Two non-hormonal menopause drugs look similar on hot flushes, differ on sleep

Two neurokinin receptor antagonists now exist for menopausal vasomotor symptoms, each established in its own placebo-controlled trials, and neither has been tested against the other. A matching-adjusted indirect comparison published in Maturitas on 24 November attempts the comparison anyway — with a method whose limits are worth understanding before its results are [s1].

Why an indirect comparison at all

Hormone therapy remains the most effective treatment for vasomotor symptoms, but many women cannot or do not wish to use it [s1]. That leaves the non-hormonal options, and the two neurokinin receptor antagonists — fezolinetant and elinzanetant — have both established efficacy in clinical trials [s1].

A clinician choosing between them has no head-to-head evidence. Matching-adjusted indirect comparison, or MAIC, is the standard technique for that situation: it uses individual patient-level data from one set of trials, reweights that population to match the published baseline characteristics of another set, and then compares outcomes across the reweighted populations, controlling for differences in treatment effect modifiers [s1].

Here, patient-level data from the fezolinetant SKYLIGHT-1 and SKYLIGHT-2 trials were used to match the population and study design of the elinzanetant OASIS-1 and OASIS-2 trials [s1]. Treatment effect modifiers were identified and the MAIC analyses performed on that basis [s1].

The outcomes examined were vasomotor symptom frequency and severity, the Patient-Reported Outcomes Measurement Information System Sleep Disturbance Short Form 8b (PROMIS SD SF 8b), and the Menopause Specific Quality of Life Questionnaire (MENQOL) [s1].

What it found

The OASIS-1 and OASIS-2 trial populations had higher mean vasomotor symptom frequency and severity at baseline, and greater sleep disturbance, than the SKYLIGHT populations [s1] — which is precisely the kind of imbalance MAIC exists to adjust for, and a reason a naive comparison of the published trial results would have been misleading.

After adjustment, fezolinetant and elinzanetant had comparable efficacy in reducing vasomotor symptom frequency and severity [s1].

Sleep disturbance improved with both drugs, but more with elinzanetant [s1]. That difference did not translate into a significant difference in overall quality of life as measured by MENQOL [s1].

The authors' conclusion is that the two have a similar efficacy profile for reduction of vasomotor symptom frequency and severity, and that the difference in sleep disturbance did not produce a difference in quality of life [s1].

The methodological caveat that governs everything above

A MAIC is not a trial. It can adjust only for characteristics that were measured and reported, and it inherits every difference between the trial protocols that reweighting cannot fix — different outcome timing, different sites, different eras, different placebo responses. Unobserved prognostic differences between the SKYLIGHT and OASIS populations remain unaddressed by construction.

The direction of any residual bias is not knowable from the analysis itself. A finding of "no meaningful difference" from a MAIC is best read as "no difference large enough to survive this adjustment," not as equivalence demonstrated.

The sleep finding is where the interesting question sits

The result that elinzanetant improved sleep disturbance more, without a corresponding quality of life difference [s1], invites two readings. Either the sleep advantage is real but too small to register on a broad quality-of-life instrument, or MENQOL is not sensitive enough to detect it. The analysis cannot distinguish them.

That matters because sleep disturbance is not a minor symptom in this population. A separate Maturitas analysis published on 1 November quantified its association with functioning, using data from peri- and postmenopausal women aged 40 to 65 in the National Health and Wellness survey — 27,621 women in the United States (2019/2021) and 20,220 across France, Germany, Italy, Spain, and the UK (2017/2020) [s2].

Among postmenopausal women, those with both sleep disturbances and vasomotor symptoms had the worst outcomes. In the US survey, adjusted estimated marginal means for that group versus women with neither symptom were 11.9% versus 9.3% for presenteeism, 12.8% versus 10.2% for work productivity impairment, and 20.3% versus 16.2% for activity impairment [s2]. In Europe the corresponding figures were 17.4% versus 12.9%, 18.8% versus 14.4%, and 28.1% versus 20.9% [s2].

Worse work-productivity outcomes were not clearly observed among perimenopausal women with symptoms [s2]. But both peri- and postmenopausal women with symptoms had more healthcare visits in the previous six months than those with neither, especially those with sleep disturbances irrespective of co-occurring vasomotor symptoms [s2].

That last clause is the link back to the drug comparison: in the burden data, sleep disturbance tracked healthcare use whether or not vasomotor symptoms were present [s2]. A treatment difference on sleep is therefore not obviously trivial, even if it did not move MENQOL [s1].

What is still missing

A direct randomised comparison of the two drugs. Until one exists, the choice between them rests on indirect evidence, tolerability, and availability — and on outcomes, like sleep, where the indirect evidence points in a direction it cannot confirm.

This article is informational and does not constitute medical advice.

Sources

• [s1] Fezolinetant compared with elinzanetant for the treatment of vasomotor symptoms associated with menopause: A matching-adjusted indirect comparison. Maturitas, 24 November 2025. https://doi.org/10.1016/j.maturitas.2025.108782
• [s2] The burden of sleep disturbances and vasomotor symptoms on work productivity, activity impairment and healthcare resource use in perimenopausal and postmenopausal women. Maturitas, 1 November 2025. https://doi.org/10.1016/j.maturitas.2025.108758