Antiemetics,
as it has actually been examined.

Because the CTZ lies in the area postrema, on the floor of the fourth ventricle, functionally outside the blood–brain barrier. It is rich in dopamine (D2) and serotonin (5-HT) receptors, so a circulating agent can activate it directly without needing to cross the barrier at all.

Physiology of vomiting →

Vomiting is a coordinated motor act: the glottis closes, the diaphragm contracts against relaxed pharyngeal sphincters, and the abdominal muscles contract to raise intra-abdominal pressure and expel stomach contents. Regurgitation is passive — the sphincters are not fully closed, so gastric contents move into the oesophagus under raised abdominal pressure alone, without that coordinated sequence, as can happen with excessive mask ventilation.

Physiology of vomiting →

Female sex, non-smoking status, a history of PONV or motion sickness, and the use of postoperative opioids.

Risk factors and the prophylaxis framework →

Several distinct neurotransmitter systems — dopaminergic, serotonergic, cholinergic, histaminergic and others — converge independently on the vomiting centre. A single class only blocks one of those inputs; escalating its dose does not address the others, while adding an agent from a different class covers an input the first drug cannot reach.

Risk factors and the prophylaxis framework →

No. They are effective for chemotherapy- and radiotherapy-induced nausea and vomiting and for PONV, but not for nausea and vomiting caused by vestibular stimulation or by dopamine agonists — those causes need a drug acting on a different receptor.

5-HT3 receptor antagonists →

Ondansetron is specific for the 5-HT3 receptor, with no meaningful activity at dopamine, histamine, adrenergic or cholinergic receptors. Droperidol acts as a dopamine (D2) antagonist, and it is D2 blockade that produces extrapyramidal effects — a receptor ondansetron does not touch.

5-HT3 receptor antagonists →

Glycopyrrolate is a synthetic quaternary amine — permanently charged — so unlike the naturally occurring tertiary amines hyoscine and atropine, it cannot cross the blood–brain barrier and has no central effects at all.

Anticholinergics →

Specifically because of its cardiovascular effects — atropine produces more marked tachycardia than hyoscine. Hyoscine, which is more sedating and less cardioactive, is the anticholinergic actually used for antiemetic purposes.

Anticholinergics →

These agents have significant anticholinergic activity alongside H1 blockade, concomitantly blocking muscarinic receptors in the vestibular system — part of their antiemetic effect comes from this second mechanism, not from H1 blockade alone.

Antihistamines →

It is not fully established. Proposed mechanisms include central inhibition of prostaglandin synthesis, control of endorphin release, and an anti-inflammatory effect that may reduce 5-HT release within the gut — say the mechanism is unsettled rather than assert one confidently.

Corticosteroids →

Metoclopramide crosses the blood–brain barrier and acts directly on central D2 receptors — the same property that produces its extrapyramidal and sedative effects. Domperidone shares the same D2-antagonist mechanism but does not cross the blood–brain barrier, so it is far less likely to cause them.

Benzamides →

Its antiemetic efficacy is modest — a trial of 30 studies found systemic metoclopramide 10 mg reduced 24-hour PONV against placebo with a number needed to treat of 7.8, and roughly half of clinical studies found it no better than placebo. Set against that modest benefit are its extrapyramidal risk, sedation, and the other problems that follow from crossing the blood–brain barrier.

Benzamides →

On the neurokinin-1 receptor, whose primary ligand is substance P, at the brainstem nuclei of the dorsal vagal complex — the nucleus tractus solitarius and area postrema.

NK1 receptor antagonists →

No — cannabinoids are effective for CINV but are not effective for managing PONV. The two indications should not be assumed to transfer.

Miscellaneous and adjuncts →

Because they act on different receptors — 5-HT3, an uncertain steroid-mediated pathway, and D2 respectively — that converge independently on the vomiting centre. NK1 antagonists are specifically described as potentiating ondansetron and dexamethasone rather than replacing them, which is the general principle: combining classes covers more of the afferent input than escalating one class ever can.

Summary and combination therapy →

Each reflects the drug acting within the CNS rather than being confined to the CTZ or a peripheral site — extrapyramidal effects from central D2 antagonism (metoclopramide, the phenothiazines, droperidol), the central anticholinergic syndrome from central muscarinic antagonism (hyoscine, atropine), and antihistamine sedation from central H1 antagonism. The unifying requirement is crossing the blood–brain barrier: domperidone (D2 antagonist) and glycopyrrolate (antimuscarinic) share the same receptor mechanism as their sister drugs but, because neither crosses the barrier, neither produces these central effects.

Anticholinergics →

Two different mechanisms in the same molecule. The prokinetic effect that makes it useful for aspiration prophylaxis — increased gastric emptying and increased lower oesophageal sphincter tone — comes from a separate cholinergic action on the gut. The antiemetic effect is dopamine (D2) receptor antagonism at the CTZ, a distinct receptor and a distinct site. Attributing both actions to the same receptor is a common examiner-flagged error.

Benzamides →

Sodium citrate, a non-particulate antacid, neutralises acid already in the stomach. Ranitidine, an H2 antagonist, reduces the volume and acidity of acid secreted by gastric parietal cells, with no effect on gastric emptying or lower oesophageal sphincter tone. Metoclopramide, a D2 antagonist and prokinetic, increases gastric emptying and lower oesophageal sphincter tone instead — three genuinely different mechanisms rather than three doses of the same idea.

Ephedrine has both direct and indirect sympathomimetic actions, and the indirect component depends on releasing noradrenaline from sympathetic nerve terminals. Repeated doses progressively deplete those noradrenaline stores, so the indirect contribution — and therefore the overall pressor effect — wanes with successive doses.

