Principles of mechanical ventilation

Mechanical ventilation supports or replaces spontaneous breathing. It is not a diagnosis and it is not a cure. It buys gas exchange, rest, and a protected airway while you treat the pneumonia, the shock, the overdose, the belly, the bleed.

Indications cluster: hypoxemic failure, hypercapnic failure / fatigue, airway protection, anticipated course (shock, OR, procedure), and impending arrest. Do not wait for a perfect ABG to intubate someone who is dying.

Brand names differ. Physiology does not. Every breath on every ventilator has four phases. If you can name which one is broken, you can fix asynchrony without memorizing a menu tree.

1. Trigger — what starts the breath: time (mandatory) or patient effort (pressure or flow trigger). Too sensitive → auto-trigger. Too insensitive or auto-PEEP → missed efforts.
2. Target / limit — what the machine is trying to deliver: a volume (VC) or a pressure (PC or PS/CPAP). This is the “control variable.”
3. Cycle — what ends inspiration: time (VC and PC, from the set Ti) or flow decay (PS/CPAP). Cycle too early → double-trigger. Cycle too late → delayed cycling, stacking.
4. Baseline — PEEP/CPAP, the pressure the lung sits on between breaths. Auto-PEEP is extra baseline the machine did not set.

The equation of motion is the whole subject in one line:

Paw = (Flow × Resistance) + (Volume / Compliance) + PEEP_total

PIP contains resistance and compliance. Pplat (no flow) is compliance only. That split is the most useful two seconds you will spend at the bedside.

A mode is a contract: the machine guarantees one variable and lets the lung decide the others. Pick the contract that matches the problem, then titrate the numbers.

When each is useful

• VC — reliable minute ventilation; you know the VT. Watch pressures. Flow is set by VT and Ti, so a long Ti leaves a hungry patient flow-starved.
• PC — variable flow meets demand; you cap alveolar pressure. VT will fall if the lung stiffens (ARDS, abdomen, PTX) — that is a feature if you notice it.
• PS/CPAP — weaning and comfort. No backup rate unless you add one. Apnea is possible. Delayed cycling is common with leaks.

A Servo-U also offers PRVC, volume support, SIMV with pressure support, Bi-Vent, and Automode. This lab uses VC, PC, and PS/CPAP — the three you need for these cases. PRVC still targets a tidal volume with a pressure the machine adjusts. It can hide a dying compliance until the volume collapses.

The hold manoeuvres

• Inspiratory hold — 0.5–2 s, no flow. Paw falls from PIP to Pplat. Gap = resistive pressure (tube, mucus, bronchospasm). Pplat itself = VT / Crs + PEEP.
• Expiratory hold — airway occluded at end-expiration. Paw rises from set PEEP to total PEEP. The difference is auto-PEEP.

Predicted body weight (PBW) from height/sex, not obesity weight. Male: 50 + 2.3 × (inches − 60). Female: 45.5 + 2.3 × (inches − 60). Start VT ≈ 6 mL/kg PBW in anyone with ARDS risk.

Pressure, flow, and volume against time. This is the ECG of the lung. Walk them in the same order every time:

1. Baseline — is PEEP where you set it? Is expiratory flow back to zero?
2. Trigger — time or patient? Extra breaths? Missed blips?
3. Inspiration — square vs decelerating flow? Scooped Paw? Overshoot?
4. Cycle — did inspiration end when the patient did?
5. Expiration — slow (obstruction)? Truncated (stacking)? Oscillating (heart / water)?

High resistance

PIP much higher than Pplat. Expiratory flow is slow and scooped. Think ETT obstruction, secretions, kink, bronchospasm. Suction and bronchodilate before you amputate VT.

Low compliance

PIP and Pplat both high. The lung or the chest wall is stiff: ARDS, edema, mainstem, pneumothorax, abdominal compartment, obesity, chest-wall burns. The hold tells you it is not the tube.

Flow starvation

Concave / scooped inspiratory Paw on VC. The patient is inhaling harder than the set flow. Raise flow or switch to a pressure mode. Paralysis without fixing flow is how you get awareness plus injury.

Loops are the same data in a different plane. Use them when the scalar has already made you suspicious.

• P–V loop — volume on x, pressure on y. Slope ≈ compliance. A beak at end-inspiration is overdistension: you are putting volume into a stiff lung. A wide hysteresis can mean recruitment or just airway resistance.
• F–V loop — volume on x, flow on y. Inspiratory limb up, expiratory limb down. A scooped expiratory limb is obstruction. A loop that does not close is a leak.

