Show Summary Details
Page of

Pulmonary thromboembolic disease 

Pulmonary thromboembolic disease
Pulmonary thromboembolic disease

Stephen Chapman

, Grace Robinson

, John Stradling

, Sophie West

, and John Wrightson

Page of

PRINTED FROM OXFORD MEDICINE ONLINE ( © Oxford University Press, 2022. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in Oxford Medicine Online for personal use (for details see Privacy Policy and Legal Notice).

date: 01 July 2022

Epidemiology and pathophysiology


A pulmonary embolism (PE) is an obstruction of part of, or the entire, pulmonary vascular tree, usually caused by thrombus from a distant site.


  • The overall annual incidence is 60–70/100, 000, with a UK annual death rate of 100/million. The estimated overall population incidence of DVT is 0.5 per 1, 000 person years

  • PE may account for up to 15% of all post-operative deaths. It is the commonest cause of death following elective surgery, and the commonest cause of maternal death

  • Post-mortem studies have consistently shown a frequency of 7–9%, and large inpatient studies have shown a frequency of around 1%, with a mortality of 0.2%. The mortality is much higher in patients with serious underlying comorbid disease

  • The incidence is likely to be stable, but improved diagnostic methods mean that it is probably reported more frequently.


  • 75% of thrombi are generated in the deep venous system of the lower limbs and pelvis, probably initiated by platelet aggregation around venous valve sinuses. Activation of the clotting cascade leads to thrombus formation, with Virchow’s triad (venous stasis, injury to the vessel wall, and increased blood coagulability) predisposing to thrombus formation. Venous stasis is increased by immobility and dehydration. In addition, coagulation factors may be altered in various disease states, e.g. in the acute phase response, malignancy, and autoimmune disease

  • 20% of leg thrombi embolize, with a higher incidence in above than below knee clots. Large clots may lodge at the bifurcation of the main pulmonary arteries, causing haemodynamic compromise. Smaller clots will travel more distally, infarcting the lung and causing pleuritic pain. These are more commonly multiple and bilateral and are found most often in the lower lobes where blood flow is greatest

  • Thrombi can also develop in the right heart following MI

  • Paradoxical emboli start within the venous system and enter the arterial circulation, usually via a patent foramen ovale (causing right-to-left shunt). They typically present with features of cerebral ischaemia and these should be considered as the cause for a cerebrovascular event in the young

  • Septic emboli are found in endocarditis, in association with intraventricular septal defects/AV shunts or central venous access.

Haemodynamic effects of PE

depend on the size of the clot and which area of the pulmonary vascular tree it subsequently obstructs, as well as the pre-existing state of the myocardium.

  • As the pulmonary vasculature in a healthy lung has a large capacitance, the mean PAP does not rise until at least 50% of the vascular bed has been occluded

  • As the PAP rises, RV afterload increases, with a resulting increase in RV end diastolic pressure. The RV will start to fail as the PAP reaches over 40mmHg acutely

  • This causes a reduction in pulmonary blood flow, leading to reduced LV filling and a reduction in systemic BP

  • Adequate blood volume for right-sided heart filling is vital. The 2° effects are much worse if right-sided filling cannot be maintained, e.g. if the patient is dehydrated, hypovolaemic, or erect

  • Arterial hypoxia results from several factors: reduced cardiac output, consequently a low mixed venous PaO2, a higher perfusion to the remaining alveoli, resulting in inadequate oxygenation of this blood

  • Hypoxia will be worse if there is a larger premorbid V/Q spread, e.g. in the elderly and in those with pre-existing lung disease. The increased blood flow, with a lower mixed venous PaO2 passing through low V/Q areas, overwhelms their oxygenating ability. It is therefore possible for a young person with healthy lungs to have a normal PaO2 and A–a gradient following a significant PE

  • Death is due to circulatory collapse from the inability of the right heart to acutely maintain an adequate cardiac output.


Risk factors can be divided into major and minor factors (see Table 39.1). This division is important for an assessment of clinical probability.

