Apr 9, 2021 This Week in Cardiology Podcast

John M. Mandrola, MD


April 09, 2021

Please note that the text below is not a full transcript and has not been copyedited. For more insight and commentary on these stories, subscribe to the This Week in Cardiology podcast.

In This Week’s Podcast

For the week ending April 9, 2021, John Mandrola, MD comments on the following news and features stories.

COVID and COVID Vaccines

Once again, there is good and bad news. Many areas of the United States are doing amazingly well. Why is Texas doing so well even though it has had some of the least strict measures? Why is Michigan doing so poorly? It’s hard to explain. I talked to a doc at a big hospital in Michigan and he says admissions are up, but not at the same pressure as during the winter surge.

Also, good news is the US vaccine rollouts. We had a rocky start, but now millions are being vaccinated each day and the data from Israel shows that getting the vulnerable vaccinated associates strongly with declining hospitalizations and deaths. Internationally, though, there are pockets of bad news. The high death toll in Brazil and rising cases in India are sad to see.

I talked about the AstraZeneca (AZ) vaccine clotting issues last week, but an update this week is that UK regulators have said they will offer an alternative vaccine to adults under age 30. They call the thrombotic events post-vaccine as “vanishingly rare but sadly very serious adverse event.”

Two points to make on this:

First the Winton Centre for Risk and Evidence Communication at Cambridge published some worthy decision support tools regarding the risks of vaccine vs risks of ICU admission from COVID-19. In short, for medium exposure-risk adults age 20-29 years, the risks of ICU admission from COVID-19 is 2.2 per 100,000 and the risk of serious clotting events from the AZ vaccine is 1.1 per 100,000. To me, if I were under-30, I can see no reason to take the vaccine. So, the UK regulators’ decision seems wise.

The second point to make about this is the asymmetry of risk. Those decision aids show that for high exposure risk adults or older adults, the risk of clotting after the vaccine is much lower than the risk of ICU admission. Now, if patients were robots who simply chose the most favorable gamble, they would always take the vaccine. But that’s not how people decide. People feel risk as much as they calculate risk. And for some, they would feel more risk from the chance of a vaccine adverse event.

It’s the same reason some people decline anticoagulation–they fear bleeding events more than stroke events, even though stroke is more common and more severe, perhaps because a relative had a severe bleed, or perhaps they are a neurosurgeon.

Hybrid Ablation for AF

I want to apologize to my listeners. A sort of big trial came and went in November of 2020 that got by me. Circulation Arrhythmia and Electrophysiology published the results of the CONVERGE trial –an RCT comparing combined epicardial surgical ablation and endocardial catheter ablation to standard endocardial ablation alone in patients with long-standing persistent atrial fibrillation (AF).

I mention this trial because patients have seen it advertised and have asked me about it, a search of YouTube returned oodles of videos from medical centers promoting the procedure as a major advance, and I’ve heard of some colleagues trying it out.

The combined surgical and catheter-based approach is called the Convergent procedure. First, a surgeon makes a sub-xiphoid incision and places ablation lesions with a device made by Atricure—the funders of the trial—onto the back wall of the left atrium for ablation and around the pulmonary veins. Complete pulmonary vein isolation via this approach is not possible due to the pericardial reflection, so an EP doc goes into the left atrium and finishes the pulmonary vein isolation with the standard catheter.

In the rationale paper, published in June, after the trial results had already been presented at the Heart Rhythm Society meeting, the authors wrote that restoration of sinus rhythm in patients with longer-standing forms of AF is an unmet need. A side bar: To my younger listeners, whenever you hear proponents of new and aggressive technology say unmet need, be especially careful about critical appraisal.

The idea behind the Convergent is twofold. In patients with longer-standing forms of persistent AF, standard pulmonary vein isolation (PVI) may not be enough. There is literally no RCT data to back that contention. In fact, the seminal STAR AF 2 trial showed that more aggressive left atrium ablation strategies were no better than standard PVI in patients with persistent AF.

And, posterior wall isolation via surgical approach directs energy away from the esophagus and might help reduce esophageal injury. There is no convincing RCT-level evidence that adding posterior wall isolation improves outcomes over standard PVI.

In CONVERGE, 100 patients were randomly assigned to the combined approach and 50 to the standard arm. The primary efficacy endpoint was a composite endpoint of freedom from AF/atrial flutter/atrial tachycardia absent new/increased dosage of previously failed/intolerant class I/III antiarrhythmic drugs through 12 months. The primary safety endpoint was major adverse events through 30 days.

The primary efficacy end point was achieved in 68% of patients in the combined hybrid procedure vs 50% in the catheter ablation group, without imputation for missing data as failures. The P value was 0.04.

