Reference
Guidelines
Summaries of widely used recommendations, written to make the reasoning visible — including where the evidence is thinner than the confident phrasing of a protocol suggests.
Read this first. These are condensed summaries for revision, not the guidelines themselves. Recommendations are revised, and this page will lag behind the source documents. Always work from the current published version and your own unit's protocol.
Severe traumatic brain injury
Based on the Brain Trauma Foundation's Guidelines for the Management of Severe Traumatic Brain Injury. A recurring feature of these guidelines is worth noting up front: many long-standing practices carry only Level III recommendations, because the randomised evidence to support something stronger does not exist.
| Parameter | Recommendation | Evidence |
|---|---|---|
| ICP treatment threshold | Treat sustained pressures above 22 mmHg. | Level IIB |
| Cerebral perfusion pressure | Target 60–70 mmHg; the individual optimum varies. | Level IIB |
| Blood pressure | Avoid hypotension — a single episode is associated with markedly worse outcome. | Level III |
| Hyperventilation | Not recommended prophylactically; a temporising measure only, and avoided in the first 24 hours. | Level IIB |
| Corticosteroids | Not recommended — associated with increased mortality in the CRASH trial. | Level I |
| Seizure prophylaxis | Phenytoin reduces early post-traumatic seizures; no benefit shown for late seizures. | Level IIA |
| Nutrition | Feed to full caloric replacement by day five to seven. | Level IIA |
Note the asymmetry: the strongest recommendation in the list is a negative one. Steroids seemed physiologically sensible for decades until a large trial showed they caused harm — a useful reminder about mechanism-based reasoning in the absence of outcome data.
Aneurysmal subarachnoid haemorrhage
Drawing on American Heart Association / American Stroke Association and Neurocritical Care Society guidance.
- Secure the aneurysm early. Treatment as soon as feasible reduces rebleeding, which carries very high mortality and peaks in the first 24 hours.
- Nimodipine for all patients. 60 mg orally every four hours for 21 days. It improves neurological outcome; notably, this appears not to be explained by any effect on angiographic vasospasm.
- Maintain euvolaemia. Prophylactic hypervolaemia and the older "triple-H" approach are no longer recommended; induced hypertension is reserved for established delayed cerebral ischaemia.
- Watch for delayed cerebral ischaemia between days four and fourteen, with serial neurological examination as the primary monitor.
- Blood pressure before securing: control extreme hypertension while preserving perfusion. The precise target is not well established by trial evidence.
Acute ischaemic stroke
- Intravenous thrombolysis within 4.5 hours of onset in eligible patients; benefit falls sharply with time, hence the emphasis on door-to-needle intervals.
- Mechanical thrombectomy for large vessel occlusion within 6 hours, and out to 24 hours in selected patients where advanced imaging shows salvageable tissue.
- Permissive hypertension in patients not receiving thrombolysis — treat only above roughly 220/120 mmHg. Lowering pressure reduces collateral flow to the penumbra.
- After thrombolysis, keep blood pressure below 180/105 mmHg to limit haemorrhagic transformation.
- Decompressive hemicraniectomy within 48 hours for malignant middle cerebral artery infarction reduces mortality substantially. Survival with significant disability is the common outcome, so this is a decision requiring explicit discussion with family.
Status epilepticus
Neurocritical Care Society guidance. The dominant failure mode in practice is under-dosing the first-line drug and then moving on too slowly.
- Emergent (0–5 min): a benzodiazepine at an adequate dose — intravenous lorazepam, or intramuscular midazolam where access is not established.
- Urgent control (5–20 min): an intravenous antiseizure medication. The ESETT trial found levetiracetam, fosphenytoin and valproate roughly equivalent, so the choice can rest on contraindications and availability.
- Refractory (20–40 min): anaesthetic infusion — midazolam, propofol or thiopentone — with continuous EEG monitoring.
- Throughout: treat the cause. Check glucose, sodium, calcium and toxicology, and image the brain.
Non-convulsive status is a real risk in any patient who does not wake as expected after convulsions stop. Motor activity ceasing is not the same as seizures ceasing — that distinction needs EEG.
Brain death determination
Criteria are jurisdiction-specific and the local legal framework governs. The common structure is:
- Prerequisites: a known, irreversible cause; exclusion of confounders — hypothermia, sedative and neuromuscular blocking drugs, severe metabolic derangement, hypotension.
- Clinical examination: unresponsive coma, and absent brainstem reflexes (pupillary, corneal, oculocephalic, oculovestibular, gag and cough).
- Apnoea test: no respiratory effort despite a documented rise in PaCO₂ to an agreed threshold.
- Ancillary testing where any component of the clinical examination cannot be completed or reliably interpreted.
Reading a recommendation critically
Three questions worth asking of any guideline statement:
- What is the evidence level? Level I and Level III sit in the same document and often in the same sentence structure, but they are not the same kind of claim.
- Who was in the trials? Exclusion criteria decide who the recommendation actually applies to — and the patient in front of you may have been excluded.
- What was the endpoint? A treatment that improves a number on a monitor has not been shown to improve the outcome that matters unless that outcome was measured.