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 February 9, 2024, John Mandrola, MD, comments on the following news and features stories: Beta-blocker withdrawal in patients with heart failure with preserved ejection fraction (HFpEF), ARCADIA (apixaban vs aspirin [ASA]), food as medicine, and bariatric surgery for hypertension (HTN).
HFpEF and Heart Rate
JAMA Cardiology has published a post-hoc analysis of the PRESERVE HF trial of beta blocker withdrawal in patients with HFpEF. As a pacemaker doctor I am intrigued by the heart rate and HFpEF story. Teaser – higher heart rate may be better.
Before I tell you about the substudy, which explored the category of patients most likely to benefit from higher heart rates (from beta blocker withdrawal), let’s review the PRESERVE HF trial.
But first, a word on chronotropic incompetence, a nebulous term but defined as the inability to increase heart rate with increased demand, such as in exercise. It’s common in HFwith reduced EF (HFrEF) and HFpEF. The mechanism of chronotropic incompetence in HFpEF is not well known, but surely it is multifactorial as HFpEF is often a condition of aging, as is chronotropic incompetence. One thing is certain: beta-blockers worsen chronotropic incompetence.
Heart rate and cardiac output in HFpEF is complicated. Obviously cardiac output = stroke volume x heart rate so some increase in heart rate is good. But too much is bad because it impairs left ventricular (LV) filling, and too slow allows for more diastolic filling which may increase LV end diastolic pressure.
So, you need a Goldilocks — just right — amount of heart rate response to exercise.
Another issue with beta blockers. This class of drugs has been shown to be beneficial in stable outpatients with HFrEF, but we don’t have any such data in HFpEF.
PRESERVE HF was published originally in the Journal of the American College of Cardiology (JACC) in November 2021, first author, Dr. Patricia Palau.
This was a cross-over study in 52 patients with HFpEF and poor heart rate response to exercise. The crossover involved cardiopulmonary exercise testing on and off beta-blockers. The two co-primary endpoints involved VO2 max.
These were 72-year-old patients, about 60% female, who mostly had New York Heart Association Class 2 symptoms. All patients were on a stable dose of beta blockers.
Both co-primary endpoints, peak VO2 and peak VO2% increased significantly in the beta blocker withdrawal group. For peak VO2 it was 14 vs 12 mL/kg/min.
Exercise heart rate was about 30 bpm higher in the beta blocker withdrawal group. Measures of quality of life were not conclusive. BNP levels did not change. No subgroups pointed to heterogeneous treatment effect.
The authors concluded, “Overall, our results highlight the role of chronotropic incompetence as a crucial pathophysiological mechanism in HFpEF and consequently open new therapeutic avenues in these patients.”
I would grade the strength of evidence from the main PRESERVE trial as modest.
The authors have now published a post-hoc study in JAMA-Cardiology in which they explore any associations between beta blocker withdrawal and changes in VO2 depending on echocardiography parameters. Specifically, the LV end diastolic volume (LVEDV), LV end systolic volume (LVESV), and LVEF.
To be clear, this was an observational correlation study. Even though it’s from a randomized controlled trial (RCT), it’s observational.
After stopping beta blocker treatment, the median increase in peak heart rate was 30 bpm (95% confidence interval [CI], 25-35; P < .001).
Beta blocker cessation was differentially associated with change of percentage of peak VO2 across the continuum of LVESV (P for interaction = .02), indicating a greater benefit in those with lower LVESV
The simple summary of this observational study helps to picture the echocardiograms of these patients. When you look at the apical four chamber of HFpEF patients: you see tiny LV cavities and hyperdynamic LVEF. These were the patients who seemed to do best with beta blocker withdrawal. The main figure plots LV systolic volume on the x – axis and change in VO2 on the y axis.
The smaller the LV, the greater the bump in VO2 max from beta blocker withdrawal.
My Comments. These observations are super-interesting. I don’t know about your zipcode, but in mine, the vast majority of these patients are on beta blockers. You know, for diastolic dysfunction because stiffer hearts need longer to fill.
But I am not so sure; maybe longer time to fill means moving higher up on the pressure volume loop. Plus, the mean exercise peak heart rate in the baseline characteristics was a measly 97 bpm. How do you tolerate activity with that?
And if the theory of longer filling times = not good for HFpEF, then you’d expect the smallest LVs and most hyperdynamic LV to get the greatest benefit. Indeed, that is exactly what this observational analysis found.
