Area postrema

In brief

A patch of brain left outside the barrier, on purpose

Almost everywhere in the brain, the capillaries are sealed. Their endothelial cells, the flat cells that line a blood vessel, are stitched together by tight junctions, so that what circulates in the blood does not simply diffuse into the nervous tissue. That seal is the blood-brain barrier, and it is the reason most drugs have no central effect at all. The area postrema is one of the handful of places where the seal is deliberately absent. It is a small, richly vascularized structure on the dorsal surface of the medulla oblongata, the lowest segment of the brainstem, sitting at the caudal end of the fourth ventricle where that cavity narrows into the central canal of the spinal cord ^1,2.

It belongs to a group called the circumventricular organs, which also includes the median eminence at the base of the hypothalamus, where the brain releases hormones into the blood supplying the pituitary; the vascular organ of the lamina terminalis; the subfornical organ; the subcommissural organ, which unlike the others is not a sensory organ of this kind; the pineal gland; and the neurohypophysis, or posterior pituitary ^1,33. Duvernoy and Risold's comparative atlas, built from histological preparations of twelve mammal and five bird species, makes the defining point: the capillaries of these organs have walls devoid of a blood-brain barrier, which permits direct exchange between blood and nervous tissue ¹. The area postrema is the only one of them in the hindbrain, and the only one positioned to act on the machinery of vomiting.

The leak is more interesting than the word suggests, because it is graded rather than wide open. In adult mice, across the sensory circumventricular organs including the area postrema, fluorescein isothiocyanate, a tracer of molecular weight 389, and a dextran of molecular weight 3,000 both diffused from the vessels into the tissue, while a dextran of molecular weight 10,000 remained caught in the perivascular space between the inner and outer basement membranes ⁴. Permeability to the smallest tracer was higher in the central part of these organs than at their edges; the tight-junction proteins occludin, claudin-5 and zonula occludens-1 were undetectable centrally but present distally; and neither small tracer passed beyond a dense network of astrocyte and tanycyte processes ⁴. In other words, the area postrema does not spill into the rest of the brain. It is sampled from the blood on one side and walled off from the brain on the other.

Tanycytes are the specialized glial cells that build that inner wall. They are elongated cells whose long processes contact the fenestrated capillaries at one end and reach toward the ventricle at the other, and in the sensory circumventricular organs they behave partly as neural stem cells, carrying the osmotic and mechanical sensor TRPV1, the lipopolysaccharide receptor TLR4 that detects a component of bacterial cell walls, and the sodium-sensing channel Nax ³³. A 2025 study in mice showed that this arrangement is not fixed: after exposure to monosodium glutamate, a neurotoxin that cannot cross the blood-brain barrier, fenestrated capillaries and vascular permeability increased in the area postrema while decreasing in the median eminence, with tanycytic expression of two growth factors that drive new vessel formation, pleiotrophin and vascular endothelial growth factor A, tracking the change, and that this plasticity was abolished in aged animals ³⁴. That is mouse anatomy, not human physiology, but it reframes the opening in the barrier as something the tissue adjusts rather than something it simply has.

Functionally, the area postrema never works alone. Together with the nucleus of the solitary tract, the brainstem's main visceral sensory nucleus, and the dorsal motor nucleus of the vagus, it forms the dorsal vagal complex, which is where vagal afferent fibers from the gut terminate ^2,35. Blood-borne information enters at the area postrema; nerve-borne information from the stomach and intestine enters a millimeter away; and the two converge. A 2008 review called the structure not simply a conduit but the initial site of integration for these signals as they enter the central nervous system ³.

Borison and Wang, and the whole history including its reversals

Before 1949, the vomiting center was said to lie in the dorsal vagal nuclei of the medulla, and it was unclear whether there were two centers, one for drugs acting directly and one for reflexes ⁷. Herbert Borison and S. C. Wang took the question apart with stereotaxic electrical stimulation, mapping where in the brainstem a current produced vomiting: first in the cat in 1949 ⁸, and then in the paper titled, with no hedging, The vomiting center; a critical experimental analysis, which placed the center in the reticular formation at a measurable distance from the dorsal vagal nuclei ^5,7. They then built the lesion method that made the claim testable, implanting radon in the medulla to produce discrete, measurable lesions ⁹, and used it to destroy the putative center in the dog medulla oblongata ¹⁰. None of those four records carries an abstract in PubMed, so each is cited for its date, authors, species, method and title; the content of the work is taken from Borison's own retrospective account of 1984 ⁷.

Their second and more durable idea was the separation of functions. They proposed that the center that coordinates the act of vomiting is not itself chemically sensitive, and that a separate emetic chemoreceptor trigger zone detects emetic agents in the blood and reports to it ⁷. Borison and Wang set the whole scheme out in Physiology and pharmacology of vomiting in Pharmacological Reviews in 1953, the paper the nickname is usually credited to ⁶. That record also carries no abstract in PubMed; the content of the concept is taken here from Borison's own retrospective account in 1984 and from the 1994 review that reassessed it ^2,7.

The trigger zone was soon identified with the area postrema, and the evidence that followed was of three kinds. Lesion experiments: destroying the area postrema prevented vomiting in response to most, though not all, emetic drugs ². That separation was drawn drug by drug, and the primary papers are Wang and Borison's own. In 1952 they set out a new concept of the organization of the central emetic mechanism from the sites of action of three agents, apomorphine, copper sulfate and the cardiac glycosides ⁴⁰; the year before they had traced copper sulfate emesis to afferent nerves from the gastrointestinal tract rather than to the trigger zone ⁴¹, and had located the central emetic action of the cardiac glycosides ⁴². None of those three records carries an abstract in PubMed either, so they are cited for their subject matter and dates. The pattern was the point: a drug in the blood acted through the trigger zone, an irritant in the gut acted through the vagus, and the two converged downstream. Electrophysiology: in 1983, Carpenter, Briggs and Strominger recorded from 122 area postrema neurons in anesthetized dogs and tested seventeen transmitters and peptides, finding a brief excitation with glutamate and a slow excitation with dopamine, apomorphine, histamine, norepinephrine, serotonin, substance P, angiotensin II, neurotensin, leucine enkephalin, vasoactive intestinal polypeptide, thyrotropin-releasing hormone, gastrin and vasopressin, and no response to acetylcholine, somatostatin or cholecystokinin ¹¹. The fit was close: every excitatory substance except serotonin and norepinephrine was emetic, and none of the three inactive ones was ¹¹. And metabolic and activity mapping: studies using 2-deoxyglucose uptake and expression of the immediate-early gene c-fos showed the area postrema lighting up after systemic emetic drugs ².

Three reversals belong in the record. The first is about motion. Wang and Chinn published a paper in 1954 titled Experimental motion sickness in dogs; functional importance of chemoceptive emetic trigger zone, whose title asserts that the trigger zone matters for motion sickness ⁴³; the record carries no abstract, so nothing beyond the title can be verified from it. The experiment that dismantled the claim was Borison's own. In 1986 Borison and Borison exposed cats to vertical sinusoidal oscillation on a spring-suspended platform and then destroyed the area postrema in two groups of five animals, one group never previously tested and one known to be susceptible. After the operation motion sickness still occurred in all five of the previously untested cats and in four of the five susceptible ones, and the operated animals were significantly more susceptible than normal cats; the single animal that lost its susceptibility had damage extending beyond the organ. Their conclusion was that the area postrema is not an essential link in the reflex arc of motion-induced vomiting ¹⁹. Eight years later Miller and Leslie's review stated the settled position: the area postrema is not essential for vomiting induced by motion or by activation of vagal afferents ². Motion sickness goes through the vestibular system, chemically triggered vomiting goes through the area postrema, and the two routes converge only downstream.

The second reversal is the vomiting center itself, and it is the one this page exists to make plain. The entity Borison and Wang localized in 1950 is not now thought to exist as a discrete structure. In 2001 Hornby wrote that it is critical to realize that there is not an isolated vomiting center, but rather groups of loosely organized neurons throughout the medulla that may be activated in sequence by a central pattern generator, with the nucleus of the solitary tract organized viscerotopically into subnuclei for swallowing, gastric sensation, laryngeal and pharyngeal sensation, baroreceptor function and respiration ¹⁶. The chemoreceptor trigger zone survived the century; its partner did not.

The third is about the scope of the organ. Borison's own 1984 summary conceded that it remained unanswered whether the trigger zone constitutes part or all of the area postrema, and that different chemosensory functions, including the control of defecation and of other autonomic responses, may be represented in different regions within it ⁷. He also noted that species unable to vomit, such as rodents, still show postrema-mediated effects such as radiation-induced gastric stasis and drug-induced conditioned taste aversion, and that in sheep, removal of the area postrema prevents digitalis from arresting rumination ⁷. The structure was never only about vomiting.

The correction: there is no vomiting center

The phrase vomiting center is still taught, still printed in textbooks and still used in clinical shorthand, and it describes something that is not there. What exists is a sequence. Chemosensitive receptors in the area postrema detect emetic agents in the blood and relay that information to the adjacent nucleus of the solitary tract; vagal afferents carrying information about the contents and tone of the stomach and intestine terminate in the same nucleus, in defined subnuclei; and from there neurons project to a central pattern generator that coordinates the sequence of behaviors that make up emesis, and directly to populations of neurons in the ventral medulla and hypothalamus ¹⁶.

A central pattern generator is a circuit that produces a fixed sequence of muscle actions once it is started, the way the circuits for walking and breathing do. Vomiting is such a sequence: a deep breath, closure of the airway, relaxation of the upper stomach, a coordinated contraction of diaphragm and abdominal wall, and retrograde flow. The point of Hornby's 2001 statement is that no single anatomical place holds that program. It is written across loosely organized groups of neurons throughout the medulla, which may be activated in sequence ¹⁶. Destroying the area postrema removes the response to most emetic drugs but leaves vomiting from motion and from vagal stimulation intact, which is what a distributed program with several inputs predicts ^2,19.

That is not the same as saying nothing is localized. Working in decerebrate, paralyzed dogs, Fukuda and Koga stimulated and partially cut the medulla and traced the route: emetic vagal afferents enter the rostral medulla, descend through the solitary complex to the area subpostrema, and project through a longitudinal reticular column to a pattern generator they placed in the Bötzinger complex, a group of neurons near the retrofacial nucleus already known for its part in breathing ⁴⁴. Emetic responses survived a cut of the solitary complex about 1 millimeter rostral to the obex but not one about 3.5 millimeters rostral, and disappeared after transection at the caudal end of the retrofacial nucleus or partial cutting of the caudal Bötzinger complex ⁴⁴. In the solitary nucleus at the level of the area subpostrema, all 40 non-respiratory neurons tested responded to stimulation of the vagus, and their firing rose after apomorphine in 6 of 10 and after glutamate in 14 of 14 applications into the fourth ventricle; cooling that part of the nucleus impaired retching driven by the vagus nerve on the same side ⁴⁵. In the Bötzinger complex itself, 78 non-respiratory neurons responded to the vagus at a mean latency of 387 milliseconds; half fired in high-frequency bursts synchronous with each retch and a third between retches, many of them building up gradually beforehand, during the period that corresponds to the prodrome of vomiting; and cooling the subpostrema part of the solitary nucleus abolished both the retching and those firings ⁴⁶. So the sequence has an organizing site; what it does not have is a single chemically sensitive center of the kind named in 1950. These are dog experiments, and the correspondence to the human medulla is assumed rather than measured.

The nucleus of the solitary tract is the structure most often miscast as the center, and the reason is that it is genuinely the convergence point. It is organized viscerotopically, meaning different body regions map to different parts of it: a subnucleus centralis for swallowing, a subnucleus gelatinosus for gastric sensation, intermediate and interstitial regions for laryngeal and pharyngeal sensation, a medial region for baroreceptor input from blood pressure sensors, and a ventrolateral region for respiration ¹⁶. The human receptor map is consistent with this: autoradiography of human medulla found the highest density of 5-HT3 receptors in the substantia gelatinosa subnucleus of the nucleus of the solitary tract, higher than in the area postrema itself ¹². A drug that treats nausea is often acting next door to the trigger zone rather than in it.

Keeping the distinction straight has practical consequences. It explains why 5-HT3 antagonists work substantially through vagal afferent endings in the gut rather than only centrally ^16,17; why neurokinin-1 antagonists add protection on day 1 and especially on days 2 to 5, the delayed phase that 5-HT3 blockers control poorly ^16,18; why the cannabinoid delta-9-tetrahydrocannabinol acts on CB1 receptors in the dorsal vagal complex ¹⁶; and why no single antiemetic has ever abolished vomiting. There is no master switch, because there is no master.

What it carries, and the drugs that find it

The reason a single small structure is pharmacologically important out of all proportion to its size is that it expresses an unusual density and variety of receptors for circulating messengers, and that drugs given by any route reach those receptors directly. The canine recording study remains the most systematic single survey: of seventeen transmitters and peptides applied by microelectrode to 122 neurons, fourteen produced excitation, glutamate in a brief high-frequency burst that was used to find the otherwise silent cells and thirteen others in slow responses alike whatever the agent, and the correspondence with emetic activity was close enough that the authors called it strong support for the central role of the area postrema in emesis ¹¹. The excitatory list ran: dopamine and apomorphine, histamine, norepinephrine, serotonin, substance P, angiotensin II, neurotensin, leucine enkephalin, vasoactive intestinal polypeptide, thyrotropin-releasing hormone, gastrin and vasopressin; acetylcholine, somatostatin and cholecystokinin did nothing, and norepinephrine also inhibited 6 of 15 cells tested and histamine 3 of 45 ¹¹.

Dopamine came first historically. Apomorphine, a dopamine agonist, is the classic experimental emetic, and the finding that destroying the area postrema abolished the vomiting caused by blood-borne emetic drugs is the observation that anchored the chemoreceptor trigger zone concept, with apomorphine the agent Wang and Borison worked with in drawing it ^2,7,40. The clinical mirror image is the dopamine D2 antagonists used as antiemetics, metoclopramide and domperidone; domperidone is generally described as acting at this structure while crossing the blood-brain barrier poorly, which would spare the rest of the brain, though no study cited on this page measured where in the brain it acts. In a reported case of area postrema syndrome caused by a medullary stroke, intravenous metoclopramide with ondansetron failed and the vomiting resolved only after oral domperidone and itopride were added to them ³⁰.