Table 39.1 Risk factors for VTE

Major risk factors (relative risk × 5–20)


Major abdominal/pelvic surgery

Orthopaedic surgery (especially lower limb)

Post-operative intensive care


Pregnancy (higher incidence with multiple births)

Caesarean section





Lower limb problems

Fracture, varicose veins

Reduced mobility


  • Institutional care

  • Long haul flight

Previous proven VTE

Minor risk factors (relative risk × 2–4)


Congenital heart disease

  • CCF

  • Hypertension

Central venous access

Superficial venous thrombosis


Oral contraceptive pill (OCP) (especially third-generation higher oestrogen-containing)

Hormone replacement therapy (HRT)


Occult malignancy

Neurological disability

Thrombotic disorders



Nephrotic syndrome


Myeloproliferative disorders

Behçet’s disease

Risk of malignancy

Occult cancer will be present in 7–12% of patients presenting with idiopathic VTE. New NICE guidance (2012) recommends considering further investigations for cancer with an abdomino-pelvic CT and a mammogram for women in all those aged >40 with a first episode of unprovoked PE or DVT. No studies so far show that this strategy leads to a reduction in cancer-related mortality.

Inherited thrombophilias

  • 25–50% of patients with VTE have an identifiable inherited thrombophilia, e.g. antiphospholipid syndrome, deficiency of antithrombin III, a prothrombin gene defect, protein C or protein S deficiency

  • These usually need to interact with an additional acquired risk factor to cause VTE

  • Factor V Leiden is present in 5% of the population and 20% of patients presenting with thrombosis

  • Current recommendations do not advocate routine screening for inheritable thrombophilias, unless in specific circumstances (see further text on thrombophilia testing), as the number needed to test to prevent an episode of VTE would be very high. In addition, detecting a heritable thrombophilia does not predict a significantly higher rate or earlier occurrence of VTE in the absence of a 2° risk factor.

Consider thrombophilia testing in:

  • Patients with recurrent venous thrombosis

  • Patients <40 with venous thrombosis with no obvious risk factors

  • Thrombosis 2° to pregnancy, OCP, HRT

  • Thrombosis at an unusual site—cerebral, mesenteric, portal, or hepatic veins

  • Do not offer testing in those continuing lifelong anticoagulation or in those with provoked clot.

All, but factor V Leiden deficiency and the prothrombin gene mutation, need to be tested for when the patient is off anticoagulants.

‘Economy class syndrome’

refers to thromboembolic disease associated with long-distance sedentary travel, with an increasing incidence of disease with increasing distance travelled. A 2001 study of >135 million passengers showed an incidence of PE of 1.5 cases/million for travel over 5, 000km, compared with 0.01 cases/million for travel under 5, 000km. For travel over 10, 000km, the incidence increased to 4.8 cases/million.

Clinical features

Acute PE

typically presents in four main ways.

  • Pulmonary infarction and haemoptysis ± pleuritic pain. ABGs may be normal and ECG changes uncommon. Localizing signs may be present, e.g. pleural rub

  • Isolated dyspnoea (in 25%) Defined as acute breathlessness in the absence of haemorrhage or circulatory collapse. The thrombus is more likely to be central, with hypoxia on blood gases. The patient may have sudden-onset and unexplained breathlessness, in the presence of risk factors for VTE. There may also be angina from increased right heart work and inadequate O2 delivery to its muscle

  • Collapse, poor reserve (in 10%) May be due to a small PE, often in an elderly patient with limited cardiorespiratory reserve. These patients can rapidly decompensate with even a relatively small PE. The clinical findings may be non-specific and reflect more the underlying disease process and thus fail to arouse suspicion of a PE

  • Circulatory collapse in a previously well patient Hypotension ± loss of consciousness in 1%. Usually due to extensive pulmonary artery occlusion from massive PE, causing marked hypoxia and hypocapnia (due to hyperventilation) and acute right heart failure, with chest pain due to right heart angina, raised JVP, and fainting on sitting up. ECG may be normal, show sinus tachycardia or right heart strain. Echo shows PHT and RV failure. These patients have the highest mortality, up to 30%.

Chronic thromboembolic disease

This typically presents with more insidious onset of breathlessness over the course of weeks to months due to increasing load of recurrent small-volume clots (see Pulmonary thromboembolic disease pp. [link][link]).

Dyspnoea and tachypnoea (RR >20) are the commonest presenting features and are absent in only 10% of patients.

