Aortic Dissection and the IRAD study

Confession time.

I have never seen a patient with an aortic dissection.

Hoepfully you would have all corrected me and said that I have never diagnosed a patient with aortic dissection.

You don’t do this type of job for 7 years without seeing one. I have no doubt some of the rule-out ACS patients I sent upstairs (to the ward not a higher metaphysical plane…) turned out to have a dissection.

I know cases where people have gone home after an ED visit and died from dissections.

This is a properly scary disease. While I can’t say I’m hoping to see one soon I am keeping my eyes open.

One of the key studies about presentations of aortic dissection is the IRAD study.

The International Registry of Acute Aortic Dissection (IRAD): new insights into an old disease. JAMA. 2000 Feb 16;283(7):897-903. PMID 10685714

This is a big registry (running from ’96) enrolling patients with a diagnosis of dissection.

So here’s a quick run through

Patients

• patients with a diagnosis of aortic dissection (A or B) referred to one of a few major centres

this is probably the most important point about this study. You only got included if you had a diagnosis. Someone makes the diagnosis and then they asked questions about the presentation, examination and work-up. So if you came in with chest pain from a dissection but no-one made the diagnosis (and we know this happens) then you’re not in the study.

Let me put it another way. You could say that this study only informs us about patients in whom we know how to make the diagnosis.

Methods

• once the diagnosis was made someone asked them lots of questions. there were 290 potential variables that they looked at
• data were collected either at presentation or by chart review

this is is the second most important part of this study. Prospective collection of data with a patient in front of you is very different from chart review. For some things this doesn’t matter. Like mortality or BP at presentation. But when it comes to things that we’re interested – like history and clinical findings there is no way to be clear about what was actually found.

Chart review can work like this:

[bemused researcher looks at chart]

“so was the pain burning or ripping or tearing?…  hmm they just wrote chest pain‚… i’ll stick that down”

This is why Jerry Hoffman goes on about methods in chart reviews so much in the abstracts. If you don’t tell us how you abstract the data and what you do with missing varables then it’s very hard to know what it means. The classic paper for chart review methods is this one:

Chart Reviews In Emergency Medicine Research: Where Are The Methods Ann Emerg Med. 1996 Mar;27(3):305-8. PMID 8599488

But even having said that, let’s say you’re asking prospective questions of the patient:

[bemused researcher to critically ill patient]

” so is your pain ripping or tearing?”

[critically ill patient to researcher]

“what does ripping feel like, like a ‘hangnail’ ripping or ‘hung drawn and quartered’ ripping?”

While it’s nice to know about character of pain, just be aware how subjective it can be and don’t hang your hat on it too much

In this study they actually don’t tell us the proportion of data which was chart review and which was prospectively collated. Just so you know.

 Results

let me just quote their discussion:

The typical patient in the IRAD registry is a male in his seventh decade with a history of hypertension who presents with abrupt onset of chest pain.

Wow thanks for that…

I’m being unfair. That’s just the nature of the disease it’s a bugger to diagnose

Some things that might be helpful:

• most of the time it’s sudden and severe
• x-rays will not let you rule it out
• they say that 12% of CXRs were normal but remember that means no abnormalities were noted. We don’t know who reported these or what they were looking for
• most had abnormalities on the CXR
• mortality overall (A and B together) was 25% and most died from aortic rupture or tamponade

My thoughts 

No single clinical feature will let you rule it out. And doesn’t that sound like any other disease you’ve heard of.

Depressing as it is and inadequate as you may feel, clinical judgement is what you have to raise your suspicion and go hunting.

And yes you will miss some, and yes you will overcall some but you’re not really that terrible at this clinical suspicion business.

Some things that would make me think of this disease (but will not let you rule it out)

• severity of pain
• a big inferior MI (when it does dissect back it will often go to the origin of the RCA)
• a pericardial effusion (this is easy to look for)
• chest pain and stroke or neuro findings
• a funny looking CXR
• palpable pulse deficit (but not BP difference, that’s a bit useless)

The September 2008 EM Cast had a great discussion with Rob Rogers on this. Vascular emergencies is his thing I think.

As a parting shot, I note they talk about D-dimer for ruling out aortic dissection. One of the big name CT surgeons for this thinks it’s wonderful. I think it’s a bad idea. Any thoughts? Anyone using D-dimer to help rule out dissection?