Serotonin came second and changed oncology. Cubeddu and colleagues reported in 1990 that ondansetron, then called GR 38032F, controlled cisplatin-induced nausea and vomiting in a placebo-controlled trial of 28 patients: median time to first emesis 11.6 hours against 2.8 hours, median episodes in 24 hours 1.5 against 5.5, and regurgitations or dry heaves per episode 1.17 against 3.2, all three comparisons at p less than 0.001, with no patient on ondansetron requiring rescue antiemetics against 12 of 14 on placebo ¹⁷. Urinary excretion of 5-hydroxyindoleacetic acid, serotonin's main metabolite, rose in every patient after cisplatin in parallel with the vomiting, which led the authors to suggest that cisplatin releases serotonin from enterochromaffin cells in the gut lining ¹⁷. Human autoradiography then showed where the receptors are: densest in the substantia gelatinosa subnucleus of the nucleus of the solitary tract, then the dorsal subnucleus, then the area postrema ¹².

A 2026 mouse experiment marks the limit of the serotonin account. Mice lacking the 5-HT3 receptor altogether showed cisplatin-induced malaise to the same extent as normal mice: the same fall in food intake and body weight, the same drop in spontaneous activity, the same pica and the same learned avoidance of a tasted flavor ⁵⁷. In both groups the number of cells expressing c-Fos, and of cells expressing both c-Fos and GFRAL, rose in the area postrema and the nucleus of the solitary tract, with plasma GDF15 raised for at least a day after treatment ⁵⁷. The authors conclude that 5-HT3 signaling is not essential for the malaise, and that GDF15 acting through GFRAL may carry it by a separate route ⁵⁷. Mice cannot vomit, so this concerns nausea-like behavior rather than emesis, which is also the clinical pattern: the 5-HT3 blockers control chemotherapy vomiting far better than they control chemotherapy nausea ¹⁷.

Substance P came third. The neurokinin-1 receptor antagonist aprepitant, added to ondansetron and dexamethasone in patients given cisplatin at 70 mg per square meter or more, raised complete response across days 1 to 5 from 52.3 percent of 260 patients to 72.7 percent of 260, p less than 0.001, with the benefit significant on day 1 and especially on days 2 to 5 ¹⁸. Hornby located the neurokinin-1 site of action in the dorsal vagal complex, and suggested part of the effect comes from blocking the receptor on vagal motor neurons, preventing the relaxation of the upper stomach that is a necessary prelude to vomiting ¹⁶.

Opioids belong on the list and are the least glamorous entry. Enkephalin excited area postrema neurons in the canine survey ¹¹, and opioid-induced nausea is a common side effect of a needed analgesic. It is worth saying plainly that the evidence base here is thinner than for the other three classes: there is no landmark trial establishing an area postrema mechanism for opioid emesis in people in the way there is for cisplatin and serotonin.

The newest entries are metabolic. The GLP-1 receptor, for the gut hormone glucagon-like peptide-1, and GFRAL, for GDF15, are both expressed on area postrema neurons ^13,15,35. The GLP-1 drugs are now in wide use, and the GDF15 drugs are in trials ^22,58. In mice, at least, their nausea is not an incidental side effect of an unrelated mechanism but comes from this structure ¹⁵.

• MRI of the dorsal medullaThin-section magnetic resonance imaging through the floor of the fourth ventricle is the only way to see the area postrema in a living person. In neuromyelitis optica it shows a linear lesion of the dorsal medulla involving the pericanal region, the area postrema and the nucleus of the solitary tract ²³.To find a structural or inflammatory cause for unexplained, unremitting nausea, vomiting or hiccupsGood enough to shorten the diagnosis: a visible dorsal medullary lesion allows area postrema syndrome to be diagnosed before the 48-hour symptom threshold is met ²¹The structure is small enough that ordinary brain imaging can read as normal; a 2026 review reports that patients are routed through gastroenterology first and immunotherapy is delayed by weeks to months ⁵⁵
• AQP4-IgG serology (cell-based assay)A blood test for the antibody against aquaporin-4, a water channel on astrocytes, which is the cause of most neuromyelitis optica spectrum disorder; a positive result turns an unexplained vomiting illness into a treatable autoimmune disease ^20,21.In anyone with unexplained intractable nausea, vomiting or hiccups, especially with a dorsal medullary lesionPrognostically important when positive: in two patients with similar presentations after COVID-19, the antibody-negative one recovered and was relapse-free at 12 months while the antibody-positive one relapsed fatally six months later ⁷⁴Seronegative cases exist and still have the syndrome; two other antibody diseases can produce it, myelin oligodendrocyte glycoprotein antibody disease, directed at a protein of the myelin sheath, and glial fibrillary acidic protein astrocytopathy, directed at a structural protein of astrocytes, and in one reported case all three antibodies were negative at first, with serum GFAP-IgG turning weakly positive only after treatment ^55,70
• Area postrema syndrome criteria and the PUQE severity scaleDiagnostic criteria and a severity grade built from 100 patients and 157 episodes, repurposing the Pregnancy-Unique Quantification of Emesis and Nausea score, with hiccup severity graded by duration ²¹.To recognize the syndrome and to measure attack severity as an outcome in clinical trialsValidated prospectively in a multicenter cohort by the group that proposed it ²¹It is a clinical scale, not a test of the tissue; it cannot separate neuromyelitis optica from MOG antibody disease, GFAP astrocytopathy or a stroke in the same place ^30,55
• Single-nucleus RNA sequencing of the area postremaA research method that reads which genes are switched on in each individual cell nucleus, used since 2020 to build cell atlases of this structure in mice and rats ^35,36,64.To find which cell types carry which receptors, and so which cells a drug acts onHigh resolution: a rat study profiled about 35,000 nuclei and resolved 19 cell subtypes, including a single population of excitatory neurons positive for both Gfral and Ret ⁶⁴The deep atlases are rodent, and a 2026 cross-species atlas adds macaque brainstem with human conservation inferred from expression; there is still no human area postrema atlas of comparable depth, so receptor maps in people rest on a 1994 autoradiography study and the human expression data in the 2017 receptor papers ^12,13,65,66
• Emesis models in species that can vomitRats and mice cannot vomit, so the behavior has to be studied in dogs, ferrets, monkeys or the musk shrew Suncus murinus, with pica and conditioned food avoidance used as stand-ins in rodents ^7,25.To tell a drug that causes true emesis from one that merely suppresses eatingIt worked in both directions: GDF15 caused emesis in musk shrews and nausea-like behavior in mice and rats ²⁵, and in nonhuman primates given platinum chemotherapy the GDF-15-neutralizing antibody mAB1 attenuated anorexia and emesis ⁴⁸Species differ, and a result in a shrew or a ferret is not a result in a person; Borison proposed that postrema-mediated behaviors in species that cannot vomit, such as drug-induced conditioned taste aversion in rodents, might serve as assays for antinauseant drugs ⁷

GDF15 and GFRAL: a hormone of sickness and the drugs built on it

GDF15, also called MIC-1, is a distant member of the transforming growth factor beta family, released into the blood by tissue that is stressed, damaged or malignant. It was first described in 1997, in activated macrophages, as macrophage inhibitory cytokine 1, a protein that limited the macrophage's own production of tumor necrosis factor alpha ⁵⁹. For years afterward it was known to rise with cancer, heart failure, kidney failure and weight loss, and to suppress appetite and body weight when given to animals, but its receptor was unknown, so the mechanism was guesswork ^13,14.

In August and September 2017, four groups published the answer independently and almost simultaneously: GDF15 binds GDNF family receptor alpha-like, GFRAL, with high affinity, and GFRAL must pair with the coreceptor RET to signal ^13,14,26,27. The striking finding was anatomical. GFRAL messenger RNA was highly expressed in the area postrema of mouse, rat and monkey, and in mouse and human in the area postrema and the nucleus of the solitary tract, and not in peripheral tissues at all ^13,14,26,27. Hsu and colleagues described the receptor as brainstem-restricted ²⁷. Mice without GFRAL were refractory to GDF15: their food intake, body weight and glucose measures did not respond, and blocking the GDF15-GFRAL interaction with a monoclonal antibody prevented GDF15's metabolic effects in rats ¹⁴. Mice lacking the receptor were hyperphagic under stress and resistant to chemotherapy-induced anorexia and weight loss ²⁷. All four papers came from pharmaceutical companies: Janssen, Lilly, Novo Nordisk, and NGM Biopharmaceuticals with Merck ^13,14,26,27. That is a commercial convergence on one target, and it should be named.

The next question was whether the appetite suppression was really nausea. In 2020 Borner and colleagues answered it across three species: GDF15 caused frank emesis in Suncus murinus, the musk shrew, and in mice and in lean and obese rats it produced anorexia and pica, the eating of non-nutritive material that rodents do when they feel sick ²⁵. Their conclusion was that GDF15 triggers anorexia through the induction of nausea and/or by engaging emetic neurocircuitry ²⁵. That sentence, with its hedge intact, is what makes the whole pathway coherent: GDF15 looks less like an appetite hormone that happens to sit in the vomiting structure than like a sickness signal whose effect on eating runs, at least in part, through making the animal feel sick.

That logic cuts both ways, and the therapeutic effort has run in both directions. Blocking GDF15 should relieve the wasting and nausea of advanced illness. In 2020, Pfizer researchers showed that circulating GDF-15 is higher in patients receiving platinum chemotherapy and positively associated with weight loss in colorectal cancer (NCT00609622), that platinum-induced anorexia and weight loss were attenuated in GDF-15 knockout mice, and that the antibody mAB1 attenuated anorexia and emesis in nonhuman primates ⁴⁸. Pfizer's clinical antibody against GDF-15, ponsegromab, followed from that preclinical work ^48,52.

The phase 2 trial, PROACC-1, randomized 187 patients with cancer cachexia and serum GDF-15 of at least 1500 picograms per milliliter to ponsegromab at 100, 200 or 400 mg or placebo, given under the skin every four weeks for three doses ^22,52. Forty percent had non-small-cell lung cancer, 32 percent pancreatic cancer and 29 percent colorectal cancer ²². At 12 weeks, median weight gain over placebo was 1.22 kg (95 percent credible interval 0.37 to 2.25) at 100 mg, 1.92 kg (0.92 to 2.97) at 200 mg and 2.81 kg (1.55 to 4.08) at 400 mg, with improvements in appetite, cachexia symptoms and physical activity at the top dose; adverse events of any cause occurred in 70 percent on drug and 80 percent on placebo ²². The trial was funded by Pfizer and its first and several other authors are Pfizer employees ^22,52. It is a phase 2 trial with weight as the primary endpoint, not survival, and a 2026 review of pancreatic cancer cachexia describes it as a potential breakthrough while noting that no drug is approved for cancer cachexia in the United States or Europe ⁶⁰. The honest reading is that the mechanism is confirmed and the clinical value is not yet established.

The other direction is obesity, where GDF15 agonism would suppress appetite and the same nausea becomes the obstacle rather than the target. The 2020 shrew work implies that any GDF15-based weight drug carries emesis in its mechanism ²⁵, and the one human trial of the idea bears part of that out while complicating it. LY3463251, a long-acting agonist at GFRAL and RET, reduced food intake and body weight in rodents and in obese nonhuman primates without signs of malaise or emesis; in a first-in-human single-dose study it caused dose-dependent nausea and emesis in a subset of healthy participants; and in a 12-week multiple ascending dose study in overweight and obese participants it significantly reduced food intake and appetite scores but produced only modest weight loss, which the authors report was independent of nausea and emesis (NCT03764774) ⁵⁰. Their own conclusion is that activating this system does shift energy balance in people but that the weight reduction was surprisingly modest, which suggests challenges in leveraging the pathway for weight loss ⁵⁰. All the authors are Eli Lilly employees. That is the honest reading: the receptor works in humans, and the drug did not do much.

Three things have happened on the blocking side since the phase 2 trial. The program moved on: a 2025 review of investigational cachexia drugs places ponsegromab in phase 3 trials and names two further agents against the same axis, AV-380 and NGM120, in development, with no drug yet approved for cancer cachexia by regulators in the United States or Europe ⁵⁸. The approach was pushed further in animals: in 2026, silencing Gfral in the brain with an antisense oligonucleotide, a short synthetic nucleic acid that switches off a chosen gene, blunted GFRAL expression and the c-Fos activation of area postrema and solitary nucleus neurons in mice with colon and pancreatic tumors, restored body weight, muscle and fat, gave better grip strength than a systemic antibody against GDF15, and extended survival ⁶¹. And the premise was questioned: a 2026 commentary argues that GDF15 is how the brain is told the body is under energetic stress, so suppressing the signal may relieve the symptoms while leaving the underlying problem, and perhaps long-term survival, worse, a concern that a 12-week weight endpoint cannot address ⁶².

Why the GLP-1 drugs make people sick, and where

Nausea is the commonest side effect of the GLP-1 receptor agonists, and gastrointestinal effects are a leading reason for stopping them in trials. In STEP 1, the trial of once-weekly semaglutide at 2.4 mg in 1,961 adults with overweight or obesity, mean body weight fell 14.9 percent against 2.4 percent with placebo, an estimated treatment difference of minus 12.4 percentage points (95 percent confidence interval minus 13.4 to minus 11.5); nausea and diarrhea were the commonest adverse events, described as typically transient and mild to moderate, and 4.5 percent of the semaglutide group discontinued because of gastrointestinal events against 0.8 percent on placebo ²⁴. The trial was funded by Novo Nordisk ²⁴.

Where that nausea comes from was uncertain for a long time, because GLP-1 receptors sit both on the vagal afferents in the gut and on neurons in the brainstem. A 2025 study in mice separated the two. Intraperitoneal exendin-4, a GLP-1 receptor agonist, produced dose-dependent retching-like behavior, and injecting the drug directly into the area postrema reproduced it ¹⁵. Cutting the vagal afferent nerves did not reduce the response. Genetically deleting the GLP-1 receptor from the area postrema abolished it completely ¹⁵. Mechanistically, exendin-4 raised the activity of area postrema GLP-1 receptor neurons through a postsynaptic pathway dependent on AMPA receptors, which carry fast excitatory signaling ¹⁵. That is a clean localization, and it is a mouse result; no equivalent experiment can be done in a person.

It fits the molecular map. The 2021 genetic map of the mouse dorsal vagal complex identified 25 neuronal populations and found that giving a GLP-1 receptor agonist altered gene expression specifically in two sets of Glp1r neurons, one in the area postrema and one in the nucleus of the solitary tract that also expresses the calcitonin receptor; transcripts and open chromatin near obesity-associated genetic variants were enriched in exactly those populations, and activating several of them reduced feeding in rodents ³⁵. The structure that makes you vomit when poisoned is also one of the structures, with the solitary nucleus next to it, through which the most effective weight-loss drugs act ³⁵. That is not a coincidence to be engineered away casually.

