Mandrola’s Top 10 Cardiology Stories of 2019 – Medscape

1. The ISCHEMIA Trial

The ISCHEMIA trial transcends the cath lab and influences much of cardiology practice. Its findings also reach into human psychology.

We have only had a few weeks to digest the results; when published, the manuscripts will spark more discussion. For now, there are clear messages.

First is to consider the two arms of the trial. The decision to randomly assign patientswith proven and significant coronary diseasebefore angiography makes this a comparison of two strategies: an early invasive strategy vs a conservative strategy with interventions only for patients who had continued symptoms.

Over 3 years, the early invasive strategy compared with optimal medical therapy did not lead to fewer cardiovascular (CV) deaths, myocardial infarctions (MIs), hospitalizations for unstable angina or heart failure, or cardiac arrest. Three observations bolster our confidence in this null result: 1) the priors from multiple similar studies, such as COURAGE and BARI-2D; 2) the fact that no component of the primary endpoint showed any significant difference, and 3) the lack of treatment effect in any subgroup.

Co-primary investigator Judith Hochman, MD, from New York University, seemed struck by the latter point. In the press conference, she made special mention that patients with three-vessel disease had higher event rates than patients with single-vessel disease, but "nevertheless, there was no advantage to the invasive approach."

We also learn a lot from the 23% crossover rate. That is, more than three quarters of patients with significant ischemic heart disease avoided revascularization and had the same outcome as those treated medicallywithout knowledge of the anatomy. (Asterisk here for the use of blinded CT to exclude left main disease.)

Now the psychology part. My interventional colleague shared this case with me. A patient had coronary angiography after a positive stress test, done for fatigue. It showed a tight proximal left-anterior-descending lesion and two other focal lesions in the circumflex distribution. He said this is right out of the ISCHEMIA trial; treat medically. No, said the referring cardiologist. The surgeon agreed; surgery tomorrow.

The psychology question: When will cardiologists, patients, and society feel safe in trusting the message of Bernard Lown, MD, who taught that chronic coronary artery disease is not a time bomb in the chest but a condition that should be treated with reassurance, lifestyle changes, and medical therapy? If Max Planck was right, it may take many generations for us to embrace ISCHEMIA's truths and stop fearing chronic coronary disease.

Rare is the accurate use of the word "gamechanger." The DAPA-HF trial, a large global randomized controlled trial (RCT), found that dapagliflozin, an inhibitor of sodium-glucose cotransporter 2 (SGLT2), reduces CV death, worsening heart failure, and overall death when used on top of optimal medical therapy in patients with heart failure due to a reduced ejection fraction.

These benefits clearly change the game of heart failure care. The New England Journal of Medicine (NEJM) published the main trial in September, and primary investigator Professor John McMurray, University of Glasgow, presented the key subgroup data at the American Heart Association (AHA) meeting: Dapagliflozin provided the same benefits in patients with and without diabetes. Dapagliflozin is no longer solely a diabetes drug. This may be just the beginning; there are many more ongoing trials of SGLT inhibitor drugs.

The presentation at the American College of Cardiology meeting of two trials of transcatheter aortic valve replacement (TAVR) in low-surgical-risk patients brought clinicians to their feet.

The PARTNER 3 trial found a much lower rate of the primary endpoint of death, stroke, and rehospitalization at 1 year, with the SAPIEN 3 system (Edwards Lifesciences) compared with surgical aortic valve replacement (SAVR). The EVOLUT trial, using the self-expanding valves from Medtronic, showed noninferiority compared with SAVR in an interim analysis of the composite primary endpoint of death or stroke at 24 months.

Later in the year, the US Food and Drug Administration approved both types of valve for use in lower-risk patients, but numerous issues should give us pause before we fully embrace TAVR over SAVRespecially in younger, healthier patients.

Both trials in low-risk patients had very short follow-up. Although it is true that we don't have perfect data on the durability of surgical valves, 1- or 2-year follow-up is inadequate for decision-making in younger patients. (See point 4 below.)

The large absolute risk reduction for TAVR in PARTNER 3 occurred because the investigators chose to include rehospitalization in the primary endpoint. No other TAVR trial did this; although stroke and death were still lower in the TAVR arm, this choice biases the results in favor of TAVR.

The EVOLUT trial enrolled 1500 patients, but the preliminary results included 2-year follow-up in only 72 patients in the TAVR group and 65 patients in the surgery group. What's more, 12 times more patients in the TAVR arm had a permanent pacemaker at 1 year.

