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 December 2, 2022, John Mandrola, MD comments on the following news and features stories.
Right before the American Heart Association (AHA) meeting last month, the New England Journal of Medicine (NEJM) published results of the EMPA-KIDNEY trial. As background, there are two important trials to mention with sodium-glucose cotransporter 2 (SGLT2) inhibitor use in patients with chronic kidney disease (CKD). (Recall, cardiology listeners, a lot of our patients with atrial fibrillation (AF), valvular disorders, and coronary artery disease (CAD) have CKD).
In CREDENCE (canagliflozin vs placebo), published in 2019, diabetic patients with CKD, a median estimated glomerular filtration rate (eGFR) of 56 mL/min/ 1.73m2 body surface area, and a median urine albumin creatinine ratio (uACR) of 923 mg/g, had a statistically significant 30% relative risk reduction (RR) in the primary outcome of composite of end stage kidney disease (ESKD), a doubling of the serum creatinine level, or death from renal or cardiovascular (CV) disease.
DAPA-CKD (dapagliflozin vs placebo), published in 2020, found that dapagliflozin (DAPA) in patients with median eGFR of 43 mL/min/ 1.73m2 and a median uACR of 965 mg/g reduced the primary outcome of sustained decline in eGFR of 50%, ESKD, or death from renal or CV causes by 39%. This was a 5.2% absolute RR and it was highly significant. DAPA also reduced a composite of CV Death/heart failure (HF) hospitalizations by a statistically significant 29%. In DAPA-CKD, a third of patients did not have diabetes.
The question was, class effect vs drug effect? Given all the cumulative data on SGLT2 inhibitors, it was likely that empagliflozin (EMPA) in CKD patients would also provide benefit. Indeed, that is what EMPA-KIDNEY found.
The 6600+ patients had a mean age of 63 years, and CKD with a median eGFR of 37 mL/min/ 1.73m2, and uACR ratio of 331 mg/g.
The primary endpoint was progression of CKD, defined specifically as ESKD, or a sustained decrease of eGFR of < 10 mL/min/ 1.73m2, a sustained decrease in eGFR of ≥ 40% from baseline or death from renal causes, or CV death
The trial was terminated early for efficacy. This may be important. Median follow-up was 2 years.
13.1% reached a primary endpoint in the EMPA group vs 16.9% in the placebo arm. The 3.8% absolute RR resulted in a 28% relative RR (hazard ratio [HR] 0.72 and confidence interval [CI] ranging from 0.64-0.82)
The benefit was driven mostly by reduction in progression of CKD; CV death was not significantly reduced.
Notably, and I really like this, total hospitalizations were reduced by nearly 5% in absolute terms, which is a lot.
There was no heterogeneity of treatment benefits based on subsets of CKD (as measured by GFR) but there was with uACR; here the benefit clustered in patients with a uACR > 300 mg/g.
There was also more benefit in patients with diabetes vs no diabetes.
Comments. EMPA-KIDNEY is what we look for, right? Statistically robust and clinically meaningful effect sizes. The trial was large and had hardly any dropouts.
Taken together with the previous data from CREDENCE and DAPA-CKD, it looks clear that SGLT2 inhibitors have a protective effect on the kidney, which may be greater in patients with more albuminuria. Lancet recently published a massive meta-analysis of the impact of diabetes on the effects of SGLT2 inhibitors on kidney outcomes. The main efficacy outcome was kidney disease progression.
Nine trials were included, with 74,000 patients
Compared with placebo, allocation to an SGLT2 inhibitor reduced the risk of kidney disease progression by 37% (RR 0.63, 95% CI 0.58–0.69) with similar RRs in patients with and without diabetes
In the four CKD trials, RRs were similar irrespective of primary kidney diagnosis. SGLT2 inhibitors reduced the risk of acute kidney injury by 23% (RR 0.77, CI 0.70–0.84) and the risk of cardiovascular death or hospitalization for heart failure by 23% (RR 0.77, CI 0.74–0.81), again with similar effects in those with and without diabetes.
I think this meta-analysis seals the deal on diabetes. It doesn’t matter. The drugs reduce the risk of renal outcomes in patients with CKD regardless of diabetes status.
JAMA-Internal Medicine published an observational study that found an association between organ donation and motorcycle rallies. Guess what the results were: organ donation increased during motorcycle rallies.
I don’t understand this study. Shall we study boating accidents around July 4th, heat illness or drownings during the summer months, or the incidence of depression during the lockdowns? Why would researchers spend time on something so obvious? Why would a journal not only publish it, but also attach an editorial?
I will tell you why: 58 news outlets picked up the story. It’s already in the top 5% of research by Altmetric.