Whether the two effects can be separated is the live pharmacological question, and the answer has moved twice. In 2021 a University of Pennsylvania and Eli Lilly collaboration showed in three species, mice, rats and the musk shrew, that activating the receptor for glucose-dependent insulinotropic polypeptide, GIPR, blocked the emesis and attenuated the illness behaviors caused by GLP-1 receptor activation while preserving the fall in food intake and body weight and the improvement in glucose tolerance, and that the area postrema and the nucleus of the solitary tract were required both for that intake suppression and for processing emetic stimuli ⁴⁹. In 2024 the separation was drawn inside the hindbrain itself: two-photon imaging in mice found that most hindbrain GLP-1 receptor neurons are tuned either to nutritive or to aversive stimuli but not both, that activating solitary nucleus GLP-1 receptor neurons produced satiety with no aversion while activating area postrema GLP-1 receptor neurons produced strong aversion along with reduced intake, that the two populations project to different downstream regions, and that a GLP-1 receptor agonist still reduced food intake when the aversion pathway was inhibited ⁵¹. Then in 2026 a second mouse study reached the opposite conclusion: silencing Glp1r neurons or restoring their signaling population by population, the authors found that solitary nucleus Glp1r neurons restrained normal intake and weight but did not mediate the weight loss produced by the drugs, while area postrema Glp1r neurons mediated both the weight loss and the aversion, which they read as preventing the nauseating and weight-lowering effects from being separated at the level of this circuit ⁵⁶. Both are mouse studies using different tools, and the disagreement is unresolved.

The gut-hormone pharmacology has kept going in parallel. Peptide YY, or PYY, suppresses eating but reliably causes nausea and emesis. In 2026, a Pennsylvania and Pfizer collaboration reported that activating GIPR attenuated PYY-induced malaise in rats while preserving its effect on food intake and weight, and that blocking GIPR instead enhanced the weight loss without worsening malaise and reduced PYY-induced c-Fos expression, a marker of neuronal activation, in the area postrema ⁶³. Since the PYY receptor NPY2R and GIPR are expressed in the same area postrema neurons, the authors propose a circuit-level explanation ⁶³. Again: rodents, and papers with industry authorship. But taken together, this line of work shows that the feeding and sickness effects of gut-hormone drugs can be pulled apart pharmacologically, with the area postrema implicated, even while the circuit experiments disagree about whether they can be pulled apart cell by cell.

What the single-cell atlases found

Until 2020 the area postrema was described by what it responded to. Since then it has been described by what it is made of, using single-nucleus RNA sequencing, a method that reads which genes are switched on in each individual cell nucleus and sorts the cells into types accordingly.

The first atlas, published online at the end of 2020 and carried in the February 2021 issue of Neuron, found that the area postrema contains only a few neuron types ³⁶. Using mouse genetic tools to switch individual types on and off, the authors identified excitatory neurons that induce nausea-related behaviors, with one neuron type mediating the aversion caused by multiple different poisons; nausea-associated responses to agonists of identified area postrema receptors were observed and were suppressed by ablating the relevant cells or deleting the relevant genes ³⁶. Anatomical mapping found a distributed network of long-range excitatory, but not inhibitory, projections, with each subtype having its own pattern ³⁶. The receptors the paper centers on include the calcium-sensing receptor, the GLP-1 receptor, GFRAL, and, downstream, the neurons of the parabrachial nucleus, a relay in the upper brainstem that carries aversive signals onward, which hold calcitonin gene-related peptide, or CGRP ³⁶.

A rat study published at the end of the following year added the pharmacological dimension. Micropunches of area postrema and nucleus of the solitary tract taken six hours after systemic GDF15 or saline yielded about 35,000 nuclei and resolved 19 transcriptomically distinct cell subtypes, including a single population of excitatory neurons positive for both Gfral and Ret, which is the direct target ⁶⁴. About 600 cell-type-specific differential expression events were identified across neurons and glia, separating the direct effects of GDF15 on the Gfral-Ret neurons from the secondary effects shared across other neuronal and glial types ⁶⁴. That separation matters: much of the transcriptional response to a drug acting here appears in cells the drug does not bind directly ⁶⁴.

The 2021 map of the whole dorsal vagal complex set the area postrema in context, identifying 25 neuronal populations with distinct transcriptional and chromatin landscapes and distinct peptide receptor profiles, and linking several of them to human obesity genetics ³⁵. And in 2026 spatial transcriptomics was applied to the area postrema of rats given aquaporin-4 antibody, in the work discussed below ⁵³.

One limitation is worth stating because it recurs through this page, though it has narrowed. The founding atlases are rodent ^35,36,64, and a 2025 harmonization pushed that further, curating mouse and rat single-nucleus data into more than 180,000 cells and 123 cell identities at five levels of resolution and describing populations not previously named, among them a leptin-receptor-expressing neuron population in the rat area postrema marked by a progenitor gene ⁶⁵. The species gap narrowed in 2026, when a cross-species atlas of more than 530,000 cells across rat, mouse and macaque caudal brainstem, with spatial profiling of the rat dorsal vagal complex, reported that the calcitonin-receptor neurons carrying prolactin-releasing hormone are conserved across rodents, macaques and humans ⁶⁶. There is still no human area postrema atlas of the depth available in rodents, so human receptor maps continue to rest on a 1994 autoradiography study and on the human expression data in the 2017 GFRAL papers ^12,13. Wherever this page says a cell type does something, the animal is named.

Area postrema syndrome: when vomiting is a brainstem sign

The clearest human demonstration that this structure does what the animal work says it does is a disease in which it is selectively attacked. Neuromyelitis optica spectrum disorder is an autoimmune disease of the central nervous system in which antibodies, usually against the astrocyte water channel aquaporin-4, damage the optic nerves, the spinal cord and particular brainstem regions ²⁰. One of its presentations is unexplained, unremitting nausea, vomiting or hiccups, and it is named for the structure: area postrema syndrome.

Misu and colleagues described it first, in 2005. Intractable hiccup and nausea occurred in 8 of 47 patients with relapsing neuromyelitis optica, 17 percent, and in none of 130 patients with multiple sclerosis; in six of the eight, MRI showed a linear medullary lesion involving the pericanal region, the tissue immediately around the central canal, which is the narrow fluid channel running down the middle of the lower brainstem and spinal cord, together with the area postrema and the nucleus of the solitary tract, and the symptoms resolved with methylprednisolone ²³. The absence of the syndrome in 130 patients with the disease it is most often confused with is what made it diagnostically useful.

The pathology came in 2011. In archival brainstem tissue from 15 patients with neuromyelitis optica, 5 with multiple sclerosis and 8 neurologically normal subjects, 6 of the neuromyelitis optica cases, 40 percent, had unilateral or bilateral lesions at the area postrema and the medullary floor of the fourth ventricle, and no patient with multiple sclerosis and no control did ⁴⁷. The lesions showed tissue rarefaction and thickened blood vessels with no obvious neuronal or axonal pathology and preserved myelin; aquaporin-4 immunoreactivity was lost or markedly reduced in all six, with lymphocytic infiltrates, microglial activation, eosinophils in three cases and complement deposition ⁴⁷. The odds of nausea or vomiting being documented clinically were 16-fold greater in cases with these lesions, with a 95 percent confidence interval of 1.43 to 437 and p = 0.02 ⁴⁷. That interval is extremely wide, which is what happens with six cases, and the direction rather than the magnitude is what it supports.

In 2015 the International Panel for NMO Diagnosis wrote area postrema syndrome into the formal diagnostic criteria as one of the six core clinical characteristics of the disease ²⁰. In 2018 an international effort measured it. Across a database of 430 AQP4-IgG-seropositive patients from several continents, isolated area postrema attacks occurred in 7.1 to 10.3 percent of patients at disease onset and in 9.4 to 14.5 percent subsequently ²¹. In 100 patients reporting 157 episodes, nausea occurred in 127 episodes, 81 percent, lasting a median of 14 days with a range from 2 to 365; vomiting in 113, 72 percent, at a median of 5 episodes a day with a range from 2 to 40, each lasting 1 to 20 minutes; and hiccups in 102, 65 percent, lasting a median of 14 days with a range from 2 to 365 ²¹. Symptoms consistently and completely resolved after immunotherapy ²¹. From these data the authors proposed diagnostic criteria and a severity scale repurposed from the Pregnancy-Unique Quantification of Emesis and Nausea score, and validated them prospectively ²¹.

The recurring clinical failure is that the symptom is read as a stomach problem. A 2026 review states that area postrema syndrome is still frequently misattributed to primary gastrointestinal disease, delaying immunotherapy by weeks to months, and characterizes a diagnostic pathway that routes patients through gastroenterology before neurology ⁵⁵. The case literature says the same thing in individual voices: a 63-year-old woman whose intractable vomiting, extensive myelitis and leptomeningeal enhancement were first read as possible cancer spread, until aquaporin-4 antibodies returned positive ⁶⁷; a case report whose authors warn that delay can lead to cumulative injury such as optic neuritis ⁶⁸. The 2026 review of intractable hiccups makes the clinical rule explicit: persistent hiccups should be treated as a possible brainstem red flag rather than a benign gastrointestinal complaint ⁶⁹.

Two refinements from 2026 are worth recording. First, the syndrome is not unique to aquaporin-4 disease: it occurs, less commonly, in myelin oligodendrocyte glycoprotein antibody-associated disease and in glial fibrillary acidic protein astrocytopathy, and the three diverge in lesion distribution, immunopathology and treatment ^55,69. Seronegative cases occur too, and one 2025 report of encephalomyelitis with area postrema symptoms and no confirmed antibody raised the possibility that tanycytes might be an immune target in addition to astrocytes, which is a hypothesis from a single case and nothing more ⁷⁰. Second, and more substantial, a 2026 study asked why the area postrema is damaged but comparatively spared. In rats given a monoclonal aquaporin-4 antibody daily for one, two or four days, and in an exploratory spatial transcriptomic analysis of four animals treated for up to five days, only a few chemokines were upregulated in the area postrema, the response was dominated by local microglia and macrophages at a low level of activation, and interleukin-1 beta was below the limit of detection, while neighboring medullary lesions recruited macrophages and neutrophils and expressed interleukin-1 beta; the area postrema strongly expressed Aldh1a2, the key enzyme for making retinoic acid, which the authors propose protects the tissue ⁵³. The same paper carries the human side: in autopsy tissue from seven patients with area postrema lesions, ALDH1A2 staining was preserved or slightly increased despite complete loss of aquaporin-4, and macrophage density in those lesions was far lower than in the same patients' spinal cord lesions, 96.2 against 429.6 cells per square millimeter, p = 0.01 ⁵³. If the retinoic acid proposal holds, it would be one explanation for the 2011 observation that these lesions show inflammation with preserved neurons and myelin ^47,53.

Treatment of the underlying disease has moved fast. A 2025 Bayesian network meta-analysis of 8 trials and 851 patients, 716 of them antibody-seropositive, compared ravulizumab, eculizumab, tocilizumab, rituximab, inebilizumab and satralizumab against control; ravulizumab had a 98 percent probability of being the best treatment for prolonging time to relapse, with a hazard ratio of 0.00 (95 percent credible interval 0.00 to 0.03) ⁷¹. The authors note that no head-to-head trials exist, so the comparison is indirect ⁷¹. The hazard ratio of zero reflects the ravulizumab trial itself, which was open-label, used the placebo arm of an earlier trial as an external comparator, and recorded no adjudicated relapses in 58 patients over 84 patient-years ⁷². France's 2026 national guideline, the only current guideline cited on this page, sets out the practical consequence: an attack is treated as an emergency with intravenous corticosteroids and plasma exchange, without waiting for antibody confirmation, and relapse prevention uses antibodies against B cells, the interleukin-6 receptor or complement C5 ⁷³. What matters for this page is simpler: the vomiting is the sign, and the treatment is immunosuppression, not an antiemetic.

Stroke, cavernoma and the rest of the differential

Inflammation is not the only thing that can damage this patch of tissue, and the general lesson holds whatever the cause: a very small lesion in the dorsal medulla can produce a syndrome that looks entirely gastrointestinal.

Lateral medullary syndrome, also called Wallenberg syndrome, is the stroke caused by occlusion of the vertebral artery or the posterior inferior cerebellar artery, and it classically produces vertigo, loss of pain and temperature sensation on one side of the face and the opposite side of the body, difficulty swallowing, hoarseness and unsteadiness. In the rare case where the infarct extends caudally and dorsomedially into the area postrema, intractable vomiting can be added and can outlast everything else. In one reported case, a woman had sustained nausea, vomiting and poor oral intake for more than a month after the index stroke, after her dizziness had resolved and her nystagmus had disappeared; intravenous metoclopramide with ondansetron did not work, and the symptoms resolved only with the addition of oral domperidone and itopride ³⁰. The authors note that area postrema syndrome from ischemic stroke is very rare ³⁰.

Structural lesions do it too. A 16-year-old boy presented with neck pain, occipital headache, paresthesias of the legs, persistent vomiting and intractable hiccups; MRI showed an intra-axial lesion within the area postrema with hemorrhagic remodeling, and surgical resection confirmed a cavernous malformation, a cluster of abnormal blood vessels ²⁹. The hiccups and vomiting resolved completely after surgery, leaving residual sensory loss on one side at three months, a reminder that this is eloquent brainstem tissue ²⁹.

Infection and post-infectious autoimmunity belong on the list. Two women developed isolated area postrema syndrome shortly after COVID-19, both with isolated dorsal medullary lesions on MRI and SARS-CoV-2 RNA detectable in cerebrospinal fluid by metagenomic sequencing; their courses then diverged entirely according to antibody status, the aquaporin-4-negative patient recovering and remaining relapse-free at 12 months, the aquaporin-4-positive patient relapsing fatally at six months on low-dose prednisone alone ⁷⁴. Two cases cannot establish that the virus caused the syndrome, and the authors frame the viral detection as supporting neuroinvasion or barrier compromise rather than proving causation ⁷⁴.

Three reports from 2026 widen the differential in a way that matters, because none of these causes responds to immunotherapy. A 13-year-old boy with recurrent vomiting, vertigo, nystagmus and failing vision had area postrema lesions and optic nerve enhancement on imaging and looked like antibody-negative neuromyelitis optica, until whole-exome sequencing found a homozygous DNAJC30 variant causing a recessive form of Leber hereditary optic neuropathy, an inherited disease of the mitochondria, the cell's energy-producing compartments; corticosteroids and immunoglobulin helped only transiently, while idebenone improved his vision, normalized his raised blood lactate and cleared the brainstem lesions ⁷⁵. A 43-year-old woman with intractable hiccups, nausea and a dorsal medullary lesion carried a disease-causing variant in GJB1, the gene behind Charcot-Marie-Tooth disease type 1X, an inherited nerve disorder that can occasionally strike the central nervous system; her lesion resolved and she remained relapse-free with no maintenance immunotherapy ⁷⁶. And a spinal dural arteriovenous fistula in the lower neck, an abnormal direct connection between an artery and a vein, produced a medullary lesion well away from the shunt that mimicked the syndrome and resolved after the fistula was closed from inside the vessel ⁷⁷. Each is a single case, so none of them changes how often the syndrome is autoimmune; together they are the reason a negative antibody panel should prompt more thought rather than less.

