M4Respiratory support, ventilation, ARDS, proning, VAP and ECMO

Clinical overview

Respiratory failure in pregnancy becomes dangerous quickly because oxygen consumption is increased, functional residual capacity is reduced and the fetus depends on maternal oxygen delivery. The causes remain broad: pneumonia, aspiration, pulmonary oedema from pre-eclampsia or cardiac disease, ARDS, asthma, pulmonary embolism, sepsis, transfusion-associated lung injury and iatrogenic fluid overload. The first task is to distinguish oxygenation failure from ventilatory failure and excessive work of breathing; the second is to support gas exchange without delaying treatment of the cause.

Pregnancy changes interpretation, not the obligation to provide evidence-based organ support. Non-invasive support, lung-protective ventilation, prone positioning and referral for ECMO are feasible when indicated. None should be withheld solely because the patient is pregnant. Fetal monitoring and delivery planning are integrated around effective maternal support; they must not prevent safe ventilation.

Core knowledge

Define the respiratory problem

Hypoxaemic failure is failure to oxygenate despite supplemental oxygen. It commonly reflects shunt or ventilation–perfusion mismatch from oedema, pneumonia, ARDS or embolic disease. Ventilatory failure is inadequate removal of CO₂ because minute ventilation is insufficient; fatigue, reduced consciousness, neuromuscular weakness, obstruction and sedative toxicity are typical mechanisms. Work-of-breathing failure precedes both when the patient can temporarily maintain gas values only through unsustainable effort.

In pregnancy, PaCO₂ is normally lower than outside pregnancy. A rising PaCO₂ toward the nonpregnant reference range in a distressed patient can signify fatigue. PaO₂ and saturation must be interpreted against inspired oxygen, altitude, haemoglobin, cardiac output and trajectory. A reassuring saturation on high inspired oxygen can coexist with severe gas-exchange failure.

First principles: why pregnancy changes the escalation decision

Every recommendation below follows from three mechanisms, not from a separate obstetric rulebook. Reason from these and you can decide about a modality nobody has studied in pregnancy.

The oxygen store is small and the demand is high. Functional residual capacity falls 10–25% by term as the diaphragm rises, while oxygen consumption rises 20–33%. FRC is the reserve a patient lives off during any interruption — an apnoea, a failed non-invasive trial, a disconnection. So the same episode that a non-pregnant patient tolerates for minutes desaturates her in seconds. Consequence: the window in which a failing trial can be converted to a controlled airway is genuinely shorter, and a skilled intubator belongs at the bedside before you start, not after it fails.

The airway is oedematous and the stomach is not empty. Mucosal engorgement narrows the glottis and worsens the view; progesterone lowers oesophageal sphincter tone and the gravid uterus raises intragastric pressure. Consequence: bilevel non-invasive ventilation — which insufflates the stomach — carries a higher aspiration cost here than in general practice, and a failed intubation is more likely to be a difficult one. That is why the module treats non-invasive support as a monitored trial with a defined stop, not a destination.

There are two patients, and the fetal one is downstream of maternal gas tensions. Fetal oxygenation depends on maternal PaO₂ and uterine blood flow; fetal CO₂ clears only down a maternal–fetal gradient. Consequence: the maternal oxygenation target is tighter than the permissive one used in non-pregnant ARDS, and delivery is not a respiratory treatment — it changes mechanics, but a lung that cannot be oxygenated still cannot be oxygenated afterwards.

Where the anatomy is a red herring. The instinct that a pregnant abdomen makes proning unsafe is wrong, and the evidence is now reasonable: proning is technically feasible and well tolerated in the third trimester, series of ventilated pregnant patients report prolonged sessions with maternal and fetal tolerance, and oxygenation gains match those in non-pregnant ARDS. What matters is how — thoracic and pelvic supports that leave the gravid abdomen hanging free, so the mechanism you are exploiting (recruiting dependent lung) is not undone by compressing the diaphragm from below.

Where the reasoning runs out. ECMO in pregnancy has no trial evidence, only accumulating series: a 2026 systematic review of mostly veno-venous runs reports maternal survival from 33% to 100%, with the larger and more recent series above 80%, and bleeding, thrombosis, infection and acute kidney injury as the main complications. That is enough to say pregnancy is not a barrier to referral, and not enough to define thresholds — so refer early to a centre rather than waiting for a number.

Escalating non-invasive support

Use the least invasive support that achieves the goal without allowing deterioration:

• Conventional oxygen suits a stable modest requirement.
• High-flow nasal oxygen provides heated humidified high flow, reduces entrainment of room air and may reduce work of breathing.
• CPAP recruits alveoli and is particularly useful in cardiogenic or pre-eclamptic pulmonary oedema when the patient can protect her airway.
• Bilevel non-invasive ventilation provides inspiratory assistance and may help ventilatory failure, but pregnancy-related aspiration risk and rapid deterioration demand close supervision.