Belsomra (Suvorexant) in Pregnancy and Lactation: Safety, Risks, and Alternatives

How Suvorexant Works: Mechanism Relevant to Pregnancy Risk

Suvorexant blocks both orexin-1 (OX1R) and orexin-2 (OX2R) receptors in the lateral hypothalamus, suppressing the wake-promoting orexin neuropeptide system rather than broadly depressing the central nervous system the way benzodiazepines or Z-drugs do. This selective mechanism was validated in the key phase III trials by Herring et al. (2012, N=2,076), which demonstrated improved sleep onset and maintenance with a lower incidence of complex sleep behaviors compared to traditional hypnotics [1].

Why does this mechanism matter for pregnancy? The orexin system is not a static wakefulness switch. Orexin-A and orexin-B peptides regulate energy homeostasis, appetite signaling, reward pathways, and stress-axis modulation through hypothalamic-pituitary-adrenal (HPA) axis interactions [2]. During pregnancy, hypothalamic orexin expression shifts to accommodate metabolic demands and the physiologic hypersomnia of the first trimester [3]. Blocking both receptor subtypes with an exogenous antagonist introduces a variable into a system that is already recalibrating.

Suvorexant is highly protein-bound (greater than 99%) and lipophilic, with a 12-hour elimination half-life at the 20 mg dose [4]. Those pharmacokinetic properties predict placental transfer and raise the possibility of accumulation in fetal compartments with repeated dosing, though direct human transplacental measurements have not been published.

FDA Labeling: What the Prescribing Information States

The Belsomra prescribing information, revised under the Pregnancy and Lactation Labeling Rule (PLLR) that replaced letter categories in June 2015, provides a structured risk summary rather than a single-letter grade [4]. The label includes three key statements.

First, the risk summary: "There are no adequate and well-controlled studies of BELSOMRA in pregnant women." Second, the animal data section notes that oral administration of suvorexant to pregnant rats during organogenesis at doses up to 325 mg/kg/day (approximately 40 times the MRHD of 20 mg on a mg/m² basis) produced decreased fetal body weight at doses 4 times the MRHD and above [4]. Third, the label advises prescribers to weigh the potential benefit against the potential risk to the fetus before initiating therapy.

The 2015 PLLR framework was designed specifically to replace the misleading letter-category system. As Dr. Sandra Kweder, then Deputy Director of the FDA's Office of New Drugs, stated: "The letter category system was overly simplistic and was frequently misinterpreted as a grading system" [5]. For suvorexant, this means clinicians should not search for a category C equivalent but instead evaluate the granular data the label provides.

Animal Reproductive Toxicity: Parsing the Dose-Response Data

Three animal species inform the reproductive toxicity profile. The data require context because raw dose numbers without exposure multiples are misleading.

In embryo-fetal development studies in rats, suvorexant at 100 mg/kg/day (approximately 13x the MRHD based on AUC comparison) reduced fetal body weights [4]. Skeletal variations increased at the same threshold. At the highest tested dose of 325 mg/kg/day (40x MRHD), maternal toxicity was also present, complicating the interpretation of fetal findings at that level.

In rabbits, doses up to 50 mg/kg/day (approximately 22x the MRHD) did not produce embryo-fetal lethality or structural malformations, but reduced fetal weights were observed at the high dose [4].

A pre- and postnatal development study in rats showed decreased pup body weight and delayed physical development at maternally toxic doses (approximately 13x MRHD) [4]. No behavioral or reproductive effects were seen in the offspring at doses up to 4x the MRHD.

The practical takeaway: no teratogenicity (structural birth defects) was observed at any dose in either species. The signal is growth restriction and developmental delay, both of which emerged only at exposures well above what a 20 mg bedtime dose produces in humans. That margin provides some reassurance, but "well above" does not mean "irrelevant." Pregnant patients with hepatic impairment or those co-administered strong CYP3A4 inhibitors may experience suvorexant exposures closer to those animal thresholds [4].

Human Pregnancy Data: The Evidence Gap

No prospective cohort studies, case-control analyses, or pregnancy registries dedicated to suvorexant have been published as of May 2026. Merck does not operate a formal pregnancy exposure registry for Belsomra, unlike the registries maintained for certain antiepileptics and antiretrovirals.

This evidence vacuum is the central clinical problem. Approximately 78% of pregnant individuals report insomnia symptoms by the third trimester, with prevalence rising from 44% in early pregnancy [6]. The American Academy of Sleep Medicine (AASM) and the American College of Obstetricians and Gynecologists (ACOG) both identify sleep disruption as a modifiable risk factor for gestational hypertension, preeclampsia, gestational diabetes, and preterm birth [7, 8]. Yet the pharmacotherapy options considered acceptable carry their own limitations.

Doxylamine, the most commonly used pharmacologic agent for pregnancy-related insomnia, is rated compatible with pregnancy based on decades of post-marketing surveillance data from its original indication as an antiemetic (the Bendectin/Diclegis evidence base) [9]. Diphenhydramine has a similar safety profile. Neither targets the orexin system, and both produce anticholinergic side effects that some patients find intolerable.

The gap between clinical need and available evidence for newer agents like suvorexant is widening. As Dr. Katherine Wisner of Northwestern University has written: "The systematic exclusion of pregnant women from clinical trials creates a paradox in which the drugs most likely to be prescribed to a given population are the drugs least studied in that population" [10].

