“List top 5 worrisome diagnoses for chest pain” – we’ve all asked this question to junior learners and medical students, and one answer that should always come up is aortic dissection (AD).
As emergency physicians we are well versed in the entity, workup and management of AD, but as a fairly challenging diagnosis, I have often found it difficult to decide who needs to be worked up for dissection. During my training I have observed significant staff variation when considering this condition; with some rarely investigating for AD, while others regularly order CT scans.
As with any disease entity, when considering our workup we need to evaluate our pretest probability to drive investigations and management. The current standard of care for AD workup is a CT scan of the chest and abdomen with contrast. A dissection protocol scan has a significant contrast and radiation load, so one must employ sound clinical judgment with his/her diagnostic decisions.
In order to interpret how to utilize this information, it is important to have a good understanding of pretest probability and likelihood ratios (LHR’s). For a refresher, check out this previous BoringEM post on interpreting likelihood ratios.
The only decision rule to date that has been validated is the AHA Aortic Dissection decision rule, which has a reasonable sensitivity of 91%, but a low specificity of 40%1,2, so its utility as a bedside assessment tool is questionable. The decision rule is based on three main components3:
- High risk conditions
- Marfan Syndrome
- Connective tissue disease
- Family history aortic disease
- Known aortic valve disease (ie: bicuspid aortic valve)
- Recent aortic manipulation
- Known thoracic aortic aneurysm
- High Risk pain features
- Chest, back or abdominal pain described as the following:
- Abrupt onset/severe in intensity AND
- Ripping/Tearing/sharp or stabbing in quality
- Chest, back or abdominal pain described as the following:
- High risk exam features
- Evidence of perfusion deficit:
- Pulse deficit
- Systolic BP differential
- Focal neurologic deficit
- Aortic insufficiency murmur
- Hypotension or shock state
- Evidence of perfusion deficit:
Risk stratification based on the above criteria:
- Patient has no features from any category above.
- In this population, the AHA suggests if an alternative diagnosis is identified you can stop there (in regards to AD workup).
- Unexplained hypotension or widened mediastinum on CXR suggests you should pursue advanced imaging.
- AHA also suggests to pursue imaging in patients who you are suspicious for AD based on the clinical scenario, regardless of their risk stratification.
- Patient has any one feature from the categories above.
- AHA suggests to seek an alternative diagnosis on CXR and ECG, but if none is identified you should proceed with CT scan of the chest.
- Patient has more than one feature
- AHA suggests that all these patients need advanced imaging.
As previously mentioned, the rule has a rather low specificity, and as a result far too many patients are scanned. In this discussion, we aim to another approach to which low risk patients may need to be worked up for AD, and which patients we can avoid imaging.
There are a few features that will help you to determine your pretest probability, with particularly important LHR’s highlighted below4,5.
- Severe pain: sensitivity of 90% (This doesn’t help you differentiate any of the chest pain syndromes, it is often a given of the presentation).
- Sudden onset chest pain, sensitivity of 84% (immediate, rather than crescendo pain):
- Not necessarily helping in making the diagnosis, LHR of 1.6
- But the LACK of sudden onset chest pain has a negative LHR of 0.3
- Think of this as the ‘thunderclap headache’ of the chest.
- Tearing or ripping pain has a LHR of 1.2-10.8.
- Migratory pain carries a LHR of 1.1-7.6.
- Past medical history of hypertension has a LHR of 1.4
- Diastolic murmur has a LHR of 1.4
- Pulse Differential or deficits has significant diagnostic variability, LHR: 2.4-47.0
- Note this is significantly different from blood pressure differential in two arms, which has a poor sensitivity of 15%.
- Focal neurologic deficit has a high LHR of 6.6-33.0.
- Hypertensive in the ED: poorly studied, but considered a significant risk factor by expert opinion6.
- Hypertension is concerning when it is unexplained within the clinical context, or persists despite appropriate analgesia.
- As a refresher; recall that according to the Stanford Classification for aortic dissections, type A involve the ascending aorta, while type B dissections involve the descending aorta.
- Typically, type A dissections have a worse prognosis, and tend to be sicker patients, while Type B often have reasonable prognosis if detected. They are more likely to be hypertensive, and won’t have many typical findings, such as neurological deficits, pulse differentials and murmurs7.
|Feature||Positive LHR||Negative LHR|
|Immediate onset chest pain||1.6||0.3|
|History of Hypertension||1.4||0.5|
|Focal Neurological deficit||6.6-33.0||0.9|
|Abnormal CXR(+)/Normal CXR(-)||2.0||0.3|
- “Abnormal” CXR has a LHR of 2.0, while normal CXR has a negative LHR of 0.34.
- Utility of CXR is to help further lower your pretest probability in the low risk patient.
- An abnormal CXR is defined by having a widened mediastinum or abnormal aortic contour, but incorporates other findings that one classically considers in aortic dissection (pleural effusion, loss of AP window, etc.).
- There may be a role for point of care ultrasound in the bedside evaluation of a patient for aortic dissection; it is suggested that emergency physicians have a sensitivity of 67% and a specificity of 99-100%8 for detecting an intimal flap on POCUS. This is of course, very user dependent and requires some further investigation, but is a useful skill for the emergency physician to augment their clinical exam, if they have an appropriate pretest probability prior to doing ultrasound.
- D-Dimer has been extensively studied in the past few years with controversial results, and let’s delve into this further:
There have been a few primary studies, and multiple meta-analyses questioning if a negative D-Dimer can rule-out aortic dissection. The most robust study was a meta-analysis that suggested that patients who are low risk according to AHA aortic dissection decision rule, and have a negative D-Dimer can safely have aortic dissection ruled out. Within this meta-analysis, they demonstrated sensitivity of 98.0%, but a rather poor specificity of 41.9%, and a false positive rate of 40.0%9,10,11.