The practical summary for a clinician is short. Nausea, vomiting or hiccups that persist without a gastrointestinal explanation, especially for 48 hours or more and especially with any other brainstem sign, warrant imaging of the dorsal medulla and antibody testing, because several of the causes are treatable and the delay is measured in weeks to months ^55,69. A negative antibody panel is a reason to look further, not to stop ^75,76,77.

A sensor of inflammation, and the biology of feeling ill

The newest line of work treats the area postrema not as a poison detector attached to a vomiting reflex, but as one of the brain's first readers of the body's inflammatory state, and so as part of the machinery that produces sickness behavior: the loss of appetite, the withdrawal, the loss of interest in things that are normally rewarding, that accompany almost any illness.

The anatomical case for this is straightforward and was made before the experiments: the tissue sits outside the barrier, its tanycyte-like stem cells carry TLR4, the receptor for bacterial lipopolysaccharide, along with the mechanical and osmotic sensor TRPV1 and the sodium-sensing channel Nax, and the organ is positioned to register blood-borne information and convey it to the rest of the brain ^3,33.

The experiments have followed in rodents. In mice with collagen-induced arthritis, a model of rheumatoid arthritis, microglia in the area postrema increased in density and changed shape during the established and chronic phases of disease but not before onset; clustering of 1,256 cells from arthritic animals against 852 from saline controls showed a rising proportion of highly activated forms, the density of microglia expressing interleukin-1 beta increased, and the animals' preference for sugar water fell as interleukin-1 beta expression rose in brain regions containing the area postrema ³⁷. Loss of interest in sweet things is the standard rodent readout for the blunting of reward that accompanies illness and depression, and here it tracked an inflammatory signal in brain tissue containing this structure ³⁷. It is a correlation within an animal model, not a demonstration that human fatigue in rheumatoid arthritis arises there.

The strongest version of the argument comes from cancer. In 2024, work at Cold Spring Harbor Laboratory showed in male mice that circulating interleukin-6, a signaling protein long suspected of driving the wasting of advanced cancer, rapidly enters the area postrema and activates its neurons and the network they feed. A peripheral tumor, which raises circulating interleukin-6, raised it in the area postrema and potentiated excitatory synaptic transmission onto area postrema neurons. Neutralizing interleukin-6 inside the brain of tumor-bearing mice attenuated the cachexia, quieted the overactive network and markedly prolonged lifespan; suppressing Il6ra, the gene for the interleukin-6 receptor, in area postrema neurons alone did the same; and silencing the Gfral-expressing area postrema neurons, the GDF15-sensing cells of the section above, also attenuated the cachectic phenotype ³⁹. In 2026 the same group turned the finding into a candidate treatment: a single dose of an antisense oligonucleotide against Il6ra, given into the cerebral ventricles after cancer onset, lowered interleukin-6 receptor levels in the brainstem, improved cachectic symptoms and extended survival in one of two mouse models, while also damping inflammatory-pathway transcription in brainstem and skeletal muscle ⁷⁸. This is mouse work and the brain-delivery route has not been tried in people, but it is the clearest evidence that the sickness of a systemic illness is produced by this structure reading the blood rather than by the illness acting on muscle and fat alone.

Metabolic disease does something similar and faster. Rodents were given streptozotocin at 200 mg per kilogram to induce type 1 diabetes, and the area postrema was examined seven days later with three-dimensional multiplex immunofluorescence and immunoelectron microscopy ³⁸. At baseline the organ already contained an unusual population of immune cells in its perivascular spaces, positive for the macrophage marker Iba1 but negative for P2RY12, which marks ordinary brain microglia, aligned to the vessel basement membranes and more numerous than in other regions ³⁸. One week after the metabolic insult the perivascular space had shrunk, its structure had become more complex and these cells had increased in number, with no comparable change in the barrier-intact solitary nucleus a fraction of a millimeter away ³⁸. The contrast between the two adjacent structures is the strength of the experiment: whatever is happening is a property of the place without the barrier. Neither the abstract nor the PubMed record names the species, and the full text was not available for this review, so the page says rodent and no more.

A 2026 multi-omic study put a name to those cells. Profiling the immune landscape of the area postrema and the median eminence, the authors found both to be border-like environments with a macrophage compartment unlike that of the rest of the brain: the microglia there have a distinct and heterogeneous phenotype shaped by age and metabolism, they are mixed in origin, with fetal monocytes entering around birth and giving rise to long-lived monocyte-derived microglia that take on a yolk-sac-derived identity, and after depletion the area postrema is repopulated both from local cells and from monocytes; the perivascular macrophages at these borders are likewise distinct in phenotype and origin ⁷⁹. The abstract does not name the species, but the fate-mapping and depletion experiments it describes are mouse methods. The claim this supports is a modest one: the immune compartment of a structure open to the blood is built differently from the rest of the brain's.

The autoimmune work completes the picture from the other side, by showing that the area postrema's immune response is unusually restrained. In the 2026 study of rats given aquaporin-4 antibody, the area postrema upregulated only two chemokines, mounted a low-grade local microglial and macrophage response with granulocytes essentially absent and interleukin-1 beta undetectable, and strongly expressed the retinoic acid-synthesizing enzyme Aldh1a2, while adjacent medullary lesions did the opposite and were tissue-destructive ⁵³. The same pattern held in human autopsy tissue: ALDH1A2 was preserved in area postrema lesions despite complete loss of aquaporin-4, and macrophage density there was far lower than in the same patients' spinal cord lesions ⁵³. A structure that must be open to the blood may need exactly such a brake.

How much of this is human is the usual question, and the answer is: a little. The aquaporin-4 work includes human autopsy tissue ⁵³; everything else here is mouse or rodent ^37,38,39,78, and the 2026 immune-niche study does not name its species ⁷⁹. One human measurement does exist on the GDF15 arm of the sickness response. In Alexander disease, a genetic astrocyte disease in which children fail to thrive and vomit, GDF15 was elevated in cerebrospinal fluid but not in blood in patients; in the rat model of the disease, mutant astrocytes made GDF15, it rose in brainstem and cerebrospinal fluid but not plasma, the GFRAL-expressing area postrema neurons were activated, the animals showed pica and anorexia with reduced lean and fat mass, and suppressing the mutant protein with an antisense oligonucleotide normalized both weight gain and GDF15 ⁸⁰. What can be said beyond that is modest. Two mouse studies, of arthritis and of cancer, found the area postrema reacting early to systemic disease and tied the change to a behavior or to survival ^37,39; a rodent study of chemically induced diabetes found its perivascular immune compartment remodeled within a week ³⁸; and a rat antibody model found its immune response unusually restrained ⁵³. Whether any of this explains human fatigue, appetite loss or low mood in chronic inflammatory disease is unestablished.

Morning sickness, and a hormone the fetus makes

Most pregnancies involve nausea and vomiting, and a small fraction involve hyperemesis gravidarum, the severe form that causes dehydration and undernutrition and that can be life-threatening ^31,32. For most of medical history it was explained by hormones in general terms, or by psychological theories that have not survived. The current explanation runs through the area postrema.

The 2024 Nature paper assembled the case. Higher GDF15 in maternal blood was confirmed to be associated with vomiting in pregnancy and with hyperemesis gravidarum, and mass spectrometry detecting a naturally labeled variant of the protein showed that the vast majority of GDF15 in maternal plasma comes from the feto-placental unit, not the mother ³¹. The twist is in the other direction. Studying rare and common genetic variants, the authors found that low levels of GDF15 in the non-pregnant state increase the risk of hyperemesis, and that women with beta-thalassemia, an inherited anemia in which GDF15 is chronically high, report very low levels of nausea and vomiting of pregnancy ³¹. In mice, the acute response to a bolus of GDF15 depended on prior circulating levels in a way that suggests the system desensitizes ³¹. So the illness is a mismatch: a fetus producing a sickness hormone into a mother whose brainstem has not been exposed to it.

The genetic evidence has since broadened. A 2026 multi-ancestry genome-wide association study of 10,974 cases of hyperemesis gravidarum and 461,461 controls across European, Asian, African and Latino ancestries identified ten loci: the four previously known, GDF15, IGFBP7, PGR and GFRAL, and six additional ones, SLITRK1, SYN3, IGSF11, FSHB, TCF7L2 and CDH9 ³². GDF15 and TCF7L2 were expressed mainly in extravillous trophoblast, the fetal cells that invade the uterine wall, with opposing effects of GDF15 between maternal and fetal genotype, while IGFBP7 and PGR were expressed mainly in maternal spiral arteries with effects limited to the maternal genome ³². That both the hormone and its area postrema receptor appear as independent genetic risk loci is the strongest available human evidence that this axis is causal rather than correlated. The study appeared first as a preprint in December 2024 and then in peer-reviewed form in 2026; the findings are consistent between the two, and the preprint is cited here as the earlier record of the same work ^32,54.

A caution is in order about where this leads. Showing that fetal GDF15 acting on maternal GFRAL explains hyperemesis does not by itself produce a treatment, and no GDF15-blocking drug has been tested in pregnancy. The desensitization finding in mice suggests a prevention strategy, raising maternal exposure before conception, that has not been tried in people ³¹. This is a mechanism established in human genetics and human biochemistry with the behavioral step inferred from animals, and the therapeutic step not yet taken.

What changed, and what is still unknown

What changed. The vomiting center, localized by electrical stimulation in 1950 and taught for half a century, is not now thought to exist as a discrete structure; vomiting is produced by a central pattern generator distributed across loosely organized groups of medullary neurons, with an organizing site in the Bötzinger complex but no chemically sensitive center ^5,16,44. The claim that the chemoreceptor trigger zone is functionally important in motion sickness, asserted in a 1954 title, was overturned by Borison's own cat experiment in 1986 and recorded as settled in the review literature by 1994 ^2,19,43. And GDF15, first described in 1997 as a macrophage cytokine ⁵⁹ and later pursued as a potential obesity drug, turned out to reduce food intake at least partly by causing nausea, which changed the whole therapeutic logic of the pathway ^13,25.

What is solid. The structure lacks a blood-brain barrier and its capillaries are fenestrated ^1,3,4. It carries receptors for dopamine, serotonin, substance P, GLP-1 and GDF15, among others ^11,12,13,15. Destroying it blocks most drug-induced vomiting in animals ². In people, destroying it by inflammation, infarction or hemorrhage produces intractable nausea, vomiting and hiccups, and that syndrome is a formal diagnostic criterion for a defined disease ^{20,21,23,29,30,47}. And the hormone that acts on its best-characterized receptor, GFRAL, is the target of a drug, the anti-GDF15 antibody ponsegromab, that worked on its primary endpoint in a randomized phase 2 trial ²².

What is not settled, first, is how much of the human picture is really human. The cell atlases are mouse and rat, with macaque added in 2026 ^{35,36,64,65,66}. The localization of GLP-1 nausea to this structure is mouse ¹⁵, and the two mouse studies that tried to separate nausea from weight loss cell by cell disagree ^51,56. The separation of feeding from malaise through GIPR is rat, with the earlier version of it in mice, rats and musk shrews ^49,63. The inflammation work is mouse and rodent ^37,38,39,78, and the aquaporin-4 antibody immunology is a rat model with human autopsy tissue alongside it ⁵³. The human evidence is of four kinds only: receptor localization in postmortem tissue ^12,13, lesion-and-syndrome correlation ^21,23,47,53, drug trials whose endpoints are symptoms, weight and food intake rather than brain activity ^{17,18,22,24,50}, and a cerebrospinal fluid measurement in one rare genetic disease ⁸⁰. No recording from a human area postrema neuron appears in the literature reviewed for this page.

Second, the boundaries of the organ's job are still open. Borison conceded in 1984 that it was unknown whether the trigger zone is part or all of the area postrema and suggested that different chemosensory functions might be represented in different regions within it ⁷. Forty years of molecular work has identified cell types but has not closed that question; the 2020 atlas found that one neuron type mediates aversion to multiple poisons, which argues for convergence rather than fine regional specialization, but the mapping was done in mice ³⁶.

Third, whether nausea can be separated from the therapeutic effect of the drugs that act here is the most consequential open question in the field, because it decides how tolerable the next generation of metabolic drugs will be. It can be done pharmacologically in animals: activating GIPR blocked GLP-1-induced emesis in mice, rats and musk shrews while preserving the weight loss in 2021 ⁴⁹, and manipulating GIPR separated PYY's effect on food intake from its malaise in rats in 2026 ⁶³. Whether it can be done cell by cell is contested: a 2024 mouse study found solitary nucleus GLP-1 receptor neurons producing satiety without aversion ⁵¹, and a 2026 mouse study found both the weight loss and the aversion running through the area postrema population ⁵⁶. No equivalent separation has been shown in a person, and the trials of the drugs built on these pathways still report nausea ^24,50.

Fourth, the inflammation story is early but no longer thin. Two mouse studies find that the area postrema registers systemic inflammation early, one in arthritis and one in cancer, where neutralizing interleukin-6 in the brain or silencing its receptor in these neurons lessened the wasting and prolonged life ^37,39,78; a rodent study of chemically induced diabetes finds its perivascular immune compartment remodeled within a week ³⁸; a 2026 rat antibody model finds its immune response restrained, with retinoic acid synthesis offered as the explanation ⁵³; and a 2026 multi-omic study finds its macrophages to be a distinct population with their own origins ⁷⁹. Whether any of this explains human fatigue, appetite loss or low mood in chronic inflammatory disease is unestablished, and the only human measurement on the hormonal arm is a cerebrospinal fluid finding in a rare genetic astrocyte disease ⁸⁰.

Fifth, and smallest only in appearance: no one has explained the hiccups. Persistent hiccups are one of the three defining symptoms of area postrema syndrome, present in 102 of 157 episodes in the largest series ²¹, and they are often what brings the patient in ⁶⁹. But the hiccup reflex involves the phrenic nerves, the diaphragm and the glottis, and nothing in the work cited here shows how a lesion of this patch of tissue produces it; the 2026 reviews describe the association and infer the mechanism from lesion location rather than demonstrating it ^55,69. A page that explains nausea and vomiting in receptor-level detail should say plainly that the third symptom is unexplained.

Finally, the clinical failure is not a research question at all and is the most fixable thing on this page. Patients with area postrema syndrome are still routed through gastroenterology, and immunotherapy is still delayed by weeks to months, for a syndrome with published diagnostic criteria, a published severity scale, a visible lesion on dedicated imaging, a blood test and, since 2026, a national guideline ^21,55,73. The hormonal hypothesis now being floated for why the area postrema becomes hyperexcitable in that disease, involving GLP-1 and its receptor, rests on a single-center study presented at a meeting and posted as a preprint, and the review that reports it calls it hypothesis-generating ⁵⁵. The strongest recommendation this page can carry is the oldest one in it: unexplained vomiting or hiccups that will not stop are a reason to look at the back of the medulla ⁶⁹.

