Eszopiclone (Lunesta) During Pregnancy and Lactation: Safety Evidence and Clinical Guidance

Eszopiclone is the brand-name drug Lunesta, a prescription nonbenzodiazepine sedative-hypnotic (a "z-drug") chemically related to zopiclone, the racemic compound sold outside the United States. Eszopiclone is specifically the S-enantiomer of zopiclone. It is FDA-approved for insomnia in adults, not for use during pregnancy or lactation, and no pregnancy indication exists for any z-drug.

At a glance

• Drug class / Nonbenzodiazepine GABA-A receptor agonist ("z-drug"), S-isomer of zopiclone
• FDA pregnancy labeling / No letter category; narrative labeling under the Pregnancy and Lactation Labeling Rule (PLLR)
• Controlled human pregnancy trials / None published as of 2026
• Animal reproductive data / Reduced fetal weight and delayed ossification at doses well above the human therapeutic range, per the FDA label
• Human lactation data / None specific to eszopiclone; inference drawn from zopiclone, which is detectable in breast milk
• Neonatal concern near delivery / Sedation, poor feeding, and respiratory depression are class effects of GABA-A agonists used late in pregnancy
• First-line alternative / CBT-I, recommended over medication by ACOG and AASM for insomnia in pregnancy and postpartum
• Verification status of this page / Pending qualified clinical review; several precise figures below require confirmation against the primary literature before clinical use

Eszopiclone (Lunesta), a nonbenzodiazepine GABA-A receptor agonist and the S-isomer of zopiclone, has no controlled human pregnancy trials as of 2026. The FDA label advises use in pregnancy "only if the potential benefit justifies the potential risk to the fetus," a position built on animal data showing reduced fetal weight and delayed skeletal development at supratherapeutic doses, not on confirmed human harm. Because eszopiclone-specific lactation studies do not exist, clinicians extrapolate from zopiclone, which is known to transfer into breast milk, while guideline groups including ACOG and AASM recommend CBT-I as the preferred first-line insomnia treatment in both pregnancy and breastfeeding rather than any sedative-hypnotic.

How eszopiclone works, and why that matters for a developing fetus

Eszopiclone binds the alpha-1 subunit of the GABA-A receptor complex, enhancing chloride ion flow and potentiating GABA-mediated inhibition in the central nervous system. That mechanism produces the sedative and sleep-promoting effect the drug is prescribed for. It is the same receptor system, GABA-A, that appears in fetal brain tissue from early gestation and that participates in neuronal migration and synapse formation. This creates a biologically plausible reason for caution around any GABA-A modulator in pregnancy, but plausibility is not the same as demonstrated harm, and no human study has isolated eszopiclone's effect on these processes.

Eszopiclone's small molecular size and moderate lipophilicity make placental crossing likely by passive diffusion, consistent with what is generally true of small-molecule CNS drugs, though no published human placental-transfer study of eszopiclone itself is available to this review. The drug's adult half-life is roughly a few hours, but a newborn's liver enzyme systems, particularly CYP3A4, are immature at birth and clear drugs more slowly than an adult's, which is the pharmacologic reason any drug crossing to a fetus or nursing infant near term can linger longer in the infant than in the mother.

What the FDA label actually says, and what it does not say

Since 2015, the FDA no longer assigns letter pregnancy categories to new or updated labels; instead it uses narrative risk statements under the Pregnancy and Lactation Labeling Rule. The Lunesta label states that there are no adequate and well-controlled studies of eszopiclone in pregnant women and instructs that it be used only if potential benefit justifies potential fetal risk (FDA prescribing information: https://www.accessdata.fda.gov/drugsatfda_docs/label/2014/021476s030lbl.pdf).

That wording reflects an evidence gap, not a documented signal of harm. A drug with negative controlled human studies would carry different language. Eszopiclone sits in the more common and less comfortable middle ground where neither safety nor teratogenicity has been established in humans. The label also carries the general sedative-hypnotic class warning about neonatal flaccidity, respiratory depression, and withdrawal-type symptoms when these drugs are used close to delivery.

Animal data: the primary evidence base, and its limits

Because controlled human pregnancy trials do not exist and are unlikely to be conducted, animal reproductive toxicology is the main quantitative evidence available, and it comes from the FDA label itself. In rat studies conducted during organogenesis, doses far above the maximum recommended human dose produced decreased fetal body weight and delayed ossification, without causing structural malformations at any tested dose. Rabbit studies similarly showed no embryotoxicity or teratogenicity at high, multiple-of-human-dose exposures. A pre- and postnatal rat study found increased pup mortality and reduced growth only at the highest maternal doses tested, with a no-effect level well above typical human exposure (FDA label, above).

These findings are reassuring in direction but limited in what they can tell a human prescriber. Rodent placentation, litter size, and drug metabolism differ meaningfully from human physiology, and an absence of structural defects in animals does not rule out subtler human effects on birth weight, gestational timing, or neurodevelopment that would only be detectable in large human cohorts, which do not exist for this drug.

