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.
In This Week’s Podcast
For the week ending September 10, 2021, John Mandrola, MD comments on the following news and features stories.
Thankfully, US COVID case numbers are on the decline again, even in Florida. But hospitals—including ours—are at capacity and it’s going to be hard for the coming weeks. Folks who work on the front lines have my respect.
Nearly every week, I think about how fortunate we were to have had mRNA vaccines available during this surge. I spent lots of time in the clinic convincing hesitant cardiac patients that the benefits are clear.
EMPEROR Preserved Update
First, an update on my coverage last week of EMPEROR-Preserved, the large trial comparing empagliflozin vs placebo in patients with heart failure with preserved ejection fraction (HFpEF). The trial found a statistically significant improvement in the composite outcome of heart failure hospitalizations (HHF) and cardiovascular (CV) death, but it was driven by HHF. CV death and overall death were not changed.
My coverage focused on the fact that HHF represented a very small proportion of total hospitalizations, and the reduction in HHF was not enough to significantly change all-cause hospitalizations. That criticism stands.
But I participated in a journal club on the trial this week and two further points came up.
The first point was the lack of overall death signal. The lack death benefit stands out because death rates were quite high at 14% in both groups.
You might wonder how this compares with other trials.
In the 1996 carvedilol HF trial, death occurred in 7.8% in the control arm, and yet the beta blocker reduced death by 65%.
In the 2011 EMPHASIS-HF trial of eplerenone vs placebo in patients with HF, the death rate in the placebo arm was 15.5% and the antimineralocorticoid (MRA) was able to reduce all-cause death by a statistically significant 24%.
Thus, it used to be commonplace for new HF therapies to reduce all-cause death; and as baseline therapies improve, it is harder to show big gains with new drugs. That’s important because new therapies will be hyped and very expensive.
The second point that came up in that journal club is that nowhere in the New England Journal of Medicine publication is listed the percent of patients on loop diuretics at baseline. Table s2 tells us about all the other classes of drugs but not loop diuretics.
There is a baseline characteristics paper in the European Journal of Heart Failure that tells us about two-thirds were on a loop or a high ceiling diuretic. Since one of the mechanisms of SGLT2 inhibitors is via diuresis and you had to have high BNP to get into the trial, I would have been very interested in a subgroup plot looking at outcome on or off loop diuretic.
STEP
Despite numerous trials looking at blood pressure (BP) targets—SPRINT, ACCORD, JATOS—the goldilocks number remains elusive. The STEP study group set out to compare CV outcomes in hypertensive Chinese patients and an intensive systolic BP target of less than 130 mmHg vs standard therapy of 130-150 mmHg. The primary outcome was a vast composite of stroke, acute coronary syndrome (ACS), HF, coronary revascularization, atrial fibrillation (AF), or death from CV causes.
About 4200 Patients were randomly assigned to each group. Mean age 66 years, mean systolic BP 146 mmHg. Office BP measurements were a bit more normal than in SPRINT: Performed by a trained trial staff member (physician or nurse). Patients were required to rest for at least 5 minutes in a seated position, and then the blood pressure was measured by trial staff (observed) three times at 1-minute intervals.
First, the two groups achieved their targets: mean systolic BP in the intensive arm was 127 mmHg vs 135 mmHg in the standard arm. Median follow-up 3.3 years.
Second, clinical outcomes: the incidence of any primary endpoint (PEP) event was 3.5% vs 4.6% and the hazard ratio (HR) was 0.74 with a confidence interval (CI) ranging from 0.6 to 0.92. Stroke, ACS, HF were all reduced with a CI that stayed below 1.
Safety was pretty good: Hypotension was 3.4% in the intensive group vs 2.6% in the standard group. But low BP was defined as less than 110 mmHg. There was no difference in fractures or kidney outcomes.
Outcomes held up in the subgroups. Visually, from the plots, the intensive arm looked better in those with higher baseline BP and higher Framingham risk.
