Manage maternal collapse and perform peri-mortem (resuscitative) caesarean section

In one line

Maternal collapse is the acute loss of circulation, consciousness or both in a pregnant or recently delivered woman, and after 20 weeks its management diverges from standard adult life support in two specific ways: the uterus must be displaced off the great vessels by continuous manual left displacement throughout chest compressions, and if there is no return of spontaneous circulation a resuscitative hysterotomy (peri-mortem caesarean) is performed where she lies — the decision taken by four minutes, aiming to deliver by five — primarily to save the mother, not the fetus.

Mechanism & pathophysiology

The whole of this topic falls out of one anatomical fact: from about 20 weeks the gravid uterus is large enough to compress the inferior vena cava and the aorta when the woman is supine. Caval compression cuts venous return; with a beating heart in shock this drops cardiac output by up to a third, and in cardiac arrest it is catastrophic. Closed-chest compression generates forward flow only by squeezing blood that is already in the thorax. If venous return is throttled by an obstructed cava, the right heart is poorly filled between compressions, so even technically perfect CPR produces a fraction of the stroke volume it would in a non-pregnant adult. Aortic compression compounds the problem by raising afterload against an already failing output. The supine pregnant woman in arrest is therefore being resuscitated against a mechanical obstruction that no drug or shock can overcome.

This is why uterine displacement is not an optional refinement but part of generating any output at all. Tilting the whole woman to the left was the historical solution, but lateral tilt steeply degrades the quality of chest compressions — the force vector is lost, the sternum is harder to reach, and depth falls — so the modern instruction is to keep her flat and have an assistant manually displace the uterus up and to the left (a two-handed cephalad-and-leftward lift from the patient's right side, or a one-handed push from her left), continuously, for the entire arrest. Delivering the fetus and emptying the uterus carries the same logic to its conclusion and abolishes aortocaval compression instantly and permanently: venous return is restored, the diaphragm drops, functional residual capacity and chest compliance improve, and roughly a litre of blood is autotransfused from the uteroplacental bed back into the maternal circulation. Resuscitative hysterotomy is, mechanistically, a resuscitation manoeuvre — it converts an unresuscitatable physiology into a resuscitatable one. Any fetal benefit is secondary.

Pregnancy also lowers the threshold at which hypoxia becomes lethal and shortens the time available. Oxygen consumption is ~20% higher and functional residual capacity ~20% lower, so the apnoeic reserve is small and desaturation is fast; the airway is oedematous, friable and prone to aspiration from a relaxed lower oesophageal sphincter and delayed gastric emptying. The dilutional anaemia and the hypercoagulable, high-output cardiovascular state mean that haemorrhage and thromboembolism — the things most likely to kill her — do so quickly. The clock that governs the four-minute rule is set by the brain: irreversible neuronal injury begins within minutes of arrest, and the fetus, dependent on already-marginal uteroplacental flow, is hypoxic from the moment maternal output fails.

The same physiology explains why the drugs of advanced life support do less in the supine pregnant woman than in a non-pregnant adult. Adrenaline and the rest are delivered to the central circulation only if compressions move blood from the periphery and the venous reservoir into the thorax; if the cava is clamped by the uterus, a drug given into a peripheral or even a femoral line may simply pool below the obstruction. This is the mechanistic reason that upper-limb venous and intraosseous access is preferred (a drug given below the diaphragm has to cross the compressed cava to reach the heart) and the deeper reason that no pharmacological or electrical intervention can substitute for mechanically relieving the obstruction. The uterus is the problem; emptying it is the treatment. It is worth being concrete about the autotransfusion: the term-pregnant uterus and its placental bed hold of the order of 500 mL to a litre of blood, and at delivery a substantial fraction of that volume returns to the maternal circulation — a meaningful preload bolus delivered exactly when the failing right heart needs filling. Mechanistically, then, resuscitative hysterotomy simultaneously removes an obstruction to venous return, restores chest compliance and diaphragmatic excursion, lowers afterload, and autotransfuses volume; no single drug does any one of those things as completely, and none does all four.

