see also:

• aortic dissection
• aortic dissection features and diagnosis
• D-Dimer and aortic dissection
• the adult patient with chest pain in the ED

• the diagnosis of non-traumatic aortic dissection is often difficult, particularly given:
  • its relatively low prevalence (2-3 cases per 100,000 person years but 2-3x prevalance of ruptured AAA)
  • only 60% make it to an ED alive
  • of those who do make it to an ED:
    • clinical presentation is often variable
    • pain may ease or abate, leading to a false reassurance on the part of the patients and physicians, despite the 1% per hour mortality after onset if they survived initially
• regarding time from onset of initial symptoms to time of presentation:
  • acute dissection is defined as occurring within 2 weeks of onset of pain;
  • subacute, between 2 and 6 weeks from onset of pain;
  • chronic, more than 6 weeks from onset of pain
• for perspective, an average ED seeing 40,000 adults per year, may see well over 10,000 patients with acute chest pain, abdo pain, back pain or syncope per year, yet only 2-3 patients will have aortic dissection and currently perhaps 1 or 2 of these will not have the diagnosis made and will die after discharge home.
• in higher risk patients, the absence of widened mediastinum on CXR, or a BP differential between arms is not sufficient to exclude a dissection, nor is a negative D-Dimer sufficient.
• many clinicians then request a CT aortogram if the clinical picture is suggestive (eg. sudden onset thoracic pain), with a risk factor (eg. PH hypertension, Marfan's or recent coronary angiography) and no obvious cause or if widened mediastinum or BP differential is found, however this will still not be pick up all cases as many present with chest pain which resolves, lower back pain or abdominal pain, pain with stroke or neuropathy, or just syncope or hypotension.
• unfortunately there is no good screening tool with high sensitivity and specificity without resorting to aortic imaging such as CT aortogram
• the ADD tool unfortunately may have very limited utility due to a likely very low specificity, and worse, with very few ED patients being able to be ruled out by a score of 0, and even then they may still need imaging.
• the ADD tool does not reference PH hypertension as a risk factor even though 75% of cases have this risk factor - this could mislead novice clinicians
• anecdotally, very few patients in the ED population have acute aortic dissection if they do not present with sudden onset severe chest or thoracic back pain and a risk factor such as hypertension, Marfan's, family history, recent aortic intervention or known aortic valve disease
• ED doctors will unfortunately continue to miss this diagnosis but hopefully they will miss fewer by having the ADD tool raise awareness
• the ADD score tool has been adapted from the 2010 American College of Cardiology/American Heart Association thoracic aortic disease guidelines
  • http://circ.ahajournals.org/content/121/13/e266.full.pdf+html
• the ADD score tool has been shown to be sensitive on retrospective data, however, it has yet to be validated in a clinical setting, and may lead to substantially higher rates of aortic imaging which may introduce its own risks of radiation and contrast.

• score 1 for each of the 3 categories that are positive:
  • category 1: a high risk factor:
    • Marfan syndrome
    • FH aortic disease
    • known aortic valve disease (especially bicuspid aortic valve)
    • recent aortic manipulation
    • known thoracic aortic aneurysm
    • NB. although not included in the ADD score, one should also consider other high risk factors such as:
      • Loeys-Dietz syndrome, vascular Ehlers-Danlos syndrome, Turner syndrome, or other connective tissue disease
      • patients with mutations in genes known to predispose to thoracic aortic aneurysms and dissection, such as FBN1, TGFBR1, TGFBR2, ACTA2, and MYH11
      • NB. other risk factor epidemiology to be aware of:
        • 60-75% of patients have hypertension, which is often uncontrolled
        • 7% of patients are aged < 40 yrs and of these, 50% had a history of Marfan syndrome
        • 13% to 19% of patients without an identified genetic syndrome with thoracic aortic aneurysms had first-degree relatives with thoracic aortic aneurysms or AoD
        • a history of extreme exertion or emotional distress may precede the onset of pain
        • use of ‘crack’ cocaine (abrupt catecholamine-induced hypertension)
  • category 2: a high risk pain feature:
    • chest, back or abdominal pain which is either:
      • sudden onset (it is classically maximal at onset)
      • severe
      • ripping or tearing
    • NB. the 2010 published version had this section as “abrupt in onset/severe intensity AND ripping/tearing/sharp or stabbing” and was “developed on the basis of an extensive review of the literature on acute AD combined with the collective experience of the writing committee” ²⁾, however, the sensitivity study used the above as ripping or tearing was only in 21% of patients with dissection ³⁾
    • NB. 6% do not have pain
  • category 3: a high risk exam feature:
    • evidence of perfusion deficit such as:
      • pulse deficit
      • systolic BP differential in each arm > 20mmHg (only in Type B dissections not Type A)
      • focal neurologic deficit with pain
    • new murmur of aortic insufficiency with pain
    • unexplained hypotension or shock state

