The planner tells you what to give and how much. This chapter is about why — why the induction agent matters far less than the dose you give it in, why "haemodynamically stable" agents are not a thing, why the paralytic is the one drug you should never reduce, and which of the drugs we were all taught to give before induction have quietly been abandoned. It is written to be read off-shift, not during a crash intubation.
A 64-year-old with community-acquired pneumonia. RR 38, SpO₂ 86% on 15 L, BP 82/48, HR 124, lactate 5.1. The registrar draws up propofol 200 mg and rocuronium 50 mg, says "she's only 70 kg", gives the propofol over five seconds, and asks for cricoid. Two minutes after the tube goes in the blood pressure is unrecordable.
By the end of this chapter you should be able to name at least five things that went wrong before the laryngoscope was ever picked up.
Drug choice is downstream of physiology. Answer these before you open a drawer.
1. What will the induction dose do to this patient's blood pressure? Every induction agent abolishes sympathetic tone to some degree, and positive-pressure ventilation then drops preload. In a patient who is only upright because of their catecholamines, that is the lethal step — not the intubation. Peri-intubation cardiovascular collapse, not failed laryngoscopy, is the commonest serious adverse event when critically ill adults are intubated. Observational
2. How badly do I need this patient not to move, and for how long? This is the paralytic question, and it is separate from the first. Full-dose neuromuscular block gives the best first-pass conditions; the argument for a short-acting agent is almost always an argument about a plan you should not be relying on anyway.
3. What happens in the ten minutes after the tube? The drugs that got the tube in are wearing off while the paralytic is not. If you have not planned sedation, analgesia and vasopressor before induction, you have planned for an awake, paralysed, hypotensive patient.
There is no haemodynamically neutral induction agent — there are only agents that are kinder about it, and doses that are small enough. A shocked patient given "the right drug" at a textbook dose will still arrest. A shocked patient given the "wrong" drug at a quarter of the dose, with noradrenaline already running, usually will not.
Onset is near-identical (one arm–brain circulation). What differs is the effect on vascular tone, myocardial contractility and the brain.
| Agent | Adult induction dose | What it does to the circulation | Where it earns its place | Where it hurts you |
|---|---|---|---|---|
| Ketamine | 1.5–2 mg/kg IV 0.5–1 mg/kg if shocked |
Indirect sympathomimetic — usually raises HR and BP. Direct myocardial depressant, which is unmasked when the catecholamine reserve is already spent. | Shock of any cause, trauma, sepsis, severe bronchospasm (it is a bronchodilator), and the patient you cannot afford to apnoeise. | Catecholamine-depleted patients — full dose can cause the collapse it is famous for preventing. Emergence phenomena if sedation is delayed. |
| Propofol | 1–2 mg/kg IV 0.5–1 mg/kg if unstable |
Venodilation, arterial dilation and negative inotropy. The most predictable blood-pressure drop of the four, and dose-dependent. | The haemodynamically well patient, status epilepticus, and anywhere a smooth, deep, reliable plane matters. | Shock. The 2 mg/kg reflex dose is the commonest avoidable cause of post-induction arrest in the ED. |
| Thiopentone | 3–5 mg/kg IV | Marked venodilation and myocardial depression — as bad as or worse than propofol in the unstable. | Raised ICP in a patient with a blood pressure to spare; refractory status. Reliable, fast, deep. | Any instability; porphyria. Tissue necrosis on extravasation. Needs reconstitution, which is a delay nobody rehearses. |
| Etomidate | 0.3 mg/kg IV | The least cardiovascular disturbance of the four at induction dose. This is real and it is the whole reason the drug persists. | Fixed cardiac output states — severe AS, cardiogenic shock, the patient you do not trust to tolerate any vasodilation. | A single dose blocks cortisol synthesis (11-β-hydroxylase) — the UK SmPC puts the suppression at up to about 8 hours after one 0.3 mg/kg dose, and it is unresponsive to ACTH. Label DAS says the drug "is not routinely used in critically ill patients". Guideline |
Randomised comparison of ketamine against etomidate for emergency intubation of critically ill adults found no difference in the primary organ-dysfunction outcome, with more biochemical adrenal insufficiency in the etomidate arm. Trial The honest reading is not "ketamine wins" — it is that among reasonable agents given at a reasonable dose, which agent is a smaller lever than how much and what else you did first.
Midazolam appears in some older RSI recipes. It is a poor induction agent: onset is slow and dose-dependent, the dose required for reliable amnesia at induction (0.2–0.3 mg/kg) is hypotensive, and it is frequently underdosed into an awareness event. If it is all you have, dose it properly and expect to support the pressure; otherwise it belongs in the post-intubation infusion, not the induction syringe. Teaching
If you remember one sentence from this page, make it that one.
