The incidence of difficult airways may be up to four times higher in the obstetric population due to various factors [6]. During pregnancy, increased swelling and vascularity in the airway can lead to poorer views and a higher risk of bleeding [7]. The cephalad displacement of the diaphragm by the gravid uterus decreases functional residual capacity, reducing apnoeic time. Most women will also experience reflux because of elevated intragastric pressures and progesterone-induced relaxation of the lower oesophageal sphincter, which puts them at high risk of aspiration during airway management; hence, they should be considered as unfasted [8]. Critical illness can present additional challenges, such as hypoxia from respiratory disease, difficulty achieving optimal positioning compared to an operating table, or situations where waking the patient in a cannot intubate scenario is not feasible.

Guidelines for managing the obstetric airway mainly focus on providing general anaesthesia for caesarean sections. The intensivist who needs to intubate a pregnant patient should be familiar with these guidelines to develop a safe airway plan. Given the increasing evidence that video laryngoscopy improves intubation success in both obstetrics and critical care, it should be used as the first choice when available [9]. Rapid sequence induction with cricoid pressure is recommended even for fasted patients to reduce the risk of aspiration [8]. In hypertensive conditions like pre-eclampsia, laryngoscopy may cause a further acute rise in blood pressure and intracranial pressure, increasing the risk of intracranial haemorrhage. The use of rapid-onset opiates (remifentanil or alfentanil) and intravenous antihypertensive medications (labetalol, esmolol) during anaesthetic induction can help reduce this response to laryngoscopy [10].

During pregnancy minute ventilation increases by 30–50%, driven by up to a 50% rise in tidal volume [11]. Respiratory rate increases to the upper limits of the normal range. These changes are mediated by progesterone, leading to respiratory alkalosis with an increase in arterial partial pressure of oxygen (pO₂) and a decrease in arterial partial pressure of carbon dioxide (pCO₂) and serum bicarbonate. As the uterus enlarges during pregnancy the diaphragm is displaced upward at rest. Relaxin-mediated effects on musculoskeletal tissue result in an increase in the ribcage’s capacity to expand laterally. Reductions in expiratory reserve volume and residual volume significantly decrease functional residual capacity by 20–30%. Most other lung volume measures, such as forced expiratory volumes, are not significantly affected. Oxygen demand at rest increases to approximately 20% above baseline by term [12].

Pregnant women may be admitted with respiratory failure caused by preexisting conditions such as asthma, interstitial lung disease, or cystic fibrosis; however, pregnancy also heightens the risk and severity of viral pneumonitis and thromboembolic disease. Pulmonary oedema may happen secondary to decompensation of cardiac disease or pre-eclampsia [13].

Blood delivered to the fetus via the umbilical vein has a lower oxygen saturation than the maternal arterial supply to the placenta; 80% vs 98% [14]. Maternal systemic hypoxaemia will therefore quickly lead to hypoxia and fetal distress. The fetus’s ability to extract oxygen from the blood depends on fetal cardiac output and a higher fetal concentration of fetal haemoglobin (HbF), which lies to the left of adult haemoglobin on the oxygen dissociation curve. Carbon dioxide diffuses freely at the placenta down a diffusion gradient, meaning maternal hypercarbia will cause fetal acidosis; this will reduce HbF’s oxygen-binding ability by shifting the curve to the right and also affect fetal oxygenation.

There is limited definitive data to guide ventilation targets in the pregnant population, but it is crucial to consider utero-placental blood flow and gas exchange. Guidelines recommend maintaining oxygen saturation above 95% and are mainly based on expert opinion [15]. Arterial pO₂ targets of 9.3 kPa (70 mmHg) are cited as necessary to ensure adequate fetal oxygenation, although these are derived from historical studies with small patient numbers [16, 17]. A lower arterial pO₂ of 8 kPa (60 mmHg) has also been suggested [18].

High-flow nasal oxygen (HFNO) and non-invasive ventilation (NIV) are recognised options for ventilatory support in pregnant patients with mild to moderate respiratory failure. The underlying cause of respiratory failure, the patient’s comorbidities, and the expected clinical course should guide the choice of initial therapy. NIV should be avoided in patients with impaired airway reflexes because of the increased risk of aspiration. HFNO provides limited positive pressure and may not achieve the same recruitment as NIV. However, HFNO is often better tolerated by patients. A sensible approach is to try HFNO and escalate to NIV if necessary in cases of mild respiratory failure, and vice versa for moderate respiratory failure if the patient improves with treatment.