Assessment

Collapse is a presentation, not a diagnosis, and the resuscitation and the search for the cause run in parallel from the first second. Confirm arrest the standard way — unresponsive, not breathing normally, no central pulse within ten seconds — and call the obstetric arrest team immediately; the response must bring obstetrics, anaesthesia, a scalpel and a neonatal team to the bedside, because the definitive treatment may be surgical and cannot wait for transfer.

The causes are the universal reversible causes of arrest plus a set specific to pregnancy. Work the 4 Hs and 4 Ts and overlay the obstetric list:

• Hypovolaemia — almost always haemorrhage in this population: antepartum (abruption, praevia, uterine rupture, accreta spectrum) or postpartum (atony, trauma, retained tissue, coagulopathy). Concealed abruption and intra-abdominal bleeding after caesarean can present as collapse with a deceptively soft abdomen. This is the commonest correctable cause and the first to exclude. See postpartum-haemorrhage for the haemorrhage drills.
• Hypoxia — failed or difficult intubation, aspiration, high neuraxial block paralysing the diaphragm, severe asthma, pulmonary oedema (commonly from pre-eclampsia or cardiac disease).
• Hyperkalaemia / metabolic — and, in pregnancy specifically, hypoglycaemia (insulin-treated diabetes) and hypocalcaemia/hypermagnesaemia from magnesium sulphate toxicity in an eclamptic woman with renal impairment, which can itself arrest the heart.
• Hypothermia — uncommon but relevant in trauma and prolonged field resuscitation.
• Thromboembolism — pulmonary embolism, the leading cause of direct maternal death in many high-income settings; also myocardial infarction and aortic dissection, both rising with maternal age.
• Toxins — local-anaesthetic systemic toxicity (LAST) from epidural or paracervical block, magnesium toxicity, illicit drugs, and iatrogenic overdose.
• Tamponade — pericardial, or after cardiac surgery/trauma.
• Tension pneumothorax.

To these add the causes that are only obstetric: eclampsia and intracranial haemorrhage (the hypertensive woman who fits and does not recover, or who collapses with a thunderclap headache — a ruptured aneurysm or hypertensive bleed); amniotic-fluid embolism (sudden collapse around the time of delivery with hypoxia, hypotension and disseminated intravascular coagulation, often heralded by a seizure or sense of impending doom); high or total spinal block; anaphylaxis (antibiotics, oxytocin, latex, suxamethonium); and sepsis, which in South Africa frequently means a background of HIV with opportunistic infection, pneumonia, or genital-tract and post-abortal sepsis. The mnemonic matters less than the discipline of running the whole list while compressions continue, because the treatable causes — haemorrhage, hypoxia, magnesium and local-anaesthetic toxicity, hypoglycaemia, tension pneumothorax — are precisely the ones a structured search will catch.

Reading the cause from the pattern of collapse separates the candidates who manage from those who merely list. A sudden collapse at delivery or in the minutes after, with hypoxia, hypotension and bleeding from every puncture site, is amniotic-fluid embolism until proven otherwise — a clinical diagnosis with no confirmatory test, in which the coagulopathy is profound and the treatment is supportive resuscitation plus aggressive correction of DIC. Collapse shortly after a regional block points to a high or total spinal (ascending block, bradycardia, apnoea, loss of consciousness) or to local-anaesthetic systemic toxicity (perioral tingling and agitation progressing to seizures and a refractory, often broad-complex, cardiac arrest); the two are distinguished by the timing and the prodrome, and LAST has a specific antidote. Collapse in a woman on a magnesium infusion, particularly with reduced urine output, is magnesium toxicity until the level is known — loss of reflexes precedes respiratory and then cardiac arrest. Collapse after a seizure that does not resolve is eclampsia complicated by, or mimicking, intracranial haemorrhage. And collapse with chest pain, breathlessness or after a long immobilisation or caesarean raises pulmonary embolism and, with increasing maternal age, myocardial infarction and aortic dissection. The history that produces the collapse is therefore as diagnostic as the examination.

