Three Part Question
[In a previously fit and well patient presenting with clinical suspicion of pulmonary embolism] is [bedside transthoracic echocardiography compared to pulmonary angiography] specific and sensitive enough as a diagnostic test [to confirm or refute submassive/massive pulmonary embolism (PE)]?
A 33 year-old male is brought into the emergency department with an episode of acute breathlessness following a syncopal episode. His left leg is in a cast and he tells us that he returned from Australia recently where he broke his lower limb during a trek. He is tachycardic, tachypnoic, hypotensive, sweaty and clammy: he is in a peri-arrest situation. You wonder if bedside transthoracic echocardiography could rapidly confirm or exclude your presumed diagnosis and support your decision-making process with regard to treatment.
Medline using the OVID interface 1950 - August 2009
EMBASE 1980 - August 2009
Cochrane Library (all the database)
Google Scholar (no time limit applied)
Medline search : exp echocardiography/ AND exp pulmonary embolism/ OR pulmonary embol$.mp. LIMIT to english language and humans, diagnosis (sensitivity , specificity, optimized)
EMBASE : exp transthoracic echocardiograhy/ AND exp lung embolism/di. LIMIT both to human and english language, diagnosis subheadings for lung embolism
Cochrane Library : search terms used were ‘echocardiography’ and ‘pulmonary embolism’
Google Scholar : searched for first 100 results using the terms, ‘transthoracic echocardiography’, ‘pulmonary embolism’ and ‘diagnosis’
114 papers were found on the Medline search, of which 1 was of the highest level of evidence, that is a meta-analysis and was summarised in the table below. Two further papers of lower level of evidence were also included in our table.
258 were found on the EMBASE search, all of them of lower level of evidence than the previous paper.
We had no yield searching the Cochrane Library.
11,900 were found using Google Scholar, one was a meta-analysis which was already included in the Medline search.
|Author, date and country
||Study type (level of evidence)
|Roy PM et al,|
|Included : prospective studies that aimed at confirming or excluding pulmonary embolism (PE). Gold standard : pulmonary angiography.|
Excluded : retrospective studies, follow-up studies with more than 5% patients lost, studies that used additional imaging to pulmonary angiography, those where crude data could not be extracted or those with specific population. Also excluded : abstracts, editorials, reviews, case reports and case series.
48 papers were included in the final analysis of the 1012 potentially eligible articles. 2 papers looked at the use of echocardiography as exclusion/confirmation strategy.
|Meta-analysis|| Positive and negative likelihood ratios for confirmation and exclusion strategies respectively || Confirmation test : in high pre-test probability, a positive echocardiographic result is associated with a post-test probability overs 85% allowing an accurate diagnosis. Moderate/low pre-test probability patients required additional imaging to confirm diagnosis. Pooled positive likelihood ratio 5.0 (95% CI 2.3- 10.6)|
Exclusion test : a negative echocardiograhy required additional testing to rule out PE regardless of the initial clinical probabilty Pooled negative likelihood ratio 0.59 (95% CI 0.41- 0.86)
| Despite being of the highest level of evidence, this paper did not use the robust Cochrane/PRISMA method for composing a meta-analysis. The included papers were published between January 1990 and September 2003 only. It was also limited to English language therefore potentially leaving out relevant literature. |
|Hsiao S-H et al,|
|150 patients with pulmonary hypertension (50 of whom had PE on computer tomography) vs. 150 patients without cardiopulmonary distress or echocardiographic signs of pulmonary hypertension as control group||Comparative study||Right ventricular myocardial performance index (MPI) and M index || RV MPI>0.55 identified PE with a sensitivity of 85% and specifificty of 78%. A M index <112 had a sensitivity of 92% and a specificity of 92%||Patients with atrial fibrillation were excluded from the sudy. Unable to provide clear echocardiographic views in patients with chronic pulmonary hypertension|
|Lodata JA et al, |
|67 patients who underwent CT for exclusion of PE and had a transthoracic echocardiography within 48 hours of CT||Comparative study|| RV/LV area ratio, RV/LV end diastolic dimension ratio, ‘McConnell sign’, ‘D-sign’, pulmonary artery diameter, triscupid regurgitaion velocity, ‘60/60 sign’. Sensitivity, specificity, positive and negative predictive values (PPV and NPPV respectively) of the above were calculated|| RV/LV EDD ratio>0.7 most accurate predictor (sensitivity 66%, specificity 77%, PPV 82%). McConnell sign : high specificity 96%, low sensitiviy 16%, PPV 86%. D-sign sensivitiy 27%, specificity 81%, PPV 67%. 60/60 sign sensitivity 22%, specificity 69%, PPV 53%. Mean TR velocities did not differ between those with and without PE (270+/- 74 vs. 294 +/-83, p 0.25). No finding had an acceptable NPPV such that echocardiography could stand alone as a rule out test for PE ||Small sample size.
This was a referral population and could therefore represent a sicker population
Pulmonary embolism (PE) still represents a major public healthcare problem and it also imposes frequent diagnostic difficulties because of its heteregenous and non-specific clinical manifestation. It is therefore often misdiagnosed, underdiagnosed or even overdiagnosed. The use of transthoracic echocardiography has become now widespread amongst emergency physicians and ultrasound devices continue to become more portable making them more available for bedside application. Right-sided cardiac signs associated with acute thromboembolism are now thought to be amenable to accurate echocardiographic analysis.
Yet surprisingly, only a small number of papers were found addressing the diagnostic role of echocardiography in the diagnosis of acute PE. Several studies were yielded, with sometimes conflicting results, that probably reflected the differences in clinical setting, population selection, severity and extent of thromboembolism and the criteria used to diagnose the condition. Some of the data, however, would suggest a valid role for the use of echocardiography as rule in tool, due to the high specificity seen (SpIN). This is reinforced with the addition of clinical probability assessment.
Of note also is the paucity of studies assessing interobserver agreement between emergency physicians and trained echocardiographers in diagnostic features of PE. It is hard to directly extrapolate evidence from the studies above to tests performed at the bedside by emergency physicians, until this work has been carried out.
EDD, end diastolic dimension; LV, left ventricle; MPI, myocardial performance index; NPPV, negative predictive value; PE, pulmonary embolism; PPV, positive predictive value; RV, right ventricle; TR, tricuspid regurgitation.
Clinical Bottom Line
A high clinical probability combined with echocardiographic findings suggestive of PE make the diagnosis of PE very likely.
Echocardiographic findings paired with low or moderate clinical probability make the diagnosis or exclusion of PE uncertain and further tests will be needed.
- Roy PM, Colombet I, Durieux P et al. Systematic review and meta-analysis of strategies for the diagnosis of suspected pulmonary embolism. BMJ. 2005 Jul 30;331(7511):259.
- Hsiao SH, Chang SM, Lee CY et al. Usefulness of tissue Doppler parameters for identifying pulmonary embolism in patients with signs of pulmonary hypertension. Am J Cardiol. 2006 Sep 1;98(5):685-90.
- Lodato JA, Ward RP, Lang RM. Echocardiographic predictors of pulmonary embolism in patients referred for helical CT. Echocardiography 2008 Jul;25(6):584-90.