Of course, it’s observational. It’s a small unblinded cohort. But let me also put in your mind that beta blockers are generic. The notion of not using beta blockers, or stopping beta blockers, or allowing for higher heart rates is one free of financial conflicts. The no-generic beta-blocker proponents in HFpEF have nothing to gain. They won’t be getting invites to industry symposia.
The ideal heart rate for these patients will surely never be algorithmic. HFpEF is defined by its variability. But the knee-jerk, dogmatic, historical notions of inducing lower heart rates may be yet another medical reversal. And boy do I love medical reversals.
ARCADIA Trial
In 2016, neurologist, Hooman Kamel, and three co-authors, wrote an incredible article describing atrial fibrillation (AF) and stroke. The gist of it was to consider a new model of AF and stroke. Their message was that AF episodes on an ECG may simply be a marker of atrial disease, and it was this atrial disease that was the important target for stroke prevention.
To quote: “An up-to-date model must emphasize systemic and atrial substrate as well as rhythm. Aging and systemic vascular risk factors cause an abnormal atrial tissue substrate, or atrial cardiopathy, that can result in AF and/or thromboembolism.”
This piece moved me. It got me thinking. It resolved many of the inconsistencies of stroke and AF. For instance, it explained why there was such poor correlation with timing of AF episodes and stroke.
What I love even more about this theory is what Kamel did next. He didn’t just lecture about it on the speaker circuit. He did a trial to test the brilliant idea.
This week, JAMA published the results of the ARCADIA RCT, a trial testing apixaban vs antiplatelet therapy (ASA) for the prevention of recurrent stroke in patients who had embolic stroke and evidence of atrial disease.
Before I tell you about the trial, we should all pause and think that this is exactly how academic medicine should work. First have great ideas, then test them in RCTs.
ARCADIA randomly assigned about 1000 patients who had what is called cryptogenic stroke and evidence of atrial cardiopathy.
They defined atrial cardiopathy using three criteria:
Abnormal terminal forces of the p-wave in lead V1;
Abnormal BNP > 250 pg/mL;
Left atrial (LA) diameter index of 3 cm/m2 or greater on echocardiogram.
These markers were chosen because of higher stroke risk seen in observational studies. Patients could not have evidence of AF.
Cryptogenic stroke was defined rigorously. Their criteria for cryptogenic stroke required CT or MRI to exclude lacunar infarcts, vascular imaging of the cervical and intracranial arteries to exclude large-artery atherosclerosis causing 50% or more stenosis of a relevant arterial lumen, and transthoracic or transesophageal echocardiography, a 12-lead ECG, and 24 hours or more of continuous heart rhythm monitoring, all to exclude major-risk cardioembolic sources.
Results: The primary efficacy endpoint was recurrent stroke of any type. The two safety endpoints were symptomatic intracranial hemorrhage (ICH) and all-cause death.
The authors estimated a 3.5% annual stroke risk after a cryptogenic stroke and a doubling of that risk to 7% in those patients with atrial cardiopathy.
They chose a hazard ratio (HR) of 0.6 (or 40% reduction) with apixaban as a minimally clinical important difference. That seems like a lot to me, but they cited a sub-study of AVERROES (apixaban vs ASA in warfarin ineligible patients) in which apixaban reduced the risk of stroke over ASA by 70% in patients with prior stroke or transient ischemic attack.
The trial was terminated early for futility. At the planned first interim look after 75 primary outcome events, the HR was 1.00. Each group (apixaban vs asa) had recurrent events in 40 patients. The P-value was 0.99. CI were wide though – from 0.64 to 1.55.
AF was diagnosed in 149 patients (14.7%) at a median 30 weeks after randomization.
In a prespecified sensitivity analysis, apixaban did not significantly reduce the rate of recurrent stroke when censoring follow-up at the time of AF diagnosis (HR, 1.05 [95% CI, 0.66 to 1.65]).
No significant evidence of heterogeneity of treatment effect on the primary efficacy outcome was found across prespecified subgroups; apixaban did not appear to be of greater benefit than ASA at higher levels of atrial cardiopathy biomarkers.
There was one very interesting safety finding: ICH occurred in zero patients in the apixaban group but in seven patients in the ASA group (annualized rate 1.1%) The authors don’t give us HR or P-values for that, for some reason.