When invasive mechanical ventilation is necessary, there is no high-quality evidence indicating that pregnant patients should be managed differently from non-pregnant patients. Ventilatory strategies are guided by case series and historical data in Acute Respiratory Distress Syndrome (ARDS); however, it is notable that there has been a significant growth in published literature on pregnant patients since the Coronavirus Disease 2019 (COVID-19) pandemic. We recommend limiting tidal volumes to 4–8 mL/kg of predicted body weight with plateau pressures below 30 cm H₂O [19]. Positive end-expiratory pressure (PEEP) should be titrated to optimise lung recruitment, recognising that venous return may already be reduced by the gravid uterus and that further PEEP adjustments might exacerbate this, leading to hypotension and reduced fetal blood flow.

The use of permissive hypercapnia and restrictive oxygen targets is controversial in pregnant patients because of their potential impact on the fetus. Hypocarbia reduces uterine blood flow, so over-ventilation should be avoided. Fetal hypercarbia may be linked to poor outcomes around delivery, but evidence on its effects beyond that is limited. Severe respiratory acidosis should be prevented; however, mild hypercarbia is likely safe when necessary for lung-protective ventilation [20]. During pregnancy, spontaneous minute ventilation increases, so higher respiratory rates are probably needed [21]. Overall, maternal hypercarbia and respiratory acidosis should be avoided where possible. Managing oxygenation to keep saturations above 95% is often challenging, and pO₂ targets may be similarly difficult in patients with severe respiratory failure. The threshold for using rescue methods typically reserved for refractory hypoxaemia or respiratory acidosis should be lowered. Continuous fetal monitoring is recommended whenever possible to detect early signs of distress.

Prone positioning in pregnancy appears to be a safe strategy, as supported by an increasing number of case reports, mainly from the COVID-19 pandemic [22]. Continuous fetal monitoring seems feasible even when in the prone position. The potential benefits of proning may be slightly diminished in pregnant patients due to the positively altered respiratory physiology in the supine position during normal pregnancy. An increased baseline cardiac output enhances apical perfusion, leading to improved ventilation-perfusion matching, while an increased anteroposterior diameter of the chest improves alveolar ventilation. Despite this, there is enough evidence to endorse its use, and a trial of prone positioning should be considered when indicated, with minor adaptations for the altered anatomy and physiology [23].

Inhaled nitric oxide (iNO) and prolonged neuromuscular blockade can be considered in cases of refractory hypoxaemia and severe ventilator desynchrony, respectively. iNO lacks a robust evidence base for reducing mortality [24], even in the non-pregnant population, and should only be regarded as a bridge to adjunctive therapies such as Extracorporeal Membrane Oxygenation (ECMO). Non-depolarising muscle relaxants are likely to cross the placenta in very small amounts, but the effects of prolonged use on the fetus remain unknown.

ECMO should be considered for pregnant patients experiencing refractory hypoxaemia or hypercapnia. Evidence from case series indicates that survival rates are favourable for both mother and fetus [25]. Early consultation with ECMO centres is recommended, as patients may benefit from tertiary intervention for acute respiratory failure without needing to proceed to ECMO.

A common topic of discussion in the case of the mechanically ventilated pregnant patient is the timing of delivery. Increased intrapleural pressure and cephalad shift of the diaphragm caused by the gravid uterus may be alleviated by delivering the fetus, although the improvement in respiratory dynamics may not necessarily lead to better outcomes [13, 26]. The haemodynamic effects of delivery, which are usually well tolerated by healthy mothers, can be poorly tolerated by patients with existing right ventricular dysfunction associated with severe respiratory failure.

Any decision regarding the timing of delivery should balance the potential benefits to the mother against the risks of iatrogenic prematurity to the fetus. In cases without standard maternal and fetal indications for delivery in pregnant patients, an individualised, multi-disciplinary team (MDT) approach; including obstetricians, anaesthetists, and intensivists, is recommended for deciding on operative delivery [13]. This may also involve seeking support from legal and/or ethical advice where appropriate.

During pregnancy, there is a 30–50% increase in cardiac output, mainly due to an increased stroke volume with a smaller rise in heart rate. Peripheral vascular resistance decreases in the first half of pregnancy and gradually returns to pre-pregnancy levels in the second half. Additionally, circulating volume increases while plasma colloid oncotic pressure decreases. The gravid uterus may compress the aorta and inferior vena cava (IVC) when in the supine position, which can significantly contribute to hypotension after 20 weeks of gestation. This is known as Supine Hypotensive Syndrome when it occurs in isolation.

