The trusted source for
healthcare information and
abstract & commentary
Synopsis: Cardiomegaly is the most common chest radiographic abnormality in acute pulmonary embolism.
Source: Elliott CG, et al. Chest 2000;118:33-38.
A chest radiograph is often the first imaging study done to assess a patient with suspected pulmonary embolism (PE) and it often influences the decision to perform additional diagnostic testing. The chest radiographic changes of PE have not been studied in a detailed fashion in previous studies due to small sample size or the lack of adequate descriptions of chest x-ray changes. Also, previous studies have not looked at the relationship between chest x-ray changes and right ventricular hypokinesis as seen on echocardiography. Right ventricular hypokinesis has previously been shown to be an important predictor of mortality associated with PE. The purpose of this study was to describe the chest x-ray findings in a large number of patients with acute PE. In addition, Elliott and associates sought to determine the sensitivity and specificity of cardiomegaly or pulmonary artery enlargement for right ventricular dysfunction as verified by echocardiography.
The International Cooperative Pulmonary Embolism Registry (ICOPER) prospectively enrolled 2454 patients who were diagnosed with an acute PE from January 1995 to November 1996. Elliott et al used this registry to prospectively identify patients diagnosed with acute PE (defined as within 31 days of symptom onset) at 52 hospitals in seven countries. They used interpretations of imaging studies provided by physicians at participating sites. The chest x-rays were first characterized as normal or abnormal; if they were abnormal, the participant physicians were asked to note the presence or absence of certain abnormalities.
The three most common modalities by which PE was diagnosed in this population were high probability V/Q scan (30%), high probability perfusion scan (24%), and pulmonary angiogram (19%). Most patients (89%) had symptoms and were hemodynamically stable, whereas only 4% exhibited hemodynamic instability. The most common symptoms included dyspnea (82%), chest pain (49%), and cough (20%).
Chest x-rays were available for 2322 patients (95%), and of those, 1759 (76%) were abnormal. The most common abnormalities were cardiac enlargement (27%), pleural effusion (23%), and elevated hemidiaphragm (20%). When looking only at patients whose PE was not diagnosed until autopsy, the results were similar except pulmonary congestion (44%) was more common than cardiac enlargement (41%), pleural effusion (36%), or elevated hemidiaphragm (26%). When looking at subsets of patients according to presenting symptoms, they found that cardiomegaly was still the most common abnormality for patients presenting with dyspnea alone (29%) and with syncope or hypotension (27%). They also evaluated patients by the different types of surgery they had undergone prior to PE. They found that those who had undergone thoracic or abdominal surgery were much less likely to have normal radiographs (4% and 16%, respectively) than those that had undergone genitourinary (37%), orthopedic (28%), or gynecologic (28%) procedures.
X-rays were available for 1084 out of 1135 patients (96%) who had an echocardiography. They demonstrated that cardiomegaly on chest x-ray had a low sensitivity (0.48) and specificity (0.63) for echocardiographic evidence of right ventricular hypokinesis. The findings were similar for enlargement of the pulmonary artery on chest x-rays (sens. 0.38, spec. 0.76).
Comment by David Ost, MD
In its clinical practice guideline regarding the diagnostic approach to acute venous thromboembolism, the American Thoracic Society concludes that the chest radiograph cannot be used to conclusively diagnose or exclude a PE. Instead, it should be used to diagnose other diseases that may mimic or coexist with PE. However, it is still of value to know which signs on a chest x-ray are more commonly seen with PE in the right clinical circumstance. Elliott et al’s study demonstrates that cardiomegaly is the most common chest x-ray abnormality associated with PE. Previous studies have come to different conclusions. Stein et al found atelectasis, pleural effusion, and pleural based opacity to be the most common abnormalities. The urokinase pulmonary embolism trial found an elevated diaphragm was the most common abnormality. However, Elliot et al’s study has the advantage of having a much larger study population than previous studies. Also, Elliott et al were not looking for abnormalities specific to PE since they did not exclude patients with known cardiopulmonary disease. This study also found that patients who are older than 70 are more likely to have abnormal radiographs than patients younger than 70, which is a different result than that obtained in the previous study. This paper also highlights how insensitive and nonspecific chest x-ray findings are for right ventricular dysfunction in acute PE. This is important because, as already mentioned, this is an important predictor of mortality. As Elliott et al point out, the sensitivity and specificity of chest x-ray findings suggestive of right ventricular hypokinesis are likely overestimations since not all patients received echocardiographs.
1. Goldhaber SZ, et al. Lancet 1999;353:386-389.
2. Tapson VF, et al. Am J Respir Crit Care Med 1999; 160:1043-1066.
3. Stein PD, et al. Chest 1991;100:598-603.
4. Sasahara AA, et al. Circulation 1973;47:1-108.
5. Stein PD, et al. J Am Coll Cardiol 1991;18:1452-1457.