Apr 12 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 April 12, 2024, John Mandrola, MD, comments on the following news and features stories from the American College of Cardiology annual meeting: DanGer Shock, (plus a sobering JAMA research letter on Impella use), REDUCE-AMI, PREVENT, and EMPACT-MI.

DanGer Shock Trial

Five reasons make the DanGer Shock trial of the micro-axial flow pump (Impella CP) vs standard care in patients with acute myocardial infarction (MI)-related cardiogenic shock one of most important trials in recent years.

Cardiogenic shock due to acute MI is highly lethal disease. Any therapy for such a bad disease is important.
The Impella support device comes with potential serious harms — limb ischemia, bleeding, hemolysis, etc.
Impella is a costly device.
Despite more than a decade of use, no previous trial has shown it to be effective. The enthusiasm for this device despite no randomized control trial (RCT)-level evidence has always surprised me. I mean, in trials, Impella underperformed against the intra-aortic balloon pump (IABP) and the IABP had previously proven no better than standard of care.
Finally, DanGer Shock showed cardiologists that is possible to do trials of terrible diseases such as acute MI-related-cardiogenic shock.

Dr Jacob Moller from Odense University in Denmark spent 11 years randomly assigning 355 patients with acute MI-related cardiogenic shock to Impella vs standard of care. The entry criteria were highly selective. Patients had to have real shock — soft blood pressure (BP), high lactate, on pressors, but they could not be neurologically too far gone.

The primary endpoint could only be mortality. Time of measurement was not 30 days, but 180 days. You probably know the results:

Patients in the Impella arm had a 26% lower risk of death; hazard ratio (HR) 0.74 (0.55-0.98). The P-value was 0.04.
There was a 4 times higher rate of serious complications in the Impella arm. This included: bleeding, 2 times the risk; limb ischemia, 5 times; renal replacement therapy, 2 times; and sepsis with positive blood cultures, 3 times.

Yet, think about it. All of these adverse effects could easily be on the causal path of death but death was still lower in the Impella arm: 46% vs 58%, -a 12% absolute risk reduction.

Comments. Congratulations to Dr. Moller and the DanGer Shock team. It’s an impressive thing to stay with a trial for 11 years.

As a neutral Martian outside observer (and rare user of Impella during ventricular tachycardia [VT] ablation) I’ve been skeptical of the device, primarily because of the disconnect in enthusiasm of the proponents despite the total lack of trial-level support.

My US colleagues say, but, John, we don’t need trials, because we take care of these patients, and we can tell you that some patients don’t make it out of the lab without this device.

Now, DanGer Shock strongly supports the contention that some patients are indeed saved, and I am happy that is true.

I am not deterred by the P-value of 0.04. The confidence intervals (CI) lend themselves to as much as a 45% lower rate of death. And I think we need to give trials that have control arm death rates of 60% at 6 months a bit more statistical leeway.

But this trial has some important caveats that affect translation.

Mainly, DanGer-shock applies to a very specific patient.
Patients had to be sick with a median arterial lactate level of 4.5 mmol/L, impaired left ventricular ejection fraction (median LVEF 26%) and could not present late (median time to randomization was 4 hours).
But they could not be too sick, say comatose or with right ventricular (RV) pump failure.

I am glad Dr. Moller took care to select patients who were most likely to benefit.

Now it’s our job as cardiologists to apply this therapy to patients similar to those in DanGer Shock. These results should have no bearing on use of the device for other indications – say, high-risk PCI. This too could be studied in RCT format.

And a word on regulatory matters.

I wrote on X today that, had the US Food and Drug Administration (FDA) required approval with evidence, DanGer Shock could have been completed in 11 months, not 11 years. Before acceptance.

Imagine if this trial was negative, as were all previous Impella trials. Think of the harm we would have caused and money we would have wasted. If DanGer Shock were negative it would have been as ugly as the antiarrhythmic drug reversal in the CAST trial.

Manufacturer Payments to Cardiologists and Use of Devices

Research reported in JAMA, led by Sanket Dhruva, MD, at the University of California San Francisco, along with Joseph Ross, MD, at Yale and others, studied the association between payments from Abiomed (the maker of Impella) to doctors and the doctors use of the device.

In their introduction, the authors note that use of the Impella support device for cardiogenic shock has doubled over just a 2-year period from 2015 to 2017.

Given the total lack of RCT-support during that time, the authors rightly considered industry payments as a potential cause.

They used the Open Payments database in the United States, which mandates documentation of any industry monies received. They then linked payments to billing in a sample of Medicare fee for service beneficiaries.

