One centre, five ways in
— and the arithmetic of who is worth treating.

• The vomiting centre is not a nucleus. It is a collection of effector neurones in the medulla — the nucleus of the tractus solitarius and part of the reticular formation — with no discrete anatomical site.
• The chemoreceptor trigger zone sits outside the blood-brain barrier. A circulating drug, toxin or hormone can therefore start the reflex without crossing anything.
• Five routes converge on one centre. That convergence is the reason no single antiemetic covers every cause, and the reason combinations work.
• Four risk factors, four percentages. Female sex, non-smoker, previous sickness, postoperative opioids — 10, 20, 40, 60 and 80% as the count rises from none to four.
• Each prophylactic agent takes about a quarter off the risk that is left. Which is why the second and third drug are worth adding at 60% baseline risk and wasted at 10%.

Keep the two apart in an answer. The trigger zone detects; the vomiting centre integrates and drives. A drug that acts only at the trigger zone cannot suppress vomiting driven by the vestibular apparatus, because that afferent never passes through the trigger zone — and that single observation explains why ondansetron does nothing for motion sickness.

The vomiting centre receives converging input from several directions at once, which is exactly why no single antiemetic is reliably effective against every cause. Dopamine, other drugs and serotonin act at the trigger zone; the limbic cortex carries psychogenic and anticipatory input; the vestibular apparatus carries motion-related input; peripheral pain pathways contribute directly; and cardiovascular and abdominal vagal afferents carry visceral input. All of them converge on the vomiting centre, which alone drives the motor act.

Acetylcholine at muscarinic receptors and histamine at H1 receptors carry the signal from the vestibular apparatus — which is the reason antimuscarinics and antihistamines have a specific role against motion-related and middle-ear surgery nausea rather than acting as general first-line agents. In the central nervous system, histamine receptors are found in the area postrema, the vomiting centre, the vestibular nucleus and the nucleus tractus solitarius. Substance P, acting at NK1 receptors in the dorsal vagal complex, is a separate transmitter again.

Vomiting begins with retrograde contractions of the ileum and jejunum, moving gut contents back towards the stomach. The glottis closes to protect the airway. The diaphragm contracts, creating negative intrathoracic pressure as the pharyngeal sphincters relax, and the abdominal muscles contract to raise intra-abdominal pressure and compress the stomach. With the upper sphincters open and the stomach under pressure, emesis occurs.

Efferent output from the vomiting centre travels by cranial nerves V, VII, IX, X and XII, through vagal parasympathetic and sympathetic chain fibres, and to skeletal muscle through α motor neurones.

Two things follow from the map, and both are worth stating explicitly because both are routinely asked. First, a class blocks a receptor, not a cause: the 5-HT3 antagonists cover the trigger zone and the vagal afferent route and cover nothing carried by acetylcholine, so they are ineffective against vestibular nausea and against vomiting induced by dopamine agonists. Second, the receptor a drug blocks exists elsewhere in the body, and the drug does not know which copy you meant — which is the whole content of the dopamine antagonists’ adverse-effect profile in the next lesson.

The two figures are the same population described at different granularity rather than a disagreement: roughly a third of an unselected list, rising towards 80% once risk factors accumulate. The second is what makes prophylaxis worth targeting rather than giving to everybody.

It matters beyond patient experience. Repeated retching and vomiting cause dehydration, electrolyte abnormality, wound dehiscence, bleeding, oesophageal rupture and airway compromise. In one large prospective series of recovery-room admissions, nausea and vomiting was the single most frequent complication recorded. And it is the leading cause of unanticipated hospital admission after outpatient surgery, which is the health-economic argument in one sentence.

Patient factors

Surgical factors

Risk rises with the duration of surgery, plausibly through prolonged exposure to emetogenic lipophilic drugs. Independent of duration, certain procedures carry a higher incidence: laparotomy, gynaecological surgery, laparoscopic surgery, and ear, nose and throat, breast, plastic and orthopaedic procedures. In children the procedure itself carries the greatest association, and herniorrhaphy, tonsillectomy and adenoidectomy, strabismus surgery and surgery on the male genitalia carry the highest risk.

