COMMENTARY

Aug 27, 2021 This Week in Cardiology Podcast

John M. Mandrola, MD

Disclosures

August 27, 2021

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 August 27, 2021, John Mandrola, MD comments on the following news and features stories.

COVID-19 Update

Not much change from last week but, fortunately, there is a deceleration in the growth of cases. I guess the bottom-line is that given the utter un-stoppability of the Delta variant, it’s fortunate that we have vaccines.

We are in for a few bad weeks. The preventability of this amount of human suffering is quite sad.

NEJM COVID Vaccine Study

A quick note on the observational study in Israel on the safety of the Pfizer mRNA vaccine and published in the New England Journal of Medicine (NEJM). Two docs came up to me today to show me articles about this study and said, Hey, Mandrola, you are wrong about the vaccine myocarditis signal. This study in the N. E. J. M. shows that myocarditis rates are higher with COVID infection than after the vaccine. This matches some of the reactions I’ve seen online. But this interpretation of the study is wrong.

The authors matched a group of vaccinated people to unvaccinated people. Then they matched a group of patients who had COVID vs those who did not. Then they measured various outcomes—from lymphadenopathy in the arm, to brain bleeding, to thrombosis, etc. But the outcome of interest for most was the myocarditis. They found 3 times the risk of myocarditis after the vaccine but 18 times the risk after COVID infection. This led to the erroneous conclusion that the myocarditis signal from vaccines was overplayed. COVID is more likely to cause myocarditis. The fatal flaw is that the authors did not stratify the myocarditis risk by age and gender.

Almost no one thinks the mRNA vaccines are not safe for older adults. But there is a clearly concentrated myocarditis risk in adolescent boys and young men, especially after the second shot. It’s a much higher risk in males than females.

For that age group, the risk of COVID is super low, and all I am saying is don’t obfuscate this signal. Instead, let’s study better ways to protect this group. For example, give one shot or a different vaccine. Let’s also rethink mandates for this age group. It boggles my mind that we are going to force a 17- or 18-year-old kid to take two shots to go to college or hold a job, given the Centers for Disease Control and Prevention (CDC) data.

Also, the fact that NEJM let this sort of paper out knowing full well about the gender and age specificity of the vaccine myocarditis signal tempts me to be cynical about science adjudication.

Anticoagulation in COVID-19

The NEJM published a large platform trial concerning the use of full-dose vs prophylactic anticoagulation (AC) in patients with COVID-19. The platform part of this study was actually the combination of three trials: ACTIV-4a, REMAP-CAP, and ATTACC. The report was published in two manuscripts—one with patients with moderate COVID disease, the other with patients who were in the ICU.

This is important work because hypercoagulability seems to be a feature of patients with COVID-19. A higher propensity of clotting makes clinicians think higher dose anticoagulation would be helpful, and observational studies published in major journals found that full-dose AC was associated with better outcomes. But regular listeners of this podcast know two things:

  • Observational studies are limited by confounding and bias. It’s very difficult if not impossible to adjust for confounders.

  • AC isn’t free: it comes with the major complication of major bleeding.

That’s why you need randomization.

AC in Patients With Moderate Severity COVID-19. The first paper included about 2200 patients with moderate COVID-19 disease. One group was randomly assigned to therapeutic AC, mostly with low molecular weight heparin (LMWH), and the other to prophylactic LMWH. 13,000 patients were screened and about 11,000 were excluded for various reasons. That’s important because it gets to application of the trial to real-world patients.

The primary endpoint is complicated: it was organ support–free days, evaluated on an ordinal scale that combined in-hospital death (assigned a value of −1) and the number of days free of cardiovascular or respiratory organ support up to day 21 among patients who survived to hospital discharge

This outcome was evaluated with the use of a Bayesian statistical model for all patients and according to the baseline D-dimer level. No P-values. Instead, we get something I love: the probability of clinical benefit given the data, not the probability of seeing the results given no benefit.