There were four main results:

56% of cardiologists who placed a LV assist device (LVAD) received payments. The median value was $164.
About 6000 cardiologists received LVAD payments in individual years. Of these cardiologists, 23% performed at least one LVAD, whereas among the 11,000 cardiologists not receiving LVAD payments, 9.5% performed at least one LVAD
In multivariable analyses, receipt of LVAD payments (vs no receipt) was associated with a 2 times higher rate of any LVAD use in the year of payment and a 1.9 times higher rate of LVAD implant in the subsequent year.
The highest risk was observed in the highest tertile of LVAD payment values.

Comments. As much as we celebrate the positive results of DanGer Shock, we should also not look away from this analysis. I am sorry, but it is unseemly.

People outside the medical guild like to think of their doctor as a trusted neutral adviser. Like a judge. Not influenced by anything other than the patient’s best outcomes. This analysis casts doubt on the presumption of neutrality.

Docs who received payments were twice as likely to use a device than docs who did not get paid.
And, the more they were paid the more devices they use.
These two associations may not prove causation, but they are both worrisome associations.
And to those who say — naively — that industry payments don’t affect their choices; how exactly do you refute such an association?
Answer: You cannot. Dhruva and colleagues have simply documented human nature. And a likely reason this procedure gained popularity despite the lack of data.

Industry money has two effects: one is to influence specific actions of single doctors. But the larger effect is to create a therapeutic fashion. That is, if big-name docs are doing it, and if the docs across town are doing it, then perhaps we should do it as well.

Sorry again, I am going to moralize. I don’t believe this is right. In the first half of my career, I was susceptible. I went to dinners, took money to speak from companies, but I don’t anymore. It’s not necessary.

I have learned conduction system pacing without taking money from industry. It is possible and it is probably better, because if education is from industry, it is not neutral.

One thing that makes me sad about my field of electrophysiology is that so much of our education comes from industry.

REDUCE-AMI Trial

Speaking of important trials, REDUCE-AMI tested one of cardiology’s strongest dogmas: use of beta-blockers after MI. The seminal trials, BHAT, ISIS-1, and COMMIT date back decades ago.

MI care now looks nothing like it did in the pre-Internet age. The likely benefit of beta-blockers after an MI stemmed from the reduction of arrhythmic death and prevention of complications due to wall stress. Acute revascularization in MI prevents much of the myocardial damage that used to occur in the era of BHAT and ISIS 1.

The investigators, led by Troels Yndigegn, MD, from Lund University/Skane University Hospital, Sweden, asked a simple question: Do beta-blockers still reduce death and recurrent MI in patients with preserved LVEF who just had an acute MI? Emphasis here on normal EF.

REDUCE-AMI investigators had priors to work with. A meta-analysis of 60 trials of post-MI beta-blockers reported that the later trials (vs the early trials) no longer found any benefit from beta-blockers after MI.

This trial was a little different than a typical industry-sponsored mega-trial. They used the Swedish registry to assess outcomes. There was no placebo tablet. Only the standard of care arm without beta-blockers.

Here are the results:

After 3.5 years of follow-up, the primary endpoint of either death or MI occurred in 7.9% of the beta-blocker arm vs 8.3% of the control arm (HR, 0.96; 95% CI, 0.79-1.16; P = 0.64).
Other secondary endpoints, such as cardiovascular (CV) death, hospital admission for atrial fibrillation, or heart failure (HF) also did not differ between the two groups.
Neither did safety endpoints such as bradycardia, hypotension, or hospitalization for lung disease.

Now, some purists may argue that the CI was wide and perhaps this trial was underpowered. They would note that the lower bound of 0.79 indicates a possible 21% benefit while the 1.16 upper bound allows for 16% harm. They might say that evidence of an absent effect is a high bar, and in this case, there may be more of an absence of evidence.

Comments. Let me first argue that the purists, in this case, are probably wrong.

First, the event rates of 2.4% and 2.5% are nearly identical. A P-value of 0.64 indicates that if we assume there was no effect of beta blockers, then this data would not be surprising at all.
Second, the reason the CI is wide is because there were fewer events than the investigators expected. That’s important because if you have low event rates, it’s harder for any therapy to lower an already low number. Again, this strengthens confidence in the null result.
Another reason I am not swayed by the purists is the prior knowledge. I realize the statistical plan here was frequentist, which does not take prior beliefs into account, but we as doctors can think in a Bayesian way and consider our pre-trial expectations: that is, that based on a 60-trial meta-analysis showing no benefit in later beta blocker trials, we expected no benefit. I wrote in my column that the results of REDUCE-AMI are like a negative stress test in a patient with no risk factors and atypical chest pain.

