Pulmonary embolism

The management of suspected pulmonary embolism is a subject in which consensus has been difficult to achieve due to the lack of firm evidence.¹ Against this background, the British Thoracic Society has suggested a “practical approach to suspected acute pulmonary embolism.”² After a comprehensive review of current literature, the report makes several recommendations, graded for evidence, and finishes with a series of charts designed for a junior doctor’s handbook covering diagnosis and management. While some of the recommendations lack top grade evidence and inevitably will be controversial, the authors achieve what they set out to do and provide a practical approach to a difficult subject using new knowledge “little of which,” as the report bluntly states, “has filtered through to clinical practice.” Certain items in this comprehensive report are worth emphasising.

Pulmonary embolism is both underdiagnosed and overdiagnosed. Ten per cent of all hospital deaths are due to acute pulmonary embolism, most of which are diagnosed only at necroscopy. It is easily missed in patients with cardiorespiratory disease (in whom even small emboli can be fatal), elderly patients, and those presenting with isolated dyspnoea. On the other hand, only a third of patients with clinically suspected pulmonary embolism have positive pulmonary angiograms. Clinical signs and symptoms are highly non-specific, but in the absence of key features—namely, dyspnoea with tachypnoea and/or pleuritic chest pain—the diagnosis is highly unlikely (less than 3% of cases). Of patients with a proved pulmonary embolus, 80%-90% have a major predisposing risk factor. The report has put together a simple scheme for assessing the clinical likelihood of an embolus based on assessment of the clinical features and the presence or absence of a major risk factor. The authors point out that oral contraception is not a major risk factor. The scheme still needs to be validated, but it should help direct immediate treatment and guide subsequent investigations, while an assessment of clinical likelihood assists interpretation of the isotope lung scan.

Basic investigations, like signs and symptoms, are non-specific, but an electrocardiogram and chest x ray should be performed if only to exclude other diagnoses. Arterial blood gases are deemed mandatory, although the evidence for performing this unpleasant test is not convincing³ and is unlikely to alter immediate management.

For a stable patient, the next diagnostic investigation is the isotope lung scan. The report, on good evidence, states that the ventilation phase rarely adds to the accuracy of the scan, which should be good news to clinicians without access to ventilation scanning. If the ventilation phase is to be used the report suggests that krypton-81m or a technetium-99m DTPA aerosol be used as they give clearer multiview images compared with the unidimensional view of the more commonly used xenon-133, perhaps resulting in fewer intermediate scan reports. Seventy per cent of lung scans are reported as being intermediate or representing a 16%-66% risk of pulmonary embolism depending on the clinical likelihood. The next recommended investigation is a Doppler ultrasound scan of the leg veins, and a consensus of opinion is developing to support this position. If the ultrasound scan is negative various management options are suggested, including pulmonary angiography. This is a difficult area for evidence based guidelines, and has recently been reviewed.⁴

For a patient presenting with collapse or hypotension, the first investigation recommended is echocardiography, which can show well defined abnormalities in patients with large central pulmonary embolism. Alternatively, it may show another cause for the clinical presentation. If it is inconclusive, pulmonary angiography or spiral computed tomography should be considered, although the report concedes that the precise investigations used in this situation will depend on local availability and expertise. Spiral computed tomography is highly sensitive for proximal emboli down to the segmental arteries, and smaller emboli are unlikely to present as collapse. As the technology becomes more widely available, spiral computed tomography will probably be the preferred test in this situation, but, as the report makes clear, in the absence of anything else a perfusion scan alone can be very helpful. Thrombolysis is the proposed treatment of choice for large central emboli presenting as collapse or hypotension. The simplest regimen recommended, though untested in this clinical situation, is an infusion of alteplase 100 mg over 2 hours.

Perhaps one of the most important recommendations made in the report is that hospitals should develop their own strategies for the diagnosis and management of pulmonary embolism particularly in unstable patients, when time is at a premium. This will require discussion and cooperation between clinicians and radiologists. The report makes a plea for more widespread access to pulmonary arteriography, but this is unlikely to occur as there is a real prospect that Doppler ultrasound and spiral computed tomography, alone or in combination, will make angiography unnecessary in most cases.⁵ The report introduces the idea that pulmonary embolism in low risk patients may be a relatively benign condition that does not require extensive investigation. Studies of cost effective non-invasive strategies for diagnosis will have to take this into account. Appropriate endpoints for studies would be the outcome for patients if treatment is withheld on the basis of negative tests rather than relating these tests to the results of angiography.

This excellent document should be read by all clinicians in emergency medicine together with their radiological colleagues, and the section for junior doctors’ handbooks would be an ideal topic for your next clinical meeting. Hopefully, this report will improve patient care, as well as stimulate further research.