Insomnia · Patient experience
Patients describe waking up paralyzed on these sleeping pills. The FDA has never required a study of this side effect.
The DEA wants to move Dayvigo and two similar sleeping pills to its lowest tier of control because it says they do not cause dependence. The agency leaves side effects out of that judgment and has called them a matter for the FDA. In the 260 reviews of Dayvigo behind a study I led, the side effect that most strongly set apart the people who quit or considered quitting was sleep paralysis.
Based on
Armanious, A. J., Dunn, A. R., Martin, C. E., & James, M. H. (2026). Perceived effectiveness, tolerability, and discontinuation of lemborexant in real-world use: Evidence from online patient reports. Sleep Medicine, 109264. https://doi.org/10.1016/j.sleep.2026.109264
A mixed-methods study of 260 self-selected online reviews of lemborexant for insomnia from June 2020 to September 2023, hand-coded and matched to each reviewer's rating where they left one.
Disclosures: None of the study's authors have received funding from or have a financial relationship with Eisai, the maker of lemborexant. Separately, this piece explains research, not medical advice; any decision about a medication is one for you and your clinician.
You take a new sleeping pill the size of a lentil and drift off. Ninety minutes later, you wake up inside a terrifying dream. You try to move, and your body will not respond.
In August 2020, a patient whose review I later studied lived through that night after their first dose. "Wanted to scream and could not make a sound," they wrote. They managed to touch their husband's arm and wake him. "He said he never saw such terror in my eyes. Sat with me [for] 2 hours until I was over it." They rated the drug a 1 out of 10 and added, "Never again."
That is sleep paralysis. The mind is awake while the body is still in rapid eye movement (REM) sleep, the stage that brings our most vivid dreams and leaves most muscles slack. People across cultures once blamed sleep paralysis on a creature that sat on the sleeper. Henry Fuseli's 1781 painting The Nightmare showed that creature as a demon perched on a sleeping woman.
The pill in that review was lemborexant, sold by Eisai as Dayvigo. It is one of three orexin blockers approved in the United States, along with Belsomra and Quviviq. All three block the signal of orexin, a brain chemical whose main job is to keep you awake. Older, better-known sleeping pills like Ambien work by boosting the effect of GABA, a chemical that quiets activity across the entire brain.
The Drug Enforcement Administration (DEA) proposed in August 2026 to move all three from Schedule IV to Schedule V, the lowest of five federal tiers of drug control. The DEA says these pills carry a potential for abuse similar to that of Ambien and other Schedule IV drugs but cause no physical or psychological dependence. By that measure, the orexin blockers are the safer ones in the eyes of the DEA. The notice itself never mentions nightmares or sleep paralysis, though, and the public's chance to comment on the proposal closed on September 10, 2026, with the DEA yet to issue a final decision when this was written.
Scheduling turns on abuse and dependence. When the DEA first scheduled Belsomra in 2014, it wrote that concerns about a drug's side effects were "outside the scope" of its scheduling authority. That leaves those side effects to the Food and Drug Administration (FDA), which has put sleep paralysis on all three labels but never required a study of that risk. The FDA should require one, whatever schedule the DEA settles on.
Patient reviews show why. For a study published in the journal Sleep Medicine, a colleague and I read all 260 reviews of lemborexant posted on Drugs.com between June 2020 and September 2023. We counted an effect only if the review reported it for lemborexant rather than for a drug the reviewer may have taken before. About a third of the reviewers described vivid dreams or nightmares, and about a fifth described some experience with sleep paralysis.
Seventy reviewers wrote about stopping the drug or about whether to keep taking it. We called them discontinuers. The other 190 were far more likely to say the drug had helped their sleep. Two in five discontinuers described sleep paralysis, against about one in seven of the rest. No other complaint split the two groups as widely.
People who write reviews have a story to tell, so they are not a fair sample of everyone on the drug. The numbers show how much the dreams and the paralysis mattered to those who wrote about them, not how often either one happens.
of the 260 I read left a rating from 1 to 10. Most of them piled up at the two ends, 111 on a 1 and sixty-two on a 10.
