Apr 30, 2021 This Week in Cardiology Podcast

John M. Mandrola, MD


April 30, 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 30, 2021, John Mandrola, MD comments on the following news and features stories.

COVID-19 Vaccination Gobally

The main thing to say about COVID is that the vaccine works. In the United States and in Israel, we see that as the numbers of vaccinated adults goes up, cases go down. Just as the trial results would predict. But that’s not the best part. The New England Journal of Medicine had a brief report on breakthrough infections after vaccination in two patients. That’s right, two of 417 employees of a university in New York City. Neither patient developed severe disease. And I think this is the prevailing message: vaccination massively, almost completely, reduces the chance of severe COVID illness.

Sadly, the global situation remains dire in South America and India. One thought I had, as a total nonexpert in public health, economics, or policy is this: What if we took the vaccines that we are giving to super-low risk young people and sent them overseas to countries in far more need? Isn’t there a debate about the benefit-harm calculus of rushing to vaccinate healthy young people against SARS-COV-2? A utilitarian or effective altruist case must be to surge vaccines to places like India and South America.

I will leave it at that – I don’t want to jump too far out of my lane.

AF Screening

At the European Heart Rhythm Association (EHRA) meeting this month, Professor Emma Svennberg presented clinical outcomes from the STROKESTOP study of screening for atrial fibrillation (AF) in two counties in Sweden. This is such a nifty and important trial. Kudos to Dr Sveenberg and colleagues for bringing outcome evidence to the issue of AF screening.

STROKESTOP is pragmatic trial of offering AF screening to 75 to 76-year-old residents of two regions in Sweden. Screening for AF was done with a handheld device twice daily for 2 weeks. In the active arm, roughly 14,000 were invited to screening; and in the control arm, around 14,000 were not invited. The combined primary endpoint was stroke, systemic thromboembolism, death, and severe bleeding over 5 years.

The investigators have already published an interim report showing that new AF was found in 3.0% of the screened population, whereas only 0.5% were found on the first ECG. No surprise there. The more you look, the more AF you find. And more than 90% of patients with newly diagnosed AF accepted initiation of oral anticoagulant (OAC) treatment. This too has been shown in other studies.

Svennberg presented the much-awaited clinical outcomes from the 5-year study in a late-breaking session. Steve Stiles from | Medscape Cardiology has excellent news coverage, which is important because the paper is not yet published.

In the two groups the mean age was 76 years, CHADSVASC score was about 3.5, and 10% had a prior stroke or TIA. A key feature of STROKESTOP was that the group randomly assigned to screening were those offered screening. The control arm comprised those who were not offered screening. These two groups constitute the main or intention-to-treat analysis. But only half of the active arm actually had the screening. The other half declined screening or could not participate in screening due to dementia or other co-morbid conditions. This group of 7000 patients (half of those offered screening) constituted the “as-treated” arm of screening.

The main result was that 4,456 events occurred in the active arm randomized to screening vs 4616 events in the control arm not offered screening. The absolute difference was 160 events. The hazard ratio (HR) was calculated at 0.96 with a lower bound of 0.92 and upper bound of—get this—0.999. And a Pvalue of 0.045. My friends, it doesn’t get much closer to the significance threshold than that.

Svennberg also presented the as-treated analysis. This was a 7000-patient subset of the 14,000 patients who were offered screening and accepted screening. The HR for the primary endpoint of these patients was much better at 0.76.

Of course, this isn’t exactly an ideal comparison because the patients who were offered screening and accepted screening were much healthier than the subset of patients who were offered screening and did not accept screening. They had a lower CHADSVASC score, fewer prior strokes, less hypertension, less vascular disease, and less diabetes. The StrokeStop authors concluded that population-based screening for AF provided a net clinical benefit in an elderly population and efforts should be made to increase participation.

We are going to have to wait for the paper but here are four comments:

First, cost efficacy. In the rationale and design paper from Europace, the STROKESTOP authors write: “Our hypothesis is that screening for AF will reduce stroke incidence in the screened population, and that this screening will prove to be cost effective.”

The cost efficacy point is crucial, because even if you accept the statistically significant result as clinically important, you have to show that this is cost effective. Oodles of prior studies of AF screening show that screening picks up AF, and leads to more office visits, and more prescriptions for meds. That is a lot of downstream stuff—all of which costs society money. Does the number needed to screen to prevent one stroke come at an acceptable cost? I have no doubt that a cost analysis is forthcoming.

Second, was it the AF detection or just enhanced care in the screened group?

One of the most interesting slides from this study showed that before screening, about 12% of the active arm had AF. After screening, this went up to 14%. So, my question is how does picking up just a 2% increase in AF episodes—many of which were brief—lead to a reduction in serious outcomes?