Where it connects

Sources

80 sources, numbered as they are cited. Every one was checked against PubMed or its publisher before it was cited here; the note under each says what it shows and what it does not.

1. 1
   Duvernoy HM, Risold PY. The circumventricular organs: an atlas of comparative anatomy and vascularization.Brain Res Rev · 2007 · 56(1):119-47doi:10.1016/j.brainresrev.2007.06.002 · PMID 17659349

   Comparative atlas built from histological preparations of 12 mammal and 5 bird species. Establishes that the area postrema is the fourth-ventricle circumventricular organ, sited at the confluence of ventricle and central canal, and that the capillaries of these organs have walls devoid of a blood-brain barrier, permitting direct blood-to-tissue exchange. An anatomical atlas, not a functional study; it gives no human measurements in the abstract.

2. 2
   Miller AD, Leslie RA. The area postrema and vomiting.Front Neuroendocrinol · 1994 · 15(4):301-20doi:10.1006/frne.1994.1012 · PMID 7895890

   The review that states the limits of the chemoreceptor trigger zone concept after 40 years: lesions of the area postrema prevent vomiting in response to most but not all emetic drugs, but the area postrema is not essential for vomiting induced by motion or by vagal afferent activation, and its role in radiation-induced vomiting remains controversial. Also summarizes the electrophysiology, 2-deoxyglucose and c-fos evidence and places the output in the nucleus of the solitary tract. A review of animal work.

3. 3
   Price CJ, Hoyda TD, Ferguson AV. The area postrema: a brain monitor and integrator of systemic autonomic state.Neuroscientist · 2008 · 14(2):182-94doi:10.1177/1073858407311100 · PMID 18079557

   Review arguing that the area postrema is not merely a conduit for blood-borne signals but the initial site of their integration, by virtue of lacking tight junctions and having fenestrated capillaries in a densely vascularized structure. A review of mostly rodent physiology.

4. 4
   Morita S, Furube E, Mannari T, Okuda H, Tatsumi K, Wanaka A, Miyata S. Heterogeneous vascular permeability and alternative diffusion barrier in sensory circumventricular organs of adult mouse brain.Cell Tissue Res · 2015 · 363(2):497-511doi:10.1007/s00441-015-2207-7 · PMID 26048259

   Tracer study in adult mice: fluorescein isothiocyanate (molecular weight 389) and dextran 3k diffused into the parenchyma while dextran 10k stayed in the perivascular space between inner and outer basement membranes; permeability was higher centrally than distally; occludin, claudin-5 and ZO-1 were undetectable at central subdivisions; neither small tracer passed the dense GFAP-positive astrocyte/tanycyte network. Mouse anatomy, not human.

5. 5
   Wang SC, Borison HL. The vomiting center; a critical experimental analysis.Arch Neurol Psychiatry · 1950 · 63(6):928-41doi:10.1001/archneurpsyc.1950.02310240087005 · PMID 15426437

   The paper that localized the vomiting center to the medullary reticular formation using stereotaxic electrical stimulation. The PubMed record carries no abstract, so it is cited for the date, the authors and the claim only; the content of the work is taken from Borison's own 1984 retrospective. The entity it named is no longer thought to exist as a discrete structure.

6. 6
   Borison HL, Wang SC. Physiology and pharmacology of vomiting.Pharmacol Rev · 1953 · 5(2):193-230PMID 13064033

   The review usually credited with the chemoreceptor trigger zone concept. The PubMed record carries no abstract and no DOI, so it is cited here for the date, authors and existence of the concept; the substance is taken from Borison's 1984 retrospective and from Miller and Leslie's 1994 review.

7. 7
   Borison HL, Borison R, McCarthy LE. Role of the area postrema in vomiting and related functions.Fed Proc · 1984 · 43(15):2955-8PMID 6542029

   Borison's own retrospective, and the clearest statement of the history: before 1949 the vomiting center was placed in the dorsal vagal nuclei; Borison and Wang used stereotaxic electrical stimulation to place it in the reticular formation at a measurable distance, and formulated the separate emetic chemoreceptor trigger zone, soon identified with the area postrema. Concedes it remains unanswered whether the trigger zone is part or all of the organ, and that different chemosensory functions may be regionally represented. Notes rodents cannot vomit but show other postrema-mediated effects (radiation-induced gastric stasis, conditioned taste aversion) and that postremectomy prevents digitalis-induced arrest of rumination in sheep.

8. 8
   Borison HL, Wang SC. Functional localization of central coordinating mechanism for emesis in cat.J Neurophysiol · 1949 · 12(5):305-13doi:10.1152/jn.1949.12.5.305 · PMID 18137708

   The first paper of the series, in cats. The PubMed record carries no abstract; it is cited for its date, authors, species and claim.

9. 9
   Borison HL, Wang SC. Quantitative effects of radon implanted in the medulla oblongata; a technique for producing discrete lesions.J Comp Neurol · 1951 · 94(1):33-55doi:10.1002/cne.900940103 · PMID 14814219

   The ablation method behind the chemoreceptor trigger zone work. The PubMed record carries no abstract; it is cited for the method and the date.

10. 10
    Wang SC, Borison HL. The vomiting center; its destruction by radon implantation in dog medulla oblongata.Am J Physiol · 1951 · 166(3):712-7doi:10.1152/ajplegacy.1951.166.3.712 · PMID 14885461

    Destruction of the putative vomiting center by radon implantation in dogs. The PubMed record carries no abstract; it is cited for the experiment's existence, species and date.

11. 11
    Carpenter DO, Briggs DB, Strominger N. Responses of neurons of canine area postrema to neurotransmitters and peptides.Cell Mol Neurobiol · 1983 · 3(2):113-26doi:10.1007/BF00735276 · PMID 6140078

    122 neurons in anesthetized dogs, 17 substances tested by iontophoresis. Excitation with histamine, norepinephrine, serotonin, dopamine, apomorphine, angiotensin II, neurotensin, leucine enkephalin, vasoactive intestinal polypeptide, thyrotropin-releasing hormone, gastrin, vasopressin and substance P; no response to acetylcholine, somatostatin or cholecystokinin; inhibition by norepinephrine in 6 of 15 and histamine in 3 of 45 cells. Responses had 2-20 s latency, lasted 30 s to 5 min, and rarely exceeded 2 Hz. All excitatory agents except serotonin and norepinephrine are emetic; none of the three inactive ones is. Anesthetized dog, single-unit recording.

12. 12
    Ohuoha DC, Knable MB, Wolf SS, Kleinman JE, Hyde TM. The subnuclear distribution of 5-HT3 receptors in the human nucleus of the solitary tract and other structures of the caudal medulla.Brain Res · 1994 · 637(1-2):222-6doi:10.1016/0006-8993(94)91236-x · PMID 8180799

    Human postmortem autoradiography with the selective 5-HT3 antagonist [3H]LY278584: highest density in the substantia gelatinosus subnucleus of the nucleus of the solitary tract throughout its extent, then the dorsal subnucleus, then the area postrema, then the commissural and medial subnuclei and the pars gelatinosus of the spinal trigeminal nucleus. Human tissue; receptor density is not the same as functional importance.

13. 13
    Mullican SE, Lin-Schmidt X, Chin CN, et al. GFRAL is the receptor for GDF15 and the ligand promotes weight loss in mice and nonhuman primates.Nat Med · 2017 · 23(10):1150-1157doi:10.1038/nm.4392 · PMID 28846097

    One of the four 2017 receptor papers. Recombinant GDF15 induced weight loss in high-fat-fed mice and in spontaneously obese nonhuman primates; GDF15 bound GFRAL with high affinity; Gfral is expressed in neurons of the area postrema and nucleus of the solitary tract in mice and humans; receptor deletion abolished GDF15's effects on intake and weight, and diet-induced obesity and insulin resistance were worse in GFRAL-deficient mice; signaling required the coreceptor RET. All authors are employees of Janssen Research and Development.

14. 14
    Emmerson PJ, Wang F, Du Y, et al. The metabolic effects of GDF15 are mediated by the orphan receptor GFRAL.Nat Med · 2017 · 23(10):1215-1219doi:10.1038/nm.4393 · PMID 28846098

    Companion 2017 paper: GFRAL is a high-affinity receptor for GDF15 and is required for GDF15's effects on body weight and food intake in mice; Gfral knockout mice were refractory to recombinant human GDF15; a monoclonal antibody blocking the interaction prevented GDF15's metabolic effects in rats; Gfral mRNA is highly expressed in the area postrema of mouse, rat and monkey. All authors are employees of Eli Lilly and Company.

15. 15
    Ding H, Wang M, Zhang J, Wan C, Huang Z, Liu L, Liu J. Exendin-4 induced retching-like behavior mediated by postsynaptic effect via AMPA receptors in the area postrema of mice.Am J Physiol Endocrinol Metab · 2025 · 329(2):E254-E265doi:10.1152/ajpendo.00174.2025 · PMID 40622911

    Intraperitoneal exendin-4 caused dose-dependent retching-like behavior in mice, reproduced by direct administration into the area postrema; vagal afferent denervation did not attenuate it, while genetic ablation of the GLP-1 receptor in the area postrema abolished it completely; the mechanism was postsynaptic and AMPA-receptor-dependent. The cleanest available localization of GLP-1 agonist nausea. Mouse only, and retching-like behavior is a proxy, since mice cannot vomit.

16. 16
    Hornby PJ. Central neurocircuitry associated with emesis.Am J Med · 2001 · 111 Suppl 8A:106S-112Sdoi:10.1016/s0002-9343(01)00849-x · PMID 11749934

    The source for the central correction on this page: it is critical to realize that there is not an isolated vomiting center, but groups of loosely organized neurons throughout the medulla activated in sequence by a central pattern generator. Also gives the viscerotopic organization of the nucleus of the solitary tract into subnuclei for swallowing, gastric sensation, laryngeal and pharyngeal sensation, baroreceptor function and respiration; places 5-HT3 antagonists' primary action at vagal afferents, neurokinin-1 antagonists in the dorsal vagal complex including on vagal motor neurons controlling fundic relaxation, and delta-9-THC at CB1 receptors in the dorsal vagal complex. A review, not new data.

17. 17
    Cubeddu LX, Hoffmann IS, Fuenmayor NT, Finn AL. Efficacy of ondansetron (GR 38032F) and the role of serotonin in cisplatin-induced nausea and vomiting.N Engl J Med · 1990 · 322(12):810-6doi:10.1056/NEJM199003223221204 · PMID 1689807

    Placebo-controlled trial in 28 patients given cisplatin: median time to first emesis 11.6 h with ondansetron versus 2.8 h with placebo; median episodes in 24 h 1.5 versus 5.5; regurgitations or dry heaves per episode 1.17 plus or minus 0.1 versus 3.2 plus or minus 0.5; the first three comparisons all p less than 0.001; 0 of 14 on ondansetron versus 12 of 14 on placebo needed rescue, for which the abstract gives no p-value. Urinary 5-hydroxyindoleacetic acid rose in all patients 2-6 h after cisplatin in parallel with emesis, supporting serotonin release from enterochromaffin cells. A small trial by current standards.

18. 18
    Hesketh PJ, Grunberg SM, Gralla RJ, et al. The oral neurokinin-1 antagonist aprepitant for the prevention of chemotherapy-induced nausea and vomiting: a multinational, randomized, double-blind, placebo-controlled trial in patients receiving high-dose cisplatin--the Aprepitant Protocol 052 Study Group.J Clin Oncol · 2003 · 21(22):4112-9doi:10.1200/JCO.2003.01.095 · PMID 14559886

    Phase 3 trial, cisplatin at 70 mg per square meter or more: complete response (no emesis, no rescue) over days 1 to 5 in 72.7 percent of 260 patients on the aprepitant regimen versus 52.3 percent of 260 on ondansetron plus dexamethasone alone, p less than 0.001, with the advantage significant on day 1 and especially days 2 to 5. Establishes the clinical value of blocking the substance P receptor; the anatomical site of action is inferred from pharmacology, not measured here.

19. 19
    Borison HL, Borison R. Motion sickness reflex arc bypasses the area postrema in cats.Exp Neurol · 1986 · 92(3):723-37doi:10.1016/0014-4886(86)90312-2 · PMID 3709744

    Two groups of five cats, motion-untested and motion sickness-susceptible, given chronic ablation of the area postrema and exposed to vertical sinusoidal oscillation on a spring-suspended platform. Motion sickness occurred postoperatively in all five previously untested cats and in four of five previously susceptible cats; the operated animals were significantly more susceptible than normal cats; the single loss of susceptibility was attributed to excessive damage around the organ. Conclusion: the area postrema is not an essential link in the reflex arc of motion-induced vomiting. The primary experiment that overturned the 1954 claim, performed by one of the authors of the original concept. Cats, and small groups.

20. 20
    Wingerchuk DM, Banwell B, Bennett JL, et al. International consensus diagnostic criteria for neuromyelitis optica spectrum disorders.Neurology · 2015 · 85(2):177-89doi:10.1212/WNL.0000000000001729 · PMID 26092914

    The International Panel for NMO Diagnosis criteria that unified the nomenclature as NMOSD, stratified by AQP4-IgG serostatus, and listed area postrema syndrome among the core clinical characteristics alongside optic neuritis, myelitis and other brainstem, diencephalic and cerebral presentations. A consensus document, not a study.

21. 21
    Shosha E, Dubey D, Palace J, et al. Area postrema syndrome: Frequency, criteria, and severity in AQP4-IgG-positive NMOSD.Neurology · 2018 · 91(17):e1642-e1651doi:10.1212/WNL.0000000000006392 · PMID 30258024

    International database of 430 AQP4-IgG-seropositive patients: isolated area postrema attacks in 7.1-10.3 percent at onset and 9.4-14.5 percent subsequently, across continents. In 100 patients with 157 episodes: nausea in 127 (81 percent) for a median 14 days (range 2-365); vomiting in 113 (72 percent), median 5 episodes/day (2-40), each 1-20 minutes; hiccups in 102 (65 percent) for a median 14 days (2-365). Symptoms consistently and completely resolved after immunotherapy. Diagnostic criteria and a severity scale repurposing the PUQE score were proposed and tested in a prospective validation cohort. Questionnaire-based, so symptom durations are patient-reported.

