May 21, 2021 This Week in Cardiology Podcast

John M. Mandrola, MD


May 21, 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 May 21, 2021, John Mandrola, MD comments on the following news and features stories.

First an Announcement

I am heading off on holiday for 2 weeks. TWIC will return June 11. Thanks for your support and please remember to take the time to give us a rating or a review b/c that helps others find us.


Nearly every day in my clinic, a cardiac clinic, mind you, people marvel at the vaccines. Our hospital is nearly empty of COVID, and Eric Topol Tweeted yesterday that US case numbers and case positivity are as low as they have been in more than a year.

Three weeks ago, we had 51,000 people at Churchill Downs for the Kentucky Derby—and there is no spike in cases. The front page of the Wall Street Journal Thursday had the European Union opening its borders. And the Heart Rhythm Society will have an in-person meeting in July. Even cases in India are going down. Globally it isn’t over, but it’s getting close. Thank goodness. What a year.


The best study at the 2021 American College of Cardiology (ACC) was clearly the Canadian-led Left Atrial Appendage Occlusion Study III (LAAOS III) trial, led by surgeon Richard Whitlock. Enrolled patients had atrial fibrillation (AF) and were having cardiac surgery for some other reason. Nearly 2400 patients were randomly assigned to two groups—surgical appendage closure vs no closure. The most important thing to remember about this trial is that both groups—closure and no-closure—remained on anticoagulation (AC).

  • The primary endpoint was stroke and systemic embolism and closure resulted in a statistically significant 2.2% absolute risk reduction (ARR) in stroke.

  • The Kaplan Meier (KM) curves widened over time, and the add-on procedure required only an extra 6 minutes of pump time and no safety issues.

  • LAAOS III showed that concomitant appendage closure reduced stroke in addition to oral AC.

  • This trial did not test the matter of coming off oral AC and thus, it doesn’t at all inform the percutaneous LAAO debate.

Recall that in the Watchman vs Warfarin trials, ischemic strokes were higher in the device arm and the only benefit (if any) from endocardial closure comes from removal of anticoagulation over time. Therefore, the positive results of LAAOS III should not enhance enthusiasm for device-based closure, which is limited by frequent leaks, a foreign body left in the arterial circulation and the need for removal of AC to achieve any possibility of net benefit.

Watchman Registry

While LAAOS III was one of the best studies of ACC, a featured clinical research paper on the 1-year outcomes of Watchman from the National Cardiovascular Disease Registry (NCDR) is one of the most flawed.

Remember the saying from Joseph Bavaria:

  • Science tells us what we can do.

  • Trials tell us what should do.

  • Registries tell us what we are actually doing.

  • Mandrola addition: If you try to extract anything more from a registry, then you are on shaky ground.

And that my friends is what happened. A look at baseline characteristics reveals what happens when regulators are lax and enthusiasm is high—indication creep: The Center for Medicare and Medicaid Services (CMS) national coverage decision calls for Watchman use in patients deemed unable to take long-term anticoagulation. But the average HAS-BLED score was 3.0. A score of 3 would predict a modest rate of bleeding of 3.74 bleeds per 100 patient years—a risk level not usually considered high enough to warrant discontinuation of oral anticoagulation. What is more, the rate of major bleeding was 69%, meaning that more than one in three patients had not had a major bleed—why are these patients unable to take oral AC?

But the major issue: the authors present a 1-year-KM rate of ischemic stroke of 1.53% and compare it to a KM rate of bleeding estimate based on historical controls of 6.2% and then show the nice graph of a 77% reduction of stroke. Three reasons why you cannot compare patients in a registry with “historical” controls:

  • In a systematic review of papers looking at patients with AF who did not receive oral AC, Quinn and colleagues showed that untreated stroke rates have massive variation, even in the same CHADVASC group. Translation: no one has any earthly idea what the rate of stroke is in untreated AF patients.

  • These were 1-year stroke rates and the Watchman protocol calls for patients to stay on oral AC for a period of time. So 1-year rates will be artificially low early in the Watchman arm.