Data presented at the Transcatheter Cardiovascular Therapeutics meeting and the European Association of Cardiothoracic Surgery meeting later in the year also suggest caution before TAVR is expanded to low-risk patients.

Vinod Thourani, MD, from Piedmont Heart in Atlanta, presented 5-year-results of the PARTNER 2A trial comparing TAVR to SAVR in intermediate-risk patients. The results upheld the 2-year finding of noninferiority, but the landmark analysis was interesting: at 0 to 2 years, death and stroke were numerically lower with TAVR, but at 2 to 5 years, the Kaplan-Meier curves switched and the primary endpoint was nearly 7% higher for TAVR. The caveat is that 25% of TAVR procedures in this trial were performed transthoracically, an access associated with a higher risk for stroke. Nonetheless, Thourani said that these findings mandate reexamination of later time points, and follow-up in PARTNER 2A has been extended to 10 years.

University College London statistician Nick Freemantle, PhD, added to the concern with PARTNER 2A when he noted the high rate of patients lost to follow-up (11.7% for TAVR and 20.4% for SAVR). He called this a "massive challenge to internal validity."

Surgeon Fabio Barili, MD, PhD, from Cuneo, Italy, presented results of a meta-analysis of six trials comparing TAVR and SAVR specifically looking at the time-varying effect of the two treatments on mortality rates. His main point was that the assumption of proportional hazards over time was not met. TAVR is better than SAVR at 0 to 12 months; at 12 to 40 months, the two procedures appear equivalent; but after 3.3 years, mortality is up to 31% higher with TAVR.

Finally, self-proclaimed TAVR proponent Joseph Bavaria, MD, presented sobering real-world evidence from the Society of Thoracic Surgeons (STS)/American College of Cardiology (ACC) TVT (Transcatheter Valve Therapies) registry. Four observations gave him concern: Despite iterations in valve technology, there have been no significant reductions in stroke rates, 30-day mortality, or severe complications (which he called "disasters"), and the rate of new pacemakers after TAVR in the past few years is nearly 12%.

I am asking: Are we sure about TAVR in the long term as it moves into younger patients?

The 2019 Bank of Sweden Prize in Economic Science in Memory of Alfred Nobel went to three researchers from Cambridge, Massachusetts, for their work in using RCTs to determine the best social interventions in poor countries.

The RCT is central to our field because use of evidence is what separates clinicians from palm readers. Professor John McMurray sent me a slide showing a list of the greatest medical inventions of our generation as voted by BMJ readersthe RCT made the short list, alongside antibiotics, vaccines, and the discovery of DNA.

Imagine for a moment if instead of implementing the Hospital Readmissions Reduction Program across the board, we assigned half the hospitals to an active arm and the other half to control. We would now know if the program worked. The endless dissections of observational studies could stop.

Another glaring example: use of the mechanical support device Impella (Abiomed) has sharply risen over the past few yearswithout any RCT-level evidence that it reduces hard outcomes. Two independent observational studies presented at this year's AHA meeting found strong associations of harm with the device. Notable in 2019 was publication of the rationale and methods paper for the DanGer Shock trialan adequately powered mortality study of Impella use in patients with cardiogenic shock. Kudos to the Danish-led research team, who have enrolled 100 of 360 patients.

The question is: Shouldn't we do these trials before acceptance of a device?

Speaking of using the RCT to answer an important question, Alex Voskoboinik, MD, and colleagues from Melbourne, Australia, decided to test the question of whether abstinence from alcohol in moderate drinkers would reduce atrial fibrillation (AF).

The Alcohol-AF trial makes this list because symptomatic patients with AF now face either antiarrhythmic drugs or a major cardiac procedureboth notable for their inelegance. This simple technique for improving AF had never before been tested in an RCT.

The Alcohol-AF study earns high marks for its lack of industry funding, strong methods, and results that have immediate clinical relevance: Abstinence or reduction in moderate drinkers reduced AF burden and AF recurrence rates, improved symptom severity, and promoted weight loss and lower blood pressure.

Finally, I must disclose a bias. In all my years of listening to investigators present their work, I have never heard a more measured and humble description of a study. On multiple occasions during his talk and in the interviews after, Voskoboinik made great effort to emphasize the limitations of this experiment. His humility toward science was intoxicating.

The nonindustry-funded, investigator-initiated German ISAR-REACT 5 study surprised the cardiology community when it showed that prasugrel trounced ticagrelor for the reduction of death, myocardial infarction, or stroke at 1 year. And prasugrel achieved this superiority without an increase in major bleeding. My interventional partner had a simple explanation. She said when you give a patient a drug that their life depends on, once-daily drugs are preferred. Ticagrelor is a twice-daily dosed drug.