22. 22
    Groarke JD, Crawford J, Collins SM, et al. Ponsegromab for the Treatment of Cancer Cachexia.N Engl J Med · 2024 · 391(24):2291-2303doi:10.1056/NEJMoa2409515 · PMID 39282907

    Phase 2, randomized, double-blind, 12-week trial in 187 patients with cancer cachexia and serum GDF-15 of 1500 pg/mL or more (40 percent non-small-cell lung, 32 percent pancreatic, 29 percent colorectal). Median between-group weight difference versus placebo: 1.22 kg (95 percent credible interval 0.37-2.25) at 100 mg, 1.92 (0.92-2.97) at 200 mg, 2.81 (1.55-4.08) at 400 mg; appetite, cachexia symptoms and physical activity improved at 400 mg; adverse events of any cause 70 percent versus 80 percent. Funded by Pfizer, NCT05546476; several authors are Pfizer employees. Weight, not survival, was the endpoint, and the trial ran 12 weeks.

23. 23
    Misu T, Fujihara K, Nakashima I, Sato S, Itoyama Y. Intractable hiccup and nausea with periaqueductal lesions in neuromyelitis optica.Neurology · 2005 · 65(9):1479-82doi:10.1212/01.wnl.0000183151.19351.82 · PMID 16275842

    Intractable hiccup and nausea in 8 of 47 cases of relapsing neuromyelitis optica (17 percent) and in none of 130 cases of multiple sclerosis; symptoms resolved with methylprednisolone; in six cases MRI showed linear medullary lesions involving the pericanal region, the area postrema and the nucleus of the solitary tract. The first description of what became area postrema syndrome. A single-center retrospective series.

24. 24
    Wilding JPH, Batterham RL, Calanna S, et al. Once-Weekly Semaglutide in Adults with Overweight or Obesity.N Engl J Med · 2021 · 384(11):989-1002doi:10.1056/NEJMoa2032183 · PMID 33567185

    STEP 1: 1,961 adults randomized 2:1 to semaglutide 2.4 mg weekly or placebo for 68 weeks. Mean weight change minus 14.9 percent versus minus 2.4 percent, estimated treatment difference minus 12.4 percentage points (95 percent CI minus 13.4 to minus 11.5), p less than 0.001. Nausea and diarrhea were the commonest adverse events, described as typically transient and mild to moderate; 4.5 percent versus 0.8 percent discontinued for gastrointestinal events. Funded by Novo Nordisk; several authors are company employees. Cited here for the human nausea burden, not for a mechanism.

25. 25
    Borner T, Shaulson ED, Ghidewon MY, et al. GDF15 Induces Anorexia through Nausea and Emesis.Cell Metab · 2020 · 31(2):351-362.e5doi:10.1016/j.cmet.2019.12.004 · PMID 31928886

    Across three mammalian species: GDF15 caused emesis in Suncus murinus (musk shrews) and induced anorexia and pica, the standard rodent signs of nausea, in mice and in lean and obese rats; mouse data implicate GDF15 in chemotherapy-induced malaise. The conclusion that GDF15's anorexia is mediated by nausea reframes every therapeutic use of the pathway. Animal work; nausea in the non-emetic species is inferred from behavior.

26. 26
    Yang L, Chang CC, Sun Z, et al. GFRAL is the receptor for GDF15 and is required for the anti-obesity effects of the ligand.Nat Med · 2017 · 23(10):1158-1166doi:10.1038/nm.4394 · PMID 28846099

    Third of the four 2017 papers: GDF15 binds GFRAL specifically and with high affinity, GFRAL requires RET to signal, GDF15's reductions in food intake and body weight were abolished in GFRAL-knockout obese mice, and GFRAL expression was limited to hindbrain neurons with none in peripheral tissues. GDF15 did not increase energy expenditure, so the anti-obesity effect is driven by reduced intake. All authors are employees of Novo Nordisk.

27. 27
    Hsu JY, Crawley S, Chen M, et al. Non-homeostatic body weight regulation through a brainstem-restricted receptor for GDF15.Nature · 2017 · 550(7675):255-259doi:10.1038/nature24042 · PMID 28953886

    Fourth 2017 paper, and the one that frames the pathway as a stress rather than a homeostatic system: GFRAL is brainstem-restricted; GDF15 activates GFRAL neurons localized exclusively in the area postrema and nucleus of the solitary tract of the mouse brainstem, which then activate the parabrachial nucleus and central amygdala; Gfral knockout mice are hyperphagic under stress and resistant to chemotherapy-induced anorexia and body weight loss. Authors are employees of NGM Biopharmaceuticals, with coauthors at Merck Research Labs and XTAL Biostructures.

28. 28
    Lindstrom PA, Brizzee KR. Relief of intractable vomiting from surgical lesions in the area postrema.J Neurosurg · 1962 · 19:228-36doi:10.3171/jns.1962.19.3.0228 · PMID 14465560

    Deliberate human surgical lesions of the area postrema, made to relieve intractable vomiting. The PubMed record carries no abstract, so it is cited for the fact that the operation was performed and for its date; no source consulted here reports what the operated patients lost.

29. 29
    Kaake A. Pediatric Area Postrema Cavernoma: Clinical Presentation With Hiccups and Vomiting.Cureus · 2026 · 18(3):e105018doi:10.7759/cureus.105018 · PMID 41970100

    A 16-year-old boy with neck pain, occipital headache, leg paresthesias, persistent vomiting and intractable hiccups; MRI showed an intra-axial area postrema lesion with hemorrhagic remodeling, and resection confirmed a cavernous malformation. Hiccups and vomiting resolved completely after surgery, with residual left-sided hypoesthesia at three months. A single case; it shows that a structural lesion of the same small patch of tissue produces the same syndrome.

30. 30
    Na S, Jung DE, Hwang E, Kim T. Area postrema syndrome caused by medullary infarction.J Stroke Cerebrovasc Dis · 2022 · 31(9):106633doi:10.1016/j.jstrokecerebrovasdis.2022.106633 · PMID 35914512

    Case of persistent intractable nausea and vomiting after a lateral medullary infarct extending caudally and dorsomedially, lasting more than a month after dizziness resolved and nystagmus disappeared. Intravenous metoclopramide with ondansetron failed; symptoms resolved with the addition of oral domperidone and itopride. The authors note area postrema syndrome from ischemic stroke is very rare. A single case; the drug response is anecdotal.

31. 31
    Fejzo M, Rocha N, Cimino I, et al. GDF15 linked to maternal risk of nausea and vomiting during pregnancy.Nature · 2024 · 625(7996):760-767doi:10.1038/s41586-023-06921-9 · PMID 38092039

    Higher maternal GDF15 was confirmed to be associated with vomiting in pregnancy and hyperemesis gravidarum; mass spectrometry of a naturally labeled variant showed the vast majority of maternal plasma GDF15 comes from the feto-placental unit; rare and common variants showed that low prepregnancy GDF15 raises hyperemesis risk; women with beta-thalassemia, in whom GDF15 is chronically high, reported very low nausea and vomiting of pregnancy; in mice the acute response to a GDF15 bolus depended bidirectionally on prior exposure, suggesting desensitization. Human genetics and biochemistry plus mouse physiology; the brainstem step is inferred from the receptor's known location. Published online 13 December 2023 in the Nature volume dated 2024; cited here by its print year.

32. 32
    Fejzo M, Wang X, Tan Q, et al. Multi-ancestry genome-wide association study of severe pregnancy nausea and vomiting.Nat Genet · 2026 · 58(4):810-820doi:10.1038/s41588-026-02564-4 · PMID 41981316

    10,974 cases and 461,461 controls across European, Asian, African and Latino ancestries. Ten loci: four already known (GDF15, IGFBP7, PGR, GFRAL) and six new (SLITRK1, SYN3, IGSF11, FSHB, TCF7L2, CDH9). GDF15 and TCF7L2 expressed primarily in extravillous trophoblast with opposing maternal and fetal effects for GDF15; IGFBP7 and PGR in maternal spiral arteries with maternal-only effects. That both the ligand and its area postrema receptor appear as independent risk loci is the strongest human genetic support for the axis. An association study: it locates risk, not mechanism.

33. 33
    Miyata S. New aspects in fenestrated capillary and tissue dynamics in the sensory circumventricular organs of adult brains.Front Neurosci · 2015 · 9:390doi:10.3389/fnins.2015.00390 · PMID 26578857

    Review of the sensory circumventricular organs, including the area postrema: size-selective and heterogeneous vascular permeability; astrocyte- and tanycyte-like neural stem cells carrying TRPV1, TLR4 and the sodium-sensing channel Nax, expressing tight-junction proteins and surrounding mature neurons; angiogenesis and neurogliogenesis regulating capillary density, with VEGF implicated. Rodent work.

34. 34
    Pham VQ, Tutunculer M, Al-Dulaimi H, Ardjmand D, Fleischmann W, Bachor TP, Xu AW. Modulation of fenestrated vasculature in the median eminence and area postrema in response to neurotoxin exposure and its impairment in aging.Front Aging Neurosci · 2025 · 17:1634283doi:10.3389/fnagi.2025.1634283 · PMID 40904345

    After exposure to monosodium glutamate, a neurotoxin that cannot cross the blood-brain barrier, fenestrated capillaries and vascular permeability fell in the median eminence but rose in the area postrema, tracking tanycytic expression of pleiotrophin and vascular endothelial growth factor A; the adaptive response was abolished in aged animals. Shows the barrier gap is adjustable, not fixed. Mouse study: young (2-4 months) and aged (21-25 months) male C57BL/6 mice, a single subcutaneous dose of monosodium glutamate at 3.5 g/kg, all experiments in males.

35. 35
    Ludwig MQ, Cheng W, Gordian D, et al. A genetic map of the mouse dorsal vagal complex and its role in obesity.Nat Metab · 2021 · 3(4):530-545doi:10.1038/s42255-021-00363-1 · PMID 33767443

    Bulk and single-nucleus gene expression plus chromatin profiling of the mouse dorsal vagal complex: 25 neuronal populations with distinct transcriptional and chromatin landscapes and peptide receptor profiles. A GLP-1 receptor agonist altered gene expression specifically in two sets of Glp1r neurons, one in the area postrema and one in the nucleus of the solitary tract also expressing the calcitonin receptor; obesity-associated genetic variants were enriched near genes in those populations, and activating several reduced feeding. Mouse; the human link is genetic enrichment, not human tissue function.

36. 36
    Zhang C, Kaye JA, Cai Z, Wang Y, Prescott SL, Liberles SD. Area Postrema Cell Types that Mediate Nausea-Associated Behaviors.Neuron · 2021 · 109(3):461-472.e5doi:10.1016/j.neuron.2020.11.010 · PMID 33278342

    The first area postrema cell atlas, by single-nucleus RNA sequencing, revealing a few neuron types. Mouse genetic tools identified excitatory neurons that induce nausea-related behaviors, with one neuron type mediating aversion imposed by multiple poisons; responses to agonists of identified receptors (including CaSR, GLP1R and GFRAL) were suppressed by targeted cell ablation or gene knockout; projections were long-range excitatory but not inhibitory, with subtype-specific patterning. Mouse only; nausea is inferred from behavior. Published online 4 December 2020 and carried in the February 2021 print issue; cited here by its print year, as elsewhere on this page.

37. 37
    Matsushita T, Otani K, Oto Y, Takahashi Y, Kurosaka D, Kato F. Sustained microglial activation in the area postrema of collagen-induced arthritis mice.Arthritis Res Ther · 2021 · 23(1):273doi:10.1186/s13075-021-02657-x · PMID 34715926

    In collagen-induced arthritis mice, microglia in the area postrema increased in density and changed morphology during the establishment and chronic phases but not pre-onset; clustering of 1,256 cells from arthritic and 852 from saline-treated animals showed a rising proportion of highly activated forms; the density of IL-1beta-positive microglia increased, and sucrose preference correlated negatively with IL-1beta expression in brain regions containing the area postrema. Mouse model of rheumatoid arthritis; a correlation within that model, not evidence about human fatigue.

38. 38
    Furube E, Tanaka Y, Yoshida S, Ohgidani M. Type 1 diabetes remodels the area postrema perivascular-immune interface.J Cereb Blood Flow Metab · 2026 · 46(8):2066-2077doi:10.1177/0271678X261427902 · PMID 41823341

    Three-dimensional multiplex immunofluorescence and immunoelectron microscopy. At baseline the area postrema held abundant Iba1-positive, P2RY12-negative cells in its perivascular spaces, aligned to basement membranes and exceeding other regions. Seven days after streptozotocin at 200 mg/kg, perivascular space area was significantly reduced, structural complexity increased and Iba1-positive cells rose in the area postrema, with no comparable change in the barrier-intact solitary nucleus. A single dose of streptozotocin and a single one-week timepoint. Neither the abstract nor the PubMed record names the species and the full text was not available for this review, so this page says rodent and no more.

39. 39
    Sun Q, van de Lisdonk D, Ferrer M, et al. Area postrema neurons mediate interleukin-6 function in cancer cachexia.Nat Commun · 2024 · 15(1):4682doi:10.1038/s41467-024-48971-1 · PMID 38824130

    Male mice. Circulating interleukin-6 rapidly entered the area postrema and activated its neurons and the network they feed; tumor-bearing mice showed raised interleukin-6 there, network hyperactivity and potentiated excitatory transmission; an anti-interleukin-6 antibody given inside the brain attenuated the cachexia, reduced the hyperactivity and markedly prolonged lifespan; CRISPR-based suppression of Il6ra in area postrema neurons and silencing of the Gfral-expressing neurons had similar effects. Mouse only, male only; the central mechanism, not the peripheral one, is what this shows.

40. 40
    Wang SC, Borison HL. A new concept of organization of the central emetic mechanism: recent studies on the sites of action of apomorphine, copper sulfate and cardiac glycosides.Gastroenterology · 1952 · 22(1):1-12PMID 12980223

    The paper that set out the organizational scheme from the sites of action of three agents, apomorphine among them. The PubMed record carries no abstract and no DOI, so it is cited for its date, authors and subject matter only.

41. 41
    Wang SC, Borison HL. Copper sulphate emesis; a study of afferent pathways from the gastrointestinal tract.Am J Physiol · 1951 · 164(2):520-6doi:10.1152/ajplegacy.1951.164.2.520 · PMID 14810961

    The contrast case to blood-borne apomorphine: emesis from an irritant in the gut, traced to gastrointestinal afferent nerves. The PubMed record carries no abstract; cited for its date, authors and subject.

42. 42
    Borison HL, Wang SC. Locus of the central emetic action of cardiac glycosides.Proc Soc Exp Biol Med · 1951 · 76(2):335-8doi:10.3181/00379727-76-18482 · PMID 14827918

    The third agent in the 1952 scheme. The PubMed record carries no abstract; cited for its date, authors and subject.