I mean no ill will to the journal or the authors, but what these sorts of papers do is highlight the business model of scientific publishing, which is attention, page views, etc. It’s the same for medical journalism. I wish it were not so. But I place less blame on the journalism industry, because they are tasked with reporting what’s in the journals.
This story makes the podcast because understanding what gets in journals is an important meta-theme in critical appraisal. Whenever you read a paper like this, you should think about what paper did not get in because of this.
Second- and Third-Order Effects of Treatments
Journalist Marlene Busko, who has an eye for important papers off the main stage at meetings, reported on a super-important abstract presented at AHA. Dr. Rebecca Woodruff, an epidemiologist at the Centers for Disease Control and Prevention (CDC), presented a paper that found a spike in deaths due to heart disease during 2020, the first year of the pandemic. This was notable for two reasons:
Previously, there had been a steady year-over-year decline since 2015.
There was greater increase in CV deaths in young non-Hispanic Black patients.
Busko noted: “Non-Hispanic Black patients and younger adults may have had more exposure to COVID-19 at their workplace, less financial stability, greater stress, and more limited access to healthcare.”
Let’s stop and think here. Why, exactly, do you think non-Hispanic Black people and younger adults had less financial stability, greater stress, and more limited access to healthcare? Ponder that question while I tell you the rest of this story.
An expert she quoted said this: "Clearly the findings of this important study should serve as a wake-up call," he said, "since we slid back many years of progress."
Comments. When doctors decide to intervene, there are first-order effects. If you order a coronary calcium (CAC) scan, a first order effect is the presence of absence of calcification. When we decide to recommend early AF ablation, first order effects are the isolation of pulmonary vein activity or the presence or absence of a complication.
But that is not all. There are other downstream effects. I will call these second and third order effects.
For example, when you order a CAC scan and it detects a hunk of calcium in the left anterior descending artery, you’ve now created a patient out of someone who was once a well-person. That person may now think twice about vigorous exercise.
When you promote widespread CAC screening, you create a societal or culture feeling of healthism — as in, you should do this or you are an irresponsible person. It creates a societal sense that we can control our life and death.
My point isn’t that this is bad or good (I do have feelings about it); my point is that interventions have second and third order effects.
It’s the same with AF ablation. A second order effect of early ablation is the creation (in this patient) of the false idea that AF is a focal disease like supraventricular tachycardia that is amenable to a procedure rather than lifestyle modification. That person may now tell his friends or post on an AF internet forum.
Then when we as a profession promote early AF ablation, we risk creating a false sense in society that AF can be treated with a catheter rather than healthy living choices.
Now back to 2020. The virus was bad. I know that. I am not minimizing it. The first order effect of shutting down the world was to minimize spread of one pathogen. Two years on, Dr. Woodruff now tells us about a second order effect, which some people tried to mention, especially the part about increasing disparities of care.
Voices that expressed these worries about second and third order effects were...let’s just say...not encouraged.
I hope the medical establishment does morbidity and mortality on this and learns from the many lessons of the pandemic. One is surely that healthcare and healthcare decisions are complex. And it’s wise to think beyond first order effects.
Paul Wang at Stanford presented the result of the ENHANCE AF trial. The Journal of the AHA published it. I like the idea of this paper because it studied, in randomized control trial (RCT) form, the use of a decision aid for anticoagulation (AC) decisions.
AC decisions are weird. On the one hand, they are one of the most common things a cardiologist or even internal medicine clinician deals with. On the surface the decision seems banal. But on the other hand, when you get into it, it gets hugely complicated right quick.
To start with, we use the CHADVASC score, a simple integer score, which may be fast and frugal, but it gives us a false sense of knowledge. For instance, there is no left atrial data in the CHADSVASC; permanent AF is considered the same as low-burden AF.
It boggles my mind that in an era of rockets that go into space and then land safely, doctors use an integer score to make AC decisions.
Another issue: even if a 2.2% annual stroke risk is even close for CHADSVASC 2, or 3.3% for CHADSVASC 3, what does that mean? How do you (vs your partner) internalize that? How do patients understand that? I am asking here.
What about pill disutility? How often do you think about the implications of taking a pill every day, forever?
These questions are why I am so drawn to decision support. Wang and colleagues used a basic digital support tool. I looked at it. It’s a brief 4-minute video, with some basic facts, a worksheet, three easy questions, and a list of questions to bring to the clinician. The clinician also has an app — more on that in the comments.
Patients with AF were randomly assigned to the support tool or usual care.
The primary endpoint was important: it was decisional conflict.
It’s measured on a validated 16-item Decisional Conflict Scale.
