Substance P Antagonists (NK-1)

Substance P is an 11-amino-acid tachykinin of about 1348 Da, released from C-fibre terminals in the dorsal horn and from peripheral sensory endings. It was the textbook 'pain neurotransmitter' for two decades, and the NK-1 receptor became one of the most heavily invested targets in pharma history. Then the analgesia trials came back negative — repeatedly and across multiple companies and pain models. Substance P turns out to be more important for the affective and autonomic dimensions of noxious input than for the sensory-discriminative signal itself, and blocking it does not stop pain in humans. Where NK-1 blockade did work is the area postrema and nucleus tractus solitarius, where substance P mediates the delayed phase of chemotherapy-induced nausea and vomiting; aprepitant and its relatives are now standard antiemetic care. There is also a partly abandoned antidepressant story — MK-869 (aprepitant) showed antidepressant activity in an early trial that later phase 3 work did not replicate. The class remains one of the most instructive failures in analgesic development.

Targets: NK-1 (tachykinin NK1) receptor, Substance P / tachykinin signalling, Area postrema and nucleus tractus solitarius

• · Given as part of a triple regimen with a 5-HT3 antagonist and dexamethasone. Aprepitant is a moderate CYP3A4 inhibitor, so the dexamethasone dose is usually halved alongside it.
• · Fosaprepitant 150 mg IV or aprepitant injectable emulsion 130 mg IV as a substitute for the oral three-day course.
• · There is no NK-1 antagonist dose for pain. Multiple well-powered trials across neuropathic pain, dental pain, osteoarthritis and migraine were negative. This entry sits in the pain class for mechanistic and historical reasons, not because the drugs treat pain.

Work out your exact syringe units →

• very commonClinically significant CYP3A4 drug interactions— This is the real hazard of the class — it changes the exposure of a long list of co-administered drugs.
• commonFatigue and asthenia— Hard to separate from the chemotherapy itself.
• commonHiccups— A curiously characteristic effect of the class.
• commonConstipation or diarrhoea
• uncommonInfusion-site reactions and phlebitis— Specific to fosaprepitant.
• rareHypersensitivity and anaphylaxis

• Concurrent pimozide, cisapride or terfenadine-type drugs, where CYP3A4 inhibition can cause fatal QT prolongation.
• Hormonal contraception is unreliable during and for about a month after aprepitant — a genuine and frequently missed interaction.
• Warfarin requires INR monitoring for two weeks after each course; aprepitant induces CYP2C9 and INR falls.

• redundantcgrp-targeting-peptides — Both aim at trigeminal neuropeptide signalling; NK-1 blockade was tested in migraine and failed where CGRP blockade succeeded.
• cautionkpv — Both dampen neurogenic inflammation by different routes; overlapping but not studied together.

• · A full medication review for CYP3A4 substrates before every course — this is the highest-yield safety step with this class.
• · INR for two weeks after each course in anyone on warfarin.
• · Nausea and vomiting episodes recorded by day, so the delayed phase is actually assessed.

Approved prescription medicines in the US and EU for antiemetic use. Not controlled substances. No NK-1 antagonist is approved for any pain indication anywhere.

• Kramer et al. 1998, Science — distinct mechanism for antidepressant activity by blockade of central substance P receptors (trial)
• Emend (aprepitant) FDA prescribing information (label)
• Hill 2000, Trends in Pharmacological Sciences — NK1 (substance P) receptor antagonists: why are they not analgesic in humans? (review)

This entry earns its place in a pain class by being the most expensive negative result in the history of analgesic development, and understanding why it failed is more useful than understanding how it works. The premise was excellent. Substance P is released from C-fibre terminals in the dorsal horn onto NK-1 receptors on second-order neurons; it is released peripherally in neurogenic inflammation; NK-1 knockout mice show blunted responses to intense noxious stimuli; and the anatomy is textbook. Every major pharmaceutical company built an NK-1 antagonist. Then the analgesia trials came back negative — in neuropathic pain, in dental pain, in osteoarthritis, in migraine, across multiple companies and multiple compounds, with adequate central receptor occupancy demonstrated by PET. That last point is what makes the failure so clean: this was not a drug-delivery problem. The receptors were occupied and the pain did not stop. The retrospective explanation is that substance P and NK-1 signalling carry the affective, autonomic and emotional dimension of noxious input rather than the sensory-discriminative signal itself. NK-1-expressing lamina I neurons project to the parabrachial nucleus and amygdala — the 'this is awful' pathway — rather than to the somatosensory cortex, and blocking that in a rodent hot-plate assay looks like analgesia while blocking it in a human with neuropathic pain does not. Substance P also appears to matter most at intense, sustained stimulus levels rather than across the ordinary range, and it is one signal among many redundant ones at the same synapse. Where NK-1 blockade did work is instructive for the same reason: in the area postrema and nucleus tractus solitarius, substance P mediates the delayed phase of chemotherapy-induced vomiting, and there it is not redundant — blocking it works, decisively, and aprepitant became standard antiemetic care. The antidepressant story followed the same arc as the pain story: Kramer's 1998 Science paper showed MK-869 with genuine antidepressant activity, the finding was widely believed, and subsequent phase 3 work failed to replicate it. Two large clinical bets, both mechanistically reasonable, both wrong, and one useful drug out the other end.