  • Of the roughly 1000 patients in the usual-care thromboprophylaxis group, 801 (76.4%) survived until hospital discharge without receipt of organ support compared with 939 of 1171 patients (80.2%) in the therapeutic-dose anticoagulation group.

  • The adjusted odds ratio was 1.27; 95% credible interval, 1.03 to 1.58—meaning the odds ratio ranged from a 3% to 58% improvement in organ-support-free days.

  • The absolute risk difference was 4 percentage points (0.5-7.2) favoring the full AC group.

What does this mean? 76% vs 80% survival without organ support seems pretty close. This is where the Bayesian analysis comes in. Bayes theorem holds that the prior probability times the likelihood ratio (the probability of data given the null divided by the probability of the data given alternate) yields the posterior probability.

In this case, the authors tell us that the posterior probability that therapeutic-dose anticoagulation increased organ support–free days compared with usual-care thromboprophylaxis was 98.6%.

  • When they separated the analysis by high and low D-dimer, they got a 97.3% probability of superiority in the high D-dimer cohort and a 92.9% in the low D-dimer cohort. 7.3% of patients in the therapeutic dose group died vs 8.2% in the prophylactic group. That calculated to an 87% probability of superiority.

  • Major bleeding occurred in 22 of 1180 patients (1.9%) in the therapeutic-dose AC group, and in 9 of 1047 (0.9%) in the usual-care thromboprophylaxis group.

Severe Disease: Full Dose vs Prophylactic. Again, therapeutic AC vs thrombophylactic AC, but in critically-ill patients.

One important caveat: standard prophylaxis was left to the discretion of clinicians. In the critically ill patients, 22% of the therapeutic-dose group did not receive a therapeutic dose, whereas 52% of those in the control group received an intermediate dose, a factor that may have diluted any benefit of therapeutic-dose anticoagulation. This issue was somewhat less important in the trial involving patients with moderate disease, in which 20.4% of the therapeutic-dose group did not receive a therapeutic dose, whereas 26.5% in the control group received an intermediate dose.

Same endpoint: organ support–free days. About 500 patients in each arm. The trial was terminated at an interim analysis due to futility.

  • The median adjusted proportional odds ratio for the effect of therapeutic-dose anticoagulation on organ support–free days was 0.83 (95% credible interval, 0.67 to 1.03).

  • That yielded a posterior probability of futility of 99.9% and a 95.0% posterior probability that full dose AC was inferior.

  • 63% of patients on full dose AC survived to hospital discharge vs 64% of patients on prophylactic dose.

Comments: It is now 18 months into a pandemic in which millions of people have been hospitalized for COVID-19. These studies measured outcomes during a hospital stay, a matter of days, not years like most cardiovascular (CV) trials. So the first message is that it is sad that it took this long to study something so important and simple as whether to give full AC.

The second message is that in May of last year, the editor of JACC was declaring full dose AC a clear winner based on a fatally flawed observational study that showed a 200% reduction in death in patients on mechanical ventilation who received full dose AC. I wrote a column on that study. These RCTs utterly refute that observational study: in the RCTs, full dose AC vs prophylactic AC in patients in the ICU had a 95% probability of being inferior.

What a lesson! My friends, observational studies are not useless, but making causal conclusions about therapies from non-randomized comparisons is extremely precarious.

The third message concerns the magnitude of benefit in the moderately ill group. Yes, the probability of “any” benefit, even a 1% relative risk improvement is 90+%, but the question clinicians want to know is what is the probability of a larger benefit, say a 10% or 20% relative risk improvement? We don’t usually accept CV drugs unless there is a 10% or 15% relative benefit. The magnitude of benefit is critical because there was a doubling of the risk of major bleeding.

  • Recall that the odds ratio was 1.27 for more organ-free-support days, But the lower bound of the credible interval was just 1.03—1.00 would be no better. So, basically, the evidence is not strong. If your prior was pessimistic, the data doesn’t change it much.