When applied to low risk patients, it appears that a D-Dimer does not significantly change one’s pretest probability for AD; therefore its utility in the workup for dissection is questionable. This sentiment was echoed in a policy statement released by ACEP in January 201512:
“D-Dimer cannot be recommended for use in the evaluation of aortic dissection.”
Based on the current data available, in my mind, utilizing D-Dimer in AD would be akin to sending off a D-Dimer in patients who are already PERC negative, when considering PE.
Take Home points
AD is a rare disease, and as emergency physicians we should be screening patients for the potential to have this disease entity. It is therefore more useful to identify features to reassure us that a patient is indeed low risk, and negate the need for further investigations, which is why negative LHR’s are an important aspect of this diagnostic consideration.
How should you apply this information?
- The AHA guidelines are too inclusive and result in the unnecessary scanning of patients.
- The decision to perform a CT scan to investigate for AD depends on your pretest probability – which can be modified using the LHR’s of various historical and physical exam features.
- Ultimately when your think that your patient is low risk, or you have a low pretest probability, you can probably stop there and consider an alternative diagnosis.
- In the patient presenting without high risk features, lacking immediate onset chest pain, vascular pain/migratory features, a normal clinical exam and CXR – aortic dissection is highly unlikely and doesn’t warrant significant consideration.
- With the exception of neurological findings, one individual historical or physical exam finding does not necessarily warrant investigation for AD, recognizing that this is a deviation from the proposed AHA guidelines, and therefore requires the clinician’s interpretation of the data available (i.e.: pretest probability).
- BP differential between arms is not useful, and instead we should be looking for pulse differences amongst extremities.
- D-Dimer adds very little to your workup of the patient with aortic dissection.
- D-Dimer is not validated in those that are intermediate to high risk.
- In low risk patients, it does not significantly change your pretest probability, and shouldn’t be utilized.
- Rogers, A. M., Hermann, L. K., Booher, A. M., Nienaber, C. a., Williams, D. M., Kazerooni, E. a., … Eagle, K. a. (2011). Sensitivity of the aortic dissection detection risk score, a novel guideline-based tool for identification of acute aortic dissection at initial presentation: Results from the international registry of acute aortic dissection. Circulation, 123(20), 2213–2218. http://doi.org/10.1161/CIRCULATIONAHA.110.988568
- Nazerian, P., Giachino, F., Vanni, S., Veglio, M. G., Castelli, M., Lison, D., … Morello, F. (2014). Diagnostic performance of the aortic dissection detection risk score in patients with suspected acute aortic dissection. European Heart Journal. Acute Cardiovascular Care. http://doi.org/10.1177/2048872614527010
- Hiratzka, L. F., Bakris, G. L., Beckman, J. a, Bersin, R. M., Carr, V. F., Casey, D. E., … Williams, D. M. (2010). 2010 ACCF/AHA/AATS/ACR/ASA/SCA/SCAI/SIR/STS/SVM Guidelines for the Diagnosis and Management of Patients With Thoracic Aortic Disease: Executive Summary. Circulation, 121(13), 1544–1579. http://doi.org/10.1161/CIR.0b013e3181d47d48
- Upadhye, S., & Schiff, K. (2012). Acute Aortic Dissection in the Emergency Department: Diagnostic Challenges and Evidence-Based Management. Emergency Medicine Clinics of North America, 30(2), 307–327. http://doi.org/10.1016/j.emc.2011.12.001
- Braverman, A. C. (2011). Aortic dissection: Prompt diagnosis and emergency treatment are critical. Cleveland Clinic Journal of Medicine, 78(10), 685–696. http://doi.org/10.3949/ccjm.78a.11053
- Sharpe, B. a. (2002). Clinical manifestations of acute aortic dissection. JAMA : The Journal of the American Medical Association, 288(7), 828; author reply 828.
- Klompas, M. (2014). Clinician’s corner: Does This Patient Have an Acute Thoracic Aortic Dissection? Clinical Scenarios, 287(17).
- Williams, J., Heiner, J. D., Perreault, M. D., & McArthur, T. J. (2010). Aortic dissection diagnosed by ultrasound. The Western Journal of Emergency Medicine, 11(1), 98–99.
- Asha, S. E., & Miers, J. W. (2015). A Systematic Review and Meta-analysis of D-dimer as a Rule-out Test for Suspected Acute Aortic Dissection. Annals of Emergency Medicine, 1–11. http://doi.org/10.1016/j.annemergmed.2015.02.013
- Nazerian, P., Morello, F., Vanni, S., Bono, A., Castelli, M., Forno, D., … Grifoni, S. (2014). Combined use of aortic dissection detection risk score and D-dimer in the diagnostic workup of suspected acute aortic dissection. International Journal of Cardiology, 175(1), 78–82. http://doi.org/10.1016/j.ijcard.2014.04.257
- Cui, J., Jing, Z., Zhuang, S., Qi, S., Li, L., Zhou, J., … Yin, Y. (2015). D-dimer as a Biomarker for Acute Aortic Dissection. Medicine, 94(4), e471. http://doi.org/10.1097/MD.0000000000000471
- Diercks, D. B., Promes, S. B., Schuur, J. D., Shah, K., Valente, J. H., & Cantrill, S. V. (2015). Clinical Policy: Critical Issues in the Evaluation and Management of Adult Patients With Suspected Acute Nontraumatic Thoracic Aortic Dissection. Annals of Emergency Medicine, 65(1), 32–42.e12. http://doi.org/10.1016/j.annemergmed.2014.11.002