Sep 20 2024 This Week in Cardiology

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, download the Medscape app or subscribe on Apple Podcasts, Spotify, or your preferred podcast provider. This podcast is intended for healthcare professionals only.

In This Week’s Podcast

For the week ending Sept 20, 2024, John Mandrola, MD, comments on the following news and features stories: Listener feedback, potassium (K+) levels after cardiac surgery, safety of cardiac devices, disparities in preventive care, stopping trials early, and an update on SHAM-PVI.

Listener Feedback

I received a nice comment on my discussion of MATTERHORN, which was transcatheter edge-to-edge repair (TEER) vs mitral valve (MV) surgery in patients with functional mitral regurgitation (MR).

The primary endpoint composite of death, hosptalizations for heart failure (HHF), MV re-intervention, assist device implant, or stroke was better at 1 year in the TEER arm —16.7% vs 22.5%.

The authors also reported safety events, which were higher in the surgery arm, due to post-op bleeding and atrial fibrillation (AF). I said reporting such was sort of ridiculous, because bleeding and AF are expected after surgery.

Stanton Rowe did not agree with my comment.

Are you suggesting that we should ignore Afib and Bleeding Complications from surgery because they are known and expected? These outcome complications matter to patients! They belong in the study as important endpoints.

This is a good question. My take of MATTERHORN is that counting outcomes at only 1 year biases the results against surgery. The Kaplan Meier curves were coming together over time.

Dr Rowe’s question highlights the difficulty in comparing totally different procedures. In MATTERHORN, these were 70 year-old patients with pretty good ejection fractions (EFs). Their outlook is longer than 12 months.

Inherent in cardiac surgery is post-op AF and a higher rate of bleeding. Okay, you could list these rates, but they come with the procedure. The question patients with this condition have is, how am I at 2, 5, and 10 years. If the outcomes are the same at 5 years, then of course, TEER is preferred. But if TEER is worse, say it leaves more MR, then patients have a decision to make.

I remind you that the original coronary artery bypass graft (CABG) vs medical therapy trials were quite close. In the Yusef meta-analysis, the life expansion was about 6 months. Patients take about 3 months to recover from surgery. And, as I mentioned, in STITCH, LVD, CAD, it took 10 years for surgery to best medicines. In these cases, I am not sure why a patient would choose surgery.

What we need to know in surgery vs device trials is what happens after a year.

Post-Cardiac Surgery AF

Speaking of AF after cardiac surgery, the European Society of Cardiology (ESC) meeting featured yet another randomized controlled trial (RCT) testing a strategy to prevent AF after the heart is operated on.

The TIGHT K RCT tested two strategies of K+ replacement: a relaxed approach to K+ replacement vs tight control.

Before I tell you the results, we should think back to trials testing aggressive vs lenient control of anything in Medicine. I can’t think of many in which more is clearly better. Maybe SPRINT trial of lower blood pressure targets, but that trial had nuances.

In TIGHT-K, led by first author, Benjamin O’Brien, whose author affiliations list Germany, London, and Cleveland, about 1700 post-CABG patients (mean age 65 years) in 23 centers (United Kingdom and Germany) were randomly assigned to a strategy of replacing K+ to keep levels above 4.5 mEq/L vs a relaxed strategy of 3.6 mEq/L. 

Since relaxed K+ replacement is easier, the authors tested this as a non-inferiority trial. That is reasonable because when using non-inferiority of a new therapy, the new therapy has to offer something less invasive, less costly, or less burdensome.

• The primary endpoint of clinically detected new onset AF in the first 5 days after CABG occurred in 26.2% of the tight K+ group and 27.8% in the relaxed group.

• This tiny difference easily reached noninferiority for the relaxed arm. In fact, the 95% confidence interval (CI) went from -2.6% better for tight group vs 5.9% worse.

• Per patient cost was significantly lower in the relaxed arm.

The authors concluded that:

For AF prevention, supplementation only when serum potassium concentration fell below 3.6 mEq/L was noninferior to the current widespread practice of supplementing potassium to maintain a serum potassium concentration greater than or equal to 4.5 mEq/L. The lower threshold of supplementation was not associated with any increase in dysrhythmias or adverse clinical outcomes.

Comments. First, I did not know, as the authors write in the introduction, that “many clinicians believe that serum potassium concentration influences risk of developing atrial fibrillation in critical illness, and frequent potassium supplementation in an effort to maintain a high-normal postoperative serum potassium concentration (≥4.5 mEq/L) is now routine practice in many centers worldwide for AF after cardiac surgery prophylaxis.”

So, good on the investigators for subjecting this plausible mechanistic dogma to empirical study with an RCT.

This work won’t win a Nobel prize, but it is important to do these studies, to subject our ideas to proper trials.

