COMMENTARY

Sep 24, 2021 This Week in Cardiology Podcast

John M. Mandrola, MD

Disclosures

September 24, 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 on Apple Podcasts, Spotify, or your preferred podcast provider.

In This Week’s Podcast

For the week ending September 24, 2021, John Mandrola, MD comments on the following news and features stories.

COVID

Overall, the US numbers are down, but the story here is a geographic one. Case numbers in Texas and Florida are dropping, but places like Alaska and Idaho are still seeing high numbers.

I hope that after the Delta variant, and with vaccinations, we see a dormant period. I can’t wait for this to be over, so perhaps we can all be friends again and go back to arguing over anticoagulation, TAVR vs SAVR, stents, biased observational data, and left atrial appendage closure.

MASTER DAPT

At the European Society of Cardiology (ESC) meeting, we learned results of the MASTER DAPT trial, an RCT comparing an abbreviated and non-abbreviated strategy of dual antiplatelet therapy (DAPT) after drug-eluting stent (DES) placement in patients who were at high bleeding risk. The New England Journal of Medicine (NEJM) published the trial, and Dr. Marco Valgimigli was the principal investigator (PI).

Some background points: Patients with high bleeding risk (HBR) represent about a third of patients who come for DES. I won’t go through all the HBR criteria, but these include things like use of oral anticoagulants, recent bleeding, age over 75 years, systemic conditions such as low-platelets, anemia, steroid or NSAID use, or a PRECISE-DAPT score above 25 (PRECISE-DAPT is a five-item risk score using clinical criteria).

The stent used in MASTER DAPT was a DES device that is not available in the United States. Patients had a mix of acute and chronic syndromes—about 50% had acute coronary syndromes (ACS); 36% were on oral anticoagulants; mean age was 76 years.

The two strategies tested started after 1 month of DAPT: one group stopped DAPT at 1 month, the other continued DAPT for 2 to 5 additional months. The mean duration of DAPT in the abbreviated arm was 34 days vs 193 days in the standard arm.

There were 3 ranked primary outcomes:

  • Net adverse clinical events. (all-cause death, MI, stroke, major bleeding);

  • Major adverse cardiac or cerebral events (death, MI, stroke);

  • Major bleeding.

The first two were tested for noninferiority (NI) and the third for superiority. In all, 4400 patients were enrolled. The type of standard APT (SAPT) in the abbreviated group included clopidogrel in 56%; aspirin (ASA) in 31%; and ticagrelor in 14%.

Results:

  • Net adverse clinical events (NACE): 7.5% in abbreviated arm vs 7.7% in standard. Definite NI.

  • Major adverse cardiac or cerebrovascular events (MACCE): 6.1% in the abbreviated arm vs 5.9% in the standard arm; NI.

  • Major bleeding: 6.5% in the abbreviated arm vs 9.4% in the standard arm; absolute risk reduction (ARR) nearly 3% and highly significant for superiority; driven by Bleeding Academic Research Consortium (BARC) 2, which was not included in NACE.

Limitations:

  • Open label study;

  • DAPT duration varied in the standard DAPT group, but that reflects clinical practice;

  • DAPT duration in both arms was longer than what is currently recommended in OAC patients. ESC guidelines favor 1 week of DAPT plus OAC;

  • Study was not designed to assess the role of type of SAPT after stopping DAPT;

  • NI margins were wider than the observed event rates were lower than expected;

  • Results may not apply to patients treated with other stents.

Discussion: This gets a little tricky, but I think it’s important—the first endpoint, NACE, had both efficacy (thrombotic events) and safety (bleeding), which you would expect to trend in different directions and always makes NI easier to reach; right? No difference is a positive finding. Trialists do this because having more endpoint events gives the trial more power and thus require fewer patients—there is always this tension in trials because of ethics (experimenting on as few as patients as possible) and costs.

Editorialist Magnus Ohman made another important point about the NI testing used in MASTER DAPT. His point was that NI was met in the first endpoint, but was driven by less bleeding. The problem was that the observed rate of events was far lower than expected. Why is that important? Because the NI margin was set as an absolute risk difference of 3.6%, meaning if the upper bound of the event rate in the abbreviated arm was less than 3.6% then it is NI.