  • Third, and this is the craziest part—at 1 year, 29% of patients were lost to follow-up. They didn’t have data on 1 in 3 patients!

Angiotensin Receptor-Neprilysin Inhibition

PARADISE MI is an industry-sponsored global trial comparing the angiotensin receptor-neprilysin inhibitor (ARNI) sacubitril/valsartan to ramipril in post-MI patients who have an ejection fraction below 40% and clinical heart failure (HF). Patients also had to have one of eight risk enhancing features.

PARADISE MI was designed to enroll higher-risk patients, and if you believe the amazing risk reduction in HF with reduced EF (HFrEF) seen in PARADIGM-HF, you would expect the ARNI to crush plain old ramipril. That’s not what happened.

In the primary endpoint (PEP) of cardiovascular (CV) death and HF events, patients treated with sacubitril/valsartan experienced 6.7 events per 100 patient-years compared with 7.4 events per 100 patient-years for ramipril-treated patients. The 0.7 absolute difference translated to a hazard ratio (HR) of 0.90, with 95% confidence intervals (Cis) ranging from 0.78 to 1.04. The P-value of 0.17 was less than 2 standard deviations from the null—not even close to the accepted standard.

To my surprise, the authors then used language and presentation techniques that seemed to distract our attention from the failure of this trial to meet its primary endpoint. First they showed the KM curve for the PEP on a truncated y-axis, thereby accentuating the nonsignificant HR of 0.90. They then showed two other analyses—the PEP with total events and investigator-adjudicated events, where the ARNI looked positive.

Presenter Marc Pfeffer said in his presentation that these non-PEP analyses were hypothesis-generating but then had this quote on his second bullet point on the conclusion slide. Note the causal language for hypothesis-generating associations:

“Pre-specified observations of reductions in both the investigator reports of the primary composite as well as in the total (recurrent) adjudicated events support incremental clinical benefits of sacubitril/valsartan.”

I wrote a column on this trial and in it I speculated that, based on the miss in PARADISE and in PARAGON, the HF with preserved EF (HFpEF) trial, perhaps we are over-estimating the benefits of sacubitril/valsartan.

Another trial presented at ACC using ARNI, the LIFE trial, enrolled patients with class IV HF due to low EF and compared sacubitril/valsartan to valsartan alone. This trial found that the ARNI drug:

  • Was not superior to valsartan alone in reducing the primary endpoint of NT-proBNP levels.

  • Did not improve the clinical composite of number of days alive, out of hospital, or free of HF events.

  • Did not decrease the risk of death from CV disease or HF hospitalizations. And there was a small but statistically significant increase in non-life-threatening hyperkalemia in the sacubitril/valsartan treatment arm.

Renal Denervation

Ajay Kirtane presented results of the RADIANCE-HTN-TRIO sham-controlled trial of ultrasound-based renal denervation (RDN) in patients with resistant hypertension (HTN). Lancet published the results in a simultaneous publication. The authors got all patients on a single pill with good doses of a calcium channel blocker, an ARB, and a thiazide--a sound methodologic move.

The trial screened 1000 patients and enrolled 136, which tells you these were highly selected patients and that true resistant HTN is likely rarer than most of us think. The sham-controlled difference was a mere 4.5 mm Hg. And this was at only 2 months. The current studies of RDN have been much better about controlling cofounding; the procedure has iterated to be better, but the trouble is now that we have better isolated the effect of RDN, it looks to be a very modest reduction.

We need a heck of lot of more data—especially on outcomes, not just the surrogate marker of systolic blood pressure—before accepting and embracing another expensive procedure.


Here’s the super common scenario – your patient has had a stent and taken dual antiplatelet therapy (DAPT) for 6-18 months. Now you want to switch to single therapy. Locally, most of us go to aspirin (ASA) but perhaps clopidogrel is better.

In the HOST-EXAM trial, South Korean trialists randomly assigned about 5500 of these patients to either clopidogrel or ASA and measured a primary endpoint of death, MI, stroke, ACS, or bleeding. The winner was clear: clopidogrel.