The pain claim is what goes wrong, and it goes wrong before anyone takes anything. Substance P is described everywhere on the internet as 'the pain neurotransmitter', and it is entirely reasonable to conclude that blocking it should relieve pain. It does not, and the evidence against is not weak or preliminary — it is multiple well-powered trials, across several pain types, from several companies, with PET-confirmed central receptor occupancy. Anyone selling an NK-1 antagonist or a substance P blocker as an analgesic is selling something that has been tested more thoroughly than most approved drugs and has failed. The second thing that goes wrong is the interaction burden, which is where this class actually hurts people. The warfarin INR drop happens after the course has ended, when nobody is watching for it. The contraceptive failure window extends about a month past the last dose. The dexamethasone dose adjustment gets missed when the antiemetic is prescribed by a different service from the steroid. And the QT-prolonging combinations are rare but lethal. Third, adherence: the three-day oral course exists to cover the delayed phase, and patients who feel well on day 2 stop taking it, which removes precisely the coverage the drug was added for. The single-dose intravenous fosaprepitant regimen exists largely to solve this. And fourth, a smaller point: hiccups are a curiously characteristic effect of this class, harmless but disconcerting, and worth knowing about so they are not investigated as something else.

1. 125 mgDay 1 of each chemotherapy cycle — Aprepitant 125 mg orally, one hour before chemotherapy. Given as part of a triple regimen with a 5-HT3 antagonist and dexamethasone — never alone.
2. 80 mgDays 2-3 of each cycle — Aprepitant 80 mg each morning. This is the part that covers the delayed phase, and it is the part that gets skipped when patients feel fine on day 2.
3. 150 mgSingle-dose intravenous alternative, day 1 — Fosaprepitant 150 mg IV infused 30 minutes before chemotherapy, or aprepitant injectable emulsion 130 mg IV, replacing the entire three-day oral course. Convenient, and it removes the day-2 adherence problem entirely.
4. —Pain indications — There is no dose. Multiple well-powered trials across neuropathic pain, dental pain, osteoarthritis and migraine were negative, with confirmed central receptor occupancy. This is not an under-dosing story and no dose exists to escalate to.

• NK-1 (tachykinin NK1) receptor — a class A GPCR — Aprepitant binds NK-1 with sub-nanomolar to low-nanomolar affinity and high selectivity over NK-2 and NK-3. Central receptor occupancy at clinical doses has been confirmed by PET, which is what makes the analgesic failure so informative — the drug reached the target and the target was the wrong one for pain.

  Competitive antagonism. Prevents substance P from activating Gq/G11 coupling with downstream phospholipase C, IP3 and calcium mobilisation. Clinically: blockade of the delayed phase of chemotherapy-induced nausea and vomiting, and nothing useful for pain.

• Area postrema and nucleus tractus solitarius NK-1 receptors — The anatomical site where the class actually works.

  Suppression of the delayed emetic response. This is the only indication in which NK-1 antagonism has ever produced a robust clinical benefit, and it is genuinely robust — aprepitant added to a 5-HT3 antagonist and dexamethasone is standard care for highly emetogenic chemotherapy.

• Dorsal horn NK-1-expressing lamina I projection neurons — Fully accessible; aprepitant crosses the blood-brain barrier.

  Blocked, with no analgesic consequence in humans. These neurons project to the parabrachial nucleus and amygdala rather than to somatosensory cortex, which is the leading explanation for why blocking them changes the affective dimension of pain in rodent assays and does nothing measurable in human pain trials.

• CYP3A4 — inhibition, and CYP2C9 — induction — Not a therapeutic target. This is the drug's most clinically consequential interaction with the body.

  Moderate CYP3A4 inhibition raises exposure of co-administered CYP3A4 substrates, which is why the dexamethasone dose is halved in the standard antiemetic regimen. CYP2C9 induction lowers warfarin exposure and drops the INR for roughly two weeks after each course. Hormonal contraception becomes unreliable during and for about a month afterwards. These are the facts about this class most likely to actually harm someone.

• Kramer et al. — MK-869 (aprepitant) for major depressive disorder Randomised, double-blind, placebo- and active-controlled · 6 weeks · 1998

  Substance P receptor antagonism produced antidepressant activity comparable to paroxetine and superior to placebo on the Hamilton depression scale. A landmark result at the time, published in Science, that subsequent phase 3 programmes failed to replicate — which is why aprepitant is an antiemetic and not an antidepressant.

Antiemetic protection begins within about an hour of the oral day-1 dose, which is why it is given an hour before chemotherapy rather than with it. The clinically important coverage is the delayed phase — days 2 to 5 — which is exactly the window 5-HT3 antagonists do not reach, and which the days 2 and 3 doses exist to cover. Rolapitant's roughly seven-day half-life covers the whole delayed phase from a single dose, and fosaprepitant achieves the same coverage from one infusion. For pain there is no onset timeline because there is no effect: the negative trials ran for adequate durations at doses producing confirmed central receptor occupancy.