Anaesthetic factors

Inhalational agents, nitrous oxide, neostigmine and opioids have all been implicated. The correlation is weaker than for patient and surgical factors, and most scoring systems do not incorporate anaesthetic factors specifically — although the duration of anaesthetic exposure is associated with sickness. One distinction is worth carrying into an answer: intraoperative opioids do not appear to increase risk, but postoperative opioids do, and it is the postoperative requirement that the risk score counts.

Learn the rounded column. The unrounded figures are what the derivation cohort produced, and they are worth having only to show that the tidy 10, 20, 40, 60, 80 sequence is a rounding of real data rather than a rule of thumb someone invented.

The two columns on the right of that table are a second table folded into the first to save space: the points run 0 to 5 and do not correspond to the factor beside them. Read the left column for what is scored, and the right pair for what the total predicts.

Only the two-, three- and four-factor figures are quoted, and only those are given here. Note how much steeper the paediatric curve is than the adult one: two factors in a child predicts 30% where two Apfel factors in an adult predict about 40%, but four factors reaches 70% on a scale where the top adult band is 80% — and the baseline is already high, because strabismus surgery has historically carried a postoperative vomiting incidence of more than 50% on its own.

What the paediatric score is for

The score exists to trigger a plan, and the plan is more aggressive than the adult one. A child with more than two risk factors — which every child having strabismus surgery already is — should receive a 5-HT3 antagonist and a steroid prophylactically: dexamethasone 0.1 to 0.2 mg/kg at the beginning of the procedure and ondansetron 0.1 mg/kg towards the end. A child with all four should additionally avoid nitrous oxide and the inhalational agents altogether, with total intravenous anaesthesia using propofol considered instead. A sub-hypnotic propofol infusion combined with antiemetics of other classes also reduces the incidence.

Two ideas hold the algorithm together, and both are quantitative rather than a matter of taste. The first is that a single prophylactic agent — ondansetron 4 mg, dexamethasone 4 mg or droperidol 1.25 mg intravenously — reduces the relative risk by about 26%, and that the three are equally effective. The second is that agents acting on different receptors combine, so the reduction compounds: at a 60% baseline, one agent brings the risk to about 44%, a second from a different class to about 33%, and a third to about 24%.

For a patient carrying three or more risk factors, the working target quoted in one national quality measure is at least two different antiemetics for every patient over 18 undergoing any procedure under inhalational general anaesthesia. That is a floor rather than a ceiling, and it is a convenient number to be able to quote.

When prophylaxis has failed

A subset of patients will need treatment in recovery despite appropriate prophylaxis. The rule is the same one the algorithm is built on: give a drug from a class that has not already been used. If an adequate dose of an antiemetic was given at an appropriate time and did not work, more of the same class is unlikely to produce any significant benefit — so a drug of the same class is not repeated within six hours of the first dose. Three agents are not redosed at all: transdermal hyoscine, dexamethasone and aprepitant.

There is also no convincing evidence that any one of the serotonin antagonists in common use is more effective than any other, so switching between them is not a rescue strategy — it is the same class again under a different name.

This follows directly from the physiology rather than from a drug list. If the centre is driven by convergent afferents, then removing an afferent is as legitimate an intervention as blocking a receptor — and the trial evidence puts the two on comparable footing.

Two of those numbers are of the same order as adding a drug, which is the point. A propofol-based technique for minor surgery, where opioid use is limited, may reduce the incidence on its own. Propofol’s own antiemetic pharmacology belongs to the intravenous induction agent module; what matters here is that the technique is part of the prophylaxis.

You should now be able to draw the reflex without notes: one centre with no discrete anatomical site, one detector sitting outside the blood-brain barrier, five afferent routes, four receptors and a fifth transmitter at the NK1 receptor. You should be able to convert a patient into a number, and a number into a plan — including the plan of giving nothing. And you should be able to say why the plan for a high-risk patient involves two drugs and a change of technique rather than a bigger dose of one drug.

Lesson 2 attaches the receptor antagonists to that map, class by class, and derives their adverse effects from where the blocked receptor sits. Lesson 3 takes the agents that do not fit the scheme at all.