Investigations in the diagnosis of pulmonary embolus

Most of the routine investigations that a patient will have when presenting to hospital are non‐specific and therefore not useful on their own in making or excluding a diagnosis of PE.

The ECG is often normal. The commonest ECG abnormality is sinus tachycardia. Other ECG changes, which occur in 70% of patients with a PE, include right heart strain, right axis deviation, depression of the ST segment, and T wave inversion in leads V1–V3. The S1Q3T3 pattern occurs in less than 10% of patients (Figure 11.2). Patients who develop bradycardia, atrial arrhythmias, new right bundle branch block, inferior Q‐waves, and anterior ST‐segment changes have a worse prognosis.

The chest X‐ray (CXR) is normal in approximately 20%, and is an essential investigation to exclude pneumothorax, consolidation, and cardiac failure. A small pleural effusion is found in 47% of patients with a PE; this is often blood‐stained if aspirated. Other radiological changes include atelectasis, pruning of the pulmonary vasculature with distal hypoperfusion, and a wedge‐shaped opacity in the lung periphery (Figure 11.3).

Arterial blood gas (ABG) analysis is not a sensitive or specific test in the diagnosis of PE, but 74% of patients will be hypoxic. Approximately 41% will have hypocapnia and a respiratory alkalosis. PE should be considered in anyone who has a normal CXR and unexplained hypoxaemia. Ventilation perfusion mismatch will result in widening of the Alveolar‐arterial (A‐a)gradient in the majority. The calculation of the A‐a gradient is described in Chapter 13. Although not helpful on its own to make a diagnosis, the ABG at presentation may be of prognostic value. As patients with an initial oxygen saturation of less than 95% have an increased risk of respiratory failure and death, it is recommended that such patients are admitted to hospital for careful monitoring while undergoing investigations and treatment.

D‐dimer is a breakdown product of cross‐linked fibrin and levels will be elevated in patients with thromboembolism. Sensitive D‐dimer testing using ELISA (enzyme‐linked immunosorbent assay) is recommended. Although it is a sensitive test, with a greater increase in those with larger PEs, it lacks specificity. D‐dimer levels will be elevated in those with any acute illness and in pregnant women. D‐dimer levels will be falsely positive in patients with chronic renal failure with an estimated glomerular filtration rate (eGFR) <60 ml min⁻¹/1.73 m². The D‐dimer level also increases gradually in patients over the age of 50 years. Age‐adjusted D‐dimer values may increase the specificity of the test, but this is not routinely done in the UK.

The D‐dimer level is only useful in excluding a PE and should not be used to make a diagnosis of PE. According to the NICE guidelines, the D‐dimer should be used in conjunction with the modified Wells score to determine the need for further investigations. If the modified Wells score is greater than or equal to 4, then the patient should go on to have further investigations to confirm the diagnosis of PE, regardless of the D‐dimer level. In those with a high clinical suspicion of PE and a normal D‐dimer level, the prevalence of PE is 20–28%.

If the probability of PE is considered unlikely (modified Wells score of less than 4), then a D‐dimer level should be obtained. If this is negative, then no further testing is required. In those in whom the D‐dimer level is elevated, a CTPA is required.

Although not sensitive or specific, serum Troponin I and T levels may indicate right ventricular dysfunction. Raised levels may be elevated in 30–50% of patients with a large PE and may be of prognostic value. The levels are rarely as high as would be after a myocardial infarction and return to normal within 2 days.

Imaging to confirm a diagnosis of PE

Computed tomography pulmonary angiogram (CTPA)

CTPA with intravenous contrast is a rapid test that is available in all hospitals in the UK. CTPA is the imaging of choice in a non‐pregnant patient with normal renal function who is haemodynamically stable and not allergic to contrast. CTPA may not be the optimal investigation in the morbidly obese patient and in women under the age of 40 because of the high dose of radiation to the breasts, which may increase the risk of breast cancer.

An algorithmic approach which combines CTPA with clinical assessment and D‐dimer levels increases the sensitivity and specificity of the test. CTPA has a sensitivity of over 90% for the diagnosis of PE, which increases to 96% when combined with a clinical probability assessment. The specificity of CTPA is 95%. When the modified Wells score is <2, the positive predictive value (PPV) of CTPA is 58%. If the Wells score is 2–6, then the PPV is 92% and for a Wells score >6, the PPV rises to 96%.

A PE will appear as a filling defect in a branch of the pulmonary artery which is otherwise filled with contrast (Figure 11.4). CTPA is most accurate for the detection of a large PE blocking the main, lobar, and segmental pulmonary arteries. It is less accurate for detecting smaller, peripheral, sub‐segmental PEs. The modern multi‐detector scanners can detect smaller, more peripheral emboli. The CTPA also has the advantage of finding other abnormalities which may be responsible for the clinical symptoms and signs.