One of listed limitations was that most patients met criteria for atrial cardiopathy by ECG or BNP and not large LAs. In fact, most patients in the study did not have large LAs.
My Comments. I reiterate, this is how medical science should work. Great ideas and great theories need to be tested in RCTs.
Even with an enriched population of patients with elevated risk of cardio-embolic causes of recurrent stroke (those with diseased LA), direct oral anticoagulants (DOACs) did not make a difference over ASA. These results were also consistent with the two negative ESUS trials of rivaroxaban and dabigatran (NAVIGATE ESUS and RESPECT ESUS).
The reason for non-significance in ARCADIA is far less important than that the trial was done, but I think it’s educational to think about potential reasons why DOAC’s failed.
One is that atrial disease was not significant enough.
There’s also this idea of atrial cardiopathy making the rounds in electrophysiology, but my take on it is that’s super uncommon. Yes, I have seen LA’s with super low-voltage and veins that look isolated but are not, mostly in women, but it’s not common.
So maybe the criteria used in ARCADIA were not strict enough. The patient population was not enriched enough to gain benefit from anticoagulation. The thing about that theory is that stricter criteria would have made the trial less useful or translatable. ECG, BNP, and echocardiogram criteria are nice because every patient with stroke gets these basic tests.
Another reason that the DOAC may have failed is competing risks of stroke. Many of these patients had vascular disease. Competing risks of the primary outcome are super-important to think about. My friends who believe in percutaneous left atrial appendage closure seem to forget that stroke is a complex problem with many upstream causes. Closure of the appendage, even when done perfectly, stops only one potential cause of stroke.
Finally, we need to think about the zero vs seven ICH events in the apixaban vs ASA groups, respectively. Sure, this could be chance, and I have emailed professor Kamel as to why there are no CI or P-values for this primary safety outcome.
But what if it is not a chance finding? There were 500 patients in each arm. All patients had enough disease to have had a cryptogenic stroke. And zero ICHs in nearly 2 years of follow-up in the apixaban arm vs seven in the ASA arm.
My friends, there is this notion that OAC is more dangerous than ASA — especially because of the risk of ICH. But this belief approaches religion. The data, from ARCADIA, AVERROES, and BAFTA, do not support that idea.
I want to congratulate the authors of ARCADIA. It is an important trial. And to my younger listeners, the global lesson is that most things we do in medicine don’t work. We should be humble about our commonsense ideas.
Food-as-Medicine
Speaking of great ideas not passing muster in RCTs, we now need to talk about the feel-good notion of using food as medicine in patients with diabetes and food insecurity.
Here is the thinking: the American diet is terrible. Patients with food insecurity are especially vulnerable to high calorie processed food. But even more vulnerable are patients with type 2 diabetes and HbA1c levels > 8%.
JAMA-Internal Medicine published the RCT of this great idea. So great an idea that both NPR and The New Yorker covered the food as medicine idea in the years before this trial.
The NPR story was especially instructive for medical learners. In celebratory prose, Allison Aubry told the story of one man’s success with the program — his HbA1C dropped from 11% to under 7%. "[They] won't go blind; [they] won't have kidney disease, amputations…. The list goes on and on," went the article. Nary a cautious tone about testing things in RCTs before declaring success.
This is just like medical practice — say, when you see a patient who was taking a therapy and did great, and these anecdotes trick your brain into thinking your intervention is beneficial overall. You know, like the doctors in the 1980s who were seeing successful suppression of premature ventricular contractions with encainide after myocardial infarction. They thought it was great to suppress those bad arrhythmias.
The Food-as-Medicine trial took place in one rural and one urban site.
The program provided healthy groceries for 10 meals per week for an entire household, plus dietitian consultations, nurse evaluations, health coaching, and diabetes education.
The control arm were people put on a wait list who were given standard advice on diet.
The primary outcome was HbA1c levels at 6 months. Secondary outcomes were biometric measures, health care use, and self-reported diet and healthy behaviors.
Trialists screened more than 3700 patients, 1000 were deemed eligible and 500 were randomized. So, right away, this was a highly selected population. Only 349 patients had lab values at 6 months.
Get this:
Both the treatment (n = 170) and control (n = 179) groups experienced a substantial decline in HbA1c levels at 6 months, resulting in a nonsignificant, between-group adjusted mean difference in HbA1c levels of −0.10 (95% CI, −0.46 to 0.25; P = .57). In other words, no difference.