• score 0 = “low risk” but may still need aortic imaging
  • 4% of dissections will be in this category, although almost half will have widened mediastinum of CXR ⁴⁾
  • if no other cause for chest pain found:
    • if widened mediastinum on CXR or unexplained hypotension then expedite aortic imaging
    • if no widened mediastinum in this group, a negative D-Dimer will accurately exclude dissection ⁵⁾
    • other patients may still need to be considered for aortic imaging
• score 1 = “intermediate risk” will probably need aortic imaging if no other cause found
  • ~37% of dissections will fall into this category
  • consider performing aortic imaging unless:
    • ECG demonstrates STEMI (but if coronary angio is nad one should search for dissection!), or,
    • CXR clearly demonstrates other cause, or,
    • clinical features strongly suggest an alternative diagnosis, or,
    • D-Dimer is negative ⁶⁾
• score 2 or 3 = “high risk”
  • ~60% of dissections fall into this category
  • contact surgeon and arrange aortic imaging ASAP

• it has not been validated prospectively for an ED population
• although it has high sensitivity of 96% for dissection it has very low specificity:
  • in an Italian “prospective” study on 1328 patients with “suspected dissection”, of whom 22% were diagnosed with dissection, and retrospectively calculating ADD scores ⁷⁾:
    • 2/3rds of the patients had an ADD score of 1 or more
    • 1/3rd had an ADD score of 0 and of these, 5.9% had dissection and widened mediastinum on CXR in this group only had a sensitivity of 17% but a specificity 86%
    • of the ~50% who had a score of 1, 27% had dissection
    • of the ~18% who had a score of 2 or 3, 39% had dissection
    • the dissection rate of 22% in this study seems far higher than one would expect if one was to apply the ADD rule to all ED patients with either syncope, chest, back or abdo pain, and if one did this, the specificity would be far worse still.
• patients with an ADD score of 0 or 1 are problematic but it appears a negative D-Dimer will exclude dissection in these groups ⁸⁾
• Some patients with aortic dissection present without any high-risk features, making early diagnosis difficult. If a clear alternative diagnosis is not established after the initial evaluation, and CXR is normal, then obtaining a diagnostic aortic imaging study, particularly in patients with advanced age (older than 70 years), syncope, focal neurologic deficit, or recent aortic manipulation (surgery or catheter based), should be considered ⁹⁾

• the ADD score is problematic in the ED as almost every ED patient with chest, abdo or back pain or syncope would get a score of at least 1 which is like trying to rule out a PE on a post-op patient by doing a D-Dimer.
• we thus need a much more specific tool
• maybe something like this for patients with syncope, chest, abdo or back pain:
• an AYTON score (score 1 for each category) of 3 or more indicating need for aortic imaging:
  • Aortic disease (eg. bicuspid aortic valve, thoracic aneurysm) or recent intervention such as coronary angiography
  • Young patient under 40 years with either Marfan's, Ehler's Danlos, Turners, FH dissection, cocaine or stimulant use, PH hypertension, arteritis, weight lifting, or other risk
  • Type of pain is either sudden onset severe pain or ripping or tearing pain
  • Old patient over 40 years with risk factors as for young patient, or older than 70 years
  • No clear explanation for the pain or syncope
• OR a WHIP score of 1 or more in the following would also indicate need for aortic imaging:
  • Wide mediastinum on CXR (CXR only needed if AYTON score > 0 or patient has chest or thoracic back pain)
  • Hypotension that is unexplained and AYTON score is 2 or more
  • Ischaemia of end organs such as bowel ischaemia without cause (ie. no AF), pancreatitis without cause (ie. no biliary, alcoholic cause), neuropathy or stroke with pain, STEMI with normal coronary angiography
  • Pulse deficit or BP differential in each arm > 20mmHg
• an AYTON score of 2 even with a WHIP score of 0 should also be considered for aortic imaging if the clinical gestalt suggests dissection is reasonably likely to explain the pain or if the clinician detects a new murmur of aortic insufficiency
• an AYTON score of 0 or 1 and a WHIP score of 0 would give a very low probability of dissection and thus aortic imaging should not be needed, particularly if the patient also has a negative D-Dimer.