  • Over 2-year follow-up, 5.7% of those in the clopidogrel arm had a PEP vs 7.7% in the ASA arm.

  • Bleeding was 30% lower in the clopidogrel arm.

  • MI, stroke, and even intracerebral hemorrhage were lower in the clopidogrel arm.

  • All-cause death was a bit higher in the clopidogrel arm, 1.9% vs 1.3%, but this was not significant and largely driven by non-cardiac death.

One concern was whether evidence generated in Asian patients can be applied to non-Asian patients. But the Korean population has a high prevalence for clopidogrel loss-of-function alleles so this would have favored ASA. That clopidogrel still won strengthens the case.

Five years ago, I would have been surprised by these results, but the more I study ASA in contemporary trials, the more I have become unimpressed with its efficacy and safety. And it’s easy to forget that in AVERROES, apixaban vs ASA 81 mg in AF, the bleeding rates for full dose apixaban were essentially the same as with ASA. As Sue Hughes mentioned in her news coverage, the 20-year-old CAPRIE trial of clopidogrel vs ASA 325 mg for secondary prevention found a modest but significant 0.5% reduction in hard endpoints.

I can’t see how this trial doesn’t cause consideration of a change in practice when it comes to transitioning from DAPT to single drug antiplatelet therapy.

Alcohol and AF

Greg Marcus and his team from the UCSF have been publishing some elegant studies surrounding the effects of alcohol consumption. At ACC, he presented results of nifty experiment studying the acute effects of alcohol on AF episodes. They took patients with known paroxysmal AF and set them up with ECG monitors and alcohol sensors. Each patient acted as their own control. They found:

  • One drink increased the odds of having an AF episode by more than 2-fold.

  • Two or more drinks increased the odds of AF by more than 3-fold.

  • The most likely duration for an AF episode after a drink was 3-4 hours.

  • There seemed to be a dose response relationship b/w concentration of alcohol and risk of AF. Every 0.1% increase in the blood alcohol concentration in the last 12 hours was associated with a 38% greater odds of an AF episode.

The key take-home from this experiment is that, in patients with AF, alcohol is a clear and modifiable risk factor for having AF episodes. This paper adds to the mountain of previous evidence linking alcohol to AF.


One of the many critiques of CASTLE-AF, the seminal RCT that found AF ablation was superior to anti-arrhythmic drugs (AAD) in patents for HFrEF was its design. CASTLE-AF took HF patients who had either failed AAD, could not tolerate AAD, or did not want AAD and randomly assigned them to more AAD vs ablation. It’s little wonder that the AAD underperformed.

RAFT-AF is a Canadian-led RCT comparing rhythm control with AF ablation vs rate control (including atrioventricular junction ablation and cardiac resynchronization therapy) in patients with HF. This trial changed the comparison to the strategy of primary AF ablation (get rid of or reduce the problem) vs rate-control (control the worst manifestation of the AF—the high ventricular rate). The investigators planned to enroll 600 patients, 300 in each group.

Alas, in 2018, data safety monitoring board recommended stopping the government-funded trial due to low enrollment, lower event rates, and perceived futility. So instead of 300 patients in each arm, there were only 200. Topline results were:

  • 32.5% in the rate control arm experienced a primary endpoint of death or HF event vs 23.4% in the ablation arm.

  • The HR was 0.71 and the CI ranged from 0.49 to 1.03, and the P-value was 0.066.

  • The KM curves started separating at 1 year. Quality of life endpoints favored the AF ablation arm, though the trial was unblinded.

All we can really take from the trial was the first bullet point in the trialists conclusion: “The numeric reduction in death/HF with the ablation-based strategy vs rate-control did not meet our accepted threshold of significance.”

RAFT-AF is inconclusive. It ended up being an under-powered trial. The effect size did seem to increase over time, and one can postulate that this is a type 2 error—which means that AF ablation is indeed better than rate control, and if there were more patients and more events, the trial would have shown that. But that remains highly speculative.


