The diagnostic difficulty associated with acute venous thromboembolism (i.e., deep venous thrombosis, pulmonary embolism, or both) is underscored by the frequency with which patients die before the diagnosis is made.¹ Pulmonary embolism (PE) occurs in hundreds of thousands of patients in the United States each year and frequently occurs in patients with cancer, chronic cardiopulmonary disease or trauma, or in postoperative patients, particularly those who are non-ambulatory. Multiple risk factors involving venous stasis, thrombophilia and venous injury (Virchow's triad) significantly increase the risk of acute venous thromboembolism.

Patients may present with PE (even massive PE) without any antecedent symptoms or signs of deep venous thrombosis (DVT), even though the vast majority of cases of acute PE result from DVT in the proximal veins of the lower extremities (above the knee). Unfortunately, the symptoms and signs associated with DVT and with PE are nonspecific. Thus, a high index of suspicion is crucial for patients who are deemed to be at risk. Because of the magnitude of this problem and the varying diagnostic approaches available, the American Thoracic Society (ATS) developed a consensus statement to outline appropriate strategies to consider in patients with suspected DVT and PE. Importantly, the medical literature was reviewed and an evidence-based approach was devised,² with recommendations made based on available data.

In September 1999, the ATS Consensus Statement and Clinical Practice Guidelines were published.³ Although the guidelines were developed by the ATS, the majority of patients with suspected DVT and PE do not initially present to pulmonologists but instead visit other subspecialists, emergency department physicians, surgeons, internists and family physicians. Thus, it would appear prudent to briefly outline some important issues regarding management of acute venous thromboembolism.

The diagnostic approach to DVT and PE requires careful attention to the presence of risk factors. Although a patient who develops shortness of breath during bedrest for a hip fracture might have aspiration pneumonia, a flare of obstructive lung disease or atelectasis causing dyspnea, PE should always be included in the differential diagnosis. Even when a patient is admitted to the hospital with community-acquired pneumonia, congestive heart failure or emphysema, the development of worsening dyspnea always merits consideration of PE despite another strong explanation for the symptoms. The presence of symptoms such as leg pain or swelling, dyspnea, chest pain, syncope or hemoptysis, particularly in the setting of risk factors for DVT, should always lead to suspicion of DVT and PE.

What should be the initial diagnostic approach to patients with suspected DVT? In most hospitals in the United States and Europe, compression ultrasound (often with Doppler and color-flow) is the most common initial approach. The sensitivity and specificity of this test are over 95 percent for identifying DVT of the proximal leg vein. For DVT of the calf, however, the test is less reliable. The use of contrast venography or, in centers with experienced staff, magnetic resonance imaging (MRI) is appropriate when clinical suspicion of DVT remains high but the ultrasound is negative or nondiagnostic.

At the present time, the standard approach to patients with suspected acute PE should include a history, physical examination, arterial blood gas analysis, chest radiography and electrocardiography. The chest radiography and electrocardiography may be useful in proving alternative diagnoses but rarely are diagnostic of PE. Unexplained hypoxemia may suggest PE but is neither sensitive nor specific for the diagnosis.

The ventilation perfusion (V/Q) scan has long been the diagnostic cornerstone. However, this technique is usually not definitively negative or positive and, generally, additional testing is indicated. Even when PE is ultimately proved by pulmonary angiography (the gold standard), the V/Q scan is commonly nondiagnostic.⁴ The V/Q scan is useful, however, when results are normal (proving the absence of PE) or high probability (proving PE except under unusual circumstances).

Contrast-enhanced spiral computed tomography (CT) with a contrast bolus is becoming increasingly used for the diagnosis of suspected PE.^5,6 This technique appears useful for PE that occurs in the main, interlobar and segmental pulmonary arterial branches but is less sensitive and specific for clots occurring in subsegmental branches. Reader expertise is crucial, but the technology is improving and further clinical trials appear to be forthcoming.⁷

Caution is recommended with interpretation of CT studies, and the threshold for pulmonary arteriography should be low. The quality of the CT images must be optimal. Some experts, however, are still reluctant to rely on CT in the absence of additional clinical research trials.⁸ When the diagnosis is uncertain after chest CT or a V/Q scan, or both, evaluation of the lower extremities appears to be a useful option.⁹ If DVT is present, the treatment is generally the same as if PE is diagnosed.

Finally, the D-dimer, a product of cross-linked fibrin degradation, has been evaluated increasingly in diagnosis of suspected DVT and PE and appears promising.¹⁰ Although the enzyme-linked immunosorbent assay (ELISA) D-dimer test has appeared to be sensitive for venous thromboembolism in several clinical trials, none of the available assays are specific for the diagnosis. Positive tests are common and therefore not helpful, particularly in persons with cancer, infections or inflammatory disease states. The development of more rapid bedside D-dimer assays, especially used in combination with other diagnostic tests, may prove useful, but the wide variability among the assays continues to limit their diagnostic utility.

Venous thromboembolism results in a substantial number of what appear to be potentially unnecessary deaths. Our increasingly sophisticated diagnostic repertoire, together with carefully designed clinical trials and interpretation of the available data in the form of clinical guidelines, should enhance our diagnostic yield of DVT and PE. Nonetheless, clinical suspicion remains an absolutely essential component of the diagnostic approach.