Access to the program did increase healthcare use. Dietician visits were 2.7 vs 0.6 visits in the control arm. More patients in the active arm had prescriptions for metformin and GLP-1 agonists and more patients in the active arm self-reported healtier diets.
The authors made three important conclusions:
One was that this program increased usage of healthcare but made no difference in HBA1c levels.
Their second conclusion is the main reason I mentioned this study: Programs targeted to individuals with elevated biomarkers require a proper control arm to demonstrate effectiveness to account for improvements that occur without intervention.
I think we should make posters of that second conclusion that is something like this: YOU ALWAYS NEED A CONTROL ARM to show something works or not. And the control arm has to be a proper control, not say a TRILUMINATE-trial type control arm.
Their third conclusion was that further research is needed to design food as medicine programs that improve health.
My Comments. I am going to write a piece about this study. It’s not directly related to cardiology, but it may be one of the most important trials of this year.
I mean, come on. Ten healthy meals per week, nurse access, dietician access, and no signal of benefit over a proper control arm.
Speaking specifically, I am no public health expert, but I do spend a lot of time counseling individual patients on healthy living. It’s hard for Americans. This isn’t Denmark. You can’t exactly walk or ride to work in most places here. It’s not safe. What’s more, Americans have gotten used to processed food. It’s hard to break the habit. Heck, it starts in schools. What they feed my grandkids is shocking.
So, I am not surprised that this well-meaning lovely idea did not work in real life on average. I don’t have the answer for better cardiometabolic health for the American public, but this data shows it won’t be simple, NPR-friendly programs.
Yet the most important message of this trial is similar to ARCADIA. It is the power of the RCT to keep us humble. Great ideas must be tested. No matter how great you think an intervention is, it has to pass muster in an RCT.
I also congratulate Dr. James Doyle and his colleagues for not resting on the praise from the New Yorker and NPR.
Weight Loss Surgery: The GATEWAY Trial
JACC has published the results of an RCT from researchers in Sao Paulo, Brazil, who compared bariatric surgery vs medications as therapy for HTN.
100 patients with HTN and an average body mass index (BMI) of 37 were randomly assigned to have surgery plus medications or medical therapy alone.
These were young patients (age 44 years), mostly women (72%).
The surgery was a Roux-en-Y gastric bypass.
The primary outcome was at least a 30% reduction in total number of antihypertensive medications while maintaining blood pressure < 140/90 mmHg.
The results were as you’d expect:
Patients in the surgery arm had a substantial reduction in BMI compared with those who did not have surgery (28 vs 36).
And, compared with medical therapy, surgery promoted a significantly higher rate of reduction in medications (80% vs 14%; relative risk: 5.91; 95% CI: 2.58-13.52; P < 0.001) and the mean number of antihypertensive medications was 3 vs 0.8.
The authors concluded that bariatric surgery represents an effective and durable strategy to control HTN and related polypharmacy in subjects with obesity.
My Comments. I highlight this trial because I laud people who do trials. I love trials. But this is a good example of how trials do not always provide an answer.
Trials have to be designed to provide an answer. Here we have one group that gets a very invasive intervention and the other group gets tablets.
Think back to the renal denervation (RDN) HTN trials. Before we did proper placebo controls (I hesitantly use the term sham procedures) we though RDN was amazing. Well, placebo-controlled trials have now shown the effects to be quite modest.
The only way to prove what the Brazilian authors sought to prove here would have been a sham-control arm. I am surprised JACC editors let this pass.
The other issue — and this is no fault of the authors because at the time the trial began GLP-1 agonists had yet to be discovered — is that all future bariatric surgery trials must be a) sham-controlled and b) compared to GLP-1 agonist.
GLP-1 agonist drugs clearly induce significant weight loss. Their role is yet to be defined. But clearly the control arm of all future surgery trials must be with this class of drugs.
It’s just like the DOAC trials. The DOAC trials did not pit a DOAC vs placebo or ASA because we had learned that warfarin was superior to placebo or ASA.
© 2024 WebMD, LLC
Any views expressed above are the author's own and do not necessarily reflect the views of WebMD or Medscape.
Cite this: Feb 09, 2024 This Week in Cardiology Podcast - Medscape - Feb 09, 2024.
Comments