A key secondary endpoint was 90% reduction in AF burden. At 18 months using 7-day Holter, 74% subjects in the hybrid Convergent group achieved this end point compared with 55% of subjects in the catheter ablation group; P = 0.04. Major adverse events occurred in 8 of 102 patients in the combined arm (one stroke, one TIA, one excessive bleeding, 3 late pericardial effusion, one phrenic nerve injury) vs none in the standard arm.

The authors concluded: “The Hybrid Convergent procedure has superior effectiveness compared to the CA for the treatment of persistent and long-standing persistent atrial fibrillation (LSPF).”

A few comments on internal validity – or trial conduct.

  • First is baseline differences. Roughly 9% of patients in the combined group were on dofetilide vs 2% in the standard arm. That’s important because dofetilide may be the most effective drug in obese patients with LSPF.

  • The combined procedure included 22% females vs 47% in the standard arm. More females in the standard arm may bias the results against the standard arm because the preponderance of the evidence suggest females with AF may fare worse after ablation. The authors say this did not impact the primary endpoint because the subgroup analysis by gender showed a similar treatment effect. But they don’t report the subgroups and there is no supplement attached. This imbalance may be chance, but it also raises the possibility of selection bias.

  • Another concern was the line in the reporting of the primary endpoint that read “without imputation for missing data.” I can’t tell from the manuscript how much data was missing. There is no CONSORT flow diagram and as I said, there is no supplement. How much data was missing?

  • A meta – point: The CONVERGE trial was sponsored and funded by AtriCure—numerous authors report compensation from the company. This is not nefarious; it is, however, notable.

The major issues mostly involve external validity or generalizability issues. The primary endpoint is a surrogate endpoint of freedom from AF. CONVERGE did not have the power to test clinical endpoints. And even looking at a surrogate endpoint of freedom from AF, the P-value barely met significance. When I look at a P-value it makes me think: how strong was the evidence? One way to look at the strength of evidence is to calculate a Bayes Factor (BF), which is akin to a likelihood ratio of a medical test. BF is the probability of the data given no effect (the null hypothesis) divided by the probability of the data given the alternate hypothesis of a true effect. A super low BF is stronger evidence because a small number implies the evidence is not consistent with the null hypothesis of no effect.

The BF for the P-value of CONVERGE is 0.19 which is close to one-fifth. Thus, there is nearly 20% chance this data is consistent with the null hypothesis being true.

Now let’s say you were a bit pessimistic. Let’s say you look at all the prior data, like STAR AF 2, and the fact that a skilled operator could accomplish a similar lesion set with a standard ablation catheter alone, and you think the chance of no effect from this procedure would be 70/30, or 2.33.

Now use the simple Bayes Theorem you learned in medical school—prior belief x likelihood ratio is equal to the posterior belief. Recall that trials like medical tests simply update one’s prior beliefs. Here you would multiply your prior belief of 2.33 x BF or LR of 0.19 and you get a 44% chance of this data being consistent with the null. In other words, the data in CONVERGE may be statistically significant but are surely not strongly against the null. Now consider that patients in the Convergent arm had surgery requiring chest tubes and a total procedure time of 293 minutes vs 171 with the endocardial procedure.

On one of the promotional YouTube videos, I heard an author say the average length of stay was 3 days in the surgery arm. That vs 1 day for standard ablation. For the longer procedure time, open chest surgery, we get added expense for a barely significant difference in a surrogate endpoint. The worst part was 8% vs 0% major adverse events in the convergent arm vs standard arm. This trial has statistically significant results but for all the reasons above I would not recommend this procedure to patients. If ever there was a need to do a confirmatory trial, this would be case in point.

Transcatheter Aortic Valve Replacement (TAVR)

Two facts about the recent TAVR vs SAVR trials: they enrolled patients with calcific aortic stenosis (AS) and compared TAVR to traditional SAVR. JACC has published an observational study looking at TAVR in patients with Rheumatic AS. TAVR in rheumatic heart disease (RHD) is more challenging for three reasons: less annular calcification means less to anchor the deployed valve; RHD tends to have more aortic insufficiency (AI); and RHD tends to involve more valves—not just the aortic. Nonetheless, if TAVR could be shown to have similar outcomes, this may have huge impact in geographies where RHD is prevalent.

First author Amgad Mentias from the Cleveland Clinic and colleagues used the Medicare database to find RHD patients who had either procedure, TAVR or SAVR. As these were nonrandomized groups, they did standard propensity matching. As you would expect, surgical patients were younger and had fewer co-morbid conditions.

The authors made two comparisons: first was TAVR vs SAVR in patients with RHD. Here they found that at 19 months (not very long), mortality was 11.2 per 100 patient years in the TAVR arm vs 7 per 100 patient years in the SAVR arm. The higher rate of death with TAVR did not reach significance.