Human pregnancy evidence: sparse, and largely borrowed from a related drug

No randomized controlled trial has studied eszopiclone in pregnant women, and ethical constraints make that unlikely to change. The available human evidence comes almost entirely from pregnancy registries and epidemiologic database studies of the broader z-drug class, chiefly zopiclone and zolpidem, not eszopiclone specifically.

Published registry analyses of zopiclone exposure in pregnancy have generally not found a statistically significant increase in major congenital malformations. Separately, at least one large population-based cohort study of zolpidem exposure reported associations with preterm birth and lower birth weight, though whether that reflects a drug effect, the underlying insomnia, or associated maternal illness is not established. These are described here in general terms because the specific effect estimates in the original source material could not be verified against the primary literature for this rewrite; a reader or clinician who needs the exact odds ratios and confidence intervals should pull the original papers before relying on those numbers.

The core limitation is that none of this evidence is eszopiclone-specific. Z-drugs share a receptor mechanism but differ in potency, isomer composition, and pharmacokinetics, so cross-drug extrapolation is a reasonable approximation, not a substitute for direct data.

Lactation: what is known about the related compound, not the drug itself

No published human lactation study of eszopiclone exists. The relevant evidence comes from zopiclone, eszopiclone's racemic parent compound, which has been measured in breast milk in a small pharmacokinetic study showing detectable transfer with a milk-to-plasma ratio in the range of roughly 0.5. Because eszopiclone is chemically the active isomer within that racemic mixture, similar or proportional transfer into breast milk is the pharmacologically expected pattern, though this has not been directly confirmed for eszopiclone in a published study.

The NIH's LactMed database (https://www.ncbi.nlm.nih.gov/books/NBK501043/) is the standard reference clinicians use for drug-in-milk data and is a reasonable starting point for an individual patient conversation, though prescribers should check it directly for the current eszopiclone-specific entry rather than relying on secondhand summaries, since database entries are updated over time.

Practically, the concern for a breastfeeding infant is sedation-related: excessive drowsiness, weak or reduced feeding, and slower-than-expected weight gain. These are the same effects seen with any sedative-hypnotic that reaches an infant through milk, and they are the specific things a pediatric clinician should be asked to watch for if a nursing mother continues eszopiclone.

Third-trimester and near-delivery use: the neonatal adaptation concern

GABA-A receptor agonists as a class, including benzodiazepines and z-drugs, are associated with a neonatal adaptation syndrome when used in the days to weeks before delivery: hypertonia or hypotonia, tremor, irritability, feeding difficulty, and, in more severe cases, respiratory depression. No published case series specifically documents this pattern with eszopiclone, but the FDA label's general class warning applies, and case reports describing this pattern after maternal zopiclone use do exist in the literature.

A pregnant patient taking eszopiclone nightly should not stop abruptly near delivery without medical guidance, because rebound insomnia can be significant. A supervised taper, coordinated with the delivery team so neonatal staff know to watch the newborn in the first one to two days, is the standard clinical approach rather than an abrupt stop or an unplanned continuation.

What should come first: CBT-I and other alternatives

Cognitive behavioral therapy for insomnia (CBT-I) is recommended as first-line treatment for chronic insomnia, and both the American Academy of Sleep Medicine and the American College of Obstetricians and Gynecologists prioritize nondrug approaches over medications like eszopiclone during pregnancy and postpartum, since these approaches avoid fetal or infant exposure. Limited research on CBT-I adapted for pregnant women demonstrates meaningful reductions in insomnia severity, though specific trial details warrant verification in the original studies before precise citation.

When medication is judged necessary, the agents with more accumulated human pregnancy and lactation data than eszopiclone include:

Low-dose doxepin. Approved for insomnia at low doses; older, higher-dose tricyclic antidepressant pregnancy data have not shown a consistent teratogenic signal, and milk transfer at low insomnia doses is generally reported as low.

Diphenhydramine. An over-the-counter antihistamine with decades of pregnancy exposure history and no consistent malformation signal in large epidemiologic studies, though anticholinergic side effects and next-day grogginess limit how useful it is for everyone.

Melatonin. A short-half-life endogenous hormone. A Cochrane-affiliated review of melatonin in pregnancy found insufficient evidence to recommend it for insomnia and did not identify a clear safety signal in the limited data available, which is a statement about absence of evidence, not proof of safety.

None of these alternatives has definitive "proven safe in pregnancy" evidence either. The honest framing is that every option, including no medication, carries some uncertainty, and the decision is about which uncertainty a given patient and clinician are willing to accept.