One of my favorite aspects of this trial comes in the attached data-sharing statement. "Will the data collected for your study be made available to others?" Answer: "Yes." This, my friends, is how it should be for all studies. (See point 10 below.)

The medical conservative award this year goes to the Dutch investigators who performed the RACE 7 ACWAS pragmatic RCT comparing early cardioversion for recent-onset AF to a wait-and-see or delayed-cardioversion strategy.

The trial's results teach us a lot about AF. The presence of sinus rhythm at 4 weeks, the primary endpoint, occurred in 91% of patients in the delayed group and 94% of the early-cardioversion group, which easily met noninferiority. The key finding: 69% of patients in the delayed-cardioversion group converted spontaneously within 48 hours.

This means that most patients we cardiovert in the emergency department for acute-onset AF could be spared an anesthetic and high-voltage shock if clinicians would simply give peace a chance and wait 48 hours. Patients with AF worldwide should send Nikki Pluymaekers, MD, and coauthors a token of appreciation.

Three studies this year confirmed my view that the best thing a cardiologist can do for a patient with advanced kidney disease is to resist the urge to do procedures.

The ISCHEMIA-CKD trial ran concurrently with the larger ISCHEMIA trial, and, despite great efforts to reduce contrast-induced kidney injury, the invasive approach was not associated with a reduction in CV death or MI. The safety endpoint of death or dialysis was 48% higher in the invasive arm, and stroke rates were nearly fourfold higher.

In patients with end-stage renal disease who underwent TAVR, an observational analysis from the STS/ACC/TVT registry reported a significantly higher 1-year mortality in patients undergoing dialysis (36.8% vs 18.7%; P < .01).

In the ICD2 trial, Dutch investigators set out to study the value and safety of an implantable cardioverter-defibrillator (ICD) in 200 patients on dialysis. The authors rightly excluded patients with a left ventricular ejection fraction below 35% because patients with heart failure on dialysis do not live long enough to benefit from primary prevention ICDs. The trial was stopped for futility after inclusion of 188 patients. Sudden cardiac death rate (the primary endpoint) did not differ significantly, and more than 30% of patients in the ICD arm had an adverse event related to the device.

The EXCEL trial, which reported 3-year results in 2016, and 5-year results this year, compared percutaneous coronary intervention (PCI) to coronary artery bypass grafting in patients with left main disease amendable to PCI. The EXCEL authors chose a composite endpoint of MI, death, and stroke. In the two reports, the top-line conclusions were that the procedures produced similar results. The 2016 findings led to PCI receiving a class I recommendation in the European guidelines for patients with low Syntax score (0 to 22) left main lesions.

But it's not that simple. Two problems with EXCEL interpretation are the following:

The definition of MI and inclusion of periprocedural MI drove the noninferiority of the composite endpoint of MI, stroke, and death.

Death rates were higher in the PCI arm, and the mortality rates were increasing over time.

There is more. This month, the BBC Newsnight program received the data set from the 2016 EXCEL study. Their televised program made three main points:

EXCEL authors did not report a secondary endpoint (MI as determined with the universal definition) that was listed in the trial protocol. NEJM editors and peer reviewers did not notice the omission.

The BBC's analysis of the unpublished secondary endpoint revealed an 80% higher rate of MI in the PCI arm.

An exploratory analysis of the primary endpoint with this definition of MI found a 44% higher rate of events in the PCI arm.

Based on this report, the European Association for Cardio-Thoracic Surgery removed their support for the left-main recommendations in the European guidelines on myocardial revascularization.

The EXCEL investigators responded in a 3500-word rebuttal, in which they deny wrongdoing and defend the results of the two EXCEL manuscripts.

The BBC Newsnight producers say they have more information. EXCEL authors say they will publish more papers from the data. In my December 13 podcast, I cite Sanjay Kaul, MD, from Cedars-Sinai, who felt that the limitations of EXCEL, combined with the failure of PCI to meet noninferiority vs CABG in the NOBLE trial, should have prevented upgrading of left-main PCI in the first place.

This controversy will extend into 2020. It's important because it elevates the debate on open science, raises serious questions about peer review, and, if not resolved quickly, decreases public trust in medical evidence. Perhaps this sounds naive, but to me the answer seems simple: More transparency and openness to sharing data.

Thank you for your support in 2019. I look forward to another year of reporting and commenting on medical science in 2020.

Read this article:

Mandrola's Top 10 Cardiology Stories of 2019 - Medscape