But that 3.6% number is based on an event rate of 12% in the control arm. That is not what was seen. Instead, the event rate was 7.7%. This is important because 3.6% represents a larger proportion of 7.7% than 12%. Sanjay Kaul also made this point on a Twitter thread from Aug 30. (By the way, this was also a problem in the Watchman vs Warfarin trials).

The low event rate also affects the NI of the second endpoint—MACCE. The NI margin here was set as 2.4% of an expected event rate of 8%. Instead, the observed event rate was 5.9%. So, the 2.4% margin is a larger proportion of 5.9%.

While the actual MACCE event rates were similar (6.1% vs 5.9%), the upper bound could have included up to a 30% higher incidence of events in the abbreviated arm. That’s pretty worrisome—30% higher rates of MACCE and still NI?

But a proponent of the abbreviated strategy could say major bleeding was significantly less. Which is true, but the difference was driven by less severe bleeding episodes. The more severe types of bleeding were similar.

Before I conclude, let’s talk about a sub-analysis of MASTER DAPT that was also published as a separate paper in Circulation. This paper analyzed the results for patients taking OAC—which made up 36% of patients.

The abbreviated and non-abbreviated arms were a bit different here. In patients with an OAC indication, the abbreviated arm was to stop DAPT after one month, then go to SAPT for 5 months, then OAC alone for more than 11 months. In the non-abbreviated arm with an OAC indication, DAPT was continued for 2 months (total of 3 months), then SAPT thereafter along with OAC.

  • First, with the abbreviated strategy, the occurrence of ischemic and net events did not differ in patients with or without oral anticoagulant therapy.

  • Second, stopping DAPT at 1 month and continuing with single antiplatelet therapy reduces bleeding significantly in patients without an indication for oral anticoagulation and numerically in patients with an indication for oral anticoagulation.

Journalist Sue Hughes has fantastic coverage, and she includes slight differences of opinion on the reduction of bleeding, mainly that the abbreviated strategy only had a significant signal of less bleeding on those WITHOUT an OAC indication.

What can we say about MASTER DAPT? I like the trial because even if you don’t do a percutaneous coronary intervention (PCI); the NI analysis makes you think.

I’d say MASTER DAPT creates many more questions than answers. Can we extrapolate these data to other stents? Do we really feel comfortable with those NI margins, given the lower-than-expected event rates and the safety signal mostly seen in non-severe bleeding cases and mostly in patients without an OAC indication?

Surely nothing about MASTERDAPT moves us toward easy algorithms of antithrombotic regimens after PCI. The choice of how best to deal with a metal cage in the coronary artery remains something that must be individualized. The tension between reducing bleeding but not increasing thrombotic events remains.

At the risk of making people mad, I will remind readers that only half of the patients in this trial had an ACS. And in stable patients, we have oodles of data that cast doubt on the treatment of focal stenoses for the reduction of myocardial infarction (MI), death, or even angina.

Of course, the best way to avoid having to balance thrombotic and bleeding complications post-PCI is to not have a PCI in the first place. For ACS syndromes, it is a no-brainer; for stable patients, there is room to think hard about the indications.

Full SAMSON Trial Published

A little less than a year ago researchers at Imperial College London presented at AHA 2020, the results of the N-of-1 SAMSON trial, a crossover trial in which patients either took the statin, a placebo, or no pill.

The Self-Assessment Method for Statin Side-effects or nocebo trial was published in letter form in the NEJM. This week, the Journal of the American College of Cardiology (JACC) has published the full paper, including the patient-level results. Last November, I felt like this was one of the most important trials of the year, and the full-publication bolsters that view.

To recap, SAMSON enrolled 60 patients who had abandoned statins and because of intolerable symptoms, these patients had no intention of restarting statins. Many of the enrolled patients had tried different statins and different doses. The most common symptom before enrollment were muscle aches (60%), fatigue, tiredness, and cramps.

The protocol was unique and elegant: Patients received twelve 1-month bottles: four bottles contained atorvastatin 20 mg, four bottles contained a matched placebo, and four bottles were empty.

Patients were instructed to record daily symptom intensity on an app and the scale was 0-100. The primary outcome of interest was the nocebo ratio: the ratio of symptoms induced by the statin that was also induced by taking the placebo.