Evidence boundary: what is established, what is plausible, what is not known

Established: Eszopiclone has no controlled human pregnancy studies. The FDA label instructs use only when benefit justifies fetal risk. Animal studies show a dose-dependent effect on fetal weight and skeletal maturation at doses far above human therapeutic exposure, without structural malformation. GABA-A agonists as a drug class can cause neonatal sedation and adaptation symptoms when used near delivery.

Plausible but unproven in humans: That eszopiclone crosses the human placenta at levels comparable to zopiclone. That eszopiclone transfers into human breast milk at a ratio similar to zopiclone. That z-drug class associations with preterm birth or low birth weight seen in zolpidem and zopiclone cohorts would also apply to eszopiclone specifically.

Not established: Whether eszopiclone causes any structural birth defect in humans. Whether in-utero or lactational eszopiclone exposure affects long-term neurodevelopment. The magnitude of infant sedation risk from a specific maternal eszopiclone dose during breastfeeding.

A decision framework for eszopiclone in pregnancy and lactation

This is not a substitute for individualized medical advice, dosing, or diagnosis. It is a structure for the conversation between a patient and her prescriber.

Step 1: Classify the insomnia.

• Mild, situational, pregnancy-related sleep disruption (physical discomfort, frequent urination) → start with sleep hygiene and CBT-I; medication is rarely the first move.
• Moderate to severe insomnia with daytime impairment, or insomnia layered on an anxiety or mood disorder → CBT-I is still first-line, but the timeline to relief matters, and a bridge medication may be reasonable while therapy takes effect.
• Insomnia that has already led to significant sleep deprivation, safety concerns, or a psychiatric crisis → this is a shared-decision conversation weighing untreated-illness risk against medication exposure, not a simple "avoid the drug" default.

Step 2: Check the trimester and delivery timing.

• First and second trimester, not close to delivery: the main concern is the theoretical animal-model signal on fetal growth and ossification at high doses; human data are absent rather than reassuring.
• Third trimester, especially the final weeks: add the neonatal adaptation syndrome concern; this is the point where tapering plans and delivery-team notification become relevant if the drug is continued.

Step 3: Weigh alternatives before defaulting to eszopiclone.

• Has CBT-I actually been tried, or only mentioned? A referral or app-based program should be attempted first when time allows.
• Has a lower-exposure-data option (low-dose doxepin, diphenhydramine) been considered and found inadequate or contraindicated?
• Is melatonin being considered as if it were proven safe? It is not; the evidence is described as insufficient, not favorable.

Step 4: If eszopiclone is continued, set explicit monitoring.

• Pregnancy: discuss growth monitoring in the third trimester if use is ongoing daily, and make sure the delivery team knows about eszopiclone use so newborn staff watch for sedation or withdrawal-type symptoms in the first one to two days.
• Lactation: watch the infant for excess drowsiness, weak feeding, or poor weight gain. If these appear, contact the pediatric clinician promptly; temporarily withholding breastfeeding with milk expressed and discarded for roughly one day has been used as a practical clearance strategy, though it has not been formally studied for eszopiclone specifically.

Step 5: Know when this becomes an urgent, not routine, conversation.

• A newborn who is unusually floppy, feeding poorly, or difficult to rouse after maternal eszopiclone exposure needs prompt pediatric evaluation, not a wait-and-see approach.
• A pregnant patient who stopped eszopiclone abruptly and is experiencing severe rebound insomnia, agitation, or other withdrawal-type symptoms should contact her prescriber the same day rather than restart or stop further medications on her own.

Common questions

Is eszopiclone proven unsafe in pregnancy? No. The evidence gap runs the other way: there are no controlled human studies at all, so the FDA label's caution reflects unknown risk, not confirmed harm.

Does taking Lunesta mean I have to stop breastfeeding? Not automatically. It means the infant should be watched for sedation and feeding problems, and the decision should be made with a prescriber who knows the maternal dose and the infant's clinical picture.

Is zopiclone data a good substitute for eszopiclone-specific data? It is the best available proxy because eszopiclone is the active isomer of zopiclone, but it is an approximation, not a direct measurement of eszopiclone in human pregnancy or milk.

What should I do if I find out I'm pregnant while taking eszopiclone? Contact the prescriber rather than stopping abruptly. A supervised transition toward CBT-I or a better-studied alternative is the standard approach, balanced against the risk of rebound insomnia.

References

• U.S. Food and Drug Administration. LUNESTA (eszopiclone) prescribing information, revised 2014. https://www.accessdata.fda.gov/drugsatfda_docs/label/2014/021476s030lbl.pdf
• National Library of Medicine. LactMed: Drugs and Lactation Database, eszopiclone entry. https://www.ncbi.nlm.nih.gov/books/NBK501043/

Several claims in the original draft of this article, including specific effect-size figures from epidemiologic studies and two attributed physician quotations, could not be verified against a confirmed primary source and were either removed, converted to unattributed general statements, or explicitly flagged above as requiring verification before clinical citation. This page is pending qualified clinical review.