If you have taken it, where would you put it? See what others have said.
| Rating | Reviewers |
|---|---|
| 1 (lowest) | 111 |
| 2 | 14 |
| 3 | 3 |
| 4 | 7 |
| 5 | 6 |
| 6 | 6 |
| 7 | 7 |
| 8 | 16 |
| 9 | 18 |
| 10 (highest) | 62 |
| Total | 250 |
No one has yet shown why orexin blockers would produce sleep paralysis, though other researchers have proposed a reason. Orexin, in addition to keeping us alert, also helps hold REM sleep back while we are awake. Blocking orexin could potentially loosen that hold. In some people, REM dreams and paralysis might then slip into the minutes around falling asleep or waking. That already happens in people with narcolepsy type 1, who have lost most of the brain cells that make orexin. A pill that blocks the same signal may bring on similar nights.
The FDA flagged such nights before it approved any of the three drugs. In a 2013 letter turning down Belsomra's first application, it said hallucinations and sleep paralysis "can be terrifying" and can worsen the insomnia the pill is meant to treat. The letter listed them among the drug's "potentially serious and life-threatening adverse effects," alongside a narcolepsy-like syndrome that included sudden attacks of muscle weakness. Partway through the two trials that won Dayvigo its approval, the FDA had Eisai ask every patient about falls at every visit. Neither the trials' study plans nor their published methods describe asking about nightmares or paralysis once people were on the drug. One trial did ask everyone about sleep paralysis on a questionnaire, but only before they joined.
So during treatment, the trials counted only the sleep paralysis people brought up on their own. About one or two in a hundred on the drug reported it, and no one given the placebo pill did. That figure is on the label, which prompts doctors to explain the paralysis to patients, and the label names nightmares among the side effects that most often made people in the trials stop. The leaflet that patients also receive lists paralysis in one line under serious side effects but never mentions dreams.
In 2022, an FDA doctor assessing Quviviq called the way the drug's trials counted sleep paralysis and hallucinations "insensitive" and said it "may be unreliable." The doctor still concluded that "additional formal study is not necessary," because the events were "potentially distressing but not otherwise intrinsically harmful or dangerous." One reviewer in our study who gave lemborexant a 3 described the trade-off:
Worked quite well, but the sleep paralysis was not worth it. … I got great sleep. It came at a price.
A patient's review on Drugs.com.
None of the follow-up studies the FDA has required of Belsomra, Dayvigo or Quviviq concern these nights. No source has thus settled how often they happen. Yet sleep paralysis stands out for all three orexin blockers in the FDA's database of side-effect reports filed after approval. That database includes only what people chose to report, so it cannot provide a rate either.
For people with insomnia, the study these drugs need has not been published. It would ask everyone it enrolled about nightmares and sleep paralysis by name from the first night, with a placebo group and a record of who stops and why. The FDA can require the makers of all three to run it. The agency called sleep paralysis potentially serious in 2013. In 2022, it listed sleep paralysis among Quviviq's "serious risks" in one assessment and called the paralysis not intrinsically dangerous in another. It has never looked closely enough to know which.
A study would answer that. Its findings should then reach the next analysis FDA staff write for the DEA. The law requires that analysis to consider "what, if any, risk there is to the public health." The 2014 rulemaking that scheduled Belsomra called concerns about side effects outside its scope yet counted the drug's lingering sedation as one such risk. The night that sent one reviewer reaching for their husband's arm should count too.
For now, the most detailed accounts of nights like theirs are the ones patients wrote down themselves.
Common questions
Doesn't sleep paralysis happen to people who take no pill at all?
It does. A review of thirty-five studies found that about one person in thirteen has at least one episode in their lifetime. That is why the study this piece calls for needs a placebo group. The two trials that won Dayvigo its approval had one. About one or two in a hundred people on the drug in those trials brought up sleep paralysis on their own, and no one given the placebo pill did. The patient reviews I studied cannot tell how many of the nights of paralysis they describe would have happened anyway.
What would the DEA's proposed move to Schedule V mean for someone who takes Dayvigo, Belsomra or Quviviq?
The DEA's notice says the companies and pharmacies that handle the drugs would keep the same registration and record-keeping duties. For patients, the main federal difference is refills. Federal rules say a Schedule IV prescription runs out after five refills or six months, whichever comes first. It must then be renewed. Those limits do not apply to Schedule V. States can set their own limits.
How can someone report a side effect such as sleep paralysis?
Through the FDA's MedWatch program. Patients and families can fill out its online form themselves. Doctors can file a report too, but they are not required to. Reports sent through MedWatch go into the FDA's database of side-effect reports filed after approval.
Contact
If this is the kind of work you care about,
I'd like to hear from you.