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. This podcast is intended for healthcare professionals only.
In This Week’s Podcast
For the week ending July 29, 2022, John Mandrola, MD comments on the following news and features stories.
Left vs Left RCT
The first topic today could represent the biggest news in electrophysiology (EP) since the major cardiac resynchronization therapy (CRT) trials showed that bi-ventricular (BiV) pacing improves outcomes in patients with left bunde branch block (LBBB) and heart failure (HF).
Listeners know that I am an ardent supporter of conduction system pacing, first in the His-bundle and now, in the area of the LBB on the left side of the septum. I admit that some of my attraction to this is technique is the aesthetics — pacing the conduction system delivers gorgeously narrow QRS complexes.
Another attraction is plausibility; if you pace the conduction system, you restore right and left ventricular synchrony. That will a) eliminate pacing induced cardiomyopathy, and b) may function as a replacement for BiV pacing and CRT.
Listeners also know how often I rail against doing things because they are plausible or make sense. I imagine they are thinking, Mandrola, you are hypocrite. As a medical conservative, you say we ought to have trials before early adoption. And my critics have a point. Except this week, we learned that there likely will be a trial.
My colleague, Mihail Chelu, now from Baylor University, and Ken Ellenbogen, have garnered 31 million dollars from the US government, via the Patient Centered Outcomes Research Institute (PCORI), to conduct the Left vs Left randomized clinical trial (RCT), which will test conduction system pacing (CSP) vs BiV pacing in those eligible for CRT.
In the press release, Chelu describes a 55-center trial in the United States and Canada, that will:
Randomly assign more than 2000 patients who are eligible for CRT to receive either CSP or standard BiV pacing.
Follow-up will be 5 years.
The endpoints will be death and hospitalization for heart failure (HHF).
I am extremely proud to say that my shop, little old Baptist Health Louisville will be a center. And I issue this challenge to the other big-gun centers. We will challenge you for top-enrolling sites. This is because while I love CSP, I also love evidence, and the CRT proponents are exactly correct, we need data. There is true equipoise here. Congratulations to Dr Chelu and Dr. Ellenbogen.
Probability and Conjunction Fallacy
Let me tell you a story that is slightly related to the topic of CRT. My partner says John, a BiV device would really help this person. And he is right. A woman with nonischemic cardiomyopathy, typical LBB, first degree atrio-ventricular (AV) block. All the good predictors.
My partner also told the patient there’s a 95% chance of benefit. That statement almost gave me a rash. You told her what? Why I am anxious? Because of the conjunction fallacy. That is the idea that probabilities don’t add they multiply. Recall that probabilities are less than one so when you multiply probabilities, they get smaller.
Before I tell you about how the conjunction fallacy relates to implanting a CRT device, let me give you the example of the conjunction fallacy, made famous by Tversky and Kahneman. They call it the Linda problem.
Linda is 31 years old, single, outspoken, and very bright. She majored in philosophy. As a student, she was deeply concerned with issues of discrimination and social justice, and also participated in anti-nuclear demonstrations.
Which is more probable?
Linda is a bank teller.
Linda is a bank teller and is active in the feminist movement.
The majority of people in their experiments chose number 2, but that is wrong. It is wrong because the probability of two events occurring together (that is, in conjunction) is always less than or equal to the probability of either one occurring alone
Tversky and Kahneman argued that most people get this problem wrong because they use a heuristic, a shortcut procedure called representativeness to make this kind of judgment. Here, option 2 seemed more "representative" of Linda from the description of her, even though it is clearly mathematically less likely. Doctors use tons of heuristics.
Now back to the CRT/BiV example: There are tons of conjunctive probabilities here:
The probability that there are adequate coronary sinus veins.
The probability that I am able to get a lead in them.
The probability that the veins have good thresholds without phrenic nerve capture.
The probability that she is CRT responder.
The probability that the incision heals without infection.
All of those are decently high probabilities but I guarantee multiplying them together does not come to 95%.
JAMA Network Open published a brilliant study led by Scott Aberegg from the University of Utah, with first author, Hal Arkes, from Potsdam Germany, in which they ask whether physicians are able to correctly estimate the overall probability of a medical outcome resulting from a sequence of clinical events.
They sent out surveys that consisted of three separate sub-studies to more than 200 docs.
Docs were asked to estimate the probability of success of each of three components of a diagnostic or prognostic sequence as well as the overall probability of the sequence.
All docs were board certified in the specialty of the scenario. They were in their 50s, so they were experienced.
Two substudies involved obstetrical cases and one substudy involved the evaluation of a pulmonary nodule.
Of the 215 docs who responded, 168 or 78% estimated the probability of the two-step process as greater than the probability of at least one or two of the component events. And it wasn’t by a little. They estimated the final probability of the process by between 12% to 20% higher than the single elements.