IVC compression should be ruled out as a contributing factor in all shocked states. In conscious patients, the left lateral decubitus position is recommended to reduce compression of abdominal vessels and enhance uteroplacental blood flow. However, during critical illness this may not be feasible and a 15–30 degree lateral tilt might be necessary. It is important to recognise that there is considerable individual variation in the position and tilt angle needed to alleviate IVC compression which can significantly influence patient care [27].

Fluid resuscitation in pregnancy demands greater caution than in the general population due to the reduced oncotic pressure associated with physiological pregnancy. The 30 mL/kg recommended by the Surviving Sepsis campaign may result in a higher incidence of clinically significant oedema [28, 29]. Measuring serum lactate is a helpful indicator of perfusion. It is important to recognise that levels may be markedly elevated during and immediately after labour but should otherwise stay within the typical range.

Vasoactive drugs all impact the uteroplacental unit and fetus as the uterine artery responds to alpha-adrenergic stimulation. The most commonly used vasopressors in pregnancy are noradrenaline and phenylephrine. Noradrenaline maintains cardiac output more effectively due to beta receptor activity and has not been shown to harm the fetus, making it the preferred first-line vasopressor in critically ill obstetric patients [28].

Most other vasopressors have undesired effects and should be used with caution. Vasopressin may induce uterine contractions and should be avoided. Ephedrine has been shown to worsen fetal acidaemia in a dose-dependent manner around delivery. Adrenaline may cause arrhythmias in the fetus but can be used in emergency situations [30, 31]. Hydrocortisone is routinely used in vasopressor-refractory shock and may be used during pregnancy. Prednisolone and hydrocortisone are preferred steroids for maternal indications, as the placenta inactivates them in significant quantities [32].

There is limited evidence for the use of most inotropic drugs, although both dobutamine and levosimendan have been utilised to treat cardiogenic shock in pregnancy [33].

Veno-arterial ECMO should be considered in cases of refractory cardiogenic shock and pregnancy-specific conditions such as amniotic fluid embolism or pre-eclampsia with pulmonary oedema and haemodynamic instability [25].

Goal-directed fluid therapy is challenging as most non-invasive cardiac output monitors are not validated in pregnancy. Most products rely on (non-pregnant) demographic data to estimate haemodynamic parameters and may therefore be inaccurate. Lithium poses an additional risk to the fetus, especially in the first trimester, so lithium dilution techniques should be avoided [34]. Predicting preload responsiveness may be possible using certain measures from pulse contour analysis, such as stroke volume variation. Serial bedside echocardiography or oesophageal Doppler can both be used to guide resuscitation, but these are more labour-intensive than other techniques and exhibit inter-operator variability, particularly in inexperienced hands. Appropriate training and accreditation processes are therefore essential.

During pregnancy the glomerular filtration rate (GFR) increases by 50% leading to a reduction in creatinine and urea levels. Mild physiological hydronephrosis is common due to compression of the ureters by the gravid uterus, but it is usually asymptomatic and of little clinical significance [35].

Pregnancy-related acute kidney injury (Pr-AKI) can be caused by normal renal pathologies as well as pregnancy-specific causes such as pre-eclampsia (PET). Although acute renal replacement therapy (RRT) is rare during pregnancy intensivists must be aware of its key differences. Firstly, haemodialysis in pregnancy requires more frequent and longer sessions due to the physiological increase in GFR. When using haemofiltration, a higher effluent flow rate than in the general population may be necessary. Secondly, urea is toxic to the fetus, so a lower serum threshold of 17–20 mmol/L should be used to initiate RRT [36]. Beyond this, the thresholds for acute RRT are the same as in the non-pregnant population.

Hydronephrosis is usually managed conservatively but occasionally requires intervention in cases of pain, infection, or acute kidney injury due to obstructive uropathy [37].

None of the commonly used creatinine-based formulas to calculate estimated GFR and creatinine clearance is validated in pregnancy. These formulas have been shown to be unreliable due to hyperfiltration and volume expansion during pregnancy. Further research is needed to develop a pregnancy-specific formula for estimating renal function in pregnancy [38, 39].

The normal physiology of pregnancy leads to a hypercoagulable state, which protects against haemorrhage but raises the risk of venous thromboembolic disease. There is an upregulation of clotting factors and a reduction in anticoagulant and fibrinolytic substrates. Additionally, an increase in venous capacitance causes stasis. Platelet function also changes, resulting in greater platelet aggregation [40].