The two drugs behave in opposite directions in shock. Low cardiac output means the induction agent reaches the brain in a smaller volume of distribution and over a slower circulation time — so the same milligrams per kilogram produce a deeper, more prolonged and more hypotensive effect. Reducing the dose to a half or a quarter is not cutting corners; it is correcting for the physiology.
The neuromuscular blocker has the opposite problem. Slow circulation delays its arrival at the neuromuscular junction, so onset is later in exactly the patient who has the least apnoea reserve. Under-dosing it buys a poor view, a coughing patient, and a second attempt. Give the full dose, or more.
In shock: induction agent down (often to 25–50% of the textbook dose), paralytic up (rocuronium at the full 1.2 mg/kg, or 1.6 mg/kg where onset matters most), vasopressor running before you push anything.
Two practical consequences. First, weight matters less than you think for the induction agent and more than you think for the paralytic — and rocuronium for RSI is dosed on actual body weight, not ideal or lean Teaching — a "70 kg" guess that is 20 kg out will under-paralyse a large patient while the reduced induction dose remains safe. Second, give the induction agent slowly enough to titrate in the unstable patient (over 30–60 seconds, watching), and then give the paralytic as a fast push with a flush behind it.
A choice that is now mostly settled in UK emergency practice, for reasons worth understanding.
| Rocuronium | Suxamethonium | |
|---|---|---|
| RSI dose | 1.2 mg/kg (1.6 mg/kg for fastest onset) | 1.5 mg/kg adult; 2 mg/kg child; 3 mg/kg infant |
| Onset | ~60 s at 1.2 mg/kg | 30–60 s |
| Duration | Long — well over an hour at RSI dose | 6–10 minutes |
| Reversal | Sugammadex 16 mg/kg for immediate reversal Label | None — you wait |
| Contraindications | Allergy only | A long list, below |
| Side effects | Essentially none at induction | ↑K⁺ ~0.5 mmol/L, bradycardia (especially repeat doses and children), fasciculation, myalgia, malignant hyperthermia trigger, ↑ICP/IOP transiently |
At 1.2 mg/kg rocuronium produces intubating conditions comparable to suxamethonium at 60 seconds, without any of the right-hand column. Head-to-head evidence, taken across all doses, still gives suxamethonium a marginal edge in intubating conditions — but that comparison is dominated by lower rocuronium doses, and the marginal edge is not worth the contraindication list in an emergency department. Trial Guideline
Given 3 minutes after 1.2 mg/kg rocuronium, 16 mg/kg sugammadex restores the train-of-four ratio to 0.9 in a median of about 1.5 minutes Label — but the induction agent is unaffected, so you get a patient with a patent neuromuscular junction and no respiratory drive. DAS is blunt about this: reversal "does not guarantee clinical improvement", its "significant preparation time impairs its utility", and reversing an NMBA "does not reverse mechanical or physiological causes of airway obstruction, does not reliably prevent, and may precipitate, CICO". Guideline It is part of the plan, it must be drawn up and its location known before induction, and it does not replace the front-of-neck decision. It also does not touch suxamethonium or the benzylisoquinoliniums.
Most of this list has been retired. Knowing why stops it coming back.
Blunts the sympathetic response to laryngoscopy. That response is a problem in exactly three situations: raised ICP, aortic dissection or a fragile aneurysm, and severe hypertensive cardiac disease. It is a liability in the shocked patient, where the same sympathetic tone is the only thing holding the pressure up. Given for effect it needs to be in three minutes before, which is often three minutes you do not have.
Given for decades to attenuate the ICP response to laryngoscopy. The evidence never supported a patient-centred benefit, and it is no longer recommended. Teaching
Used to blunt suxamethonium fasciculation and the associated potassium rise. It delays the block, complicates the sequence, and in a department that has moved to rocuronium the question does not arise.
Reserve it for anticipated bradycardia: suxamethonium in infants or a repeat dose, or a strongly vagotonic procedure. The old teaching of a 0.1 mg "minimum dose" to avoid paradoxical bradycardia is not supported by evidence, and in an infant under 5 kg it is an overdose — though note some paediatric resuscitation courses still teach it. Teaching
The drug choice that most often saves the patient is the one hanging on the drip stand.
Post-induction hypotension is common, predictable and predictive of death. The patients who get it are the ones you can identify in advance: a raised shock index (HR ÷ systolic BP — around 0.8 is the best cut-point for post-intubation hypotension, and ≥0.9 is the value associated with peri-intubation cardiac arrest), a rising lactate, a patient already on pressors, and the patient whose "normal" blood pressure is being held up by a heart rate of 130. Observational
There are two ways to have a vasopressor available, and they answer different questions.