Bedside information is gathered without interrupting CPR: the drug chart and infusions (is magnesium running? was a block just sited?), the blood loss, the gestation (a fundus at or above the umbilicus is the trigger for hysterotomy and is a clinical judgement, not an ultrasound exercise), point-of-care glucose, and a focused look for the obvious — a soaked drape, a tense abdomen, the rash and wheeze of anaphylaxis. Investigations follow only as they can be obtained alongside resuscitation: arterial gas and lactate, a full blood count and coagulation screen, electrolytes, calcium and magnesium, a bedside echo in skilled hands to identify tamponade or a dilated right heart suggesting pulmonary embolism, and a viscoelastic test where available to direct massive transfusion. None of this delays the manual displacement of the uterus or, when indicated, the knife.

Management

The framework is the standard chain of high-quality CPR and advanced life support with pregnancy-specific modifications layered on top, and it is best held as immediate, then the decision and procedure of resuscitative hysterotomy, then ongoing post-delivery care.

Immediate: maternal advanced life support, modified

Begin compressions immediately, 100–120 per minute, the standard hand position on the lower half of the sternum, depth 5–6 cm, minimal interruptions. Two evidence-based modifications define obstetric CPR:

• Continuous manual left uterine displacement from 20 weeks (or a uterus palpable at/above the umbilicus), maintained throughout, by an assistant whose only job this is. Left lateral tilt is inferior because it degrades compressions; if displacement is impossible and tilt is used as a fallback, 15–30° on a firm surface is the compromise, but flat-with-manual-displacement is the recommended technique in the AHA, ERC and RCOG guidance.
• Early airway control. Anticipate a difficult airway, use a smaller endotracheal tube, apply cricoid and aspiration precautions, and have a senior anaesthetist and a difficult-airway plan ready. Hypoxia kills the pregnant woman fast.

Everything else is standard adult ALS, and the guidelines are emphatic that it should not be diluted: the same defibrillation energies, the same shockable/non-shockable algorithm, the same adrenaline 1 mg every 3–5 minutes, the same hand position. Pads can be placed normally; defibrillation is safe in pregnancy and must not be withheld, and the fetal monitor — if attached — is removed before shocking and otherwise ignored, because fetal assessment never directs maternal resuscitation. Secure intravenous access above the diaphragm (the obstructed cava makes lower-limb access unreliable). Run the reversible causes and treat what you find:

• Magnesium toxicity — stop the infusion and give calcium gluconate 10%, 10 mL intravenously (or calcium chloride), the direct antidote.
• Local-anaesthetic systemic toxicity — 20% intravenous lipid emulsion, a bolus of ~1.5 mL/kg followed by an infusion, alongside prolonged ALS; LAST arrests are notoriously refractory and resuscitation may need to continue far longer than usual.
• Anaphylaxis — adrenaline (intramuscular, or titrated intravenously in arrest), fluids, and removal of the trigger.
• Haemorrhage — activate the massive haemorrhage protocol with early tranexamic acid 1 g and balanced blood products, and find and stop the source (uterotonics, compression, surgery).
• Hypoglycaemia — correct with intravenous dextrose.
• Pulmonary embolism — in arrest, thrombolysis is a considered option despite pregnancy when PE is the likely cause.

The decision and the procedure: resuscitative hysterotomy

If there is no return of spontaneous circulation, the single decision that changes outcome is to perform a resuscitative hysterotomy. The rule is conventionally stated as: in a pregnant woman in cardiac arrest with a uterus at or above the umbilicus (roughly ≥20 weeks), decide by four minutes of arrest without ROSC and aim to deliver by five minutes.

