Friday, November 20, 2015

Validation of Smith Modified Sgarbossa Criteria Published in American Heart Journal

We have completed and published the external validation of the Modified Sgarbossa Criteria for Diagnosis of Acute Coronary Occlusion in the Presence of Left Bundle Branch Block.

H. Pendell Meyers had just graduated from college when he took this project on.  Now he is a 4th year student at Duke, interviewing for Emergency Medicine Residency Positions.

He did amazing work on this project.

Here is a link to the abstract:
http://www.sciencedirect.com/science/article/pii/S0002870315005918

A quick summary:

Sgarbossa criteria:

1. at least 1 mm of concordant ST elevation in at least one lead (5 points)
2. at least 1 mm of concordant ST depression in at least one of leads V1-V3 (3 points)
3. at least 5 mm of discordant ST elevation in at least one lead (2 points)

Weighted criteria (the actual Sgarboss criteria): At least 3 points required to make the diagnosis of acute MI.  Thus, criterion 3 is not sufficient.

Unweighted criteria: any one of the above

Derived Smith-Modified Criteria, published in 2012 in Annals of EM (amazing work on this was done by Dr. Ken Dodd, who was a medical student at the time and is one of our emergency medicine/internal medicine residents now):

1. at least 1 mm of concordant ST elevation in at least one lead (5 points)
2. at least 1 mm of concordant ST depression in at least one of leads V1-V3 (3 points)
3. at least 1 mm of discordant ST elevation AND an ST elevation to S-wave ratio of at least 25% in at least one lead.

Validation Results:

There were 45 patients with LBBB and acute coronary occlusion and 249 controls:
The Modified Criteria (ST/S ratio of greater than or equal to 25%) was far more sensitive than either the weighted or unweighted Sgarbossa criteria 

15 comments:

  1. Excellent work! Congrats! You seldom see those odds ratios in publication these days!
    Jacob

    ReplyDelete
  2. Congrats Master for your success!
    I can say that so far i've learned more things from you than many other books and even cardiologist.
    Your labor and work are priceless.
    Cheers :)

    ReplyDelete
  3. Dr smith, Congrats on the results. This shows your passion for the art of electrocardiography. I have learned so much from you over the past few years. I have been very fortunate to apply what you have taught me in the field as a Paramedic and have impacted many lives. Take care

    ReplyDelete
  4. congrats. love this ECG blog and great to see a well deserved publication.

    ReplyDelete
  5. Dr. Smith,
    Great work as always. I've run across other references to the Smith Modification where the J point discordance is measured against the S wave OR the R wave- whichever is more prominent. In this validation it looks like only ST/S wave was measured, is the ST/R wave ratio not valid?
    Thanks!

    ReplyDelete
    Replies
    1. Rory,
      In our original derivation, if there was one lead with proportionally excessive ST elevation or ST depression, that was 100% sensitivie and 88% specific.
      However, when we tested this rule in the validation group, the sensitivity was much lower.
      So we have abandoned it.
      Thanks!
      Steve

      Delete
  6. Dr. Smith --

    An independent validation study not including the original authors (Electrocardiol 2013 Nov-Dec;46(6):528) did show sensitivity improvement of the Smith-Modified criteria over the original Sgarbossa rule, but far less than reported in this study. It also reported the superior sensitivity of the Selvester 10% RS rule as compared to other modifications.

    How do you see this reference fitting in with the literature?

    Thank you.

    ReplyDelete
    Replies
    1. Yes, I have read this.
      1. I think the idea of area under the LBBB curve is a good one. I did not have digital data to work with. It is necessary for such calculations.
      2. they did not use angiographic endpoints. Only acute MI. That includes NonSTEMI. Most studies have 2-3 x as many NonSTEMIs as STEMIs. I'm trying to diagnose acute coronary occlusion (needs emergent PCI), not NonSTEMI (which is easily diagnosed with troponin and can wait until tomorrow for angiogram, if remains pain free).
      make sense?
      Steve Smith

      Delete
  7. If I'm reading it right, your original paper didn't include sensitivity/specificity for a ST/S ratio of 0.2. Based on your data and the data from Meyers et al, do you think it's a safer bet to use 0.2 as opposed to 0.25? (Based on the bias to over-triage as opposed to miss an ACO).

    Congrats on having your data validated. Strong work as always.

    ReplyDelete
    Replies
    1. Bryan,
      It would be more sensitive, but slightly less specific.
      Thanks!
      Steve

      Delete

DEAR READER: I have loved receiving your comments, but I am no longer able to moderate them. Since the vast majority are SPAM, I need to moderate them all. Therefore, comments will rarely be published any more. So Sorry.