COMMENTARY

Mandrola's Top 10 in Cardiology for 2020

John Mandrola, MD

Disclosures

December 15, 2020

1. COVID-19 and Cardiology

Before 2020, heart disease and cancer dominated medical news. The COVID-19 pandemic reset all that. Now, attention has turned to virologists, epidemiologists, hospitalists, and the emergency medicine and intensive care unit doctors.

I lead the This Week in Cardiology podcast with pandemic news each week because the lessons learned in the science of COVID-19 inform the most important aspect of cardiology: how to translate evidence to patient care.

The blistering pace of science has been both a blessing (think vaccines) and a curse (think retractions and flawed observational studies that have been viewed more than 800,000 times).

2. SGLT2 Inhibitors

In patients with heart failure with reduced ejection fraction, the placebo-controlled EMPEROR-Reduced trial of empagliflozin largely confirmed the benefits seen with dapagliflozin in DAPA-HF.

EMPEROR found a statistically significant 25% reduction in the combined endpoint of cardiovascular (CV) death and hospitalization for heart failure. A meta-analysis of the two trials confirmed a benefit for the drugs in all-cause death, CV death, heart failure, hospital admissions, and a composite renal outcome. And both sodium-glucose transport protein (SGLT) 2 inhibitors produced similar benefits in patients with and without diabetes.

The novel inhibitor of both SGLT1 and 2, sotagliflozin, was also shown to prevent CV outcomes in two trials: SOLOIST enrolled patients with diabetes and worsening heart failure, and SCORED comprised patients with diabetes and chronic kidney disease (CKD).

There was more good news for CKD when the DAPA-CKD trial showed that dapagliflozin reduced a composite renal outcome by 39% compared with placebo. These findings were similar to those seen in the CREDENCE and CANVAS studies.

If only all our therapies had this much supportive evidence. I predict that this class of drug may break the string of failures in therapies for heart failure with preserved ejection fraction.

My favorite part of this story is that we may not have discovered this class of drugs if Steve Nissen, MD, and Kathy Wolski, MPH, did not have access to unpublished data involving rosiglitazone. Their meta-analysis prompted the FDA to mandate cardiovascular outcome trials for new diabetes drugs.

3. Aortic Valve Disease

Until January, standard practice for patients with severe aortic stenosis (AS) was to wait for symptoms to develop before intervening. This year, however, a South Korean group published a randomized controlled trial (RCT) comparing early aortic valve surgery vs conservative therapy in asymptomatic patients with severe AS and found a massive 91% reduction in death with early surgery.

That trial only added to the enthusiasm of treating patients with AS. At the European Association for Cardio-Thoracic Surgery (EACTS) annual meeting, Joseph Bavaria, MD, reported that the number of transcatheter aortic valve replacement (TAVR) procedures performed in the United States has now surpassed the number of surgical aortic valve replacement (SAVR) procedures in all its forms.

 

Yet 2020 brought extremely sobering data on TAVR relative to SAVR. PARTNER 2 authors reported 5-year data in intermediate-risk patients with AS. The results are complicated, but the early advantage of TAVR clearly decreased over longer follow-up.

This catch-up effect was also seen in the 2-year results of PARTNER 3, a trial that compared TAVR vs SAVR in lower-risk patients. Unlike all other TAVR trials, PARTNER 3 included rehospitalization in its primary endpoint. This led to markedly positive results for TAVR at 1 year; but in the 2-year results, the rates of stroke and death did not differ significantly in the two groups. Valve thrombosis was significantly higher in the TAVR arm (2.6% vs 0.7% for SAVR; P = .02).

History has shown that many cardiac procedures begin in a surgeon's hands (revascularization, ablation, pacing, etc.) but transition to the cardiologist with the aid of technology. The thing about TAVR is that despite substantial iterations in the procedure, real-world data only add to the concern raised from trial results. Bavaria noted at EACTS that over the past 5 years, the 30-day rates of stroke and need for pacemaker have not budged. "We have to do better," he said.

TAVR is a great advance, but as we expand this procedure to lower-risk patients and those with the more common pathology of bicuspid valve disease, we should have a much higher bar of evidentiary support. Do we have it? I think not.

4. Left Main Coronary Artery Disease

In January, NOBLE investigators published 5-year results of their trial comparing percutaneous coronary intervention (PCI) with drug-eluting stents vs coronary artery bypass grafting (CABG) for patients with left main coronary artery disease. PCI was inferior to CABG for the primary endpoint of death, nonprocedural myocardial infarction (MI), and repeat revascularization.