Thromboembolic disease is a major cause of maternal mortality and morbidity. Pregnant patients should receive prophylactic low molecular weight heparin (LMWH) during their ICU stay. The increased thrombotic risk extends beyond pregnancy and, although UK national guidance does not specify critical illness as a risk factor, prolonged prophylaxis for six weeks postpartum should be considered on an individual basis [41]. If deep vein thrombosis (DVT) or pulmonary embolus (PE) occurs, treatment-dose LMWH should be started.

Evidence regarding the treatment of high-risk (massive) and intermediate-high-risk (sub-massive) PE is limited to case series. Stratification of risk can be carried out similarly to the non-obstetric population, but it is important to remember fetal distress as an additional marker of end-organ failure. All reperfusion strategies should be considered, including systemic and catheter-directed thrombolysis. ECMO has been used in cases of refractory shock [42]. In a systematic review of more than 100 pregnant women with high-risk PE, of whom 83 received systemic thrombolysis, survival was 94%, although this favourable rate may reflect reporting bias. The risk of major bleeding was significantly higher (58%) in the postpartum period than during pregnancy (18%), and careful multidisciplinary decision-making is required to balance these risks against the dangers of deterioration or death from pulmonary embolism [43].

There is a lack of evidence regarding critical care nutrition during pregnancy, so exact requirements can be difficult to determine. There is no increase in calorie needs in the first trimester, but as the metabolic rate rises by about 15%, it is recommended to provide an additional 340 kcal/day in the second trimester and 452 kcal/day in the third trimester [54].

The protein content of feed should be increased during pregnancy and breastfeeding, and micronutrients such as folic acid, iron, calcium, and zinc may need to be supplemented [54]. Propofol can contribute significantly to calorie intake when used for continuous sedation and should be accounted for.

Delayed gastric emptying is common during pregnancy, as is constipation. These should be anticipated, and prokinetic and laxative drugs should be used as needed. Proton pump inhibitors are frequently prescribed during pregnancy and can be utilised to reduce aspiration risk and prevent ulcers. Parenteral nutrition has been effectively used in pregnancy, although it has been studied extensively only in cases of chronic intestinal failure [55].

Lactation support and expression of breastmilk should be available to patients in the ICU. When patient preferences are known, it may be in the patient’s best interest to initiate breastmilk expression while sedated [56]. It is important to remember that drugs given to the mother may affect the neonate, so consultation with a lactation pharmacist may be necessary.

There is no consensus on targets for blood glucose during critical illness in pregnancy. Intensive glucose control can be potentially harmful during critical illness, and UK best practice guidelines for general ICU patients support a more liberal target of 7.5–10 mmol/L [57]. However, in pregnancy (outside of critical illness), even mild hyperglycaemia is linked to worse outcomes, and National Institute for Health and Care Excellence (NICE) guidelines recommend maintaining glucose levels between 4 and 7.8 mmol/L [58, 59]. Therefore, it is reasonable to aim for the upper end of this range (i.e., 6–8 mmol/L) for critically ill pregnant patients.

The most common reason for maternal ICU admission in the UK is post-partum haemorrhage (PPH) [5]. There are no standard definitions of major obstetric haemorrhage (MOH), although thresholds of 500 mL in vaginal delivery and 1000 mL in caesarean are often used. Blood volume increases in pregnancy to about 100 mL/kg, but this shows significant variability between individuals so shock can occur at much lower volumes than expected [69]. The causes of PPH are usually classified as the “four Ts”: tone (uterine), trauma to the birth canal, tissue (retained products), or thrombin (coagulopathy).

Uterine atony is the most common cause of postpartum haemorrhage, and treatment involves a combination of pharmacological and surgical interventions. The involvement of the obstetric team is essential and management in theatre is often necessary. Intensivists should be familiar with the main classes of uterotonic drugs: oxytocin analogues, ergot alkaloids, and prostaglandin analogues. In cases of persistent uterine bleeding a Bakri balloon (a device used to achieve intra-uterine tamponade) may be employed and can remain in situ for up to 24 hours [70]. Interventional radiology techniques such as uterine artery embolisation may be required if bleeding is uncontrolled. ICU staff may be unfamiliar with palpating the uterus to confirm adequate contraction therefore serial measurement of symphysis-fundal height may provide an objective measure to support this.

Trauma to the birth canal or perineum can also cause significant blood loss. Treatment is mainly surgical but might include leaving vaginal packs in place. Retained packs have led to considerable morbidity; therefore it is advised that the presence of a vaginal pack be documented during patient handover. ICUs should be aware of local safety protocols [71]. Retained tissue may cause early or delayed bleeding. Small amounts of bleeding are normal within the first four days but ongoing bleeding, foul discharge, or infection with an unclear cause should prompt further examination [72].