Before I tell you the results, we should review five background points:

  • In blinded RCTs of statins, the incidence of nuisance side effects—mostly statin-associated muscle symptoms (SAMS)—are similar in the placebo and statin arms.

  • In observational studies, which are not blinded, the incidence of SAMS are higher in those who take statins.

  • While we can argue about the effect size of statins in absolute terms, these are the most-studied drugs, and they deliver a remarkably consistent 20% to 25% reduction in future cardiac events. But most people who start a statin end up stopping the drug, usually because of side effects.

  • Regression to the mean is an important concept. I mean, we docs often do mini-trials of statins. But the problem with our experiments is that the statins are often stopped when symptoms are maximal and naturally tend to decline and are restarted when symptoms have resolved and can only get worse.

  • Middle-aged people, I can testify, have lots of fluctuating aches and pains;

SAMSON findings: The marginal mean symptom score was 8.0 during the no-tablet months, 15.4 during the placebo months—significantly higher than no pill months at a P < 0.001 vs no-tablet months—and 16.3 during statin months, which was also significantly higher than no-pill. But the difference between the placebo score of 15.4 and statin score of 16.3 were not at all significantly different, P = 0.39 vs placebo.

Imagine a bar graph with the symptom score on the y axis. The score of 15.4 is about the same as 16.3 and the nocebo ratio was 0.9. In other words, nearly all of the ill effects of taking the statin come from taking the pill as much as the statin chemical.

  • We knew this from the NEJM letter last November, but the full paper had some neat additional features.

  • Similar numbers of participants stopped taking the statin as the placebo.

  • Intensity of symptom onset did not predict whether the pill was a statin or placebo.

  • The magnitude of symptom relief did not predict whether the pill was a statin or placebo.

  • After the trial was over, 30 of the 60 participants in the trial had restarted their statin.

  • Review of individual patient data revealed three patterns: some patients experienced no adverse effects during the trial (despite being intolerant before the trial); some patients had intense symptoms throughout regardless of whether they were taking pills or not; and the most common pattern was that symptoms increased with both statin and placebo.

Comments: This study has three big implications. First, it shows the power of the nocebo effect. Clearly the major source of adverse effects was simply the act of expecting it, not so much from the statin. This has major repercussions for how we interact with patients. Our words can harm, or they can heal. While it is right to explain potential adverse effects of therapies to patients, it is also important to understand that expectations matter. I for one have discussed this trial multiple times with patients. The same issue comes up with other drugs, such as beta-blockers. Negative expectations surely play a role in some of the side effects of these meds as well.

A second important implication revolves around regression to the mean and the dangers of current practice of informal experiments taking patients off and on the statins. The problem is that statins are often stopped when symptoms are maximal and naturally tend to decline, and statins are restarted when symptoms have resolved and can only get worse.

Thus, the prompt onset and offset of symptoms after starting and stopping pills is often interpreted by clinicians and patients as evidence of causation. SAMSON shows that this was true, but the patterns were similar for statin and placebo. This is why the no-tablet months were so critical. Current practice, therefore, could be entraining a nocebo response.

Finally, there is a lot of background noise when it comes to generalized complaints like muscle aches and fatigue. In SAMSON, sometimes intolerant patients could restart without trouble, and other times, patients felt bad no matter what they were talking.

Senior author Darrel Francis has told me that placebo pills are super-expensive, and drug companies have no incentive to make them, so SAMSON-like trials are not going to be practical for individual docs and patients. That said, SAMSON is a worthy trial to keep front and center in our brains and in our clinic drawers, because it reminds us of the massive complexity of the doctor-patient interaction.

Nightshift Work and AF

A large group of authors from the United States, China, and Sweden have published an interesting paper in the European Heart Journal on the association of nightshift work and atrial fibrillation (AF) and coronary heart disease (CHD). The data source was the UK Biobank. The analysis included about 276,000 individuals who did not have CHD, stroke, heart failure (HF), or AF at baseline. Follow-up was about 10 years, and there were a lot of AF episodes—almost 6000. The authors created four categories: day workers, shift workers with rare night shifts, workers with some night shifts, and usual or permanent nightshift workers. The control arm was day workers. Those who worked at night more often had lower socioeconomic status, lower education level, worked longer hours, and reported shorter sleep durations.