My friends, let me remind you that these are formally illogical and mathematically incoherent answers. “This biased estimation, consistent with the conjunction fallacy, may have substantial implications for diagnostic and prognostic decision-making,” the authors concluded.
Adam Rodman, from Beth Israel in Boston, commented, “If, as the paper shows, physicians fundamentally misunderstand the multiplication rule of probability due to innumeracy, they are dramatically overestimating the post-test probabilities of disease ― and drastically overtreating people.”
This idea is backed by the Gerd Gigerenzer example, when he asked docs to calculate the probability that a woman had breast cancer from a positive mammogram. More than 90% of docs responded with an estimate 10-fold higher than the correct answer.
Senior author Scott Aberegg said this about his paper: “It really seems as though ― in this study and in my experience in medicine ― that when asked to predict a probability, physicians either are completely unwilling to do so or provide nothing more than a wild guess.
“That sounds pejorative. It sounds like I’m poking at my fellow physicians, and I don’t want it to seem that way. But I think that’s just reflective of a general discomfort with or lack of facility with probabilities and simple numbers.”
Here is my take: I agree that physicians, including me, are not that facile with probability. It’s why we are so easily bamboozled by studies. But the main problem with our widespread allergy to probability is that it makes us way too confident and way too optimistic.
My partner earnestly telling the patient there was a 95% chance of CRT benefit, for instance. Or the idea that something like atrial fibrillation (AF) screening or vascular screening will lead to benefits. Think of the multiple components on that chain of probability from a rhythm screening device to a longer life.
As for the solution. Aberegg is right when he says medical education should redouble it’s focus on probabilities.
I think Kahneman’s Thinking, Fast and Slow ought to be taught in place of the Kreb’s cycle. Kahneman wasn’t right about everything, but his work helps us understand our human weakness of thinking. And this would make us better doctors.
Assist Devices in Cardiogenic Shock
Cardiogenic shock is bad. Simply stated, it occurs when the cardiac output is not sufficient for the organs to function. Acute myocardial infarction (MI) or heart attack is the most common cause of cardiogenic shock. Doctors sometimes use assist devices, which makes perfect sense, right?
The old one is called an intra-aortic balloon pump (IABP). The IABP works by inflating and deflating in the aorta. This acts to enhance cardiac output but by less direct means. When it inflates during diastole, it improves coronary blood flow; when it deflates during systole, it acts sort of like a suction pulling blood out of the heart — so called afterload reduction.
A commonly used newer assist device is called an intravascular microaxial LV assist device (LVAD). The device goes across the aortic valve in the left ventricle (LV). It grabs blood from the LV and pumps it out to the aorta and basically adds to the cardiac output – up to 3.8 liters, which is a lot.
This podcast shuns brand names but the most common micoraxial LVAD is made by Abiomed and called Impella. JAMA-Internal Medicine has published another observational study comparing the LVAD – the Impella — and the IABP in patients with cardiogenic shock.
The Yale-led team of authors looked at short-term clinical outcomes. They used administrative insurance claims data. Obviously these two groups — LVAD vs IABP — were not random, so they used propensity matching in an attempt to match as many baseline variables as possible.
The sample size was about 3000 patients. Of these, investigators were able to find about 800 matched pairs.
Patients treated with LVAD were 63% more likely to die in the hospital. The confidence interval (CI) of the odds ratio (OR) ranged from 1.32 to 2.0, so it was significant.
Patients treated with LVAD were 71% more likely to die at 30 days and 44% more likely to die at 1 year. Both of these were significant as well.
Bleeding at 30 days was also higher in the LVAD arm; OR, 1.35; 95% CI, 1.04-1.76.
Kidney replacement therapy or dialysis was 88% higher in the LVAD group and was also significant.
Costs were $50,000K higher in the LVAD group.
Comments. On the surface, this sure looks bad. The LVAD cost more, and was associated with more death, more kidney failure, and more bleeding.
What is more, it’s compared against a therapy — IABP — found in meta-analyses to be no better than medical therapy. Pause there. The Impella was worse than a device that was no better than meds.
Thus far, no RCTs of LVAD in cardiogenic shock that look at outcomes. There is however one that is ongoing, called DanGer Shock, which will randomize LVAD vs Guideline driven therapy and look at mortality at 6 months. For now, all we have are small trials of LVAD vs IABP which don’t have the power to sort out outcomes.
Previous observational studies have found worse outcomes with Impella/LVAD relative to IABP, just like this one. And it’s plausible. The LVAD device requires a large bore arterial access. This may increase the risk for bleeding. It’s in the LV and could increase stroke or emboli to the kidneys. Also, the LVAD placement may delay revascularization. Dorking around with it may distract from opening the artery. So, there are reasons why these observational trials may be true. But — and I hope you all know what is coming — you cannot make that causal leap that these studies are enough to tell you LVAD/Impellas are worse than IABP, which is no better than meds. That is because it is possible that sicker patients receive the LVAD and that is why they fare worse. Propensity matching tries to balance these differences out, but it is not a substitute for randomization.