As it so often is, the two main take-home messages are a) beta-blockers likely offer no benefit to post-MI patients with a preserved EF; that’s great to know because beta-blockers come with potential side effects, and reduction of pill burden is always a good thing; and b) evidence-based practice needs expiration dates.

Beta-blockers exert the same pharmacodynamic effect now as they did in the 1980s; what has changed is the disease. When diseases change (say the risk of sudden death in heart failure), we need new trials.

Listeners also note that there are at least three, maybe more, large RCTs testing a similar question. These trials, like DANBLOCK, BETAMI, and REBOOT will likely confirm this trial.

PREVENT Trial

This is a fun topic. We learn early on that the most likely lesion to cause an acute coronary syndrome (ACS) is a young, thin walled, friable, but non-flow-limiting lesion. A vulnerable plaque if you will. Highly stenotic, old, chronic lesions are less apt to rupture or clot off, and more likely to cause angina.

The Holy Grail of cardiology is to a) identify and b) treat these vulnerable plaques. South Korean investigators, led by Seung-Jung Park, MD, attempted to study one approach to this foundational problem.

To date, all of the evidence we have, suggest that percutaneous coronary interventions (PCI) of lesions should be reserved for highly obstructive, symptomatic, medically refractory lesions or during an ACS.

Park had another idea. Let’s take those lesions that are visible (more than 50% on angiography), but non-flow-limiting — negative fractional flow reserve greater than 0.8 — and do intracoronary imaging. If the imaging shows two of four high risk features for plaque vulnerability, then let’s randomly assign those patients to preventive PCI vs medical care.

Patients had to get to a cath lab to be randomized. The target lesion could have been the only lesion, or it could have been a second one, say after an ACS-lesion was stented.
They screened over 5600 patients in South Korea, Japan, Taiwan, and New Zealand to enroll about 1600 in the trial.
The primary endpoint was a composite of major adverse cardiac events (MACE) — CV death, target-vessel MI, ischemia-driven target-vessel revascularization, or hospitalization for unstable or progressive angina.
The trial had up to 7.9 years of follow-up (median, 4.3 years), but the primary endpoint was assessed at 2 years.
The results were positive. At 2 years, a primary outcome event occurred in three patients (0.4%) in the PCI arm and 27 patients (3.4%) in the control arm. The authors reported this as an absolute risk difference of –3.0% (95% CI, -4.4 to -1.8; P =·.0003).
Converting to relative risk yields an HR of 0.11 (95% CI, 0.03-0.36). Each component of the primary endpoint favored the PCI arm.
A secondary composite endpoint of death, any MI, or any revascularization also favored PCI (risk difference, -2.2%; 95% CI, -4.1 to -0.2).
The Kaplan Meier curves separated early then remained parallel, suggesting all the benefit occurred in the first months.

I wrote about this study, and I started by saying how much I liked it. It was so bold. And novel.

But, it’s early. There are many reasons to be cautious. Sadly, the authors were not cautious in their conclusion, and their enthusiastic tone in the discussion.

They wrote that their results “support consideration to expand indications for PCI to include non–flow-limiting high-risk vulnerable plaques."

I don’t think so. For many reasons:

There were small numbers of events, three vs 27, mostly non-fatal events, in a trial of 1600.
The significant results at 2 years did not remain significant at 4 and 7 years. You’d think that if this intervention — PCI for vulnerable plaques — was so great, the accrual of more events would strengthen, not weaken, the effect size.
The other issue, and this is not their fault, is the open-label design. Two of the endpoints are susceptible to human decision-making. There is an issue of subtraction anxiety. In other words, those patients who did not receive a stent (or get fixed) may have been more apt to be admitted for unstable angina,or have a ‘target vessel revascularization.’
There were also more patients taking P2Y12 inhibitors in the PCI arm. Another reason they may have had fewer events.
There was missing data in 50 patients; that’s not a small number given that the delta was only 24 events.
Finally, and perhaps most important, there is the matter of translating this data to real life. Imaging of coronary lesions is not as common in some geographies. What’s more, two of the four imaging techniques they used involve complicated devices, that are not universally available to many interventional cardiologists.

All this makes me say this is a great early study. It gives me some hope. But it’s very early. Medical science is hard. Major achievements are rare. We need a lot more confirmatory evidence.

It would be a terrible mistake for interventional cardiology to start intervening on what they consider vulnerable plaques based on this evidence. We need more trials.

