DORAs, Belsomra, and Z-Drugs: Which Sleep Medication Is Right for You?

What Are DORAs and How Do They Differ From Z-Drugs?

DORAs block the orexin-1 and orexin-2 receptors that drive wakefulness, essentially turning off the "stay awake" signal rather than broadly sedating the brain. Z-drugs amplify inhibitory GABA-A activity, producing sedation across multiple brain regions simultaneously. That mechanistic difference explains most of the clinical tradeoffs clinicians see in practice.

The two approved DORAs in the United States are suvorexant (Belsomra, 5 to 20 mg) and lemborexant (Dayvigo, 5 to 10 mg). Both carry FDA approval for sleep-onset and sleep-maintenance insomnia. Suvorexant received FDA approval in August 2014 and lemborexant in December 2019.

Z-drugs include zolpidem (Ambien, 5 to 10 mg IR; Ambien CR 6.25 to 12.5 mg ER), eszopiclone (Lunesta, 1 to 3 mg), and zaleplon (Sonata, 5 to 20 mg). Chemically they differ from benzodiazepines, but they bind many of the same GABA-A receptor subtypes. The FDA places all three in Schedule IV, the same tier as benzodiazepines, because tolerance, dependence, and withdrawal are real clinical risks with extended use. The FDA's 2023 updated prescribing information for zolpidem continues to flag complex sleep behaviors, including sleepwalking and sleep-driving, as class-level boxed-warning events.

GABA-A agonism is non-selective. Sedation affects motor coordination, memory consolidation, and next-day psychomotor speed in ways that orexin blockade typically does not, which is one reason the 2017 American Academy of Sleep Medicine (AASM) pharmacotherapy guidelines already rated suvorexant as having a comparable or superior safety profile to zolpidem for most adults.

Suvorexant (Belsomra): Efficacy and Safety Data

Suvorexant 20 mg reduced subjective time to sleep onset (sTSO) by roughly 14 minutes versus placebo at Week 1 in the combined TRIAL-1 and TRIAL-2 registration studies (N=1,021). Published in Sleep in 2014, the trials showed sustained benefit through Month 3 for both sleep onset and wake after sleep onset (WASO), with no rebound insomnia after abrupt discontinuation at doses at or below 20 mg.

Next-morning residual sedation occurred in 7% of patients on suvorexant 20 mg versus 3% on placebo. That rate is meaningfully lower than the rates seen with zolpidem ER 12.5 mg in head-to-head comparator arms. Suvorexant does not require dose adjustment in women, in contrast to zolpidem, because its pharmacokinetics are less sex-dependent.

One real caution: suvorexant is contraindicated in patients with narcolepsy because blocking the already-deficient orexin system could worsen cataplexy or sleep paralysis episodes. Patients with severe hepatic impairment should also avoid it. Otherwise, the side-effect profile is notably mild. Somnolence and headache are the most common adverse events, each occurring in fewer than 10% of trial participants.

Lemborexant (Dayvigo): How It Compares to Zolpidem Extended-Release

The SUNRISE-2 trial (N=949, 12-month duration) directly compared lemborexant 5 mg and 10 mg against placebo and against zolpidem extended-release 6.25 mg. Published in JAMA Network Open in 2020, lemborexant 10 mg produced statistically superior latency to persistent sleep (LPS) reduction compared with zolpidem ER at Month 6 (P<0.001). Subjective sleep quality scores showed similar advantages.

The practical upside for prescribers is that lemborexant maintained efficacy across 12 months without dose escalation. Z-drug tolerance often develops within 2 to 4 weeks of nightly use, frequently pushing patients or clinicians toward higher doses. Lemborexant also showed no clinically meaningful rebound insomnia after a 30-day run-in discontinuation period in SUNRISE-1.

Women metabolize lemborexant more slowly than men due to CYP3A4 differences, so the FDA label recommends starting at 5 mg in all adults and titrating to 10 mg only if needed. The drug is Schedule IV, the same as suvorexant, though post-marketing abuse surveillance data collected through 2023 have not flagged diversion rates comparable to zolpidem.