  • If you believed the probability that full dose AC had no effect was 75%, the data would only change that to a 23% probability of no effect.

The fourth message relates to selection of these patients; 80% of screened patients were excluded. This means the benefit of full dose AC was reserved for a very select group.

In sum, regarding preventive AC, we can say don’t use full dose in the ICU, but outside the ICU, we have oodles of uncertainty: we have less than robust evidence for a large effect size in an ordinal outcome, little difference in survival, higher bleeding, an unexplained difference in effect depending on severity of disease, and strict inclusion criteria. Full dose AC may benefit some patients, but gosh, clinicians will have to be super careful in applying these results at the bedside.

Exercise in AF

At the European Society of Cardiology (ESC) meeting, Adrian Elliot, an exercise scientist in the group of Prash Sanders in Adelaide, Australia, presented a small but elegant randomized controlled trial of structured exercise in patients with atrial fibrillation (AF). The clinical question was whether a structured exercise program would reduce clinical outcomes in patients with symptomatic AF.

Observational studies associated cardiorespiratory fitness with less AF, but observational studies may be confounded—patients able to exercise may have less severe atrial disease.

This research team randomly assigned 60 patients to a combined supervised and home exercise program or usual care. The exercise group had an exercise program tailored to their capability; this included 4x4 minute efforts at 90% effort. It was progressively increased by 20% each week and monitored with detailed physical activity diary. The control group got education and recommendations to do 150 minutes of physical activity weekly.

The primary endpoints were freedom from AF and AF symptom severity.

  • In the first endpoint, at 12 months, AF recurred in 60% of the exercise arm vs 80% of the control arm. The hazard ratio was 0.50 with a confidence interval ranging from .33 to .78.

  • In the second endpoint, also at 12 months, there was a statistically significant improvement in AF symptom severity as measured on an atrial fibrillation severity scale (AFSS) questionnaire.

  • Exercise capacity as measured by VO2 peak modestly increased at 6 and 12 months in the exercise group, but there was no significant weight loss, blood pressure effect, or structural measurements.

With modest exercise for just 6 months, you wouldn’t expect big losses in weight or improvements in echocardiogram parameters. There are limitations. The trial cannot be blinded, the effect size is small, and many patients can’t or won’t exercise. But this is a signal that if we try, and exercise patience, and don’t rush to ablate, we can help some motivated patients with an extremely safe and beautiful therapy.

EMPEROR-PRESERVED – EMPAGLIFLOZIN in HFpEF

The big trial, along with a pooled analysis of EMPEROR-Preserved and EMPEROR-Reduced will be presented this evening in Europe, which is midday on the East Coast here in the United States. I tape this podcast Friday morning before clinic. I have only the preliminary slides from the press release. But it’s such a big deal that I have to mention it:

EMPEROR-Reduced and EMPEROR-Preserved were sister trials, with very similar protocols, case report forms, investigators, endpoints, statistical plans, and administrative structures. They were carried out during a similar time period and separated by an ejection fraction of 40%. Briefly, EMPEROR-Reduced studied roughly 1800 pts with heart failure with reduced ejection fraction (HFrEF). Empagliflozin (Empa) 10 mg reduced the composite endpoint of heart failure hospitalizations (HHF)/CV death by 5% in absolute terms and 25% in relative terms. The effect was driven mostly by HHF. The 8% relative risk reduction in CV death was not statistically significant. Same with overall mortality: numerically lower but not statistically significant.

  • The EMPEROR-Preserved trial studied roughly 6000 patients with HF with preserved EF (HFpEF).

  • Empa 10 mg daily reduced the composite endpoint of HHF/CV death by 3.3% in absolute terms and 21% in relative terms.

  • The effect was driven mostly by HHF not CV death.

  • The 9% relative risk reduction in CV death was not statistically significant.

  • Overall mortality was not different. Hazard ratio was right at 1.00.

  • About one-third of the patients in EMPEROR-Preserved had an EF of 40% to 50%. That’s important, because these were not all ‘normal’ EF patients. Both trials found that the effect appeared consistent in diabetics and nondiabetics.