Remember to consider PE in the differential diagnosis of:

  • Unexplained SOB

  • Collapse

  • New-onset AF

  • Signs consistent with acute right heart failure

  • Pleural effusion.

Examination of a patient with PE

  • May be normal

  • Tachycardia and tachypnoea are common

  • AF

  • Reduced chest movement (due to pain)

  • Pleural rub

  • Classically loud P2 and splitting of the second heart sound, with a gallop rhythm (acute right heart strain)

  • Hypoxia (with hypocapnia due to hyperventilation, and an increased A–a gradient), but PaO2 may be in the normal range in young healthy individuals

  • Low-grade fever

  • Signs of DVT (common, in around 25%)

  • Acute right heart failure—low cardiac output and raised JVP, with reduced BP and perfusion pressure

  • Deterioration in cardiac output on sitting up, when filling pressure falls.

Diagnosis of acute PE

The diagnosis of a PE can be difficult and involves a clinical assessment of probability. This takes risk factors, clinical presentation, and clinical signs into account. Investigations are then performed, that may add weight to the clinical decision, rather than being stand-alone diagnostic tests. Therefore, the estimation of the pre-test clinical probability of DVT and PE is of vital importance in interpreting the results of the tests performed.

Pre-test clinical probability scoring systems

(See Tables 39.2 and 39.3.)

Table 39.2 Wells’ score for DVT

Clinical feature


Active cancer (treatment ongoing, within 6 months, or palliative)


Paralysis, paresis, or recent plaster immobilization of the lower extremities


Recently bedridden for 3 days or more or major surgery within 12 weeks requiring general or regional anaesthesia


Localized tenderness along the distribution of the deep venous system


Entire leg swollen


Calf swelling at least 3cm larger than asymptomatic side


Pitting oedema confined to the symptomatic leg


Collateral superficial veins (non-varicose)


Previously documented DVT


An alternative diagnosis is at least as likely as DVT


Likelihood of DVT

DVT likely

≥2 points

DVT unlikely

≤1 point

Table 39.3 Wells’ score for PE

Clinical feature


Clinical signs and symptoms of DVT (minimum of leg swelling and pain with palpation of the deep veins)


An alternative diagnosis is less likely than PE


HR >100bpm


Immobilization for >3 days or surgery in the previous 4 weeks


Previous DVT or PE




Malignancy (on treatment, treated in the past 6 months, or palliative)


Likelihood of PE

PE likely

>4 points

PE unlikely

≤4 points

Local alternative scoring systems may be in place.

These scoring systems should always be used with the D-dimer result.


has an important role in diagnosing and excluding PE and should only be used with a pre-test clinical probability assessment following careful clinical evaluation by an experienced clinician. D-dimers are sensitive for DVT and thromboembolism but not specific. They are rarely in the normal range in cases of acute thromboembolism but are not a valid screening test for PE alone. D-dimers are generated as a result of fibrinolysis, which occurs in many clinical situations, including sepsis, post-surgery, pneumonia, neoplasia, inflammatory disease, pregnancy, and advanced age.

  • Only a normal result (which virtually excludes PE) is of clinical value

  • An abnormal result (however high) does not necessarily imply a significantly increased probability of PE

  • The sensitivity ranges from 87% to 99%, depending on the assay used; these should be known before incorporating into diagnostic algorithms

  • D-dimer testing for excluding PE has been validated as an outpatient test but not in inpatient groups.

Assessment and documentation of pre-test clinical probability in PE is paramount. This enables accurate clinical assessment and may obviate the need for imaging.

An alternative explanation for the symptoms should be sought when a PE is excluded.


  • ECG Non-specific changes are frequent. Most commonly, sinus tachycardia. AF, RBBB, anterior T-wave inversion (indicating RV strain) are common. The S1Q3T3 pattern is uncommon

  • CXR A good-quality departmental CXR is required. No specific features are characteristic in PE, but it may reveal another pathology. Small effusions are present in 40% (80% are exudates, 20% transudates). Focal infiltrates, segmental collapse, and a raised hemidiaphragm can also occur

  • ABG may be normal, especially in the young and healthy. Hypoxia and hypocapnia, due to hyperventilation, with an increased A–a gradient are more common

  • D-dimer (see Pulmonary thromboembolic disease pp. [link][link])