One is that more use of OAC helps reduce stroke, but the other possibility is performance bias. Patients who had AF detected had more clinical encounters. This could mean better blood pressure control, attention to diuretics, medication adherence, and the like. (By the way, this was the same issue in the EAST AFNET 4 trial that found early rhythm control was better than rate control. Rhythm control group had way more clinical care.)

Performance bias is not a knock on the investigators; it is just a part of pragmatic strategy trials. When the paper comes out, we will have to assess for differences in care delivered.

Third, this was a very small effect size that barely reached statistical significance. That doesn’t mean it’s not a true effect, but let’s look at the strength of the evidence. That P value is about 2 standard deviations from null. An online calculator gives it a Z score of 1.97. Now if you go to another online calculator you can convert that P value to a Bayesian factor (BF) which is simply a likelihood ratio (LR) of the probability of getting this data if there was no difference in the two arms, divided by the probability of the data given a true effect of being offered screening. The LR or BF calculates to 0.14. That is considered weak to moderate evidence. You want your LR or BF to be super low, so that even if you had a pessimistic prior belief, the strength of the evidence would sway your beliefs after the trial. For context, the Pfizer and Moderna COVID-19 vaccine trials had an LR or BF that was 0.0000 with 30 zeros.

Recall that trials are like medical tests—rather than give you a definite answer, they update your prior beliefs or knowledge. What should our prior belief about screening be? Well, Saquib and colleagues found that most screening for serious diseases—like cancer, which is far more deadly than AF—rarely finds overall mortality benefits. So, let’s be generous and say our prior for AF screening reducing hard outcomes would be 80-20 odds of no effect; 80/20 is 4.

By Bayes Theorem, which states that the prior times the LR (or BF) updates your belief after the evidence, you get 4 x 0.14 or 0.56. That means that if you held an 80-20 pessimistic prior, the probability of these data given the null hypothesis is 56%.

But let’s say you are a Martian, and you don’t know that most screening trials rarely find reductions in hard outcomes, like death, so you have a neutral prior belief of 50-50 odds of reducing a hard endpoint. 1 x. 0.14 still yields a 14% probability of this data given no effect.

Fourth, proponents of screening might say that technology will soon render AF detection akin to a vital sign, like a blood pressure. For instance, older patients with pacemakers are essentially screened for AF all the time. The advent of smart watches and devices will soon reduce the burdens of AF screening.

That means a lot of this wrangling over offering population screening will gradually become a moot point, because many if not most patients will be able to do it without doctors. The next step is not so much whether we can detect AF, but whether treating it reduces outcomes. It’s a huge question. One for the next generation of clinicians.

I can’t help but wonder whether we could accomplish the same small effect size seen in STROKESTOP by simply choosing a random sample of 75 to 76-year-old patients with CHADSVASC of 3 to 4 and testing OAC vs nothing, regardless of AF episodes.

Conduction System Pacing

What’s up with the cardiology in Copenhagen? It seems nearly every week I am discussing important science from this city. Kudos to you all. Dr. Michael Vinther presented results of the His-Alternative trial at the EHRA meeting. JACC-Electrophysiology has a simultaneous publication. The study question involves use of His bundle pacing to achieve effective cardiac resynchronization in patients eligible for cardiac resynchronization therapy (CRT).

Left bundle branch block (LBBB) can worsen heart failure (HF) in patients with systolic dysfunction because the right ventricle (RV) contracts before the left ventricle (LV), and the dys-synchrony exacerbates the already low cardiac output (CO). Numerous trials have shown that patients with LV systolic dysfunction and LBBB have improved outcomes with a biventricular (BiV) pacing device. The RV and LV leads eliminate the LBBB by pacing the heart at the same time. CRT with BiV devices is basically electrical therapy for the LBBB.

Recently, many electrophysiology (EP) docs, including me, have become interested in using conduction system pacing to accomplish simultaneous activation of the ventricles in patients with LBBB and heart failure with reduced ejection fraction (HFrEF). One of the aesthetically intoxicating aspects of pacing the His-bundle or left bundle is that conduction occurs with a narrow QRS—indicating simultaneous activation of both ventricles.

I was part of the His-Sync RCT, which showed that His bundle pacing (HBP) was not statistically superior to CRT in improving EF, but there was a signal towards benefit in EF improvement. One key lesson learned from His-Sync and other studies is that conduction system pacing can only correct “true” LBBB. A lead in the conduction system cannot correct an electrical delay of the LV when the problem is in the ventricular muscle.