  • In both trials, Empa reduced HHF across all EFs, except for an attenuation in those with an EF of 65% or greater.

NEJM published a second research letter on renal outcomes in the two trials, called EMPEROR-Pooled.

  • The major renal outcome was a composite of three things: sustained decrease in estimated glomerular filtration rate (eGFR) of > 40%, or decrease in the eGFR to less than 10 to 15, or dialysis.

  • Empa reduced major renal events in EMPEROR-Reduced (HR 0.51), but not in EMPEROR-Preserved (HR 0.95). This indicates that EF influences the effect of Empa on the kidney.

In the EMPEROR-Pooled presentation, Dr. Milton Packer offered some comparisons to PARAGON HF. Recall that PARAGON was a trial comparing sacubitril/valsartan to valsartan alone in patients with HFpEF. The baseline characteristics in EMPEROR and PARAGON were similar. The PARAGON-HF trial randomly assigned 4800 patients who were followed for a median of 35 months, whereas the EMPEROR-Preserved trial randomly assigned 6000 patients who were followed for a median of 26 months.

The incidence rates of the composite primary outcome in the control arm were similar: 14.6% in PARAGON and 12.5% in EMPEROR-Pooled. But because PARAGON had longer follow-up, it had a lot more events, and thus more statistical power.

In PARAGON, the neprilysin inhibitor reduced the relative risk of the composite endpoint of CV death and total HHF by 13%. HR 0.87 (0.75-1.01). The P-value was painful at 0.06. But in EMPEROR, Empa reduced the same composite endpoint by 21%, RR 0.79 (0.69-0.92); the P-value was 0.003.

Dr Packer showed a slide and wrote up an editorial in Circulation showing a comparison across the two trials, which he admits is fraught. For the EF in the mildly reduced ranged, Empa vs placebo had a slightly better relative risk reduction than did sacubitril/valsartan vs valsartan.

Comments: I wish I had more time to think about this. In EMPEROR-Preserved, I see a statistically robust signal of the composite endpoint of HHF/CVD. This is an important endpoint. To date, no drug has proven effective in HFpEF.

That is a win for Empa. But there was not a significant reduction in CV death (in either of the EMPEROR trials) or overall death. Recall that in DAPA-HF, dapagliflozin did reduce CV death in patients with HFrEF.

We also learned that renal outcomes with Empa were not reduced in patients with HFpEF whereas they were in patients with HFrEF.

  • My question is, if Empa is so great at reducing HHF, why doesn’t it reduce CV death or overall death or renal outcomes? Was the trial too short? Would longer follow-up and the accumulation of more events strengthen the effect size?

  • The investigators and proponents will focus on HHF and the composite but clinicians and patients are going to have to weigh the value of a drug that did not reduce CV death or mortality or renal outcomes.

  • For a patient with HF and an EF of 40%, Empa seems quite reasonable. But for the extremely common elderly woman with a small hyperdynamic left ventricle and HFpEF, probably not so much.

The other question facing clinicians is Empa vs sacubitril/valsartan. For many patients, either cost or blood pressure concerns will force a choice between classes of drugs. You could go by Dr. Packer’s analysis and say Empa had more of a reduction in the composite endpoint, but of course, Empa was compared with placebo and sacubitril/valsartan was compared against the active comparator of valsartan. The tougher control arm may have blunted the response of sacubitril/valsartan. We will learn a lot from the DELIVER trial which looks at dapagliflozin in HFpEF. Was the Empa signal a class effect? Is one drug the superior choice for HF?

In sum, looking at the vast health benefits, in diabetes, in chronic kidney disease, and now in both types of HF, the SGLT2 inhibitor drugs look to be a lot like statins—a major breakthrough in the prevention of bad outcomes for major organ diseases.

Comments

3090D553-9492-4563-8681-AD288FA52ACE
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.
Post as:

processing....