  • Brain natriuretic peptide (BNP) levels in acute PE reflect severity of RV strain and haemodynamic compromise, providing additional prognostic information to that of echo

  • Elevated cardiac troponin levels are associated with worse short-term prognosis in acute PE. Heart-type fatty acid-binding protein (H-FABP), an early marker of myocardial injury, is reported to be superior to troponin or myoglobin measurements for risk stratification of PE on admission. There are currently no universally accepted criteria for the measurement of myocardial injury in acute PE

  • CTPA is the gold standard investigation and is recommended as the initial imaging technique in suspected non-massive PE. It has a sensitivity of >95% and may enable an alternative diagnosis to be made if PE is excluded. Advances in imaging mean that a 16-slice multi-detector row scanner can image the entire chest with resolution approaching 1mm, requiring a breath-hold of <10s. Emboli can be detected in sixth-order pulmonary vessels, which are so small that their clinical relevance is uncertain. CTPA should be performed within 1h in suspected massive PE and within 24h of suspected non-massive PE. The sensitivity and specificity of CTPA depends on the location of the emboli, with lower sensitivity for clot confined to the segmental or subsegmental pulmonary vessels, compared with more central clot. CTPA needs specialist reporting.

In those with a high clinical probability, but negative CTPA, the options are:

  • PE has been excluded; stop anticoagulation, or

  • Perform further imaging (leg US, conventional pulmonary angiogram, venous phase CT to include the legs).

In one large prospective multicentre study, with all patients investigated with CTPA and leg US, those with negative tests and low or intermediate clinical probability were not anticoagulated. Only 0.2% had a definite PE after 3 months of follow-up. Those with negative tests, but high clinical probability, were investigated further, and PE was identified in 5% (Musset D et al. Lancet 2002;360:1914).

A volume of 100–150mL contrast media is required for CTPA, which poses a substantial risk of nephropathy (in patients with renal insufficiency and diabetes) and sometimes fluid overload (patients with impaired LV function). In these patients, leg US and/or isotope lung scanning might be safer first-line investigations.

  • Isotope lung scanning (V/Q scan)—mostly now superseded by CTPA. Some units may just perform the Q (perfusion) part of the scan. May be useful as a first-line imaging investigation in patients with a normal CXR and with no concurrent cardiopulmonary disease, in whom a negative scan can reliably exclude a PE. Scans are reported as low, intermediate, or high probability, and the report’s meaning must be interpreted in light of the pre-test clinical probability score. Further imaging is necessary for those in whom:

    • The scan is indeterminate

    • There is a discordant scan result and clinical probability.

The clinical significance of the V/Q scan report is:

  • Normal = no PE

  • Low or intermediate pre-test clinical probability plus low probability scan = PE excluded

  • High pre-test clinical probability plus high probability scan = PE diagnosed

  • Any other = need further imaging.

Other imaging techniques

  • Leg US Around 70% of patients with a proven PE have a proximal DVT; hence, leg imaging can be used as an alternative to lung imaging in those with clinical DVT. A single examination is not adequate to exclude subclinical DVT (venography is more sensitive). It is safe to withhold anticoagulation in patients with suspected DVT and a single negative leg US, but these data cannot yet be extrapolated to those presenting with suspected PE. If a leg US is positive in a patient with clinical features of PE, this excludes the need for further imaging. Up to 50% of patients with a clinically obvious DVT will have a high-probability V/Q scan

  • Conventional pulmonary angiogram is available in a few specialist centres only where catheter fragmentation of large clots may be of therapeutic benefit. Now mostly superseded by CTPA

  • CT venography is an emerging area. It can be combined with CTPA to image the pelvic leg veins simultaneously

  • Echo is diagnostic in submassive and massive PE. The transoesophageal route is more sensitive, enabling visualization of intrapulmonary and intracardiac clot. Gives prognostic information and aids risk stratification

  • Transthoracic US is used uncommonly. May show peripheral infarcts with peripheral PEs.


Risk stratification

The PE severity index (PESI score; see Table 39.4) allows stratification of immediate risk following PE, dividing patients into high risk and non-high risk, and clearly defines a low-risk population suitable for outpatient management.