A positive CTPA will confirm a diagnosis of PE and a negative CTPA means that a PE is unlikely. When the clinical suspicion is high but the CTPA is negative, 5% will have a PE. Therefore, patients with a high Wells score and a negative CTPA may require further investigations, which may include a VQ scan or a contrast‐enhanced pulmonary angiogram.

Ventilation perfusion scan (VQ scan)

A VQ scan should be considered in any patient in whom a CTPA is contraindicated as discussed above. It should also be considered in women under the age of 40. A normal CXR is necessary when interpreting the VQ images and is, therefore, not a suitable investigation in a patient with chronic lung disease. A VQ scan may occasionally be indicated if the clinical suspicion of a PE is high but the CTPA is negative. A VQ scan is a nuclear medicine scan that is not available in all centres and not available out of hours because radioactive isotopes are required.

The PIOPED 11 study is the largest study to date which looked at the sensitivity and specificity of VQ scanning. A VQ scan has a moderately high sensitivity but a poor specificity, with a high number of false positive test results. As with CTPA, diagnostic accuracy was greater when the results of the VQ scan was combined with a clinical probability score.

A VQ scan is reported according to whether there are areas which have normal ventilation but abnormal perfusion (VQ mismatch). Patients with underlying lung disease, for example, COPD, will have matched ventilation and perfusion defects. A VQ scan can be reported as normal, low‐probability of PE, intermediate probability of PE, or high‐probability of PE.

A normal VQ scan means that a PE is unlikely, and no further investigations are required. A patient with a Wells score <2 and a normal or low‐probability VQ scan will have <4% chance of having a PE. If this is combined with a normal D‐dimer level, then the chance of a PE is <3%. A high probability VQ scan in a patient with a high clinical probability score means that there is a 96% chance of a PE (Figure 11.5). Patients with a low probability or inconclusive VQ scan will need further investigations (Figure 11.6).

Patients with a high clinical suspicion of PE in whom a CTPA is either negative or contra‐indicated and in whom the VQ scan is inconclusive will require further imaging.

A contrast‐enhanced pulmonary angiogram is historically the definitive test for diagnosing PE and has a good sensitivity and specificity. Although it is an invasive test and is associated with a small risk of harm, it is safe and well tolerated in haemodynamically stable patients with <2% mortality. Complications include catheter‐related events, contrast‐related complications, and cardiac complications. One advantage of this test is that if a clot is directly visualised, it can be lysed by embolectomy and/or thrombolysis if anticoagulation is contra‐indicated.

A magnetic resonance pulmonary angiogram (MRPA) is less sensitive and specific and is rarely used. Proximal vein compressive lower‐extremity ultrasound (CUS) can detect a DVT so can indirectly make a diagnosis of PE. It is not recommended in routine practice for diagnosing PE as only 9–12% of patients with PE are found to have a DVT by this method. However, in those in whom other investigations are contra‐indicated, serial CUS done weekly for several weeks could be useful to detect DVT if the clinical suspicion is high. It is a valuable test in pregnant women as there is no exposure to radiation.

A transthoracic echocardiogram cannot make a diagnosis of PE, but in 30–40% of patients with a PE there will be changes consistent with right ventricular strain, which includes regional wall motion abnormalities that spare the right ventricular apex. In severe PE, there may be evidence of elevated right ventricular pressures, an increase in right ventricular size, tricuspid regurgitation, and pulmonary hypertension. In 4% of cases, a thrombus may be seen in the right ventricle, which confers a poor prognosis. The echo changes may be of prognostic value and resolution of changes can be used to monitor improvement with anticoagulation and, sometimes, to guide the length of anticoagulation. An echocardiogram can also diagnose other causes of hypotension and cardiovascular collapse, including aortic dissection and pericardial tamponade.

PE is a leading cause of mortality during pregnancy and in the 6 weeks post‐partum, accounting for 20–30% of maternal deaths. It is difficult for the clinician to calculate the clinical probability of a pregnant woman having a PE as there are no validated scores in this group of patients. The imaging to use in a pregnant woman often causes much concern for the doctor and the patient. All pregnant women who present with possible PE should have a CXR (with lead protection for the foetus) which may suggest an alternative diagnosis. Both CTPA and VQ scan will expose the foetus to some radiation. CTPA exposes the mother’s breasts to a significant dose of radiation at a time when they are particularly metabolic, thus increasing the future risk of breast malignancy. It is recommended that a CUS of legs and pelvis is a useful initial investigation in a pregnant woman if PE is suspected. If this is normal but the presentation is suggestive of a PE, then a half‐dose perfusion scan is recommended. CTPA is reserved for pregnant women who are clinically unwell and in whom other investigations are indeterminate.