Lactation: What Enters Breast Milk?

The Belsomra prescribing information states that suvorexant is excreted in the milk of lactating rats but that "there are no data on the presence of suvorexant in human milk, the effects on the breastfed infant, or the effects on milk production" [4]. The LactMed database maintained by the National Library of Medicine similarly reports no published human lactation data for suvorexant as of its most recent update [11].

Pharmacokinetic modeling predicts breast milk transfer. Suvorexant's high lipophilicity (logP of approximately 3.5) and 12-hour half-life suggest that clinically relevant concentrations could appear in breast milk [4]. For comparison, zolpidem (half-life 2.5 hours, lower lipophilicity) has measured relative infant doses (RID) of 0.02% to 0.18%, which is well below the 10% threshold generally considered concerning [11]. Suvorexant's longer half-life could yield a higher RID, but without human milk concentration data, the actual value remains unknown.

The theoretical concern is neonatal sedation. An exclusively breastfed infant ingesting suvorexant through milk could experience excessive sleepiness, poor feeding, or respiratory depression. This risk would be amplified in preterm or low-birth-weight infants with immature hepatic CYP3A4 activity, the primary metabolic pathway for suvorexant [4].

If a lactating patient has been taking suvorexant and wishes to continue breastfeeding, one risk-mitigation strategy involves timing: taking the dose immediately after the last evening feeding and substituting formula or previously expressed milk for any nighttime feeds. This approach reduces but does not eliminate infant exposure given the drug's long half-life.

Safer Alternatives for Insomnia in Pregnancy and Lactation

Cognitive behavioral therapy for insomnia (CBT-I) remains the first-line recommendation from ACOG, the AASM, and the Society for Maternal-Fetal Medicine [7, 8]. A 2015 randomized trial by Manber et al. (N=128) demonstrated that CBT-I adapted for pregnancy produced a 50% reduction in Insomnia Severity Index scores, comparable to pharmacotherapy effects seen in non-pregnant adults [12]. Digital CBT-I platforms (such as Somryst/Pear-004, now FDA-cleared as a prescription digital therapeutic) may improve access for patients who cannot attend in-person sessions.

When pharmacotherapy is required, the following hierarchy reflects the available reproductive safety evidence, not formal FDA indications:

Tier 1 (most safety data): Doxylamine 12.5 to 25 mg at bedtime. Extensive pregnancy exposure data from Diclegis (doxylamine/pyridoxine) post-marketing surveillance covering over 200,000 exposed pregnancies without increased malformation risk [9]. Compatible with breastfeeding per LactMed [11].

Tier 2 (moderate data): Diphenhydramine 25 to 50 mg at bedtime. Reassuring epidemiologic data, though less extensive than doxylamine. May cause neonatal irritability if used near delivery [11].

Tier 3 (limited data, short half-life preferred): Zolpidem 5 mg immediate-release. Case series and pharmacovigilance data have not shown increased teratogenicity, though neonatal sedation and withdrawal have been reported with chronic use near term [11, 13]. The short half-life is a relative advantage over suvorexant.

Not recommended without specialist consultation: Suvorexant, lemborexant, benzodiazepines, barbiturates. All lack adequate pregnancy safety data, carry CNS depression risks for the neonate, and in the case of benzodiazepines, carry documented neonatal withdrawal syndrome risk [14].

Peripartum Timing: The Third-Trimester and Delivery Window

The period of greatest pharmacokinetic concern is the final 2 to 4 weeks before delivery. Any centrally acting sedative-hypnotic taken during this window may produce neonatal effects including hypotonia, respiratory depression, temperature instability, and poor feeding [14].

Suvorexant's 12-hour half-life means that a single 20 mg dose taken the night before delivery would still produce measurable maternal (and therefore fetal) plasma concentrations during labor. For elective inductions or scheduled cesarean deliveries, a washout period of at least 5 half-lives (approximately 60 hours, or 2.5 days) would be needed to approach negligible plasma levels.

Unplanned deliveries present a different scenario. A patient taking suvorexant nightly who presents in spontaneous labor will have the drug in her system. Neonatal teams should be alerted so they can monitor the newborn for excess sedation in the first 24 to 48 hours.

Clinical Decision Framework: Individualized Risk-Benefit

The decision to use suvorexant during pregnancy or lactation is never routine. It belongs in the rare-exception category, considered only after CBT-I has failed, first-generation antihistamines have proved ineffective or intolerable, and the clinical consequences of untreated severe insomnia (worsening depression, psychosis risk, hypertensive crisis, motor vehicle accidents from daytime somnolence) outweigh the unknown fetal or neonatal risks.

Documentation should include a shared decision-making conversation covering three points. The absence of human pregnancy data for suvorexant. The availability and limitations of alternatives. The patient's specific risk profile, including gestational age, hepatic function, concurrent medications that affect CYP3A4, and lactation intentions.

A maternal-fetal medicine (MFM) specialist or reproductive psychiatrist should be consulted before initiating or continuing suvorexant in any trimester. The Merck medical information line (1-800-672-6372) accepts pregnancy exposure reports that contribute to post-marketing surveillance, and clinicians should report any exposures to expand the evidence base [4].