I believe this work is clinically actionable. They found no other disadvantages to the relaxed strategy, it cost less and was less burdensome for staff.

Finally, I realize it’s only one study, but this is yet another example wherein policies that make sense are implemented. Sure, it’s reasonable to conclude that K+, being an electrically active ion, one that is clearly important in the genesis of arrhythmia, might be an important thing to maximize. But. While low K+ and high K+ are clearly bad for the heart, the TIGHT K trial shows that the difference between low normal and high normal is not important. That is good to know.

Evidence Base for New Cardiac Devices

New things in cardiology scare me, a little. I lived through some terrible disasters with implantable cardioverter-defibrillator (ICD) leads. The worst was the Medtronic Sprint Fidelis lead. It was bright blue and thinner than the bulky ICD leads we were using.

The problem was that they failed at higher rates than the bulky tried and true leads. I know; that sounds like a benign sentence. But failure meant sensing noise and the device thinking it was ventricular fibrillation. When that happens patients get wide-awake shocks. Pacer dependent patients don’t pace at all and have syncope.

Back in my younger days, I was a rapid and early adopter, and we had tons of these bad leads. I don’t mean to pick on only Medtronic. All the device companies have had recalls.

The Annals of Internal Medicine has published a survey observational study that set out to characterize US Food and Drug Administration (FDA) Class 1 recalls — the most serious recalls — of cardiovascular devices and to describe the evidence supporting device approval. The first author is Claudia See; senior author, Kushal Kadakia.

It was not good news for early adopters.

• The authors write: “The agency grades recalls according to their severity, with Class I designations reserved for situations where the FDA deems that use of a device would place a patient at increased risk for serious adverse health consequences or death. Although only 1% of FDA-authorized devices are subject to Class I recalls, these recalls affect millions of patients and take years to resolve.”

• Second background point: devices are approved via two pathways, a 510(k) pathway wherein the maker only needs to show that the device is similar to approved devices. This is the softer way to get approval.

• The other way to get a device approved is with the premarket approval pathway wherein there is usually evidence of safety and efficacy.

• The FDA can require post-approval studies at the time of a pre-market approval. This is akin to ordering surveillance after approval.

The authors studied Class 1 recalls from 2013 to 2022. They noted 137 recalls involving 157 devices. Nearly three-fourths were moderate risk devices that were approved through the softer 510(k) pathway.

About one-third were high risk pre-market approval devices.  The most common reason for recall was a device design issue.

About 1 in 4 devices had multiple Class 1 recalls.

Here is the important finding: most studies used surrogate (27 [79.4%]) and composite (24 [70.6%]) measures as primary end points. Two-thirds of studies were non-randomized and open label.

• Twenty-two (48.9%) premarket approval devices had required post- approval studies, with 14 reporting delays.

• No 510(k) devices were subject to postmarket surveillance.

• I just told you that the majority of Class 1 recalls were devices approved through the 510k pathway and now there is this: no 510(k) devices were subject to post-market surveillance.

A summary sentence that captures the essence of the paper is that:

Medical devices later recalled due to safety issues often had little clinical evidence supporting their original authorization.

The authors offered an example of how dubious pre-market evidence leads to trouble:

Boston Scientific’s VICI VENOUS STENT System was approved in 2019 based on a pivotal study using the surrogate measure of vessel patency 12 months after implantation of an iliofemoral stent. In 2021, nearly 35,000 units were recalled after reports of stent migration, with clinicians requested to cease device implantation.

Comments. This paper has a policy implications that are beyond the scope of clinical medicine. I know little about policy, but the fact that FDA did not use its authority to require post-market surveillance for many devices is worrisome.

In fact, the authors note that 80% of surveyed doctors report only a limited understanding of FDA’s processes.

I am not sure it’s feasible for us to become facile in how the FDA works, but this study, and my experiences with horrific ICD lead issues over the past decades drives my medically conservative approach to new stuff.

When new devices come to market, there is often great enthusiasm…like a tailwind on a bike ride. But we should have questions in our minds: what was the pre-market evidence? And what is the post-market surveillance like?

This paper clearly shows that caution with new devices is wise.

Disparities in Care

JAMA Network Open has published a large observational study of 1.5 million patients seeking preventive care. The primary question and outcome was the frequency of insurer denials for preventive services. Subgroup analyses were done to assess outcomes across patient household income, education, race and ethnicity.

This was a Toronto-based team, first author, Alex Hoagland, PhD. For each enrollee, they identified the use of seven preventive services recommended by the US Preventive Services Task Force that would have been subject to the Affordable Care Act (ACA) preventive care provision. 

These were contraceptive administration, breast cancer screening, cholesterol screening, colorectal cancer screening, depression screening, diabetes screening, and wellness visits.