Anticoagulation

Anticoagulation is the main treatment for PE. The risk of PE recurrence is 25% in patients with a high probability score and anticoagulation has been shown to reduce this. The main complication of anticoagulation is bleeding, and intracranial bleeding may be life‐threatening. The risk of bleeding is estimated to be 1.6% in the first 3 months in those with no risk factors for bleeding and will be up to 3% in those with risk factors. Minor haemoptysis, epistaxis, and menstruation are not contraindications to anticoagulation. Anticoagulation has also been shown to reduce mortality, the benefits outweighing the risk of major bleeding.

The aim of anticoagulation is to reach a therapeutic level within 24 hours of treatment using either LMWH, subcutaneous fondaparinux, intravenous unfractionated heparin (IVUFH), or subcutaneous unfractionated heparin (SCUFH). A patient diagnosed with PE with a high risk of haemorrhage should be discussed with an expert prior to anticoagulation.

LMWH is recommended in haemodynamically stable patients with normal renal function. It is not indicated in patients who are morbidly obese as there is decreased absorption of medication given subcutaneously. Advantages of LMWH over IVUFH include lower mortality, fewer recurrent thromboembolic events, fewer major bleeding episodes, and a lower incidence of heparin induced thrombocytopaenia (HIT). LMWH has more predictable pharmacokinetics than UFH, requires twice daily administration of a fixed dose, and monitoring of anti‐Xa levels is not required. The choice of which LMWH to use will be dictated by the cost and clinical experience. The dose is calculated according to the patient’s weight and given subcutaneously by injection.

LMWH is also recommended for the treatment of PE in a pregnant woman and more careful monitoring is recommended. Anticoagulation should be continued for 3 months after birth if pregnancy is the only risk factor for developing the PE. Those with other risk factors may need a longer period of anticoagulation. Warfarin is teratogenic so is contraindicated in pregnancy, particularly in the first trimester. Warfarin is, however, considered to be safe in breastfeeding mothers.

IVUFH is recommended in patients with massive PE and hypotension as they may require thrombolysis and the effects of the UFH can be reversed with protamine sulphate more rapidly than when patients receive LMWH or fondaparinux. IVUFH is also indicated in those in whom there is an increased risk of bleeding, those with renal failure (creatinine clearance less than 30 ml min⁻¹) and in the morbidly obese. Patients on UFH will require monitoring of their activated partial thromboplastin time (APTT).

Oral anticoagulation

Warfarin, a vitamin K antagonist, which blocks the production of the vitamin‐K dependent clotting factors (11, V11, 1X and X), is the drug most commonly used for the long term treatment of PE. It is effective in preventing recurrent PEs and DVTs. Warfarin can be started as soon as the diagnosis of PE is confirmed while the patient is on the treatment dose of LMWH but should not be started without prior treatment with LMWH as there is evidence that this may increase the incidence of PE and/or DVT. It is recommended that LMWH treatment should continue for at least 5 days after starting treatment and until the International Normalised Ratio (INR) has been therapeutic (between 2 and 3) for at least 24 hours. This is because it takes at least 5 days for the intrinsic clotting pathway activity to be suppressed. It is recommended that the starting dose of warfarin should be 5 mg for 2 days and then the dose calculated according to the INR. The effects of warfarin can be reversed by giving vitamin K. Fresh frozen plasma can also be given if necessary.

Warfarin is a cheap drug but has a narrow therapeutic range and requires monitoring. Warfarin is a drug that has interactions with other commonly used drugs which are metabolised through the cytochrome P450 system. Doctors should be aware of these interactions. Certain food items, particularly those containing vitamin K, can also alter warfarin levels, so patients should be given information booklets with details of foods to avoid.

Increasingly, Factor Xa inhibitors, such as rivoroxaban, apixaban, and edoxaban are being used. Dabigatran, a direct thrombin inhibitor, is also being used in certain circumstances. These are fixed dose agents that do not require monitoring. However, the effects cannot be easily reversed. It is not within the scope of this book to discuss these newer anticoagulants in detail.

Patients with PE who are haemodynamically stable, not hypoxaemic, not in respiratory distress, who do not have significant co‐morbidities, no increased risk of bleeding, and who do not live alone can be safely anticoagulated at home.

Duration of anticoagulation

The length of time that anticoagulation should be continued depends on the underlying cause of the PE. Rates of clot resolution with anticoagulant therapy are variable. It is estimated that there is resolution of the PE in 40% of patients within 1 week, in 50% within 2 weeks and in 73% within 4 weeks. Thrombi can also move during anticoagulation.

If the DVT and/or PE is due to an identifiable risk factor, such as immobility or surgery, the guidelines recommend 3 months of anticoagulation, so long as the INR is therapeutic during this period. The patient should be reviewed after this period to ensure that the symptoms have resolved and that there is no evidence of pulmonary hypertension.

