A negative D-dimer result in a patient presenting with signs and symptoms in keeping with venous thromboembolism but with a low clinical probability is sufficient to exclude the diagnosis Ventilation perfusion isotope scanning, although widely available, should only be used in patients with suspected pulmonary embolism who have a normal chest radiograph or no previous chronic cardiopulmonary disease Studies suggest good sensitivity and specificity of spiral computerised tomography (CT) in the diagnosis of central or segmental pulmonary emboli All hospitals receiving acute medical emergencies should devise their own algorithm for diagnosis of venothromboembolism that reflects local expertise and availability of investigations
Key Pointsare no absolute contraindications, although particular care should be exercised in patients with known sensitivity to contrast media and in those with severe pulmonary hypertension, renal impairment or following acute MI. The femoral vein approach is commonly used, but some prefer the internal jugular or subclavian approach because of the reduced risk of disturbing thrombi and the ability to maintain venous access for pressure monitoring and the administration of thrombolytic drugs where indicated. Good liaison between the radiologist and intensive care unit is recommended. Pigtail catheters of sufficient size (7F) to enable high-flow injections of non-ionic contrast media should be used. Volume and flow will depend on facilities available. Superselective injections may be necessary. A main PA injection may be sufficient when prior echocardiography suggests the possibility of a large centrally placed clot. Where prior V/Q scan is non-diagnostic, angiography can be confined to the more abnormal side 65 .Minor complications have been reported in 2% and major or fatal complications in 0.5-1.3% of investigations 66 using these techniques, mainly in patients already severely ill. The introduction of low osmolar non-ionic contrast media has led to a reduction in complications, a recent report finding only one major (non-fatal) complication in every 300 patients 67 .There may be difficulties in interpretation of PA results, even by experienced radiologists. The PIOPED study found 19% interobserver disagreement, varying from 2% for central to 34% for subsegmental abnormalities 68 , and 11% within-observer discrepancy. A recent study with consensus review in patients with non-diagnostic lung scans led to a change in initial diagnosis in 20%. There was better agreement when digital subtraction was used, but this is not widely available.
RecommendationsThe algorithm shown in Fig 1 outlines one approach to the investigation of a patient presenting with suspected VTE.
ConclusionsThe investigation of suspected VTE will continue to be refined by further evidence from prospective studies using D-dimers and spiral CT. Moreover, newer modalities such as magnetic resonance imaging (MRI) are beginning to demonstrate diagnostic value, particularly in North America. However, at present, routine use of M...