The two classic primary-care misses in this presentation are raised intracranial pressure and bowel obstruction, and both are missed for the same reason: the vomiting dominates the consultation and the accompanying feature is not sought. Posture-dependent vomiting — worse on waking, easing after being upright for an hour — is the cardinal history in raised ICP and is almost never volunteered spontaneously. In obstruction, the giveaway is the combination of vomiting with absolute constipation (no stool and no flatus); diarrhoea makes gastroenteritis more likely but does not exclude obstruction with overflow. Silent myocardial infarction deserves its place on the list because in older patients and those with diabetes, autonomic neuropathy can reduce a coronary event to nausea and vomiting alone — the threshold for an ECG in an older person with unexplained vomiting should be very low. Finally, any patient on long-term steroids who has been vomiting cannot absorb their replacement dose and is at risk of an adrenal crisis within hours.

Timing and content do more diagnostic work here than any test available in primary care. Vomiting undigested food eaten many hours earlier is close to pathognomonic of delayed gastric emptying — either gastroparesis (classically in long-standing diabetes) or mechanical outlet obstruction from a peptic stricture or tumour. Bilious vomiting indicates duodenal contents and suggests the pylorus is patent, but its significance depends on distension, bowel function and overall illness severity. Cannabinoid hyperemesis syndrome is now common enough to belong on every list: cyclical severe vomiting in a regular cannabis user, characteristically relieved by hot showers, and frequently investigated for years before the connection is made. GLP-1 receptor agonists have become one of the most common drug causes of nausea in UK practice, and a recent dose escalation is usually the answer.

Grouping by mechanism is not academic tidiness — it directly determines which antiemetic will work, because each group acts through a different receptor pathway. Vestibular nausea is histamine and muscarinic driven, so cyclizine works and ondansetron largely does not. Chemical and metabolic nausea acts on the chemoreceptor trigger zone via dopamine and 5-HT3, so ondansetron or haloperidol are effective. Gastric stasis needs a prokinetic. Choosing an antiemetic by mechanism rather than by local habit is one of the clearest examples in general practice of pathophysiology changing a prescription. Acute angle-closure glaucoma is worth remembering as the ophthalmic cause that reaches the surgery as "vomiting and headache" — the red painful eye with a hazy cornea and fixed semi-dilated pupil is the clue, and it is sight-threatening within hours.

Fundoscopy is the examination most often skipped and most often decisive. Papilloedema in a patient with morning vomiting and headache converts a symptomatic prescription into urgent same-day neurological/acute assessment (which decides imaging), and it takes under a minute. A succussion splash — audible gastric fluid four hours or more after the last meal — is a genuinely useful bedside sign of gastric outlet obstruction that has largely fallen out of use. In acute vertigo with vomiting, the HINTS examination (head impulse, nystagmus, test of skew) outperforms early MRI for distinguishing vestibular neuronitis from posterior circulation stroke, and the single most important discriminator in practice is gait: a patient with vestibular neuronitis can walk unaided if unsteadily, whereas one who cannot stand without falling should be assumed to have a central cause. Dental erosion and parotid enlargement may be the only objective findings in concealed self-induced vomiting.

The pregnancy test is first because getting it wrong has two distinct consequences: a missed ectopic pregnancy, and the prescription of a teratogen to a woman who did not know she was pregnant. Calcium is included in the baseline panel because hypercalcaemia is a treatable and easily overlooked cause of nausea, constipation, thirst and confusion, and it is frequently the first sign of malignancy or hyperparathyroidism. Hypokalaemia in a persistent vomiter is both a consequence and a compounding cause — it worsens ileus and gastric stasis, creating a self-perpetuating cycle that has to be broken by replacement. On the NG12 wording: nausea and vomiting sits in the "consider non-urgent direct-access endoscopy" group in people aged 55 and over with additional features, whereas dysphagia at any age, or age 55 and over with weight loss plus upper abdominal pain, reflux or dyspepsia, is a suspected cancer pathway referral (the urgent route) — these two thresholds are commonly confused, and quoting the wrong one either delays a cancer diagnosis or floods the endoscopy service.

Suspected bowel obstruction goes to the surgical take, not to an outpatient scan, because the decision needed is fluid resuscitation, nasogastric decompression and imaging within hours; arranging an outpatient CT for a patient with bilious vomiting and a distended abdomen is a well-recognised route to a strangulated bowel. The brain tumour pathway is worth naming explicitly: NG12 directs consideration of an urgent direct-access MRI (or CT if MRI is contraindicated), in line with the local pathway, for adults with progressive sub-acute loss of central neurological function, and morning vomiting with headache in an adult with no migraine history is exactly that picture. Hypercalcaemia earns a mention on the cancer list in its own right because in an older patient with vomiting, constipation, thirst and confusion, it is often the presenting biochemical abnormality of myeloma — and NG12 names FBC, calcium, plasma viscosity or ESR, serum protein electrophoresis and serum free light chains (urine Bence Jones only if light chains are unavailable) as the myeloma tests, sent together.

The MHRA restrictions on metoclopramide (2013) and domperidone (2014) changed prescribing substantially and are frequently overlooked: metoclopramide is limited to a maximum of five days because of extrapyramidal reactions and tardive dyskinesia, with particular caution in young people, and domperidone is limited to short courses at the lowest effective dose because of QT prolongation and sudden cardiac death. Giving a prokinetic in mechanical obstruction risks perforation and is an absolute contraindication. The route matters as much as the drug — an oral tablet given to a patient who is actively vomiting is a wasted prescription, and buccal or rectal alternatives are underused in general practice. The most important line in this step is the last one: failure of a well-chosen antiemetic should prompt a return to step 1 first, because a common reason for treatment failure is an unrecognised cause; if the cause still fits, a second agent from a different class is reasonable, avoiding unnecessary polypharmacy.

Sick-day rules are the single highest-value piece of non-drug advice in this pathway because they prevent two specific emergencies. In type 1 diabetes, the instinct to omit insulin when not eating precipitates ketoacidosis; the correct advice is to continue insulin, test glucose and ketones more often, and maintain carbohydrate intake in liquid form. In steroid-dependent patients, vomiting means the oral replacement dose is not absorbed, and an adrenal crisis can develop within hours — these patients need an emergency injection kit and a clear plan. Advising against brushing teeth immediately after vomiting is a small but genuinely useful detail: gastric acid demineralises enamel, and mechanical brushing in the following hour accelerates irreversible erosion, which is why dental damage is often the most visible long-term consequence of recurrent vomiting.

The 48-hour review is where most diagnostic saves happen in this pathway. Acute vomiting from gastroenteritis is nearly always improving by then, so persistence is a strong signal that the original working diagnosis was wrong — and the conditions that masquerade as gastroenteritis for 48 hours (obstruction, raised ICP, DKA, Addison's, hypercalcaemia) are precisely the dangerous ones. The instinct at that review is to add or change antiemetic; the discipline is to go back to the history and examination. Equally important is the active management of short-course antiemetics: metoclopramide and domperidone prescriptions issued in an acute episode have a tendency to become repeat items, which is how MHRA duration limits are quietly breached over months. Every antiemetic prescription should carry a review date from the moment it is issued.