What’s more, and I am no expert on this, I have trouble believing that insurance claims data is granular enough to make these adjustments. Now, you may counter this and say, well, Mandrola, this isn’t the only observational study showing associated harm, there are others. And I would agree but counter that it could be like digoxin, meaning that all the observational studies are likely confounded by selection bias.
Here is the core existential problem with observational studies: Sometimes they align perfectly with RCTs, you just can’t know which ones. I got that from Dr David Cohen.
Finally, the excellent editorial by Gene Hu and colleagues, notes that the American Heart Association (AHA) has released a policy statement for patients with MI saying that, for “[those] with cardiogenic shock at presentation or during hospitalization, treatment with advanced mechanical circulatory support devices should be initiated early,” and that centers without these capabilities should have means to transfer patients.
While I don’t think this and other observational studies prove harm, I would also say there is no compelling evidence that the far more costly device is worthy of such a statement.
I also agree with Hu and colleagues that there was regulatory failure in this space. LVAD use has ballooned in recent years; there is a billing code, and doctors get more relative value units for doing it. But why wouldn’t regulators make proponents show that a more expensive more invasive device works before allowing it?
Imagine a counterfactual world where regulators said, sure, you can use this device, and we will pay you for it only if it is part of a trial.
Now, if the DanGer trial comes out null, or worse for mortality, think of the money and potentially lives that would have been saved with proper evidence before adoption.
Same story with Watchman. If regulators had required evidence before re-imbursement, we would know if this device works.
Vitamin D Again
The New England Journal of Medicine has published yet another RCT looking at supplemental vitamin D to improve health. This Week in Cardiology has covered vitamin D over and over again and all these trials have failed to show benefits in cardiovascular health.
The one place where vitamin D supplements would seem most likely to help is in bone health. In fact, previous trials have shown conflicting results. But last year the neutral judges from the US Preventive Services Task Force found no effect of vitamin D on fracture incidence among adults with low vitamin D levels.
The most recent trial was an ancillary study of the VITAL trial, which was published in 2019.
More than 25,000 patients.
2000 units daily vitamin D and/or fish oil vs placebo in older adults.
Vitamin D led to no reductions in major adverse cardiac events or invasive cancer.
I could give you the HR and CI but suffice to say there were no differences in fractures. And, not even a modification of treatment effect according to baseline variables such as low baseline vitamin D levels
The editorialists have this helpful sentence: Results of analyses from VITAL published in peer-reviewed journals have shown that vitamin D supplementation
Did not prevent cancer or cardiovascular disease,
Did not prevent falls,
Did not improve cognitive function,
Did not reduce atrial fibrillation, reduce migraine frequency, improve stroke outcomes, decrease age-related macular degeneration, or even reduce knee pain.
The editorialists also remind that more than 10 million vitamin D levels are measured annually in the United States. They suggest that this practice of screening should stop. As should treating to a target. I would add that we could stop burning money researching this question. The only remaining question is why this therapeutic fashion has been so resistant to evidence.
Peripheral Artery Disease
Most of the atherosclerosis I discuss on TWIC is in the coronaries. But so-called peripheral artery disease (PAD) is also common. We call it claudication when stenoses in the leg arteries cause exercise-induced pain that resolves with stopping. I have a bunch of patients whose exercise is limited by claudication. These stenoses can by bypassed or stented and of course medical therapy such as statins helps as well.
The Journal of the AHA has published results of an RCT looking at another way to manage these patients: exercise. And not just any exercise but exercising through the pain of claudication.
The LITE RCT, led by a team at Northwestern University, randomly assigned patients to a home-based walking exercise program that induced ischemic pain vs exercise without ischemic pain or non-exercising control arm.
The primary outcome was a change in walking speed and a performance questionnaire.
The group that walked with pain did better.
The group that walked – even longer distances — but without pain did no better than the non-exercise control arm.
I think I knew this but still, when you look at the graphs, there was a significant difference in the pain group. I used my phone-a-friend policy and talked with our vascular surgeons: They are funny.
One said walking programs work but no pain no gain
Another said: It’s amazing that cardiology is doing a study that has been proven in vascular surgery for 50 years. Probably the reason this is being studied is the change in approach brought on by endovascular therapy – angioplasty and stents.
A third said, walk until it hurts, then walk some more, then rest. At least 30 minutes, at least three times per week, is decades old advice.
One of our surgeons told me that their group is highly reticent to intervene on disease below the inguinal ligament. It relieves claudication quickly and effectively but they worry that the future risk of amputation is greater. Why? Because stents in the periphery are smaller and it doesn’t take much hyperplasia to shut it down – then the patient is worse.
He also emphasized how strongly they optimize medically. They get people to stop smoking, and they use high dose statins.
I close with this topic because as a retired bike racer, the fact that we have strong evidence that exercise to the point of pain actually is beneficial makes me happy.