The second question was TAVR in RHD vs TAVR in non-RHD. Here they had 89,000 patients without RHD and about 600 with RHD. After adjustment, they found 15.2 vs 17.7 deaths per 100 patient years in RHD vs non-RHD. Again, this was a nonsignificant difference.

The authors concluded that, “compared with SAVR, TAVR could represent a viable and possibly durable option for patients with rheumatic AS.” I love the modifiers ‘could’ and ‘possibly.’ Thank you Dr. Mentias.

The BMJ Open Heart journal published the second TAVR study – a meta-analysis of seven observational studies comparing TAVR to minimally invasive SAVR (miAVR). In their introduction, first author Ahmed Sayed from Ain Shams University in Cairo along with colleagues in Egypt and Boston explained that a few years ago, miAVR was increasingly being done. I am not sure what the percentage of the miAVR is in more recent years. Potential advantages of miAVR over TAVR include the ability to directly observe the valve during the operation, removal of annular calcifications, and lower incidence of paravalvular leakage.

Sadly, none of the seven studies that compared the two techniques were trials; all were observational. But, as a colleague, once said, you do your best with the data you have, not the data you wish for. Nonetheless, the main findings are provocative.

  • At one year, TAVR was associated with a higher mortality that ranged from 16% higher to more than 3 times higher.

  • A subgroup analysis of the two studies that used only transfemoral TAVR vs miAVR revealed a non-significant difference of 1.71, with a confidence interval from 0.86-3.43. This could either be due to a real difference or a reduced power to detect a difference.

  • Stroke and bleeding were not significantly different.

  • As for secondary outcomes, patients undergoing TAVR were more likely to experience paravalvular leakage, but less likely to suffer acute kidney injury compared with miAVR.

We must not make causal conclusions from these observations. And all the studies were likely biased from unmeasured confounding factors; eg, healthier patients get miAVR. Nonetheless, the observation of higher mortality of TAVR vs miAVR combined with the trends for better longterm performance of SAVR in the PARTNER series, only adds to the caution I have regarding using TAVR in lower risk groups. I highlight two observational studies because both point to important areas for proper randomized clinical trials, which is really the only way to settle the question.

Doctor” and Medical Degrees

Medscape ran a poll on who should get to be called a doctor. As many of you know, the Wall Street Journal ran an Op-Ed 4 months ago saying Jill Biden shouldn’t be called doctor because she is not a medical doctor. I discuss this topic for a few reasons.

One is that the first thing you have to realize is the business model of media. The currency of media now is attention. Making you mad is good for business. Nuance and balance is bad for business. So, the first thing you should think about when an Op-Ed, or Tweet, or a biased news story makes you mad, is that this is by design. In the same way you consider conflicts of interest in evaluating a paper, I also urge you to be mindful of the business plan of media and platforms like Twitter. Resist the urge to be enraged.

Now to the notion of who gets to be called a doctor. I strongly feel that anyone who wants to be called doctor should be. Four points here:

In my mind, a self-directed PhD would be far harder to finish than complying with the concrete set-out tasks in medical school. Yes, it was hard to memorize all that stuff, take all those tests, and work long hours, but it was set out for you. Do this... and you are a doctor. And by the way, it doesn’t mean you are a good doctor. In fact, superb test takers don’t always make good clinicians. Contrast the well laid-out plans of medical school to the open-ended tasks of a PhD. You get your doctorate when you contribute enough to science. I’ve done enough research to that that is much harder than taking tests and memorizing differential diagnoses.

Second, while I love being a doctor, I’ve long gotten over the romance of it. It’s a great job, but, in general, we’d do well to get over ourselves. Sanctifying our MD is unbecoming. We don’t control outcomes as much as we think. You come to realize from practicing over decades that people often heal and live well despite us, not because of us. Understanding how little control we have is very important for avoiding iatrogenesis.

Don’t mistake this for nihilism. I of all people believe we should dedicate ourselves to our craft, but we should also recognize that an advanced practice nurse with a decade of experience in one field is surely as good or better than any newly minted MD. We should also realize that many therapies we consider class I exert quite small effect sizes on individuals. Another reason to get over the ‘holiness’ of the MD is that when you get employed, you learn you can be off-boarded in two nanoseconds. You can have decades of service, and boom, you could be gone.

Third point: getting hacked off about titles, like ‘provider,’ is not worth the trouble. I used to care, but now, I try to ignore the BS. Just revel in the notion you are paid to help people. People who you help don’t care what your title.

Fourth. As a kid who loved basketball, one of my greatest childhood heroes was Doctor J. A legend who surely earned his doctorial status.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.