  • High-risk PE Shock or hypotension, with positive biomarkers of RV dysfunction, is a life-threatening emergency and has a mortality of >15%

  • Non-high-risk PE can be stratified with the use of cardiac biomarkers of RV dysfunction or myocardial injury into intermediate- (one or more positive markers but no shock, mortality 3–15%, requires in-hospital management) and low-risk PE (negative markers, mortality <1%).

Table 39.4 PESI score*

Clinical feature


Age—add 1 point per year of age


♂ patient


History of cancer


History of heart failure


History of chronic lung disease


HR ≥110


Systolic BP <100mmHg


RR ≥30/min


Temperature <36°C


Altered mental status? (disoriented, lethargy, stupor, or coma)


SaO2 on room air <90%


≤65 Class 1 Very low risk

66–85 Class 2 Low risk

86–105 Class 3 Intermediate risk

106–125 Class 4 High risk

>125 Class 5 Very high risk

* Predicts 30-day outcome of patients with PE (Aujesky D et al. Am J Respir Crit Care Med 2005;172:1041).

Echocardiographic features of RV dysfunction occur in up to 25% of all patients with PE and are associated with a 2-fold increased risk of death. There is no agreed definition of the echo features of RV dysfunction.


is as effective as standard unfractionated IV heparin and should be given to patients with intermediate or a high pre-test clinical probability immediately, prior to imaging.

Unfractionated heparin

should be considered in massive PE (faster onset of action); first dose bolus prior to commencement of LMWH. Renal impairment (eGFR <30); use either unfractionated heparin or LMWH with anti-factor Xa monitoring. Use unfractionated heparin if risk of bleeding.

Oral anticoagulation

should only be commenced once PE is proven, after initial heparin treatment. Target INR 2.0–3.0 (heparin can be stopped after 5 days or once INR >2). Some centres now advocate outpatient anticoagulation for PE as well as DVT. Recent data suggest that this is as safe as inpatient anticoagulation in a carefully selected population in centres with a well-established outpatient DVT service.

Length of warfarin anticoagulation

  • Temporary provoking risk factor: 3 months

  • First episode of idiopathic PE—review anticoagulation at 3 months. Discuss with patient, including consideration of risk of bleeding and risk of recurrence. Some advocate 6 months’ treatment or lifelong treatment—in which case this decision should be reviewed annually, as the relative risks and benefits of anticoagulation may change

  • For patients with active cancer—6 months of LMWH before a decision as to whether to continue with a vitamin K antagonist (VKA) long term

  • Recurrent idiopathic PE—no guidelines exist; length of treatment depends on individual circumstances, with risk of bleeding balanced with risk of recurrent event, and often long-term anticoagulation

  • Persisting risk factors: lifelong anticoagulation may be recommended

  • New direct thrombin inhibitors, e.g. dabigatran, and factor Xa inhibitors, e.g. rivaroxaban (predictable dose response curves and no need for laboratory monitoring), are now licensed for PE treatment.

Side effects

  • The risk of bleeding increases with age and concurrent illness

  • Higher bleeding rate with concomitant aspirin use and previous GI bleed

  • Risk of bleeding relates to duration and intensity of anticoagulation.


There is emerging evidence to support the use of thrombolysis in certain subgroups of patients with PE; however, this is a controversial area, and the risk/benefit analysis of this treatment must always be carefully considered.

Massive PE

causing circulatory collapse (systolic BP <90mmHg or a pressure drop of 40mmHg with no other explanation). Current NICE guidance (2012) recommends unfractionated heparin and subsequent systemic thrombolysis. In practice, thrombolysis is usually given to the acutely unwell/peri-arrest patient, when the history and physical findings are suggestive of massive PE, in the absence of another reasonable explanation. There is rarely time for imaging or investigations in this situation. See Box 39.1.

Non-massive PE

is more controversial. Most would only recommend thrombolysis for patients with clinically massive PE. Increasing evidence suggests that individuals with a large clot volume, in the absence of haemodynamic compromise, have better clinical outcomes with thrombolysis. This is due to the prevention of chronic thromboembolic disease, as larger clot volume is a risk factor for this. More data are required.


None absolute; rarely a consideration in the life-threatening situation. Risk of major haemorrhage is 3–4 times that of heparin (around 13% in large studies), with a higher incidence of bleeding in the elderly. Active bleeding or recent intracerebral bleed are contraindications.