The stacking story for this class is almost entirely about cytochrome P450, and it runs in both directions. Aprepitant is a CYP3A4 substrate, so strong CYP3A4 inhibitors (ketoconazole, clarithromycin, ritonavir) raise its exposure substantially and strong inducers (rifampicin, carbamazepine, phenytoin) drop it. It is simultaneously a moderate CYP3A4 inhibitor, so it raises exposure of anything else cleared that way — which is why dexamethasone is routinely halved when given in the same antiemetic regimen, and why pimozide, cisapride and terfenadine-type drugs are absolutely contraindicated on QT grounds. And it induces CYP2C9, which is why the INR falls in warfarin patients for about two weeks after each course, and why hormonal contraception is unreliable during and for roughly a month afterwards. On the antiemetic side the intended stack is well established and genuinely additive: NK-1 antagonist plus a 5-HT3 antagonist plus dexamethasone, with olanzapine added as a fourth agent for highly emetogenic regimens. Each covers a different receptor and a different phase, and the combination is much better than any component. Within this site's class, stacking an NK-1 antagonist with a CGRP-targeting agent for migraine makes no sense — NK-1 blockade was tested in migraine and failed where CGRP blockade succeeded, so you would be adding an interaction-heavy drug for no benefit. Pairing with KPV or other anti-inflammatory peptides overlaps mechanistically on neurogenic inflammation but has never been studied.

The most useful comparison is with the CGRP class two entries above, because the two are near-perfect mirror images. Both targeted a neuropeptide released from sensory nerve terminals. Both had strong preclinical support and enormous pharmaceutical investment. CGRP blockade worked in humans and produced an entire generation of approved migraine drugs; NK-1 blockade did not work in humans and produced an antiemetic. The difference was not effort, delivery or receptor occupancy — it was that CGRP turned out to be causally necessary for the clinical phenomenon and substance P turned out to be one redundant signal among several. That is a lesson about target validation that applies to every speculative mechanism on this site: human provocation data, of the kind CGRP had and substance P never did, is worth more than any amount of rodent pharmacology. Within its actual indication, aprepitant compares well — adding an NK-1 antagonist to a 5-HT3 antagonist and dexamethasone meaningfully improves control of delayed chemotherapy-induced vomiting, and that is a real quality-of-life benefit for people going through something genuinely awful. Just do not let the analgesic hypothesis borrow the antiemetic evidence. They are separate claims and only one of them survived.

Not priced here, because the unit that matters is a chemotherapy cycle rather than a month and because generic aprepitant is now widely available at a fraction of the original brand price. In practice this is dispensed as part of an oncology supportive-care regimen and billed accordingly. There is no monthly self-administration pattern to cost out, and there is certainly no pain indication to cost out.

• ?This entry covers a drug class rather than a single agent, and the pharmacokinetic fields are aprepitant's. Rolapitant differs radically — a half-life of roughly seven days versus 9-13 hours — and netupitant and fosnetupitant differ again. Only the aprepitant label was retrieved in this session.
• ?The participant count for the Kramer 1998 MK-869 depression trial is left null; I confirmed the publication but did not resolve the enrolment figure.
• ?The claim that NK-1-expressing lamina I neurons project preferentially to parabrachial nucleus and amygdala rather than somatosensory cortex, and that this explains the rodent-to-human analgesic gap, is the leading explanation in the field rather than a settled fact.
• ?Time to steady state is left null — aprepitant is given as a three-day course per cycle rather than continuously, so steady state is not a meaningful concept for its clinical use.
• ?Aprepitant's NK-1 binding affinity is described qualitatively; I did not resolve a primary source for a specific Ki in this session.
• ?The tmax of 3.5 hours is the midpoint of the label's stated 3-4 hour range, and the 62% bioavailability is the midpoint of the stated 60-65% range.

• Distinct mechanism for antidepressant activity by blockade of central substance P receptors Kramer MS, Cutler N, Feighner J, et al., Science, 1998 · PMID 9733503

  The paper that launched a thousand NK-1 programmes. Read it alongside the knowledge that the antidepressant finding did not replicate in phase 3 — it is a case study in how a clean early result in a high-profile journal can direct a decade of investment.

• NK1 (substance P) receptor antagonists — why are they not analgesic in humans? Hill R, Trends in Pharmacological Sciences, 2000 · PMID 10871891

  The single most useful paper on this page for anyone interested in why analgesic drug development fails. Written from inside the industry, at the moment the failure became undeniable, and it explains the rodent-to-human gap better than any retrospective.

• Aprepitant capsules — US prescribing information Torrent Pharmaceuticals Limited, DailyMed, US National Library of Medicine

  Source of the pharmacokinetics quoted here: absolute oral bioavailability approximately 60-65%, tmax 3-4 hours, apparent steady-state volume of distribution approximately 70 L, plasma protein binding greater than 95%, terminal half-life 9-13 hours, CYP3A4-dominant metabolism with minor CYP1A2 and CYP2C19, no renal excretion of intact drug, and explicit confirmation that aprepitant crosses the blood-brain barrier in humans.