They sought to examine the association with patients’ social determinants of health and insurance claims denial for preventive services.

Major findings:

• First, patients with lower incomes experienced denials more frequently than patients with higher incomes, as did those with less education or those from minoritized racial and ethnic groups.

• The most common denials were for diabetes screening, depression screening, and cholesterol screening.

• Second, differences across income and race and ethnicity remained significant even after adjusting for patient geography, service type, and insurer.

• Third, services for at-risk patient populations, including contraceptive administration or mental health screenings, were associated with higher denial rates than other preventive services.

Comments. We all know that there are disparities in health outcomes. Rich people have better outcomes than poor people. I don’t know if that can ever be eliminated, but it seems right to try to reduce these gaps.

I am not an epidemiologist, and this paper had some serious statistical maneuvers. There were a lot of adjustments made. In other words, the research team had a lot of choices of covariates to adjust for. But I have little doubt that there are unequal insurance denials based on income, race, and education level.

But the reason I highlight this study is less about the associations and more about the concepts of preventive care. The ACA allows for free preventive care. That assumes preventive care improves health. Does it?

I know of no RCTs for depression screening or diabetes screening. Wellness visits have been studied in RCTs and there was no benefit in outcomes. Colorectal cancer screening may result in a tiny improvement in overall mortality with flex sigmoidoscopy, but the big colonoscopy trial was negative for overall mortality. And mammography may reduce breast cancer deaths, but pooled data show no difference in overall mortality.

• My point is that the way to fewer health disparities is unlikely to come through better access to preventive screening.

• In fact, I have long believed that healthcare has little role in improving disparate health outcomes.

• Rather, this seems more like a job for society and policy decisions.

If we somehow could lessen the gaps between rich and poor, a certain side effect of that would be less health outcome disparity.

Of course, how best to reduce gaps between rich and poor is many levels higher than this podcast could ever sort out.

Stopping Trials Early

The Journal of the American College of Cardiology has published a very nice piece from Drs. Sanjay Kaul and Javeed Butler on stopping trials early, insights from the recent pivotal trials in chronic kidney disease (CKD).

I learned a lot reading it.

They focused on three trials looking at SGLT2 inhibitors (CREDENCE, DAPA-CKD and EMPA-KIDNEY) and one with semaglutide (FLOW).

All of these studies were stopped early. Published results were based on accrual of planned events in CREDENCE (69%), DAPA-CKD (75%), EMPA-KIDNEY (93%), and FLOW (87%). I did not know that.

• The first point they made is that stopping trials makes it harder to sort out important signals, such as the higher amputation risk in CANVAS, which was not confirmed in CREDENCE but not ruled out either because of the low number of events due to early stopping.

• Second, DAPA-CKD was stopped after only 60% of planned events. So, when the results came out, there was a 39% risk reduction in primary outcome driven mostly by kidney components.

• But there was also a statistically significant 31% reduction in all-cause deaths, driven primarily by a 45% reduction in non-cardiovascular (CV) deaths (36 vs 66 events), largely related to infection and malignancy. CV deaths were not reduced. Is non-CV mortality reduction related to infection and malignancy plausible or a random-high of early stopping? My answer is it is likely a random high due to early stopping.

• They then list the four major reasons for stopping trials early:

  1. unequivocal benefit;

  2. unacceptable harm;

  3. futility; and

  4. administrative reasons (enrollment or funding concerns).

The tension comes in getting the information we need for estimating true effect size vs protecting patients. They site an important systematic review from Bassler and colleagues that finds that most trials that stopped early, especially those with fewer than 500 events, fail to provide reliable and valid estimates of treatment effect, often overestimating it by nearly 30%.

Finally, they write:

Other unintended consequences of stopping trials early include inconclusive data on outcomes that are important to patients such as survival, quality of life, or adverse event, compromised ability in the future to conduct more definitive studies, inappropriate comparison of benefit (or harm) of drugs within a class across completed vs truncated trials, unwarranted publicity and market share, or potentially undue skepticism regarding reimbursement or prescription decisions.

Comments. I highlight this important piece for its details regarding the CKD trials, but more for the important consideration of trials that stop early.

When interpreting such studies we need to consider that having fewer events increases uncertainty. Sure, this may be reflected in wider CIs, but still, we should approach such trials with caution. Likely this is the reason we see overall mortality benefit in DAPA_CKD but no CV death reduction.

There will always be a debate about early stopping of trials, but having thought about this, I come down on the side of avoiding it and trying to maximize the evidence.

SHAM PVI

I discussed this important trial on the September 6 podcast. I wrote about it this week. My column is up on the site, and right after recording this podcast, I will speak with the investigators in an interview. Look for that coming soon.