The operational details are unforgiving and each is a place candidates go wrong:

• It is done where she lies. There is no transfer to theatre, no scrub, no skin preparation, no waiting for an anaesthetist or for anaesthesia to be given (an arrested patient does not perceive pain; an antiseptic wipe at most). Moving her wastes the only resource that matters, which is time, and interrupts compressions.
• The fastest route in. A vertical midline (or generous transverse) skin incision and whatever uterine incision delivers the fetus quickest; the classical vertical uterine incision is often the fastest in inexperienced hands and through a contracted lower segment. Bleeding is minimal during arrest because there is no circulation — haemostasis is dealt with after ROSC.
• Compressions never stop. Chest compressions and the ALS algorithm continue throughout the procedure and after delivery.
• The goal is maternal survival. This reframing — from "peri-mortem caesarean to save the baby" to "resuscitative hysterotomy to save the mother" — is the conceptual core. Relieving aortocaval compression and autotransfusing the uteroplacental volume is what may restore an output that drugs and shocks could not. Delivery is therefore indicated even if the fetus is already dead, even below the conventional viability threshold, and even if the four- or five-minute window has passed: late is better than not at all, because the maternal benefit persists.

A neonatal team should be present, but their needs never delay the hysterotomy and never override the maternal indication.

Ongoing: after delivery

If ROSC is achieved, the work shifts to definitive treatment of the cause and damage control. The open uterus and abdomen now need surgical haemostasis — uterotonics, repair, and the full postpartum haemorrhage escalation if atony or accreta is found; transfusion guided by viscoelastic testing; correction of coagulopathy, especially in amniotic-fluid embolism where DIC is profound. Targeted post-cardiac-arrest care follows: controlled oxygenation and ventilation, haemodynamic support, treatment of the precipitant (anticoagulation or embolectomy for confirmed pulmonary embolism, source control and antibiotics for sepsis, neurosurgical referral for intracranial haemorrhage), and consideration of targeted temperature management. The neonate, delivered in poor condition after a period of maternal arrest, needs full resuscitation and, where the gestation and timing allow, therapeutic hypothermia for hypoxic-ischaemic encephalopathy. Debriefing and incident review are part of management, not an afterthought, and every such event feeds the maternal-mortality audit cycle.

The South African collapse profile and the systems that decide outcome

What collapses a South African mother is not quite the high-income picture, and the plan has to be built around it. The NCCEMD Saving Mothers reports place non-pregnancy-related infections — largely HIV and TB-related disease — among the leading causes of maternal death, alongside obstetric haemorrhage, hypertensive disorders (with eclampsia and intracranial haemorrhage their lethal end-points), medical and surgical conditions, pregnancy-related sepsis, acute collapse and embolism (including amniotic-fluid and pulmonary embolism), and anaesthesia-related deaths. The anaesthetic and haemorrhagic risk concentrates around caesarean section: the NCCEMD found the risk of a woman dying as a result of caesarean to be roughly three times that of vaginal delivery, with deaths clustering in the provinces with the lowest caesarean rates — a signal that the danger lies less in the operation than in where and by whom it is performed, often under spinal anaesthesia by a junior operator at a district hospital without immediate senior or blood-bank support. A high or total spinal, an unrecognised post-spinal hypotension, or a torrential post-caesarean bleed in such a setting is exactly the collapse this chapter is about.

The corollary for management is that the determinant of survival in South Africa is rarely knowledge of the four-minute rule and almost always the system around the bedside. The interventions that move outcomes are mundane and repeatable: a pre-stocked resuscitative-hysterotomy or maternal-arrest pack on every labour ward (scalpel, cord clamps, suction, large swabs), a written obstetric-arrest call protocol that brings the right people in seconds, regular skills-and-drills and simulation of maternal collapse so that manual uterine displacement, the timing and the decision to cut are rehearsed rather than improvised, and an honest audit of every near-miss and death through the Saving Mothers framework. These are precisely the avoidable factors the confidential enquiries identify again and again — recognition too late, no plan, no equipment, no one who has done it before. Naming the system as the lever, not the individual's recall of a guideline, is the mature management position.

Guidelines compared

The major bodies agree on the essentials and differ mainly in emphasis and in how prescriptively they state the timing.

The substantive shift over the last decade, common to all of them, is the renaming and reframing of the procedure from peri-mortem caesarean (fetus-centred) to resuscitative hysterotomy (mother-centred), and the move away from lateral tilt towards manual displacement. The remaining tension is in how rigidly to hold the five-minute target: the guidelines retain it as an aspiration while acknowledging the evidence does not support a hard cut-off, which is the substance of the next section.