You don’t hear much about that endpoint. Decisional conflict is a state of uncertainty about a course of action. Such uncertainty is more likely when a person is confronted with decisions involving risk or uncertainty of outcomes, when high-stakes choices with significant potential gains and losses are entertained, when there is a need to make value tradeoffs in selecting a course of action, or when anticipated regret over the positive aspects of rejected options is probable. The results were clear:
A total of 1001 participants were enrolled and followed at five different sites in the United States.
The main result was a clinically meaningful reduction in decisional conflict: a 7-point difference in median scores in the support arm.
This was not a clinical outcomes study, but there were no differences in stroke or bleeding outcomes.
Most interesting was the fact that reducing decisional conflict did not alter AC acceptance.
Equal numbers of patients in each arm accepted AC. That is different from some previous studies that have found as decision quality increases, acceptance of warfarin decreased.
Comments. I highlight this study because it gets close to empirically studying decisions. Again, I want to highlight the notion of second order effects. The first order effect of oral ACs is providing a net benefit in stroke reduction.
Measuring outcomes like decisional conflict underscore the idea that a second order effect of AC recommendations may be to improve patient education, and if done on a wide scale, improve societal education, which is a third order effect.
My friends, not every endpoint in cardiology need be a clinical endpoint. Reducing decisional conflict and decisional regret is huge. I’ve often wondered if an implanted cardiac device quality measure ought to be decision quality.
This is important because cardiology is replete with preference-sensitive decisions that turn on patient education and partnership with a skilled physician. The ideal situation is that patients take the most effective therapy that also aligns with their preferences.
I have some criticisms of the paper.
One is the decision support tool — specifically, the icon array graphic that depicts the number of people with and without a stroke. These are helpful for patients to see absolute risks and risk reduction. But they were only available in the clinician app. I think those need to be in the patient app so patients can appreciate the effect size.
Second, this looks like an education intervention that increased patients’ confidence in their medication. Table 1 shows that over half of the patients were already on anticoagulation so they didn’t really target people who were in the throes of decision making.
That’s an important distinction because this may be more of a good intervention for helping patients be comfortable with their anticoagulation, but it may not really be “shared decision making” as it isn’t clear how it actually influenced a discussion.
Still though, we need more studies like this. Decision quality is important, and so is empirical testing of ways to get there.
Don’t stop listening. I will not talk about the seminal trials again. Instead, I want to highlight a disturbing trend in real-world application of this procedure. A new paper is out with sobering findings, regarding patient selection and outcomes.
This paper is in the Journal of Clinical Medicine. First author Ziad Anow, who led a group from Israel who used a national electronic health record (EHR) database of patients with AF who had left atrial appendage closure (LAAC) from 2016 to June 2021.
The primary outcomes included hemorrhagic and ischemic stroke.
The sample size was 389 patients. Follow-up was short at 2 years.
Before I tell you the results, stop and think, and guess what you think the ischemic and hemorrhagic stroke rates are. As a hint, I will tell you what they were in the device arm of PROTECT, PREVAIL, and PRAGUE-17.
For ischemic stroke, in PROTECT, stroke rate was 1.4%, in PREVAIL it was 1.9%, and in PRAGUE-17 it was 4.4%.
For hemorrhagic stroke, the rate in PROTECT was 0.2%, 0.4% in PREVAIL, and not reported in PRAGUE-17.
In the observational study in Israel, ischemic stroke occurred in 13% of patients and hemorrhagic stroke in 4.4%.
The Israel paper also reported a GI bleed rate of 8%, which was higher than PROTECT and PREVAIL.
One possible explanation was that the mean age of patients in the Israel experience is 77 and mean CHADSVASC is 5.5, both of which are older than the seminal trials.
Comments. This paper pairs well with a paper I reported on in the September 17th podcast. This was also an observational study of 800 patients from Spain, Italy, and Canada.
These authors showed that one in six or 15.5% of patients in this series died within the first year. And it looks similar here in the United States.
I was part of an abstract that looked at Medicare claims data and found that the average age of patients who were having LAAC was 78 (similar to the Israel data) and had higher rates of co-morbid conditions compared with PROTECT and PREVAIL.
The most recent National Cardiovascular Data Registry data from Watchman, published in Circulation Outcomes, first author, Usama Daimee, also finds a mean age of 76, and mean CHADSVASC 4.8.
No matter how you feel about percutaneous closure, even if you are a proponent, the use of the device in older patients with greater co-morbidities is worrisome. That’s because procedural risks are higher, competing causes of stroke are higher, and the risks of bleeding from antiplatelets is higher.
The answer, to me, is so obvious. You randomize. Then we would know. We just don’t know the benefits in these sicker patients
© 2022 WebMD, LLC
Any views expressed above are the author's own and do not necessarily reflect the views of WebMD or Medscape.
Cite this: Dec 2, 2022 This Week in Cardiology Podcast - Medscape - Dec 02, 2022.