Amniotic fluid embolism occurs when some amniotic fluid gains access to the maternal circulation, often in massive volumes and with catastrophic consequences. It was discovered for the first time by Ricardo Juvenal Meyer in 1926, who was extremely surprised to find whole chunks of foetal tissue (skin cell, lanugo hairs, intestinal mucin) in the pulmonary vessels of dead mothers. Clearly that was an abnormal finding, but nobody really put two and tow together until a whole case series of sudden maternal deaths was linked to pulmonary embolism of amniotic fluid by Steiner and Lushbaugh (1941). Death occurs typically due to circulatory collapse, or (if that doesn't get you) respiratory failure and severe hypoxia.
Though Question 28 from the first paper of 2017 was the first time the college had asked for some detail about this condition, it comes up now and again as a differential in the lists of possible reasons for DIC (for example), or sudden peri-Caesarian cardiovascular catastrophe, or possibly as a part of a particularly vicious viva station. The question specifically asked for clinical features and risk factors. Potentially, vicious viva scenarios may in the future ask the candidates to make an echo diagnosis, or to manage the situation (which may or may not progress to cardiac arrest).
As far as published literature goes, one cannot go past Moore's article from Critical Care Medicine (2005). That would probably be enought to satisfy the requirements of the time-poor exam candidate who does not wish to pay UpToDate subscription fees.
Clinical features amniotic fluid embolism
Question 28 from the first paper of 2017 asked the candidates to outline the important clinical features of amniotic fluid embolism. The college answer listed the following features, making them canonically "important":
- The onset of the symptoms and signs of amniotic fluid embolism syndrome (AFES) most commonly occurs during labour and delivery, or immediately postpartum
- Non-specific symptoms – chills, nausea, vomiting, agitation
- Hypotension due to cardiogenic shock
- Hypoxemia and respiratory failure
- Disseminated intravascular coagulation
- Coma or seizures
This list likely represents the maximum expected of a candidate. The British AFE Register (Tuffnell et al, 2005) used something very similar to identify cases for record-keeping purposes. The hypoxia hypotension and DIC are what you migth call "cardinal signs". Other associated features are listed by Moore et al (2005):
- Foetal distress
- Nausea / vomiting
Differentials are also offered by the same article, which is helpful:
- Pulmonary thromboembolism
- Air embolism
- Aspiration of gastric contents
- Anesthetic complications
- Sepsis/systemic inﬂammatory response
- Myocardial infarction
- Transfusion reactions
How does this happen?
Amniotic fluid gains access to the circultaion by a number of possible routes:
- Endocervical veins
- Placental insertion site
- Uterine injury (i.e. open wound on the uterus, such as a Caesarian incision)
Let's say that there is some sort of direct venous blood/amniotic fluid interface, at the breach of some venous blood vessel somewhere in the uterus. Amniotic fluid will get into the venous circulation though this opening by a variety of mechanisms:
- Negative venous pressure: venous blood returns to the heart, and therefore there is a constant flow through this side of the circulation. This is exacerbated by the deep breaths characteristic of labour. Ergo, occasionally during particularly deep inspiration the venous pressure may become negative (due to extremely negative intrathoracic pressure) entraining amniotic fluid through those breahed venules.
- Positive amniotic fluid pressure. Amniotic fluid pressure during pregnancy is usually positive anyway (5-10mmHg). In labour, it can increase substantialy with uterine contractions. Imagine then if the uterus contracts while the uterine blood vessels are open. Amniotic fluid will very likely choose a path of least resistance, and funnel massively into the inferior vena cava.
Now, amniotic fluid is not some sort of crystal-clear mountain stream straight from some sanitised fairytale description of pregnancy. A sopering realism is introduced by Doke Uyeno (1919) who described its properties in horriic detail.
The amniotic fluid is slimy, yellowish white or pale yellow, and cloudy like soap water (rarely clear); it almost always contains much mucous flocks and has a peculiar odor.
This magical material has interesting physicochemical properties and a complex composition, largely consiting of:
- Maternal plasma transudate
- Foetal urine
- Foetal intestinal mucus
- Foetal skin flakes and lanugo hair
- Fatty acids and cholesterol
- Zinc coproporphyrin
- Arachidonic acid metabolites
So, a volume of this material entering the circulation can only the most disastrous of consequences, being equivalent to the intravenous injection of sewer water.