• although this score makes logical sense it has NOT been validated
• scoring tools for such uncommon conditions with variable presentations are very difficult to validate prospectively
• we need a relatively specific tool to avoid unnecessary CT aortograms with contrast and their inherent patient risks and costs, but at the same time sensitive enough to detect most cases of dissection
• perhaps more importantly, the tool may provide better clinician awareness of this difficult to diagnose condition which is often like a needle in a haystack of the plethora of ED presentations
• features of the WHIP score either suggest relatively specific phenomena for dissection (such as BP differential or wide mediastinum) or serious illness which warrants invasive investigation even though probability may be low that dissection is a cause.
  • a negative WHIP score alone is insufficient to exclude dissection as sensitivity would be < 50%, hence the need to have an additional score such as the AYTON score
  • unexplained hypotension alone would be a poor discriminator leading to excessive CT scans, hence the need to qualify it with some other risk factors.
• at least 80% of patients with dissection will have either sudden onset severe pain or ripping or tearing pain, hence:
  • even without any risk factors or WHIP features, as long as there is no other clear cause they would score AYTON 2 and still be considered for aortic imaging
  • as 75% of patients with dissection have a PH hypertension, the additional presence of this or any other risk factor in such a patient would give an AYTON score of 3 or more and indicate aortic imaging
• dissection is quite rare in those under 40yrs unless they have a risk factor such as Marfan's which is present in ~50% of these cases, so young adults with a risk factor and no clear cause for the pain also get at least AYTON 2
• a person with presumed ischaemic chest pain but normal findings and normal troponin would get an AYTON score of at least 2 if they are over 70 or have a risk factor such as hypertension and thus should still be considered for aortic imaging
• patients with renal colic, biliary colic, ruptured ovarian cyst, etc with sudden onset severe pain are unlikely to have a positive WHIP score or an AYTON score > 2 and thus can quite rightly avoid aortic imaging
• given the above it would be reasonable to expect sensitivity of at least 80% which is not bad for a very uncommon condition if the specificity is also not too low.
• whilst ~24% of patients with dissections have a new murmur of aortic insufficiency, this has intentionally not been included in the above score as it is unlikely to significantly increase sensitivity and likely to have considerable inter-observer variability in its detection, and debate as to whether or not it is there may delay imaging - and every hour of delay increases mortality risk by 1%, nevertheless, if found it would suggest aortic imaging is indicated.
• given the ED population this potentially applies to, it would be absurd to mandate a CXR in EVERY patient including young patients with no risk factors for dissection who present with vasovagal syncope or abdo pain or back pain, hence the advice that a CXR only need to be done for the score if AYTON score is > 0 or there is chest pain as the prime presentation.

• in order to increase ante-mortem diagnosis rates for dissection, clinician awareness needs to be raised, and with or without the assistance of tools such as these, inevitably, rates of aortic imaging will need to increase
• the aim is to find an acceptable balance between over-imaging, under-diagnosing and complexity
• the ADD score would appear to diagnose 96% but at risk of very high numbers needing to be imaged
• the AYTON-WHIP score aims to improve upon this by accepting a slightly lower sensitivity while still being relatively simple but hopefully gaining a much lower imaging rate, nevertheless, certain subgroups may still end up in the excessive imaging category:
  • patients with sudden onset severe low back pain or abdominal pain who have no clear cause but who are over 70yrs or have hypertension, would have an AYTON score of 3 which is potentially problematic as this group is probably fairly low risk for dissection given the multitude of other causes which could explain this picture - and thus the score may not be specific enough for this subgroup to warrant CT aortograms from chest to pelvis - although a contrast abdo CT may be indicated so perhaps extending this to an aortogram is not as problematic as it may seem.