Asynchrony is a mismatch at trigger, flow, or cycle. It is associated with longer ventilation, diaphragm injury, and in ARDS with occult volutrauma. Diagnose it on the scalars. Treat the phase that is broken.

Ineffective (missed) triggering

Small negative deflections in Paw or flow during expiration that do not start a breath. The usual tax is auto-PEEP: the patient must first decompress trapped gas to reach the trigger threshold. Fix trapping, then apply PEEP toward measured auto-PEEP so the effort can land. Do not just make the trigger hair-trigger on a stacking COPD patient.

Auto-triggering

Extra breaths without a real effort. The liar is usually water in the circuit, a leak that depressurizes the trigger, or cardiogenic oscillations (big stroke volume, low drive, sensitive flow trigger — classic after cardiac surgery). Clue: deeply sedated, total rate far above set, oscillations on expiratory flow. Empty the water, desensitize, fix the leak. Do not paralyze a pulse.

Double triggering

The patient’s neural inspiration outlasts the mechanical Ti. The vent cycles off, the effort is still there, and a second breath stacks. You see a second flow pulse with no full expiration. Each stack is roughly a double VT. Lengthen Ti, give a bit more VT if it is too small, switch to PC, or deepen sedation if the lung is at risk.

Reverse triggering

The opposite story: a mandatory breath entrains a patient effort that was not the trigger. Effort starts during or just after inflation, often 1:1 with the machine, and may double-cycle. It is a reason “passive” ARDS patients still get VILI. Lengthening Ti does not fix this the way it fixes ordinary double-trigger. Confirm on the scalar, then sedation ± neuromuscular blockade if driving pressure is already high.

Auto-PEEP (intrinsic PEEP) is alveolar pressure still above set PEEP when the next breath starts. The flow scalar never returns to zero. An expiratory hold quantifies it.

Causes: obstruction (asthma, COPD, secretions), a rate that does not leave enough Te, a long Ti, a small ETT, delayed cycling. Consequences: missed triggers, stacked VT, barotrauma, and obstructive shock — high intrathoracic pressure, no venous return, falling MAP, rising CVP, PEA if you keep bagging.

• Lower the rate. Shorten Ti. Bronchodilate. Suction.
• Permissive hypercapnia is safer than stacking.
• If MAP is collapsing: disconnect, allow a full expiration, then reconnect on a slower vent. Pressors are a bridge.
• Applied PEEP ≈ 80% of measured auto-PEEP can help triggering. It does not empty the lung by itself.

Inspired VT > expired VT. Volume-time does not return to baseline. The F–V loop will not close. Then ask where.

A fistula wants the lowest mean airway pressure that keeps gas exchange survivable. Climbing PEEP increases the driving pressure across the hole. Independent-lung ventilation or a bronchial blocker is a later conversation; the first conversation is “stop feeding it.”

Sudden desat or high-pressure alarm after a stable stretch. Hands on the patient, eyes on the scalars, then:

If they are dying: disconnect, bag with 100% O₂, feel the compliance of the bag. A brick that empties when you wait is auto-PEEP. A brick that stays a brick is PTX, mainstem, or a stiff chest wall / abdomen.

The baby lung is small. 6 mL/kg PBW is the starting VT, not a trophy. If driving pressure is still high, go lower and raise RR. Pplat ≤ 30. ΔP generally < 15. PEEP opens the lung for oxygenation; it is not a personality.

• Wean FiO₂ toward ≤ 0.6 once SpO₂ is 92–96%. Hyperoxia is not a goal.
• Prone for moderate–severe ARDS (P/F ≲ 150) when the shop can do it safely.
• NMB for the first 48 hours if severe asynchrony or refractory hypoxemia.
• Permissive hypercapnia if pH is survivable (~≥ 7.20, context-dependent).
• Recruitment manoeuvres are not routine. They raise IAP and RV afterload.

The dangerous ventilator in obstructive disease is a fast one. These patients die of stacking and hypotension, not of a PaCO₂ of 70.

• Low RR (10–14), modest VT (6–8 mL/kg), short Ti so Te is long (I:E 1:3–1:5).
• Bronchodilators, steroids, treat the trigger. Heliox is a footnote.
• Watch flow to zero. Hold for auto-PEEP. Watch MAP.
• Arresting? Disconnect. Empty. Then a slow vent, not a faster one.