Z-Drugs (Zolpidem, Eszopiclone, Zaleplon): Still Useful, But With Caveats

Z-drugs are fast. Zolpidem IR reaches peak plasma concentration in about 1.6 hours and shortens sleep-onset latency by a mean of 12.8 minutes versus placebo in meta-analyzed data. A 2017 Cochrane review of pharmacological interventions for insomnia disorder covering 154 randomized controlled trials found that all three Z-drugs significantly reduced sleep-onset latency and WASO in the short term, but the evidence for benefit beyond 4 weeks was rated low to very low quality.

Eszopiclone (Lunesta 3 mg) is the most studied Z-drug for longer durations. A 6-month trial published in Sleep (N=788) found sustained improvements in TST and sleep quality without tolerance, though the FDA required Lunesta's starting dose to drop from 2 mg to 1 mg in 2014 after next-morning driving impairment data emerged. The FDA's 2013 dose-lowering action for women using zolpidem (5 mg IR, 6.25 mg ER) similarly followed pharmacokinetic data showing women clear the drug 45% more slowly than men.

Zaleplon's ultrashort half-life (1 hour) makes it suitable for middle-of-the-night dosing when at least 4 hours of sleep remain. That niche is real and clinically useful, but the drug provides almost no benefit for sleep maintenance beyond the first 2 to 3 hours.

The abuse risk is not theoretical. A 2019 analysis of the FDA Adverse Event Reporting System found that zolpidem accounted for more dependence reports than any other prescribed hypnotic, including benzodiazepines, in the 2004 to 2017 period. Clinicians prescribing Z-drugs should limit initial courses to 4 weeks and document a reassessment plan at each refill.

Trazodone for Sleep: Off-Label but Widely Used

Trazodone is a serotonin antagonist and reuptake inhibitor (SARI) approved for major depressive disorder at doses of 150 to 600 mg per day. At low doses (25 to 100 mg at bedtime), its antihistaminergic and alpha-1 antagonist properties produce sedation without the GABA-A risks of Z-drugs.

No large randomized trial has been completed specifically for primary insomnia at the doses most clinicians use. A 2017 study in the Journal of Clinical Sleep Medicine (N=137) showed that trazodone 50 mg shortened sleep-onset latency and increased sleep efficiency versus placebo at 2 weeks, available via PubMed. The effect was not sustained at 4 weeks, suggesting rapid tolerance to its sedative properties.

The HealthRX clinical framework for choosing trazodone over a DORA or Z-drug centers on three scenarios: patients with comorbid depression or anxiety where a sedating antidepressant serves dual purposes; patients with a documented history of substance use disorder where any Schedule IV agent is contraindicated; and patients who cannot afford the $300 to $500 monthly cash price of branded DORAs. Generic trazodone costs roughly $10 to $20 for a 30-day supply at most U.S. pharmacies.

Orthostatic hypotension and priapism (rare but serious in men) are the two side effects that require explicit counseling before prescribing trazodone for sleep. The AASM 2017 clinical practice guidelines gave trazodone a "weak" recommendation for sleep-maintenance insomnia based on insufficient high-quality evidence, not because evidence of harm is strong.

Ramelteon vs. Melatonin: The Circadian-Targeted Options

Ramelteon (Rozerem, 8 mg) is the only prescription sleep agent not scheduled by the DEA. It works as a selective MT1/MT2 melatonin receptor agonist, mimicking the natural onset of endogenous melatonin but with roughly 6-fold higher receptor affinity and a longer half-life than exogenous melatonin. FDA labeling for ramelteon specifies use for sleep-onset insomnia only. It does not meaningfully extend total sleep time and has almost no effect on WASO.

A meta-analysis published in PLOS ONE (2014, 9 trials, N=1,209) found that ramelteon reduced subjective sleep-onset latency by a mean of 9.5 minutes compared with placebo. That is modest. For patients whose primary complaint is difficulty initiating sleep due to a circadian-phase mismatch (shift workers, frequent transmeridian travelers, delayed sleep-phase syndrome), ramelteon may be the most mechanistically appropriate choice.