The evidence & the controversy

This is one of the few areas of obstetrics where the headline recommendation rests on no randomised trial and never will — randomising a woman in cardiac arrest to delayed delivery is neither feasible nor ethical. The evidence base is case reports, case series and registries, and it must be read with that limitation stated plainly rather than dressed up.

The four-minute rule originates from a 1986 review and was adopted by the resuscitation and obstetric bodies on physiological reasoning and the best available observational data. When that data was systematically re-examined by Einav and colleagues, the picture was more nuanced than the rule implies. Across 94 reported cases, maternal survival to discharge was 54.3%, and — the crucial finding — the procedure was judged to have benefited the mother in about a third of cases and harmed her in none. But the timing data did not show a clean cliff at four or five minutes: survivors were delivered well beyond the window, and the relationship between delay and outcome was a gradient, not a threshold. The reasonable reading is that the five-minute target is a goal to drive urgency, not a deadline after which the procedure becomes pointless — delivering at eight or ten minutes is still worthwhile, both because some mothers survive and because uterine decompression aids ongoing resuscitation regardless.

Rose and colleagues pushed this further, arguing for renaming the procedure resuscitative hysterotomy and for preparing to deliver from the moment of arrest rather than counting down four minutes — if the uterus is at or above the umbilicus and the mother is not rapidly reversible, the knife should be ready immediately, because the time lost locating equipment, assessing the fetal heart, or moving to theatre is the time that decides the outcome. This is now the dominant conceptual position. The honest caveat, reinforced by the most recent systematic review of out-of-hospital cases (Leech 2024), is that survival figures depend heavily on the setting: in a witnessed in-hospital arrest with a team at the bedside, outcomes can be good, but in unwitnessed out-of-hospital arrest maternal survival was only 4.5% — a reminder that the procedure rescues a physiology, not a corpse, and that prevention and early recognition of collapse do far more good than heroic late surgery.

A live and genuinely unsettled thread is what to add for the small group of women who arrest from a reversible but persistent cause — a massive pulmonary embolism, an amniotic-fluid embolism with refractory cardiogenic shock — and who do not regain an output after hysterotomy and standard ALS. Extracorporeal CPR (veno-arterial ECMO instituted during arrest) has supported maternal survival in case reports and is increasingly discussed for amniotic-fluid embolism and massive PE, and resuscitative endovascular balloon occlusion of the aorta (REBOA) is being explored for exsanguinating obstetric haemorrhage. Both are tertiary, resource-intensive and evidence-thin, and in most of the world — certainly in most South African facilities — they are not available; raising them is appropriate as the direction of travel for a referral centre, but presenting either as standard would overstate a literature that is still individual cases. The honest position is that they are rescue options for a reversible cause in a centre that has them, not a substitute for the immediate, universal interventions of displacement, standard ALS and timely hysterotomy.

The ethics of the procedure deserve a clear framework rather than an evasion. Resuscitative hysterotomy is performed on a woman in cardiac arrest, without her consent and usually without her family's, and it may deliver a previable or already-dead fetus. The justification is that it is a maternal resuscitation manoeuvre undertaken in her best interests when she cannot consent and death is otherwise certain — the same emergency-doctrine logic that authorises any life-saving intervention in an unconscious patient — and that the bar for acting is therefore low, not high. The defensible approach is to act on that doctrine, perform the procedure when indicated, and address consent, disclosure and grief afterwards; hesitating over consent in an arrest is itself the error.

The deeper controversy in the South African context is not the four-minute rule at all but whether the conditions for success exist. The procedure demands a witnessed arrest, a team that arrives in seconds, a person whose sole task is uterine displacement, and an operator and scalpel at the bedside — a chain that holds in a tertiary labour ward but frequently breaks at district level, where a single doctor may be covering the hospital and the arrest is recognised late. The defensible position is that the rule is correct and that the realistic lever in South Africa is systems: obstetric-arrest drills, simulation, pre-positioned resuscitative-hysterotomy packs on labour wards, and a clear escalation protocol — the same skills-and-drills approach that the maternal-mortality audit repeatedly identifies as the gap.