Hypoxia and shock
These are due to several components, elegantly illustrated in these diagrams from Moore et al (200%):
The hypoxia is due to the crude effects of V/Q mismatching which occurs when amniotic fluid emboli clog the lungs. Several pathophysiologies contribute to the hypoxia:
- PE-like VQ mismatch
- Cardiogenic pulmonary oedema
- Non-cardiogenic "capillary leak" pulmonary oedema
- Bronchospasm (similar mechanism to anaphylaxis)
The shock is initially obstructive (giving rise to characteristic intraopetaive TOE findings) and subsequently distributive with a sepsis-like capillary leak.
Disseminated intravascular coagulation (DIC)
This is though to be due to a systemic inflammatory response with activation of complement and the activation of the coagulation cascade throughout the circulation. It can start rapidly, within 10-30 minutes.
Agitation, delirium, seizures, coma
Moore et al (2005) attributes the neurological findings in AFE to severe hypoxia.
Risk factors for amniotic fluid embolism
- Precipitous or tumultuous labour.
- Advanced maternal age.
- Caesarean and instrumental delivery.
- Placenta previa and abruption.
- Grand multi-parity (≥5 live births or stillbirths),
- Cervical lacerations.
- Foetal distress.
- Medical induction of labour.
Knight et al (2010 and 2012) adds a few more risk factors:
- Ethnic minority background
- Emergency delivery
- Smoking during pregnancy
- Socioeconomic disadvantage
- Age over 35 (what the college describes as "advanced maternal age", otherwise known as "checkout time" )
- Prematurity (RR 1.7 as compared to full term)
- Any sort of hypertensive disorder of pregnancy (not merely preeclampsia)
- Previous caesarian (but only in Australia! in other countries, a previous caesarian actually decreases the risk. Are we doing it wrong?)
- Of the methods of induction of labour, vaginal prostaglandin seems to be associated with the greatest risk.
- Manual removal of the placenta (has an RR of 19.6)
A specific risk factor labor which UpToDate and the college describe as "tumultuous", implying bizarrely that all other labour is a calm and peaceful process.
Diagnosis of amniotic fluid embolism
This is usualy a diagnosis of exclusion, and based on a strong clinical suspicion.
Gold standard of diagnosis is the totally unrealistic manoeuvre of sucking some amniotic fluid debris out of a distal port of the PA catheter.
Foetal debris in maternal sputum is a more civilised alternative, but again unlikely to be game-changing if you already strongly suspect AFE.
Foetal antigen serology looking for TKH-2 antibodies ( or insulin-like growth factor binding protein-1) are apparently not well validated
- FBC (thrombocytopenia)
- Coags/fibrinogen (DIC)
- Troponin (cardiac strain)
- ABG (hypoxia, lactic acidosis)
- Imaging etc:
- TTE/TOE (right heart strain, evidence of obstructive shock)
- ECG (RV strain)
- CXR (pulmonary oedema)
Management of amniotic fluid embolism
- Maintain normoxia.
- VV ECMO is an option.
- Ventilate with 100% FiO2
- Ensure adequate filling to improve shunt fraction:
- Encourage pulmonary arterial vasodilation:
- Inhaled nitric oxide
- Inhaled prostacycline
- Maintain satisfactory cardiac output
- Inotropes and vasopressors
- VA ECMO
- Decontaminate the circulation
- CVVHDF to clear out the circulating proinflammatory foetus chunks
(Kaneko et al, 2001)
- Plasma exchange, which frequently happens in an uncontrolled manner as the patient tries to bleed to death and receives a massive transfusion "in exchange"(Awad et al, 2001).
- Cell salvage and blood filtration may work. Water et al (2000) reported a case where much of the foetal debris was removed from the blood by a cell saver device.
- CVVHDF to clear out the circulating proinflammatory foetus chunks
- Intubation is likely to be required at some stage, for decreased LOC, to control the FiO2 more precisely and to send some positive pressure against the APO
- Sedation for tube tolerance and also to decrease the likelihood of seizures.
- Attention to electrolytes to ensure ionised calcium is well corrected
- Fluid resuscitation to account for third-space losses (leaky capillaries)
- Activated Factor VIIa for DIC-induced haemorrhagic complications (Lim et al, 2004)
- Delivery must be urgent
- Perimortem caesarian may be required (65% of the cases occur before delivery)