The clarity of NOBLE stands in contrast to the opacity of EXCEL, the larger trial comparing PCI to CABG in patients with left main disease. The 5-year results of EXCEL may have been published in late 2019, but the discussion surrounding this trial and how to manage patients with left main disease continued in earnest this year.

I see two important facets of this story. First is trust: the EXCEL controversy began in 2019 when someone leaked the trial data set to the BBC. The key feature of those data were rates of MI according to the universal definition—a prespecified secondary endpoint that was not reported in the 3-year or 5-year results.

EXCEL authors initially called the leaked data "fake information." But we learned in 2020 that it wasn't fake. It was nearly exactly the same data published in their  New England Journal of Medicine (NEJM) letter and in a reanalysis of the trial published in the Journal of the American College of Cardiology.

The clinical issue of treating patients with left main disease is easier. In June, James Brophy, MD, PhD, published a Bayesian reanalysis of the EXCEL trial. Bayesian analyses are useful because they offer the probability of a hypothesis (benefit or harm) given the data.

Using EXCEL data alone, Brophy calculated a 95% probability that the 5-year primary outcome difference was increased with PCI vs CABG and a 99% probability that there was excess mortality with PCI. When he incorporated previous trials of PCI vs CABG, the estimated probability of excess mortality with PCI decreased to 85%.

Isolated left main disease is uncommon, but when it occurs, patients ought to be counseled that outcomes with PCI are inferior over the long run.

5. Early Rhythm Control in AF

The top-line results of three trials published this year tempt you to believe early rhythm control produces better outcomes. I encourage a deeper look.

The EAST-AFNET 4 trial was a pragmatic but unblinded trial comparing early rhythm control with either drugs or ablation vs usual care in patients with atrial fibrillation (AF). Early rhythm control led to a statistically significant 21% reduction in the composite primary outcome of CV death, stroke, heart failure admission, or acute coronary syndrome.

The STOP AF and EARLY-AF trials both compared ablation with a cryoballoon vs antiarrhythmic drugs in patients who had not previously received rhythm control. Ablation beat drugs in the surrogate endpoint of arrhythmia episodes in both trials.

I wrote about the complexities of EAST-AFNET 4 and the caveats of both early ablation trials. Rhythm control has a role. It can help patients, but it comes with the hazards of drug-induced proarrhythmia and procedural complications. We'd also be wise to remember that AF often stems from causes outside the atria.

6. Statin Nocebo Effects

Researchers at Imperial College London have once again shown that advances in knowledge do not always require mega-trials.

In 2017, their 200-patient ORBITA trial showed that PCI resolved ischemic myocardium but was no better at improving exercise time than a placebo or sham procedure. The obvious conclusion: the improvement in quality of life that patients feel after a PCI may stem in large part to the positive expectations of being cared for or being "fixed."

The opposite of the placebo effect is the nocebo effect—when negative expectations lead to negative effects. If people think a statin will make them feel poorly, they may be more apt to feel poorly.

The SAMSON trial enrolled 60 patients who had previously stopped statins. Using an of N-of-1 design, the researchers elegantly showed that, yes, patients do have significant side effects from statins, but it is not due to the statin chemical but from the act of taking a statin pill.

The importance of this work is twofold: If we use the results of SAMSON to teach patients, many will restart a beneficial drug—as did 30 of the 60 patients in the trial. The second benefit is that it confirms the importance of our words and actions at the bedside. Our ability to help or harm transcends biochemistry and physiology.

7. Misuse of Noninferiority Trials

Two trials this year clearly show how noninferiority trial designs can be misused to deliver favorable results. Medical advances ought to be superior; a noninferior therapy makes sense only if it is more convenient or less invasive/toxic/costly than the current standard. Such is the case with direct-acting oral anticoagulants (DOACs) vs warfarin.

The PRAGUE-17 trial compared percutaneous left atrial appendage occlusion (LAAO) to DOACs in patients with AF. LAAO is a stroke prevention strategy, and its potential advantages include avoidance of long-term anticoagulation (safety and convenience). To show noninferiority, you would measure efficacy (stroke prevention) in the two arms. If the difference in stroke rates between the two study arm was not greater than a prespecified margin, you could declare the device noninferior. Safety (bleeding and procedure complications) are measured separately.