  • Compared with day workers, increasing categories of night shifts were associated with a higher risk of incident AF in age- and sex-adjusted models (P for trend 0.001), and rotating shifts.

  • Those with usual or permanent night shifts had the highest risk (HR 1.16, 95% CI 1.02–1.32).

  • The association remained significant after adjustments for demographics.

  • Usual/permanent current night shifts were significantly associated with a higher risk of CHD (HR 1.22, 95% CI 1.11–1.35) but not HF and stroke.

  • The authors then calculated a population attributable risk (PAR)—the amount of AF that, if this were a causal relationship, would be eliminated by eliminating shift work—and the PAR was only 1.7% for AF over 10 years. By comparison, the PAR for smoking exposure and lung cancer is 80%.

I am surprised the association was so weak. I have seen a lot of patients whose AF improves when they either go to day shift or retire. The CHD signal looked stronger. Also, disordered sleep has a clear association with AF, and nightshift work often leads to sleep abnormalities.

When I saw the headline, I thought, this is going to be a big association. But it ended up being a weak signal, one that barely reached significance after adjustments. The message is that there are likely many causal factors in AF, and shiftwork is likely one of the many.

The normal caveats of observational studies apply: the authors can only adjust for things on a spreadsheet. People who work nights may have characteristics not captured on a spreadsheet. Also, these sorts of studies use self-reported data, so there may be recall biases. The UK Biobank also may have a degree of healthy volunteer bias and may not represent the average shift worker.

Therapeutic Fashion Changing in HF

Journalist Mitchel Zoler covered a session at the Heart Failure Society of America meeting in which a number of trialists and noted HF experts opined on expanded use of sacubitril/valsartan (Sac/Val)and SGLT2 inhibitors in HF.

  • First point: Ejection fraction (EF) is not valuable. Forget about it, said Milton Packer. Specifically, the 40% cutoff. Others agreed. The suggestion came forth: Wouldn't it be easier if we said that every patient with HF needs to receive one agent from each of the four [pillar] drug classes.

  • Second Point: Strain is a better measure of systolic and diastolic HF.

  • Third Point: Never mind trials with nonsignificant primary endpoints, like PARAGON-HF, CHARM-Preserved, and Topcat. If we instead focus on post-hoc analyses of certain patients, we can find significant benefits of the drug.

I worry about this sort of argument. Here is why:

  • First, the drugs they are proposing to use are costly and, I would argue, based on the trial results, low value.

  • Second, I am learning about strain, but its problem is that no trial used it as an entry criteria. For better or worse, EF is the established measure. Strain may become a useful surrogate, but it isn’t there yet.

You might say, come on Mandrola, don’t you want what is best for HF patients? I do, but we also have to stay clear-eyed about the evidence. Here is how weak some of it is:

Sac/Val in HF with reduced EF (HFrEF): The comparator in PARADIGM-HF was medium-dose enalapril. Did Sac/Val crush it because it was great, or because enalapril was weak. Packer and other HF experts shun this argument, but since then, every time Sac/Val gets compared to Val alone, it looks a lot less impressive. I’d love to see a PARADIGM-HF repeated with Val as the comparator.

Sac/Val in HF with preserved EF (HFpEF): Remember that PARAGON-HF missed its primary endpoint. The US Food and Drug Administration approval rests on subgroup analyses of a trial with a nonsignificant endpoint, which can be justified, but mostly for hypothesis-generation for future trials. What’s more, there were no differences in cardiovascular (CV) death or overall death, despite a high death rate of 14% in both arms.

Empagliflozin in HFpEF: Again, the only thing this drug did was reduced hospitalizations for HF (HHF), no diff in CV death, death, or quality of life. Also, we weren’t given baseline loop diuretic use in the groups, and HHF was just a small fraction of total hospitalizations.

Thus, we can say that for those with HFrEF, RAS inhibition, beta-blockade, SGLT2 inhibitors, and MRA use along with cardiac resynchronization therapy and implantable cardioverter-defibrillator therapy are clearly beneficial. Are the new drugs as good as the old? I am not totally convinced. And in the larger group of patients with HF and a better EF, compelling data for use of super-expensive drugs---as a routine—is sparse. Yes, specific patients may benefit, but I don’t think we are anywhere near instituting algorithmic pathways or quality measures.

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....