A perfect Cormack-Lehane view does not save a patient whose physiology arrests during the pause
The Mallampati score predicts a hard laryngoscopy. It says nothing about what happens in the sixty seconds after you push the drugs.
Every difficult airway algorithm most of us trained on is built around anatomy: mouth opening, neck mobility, thyromental distance, Mallampati class. That framework answers one question: will I see the cords? It says nothing about a second, independent question: will this patient's circulation and gas exchange survive the apneic period and the drugs I'm about to give? Four physiologic states answer that second question badly, and they kill patients with grade 1 views just as reliably as a genuinely difficult anatomic airway kills patients with a hidden epiglottis.
Shunt physiology means there is no reserve left to spend during the apneic pause.
Pneumonia, ARDS, and cardiogenic edema all produce true shunt: blood passing through unventilated lung. Shunt does not respond to supplemental oxygen the way dead space or V/Q mismatch does, which is why these patients can look stable on high-flow oxygen and then crater within seconds of losing their respiratory effort.
The PREOXI trial randomized 1,301 critically ill adults to noninvasive ventilation versus an oxygen mask for preoxygenation before intubation and found severe hypoxemia (SpO2 <85%) in 9.1% of the NIV group versus 18.5% of the oxygen mask group.[1] NIV or tight-fitting BVM with PEEP, not a nonrebreather alone, is the evidence-based default when SpO2 won't climb toward 95%.
Nasal cannula apneic oxygenation is low-risk and reasonable to leave running, but the FELLOW trial found no difference in lowest SpO2 with apneic oxygenation when preoxygenation was already adequate.[2] The original benefit signal came from operating room and prehospital series, not from ICU patients preoxygenated with NIV or BVM. The graphic's own caveat, that ApOx may add little on top of good NIV preoxygenation, tracks the trial data more accurately than the framing above it in the same carousel.
Positive pressure plus induction agents remove sympathetic drive and venous return at the same moment. An already-failing circulation has nothing left to compensate with.
Shock index (HR/SBP) trends higher in patients who arrest peri-intubation, but the correlation is weaker than a single threshold suggests. In the NEAR registry case-control data, mean shock index was 0.97 in arrest cases versus 0.83 in matched controls, a real but modest separation, and other cohorts have called shock index in isolation a poor discriminator on its own.[4,5] Modified shock index (HR / MAP) ≥1.3 and reverse shock index <1 have shown stronger associations with peri-intubation arrest in more recent case-control work.[6,7] Treat >0.8 as a flag to look harder at the patient, not as a validated decision threshold on its own.
PrePARE (2019) and its confirmatory multicenter trial PREPARE II (2022, JAMA, 337 patients) both found that a routine 500 mL crystalloid bolus before intubation did not reduce the composite rate of cardiovascular collapse.[8,9] Push-dose pressor on the pole and vasopressor readiness before induction, not a reflexive bolus, is what the trial data actually supports.
The RSI trial, published in the New England Journal of Medicine December 9, 2025, randomized 2,365 critically ill adults across 14 US EDs and ICUs to ketamine or etomidate for induction.[10] Ketamine showed no mortality advantage over etomidate and was associated with more cardiovascular collapse during intubation. A companion 2025 meta-analysis of RCTs reached the same direction of effect (RR 1.29 for hemodynamic instability with ketamine vs etomidate).[11] The carousel's framing, that neither agent is safe in shock and dose reduction plus pressor readiness matters more than the drug choice, holds up. The claim that ketamine is uniformly more "cardiostable" than etomidate does not, and the newest large trial argues the opposite direction for this specific outcome.
A compensated acidosis is a patient breathing at a rate that keeps them alive. Take that rate away and the compensation goes with it.
