A drug that works for reasons nobody has settled
— and the case for using more than one.

• Dexamethasone prevents; it does not treat. It has an established role given at the start of anaesthesia and is not effective against established vomiting.
• Its mechanism is not established, and saying so is the correct answer. Three proposals exist and none is settled.
• Its onset is delayed, which is why it is given as early as possible after induction rather than at the end.
• NK1 antagonists block substance P in the dorsal vagal complex — a transmitter no other class touches, which is why they potentiate rather than duplicate.
• Cannabinoids work for chemotherapy and not for surgery. One emetogenic indication does not imply the other.
• A second drug from a different class is worth about as much as the first. A second dose from the same class is worth nothing.

There is a second reason the uncertainty is worth sitting with. Every other class can be placed on the receptor map from lesson 1 — the drug blocks a named receptor at a named site. Dexamethasone cannot. That is not a gap in the map; it is a genuine gap in the pharmacology, and it is the clearest illustration in the topic that clinical efficacy and mechanistic understanding are separate things.

The role of the tachykinins in the emetic pathway was first established by immunohistological work identifying substance P in the dorsal vagal complex of the ferret, an area regarded as essential to the vomiting reflex. Chemotherapeutic drugs induce substance P release, producing both acute and delayed nausea, and NK1 antagonists attenuate both — which is why they are particularly useful for delayed chemotherapy-induced sickness. Their effectiveness there led to their use in the perioperative period, where they potentiate the effect of other antiemetics such as ondansetron and dexamethasone.

Fosaprepitant is the intravenous prodrug: the same molecule with a phosphonic acid group on the triazolone nitrogen, cleaved in the body back to aprepitant. In small numbers of patients undergoing gynaecological or lower-limb surgery it was more effective than ondansetron in preventing vomiting but not nausea, with no effect on complete response rates. That the two effects separate is consistent with aprepitant’s own profile, and it is a reminder that nausea and vomiting are distinct outcomes that a single drug may treat unequally.

A third, orally administered NK1 antagonist. Its defining property is an exceptionally long half-life of 180 hours. A multicentre study found it similarly effective to placebo and to ondansetron given at induction for early control of symptoms, with a suggestion of prolonged protection against emetic symptoms afterwards. It does not appear to have been further evaluated for this indication since, and cost remains a barrier for this class as a whole.

Midazolam’s possible antiemetic mechanism is a decrease in the synthesis and release of dopamine within the chemoreceptor trigger zone — which, if correct, makes it an indirect dopamine story rather than a new receptor. Perioperative midazolam may reduce postoperative sickness by 38 to 55%, and 2 mg intravenously given 30 minutes before the end of surgery may be as effective as ondansetron 4 mg for treatment — a finding confirmed across a meta-analysis of 16 studies. That is a substantial effect for a drug most people would not list among the antiemetics at all.

Lorazepam is used as an antiemetic during chemotherapy, drawing on its amnesic and sedative properties. Its antiemetic mode of action is uncertain, but it may modify central connections to the vomiting centre and prevent the anticipatory nausea seen with repeated doses of chemotherapy. That is the limbic route from lesson 1’s map — the one afferent with no receptor to block — and it is the reason a benzodiazepine appears on that row of the map at all.

Midazolam as an intravenous agent, with its structure, its ring chemistry and its kinetics, belongs to the intravenous induction agent module; only its antiemetic action is taught here.

Nabilone acts at the vomiting centre and has been used as an antiemetic following chemotherapy; dronabinol, synthetic delta-9-tetrahydrocannabinol, is the other orally available cannabinoid used for this purpose. The proposed mechanism is binding of the cannabinoid type-1 receptor, peripherally — where it suppresses intestinal motility — and centrally in the nucleus of the solitary tract.

Two further non-drug measures belong in the same answer, and both were quantified in lesson 1: a propofol-based technique avoiding nitrous oxide, which is worth about as much as adding one antiemetic, and intravenous crystalloid — of the order of 1 to 2 litres, or 20 mL/kg — which reduces the incidence and severity of symptoms and also reduces dizziness and drowsiness. A complete answer to “how would you manage this patient’s risk” contains at least one item that is not a drug.

The evidence behind the left-hand column is a single large factorial trial: ondansetron 4 mg, droperidol 1.25 mg and dexamethasone 4 mg were equally effective, each reducing the risk by about a quarter, and using them together was approximately additive. Applied to a 60% predicted risk, one drug brings it to about 44%, two to about 33% and three to about 24%. In the same trial, a total intravenous technique with propofol and the avoidance of nitrous oxide had similar equivalence to the use of one antiemetic.

The third column is the part worth carrying into an answer. Each added agent buys less than the one before it, because it acts on the risk that is left rather than on the original risk. That is why the argument for a third drug is strong at 80% baseline risk and weak at 20% — and it is a better answer than the word “multimodal” on its own.

The rule first. There is no value in prescribing an antiemetic drug that has already been used within 6 hours of the previous administration. If an adequate dose was given at an appropriate time and the patient vomited anyway, that receptor is blocked and more of the same buys nothing. Three agents are not redosed at all: dexamethasone, transdermal hyoscine and aprepitant. And switching between members of the 5-HT3 class is not a rescue strategy — there is no convincing evidence that any one of them in common use is more effective than another.