On diagnostic procedures in pulmonary embolism

Abstract

Pulmonary embolism (PE) and deep venous thrombosis (DVT) are two entities of venous thromboembolism (VTE). PE is a common cause of mortality and morbidity. The symptoms and signs of PE are difficult to interpret, and the diagnosis is difficult to establish. Other diseases can have a similar clinical presentation as PE, increasing the risk for misdiagnosis and inappropriate treatment. A rapid and prompt diagnosis of PE is crucial and has an impact on the mortality rate.

In Paper I, we studied 90 hemodynamically stable patients with suspected PE. Computed tomography of the pulmonary arteries (CTPA), and pulmonary arteriography (PA) were performed, the D-dimer was measured, and the diagnostic accuracy of these tests was compared. The prevalence of PE was 37%. CTPA had a positive predictive value (PPV) of 94% and a negative predictive value (NPV) of 95%. With a cut-off of D-dimer of 0.25 mg/L, the PPV was 63% and the NPV was 92%. We concluded that CTPA has a higher sensitivity than D-dimer for diagnosing PE, and that a cut-off value of 0.25 mg/L is sufficient as a screening method, although CTPA is recommended to exclude false positive cases.

In Paper II, we retrospectively examined the symptoms and signs in a group of patients with PE (PE/infection, n=25), who were misdiagnosed initially with pneumonia and compared these with the symptoms and signs of patients with correct diagnoses of PE, (PE/medicine, n=64) and community-acquired pneumonia, (CAP, n=54). Dyspnoea and pleuritic chest pain as initial symptoms contributed to 76% (PE/infection), 81% (PE/medicine) and 9% (CAP). Fever, chills, or cough contributed to 8%, 16%, and 80%, respectively. Body temperature, Creactive protein (CRP), and the presence of pulmonary infiltrates on chest X-rays were higher in PE/infection and CAP than in PE/medicine. The symptoms did not differ significantly between the two PE groups. Dyspnoea and pleuritic chest pain were the predominant initial symptoms in patients with PE whereas fever and cough were the predominant symptoms in patients with CAP. High levels of CRP and the presence of pulmonary infiltrates did not exclude PE.

In Paper III, we studied 151 patients with suspected DVT and PE with a Wells pretest probability score ≤1.5 points and a D-dimer <0.5 mg/L, to determine whether the score and test safely exclude VTE and whether a follow-up D-dimer test measured 3-7 days after admission adds extra information. The patients were followed for three months. A follow-up D-dimer was available for 67% of the patients and was elevated in 13%. None of these patients had a VTE diagnosed. All patients were contacted after three months, and none had clinical signs of VTE. We concluded that a low Wells score and a normal D-dimer safely exclude VTE in patients at the emergency department (ED) and that a follow-up D-dimer test adds no extra information.

In Paper IV, we evaluated the diagnostic accuracy of the Wells score, D-dimer, and PA or CTPA to diagnose 120 patients with an intermediate to high risk of PE. The cut-off D-dimer level of 0.5 mg/L was adequate with an NPV of 92%. The combination of the Wells score and D-dimer gave an NPV of 94%, and a cut-off of Wells score of four points was adequate. We concluded that D-dimer and Wells score are safe methods of ruling out PE in patients with intermediate to high risk of PE, although the specificity is low. Both CTPA and PA can yield false negative and false positive results, which can be difficult to interpret.

In Paper V, we compared (a) the level of von Willebrand factor antigen (vWF) on inclusion and after three months in 46 patients with symptomatic PE without signs of DVT, with (b) the level in 45 patients with symptomatic DVT without signs of PE. The mean level of vWF was 1.87 IU/mL in the PE group and 1.64 IU/mL in the DVT group at inclusion. Patients with proximal DVT had a mean level of 1.88 IU/mL and patients with distal DVT had a mean level of 1.52 IU/mL. After three months, the mean level of vWF was 1.45 IU/mL (PE), 1.41 (all DVT), 1.65 (proximal DVT) and 1.28 (distal DVT), respectively. These findings suggest that the level of vWF differs between PE and DVT. The levels of vWF is similar in patients with PE and proximal DVT, whose vWF level is significantly higher than in patients with distal DVT. This difference may indicate that vWF reflects the extention of a DVT.

Conclusions These studies show that the diagnostic accuracy of CTPA is high and better than D-dimer in diagnosing PE, and that PA and CTPA can produce different results. The initial symptoms of PE are often typical and can be used to discriminate PE from pneumonia. High levels of CRP and body temperature, and findings of pulmonary infiltrates are common in PE. Wells score and D-dimer have a high NPV and can safely exclude PE in patients with suspected PE and DVT. The cutoff of 0.5 mg/L in D-dimer and four points in the Wells score are ade-quate and can be recommended. A follow up measurement of D-dimer after 3-7 days adds no extra information. The vWF level is higher in patients with PE and proximal DVT than in those with distal DVT, indicating the similarity of PE and proximal DVT. The vWF levels can be used to indicate the extension of the thrombus. The levels of vWF decreases after three months in patients with PE and DVT.

The diagnostic work-up of patients with suspected PE should comprise an evaluation of symptoms, signs, risk factors, and alternative diagnoses. When the suspicion of PE is high, an evaluation of Wells score is recommended, followed by a D-dimer if the score is low. A low Wells score and a low D-dimer can safely exclude PE. If the Wells score or D-dimer is elevated, a diagnostic test is mandatory. CTPA and PA produce equivalent results, although CTPA is recommended because of its availability.