News · Sleep
A menopause sleep drug worked at week 4, then didn't
A SLEEP trial gave 110 postmenopausal women elinzanetant or placebo and measured sleep in a lab. The 22-minute advantage at four weeks had vanished by twelve.
Based on a peer-reviewed randomized trial in SLEEP
- The NIRVANA phase II trial randomized 110 postmenopausal women aged 40 to 65 to elinzanetant 120 mg or placebo for 12 weeks, publishing in SLEEP.
- All had moderate-to-severe hot flushes (at least 20 a week) and measured sleep disruption of at least 30 minutes awake after falling asleep.
- Sleep was measured three ways: laboratory polysomnography, a daily diary, and contactless home monitoring in a subgroup.
- At week 4, elinzanetant reduced lab-measured time awake after sleep onset by 22.50 minutes versus placebo (90% CI, -35.47 to -9.52).
- At week 12 the same measure showed a difference of -0.94 minutes (90% CI, -13.96 to 12.09). Effectively nothing.
- Effects favoring the drug across the full 12 weeks came from post hoc modeling, not the prespecified endpoint.
- Intervals here are 90%, not the usual 95%, which is a lower bar for statistical clarity.
- Phase II, 110 women, 12 weeks. This is an early signal, not a verdict on the drug.
Menopausal sleep disruption is one of the most common complaints in medicine and one of the least well served. So a trial that puts a new non-hormonal drug in a sleep laboratory is worth reading closely, including the part that did not work.
A study in SLEEP did exactly that. The NIRVANA trial explored elinzanetant’s effect on multimodal measures of objective and subjective sleep disturbance in postmenopausal women. The four-week result was good. The twelve-week result was not.
Why a sleep lab, when earlier trials already looked positive
Elinzanetant has been through larger trials already, and they were encouraging. Elinzanetant reduced self-reported sleep disturbance in postmenopausal women with moderate-to-severe vasomotor symptoms in the Phase III OASIS trials.
Self-reported is the operative word. How well someone thinks they slept and how much they were actually awake are different measurements, and they come apart more often than anyone would like. The authors note that menopausal sleep disturbances can negatively impact quality of life, which is the reason to check whether the subjective gains show up on instruments.
This trial was built to check. It used polysomnography, the full laboratory setup, alongside a diary and a home monitor.
Who was in it
Entry was restricted to women with a genuine problem on both fronts.
Postmenopausal women aged 40-65 with moderate-to-severe hot flushes of at least 20 per week and sleep disturbance of at least 30 minutes awake after sleep onset were randomized 1:1 to receive elinzanetant 120 mg or placebo for 12 weeks. There were 110 of them.
Sleep was assessed three ways: by polysomnography, a daily Sleep Diary, and nightly contactless home monitoring in a subgroup. Measuring the same thing three ways is a genuine strength, because each method fails differently.
The four-week result
Early on, the drug did what it was supposed to do.
Elinzanetant was estimated to reduce lab-measured wakefulness after sleep onset at Week 4 versus placebo, by 22.50 minutes. For a woman lying awake in the middle of the night, twenty-two minutes is not a trivial amount of her night back.
The twelve-week result
Then it went away.
Minimal group differences between treatment arms were estimated at Week 12 for lab-measured wakefulness after sleep onset - a difference of -0.94 minutes. Less than a minute, in a measure where the four-week gap had been twenty-two.
That is the prespecified comparison at the trial’s endpoint, and it is as close to nothing as a number gets.
What the authors did next, and how much it counts
Faced with a null endpoint, the researchers pooled everything.
When including all available polysomnography data across the 12-week period, post hoc random coefficient modeling indicated treatment effects favoring elinzanetant for wakefulness after sleep onset, total sleep time, sleep efficiency, latency to persistent sleep, number of awakenings, arousal index, and % Stage N2 sleep. The other two measurement methods agreed: treatment effects on wakefulness and awakenings favoring elinzanetant were observed with the home monitor and Sleep Diary across 12 weeks.
Two things are true about this. Several independent measurement methods pointing the same way is real supporting evidence. And an analysis chosen after seeing the data is weaker than the endpoint agreed beforehand, which is precisely why endpoints are agreed beforehand.
It is also worth noting the intervals throughout are 90%, not the customary 95%. That is a deliberately lower bar for calling something clear.
What the trial cannot show
One hundred and ten women, twelve weeks, phase II. This is a signal-finding study, not a verdict.
The authors’ conclusion is carefully hedged: in this sleep-focused study, elinzanetant reduced wakefulness after sleep onset at Week 4 and led to estimated improvements in sleep continuity over 12 weeks, supported by multimodal subjective and objective sleep outcomes.
The more useful lesson sits underneath the drug. Earlier trials showed women reporting better sleep; this one found the objective effect hard to sustain past a month. Both can be true, and the gap between them is the thing to watch as this drug moves forward.
People also ask
What is elinzanetant?
A non-hormonal drug that blocks two neurokinin receptors, NK-1 and NK-3. Those receptors sit in the brain circuit that regulates body temperature, which is why blocking them reduces hot flushes without using estrogen. It has already shown reductions in self-reported sleep disturbance in the larger phase III OASIS trials in women with moderate-to-severe hot flushes, which is what made a sleep-focused study worth running.
What does 'wakefulness after sleep onset' measure?
How many minutes you spend awake after first falling asleep. It is the standard laboratory measure of fragmented sleep, and it captures the specific complaint most menopausal women describe: getting to sleep is manageable, staying asleep is not. Because it is recorded by polysomnography rather than self-report, it is not vulnerable to how a person felt about their night.
Why did the effect disappear by week 12?
The trial does not say, and the honest answer is that nobody knows from this data. Possibilities include the placebo group catching up, the drug's effect on sleep fragmentation genuinely fading, or a 110-person trial being too small to hold a modest difference steady over three months. What can be said is that the prespecified comparison at week 12 was -0.94 minutes, which is indistinguishable from no difference at all.
So do the 12-week results mean anything?
They mean something less than a prespecified result would. When the researchers pooled all available lab data across the 12 weeks using post hoc random coefficient modeling, treatment effects favored elinzanetant across several sleep measures, and the diary and home-monitoring data pointed the same way. An analysis chosen after seeing the data carries less weight than the endpoint agreed in advance, so this is supporting evidence rather than proof.
Should anyone act on this?
This is general information rather than advice, and a phase II trial in 110 women is not a basis for a treatment decision. What it usefully establishes is that measured sleep and reported sleep are not the same thing: earlier trials found improvement in what women said, and this one found the objective effect harder to sustain. Anyone weighing options for menopausal sleep disruption should have that conversation with a clinician.