That is not what the investigators did. Instead, they combined efficacy and safety into a single composite endpoint. By combining endpoints that you expect to go in opposite directions (thrombosis and bleeding), you bias to the null. Indeed, that is what happened.

The PRAETORIAN trial compared the subcutaneous implantable cardioverter defibrillator (S-ICD) to the standard transvenous ICD. The potential advantage of the S-ICD is that its subcutaneous location may avoid lead complications. But this feature could compromise efficacy.

Transvenous ICDs have a long-proven efficacy record, so a proper noninferiority trial would compare efficacy of the two devices—which is sensing and termination of ventricular tachycardia. Safety (device complications) would be measured separately.

But again, PRAETORIAN authors used a combined endpoint of device-related complications and inappropriate shocks—components that everyone knows will go in opposite directions. As expected, the rates of events in the two arms cancel each other out, and the authors declared noninferiority.

Both LAAO and S-ICD have become normalized in electrophysiology practice. Yet the evidence for both is dubious.

8. Postpublication Peer Review via Social Media

Scientific studies have had to pass trials by Twitter for nearly a decade. Yet the secretive insular practice of peer review still holds sway.

In 2020, public peer review once again exposed the weakness of traditional peer review and perhaps also the problem with metrics as currency in academic publishing.

Early in the pandemic, two medical debates garnered attention. One was the possible association of renin-angiotensin inhibitor use and mortality; the other, use of hydroxychloroquine therapy.

A company called Surgisphere purported to have culled big data from the electronic health records of hundreds of hospitals globally. The Lancet and NEJM published two observational studies, presumably after careful editorial and peer review. The papers quickly generated both citations and attention in the media.

The trial by Twitter (and blogs) was short, the guilty verdict unanimous, the punishment harsh.

On Twitter, Boback Ziaeian, MD, detailed the obvious statistical flaws in the papers. On his influential blog, Columbia University statistician Andrew Gelman mocked the editor of the Lancet for publishing such a flawed study after Horton had opined 5 years previously about the sickly condition of scientific publication.

Within days, the top two journals in medicine retracted both papers. University of Pittsburgh statistician Andrew Althouse, PhD, had a nice explainer on how this can happen.

I will not argue that context-specific peer review is without merit. But the need for speed in this year's pandemic has further strengthened the case for a far more transparent open and public review of science.

9. Disparities in Health Outcomes

The high costs and inequities of the U.S. healthcare system were well known before COVID-19. The pandemic laid bare the consequences of such an unjust system.

At the end of November, data from the Centers for Disease Control and Prevention revealed what frontline clinicians know: American Indian, Hispanic, and Black patients have higher rates of hospitalization and death from COVID-19 relative to White, non-Hispanic people.

In April, cardiologist Clyde Yancy, MD, argued in JAMA that although some of these disparate outcomes were due to higher rates of comorbid conditions, such as diabetes, hypertension, and obesity, the "concerns go beyond these comorbidities." Yancy noted that "the communities where many black people reside are in poor areas characterized by high housing density, high crime rates, and poor access to healthy foods."

I've seen many patients who were sickened by COVID-19 because their jobs could not be done via a laptop from home or because they were forced to live in cramped conditions.

When there was a booming economy, disproportionate health outcomes based on status in society received cursory attention. The pandemic has heightened awareness of the need for structural change in society.

Such changes are complicated policy matters that surely fall outside the realm of clinical medicine. But I agree with Yancy's take that 2020 ought to be a trigger to muster the most important factor in making change: the will to do it.

10. Medical Book of the Year

The 15th-century thinker Blaise Pascal  wrote that, "all our dignity consists in thought. It is by this that we must raise ourselves up…. Let us work, therefore, to think well: for such is the principle of morality."

In a year where intolerance of ideas is spreading as fast as a respiratory virus, Theodore Dalrymple's False Positive is a must read.

The book began as an exercise to help the author's nephew, a medical student, with an exam on critical appraisal. Each of its 52 short chapters is dedicated to a weekly issue of NEJM from 2017. The author chooses the articles that he feels suffer most from a dearth of reflection.

You may disagree with some of Dalrymple's views, but engaging with how a recently retired former prison doctor explores what is left unsaid in a year's worth of articles in the NEJM has never had more relevance.

Comments

3090D553-9492-4563-8681-AD288FA52ACE
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.
Post as:

processing....

Recommendations