CRT Super-Response and the Power of Twitter

CRT Super-Response and the Power of Twitter

An unusual example of super-response to biventricular pacing and a commentary about real time academic collaboration via social media.
August 30, 2013

  1. What follows is a two day exchange on twitter earlier this week.  What was intended as a quick unknown case for the heart failure device community turned into something a lot more interesting.  Below I’ve organized the twitter posts and responses in the order in which they occurred and then follow with a discussion.  The topic is very specific to the pacemaker/ICD community, but the implications are much more broad.  Stay with me.

    First up is a ECG and electrogram recording from a biventricular ICD case that demonstrated some unusual findings.  I posed the question “What is interesting about this tracing?”
  2. Does anybody with interest in biventricular pacing wanna guess why I think this tracing is unique?…
  3. Shortly after my post, a few folks took the bait and @ replied with their thoughts.
  4. @EJSMD RBBB with RV-EGM coming first… could explain why pt responded? but I don’t know why “super”
  5. @EJSMD despite underlying RBBB, LV activation is late in QRS duration.…
  6. Jeff is one of my industry support folks.  He and I have talked about this topic a lot.  He’s really smart and got it right away.
  7. @jvober Jeff, you’re like the smart kid in the front row going “Ooh Ooh, call on me!!” Hang back, let the other kids have a chance. 🙂
  8. @EJSMD why do u Measure the RV – LV interval from onset of RV to peak of LV? Interval seems pretty long
  9. @PAC_Aware The interval measured is called the qLV: Onset of surface QRS to LV lead activation. The LV mark may be a little late.
  10. I offered a study citation for those interested.
  11. Those interested in my CRT tracing may wish to review this study on qLV measurements @PAC_Aware:…
  12. @PAC_Aware So do you see what is interesting about this particular tracing?
  13. Below comments reply to those from @tobymarkowitz.  As his account is locked, I am not able to add these to the Storify narrative.
  14. @tobymarkowitz Interesting. We have not systematically evaluated CRT withdrawal in super-responders, but our anecdotal exp. has been poor
  15. The next morning I posted an annotated form of the tracing with an explanation of what was going on that I found interesting.
  16. F/U to CRT case: Here’s a RBBB CHF pt with normalized EF after CRT. Note evidence of LB delay by QLV & old ECG:…
  17. Answer to yesterday’s CRT case up on last tweet: @tobymarkowitz @PAC_Aware @jvober @glibaudio @amcj1 Thanks for you contributions!
  18. Follower @amcj1 found this intriguing and we then passed around a few ideas about these findings and their broader application.  This exchange got me really excited about the power of twitter as a tool for real time academic collaboration.
  19. @EJSMD thank you, it’s very interesting. I wonder if it’s the case to measure Q-LV in every RBBB with low EF to decide BiV implantation
    • @amcj1 BINGO! Sounds like a hypothesis we could test.
  20. @EJSMD I was thinking about that… which catheter though? EP-CS and you measure Q-LV on the lateral/posteroL dipoles? or use a LV catheter?
  21. @amcj1 Currently we use the LV lead. Years ago there was .014 guide wire with bare end that could record EGMs. Very thin EP cath maybe.
  22. @EJSMD using the LV lead is expensive (in Italy, at least) and not so easy to do (long & short sheaths, guidewire, possibly contrast…)
  23. @amcj1 Agree, LV lead should only be used if you are committed to using it anyway. Not good tool for deciding CRT vs. no CRT.
  24. @EJSMD do you think/know if LV on the CS catheter (body) is approximate/correlates with the one on the LV lead (branch)
  25. @amcj1 Not sure, but suspect you’d have to get into lateral branch to know if LV is late. I’ll take a look in future cases.
  26. @EJSMD I’ll check as well we use an EP catheter to engage CS, I’ll record it and compare to the final LV; theoretically, it’s base vs. mid..
  27. @EJSMD .. if you stay at the same level (I mean, posterior vs PL vs lateral).
  28. @EJSMD ok, I better stop to bother you – you intrigued me, though! 🙂
  29. @EJSMD I’ll keep you updated, if you like!
  30. @amcj1 Definitely. Look at QLV times in the main body of CS and validate w lat veins. May be easy way to predict CRT resp in RBBB/IVCD.
  31. Real time twitter collaboration w @amcj1 could lead to a new tool to predict CRT response in non-LBBB. How cool is that? I’ll post storify.
  32. Discussion:

    The case presented is an example of CRT super-response in a patient with right bundle branch block (RBBB).  Let’s talk about the specifics of this interesting case, and how Twitter might be harnessed as a tool for real-time academic collaboration.

    Much of what I present here is true EP weenie stuff and I doubt I’ll be able to make it interesting to all but hardcore device/CHF folks.  Hang with me, though, because there’s a important lesson in here.