His-Alternative randomized one group of CRT-eligible patients to receive His-CRT and the other, BiV-CRT. The primary endpoint was the success rate of obtaining capture of the His-bundle with narrowing of the QRS-duration. The secondary outcome was a fall in end-systolic volume > 15%. The key inclusion criteria was a “true LBBB” by the Straus criteria, or greater than 90% RV pacing, which of course activates the RV first, as in a typical LBBB.

  • His-corrective pacing was achieved in 72% of the patients in the His-CRT group. As it was in His-Sync, this study had a lot (28%) of crossovers from His-CRT to BiV-CRT.

  • At 6-months, QRS durations were similar in His-CRT vs BiV-CRT (131 vs 134 msec).

  • In the intention-to-treat (ITT) 6-month follow-up, LVEF increased by 16 in His-CRT vs 13 in BiV-CRT (not significant).

  • Pacing thresholds were higher for His-CRT compared with BiV-CRT both at implantation (1.8V vs 1.2V, p < 0.01) and at 6-month follow-up (2.3V vs 1.4V, P < 0.01).

  • The per-protocol LVEF was significantly higher at 6 months (48% vs. 42%, p < 0.05) and the end-systolic volume was lower (65 ml vs. 83 ml, p < 0.05) in His-CRT treated patients compared with BiV-CRT.


As a conduction system pacing enthusiast, I worry here about sounding too much like the content experts I criticize for being too rosy about the data. So, let’s be clear: This was a small trial with a surrogate short-term endpoint. What’s more, the higher threshold to correct the LBBB with HBP is important; over the long-term it means more generator changes and this is not a positive. CRT with BiV pacing has been proven beneficial in multiple trials, and data like this should not change our baseline practice.

But this study shows that it is feasible to correct LBBB with HBP. It shows that if you do, EF and QRS improve similarly. But these are surrogate outcomes. Progress is usually iterative; 95% effective COVID vaccines are not the rule.

My take from doing HBP for 5 years is that it is hard. The tools are still primitive. For HBP there is essentially one lead and two sheaths. For BiV pacing there are numerous leads, dozens of sheaths, and wires and inner sheaths, and snares that have all been developed to better implant BiV devices. HBP now is where BiV pacing was 10+ years ago.

Two more comments:

  • HBP can be a great fallback when there are no coronary sinus branches available. But you can’t just do HBP without practice. It’s too hard. My suggestion therefore is that operators learn to do HBP on easier cases of say sinus node dysfunction or heart block. You hone your skills so that you it is in your bag of tricks for tough BiV cases.

  • The new technique on the block is LBB pacing. Here you use the same lead, same sheath, but burrow the lead through the septum to capture the LBB. This results in a RBB, and remember, RBB means LV activation is not late. LBB is feasible – I have done it, and the thresholds are far better. It too is intoxicating in its aesthetics.

We need to study CRT with LBBB pacing instead of HBP but, alas, industry has no confluence of interests here, because if conduction system pacing works, it means fewer expensive CRT devices.

FDA Wants to Ban Menthol Cigarettes

The US Food and Drug Administration (FDA) has announced that within a year it will ban menthol in cigarettes as well as all flavors including menthol in cigars. Menthol makes it easier to start smoking, and also enhances the effects of nicotine, making it more addictive and harder to quit, said the FDA in a statement Thursday. Flavored mass-produced cigars and cigarillos are disproportionately popular among youth, especially non-Hispanic Black high school students, who in 2020 reported past 30-day cigar smoking at levels twice as high as their White counterparts, said the FDA.

Anything that reduces smoking rates—especially in the young-- is a huge plus for public health. I’ve read online that some people worry about this creating a black market for this product. And this could be problematic.

Yes, please, it is always important to think about tradeoffs. Remember the wisdom of Thomas Sowell who said there are no solutions, only tradeoffs. That said, smoking is just so hazardous, that to protect public safety, government should act. A world in which it is harder for people to get addicted to smoking is a better one.

Heartfelt Essay

My colleague Melissa Walton-Shirley has penned one of the best essay’s I’ve ever read on | Medscape Cardiology. You should go to a quiet place, turn off your notifications, and slowly read “Losing Both Parents During COVID: A Physician's 'Unwitnessed Grief'.”

Melissa Walton-Shirley is a force of nature; in the last few years of our gatherings at Medical meetings, we knew her parents were struggling with health issues. Her essay poignantly tells the story of how this terrible pandemic has robbed families and patients of having a good death.

Here is one quote: “In early May, I would get the closest to my father that I would ever be again—6 feet away, as he sat on the swing on their front porch. It was there that he told me what I already knew: he was dying. There was no hug to follow.”

Look, I’m not the most religious person in the world, but the goodness of Melissa Walton-Shirley surely increases the LR of there being a benevolent force in this Universe.


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.