43. 43
    Wang SC, Chinn HI. Experimental motion sickness in dogs; functional importance of chemoceptive emetic trigger zone.Am J Physiol · 1954 · 178(1):111-6doi:10.1152/ajplegacy.1954.178.1.111 · PMID 13180723

    Cited for its title's claim only: the PubMed record carries no abstract, so nothing in the paper can be verified from the database. Listed because the claim that the trigger zone is functionally important in motion sickness was later reversed by the review literature, which concluded the area postrema is not essential for motion-induced vomiting.

44. 44
    Fukuda H, Koga T. The Bötzinger complex as the pattern generator for retching and vomiting in the dog.Neurosci Res · 1991 · 12(4):471-85doi:10.1016/s0168-0102(09)80001-1 · PMID 1665553

    Systematic stimulation and partial cutting of the medulla in decerebrate, paralyzed dogs. Emetic responses survived cuts of the solitary complex about 1 mm rostral to the obex but not about 3.5 mm, and disappeared after transection at the caudal end of the retrofacial nucleus or partial cutting of the caudal Bötzinger complex. Proposes that emetic vagal afferents descend through the solitary complex to the area subpostrema, which projects through a reticular column to a pattern generator in the Bötzinger complex. Decerebrate dog; the correspondence to the human medulla is assumed.

45. 45
    Koga T, Fukuda H. Neurons in the nucleus of the solitary tract mediating inputs from emetic vagal afferents and the area postrema to the pattern generator for the emetic act in dogs.Neurosci Res · 1992 · 14(3):166-79doi:10.1016/0168-0102(92)90078-q · PMID 1331921

    Cooling the subpostrema portion of the solitary nucleus impaired retching driven by the vagus nerve on the same side; all 40 non-respiratory neurons tested there responded to vagal stimulation, with a mean latency of 306.5 ms; their discharge rose after apomorphine in 6 of 10 and glutamate in 14 of 14 applications into the fourth ventricle; 20 of 289 showed antidromic responses to Bötzinger stimulation. Decerebrate dog.

46. 46
    Fukuda H, Koga T. Non-respiratory neurons in the Bötzinger complex exhibiting appropriate firing patterns to generate the emetic act in dogs.Neurosci Res · 1992 · 14(3):180-94doi:10.1016/0168-0102(92)90079-r · PMID 1331922

    78 non-respiratory Bötzinger neurons responded to vagal stimulation with a mean latency of 387 ms; half burst in synchrony with each retch and a third between retches, many with gradually augmenting firing beforehand; cooling the subpostrema solitary nucleus abolished the retching and those firings for stimulation of the vagus on the same side but not the other. Decerebrate dog.

47. 47
    Popescu BFGh, Lennon VA, Parisi JE, et al. Neuromyelitis optica unique area postrema lesions: nausea, vomiting, and pathogenic implications.Neurology · 2011 · 76(14):1229-37doi:10.1212/WNL.0b013e318214332c · PMID 21368286

    Neuropathology of archival brainstem tissue from 15 patients with neuromyelitis optica, 5 with multiple sclerosis and 8 normal subjects. Six neuromyelitis optica cases (40 percent) and no others had area postrema and fourth-ventricle-floor lesions: tissue rarefaction, thickened vessels, no obvious neuronal or axonal pathology, preserved myelin, AQP4 immunoreactivity lost or markedly reduced in all six, lymphocytic infiltrates, microglial activation, eosinophils in three, complement deposition. Odds of documented nausea or vomiting 16-fold greater with these lesions (95 percent CI 1.43-437, p = 0.02). The confidence interval is very wide because the numbers are small; the direction is what it supports.

48. 48
    Breen DM, Kim H, Bennett D, et al. GDF-15 Neutralization Alleviates Platinum-Based Chemotherapy-Induced Emesis, Anorexia, and Weight Loss in Mice and Nonhuman Primates.Cell Metab · 2020 · 32(6):938-950.e6doi:10.1016/j.cmet.2020.10.023 · PMID 33207247

    Circulating GDF-15 was higher in patients receiving platinum chemotherapy and positively associated with weight loss in colorectal cancer (NCT00609622); high-emetogenicity agents raised GDF-15 and caused weight loss in mice; platinum-induced anorexia and weight loss were attenuated in GDF-15 knockout mice and neutralization with mAB1 improved survival; in nonhuman primates mAB1 attenuated anorexia and emesis. Pfizer preclinical work preceding its clinical anti-GDF-15 antibody, ponsegromab; the antibody used here is named only as mAB1. All authors are Pfizer employees.

49. 49
    Borner T, Geisler CE, Fortin SM, et al. GIP Receptor Agonism Attenuates GLP-1 Receptor Agonist-Induced Nausea and Emesis in Preclinical Models.Diabetes · 2021 · 70(11):2545-2553doi:10.2337/db21-0459 · PMID 34380697

    Mice, rats and musk shrews: GIPR signaling blocked emesis and attenuated the illness behaviors elicited by GLP-1 receptor activation while maintaining reduced food intake, weight loss and improved glucose tolerance. The area postrema and nucleus of the solitary tract were required both for the intake suppression and for processing emetic stimuli, and single-nuclei sequencing placed GIPR and GLP-1R on distinct inhibitory and excitatory neurons there, with GIPR expressed most in GABAergic cells. Preclinical only; coauthored with Eli Lilly employees. The earlier demonstration that a gut hormone's feeding and sickness effects can be pulled apart.

50. 50
    Benichou O, Coskun T, Gonciarz MD, et al. Discovery, development, and clinical proof of mechanism of LY3463251, a long-acting GDF15 receptor agonist.Cell Metab · 2023 · 35(2):274-286.e10doi:10.1016/j.cmet.2022.12.011 · PMID 36630958

    A GDF15 analog and potent agonist at GFRAL and RET. In rodents and obese nonhuman primates it reduced food intake and body weight without signs of malaise or emesis; in a first-in-human single-dose study it produced dose-dependent nausea and emesis in a subset of healthy participants; in a 12-week multiple ascending dose study in overweight and obese participants it significantly decreased food intake and appetite scores with modest weight reduction that the authors report was independent of nausea and emesis (NCT03764774). Their own conclusion is that the weight loss was surprisingly modest, suggesting challenges in leveraging the GDF15 system for clinical weight loss. All authors are Eli Lilly employees. The only human trial of GFRAL agonism on this page.

51. 51
    Huang KP, Acosta AA, Ghidewon MY, et al. Dissociable hindbrain GLP1R circuits for satiety and aversion.Nature · 2024 · 632(8025):585-593doi:10.1038/s41586-024-07685-6 · PMID 38987598

    In mice, only hindbrain GLP-1 receptor neurons were required for the efficacy of GLP-1-based obesity drugs; two-photon imaging found most of those neurons tuned to either nutritive or aversive stimuli but not both; area postrema GLP-1 receptor neurons were broadly responsive and drove aversion, solitary-nucleus ones were biased to nutritive stimuli and drove satiety without aversion; the two populations project to different downstream regions, and a GLP-1 receptor agonist still reduced intake with the aversion pathway inhibited. Mouse only. Its conclusion is contradicted by the 2026 study below.

52. 52
    Groarke JD, Crawford J, Collins SM, et al. Phase 2 study of the efficacy and safety of ponsegromab in patients with cancer cachexia: PROACC-1 study design.J Cachexia Sarcopenia Muscle · 2024 · 15(3):1054-1061doi:10.1002/jcsm.13435 · PMID 38500292

    The protocol paper for the trial above: approximately 168 adults planned, randomized 1:1:1:1 to ponsegromab 100, 200 or 400 mg or placebo subcutaneously every 4 weeks for 12 weeks, primary endpoint mean change in body weight at week 12, secondary endpoints including nausea and vomiting from the Cancer-Related Cachexia Symptom Diary. Cited for the design and the pre-specified endpoints. All authors but one (Crawford, Duke Cancer Institute) are Pfizer employees.

53. 53
    Yu Q, Takai Y, Yamazaki N, et al. The immunological landscape of the area postrema in neuromyelitis optica spectrum disorders.Brain Pathol · 2026 · e70127doi:10.1111/bpa.70127 · PMID 42503643

    Three bodies of evidence in one paper. (1) Lewis rats given the monoclonal AQP4 antibody E5415A intraperitoneally for 1, 2 or 4 days (14 new animals plus age-matched archival tissue): in the area postrema only the chemokines Ccl2 and Ccl6 were upregulated, the response was dominated by local microglia and macrophages at a low level of activation, granulocytes were essentially absent and IL-1beta was undetectable, while perivascular lesions in the adjacent medulla carried many chemokines, neutrophils and IL-1beta. (2) An exploratory spatial transcriptomic analysis of four rats, one untreated and one each after 1, 2 and 5 days of antibody. (3) Human autopsy tissue from 18 NMOSD cases, 7 with area postrema lesions, plus controls: ALDH1A2 immunoreactivity was preserved or slightly increased in those lesions despite complete AQP4 loss, and macrophage density was significantly lower than in the same patients' spinal cord lesions (96.2 plus or minus 66.7 against 429.6 plus or minus 591.5 cells per square millimeter, p = 0.01). Aldh1a2, the key enzyme for retinoic acid synthesis, was strongly expressed in the area postrema but not the medulla, which the authors propose explains the relative tissue preservation; they state that overexpression and knockdown experiments are still needed to prove it, so the retinoic acid protection is a proposal, not a demonstration. Female rats only. The authors declare no conflicts of interest.

54. 54
    Fejzo M, Wang X, Zöllner J, et al. Multi-ancestry GWAS of severe pregnancy nausea and vomiting identifies risk loci associated with appetite, insulin signaling, and brain plasticity.Res Sq · 2024 · preprint rs.3.rs-5487737/v1doi:10.21203/rs.3.rs-5487737/v1 · PMID 39764105

    The December 2024 preprint of the study above, posted before peer review. Cited only as the earlier public record of the same analysis and the same ten loci; findings are consistent with the 2026 peer-reviewed version, which should be preferred.

55. 55
    Shosha E, Madani J. Recognizing and Differentiating Area Postrema Syndrome Across NMOSD, MOGAD, and GFAP Astrocytopathy.Int J Mol Sci · 2026 · 27(19):8661doi:10.3390/ijms27198661 · PMID 42883907

    Current narrative review. States that area postrema syndrome is still frequently misattributed to primary gastrointestinal disease, delaying immunotherapy by weeks to months, and that its pathophysiology has been inferred largely from lesion topography. Reports a single-center study presented at the 2025 American Neurological Association meeting and posted as a preprint proposing that elevated GLP-1 and GLP-1R correlate with the presence and severity of the syndrome in AQP4-IgG-positive disease, with receptor blockade normalizing neuronal hyperexcitability in animal models; the review's own authors state this has not completed independent replication or full peer review and treat it as hypothesis-generating. Note that the definition given in this review's abstract (symptoms persisting 48 hours or less) conflicts with the 48-hours-or-more threshold used elsewhere in the literature; the longer threshold is the one used on this page.

56. 56
    Yacawych WT, Wang Y, Zhou G, et al. A single dorsal vagal complex circuit mediates the aversive and anorectic responses to GLP1R agonists.J Clin Invest · 2026 · advance publication, article 193743doi:10.1172/JCI193743 · PMID 42742987

    The counterpart to the 2024 Nature paper, using silencing and signaling-restoration rather than imaging and activation. In mice, solitary-nucleus Glp1r neurons contributed to the normal restraint of food intake and body weight but did not mediate drug-induced weight loss, while area postrema Glp1r neurons mediated both the weight-lowering and the aversive effects, which the authors read as preventing the two from being separated at the level of this circuit. Mouse only; several authors are Novo Nordisk employees. Cited here because it disagrees with the earlier result, not because it settles the question.

57. 57
    Fujii R, Kawai H, Kai M, Kondo M. Integrated behavioral and biological assessment reveals 5-HT3 receptor-independent nausea/malaise-like responses in cisplatin-treated mice.Biochem Biophys Res Commun · 2026 · 819:153812doi:10.1016/j.bbrc.2026.153812 · PMID 42025017

    Cisplatin produced reduced food intake, weight loss, reduced spontaneous activity, pica and conditioned taste aversion to a similar extent in wild-type and 5-HT3 receptor knockout mice; c-Fos-positive and c-Fos/GFRAL double-positive cells increased in the area postrema and nucleus of the solitary tract in both, with plasma GDF15 elevated for at least one day. Suggests GDF15-GFRAL signaling contributes to cisplatin malaise independently of 5-HT3. Mouse; nausea inferred from behavior, since mice cannot vomit.

58. 58
    Muscaritoli M, Molfino A, Orlando S, Tambaro F. Advancements of investigational agents for cancer cachexia: what clinical progress have we seen in the last 5 years?Expert Opin Investig Drugs · 2025 · 34(11):855-867doi:10.1080/13543784.2025.2588640 · PMID 41222020

    Review: cachexia affects up to 80 percent of patients with advanced cancer and no drug is approved for it by the FDA or EMA; GDF-15-directed agents lead the pipeline, with ponsegromab entering phase 3 and AV-380 and NGM120 in development; also covers anamorelin, ACM-001, ART27.13 and low-dose olanzapine. A narrative review, so its percentages are summarized from the literature it cites.

59. 59
    Bootcov MR, Bauskin AR, Valenzuela SM, et al. MIC-1, a novel macrophage inhibitory cytokine, is a divergent member of the TGF-beta superfamily.Proc Natl Acad Sci U S A · 1997 · 94(21):11514-9doi:10.1073/pnas.94.21.11514 · PMID 9326641

    The first description of the protein now called GDF15, cloned from activated macrophages as macrophage inhibitory cytokine 1 and shown to limit the macrophage's own production of tumor necrosis factor alpha. Cited for what the molecule was originally described as: an immune regulator, not a weight-regulating factor.

60. 60
    Klus W, Ossowska J, Kowalcze K, Kiliszczyk A, Paziewska A. Pathogenesis, Diagnostic Pathways, and New Therapeutic and Nutritional Strategies for Pancreatic Cancer-Associated Cachexia.Cancers (Basel) · 2026 · 18(7):1060doi:10.3390/cancers18071060 · PMID 41976282

    Current review: cachexia affects about 80 percent of patients with pancreatic ductal adenocarcinoma and directly accounts for about 30 percent of cancer-related deaths; GDF-15 is among the drivers and a candidate biomarker; no drug is specifically approved for cancer cachexia in the United States or Europe, and current guidance is multimodal supportive care, with the GDF-15 inhibitor ponsegromab identified as a recent phase 2 advance. A narrative review, so its percentages are summarized from the literature it cites rather than newly measured.

61. 61
    Hong HJ, Lee MH, Park M, et al. Therapeutic Gfral silencing via antisense oligonucleotides ameliorates cancer-associated cachexia and extends survival in tumor-bearing mice.Cell Rep Med · 2026 · 7(8):102939doi:10.1016/j.xcrm.2026.102939 · PMID 42508405

    A single intracerebroventricular dose of a Gfral-directed antisense oligonucleotide in orthotopic mouse colon and pancreatic cancer models blunted GFRAL expression and c-Fos activation in the area postrema and nucleus of the solitary tract, restored body weight, muscle mass and fat, gave better grip strength than a systemic anti-GDF15 antibody, and extended survival. Mouse; delivery into the brain; authors include employees of a company developing the approach.