And if this pans out, it could be a Nobel Prize level advance. And Dr. Park and his colleagues would get the credit as trailblazers.

EMPACT MI trial of Empagliflozin in the Post-MI Setting

Empagliflozin (E) is a sodium–glucose cotransporter 2 (SGLT2) inhibitor drug and this new class of medication benefits patients with diabetes, heart failure, and chronic kidney disease.

But as I described earlier in the REDUCE-AMI trial of beta-blockers, the post-MI setting is a much higher bar. It’s a higher bar because our MI therapies are really good — even when there is LV dysfunction. Think ACE inhibitor.

The trial was simple. EMPACT randomly assigned one group to E the other to placebo. The primary endpoint was hospitalization for HF or death from any cause.

EMPACT entry criteria required either an LVEF ≤ 45% or “congestion” requiring therapy. Trialists also added an enrichment factor, such as age > 65 or diabetes, to enrich the population of patients who could benefit.
Patients in the trial were on excellent background therapy of renin-angiotensin blockers, mineralocorticoid receptor antagonists, and beta blockers.
Similar to REDUCE AMI, the main results were non-significant. After 17 months follow-up, a primary outcome event occurred in 8.2 % of patients in the E arm vs 9.1% in the placebo arm (HR 0.90, 95% CI 0.76-1.06; P = 0.21.
The actual primary events were split nearly evenly between death and HF hospitalization.
All-cause death did not differ in either arm (5.5 vs 5.2%, HR 0.97 CI 0.78-1.19). HF events were lower but not enough to affect the primary endpoint. For instance, first heart failure events were 23% lower in the E arm (HR 0.77, CI 0.60-0.98). And the total number of heart failure hospitalizations were 33% lower in the E arm (HR 0.67, CI 0.51-0.89).
The rate of death due to cardiovascular causes was 4% in both arms.

Of note, the authors published, in a separate manuscript, the results of the HF hospitalization endpoints. Adverse events were similar in both groups.

Comments. The SGLT2 inhibitor story after MI is similar to the beta-blocker story. As the risk of the primary outcome goes down, new therapies struggle to make clinically important incremental gains.

EMPACT also comports well with the non-significant DAPA-MI trial of dapagliflozin in the post-MI setting.

DAPA-MI, published in NEJM -Evidence, may technically, have a positive endpoint because the authors changed it to a win-ratio, with soft endpoints that drove the positive result, but like EMPACT, the hard endpoints of HF and death were not different.

The failure of SGLT2 inhibitors in post-MI care is also similar to the nonsignificant PARADISE MI trial of the angiotensin-receptor-neprilysin inhibitor sacubitril/valsartan vs ramipril. In that trial of post-MI patients with a mean LVEF of 36%, sacubitril/valsartan did not significantly reduce the composite primary outcome of CV death or incident HF.

As I wrote in an upcoming column, the problem in all three of these cases is not the drugs. The problem is that the comparators are good. When an MI occurs, acute revascularization provides the majority of the benefit. Antiplatelet drugs reduce stent thrombosis, statins reduce future atherosclerotic events, and when there is LV dysfunction, basic afterload reductions with ACE-inhibitors provide the majority of benefit. This is all good news.

SGLT2 inhibitors have shown efficacy in certain types of patients. Two large trials now show that post-MI patients are not one of these categories, which is fine; SGLT2 inhibitors help plenty of patients.

My friend and mentor Sanjay Kaul had a very interesting thread on empagliflozin on X. He noted that we now have three HF trials (EMPEROR-Preserved, EMPEROR-Reduced, EMPULSE), one chronic kidney disease trial (EMPA-Kidney), and one post-MI trial (EMPACT-MI) that show no mortality benefit with E.

By contrast, Kaul writes, a large and clinically important mortality benefit was seen in EMPA-REG OUTCOME (with a lower risk cohort). I think we can now clearly say the latter was an outlier!

Kaul writes that “Large effect size, persuasive P values, and high replication probabilities be damned. If it is too good to be true, then it probably is.”

He then links to an opinion paper he wrote in Circulation that is open access: Is the mortality benefit with E in patients with type 2 diabetes too good to be true?

It’s excellent reading. It was written 8 years ago. He makes a Bayesian case that the remarkable 2.6% absolute risk reductions in mortality of EMPA-REG are likely to be real. But. This is hard to explain mechanistically. And we need future trials to confirm.

Now we have them — in higher risk patients. And boom. They don’t seem to confirm the big effect size and persuasive P-value.

It's all so sobering, and I like that about medical science.

Comments

Commenting is limited to medical professionals. To comment please Log-in.

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.