DKA, severe lactic acidosis, and salicylate toxicity share one mechanism: a falling pH held up entirely by a high minute ventilation blowing off CO2. The moment a paralytic stops that ventilation, CO2 rises uncontested, pH falls further on top of the metabolic acidosis already present, and the patient can arrest from acidemia alone, independent of oxygenation. Salicylate poisoning is the most dramatic version: apnea in a salicylate-toxic patient can precipitate cardiovascular collapse within minutes as unbound salicylate redistributes into the CNS with falling pH.[12]
Match the ventilator to the patient's own pre-intubation minute ventilation and respiratory rate immediately after tube placement, not a default 12-16. Under-ventilating a compensating acidotic patient recreates the exact physiology that just killed the last one. Ketamine is favored here less for a proven mortality benefit and more because it preserves spontaneous respiratory drive and hemodynamics through the induction window better than propofol, buying time to correct pH before or during the apneic period where possible.[3]
A dilated, pressure-loaded right ventricle depends on coronary perfusion pressure and low pulmonary vascular resistance to keep beating. Both are targets you can protect or destroy with your choice of pressor.
Pulmonary hypertension, massive PE, and RV infarct all produce a thin, pressure-sensitive RV living close to its afterload limit. Positive pressure ventilation drops preload and can raise PVR in the same breath (see our companion piece on heart-lung interactions in ARDS for the full mechanism), while hypoxia, hypercapnia, and acidosis each independently vasoconstrict the pulmonary bed and load the RV further.
Phenylephrine is a pure alpha-1 agonist: it raises systemic vascular resistance and MAP, but alpha-1 receptors are also present in the pulmonary vasculature, and phenylephrine raises PVR while doing it, worsening RV afterload at the exact moment the RV can least afford it.[13] Norepinephrine raises MAP and RV coronary perfusion pressure with comparatively less pulmonary vasoconstrictive effect and is the more commonly favored first agent in acute cor pulmonale. Vasopressin at low dose raises systemic vascular resistance through V1 receptors while having a neutral to mildly favorable effect on the pulmonary circulation, making it a reasonable adjunct when norepinephrine alone isn't holding MAP.[14]
The spiral is self-reinforcing: positive pressure and induction drop preload and MAP, hypotension drops RV coronary perfusion pressure below the elevated RV wall tension needed to perfuse it, RV ischemia drops contractility further, and any transient hypoxia or hypercapnia during the apneic window adds a PVR spike on top. Each piece of the spiral makes the next piece worse. Breaking it requires treating preload, contractility, and PVR simultaneously rather than sequentially.[15]
| Profile | Induction approach | Key move | Evidence caveat |
|---|---|---|---|
| Hypoxaemia | Ketamine (DSI when preoxygenation is limited by agitation/intolerance) | NIV or tight BVM+PEEP preoxygenation to SpO2 ≥95% before pushing drugs | ApOx via nasal cannula is low-risk but adds little once NIV preoxygenation is adequate[2] |
| Hypotension | Reduced-dose ketamine or etomidate | Pressor ready before induction; resuscitate with pressor, not reflex fluid | Routine fluid bolus does not reduce collapse[8,9]; ketamine no longer the clearly "safer" agent[10] |
| Metabolic acidosis | Ketamine | Set post-intubation ventilator rate to the patient's pre-intubation minute ventilation | Under-ventilation, not the drug choice, is the lethal error |
| RV failure | Ketamine or etomidate, low airway pressure technique | Norepinephrine ± vasopressin before induction; avoid phenylephrine; avoid hypoxia/hypercapnia/acidosis | Phenylephrine's alpha-1 effect extends to the pulmonary bed and raises RV afterload[13] |
Rocuronium 1.2 mg/kg for first-pass success holds across all four profiles and is well supported independent of hemodynamic status.[16] Dose reduction and pressor readiness belong to the induction agent regardless of which one is chosen; no agent is hemodynamically free in true shock.
Two claims in the original carousel were softened here against the primary data: the shock index >0.8 threshold is a real association but a weak standalone predictor, and the framing of ketamine as the more "cardiostable" agent is now contradicted by the largest RCT on the question (RSI trial, NEJM 2025). The apneic oxygenation and RV pressor-selection points held up well against the trial and physiology literature.