Rarely done, and only in life-threatening massive PE. Options include surgical embolectomy (few regional centres only) and mechanical clot fragmentation via RHC.

IVC filter placement

There is little evidence to show improved survival or reduction in recurrent PE rate with IVC filters, and changing to LMWH may be as effective. They are potentially pro-thrombotic and should be removed as soon as possible once no longer required. IVC filters may be indicated in:

  • Acute VTE in patients with an absolute contraindication to anticoagulation

  • Patients with recent massive PE who survive (a second PE may be fatal)

  • Recurrent VTE despite adequate anticoagulation

  • Post-pulmonary thromboendarterectomy in PHT.

Further information

Konstantinides S et al. Heparin plus alteplase compared with heparin alone in patients with submassisve pulmonary embolism. N Engl J Med 2002;347:1143–50.Find this resource:

Torbicki A et al. European Society of Cardiology acute PE guidelines. Eur Heart J 2008;29:2276–315. Pulmonary thromboembolic disease this resource:

Special circumstances

Pregnancy and thromboembolic disease

  • The incidence of DVT ± PE in pregnancy is 1 in 1, 000, rising to 2 in 1, 000 in the puerperium. The risk of PE in pregnancy is greater with increasing maternal age and with increasing gestational age. More PEs occur during pregnancy than after delivery. There is a 20–30 times increased risk with Caesarean section, compared with normal vaginal delivery. PE is one of the commonest causes of maternal death in pregnancy (1/100, 000 pregnancies)

  • D-dimers are raised in the normal pregnancy and so are unhelpful in the investigation of thromboembolic disease, unless negative

  • The CTPA whole body radiation dose is 2–4mGy, with an absorbed dose to the foetus of 0.01mSv. This equates to a risk of fatal cancer to age 15 of <1 in 1 million. The absorbed dose to the breast is 10mSv (higher in pregnancy). CTPA increases the lifetime breast cancer risk in premenopausal women from 10% to 11.4%, with an even higher risk in pregnancy

  • The V/Q scan whole body radiation dose is 1.5–2mGy, with an absorbed dose to the foetus of 0.12mSv. This equates to a risk of fatal cancer to age 15 of 1 in 280, 000. The absorbed dose to the breast is 0.28mSv

  • The overall radiation risk depends on the gestation of the foetus and the metabolic activity of the pregnant breast tissue. There is considerable debate as to which imaging technique is best in pregnancy, in terms of radiation risk to both the mother (including breast tissue) and the foetus. The lowest overall risk favours a Q scan as the first-line investigation, especially as this young healthy population are likely to have normal lungs. Some experts suggest a leg US first (see Pulmonary thromboembolic disease p. [link])

  • In those with antenatal thromboembolic disease, LMWH is used. Close to delivery, this is changed to unfractionated heparin, as it is easier to monitor and to reverse its effects. It is unclear whether heparin should be stopped or the dose reduced at the time of delivery. LMWH levels can be monitored with anti-Xa levels

  • There are case reports of successful thrombolysis, catheter-directed thrombolysis, and embolectomy in massive PE, but no relevant trials

  • Warfarin is teratogenic and is contraindicated in pregnancy, although it is safe in breastfeeding

  • Anticoagulation should be continued for 6 weeks after delivery or 3 months following the initial episode, whichever is longer.

Thromboembolic disease and the OCP/HRT

  • Oestrogen-containing OCPs, pregnancy, and HRT increase the risk of PE, but the incidence of fatal PE is low—estimated at 1/100, 000 OCP users, with a median age of 29

  • Risk of fatal PE is twice as high in those taking third-generation pills

  • Previous history of DVT or PE is a contraindication to the OCP

  • Meta-analyses show a relative risk of VTE of 2.1 in HRT users, which is highest in the first year of use.

Flight prophylaxis for thromboembolic disease

  • For patients with high risk of a PE, i.e. previous VTE, within 6 weeks of surgery, or current malignancy, the 1997 BTS guidelines recommend low-dose aspirin, LMWH, or formal anticoagulation (INR 2–3) prior to flying

  • For those with moderate or low risk, graduated or compression stockings, with or without pre-flight aspirin, is suggested.