The ventilator does not know the abdomen exists. It only sees a stiff respiratory system. Abdominal compartment syndrome (ACS) is IAP > 20 mmHg with new organ failure — typically oliguria, rising pressors, and high PIP and Pplat. The chest wall is the problem, and the abdomen is the chest wall.

• Measure bladder pressure in any tight abdomen with high Pplat and dying kidneys. Physical exam alone is a poor test.
• Massive resuscitation, pancreatitis, trauma, tight fascia, ascites, and high PEEP / prone all raise IAP.
• Medical adjuncts: NMB, NG decompression, drain ascites, limit further volume, modest (not heroic) PEEP.
• Definitive: decompressive laparotomy (or large-volume paracentesis if the contents are liquid). Pplat often falls for free once the chest wall can move.
• Do not recruit these patients. Recruitment and high PEEP transmit into the abdomen and can make ACS worse. That is the trap in the advanced cases.

Overdistended alveoli compress extra-alveolar vessels and raise pulmonary vascular resistance. Hypoxia and acidosis do the same. The ARDS RV is fragile. A beautiful SpO₂ with a falling MAP, rising CVP, and a D-sign is not a victory.

Best PEEP is the one that opens the lung without emptying the LV. Echo before you climb. Ease PEEP, keep VT small, fix the blood gas, support the RV (ino / inhaled pulmonary vasodilator). ECMO is the next platform, not another 4 cm of PEEP into a D-sign.

These cases are rare and the delay is what kills. The ventilator often speaks first.

Malignant hyperthermia is a ryanodine-receptor (RYR1) crisis: uncontrolled sarcoplasmic-reticulum calcium release, muscle rigidity, a furnace of CO₂ and heat, mixed acidosis, hyperkalemia, rhabdomyolysis. Triggers: succinylcholine and volatile anesthetics (sevoflurane, desflurane, isoflurane). Not propofol, not rocuronium, not opioids.

Two doorways into the ICU

• Post-RSI with succinylcholine — masseter rigidity, then a rigid chest, EtCO₂ that will not fall, fever that lags the CO₂.
• Post-op / PACU after volatiles — they leave the OR still hot, still producing CO₂, sometimes still on a circuit that smells like gas.

What the vent shows

• EtCO₂ and PaCO₂ rising despite a high minute ventilation. That is production, not dead space (PE would raise the PaCO₂–EtCO₂ gap).
• Mixed acidosis: respiratory (CO₂) plus metabolic (lactate). HCO₃ falls.
• High PIP and Pplat from rigidity — a stiff chest wall, not ARDS. NMB will not relax it. Dantrolene will.
• Tachycardia, rising core temperature, rising K⁺, cola urine later.

What you actually do

1. Call for help. In real life: MH hotline 1-800-MH-HYPER (1-800-644-9737).
2. Stop the trigger. Volatiles off. Fresh circuit or charcoal filter. High-flow 100% O₂.
3. Dantrolene 2.5 mg/kg IV, repeat until EtCO₂, rigidity, and temperature break.
4. Hyperventilate. Cool (stop ~38 °C). Treat hyperkalemia (calcium, insulin/glucose — not a CCB).
5. Do not give more succinylcholine. Do not give a calcium-channel blocker with dantrolene.

Differential that fools people: inadequate ventilation, thyroid storm, NMS (slow, neuroleptics, not succs), sepsis, CO₂ insufflation. The time course plus the trigger plus EtCO₂ outrunning VE is MH.

Same language as the other labs. Read the handover. Look at the scalars before you touch a slider. Holds are free information. Harmful clicks cost score; preferred ones pay a little back.

1. Name the mode and the three scalars. Is expiratory flow at zero?
2. Split PIP vs Pplat with an inspiratory hold if pressures are high.
3. If the rate is “wrong,” decide: double-trigger, auto-trigger, reverse trigger, or missed efforts.
4. If they are crashing: DOPE, then dump auto-PEEP or drain a tension, then titrate.
5. If the lung is stiff and the belly is tight: measure IAP before you recruit.
6. If EtCO₂ outruns VE after succs or volatiles: that is MH. Dantrolene, not another NMB.

Guided mode asks three questions (assess → intervene → synthesize) and then walks the preferred path so you can watch the traces change. Free play leaves you with the console. Uncorrected hypoxemia, auto-PEEP shock, tension, ACS, or untreated MH progresses to pre-arrest and a fail. MH cases sit with the other advanced scenarios.