Coagulopathy is rarely the primary cause of MOH but may complicate any of the above causes. Large volume red cell transfusions may have been administered resulting in a dilutional coagulopathy. Local approaches to MOH treatment will vary but it is essential to rule out fibrinogen deficiency as a cause of persistent bleeding. Measurement of fibrinogen through direct laboratory testing or viscoelastic testing can reliably identify this. During pregnancy fibrinogen levels are higher than in the general population, so a higher target of $>$2 g/L is used in resuscitation. Fibrinogen concentrate or cryoprecipitate, rather than fresh frozen plasma, should be used where available [70]. Tranexamic acid is currently recommended in all PPH [73].

Maternal sepsis remains one of the leading causes of maternal mortality [60]. It can be difficult to identify intrapartum, as increases in temperature, white cell count, and lactate may occur temporarily with normal vaginal delivery. There should be a low threshold for initiating antibiotic therapy and investigating for an underlying source of infection. Intra-uterine infection can occur at any point during the peripartum period. Most antibiotics are safe to use during pregnancy and breastfeeding.

Pre-eclampsia (PET) is a complex condition characterised by placental hypoperfusion/hypoxia and the subsequent release of antiangiogenic markers, which leads to widespread endothelial dysfunction, vasoconstriction, and immune dysregulation [74]. The condition is defined as hypertension exceeding 140/90 mmHg alongside signs of end-organ dysfunction after 20 weeks of pregnancy [75].

Severe features such as cardiac, respiratory, renal, hepatic, neurological, and haematological failure may all develop as a result of microvascular changes [76]. HELLP syndrome (haemolysis, elevated liver enzymes, and low platelets) is a recognised variant of PET reflecting this widespread organ dysfunction and can occur in the absence of hypertension.

Supportive management focuses on controlling blood pressure, which is vital to prevent cerebral bleeds. Commonly used antihypertensive drugs include labetalol (oral or intravenous), hydralazine (intravenous), and nifedipine (oral). Due to the risk of pulmonary oedema, patients with PET should be fluid restricted to 80 mL/hr, which can be challenging in an ICU setting. Spikes in blood pressure must be avoided, especially during invasive procedures or airway management. A frequently adopted strategy for induction of anaesthesia and intubation is to use a high dose, short-acting opioid to minimise the hypertensive response to stimuli such as laryngeal manipulation.

Eclampsia refers to new-onset seizures associated with hypertension during pregnancy. The mainstay of treatment for eclampsia is intravenous magnesium sulphate. Conventional anti-epileptic drugs are generally avoided; in fact, they are rarely needed, and prolonged seizures despite magnesium treatment should prompt consideration of an alternative diagnosis [77]. Definitive treatment involves delivering the placenta; however, PET may develop or persist up to 6 weeks postpartum.

Acute Fatty Liver of Pregnancy (AFLP) is a rare but serious condition. Immediate delivery of the fetus is crucial for survival and mortality is then around 4%. It presents with subacute abdominal discomfort and malaise but can progress to acute liver failure, jaundice, coagulopathy, encephalopathy, and hypoglycaemia. Although a liver biopsy showing microvesicular steatosis is required for a definitive diagnosis, it is rarely performed. Clinical suspicion and the use of the Swansea criteria scoring system are used to guide diagnosis and delivery [78].

Amniotic fluid embolism (AFE) is a rare condition but one with a high mortality rate. It is usually described during labour and delivery but may occur up to 6 hours postpartum. It is characterised by sudden cardiovascular collapse, hypoxaemic respiratory failure, and disseminated intravascular coagulation, often with altered mental status. The pathophysiology is not fully understood but is believed to be caused by amniotic fluid and fetal debris entering the maternal circulation. Evidence from animal models suggests this leads to severe acute pulmonary hypertension that causes cardiovascular collapse. Widespread activation of clotting factors may subsequently result in disseminated intravascular coagulation (DIC) and microvascular occlusion, leading to end-organ failure [79].

Management is supportive, focusing on maintaining cardiac output and oxygen delivery, as well as correcting coagulopathy. In the event of cardiac arrest, a perimortem caesarean section should be performed. Patients who survive the initial event will typically require intensive care for cardiorespiratory support. Mortality rates can be as high as 30%, and 9–15% of survivors sustain permanent neurological injuries [80].