62. 62
    Cohen AA, Picard M. Potential Risks of Blocking GDF15-Based Brain Energy Sensing.J Am Geriatr Soc · 2026 · 74(5):1516-1519doi:10.1111/jgs.70399 · PMID 41839791

    Commentary: GDF15 signals energetic stress to the brain, and suppressing that signal in frailty or cancer may improve symptoms at a possible cost to long-term health and survival. An argument rather than data, included because this page should not present symptom relief as unambiguous benefit.

63. 63
    Borner T, Pataro AM, Curtis GR, et al. GIPR signaling modulates PYY-induced hypophagia and malaise in rodents.Mol Metab · 2026 · 106:102334doi:10.1016/j.molmet.2026.102334 · PMID 41679433

    In rats, systemic GIPR agonism attenuated PYY-induced malaise while preserving its effects on food intake and body weight, and GIPR antagonism enhanced PYY-induced hypophagia and weight loss without worsening malaise; inhibiting GIPR signaling significantly reduced PYY-induced c-Fos expression in the area postrema, where NPY2R and GIPR are expressed in the same neurons. Extends the group's 2021 finding for GLP-1 receptor agonists (Diabetes 2021, PMID 34380697) to PYY. Rat study with substantial Pfizer authorship.

64. 64
    Reiner BC, Crist RC, Borner T, Doyle RP, Hayes MR, De Jonghe BC. Single nuclei RNA sequencing of the rat AP and NTS following GDF15 treatment.Mol Metab · 2022 · 56:101422doi:10.1016/j.molmet.2021.101422 · PMID 34942400

    Micropunches of area postrema and nucleus of the solitary tract from Sprague Dawley rats 6 hours after systemic GDF15 or saline: about 35,000 nuclei, 19 transcriptomically distinct cell subtypes including a single population of Gfral- and Ret-positive excitatory neurons, and about 600 cell-type-specific differential expression events across neurons and glia, separating direct from secondary effects. Rat; a single 6-hour timepoint. Published online 21 December 2021 and carried in the February 2022 print volume; cited here by its print year.

65. 65
    Hes C, Tomlinson AJ, Michielsen L, et al. A unified rodent atlas reveals the cellular complexity and evolutionary divergence of the dorsal vagal complex.eLife · 2025 · 14:e106217doi:10.7554/eLife.106217 · PMID 41055205

    Curated and harmonized mouse and rat single-nucleus data: more than 180,000 cells and 123 cell identities at five levels of resolution, with calcium-permeable features in astrocytes, neurons co-expressing markers previously thought separate, and a leptin-receptor-expressing neuron population in the rat area postrema marked by a progenitor gene. Rodent only, and an integration of existing datasets as well as new data.

66. 66
    Ludwig MQ, Coester B, Gordian D, et al. A cross-species atlas of the dorsal vagal complex reveals neural mediators of the effects of cagrilintide on energy balance.Nat Metab · 2026 · 8(6):1350-1367doi:10.1038/s42255-026-01539-3 · PMID 42260119

    More than 530,000 cells and 80 neuronal populations across rat, mouse and macaque caudal brainstem, with spatial profiling of the rat dorsal vagal complex. Cagrilintide, an amylin receptor agonist, acts on two conserved calcitonin-receptor populations: acute treatment altered gene expression in area postrema Calcr/Ramp3 neurons but chemogenetic activation of those neurons did not change long-term intake or weight, while long-term treatment raised prolactin-releasing hormone expression in solitary-nucleus Calcr/Prlh cells that are conserved across rodents, macaques and humans, and knocking that gene down abolished cagrilintide's effect but not semaglutide's. Rodent and macaque tissue, with human conservation inferred from expression rather than measured in human brain. Several authors are Novo Nordisk employees. This page cites the peer-reviewed version rather than the 2025 bioRxiv preprint of the same work.

67. 67
    Nittala S, Gonzalez RF, Mahmoudi F, Nelson F. Area postrema syndrome and longitudinally extensive transverse myelitis masquerading as leptomeningeal carcinomatosis.Neurol Sci · 2026 · 47(8)doi:10.1007/s10072-026-09275-y · PMID 42527715

    A 63-year-old woman with treated breast cancer, one week of intractable nausea and vomiting with leg weakness; MRI showed longitudinally extensive transverse myelitis, an enhancing dorsal medullary lesion and diffuse leptomeningeal enhancement initially raising concern for cancer spread; she developed respiratory failure, and AQP4-IgG returned positive. Partial recovery after methylprednisolone and plasma exchange. A single case, cited for how the syndrome presents and is mistaken.

68. 68
    Manole AG, Ion D, Bolocan A, Andronic O, Păduraru DN. Diagnostic Challenges and Clinical Management of Area Postrema Syndrome: A Case Report on Persistent Nausea and Vomiting.Clin Case Rep · 2026 · 14(7):e72472doi:10.1002/ccr3.72472 · PMID 42434537

    Case report emphasizing that persistent nausea and vomiting are frequently misdiagnosed as gastrointestinal disease, that MRI and AQP4-IgG testing enable prompt corticosteroid and long-term immunomodulatory treatment, and that delay can lead to cumulative neurological injury including optic neuritis. A single case; it carries no measurements.

69. 69
    Eliaçık S. Intractable hiccups as a brainstem red flag: demyelinating disorders with emphasis on neuromyelitis optica spectrum disorder and area postrema syndrome.Neurol Sci · 2026 · 47(9)doi:10.1007/s10072-026-09288-7 · PMID 42603218

    Structured narrative review of PubMed, Scopus and Web of Science from inception to January 2026. Defines area postrema syndrome as unexplained hiccups, nausea or vomiting lasting 48 hours or more, a core diagnostic feature of NMOSD; states that NMOSD is the prototypical demyelinating cause of intractable hiccups through selective astrocytic injury in the area postrema, that the syndrome may precede optic neuritis or myelitis, and that multiple sclerosis and MOGAD differ in lesion distribution and immunopathology. A review; its recommendation is clinical judgment, not trial evidence.

70. 70
    Prentice DA, Ambati R, Kho LK, Jenkins T, Parizel PM. Seronegative Autoimmune Encephalomyelitis with Area Postrema Symptoms.Case Rep Neurol · 2025 · 17(1):50-56doi:10.1159/000545402 · PMID 40336575

    A man in his 70s with headache, fever, confusion, seizure and persistent hiccups; MRI showed brainstem involvement and extensive transverse myelitis; AQP4, MOG and GFAP antibodies were initially negative, with serum GFAP-IgG weakly positive after treatment. The authors hypothesize that tanycytes may be an additional immune target. A single case and an explicitly speculative hypothesis; cited as such.

71. 71
    John N, Lim A, Sunthar SR, et al. Immunotherapies in neuromyelitis optica: Bayesian network meta-analysis.J Neurol · 2025 · 272(9):563doi:10.1007/s00415-025-13279-7 · PMID 40775082

    Eight randomized trials, 851 patients, 716 (84 percent) seropositive, comparing ravulizumab, eculizumab, tocilizumab, rituximab, inebilizumab, satralizumab and control. Ravulizumab ranked first for time to relapse, hazard ratio 0.00 (95 percent credible interval 0.00-0.03), with a 0.99 score on the surface under the cumulative ranking curve, a summary of how often a treatment ranks best across the network, and a 98 percent probability of superiority, supported by annualized relapse rate and a seropositive sensitivity analysis. Indirect comparison: no head-to-head trials exist. The ravulizumab estimate comes from an open-label, externally controlled trial with no adjudicated relapses in 58 patients (Ann Neurol 2023, PMID 36866852), not from a randomized arm, so not every one of the eight included trials was randomized against a concurrent control.

72. 72
    Pittock SJ, Barnett M, Bennett JL, et al. Ravulizumab in Aquaporin-4-Positive Neuromyelitis Optica Spectrum Disorder.Ann Neurol · 2023 · 93(6):1053-1068doi:10.1002/ana.26626 · PMID 36866852

    CHAMPION-NMOSD (NCT04201262): a phase 3, open-label, externally controlled trial. Because eculizumab was already approved, a concurrent placebo group was not used and the placebo arm of the earlier PREVENT trial (n = 47) served as an external comparator. No adjudicated relapse occurred in 58 patients on ravulizumab over 84.0 patient-years, against 20 relapses in the external comparator over 46.9 patient-years; two patients had meningococcal infections and both recovered. This is the trial behind the hazard ratio of zero in the 2025 network meta-analysis, and it was not randomized against a concurrent control. Several authors are employees of Alexion, AstraZeneca Rare Disease.

73. 73
    Giorgi L, Marignier R, Pique J, et al. French guidelines for the diagnosis and management of neuromyelitis optica spectrum disorder.Rev Neurol (Paris) · 2026 · advance publicationdoi:10.1016/j.neurol.2026.06.004 · PMID 42481362

    The only current national guideline cited on this page. Lists area postrema syndrome, with nausea, vomiting and severe hiccups often resistant to conventional treatment, among the presentations; states that attacks should be treated as an emergency with intravenous corticosteroids and plasma exchange without waiting for antibody confirmation; and sets out maintenance therapy with antibodies against B cells, the interleukin-6 receptor or complement C5. A consensus document, not a study.

74. 74
    Zhu W, Qian J, Peng M, Li Y, Hu J. Post-COVID-19 Area Postrema Syndrome With SARS-CoV-2 in CSF: A Dual-Case Report and Review of the Literature.Immun Inflamm Dis · 2026 · 14(4):e70421doi:10.1002/iid3.70421 · PMID 41943240

    Two women developed isolated area postrema syndrome after COVID-19, both with isolated T2/FLAIR dorsal medullary lesions and SARS-CoV-2 RNA in cerebrospinal fluid by metagenomic sequencing. The AQP4-IgG-negative patient recovered with immunoglobulin, corticosteroids and mycophenolate and was relapse-free at 12 months; the AQP4-IgG-positive patient relapsed fatally at 6 months on low-dose prednisone alone. Two cases: they show association and prognostic contrast, not that the virus caused the syndrome.

75. 75
    Dzwilewski K, Krygier M, Szymarek J, et al. When LHON Mimics Demyelination: Area Postrema Syndrome in Biallelic DNAJC30 Variants.J Clin Med · 2026 · 15(13):5289doi:10.3390/jcm15135289 · PMID 42452748

    A 13-year-old boy with recurrent vomiting, vertigo, nystagmus and subacute visual loss; T2 and FLAIR lesions of the area postrema with optic nerve enhancement suggested seronegative NMOSD; AQP4 and MOG antibodies were negative; whole-exome sequencing found the homozygous DNAJC30 variant c.152A>G p.(Tyr51Cys), causing a recessive form of Leber hereditary optic neuropathy; idebenone improved vision, normalized blood lactate and resolved the brainstem lesions. A single case. The gene symbol is italicized in the published title.

76. 76
    Cheng PC, Sun YT. Reversible area postrema syndrome in a patient with GJB1 mutation mimicking neuromyelitis optica spectrum disorder.Mult Scler · 2026 · advance publication, article 13524585261436692doi:10.1177/13524585261436692 · PMID 42265941

    A 43-year-old woman with intractable hiccups, nausea and a dorsal medullary lesion carried the pathogenic GJB1 variant p.Arg22Gln, which causes Charcot-Marie-Tooth disease type 1X; the lesion resolved and she remained relapse-free with no maintenance immunotherapy. A single case. The gene symbol is italicized in the published title.

77. 77
    Takenobu Y, Yukawa K, Sugiyama J, et al. Remote brainstem lesion associated with lower cervical spinal dural arteriovenous fistula-venous outflow bottleneck: a case report and literature review.Ther Adv Neurol Disord · 2026 · 19:17562864261456509doi:10.1177/17562864261456509 · PMID 42344943

    A right C6 spinal dural arteriovenous fistula with cranially directed venous drainage produced a medullary lesion remote from the shunt that mimicked area postrema syndrome; fusion imaging showed focal narrowing at the cerebellomedullary fissure as a venous outflow bottleneck; endovascular embolization achieved complete obliteration and clinical recovery. A single case with a literature review; the anatomical mechanism is proposed, not established.

78. 78
    Wang Y, Sun Q, Zheng W, et al. Antisense oligonucleotides against Il6ra ameliorate cancer cachexia in mice.Mol Ther · 2026 · advance publicationdoi:10.1016/j.ymthe.2026.08.028 · PMID 42619263

    Splice-switching antisense oligonucleotides given by intracerebroventricular injection after cancer onset reduced brainstem interleukin-6 receptor levels and ameliorated cachexia in two mouse models, extending survival in one, and reduced inflammatory-pathway transcription in brainstem and skeletal muscle. Oligonucleotides against the human receptor were made but not tested in people. Mouse; delivery into the brain, which has no clinical precedent for this indication.

79. 79
    Van Hove H, Bastos J, Kancheva D, et al. Microglia with distinct origins and non-canonical phenotypes coexist within the border-like niche of the brain's circumventricular organs.Immunity · 2026 · advance publicationdoi:10.1016/j.immuni.2026.09.002 · PMID 42810333

    Multi-omic characterization of the immune landscape of the area postrema and median eminence, which the authors describe as brain-border-like environments with a macrophage compartment unlike the rest of the brain's: a distinct and heterogeneous microglial phenotype shaped by aging and metabolism, mixed ontogeny with fetal monocytes entering perinatally and giving rise to long-lived monocyte-derived microglia that adopt a yolk-sac-derived identity, repopulation of the area postrema after depletion from both local cells and monocytes, and perivascular border-associated macrophages of distinct phenotype and origin. The abstract does not name the species; the fate-mapping and depletion experiments it describes are mouse methods.

80. 80
    Hagemann TL, Sonsalla MM, Luzinski C, et al. Growth Differentiation Factor 15 Elevation in the Central Nervous System Is Associated With Failure to Thrive in Alexander Disease.Ann Clin Transl Neurol · 2025 · 13(1):144-156doi:10.1002/acn3.70209 · PMID 41024483

    R237H rat model of Alexander disease: pica and anorexia at an early age, reduced lean and fat mass and muscle atrophy despite reduced energy expenditure, GDF15 made by mutant astrocytes and raised in brainstem and cerebrospinal fluid but not plasma, GFRAL-expressing area postrema neurons activated, and GFAP suppression with antisense oligonucleotides normalizing both weight gain and GDF15. In patients with Alexander disease, GDF15 was elevated in cerebrospinal fluid but not in blood. A rat model plus a human biomarker measurement; the human part is a measurement, not an intervention.