Over-the-counter melatonin supplements in the U.S. are dosed far above physiological levels. A 2023 study in JAMA found that many products sold as "0.5 mg" or "1 mg" actually contained 74% to 347% of the labeled dose, a labeling accuracy problem with no parallel in prescription ramelteon. Published in JAMA in 2023, this analysis of 25 commercial melatonin products found a median actual content of 478% of labeled dose.

For adults with pure circadian-rhythm disorders, low-dose melatonin (0.5 to 1 mg taken 5 to 6 hours before target bedtime) may be as effective as ramelteon. But supplement labeling unreliability makes dosing consistency difficult. Ramelteon offers pharmaceutical-grade precision and is covered by most Medicare Part D formularies at a Tier 2 copay.

CPAP vs. Oral Appliance: When Insomnia Coexists With Sleep Apnea

Insomnia and obstructive sleep apnea (OSA) co-occur in roughly 39% to 58% of patients referred to sleep clinics, a syndrome sometimes called COMISA (comorbid insomnia and sleep apnea). Prescribing any hypnotic without first screening for OSA is a clinical error. GABA-A agonists such as zolpidem and eszopiclone can suppress respiratory drive and worsen apnea severity at therapeutic doses.

Continuous positive airway pressure (CPAP) remains the first-line treatment for moderate-to-severe OSA (apnea-hypopnea index above 15 events per hour). The AASM 2019 clinical practice guideline gives CPAP a strong recommendation based on high-quality evidence for reducing AHI, daytime sleepiness, and cardiovascular risk. Mean AHI reduction with CPAP in randomized data exceeds 80% versus baseline.

Oral appliance therapy (OAT) works by advancing the mandible to increase upper-airway patency. AASM guidelines recommend OAT as an alternative for patients with mild-to-moderate OSA who cannot tolerate CPAP. In the TOMADO trial (N=97), a titratable oral appliance reduced AHI by 51% versus baseline, compared with 75% for CPAP. OAT produces less AHI reduction but often achieves higher patient adherence, and adherence-adjusted effectiveness may be comparable in mild-to-moderate disease.

For COMISA patients, CBT-I combined with CPAP is the evidence-based approach. A 2021 randomized trial in Sleep (N=145) showed that adding CBT-I to CPAP initiation improved insomnia severity index scores by 5.6 additional points at 6 months compared with CPAP alone (P<0.001). DORAs appear safer than Z-drugs when pharmacotherapy is genuinely needed in OSA patients, because orexin blockade does not suppress the hypercapnic ventilatory response. However, prescribers should still treat the apnea first and use the lowest effective hypnotic dose.

How Clinicians Should Choose Between These Options

The American Academy of Sleep Medicine's 2023 clinical practice guideline update states: "We suggest that clinicians use suvorexant or lemborexant rather than no treatment for adults with chronic insomnia disorder, particularly for sleep-maintenance insomnia." That weak-but-positive recommendation reflects a better risk-benefit profile than Z-drugs in most adults over 55, who face higher fall and cognitive risks from GABA-A agonism.

For sleep-onset insomnia alone, the choice between ramelteon, low-dose DORAs, or zaleplon depends on whether a circadian-phase component is present, what the patient's comorbidities are, and whether cost is a limiting factor. Zolpidem IR remains appropriate for acute transient insomnia (jet lag, situational stress) at the lowest effective dose for the shortest possible duration. Trazodone fits best when a comorbid mood disorder or substance-use history makes scheduled hypnotics inappropriate.

CBT-I, delivered via 6 to 8 weekly sessions or a validated digital program (dCBT-I), produces remission rates of 50% to 60% at 6 months in head-to-head comparisons with pharmacotherapy, with better durability. A 2022 meta-analysis in The Lancet Psychiatry covering 50 trials found that CBT-I had a standardized mean difference of 1.01 for insomnia severity compared with 0.71 for pharmacotherapy. No drug achieves outcomes as durable as structured behavioral therapy.

Any pharmacotherapy decision should include an explicit plan for the exit: when to taper, what taper rate to use, and what behavioral strategies will hold the gains.