This is clearly a positive trial, but I think there are some important limitations. Let’s look at the inclusion/exclusion criteria. Only patients of Han ethnicity were included. The authors state explicitly that this limits generalizability. Exclusion: patients with stroke, any myocardial infarction (MI), revascularization in the previous year, or diastolic BP less than 60 mmHg were excluded. This enrolls a pretty robust group of older patients. The evidence supporting that was the death rate of only 1.5%. (Contrast that with the 14% death rate in EMPEROR-Preserved.)
Combine the robustness of the enrolled population with the fact that these patients are in a tightly regulated clinical trial. I suspect that it’s hard to duplicate the careful follow-up of an RCT to the real world. Pushing on BP in more fragile patients in real life may be harder.
One criticism I heard that I don’t think is that important is that patients in the control arm who had systolic BPs below 120 mmHg had their meds reduced to get back up to a higher target. We wouldn’t do that in practice, goes the critique. My comment is that this is a scientific experiment set out to test a BP target, and if there weren’t differences in BP, there would be little knowledge gained.
So, I guess if you had a patient like those in STEP or SPRINT and they really wanted to do everything to absolutely maximize BP then it’s probably worth trying to go a little lower. But I would be very cautious in translating this data. The absolute risk reductions were small. The potential downsides are significant.
GUIDE-HF
Whenever I hear the statement: If only we had more data, we’d be better able to manage this patient, I think back to the hours I spent studying Swan-Ganz tracings and cases. More data from the pulmonary artery (PA) catheter would surely help us manage sick patients in the ICU. Well, you know how that went: RCTs showed the invasive monitor that provided hemodynamic data didn’t change outcomes.
Now there is device made by Abbot called CardioMEMS that goes into the pulmonary artery and provides feedback that is designed to help manage patients with HF. The idea is that daily weights and observation for edema, etc, are not enough. The monitor in the PA must be better. Give me granular data and I will fine tune those diuretics—perfectly!!!
And, way back in 2011, the CHAMPION trial found that device-managed HF patients had 28% fewer HHF. The problem with that trial was that it enrolled a narrow group of New York Heart Association (NYHA) Class 3 patients with a previous HHF admission and was not powered for outcomes like mortality.
Of course, given the expense and added workflow needed to manage the data from this device, one should have a properly powered trial that enrolls a broader population of HF patients. At the European Society of Cardiology (ESC) meeting we learned results of the GUIDE-HF RCT comparing use of PA pressure data with standard care. Both groups got the monitor. The endpoint was a composite of total HF events and all-cause mortality. Crucially, the trial enrolled a broader group of patients who could have Class 2-4 NYHA symptoms and either a recent HF admission or elevated BNP.
About 1000 patients were enrolled at more than 100 centers.
The rate of the PEP was 0.563 per patient year in the monitor-on group and 0.64 per patient year in the standard care group. The HR was 0.88 and the CI ranged from 0.74 to 1.05 with a p-value of 0.016.
40 patients died in the monitoring group vs 37 in the standard care. So obviously no signal of benefit.
Nonsignificant trial, right? We can relegate this PA monitor to the bin with Swan-Ganz catheters. Nope. In perhaps one of the boldest attempts at spin I have seen in a good while, investigators presented a “pre-COVID-19 impact analysis that compared outcomes before March 2020.” By that time, all of the trial’s patients had been followed for at least 3 months, and about three-quarters of the endpoints were counted.
The post-hoc analysis pointed to a possible benefit for the CardioMEMS-guided strategy. This was a barely significant 19% drop in risk (P = .049) for the primary endpoint of death, HF hospitalization, or urgent HF hospital visit. The effect was driven by a 24% decline in HF events (P= .014), with no significant contribution from mortality.
Steve Stiles has an excellent report: he cites Larry Allen, an HF specialist at the University of Colorado. Allen has it right I think:
“While the technology is pretty amazing, the real question is whether it tells us something that we didn’t already know that leads to improved care. Unfortunately, as tested here, it doesn’t, or at least not enough to make a big difference."
This is the same message as the Swan-Ganz catheter years ago. It is fancy, outputs lots of data, but that is not enough. If you propose an invasive device, intensive data monitor, and significant costs, you must show clear benefit. GUIDE HF did not demonstrate that. Do not be distracted by an after-the-fact analysis of data.