Fluid is not automatically the answer. PrePARE randomised a 500 mL bolus against none before intubation and was stopped for futility — cardiovascular collapse in 20% versus 18%. Trial Note this cuts across the older UK guideline, which suggests that "in the absence of cardiac failure, rapid infusion of 500 ml crystalloid before or during intubation may mitigate hypotension" Guideline — and note that the guideline's own caveat is the point. Give fluid or blood for a reason — haemorrhage, true hypovolaemia, a fluid-responsive septic patient — not as a reflex, and not into a failing ventricle or a wet chest.
The planner applies these automatically once you set the indication and concerns in Steps 1–2. Here is the reasoning behind each.
| Physiology | Induction | Paralysis | The thing that actually matters |
|---|---|---|---|
| Haemorrhagic shock | Ketamine 0.5–1 mg/kg, slowly | Rocuronium 1.2–1.6 mg/kg | Blood running before induction. The induction dose is a quarter of the textbook. Expect a further drop when positive pressure starts. |
| Septic shock | Ketamine, reduced; avoid etomidate | Rocuronium | Noradrenaline running and titrated first. Post-intubation hypotension here is the rule, not the exception. |
| Traumatic brain injury | Ketamine or propofol; thiopentone if the pressure allows | Rocuronium (suxamethonium contraindicated beyond ~72 h from injury) | Avoid hypotension and hypoxia — either one worsens secondary injury. Aim for a normal PaCO₂; hyperventilation is a short-term rescue for herniation, not a setting. |
| Severe asthma / COPD | Ketamine 1.5–2 mg/kg (bronchodilator) | Rocuronium | After the tube: low rate, long expiratory time, low PEEP. Disconnect and decompress the chest if the pressure drops — think gas trapping and tension pneumothorax before you think drugs. |
| Cardiogenic shock / fixed output | Etomidate, or ketamine at a reduced dose | Rocuronium | Avoid propofol and thiopentone. These patients tolerate vasodilation worst of all. |
| Severe metabolic acidosis (DKA, salicylate, sepsis) | Ketamine, reduced | Rocuronium, full dose, fast | The apnoeic period removes the respiratory compensation that is keeping the pH survivable. Minimise apnoea time and match their minute ventilation on the ventilator, not a textbook one. |
| Status epilepticus | Propofol or thiopentone (both anticonvulsant) | Rocuronium | Paralysis stops the movement, not the seizure. Arrange EEG or continue anticonvulsant treatment — otherwise you have hidden it. |
| Hyperkalaemia | Any, dosed for the circulation | Rocuronium — suxamethonium is contraindicated | Treat the potassium in parallel. Do not let the RSI displace calcium, insulin–dextrose and a dialysis referral. |
Rocuronium at 1.2 mg/kg outlasts every induction agent by a wide margin. The gap between the induction agent wearing off and the sedation infusion running at an effective rate is where accidental awareness happens, and emergency airway management is among the higher-risk settings for it. Observational
Step 6 of the planner will calculate all of these in mL/h against the bag your department stocks, and Step 7 prints the lot.
Based on small, old studies in spontaneously breathing patients with obstructive hydrocephalus. In a controlled, ventilated RSI, current evidence does not support avoiding it — and the hypotension caused by the alternative is a documented harm in TBI. Caution still applies with significant mass effect.
The induction agent outlasts it, and a shunting patient desaturates faster than the block wears off. This reasoning substitutes a drug for a failed-airway plan.
The adrenal suppression is real but time-limited (up to ~8 h per the UK label), and the mortality signal has never been clean — KETASED found no difference in organ dysfunction. The UK position is narrower than a ban and stronger than a shrug: DAS says etomidate "is not routinely used in critically ill patients". Treat it as a drug you need a specific reason to reach for, such as a fixed cardiac output state, not one you have been forbidden.
Neither. DAS 2018 still recommends it — aspiration risk "may be reduced … whilst still debated, by cricoid force application by a trained assistant", at 10 N awake rising to 30 N once unconscious. What is not negotiable is letting go: reduce or remove it "if there is difficulty with laryngoscopy, passage of the tracheal tube, facemask ventilation or active vomiting", and always before inserting a supraglottic airway. Guideline
PreVent randomised it: severe hypoxaemia 10.9% with bag-mask ventilation versus 22.8% without, and operator-reported aspiration 2.5% versus 4.0% (p=0.41). Trial DAS goes further and recommends facemask ventilation with CPAP before intubation and between attempts where hypoxia is likely. Guideline
A routine pre-induction bolus has not been shown to prevent cardiovascular collapse, and it is harmful in the patient with a wet chest or a failing ventricle. Give fluid for an indication.
None of these is a laryngoscopy skill. All of them are decisions made while the drugs were being drawn up.
Doses and cautions on this page mirror the planner and are drawn from UK product literature and national airway guidance. The evidence tags above indicate the kind of source, not its strength.