Well established is that COVID-19 infection has been associated with thromboembolic complications. Early in the pandemic, there was much interest in using various anticoagulation regimens. Top-line results from a platform of three linked clinical trials — REMAP-CAP, ACTIV-4, and ATTACC — were announced in January and said to show that a full therapeutic dose of heparin was superior to low prophylactic doses for the PEP of need for ventilation or other organ supportive interventions at 21 days after randomization. Five months on, these trials remained unpublished.

At ACC, Dr. Renato Lopes, presented results of the ACTION trial comparing full-dose rivaroxaban with standard prophylactic AC in patients hospitalized with COVID-19 infection and who had an elevated d-dimer. Of importance, ACTION measured outcomes at 30 days, not in-hospital.

After randomly assigning more than 600 patients:

  • The trial did not show a statistically significant reduction in a combined primary endpoint of death, duration of hospital stay, and duration of oxygen use.

  • Bleeding rates were 3.6 times higher and death rates were numerically higher with rivaroxaban.

ACTION delivers two messages: one is don’t use oral rivaroxabanin hospitalized COVID-19 patients who do not have an obvious venous or pulmonary thromboembolism. The second is the power of an RCT for answering clinical questions. Yes, patients with COVID-19 who have elevated d-dimers may have an increased risk for clotting complications, but this does not mean full-dose oral AC will provide benefit.


Here is another common scenario: an older patient with frailty presents with acute decompensated heart failure. Everyone is focused on medical and possibly device management. None of the guidelines address physical functioning.

A group of docs have designed an innovative rehab program—things like helping patients with rising from a chair without hand support, and improving balance, endurance, and gait speed. The program starts in hospital and transitions to the outpatient setting. A key goal of the intervention during the first 3 months (the outpatient phase) was to prepare the patient to transition to the independent maintenance phase (months 4 through 6). The problem is that most such intervention trials have been neutral. Trialists in the REHAB-HF trial tested this intervention. Their primary endpoint was a battery of short physical performance measures. These have been used previously in other trials, shown reliability, and seem to predict clinical outcomes. The results were clear and positive. Patients in the active arm had significantly improved measures in physical functioning at 3 months.

  • Six-minute walk time increased.

  • Frailty scores decreased.

  • Kansas City Cardiomyopathy Questionnaire scores improved.

  • Depression scores decreased.

In the discussion section the authors note the magnitude of the effect was greater than previous trials, and the positive effects held up in most subgroups. A secondary endpoint of rehospitalization showed no difference but I would emphasize that the trial was not powered for a binary clinical outcome but continuous measures of physical function.

Some caveats:

  • Trialists screened 27,000 patients to enroll less than 400; these are highly selected patients.

  • Patients could not be blinded—there may have been performance bias.

  • We need to learn more about the actual interventions.

  • We need to be careful. For many patients, HF is an end-of-life condition and for these patients, palliative care and hospice may be the best approach.

Finally – 81 mg vs 325 mg of ASA

ACC featured a large pragmatic trial of ASA 81 mg vs 325 mg for secondary prevention of major adverse cardiac events (MACE). The main results of the ADAPTABLE trial were that there were no significant differences in MACE or bleeding outcomes with the two doses.

The trial featured novel methods in which researchers used the (PCORnet), a group of 40 US centers committed to compiling data in a common format; invitations to enroll in the study were sent to eligible patients identified from medical records. Consent and randomization took place on the patient web portal. Participants then purchased aspirin at the assigned dose themselves, and all follow-up was done virtually or on the phone, with outcomes ascertained remotely (from patient reports, electronic medical records, and insurance claims) without adjudication.

The good news is that this is a less expensive way to do trials. And there were lessons learned that will guide the way to doing more pragmatic trials. The bad news is that 15,000 patients were enrolled and we still have no idea which dose is better. About 85% of patients began the trial on the 81 mg dose. Was there really equipoise here? Then during the trial 41% of those randomized to 325 switched to 81 mg and only 7% switched the other way. Since switching was not likely at random, we cannot make definitive clinical conclusions about the best ASA dose.


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.