It binds the ryanodine receptor (RYR1) on the sarcoplasmic reticulum of striated muscle, uncoupling the excitation–contraction process and preventing the excessive release of Ca²⁺ that drives MH’s generalised muscle rigidity. Vascular smooth muscle and cardiac muscle are not primarily dependent on sarcoplasmic-reticulum Ca²⁺ release for contraction, so they are relatively spared — dantrolene has little effect on the muscle action potential or on non-depolarising block duration.

An initial 2.5 mg/kg IV, followed by 1 mg/kg every 5 minutes until the metabolic signs begin to resolve. There is no fixed upper limit, but little additional benefit is seen above a total of 10 mg/kg. Treatment continues on intensive care and should not stop until symptoms have fully resolved, since MH may recur.

She scores four out of four: female sex, non-smoker, a history of motion sickness, and postoperative opioids expected. That puts her predicted risk at about 80% — the top band — and laparoscopic surgery is independently associated with a higher incidence on top of that.

At that risk I would give at least two agents from different classes, and I would treat the anaesthetic technique as part of the prophylaxis rather than as a fallback: a propofol-based maintenance rather than a volatile agent, nitrogen rather than nitrous oxide, and multimodal analgesia to keep the postoperative morphine requirement down. Each drug takes roughly a quarter off the risk that is left, so two agents bring her from about 80% to about 44%, which is a great deal more useful than doubling the dose of one.

I would also be explicit that she is a day case, so the four-factor score overestimates her — but she is at the top of it either way, and age under 50 is an additional independent factor for symptoms after discharge, which she also has.

Likely follow-ups

• Which of her factors would you drop if she were a smoker, and what does that do to the number?
• She is going home the same day. Does that change the score you should be using?
• How much does one prophylactic agent actually buy her?

Antiemetics 1: The vomiting reflex and prophylaxis →

I would start from the mechanism, because every disadvantage follows from it. Metoclopramide’s antiemetic action is D2 antagonism at the chemoreceptor trigger zone. It also antagonises 5-HT3 and is a 5-HT4 agonist peripherally, but the antiemetic effect is essentially the dopamine one.

The first disadvantage is coverage. The trigger zone is one of several afferents converging on the vomiting centre. Vestibular input runs through muscarinic and H1 receptors, the dorsal vagal complex through NK1, and the limbic cortex through neither — so a drug that antagonises D2 alone leaves most of the afferent supply untouched.

The second is that it crosses the blood-brain barrier. That is what lets it reach the trigger zone and the vomiting centre directly, but it also brings sedation, agitation, dry mouth and the whole extrapyramidal picture — acute dystonia with oculogyric crisis, opisthotonus, trismus and torticollis, akathisia, and rarely the neuroleptic malignant syndrome. Those can appear up to 72 hours later, and they are commoner in young women. Akathisia after an intravenous dose can be severe enough to lead to cancellation of surgery.

Third, haemodynamic effects: hypotension, tachycardia and bradycardia after rapid intravenous administration.

Fourth, the efficacy evidence is weak. About half of clinical studies have found placebo as effective. A meta-analysis of 10 mg gives an odds ratio of 0.58 with a number needed to treat of about 8 — real, but modest — and current guidance does not list it first-line. It works better at 20 mg given at the end of anaesthesia rather than at induction, and better still at 20 to 25 mg, but that is exactly the dose range that produces more akathisia.

And it should be avoided, or used with real caution, in Parkinson’s disease, restless legs syndrome and other dopamine-related movement disorders.

Domperidone answers the counterfactual: same D2 antagonism, same useful site, but it does not cross the barrier, so the extrapyramidal effects largely disappear — while the hyperprolactinaemia does not, because the pituitary is outside the barrier too. For a high-risk patient I would use a 5-HT3 antagonist with dexamethasone, from two different classes, rather than metoclopramide at all.

Likely follow-ups

• Which receptors does it act on, and in which direction at each?
• How would domperidone differ, and why?
• What would you give instead, and what would you give as well?

Antiemetics 2: The receptor antagonists →

I would classify them by the receptor they act on in the control of vomiting, because that is what makes the classification predictive rather than a list. 5-HT3 antagonists, such as ondansetron, acting both peripherally on vagal afferents and centrally at the trigger zone. Dopamine D2 antagonists, which subdivide by structure into phenothiazines such as prochlorperazine, butyrophenones such as droperidol, and benzamides such as metoclopramide, all acting at the trigger zone. Histamine H1 antagonists, such as cyclizine, acting on the vestibular pathway and in the brainstem. Antimuscarinics, such as hyoscine, on the vestibular pathway. Corticosteroids — dexamethasone — whose mechanism is not established. And NK1 antagonists, such as aprepitant, blocking substance P in the dorsal vagal complex. To those I would add benzodiazepines and cannabinoids as adjuncts, and propofol’s own antiemetic action at sub-hypnotic concentration.

The reason for combining them is anatomical rather than pharmacokinetic. Several distinct afferent systems converge on one vomiting centre, so a drug that blocks one receptor leaves every other afferent open. Combining agents from different classes therefore adds cover, while a second dose of the same class adds nothing — the receptor is already blocked.

And I can put a number on it. Ondansetron 4 mg, droperidol 1.25 mg and dexamethasone 4 mg are each about equally effective, reducing relative risk by about a quarter, and they are approximately additive: from a 60% predicted risk, one drug gives about 44%, two about 33% and three about 24%. That is why a four-factor patient gets two or three agents and a no-factor patient gets none.

Likely follow-ups

• Which of your classes will not work for motion sickness, and why?
• What is dexamethasone's mechanism?
• Your patient has had ondansetron and dexamethasone and is still vomiting. What now?

Antiemetics 3: Steroids, NK1 antagonists and combinations →