Future developments

Low-intensity warfarin therapy

A reduced target INR of 1.5–2.0 may be used. A study of a cohort of patients treated for up to 4.3y, following 6 months of standard warfarin therapy for idiopathic VTE, led to a 48% reduction in recurrent VTE, major haemorrhage, or death, compared with placebo (Ridker PM et al. N Engl J Med 2003;348:1425).

Further information

NICE Guidance GC144 June 2012. Venous thromboembolic diseases: the management of venous thromboembolic diseases and the role of thrombophilia testing. Pulmonary thromboembolic disease and summary in BMJ 2012;345:43–5.Find this resource:

Kaeron et al. Chest 2012;141(2)Suppl:e419S–e494S.Find this resource:

Goldhaber SZ. Seminar: pulmonary embolism. Lancet 2004;363:1295–305.Find this resource:

Rare causes

Air embolism

Air within the arterial or venous circulation. Small amounts of air can be tolerated, but large amounts can lodge in the pulmonary vasculature and cause mechanical obstruction and death. This is rare.


Neck vein cannulation, intrauterine manipulations (such as criminal abortion where a frothy liquid is passed under pressure into the uterus), bronchial trauma, or barotrauma causing air to enter the pulmonary vein and left heart. Air in the LV causes impairment to venous filling and subsequent poor coronary perfusion as air enters the coronary arteries.


Arterial air emboli may cause dizziness, loss of consciousness, and convulsions. Air may be seen in the retinal arteries or from transected vessels. Venous air emboli may cause raised venous pressure, cyanosis, hypotension, tachycardia, syncope, and a ‘mill-wheel’ murmur over the praecordium.


Patients should lie on their right side, with head down and feet up, to allow air to collect and stay at the cardiac apex. From here, it can be aspirated via thoracotomy.

Amniotic fluid embolism

is estimated to occur in 1 in 25, 000–80, 000 live births. It is the third commonest cause of maternal death, and the most common cause of death in the immediate post-partum period. Usually catastrophic, 80% of women die, 20–50% of these in the first hour. An anaphylactic-type response to amniotic fluid entering the circulation is seen. Amniotic fluid enters the circulation because of torn foetal membranes, which can occur in Caesarean section, uterine or cervical trauma, or uterine rupture. It has a thromboplastic effect, causing DIC and thrombi to form in pulmonary vessels. Not all women react in this way to amniotic fluid. It is more common in older multiparous mothers, who have had short tumultuous labour, often involving uterine stimulants.


presents with sudden-onset respiratory distress, hypoxia, bronchospasm, cyanosis, cardiovascular collapse, pulmonary oedema, convulsions, coma, and cardiac arrest. Coagulopathy with intractable uterine bleeding and uterine atony is seen.


is clinical. Foetal debris/cells can be identified in blood sampled from the maternal pulmonary artery, but this is not pathognomonic.


is supportive, whilst the thrombi clear from the maternal lungs. Maintain the circulation with fluids and inotropes. Respiratory support with O2 and ventilation may be needed. Correct coagulopathy with fresh frozen plasma and packed cells. Control placental bleeding.

Fat embolism

Common pathological finding following long bone fractures. Occurs especially with lower limb fractures—pelvis and femur. Commoner in fractures that have not been immobilized. Can also occur after prosthetic joint replacement, cardiac massage, liver trauma, burns, bone marrow transplant, rapid high-altitude decompression, and liposuction. Generally occurs in the young and previously healthy. Presents 24–72h post-fracture. Marrow fat enters the circulation and lodges in the lungs, causing mechanical obstruction.


presents with hypoxia, coagulopathy, with a transient petechial rash on the neck, axillae, and skinfolds, and neurological disturbance such as confusion, disorientation, or sometimes coma. Stable patients may deteriorate with low-grade fever, petechial rash, hypoxia, and confusion. Jaundice and renal dysfunction are possible.


is usually made clinically in a patient with a lower limb fracture presenting with tachypnoea and hypoxia. Fat globules can be identified in the urine. CXR shows bilateral alveolar infiltrates. ARDS can develop.


is with early immobilization of fracture, fluid replacement, O2, and supportive care.

Septic, hydatid, and tumour emboli

are also rare causes. Uterine leiomyosarcoma has vascular tropism and can invade the IVC and obstruct the pulmonary arteries. Teratomas can invade the IVC and pulmonary arteries.