Patients who have an ongoing risk of thromboembolism, such as an inherited clotting disorder, will require lifelong anticoagulation. Patients with an unprovoked PE, with no obvious risk factors, should have a thorough clinical assessment and appropriate investigations to exclude malignancy. They may require life‐long anticoagulation as the risk of recurrence is 25% at 5 years without anticoagulation. The risk of bleeding is estimated to be 1.2% at 5 years. The risk of PE recurrence if anticoagulation is stopped, together with the risk of bleeding with continuing anticoagulation, should be discussed.

Patients with malignancy have an increased risk of PE. LMWH is recommended for patients with malignancy who develop PE. These patients also have an increased risk of bleeding, so the decision as to which anticoagulation to use, and for how long, must be made after weighing up the pros and cons for each patient.

Inferior vena cava filter

An IVC filter should be considered in anyone with a diagnosis of PE who has a significant risk of haemorrhage if commenced on anticoagulation. This includes those who are more than 65 years old, those with recent surgery, known haematological risk factors, liver failure, and malignancy. Patients who have extensive DVT and pelvic malignancy may also have recurrent episodes of PE as the clot breaks off and travels to the lungs. Retrievable filters are recommended, and the majority are placed infra‐renally to prevent further emboli from reaching the lungs. This is usually a temporary solution and the filter will need to be removed once anticoagulation has been optimised.

Management of massive life‐threatening pulmonary embolus

Approximately 8% of patients with a PE present with shock and collapse. Patients who have a systolic blood pressure of less than 90 mmHg may not be well enough for a CTPA to confirm the diagnosis but must rely on a bedside transthoracic echocardiogram which will show signs of right heart strain. Patients who present with a suspected massive pulmonary embolus with haemodynamic compromise, signs of right heart strain on transthoracic echocardiogram and or bilateral or saddle embolus on CTPA should be thrombolysed.

Patients should have immediate, but careful, intravenous fluid resuscitation, oxygen therapy to maintain the oxygen saturation between 94% and 98% and vasopressor support. If the intravenous fluid is given too aggressively, there is a risk of right ventricular failure. While waiting for confirmation of a PE, the patient should be commenced on IVUFH. Patients should receive 50 mg of alteplase as a bolus via a peripheral vein followed by intravenous heparin infusion. The activated partial thromboplastin time (APTT) should be maintained at between 1.5–2.5 times normal. Analgesia will be required for pain. Oral anticoagulation, usually with warfarin, should be started with a loading dose of 10 mg, with an aim to maintain the INR between 2 and 3.

Thrombolysis can increase the risk of cerebral and pulmonary haemorrhage. The decision to thrombolyse should be made by a senior respiratory physician after consultation with a radiologist and intensivists. Such patients should be managed in a high dependency unit (HDU) or intensive care unit (ICU).

If thrombolysis is contra‐indicated, for example, in a pregnant patient, or it fails, then catheter‐directed embolectomy or surgical embolectomy should be considered. These procedures are only available in tertiary centres and are associated with a high mortality.

Prognosis after acute PE

The overall mortality without treatment is 30% but reduces to 2–11% with anticoagulation. The risk of death is highest in the first week due to cardiogenic shock. The risk of recurrent PE is also greatest in the first 2 weeks. In the longer term, mortality is due to the underlying condition that caused the PE, such a malignancy.

The Pulmonary Embolism Severity Index (PESI) can be used to calculate the risk of death. Poor prognostic factors include age more than 65 years, co‐morbid conditions, shock, right ventricular failure, hypoxaemia, thrombus in the right ventricle, elevated brain natriuretic peptide and N‐terminal pro‐brain natriuretic peptide, and elevated troponin I and T levels.

Recurrent pulmonary emboli

Patients with recurrent, acute PE may require life‐long anticoagulation. Compliance with treatment should be checked, ensuring that the INR is therapeutic. Some patients, especially those with malignancy or pelvic DVTs, may have recurrent pulmonary emboli despite anticoagulation. In some patients it can be difficult to maintain the INR in the therapeutic range. In these patients, and those in whom anticoagulation is contra‐indicated, an inferior vena cava filter should be considered. If a patient with a known diagnosis of PE, who is already being anticoagulated, presents with symptoms and signs of a PE, the same diagnostic approach should be taken. Images should be carefully reviewed by the radiologist as interpretation may be difficult.

Chronic pulmonary emboli

Patients with chronic pulmonary emboli will present with progressively worsening breathlessness and clinical features of pulmonary hypertension, which includes raised JVP, peripheral oedema, and parasternal heave. The ECG may show right ventricular hypertrophy and right axis deviation. The CXR will show prominent pulmonary arteries. A VQ scan will demonstrate unmatched defects. These patients are at risk of developing chronic thromboembolic pulmonary hypertension (CTEPH), which will be discussed in the next section.