    CRT super-response has been defined in the literature as a >20% improvement in EF ( or residence in the top quartile of EFs seen in a study population getting CRT (  For my purposes, I like to keep it simple and describe CRT super-response as normalization or near-normalization of LV function after placement of a biventricular pacemaker or ICD.
    Seeing a CRT super-responder is one of the great pleasures in my practice, and I suspect most EPs and CHF doctors who work in this arena would agree with me.  Having been involved in management of heart failure patients for years before CRT was developed, I still find it amazing that you can take someone so sick and make them so much better with comparatively little effort.
    Factors that favor CRT super-response in the previously cited studies include female sex, nonischemic cardiomyopathy, LBBB and wide (>150 msec) QRS interval.  In practice, that is what I’ve seen.  I bet I could host a big picnic with all the happy LBBB non-ischemics we’ve helped by getting their EF over 50%.  On the other hand, the RBBB population doesn’t do nearly as well.  This is reflected in the latest CRT guidelines ( html) which assign non-LBBB with QRS <150 to a class IIB indication.
    Intuitively it makes sense that RBBB patients would not do so well.  We believe the main mechanism of CRT to be relief of intraventricular dyssynchrony by pre-exciting a late activating LV wall.  RBBB patients need not have any left sided delay, so it makes sense that many won’t response to LV lateral wall pacing.  We know that some RBBB patients also have co-exiting left bundle delay, and perhaps these will be the patients that respond the best to CRT.
    In the lab at The Christ Hospital, we’ve been using QLV (onset of QRS on surface to LV activation) measurements to define late electrical activation during LV lead placement for years.  We hook up our leads to the EP recording system and display them under the ECG recording.  This allows to easily see what is late when we are picking pacing sites in the coronary veins.  We try to favor long QLV measurements when settling on the pacing site.  It turns out the maximizing QLV with LV lead placement correlates well with favorable long term response to CRT.  I first saw an abstract  on this response from Jag Singh at Mass General.  In the tweets above I cited similar work from Michael Gold and others in the SMART-AV trial (including my old Cleveland Clinic mentor Pat Tchou).
    I’ve been paying close attention to QLV measurements at all of our cases, so when the patient presented above came though the lab for a pulse generator replacement, I was really surprised by what we saw.  Here was a male ischemic RBBB patient who super-responded (EF went from 20% pre-implant to 50% on his most recent echo).  That alone is pretty unusual.  Moreover, when we measure QLV in RBBB patients, we rarely see the LV lateral wall activate very late.  This patient had really late activation of the LV with a QLV, measuring 106 msec or 62% of the total QRS.  This really looked more like a true LBBB patient to me.  When I pulled up the old ECG (shown in the answer slide), it made sense.  This patient presented first as a LBBB patient, but over the years has developed additional RB delay.  In effect the LBBB is hidden by the RBBB.  Certainly the left axis is a clue, that this would be going on, but when I look at QLV in RBBB/LAFB patients, they are still typically very short.
    I put this up on twitter as a fun challenge and to get some input from some of the smart EP folks who follow me.  I’ve enjoyed participating in the unknown cases in the blogosphere (particularly those from Dr. Wes (  For such an obscure topic, I was thrilled that a few people took the time to respond.
    After the answer was posted @amcj1 and I traded a few tweets that I thought were really cool.  Bear in mind I have never met or spoken to @amcj1.  Frankly all I know is the he or she goes by CJ and works in Italy.  He or she thought it would be interesting to use this as a tool to predict CRT response in non-LBBB patients.  That was exactly my thought as I prepared the case presentation.  Would the presence of a long QLV be a good way to decide whether or not to place an LV lead in these borderline patients?  We both think that this would be worth exploring.  Over a few tweets CJ and I worked out what might be a practical way to explore this hypothesis.  Maybe getting a QLV from the main body of coronary sinus with an EP recording catheter would give us the information we need to decide whether to go on to commit to an LV lead.  We could get that data quickly and relatively cheaply.  Clearly much more work needs to be done, but I can see the germ of a valuable research study here.
    I’m certain social media has an untapped role in the advancement in science.   In the world of healthcare, we are really only scratching the surface of what is possible.  Imagine Twitter exchanges like we covered above on other difficult topics in medicine.  Through real time collaboration, we could harness the collective intelligence of the many to reach conclusions and create ideas that none of us could accomplish on our own.  No topic is too obscure if the reach of the media is broad.  Social media is democratic in the best sense of the word.  The best ideas will percolate to the top, independent of the usual barriers such as academic status or job title.
    Edward J. Schloss MD FACC FHRS
    Medical Director, Cardiac Electrophysiology
    The Christ Hospital
    Cincinnati, OH

Read next page

Did you find this story interesting? Be the first to
or comment.