative with 20 years of post-marketing safety data 5.