Why Nitroglycerin Failed to Stop Menopausal Hot Flashes

A failed heart-drug experiment clarified where hot flashes begin

Hot flashes look like a blood-vessel problem: skin flushes, sweat appears, and heat seems to pour out through the body's surface. A 2023 randomized trial tested that idea directly with transdermal nitroglycerin, a vasodilator used in cardiovascular care. The drug produced the vascular tolerance investigators intended to create. It did not reduce hot flashes.

That failure matters because the trial did more than test an off-label treatment. It tested a mechanistic claim. If peripheral blood-vessel signaling drove hot flashes, making vessels tolerant to nitric-oxide-mediated dilation might have reduced episodes. Instead, the study points upstream, toward the brain's temperature-control circuitry.

The practical map is less mysterious than the physiology: peripheral vascular manipulation has failed; central neurokinin-3 (NK3) receptor blockade has strong randomized-trial support; and wrist cooling remains a plausible but early-stage adjunct.

Nitroglycerin induced tolerance, but hot flashes continued

In the Huang et al. randomized clinical trial, 141 perimenopausal and postmenopausal women reported an average of 10.8 total hot flashes and 8.4 moderate-to-severe hot flashes per day at baseline. Participants used transdermal nitroglycerin titrated from 0.2 to 0.6 mg per hour or placebo for 12 weeks. The primary endpoint was change in total and moderate-to-severe hot-flash frequency at week 12.

Nitroglycerin did not beat placebo at that endpoint. The between-group difference was −0.1 total flashes per day (95% CI, −1.2 to 0.4) and −0.5 moderate-to-severe flashes per day (95% CI, −1.6 to 0.7); both results were nonsignificant. At week 5, moderate-to-severe flashes differed by −1.1 episodes per day (95% CI, −2.2 to 0; P=0.05), but that borderline result had disappeared by week 12.

The striking part was headache. Nitroglycerin commonly causes headache through nitric-oxide-mediated vasodilation, so headache offered a participant-reported check on whether vascular responsiveness was changing. At week 1, 67.1% of nitroglycerin users reported headache, versus 5.6% of placebo users. By week 12, headache had fallen to roughly one participant in each group. The drug's vascular effect appeared to become tolerant over time, while hot-flash frequency did not move with it.

This does not prove every step of hot-flash biology. It does show that inducing peripheral nitrate tolerance did not provide a clinical benefit in this population over 12 weeks. The cleanest interpretation is that cutaneous vasodilation is an output of a centrally triggered heat-loss response, rather than the primary switch that starts a hot flash.

Why the negative result still has limits

The trial took place at one academic center in northern California and enrolled 141 people, so it could miss a modest effect. Participant-directed dose titration may also have produced uneven exposure, and the study did not report blood nitrate measurements. Headache resolution supports tolerance, but selective withdrawal by people with persistent or severe headaches could partly explain the decline. If dropout, rather than pharmacologic tolerance, drove that pattern, the mechanistic inference becomes weaker. The week-5 P=0.05 result also did not survive to the primary week-12 endpoint and lacked multiple-comparison correction.

The leading target is central thermoregulation

The dominant model places the initiating problem in the hypothalamus. Estrogen withdrawal reduces suppression of KNDy neurons—named for kisspeptin, neurokinin B, and dynorphin—in the arcuate nucleus. Higher neurokinin B signaling through NK3 receptors appears to narrow the thermoneutral zone, so small temperature changes can trigger sweating and skin vasodilation.

Fezolinetant blocks NK3 receptors at this central node. In SKYLIGHT 1, a phase 3 double-blind randomized trial of fezolinetant 30 mg or 45 mg versus placebo in postmenopausal women with intact uteri and no active liver disease, both doses significantly reduced the frequency and severity of moderate-to-severe vasomotor symptoms at weeks 4 and 12. The 45 mg dose is the FDA-approved regimen.

The result was not confined to a single short trial. A pooled analysis of SKYLIGHT 1, 2, and 4, involving about 3,000 participants, found that reductions in symptom frequency and severity persisted through 52 weeks. A phase 3b trial in women unsuitable for hormone therapy also found fezolinetant effective and tolerable, extending the evidence to a population with higher comorbidity than the main trials.

For patients, a drug's average result matters less than whether improvement feels meaningful. Pooled SKYLIGHT 1 and 2 analyses have defined patient-anchored thresholds for meaningful symptom-frequency change, and a separate responder analysis examined which participants achieved substantial reductions. Another pooled analysis found that benefit emerged in the first weeks and included both daytime and nighttime symptoms.

Effective does not mean casual

Fezolinetant requires a prescription and liver monitoring. Rare but serious liver injury has been identified, and pre-existing liver disease is a contraindication. In the 52-week pooled safety analysis, common treatment-emergent events, including upper respiratory infection, headache, and COVID-19, occurred at similar rates in fezolinetant and placebo groups. The randomized program supports efficacy through one year; it does not answer questions about efficacy or safety beyond that period.

A 2024 systematic review and meta-analysis also pooled fezolinetant efficacy and safety data, but the abstract indexed under its PubMed record appears to contain unrelated maternal-sepsis content. The paper's title identifies it as a fezolinetant review, yet readers should verify claims against the full text before relying on that citation.

How does fezolinetant compare with other nonhormonal drugs?

A 2026 network meta-analysis ranked NK3/NK1 antagonists, including fezolinetant and elinzanetant, above SSRIs, SNRIs, and gabapentin for reducing moderate-to-severe vasomotor symptoms. SSRIs, SNRIs, and gabapentin still have evidence-supported roles, particularly when they fit a patient's other symptoms or medication constraints. Their smaller average reductions do not establish that they act on an inferior mechanism.

The ranking is indirect. Network meta-analysis compares separate trials with different eligibility criteria, placebo-response rates, and outcome methods; it is not a head-to-head randomized comparison. A matching-adjusted indirect comparison found broadly comparable efficacy for fezolinetant, which blocks NK3 selectively, and elinzanetant, which blocks both NK1 and NK3. That result can fit an NK3-dominant pathway if NK1 blockade adds little, but it also leaves room for a more distributed neurokinin system. Cross-trial matching cannot settle that question.

Wrist cooling is reasonable to test, but the evidence is preliminary

Cooling the glabrous skin of the wrist offers a different idea: sensory input from cool-sensitive thermoreceptors may influence the hypothalamus's perception of temperature. A small crossover pilot reported about a 46% reduction in severe hot-flash episodes with wrist cooling to 8.3°C, with no adverse events reported among postmenopausal women, prostate-cancer patients receiving androgen-deprivation therapy, and breast-cancer patients.

Those findings are encouraging, not established care. The study was single-site, small, difficult to blind, and potentially affected by manufacturer-adjacent interests. Its proposed mechanism remains speculative. Cooling the wrists is a low-burden personal experiment for many people, but it should not replace an evidence-based discussion of hormone therapy or prescription nonhormonal treatment.

The useful conclusion

The nitroglycerin trial did not identify a new treatment. It ruled out a specific strategy: creating peripheral vascular nitrate tolerance did not reduce hot flashes at 12 weeks, even though the expected headache pattern changed dramatically. Fezolinetant has the strongest nonhormonal evidence among the options discussed here, with multiple phase 3 trials and follow-up through 52 weeks, balanced against liver-safety monitoring. Wrist cooling may be worth trying as an adjunct, provided its pilot evidence is treated as pilot evidence.

Hot flashes involve blood vessels, but the best current evidence says the command to dilate them begins higher up—in the brain's thermoregulatory circuitry.