Landmark trials & key evidence

There are no randomised controlled trials in this field, and saying so is part of a defensible answer; the evidence is observational and consensus-derived, and the table reflects that.

Exam traps & red flags

• Tilting instead of displacing. Lateral tilt degrades chest compressions; the correct manoeuvre is continuous manual left uterine displacement with the woman flat. Reaching for a wedge is the common error.
• Counting to four minutes before doing anything. The four-minute mark is the decision point and the five-minute mark the delivery target, but preparation for hysterotomy begins at the moment of arrest in a uterus at/above the umbilicus — waiting passively until four minutes have elapsed before sending for a scalpel guarantees the window is missed.
• Moving her to theatre. The procedure is done where she lies. Transfer, scrubbing, draping and waiting for anaesthesia all forfeit the only resource that matters.
• Treating it as a caesarean for the baby. The indication is maternal — delivery is performed even if the fetus is dead, even below viability, and even after the five-minute window, because relieving aortocaval compression aids the mother's resuscitation.
• Diluting standard ALS. Withholding defibrillation, reducing energies, or omitting adrenaline "because she is pregnant" is wrong; the drugs, energies and algorithm are unchanged.
• Missing the obstetric causes. Magnesium toxicity (give calcium), local-anaesthetic systemic toxicity (give lipid emulsion), high spinal, amniotic-fluid embolism with DIC, eclampsia/intracranial bleed, anaphylaxis and concealed haemorrhage are the reversible causes that a standard 4-H/4-T sweep can miss if the obstetric overlay is forgotten.
• Forgetting the autotransfusion and ongoing bleeding. After ROSC the decompressed, atonic uterus and the open abdomen can bleed catastrophically — the haemorrhage protocol must follow delivery, not be assumed to be over.
• Assuming the system will work. In a setting without drills, packs and an arrest team, the procedure is recognised and performed too late; the avoidable factor is the system, and naming it is part of the answer.

Evidence anchors

• Einav S, Kaufman N, Sela HY. Maternal cardiac arrest and perimortem caesarean delivery: evidence or expert-based? Resuscitation 2012;83(10):1191-200
• Rose CH, Faksh A, Traynor KD, et al. Challenging the 4- to 5-minute rule: from perimortem cesarean to resuscitative hysterotomy. Am J Obstet Gynecol 2015;213(5):653-6
• Jeejeebhoy FM, Zelop CM, Lipman S, et al. Cardiac Arrest in Pregnancy: A Scientific Statement From the American Heart Association. Circulation 2015;132(18):1747-73
• Lipman S, Cohen S, Einav S, et al. The SOAP consensus statement on the management of cardiac arrest in pregnancy. Anesth Analg 2014;118(5):1003-16
• Lott C, Truhlář A, Alfonzo A, et al. European Resuscitation Council Guidelines 2021: Cardiac arrest in special circumstances. Resuscitation 2021;161:152-219
• Leech C, Nutbeam T, Chu J, et al. Maternal and neonatal outcomes following resuscitative hysterotomy for out-of-hospital cardiac arrest: a systematic review. Resuscitation 2024;207:110479
• Chu J, Hinshaw K, Paterson-Brown S, et al. Perimortem caesarean section – why, when and how. The Obstetrician & Gynaecologist 2018;20(3):151-8
• Gebhardt GS, Fawcus S, Moodley J, Farina Z. Maternal death and caesarean section in South Africa: results from the 2011–2013 Saving Mothers Report of the NCCEMD. S Afr Med J 2015;105(4):287-91
• RCOG Green-top Guideline No. 56 — Maternal Collapse in Pregnancy and the Puerperium (1st edition 2011; reviewed/extended December 2024).
• NCCEMD Saving Mothers reports (NDoH, South Africa) — the leading causes of maternal death in South Africa: non-pregnancy-related infections (largely HIV/TB-related), obstetric haemorrhage, hypertensive disorders, medical and surgical conditions, pregnancy-related sepsis, acute collapse and embolism (including amniotic-fluid embolism and pulmonary embolism), and anaesthesia-related deaths.