I want to be careful with my words here, but the attempt to sell doctors on such a re-analysis tempts me toward cynicism. Why? Because Abbott funded the trial and stands to gain sales from a positive trial. And, perhaps worse, many of the authors have financial conflicts of interest. When people see this stuff in a place like the Lancet, they begin to worry about other science. And that is sad.
IAMI Trial
COVID has made us forget about influenza. But long before SARS CoV-2 and surely in years to come, when travel patterns return to normal, influenza will again threaten older patients with CV disease. At ESC, Prof Ole Frobert presented results of the IAMI trial comparing flu vaccine vs placebo in patients while they were admitted to the hospital for MI. The primary endpoint was a composite of all-cause death, MI, or stent thrombosis at 12 months. Guidelines have recommended flu vaccine for these patients, but the recommendation was based on observational studies and small RCTs.
IAMI enrolled about 2500 patients at 30 centers, half received flu vaccine and half received placebo. Sadly, due to the pandemic, the Data Safety Monitoring Board recommended halting the trial before it attained the prespecified sample size. This is different than the previous study, because, with the dominance of SARS CoV-2, influenza incidence, along with other respiratory viruses, plummeted. Plus, the increasing illness with COVID would have made adjudication of events difficult.
The authors calculated that they would need 386 events to reach 80% power to detect a 25% reduction in the composite PEP.
There were only 158 events—5.3% in the vaccine arm, and 7.2% in the placebo arm. The HR was 0.72, so a 28% reduction. The CI ranged from 0.52 to 0.99; P=0.040.
The rates of all-cause death were 2.9% in the influenza vaccine group and 4.9% in the placebo group (HR 0.59 [95% CI 0.39 to 0.89], P=0.010).
CV death was 2.7% in the vaccinated vs 4.5% in the unvaccinated (HR 0.59 [95% CI 0.39 to 0.90], P=0.014).
Strengthening the finding a bit was that 13% of the placebo arm crossed over to get the vaccine—which would have biased the trial towards the null.
Serious adverse events were rare and not different.
The authors also included in their simultaneous publication in Circulation, a meta-analysis of prior trials, which included three other RCTs of flu vaccine vs placebo in high-risk CV patients. The pooled estimate of cardiovascular death of the HR from the fixed-effect meta-analysis of all four trials was 0.51; 95% CI, 0.36 to 0.71; P=0.0001. There was no evidence of between-study heterogeneity (P=0.48, I-squared=9.7%).
During the trial discussion, Professor Barbara Casadei emphasized the main limitation of the paper—early termination reduced the power to detect differences. The authors also led with this limitation. The point about early termination is not so much the possibility of false positive result, but more an accentuation of effect size. There’s an important paper on this by Bassler and colleagues in the Journal of the American Medical Association in 2010. It’s a systematic review and meta-analysis of truncated vs non-truncated trials.
One of the main findings was that large differences in treatment effect size between truncated and nontruncated RCTs (ratio of relative risks <0.75) occurred with truncated RCTs having fewer than 500 events.
IAMI had a HR lower than 0.75 and far fewer than 500 events.
It’s possible that the effect size was exaggerated but even so, the adverse effects are minimal to none, and given the strong probability of any benefit, it seems wise to vaccinate older adults admitted with MI.
Some may wonder how the flu vaccine reduces CV outcomes. I do too, but I would also add, as Prof Casedei said, it doesn’t matter how it works. That it has been shown to work in multiple RCTs is enough.
I intentionally bookended the IAMI with EMPEROR-Preserved. In IAMI we have an extremely low-cost, low-burden intervention that likely reduces important outcomes like death and CV death. In EMPEROR-Preserved, we have an extremely costly pill to be taken every day, that reduced a surrogate outcome of HHF and had no effect on mortality, nor renal outcomes, nor quality of life.
There were a few other notable studies that I will cover next week in ESC #3, including MASTERDAPT and STOP-DAPT-ACS.
© 2021 WebMD, LLC
Any views expressed above are the author's own and do not necessarily reflect the views of WebMD or Medscape.
Cite this: Sep 10, 2021 This Week in Cardiology Podcast - Medscape - Sep 10, 2021.
Comments